Federal Reserve Bulletin, 2006-01

Volume 92 2006 Compilation Federal Reserve BULLETIN Board of Governors of the Federal Reserve System, Washington, D.C.

Publications Committee RosannaPianaltoCameron,Chair ScottG.Alvarez SandraF.Braunstein RogerT.Cole MarianneM.Emerson JenniferJ.Johnson KarenH.Johnson StephenR.Malphrus H.FayPeters VincentR.Reinhart LouiseL.Roseman MichelleA.Smith DavidJ.Stockton TheFederalReserveBulletinisissuedannuallyunderthedirectionofthestaffpublicationscommittee.Thiscommitteeisresponsibleforopinionsexpressedexceptin officialstatementsandsignedarticles.ItisassistedbythePublicationsDepartmentunderthedirectionofLucretiaM.Boyer.

Table of Contents PREFACE ARTICLES Recent Changes in U.S. Family Finances: Evidence from the 2001 and 2004 Survey of Consumer Finances........................................................... A1 Brian K. Bucks, Arthur B. Kennickell, and Kevin B. Moore March22 Industrial Production and Capacity Utilization: The 2005 Annual Revision..................... A39 Kimberly Bayard and Charles Gilbert March16 Understanding U.S. Cross-Border Securities Data............................................ A59 Carol C. Bertaut, William L. Griever, and Ralph W. Tryon May4 Profits and Balance Sheet Developments at U.S. Commercial Banks in 2005 ................... A77 Elizabeth C. Klee and Gretchen C. Weinbach June14 Credit Card Disclosures, Solicitations, and Privacy Notices: Survey Results of Consumer Knowledge and Behavior...................................................A109 Thomas A. Durkin August7 Higher-Priced Home Lending and the 2005 HMDA Data.....................................A123 Robert B. Avery, Kenneth P. Brevoort, and Glenn B. Canner September8 Financial Services Used by Small Businesses: Evidence from the 2003 Survey of Small Business Finances.............................................................A167 Traci L. Mach and John D. Wolken October24 REPORTS ON THE CONDITION OF THE U.S. BANKING INDUSTRY Third Quarter, 2005 ...................................................................... B1 May2 Fourth Quarter, 2005 ..................................................................... B7 October18 LEGAL DEVELOPMENTS Fourth Quarter, 2005 ..................................................................... C1 March22 First Quarter, 2006....................................................................... C65 July13 Second Quarter, 2006..................................................................... C143 August21 Third Quarter, 2006 ...................................................................... C159 December22 INDEX.................................................................................. D1

Preface The Federal Reserve Bulletin was introduced in 1914 as a vehicle to present policy issues developed by the Federal Reserve Board. Throughout the years, the Bulletin has been viewed as a journal of record, serving to provide the public with data and research results generated by the Board. Authors from the Board’s Research and Statistics, Monetary Affairs, International Finance, Banking Supervision and Regulation, Consumer and Community Affairs, Reserve Bank Operations, and Legal divisions contribute to the Bulletin, which includes topical research articles, the Report on the Condition of the U.S. Banking Industry, orders on banking applications,andenforcementactions. Starting in 2004, the Bulletin was published quarterly rather than monthly. In 2006, in response to the increaseduseoftheInternet—andinordertoreleasearticlesandreportsinamoretimelyfashion—theBoard discontinued the quarterly print version of the Bulletin and began to publish the contents of the Bulletin on its publicwebsiteastheinformationbecameavailable.Allarticles,bankingconditionreports,ordersonbanking applications, and enforcement actions that were published in the online Bulletin in 2006 are included in this printcompilation. The tables that appeared in the Financial and Business Statistics section of the Bulletin from 1914 through 2003 are now published monthly as a separate print publication, the Statistical Supplement to the Federal Reserve Bulletin. All statistical series are published with the same frequency as they were in the Bulletin, and thenumberingsystemforthetablesremainsthesame. Online access to the Bulletin and the Statistical Supplement is free. A free e-mail notification service is available to alert subscribers to the release of articles and reports in the Bulletin; the monthly Statistical Supplement;andpressreleases,testimonies,andspeeches.Themessageprovidesabriefdescriptionandalink totherecentposting. FederalReserveBulletin: www.federalreserve.gov/pubs/bulletin StatisticalSupplementtotheFederalReserveBulletin: www.federalreserve.gov/pubs/supplement/default.htm Subscribetoe-mailnotificationservice: www.federalreserve.gov/generalinfo/subscribe/notification.htm.

Articles

Recent Changes in U.S. Family Finances: Evidence from the 2001 and 2004 Survey of Consumer Finances Brian K. Bucks, Arthur B. Kennickell, and Kevin B. 2. Change in median and mean net worth, 1995–2004 SCF Moore, of the Board’s Division of Research and Statistics, prepared this article with assistance from Percent GerhardFriesandA.MichaelNeal. Median Mean 30 The Federal Reserve Board’s Survey of Consumer 25 Finances for 2004 provides insights into changes in family income and net worth since the 2001 survey. 20 The survey shows that, over the 2001–04 period, the 15 median value of real (inflation-adjusted) family income before taxes continued to trend up, rising 10 1.6 percent, whereas the mean value fell 2.3 percent. Patterns of change were mixed across demographic 5 groups. These results stand in contrast to the strong andbroadgainsseenfortheperiodbetweenthe1998 1995–98 1998–2001 2001–04 and 2001 surveys and to the smaller but similarly SOURCE: Federal Reserve Board, Survey of Consumer Finances. broad gains between the 1995 and 1998 surveys (figure1). Much like median income, median real family net worth in the 2001−04 period increased 1.5 percent, and decreases in wealth; in some instances, median but mean net worth rose 6.3 percent. The increase in and mean values moved in opposite directions, a wealth appears to have been clearest in the middle pattern that signals distributional changes within incomegroup.Overmanyotherdemographicgroups, groups.Incontrast,thegrowthinwealthbetweenthe the data show a complex pattern of mixed increases 1998 and 2001 surveys and between the 1995 and 1998 surveys was stronger both in the mean and in the median, and the growth was shared by most demographicgroups(figure2). 1. Change in median and mean incomes, 1995–2004 SCF Threekeyshiftsinthe2001–04periodunderliethe changesinnetworth.First,thestrongappreciationof Percent housevaluesandariseintherateofhomeownership Median produced a substantial gain in the value of holdings Mean 20 of residential real estate. Second, despite the general recovery of prices in equity markets since 2001, the 15 direct and indirect ownership of stocks declined, as did the typical amount held. Third, the amount of 10 debt relative to total assets increased markedly, and the largest part of that increase was attributable to 5 debtsecuredbyrealestate. + As debt rose over the period, families devoted _0 moreoftheirincomestoservicingtheirdebts,despite a general decline in interest rates. Also, the fraction 1995–98 1998–2001 2001–04 offamilieswithlargerequireddebtservicepayments SOURCE: Federal Reserve Board, Survey of Consumer Finances. relativetotheirincomesroseasmallamount,andthe

A2 Federal Reserve Bulletin 2006 fraction of families that had payments that were late The national house price index produced by the sixty days or more in the year preceding the survey Office of Federal Housing Enterprise Oversight rose more substantially. These increases affected increasednearly27percentfrom2001to2004.Price mainly the bottom 80 percent of the income increases varied sharply across the country; by area, distribution. the largest gains were in the New England, Middle This article reviews these and other changes in the Atlantic, and Pacific sections of the country—all financialconditionofU.S.familiesbetween2001and more than 35 percent; average gains were consider- 2004.1ThediscussiondrawsondatafromtheFederal ably smaller in parts of the South. Homeownership Reserve Board’s Survey of Consumer Finances ratescontinuedagradualclimb. (SCF) for those years; it also uses evidence from Other institutional factors also affected family earlier years of the survey to place the 2001–04 finances. Tax cuts enacted by the Jobs and Growth changesinabroadercontext. Tax Relief Reconciliation Act of 2003 increased the child tax credit, provided some concessions for married couples, and expanded the proportion of taxpay- ECONOMIC BACKGROUND ers covered by the lowest tax-rate bracket. A major element of the 2003 tax act was the decrease in tax The U.S. economy was in a mild recession through rates on capital gains coupled with the change to much of 2001, and real gross domestic product was taxingdividendsatthesamerateascapitalgains.The flatfortheyear.However,thispauseinthegrowthof proportionoffamiliesthatusetheInternetasasource real GDP was followed by some pickup in 2002 and forfinancialservices,tools,orinformationcontinued sharper gains of 4.1 percent in 2003 and 3.8 percent to grow; according to the SCF, it rose from 32.5 perin 2004. The unemployment rate, which had peaked centin2001to46.5percentin2004. at 6.5 percent in mid-2003, fell to 5.1 percent in Several demographic shifts had important conse- 2004. The rate of inflation, as measured by the con- quencesforthestructureofthepopulation.Theaging sumerpriceindexforallurbanconsumers(CPI),was of the baby-boom population from 2001 to 2004 moderate by historical standards over the 2001–04 drovea2percentagepointincreaseintheshareofthe period; for 2004, it was 2.7 percent, nearly the same population aged 55 to 64. Overall population growth rateasfor2001. was about 3 percent, and, according to figures from Developments in financial markets over the three- the Bureau of the Census, 58 percent of the growth year period were varied. The major stock market was due to net immigration. Also according to Cenindexes declined before erasing most of the losses sus estimates, the number of households increased with an increase in 2004. Most interest rates had 3.6 percent—a rate slower than the 5.5 percent pace initially declined but began to rise by the end of inthe1998–2001period—andtheaveragenumberof 2004. For example, the interest rate on a thirty-year people per household remained close to two and a fixed-ratemortgageaveraged6.82percentinSeptem- half. ber2001,whenabouthalftheinterviewsforthe2001 SCFhadbeencompleted,andwas5.75percentthree years later. Lower interest rates also brought lower INCOME yields on liquid deposits, time deposits, and bonds; for example, the rate on a three-month certificate of Thechangeinrealbefore-taxfamilyincomebetween deposit had dropped from an average of 3.69 percent 2001and2004standsinstrongcontrasttothechange for 2001 to slightly more than 1 percent in early for the preceding three-year period.2 Over the more 2004, although the rate climbed to 2.45 percent by theendoftheyear. 2. Tomeasureincome,theinterviewersrequestinformationonthe 1. Seebox‘‘TheDataUsedinThisArticle’’forageneraldescrip- family’scashincome,beforetaxes,forthefullcalendaryearprecedtionofthedata.Theappendixtothisarticleprovidesasummaryof ing the survey. The components of income in the SCF are wages; keytechnicalaspectsofthesurvey.Foradetaileddiscussionofthe self-employmentandbusinessincome;taxableandtax-exemptinter- 1998 and 2001 surveys as well as references to earlier surveys, see est;dividends;realizedcapitalgains;foodstampsandother,related AnaM.Aizcorbe,ArthurB.Kennickell,andKevinB.Moore(2003), supportprogramsprovidedbygovernment;pensionsandwithdrawals ‘‘RecentChangesinU.S.FamilyFinances:Evidencefromthe1998 fromretirementaccounts;SocialSecurity;alimonyandothersupport and2001SurveyofConsumerFinances,’’FederalReserveBulletin, payments; and miscellaneous sources of income for all members of vol.89(January),pp.1–32. theprimaryeconomicunitinthehousehold.

Recent Changes in U.S. Family Finances: Evidence from the 2001 and 2004 Survey of Consumer Finances A3 TheDataUsedinThisArticle DatafromtheSurveyofConsumerFinances(SCF)arethe ReserveBulletinarticlesareattributabletoadditionalstatisbasisoftheanalysispresentedinthisarticle.TheSCFisa tical processing, correction of minor data errors, revisions triennialinterviewsurveyofU.S.familiessponsoredbythe tothesurveyweights,conceptualchangesinthedefinitions BoardofGovernorsoftheFederalReserveSystemwiththe ofvariablesusedinthearticles,andadjustmentsforinflacooperation of the U.S. Department of the Treasury. Since tion. In this article, all dollar amounts from the SCF are 1992, data for the SCF have been collected by NORC, a adjustedto2004dollarsusingthe‘‘currentmethods’’verresearchorganizationattheUniversityofChicago,roughly sion of the consumer price index (CPI) for all urban betweenMayandDecemberofeachsurveyyear. consumers.2 The majority of statistics included in this article are The principal detailed tables describing asset and debt related to characteristics of ‘‘families.’’ As used here, this holdings focus on the percent of various groups that have termismorecomparabletotheU.S.BureauoftheCensus such items and the median holding for those that have definition of ‘‘households’’ than to its use of ‘‘families,’’ them.3Thisconditionalmedianischosentogiveasenseof which excludes the possibility of one-person families. The the‘‘typical’’holding.Generally,whenonedealswithdata appendixprovidesfulldefinitionsof‘‘family’’fortheSCF thatexhibitverylargevaluesforarelativelysmallpartof andtheassociatedfamily‘‘head.’’Thesurveycollectsinfor- the population—as is the case for many of the items mationonfamilies’totalincomebeforetaxesforthecalen- considered in this article—estimates of the median are daryearprecedingthesurvey.Butthebulkofthedatacover often statistically less sensitive to such outliers than are the status of families as of the time of the interview, estimatesofthemean. including detailed information on their balance sheets and Oneliabilityofusingthemedianasadescriptivedevice useoffinancialservicesaswellasontheirpensions,labor is that medians are not ‘‘additive’’; that is, the sum of the forceparticipation,anddemographiccharacteristics.Except medians of two items for a common population is not inasmallnumberofinstances(seetheappendixfordetails), generally equal to the median of the sum (for example, the survey questionnaire has changed in only minor ways medianassetslessmedianliabilitiesdoesnotequalmedian since 1989, and every effort has been made to ensure the networth).Incontrast,meansforacommonpopulationare maximumdegreeofcomparabilityofthedataovertime. additive.Whereacomparablemedianandmeanaregiven, The need to measure financial characteristics imposes thegrowthofthemeanrelativetothemedianmayusually special requirements on the sample design for the survey. betakenasindicativeofchangeatthetopofthedistribu- TheSCFisexpectedtoprovidereliableinformationbothon tion;forexample,whenthemeangrowsmorerapidlythan attributes that are broadly distributed in the population themedian,itistypicallytakentoindicatethatthevalues (such as homeownership) and on those that are highly comprisedbythetopofthedistributionrosemorerapidly concentrated in a relatively small part of the population thanthoseinthelowerpartofthedistribution. (such as closely held businesses). To address this require- Toprovideameasureofthesignificanceofthedevelopment, the SCF employs a sample design, essentially ments discussed in this article, standard errors due to unchangedsince1989,consistingoftwoparts:astandard, sampling and imputation for missing data are given for geographically based random sample and a special over- selectedestimates.Spacelimitspreventtheinclusionofthe sample of relatively wealthy families. Weights are used to standard errors for all estimates. Although we do not combine information from the two samples to make esti- directly address the statistical significance of the results, mates for the full population. In the 2004 survey, 4,522 the article highlights findings that are significant or are families were interviewed, and in the 2001 survey, 4,449 interestinginabroadercontext. wereinterviewed. Thisarticledrawsprincipallyuponthefinaldatafromthe 2004 and 2001 surveys. To provide a larger context, some 2. Inanongoingefforttoimproveaccuracy,theBureauofLaborStatisinformation is also included from the final versions of ticshasintroducedseveralrevisionstoitsCPImethodology.Thecurrentearliersurveys.1Differencesbetweenestimatesfromearlier methodsindexattemptstoextendthesechangestoearlieryearstoobtaina seriesasconsistentaspossiblewithcurrentpracticesintheofficialCPI.For surveysasreportedhereandasreportedinearlierFederal technicalinformationabouttheconstructionofthisindex,seeKennethJ. Stewart and Stephen B. Reed (1999), ‘‘Consumer Price Index Research SeriesUsingCurrentMethods,1978–1998,’’MonthlyLaborReview,vol. 1. Additional tabular information from the survey is available at 122(June),pp.29–38.Toadjustassetsandliabilitiesto2004dollars,the www.federalreserve.gov/pubs/oss/oss2/scf2004home.html. These tables earlier survey data were multiplied by the following amounts: for 1995, includedatacomparabletothefiguresshowninthisarticleforallofthe 1.2311;for1998,1.1593;andfor2001,1.0651.Toadjustfamilyincomefor surveys from 1989 to 2004. For some assets and debts by demographic the previous calendar year to 2004 dollars, the following factors were group,thesetablesreportmeansaswellasmediansforeachgroup.The applied:for1995,1.2610;for1998,1.1757;for2001,1.0948;andfor2004, estimatesofthemeans,however,aremorelikelytobeaffectedbysampling 1.0269. errorthanaretheestimatesofthemedians.Thetablesalsoincludesome 3. Themedianofadistributionisdefinedasthevalueatwhichequal alternativeversionsofthetablesinthisarticle. partsofthepopulationconsideredhavevalueslargerorsmaller.

A4 Federal Reserve Bulletin 2006 1. Before-taxfamilyincome,percentageoffamiliesthatsaved,anddistributionoffamilies,byselectedcharacteristicsof families,1995–2004surveys Thousandsof2004dollarsexceptasnoted 1995 1998 cha F ra a c m te il r y istic Income Perce o n f tage Percentage Income Perce o n f tage Percentage of of families families Median Mean thatsaved families Median Mean thatsaved families Allfamilies..................... 37.8 54.9 55.2 100 38.8 61.7 55.9 100 (.9) (.9) (.9) (1.3) Percentileofincome Lessthan20 .................... 8.5 8.2 31.6 20.0 9.6 9.2 32.1 20.0 20–39.9......................... 21.7 21.6 43.4 20.0 23.5 23.4 45.5 20.0 40–59.9......................... 37.8 37.1 57.2 20.0 38.8 39.4 56.1 20.0 60–79.9......................... 56.1 57.0 66.8 20.0 61.8 63.0 67.9 20.0 80–89.9......................... 84.5 85.7 69.9 10.0 91.6 92.2 73.7 10.0 90–100 ......................... 138.6 215.8 84.2 10.0 151.5 254.5 82.0 10.0 Ageofhead(years) Lessthan35 .................... 31.5 38.4 56.4 24.8 31.8 41.9 53.0 23.3 35–44 .......................... 47.2 60.0 54.3 23.0 48.8 69.6 57.3 23.3 45–54 .......................... 49.6 81.4 58.0 17.9 58.8 80.9 57.8 19.2 55–64 .......................... 41.6 66.4 58.0 12.5 44.7 83.2 61.1 12.8 65–74 .......................... 23.7 46.1 50.0 12.0 28.2 54.2 56.3 11.2 75ormore...................... 19.7 32.7 51.7 9.8 19.4 33.9 48.6 10.2 Educationofhead Nohighschooldiploma ......... 17.9 25.8 42.8 18.5 18.0 25.2 39.5 16.5 Highschooldiploma ............ 32.1 43.0 50.6 31.7 33.9 42.9 53.7 31.9 Somecollege ................... 37.8 49.9 54.1 19.0 41.1 58.9 56.7 18.5 Collegedegree .................. 56.3 87.9 68.2 30.7 63.7 99.3 65.6 33.2 Raceorethnicityofrespondent Whitenon-Hispanic ............. 40.7 60.4 59.1 77.6 44.2 68.7 60.0 76.8 NonwhiteorHispanic ........... 24.4 36.0 41.7 22.4 27.0 38.5 42.3 23.2 Currentworkstatusofhead Workingforsomeoneelse ....... 45.4 59.6 60.4 58.3 47.0 62.1 59.8 59.2 Self-employed .................. 46.7 98.5 63.4 10.3 61.1 126.8 61.1 11.3 Retired ......................... 20.7 34.5 46.0 25.1 22.3 38.2 48.7 24.4 Othernotworking............... 13.9 22.2 30.8 6.4 13.5 25.2 33.3 5.1 Region Northeast ....................... 37.8 60.6 52.6 19.8 41.1 70.7 53.5 19.3 Midwest ........................ 38.6 56.0 59.2 23.9 38.2 56.8 58.3 23.6 South ........................... 35.0 50.9 54.6 35.1 36.6 57.3 55.0 35.7 West............................ 39.1 55.2 54.0 21.2 42.0 66.2 56.9 21.3 Housingstatus Owner .......................... 46.7 68.0 61.3 64.7 50.7 77.3 62.2 66.2 Renterorother.................. 22.7 30.9 44.0 35.3 23.5 31.0 43.4 33.8 Percentileofnetworth Lessthan25 .................... 17.8 22.9 35.7 25.0 18.5 23.6 36.3 25.0 25–49.9......................... 35.3 38.6 51.4 25.0 35.3 39.3 50.3 25.0 50–74.9......................... 43.6 50.1 59.4 25.0 47.0 54.3 61.8 25.0 75–89.9......................... 52.6 65.0 68.8 15.0 65.8 78.4 71.9 15.0 90–100 ......................... 99.1 172.8 82.4 10.0 102.4 206.3 80.0 10.0 Note: For questions on income, respondents were asked to base their indexforallurbanconsumers(seetextbox‘‘TheDataUsedinThisArticle’’). answersonthecalendaryearprecedingtheinterview.Forquestionsonsav- Seetheappendixfordetailsonstandarderrors(showninparenthesesbelowthe ing,respondentswereaskedtobasetheiranswersonthetwelvemonthspreced- firstrowofdataforthemeansandmedianshereandintable3)andfordefiniing the interview. For discussion of racial and ethnic designations, see the tionsoffamilyandfamilyhead. appendix. Percentagedistributionsmaynotsumto100becauseofrounding.Dollars havebeenconvertedto2004valueswiththecurrent-methodsconsumerprice recentperiod,medianincomerose1.6percent,while ing three-year period, the median had increased the mean fell 2.3 percent (table 1).3 Over the preced- 9.5percent,andthemeanhadincreased17.3percent. 3. Overthe2001–04period,estimatesofinflation-adjustedhouseholdincomeforthepreviousyearfromtheCurrentPopulationSurvey respondentsinthatsurveymightbeidentifiable.Asdiscussedinmore (CPS)oftheBureauoftheCensusshowadeclineinboththemedian detailintheappendix,thetwosurveysalsodifferintheirdefinitions (1.5percent)andthemean(2.6percent).Typically,theSCFshowsa oftheunitsofobservationandinotheraspectsoftheirmethodologies. higherlevelofmeanincomethandoestheCPS;for2004,theSCF Thenationalincomeandproductaccounts(NIPA)provideaggregate yields an estimate of $70,700, while the CPS yields an estimate of information on the incomes of households. If NIPA estimates of $62,200. This difference in mean levels is largely the result of the personalincomeareadjustedforinflationandgrowthinthenumberof truncation of large values in the CPS data above a certain amount, householdsoverthe2001–04period,theyimplyvirtuallynochange which is done with the intent of minimizing the possibility that inhouseholdincome.

Recent Changes in U.S. Family Finances: Evidence from the 2001 and 2004 Survey of Consumer Finances A5 1.—Continued Thousandsof2004dollarsexceptasnoted 2001 2004 cha F ra a c m te il r y istic Income Perce o n f tage Percentage Income Perce o n f tage Percentage of of families families Median Mean thatsaved families Median Mean thatsaved families Allfamilies..................... 42.5 72.4 59.2 100 43.2 70.7 56.1 100 (.8) (2.0) (.8) (1.2) Percentileofincome Lessthan20 .................... 10.9 10.7 30.0 20.0 11.1 10.8 34.0 20.0 20–39.9......................... 26.0 25.7 53.4 20.0 25.7 26.1 43.5 20.0 40–59.9......................... 42.5 42.9 61.3 20.0 43.2 43.4 54.4 20.0 60–79.9......................... 69.0 69.4 72.0 20.0 68.1 69.1 69.3 20.0 80–89.9......................... 105.1 104.4 74.9 10.0 104.7 106.5 77.8 10.0 90–100 ......................... 180.6 322.4 84.3 10.0 184.8 302.1 80.6 10.0 Ageofhead(years) Lessthan35 .................... 35.6 47.1 52.9 22.7 32.9 45.1 55.0 22.2 35–44 .......................... 54.7 82.1 62.3 22.3 49.8 73.8 58.0 20.6 45–54 .......................... 58.0 99.3 61.7 20.6 61.1 94.4 58.5 20.8 55–64 .......................... 48.2 92.6 62.0 13.2 54.4 100.3 58.5 15.2 65–74 .......................... 29.6 61.9 61.8 10.7 33.3 59.6 57.1 10.5 75ormore...................... 23.8 39.1 55.5 10.4 23.7 40.9 45.7 10.7 Educationofhead Nohighschooldiploma ......... 18.1 26.7 38.7 16.0 19.4 25.9 35.9 14.4 Highschooldiploma ............ 36.1 47.7 56.7 31.7 35.6 44.8 54.0 30.6 Somecollege ................... 43.6 59.1 61.7 18.3 41.1 56.0 51.0 18.4 Collegedegree .................. 72.3 124.2 70.0 34.0 73.0 117.5 68.3 36.6 Raceorethnicityofrespondent Whitenon-Hispanic ............. 48.2 81.9 63.1 75.4 49.4 80.7 60.1 72.2 NonwhiteorHispanic ........... 27.4 43.3 47.4 24.6 29.8 44.9 45.6 27.8 Currentworkstatusofhead Workingforsomeoneelse ....... 50.4 71.7 61.6 60.9 49.3 70.1 59.2 60.1 Self-employed .................. 67.4 147.3 70.4 11.7 66.7 141.5 68.7 11.8 Retired ......................... 22.4 42.6 50.6 22.9 24.4 43.2 44.0 23.7 Othernotworking............... 17.6 38.8 42.3 4.5 20.5 37.4 44.9 4.4 Region Northeast ....................... 44.0 82.7 58.1 19.0 50.9 87.5 59.5 18.8 Midwest ........................ 46.7 68.9 63.0 23.0 45.2 67.4 59.9 22.9 South ........................... 38.3 65.4 57.3 36.2 37.0 61.9 52.5 36.3 West............................ 43.4 78.9 59.5 21.8 46.2 74.5 55.2 22.0 Housingstatus Owner .......................... 55.5 90.6 66.7 67.7 55.2 87.3 62.3 69.1 Renterorother.................. 26.3 34.3 43.6 32.3 24.6 33.7 42.3 30.9 Percentileofnetworth Lessthan25 .................... 21.0 25.5 34.5 25.0 20.5 25.1 34.8 25.0 25–49.9......................... 37.2 42.3 54.3 25.0 37.0 42.2 53.6 25.0 50–74.9......................... 54.2 62.2 68.0 25.0 52.4 60.6 62.2 25.0 75–89.9......................... 74.6 83.9 77.7 15.0 77.0 87.8 72.4 15.0 90–100 ......................... 136.9 273.1 83.9 10.0 143.8 256.0 76.0 10.0 The change over the 2001–04 period was strongly persons who have a college degree are substantially influenced by a 6.2 percent decline in the overall higher than for those with any lesser amount of medianamountofwagesmeasuredinthesurveyand schooling. Incomes of white non-Hispanic families a 3.6 percent decline in the mean (data not shown in are substantially higher than those of other families.4 the tables); wages represent the largest share of fam- Families headed by self-employed workers consisilyincome.Investment-relatedincomesalsodeclined. tently have the highest median and mean incomes of Some patterns of income distribution hold gener- all work-status groups. Income is also higher for ally across the years of SCF data shown in table 1. homeownersthanforotherfamilies,anditisprogres- Across age classes, median and mean incomes show sively higher for groups with greater net worth. alife-cyclepattern,risingtoapeakinthemiddleage Across the four regions of the country as defined by groups and then declining for groups that are older the Bureau of the Census, the ordering of median and increasingly more likely to be retired. Income also shows a strong positive association with educa- 4. Seetheappendixforadiscussionofracialandethnicidentification; in particular, incomes for families headed by tionintheSCF.

A6 Federal Reserve Bulletin 2006 incomes over time has varied, but the means gener- tiongroups.Intheprecedingthree-yearperiod,mean ally show higher values for the Northeast and the incomes had increased markedly for all education WestthanfortheMidwestandtheSouth. groups except the some-college group, and median incomes had increased notably for all groups except the no-high-school-diploma group; the median had Income by Demographic Category increasedmoststronglyforthecollege-degreegroup. In the 2001–04 period, the median income of non- Across the lowest 90 percent of the income distribu- white or Hispanic families rose 8.8 percent, and the tion between 2001 and 2004, changes in median and mean rose 3.7 percent. In contrast, the median for meanincomesvariedindirection,butallthechanges whitenon-Hispanicfamiliesrose2.5percent,andthe were 2 percent or less in absolute value.5 For the top meandeclined1.5percent.However,boththemedian 10 percent, changes in the median and mean were and the mean for nonwhites or Hispanics were about more substantial, but changes in the two statistics 60 percent of the corresponding figures for nonwere in opposite directions; the median rose 2.3 per- Hispanic whites in 2004. Between 1998 and 2001, cent, while the mean fell 6.3 percent. The decline in the median income of nonwhite or Hispanic families the mean for this group appears to be a result of a had been about unchanged, while the median had declinesince2001ininvestmentincome,whichtends increased9.0percentforotherfamilies;themeanhad to be concentrated among high-income families. The risenforbothgroups.7 changes throughout the income distribution contrast Of the work-status groups, only the retired group with the broad and substantial gains in both median had an increase in both median and mean incomes and mean incomes that had been seen over the pre- between 2001 and 2004; the median rose 8.9 percent ceding three-year period, when both measures had andthemean1.4percent.8Fortheother-not-working risen10percentormoreformostgroups. group, the median rose 16.5 percent, and the mean The income changes across almost all age groups declined 3.6 percent; since before the 1995 survey, were substantial. For the groups under age 45, both this group has had the lowest measures of income of median and mean incomes dropped. For the remain- any of the work-status groups. For the other workingagegroups,medianincomesrosestronglyexcept status groups, both median and mean incomes fell forthe75-or-oldergroup,butthemeansroseonlyfor from 2001 to 2004. Over the 1998–2001 period, the 55–64 group and the 75-or-older group. Over the median income had increased most for the selfpreceding three-year period, median income had employed and the other-not-working groups; mean increased for most age groups, particularly for the incomes were higher for all groups, especially the oldest, while the mean rose for all groups but espe- other-not-workinggroup. ciallyforthe45–54group. By region, the only growth in both median and Across education groups, median incomes rose mean incomes between 2001 and 2004 was in the only for families headed by persons with less than a Northeast. In the West, only the median rose, and in high school diploma and for families headed by per- the Midwest and South, median and mean incomes sons with a college degree; growth was particularly both fell. Over the 1998–2001 period, regional strong for the former group—7.2 percent—but that median income increased at the highest rate in the groupstillhasthelowestmedianincomeofalleduca- Midwest; growth in the mean was similar for all tion groups.6 Mean incomes declined for all educa- regions except the South, where it was somewhat lower. 5. Hereareselectedpercentilesoftheincomedistributionforthe pastfoursurveys,in2004dollars: 7. IftheinformationonHispanicorLatinoethnicidentificationis used in the classification of the 2004 results, the median income of Percentile nonwhitesorHispanicswas$30,000,andthemeanwas$45,400;for 1995 1998 2001 2004 ofincome otherfamilies,themedianwas$49,900,andthemeanwas$81,200. Thesefiguresdifferonlyslightlyfromthosegivenintable1. 20..................... 15,100 16,100 17,900 18,900 8. Tobeincludedintheretiredgroup,thefamilyheadmustreport 40..................... 28,100 30,600 32,800 33,900 beingretiredandnotcurrentlyworkingatanyjoborreportbeingout 60..................... 45,600 49,400 54,700 53,600 80..................... 73,800 79,100 87,600 89,300 ofthelaborforceandovertheageof65.Theother-not-workinggroup 90..................... 101,100 108,900 126,600 129,400 comprisesfamilyheadswhoareunemployedandthosewhoareoutof the labor force but who are neither retired nor over age 65; the 6. Overthe2001−04period,theshareoffamilieswithaheadwith compositionofthisgroupshiftedfrom2001to2004toincludemore less than a high school diploma declined 1.6 percentage points, to familieswithaheadwhohadacollegedegree.In2004,62.1percent 14.4percent.Comparedwith2001,alargershareofthe2004group ofthegroupwasunemployed,andtheremainderwasoutofthelabor wasnonwhiteorHispanic,andtheshareyoungerthan45increased force;in2001,52.5percentofthegroupwasunemployed(datanot slightly(datanotshowninthetables). showninthetables).

Recent Changes in U.S. Family Finances: Evidence from the 2001 and 2004 Survey of Consumer Finances A7 By housing status, median and mean incomes fell 2. Reasonsrespondentsgaveasmostimportantfortheir both for homeowners and for other families from families’saving,distributedbytypeofreason, 1995–2004surveys 2001 to 2004. The decline in the median for home- Percent ownerswasonly0.5percent,butthedeclineforother familieswas6.5percent.Thefactthatthemedianfor Reason 1995 1998 2001 2004 these groups dropped while the overall median Education ................... 10.8 11.0 10.9 11.6 showedagainmaybeexplained,inpart,byaninflux Forthefamily .............. 2.7 4.1 5.1 4.7 Buyingownhome........... 5.1 4.4 4.2 5.0 of new homeowners, who tend, on the one hand, to Purchases ................... 12.8 9.7 9.5 7.7 haveincomeslowerthanthoseofpreviouslyexisting Retirement .................. 23.7 33.0 32.1 34.7 Liquidity.................... 33.0 29.8 31.2 30.0 homeownersand,ontheotherhand,tohaveincomes Investments ................. 4.2 2.0 1.0 1.5 Noparticularreason ......... .8 1.3 1.1 .7 higher than those of remaining renters. Over the Whenaskedforareason, preceding three-year period, median and mean reporteddonotsave .... 6.8 4.9 4.9 4.0 Total ................... 100 100 100 100 incomes had risen both for homeowners and for Note: Seenotetotable1andtextnote10. others. By percentile of net worth, median income increased from 2001 to 2004 only for families above In contrast to the SCF measure, estimates of the the75thpercentileofthewealthdistribution;itfellor personal saving rate from the national income and was little changed for other groups.9 Mean income product accounts (NIPA) show no change on an rose only between the 75th and 90th percentiles of annual basis from 2001 to 2004. However, the SCF the wealth distribution. From 1998 to 2001, the two and NIPA concepts of saving differ in some imporincome measures had increased for all groups but tant ways. First, the underlying SCF question asks particularlyforthetopdecile. only whether the family’s spending has been less than,morethan,oraboutthesameasitsincomeover thepastyear.Thus,fewerfamiliesmaybesaving,but Saving those that are doing so may be saving more. Second, the NIPA measure of saving relies on definitions of Becausesavingoutofcurrentincomeisanimportant incomeandconsumptionthatmaynotbethesameas determinant of family net worth, the SCF asks those that respondents had in mind when answering respondents whether, over the preceding year, the thesurveyquestions.Forexample,theNIPAmeasure family’s spending was less than, more than, or about of personal income includes payments employers equal to its income. Though only qualitative, the make to their employees’ defined-benefit pension answersareausefulindicatorofwhetherfamiliesare plans but not the payments made from such plans to saving. Asking instead for a specific dollar amount families, whereas the SCF measure includes only the wouldrequiremuchmoretimefromrespondentsand latter.TheSCFmeasurealsoincludesrealizedcapital wouldlikelylowertherateofresponsetothesurvey. gains, whereas the NIPA measure excludes capital Overall,from2001to2004theproportionoffami- gainsofallforms,realizedandunrealized. liesthatreportedthattheyhadsavedinthepreceding A separate question in the survey asks about famiyear fell 3.1 percentage points, to 56.1 percent, lies’ more typical saving habits. In 2004, 7.0 percent although the proportion remained higher than in the of families reported that their spending usually 1995 and 1998 surveys. Across most of the demo- exceeds their income; 16.1 percent reported that the graphic groups over the recent three-year period, the twoareusuallyaboutthesame;36.1percentreported predominant pattern is also one of a decline in the thattheytypicallysaveincome‘‘leftover’’attheend proportion of families that saved. In contrast, the oftheyear,incomeofonefamilymember,orunusual 2001 survey had predominantly shown increases additional income; and 40.8 percent reported that from1998. they save regularly (data not shown in the tables). These figures are not much changed over the last threesurveys. 9. Hereareselectedpercentilesofthedistributionofnetworthfor thepastfoursurveys,in2004dollars: The SCF also collects information on families’ most important motivations for saving (table 2).10 Percentile 1995 1998 2001 2004 ofnetworth 25..................... 12,300 11,500 13,500 13,300 10. Althoughfamilieswereaskedtoreporttheirmotivesforsaving 50..................... 70,800 83,100 91,700 93,100 regardless of whether they were currently saving, some families 75..................... 197,800 242,100 301,100 328,500 90..................... 469,000 572,100 780,900 831,600 reportedonlythattheydonotsave.Theanalysishereisconfinedto thefirstreasonreportedbyfamilies.

A8 Federal Reserve Bulletin 2006 3. Familynetworth,byselectedcharacteristicsoffamilies,1995–2004surveys Thousandsof2004dollars 1995 1998 2001 2004 Family characteristic Median Mean Median Mean Median Mean Median Mean Allfamilies..................... 70.8 260.8 83.1 327.5 91.7 421.5 93.1 448.2 (2.4) (6.4) (3.2) (10.7) (3.3) (7.1) (4.3) (9.7) Percentileofincome Lessthan20 .................... 7.4 54.7 6.8 55.4 8.4 56.1 7.5 72.6 20–39.9......................... 41.3 97.4 38.4 111.5 39.6 121.8 34.3 122.0 40–59.9......................... 57.1 126.0 61.9 146.6 66.5 171.4 71.6 193.8 60–79.9......................... 93.6 198.5 130.2 238.3 150.7 311.3 160.0 342.8 80–89.9......................... 157.7 316.8 218.5 377.1 280.3 486.6 311.1 485.0 90–100 ......................... 436.9 1,338.0 524.4 1,793.9 887.9 2,406.7 924.1 2,534.4 Ageofhead(years) Lessthan35 .................... 14.8 53.2 10.6 74.0 12.3 96.6 14.2 73.5 35–44 .......................... 64.2 176.8 73.5 227.6 82.6 276.4 69.4 299.2 45–54 .......................... 116.8 364.8 122.3 420.2 141.6 517.6 144.7 542.7 55–64 .......................... 141.9 471.1 148.2 617.0 193.3 775.4 248.7 843.8 65–74 .......................... 136.6 429.3 169.8 541.1 187.8 717.9 190.1 690.9 75ormore...................... 114.5 317.9 145.6 360.3 161.2 496.2 163.1 528.1 Educationofhead Nohighschooldiploma ......... 27.9 103.7 24.5 91.4 27.2 109.7 20.6 136.5 Highschooldiploma ............ 63.9 163.7 62.7 182.9 61.8 192.5 68.7 196.8 Somecollege ................... 57.6 232.3 85.6 275.5 76.3 303.8 69.3 308.6 Collegedegree .................. 128.6 473.7 169.7 612.3 227.2 845.7 226.1 851.3 Raceorethnicityofrespondent Whitenon-Hispanic ............. 94.3 308.7 111.0 391.1 129.6 518.7 140.7 561.8 NonwhiteorHispanic ........... 19.5 94.9 19.3 116.5 19.1 123.8 24.8 153.1 Currentworkstatusofhead Workingforsomeoneelse ....... 60.3 168.4 61.2 194.8 69.3 240.1 67.2 268.5 Self-employed .................. 191.8 862.8 288.0 1,071.3 375.2 1,340.6 335.6 1,423.2 Retired ......................... 99.9 277.2 131.0 356.5 120.4 479.2 139.8 469.0 Othernotworking............... 4.5 70.1 4.1 85.8 9.5 191.7 11.8 162.3 Region Northeast ....................... 102.0 308.9 109.3 351.3 98.3 480.0 161.7 569.1 Midwest ........................ 80.8 244.7 93.1 288.5 111.3 361.6 115.0 436.1 South ........................... 54.2 229.5 71.0 309.6 78.6 400.4 63.8 348.0 West............................ 67.4 286.1 71.1 379.1 93.3 468.8 94.8 523.7 Housingstatus Owner .......................... 128.1 373.7 153.2 468.7 182.9 594.8 184.4 624.9 Renterorother.................. 6.0 53.8 4.9 50.4 5.1 58.5 4.0 54.1 Percentileofnetworth Lessthan25 .................... 1.2 −.2 .6 −2.1 1.2 † 1.7 −1.4 25–49.9......................... 34.7 37.6 37.9 41.6 43.4 47.0 43.6 47.1 50–74.9......................... 117.1 122.6 139.7 149.1 166.8 176.6 170.7 185.4 75–89.9......................... 272.3 293.6 357.7 372.6 458.2 478.6 506.8 526.7 90–100 ......................... 836.7 1,766.7 1,039.1 2,244.2 1,386.6 2,936.1 1,430.1 3,114.2 Note: Seenotetotable1. †Lessthan0.05($50). In 2004, the most frequently reported motive was NET WORTH retirement-related (34.7 percent of families), and the next most frequently reported was liquidity-related From 2001 to 2004, real net worth (wealth)—the (30.0percentoffamilies),aresponsethatisgenerally difference between families’ gross assets and their taken to be indicative of saving for precautionary liabilities—rose, though the mean rose notably more reasons.11 At least since 1995, these have been the strongly than the median (table 3). The median rose dominant reported reasons, but saving for retirement 1.5 percent, while the mean rose 6.3 percent; the has increased notably in importance. The education- corresponding values for the period from 1998 to related motive also appears to be important; in 2004, 2001were10.3percentand28.7percent.12 11.6 percent of families reported it as their primary motive. The importance of saving for purchases has 12. The Federal Reserve’s flow of funds accounts provide an fallenovertime. estimateofthetotalnetworthofthehouseholdsector,whichincludes both households and nonprofit institutions. Between year-end 2001 andyear-end2004,theflowoffundsestimateofrealnetworthrose 12.3percent.Accountingforthe3.6percentincreaseinthenumberof households over this period produces a change in net worth about 11. Liquidity-relatedreasonsinclude‘‘emergencies,’’thepossibili- 2percentagepointshigherthantheSCFestimateoftheincreaseinthe tiesofunemploymentandillness,andtheneedforreadymoney. mean.

Recent Changes in U.S. Family Finances: Evidence from the 2001 and 2004 Survey of Consumer Finances A9 By age group, median and mean net worth show a median net worth rose for most income groups; the ‘‘hump’’ pattern that generally has its peak in the mean rose for all income groups, but the increases 55–64agegroup.Thispatternreflectsbothlife-cycle werestrongestforthetoptwoincomequintiles. saving behavior and growth in real wages over time. The survey shows some substantial movements of Themedianandmeanvaluesofwealthriseintandem wealthbyagegroupbetween2001and2004.Median with income, a relationship reflecting both income wealth rose most strongly—28.7 percent—for the earned from assets and a higher likelihood of saving 55–64 age group, which had also experienced the among higher-income families. Wealth shows strong largest median gain over the previous three-year differentialsacrossgroupsdefinedintermsofeduca- period. The less-than-35 age group also saw a subtion, racial and ethnic background, work status, and stantial gain in the median—15.4 percent—over the housing status; these differentials generally mirror more recent period; at the same time, median wealth those for income, but the wealth differences are fell 16.0 percent for the 35–44 age group. Mean larger. wealth rose for all age groups except for the lessthan-35groupandthe65–74group. More than offsetting gains over the 1998 to 2001 Net Worth by Demographic Category period,mediannetworthfell24.3percentfrom2001 to2004forfamiliesheadedbypersonswithlessthan Analysis by demographic group for the 2001–04 a high school diploma or equivalent; the median for period shows a complicated pattern of gains and the group with some college education also fell, by lossesinmedianandmeannetworth,withchangesin 9.2percent.Themedianroseonlyforfamiliesheaded the median often opposing those in the mean. The by persons with a high school diploma or equivapatterns suggest correspondingly complex change in lent. Mean wealth rose or held about steady for all the underlying ownership and values of assets and education groups. For the no-high-school-diploma debts and the distribution of wealth within demo- group,meanwealthrose24.4percent;giventhelarge graphic groups; to some degree, movements of fami- decline in the median for this group, this result suglies between groups may also explain some of the gestsashiftinthedistributionofnetworthwithinthe shiftsinwealth. group. The college-degree group, which had experi- Median and mean net worth rose or held about enced the largest gains in the preceding three-year steady for all percentile groups of the distribution of period, saw little change in its median or mean networthexceptforfamiliesinthelowest25percent wealth. of the distribution of net worth. In that group, the The data show gains from 2001 to 2004 in median median rose from $1,200 to $1,700, up from $600 in andmeanwealthforwhitenon-Hispanicfamiliesand 1998; the mean fell from near zero to negative for nonwhite or Hispanic families, but the gains for $1,400, closer to its 1998 value of negative $2,100. thelatterweremuchlargerinpercentageterms.13For For the rest of the net worth distribution, growth in white non-Hispanics, the median rose 8.6 percent, the median and mean over the recent three-year and the mean rose 8.3 percent; for nonwhites or period was notable for the groups above the 50th Hispanics, the median rose 29.8 percent, and the percentileandparticularlysoforthoseinthe75–89.9 mean rose 23.7 percent. However, as was the case percentile group; the gains for the groups in the top with income, these measures of the wealth of nonhalfofthedistributioncontinuedauniformpatternof whitesorHispanicsarefarlowerthanthoseforother gainsatleastbackto1995. families, and the differences are even larger than Overtherecentperiod,mediannetworthincreased those in the case of family income; in 2004, the for all income groups above the 40th percentile and median wealth of nonwhite or Hispanic families was especiallyforthe80–89.9percentilegroup,forwhich only 17.6 percent of that for other families. In conthemedianrose11.0percent;themeanforthisgroup trast to the whole group of nonwhite or Hispanic was little changed. The mean for the lowest quintile families, the subgroup of African American families hadthelargestproportionalincrease—29.4percent— saw virtually no change in their median net worth but the rise appears to be due to an increase in the from 2001 ($20,300) to 2004 ($20,400), but their fraction of the group consisting of relatively wealthy families with incomes that are likely to have been 13. IftheinformationonHispanicorLatinoethnicidentificationis temporarily low. The mean increased or held about usedintheclassificationofthe2004results,themediannetworthof steady for the other income groups, and it rose par- nonwhitesorHispanicswas$27,100,andthemeanwas$162,500;for otherfamilies,themedianwas$142,700,andthemeanwas$566,600. ticularly for the 40–59.9 percentile group—a These figures are slightly higher than the corresponding values 13.1 percent gain. Over the preceding years shown, reportedintable3.

A10 Federal Reserve Bulletin 2006 mean net worth rose 37.1 percent, from $80,700 to less than 35 and by those aged 65 or older, nonwhite $110,600 (data not shown in the tables). Over the or Hispanic families, families with a head who was 1998–2001 period, the growth of wealth for non- notworking,renters,andfamiliesinthebottomquarwhites or Hispanics had substantially lagged that for tile of the wealth distribution. The 2001 ownership otherfamilies. levels for other groups were already at or near Among work-status groups, both of the groups in 100percent. which the family head was currently working saw a Over the recent three-year period, the median real decline in median wealth from 2001 to 2004, while value of assets among families having any asset rose the median rose substantially for the other groups. 10.3 percent, from $156,800 to $172,900 (second However,themeansshowtheoppositepattern:gains halfoftables8.Aand8.B,lastcolumn).Thatgainfar for those working and losses for the other groups. exceeds the 1.5 percent rise in median net worth Over the preceding three years, both median and computed for all families regardless of whether they mean wealth had risen for all work-status groups haveassets.Thisdivergencesuggeststhatchangesin exceptfortheretiredgroup,whichhadseenadecline debt holdings, which in some cases appear to have a initsmedianwealth. direct connection to the increased assets of families, Between 2001 and 2004, the mean and median are a key factor. Median assets rose substantially for wealth of families increased in all regions of the most demographic groups. However, declines were country except for the South, where the median notable for almost all the groups that saw the largest declined18.8percentandthemeanfell13.1percent. increases in ownership levels—the lowest quintile of The most striking gain is the 64.5 percent rise in theincomedistribution,theyoungestagegroup,nonmedian wealth for the Northeast region, where it had white or Hispanic families, the other-not-working declinedinthe1998–2001period. group, renters, and the lowest quartile of the wealth By housing status, the mean net worth of home- distribution. One particularly noteworthy increase in owners rose 5.1 percent from 2001 to 2004. The the median value of assets was in the 55–64 age medianforhomeownerswaslittlechanged;forother group,whichsawariseof45.7percent.Theprevailfamilies, the median fell 21.6 percent, and the mean ing impression from the preceding three years had fell 7.5 percent. This pattern is likely explained in been one of broad increases in the median. In the partbythegrowthinhomeownershipovertheperiod, recentthree-yearperiod,meanassetsrose8.6percent as discussed later in this article. New homeowners (secondhalfoftables8.Aand8.B,memoline). tend to have less wealth than previously existing homeowners, having had less time to benefit from Financial Assets appreciation of home prices. At the same time, the wealth of the remaining renters will tend to be Financialassetsasashareoftotalassetsfell6.3perdepressed by rising homeownership because the centage points from 2001 to 2004, to 35.7 percent renter group will have fewer families with assets (table 4, memo line); the decline is from a level in sufficient to initiate a home purchase. In the preceding three-year period, median and mean wealth had increasedforbothgroups. 4. Valueoffinancialassetsofallfamilies,distributed bytypeofasset,1995–2004surveys ASSETS Percent Typeoffinancial 1995 1998 2001 2004 asset Movements in the dollar value of families’ financial assets (tables 4, 5, and 6) and nonfinancial assets Transactionaccounts .......... 13.9 11.4 11.5 13.2 Certificatesofdeposit ......... 5.6 4.3 3.1 3.7 (tables7and8)are,bydefinition,aresultofchanges Savingsbonds ................ 1.3 .7 .7 .5 in valuation and in the patterns of ownership. The Bonds ........................ 6.3 4.3 4.6 5.3 Stocks ........................ 15.6 22.7 21.7 17.6 overall proportion of families with any asset rose Pooledinvestmentfunds (excludingmoneymarket 1.2 percentage points, to 97.9 percent, in 2004 (first funds).................... 12.7 12.4 12.2 14.7 Retirementaccounts ........... 28.1 27.6 28.4 32.0 half of tables 8.A and 8.B, last column); this rise Cashvaluelifeinsurance ...... 7.2 6.4 5.3 3.0 continues a trend, at work at least since 1995, that Othermanagedassets ......... 5.9 8.6 10.6 8.0 Other ......................... 3.3 1.7 2.0 2.1 had been interrupted by a pause in 2001 (data not Total ..................... 100 100 100 100 shown in the tables). The largest increases in the Memo proportion holding any asset were in the following Financialassetsasa shareoftotalassets ....... 36.7 40.7 42.0 35.7 demographic groups: the lowest quintile of the Note: For this and following tables, see text for definition of asset incomedistribution,familiesheadedbypersonsaged categories.Alsoseenotetotable1.

Recent Changes in U.S. Family Finances: Evidence from the 2001 and 2004 Survey of Consumer Finances A11 5. Familyholdingsoffinancialassets,byselectedcharacteristicsoffamiliesandtypeofasset,2001and2004surveys A. 2001SurveyofConsumerFinances Pooled Cash Trans- Certifi- Retire- Other Any Family Savings invest- value action catesof Bonds Stocks ment managed Other financial characteristic bonds ment life accounts deposit accounts assets asset funds insurance Percentageoffamiliesholdingasset Allfamilies..................... 91.4 15.7 16.7 3.0 21.3 17.7 52.2 28.0 6.6 9.4 93.4 Percentileofincome Lessthan20 .................... 71.6 10.0 3.8 * 3.8 3.6 13.2 13.8 2.2 6.2 75.5 20–39.9......................... 90.3 14.7 11.0 * 11.2 9.5 33.3 24.7 3.3 10.2 93.6 40–59.9......................... 96.6 17.4 14.1 1.5 16.4 15.7 52.8 25.6 5.4 9.9 98.3 60–79.9......................... 99.1 16.0 24.4 3.7 26.2 20.6 75.7 35.7 8.5 9.2 99.6 80–89.9......................... 99.7 18.3 30.3 3.9 37.0 29.0 83.7 38.6 10.7 10.8 99.8 90–100 ......................... 99.2 22.0 29.7 12.7 60.6 48.8 88.3 41.8 16.7 12.5 99.7 Ageofhead(years) Lessthan35 .................... 87.1 6.3 12.7 * 17.4 11.5 45.1 15.0 2.1 10.5 89.7 35–44 .......................... 91.1 9.8 22.6 2.1 21.6 17.5 61.4 27.0 3.1 9.7 93.5 45–54 .......................... 92.7 15.2 21.0 2.8 22.0 20.2 63.4 31.1 6.4 8.5 94.7 55–64 .......................... 93.8 14.4 14.3 6.1 26.7 21.3 59.1 35.7 13.0 10.6 95.0 65–74 .......................... 93.8 29.7 11.3 3.9 20.5 19.9 44.0 36.7 11.8 8.5 94.6 75ormore...................... 93.7 36.5 12.5 5.7 21.8 19.5 25.7 33.3 11.2 7.7 95.1 Raceorethnicityofrespondent Whitenon-Hispanic ............. 95.3 18.5 19.5 3.8 24.7 21.0 57.1 29.9 8.2 9.4 96.7 NonwhiteorHispanic ........... 79.4 6.8 8.1 .4 11.0 7.4 37.4 22.0 1.8 9.6 83.2 Currentworkstatusofhead Workingforsomeoneelse ....... 92.9 11.3 19.4 2.0 20.9 17.3 61.5 27.4 5.3 9.5 94.9 Self-employed .................. 95.9 18.7 16.6 6.1 29.8 22.9 58.9 34.6 6.9 12.4 97.6 Retired ......................... 89.0 27.1 11.4 4.5 19.6 17.3 29.2 29.0 10.4 8.1 90.9 Othernotworking............... 72.2 7.8 7.5 * 13.3 10.9 26.8 12.9 5.6 6.5 74.2 Housingstatus Owner .......................... 96.7 20.0 21.2 4.0 27.0 22.7 62.6 34.5 8.9 8.8 97.8 Renterorother.................. 80.3 6.7 7.2 .7 9.3 7.1 30.4 14.3 2.0 10.6 84.1 Percentileofnetworth Lessthan25 .................... 73.7 1.8 4.3 * 5.0 2.5 18.9 6.9 * 8.1 78.0 25–49.9......................... 94.2 8.8 12.8 * 9.5 7.2 45.3 26.0 1.3 8.7 96.7 50–74.9......................... 98.2 23.2 23.6 * 20.3 17.5 63.2 34.6 6.2 8.7 98.9 75–89.9......................... 99.6 30.1 25.9 5.3 41.2 36.0 77.6 41.7 13.9 9.6 99.8 90–100 ......................... 99.6 26.9 26.3 18.4 64.3 54.8 87.4 48.6 26.4 16.2 100.0 2001 that marked the high point observed in the who were not working and declined for other worksurvey.Therelativesharesofvariousfinancialassets statusgroups.Ownershipalsorosenotablyforrenters also shifted. Declines in the percentage shares of andfornonwhiteorHispanicfamilies. direct holdings of publicly traded stocks, cash value Parallelingthedropintheoverallratiooffinancial life insurance, and ‘‘other managed assets’’ were assets to total assets over the recent period, the approximately balanced by increases in the shares of medianholdingoffinancialassetsforfamilieshaving retirement accounts, pooled investment funds, and such assets fell 22.8 percent (second half of transaction accounts.14 After showing a declining tables 5.A and 5.B, last column), while the mean fell trend in earlier survey years, the share of certificates 6.9 percent (memo line). The change in the median ofdepositedgedup. morethanoffsettheincreaseoverthepreviousthree- Overall, the ownership of any financial asset over year period. The picture is one of declines in the therecentperiodedgeduponly0.4percentagepoint, medians over the recent period for almost every to 93.8 percent (first half of tables 5.A and 5.B, last demographic group; exceptions were the eighth column). However, the recent data show some pro- income decile and the 55–64 age group. Mean holdnounced patterns of change for some demographic ings declined for every group (means of groups not groups. By income, the largest changes in ownership showninthetables). were a rise for the lowest quintile and a fall for the second quintile; by age, notable increases appeared onlyforthegroupsofthose65orolder;andbywork Transaction Accounts and Certificates of Deposit status, ownership rose for families headed by people In 2004, 91.3 percent of families had some type of transaction account—a category comprising 14. Thedefinitionsofassetcategoriesintable4aregivenbelow,in thesectionsoftextdevotedtothem. checking, savings, and money market deposit

A12 Federal Reserve Bulletin 2006 5.—Continued A. 2001SurveyofConsumerFinances—Continued Pooled Cash Trans- Certifi- Retire- Other Any Family Savings invest- value action catesof Bonds Stocks ment managed Other financial characteristic bonds ment life accounts deposit accounts assets asset funds insurance Medianvalueofholdingsforfamiliesholdingasset(thousandsof2004dollars) Allfamilies..................... 4.2 16.0 1.1 46.3 21.3 37.3 30.9 10.7 74.6 4.3 29.8 Percentileofincome Lessthan20 .................... .9 10.7 1.1 * 8.0 22.4 4.8 3.8 25.8 1.8 2.1 20–39.9......................... 1.9 14.9 .6 * 10.7 25.6 8.5 6.6 38.3 3.2 8.4 40–59.9......................... 3.0 13.8 .5 10.7 8.5 25.6 14.5 7.5 74.6 3.2 18.2 60–79.9......................... 5.5 16.0 1.1 42.6 18.1 32.0 32.0 12.8 63.9 3.2 59.1 80–89.9......................... 10.1 13.8 1.1 53.3 21.3 29.8 58.6 10.7 74.6 7.5 103.4 90–100 ......................... 27.7 26.6 2.1 94.5 53.3 93.2 138.5 25.6 119.3 16.0 387.7 Ageofhead(years) Lessthan35 .................... 1.9 4.3 .3 * 6.1 9.6 7.0 10.7 42.6 1.7 6.6 35–44 .......................... 3.6 6.4 1.1 14.5 16.0 18.6 30.4 9.6 53.3 2.1 28.6 45–54 .......................... 4.8 12.8 1.1 63.9 16.0 41.0 51.1 11.7 63.9 5.3 48.0 55–64 .......................... 5.9 20.2 2.7 63.9 42.6 63.9 58.6 10.7 58.6 10.7 59.8 65–74 .......................... 8.5 21.3 2.1 76.1 90.5 74.6 63.9 9.3 127.8 8.5 54.7 75ormore...................... 7.8 26.6 3.2 37.3 63.9 74.6 49.0 7.5 106.5 19.2 42.6 Raceorethnicityofrespondent Whitenon-Hispanic ............. 5.1 16.0 1.1 53.3 23.4 42.6 37.3 10.7 74.6 5.3 41.3 NonwhiteorHispanic ........... 1.8 9.6 .7 8.1 8.5 18.6 10.7 9.3 47.9 1.6 7.6 Currentworkstatusofhead Workingforsomeoneelse ....... 3.4 9.6 1.1 27.7 11.7 21.3 26.0 10.1 58.6 2.7 25.7 Self-employed .................. 8.5 17.0 2.1 76.5 37.3 104.4 58.2 18.1 116.1 12.8 65.0 Retired ......................... 5.3 26.6 4.3 53.3 63.9 74.6 57.5 9.6 106.5 10.7 34.4 Othernotworking............... 1.9 42.6 .3 * 8.5 42.6 21.3 10.7 41.5 2.1 5.6 Housingstatus Owner .......................... 6.2 16.0 1.3 53.3 24.5 42.6 40.7 10.7 74.6 6.4 53.5 Renterorother.................. 1.3 10.7 .4 31.5 6.7 10.7 7.2 8.0 42.6 2.1 4.2 Percentileofnetworth Lessthan25 .................... .7 1.6 .2 * 1.4 2.1 2.1 1.9 * 1.1 1.4 25–49.9......................... 2.3 5.3 .5 * 3.4 5.3 8.0 5.5 10.7 2.4 11.2 50–74.9......................... 5.9 12.2 1.2 * 8.8 16.0 32.0 9.6 23.4 4.8 56.5 75–89.9......................... 14.5 21.3 2.1 21.3 27.6 39.9 81.5 12.8 74.6 10.7 214.8 90–100 ......................... 38.3 42.6 2.1 95.9 129.9 149.1 202.4 32.0 213.0 35.1 753.5 Memo Meanvalueofholdingsfor familiesholdingasset ....... 25.3 39.9 8.4 310.2 204.8 139.3 109.3 38.4 321.5 41.9 215.6 Note: Seenotetotable1. *Tenorfewerobservations. accounts, money market mutual funds, and call typesoftransactionaccountdeclined,asshowninthe accounts at brokerages. The ownership rate, essen- followingtable: tially unchanged from 2001, was 90.6 in the 1998 survey and notably lower before then. Families that Familiesholding Typeof did not have any type of transaction account in 2004 transactionaccount 2004 Change,2001−04 were disproportionately likely to have low incomes, (percent) (percentagepoints) to be headed by a person younger than 35, to be Checking.............. 89.4 2.1 nonwhite or Hispanic, to be headed by a person who Savings ............... 47.1 −8.1 Moneymarket ......... 21.1 −.6 was neither working nor retired, to be renters, or to Call................... 2.5 −.7 have relatively low levels of wealth. Over the 2001–04 period, ownership rose notably for families Mostofthechangeappearstoreflectconsolidation atthebottomoftheincomeandwealthdistributions, ofmultipletypesofaccountintoacheckingaccount; familiesheadedbypersonsaged75orolder,families for many such families, the relatively low interest with heads who were not working, nonwhites or rates on deposits may have been insufficient to com- Hispanics,andrenters. pensate for the effort of managing multiple accounts. Underlyingthelevelingoffinthegrowthofowner- See box ‘‘Families without a Checking Account’’ for ship of transaction accounts in the recent three-year a discussion of reasons that some families do not period was a larger shift in the types of account haveacheckingaccount. families used. The share of families with a checking Median holdings in transaction accounts for those accountrose,andthesharesoffamilieswithallother who had such accounts fell 9.5 percent from 2001 to

Recent Changes in U.S. Family Finances: Evidence from the 2001 and 2004 Survey of Consumer Finances A13 5. Familyholdingsoffinancialassets,byselectedcharacteristicsoffamiliesandtypeofasset,2001and2004surveys—Continued B. 2004SurveyofConsumerFinances Pooled Cash Trans- Certifi- Retire- Other Any Family Savings invest- value action catesof Bonds Stocks ment managed Other financial characteristic bonds ment life accounts deposit accounts assets asset funds insurance Percentageoffamiliesholdingasset Allfamilies..................... 91.3 12.7 17.6 1.8 20.7 15.0 49.7 24.2 7.3 10.0 93.8 Percentileofincome Lessthan20 .................... 75.5 5.0 6.2 * 5.1 3.6 10.1 14.0 3.1 7.1 80.1 20–39.9......................... 87.3 12.7 8.8 * 8.2 7.6 30.0 19.2 4.9 9.9 91.5 40–59.9......................... 95.9 11.8 15.4 * 16.3 12.7 53.4 24.2 7.9 9.3 98.5 60–79.9......................... 98.4 14.9 26.6 2.2 28.2 18.6 69.7 29.8 7.8 11.2 99.1 80–89.9......................... 99.1 16.3 32.3 2.8 35.8 26.2 81.9 29.5 12.1 11.4 99.8 90–100 ......................... 100.0 21.5 29.9 8.8 55.0 39.1 88.5 38.1 13.0 13.4 100.0 Ageofhead(years) Lessthan35 .................... 86.4 5.6 15.3 * 13.3 8.3 40.2 11.0 2.9 11.6 90.1 35–44 .......................... 90.8 6.7 23.3 .6 18.5 12.3 55.9 20.1 3.7 10.0 93.6 45–54 .......................... 91.8 11.9 21.0 1.8 23.2 18.2 57.7 26.0 6.2 12.1 93.6 55–64 .......................... 93.2 18.1 15.2 3.3 29.1 20.6 62.9 32.1 9.4 7.2 95.2 65–74 .......................... 93.9 19.9 14.9 4.3 25.4 18.6 43.2 34.8 12.8 8.1 96.5 75ormore...................... 96.4 25.7 11.0 3.0 18.4 16.6 29.2 34.0 16.7 8.1 97.6 Raceorethnicityofrespondent Whitenon-Hispanic ............. 95.5 15.3 21.1 2.5 25.5 18.9 56.1 26.8 9.2 10.2 97.2 NonwhiteorHispanic ........... 80.6 6.0 8.5 * 8.0 5.0 32.9 17.4 2.1 9.4 85.0 Currentworkstatusofhead Workingforsomeoneelse ....... 92.2 9.8 20.1 .8 19.6 13.5 57.1 21.8 5.4 9.5 94.5 Self-employed .................. 94.4 14.2 18.7 4.3 31.6 22.3 54.6 29.8 7.6 15.1 96.1 Retired ......................... 90.4 20.2 11.4 3.5 19.0 16.2 32.9 29.7 12.8 8.4 93.6 Othernotworking............... 76.2 7.9 14.5 * 14.3 10.2 24.9 10.7 * 11.5 79.6 Housingstatus Owner .......................... 96.0 15.9 21.2 2.6 25.8 19.2 60.2 30.1 9.6 9.6 97.5 Renterorother.................. 80.9 5.6 9.5 .2 9.1 5.7 26.2 11.0 2.0 10.9 85.5 Percentileofnetworth Lessthan25 .................... 75.4 2.2 6.2 * 3.6 2.0 14.3 7.7 * 6.9 79.8 25–49.9......................... 92.0 6.5 13.2 * 9.3 7.2 43.1 19.3 2.3 9.5 96.1 50–74.9......................... 98.0 16.0 22.7 * 21.0 12.5 61.8 30.1 8.8 10.2 99.4 75–89.9......................... 99.7 24.2 28.5 3.2 39.1 32.4 77.6 36.7 15.6 11.2 100.0 90–100 ......................... 100.0 28.8 28.1 12.7 62.9 47.3 82.5 43.8 21.0 16.4 100.0 2004,whilethemeanrose7.1percent.Acrossdemo- sametime,theoverallrealmedianvaluefell6.3pergraphic groups, the patterns of change in the median cent. Across income groups, declines in the median are mainly a mixture of substantial increases and were concentrated in the groups below the 60th perdeclines. Median balances fell for the lowest two centile, whereas the medians for the higher-income income groups and the lowest three wealth groups groups increased; along with the fact that the overall but rose or held steady for the other income and mean holding rose 37.6 percent, this result suggests wealth groups. Across age groups, the median thattheconcentrationofCDbalancesroseamongthe increased only for the 55–64 group and fell or was higher-incomegroups.ThemedianvalueofCDsrose unchangedforotherfamilies.Byworkstatus,median forallwealthgroupsexceptthethirdquartile. balances rose substantially for the self-employed group. Holdings declined for both of the racial and ethnic groups and for both of the housing-status Savings Bonds and Other Bonds groups. Certificates of deposit (CDs)—interest-bearing Savingsbondsareowneddisproportionatelybyfamideposits with a set term—are traditionally viewed as lies with incomes in the highest 40 percent of the a low-risk saving vehicle, and they are often used by distribution and by families in the top half of the people who desire a safe haven from the volatility of distribution of net worth. Over the 2001–04 period, financial markets. Over the 2001–04 period, the the ownership of savings bonds rose 0.9 percentage attractivenessofCDsdeclinedastheinterestrateson point overall, and it rose for most demographic them fell. The resulting 3.0 percentage point decline groups; these gains partially offset declines in the in ownership broke the slow upward movement seen preceding three-year period. Median holdings fell since 1998. Over the more recent period, ownership 9.1 percent, and the mean fell 31.0 percent between declined among most demographic groups. At the 2001and2004.

A14 Federal Reserve Bulletin 2006 5.—Continued B. 2004SurveyofConsumerFinances—Continued Pooled Cash Trans- Certifi- Retire- Other Any Family Savings invest- value action catesof Bonds Stocks ment managed Other financial characteristic bonds ment life accounts deposit accounts assets asset funds insurance Medianvalueofholdingsforfamiliesholdingasset(thousandsof2004dollars) Allfamilies..................... 3.8 15.0 1.0 65.0 15.0 40.4 35.2 6.0 45.0 4.0 23.0 Percentileofincome Lessthan20 .................... .6 10.0 .4 * 6.0 15.3 5.0 2.8 22.0 2.5 1.3 20–39.9......................... 1.5 14.0 .6 * 8.0 25.0 10.0 3.9 50.0 2.0 4.9 40–59.9......................... 3.0 10.0 .8 * 12.0 23.0 17.2 5.0 36.0 2.5 15.5 60–79.9......................... 6.6 18.0 1.0 80.0 10.0 25.5 32.0 7.0 35.0 4.0 48.5 80–89.9......................... 11.0 20.0 .8 26.7 15.0 33.5 70.0 10.0 50.0 5.0 108.2 90–100 ......................... 28.0 33.0 2.0 160.0 57.0 125.0 182.7 20.0 100.0 20.0 365.1 Ageofhead(years) Lessthan35 .................... 1.8 4.0 .5 * 4.4 8.0 11.0 3.0 5.0 1.0 5.2 35–44 .......................... 3.0 10.0 .5 10.0 10.0 15.9 27.9 5.0 18.3 3.5 19.0 45–54 .......................... 4.8 11.0 1.0 30.0 14.5 50.0 55.5 8.0 43.0 5.0 38.6 55–64 .......................... 6.7 29.0 2.5 80.0 25.0 75.0 83.0 10.0 65.0 7.0 78.0 65–74 .......................... 5.5 20.0 3.0 40.0 42.0 60.0 80.0 8.0 60.0 10.0 36.1 75ormore...................... 6.5 22.0 5.0 295.0 50.0 60.0 30.0 5.0 50.0 22.0 38.8 Raceorethnicityofrespondent Whitenon-Hispanic ............. 5.0 16.0 1.0 80.0 18.0 45.0 41.0 7.0 45.0 5.0 36.0 NonwhiteorHispanic ........... 1.5 12.0 .6 * 5.3 18.0 16.0 5.0 40.0 2.5 5.0 Currentworkstatusofhead Workingforsomeoneelse ....... 3.1 10.0 .7 25.0 10.0 25.0 30.0 5.4 50.0 3.0 20.5 Self-employed .................. 10.0 20.0 1.9 130.0 25.0 60.0 60.0 10.5 42.0 6.0 53.2 Retired ......................... 4.2 25.0 3.0 90.0 45.0 75.0 47.0 5.0 45.0 10.0 26.5 Othernotworking............... 2.0 8.0 2.0 * 5.0 15.9 31.0 8.4 * 3.0 5.0 Housingstatus Owner .......................... 6.0 20.0 1.0 65.0 20.0 50.0 46.0 7.0 45.0 6.0 47.9 Renterorother.................. 1.1 7.0 .7 130.0 4.5 10.0 11.0 3.0 42.0 2.0 3.0 Percentileofnetworth Lessthan25 .................... .5 2.0 .3 * 1.9 2.0 2.9 .8 * .7 1.0 25–49.9......................... 2.0 5.8 .5 * 3.5 7.4 11.8 4.0 9.4 2.0 9.9 50–74.9......................... 5.8 10.4 1.0 * 8.0 16.0 33.5 5.0 22.0 5.0 47.2 75–89.9......................... 15.8 31.0 2.0 25.0 20.0 50.0 95.7 10.0 50.0 7.0 203.0 90–100 ......................... 43.0 46.0 2.5 111.1 110.0 160.0 264.0 20.0 135.0 40.0 728.8 Memo Meanvalueofholdingsfor familiesholdingasset ....... 27.1 54.9 5.8 547.0 160.3 184.0 121.3 23.1 207.0 39.5 200.7 Note: Seenotetotable1. *Tenorfewerobservations. Other bond types tend to be very narrowly held, that had them rose substantially; the median went up and the ownership rate, which had been flat since 40.4percentandthemean76.3percent. 1995, fell to 1.8 percent in 2004, a drop of 1.2 percentage points from 2001.15 The underlying data in the survey suggest that, among families that owned Publicly Traded Stock bonds, the proportion that owned mortgage-backed bonds and corporate or foreign bonds rose in the The direct ownership of publicly traded stocks is recent period, while ownership of tax-exempt and morewidespreadthanthedirectownershipofbonds, other government bills and bonds fell somewhat. but, as with bonds, it is also concentrated among Ownership of any type of bond is notably concen- high-income and high-wealth families. The share of trated among the highest tiers of the income and families with any such stock holdings declined wealth distributions, and these groups saw compara- 0.6 percentage point from 2001 to 2004 after having tivelylargedeclinesinownershipfrom2001to2004. risen steadily since the 1995 survey. Over demo- At the same time, the value of bonds for families graphic groups, the decline was most marked for the highestdecileoftheincomedistribution. Although the major stock price indexes had 15. ‘‘Otherbonds’’asreportedinthesurveyarehelddirectlyand declined in 2001 to about the levels of 1998 and had include corporate and mortgage-backed bonds; federal, state, and recoveredbythetimeofthe2004survey,themedian local government bonds; and foreign bonds. In the survey, financial amount of directly held stock for families with such assets held indirectly are those held in retirement accounts and in othermanagedassets. assets was 29.6 percent lower in 2004 than in 2001;

Recent Changes in U.S. Family Finances: Evidence from the 2001 and 2004 Survey of Consumer Finances A15 FamilieswithoutaCheckingAccount Between 2001 and 2004, the proportion of families with Whenattentionisfurtherrestrictedtofamiliesthatonce any type of transaction account barely changed (table 5), had a checking account (data not shown in the table), the whilethesharewithoutacheckingaccountfell2.1percent- general pattern of responses is similar to that for all age points, from 12.7 percent to 10.6 percent (data not families without a checking account, but there were some showninthetables).Thedeclineinthefractionoffamilies notable changes over the period. For families that once withoutacheckingaccountfollowsalongertrend;in1992, hadacheckingaccount,theproportionreportingthatthey thesharewas16.6percent.1 couldnotmanageacheckingaccountordidnotlikebanks Among families without a checking account in 2004, both rose from 2001. These increases are offset by 52.1 percent had held such an account in the past, 55.1 decreases in the proportion reporting that they found serpercenthadincomesinthelowest20percentofthatdistri- vice charges too high, did not write enough checks, had bution,56.6percentwereheadedbypersonsyoungerthan credit problems, or did not have enough money for an 45,and61.0percentwerenonwhiteorHispanic. accounttobeworthwhile. TheSCFaskedallfamiliesthatdidnothaveachecking account to give a reason for not having an account (see Distributionofreasonscitedbyrespondentsfortheir table).Themostcommonlyreportedreason—givenby27.9 families’nothavingacheckingaccount,byreason, percent of families—was that the family did not write 1995–2004surveys enough checks to make account ownership worthwhile. Percent Another 14.4 percent said that they did not have enough money to make account ownership worthwhile, and 22.6 Reason 1995 1998 2001 2004 percentsaidthattheydidnotlikedealingwithbanks.The Donotwriteenoughchecks pattern of the reported reasons differs only slightly from tomakeitworthwhile ..... 25.3 28.4 28.6 27.9 thatin2001. Minimumbalanceistoohigh ... 8.8 8.6 6.5 5.6 Donotlikedealingwithbanks.. 18.6 18.5 22.6 22.6 Servicechargesaretoohigh .... 8.4 11.0 10.2 11.6 Cannotmanageorbalance acheckingaccount ........ 8.0 7.2 6.6 6.8 1. Forthedefinitionof‘‘transactionaccount,’’seethemaintext.Fora Nobankhasconvenienthours discussionofthewaysthatlower-incomefamiliesobtaincheckingandcredit orlocation ................ 1.2 1.2 .4 1.1 services and the effects that developments in electronic transactions may Donothaveenoughmoney..... 20.0 12.9 14.0 14.4 have on such families, see Jeanne M. Hogarth and Kevin H. O’Donnell Creditproblems................ 1.4 2.7 3.6 2.4 Donotneed/wantanaccount ... 4.9 6.3 5.1 5.2 (1999),‘‘BankingRelationshipsofLower-IncomeFamiliesandtheGovern- Other ......................... 3.5 3.1 2.4 2.4 mentalTrendtowardElectronicPayments,’’FederalReserveBulletin,vol.85 Total ..................... 100 100 100 100 (July),pp.459–73. themeanwas21.7percentlower.Thedeclinesinthe 15.0 percent of families.16 Typically, the pattern of median and mean were shared by most demographic ownershipofpooledinvestmentfundsresemblesthat groups (means for groups not shown in the tables); a of stocks, but in contrast to the mixed changes in notable exception was the increase in the median for stock ownership over this period, ownership of thethirdincomequintile,41.2percent. pooled investment funds declined for almost every Thegreatmajorityoffamiliesownedstockinonly demographic group. Both the overall change and a small number of companies. In 2004, 34.6 percent the changes for demographic groups break an earlier had stock in only one company, 59.5 percent had trend toward broadly increased ownership of this stock in three or fewer companies, and 9.5 percent asset. Among families owning pooled investment had stock in fifteen or more companies (data not funds, the survey indicates that ownership shifted shown in the tables). For 37.1 percent of stock own- over the recent period from funds largely invested in ers, at least one of the companies was one that eitherstocksorgovernmentbondstowardfundsdediemployed or had employed the family head or that cated to a balance between stocks and bonds of any person’s spouse or partner. The 2001 data show a type. For 2004, the survey for the first time provides similarpattern. separate information on a miscellaneous category of funds,whichiscomposedofhedgefunds,exchange- Pooled Investment Funds 16. Pooled investment funds in this article are taken to exclude moneymarketmutualfundsandindirectlyheldmutualfundsandto includeallothertypesofdirectlyheldpooledinvestmentfunds,such From 2001 to 2004, direct ownership of pooled astraditionalopen-endandclosed-endmutualfunds,realestateinvestinvestment funds fell 2.7 percentage points, to menttrusts,andhedgefunds.

A16 Federal Reserve Bulletin 2006 traded funds, and similar instruments; the survey accounts accumulated from previous employment estimates that 4.3 percent of families with pooled to purchase an annuity at retirement; annuities are investment funds (0.7 percent of all families) had treated in this article as a separate type of managed fundsofthistype(datanotshowninthetables). asset. Among families owning pooled investment funds, From 2001 to 2004, the fraction of families with the value of holdings has continued an increase seen retirement accounts fell 2.5 percentage points; the over the preceding decade; in the recent period, the drop offset most of the 3.3 percentage point increase median holding rose 8.3 percent, and the mean rose of the preceding three years. In the recent period, 32.1 percent. Among the top quintile of the income more than 60 percent of families with some type of distribution, where ownership is most prevalent, the account plan had one associated with a current or median holding rose substantially over the recent past job, and nearly as many had an IRA or Keogh period; holdings fell for the other income groups. At account;aboutone-fourthoffamilieswithretirement thesametime,medianholdingsacrosswealthgroups accounts had both types (data not shown in the fellonlyforthelowestquartile.Byage,holdingsrose tables).Overthistime,ownershipdeclinedfornearly onlyforthe45–64agegroups.Medianholdingsrose all groups; key exceptions were families with a for white non-Hispanic families and fell for other retired head and families headed by persons aged families. 55 to 64 or aged 75 or older. In the preceding three years, ownership had been up in almost every demographicgroup. Retirement Accounts In a continuation of the trend over the preceding decade, holdings in retirement accounts increased Ownership of tax-deferred retirement assets such markedly in the 2001–04 period; for those having as individual retirement accounts (IRAs) tends to retirement accounts, the median rose 13.9 percent, increase with families’ income and net worth.17 For and the mean rose 11.0 percent. Gains also appeared severalreasons,ownershipisalsomorelikelyamong in the median holdings of most demographic groups families headed by persons less than 65 years of age over the recent period; one of the largest increases than among the older groups. First, even though was among nonwhite or Hispanic families, a group retirement accounts have been in existence for about for which ownership of such accounts declined subtwenty years, they may not have become common stantially in 2004. The 75-or-older age group saw a until relatively late in the careers of people in the sizabledeclineinitsmedian. older groups. Second, beginning in the year that a Althoughtax-deferredretirementassetsareclearly personreachesage591⁄ 2 ,fundsheldbythatpersonin animportantelementinretirementplanning,families retirement accounts may be withdrawn without pen- mayholdavarietyofotherassetsthatareintended,at alty,andsomeinthegroupmayhavedoneso.Third, least in part, to finance retirement. Such other assets families may have used funds from retirement might also be used for contingencies as necessary. Similarly,aneedforliquiditymightdriveafamilyto liquidate or borrow against a tax-deferred retirement 17. Tax-deferred retirement accounts consist of IRAs, Keogh asset, even if it will be assessed a penalty for doing accounts, and certain employer-sponsored accounts. Employersponsored accounts consist of 401(k), 403(b), and thrift saving so. accountsfromcurrentorpastjobs;othercurrentjobplansfromwhich Two common and often particularly important loansorwithdrawalscanbemade;andaccountsfrompastjobsfrom typesofretirementplanarenotincludedintheassets whichthefamilyexpectstoreceivetheaccountbalanceinthefuture. Thisdefinitionofemployer-sponsoredplansisintendedtoconfinethe described in this section: Social Security (the federanalysis to accounts that are portable across jobs and for which ally funded Old-Age and Survivors’ Insurance profamilieswillultimatelyhavetheoptiontowithdrawthebalance. gram, or OASI) and employer-sponsored defined- IRAsandKeoghsmaybeinvestedinvirtuallyanyasset,including stocks, bonds, pooled investment funds, options, and real estate. In benefit plans. OASI is well described elsewhere, and principle, employer-sponsored plans may be invested in a similarly it covers the great majority of the population.18 The broad way, but, in practice, individuals’ choices for investment are retirement income provided by defined-benefit plans often limited to a narrower set of assets. The 2004 SCF introduced a new sequence of questions to cover employer-sponsored pensions is typically based on workers’ salaries and years of associated with the current jobs of the survey respondent and the work with an employer, a group of employers, or a spouse or partner of that person. The goal of this redesign was to union.Unfortunately,incomestreamsfromOASIand better cope with the proliferation of complex plans and with the confusionmanypeopleappeartohaveabouttheexacttypesoftheir plans.Althoughthenewsequencewasdesignedtocontaintheearlier questions,itisstillpossiblethatthenewcontextmayhavechanged 18. ForadetaileddescriptionofOASI,seeSocialSecurityAdminpatternsofresponseforsometypesofrespondentinwaysnotcompat- istration, ‘‘Online Social Security Handbook,’’ Publication 65-008, iblewiththeearlierdata. www.ssa.gov/OP_Home/handbook/ssa-hbk.htm.

Recent Changes in U.S. Family Finances: Evidence from the 2001 and 2004 Survey of Consumer Finances A17 defined-benefit plans cannot be translated directly three years (data not shown in the tables). Participainto a current value because valuation depends tionbyeligibleworkersisusuallyvoluntary.In2004, critically on assumptions about future events and 84.1 percent of family heads who were eligible to conditions—workdecisions,earnings,inflationrates, participate elected to do so, down from 85.2 percent discount rates, mortality, and so on—and no widely in 2001.21 The choice to participate appears to be agreed-upon standards exist for making these related strongly to income. Of heads of families with assumptions.19 incomesinthelowest20percentofthedistributionin However, the SCF does contain substantial infor- 2004, 50.6 percent who were eligible declined to mationforfamilyheadsandtheirspousesorpartners participate;incontrast,amongheadsoffamilieswith regarding their defined-benefit plans and plans with incomes in the highest 10 percent of the distribution, some type of account feature to which they have only 5.0 percent of eligible workers declined to rightsfromacurrentorpastjob.20In2004,57.5per- participate. centoffamilieshadrightstosometypeofplanother than OASI through the current or past work of either the family head or that person’s spouse or partner, a Cash Value Life Insurance level nearly the same as in 2001. Of this group of families, 57.4 percent had a plan that was a standard Cash value life insurance combines an investment defined-benefit plan with an annuity payout scheme, vehicle with insurance coverage in the form of a 62.8 percent had a plan with at least some account death benefit.22 Some cash value life insurance polifeature,and20.1percenthadbothtypesofplan(data cies offer a high degree of choice in the way the notshowninthetables). policy payments are invested. Investment returns on In many pension plans with account features, con- such policies are typically shielded from taxation tributions may be made by the employer, the worker, until the money is withdrawn; if the funds remain or both. In some cases, these contributions represent untapped until the policyholder dies, the beneficiary a substantial amount of saving, though workers may of the policy may receive, tax-free, the death benefit offset this saving by reducing their saving in other or the cash value, whichever is greater. In contrast, forms. An employer’s contributions also represent term insurance, the other popular type of life insuradditional income for the worker. In 2004, 88.5 per- ance, offers only a death benefit. One attraction of cent of families with account-type plans on a current cash value policies for some people is that they job of either the family head or the spouse or partner promote regular saving funded through the required of the family head had employers who made contri- policypremium. butionstotheplan,and89.4percentoffamilieswith Ownership of cash value insurance is broadly such plans made contributions themselves (data not spread across demographic groups, with a tendency showninthetables).Themedianannualcontribution toward increasing rates among families with higher by employers who contributed to such accounts was levels of income and wealth and those with older $2,400, and the median contribution of families that family heads. Ownership of cash value policies over contributedwas$2,700. the 2001–04 period continued a declining trend, The eligibility of working heads of families to decreasing 3.8 percentage points, to 24.2 percent of participate in any type of job-related pension fell families. The decline was shared by nearly all the from 57.2 percent in 2001 to 54.8 percent in 2004; it demographic groups. Over this time, the ownership had risen 2.0 percentage points over the preceding of either cash value or term life insurance also fell— from 69.3 percent to 65.4 percent of families (data 19. For one possible calculation of net worth that includes the annuityvalueofdefined-benefitpensionbenefitsandOASIpayments, 21. AnanalysisoftheMarchCPSwithadefinitionoffamilyhead seeArthurB.KennickellandAnnikaE.Sundén(2005),‘‘Pensions, thatisclosesttothatinthisarticleshowsasimilartrendinpension Social Security, and the Distribution of Wealth,’’ Finance and Eco- eligibilityforemployedfamilyheads,butthattrendisatasomewhat nomicsDiscussionSeries1997-55(Washington:BoardofGovernors higherlevelthanintheSCF.TheCPSeligibilityestimateforfamily of the Federal Reserve System, October), www.federalreserve.gov/ heads with a job in the past year was 61.9 percent in 2001 and pubs/feds/1997/index.html. 57.8 percent in 2004. Differences in the definition of the relevant 20. The definition of account plan here differs slightly from that employmentmayexplainsomeofthedifferenceinthelevelsinthe usedincomputingthesurveywealthmeasure,whichincludesaccount twosurveys.UnliketheSCF,theCPSshowsasmallincreaseinthe balancesonlyifthefamilyhastheabilitytomakewithdrawalsfrom, uptakerateforsucheligibleworkers—from82.9percentin2001to or borrow against, the account. Here the only criterion used in 83.4percentin2004. classificationiswhetherthereisanyaccountbalance.Forexample,a 22. The survey measures the value of such policies according to defined-benefitplanwithaportablecashoption,whichwouldallow their current cash value, not their death benefit. The cash value is thecoveredworkertoreceivealumpsuminlieuofregularpayments includedasanassetinthisarticleonlywhentherewasanonzerocash inretirement,wouldbetreatedasanaccountplanhere. valueatthetimeoftheinterview.

A18 Federal Reserve Bulletin 2006 notshowninthetables).Ofthosefamilieswithsome Between2001and2004,themedianvalueofother type of life insurance, the proportion with term poli- managed assets fell 39.7 percent, and the mean fell cies rose, while the proportion with cash value poli- 35.6percent.Duringtheprecedingthree-yearperiod, cies fell; these changes follow earlier trends in the the median had more than doubled. Over the recent survey. period, median holdings declined for almost all After rising fairly strongly over the period from demographic groups. The declines reflect substantial 1992 to 2001, the median value of cash value insur- reductionsbothinannuitiesandintrustsormanaged ance for families that had any fell 43.9 percent investment accounts. For families with an equity between 2001 and 2004, and the mean fell 39.8 per- interest in an annuity, the median holding fell cent. The median showed sizable declines in every 30.6percent,to$37,000,in2004;forfamilieswitha demographicgroupshown.Percentagedeclineswere trust or managed investment account as defined in mostnotableamongfamiliesinthebottomquartileof this article, the median holding fell 37.4 percent, to the wealth distribution, among younger families, and $100,000(datanotshowninthetables). amongrenters. As noted in the discussion of retirement accounts, some families use settlements from retirement accounts to purchase an annuity. In 2004, 26.7 per- Other Managed Assets cent of families with annuities had done so (data not shown in the tables). Of these families, 91.6 percent Ownership of other managed assets—personal annu- hadanequityinterestintheirannuities. ities and trusts with an equity interest and managed investment accounts—is concentrated among families with higher levels of income and wealth and Other Financial Assets among families headed by persons who are aged 55 or older or who are retired.23 Ownership of these For other financial assets—a heterogeneous category assets rose 0.7 percentage point between 2001 and including oil and gas leases, futures contracts, royal- 2004 after a similarly small increase over the previ- ties, proceeds from lawsuits or estates in settlement, ousthreeyears.Acrossdemographicgroups,changes and loans made to others—ownership rose 0.6 perin ownership were mixed; ownership increased centage point between 2001 and 2004, to 10.0 permost—5.5 percent—for the oldest age group, and cent. Ownership of such assets tends to be more it decreased most—5.4 percent—for the highest common among higher income and wealth groups, wealth group. Of families having such accounts in youngeragegroups,andfamiliesheadedbyaperson 2004, 26.3 percent had only a trust or managed who is self-employed. Changes in ownership across investment account, 68.9 percent had only an annu- demographic groups were generally positive, while ity, and 4.9 percent had both (data not shown in the themedianholdingforthosewhohadsuchassetsfell tables). 7.0percent,to$4,000. Some publicly traded companies offer stock options to their employees as a form of compensa- 23. The survey encourages respondents who have trusts or mantion.24 Although stock options, when executed, may agedinvestmentaccountsthatareheldinrelativelycommoninvestments to report the components. Of the 4.2 percent of families that representanappreciablepartofafamily’snetworth, reportedhavinganykindoftrustormanagedinvestmentaccountin the survey does not specifically ask for the value of 2004, 45.1 percent of them reported at least one of the component these options.25 Instead, the survey asks whether the assets separately. Of families that detailed the components in 2004, 87.2 percent reported some type of financial asset, 11.0 percent family head or that person’s spouse or partner had reportedaprimaryresidence,13.4percentreportedotherrealestate, been given stock options by an employer during the 3.6percentreportedabusiness,and2.7percentreportedanothertype preceding year. In 2004, 9.3 percent of families of asset (data not shown in the tables). Comparable figures are not availablefor2001. reported having received stock options, a share Inthisarticle,thetrustormanagedinvestmentaccountsincludedin 2.1 percentage points below the level in 2001 (data other managed assets are those in which families have an equity notshowninthetables). interestandforwhichcomponentpartswerenotseparatelyreported. Typically,suchaccountsarethoseinwhichtheownershipiscomplicated or the management is undertaken by a professional. In 2004, 24. SeeDavidLebow,LouiseSheiner,LarrySlifman,andMartha 79.0percentoffamilieswithtrustsormanagedinvestmentaccounts Starr-McCluer (1999), ‘‘Recent Trends in Compensation Practices,’’ had an equity interest in those accounts. Annuities may be those in Finance and Economics Discussion Series 1999-32 (Washington: whichthefamilyhasanequityinterestintheassetorinwhichthereis BoardofGovernorsoftheFederalReserveSystem,July). anentitlementonlytoastreamofincome.Thewealthfiguresinthis 25. Becausesuchoptionsaretypicallynotpubliclytradedortheir articleincludeonlytheannuitiesinwhichthereisanequityinterest. execution is otherwise constrained, their value is uncertain until the In2004,7.2percentoffamiliesreportedhavinganytypeofannuity, exercisedate;untilthen,meaningfulvaluationwouldrequirecomplex andofthesefamilies,81.8percentreportedhavinganequityinterest. assumptionsaboutthefuturebehaviorofstockprices.

Recent Changes in U.S. Family Finances: Evidence from the 2001 and 2004 Survey of Consumer Finances A19 6. Directandindirectfamilyholdingsofstock,byselectedcharacteristicsoffamilies,1995–2004surveys Percentexceptasnoted Medianvalueamongfamilies Totalstockholdingsasashare Familieswithholdings withholdings Family (thousandsof2004dollars) oftotalfinancialassets characteristic 1995 1998 2001 2004 1995 1998 2001 2004 1995 1998 2001 2004 Allfamilies............. 40.4 48.9 51.9 48.6 18.0 29.0 36.7 24.3 40.1 54.0 56.0 47.4 Percentileofincome Lessthan20 ............ 6.5 10.0 12.4 11.7 4.6 5.8 7.4 7.0 14.2 20.4 36.9 31.3 20–39.9................. 24.7 30.8 33.5 28.8 7.8 11.6 8.0 8.8 26.7 29.8 34.9 29.6 40–59.9................. 41.5 50.2 52.1 49.2 7.7 13.9 16.0 11.6 28.5 38.1 46.5 41.0 60–79.9................. 54.3 69.3 75.7 66.5 15.6 22.0 30.4 20.0 35.6 45.8 51.7 37.5 80–89.9................. 69.7 77.9 82.0 82.5 30.8 52.2 68.8 34.6 41.3 50.4 57.4 43.2 90–100 ................. 80.0 90.4 89.6 91.0 73.9 156.5 263.8 169.9 45.7 62.5 60.5 53.6 Ageofhead(years) Lessthan35 ............ 36.6 40.8 48.9 38.8 6.3 8.1 7.5 5.2 27.2 44.9 52.5 30.0 35–44 .................. 46.4 56.7 59.5 52.3 12.3 23.2 29.3 12.7 39.5 55.0 57.3 47.7 45–54 .................. 48.9 58.6 59.2 54.4 31.9 44.1 53.3 30.6 43.1 55.7 59.1 46.8 55–64 .................. 40.0 55.9 57.1 61.6 38.2 54.5 86.5 59.5 44.5 58.4 56.2 51.1 65–74 .................. 34.4 42.7 39.2 45.8 41.9 64.9 159.8 75.0 35.8 51.3 55.2 51.1 75ormore.............. 27.9 29.4 34.2 34.8 24.6 69.6 127.8 85.9 39.8 48.7 51.4 39.1 Housingstatus Owner .................. 48.8 59.8 62.0 59.1 22.2 39.4 53.3 34.4 41.1 55.1 56.7 48.0 Renterorother.......... 25.0 27.5 30.7 25.1 7.9 8.7 7.5 6.7 32.4 40.5 46.2 35.5 Note: Indirect holdings are those in retirement accounts and other managedassets.Seealsonotetotable1. Direct and Indirect Holdings of Publicly Traded stocks,29.4percentthroughdirectholdingsofpooled Stocks investmentfunds,and9.7percentthroughamanaged investmentaccountoranequityinterestinanannuity Families may hold stocks in publicly traded compa- or trust (data not shown in the tables); 44.0 percent niesdirectlyorindirectly,andinformationabouteach had ownership through more than one such means. of these forms of ownership is collected separately Regarding the distribution of the amount of directly in the SCF. When direct and indirect forms are and indirectly held equities, 30.8 percent was held combined, the 2004 data show a break in a trend of in tax-deferred retirement accounts, 37.1 percent as increasingstockownershipdatingtobeforethe1995 directly held stocks, 24.1 percent as directly held survey(table6).Between2001and2004,thefraction pooled investment funds, and 8.0 percent as other offamiliesholdinganysuchstockfell3.3percentage managedassets. points,to48.6percent,alevelapparentlylastreached sometimebetweenthe1995and1998surveys.Much like ownership of directly held stock, ownership of Nonfinancial Assets direct and indirect holdings is more common among higher-income groups and among families headed By definition, a rise in nonfinancial assets as a share by persons aged 35 to 64. Over the recent three-year of total assets must exactly offset the 6.3 percentage period, ownership declined for all income groups except the top two deciles and for the age groups 55orolder. 7. Valueofnonfinancialassetsofallfamilies,distributed bytypeofasset,1995–2004surveys At the same time, the overall median value of Percent direct and indirect stock holdings dropped 33.8 percent. The decline was shared by all the demographic Typeofasset 1995 1998 2001 2004 groupsshownexceptforfamiliesinthesecondquin- Vehicles ...................... 7.1 6.5 5.9 5.1 tile of the income distribution, a group with a rate Primaryresidence ............. 47.5 47.0 46.9 50.3 of ownership that is much below the average. As a Otherresidentialproperty ...... 8.0 8.5 8.1 9.9 Equityinnonresidential proportion of financial assets, holdings declined property.................. 7.9 7.7 8.2 7.3 Businessequity ............... 27.2 28.5 29.3 25.9 8.6 percentage points overall and also fell substan- Other ......................... 2.3 1.7 1.6 1.5 Total ..................... 100 100 100 100 tiallyforeverydemographicgroupshown. Amongfamiliesthatheldstocksin2004,78.2per- Memo Nonfinancialassetsasa cent held them through a tax-deferred retirement shareoftotalassets ....... 63.3 59.3 58.0 64.3 account, 42.5 percent through direct holdings of Note: Seenotetotable1andtextnote26.

A20 Federal Reserve Bulletin 2006 8. Familyholdingsofnonfinancialassetsandofanyasset,byselectedcharacteristicsoffamiliesandtypeofasset,2001and 2004surveys A. 2001SurveyofConsumerFinances Other Equityin Any Family Primary Business Any Vehicles residential nonresidential Other nonfinancial characteristic residence equity asset property property asset Percentageoffamiliesholdingasset Allfamilies..................... 84.8 67.7 11.3 8.2 11.9 7.5 90.7 96.7 Percentileofincome Lessthan20 .................... 56.8 40.6 3.1 2.8 2.5 2.9 67.7 85.6 20–39.9......................... 86.7 57.3 5.4 6.7 7.1 5.8 93.1 98.3 40–59.9......................... 91.6 66.0 7.9 6.7 8.8 6.2 95.6 99.8 60–79.9......................... 94.8 81.8 14.2 7.0 12.0 8.7 97.8 100.0 80–89.9......................... 95.4 90.9 19.7 12.1 18.7 9.4 99.4 100.0 90–100 ......................... 92.8 94.4 32.8 23.9 39.0 17.9 99.5 100.0 Ageofhead(years) Lessthan35 .................... 78.8 39.9 3.4 2.8 7.0 6.8 83.0 93.2 35–44 .......................... 88.9 67.8 9.2 7.4 14.2 7.8 93.2 97.4 45–54 .......................... 90.5 76.2 14.7 10.0 17.1 7.2 95.2 98.1 55–64 .......................... 90.7 83.2 18.3 12.3 15.6 7.9 95.4 98.4 65–74 .......................... 81.3 82.5 13.7 12.9 11.7 9.7 91.6 97.1 75ormore...................... 73.9 76.2 15.2 8.3 2.4 5.8 86.4 97.8 Raceorethnicityofrespondent Whitenon-Hispanic ............. 89.2 74.3 13.0 9.6 13.9 8.9 94.7 99.0 NonwhiteorHispanic ........... 71.4 47.3 6.3 3.9 5.5 3.1 78.4 89.8 Currentworkstatusofhead Workingforsomeoneelse ....... 88.5 64.7 10.0 6.7 6.1 7.3 92.5 97.8 Self-employed .................. 88.6 80.3 19.5 17.9 60.8 14.0 97.1 98.6 Retired ......................... 77.1 73.8 12.0 8.2 3.3 5.3 86.7 95.8 Othernotworking............... 63.8 43.6 4.9 3.8 5.8 * 70.3 82.2 Housingstatus Owner .......................... 92.2 100.0 14.9 10.9 15.5 8.7 100.0 100.0 Renterorother.................. 69.3 ... 3.9 2.5 4.2 4.9 71.3 89.9 Percentileofnetworth Lessthan25 .................... 64.8 14.3 * * 1.2 3.0 68.2 87.0 25–49.9......................... 86.8 69.6 4.5 3.6 4.0 5.0 96.3 100.0 50–74.9......................... 94.1 91.4 12.7 8.0 11.5 6.6 98.7 100.0 75–89.9......................... 93.1 95.1 19.5 15.3 22.4 10.2 99.6 100.0 90–100 ......................... 94.1 95.8 39.0 30.0 42.8 22.7 99.7 100.0 point drop in the share of financial assets from 2001 shown,the2004ratewasabout90percentormore— to 2004 discussed earlier in this article (table 4). The exceptions were the lowest income and wealth changes in these shares may have been driven by groups, nonwhite or Hispanic families, families changes in portfolio choices, portfolio valuation, or headed by persons who were neither working nor both. Over the six most recent surveys, the 2001 retired,andrenters.Overthe2001–04period,ownerestimate of the value of nonfinancial assets as a ship rose most for the lowest income and wealth share of total assets, 58.0 percent, appears to be the groups, the youngest and the two oldest age groups, low point; the 2004 level, 64.3 percent, is about the nonwhite or Hispanic families, renters, and families same as the level seen in the 1995 survey (table 7). headed by persons who were neither working nor Over the recent three-year period, the value of pri- retired. The only substantial declines in ownership mary residences as a share of nonfinancial assets were seen by the 55–64 age group and the second increased 3.4 percentage points, to 50.3 percent, the quintileoftheincomedistribution. largest share ever recorded in the survey. The share Over the recent period, the median holding of of other residential property also rose. The largest nonfinancial assets for families having any such offsettingdeclinewasintheshareofbusinessequity, assets rose 22.2 percent, and the mean rose 19.5 perwhich fell 3.4 percentage points. Smaller declines cent. Across demographic groups, substantial gains wereseeninthesharesoftheremainingnonfinancial far outnumbered declines in the median. Over this assets. time, the median fell only for some groups that saw In 2004, the level of ownership of nonfinancial gains in ownership; this result suggests that the fall assets was 92.5 percent, 1.8 percentage points higher in the median may have been driven, at least in part, thanin2001(firsthalfoftables8.Aand8.B,next-to- by the influx of new owners with relatively small lastcolumn).Acrossmostofthedemographicgroups holdings.

Recent Changes in U.S. Family Finances: Evidence from the 2001 and 2004 Survey of Consumer Finances A21 8.—Continued A. 2001SurveyofConsumerFinances—Continued Other Equityin Any Family Primary Business Any Vehicles residential nonresidential Other nonfinancial characteristic residence equity asset property property asset Medianvalueofholdingsforfamiliesholdingasset(thousandsof2004dollars) Allfamilies..................... 14.4 131.0 85.2 52.7 106.5 12.8 120.9 156.8 Percentileofincome Lessthan20 .................... 5.7 69.2 26.6 34.6 60.0 6.4 36.5 24.4 20–39.9......................... 8.9 85.2 79.9 32.0 37.3 6.4 60.7 71.5 40–59.9......................... 13.4 101.2 53.3 32.0 65.7 10.7 98.2 122.5 60–79.9......................... 18.7 138.5 74.6 53.3 66.6 10.7 161.5 245.0 80–89.9......................... 24.2 186.4 66.6 49.0 106.5 21.3 239.2 401.6 90–100 ......................... 31.9 319.5 213.0 155.8 285.7 53.3 510.8 1,075.1 Ageofhead(years) Lessthan35 .................... 12.1 101.2 79.9 35.5 53.3 10.7 31.7 41.4 35–44 .......................... 15.8 133.1 79.9 42.1 106.5 9.6 125.5 167.9 45–54 .......................... 16.7 143.8 69.2 60.3 108.6 11.7 150.8 225.7 55–64 .......................... 16.1 138.5 85.2 83.6 106.5 32.0 157.5 241.1 65–74 .......................... 14.5 137.4 154.4 53.3 106.5 21.3 158.9 228.6 75ormore...................... 9.4 118.2 85.2 29.8 544.2 12.8 130.6 180.6 Raceorethnicityofrespondent Whitenon-Hispanic ............. 15.6 138.5 85.2 53.3 106.5 16.0 141.4 197.7 NonwhiteorHispanic ........... 10.6 99.1 63.9 32.0 53.3 4.8 62.8 61.3 Currentworkstatusofhead Workingforsomeoneelse ....... 14.6 127.8 74.6 42.1 53.3 10.7 108.6 137.4 Self-employed .................. 20.5 213.0 159.8 109.1 140.9 32.0 356.8 467.8 Retired ......................... 10.7 106.5 90.5 61.8 69.8 21.3 111.9 148.5 Othernotworking............... 10.9 106.5 117.2 35.1 117.2 * 80.6 45.9 Housingstatus Owner .......................... 17.2 131.0 85.2 53.3 114.7 16.0 167.2 255.8 Renterorother.................. 8.1 ... 63.9 34.6 37.3 6.4 9.4 14.2 Percentileofnetworth Lessthan25 .................... 6.7 52.7 * * 10.7 4.3 8.8 8.7 25–49.9......................... 12.5 74.6 25.6 9.6 16.0 10.7 66.7 79.9 50–74.9......................... 16.2 127.8 53.3 26.6 53.3 10.7 154.3 229.7 75–89.9......................... 20.2 213.0 85.2 55.7 127.8 19.2 300.1 541.6 90–100 ......................... 30.7 372.8 223.7 225.5 532.6 42.6 758.9 1,531.7 Memo Meanvalueofholdingsfor familiesholdingasset ....... 19.5 192.6 198.4 277.2 687.5 60.2 306.6 495.6 Note: Seenotestotable7. *Tenorfewerobservations. ... Notapplicable. Vehicles 2004,whilethemeanrose3.1percent.27Themedian value of vehicle holdings fell notably for the lowest Vehicles continue to be the most commonly held twoincomeandwealthgroups,thetwooldestandthe nonfinancial asset.26 Over the recent three-year youngest age groups, nonwhite or Hispanic families, period,theshareoffamiliesthatownedsometypeof renters, and families having a head who was retired; vehicle rose 1.5 percentage points, to 86.3 percent. for most other families, the median rose. Continuing Ownership rose for most demographic groups but atrend,theshareofthetotalvalueofownedvehicles particularly for families in the lowest income and attributable to sport-utility vehicles rose over the wealthgroups,familiesheadedbypersonsaged65to recent period from 14.0 percent to 19.1 percent (data 74,andnonwhiteorHispanicfamilies. notshowninthetables). Themedianmarketvalueofvehiclesforthosewho Somefamilieshavevehiclesthattheyleaseorthat ownedatleastonedeclined1.4percentfrom2001to are provided to them by an employer for personal 26. Thedefinitionofvehicleshereisabroadonethatincludescars, 27. Survey respondents are asked to provide the year, make, and vans, sport-utility vehicles, trucks, motor homes, recreational vehi- model of each of their cars, vans, sport-utility vehicles, and trucks. cles,motorcycles,boats,airplanes,andhelicopters.Offamiliesown- Thisinformationisusedtoobtainmarketpricesfromdatacollected ing any type of vehicle in 2004, 99.8 percent had a car, van, sport- bytheNationalAutomobileDealersAssociationandavarietyofother utilityvehicle,motorcycle,ortruck.Theremainingtypesofvehicle sources.Forothertypesofvehicle,therespondentisaskedtoprovide wereheldby13.3percentoffamilies. abestestimateofthecurrentvalue.

A22 Federal Reserve Bulletin 2006 8. Familyholdingsofnonfinancialassetsandofanyasset,byselectedcharacteristicsoffamiliesandtypeofasset, 2001and2004surveys—Continued B. 2004SurveyofConsumerFinances Other Equityin Any Family Primary Business Any Vehicles residential nonresidential Other nonfinancial characteristic residence equity asset property property asset Percentageoffamiliesholdingasset Allfamilies..................... 86.3 69.1 12.5 8.3 11.5 7.8 92.5 97.9 Percentileofincome Lessthan20 .................... 65.0 40.3 3.6 2.7 3.7 3.9 76.4 92.2 20–39.9......................... 85.3 57.0 6.9 3.8 6.7 4.4 92.0 97.8 40–59.9......................... 91.6 71.5 10.0 7.6 9.5 7.5 96.7 99.8 60–79.9......................... 95.3 83.1 14.0 10.6 12.0 10.4 98.4 100.0 80–89.9......................... 95.9 91.8 19.3 12.8 16.0 8.3 99.1 99.8 90–100 ......................... 93.1 94.7 37.2 20.8 34.7 16.7 99.3 100.0 Ageofhead(years) Lessthan35 .................... 82.9 41.6 5.1 3.3 6.9 5.5 88.6 96.5 35–44 .......................... 89.4 68.3 9.4 6.4 13.9 6.0 93.0 97.7 45–54 .......................... 88.8 77.3 16.3 11.4 15.7 9.7 94.7 98.3 55–64 .......................... 88.6 79.1 19.5 12.8 15.8 9.2 92.6 97.5 65–74 .......................... 89.1 81.3 19.9 10.6 8.0 9.0 95.6 99.5 75ormore...................... 76.9 85.2 9.7 7.7 5.3 8.5 92.5 99.6 Raceorethnicityofrespondent Whitenon-Hispanic ............. 90.3 76.1 14.0 9.2 13.6 9.3 95.8 99.3 NonwhiteorHispanic ........... 76.1 50.8 8.9 5.8 5.9 3.8 84.0 94.4 Currentworkstatusofhead Workingforsomeoneelse ....... 89.7 66.5 10.4 6.8 5.8 7.1 93.8 98.4 Self-employed .................. 91.2 79.1 25.8 18.7 58.1 12.9 97.5 99.1 Retired ......................... 79.0 75.8 12.8 7.9 3.5 7.1 89.8 97.7 Othernotworking............... 66.9 40.0 5.4 * 6.9 6.4 76.3 89.6 Housingstatus Owner .......................... 92.3 100.0 15.7 11.0 14.7 9.2 100.0 100.0 Renterorother.................. 73.0 ... 5.4 2.4 4.3 4.6 75.9 93.3 Percentileofnetworth Lessthan25 .................... 69.8 15.2 * * * 2.9 73.7 91.7 25–49.9......................... 89.2 71.2 4.9 4.1 5.6 5.4 97.5 100.0 50–74.9......................... 92.0 93.4 12.7 8.3 11.2 7.8 99.0 100.0 75–89.9......................... 95.2 96.2 23.1 15.1 19.9 12.3 99.8 100.0 90–100 ......................... 93.1 96.9 45.6 28.8 40.8 18.8 99.9 100.0 use. The share of families having a vehicle from any rate less than the overall rate included nonwhite or source rose 1.3 percentage points over the recent Hispanic families, families whose head was neither period,to89.2percent(datanotshowninthetables). working nor retired, families with relatively low Thesmalldifferencebetweenthisrateandtheowner- income or wealth, and families headed by persons shiprateforpersonallyownedvehiclesbeliesalarger agedlessthan35.Overthethree-yearperiod,ownerchange in the rates of holding for leased and ship rose most for families in the middle of the employer-provided vehicles. The proportion of fami- income and wealth distributions, for families headed lies with a leased vehicle fell from 5.8 percent to by persons aged 75 or older, and for nonwhite or 4.0 percent, while that with an employer-provided Hispanic families; the rate fell notably for the 55–64 vehiclefellfrom9.1percentto7.7percent. age group and for the self-employed and the othernot-working work-status groups. Despite the aboveaverage rise in ownership for nonwhite or Hispanic Primary Residence and Other Residential Real families, their ownership rate remained well below Estate thatforotherfamilies. As would be expected from the large increase in The homeownership rate over the 2001−04 period boththeshareoftotalassetsattributabletononfinancontinued its upward trend, rising 1.4 percentage cial assets and the share of nonfinancial assets attribpoints, to 69.1 percent.28 In 2004, groups that had a utable to primary residences, the median and mean valuesoftheprimaryresidencesofhomeownersrose 28. This measure of primary residences comprises mobile homes sharply over the recent period; overall, the median andtheirsites,thepartsoffarmsandranchesnotusedforafarmingor rose 22.1 percent, and the mean rose 28.1 percent. ranching business, condominiums, cooperatives, townhouses, other single-familyhomes,andotherpermanentdwellings. Because housing wealth is typically the largest com-

Recent Changes in U.S. Family Finances: Evidence from the 2001 and 2004 Survey of Consumer Finances A23 8.—Continued B. 2004SurveyofConsumerFinances—Continued Other Equityin Any Family Primary Business Any Vehicles residential nonresidential Other nonfinancial characteristic residence equity asset property property asset Medianvalueofholdingsforfamiliesholdingasset(thousandsof2004dollars) Allfamilies..................... 14.2 160.0 100.0 60.0 100.0 15.0 147.8 172.9 Percentileofincome Lessthan20 .................... 4.5 70.0 33.0 11.0 30.0 4.5 22.4 17.0 20–39.9......................... 7.9 100.0 65.0 30.0 30.0 7.5 71.1 78.3 40–59.9......................... 13.1 135.0 55.0 36.0 62.5 10.0 131.2 154.4 60–79.9......................... 19.8 175.0 100.0 47.0 150.0 10.0 197.2 289.4 80–89.9......................... 25.8 225.0 98.0 60.0 100.0 17.5 281.8 458.5 90–100 ......................... 33.0 450.0 268.3 189.0 350.0 50.0 651.2 1,157.7 Ageofhead(years) Lessthan35 .................... 11.3 135.0 82.5 55.0 50.0 5.0 32.3 39.2 35–44 .......................... 15.6 160.0 80.0 42.2 100.0 10.0 151.3 173.4 45–54 .......................... 18.8 170.0 90.0 43.0 144.0 20.0 184.5 234.9 55–64 .......................... 18.6 200.0 135.0 75.0 190.9 25.0 226.3 351.2 65–74 .......................... 12.4 150.0 80.0 78.0 100.0 30.0 161.1 233.2 75ormore...................... 8.4 125.0 150.0 85.8 80.3 11.0 137.1 185.2 Raceorethnicityofrespondent Whitenon-Hispanic ............. 15.7 165.0 105.0 66.0 135.0 16.5 164.8 224.5 NonwhiteorHispanic ........... 9.8 130.0 80.0 30.0 66.7 10.0 64.1 59.6 Currentworkstatusofhead Workingforsomeoneelse ....... 14.9 160.0 88.0 40.0 50.0 10.0 141.9 161.2 Self-employed .................. 21.9 248.0 141.5 125.0 174.0 30.0 335.4 468.3 Retired ......................... 10.1 130.0 100.0 60.0 120.0 25.0 131.7 165.6 Othernotworking............... 10.7 130.0 86.0 * 25.0 20.0 60.0 30.3 Housingstatus Owner .......................... 17.5 160.0 100.0 62.0 122.8 17.5 201.6 289.9 Renterorother.................. 7.2 ... 80.0 56.0 50.0 8.0 8.4 12.2 Percentileofnetworth Lessthan25 .................... 5.6 65.0 * * * 3.0 7.4 7.7 25–49.9......................... 11.9 85.0 25.6 14.9 17.5 6.0 72.4 84.5 50–74.9......................... 17.4 159.3 65.0 25.0 55.0 10.0 188.1 257.3 75–89.9......................... 22.6 250.0 100.0 73.9 150.0 25.0 360.8 600.2 90–100 ......................... 30.6 450.0 325.0 250.0 527.4 80.0 907.7 1,572.6 Memo Meanvalueofholdingsfor familiesholdingasset ....... 20.1 246.8 267.3 298.1 765.5 66.6 366.3 538.4 Note: Seenotestotable7. *Tenorfewerobservations. ... Notapplicable. ponentoffamilies’fungiblewealth,thelargepercent- withprimaryresidences,themedianandmeanvalues age gains in the median and mean produced large for owners increased sharply over the recent period; dollargains:$29,000forthemedianand$54,200for the median rose 17.4 percent and the mean 34.7 perthe mean. Homeowners in all demographic groups cent. Most of the demographic groups saw substansaw gains in the median, most of them substantial. tialgainsinthemedian;onlyafewsawdeclines,but Oneofthelargestincreaseswasthe31.2percentrise wheretheyoccurredtheytendedtobesubstantial. in the median value of primary residences for nonwhite or Hispanic families; in contrast, the median forotherfamiliesrose19.1percent. Net Equity in Nonresidential Real Estate In2004,12.5percentoffamiliesownedsomeform of residential real estate besides a primary residence Theownershipofnonresidentialrealestatewasabout (second homes, time shares, one- to four-family unchangedat8.3percentoffamiliesin2004.29Ownrental properties, and other types of residential propershipfollowsapproximatelythesamerelativedistrierty),alevelup1.2percentagepointsfromthefigure bution over demographic groups as does the ownerin 2001 but approximately the same as the 1998 estimate. Ownership is much more common among thehighestincomeandwealthgroups,amongtheage 29. Nonresidential real estate comprises the following types of propertyunlessitisownedthroughabusiness:commercialproperty, groups between 45 and 74, and among families rentalpropertywithfiveormoreunits,farmandranchland,undevelheaded by self-employed persons. As was the case opedland,andallothertypesofnonresidentialrealestate.

A24 Federal Reserve Bulletin 2006 ship of other residential real estate. Changes in The SCF classifies privately owned business interownership during the recent period were mixed ests into those in which the family has an active across demographic groups. Among the income management role and those in which it does not. Of groups with substantial ownership in 2001, the key families having any business interests in 2004, changes were a decline in ownership among the 92.8percenthadanactiveroleand12.3percenthada highestdecileandanincreaseamongthefourthquin- passive role; 5.1 percent had interests in which they tile. Overall, the median value of such property for had each type of role (data not shown in the tables). ownersrose13.9percent,andthemeanrose7.5per- In terms of assets, the actively managed interests cent. Among income groups, the largest gains in the accounted for 89.1 percent of total privately owned median were in the top two deciles, which also had business interests. The median number of actively thehighestratesofownership;declinesinthemedian managed businesses was 1. The businesses reported appearedforallotherincomegroupsexceptthethird inthesurveywereamixtureofverysmallbusinesses quintile. withmoderatevaluesandsubstantiallymorevaluable businesses. Families with more than one business are asked to Net Equity in Privately Held Businesses report which business is most important; that business is designated as the primary one.32 The vast The share of families that owned a privately held majority of primary businesses operated in an indusbusiness interest edged down 0.4 percentage point tryotherthanmanufacturing;themostcommonorgaduring the recent period, to 11.5 percent.30 The pro- nizational form of those businesses was sole propriportion has changed little over the past several sur- etorship,andthemediannumberofemployeeswas2. veys.Ownershipofthistypeofassettendstoincrease However, primary actively managed businesses with with income and wealth and to be the highest for morethantwoemployeesaccountedfor83.7percent families headed by persons aged between 45 and 64; of the value of all such businesses, and the largest over the recent three-year period, declines in owner- share of value (40.6 percent) was attributable to ship were largely concentrated in the highest income businessesorganizedassubchapterScorporations. and wealth groups. Continuing a pattern seen in the preceding three years, ownership also declined Other Nonfinancial Assets among families with a head who was selfemployed.31 The median holding of business equity for those Ownership of the remaining nonfinancial assets (tanhaving any declined 6.1 percent, while the mean gibleitemsincludingartwork,jewelry,preciousmetincreased 11.3 percent. These changes follow a jump als, antiques, hobby equipment, and collectibles) of53.0percentinthemedianand21.8percentinthe increased marginally in the recent period, to 7.8 permean between the 1998 and 2001 surveys. Across cent.Amongwealthgroups,thenotablechangewasa income groups over the recent three-year period, decline of 3.9 percentage points in ownership in the gains in the median were seen in the top decile and highest wealth group; this change entirely offset a the fourth quintile. Growth rates in median holdings gain for the group over the previous three years. For were similar across racial or ethnic groups; however, families having such assets, the median value themedianlevelfornonwhitesorHispanicsremained rose 17.2 percent over the recent period, and the roughly half that of other families with business mean rose 10.6 percent. Across wealth groups, assets. median holdings rose substantially in the top two wealthgroupsanddeclinedamongtherest. 30. Theformsofbusinessinthiscategoryaresoleproprietorships, limitedpartnerships,othertypesofpartnership,subchapterScorpora- Unrealized Capital Gains tions and other types of corporation that are not publicly traded, limitedliabilitycompanies,andothertypesofprivatebusiness.Ifthe familysurveyedlivedonafarmorranchthatwasusedatleastinpart Changes in the values of assets such as stock, real foragriculturalbusiness,thevalueofthatpartnetofthecorrespondestate, and businesses are a key determinant of ingshareofassociateddebtsisincludedwithotherbusinessassets. 31. Inthesurvey,self-employmentstatusandbusinessownership changes in families’ net worth. Unrealized gains are are independently determined. Among the 11.5 percent of families withabusinessin2004,69.9percenthadafamilyheadorthespouse orpartneroftheheadwhowasself-employed;amongthe15.0percent 32. Forfamilieswithonlyonebusiness,thatbusinessis,bydefault, of families in which either the head or the spouse or partner of the considered the primary one. In 2004, the primary actively managed head was self-employed, 53.5 percent owned a business (data not business accounted for 78.7 percent of the value of all actively showninthetables). managedbusinesses.

Recent Changes in U.S. Family Finances: Evidence from the 2001 and 2004 Survey of Consumer Finances A25 9. Familyholdingsofunrealizedcapitalgains,byselectedcharacteristicsoffamilies,1995–2004surveys Thousandsof2004dollars 1995 1998 2001 2004 Family characteristic Median Mean Median Mean Median Mean Median Mean Allfamilies................. 6.8 83.3 12.5 111.9 16.0 138.2 23.0 161.7 Percentileofincome Lessthan20 ................ † 19.1 † 20.9 † 18.6 † 31.1 20–39.9..................... .4 31.2 2.1 34.0 1.5 44.0 3.0 52.1 40–59.9..................... 4.6 40.4 10.4 51.1 10.1 49.1 21.0 74.3 60–79.9..................... 16.4 56.9 23.4 76.3 29.8 91.8 46.7 120.3 80–89.9..................... 33.2 85.5 39.6 110.5 58.6 151.0 70.0 155.5 90–100 ..................... 80.0 452.6 112.5 643.7 170.4 824.1 221.9 905.8 Ageofhead(years) Lessthan35 ................ † 11.8 † 17.9 † 30.3 † 28.1 35–44 ...................... 4.9 44.7 8.2 73.4 11.7 99.0 19.4 119.6 45–54 ...................... 22.9 117.7 25.9 146.0 29.8 164.5 39.0 205.4 55–64 ...................... 34.0 169.5 40.8 216.8 43.6 237.2 58.0 286.7 65–74 ...................... 36.9 145.3 53.9 190.2 51.1 254.9 50.0 231.4 75ormore.................. 40.2 106.0 41.7 131.8 53.3 159.8 58.1 189.0 Note: Seenotetotable1. †Lessthan0.05($50). increases or decreases in the value of assets that are 10. Amountofdebtofallfamilies,distributed yet to be sold. To obtain information on this part of bytypeofdebt,1995–2004surveys net worth, the survey asks about changes in value Percent from the time of purchase for certain key assets— Typeofdebt 1995 1998 2001 2004 publicly traded stocks, pooled investment funds, the Securedbyresidentialproperty primary residence, other real estate, and the current Primaryresidence ............. 73.1 71.4 75.2 75.2 taxbasisofbusinesses.33Themedianunrealizedcapi- Other......................... 7.6 7.5 6.2 8.5 Linesofcreditnotsecuredby tal gain in these assets over the 2001–04 period residentialproperty ......... .6 .3 .5 .7 Installmentloans ................ 12.0 13.1 12.3 11.0 moved up 43.8 percent, and the mean moved up Creditcardbalances ............. 3.9 3.9 3.4 3.0 17.0 percent (table 9); during the 1998–2001 period, Other ........................... 2.9 3.7 2.3 1.6 Total ....................... 100 100 100 100 the median had risen 28.0 percent, and the mean had Memo risen 23.5 percent. The recent gains predominantly Debtasapercentage oftotalassets .............. 14.6 14.2 12.1 15.0 accrued to the middle income groups and to age groupsotherthantheyoungestandthe65–74groups. Note: Seenotetotable1andtextnote38. The rise in unrealized gains reflects strong appreciation of residential real estate over the period as well astherelativeilliquidityofrealestateandbusinesses. differenttypesofdebtasashareoftotaldebtwerean Of the total amount of unrealized capital gains in increase in the share of loans for other residential 2004, 44.6 percent was due to appreciation of pri- property and a decrease in the share of installment mary residences; the comparable figure for 2001 had loans (table 10). The largest share of total debt was been 35.6 percent (data not shown in the tables). In debtsecuredbytheprimaryresidence,theamountof 2004, unrealized gains measured in the SCF whichkeptpacewiththeincreaseintotaldebt. accounted for 30.7 percent of the assets of all fami- Because liabilities increased faster than assets, the lies; the median share of such gains relative to assets ratiooftheoverallsumoffamilydebtstothesumof overallfamilieswas11.2percent. their assets (the leverage ratio) rose 2.9 percentage points, from 12.1 percent to 15.0 percent (table 10, memo line).34 This increase follows a 2.1 percentage LIABILITIES point decrease over the preceding three years. If the calculation is restricted to families that had debt, the Liabilities and assets increased substantially from leverage ratio was 19.9 percent in 2004, an increase 2001 to 2004, but the rise in liabilities was more of 3.4 percentage points from 2001 (data not shown rapidoverall.Overthistime,theprincipalchangesin inthetables). 33. The survey does not collect information on capital gains on 34. Data from the flow of funds accounts show that the leverage everyassetforwhichsuchgainsarepossible.Mostnotably,itdoes ratioforthehouseholdsectorincreasedfrom16.3percentin2001to notcollectsuchinformationforretirementaccounts. 18.1percentin2004.

A26 Federal Reserve Bulletin 2006 11. Familyholdingsofdebt,byselectedcharacteristicsoffamiliesandtypeofdebt,2001and2004surveys A. 2001SurveyofConsumerFinances Securedbyresidential Linesof property creditnot Credit Family Installment Any securedby card Other characteristic loans debt Primary residential balances Other residence property Percentageoffamiliesholdingdebt Allfamilies....................... 44.6 4.6 1.5 45.2 44.4 7.2 75.1 Percentileofincome Lessthan20 ...................... 13.8 * 1.3 25.5 30.3 5.9 49.3 20–39.9........................... 27.0 1.8 1.5 43.2 44.5 5.6 70.2 40–59.9........................... 44.4 3.2 1.5 51.9 52.8 7.7 82.1 60–79.9........................... 61.8 5.3 1.5 56.7 52.6 7.7 85.6 80–89.9........................... 76.9 10.3 2.6 55.7 50.3 9.3 91.4 90–100 ........................... 75.4 14.2 1.4 41.2 33.1 8.8 85.3 Ageofhead(years) Lessthan35 ...................... 35.7 2.7 1.7 63.8 49.6 8.8 82.7 35–44 ............................ 59.6 4.9 1.7 57.1 54.1 8.0 88.6 45–54 ............................ 59.8 6.4 1.5 45.9 50.4 7.4 84.6 55–64 ............................ 49.0 7.4 3.1 39.3 41.6 7.4 75.4 65–74 ............................ 32.0 3.4 * 21.1 30.0 5.0 56.8 75ormore........................ 9.5 2.0 * 9.5 18.4 3.6 29.2 Raceorethnicityofrespondent Whitenon-Hispanic ............... 47.6 5.3 1.7 45.4 43.3 7.4 75.8 NonwhiteorHispanic ............. 35.6 2.4 1.2 44.4 47.6 6.5 73.0 Currentworkstatusofhead Workingforsomeoneelse ......... 52.5 5.3 1.4 57.0 53.2 8.2 86.5 Self-employed .................... 59.1 7.3 3.5 39.8 42.8 8.1 81.7 Retired ........................... 19.6 1.9 * 17.2 24.0 4.4 44.2 Othernotworking................. 28.1 * * 41.5 32.3 6.1 61.9 Housingstatus Owner ............................ 66.0 5.8 1.0 45.5 44.4 6.9 79.9 Renterorother.................... ... 2.0 2.8 44.5 44.3 7.8 65.0 Percentileofnetworth Lessthan25 ...................... 11.2 * 2.4 48.9 45.5 8.3 68.7 25–49.9........................... 49.4 2.0 1.3 51.0 55.1 7.2 80.8 50–74.9........................... 59.1 5.4 * 48.2 44.6 7.1 78.0 75–89.9........................... 61.1 7.8 * 37.2 38.9 4.9 74.8 90–100 ........................... 55.5 14.2 2.1 25.6 22.4 8.2 70.2 Holdings of Debt groups is driven in large part by the paying off of mortgages on primary residences. Over the recent The share of families with any type of debt climbed three-year period, the median amount of outstanding 1.3 percentage points during 2001–04, to 76.4 per- debtroseforallgroupsexceptforfamiliesheadedby cent (first half of tables 11.A and 11.B, last column); persons who were neither working nor retired. The the share had risen 1.0 percentage point over the increases in the median were particularly notable for preceding three years. Borrowing is less prevalent familiesheadedbypersonsaged65orolder,buttheir among families in the lowest income and wealth medianremainedmuchbelowtheoverallmedian. groups and in age groups 65 or older. Over the 2001–04 period, the prevalence of borrowing declined for renters, the youngest age group, and the Mortgages and Other Borrowing on the Primary lowest quartile of the wealth distribution and Residence increased or held about steady for the other groups. The largest increase was the 11.1 percentage point Continuing an earlier trend of increases, the proporriseforfamiliesheadedbypersonsaged75orolder. tion of families with debt secured by the primary The overall median and mean values of total out- residence(hereafter,home-secureddebt)rose3.3perstanding debt for families that had any each rose centage points, to 47.9 percent (the share of home- 33.9 percent from 2001 to 2004; from 1998 to 2001, owners with such debt in 2004 was 69.3 percent).35 median debt had increased 9.5 percent and the mean 5.2percent.Acrossdemographicgroups,mediandebt 35. Home-secureddebtconsistsoffirst-lienandjunior-lienmorttends to rise with income and wealth and to rise and gages and home equity lines of credit secured by the primary resithen decline with age. The decline among older age dence. For purposes of this article, first- and junior-lien mortgages

Recent Changes in U.S. Family Finances: Evidence from the 2001 and 2004 Survey of Consumer Finances A27 11.—Continued A. 2001SurveyofConsumerFinances—Continued Securedbyresidential Linesof property creditnot Credit Family Installment Any securedby card Other characteristic loans debt Primary residential balances Other residence property Medianvalueofholdingsforfamiliesholdingdebt(thousandsof2004dollars) Allfamilies....................... 74.6 42.6 4.2 10.3 2.0 3.2 41.3 Percentileofincome Lessthan20 ...................... 29.8 * .6 4.9 1.1 1.1 5.5 20–39.9........................... 42.6 32.0 1.1 7.0 1.3 3.2 12.2 40–59.9........................... 59.8 41.3 .7 10.3 2.1 2.1 31.0 60–79.9........................... 80.5 44.7 4.3 12.7 2.4 3.2 66.4 80–89.9........................... 96.9 33.2 8.3 15.4 4.0 4.3 103.1 90–100 ........................... 142.7 83.1 10.7 14.3 3.0 22.4 155.9 Ageofhead(years) Lessthan35 ...................... 82.0 55.4 .5 10.2 2.1 2.1 26.5 35–44 ............................ 85.2 52.2 .7 11.8 2.1 3.3 65.5 45–54 ............................ 79.9 35.7 5.7 10.3 2.4 5.3 57.8 55–64 ............................ 58.6 41.3 21.8 9.5 2.0 5.3 36.9 65–74 ............................ 41.5 82.0 * 7.5 1.0 2.7 14.0 75ormore........................ 47.7 44.7 * 6.2 .8 2.7 5.3 Raceorethnicityofrespondent Whitenon-Hispanic ............... 79.9 42.6 4.3 10.7 2.1 3.9 47.7 NonwhiteorHispanic ............. 63.9 42.6 .7 8.8 1.6 2.1 21.3 Currentworkstatusofhead Workingforsomeoneelse ......... 78.8 39.9 3.2 10.6 2.1 2.2 45.3 Self-employed .................... 106.5 93.2 16.0 10.8 2.7 12.7 82.9 Retired ........................... 33.6 46.6 * 7.4 .9 3.5 10.4 Othernotworking................. 76.7 * * 10.4 2.1 2.7 36.0 Housingstatus Owner ............................ 74.6 42.6 16.0 11.2 2.2 4.3 73.9 Renterorother.................... ... 40.0 1.1 7.5 1.3 2.1 6.4 Percentileofnetworth Lessthan25 ...................... 60.7 * .6 8.8 1.7 2.1 9.3 25–49.9........................... 60.2 21.3 1.9 10.0 2.0 1.3 41.0 50–74.9........................... 73.5 50.1 * 10.7 2.1 4.3 63.9 75–89.9........................... 91.6 32.0 * 12.5 2.2 7.5 85.2 90–100 ........................... 143.8 83.1 21.8 12.1 2.1 32.0 130.7 Memo Meanvalueofholdingsfor familiesholdingdebttype..... 97.7 78.9 19.2 15.9 4.4 18.8 77.2 Note: Seenotetotable10. *Tenorfewerobservations. ... Notapplicable. In2004,45.0percent(42.3percentin2001)offami- lies in the 45–54 group and declines sharply among lieshadafirst-lienmortgage,4.2percent(5.8percent older age groups, a pattern also seen in earlier years. in 2001) had a junior-lien mortgage, and 8.6 percent Over the recent period, the prevalence of home- (4.8percentin2001)hadahomeequitylineofcredit secured debt increased for nearly every demographic withacurrentbalance(datanotshowninthetables). group. Of the types of debt considered in this article, home- Overall, the median amount of home-secured debt secured debt had the largest change in overall preva- rose 27.3 percent from 2001 to 2004, and the mean lence.Theuseofsuchdebttendstorisewithincome. rose 27.0 percent; the median had increased 3.8 per- Across wealth groups, it is more nearly equal for centovertheprecedingthreeyears,andthemeanhad groups above the bottom quartile; however, home- increased 8.4 percent. In the recent period, median owners in the lowest wealth group in 2004 had the borrowingrosesubstantiallyforeverygroupbutone. highest rate of such borrowing, 81.6 percent. Over It declined for the 75-or-older age group even as the agegroups,therateofborrowingpeaksamongfami- group had an unusually large increase in the fraction of families having such debt. This result indicates thatthedeclineinthemedianwasdrivenbyarisein consist only of closed-end loans, that is, loans typically with a one-time extension of credit and a prearranged payment size and the number of smaller home-secured loans. Overall, frequency.Asatypeofopen-endcredit,homeequitylinestypically in 2004, 91.3 percent (92.4 percent in 2001) of total allow credit extensions at the borrower’s discretion subject to a home-secureddebtwasowedonfirst-lienmortgages, prearrangedlimitandallowrepaymentsattheborrower’sdiscretion subjecttoaprearrangedminimumsizeandfrequency. 3.0percent(4.4percentin2001)wasowedonjunior-

A28 Federal Reserve Bulletin 2006 11. Familyholdingsofdebt,byselectedcharacteristicsoffamiliesandtypeofdebt,2001and2004surveys—Continued B. 2004SurveyofConsumerFinances Securedbyresidential Linesof property creditnot Credit Family Installment Any securedby card Other characteristic loans debt Primary residential balances Other residence property Percentageoffamiliesholdingdebt Allfamilies....................... 47.9 4.0 1.6 46.0 46.2 7.6 76.4 Percentileofincome Lessthan20 ...................... 15.9 * * 26.9 28.8 4.6 52.6 20–39.9........................... 29.5 1.5 1.5 39.9 42.9 5.8 69.8 40–59.9........................... 51.7 2.6 1.8 52.4 55.1 8.0 84.0 60–79.9........................... 65.8 4.1 1.8 57.8 56.0 8.3 86.6 80–89.9........................... 76.8 7.5 2.6 60.0 57.6 12.3 92.0 90–100 ........................... 76.2 15.4 2.5 45.7 38.5 10.6 86.3 Ageofhead(years) Lessthan35 ...................... 37.7 2.1 2.2 59.4 47.5 6.2 79.8 35–44 ............................ 62.8 4.0 1.5 55.7 58.8 11.3 88.6 45–54 ............................ 64.6 6.3 2.9 50.2 54.0 9.4 88.4 55–64 ............................ 51.0 5.9 .7 42.8 42.1 8.4 76.3 65–74 ............................ 32.1 3.2 .4 27.5 31.9 4.0 58.8 75ormore........................ 18.7 1.5 * 13.9 23.6 2.5 40.3 Raceorethnicityofrespondent Whitenon-Hispanic ............... 51.9 4.4 1.7 47.0 46.0 7.8 78.0 NonwhiteorHispanic ............. 37.4 3.0 1.1 43.2 46.7 7.3 72.5 Currentworkstatusofhead Workingforsomeoneelse ......... 56.1 4.1 1.9 55.7 54.9 9.8 86.1 Self-employed .................... 59.5 10.2 3.0 43.5 44.3 5.8 81.5 Retired ........................... 24.6 1.2 * 22.8 25.9 3.9 50.7 Othernotworking................. 30.3 * * 45.6 41.0 * 70.4 Housingstatus Owner ............................ 69.4 5.1 1.3 46.6 48.8 7.7 82.3 Renterorother.................... ... 1.7 2.1 44.6 40.4 7.3 63.4 Percentileofnetworth Lessthan25 ...................... 12.4 * 1.3 47.5 40.3 6.2 64.9 25–49.9........................... 52.8 1.4 1.7 52.4 57.9 9.4 83.8 50–74.9........................... 66.1 4.5 1.9 49.1 52.8 7.0 83.2 75–89.9........................... 61.6 5.7 1.3 40.2 40.5 7.1 74.6 90–100 ........................... 58.4 16.6 1.4 27.2 23.5 9.1 72.7 lienmortgages,and5.7percent(3.2percentin2001) secondary residences, the Tax Reform Act of 1986 was owed on home equity lines of credit (data not created an incentive for homeowners with a need showninthetables). for additional liquid funds to borrow against their The rising values of primary residences over the homeequity.Overthe2001–04period,somefamilies 2001−04 period outpaced the increases in home- may have felt an important additional incentive from secured debt and thus raised the typical amount of low mortgage interest rates, rapidly appreciating home equity held by families. Median home equity homevalues,andtechnologicalchangesthatreduced among those with home-secured debt rose from the time and cost of mortgage refinancing. Such $61,900 to $70,000 over the period, a 13.1 percent borrowing against home equity may take the form increase (data not shown in the tables).36 Among ofrefinancinganexistingfirst-lienmortgageformore those with such debt, the median ratio of home- than the outstanding balance, obtaining a juniorsecured debt to the value of the primary residence lien mortgage, or accessing a home equity line of held steady at 56.0 percent, down from 58.8 percent credit. in 1998. Over the recent three-year period, an SCF- The survey provides detailed information on all basedestimateoftheaggregateratioofhome-secured these options for home equity borrowing. In 2004, debttohomevaluesforallhomeownersrose1.4per- 44.9 percent of homeowners with a first-lien mortcentagepoints,to34.9percent. gage had refinanced their current first-lien mortgage By eliminating the deductibility of interest pay- in the preceding three years (20.8 percent in 2001), mentsonmostloansotherthanthoseonprimaryand and 34.0 percent of such refinancers had borrowed money beyond the amount refinanced (35.2 percent in 2001); the median amount of additional equity 36. Among all homeowners in 2004, median home equity was $86,000;in2001ithadbeen$74,600. extracted by those who had done so was $20,000

Recent Changes in U.S. Family Finances: Evidence from the 2001 and 2004 Survey of Consumer Finances A29 11.—Continued B. 2004SurveyofConsumerFinances—Continued Securedbyresidential Linesof property creditnot Credit Family Installment Any securedby card Other characteristic loans debt Primary residential balances Other residence property Medianvalueofholdingsforfamiliesholdingdebt(thousandsof2004dollars) Allfamilies....................... 95.0 87.0 3.0 11.5 2.2 4.0 55.3 Percentileofincome Lessthan20 ...................... 37.0 * * 5.6 1.0 2.0 7.0 20–39.9........................... 53.3 32.5 .3 8.0 1.9 2.7 16.1 40–59.9........................... 78.0 66.0 1.0 10.8 2.2 2.3 44.7 60–79.9........................... 97.0 62.0 7.0 13.9 3.0 3.5 93.4 80–89.9........................... 133.0 78.0 14.0 15.1 2.7 5.0 136.0 90–100 ........................... 185.0 159.0 40.0 18.0 4.0 9.4 209.0 Ageofhead(years) Lessthan35 ...................... 107.0 62.5 1.0 11.9 1.5 3.0 33.6 35–44 ............................ 110.0 75.0 1.9 12.0 2.5 4.0 87.2 45–54 ............................ 97.0 87.0 7.0 12.0 2.9 4.0 83.2 55–64 ............................ 83.0 108.8 14.0 12.9 2.2 5.5 48.0 65–74 ............................ 51.0 100.0 4.0 8.3 2.2 5.0 25.0 75ormore........................ 31.0 39.0 * 6.7 1.0 2.0 15.4 Raceorethnicityofrespondent Whitenon-Hispanic ............... 98.0 87.0 4.0 12.4 2.5 4.0 69.5 NonwhiteorHispanic ............. 83.0 66.0 .4 9.6 1.6 3.0 30.5 Currentworkstatusofhead Workingforsomeoneelse ......... 100.0 83.0 4.0 12.0 2.3 3.5 71.8 Self-employed .................... 119.8 100.0 2.2 15.4 2.7 7.0 93.4 Retired ........................... 42.0 79.0 * 7.3 1.4 3.0 15.4 Othernotworking................. 78.0 * * 7.5 2.5 * 21.1 Housingstatus Owner ............................ 95.0 90.0 8.0 12.9 2.5 4.0 95.8 Renterorother.................... ... 83.0 .5 8.7 1.5 3.0 7.8 Percentileofnetworth Lessthan25 ...................... 71.0 * .3 10.5 1.8 4.0 11.4 25–49.9........................... 75.0 26.3 1.0 9.3 2.0 2.0 44.2 50–74.9........................... 97.0 47.0 8.0 13.3 2.5 4.0 90.1 75–89.9........................... 115.0 99.0 22.0 12.9 3.0 5.0 110.7 90–100 ........................... 186.1 148.0 50.0 17.5 3.0 20.0 190.8 Memo Meanvalueofholdingsfor familiesholdingdebttype..... 124.1 166.7 36.6 18.8 5.1 17.1 103.4 Note: Seenotetotable10. *Tenorfewerobservations. ... Notapplicable. (data not shown in the tables).37 Junior-lien mort- balances attributable to equity extraction, and debt gagesnotusedtofinanceahomepurchasewereused consolidation accounted for 31.0 percent (data not by4.7percentofhomeownersin2004(7.2percentin showninthetables). 2001), and the median amount owed on such loans With house prices rising over the past three years, forthosehavingonewas$16,000($20,200in2001). much discussion has centered on how families have The proportion of homeowners with home equity managed to finance the purchase of a new home. linesofcreditin2004was17.8percent(11.2percent Interest rates are a key determinant of the size of the in 2001), and the proportion borrowing against such regular payment that families must make to service lines was 12.4 percent (7.1 percent in 2001); the their mortgages. The median rate on the stock of median balance for those borrowing against such outstanding first-lien mortgages on primary resilines was $22,000 ($16,000 in 2001). For 2004, the denceswas5.90percentin2004(themeanwas6.19) major uses of extracted equity were for home and 7.25 percent in 2001 (the mean was 7.59). Some improvementanddebtconsolidation.Homeimprove- families select a mortgage with a variable interest ment accounted for 45.0 percent of the outstanding rate,typicallybecausesuchloanshavealowerinitial rate than a fixed-rate loan. In 2004, 15.0 percent of 37. Of those with a first-lien mortgage in 2004, 56.7 percent are homeowners with a first-lien mortgage on the prirecordedinthesurveyashavingrefinanceditatleastonce(42.8per- mary residence had an interest rate on their loan that centin2001);35.0percentoftheserefinancersextractedequityinthe could vary; the comparable figure for 2001 was mostrecentinstance(36.2percentin2001),andthemedianamount extractedwas$20,000(datanotshowninthetables). 11.4percent.

A30 Federal Reserve Bulletin 2006 Another key determinant of mortgage payments is Installment Borrowing the length of time over which the loan must be repaid. Mortgages with an initial term of thirty years Installment borrowing is about as common as homeor longer accounted for 57.5 percent of fixed-term secured borrowing.39 In 2004, 46.0 percent of famifirst-lienmortgagesontheprimaryresidencein2004, lies had installment debt, an increase of 0.8 percentand those with a term of fifteen years or less age point over 2001. Although the use of installment accountedfor32.9percent;in2001,62.9percenthad borrowing has increased in each of the past two atermofthirtyyearsormore,and28.6percenthada surveyintervals,theoverallrateofuseiscomparable term of fifteen years or less. Some purchasers take to the levels seen in the 1992 and 1995 surveys. The out mortgages that do not require them to pay back use of installment borrowing is broadly distributed the entire principal over the contract period of the across demographic groups, with notably lower use loan; in such cases, a payment of any remaining only in the lowest income group, the highest wealth principal is required at the end of the loan term. In group, and families headed by retired persons or 2004, 4.1 percent of first-lien mortgages on primary persons aged 65 or older. From 2001 to 2004, the residences had such a loan feature; in 2001, the median amount owed on installment loans rose comparablefigurehadbeen2.1percent. 11.7percent,andthemeanrose18.2percent.Mostof thedemographicgroupssharedintheoverallincrease inthemedian.Themajorityofinstallmentborrowing Borrowing on Other Residential Real Estate isrelatedtothepurchaseofavehicle(datanotshown inthetables);in2004,suchborrowingaccountedfor From 2001 to 2004, the proportion of families that 55.5 percent of the total amount owed (54.8 percent owned other residential property rose, but the pro- in 2001). The second-largest use of installment borportion with outstanding loans on such properties rowing is for education-related expenses. Balances declined 0.6 percentage point, to 4.0 percent. Only on loans for this purpose in 2004 made up 26.0 perabout one-third of owners in 2004 also had a mort- cent of total installment debt; the comparable figure gage on the property. As with the ownership of for2001hadbeen22.2percent. such property, the associated borrowing is most prevalentamongfamilieswithrelativelyhighincome or wealth. Use of such borrowing declined for most Credit Card Balances demographic groups over the three-year period. But as would be expected from the increased share of As with installment borrowing, the carrying of credit total debt attributable to this type of borrowing, the cardbalancesiswidespreadbutnotablyloweramong amount outstanding rose substantially. Both the the highest and lowest income groups, the highest median and the mean amounts owed more than wealth group, and families headed by persons who doubled. Median and mean amounts also rose are aged 65 or older or are retired.40 From 2001 to substantially among families with mortgages on 2004, the proportion of families carrying a balance other residential real estate in most demographic rose 1.8 percentage points, to 46.2 percent. The pregroups. ceding three years had seen a much smaller increase inuse.Therecentincreasewassharedbymostdemographic groups; the proportion carrying a balance Borrowing on Other Lines of Credit declined for the lowest two income groups, the lowest wealth group, the youngest age group, nonwhite Only 3.3 percent of families had an available line of orHispanicfamilies,andrenters. credit other than a home equity line in 2004 (data not shown in the tables).38 Even fewer families— 39. The term ‘‘installment borrowing’’ in this article describes 1.6 percent—had a balance on such a line, a propor- closed-end consumer loans, that is, those that typically have fixed payments and a fixed term. Examples are automobile loans, student tion virtually unchanged from 2001. The median loans,andloansforfurniture,appliances,andotherdurablegoods. amount outstanding on these lines fell 28.6 percent 40. In this article, credit card balances consist of balances on over this three-year period, while the mean rose bank-typecards(suchasVisa,MasterCard,andDiscover,andOptima and other American Express cards that routinely allow carrying a 90.6percent. balance), store cards or charge accounts, gasoline company cards, so-called travel and entertainment cards (such as American Express cardsthatdonotroutinelyallowcarryingabalanceandDinersClub), othercreditcards,andrevolvingstoreaccountsthatarenottiedtoa 38. Inthisarticle,borrowingonlinesofcreditexcludesborrowing credit card. Balances exclude purchases made after the most recent oncreditcards. billwaspaid.

Recent Changes in U.S. Family Finances: Evidence from the 2001 and 2004 Survey of Consumer Finances A31 Overall, the median balance for those carrying a onallbank-typecardsheldbythefamilyroseslightly balance rose 10.0 percent, to $2,200; the mean rose over the recent three-year period, from $210 in 2001 15.9 percent, to $5,100. Over the preceding three to $250 in 2004. For families having any bank-type years, the median had been little changed, but the cards, the median number of such cards remained at meanhadfallen8.3percent.Intherecentperiod,the 2; the median credit limit on all such cards rose median balance rose strongly for most demographic 26.2percent,to$13,500;andthemedianinterestrate groups; but borrowing declined notably for the low- onthecardwiththelargestbalance(oronthenewest est and next-to-highest income groups and for the card, if there were no outstanding balances) fell youngestagegroup. 3.5percentagepoints,to11.50percent. Many families with credit cards do not carry balances.41 Of the 74.9 percent of families with credit cards in 2004, only 58.0 percent had a balance at the Other Debt time of the interview; in 2001, 76.2 percent had cards, and 55.4 percent of these families had an From 2001 to 2004, the proportion of families that outstanding balance on them (data not shown in the held other types of debt edged up 0.4 percent, to tables). The proportion of cardholders who had a 7.6 percent.42 In 2004, 0.5 percent of families had a bank-type card was unchanged over this three-year marginloan,3.5percenthadaloanagainstapension period,whereastheproportionofcardholdershaving fromacurrentjobofthefamilyheadorthatperson’s most other card types declined, as shown in the spouse or partner, 1.6 percent had a loan against a followingtable: cash value life insurance policy, and 2.7 percent had another miscellaneous type of loan (data not shown inthetables). Cardholdersholding Typeof The use of other debt is spread broadly across creditcard 2004 Change2001−04 (percent) (percentagepoints) demographic groups, but rates of use are notably lower for families headed by those who are 65 years Bank.......................... 95.4 .0 Store.......................... 58.4 −1.0 of age or older and by those who are retired. Across Gasoline ...................... 17.3 −3.8 Travelandentertainment ....... 10.0 −3.8 income groups, use of such debt fell from 2001 to Miscellaneous ................. 2.6 .2 2004 only for the lowest group. The median amount owedbyfamilieswiththistypeofdebtrose25.0per- Thedeclinesincardownershipprobablyreflect,at cent, to $4,000, between 2001 and 2004; over the least in part, a rise during the period in the issuance same time, the mean fell 9.0 percent. In 2004, of bank-type cards under the brand names of stores 50.4 percent of the total amount of this type of debt and gasoline companies and in the issuance of new was attributable to margin loans, 21.2 percent to types of American Express card that routinely allow loans against a pension from a current job of the carryingabalance. family head or that person’s spouse or partner, As the most widely held type of card, bankcards 9.8 percent to loans against cash value life insurance hold particular importance in any examination of policies, and the remaining 18.7 percent to miscellafamily finances. Indeed, balances on such cards neousloans(datanotshowninthetables). accounted for 84.9 percent of outstanding credit card balances in 2004, up from 82.1 percent in 2001. As reflected in the overall movement for credit cards Reasons for Borrowing from 2001 to 2004, the proportion of bankcard holderswhohadabalancewentup2.5percentagepoints, The SCF provides information on the reasons that to 56.2 percent; the proportion of bankcard holders families borrow money (table 12). One subtle probwho reported that they usually pay their balances in lem with the use of these data is that, even though fullwasaboutunchangedin2004at55.7percent.For moneyisborrowedforaparticularpurpose,itmaybe themonthprecedingtheinterview,themediancharge employed to offset some other use of funds. For example, a family may have sufficient funds to purchase a home without using a mortgage but may 41. The remaining discussion of credit cards excludes revolving storeaccountsthatarenottiedtoacreditcard.In2004,6.0percent (5.5 percent in 2001) of families had such an account, the median outstandingbalanceforfamiliesthathadabalancewas$700($600in 42. The‘‘otherdebt’’categorycomprisesloansoncashvaluelife 2001), and the total of such balances accounted for 4.3 percent insurance policies, loans against pension accounts, borrowing on (5.2percentin2001)ofthetotalofbalancesoncreditcardsandsuch margin accounts, and a miscellaneous category largely comprising storeaccounts(datanotshowninthetables). personalloansnotexplicitlycategorizedelsewhere.

A32 Federal Reserve Bulletin 2006 12. Amountofdebtofallfamilies,distributedbypurpose 13. Amountofdebtofallfamilies,distributed ofdebt,1995–2004surveys bytypeoflendinginstitution,1995–2004surveys Percent Percent Purposeofdebt 1995 1998 2001 2004 Typeofinstitution 1995 1998 2001 2004 Primaryresidence Commercialbank ............... 34.9 32.8 34.1 35.1 Purchase ..................... 70.3 67.9 70.9 70.2 Thriftinstitution1................ 10.8 9.7 6.1 7.3 Improvement ................. 2.0 2.1 2.0 1.9 Creditunion .................... 4.5 4.3 5.5 3.6 Otherresidentialproperty ........ 8.2 7.8 6.5 9.5 Financeorloancompany ........ 3.2 4.1 4.3 4.1 Investmentsexcludingrealestate. 1.0 3.3 2.8 2.2 Brokerage....................... 1.9 3.8 3.1 2.5 Vehicles ........................ 7.6 7.6 7.8 6.7 Mortgageorrealestatelender ... 32.8 35.6 38.0 39.4 Goodsandservices.............. 5.7 6.3 5.8 6.0 Individuallender ................ 5.0 3.3 2.0 1.7 Education ....................... 2.7 3.5 3.1 3.0 Othernonfinancial............... .8 1.3 1.4 2.0 Unclassifiableloansagainst Government..................... 1.2 .6 1.1 .7 pensionaccounts ........... .2 † † † Creditcardissuer ............... 3.9 3.9 3.7 3.0 Other ........................... 2.2 1.5 1.1 .6 Pension ......................... .2 .4 .3 .3 Total ......................... 100 100 100 100 Other ........................... .7 .3 .5 .2 Total ....................... 100 100 100 100 Note: Seenotetotable7. †Lessthan0.05percent. Note: Seenotetotable1. 1. Savingsandloanassociationorsavingsbank. insteadchoosetofinancethepurchasetofreeexisting tern. The share for commercial banks moved up funds for another purpose. Thus, trends in the data 1.0percentagepoint,whiletheshareforcreditunions can only suggest the underlying use of funds by fell 1.9 percentage points. Other smaller changes families. accounted for the rest of the pattern of changes in Althoughthesurveyinformationonuseissubstan- 2004. tial, it is not exhaustive. Most importantly, for the In some cases, loans may have been held at the case of credit cards it was deemed impractical to ask time of the interviews by institutions other than the about the purposes of borrowing that might well be onesthatoriginallymadetheloans.Resaleofloansis heterogeneous for individual families. For the analy- particularly important for mortgage debt. According sis here, all credit card debt is included in the cate- to the 2004 survey, 41.5 percent of the first-lien gory ‘‘goods and services.’’ The surveys before 2004 mortgages on primary residences were held by lendlack information on the use of funds borrowed ersotherthantheonesthatmadetheoriginalloans,a through a first-lien mortgage; therefore, for purposes figure only slightly changed from 2001.43 In dollarof this calculation, all funds owed on a first-lien weighted terms, the results are similar; mortgages mortgageonaprimaryresidenceareassumedtohave withnon-originatinglendersaccountfor43.3percent been used for the purchase of the home, even when of the outstanding balances on first-lien mortgages thehomeownerhasrefinancedandextractedequity. forprimaryresidencesin2004(datanotshowninthe The great majority of family debt is attributable to tables). the purchase of a primary residence; from 2001 to 2004, the share of debt for this purpose declined a fractionofapercentagepoint.Borrowingforresiden- Debt Burden tial real estate other than a primary residence, the second-largest purpose for borrowing, rose notably. The ability of individual families to service their Theshareofborrowingforvehicles,thethird-largest loans is a function of two factors: the level of their share, fell 1.1 percentage points. The shares of bor- loan payments and the income and assets they have rowingforotherpurposesheldaboutsteady. available to meet those payments. In planning their borrowing, families make assumptions about their future ability to repay their loans. Problems may Choice of Lenders occur when events turn out to be contrary to those assumptions. If such misjudgments are sufficiently The survey provides information on the types of large and prevalent, a broad pattern of default, lender to which families owe money at the time of the interview (table 13). The share of total family debt held by thrift institutions—savings and loan 43. Mortgagesandotherloansmayalsobeservicedbyaninstituinstitutions and savings banks—rose 1.2 percentage tionotherthanthecurrentlender,andsomerespondentsmaymistakpointsfrom2001to2004,reversingaprevioustrend; enly report their loan as having been sold even though it is simply beingservicedbyaninstitutionotherthanthecurrentlender.Because the1.4percentagepointincreaseintheshareforreal aloancanalsobesoldwithoutchangingtheservicer,someborrowers estate or mortgage lenders continued an earlier pat- maymistakenlyreportthattheirloanhasnotbeensold.

Recent Changes in U.S. Family Finances: Evidence from the 2001 and 2004 Survey of Consumer Finances A33 restraint in spending, and financial distress in the total debt payments for all families to total family widereconomymightensue. income of all families. From 2001 to 2004, the SCF- From the third quarter of 2001 to the same period based estimate rose more than the aggregate-level in 2004, inflation-adjusted aggregate household debt measure, increasing 1.5 percentage points, to reported in the Federal Reserve’s flow of funds 14.4 percent; in the previous three-year period, the accounts increased 26.3 percent.44 At the same time, SCF measure had declined while the aggregate-level incomewasrelativelyflat,andinterestratestendedto measurehadrisen.Iftotalpaymentsandincomesare belowerattheendoftheperiod.Thetypicalcontract computed from the survey data using only families periods of various types of loan appear to have been with debt payments, the results for the recent period largely unchanged, but borrowers may have substi- show a slightly larger increase, from 16.0 percent in tuted longer-term home-secured debt for other debts 2001to17.7percentin2004;iftheratioiscomputed that typically have shorter contract periods. Thus, using only families with home-secured debt, the data whetherthegrowthindebttranslatedintoachangein showarisefrom18.2percentin2001to20.1percent families’ ability to service their debts is not clear in2004(datanotshowninthetables). a priori. The net consequences of these movements The ability to look at the distribution of payments on the ratio of payments to income can only be relative to income at the level of families potentially assessed by looking at how these factors vary offers insights that are not available from any of the togetheroverfamilies. aggregate-level figures. In particular, the survey The Federal Reserve staff has constructed an allows a detailed look at the spectrum of payments aggregate-level debt service ratio, defined as an esti- relative to income across all families with debts. mate of total scheduled loan payments (interest plus Over the recent period, the median of the ratios for minimum repayments of principal) for all house- individualfamiliesthathadanydebtrose1.3percentholds, divided by disposable personal income. From age points, to 18.0 percent, in 2004; the increase the third quarter of 2001 to the same period in 2004, reversed a 1.2 percentage point decline in the ratio the aggregate-level measure edged up about 0.4 per- over the preceding three years. The median also rose centagepoint,to13.2percent.45 at least slightly in the recent period for all demo- Thesurveydatamaybeusedtoconstructasimilar graphicgroupsshownexceptforthe65–74agegroup estimate of the debt-burden ratio and to construct andrenters,forwhichgroupstheratiofellslightly.47 such an estimate for various demographic groups A limitation of the median ratio is that it may not (table 14).46 The SCF-based estimate is the ratio of be indicative of distress because it reflects the situation of only a typical family. Unless errors of judgment by both families and lenders are pervasive, one 44. Seehttp://www.federalreserve.gov/releases/Z1/Current/. 45. Dataonthismeasure,the‘‘debtserviceratio,’’andadescrip- would not expect to see signs of financial distress at tion of the series are available at www.federalreserve.gov/releases/ the median. Thus, a more compelling indicator of housedebt/default.htm. See Karen Dynan, Kathleen Johnson, and distress is the proportion of families with unusually KarenPence(2003),‘‘RecentChangestoaMeasureofU.S.Household Debt Service,’’ Federal Reserve Bulletin, vol. 89 (October), large total payments relative to their incomes. From pp.451–60. 2001 to 2004, the proportion of debtors with pay- 46. Thesurveymeasureofpaymentsrelativetoincomemaydiffer mentsexceeding40percentoftheirincomeedgedup fromtheaggregate-levelmeasureforseveralreasons.First,thedebt paymentsincludedineachmeasurearedifferent.Theaggregate-level 0.4 percentage point, to 12.2 percent; in the precedmeasureincludesonlydebtsoriginatedbydepositories,financecom- ingthreeyears,theproportionhadfallen1.8percentpanies,andotherfinancialinstitutions,whereasthesurveyincludes,in age points. The survey shows an interesting pattern principle,debtsfromallsources. Second, the aggregate-level measure uses a NIPA estimate of of increases and decreases in the proportion of famidisposable personal income for the period concurrent with the esti- lies with debt that had relatively high payments matedpaymentsasthedenominatoroftheratio,whereasthesurvey across demographic groups in the recent three-year measureusestotalbefore-taxincomereportedbysurveyfamiliesfor theprecedingyear;thedifferencesinthesetwoincomemeasuresare period. The share fell for families in the lowest and complex. the two highest income and wealth groups and for Third, the payments in the aggregate-level measure are estimated using a formula that entails complex assumptions about minimum payments and the distribution of loan terms at any given time; the Current Population Survey. Over the 2001–04 period, however, the surveymeasureofpaymentsisdirectlyaskedofthesurveyrespon- SCF shows more growth in the aggregate level of debt than the dentsbutmayalsoincludepaymentsoftaxesandinsuranceonreal Federal Reserve’s flow of funds accounts; timing and conceptual estateloans. differences may explain some of the difference. Finally, the survey Fourth,becausethesurveymeasuresofpaymentsandincomeare measureexcludesdebtpaymentsofhouseholdmemberswhoarenot based on the responses of a sample of respondents, they may be membersofthefamilyunitanalyzedinthisarticle. affectedbothbysamplingerrorandbyvarioustypesofresponseerror. 47. Themedianoftheratioforfamilieswithhome-secureddebtin Asmentionedearlierinthisarticle,thesurveyincomemeasuretracks 2004was24.2percent,upfrom22.2percentin2001(datanotshown the most comparable measure of income in the Census Bureau’s inthetables).

A34 Federal Reserve Bulletin 2006 14. Ratioofdebtpaymentstofamilyincome(aggregateandmedian),shareofdebtorfamilieswithratiogreaterthan40 percent,andshareofdebtorswithanypaymentpastduesixtydaysormore,1995–2004surveys Percent Aggregate Medianfordebtors Family characteristic 1995 1998 2001 2004 1995 1998 2001 2004 Allfamilies..................... 14.1 14.9 12.9 14.4 16.2 17.9 16.7 18.0 Percentileofincome Lessthan20 .................... 19.1 18.7 16.1 18.2 13.3 18.8 19.2 19.7 20–39.9......................... 17.0 16.5 15.8 16.7 17.5 17.5 16.7 17.4 40–59.9......................... 15.6 18.6 17.1 19.4 15.7 19.4 17.6 19.5 60–79.9......................... 17.9 19.1 16.8 18.5 18.9 19.5 18.1 20.6 80–89.9......................... 16.6 16.8 17.0 17.3 16.8 17.8 17.3 18.1 90–100 ......................... 9.5 10.3 8.1 9.3 12.6 13.7 11.2 12.7 Ageofhead(years) Lessthan35 .................... 17.8 17.2 17.2 17.8 16.8 16.9 17.7 18.0 35–44 .......................... 17.2 17.7 15.1 18.2 18.3 20.0 17.8 20.6 45–54 .......................... 15.1 16.3 12.8 15.3 16.6 17.9 17.4 18.4 55–64 .......................... 11.8 13.4 10.9 11.5 14.2 17.6 14.3 15.8 65–74 .......................... 7.2 8.8 9.2 8.7 12.3 13.2 16.0 15.6 75ormore...................... 2.5 4.1 3.9 7.1 2.9 8.1 8.0 12.8 Percentileofnetworth Lessthan25 .................... 13.4 15.0 13.4 13.0 11.7 13.6 11.5 13.0 25–49.9......................... 18.5 20.1 18.0 19.5 19.0 20.0 20.1 21.2 50–74.9......................... 18.0 18.3 16.8 20.6 19.3 20.2 18.3 21.4 75–89.9......................... 14.0 14.8 15.4 15.1 15.3 17.8 16.8 17.9 90–100 ......................... 9.0 10.2 7.5 8.5 12.7 14.0 11.2 12.6 Housingstatus Owner .......................... 15.6 16.2 13.9 15.6 20.1 21.2 20.0 21.5 Renterorother.................. 7.9 8.2 7.4 7.2 8.1 8.5 8.3 8.2 Note: The aggregate measure is the ratio of total debt payments to total incomeforallfamilies.Themedianisthemedianofthedistributionofratios calculatedforindividualfamilieswithdebt.Alsoseenotetotable1. families headed by persons older than 55; it rose for asked whether they have been behind in any of their themiddleoftheincomeandwealthdistributionsand payments in the preceding year. This measure differs for younger families. Both for homeowners and for conceptuallyfromtheaggregatedelinquencyratesin renters, the proportion with high payments was only that the survey counts multiple occasions of late slightlychanged.48 payments as one, counts families instead of balances Other commonly used indicators of debt- oraccounts,andincludesalltypesofloan;becauseit repaymentproblemsareaggregatedelinquencyrates, counts individual families, not their balances, it is that is, the number of delinquent accounts or the closer in spirit to aggregate measures based on the percentage of total balances on which payments are numbers of delinquent accounts than to those based late. The measures based on numbers of delinquent on the amounts of delinquent balances. Over the accounts tend to show increases or small declines 2001–04 period, the survey shows an increase of overtherecentthree-yearperiod,whilethemeasures 1.9 percentage points in the proportion of debtors basedondollarvolumesshowadecline.49 whoweresixtyormoredayslatewiththeirpayments ArelatedmeasureiscollectedintheSCF.Families onanyoftheirloansintheprecedingyear,to8.9perthat have any debt at the time of their interview are cent. This measure showed increases for all the demographic groups except for the highest two income groups, for the third quartile and highest 48. Of families with home-secured debt, the proportion that had decile of the wealth distribution, and for families totalpaymentsofmorethan40percentoftheirincomewas17.1per- headedbypersonsaged55–64.50Someofthelargest centin2004,afigurevirtuallyunchangedfromthatin2001(datanot increases were seen by groups that had more modest showninthetables). 49. Several measures of credit delinquency are commonly used. or even negative changes in the other survey-based DatafromtheCallReportandMoody’sInvestorsServicearebased measuresofdebtburden. ondollarvolumesofdelinquentloans.Thosedatasuggestthatdelinquenciesgenerallydeclinedbetween2001and2004oncreditcards, on closed-end consumer credit, and on mortgages. Over the same period, however, data from the American Bankers Association on 50. Forfamilieswithhome-secureddebt,theresultisverysimilar numbersofdelinquentaccountsshowasmallerdeclineindelinquen- to that for homeowners overall. The proportion with payments late ciesforclosed-endconsumerloans,littlechangeformortgages,and sixtydaysormorein2004was5.7percent,upfrom4.5percentin anincreaseforcreditcards. 2001(datanotshowninthetables).

Recent Changes in U.S. Family Finances: Evidence from the 2001 and 2004 Survey of Consumer Finances A35 14.—Continued Percent Debtorswithratiogreaterthan40percent Debtorswithanypaymentpastduesixtydaysormore Family characteristic 1995 1998 2001 2004 1995 1998 2001 2004 Allfamilies..................... 11.7 13.6 11.8 12.2 7.1 8.1 7.0 8.9 Percentileofincome Lessthan20 .................... 27.5 29.9 29.3 27.0 10.2 12.9 13.4 15.9 20–39.9......................... 18.0 18.3 16.6 18.6 10.1 12.3 11.7 13.8 40–59.9......................... 9.9 15.8 12.3 13.7 8.7 10.0 7.9 10.4 60–79.9......................... 7.7 9.8 6.5 7.1 6.6 5.9 4.0 7.1 80–89.9......................... 4.7 3.5 3.5 2.4 2.8 3.9 2.6 2.3 90–100 ......................... 2.3 2.8 2.0 1.8 1.0 1.6 1.3 .3 Ageofhead(years) Lessthan35 .................... 12.1 12.8 12.0 12.8 8.7 11.1 11.9 13.7 35–44 .......................... 9.9 12.5 10.1 12.6 7.7 8.4 5.9 11.7 45–54 .......................... 12.3 12.9 11.6 13.1 7.4 7.4 6.2 7.6 55–64 .......................... 15.1 14.0 12.3 10.2 3.2 7.5 7.1 4.2 65–74 .......................... 11.3 18.1 14.7 11.6 5.3 3.1 1.5 3.4 75ormore...................... 7.4 21.4 14.6 10.7 5.4 1.1 .8 3.9 Percentileofnetworth Lessthan25 .................... 10.1 13.0 11.6 10.6 14.5 16.1 17.7 22.9 25–49.9......................... 12.9 15.9 14.1 15.8 8.2 9.8 7.2 11.0 50–74.9......................... 12.7 13.0 11.3 12.8 4.4 5.5 3.6 3.2 75–89.9......................... 9.9 12.2 10.7 9.6 2.4 1.0 .7 1.1 90–100 ......................... 11.6 12.4 8.5 7.6 .7 2.4 .3 .1 Housingstatus Owner .......................... 14.3 16.5 14.7 14.9 5.1 6.1 4.3 5.6 Renterorother.................. 5.8 6.4 4.2 4.4 11.5 12.8 14.0 18.6 SUMMARY another type of managed asset account fell to about 49percentin2004afterhavingreachedanSCFhigh Data from the SCF show that despite small changes ofalmost52percentin2001. in real family income over the 2001–04 period—an Logically balancing the decline in the share of increaseof1.6percentinthemedianandadeclineof financial assets in families’ portfolios was the rise in 2.3 percent in the mean—families overall still saw the share of their nonfinancial assets. The most someincreaseintheirnetworth.Themedianvalueof importantfactorinthisrisewasresidentialrealestate. net worth rose 1.5 percent, while the mean rose The homeownership rate went up 1.4 percentage more—6.3 percent. However, the measured gains in points, and the ownership rate for other residential wealth in the 2001–04 period pale in comparison real estate (including both second homes and investwith the much larger increase of the preceding three ment properties) went up 1.2 percentage points. At years; from 1998 to 2001, median net worth rose the same time, the value of real estate increased 10.3 percent and the mean 28.7 percent. In the more dramaticallyinmanyareas. recent period, median wealth declined for families in Overall, asset ownership and debt use increased in the bottom 40 percent of the income distribution and both prevalence and amount. The net effect was an rose for those higher in the distribution; in contrast, increaseintheproportionoffamilies’assetsoffsetby mean net worth rose or held about steady for all debts—from about 12 percent in 2001 to 15 percent incomegroups. in 2004. The most important factor in the increase In the three years after the 2001 survey, interest was a rise in the amount of debt associated with rates moved generally lower; indexes of equity mar- residential real estate. The amount of other types of ket performance trended generally downward over debtalsorose. the early part of the period but made up the losses Even with interest rates lower in 2004 than in withgainsin2004;andresidentialrealestateappreci- 2001, the SCF data show a moderate increase in atedstrongly.Againstthisbackdrop,theoverallshare measuresofdebtburden.Theperiodsawincreasesin of financial assets in families’ portfolios, as defined the proportion of families that had been delinquent in this article, declined despite substantial gains in with their payments in the year preceding the survey holdings for some groups. Of particular note, the and in the median ratio of loan payments to family share of families that held stocks either directly or income.Theincreaseindelinquencieswassomewhat indirectlythroughanaccount-typeretirementplanor less broadly spread across demographic groups than

A36 Federal Reserve Bulletin 2006 was the increase in the median ratio. At the same Definition of ‘‘Family’’in the SCF time, the proportion of families with high values of theratiowasonlymarginallyhigher. The definition of ‘‘family’’ used throughout this article differs from that typically used in other government studies. In the SCF, a household unit is APPENDIX: SURVEY PROCEDURES AND divided into a ‘‘primary economic unit’’ (PEU)— STATISTICAL MEASURES the family—and everyone else in the household. The PEU is intended to be the economically domi- Detailed documentation of the SCF methodology is nant single individual or couple (whether married available elsewhere.51 The 2004 data used here are or living together as partners) and all other persons derived from the final internal version of the survey in the household who are financially interdepeninformation.Datafromthissurvey,suitablyalteredto dent with that economically dominant person or protect the privacy of respondents, along with addi- couple. tional tabulations of data from the surveys beginning This report also designates a head of the PEU, not with 1989, will be available in February 2006 on the toconveyajudgmentabouthowanindividualfamily FederalReserve’swebsiteatwww.federalreserve.gov/ is structured but as a means of organizing the data pubs/oss/oss2/scf2004home.html. Links to the data consistently. If a couple is economically dominant in usedinthisarticleforearlierperiodsareavailableon the PEU, the head is the male in a mixed-sex couple that site. Results reported in this article for earlier andtheolderpersoninasame-sexcouple.Ifasingle surveysmaydifferfromtheresultsreportedinearlier individual is economically dominant, that person is articles because of additional statistical processing, designatedasthefamilyheadinthisreport. correction of data errors, revisions to the survey weights, conceptual changes in the definitions of variables used in the articles, and adjustments for Racial and Ethnic Identification inflation. Asapartofthegeneralreconciliationsrequiredfor Inthisarticle,theraceandethnicityofafamilyinthe this article, the survey data were compared with SCFareclassifiedaccordingtotheself-identification many external estimates, a few of which are men- of that family’s original respondent to the SCF intertioned in the text. Generally, the survey estimates view.Thequestionsunderlyingthemethodofclassicorrespond fairly well to external estimates. One ficationusedinthesurveywerechangedinboth1998 particularly important comparison is between the and 2004. Starting in 1998, SCF respondents were SCF and the Federal Reserve’s flow of funds allowed to report more than one racial identification; accounts for the household sector. This comparison in surveys before then, only one response was suggests that when the definitions of the variables in recorded. For maximum comparability with earlier thetwosourcescanbeadjustedtoacommonconcep- data, respondents reporting multiple racial identificatual basis, the estimates of totals in the two systems tions were asked to report their strongest racial tendtobeclose.ThedataseriesintheSCFandinthe identification. flow of funds accounts usually show very similar Beginning with the 2004 survey, the question on growthrates.52Ingeneral,thedatafromtheSCFcan racial identification is preceded by a question on be compared with those of other surveys only in whether respondents consider themselves to be Histermsofthemediansbecauseofthespecialdesignof panic or Latino in culture or origin; previously, such theSCFsample. ethnic identification was captured only to the extent that it was reported as a response to the question on racialidentification.Thesequenceofthesetwoques- 51. SeeArthurB.Kennickell(2000),‘‘WealthMeasurementinthe SurveyofConsumerFinances:MethodologyandDirectionsforFuture tions in the 2004 SCF is similar to that in the CPS. Research’’(Washington:BoardofGovernorsoftheFederalReserve When families in the March 2004 CPS are classified System,May);ArthurB.Kennickell(2001),‘‘ModelingWealthwith inthewaymostcompatiblewiththeSCF,thepropor- Multiple Observations of Income: Redesign of the Sample for the 2001SurveyofConsumerFinances’’(Washington:BoardofGover- tion of Hispanic families is 10.5 percent; the 2004 norsoftheFederalReserveSystem,October),www.federalreserve.gov/ SCF estimate is 11.2 percent. Differences in these pubs/oss/oss2/method.html;andreferencescitedinthesepapers. proportions are attributable to sampling error and 52. For details on how these comparisons are structured and the results of comparisons for earlier surveys, see Rochelle L. possibly to differences in the wording and context of Antoniewicz(2000),‘‘AComparisonofFlowofFundsAccountsand thequestions. theSurveyofConsumerFinances’’(Washington:BoardofGovernors For greater comparability with the earlier SCF of the Federal Reserve System, October), www.federalreserve.gov/ pubs/oss/oss2/method.html. data,thedatareportedinthisarticleignoretheinfor-

Recent Changes in U.S. Family Finances: Evidence from the 2001 and 2004 Survey of Consumer Finances A37 mation on ethnic identification available in 2004, but sents 112.1 million families, and the 2001 survey respondents reporting multiple racial identifications represents106.5millionfamilies.54 in the surveys starting with 1998 are classified as ‘‘nonwhite or Hispanic.’’ For the 1995 survey, only the single recorded response to the racial classifica- The Interviews tion question is used to classify families. In the 2004 SCF, 2.3 percent of respondents reported more than The survey questionnaire has changed in only minor oneracialidentification,upfrom1.5percentin2001. ways since 1989, except in a small number of Ofthosewhorespondedaffirmativelytothequestion instancesinwhichthestructurewasalteredtoaccomonHispanicorLatinoidentificationin2004,85.7per- modate changes in financial behaviors, in types of cent also reported ‘‘Hispanic or Latino’’ as one of financial arrangements available to families, and in their racial identifications, and 82.1 percent reported regulations covering data collection. In the 2004 surit as their primary racial identification. Because the vey, the most important changes were made in the questiononHispanicorLatinoethnicityprecedesthe way data are collected on pensions associated with one on racial identification in the 2004 survey, the current jobs and in the way information is solicited answer to the second of these two questions may about the racial and ethnic identification of families. havebeeninfluencedbytheanswertothefirst.53 Inthesecasesandinallearlierones,everyefforthas beenmadetoensurethemaximumdegreeofcomparability of the data over time. Except where noted in thearticle,thedataarehighlycomparableovertime. The Sampling Techniques The generosity of families in giving their time for interviews has been crucial to the SCF. In the 2004 The survey is expected to provide a core set of data SCF, the median interview length was about eighty on family income, assets, and liabilities. The major minutes. However, in some particularly complicated aspectsofthesampledesignthataddressthisrequirecases, the amount of time needed was substantially ment have been constant since 1989. The SCF commore than two hours. The role of the interviewers in bines two techniques for random sampling. First, a this effort is also critical. Without their dedication standard multistage area-probability sample (a geoandperseverance,thesurveywouldnotbepossible. graphically based random sample) is selected to pro- The SCF interviews were conducted largely vide good coverage of characteristics, such as homebetween the months of May and December in each ownership, that are broadly distributed in the survey year by NORC, a social science and survey population. research organization at the University of Chicago. Second, a supplemental sample is selected to dis- The majority of interviews were obtained in person, proportionately include wealthy families, which hold although interviewers were allowed to conduct telea relatively large share of such thinly held assets as phone interviews if that was more convenient for the noncorporate businesses and tax-exempt bonds. respondent.Inthesurveysbeginningwith1995,each Called the ‘‘list sample,’’ this group is drawn from a interviewerusedaprogramrunningonalaptopcomlist of statistical records derived from tax returns. putertoadministerthesurveyandcollectthedata. These records are used under strict rules governing The use of computer-assisted personal interviewconfidentiality, the rights of potential respondents to ing has the great advantage of enforcing systematic refuse participation in the survey, and the types of collection of data across all cases. The computer information that can be made available. Individuals listed by Forbes magazine as being among the wealthiest 400 people in the United States are Information:ResultsoftheCurrentPopulationSurveySupplementon excludedfromsampling. Race and Ethnicity,’’ BLS Statistical Notes 40, CPS Publications (Washington:BureauofLaborStatistics,June),www.bls.census.gov/ Of the 4,522 interviews completed for the 2004 cps/racethn/1995/stat40rp.htm. SCF, 3,007 were from the area-probability sample, 54. In the development of weights for the SCF, population estiand 1,515 were from the list sample; the figures for matesoftheBureauoftheCensusareakeyinput.Afterthedatafor the 2001 SCF were processed, the Bureau of the Census altered its 2001 are 2,917 from the area-probability sample and population estimates in a way that increases the number of family 1,532 from the list sample. The 2004 survey repre- units relevant for the 2001 SCF to 108.2 million. Pending a more detailedinvestigationintothechangeinthepopulationestimate,the 2001SCFestimatesreportedinthisarticlearecalculatedwithweights 53. Forareviewoftheeffectsofvariousapproachestomeasuring basedontheoriginal,lowerCensuspopulationfigure.Theuseofa race and ethnicity, see Clyde Tucker, Ruth McKay, Brian Kojetin, different number of families does not affect the median and mean Roderick Harrison, Manuel de la Puente, Linda Stinson, and Ed estimates reported in this article. The 1998 survey represents 102.6 Robinson(1996),‘‘TestingMethodsofCollectingRacialandEthnic millionfamilies,andthe1995surveyrepresents99.0millionfamilies.

A38 Federal Reserve Bulletin 2006 program developed to collect the data for the SCF estimatesbasedonasampleinsteadofacensus—isa wastailoredtoallowthecollectionofpartialinforma- particularly important source of error. Such error can tion in the form of ranges whenever a respondent be reduced either by increasing the size of a sample eitherdidnotknowordidnotwanttorevealanexact or, as is done in the SCF, by designing the sample to dollarfigure. reduce important sources of variability. Sampling The response rate in the area-probability sample is error can be estimated, and for this article we use morethandoublethatinthelistsample.Inboth2001 replicationmethodstodoso. and 2004, about 70 percent of households selected Replication methods draw samples, called replifor the area-probability sample actually completed cates,fromthesetofactualrespondentsinawaythat interviews. The overall response rate in the list incorporatestheimportantdimensionsoftheoriginal sample was about 30 percent; in the part of the list sample design. In the SCF, weights were computed sample likely containing the wealthiest families, the for all the cases in each of the replicates.56 For each responseratewasonlyabout10percent. statisticforwhichstandarderrorsarereportedinthis article, the weighted statistic is estimated using the replicate samples, and a measure of the variability of Weighting these estimates is combined with a measure of the variabilityduetoimputationformissingdatatoyield To provide a measure of the frequency with which thestandarderror. families similar to the sample families could be Other errors include those that interviewers may expectedtobefoundinthepopulationofallfamilies, introduce by failing to follow the survey protocol or ananalysisweightiscomputedforeachcaseaccountmisunderstandingarespondent’sanswers.SCFintering both for the systematic properties of the sample viewersaregivenlengthy,project-specifictrainingto design and for differential patterns of nonresponse. minimizesuchproblems.Respondentsmayintroduce The SCF response rates are low by the standards of error by interpreting a question in a sense different other major government surveys, and analysis of the fromthatintendedbythesurvey.FortheSCF,extendataconfirmsthatthetendencytorefuseparticipation sive pretesting of questions and thorough review of is highly correlated with net worth. However, unlike thedatatendtoreducethissourceoferror. other surveys, which also almost certainly have dif- Nonresponse—either complete nonresponse to the ferential nonresponse by wealthy households, the survey or nonresponse to selected items within the SCFhasthemeanstoadjustforsuchnonresponse.A survey—may be another important source of error. major part of SCF research is devoted to the evalua- As noted in more detail above, the SCF uses weighttionofnonresponseandadjustmentsfornonresponse ing to adjust for differential nonresponse to the surintheanalysisweightsofthesurvey.55 vey. To address missing information on individual questions within the interview, the SCF uses statisti- Sources of Error cal methods to impute missing data; the technique makesmultipleestimatesofmissingdatatoallowfor Errorsmaybeintroducedintosurveyresultsatmany anestimateoftheuncertaintyattributabletothistype stages. Sampling error—the variability expected in ofnonresponse. 55. The weights used in this article are adjusted for differential rates of nonresponse across racial and ethnic groups by homeownershipstatus.SeeArthurB.Kennickell(1999),‘‘Revisionsofthe SCF Weighting Methodology: Accounting for Race/Ethnicity and 56. SeeArthurB.Kennickell(2000),‘‘RevisionstotheVariance Homeownership’’ (Washington: Board of Governors of the Federal EstimationProcedurefortheSCF’’(Washington:BoardofGovernors Reserve System, December), www.federalreserve.gov/pubs/oss/oss2/ of the Federal Reserve System, October), www.federalreserve.gov/ method.html. pubs/oss/oss2/method.html.

A39 Industrial Production and Capacity Utilization: The 2005 Annual Revision Kimberly Bayard and Charles Gilbert, of the Board’s Besides the revisions to the monthly data for IP Division of Research and Statistics, prepared this and capacity utilization starting in 1972, the compari­ article. Vanessa Haleco provided research assistance. son base year for all production and capacity indexes was changed: The indexes are now expressed as On November 7, 2005, the Board of Governors of the percentages of output in 2002 instead of 1997.2 The Federal Reserve System issued revisions to its index rebasing affects all series from their start dates: 1919 of industrial production (IP) and the related measures for total IP and manufacturing IP, 1948 for manufac­ of capacity and capacity utilization for the period turing capacity, and 1967 for total industrial capacity. from January 1972 through September 2005. For this Table 1 summarizes the changes to industrial pro­ period, both the levels and the rates of change were duction, capacity, and capacity utilization from 2001 revised. For years before 1972, the levels, but not forward. Measured from fourth quarter to fourth quar­ the rates of change, were also revised. Overall, the ter, industrial output since 2001 is now reported to changes to total industrial production were small have increased a little more overall than reported (figure 1).1 previously. The contraction in 2001 is now shown to be a bit steeper than it was earlier, and the gains in Note: Charles Gilbert directed the 2005 revision and, with Kim­ berly Bayard, David Byrne, Christopher Kurz, Paul Lengermann, Maria Otoo, Dixon Tranum, and Daniel Vine, prepared the revised the data published in the G.17 Federal Reserve Statistical Release estimates of industrial production. Norman Morin, John Stevens, and ‘‘Industrial Production and Capacity Utilization’’ on February 15, Daniel Vine prepared the revised estimates of capacity and capacity 2006. Statements about previously reported estimates refer to the data utilization. David Byrne, Carol Corrado, and Aditya Bhave prepared published on October 14, 2005. the improved estimates for communications equipment. 2. For comparisons in this article between the revised and previous 1. The production and capacity indexes and the utilization rates indexes, the previous indexes are implicitly recomputed to have a base referred to in the text and shown in the tables and charts are based on year of 2002. 1. Industrial production, capacity, and capacity utilization: Total industry, January 1999–January 2006 Production and capacity Ratio scale, 2002 output = 100 Capacity utilization Percent Revised Earlier 140 84 130 82 Capacity 80 120 78 110 Production 76 100 74 2000 2002 2004 2006 2000 2002 2004 2006 NOTE: Here and in the following figures, the shaded areas are periods of Utilization” on February 15, 2006. Data labeled “earlier” reflect those business recession as defined by the National Bureau of Economic Research. published before the November 7, 2005, annual revision. The “earlier” line Data labeled “revised” are the corresponding data published in the G.17 for capacity extends through the end of 2005 because the capacity indexes are Federal Reserve statistical release “Industrial Production and Capacity based on annual projections that are converted to a monthly basis.

A40 Federal Reserve Bulletin 2006 1. Revised rates of change in industrial production and capacity and the revised rates of capacity utilization, 2001–05 Difference between rates of change: Revised rate Memo: revised minus earlier (percent) 2005 (percentage points) Item pro­ portion 2001–05 2001–05 2001 2002 2003 2004 2005 2001 2002 2003 2004 2005 avg. avg. Production Total index ......................... 100.0 1.0 −5.3 2.3 1.5 4.3 2.3 .2 −.2 .8 .3 .0 .2 Manufacturing .................... 80.8 1.2 −5.6 2.2 1.7 5.1 2.6 .2 −.2 .9 .2 .1 .2 Excluding selected high-tech industries .................. 76.0 .5 −5.2 1.9 .4 4.2 1.2 .2 −.3 1.2 .0 .0 .0 Selected high-tech industries .... 4.8 12.0 −9.8 4.8 21.1 18.4 25.2 .0 .3 −3.3 2.4 −.3 1.0 Mining and utilities ............... 19.2 .2 −3.4 2.7 .6 .5 .9 .2 −.1 −.1 1.0 .0 .0 Capacity Total index ......................... 100.0 1.1 2.9 .7 −.2 .6 1.5 .1 .2 .3 .0 −.5 .4 Manufacturing .................... 82.9 1.1 2.8 .4 −.1 .5 2.0 .1 .3 .3 .0 −.6 .6 Excluding selected high-tech industries .................. 77.4 .2 .8 −.2 −.5 .1 .6 .2 .3 .2 .1 .0 .3 Selected high-tech industries .... 5.5 14.3 27.8 9.5 8.0 6.8 19.7 −.0 .3 1.5 −.4 −6.6 5.1 Mining and utilities ............... 17.1 1.5 3.1 2.3 1.3 1.2 −.2 −.1 .1 −.3 .2 .2 −.6 Capacity utilization Total index ......................... 100.0 77.1 74.2 75.3 76.6 79.4 79.8 .1 −.4 −.1 .1 .6 .5 Manufacturing .................... 82.9 75.4 72.1 73.4 74.7 78.2 78.5 .0 −.5 −.1 .0 .5 .3 Excluding selected high-tech industries .................. 77.4 76.2 73.2 74.8 75.4 78.6 78.9 −.1 −.6 .1 .1 .1 −.1 Selected high-tech industries .... 5.5 66.7 61.1 58.5 65.7 72.8 75.3 .2 .3 −2.3 −1.0 3.0 1.1 Mining and utilities ............... 17.1 86.8 87.0 87.3 86.7 86.1 86.7 .7 .2 .3 1.0 .8 1.2 Note: The rates of change for years are calculated from the fourth quarter of The selected high-tech industries are semiconductors and related elec­ the previous year to the fourth quarter of the year specified. The capacity utiliza­ tronic components, computers and peripheral equipment, and communications tion rates are for the last quarter of the year. equipment. The difference between the revised and earlier rates of change for produc­ tion and capacity for 2005 is calculated for the period 2004:Q4 to 2005:Q3. The difference in capacity utilization for 2005 refers to 2005:Q3. 2002, 2003, and 2005 are stronger. The increase in comprehensive source data and improved methods output in 2004 is the same as previously reported.3 for compiling a few series. The new annual source As of the third quarter of 2005, industrial capacity data are generally for 2003 and 2004, and the modi­ utilization—the ratio of production to capacity—was fied methods affected indexes largely from 1997 79.8 percent, a little higher than previously stated but forward. 1.2 percentage points below the 1972–2004 average.4 The statistical revisions to the IP index were For the fourth quarter of 2004, capacity utilization derived principally from the inclusion of information was revised up 0.6 percentage point, to 79.4 percent; contained in annual reports issued by the U.S. Census upward revisions to the operating rates for manufac­ Bureau—namely, the 2003 Annual Survey of Manu­ turing and mining more than offset a downward revi­ factures, the revised 2002 Census of Manufactures, sion to the operating rate for utilities.5 and selected 2004 Current Industrial Reports. New The revision indicates that industrial capacity government source data included annual data on min­ increased at a faster rate in 2001, 2002, and 2005 erals for 2003 and 2004 from the U.S. Geological than reported earlier. The rate of change for 2003 was Survey (USGS) and updated deflators from the unaltered, and the current estimates for 2004 point to Bureau of Economic Analysis. Also, the new monthly a slower expansion than reported earlier. production estimates reflect updated seasonal factors The updated measures of production and capacity and include monthly source data that became avail­ reflect the incorporation of newly available, more­ able, or were revised, after the closing of the regular four-month reporting window. The capacity indexes and capacity utilization rates 3. For 2005, the rates of change are calculated from the fourth incorporate the revised production indexes; results quarter of 2004 to the fourth quarter of 2005, but any comparisons from the Census Bureau’s 2004 Survey of Plant made between the current industrial production data and the prerevision data are based on annualized rates of change from the fourth Capacity for the fourth quarter of the year; and newly quarter of 2004 to the third quarter of 2005. available data on industrial capacity from the USGS, 4. These comparisons use quarterly average data. the Energy Information Agency of the Department of 5. The fourth quarter of 2004 is the most recent quarter with available survey data on capacity utilization. Energy, and other organizations.

Industrial Production and Capacity Utilization: The 2005 Annual Revision A41 2. Industrial production: Market groups, January 1989–January 2006 Products Ratio scale, 2002 output = 100 Equipment Ratio scale, 2002 output = 100 155 110 Defense and space 135 100 115 95 Consumer goods 90 Business 75 80 Final products and 55 nonindustrial supplies 70 Nonindustrial supplies Industrial materials 110 115 Energy 100 100 Construction 85 90 Business 70 80 Non-energy 55 70 1991 1994 1997 2000 2003 2006 1991 1994 1997 2000 2003 2006 RESULTS OF THE REVISION Industrial Production For the third quarter of 2005, total industrial produc­ The revision indicated that the overall path of indus­ tion was reported to be 108.0 percent of output in trial production was much the same as stated earlier. 2002, and capacity stood at 135.3 percent of output in The most significant features of this revision—the 2002. Both indexes are higher than reported previ­ incorporation of the 2003 Annual Survey of Manu­ ously. The capacity utilization rate for total industry, factures and of the revision to the 2002 Census of at 79.8 percent, was slightly higher than earlier Manufactures—had little effect on the top-line esti­ reports indicated. Results of the revision can be found mates. From 1992 through 2000, total IP increased in the appendix tables.6 at an average annual rate of 4.7 percent. The index declined 3.5 percent in 2001. After the trough, IP registered moderate gains in 2002, 2003, and 2005 and advanced strongly in 2004 (figure 1). 6. Table A.1 shows the revised data for total industrial production, and table A.2 shows the revised data for capacity and capacity utilization for total industry. Tables A.3 and A.4 show the revised rates of change (fourth quarter to fourth quarter) of industrial pro­ Market Groups duction for market groups, industry groups, special aggregates, and selected detail for the years 2001 through 2005. Table A.5 shows the revised rates of change of annual industrial production indexes for Among major market groups (figure 2 and table A.3), market and industry groups for the years 2001 through 2005. the revision shows little change compared with pre­ Tables A.6 and A.7 show the revised figures for capacity and capacity utilization. Tables A.3, A.4, A.5, and A.6 also show the difference vious estimates in production for final products and between the revised and earlier rates of change. Table A.7 shows the nonindustrial supplies for recent years. This index difference between the revised and previous rates of capacity utiliza­ declined in 2001, posted moderate gains in 2002 and tion for the final quarter of the year. Table A.8 shows the annual proportions of market groups and industry groups in total IP. 2003, and increased more rapidly in 2004 and 2005

A42 Federal Reserve Bulletin 2006 (measured from the fourth quarter of the preceding The increases in 2003, 2004, and 2005 were slightly year to the fourth quarter of the year indicated). greater than earlier estimates. Overall, the revisions to consumer goods were small Compared with the previous estimates, the revision for recent years, and the index still shows a general to durable manufacturing shows a larger rate of climb since having dropped back in 2001. The revi­ decline in 2001 and a more rapid rate of increase sions to most of the components of consumer goods since then. The overall contour of this index shows were small; however, new data yielded a sizable solid gains for the past few years—more than 5 per­ downward revision to home electronics for 2003, cent per year, on average, from 2002 forward. All which resulted in less of an increase than stated major durable goods industries posted gains in 2004, earlier. and many showed continued strength in 2005. The The production of business equipment is now only major industries with notable decreases in pro­ reported to have been somewhat weaker, on balance, duction in 2005 were primary metals and furniture in the 2001–05 period than estimated earlier; a down­ and related products. ward revision to the index for information processing Overall, the index for nondurable manufactur­ equipment is largely responsible for the lower 2003 ing was little changed from earlier estimates. The estimate. The output of business equipment declined output of nondurables declined markedly in 2001 and in 2001 and 2002 but has risen since then. The advanced strongly in 2004; the swings in other recent production of defense and space equipment has years were less pronounced. On balance, the output increased in recent years, particularly in 2004 and indexes for textile and product mills and for apparel 2005; however, the revision indicates smaller gains in and leather have registered sharp declines in recent 2002 and 2005 than stated previously and a larger years, whereas the indexes for food, beverage, and gain in 2004.7 The indexes for construction supplies tobacco products; chemicals; and plastics and rubber and business supplies were revised little for recent products all posted gains. years; output in these two market groups dropped The revision indicated lower output in recent years in 2001 but has risen in each year since. From 2002 for the industries that have historically been defined through 2004, the gains in the output of materials as manufacturing (namely, publishing and logging) were revised up about 0.7 percentage point per year; but that are classified elsewhere under the North 2005 had a smaller upward revision, and 2001 had a American Industry Classification System (NAICS). downward revision of 1⁄ 2 percentage point. Produc­ In 2003 and 2004, the rates of change are now, on tion of materials has advanced in every year from average, about 2 percentage points lower than previ­ 2002 on; the largest increases were in 2002 and 2004. ously published. The production of non-energy materials was revised up, overall, in the 2001–05 period. Within this group, the index for durable materials was revised up a bit. Special Aggregates The index for nondurable materials revised up in 2002 and was little changed, on balance, in subse­ A number of special industrial production aggregates quent years. The output of energy materials was are published to help users understand changes in nearly unchanged for 2001–03; the rate of change the industrial sector (table A.4). The high-technology was revised up a bit in 2004 and down a similar industries are important contributors to growth in the amount in 2005. manufacturing sector (figures 3 and 4). The revision shows little change to the aggregate for selected high-technology industries—computers and periph­ Industry Groups eral equipment, communications equipment, and semiconductors and related electronic components. Relative to earlier reports, the current estimates for The small revisions to the aggregate, however, mask manufacturing IP indicate a slightly steeper upward somewhat larger revisions to each of its lower-level trajectory for 2002 through 2005 (table A.3). Like the industries. The production of computers and periph­ revisions to total industrial production, the current eral equipment is now estimated to have declined in estimates for manufacturing IP show a marginally 2002, rather than to have increased, and to have risen larger decline in 2001 and a faster increase in 2002. less in 2003 and 2004 than was reported earlier; the output gain in 2005 is nearly the same as estimated earlier. For communications equipment, the revision shows a steeper drop in production in 2002 and a 7. For 2005, the third quarter of 2005 was used to calculate the extent of the revisions. more moderate recovery in 2003 than previously

Industrial Production and Capacity Utilization: The 2005 Annual Revision A43 3. Industrial production: Manufacturing, and manufacturing 4. Industrial production: Selected high-technology excluding selected high-technology industries, industries, January 1999–January 2006 January 1989–January 2006 Ratio scale, 2002 output = 100 Level Ratio scale, 2002 output = 100 275 225 115 Communications equipment Excluding selected 175 105 high-technology 125 industries 95 Computers 100 85 75 Manufacturing 75 50 Semiconductors 35 65 1999 2000 2001 2002 2003 2004 2005 2006 NOTE: Refer to general note in chart 3. Change from year earlier Percent Manufacturing rapidly in the second half of the 1990s, then deceler­ 10 ates through 2004 (with a small decline in 2003), and picks up moderately in 2005 (figure 1). The expan­ 5 sion in 2004 is now reported to have been less than estimated earlier, and the increase in 2005 is some­ + 0_ what stronger (table A.6). Among selected hightechnology industries, the revision suggests a slower Excluding selected expansion of capacity in 2003 and 2004 than reported high-technology industries 5 earlier; however, for 2005, high-technology capacity expanded more rapidly than stated previously— 1991 1994 1997 2000 2003 2006 20.8 percent. Capacity in mining is now estimated to have NOTE: Manufacturing comprises North American Industry Classification System (NAICS) manufacturing industries (sector 31-33) plus the logging decreased in each of the past four years and has industry and the newspaper, periodical, book, and directory publishing declined, on balance, more than previously estimated. industries. Logging and publishing are classified elsewhere in NAICS (under agriculture and information respectively), but historically they were In contrast, capacity at electric and gas utilities considered to be manufacturing industries and were included in the industrial sector under the Standard Industrial Classification (SIC) system. In December accelerated sharply from 2001 to 2004 and flattened 2002 the Federal Reserve reclassified all its industrial output data from the out in 2005; the current estimates for 2005 are lower SIC system to NAICS. The selected high-technology industries are semiconductors and related than previously reported. electronic components (NAICS 334412–9), computers and peripheral By stage of process, capacity in the crude stage, equipment (NAICS 3341), and communications equipment (NAICS 3342). which has contracted since 2001, is now estimated to have been somewhat weaker, on balance, in the stated; output is now estimated to have risen more 2001–05 period than reported earlier. Relative to rapidly in 2004 and 2005. The index for semiconduc­ previous reports, the capacity index for the primary tors has risen rapidly since a small decline in 2001. and semifinished stages increased more in 2001, Compared with previous estimates, the revision to 2002, and 2005; declined less in 2003; and increased semiconductors shows a slower increase in 2004 but less in 2004. Among finished goods producers, capac­ a more rapid increase in 2003. ity expanded throughout the 2001–05 period. Rela­ The output gain for motor vehicles and parts is tive to earlier reports, the revised estimates show now reported to have been stronger in 2002 and more acceleration in 2002, 2004, and 2005 and a little 2005—by about 1 percentage point—than was less in 2001; the estimates for 2003 were unrevised. reported earlier. The estimates for other recent years are similar to previous reports. Capacity Utilization Capacity Capacity utilization for total industry was revised up The revision did not change the overall contour of in recent years, but the revisions were relatively manufacturing capacity. Capacity still accelerates small; for the third quarter of 2005, utilization stood

A44 Federal Reserve Bulletin 2006 at 79.8 percent, a rate 0.5 percentage point higher 6. Capacity utilization: Selected high-technology industries, than previous estimates suggested but 1.2 percentage January 1996–January 2006 points below its long-term (1972–2004) average Computers and peripheral equipment Percent (table A.7). Revised The factory operating rate reached 78.5 percent in Earlier the third quarter of 2005 after an upwardly revised 90 reading of 78.2 percent in the fourth quarter of 2004 70 and an unrevised 74.7 percent in the fourth quarter of 2003. Within manufacturing, the current revision 50 places the overall utilization rates in recent years for durable goods manufacturers somewhat higher than previously stated, especially in the fourth quarter of Communications equipment 2004 and the third quarter of 2005. The utilization rates for manufacturers of nondurables were little 90 changed from earlier estimates. Capacity utilization in the other (non-NAICS) manufacturing industries is 70 now lower than previously reported, and the recent increases are smaller than those stated earlier. 50 Among selected high-technology industries, utili­ zation is now reported to have been lower in the Semiconductors and related electronic components fourth quarters of 2002 and 2003 but higher in the fourth quarter of 2004 and the third quarter of 90 2005 (figures 5 and 6). For 2002, a downward revi­ sion to the utilization rate for semiconductors and 70 related electronic components accounts for most of the lower estimate. For 2003, the downward revision 50 is largely attributable to lower utilization in the com­ munications equipment industry. For 2004 and 2005, 1996 1998 2000 2002 2004 2006 higher utilization rates for producers of semiconduc­ tors and related electronic components account for much of the upward revisions. Excluding these high- tively, about 1 percentage point above and below the technology industries, capacity utilization for manu­ index’s long-term average. Improved capacity esti­ facturing is little changed. mates of support activities for mining accounted for Capacity utilization in mining was revised up, much of the recent revisions. The utilization rate for to 88.3 in the fourth quarter of 2004 and to 86.1 in electric and gas utilities in the third quarter of 2005 the third quarter of 2005; these estimates are, respec­ was little changed by the revision and stood at 88.1 percent, a rate 1.3 percentage points above its 5. Capacity utilization: Selected high-technology industries long-term average. and manufacturing excluding selected high-technology industries, January 1989–January 2006 TECHNICAL ASPECTS OF THE REVISION Percent Manufacturing The revision incorporates updated comprehensive excluding selected 95 annual data and revised monthly source data used in high-technology industries the estimation of production, capacity, and utiliza­ 85 tion. As noted earlier, this revision includes informa­ tion drawn from the 2003 Annual Survey of Manu­ 75 factures. This revision also incorporates the 2004 Selected Survey of Plant Capacity, other annual industry high-technology industries 65 reports on output and capacity, recent information on prices, and revised monthly source data on physical 55 products and on labor inputs. Along with the indi­ vidual production series and seasonal factors, the 1991 1994 1997 2000 2003 2006 annual value-added weights used in aggregating the

Industrial Production and Capacity Utilization: The 2005 Annual Revision A45 indexes to market and industry groups were also of the Federal Reserve’s index, the rate of change updated. for the output of semiconductors would have been 33 percentage points lower per year, on average, from 1997 to 2003. Changes to Benchmark Indexes The price indexes used for most components of communications equipment are also constructed by The benchmark indexes for manufacturing—defined the Federal Reserve and were updated in this revi­ for each six-digit NAICS industry as nominal gross sion. Price and production indexes for networking output divided by a price index—were updated to equipment (routers, switches, and hubs) are discussed include new information from the 2003 Annual later. Among non-networking equipment, industry Survey of Manufactures (ASM) and revisions to the and government sources on prices for central office estimates from the 2002 Census of Manufactures. equipment, fiber-optic equipment, PBX (private The benchmark indexes for most industries incor­ branch exchange) equipment, consumer telecom­ porated updated price indexes from the industry out­ munications equipment, and wireless infrastructure put program of the Bureau of Economic Analysis equipment were used to extend previous work (BEA). One exception is the benchmark index for through 2004.9 semiconductor manufacturing, which comprises five Another change to the benchmark indexes for 2003 subindustry indexes (figure 7). The price indexes for involved the recalculation of nominal gross output. these subindustry indexes are constructed from infor­ Before the 2004 annual revision, nominal gross out­ mation issued by trade associations, private research put for an industry was calculated as the cost of companies, company reports, and producer price materials plus value added. In the 2004 revision, the indexes from the U.S. Bureau of Labor Statistics.8 benchmark index from 1997 to 2002 was instead Because the overall benchmark price index for semi­ calculated with gross output defined as the traditional conductor manufacturing falls faster than the BEA figure less the cost of resold goods (those goods index, the Federal Reserve’s measure of real output purchased by a manufacturer and then resold without in this industry rises faster than the comparable BEA any material transformation). The 2003 ASM, how­ measure. If the BEA price index were used in place ever, did not include detailed data for the cost of resold goods; in this revision, the nominal gross 8. Carol Corrado (2003), ‘‘Industrial Production and Capacity output used to compute the changes in the benchmark Utilization: The 2002 Historical and Annual Revision,’’ Federal indexes for 2003 was calculated with the traditional Reserve Bulletin, vol. 89 (April), pp. 151–76. method.10 The 2003 ASM also featured a reduction in its 7. Annual price index: Semiconductors, 1997–2004 industry detail. Previous ASMs reported results for every six-digit NAICS industry (473 in manufactur­ Ratio scale, 2002 price = 100 ing under the 2002 NAICS). For 2003, however, the 1,525 reports combined 239 of those six-digit industries 1,025 into 88 higher-level industry aggregates. The bench­ mark indexes for manufacturing IP are calculated 525 from gross output for six-digit industries and then 300 aggregated to the IP industry level (210 industries, each one comprising one or more six-digit NAICS 150 industries) using proportions based on value added. To maintain benchmark references that were consis­ tent over time, the Federal Reserve imputed estimates 50 of gross output and value added for those six-digit industries that were dropped from the ASM. For each 1997 1998 1999 2000 2001 2002 2003 2004 unreported industry, the imputations were based on values for the aggregate industry that contained it and Year Index Percent change the shares of the unreported industry in the aggregate 1997 ................. 1,661.69 −42.00 1998 ................. 849.73 −48.86 in 2002. 1999 ................. 485.02 −42.92 2000 ................. 284.08 −41.43 2001 ................. 148.48 −47.73 9. Corrado, ‘‘Industrial Production and Capacity Utilization,’’ 2002 ................. 100.00 −32.65 p. 159. 2003 ................. 72.21 −27.79 2004 ................. 60.63 −16.04 10. The annual changes in the benchmark indexes between 1997 and 2002 continued to be calculated as they were in the 2004 revision.

A46 Federal Reserve Bulletin 2006 The loss of the detail in the ASM probably had based on manufacturers’ shipments (the implicit only a small effect on the IP industry benchmarks. assumption being that the factory inventories do not All but eight of the IP industries are made up entirely change). In this revision, the procedure that was of industries or industry combinations that were introduced in the 2004 annual revision for estimat­ reported in the ASM. Because the other eight indus­ ing inventories of the machine tool industry was tries, which constitute about 3 percent of total IP, are expanded to the following twenty-one industries all six-digit NAICS industries, any effects of the (with a total weight in IP of 6.6 percent in 2004) for imputation of nominal gross output mostly cancel out which shipments are the high-frequency indicator: at the five-digit NAICS industry level. 1. corn syrup and starch (NAICS 311221 pt.) 2. reconstituted wood products (321219) Changes to Individual Production Series 3. paperboard containers (32221) 4. pharmaceutical preparations (325412) With this revision, the monthly production indicators 5. cement (32731) for some series have changed. The source data for 6. aluminum foundries (331521,4) production indexes for the following twenty indus­ 7. metal cans, boxes, and other metal containers tries, which constituted 10.9 percent of IP in 2004, (light gauge) (33243) have been changed from electric power use to 8. burners and other parts (333414 pt.) production-worker hours for 1997 to the present: 9. boilers, heaters, and furnaces (333414 pt.) 10. warm air furnaces (333415 pt.) 1. fruit and vegetable preserving and specialty food 11. electron tubes (334411) (NAICS 3114) 12. electric housewares and household fans (335211) 2. other food (31193-9) 13. household vacuum cleaners (335212) 3. leather and hide tanning and finishing (3161) 14. electric water heaters (335228 pt.) 4. printing (32311) 15. gas water heaters (335228 pt.) 5. soap, cleaning compound, and toilet preparations 16. storage batteries (335911) (3256) 17. truck trailers (336212) 6. other chemical product and preparations (3259) 18. motor homes (336213) 7. rubber and plastics hoses and belting (32622) 19. travel trailers and campers (336214) 8. other rubber products (32629) 20. mattresses (33791) 9. other pressed and blown glass and glass-ware 21. book publishers (51113) (327212) 10. glass product made of purchased glass (327215) The model underlying the estimates of invento­ 11. lime (32741) ries assumes that manufacturers target a specific 12. other nonmetallic mineral products (3279) inventory–sales ratio. In response to surprises in 13. ferrous metal foundries (33151) demand, manufacturers are assumed to adjust produc­ 14. hardware (3325) tion plans to partially accommodate the surprise in 15. machine shops; turned product; and screws, nuts, the same period, and the remainder of the accommo­ and bolts (3327) dation takes place in subsequent periods. Figure 8 16. air purification equipment, fans, and blowers illustrates the high and low frequency effects of the (333411, 2) model-based inventory adjustment on the shipments 17. electric lighting equipment (3351) indicators. As a representative example, seasonally 18. electrical equipment (3353) adjusted shipments of cement are shown as the thick 19. other electrical equipment (33593, 9) green line in the top panel. The shipments indicator 20. office furniture (including fixtures) (3372) adjusted for inventory change—the new production indicator shown as the thin black line—has the same The decision to switch the monthly indicators for basic contour as shipments, with some of the extreme these series resulted from deterioration in the sample swings tempered. Similarly, shipments of truck trail­ of electric utilities that report power use for these ers, with and without the model-based inventory industries. The IP indexes no longer contain any adjustment, are shown in the bottom panel. From series that are based on electric power use as their the fourth quarter of 2000 through mid-2002, the monthly indicator for the period 1997 to the present. new production indicator is lower than shipments as The IP indexes based on product data usually manufacturers try to keep their inventories in line reflect measures of production, but some have been with slumping demand.

Industrial Production and Capacity Utilization: The 2005 Annual Revision A47 8. Estimated production and shipments: Cement and truck ment from Semiconductor Equipment and Materials trailers, January 1999–December 2002 International, an industry association; the billings data are deflated by the producer price index for the Cement Millions of metric tons industry. The indicators for bare printed circuit boards Estimated production Shipments 9.8 (NAICS 334412) and for printed circuit assemblies 9.6 (NAICS 334418) are now constructed from a 9.4 weighted average of shipment indexes of flexible and 9.2 rigid circuit boards that is deflated by a producer price index.11 The output indicator for boats (NAICS 9.0 336612) is derived from data from the National 8.8 Marine Manufacturers Association for the period 8.6 from 2002 forward; a Fisher quantity index is com­ 8.4 puted from unit shipments and values for the follow­ 8.2 ing eleven types of watercraft: Truck trailers Ratio scale, 2002 output = 100 1. cruisers 2. skiboats 250 3. inflatables 200 4. fiberglass outboard boats 5. aluminum outboard boats 150 6. total sterndrive boats 7. canoes 8. personal watercraft 100 9. jet drive boats 10. boat trailers 11. kayaks 60 The new product-based production indexes consti­ 1999 2000 2001 2002 tuted 1.1 percent of IP in 2004. Table 2 summarizes NOTE: Estimated production measures for cement and truck trailers are the type of data (measured as a percentage of value calculated from shipments adjusted for model-based changes in inventory. Truck trailer data comprise three categories: trailers, containers, and chassis. added in 2004) available in each month of the four- Shipments of each category are weighted by relative prices and summed; the month IP publication window. The first estimate of sum is then indexed. output for a month is preliminary and is subject to This revision also incorporated new indicators revision in each of the subsequent three months as based on product shipments for four industries. new source data become available. By the third revi­ Previously, these industries were estimated from sion (the fourth month of estimate), the productproduction-worker hours. A model-based estimate based content of IP is 51.8 percent. of the change in inventories is added to the index The revision incorporated refined methods for a derived from shipments to compute a production few series. The production indicator for boilers, heat­ index for the following four industries: ers, and furnaces (NAICS 333414 pt.) is now based 1. semiconductor manufacturing equipment (NAICS 11. The shipments indexes are from IPC, an association of electron­ 333295) ics manufacturers that emphasizes the importance of printed circuit boards. 2. bare printed circuit boards (334412) 3. printed circuit assemblies (334418) 4. boats (336612) 2. Proportion of industrial production data by type available in reporting window, 2004 The total number of individual output indexes that Percentage of value added make up industrial production remains at 300 for the Month of estimate period from 1992 forward. Type of data 1st 2nd 3rd 4th With this revision, the monthly indicator for semiconductor manufacturing equipment (NAICS Product-based ................... 24.8 40.4 50.7 51.8 Production-worker hours ......... 44.9 44.9 44.9 44.9 333295) from 1992 and forward was derived from Total available ................... 69.7 85.3 95.6 96.7 Federal Reserve estimates ........ 30.3 14.7 4.4 3.3 data on billings for front-end semiconductor equip­

A48 Federal Reserve Bulletin 2006 on the quarterly output of boilers (data from the Gas production indicator for IP. A correction factor to Appliance Manufacturers Association). The data for align this indicator with the benchmark index is then heaters and furnaces that were previously used to applied. construct the index are no longer available. The production indicator for machine tools (NAICS 333512, 3) is computed based on a Fisher Networking Equipment index of quarterly shipments data from the Census Bureau’s Current Industrial Report on metalworking The Federal Reserve updated estimates for commu­ machinery (MQ333W). The report has nominal ship­ nications equipment manufacturing (NAICS 3342) ments and unit shipments for numerous categories from 2001 on. The 2000 revision introduced a new of machine tools. Previously, the implicit unit values IP series for the production of networking equipment from the report were used as the deflators in the (routers, switches, and hubs).12 The series is not Fisher index calculation. Because the recent data published in the monthly statistical release, but it is for some of these unit values were highly volatile, a included in the broader IP aggregate for communica­ heterogeneity problem in the categories may exist. tions equipment and updated on an ongoing basis. In this revision, the unit values were replaced by Table 3 reports the overall price index for communi­ producer price indexes in the Fisher index calculation cations equipment manufacturing, the price index for for the following categories of machine tools: lathes; networking equipment and service provider routers, grinding, polishing, buffing, honing, and lapping and the price index for all other communications machines; machining centers; miscellaneous metal equipment. cutting machines valued at more than $3,025; punch­ Price indexes were constructed for all types of ing, shearing, bending, and forming machines; networking equipment from detailed data available presses valued at more than $3,025; and miscella­ from Gartner and from the Telecommunications neous metal forming machines. The following cate­ Industry Association (TIA). Estimates of the annual gories still use unit values as deflators: gear cutting value of U.S. production by equipment type were machines, milling machines, miscellaneous metal developed using information from Current Industrial cutting machines valued at less than $3,025, and Reports issued by the U.S. Census Bureau and from presses valued at less than $3,025. consultation with industry analysts. The procedure for estimating the index for civilian As usual, the annual revision incorporated source aircraft was changed to better capture the occasional data on networking equipment and service provider severe disruptions to production that result from labor routers. In addition to detailed information on routers actions. Civilian aircraft production is estimated from data on deliveries of large commercial aircraft (which 12. Carol Corrado (2001), ‘‘Industrial Production and Capacity account for most of the total value of commercial Utilization: The 2000 Annual Revision,’’ Federal Reserve Bulletin, vol. 87 (March), pp. 132–48. The improvements to industrial produc­ aircraft produced in the United States) and data on tion outlined in Corrado (2001) were based on research documented production-worker hours for the aircraft assembly by Mark Doms and Chris Forman (2005), ‘‘Prices for Local Area industry (NAICS 336411). Every large plane com­ Network Equipment,’’ Information Economics and Policy, vol. 17 (July), pp. 365–88. pleted in a month is the result of production activity in that month and in a number of earlier months. A preliminary estimate of production is made by assum­ 3. Price indexes for communications equipment manufacturing, 1997–2004 ing that the production embodied in a plane occurred in the month it was delivered and in the nine previous 2002 price = 100 months, with the progress toward completion higher Local area network in the last few months before the plane is completed. Other equipment Year Total communications Previously, the preliminary estimate of output for a and service equipment provider period was derived entirely from the delivery levels routers or scheduled deliveries for planes in subsequent 1997 ................... 157.8 310.4 134.8 periods. With this revision, the preliminary estimate 1998 ................... 142.9 223.9 128.8 1999 ................... 129.9 183.3 119.2 of output also assumes that the amount of progress 2000 ................... 119.9 163.0 111.0 2001 ................... 109.6 128.2 105.5 that occurred in any one month is dependent on the 2002 ................... 100.0 100.0 100.0 relative production-worker hours in that month. The 2003 ................... 92.8 75.2 97.4 2004 ................... 86.8 57.0 95.5 implicit productivity series from the preliminary esti­ Memo mates is smoothed and then applied to the production- Average percent change, worker hour series to determine the estimate of the 1997–2004 ........ −8.2 −21.5 −4.8

Industrial Production and Capacity Utilization: The 2005 Annual Revision A49 9. Quarterly price indexes: Networking equipment, 2001–04 equipment account for about two-thirds of worldwide sales of fiber-optic equipment. Because data on prices Ratio scale, 2001:Q1 price = 100 of wave division multiplex (WDM) equipment were not available this year, the price index for SONET 100 equipment was applied to WDM equipment (access 80 multiplexers are covered under access systems). Routers PBX equipment was an area of particular focus in 60 this revision because of the ongoing transition away from circuit-switched PBX equipment and toward Switches 40 equipment based on the internet protocol (IP-PBX equipment), which handles voice-over-IP activity. Wireless networking equipment 30 Data on prices and quantities from Gartner for six­ teen types of equipment were used to extend this index. Figure 10 shows price indexes for IP-PBX 2001 2002 2003 2004 equipment, circuit-switched PBX equipment, and Average annual percent change, 2001–04 PBX systems that use intermediate technologies. Prices for IP-PBX equipment fell, on average, Networking equipment ................................. −23.6 Routers ............................................. −24.1 5.8 percent per year from 1999 to 2004, somewhat Switches ............................................ −22.5 Wireless networking equipment ....................... −31.4 faster than the average of nearly 4 percent for circuitswitched PBX equipment and the average of almost 5 percent for intermediate technologies. and switches, this year’s estimates include new data For other types of communications equipment, on wireless networking equipment from Gartner start­ Gartner data were used to construct newly introduced ing in 2001. Products covered include networking or updated price indexes for access systems, such as switches and controllers, add-on adapters for wireless modems and access multiplexers, ATM (asynchro­ networking, and wireless access points. Figure 9 nous transfer mode) equipment, frame relay equip­ shows quarterly price indexes for routers, switches, ment, and ISDN (integrated services digital network) and wireless networking equipment. equipment. Also, Gartner data were used to construct price indexes for the voice-processing equipment and automatic call distributors for the 1992–2004 period. Other Communications Equipment Information was drawn from the TIA’s 2005 Mar­ ket Review and Forecast to generate price indexes for The Federal Reserve also updated price and quantity indexes for other types of communications equip­ ment using detailed information found in reports 10. Annual price indexes: Private branch exchange (PBX) from Gartner and the TIA. The types of equipment equipment, 1999–2004 covered included fiber-optic equipment, PBX equip­ ment, voice processing equipment, and network Ratio scale, 1999 price = 100 access systems. 100 The price index for fiber-optic equipment, intro­ duced with the 2002 IP revision, was extended this Intermediate technologies 95 year.13 Information from Gartner on five classes of 90 synchronous optical network (SONET) equipment Circuit switched and three types of digital cross-connect equipment 85 were used to create price indexes for these products. During the 2000–04 period, prices for SONET equip­ 80 ment fell 11 percent and prices for digital cross- 75 connect equipment fell 5 percent. These types of Internet protocol 13. The price indexes for communications equipment introduced 1999 2000 2001 2002 2003 2004 in the 2002 revision were based on work described in Mark Doms Average annual percent change, 1999–2004 (2005), ‘‘Communications Equipment: What Has Happened to Prices?’’ in Carol Corrado, John Haltiwanger, and Daniel Sichel, eds., PBX equipment ........................................ −4.3 Measuring Capital in the New Economy, National Bureau of Eco­ Circuit switched ...................................... −3.9 Internet protocol ..................................... −5.8 nomic Research Studies in Income and Wealth (Chicago, University Intermediate technologies ............................. −4.8 of Chicago Press), pp. 323–62.

A50 Federal Reserve Bulletin 2006 consumer telecommunications equipment, office fac­ industry capacity utilization (taken from the SPC) simile equipment, and ATM equipment. and measures of industry capital input constructed The remaining types of equipment without specific from data in the ASM. information on prices and quantities were assigned appropriate producer price indexes. These included broadcast studio equipment, alarm systems, vehic­ Weights for Aggregation ular and pedestrian signal equipment, and paging equipment. The IP index is a Fisher index. This revision uses information from the Census of Manufactures to obtain updated estimates of the industry value-added Discontinuance of the Survey weights used in the aggregation of IP indexes and of Electric Power Use capacity utilization rates. The Federal Reserve derives estimates of value added for the electric The Federal Reserve discontinued its use of survey and gas utility industries from annual revenue and data of industrial electric power use. The survey was expense data issued by other organizations. The initiated in the 1960s to collect data on electricity use weights for aggregation, expressed as unit value as alternative indicators of output; electricity use added, were estimated for recent years using the tends to be highly correlated with production in latest data on producer prices. Table A.8 shows capital-intensive industries. Survey data were used the annual value-added proportions incorporated in directly as the production indicators for industries the IP index from 1997 through 2004. (particularly those with highly automated assembly operations and a diverse product mix) for which product-based data were unavailable. However, by Revised Monthly Data 2004, the electric power usage covered by the survey had dropped nearly 40 percent from its peak in 1993, This revision incorporates product data that became and several regions and industries had significant available after the regular four-month reporting win­ gaps. The surveys, conducted by the Federal Reserve dow for monthly IP closed. For example, the data on District Banks, had especially poor response rates tobacco issued by the Department of the Treasury’s for District 6 (Atlanta) and District 11 (Dallas). The Alcohol and Tobacco Tax and Trade Bureau are industries with a substantial presence in those Dis­ released with too great a lag to be included with tricts include textiles and chemicals. As noted before, monthly IP estimates; however, the data are available the twenty industry indexes that relied on electric for inclusion in the annual revision. power use now have production-worker hours as the underlying source data for the period 1997 to the present. Revised Seasonal Factors In the future, the Federal Reserve hopes to incorpo­ rate information from the Census Bureau’s Survey of Seasonal factors for all series were reestimated using Plant Capacity (SPC) into its estimation of produc­ data that extended into 2005. Factors for productiontion indexes for capital-intensive industries. The SPC worker hours—which adjust for timing, holiday, and collects data on the factory workweek; like electric monthly seasonal patterns—were updated with data power use, the workweek is an indicator of the level through September 2005 and were prorated to corre­ of operations in capital-intensive industries. Cur­ spond with the seasonal factors for hours aggregated rently, the SPC data are collected only in the fourth to the three-digit NAICS level. The updated factors quarter of the year, so they are insufficient for use as for the physical product series, which include adjust­ high-frequency indicators. ments for holiday and workday patterns, used data Beyond their use as a direct monthly indicator, through at least June 2005. Seasonal factors for unit electric power data were used in a model that extrapo­ motor vehicle assemblies have been updated, and lates productivity in industries for which the monthly projections through June 2006 are on the Board’s output index was based on production-worker hours. website at www.federalreserve.gov/releases/g17/ In the absence of electric power data, the model now mvsf.htm. relies on a proxy derived from survey measures of Appendix tables start on page A51

Industrial Production and Capacity Utilization: The 2005 Annual Revision A51 APPENDIX A: TABLES BASED ON THE G.17 RELEASE, FEBRUARY 15, 2006 A.1. Revised data for industrial production for total industry Seasonally adjusted data except as noted Quarter Annual Year Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec. avg.1 1 2 3 4 Industrial production (percent change) 1975 ........ −1.3 −2.4 −1.1 .1 −.2 .7 1.1 1.0 1.3 .4 .3 1.2 −24.0 −5.3 10.5 8.9 −8.9 1976 ........ 1.5 .9 .1 .7 .4 .0 .6 .7 .2 .1 1.5 1.0 12.6 5.2 5.1 7.8 7.8 1977 ........ −.6 1.5 1.3 .9 .8 .7 .3 .1 .5 .3 .0 .2 8.4 12.6 4.9 2.8 7.7 1978 ........ −1.3 .4 1.8 2.1 .3 .7 .0 .3 .2 .9 .8 .6 −1.3 16.6 3.4 7.7 5.5 1979 ........ −.7 .6 .3 −1.0 .7 .0 −.2 −.7 .1 .5 −.1 .1 1.9 −.5 −1.5 1.3 3.1 1980 ........ .4 .0 −.3 −2.0 −2.5 −1.3 −.7 .3 1.6 1.2 1.7 .6 1.6 −15.9 −6.3 16.1 −2.6 1981 ........ −.6 −.5 .6 −.5 .7 .5 .7 .0 −.6 −.8 −1.1 −1.1 .9 1.4 4.2 −8.6 1.3 1982 ........ −1.9 1.9 −.7 −.9 −.7 −.4 −.4 −.9 −.4 −.8 −.4 −.8 −7.6 −5.1 −5.9 −7.5 −5.1 1983 ........ 1.9 −.6 .9 1.2 .8 .6 1.6 1.1 1.5 .9 .3 .5 4.3 9.4 14.7 10.9 2.6 1984 ........ 2.1 .5 .5 .6 .5 .4 .3 .1 −.2 −.1 .4 .1 12.4 6.4 2.9 .4 9.0 1985 ........ −.3 .4 .2 −.2 .1 .0 −.6 .5 .4 −.4 .3 1.0 1.1 .6 −.6 2.5 1.3 1986 ........ .5 −.8 −.6 .0 .2 −.3 .6 −.2 .2 .4 .5 .9 2.3 −2.4 1.7 4.5 1.0 1987 ........ −.3 1.2 .2 .6 .7 .5 .6 .7 .3 1.5 .5 .5 5.4 7.0 7.1 9.9 5.1 1988 ........ .1 .3 .3 .5 .0 .2 .2 .5 −.3 .6 .2 .4 3.4 3.2 2.1 3.3 5.0 1989 ........ .3 −.5 .3 −.1 −.7 .0 −.9 .9 −.3 −.1 .3 .7 1.5 −1.9 −2.5 1.8 .9 1990 ........ −.6 .9 .5 .0 .1 .3 −.1 .3 .2 −.7 −1.2 −.7 2.9 2.9 1.3 −5.9 .9 1991 ........ −.5 −.7 −.5 .2 1.0 1.0 .0 .2 .9 −.2 −.1 −.3 −7.6 2.7 5.7 1.0 −1.5 1992 ........ −.5 .7 .7 .7 .4 .0 .8 −.5 .2 .7 .4 .1 −.3 6.9 2.8 3.9 2.9 1993 ........ .5 .3 .0 .3 −.4 .3 .3 .0 .5 .7 .4 .5 3.7 1.0 2.4 6.1 3.3 1994 ........ .5 .0 1.0 .5 .6 .7 .2 .5 .2 .9 .6 1.1 5.3 7.5 5.3 7.9 5.4 1995 ........ .3 .0 .2 .0 .2 .3 −.4 1.4 .4 −.2 .3 .4 5.3 1.1 3.7 3.5 4.8 1996 ........ −.8 1.5 −.2 .9 .7 .8 −.2 .7 .6 .0 1.0 .7 1.8 8.4 5.1 6.2 4.2 1997 ........ .2 1.2 .8 −.1 .6 .5 .5 1.2 .9 .7 .9 .4 8.5 5.7 8.9 10.6 7.3 1998 ........ .4 .1 .1 .5 .6 −.5 −.3 2.2 −.2 .7 −.1 .3 4.4 3.1 3.6 5.6 5.9 1999 ........ .5 .5 .2 .2 .8 −.1 .7 .5 −.4 1.3 .6 .9 4.1 4.2 4.3 7.7 4.5 2000 ........ .2 .4 .4 .7 .3 .1 −.3 −.3 .4 −.3 .0 −.4 5.4 5.2 −.9 −1.3 4.3 2001 ........ −.8 −.7 −.3 −.1 −.8 −.5 −.4 −.3 −.4 −.5 −.4 .0 −6.6 −4.9 −5.1 −4.5 −3.5 2002 ........ .6 −.1 .9 .4 .4 .9 −.3 .1 .0 −.4 .3 −.5 2.9 6.1 1.7 −1.6 .1 2003 ........ .5 .0 −.2 −.8 −.1 .3 .5 .1 .7 .1 .9 .2 .8 −3.3 3.6 5.1 .6 2004 ........ .4 .8 −.3 .8 .9 −.6 .6 .3 −.2 .7 .2 .7 5.3 5.2 2.6 4.2 4.1 2005 ........ .2 .4 .0 −.1 .2 .8 .0 .3 −1.3 1.1 1.1 .9 3.8 1.6 1.4 5.6 3.3 2006 ........ −.2 ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... Industrial production (2002 = 100) 1975 ........ 46.3 45.2 44.7 44.7 44.6 44.9 45.4 45.8 46.4 46.6 46.7 47.3 45.4 44.8 45.9 46.9 45.7 1976 ........ 48.0 48.4 48.4 48.8 49.0 49.0 49.2 49.6 49.7 49.8 50.5 51.0 48.3 48.9 49.5 50.5 49.3 1977 ........ 50.8 51.5 52.2 52.6 53.0 53.4 53.6 53.6 53.9 54.0 54.0 54.1 51.5 53.0 53.7 54.1 53.1 1978 ........ 53.4 53.6 54.6 55.7 55.9 56.3 56.3 56.5 56.6 57.1 57.6 57.9 53.9 56.0 56.5 57.5 56.0 1979 ........ 57.5 57.8 58.0 57.5 57.8 57.8 57.7 57.3 57.4 57.7 57.7 57.7 57.8 57.7 57.5 57.7 57.7 1980 ........ 58.0 58.0 57.8 56.7 55.2 54.5 54.2 54.3 55.2 55.9 56.9 57.2 57.9 55.5 54.6 56.6 56.2 1981 ........ 56.9 56.6 56.9 56.6 57.0 57.3 57.7 57.7 57.3 56.9 56.3 55.7 56.8 57.0 57.6 56.3 56.9 1982 ........ 54.6 55.7 55.3 54.8 54.4 54.2 54.0 53.6 53.3 52.9 52.7 52.2 55.2 54.5 53.6 52.6 54.0 1983 ........ 53.2 52.9 53.4 54.0 54.4 54.7 55.6 56.2 57.0 57.5 57.7 58.0 53.2 54.4 56.3 57.7 55.4 1984 ........ 59.2 59.4 59.7 60.1 60.4 60.6 60.8 60.9 60.8 60.7 60.9 61.0 59.5 60.4 60.8 60.9 60.4 1985 ........ 60.8 61.1 61.2 61.1 61.1 61.2 60.8 61.1 61.3 61.1 61.3 61.9 61.0 61.1 61.0 61.4 61.2 1986 ........ 62.2 61.7 61.4 61.4 61.5 61.3 61.7 61.6 61.7 61.9 62.2 62.8 61.8 61.4 61.6 62.3 61.8 1987 ........ 62.6 63.4 63.5 63.8 64.3 64.6 65.0 65.4 65.6 66.6 66.9 67.2 63.2 64.2 65.3 66.9 64.9 1988 ........ 67.2 67.5 67.6 68.0 67.9 68.1 68.2 68.5 68.3 68.7 68.8 69.1 67.4 68.0 68.3 68.9 68.2 1989 ........ 69.3 69.0 69.2 69.1 68.7 68.7 68.0 68.7 68.5 68.4 68.6 69.1 69.2 68.8 68.4 68.7 68.8 1990 ........ 68.7 69.3 69.6 69.6 69.6 69.8 69.7 69.9 70.0 69.5 68.7 68.2 69.2 69.7 69.9 68.8 69.4 1991 ........ 67.9 67.4 67.1 67.2 67.9 68.6 68.6 68.7 69.3 69.2 69.1 68.9 67.5 67.9 68.9 69.0 68.3 1992 ........ 68.5 69.0 69.5 70.0 70.2 70.3 70.8 70.5 70.6 71.1 71.4 71.5 69.0 70.2 70.7 71.3 70.3 1993 ........ 71.8 72.1 72.1 72.3 72.0 72.2 72.5 72.5 72.8 73.3 73.6 74.0 72.0 72.2 72.6 73.7 72.6 1994 ........ 74.4 74.4 75.1 75.5 76.0 76.5 76.7 77.0 77.2 77.9 78.3 79.2 74.6 76.0 77.0 78.5 76.5 1995 ........ 79.4 79.4 79.6 79.5 79.7 79.9 79.6 80.7 81.0 80.9 81.1 81.4 79.5 79.7 80.4 81.1 80.2 1996 ........ 80.7 82.0 81.8 82.5 83.1 83.8 83.7 84.2 84.7 84.7 85.6 86.1 81.5 83.2 84.2 85.5 83.6 1997 ........ 86.3 87.3 88.0 88.0 88.5 88.9 89.3 90.4 91.3 92.0 92.8 93.2 87.2 88.5 90.4 92.7 89.7 1998 ........ 93.6 93.7 93.7 94.2 94.7 94.2 93.9 96.0 95.8 96.5 96.4 96.7 93.7 94.4 95.2 96.5 94.9 1999 ........ 97.1 97.6 97.8 98.0 98.8 98.7 99.4 99.9 99.4 100.7 101.3 102.2 97.5 98.5 99.6 101.4 99.3 2000 ........ 102.4 102.8 103.1 103.9 104.2 104.2 103.9 103.6 104.0 103.7 103.6 103.2 102.8 104.1 103.8 103.5 103.5 2001 ........ 102.4 101.6 101.3 101.2 100.4 99.9 99.5 99.2 98.8 98.3 97.9 97.9 101.8 100.5 99.2 98.1 99.9 2002 ........ 98.6 98.4 99.3 99.7 100.1 101.0 100.6 100.7 100.7 100.3 100.5 100.1 98.8 100.3 100.7 100.3 100.0 2003 ........ 100.5 100.6 100.4 99.6 99.5 99.8 100.3 100.4 101.0 101.1 102.0 102.3 100.5 99.7 100.5 101.8 100.6 2004 ........ 102.7 103.5 103.2 104.0 105.0 104.4 105.0 105.3 105.1 105.8 106.0 106.7 103.1 104.4 105.1 106.2 104.7 2005 ........ 106.9 107.4 107.3 107.2 107.4 108.3 108.3 108.6 107.2 108.4 109.5 110.5 107.2 107.6 108.0 109.5 108.2 2006 ........ 110.3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Note: Monthly percent change figures show the change from the previous Estimates from November 2005 to January 2006 are subject to further month; quarterly figures show the change from the previous quarter at a revision in the upcoming monthly releases. compound annual rate of change. Production and capacity indexes are expressed 1. Annual averages of industrial production are calculated from not seasonas percentages of output in 2002. ally adjusted indexes. . . . Not available as of February 15, 2006.

A52 Federal Reserve Bulletin 2006 A.2. Revised data for capacity and utilization for total industry Seasonally adjusted data except as noted Quarter Annual Year Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec. avg. 1 2 3 4 Capacity (percent of 2002 output) 1975 ........ 59.8 59.9 60.0 60.2 60.3 60.4 60.5 60.6 60.7 60.8 61.0 61.1 59.9 60.3 60.6 61.0 60.4 1976 ........ 61.2 61.3 61.5 61.6 61.7 61.9 62.0 62.1 62.3 62.4 62.6 62.7 61.3 61.7 62.1 62.6 61.9 1977 ........ 62.9 63.0 63.2 63.4 63.5 63.7 63.9 64.1 64.3 64.5 64.6 64.8 63.0 63.5 64.1 64.6 63.8 1978 ........ 65.0 65.2 65.4 65.6 65.7 65.9 66.1 66.2 66.4 66.6 66.7 66.9 65.2 65.7 66.2 66.7 66.0 1979 ........ 67.0 67.2 67.3 67.5 67.7 67.8 68.0 68.1 68.2 68.4 68.5 68.7 67.2 67.7 68.1 68.5 67.9 1980 ........ 68.8 68.9 69.1 69.2 69.3 69.5 69.6 69.8 69.9 70.0 70.2 70.3 68.9 69.3 69.8 70.2 69.6 1981 ........ 70.5 70.6 70.8 70.9 71.1 71.3 71.4 71.6 71.8 71.9 72.1 72.3 70.6 71.1 71.6 72.1 71.4 1982 ........ 72.5 72.6 72.8 72.9 73.1 73.3 73.4 73.5 73.6 73.7 73.8 73.9 72.6 73.1 73.5 73.8 73.3 1983 ........ 74.0 74.0 74.0 74.1 74.1 74.1 74.2 74.2 74.2 74.3 74.3 74.4 74.0 74.1 74.2 74.3 74.2 1984 ........ 74.5 74.5 74.6 74.7 74.9 75.0 75.1 75.3 75.4 75.6 75.7 75.9 74.6 74.9 75.3 75.7 75.1 1985 ........ 76.1 76.3 76.5 76.6 76.8 77.0 77.2 77.4 77.5 77.7 77.8 77.9 76.3 76.8 77.4 77.8 77.1 1986 ........ 78.1 78.2 78.3 78.4 78.5 78.6 78.6 78.7 78.8 78.9 79.0 79.1 78.2 78.5 78.7 79.0 78.6 1987 ........ 79.3 79.4 79.5 79.7 79.8 79.9 80.1 80.2 80.3 80.4 80.5 80.6 79.4 79.8 80.2 80.5 80.0 1988 ........ 80.6 80.7 80.8 80.8 80.9 80.9 80.9 81.0 81.1 81.1 81.2 81.3 80.7 80.9 81.0 81.2 80.9 1989 ........ 81.4 81.5 81.7 81.8 82.0 82.1 82.3 82.5 82.6 82.8 83.0 83.2 81.5 82.0 82.5 83.0 82.2 1990 ........ 83.3 83.5 83.7 83.9 84.0 84.2 84.3 84.5 84.6 84.8 84.9 85.0 83.5 84.0 84.5 84.9 84.2 1991 ........ 85.2 85.3 85.4 85.5 85.6 85.8 85.9 86.0 86.1 86.2 86.3 86.5 85.3 85.6 86.0 86.3 85.8 1992 ........ 86.6 86.7 86.9 87.0 87.2 87.3 87.5 87.6 87.8 88.0 88.1 88.3 86.7 87.2 87.6 88.1 87.4 1993 ........ 88.4 88.5 88.7 88.8 88.9 89.1 89.2 89.3 89.5 89.6 89.8 90.0 88.5 88.9 89.3 89.8 89.2 1994 ........ 90.1 90.3 90.5 90.8 91.0 91.2 91.5 91.8 92.1 92.4 92.8 93.1 90.3 91.0 91.8 92.8 91.5 1995 ........ 93.5 93.8 94.2 94.6 95.0 95.4 95.8 96.2 96.6 97.0 97.5 97.9 93.8 95.0 96.2 97.5 95.6 1996 ........ 98.3 98.7 99.2 99.6 100.0 100.4 100.9 101.3 101.7 102.2 102.7 103.1 98.7 100.0 101.3 102.7 100.7 1997 ........ 103.6 104.2 104.7 105.2 105.8 106.4 107.0 107.7 108.4 109.0 109.7 110.5 104.2 105.8 107.7 109.7 106.9 1998 ........ 111.2 111.9 112.6 113.3 114.0 114.6 115.2 115.8 116.4 116.9 117.5 118.0 111.9 114.0 115.8 117.5 114.8 1999 ........ 118.5 119.0 119.5 120.0 120.5 120.9 121.4 121.8 122.3 122.8 123.2 123.7 119.0 120.4 121.8 123.2 121.1 2000 ........ 124.2 124.6 125.1 125.5 126.0 126.4 126.8 127.2 127.7 128.1 128.5 128.9 124.6 126.0 127.2 128.5 126.6 2001 ........ 129.2 129.6 130.0 130.3 130.6 130.9 131.2 131.5 131.8 132.0 132.2 132.5 129.6 130.6 131.5 132.2 131.0 2002 ........ 132.6 132.8 132.9 133.0 133.1 133.2 133.2 133.2 133.2 133.2 133.2 133.2 132.8 133.1 133.2 133.2 133.1 2003 ........ 133.1 133.1 133.0 133.0 133.0 132.9 132.9 132.9 132.9 132.9 132.9 132.9 133.1 133.0 132.9 132.9 133.0 2004 ........ 133.0 133.0 133.0 133.1 133.1 133.2 133.3 133.4 133.5 133.6 133.7 133.9 133.0 133.1 133.4 133.7 133.3 2005 ........ 134.0 134.2 134.3 134.5 134.7 134.9 135.1 135.3 135.5 135.7 135.9 136.1 134.2 134.7 135.3 135.9 135.0 2006 ........ 136.3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Utilization (percent) 1975 ........ 77.4 75.3 74.4 74.3 74.1 74.4 75.1 75.6 76.5 76.6 76.6 77.4 75.7 74.3 75.7 76.9 75.6 1976 ........ 78.4 78.9 78.8 79.2 79.3 79.2 79.4 79.8 79.8 79.8 80.8 81.4 78.7 79.2 79.7 80.6 79.6 1977 ........ 80.7 81.7 82.5 83.1 83.5 83.8 83.8 83.7 83.8 83.8 83.6 83.5 81.7 83.5 83.8 83.6 83.1 1978 ........ 82.1 82.3 83.5 85.0 85.1 85.5 85.2 85.3 85.3 85.8 86.3 86.6 82.6 85.2 85.3 86.2 84.8 1979 ........ 85.8 86.1 86.2 85.1 85.5 85.3 85.0 84.2 84.1 84.4 84.1 84.1 86.0 85.3 84.4 84.2 85.0 1980 ........ 84.2 84.1 83.7 81.9 79.6 78.5 77.8 77.9 79.0 79.8 81.0 81.3 84.0 80.0 78.2 80.7 80.7 1981 ........ 80.7 80.1 80.4 79.8 80.2 80.4 80.7 80.6 79.9 79.1 78.1 77.0 80.4 80.1 80.4 78.1 79.7 1982 ........ 75.4 76.6 75.9 75.1 74.4 74.0 73.6 72.9 72.5 71.8 71.4 70.7 76.0 74.5 73.0 71.3 73.7 1983 ........ 72.0 71.5 72.1 72.9 73.4 73.8 74.9 75.7 76.8 77.5 77.6 77.9 71.8 73.4 75.8 77.7 74.7 1984 ........ 79.5 79.7 80.0 80.4 80.7 80.8 80.9 80.9 80.6 80.3 80.5 80.3 79.7 80.7 80.8 80.4 80.4 1985 ........ 79.9 80.1 80.0 79.7 79.6 79.4 78.7 78.9 79.1 78.6 78.8 79.4 80.0 79.6 78.9 78.9 79.4 1986 ........ 79.7 79.0 78.4 78.3 78.4 78.0 78.4 78.2 78.3 78.5 78.7 79.3 79.0 78.2 78.3 78.9 78.6 1987 ........ 79.0 79.8 79.8 80.1 80.6 80.8 81.2 81.6 81.7 82.8 83.1 83.4 79.6 80.5 81.5 83.1 81.2 1988 ........ 83.4 83.6 83.7 84.1 84.0 84.2 84.3 84.6 84.3 84.7 84.8 85.0 83.6 84.1 84.4 84.8 84.2 1989 ........ 85.1 84.6 84.7 84.5 83.8 83.6 82.7 83.3 82.8 82.6 82.7 83.1 84.8 84.0 82.9 82.8 83.6 1990 ........ 82.4 82.9 83.2 83.0 82.9 82.9 82.7 82.8 82.8 82.0 80.9 80.2 82.8 82.9 82.7 81.1 82.4 1991 ........ 79.7 79.1 78.6 78.6 79.3 80.0 79.9 79.9 80.5 80.2 80.0 79.7 79.1 79.3 80.1 80.0 79.6 1992 ........ 79.1 79.5 80.0 80.4 80.6 80.5 81.0 80.4 80.5 80.9 81.0 81.0 79.6 80.5 80.6 81.0 80.4 1993 ........ 81.2 81.4 81.3 81.4 81.0 81.1 81.2 81.1 81.4 81.8 82.0 82.3 81.3 81.1 81.2 82.0 81.4 1994 ........ 82.5 82.3 83.0 83.2 83.5 83.8 83.8 83.9 83.8 84.2 84.5 85.0 82.6 83.5 83.8 84.6 83.6 1995 ........ 85.0 84.7 84.5 84.1 83.9 83.8 83.1 83.9 83.8 83.3 83.2 83.2 84.7 83.9 83.6 83.2 83.9 1996 ........ 82.1 83.0 82.5 82.9 83.1 83.5 83.0 83.2 83.3 82.9 83.3 83.5 82.5 83.2 83.1 83.3 83.0 1997 ........ 83.3 83.9 84.1 83.6 83.6 83.5 83.5 84.0 84.3 84.3 84.6 84.4 83.7 83.6 83.9 84.4 83.9 1998 ........ 84.2 83.7 83.3 83.1 83.1 82.2 81.5 82.8 82.3 82.5 82.1 81.9 83.7 82.8 82.2 82.2 82.7 1999 ........ 82.0 82.0 81.8 81.7 82.0 81.7 81.9 82.0 81.3 82.1 82.2 82.6 81.9 81.8 81.7 82.3 81.9 2000 ........ 82.5 82.5 82.5 82.7 82.7 82.5 82.0 81.4 81.5 80.9 80.7 80.1 82.5 82.6 81.6 80.6 81.8 2001 ........ 79.2 78.4 77.9 77.6 76.9 76.3 75.9 75.4 75.0 74.5 74.0 73.9 78.5 76.9 75.4 74.2 76.3 2002 ........ 74.3 74.1 74.7 74.9 75.2 75.8 75.6 75.6 75.6 75.3 75.5 75.2 74.4 75.3 75.6 75.3 75.1 2003 ........ 75.5 75.6 75.4 74.9 74.9 75.1 75.4 75.5 76.0 76.1 76.8 76.9 75.5 74.9 75.6 76.6 75.7 2004 ........ 77.2 77.8 77.6 78.1 78.8 78.4 78.8 79.0 78.7 79.2 79.3 79.7 77.5 78.4 78.8 79.4 78.6 2005 ........ 79.8 80.0 79.9 79.7 79.8 80.3 80.2 80.3 79.1 79.9 80.6 81.2 79.9 79.9 79.8 80.6 80.1 2006 ........ 80.9 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Note: Refer also to the general note in table A.1. . . . Not available as of February 15, 2006.

Industrial Production and Capacity Utilization: The 2005 Annual Revision A53 A.3. Rates of change in industrial production, by market and industry groups, 2001–051 Difference between rates of change: Revised rate of change revised minus earlier Item NAICS (percent) (percentage points) code 2 2001 2002 2003 2004 2005 2001 2002 2003 2004 2005 Total industry .......................... ... −5.3 2.3 1.5 4.3 3.1 −.2 .8 .3 .0 .2 Market Group Final products and nonindustrial supplies . . . . . −4.6 1.5 1.7 4.3 4.7 .0 .8 .0 −.4 .2 Consumer goods ...................... . . . −1.3 2.7 1.3 2.0 2.3 .2 1.1 .1 −1.0 .5 Durable ............................ . . . −1.3 7.9 4.3 1.3 3.1 .0 1.5 1.0 −.2 .5 Automotive products .............. . . . 2.7 12.4 6.5 .4 2.5 .4 2.3 1.3 −.7 −.9 Home electronics ................. . . . 6.3 −7.7 18.5 −3.7 16.4 .5 −3.7 −16.4 4.4 15.3 Appliances, furniture, carpeting .... . . . −2.9 4.0 2.2 2.4 2.4 .5 2.2 .8 −.6 .5 Miscellaneous goods .............. . . . −7.5 4.9 −.7 3.0 3.2 −.9 .6 2.8 .4 1.1 Nondurable ......................... . . . −1.3 .8 .1 2.3 2.0 .2 1.0 −.3 −1.4 .4 Non-energy ...................... . . . −.5 −1.0 .6 2.5 1.9 .3 1.3 −.2 −1.8 .0 Foods and tobacco ............. . . . −.4 −3.1 2.1 1.8 2.4 .8 .6 −.3 −2.6 .4 Clothing ....................... . . . −21.5 −7.1 −9.3 −2.8 −1.0 −.7 2.6 5.6 2.3 5.1 Chemical products ............. . . . 6.5 4.0 1.0 3.0 .4 −.6 3.1 .4 −1.1 −1.4 Paper products ................. . . . −2.3 −.5 −3.5 5.9 3.5 .4 .3 −4.1 −1.1 .3 Energy ........................... . . . −5.3 9.5 −1.9 1.7 2.4 −.1 −.6 −.5 .7 2.1 Business equipment ................... . . . −13.6 −2.3 2.8 10.8 10.4 −.3 .3 −1.9 1.0 −1.1 Transit ............................. . . . −4.6 −10.8 3.6 9.7 14.1 −1.2 1.9 3.5 −1.2 1.7 Information processing .............. . . . −15.5 −8.9 6.5 14.6 19.4 1.9 −5.2 −9.8 4.6 .3 Industrial and other ................. . . . −15.3 5.4 .4 9.2 4.4 −1.6 3.4 .9 −.2 −2.4 Defense and space equipment .......... . . . 8.4 .8 5.4 9.7 9.2 .4 −3.0 .1 3.6 −1.2 Construction supplies .................. . . . −5.0 1.2 1.7 4.6 6.8 .0 1.1 1.1 .7 −.1 Business supplies ..................... . . . −5.5 2.5 .9 3.9 3.8 .0 1.1 .8 −1.0 .8 Materials ................................ . . . −6.1 3.4 1.2 4.3 .9 −.5 .7 .7 .7 .1 Non-energy ........................... . . . −7.2 4.4 1.7 6.0 3.4 −.7 .9 .9 .5 .7 Durable ............................ . . . −8.3 5.4 3.8 7.6 7.2 −1.0 .9 1.7 .6 1.5 Consumer parts ................... . . . −7.7 6.0 −1.2 2.0 2.1 .2 −1.0 −3.9 −.2 .2 Equipment parts .................. . . . −9.4 7.4 11.8 14.1 16.5 −1.5 1.2 6.0 −2.0 .9 Other ............................ . . . −7.3 3.4 .0 4.9 2.2 −1.1 1.4 .9 2.3 2.1 Nondurable ......................... . . . −5.1 2.7 −1.7 3.2 −3.0 .0 1.0 −.4 .3 −.7 Textile ........................... . . . −10.3 4.9 −6.5 −6.4 −4.7 −.8 2.9 6.5 −1.9 .8 Paper ............................ . . . −6.2 .8 −6.5 4.6 −.9 .1 −1.3 −2.1 1.4 −.8 Chemical ........................ . . . −4.5 4.0 1.6 5.3 −8.2 .0 1.8 −.5 .4 −2.1 Energy ............................... . . . −2.8 .4 −.1 −.2 −5.1 .0 .0 .2 1.0 −1.2 Table A.3. continues on page A54.

A54 Federal Reserve Bulletin 2006 A.3. Rates of change in industrial production, by market and industry groups, 2001–051—Continued Difference between rates of change: Revised rate of change revised minus earlier Item NAICS (percent) (percentage points) code 2 2001 2002 2003 2004 2005 2001 2002 2003 2004 2005 Industry Group Manufacturing 3 ......................... . . . −5.6 2.2 1.7 5.1 4.2 −.2 .9 .2 .1 .2 Manufacturing (NAICS) ............... 31–33 −5.6 2.5 2.0 5.2 4.3 −.3 .9 .4 .1 .2 Durable manufacturing .............. . . . −7.7 3.8 4.0 7.1 7.8 −.5 .7 .7 .5 .6 Wood products ................... 321 −1.3 .9 4.0 3.0 7.8 .4 .9 .8 2.2 4.5 Nonmetallic mineral products ..... 327 −3.4 1.0 2.2 5.1 3.2 −.9 .7 .5 .7 −1.8 Primary metal .................... 331 −9.3 6.6 1.0 3.9 −1.6 −.6 −.4 .3 .6 2.6 Fabricated metal products ......... 332 −10.5 1.2 −.7 5.2 4.1 −1.9 1.5 2.2 2.0 1.2 Machinery ....................... 333 −18.6 4.2 1.0 11.5 6.4 −1.8 2.9 .3 −.4 −3.6 Computer and electronic products . . 334 −8.9 2.9 15.7 16.1 23.3 .1 −2.8 1.2 .9 .9 Electrical equipment, appliances, and components ............. 335 −14.8 −2.2 −.7 5.2 6.8 −.2 3.1 −1.9 −.8 4.5 Motor vehicles and parts .......... 3361–3 −1.8 12.2 4.7 2.6 2.3 .2 1.0 −.1 −.3 .8 Aerospace and miscellaneous transportation equipment ..... 3364–9 5.1 −7.3 −.4 5.3 11.5 .5 .2 −1.2 .8 −1.0 Furniture and related products ..... 337 −6.3 7.3 .3 2.2 −2.0 −.1 3.1 2.1 −.1 −.8 Miscellaneous .................... 339 −1.5 9.6 .6 3.9 4.8 .1 2.2 2.8 −.4 .9 Nondurable manufacturing .......... . . . −2.9 .8 −.4 2.8 −.1 .1 1.2 .0 −.3 −.4 Food, beverage, and tobacco products ..................... 311,2 −.2 −2.2 1.7 1.7 3.0 .7 .7 −.4 −2.2 .5 Textile and product mills .......... 313,4 −9.9 2.2 −4.2 −3.9 .5 −.4 1.8 4.3 −1.4 1.6 Apparel and leather ............... 315,6 −21.9 −6.9 −9.4 −2.2 −.2 −.8 2.4 4.9 2.4 5.7 Paper ............................ 322 −5.0 3.4 −6.0 4.5 −.7 .8 −.7 −2.7 1.2 −.9 Printing and support .............. 323 −9.0 −3.2 −3.0 1.5 1.6 −.9 .0 .5 .6 .1 Petroleum and coal products ....... 324 1.0 2.9 .3 6.2 −6.3 .2 −1.2 −.9 2.2 −.8 Chemical ........................ 325 .1 3.6 .7 4.2 −3.3 −.1 2.6 −.5 .0 −2.1 Plastics and rubber products ....... 326 −5.1 4.4 −.2 3.2 3.6 −.6 1.9 1.9 1.6 1.2 Other manufacturing (non-NAICS) ..... 1133,5111 −5.5 −2.6 −3.0 3.7 1.7 .8 1.4 −3.2 −1.4 .3 Mining .................................. 21 −.7 −3.8 .5 −.4 −6.3 .0 −.1 .3 1.6 −1.6 Utilities ................................. 2211,2 −5.2 7.0 .7 1.2 3.3 −.1 −.1 1.3 −1.5 2.2 Electric ............................... 2211 −3.8 5.6 1.9 2.0 4.8 −.1 −.1 1.4 −1.7 1.7 Natural gas ........................... 2212 −12.9 15.5 −5.5 −2.9 −3.6 −.1 .1 .7 −.4 4.6 Note: Estimates for the fourth quarter of 2005 are subject to further revision 3. Manufacturing comprises North American Industry Classification System in the upcoming monthly releases. (NAICS) manufacturing industries (sector 31-33) plus the logging industry and 1. Rates of change are calculated as the percent change in the seasonally the newspaper, periodical, book, and directory publishing industries. Logging adjusted index from the fourth quarter of the previous year to the fourth quarter and publishing are classified elsewhere in NAICS (under agriculture and infor­ of the year specified in the column heading. For 2005, the difference between mation repectively), but historically they were considered to be manufacturing the rates of change are calculated from annualized rates of change between the industries and were included in the industrial sector under the Standard Indus­ fourth quarter of 2004 and the third quarter of 2005. trial Classification (SIC) system. In December 2002 the Federal Reserve reclas­ 2. North American Industry Classification System. sified all its industrial output data from the SIC system to NAICS. . . . Not applicable.

Industrial Production and Capacity Utilization: The 2005 Annual Revision A55 A.4. Rates of change in industrial production, special aggregates and selected detail, 2001–051 Difference between rates of change: Revised rate of change revised minus earlier Item NAICS (percent) (percentage points) code 2 2001 2002 2003 2004 2005 2001 2002 2003 2004 2005 Total industry .......................... ... −5.3 2.3 1.5 4.3 3.1 −.2 .8 .3 .0 .2 Energy .................................. . . . −3.3 2.8 .5 .7 −2.1 .0 −.2 .7 .2 −.1 Consumer products .................... . . . −5.3 9.5 −1.9 1.7 2.4 −.1 −.6 −.5 .7 2.1 Commercial products .................. . . . −1.4 4.5 5.1 2.4 3.0 −.2 −.2 5.1 −5.0 1.3 Oil and gas well drilling ............... . . . −7.8 −15.2 21.2 8.3 11.8 .3 .3 .2 −.4 1.2 Converted fuel ........................ . . . −8.0 4.1 .6 1.6 −1.7 .1 .1 .6 −.1 −1.3 Primary materials ..................... . . . .0 −1.6 −.4 −1.0 −6.5 .0 −.1 .0 1.6 −.9 Non-energy ............................. . . . −5.7 2.2 1.7 5.1 4.4 −.2 1.0 .2 .0 .2 Selected high-technology industries ..... . . . −9.8 4.8 21.1 18.4 26.1 .3 −3.3 2.4 −.3 1.0 Computers and peripheral equipment . 3341 −2.7 −2.6 5.8 4.6 11.3 .9 −3.5 −16.0 −2.3 −1.2 Communications equipment ......... 3342 −27.4 −22.6 9.9 22.3 25.7 2.9 −8.3 −12.6 12.8 3.4 Semiconductors and related electronic components .......... 334412–9 −.6 26.0 34.1 21.4 30.7 −2.3 .8 17.9 −8.5 −1.0 Excluding selected high-technology industries ........................ . . . −5.2 1.9 .5 4.2 3.0 −.3 1.3 .1 .0 .1 Motor vehicles and parts ............ 3361–3 −1.8 12.2 4.7 2.6 2.3 .2 1.0 −.1 −.3 .8 Motor vehicles ................... 3361 2.4 14.4 10.4 1.6 −.2 −.1 3.3 3.7 −.8 .2 Motor vehicle parts ............... 3363 −4.0 10.3 −1.5 2.2 3.3 1.0 −.5 −4.2 .5 .2 Excluding motor vehicles and parts . . . . . . −5.5 .9 .0 4.4 3.1 −.3 1.3 .1 .0 .0 Consumer goods .................. . . . −1.5 .0 .7 2.3 2.1 .2 1.2 .0 −1.4 .3 Business equipment ............... . . . −11.3 .4 .5 9.0 9.5 −1.2 2.5 .4 −.3 −2.4 Construction supplies ............. . . . −4.8 1.4 1.7 4.6 6.7 .0 1.1 1.1 .7 −.1 Business supplies ................. . . . −6.4 1.4 −1.6 3.3 2.6 .0 1.6 −.7 .2 .4 Materials ......................... . . . −7.8 1.8 −.8 4.7 .8 −.6 1.2 .1 .9 .4 Measures excluding selected hightechnology industries Total industry ........................... . . . −4.9 2.1 .5 3.6 1.9 −.2 1.0 .2 .0 .0 Manufacturing 3 ....................... . . . −5.2 1.9 .4 4.2 2.8 −.3 1.2 .0 .0 .0 Durable ............................ . . . −7.2 3.5 1.7 5.6 5.5 −.7 1.3 .5 .5 .3 Measures excluding motor vehicles and parts Total industry ........................... . . . −5.5 1.5 1.3 4.5 3.2 −.2 .8 .3 .1 .1 Manufacturing 3 ....................... . . . −5.9 1.3 1.4 5.4 4.4 −.3 .9 .2 .1 .1 Durable ............................ . . . −8.6 2.2 3.8 8.0 8.9 −.7 .6 .9 .7 .6 Measures excluding selected hightechnology industries and motor vehicles and parts Total industry ........................... . . . −5.1 1.3 .1 3.6 1.9 −.3 1.0 .2 .0 .0 Manufacturing 3 ....................... . . . −5.5 1.0 .0 4.4 2.9 −.3 1.3 .0 .1 −.1 Measures of non-energy material inputs to Finished processors ...................... . . . −8.5 5.7 3.6 7.9 8.4 −.8 .2 1.7 −.4 .8 Primary and semifinished processors ...... . . . −5.8 3.3 .2 4.3 −.5 −.5 1.5 .2 1.3 .7 Stage-of-process groups Crude ................................... . . . −2.8 −1.1 −1.7 1.9 −8.2 .2 −.1 −.3 1.8 −1.9 Primary and semifinished ................ . . . −6.3 4.2 1.2 4.4 4.4 −.4 .8 .9 .1 1.2 Finished ................................ . . . −4.5 .6 2.8 4.9 5.3 .0 1.0 −.4 −.5 −.4 Note: Estimates for the fourth quarter of 2005 are subject to further revision the rates of change are calculated from annualized rates of change between the in the upcoming monthly releases. fourth quarter of 2004 and the third quarter of 2005. 1. Rates of change are calculated as the percent change in the seasonally 2. North American Industry Classification System. adjusted index from the fourth quarter of the previous year to the fourth quarter 3. Refer to footnote 3 in table A.3. of the year specified in the column heading. For 2005, the difference between . . . Not applicable.

A56 Federal Reserve Bulletin 2006 A.5. Rates of change for annual industrial production indexes, 2001–051 Difference between rates of change: Revised rate of change revised minus earlier (percent) Item (percentage points) 2001 2002 2003 2004 2005 2001 2002 2003 2004 2005 Total industry .............................. −3.5 .1 .6 4.1 3.2 .0 .4 .6 −.1 .1 Market Groups Consumer goods ............................. −1.1 2.2 1.0 2.1 2.1 .0 .8 .8 −.6 −.4 Durable ................................... −4.3 6.3 4.0 2.8 1.9 .0 .9 1.9 −.3 .2 Nondurable ............................... .1 .7 −.2 1.8 2.2 .0 .7 .4 −.8 −.6 Business equipment .......................... −7.0 −7.8 .0 9.3 9.0 .1 .1 −1.2 −.1 .2 Defense and space equipment ................ 8.6 −.1 5.0 7.8 10.6 −.3 −1.3 −2.3 3.1 1.1 Construction supplies ........................ −4.5 −.2 −.9 5.5 4.0 .1 .7 1.4 .4 .0 Business supplies ............................ −4.0 .2 .7 3.2 3.5 .0 1.0 1.1 −1.3 .4 Materials .................................... −4.5 1.0 .4 4.2 2.0 .0 .1 .9 .5 .3 Non-energy ............................... −5.6 1.4 .6 5.8 3.5 .0 .1 1.1 .6 .5 Energy ................................... −1.1 .0 −.4 .0 −1.7 .0 .0 .3 .4 −.2 Industry Groups Manufacturing 2 ............................. −4.2 .1 .5 4.8 3.9 .0 .5 .6 .0 .1 Manufacturing (NAICS) ................... −4.1 .3 .7 5.0 3.9 .0 .4 .7 .2 .2 Durable manufacturing .................. −4.8 −.2 2.3 7.3 6.4 .1 .0 1.0 .3 .7 Nondurable manufacturing .............. −3.2 1.0 −1.1 2.2 .7 .0 .8 .4 .0 −.6 Other manufacturing (non-NAICS) ......... −5.9 −3.0 −3.0 1.8 3.0 −.2 2.0 −2.0 −2.4 −.4 Mining ...................................... .9 −4.3 −.2 −.2 −2.0 .0 −.1 .2 .7 −.4 Utilities ..................................... −.5 3.1 2.0 1.2 2.6 .0 .0 1.6 −1.4 .5 Note: Estimates for the fourth quarter of 2005 are subject to further revision one year and the fourth quarter of the next. The difference between revised and in the upcoming monthly releases. earlier changes for 2005 are computed from annualized rates of change between 1. The rates of change are calculated from annual averages of seasonally the full year 2004 and the first three quarters of 2005. adjusted industrial production indexes, rather than between the fourth quarter of 2. Refer to footnote 3 in table A.3. A.6. Rates of change in capacity, by industry groups, 2001–051 Difference between rates of change: Revised rate of change revised minus earlier (percent) Item (percentage points) 2001 2002 2003 2004 2005 2001 2002 2003 2004 2005 Total industry .............................. 2.9 .7 −.2 .6 1.6 .2 .3 .0 −.5 .4 Manufacturing 2 ............................. 2.8 .4 −.1 .5 2.1 .3 .3 .0 −.6 .7 Manufacturing (NAICS) ................... 3.1 .5 .1 .5 2.2 .3 .4 .0 −.7 .7 Durable manufacturing .................. 5.1 1.1 1.2 1.4 4.0 .2 .5 .0 −.8 1.0 Nondurable manufacturing .............. .5 −.3 −1.1 −.7 −.2 .4 .2 .2 −.5 .3 Other manufacturing (non-NAICS) ......... −1.1 −2.7 −3.0 .4 .5 .2 −.2 −.1 .3 .2 Mining ...................................... 2.1 −1.3 −1.0 −.6 −.6 .0 −.7 .7 −.3 .1 Utilities ..................................... 3.9 4.5 3.1 2.6 .0 .0 .0 .1 .7 −1.2 Selected high-technology industries ........... 27.8 9.5 8.0 6.8 20.8 .3 1.5 −.4 −6.6 5.9 Manufacturing except selected high-technology industries 2 .............. .8 −.2 −.5 .1 .6 .3 .2 .1 .0 .3 Stage-of-process groups Crude ....................................... 1.5 −1.0 −2.1 −1.1 −.9 .6 −.3 .2 −.9 −.2 Primary and semifinished .................... 3.4 .9 −.1 .9 2.5 .5 .2 .1 −1.1 .7 Finished .................................... 2.3 .6 .6 .8 1.2 −.1 .4 .0 .5 .3 1. Rates of change are calculated as the percent change in the seasonally 2. Refer to footnote 3 in table A.3. adjusted index from the fourth quarter of the previous year to the fourth quarter of the year specified in the column heading.

Industrial Production and Capacity Utilization: The 2005 Annual Revision A57 A.7. Capacity utilization rates, by industry groups, 1972–2005 Difference between rates: Revised rate revised minus earlier (percent of capacity, seasonally adjusted) NAICS (percentage points) Item code1 1972–2004 2002:Q4 2003:Q4 2004:Q4 2005:Q3 2002:Q4 2003:Q4 2004:Q4 2005:Q3 avg. Total industry .............................. . . . 81.0 75.3 76.6 79.4 79.8 −.1 .1 .6 .5 Manufacturing 2 ............................. . . . 79.8 73.4 74.7 78.2 78.5 −.1 .0 .5 .3 Manufacturing (NAICS) ................... 31–33 79.6 73.0 74.3 77.8 78.1 −.2 .1 .7 .4 Durable manufacturing .................. . . . 78.1 70.3 72.2 76.3 77.1 −.3 .1 1.1 1.0 Wood products ....................... 321 80.1 74.4 78.4 81.0 80.9 .0 1.1 2.5 5.1 Nonmetallic mineral products ......... 327 79.4 77.7 79.4 82.4 80.8 .9 1.4 1.6 .5 Primary metal ........................ 331 80.4 78.2 80.6 85.0 79.4 −.4 1.3 2.4 4.2 Fabricated metal products ............. 332 77.1 70.4 70.0 73.9 74.7 1.0 2.3 3.8 3.7 Machinery ........................... 333 78.7 67.5 69.4 78.3 80.1 −.3 −.5 .0 −2.2 Computer and electronic products ...... 334 78.6 60.9 67.0 72.9 76.5 −1.8 −.8 1.7 .6 Electrical equipment, appliances, and components ................. 335 83.2 74.0 75.8 81.1 84.9 1.4 1.0 .8 2.7 Motor vehicles and parts .............. 3361–3 77.5 80.4 80.3 80.5 80.9 −.5 −1.6 −2.1 −1.3 Aerospace and miscellaneous transportation equipment ......... 3364–9 72.6 63.0 62.3 65.2 65.8 −.2 −.9 −.3 −.6 Furniture and related products ......... 337 78.7 73.1 73.8 75.1 73.8 .5 2.6 1.6 .9 Miscellaneous ........................ 339 76.6 75.0 74.8 76.7 78.2 −.9 .6 −.5 −.8 Nondurable manufacturing .............. . . . 81.7 76.6 77.2 79.9 79.5 .0 −.1 .0 −.4 Food, beverage, and tobacco products . . 311,2 81.8 77.0 78.2 79.4 80.7 .4 −.4 −2.3 −1.8 Textile and product mills .............. 313,4 82.8 74.7 74.6 75.1 77.2 −2.1 1.2 .5 .4 Apparel and leather ................... 315,6 79.3 65.7 66.7 72.9 77.3 −1.0 1.8 2.4 4.5 Paper ................................ 322 87.8 84.9 81.4 85.5 83.2 .1 −2.1 −1.2 −1.6 Printing and support .................. 323 84.0 72.5 71.6 74.9 76.5 −.5 .0 2.8 3.3 Petroleum and coal products ........... 324 85.9 86.5 88.4 93.4 90.1 −.5 −.5 2.1 1.4 Chemical ............................ 325 78.3 74.4 74.4 77.4 74.7 1.0 .4 .8 −.9 Plastics and rubber products ........... 326 83.5 79.0 81.3 85.0 86.6 −2.0 .1 1.5 1.8 Other manufacturing (non-NAICS) ......... 1133,5111 84.7 82.1 82.2 84.9 85.5 1.5 −1.2 −2.6 −2.6 Mining ...................................... 21 87.3 86.8 88.2 88.3 86.1 1.4 1.1 2.7 1.5 Utilities ..................................... 2211,2 86.8 87.6 85.6 84.4 88.1 −.3 .8 −1.0 1.0 Selected high-technology industries ........... . . . 78.2 58.5 65.7 72.8 75.3 −2.3 −1.0 3.0 1.1 Computers and peripheral equipment ....... 3341 78.2 70.4 74.9 76.3 79.0 −.5 .8 .4 −1.4 Communications equipment ................ 3342 76.0 42.2 47.1 58.6 68.9 −.6 −5.8 −.1 .7 Semiconductors and related electronic components .......................... 334412–9 80.6 64.3 74.3 79.4 77.3 −5.5 −.6 3.8 .3 Measures excluding selected high-technology industries Total industry ............................... . . . 81.2 76.6 77.3 79.8 80.2 .1 .2 .2 .1 Manufacturing 2 ........................... . . . 79.9 74.8 75.4 78.6 78.9 .1 .1 .1 −.1 Stage-of-process groups Crude ....................................... . . . 86.4 84.1 85.0 87.8 84.2 .6 .1 2.3 1.3 Primary and semifinished .................... . . . 82.2 77.5 78.6 81.3 81.8 −.5 .2 1.1 1.5 Finished .................................... . . . 77.9 71.1 72.4 75.3 76.9 .3 .0 −.7 −1.1 1. North American Industry Classification System. . . . Not applicable. 2. Refer to footnote 3 in table A.3.

A58 Federal Reserve Bulletin 2006 A.8. Annual proportion in industrial production, by market groups and industry groups, 1997–2005 NAICS Item 1997 1998 1999 2000 2001 2002 2003 2004 2005 code1 Total industry .............................. . . . 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 Market groups Final products and nonindustrial supplies ...... . . . 56.9 58.1 57.6 57.6 59.1 59.0 58.6 57.8 57.6 Consumer goods .......................... . . . 27.6 28.0 28.2 28.5 30.1 31.0 31.2 30.2 29.7 Durable ................................ . . . 7.9 7.9 8.0 7.9 8.1 8.9 9.0 8.7 8.3 Automotive products .................. . . . 3.7 3.7 3.9 3.7 4.0 4.7 4.9 4.7 4.6 Home electronics ..................... . . . .4 .4 .4 .4 .4 .3 .3 .3 .2 Appliances, furniture, carpeting ........ . . . 1.4 1.4 1.4 1.4 1.4 1.5 1.4 1.4 1.4 Miscellaneous goods .................. . . . 2.4 2.4 2.4 2.4 2.3 2.4 2.3 2.3 2.2 Nondurable ............................. . . . 19.7 20.1 20.2 20.7 22.0 22.1 22.1 21.6 21.3 Non-energy .......................... . . . 16.3 16.9 16.7 16.9 18.1 18.3 18.0 17.4 16.9 Foods and tobacco ................. . . . 8.7 9.2 9.1 9.3 10.0 9.8 9.8 9.5 9.1 Clothing ........................... . . . 1.6 1.5 1.3 1.2 1.1 1.0 .9 .8 .7 Chemical products ................. . . . 3.7 3.8 3.8 3.9 4.5 4.9 4.9 4.8 4.7 Paper products ..................... . . . 1.8 1.9 1.9 1.9 2.0 2.0 1.9 1.9 1.9 Energy ............................... . . . 3.4 3.2 3.5 3.7 3.8 3.9 4.1 4.2 4.4 Business equipment ....................... . . . 11.8 12.3 11.9 11.7 11.2 10.3 9.9 10.0 10.1 Transit ................................. . . . 2.1 2.5 2.3 2.0 2.0 1.9 1.8 1.9 2.0 Information processing .................. . . . 4.0 4.0 4.1 4.1 3.8 3.1 2.9 2.7 2.8 Industrial and other ..................... . . . 5.8 5.8 5.5 5.6 5.4 5.3 5.2 5.4 5.3 Defense and space equipment .............. . . . 1.9 1.9 1.8 1.5 1.8 1.8 1.9 1.9 2.1 Construction supplies ...................... . . . 4.2 4.3 4.3 4.3 4.3 4.3 4.3 4.4 4.5 Business supplies ......................... . . . 11.1 11.1 11.1 11.2 11.2 11.2 11.1 10.8 10.6 Materials .................................... . . . 43.1 41.9 42.4 42.4 40.9 41.0 41.4 42.2 42.4 Non-energy ............................... . . . 33.8 33.3 33.1 32.3 30.9 30.7 30.0 30.2 29.4 Durable ................................ . . . 21.7 21.5 21.4 20.9 19.6 19.1 18.7 19.0 18.8 Consumer parts ....................... . . . 4.2 4.2 4.4 4.1 3.8 4.0 3.8 3.7 3.5 Equipment parts ...................... . . . 8.3 8.2 8.1 8.2 7.3 6.7 6.6 6.7 6.7 Other ................................ . . . 9.2 9.1 8.9 8.6 8.4 8.4 8.4 8.7 8.5 Nondurable ............................. . . . 12.1 11.8 11.7 11.4 11.2 11.6 11.3 11.2 10.7 Textile ............................... . . . 1.1 1.0 1.0 .9 .8 .8 .8 .7 .6 Paper ................................ . . . 2.9 2.8 2.9 2.8 2.8 2.7 2.5 2.4 2.3 Chemical ............................ . . . 4.9 4.6 4.5 4.3 4.2 4.5 4.5 4.7 4.4 Energy ................................... . . . 9.3 8.6 9.2 10.1 10.0 10.3 11.4 12.0 13.0 Industry groups Manufacturing 2 ............................. . . . 85.7 86.5 85.8 84.5 84.1 83.9 82.6 82.0 80.8 Manufacturing (NAICS) ................... 31–33 81.3 81.8 81.0 79.7 79.2 79.1 77.9 77.5 76.4 Durable manufacturing .................. . . . 46.5 47.1 46.7 45.6 44.3 43.6 42.9 43.1 42.9 Wood products ....................... 321 1.5 1.5 1.6 1.4 1.4 1.5 1.5 1.6 1.5 Nonmetallic mineral products ......... 327 2.2 2.3 2.3 2.2 2.3 2.3 2.2 2.3 2.3 Primary metal ........................ 331 3.1 2.9 2.8 2.5 2.3 2.3 2.3 2.6 2.4 Fabricated metal products ............. 332 6.0 6.1 6.0 6.0 5.9 5.7 5.6 5.7 5.8 Machinery ........................... 333 6.3 6.2 5.9 6.0 5.7 5.3 5.1 5.4 5.3 Computer and electronic products ...... 334 10.4 10.2 10.3 10.3 9.2 8.0 7.8 7.7 7.9 Electrical equipment, appliances, and components ................. 335 2.6 2.6 2.5 2.5 2.4 2.2 2.1 2.1 2.1 Motor vehicles and parts .............. 3361–3 6.7 6.6 7.0 6.6 6.5 7.4 7.5 7.3 7.1 Aerospace and miscellaneous transportation equipment ......... 3364–9 3.4 4.1 3.8 3.3 3.8 3.6 3.5 3.5 3.7 Furniture and related products ......... 337 1.6 1.7 1.7 1.7 1.7 1.8 1.8 1.7 1.6 Miscellaneous ........................ 339 2.8 2.8 2.8 2.9 3.1 3.3 3.3 3.2 3.2 Nondurable manufacturing .............. . . . 34.7 34.7 34.4 34.1 35.0 35.5 35.0 34.4 33.5 Food, beverage, and tobacco products . . 311,2 10.1 10.6 10.4 10.7 11.4 11.4 11.4 11.0 10.7 Textile and product mills .............. 313,4 1.7 1.6 1.5 1.4 1.3 1.4 1.3 1.2 1.1 Apparel and leather ................... 315,6 1.8 1.6 1.4 1.3 1.2 1.0 .9 .8 .7 Paper ................................ 322 3.2 3.2 3.2 3.2 3.1 3.1 2.9 2.8 2.7 Printing and support .................. 323 2.7 2.6 2.6 2.6 2.6 2.4 2.3 2.1 2.0 Petroleum and coal products ........... 324 1.6 1.5 1.7 1.9 1.7 1.7 2.0 2.3 2.5 Chemical ............................ 325 10.1 9.9 9.6 9.4 9.8 10.6 10.6 10.5 10.2 Plastics and rubber products ........... 326 3.7 3.7 3.8 3.7 3.7 3.8 3.7 3.6 3.6 Other manufacturing (non-NAICS) ......... 1133,5111 4.4 4.7 4.8 4.8 4.9 4.8 4.7 4.5 4.4 Mining ...................................... 21 5.4 4.8 5.5 6.5 6.4 6.4 7.5 8.5 9.8 Utilities ..................................... 2211,2 8.9 8.7 8.6 9.0 9.5 9.7 9.9 9.5 9.5 Electric ................................... 2211 7.7 7.5 7.4 7.6 8.1 8.2 8.3 7.9 7.8 Natural gas ............................... 2212 1.3 1.2 1.2 1.4 1.4 1.4 1.6 1.6 1.7 Note: The IP proportion data are estimates of the industries’ relative contri- 1. North American Industry Classification System. butions to the overall IP change between the reference year and the following 2. Refer to footnote 3 in table A.3. year. For example, a 1 percent increase in durable goods manufacturing between . . . Not applicable. 2005 and 2006 would account for a 0.429 percent increase in total IP.

A59 Understanding U.S. Cross-Border Securities Data Carol C. Bertaut, William L. Griever, and Ralph W. mately 40 percent higher than that of U.S. holdings; Tryon, of the Board’s Division of International by the end of 2005, it was approximately 70 percent Finance, prepared this article. Stephen S. Gardner higher.Themore-rapidgrowthofforeignholdingsof andJonasJ.Robisonprovidedresearchassistance. U.S. securities over the past ten years is the counterpart to the record U.S. trade and current account In recent years, foreign holdings of U.S. securities deficits incurred over the period, as the financial havegrownmarkedly.During2005,reportedforeign inflows associated with the deficits have occurred holdings increased nearly $1 trillion for the second largelythroughforeignpurchasesofU.S.securities. consecutiveyear,bringingtheestimatedtotaltoabout ThetrendinforeignholdingsrelativetoU.S.hold- $7.3 trillion, or roughly 16 percent of all U.S. long- ings varies by type of security. In recent years, U.S. term securities outstanding at year-end. These large holdingsofforeignequityhavebeensomewhatlarger numbersareunderstandablyattractingagreatdealof than foreign holdings of U.S. equity (figure 1). For attention,asexternaldeficitsareasubjectofgrowing holdings of long-term debt, however, the situation concernintoday’sglobaleconomy. has been very different, as foreign holdings have In this article, we present current data on U.S. exceededU.S.holdingsbyawidemargin.Thedisparcross-border securities holdings and transactions and itycanbepartlyexplainedbytheholdingsofforeign describethesystemthatcollectsthedata.Wediscuss officialinstitutions,whicharediscussedindetaillater howtomakethebestuseoftheinformationavailable inthisarticle. by avoiding common misinterpretations of the data AnincreaseinthelevelofforeignholdingsofU.S. and by adjusting the published figures to improve securitieshasalsoresultedinanincreaseintheshare their accuracy and comprehensiveness. We also dis- of U.S. securities that are foreign held. Since 1974, cuss how to construct monthly estimates of cross- whensurveysbegantocollectdataonforeignownerborder securities holdings by country, combining shipofU.S.long-termsecurities,theshareofthetotal monthly transactions data with less frequently value of U.S. long-term securities held by foreigners reported positions data. Besides providing more- has more than tripled, from less than 5 percent to timely measures of holdings of securities, these esti- 16percentasofJune2005(table1).Asafractionof mates incorporate a number of adjustments that improve our overall picture of cross-border portfolio positions. Finally, to improve our ability to correctly attribute U.S. liabilities to foreign holders, we com- 1. Foreign holdings of U.S. long-term securities and U.S. holdings of foreign long-term securities, by type of pareourestimatesofforeignholdingsofU.S.securisecurity, 1994–2005 ties with estimates obtained from asset surveys conductedbyothercountries. Billions of U.S. dollars 4,500 4,000 INCREASING IMPORTANCE OF FOREIGN Foreign debt holdings 3,500 HOLDINGS OF U.S. SECURITIES 3,000 U.S. equity holdings 2,500 The increasing importance of foreign holdings of U.S. securities can be seen by comparing the growth 2,000 of these holdings with the growth of U.S. ownership 1,500 Foreign equity holdings of foreign securities. Since 1994, when the first sur- 1,000 vey of U.S. holdings of foreign long-term securities U.S. debt holdings 500 was conducted, foreign ownership of U.S. long-term 0 securities has consistently exceeded U.S. ownership 1995 1997 1999 2001 2003 2005 of foreign long-term securities. At the end of 1994, SOURCE: Treasury International Capital reporting system and staff the market value of foreign holdings was approxi- estimates.

A60 Federal Reserve Bulletin 2006 1. ForeignholdingsofU.S.long-termsecuritiesasashare 3. Total foreign holdings and foreign official holdings of ofsuchsecuritiesoutstanding,bytypeofsecurityand U.S. short-term Treasury securities, 1994–2005 forsurveydates,1974–2005 Percent Billions of U.S. dollars Typeofsecurity 350 Month Debt and 300 year All Equity1 U.S. Total foreign holdings U.S. govern- Other3 Treasury2 ment 250 agency 200 Dec.1974 .. 5 4 15 n.a. n.a. Dec.1978 .. 4 5 12 3 1 Dec.1984 .. 6 5 14 3 3 150 Dec.1989 .. 9 6 22 4 7 Dec.1994 .. 8 5 19 5 8 Foreign official holdings Mar.2000 .. 10 7 35 7 12 100 June2002... 12 8 41 10 16 June2003... 14 9 46 11 16 June2004... 14 9 52 11 17 June2005... 16 10 52 14 20 1995 1997 1999 2001 2003 2005 Note: Percentagesshouldbeviewedasapproximate,asdataonthetotal SOURCE: Treasury International Capital reporting system. valueofU.S.long-termsecuritiesoutstandingbysecuritytypeareunavailable onthesamebasisasthatusedincollectingthesurveydataonforeignholdingsofsuchsecurities.Forexample,whereasdataontotalU.S.long-termdebt A similar relationship holds for relative sizes of securitiesoutstandingarebasedonthefacevalueofthesecurities,dataonforforeign and U.S. holdings of short-term securities, eignholdingsarebasedontheirmarketvalue.However,thepercentagesshould stillbeusefulforshowinglong-termtrends. althoughthemagnitudeoftheseholdingsisconsider- 1. Bothcommonandpreferredstockaswellasalltypesofinvestmentcomably smaller. Total foreign holdings of U.S. shortpanyshares,suchasopen-end,closed-end,andmoneymarketmutualfunds. 2. MarketableTreasurysecuritiesonly. term debt securities are more than twice as large as 3. U.S. debt securities issued by all other institutions, primarily corporate U.S. holdings of foreign short-term debt securities, issuers. n.a. Notavailable. in large part because of the sizable holdings of for- Source: U.S.DepartmentoftheTreasury,ReportonForeignPortfolioHoldeignofficialinstitutions(figure2).Theimportanceof ingsofU.S.Securities,variousdates. holdings by foreign official institutions is especially striking for short-term Treasury securities (figure 3). thetotaloutstanding,holdingsaregreatestinTreasur- As shown in the figure, foreign official holdings ies:MorethanhalfofallmarketableTreasurysecuri- account for more than three-fourths of short-term ties held by the public are foreign owned. In terms Treasurysecuritiesheldbyforeigners. ofmarketvalue,thelevelofforeignholdingsofU.S. long-term securities increased from $67 billion as of year-end1974to$6.3trillionasofJune2005. THE TIC REPORTING SYSTEM The data that underlie these estimates of U.S. cross- 2. Total foreign and foreign private holdings of U.S. borderfinancialactivityarecollectedbytheTreasury short-term debt securities, and total U.S. holdings of International Capital (TIC) reporting system.1 This foreign short-term debt securities, 1994–2005 system is more comprehensive than many users realize. Users often assume that the TIC system collects Billions of U.S. dollars only monthly data on cross-border transactions in long-termsecurities.Althoughthesedatareceivecon- 600 siderable attention in the financial press, they consti- Total foreign holdings 500 tuteonlyasmallpartoftheTICsystem. 400 Besides the transactions data, which cover only long-term securities (that is, securities with an origi- 300 nal maturity of more than one year), the TIC system Foreign private holdings 200 includes monthly and quarterly cross-border data (including holdings of short-term securities) reported 100 Total U.S. holdings 0 1. TIC data are published on the Treasury Department’s website at www.treas.gov/tic/. The website includes past and present data, 1995 1997 1999 2001 2003 2005 articles about the TIC system, TIC forms and instructions, related SOURCE: Treasury International Capital reporting system. non-TICwebsites,andTICcontactinformation.

Understanding U.S. Cross-Border Securities Data A61 by banks and broker–dealers; periodic (now annual) Holdings of Short-Term Securities in-depth surveys of cross-border holdings of both long- and short-term securities; and quarterly posi- Selected data on cross-border holdings of short-term tion data reported by nonbank respondents such as securitiesarecollectedmonthlyorquarterly,butthese commercial concerns, exporters and importers, and datamaybelesswellknownthanthedataontransacother financial institutions. In 2005, the TIC system tionsinlong-termsecuritiesbecausetheyarereported alsobegantocollectdataoncross-borderderivatives and released with the banking data collected on the positions.2 TIC B forms. The B data include foreign holdings of U.S. short-term securities—such as U.S. Treasury billsandcertificates,U.S.governmentagencysecuri- Transactions in Long-Term Securities ties, commercial paper, and negotiable certificates of deposit (collected in the banking liabilities data)— Information on cross-border transactions in U.S. and as well as U.S. holdings of similar types of foreign foreign long-term securities is collected monthly on short-termsecurities(collectedinthebankingclaims the TIC S form. Data are collected by country, at data). The primary respondents for these position marketvalue,andarepublishedwithaforty-five-day data are U.S.-resident custodians that report their lag. The primary respondents for these transactions holdingsonbehalfoftheircustomers. data are U.S.-resident brokers and dealers, although Like the S data on long-term securities transacsomeendinvestorsandsecurityissuersalsoreporton tions,theBdataonshort-termsecuritiesholdingsare theTICS. collected by country and by broad class of security ForU.S.securities,dataarecollectedseparatelyfor type, such as U.S. Treasury securities; these data are four types of securities: equity, U.S. Treasury debt, alsoreportedbymajorforeigncounterparties,includ- U.S. government agency debt, and debt issued by all ing foreign official institutions, foreign banks, and other institutions (primarily corporate issuers). For other private foreigners. The short-term securities foreignsecurities,onlytwosecuritytypes,equityand data are reported at face value; data on U.S.-dollarlong-termdebt,areseparatelymeasured.Information denominated and foreign-currency-denominated onforeignofficialpurchasesofU.S.securitiesisalso securities are reported separately. Because the shortcollected separately from information on purchases term securities data are reported as positions, net byotherforeigners. transactions in these securities must be calculated as For analytical purposes, the sales of each type of the change in position from one period to another. security are usually subtracted from gross purchases The S forms and the B forms provide much less to measure net transactions. The S form follows detail than do the periodic surveys, which are disinternationalreportingconventionsformeasuringthe cussed in the next section. Nevertheless, because the balanceofpayments:Itreportsforeignnetpurchases dataarereleasedaboutforty-fivedaysaftertheendof of U.S. long-term securities with a positive sign agivenmonth,theyofferatimelyandfairlycomprebecause they are a source of capital inflow to the hensivemeasureofcross-bordersecuritiesflows. United States, and it reports U.S. net purchases of foreign long-term securities with a negative sign becausetheyareasourceofcapitaloutflowfromthe Annual Surveys of Holdings of Long- and UnitedStates.3 Short-Term Securities More-comprehensive data on the level of both foreign holdings of U.S. securities (U.S. liabilities) and U.S. holdings of foreign securities (U.S. assets) are 2. ForallmonthlyandquarterlyTICforms,reportingisrequired measured in the annual surveys. As noted earlier, the by law as long as the reporter has cross-border activity above the surveys now collect data on both long- and shortexemptionlevelsetforthatform. term securities.4 Whereas in other parts of the TIC 3. The TIC S form reports all data from the perspective of the foreignresidentinvolvedinthecross-bordertransaction.Thus,when system the respondents report data in aggregate by aU.S.investorpurchasesaforeignsecurity,thetransactionisreported countryandbybroadinstrumenttype,respondentsto asaforeignsaleofaforeignsecurity.Likewise,whenaU.S.investor sells a foreign security, the transaction is recorded as a foreign purchaseofaforeignsecurity.Therefore,netforeignsalesofforeign securities are equivalent to net U.S. purchases of foreign securities. 4. Theannualsurveyscollecteddataononlylong-termsecurities Thedataontransactionsinforeignsecuritiesarealsoreportedinthis untiltheDecember2001surveyofU.S.holdingsofforeignsecurities. way in the TIC system’s online files of gross purchases and gross Data on foreign holdings of U.S. short-term securities were first sales. collectedintheJune2002survey.

A62 Federal Reserve Bulletin 2006 the surveys report information on cross-border hold- ratecross-borderdatafromadiverseandevolvingset ingsonasecurity-by-securitybasis. ofparticipants.6 Collectingdataonholdingsofindividualsecurities allows for much more detailed data reporting and significantly improves survey accuracy, but it also USE OF TIC DATA IN THE BALANCE OF requires the processing of a large number of records PAYMENTS AND FLOW OF FUNDS ACCOUNTS (more than 500,000 for the asset surveys and almost 2.8 million for the liabilities surveys). The surveys The most comprehensive measures of cross-border thus take much longer to complete than do other financialflowsandpositionsarethosethattheBureau reports for the TIC system: Preliminary results are ofEconomicAnalysis(BEA)reportsinthequarterly usually available after nine months and final data balance of payments accounts and in the annual net after twelve months. However, the greater detail in international investment position.7 The portfolio stathe data collected permits the surveys to produce tisticsintheseinternationalaccountsarebasedonthe information that is otherwise unavailable, such as monthly and quarterly TIC securities data and on the currency composition, maturity structure, industry annual surveys. However, the balance of payments sector, both face and market value of holdings, and accounts also include flows and positions calculated the specific securities held. Liabilities surveys mea- from the remaining TIC bank- and nonbank-reported sure positions as of June 30, and asset surveys mea- data, as well as information on direct investment surethemasofDecember31.5 collectedandcompiledbytheBEA. The TIC data are also used as inputs in the estimates for the ‘‘rest of the world’’ sector, included in Banking and Nonfinancial Corporate Data the flow of funds accounts compiled by the Board of Governors of the Federal Reserve System. In most Besidesdataonholdingsofshort-termsecurities,the estimates of financial flows and holdings for that B forms collect data on cross-border positions in sector, the flow of funds accounts incorporate the the form of deposits, loans, brokerage balances, and BEA’s official balance of payments statistics, and repurchaseagreements.Althoughthesedataarecom- thus the flow of funds statistics are based only indimonly referred to as the TIC ‘‘banking’’ data, they rectly on the TIC data. However, if the balance of include positions reported by entities other than payments statistics are not yet available, the estibanks,suchasotherdepositoryinstitutions,bankand mates for the rest of the world in the preliminary financial holding companies, and securities brokers release of the flow of funds accounts for a given anddealers. quarterarederiveddirectlyfromtheTICdata.8 Cross-border positions of ‘‘nonbanks’’ (including entities such as exporters and importers, industrial firms, insurance companies, and pension funds) are ISSUES IN THE COLLECTION AND collected quarterly, by country, on the TIC C forms. INTERPRETATION OF THE TIC SECURITIES The C forms distinguish between ‘‘financial’’ claims DATA andliabilities(suchasdeposits,short-termsecurities, and loans) and ‘‘commercial’’ claims and liabilities WhilerecognizingthattheTICsystemcoversavari- (such as accounts receivable or payable arising from ety of cross-border financial transactions, we will import or export activities). Compared with the data focus in the remainder of this article on interpreting reported on the B forms, the C data report much theTICdataonsecurities—thatis,themonthlytranssmaller cross-border positions. As of year-end 2004, total bank-reported claims and liabilities (excluding short-term securities) were about $2 trillion and 6. Thisproblemaffectscross-borderdatacollectionnotonlyinthe $2.4trillionrespectively.Incontrast,thecorrespond- United States but also in other countries. For example, an International Monetary Fund conference on capital flow and debt statistics ing amounts for nonbanks were only about $200 bilpointedtoageneraldifficultyinobtainingaccurateandtimelyinforlion and $100 billion. In part, these smaller reported mationonthecross-borderactivityofnonbankcommercialconcerns. positions illustrate the difficulty of collecting accu- Refertotheconferencesummary,note13,atwww.imf.org/external/ pubs/ft/seminar/2000/capflows/summary.htm. 7. TheBEA’sdataoninternationalaccounts,includingthebalance of payments accounts and the international investment position, are publishedintheSurveyofCurrentBusinessandattheBEA’swebsite 5. Thedatesofthesurveysarestaggeredprimarilytoreducethe (www.bea.gov/bea/di1.htm). year-end reporting burden on the institutions that report the survey 8. The flow of funds accounts are published at data. www.federalreserve.gov/releases/z1/current/default.htm.

Understanding U.S. Cross-Border Securities Data A63 actions data on long-term securities, the monthly survey, foreign holdings attributed to that category position data on short-term securities, and the annual amountedtoalmost$200billion. survey data. The following sections discuss topics Another problem of country attribution occurs in related to the design and accuracy of the TIC system the reporting of monthly transactions data. The thatshouldbeunderstoodforproperinterpretationof monthly transactions data, by design, record purthesedata.Butbecausecross-borderfinancialinflows chases and sales against the country from which can come through various means, including through transactions are made, which is not necessarily the the banking system and through direct investment, it country of the ultimate purchaser or actual seller (in is important to view the cross-border securities data the case of foreign transactions in U.S. securities) or inthisbroadercontext. the country of issuance (in the case of U.S. transactionsinforeignsecurities).Thisreportingconvention means that if, for example, a resident of Germany Country Attribution buys a U.S. Treasury bond and the transaction is bookedthroughaLondonbroker,theTICSdatawill The collection of accurate country-level data on show a net purchase of a Treasury bond recorded cross-border financial activity ranges from straight- againsttheUnitedKingdom,notGermany.Likewise, forward to virtually impossible, depending on the if a U.S. investor purchases French equity from a type of data to be collected and the method of dealer in Switzerland, the TIC S data will report a collection. U.S.netpurchaseofforeignequityfromSwitzerland. The country attribution of the portfolio asset sur- As transactions tend to be concentrated in major veys should be extremely accurate. The annual posi- international financial centers, such as the United tion surveys, by design, attempt to collect informa- Kingdom and the Cayman Islands, the monthly data tion by country of issuer for foreign securities and show a significant financial center ‘‘transactions by country of foreign owner for U.S. securities. And bias’’ that often gives an inaccurate picture of the because the surveys collect data at the level of indi- nationalityoftheactualforeignbuyersandsellers. vidual securities, precisely identifying each security issuer’scountryofresidence—frominformationsuppliedbysurveyreportersaswellasfromcommercial Foreign Official Institutions datasources—isarelativelystraightforwardtask. Intheliabilitiessurveys,however,theinvolvement Data on foreign ownership of U.S. securities are ofchainsofintermediariesinthecustodyormanage- divided into holdings of foreign official institutions mentofsecuritiesfrequentlymakesaccurateidentifi- andholdingsofforeignprivateinvestors.Contraryto cation of the actual owners of U.S. securities impos- the assumptions of many data users, the holdings of sible.Forexample,aresidentofItalymaybuyaU.S. foreign official institutions as reported in the TIC securityandentrustittoacustodianbankinSwitzer- system consist of more than the foreign reserve asset land. The Swiss bank, in turn, will typically employ holdings of central banks and of other foreign govthe services of a U.S.-resident custodian to facilitate ernment institutions involved in the formulation of settlementandcustodyoperations.Whensurveysare international monetary policy. They also include the conducted, information is collected only from U.S.- holdings of foreign government-sponsored investresident entities. Thus, the U.S.-resident bank, acting ment funds and other foreign government instituasthesubcustodianoftheSwissbank,willreportthis tions, and thus they may differ from data on reserve security on the survey. Because the U.S. bank will assetholdingsfoundelsewhere. typically know only that it is holding the security on The distinction between foreign official and other behalf of a Swiss bank, it will report the security as foreign investors is made because the motivations of Swissheld.Thispracticetendstocreatea‘‘custodial official investors for holding U.S. securities may difbias’’ in the liabilities surveys by attributing exces- ferfromthoseofprivateinvestors.Therapidbuildup sively large holdings to countries that are major cus- in U.S. liabilities since 2001 is due in part to the todial, investment management, or security deposi- substantial acquisition by foreign official institutions tory centers, such as Belgium, the Cayman Islands, ofU.S.long-termsecurities,especiallylong-termU.S. Luxembourg, and Switzerland. An additional prob- Treasury and U.S. government agency securities. By lem is caused by bearer, or unregistered, securities. year-end 2005, foreign official institutions are esti- Because no information is typically available on the matedtohaveheldapproximately$1.8trillionofthe ownership of these securities, they are listed on the total$6.7trillioninU.S.long-termsecuritiesheldby surveys as ‘‘country unknown.’’ In the June 2005 allforeigninvestors(figure4).

A64 Federal Reserve Bulletin 2006 4. Total foreign, foreign private, and foreign official 2. MarketvalueofforeignofficialholdingsofU.S. holdings of U.S. long-term securities, 1994–2005 long-termTreasurybonds:Comparisonofsurveyresults withestimatedvalues,forsurveydates,2000–05 Billions of U.S. dollars Billionsofdollars 7,000 Monthandyear Survey1 Estimate2 6,000 Mar.2000 ........................ 492 436 Total foreign holdings June2002 ........................ 561 454 5,000 June2003 ........................ 653 605 June2004 ........................ 923 846 4,000 June2005 ........................ 1,073 1,028 3,000 Note: Foreign official holdings consist of foreign reserve asset holdings, Foreign private holdings holdingsofforeigngovernment-sponsoredinvestmentfunds,andholdingsof 2,000 foreigngovernmentinstitutionsnotinvolvedintheformulationofinternational monetarypolicy. Foreign official holdings 1,000 1. SurveysofforeignholdingsofU.S.securities(liabilities). 2. Resultofaddingmonthlytransactionsatmarketvaluetothemarketvalue 0 amountsmeasuredbythemostrecentsurvey. Source: Forsurveyresults(exceptforJune2005),U.S.Departmentofthe Treasury (2004), Report on Foreign Portfolio Holdings of U.S. Securities, 1995 1997 1999 2001 2003 2005 June30;resultforJune2005isfromthepreliminaryreleaseofthe2005survey SOURCE: Treasury International Capital reporting system and staff offoreignportfolioholdingsofU.S.securities.Forestimatedvalues,Treasury estimates. International Capital reporting system, ‘‘Major Foreign Holders of Treasury Securities’’(table),www.treas.gov/tic/mfh.txt. Accurately distinguishing official from private calendar month or against the same country as was holders,however,isdifficultforthesamereasonsthat themovementintoU.S.custody.9 obtaining accurate information on the country of foreign owners of U.S. securities is difficult. Chains of intermediaries can obscure not only the country Effects of Exchange Rate Changes and but also the type of foreign holder. Thus, foreign Other Valuation Adjustments officialholdingsarealmostcertainlyundercountedin theTICdata,thoughthedegreeofundercountisless Although many users of the TIC data assume that intheannualsurveysthaninthemonthlytransactions foreign securities held by U.S. investors are excludata. The undercount in the transactions data is evi- sively denominated in foreign currencies and that dentwhentheresultsofanewannualsurveybecome U.S. securities held by foreign investors are excluavailable: Frequently, measured official holdings in sively denominated in U.S. dollars, such is not the the new survey exceed those derived from summing case.Accordingtothemostrecentdataavailable(for the monthly transactions since the previous survey 2004),74percentofthe$1.2trillioninU.S.holdings (table2). of foreign debt securities were denominated in U.S. To understand how foreign official acquisitions of dollars, whereas 12 percent of the $4.1 trillion in U.S. securities may be undercounted, consider the foreignholdingsofU.S.debtsecuritiesweredenomifollowingexample.Supposeaforeignofficialinstitu- natedinforeigncurrencies.10 tion acquires a U.S. security, such as a U.S. Treasury Accounting correctly for the currency denominabond, from a private foreign entity on a foreign tions of U.S. holdings of foreign securities and of securities exchange and then has the security moved foreign holdings of U.S. securities allows for moreto the United States to be held in custody at the accurate measurement of valuation changes resulting Federal Reserve Bank of New York. In this case, the from exchange rate fluctuations. In any given year, surveys will report increased holdings of Treasury theeffectsofsuchfluctuationscanbelargerthantotal securitiesbyforeignofficialinstitutions,butnocorre- net securities flows during the year. For example, sponding foreign official purchase will be recorded given the level of cross-border holdings of both on the TIC S because the acquisition by the foreign officialinstitutionfromanotherforeignerisnotaU.S. 9. For more information, including a comparison of TIC data on cross-border transaction; rather, it is a foreign-to- foreignofficialholdingsofTreasuryandagencysecuritieswithFedforeign transaction. Note, however, that when the eral Reserve data on such securities held in custody at the Federal ReserveBankofNewYorkforofficialaccounts,referto‘‘Frequently private foreigner first acquired the Treasury security, Asked Questions (FAQs) regarding the TIC System and TIC data,’’ a U.S. cross-border transaction should have been FAQ10,www.treas.gov/tic/faq1.html. reported in the TIC system. But it would not have 10. For the currency denominations of U.S. holdings of foreign securities,refertowww.treas.gov/tic/shc2004r.pdf,p.11;forthecurbeen recorded as a foreign official purchase, nor rency denominations of foreign holdings of U.S. securities, refer to would it necessarily have been recorded in the same www.treas.gov/tic/shl2004r.pdf,p.117.

Understanding U.S. Cross-Border Securities Data A65 equity and debt securities at the end of 2004, a cent respectively, bringing the corresponding market 10percentappreciationinthevalueoftheU.S.dollar valuesofsuchholdingsto$259billionand$458bilwould have resulted in a net loss of $250 billion to lion. Estimates of the repayment flows associated theU.S.balancesheet,asitwouldhavedecreasedthe with foreign holdings of asset-backed securities are value of U.S. holdings of foreign securities (U.S. published on the TIC website.12 These repayment assets) denominated in foreign currencies approxi- flows can be sizable: For 2005, they are estimated to mately $300 billion while decreasing the value of have reduced foreign holdings of U.S. agency bonds foreign holdings of U.S. securities (U.S. liabilities) by $48 billion and such holdings of corporate bonds only about $50 billion (because most U.S. securities by $38 billion. U.S. holdings of foreign asset-backed are U.S.-dollar-denominated and thus unaffected).11 securities are still relatively small but have been Other asset price changes, such as changes in the increasing. value of U.S. or foreign equities, can also have siz- Unlike asset-backed securities, which gradually ableeffectsonthevalueofsecuritiesholdings. decline in value, zero-coupon securities gain value over time as they accumulate implicit interest payments. Again, these increases in value will not be Asset-Backed and Zero-Coupon Securities captured by the monthly transactions system but will be measured by the annual surveys. Cross-border When estimating the value of cross-border securities holdings of zero-coupon securities are much smaller holdings,analystsshouldtakeintoaccountmorethan than such holdings of asset-backed securities. the measured levels of holdings, the measured levels According to the asset and liabilities surveys conof transactions, and fluctuations in prices and ducted in 2004, foreign investors held zero-coupon exchange rates. The reason is that some securities— U.S.securitieswithamarketvalueof$20billionand namely, asset-backed and zero-coupon securities— a face value at maturity of $30 billion. The corregain or lose value over time even if there are no spondingfiguresforU.S.investors’holdingsofzerocross-border transactions and prices and exchange coupon foreign securities were $10 billion and ratesremainstable. $16billionrespectively. Because many asset-backed securities repay principal on a regular basis, they decrease in value over time. These principal repayment flows are not Offshore Financial Centers recorded by the TIC S monthly transactions system, but the effects of these principal repayments on the An institution is considered to be resident in the value of asset-backed holdings are measured by the country in which it is incorporated or otherwise annual position surveys. Asset-backed securities legally created. In many cases, residency and the account for a growing share of foreign holdings of center of economic activity coincide. But when they U.S. government agency debt and U.S. corporate differ, problems of data interpretation arise. For debt: As of June 30, 2002, asset-backed securities example, companies frequently create corporations represented 25 percent of foreign holdings of U.S. and ‘‘special-purpose vehicles’’ (SPVs) in so-called agency securities and 15 percent of foreign holdings offshorefinancialcenterstotakeadvantageofthetax of corporate debt securities. By June 30, 2005, these or regulatory benefits that these countries offer.13 proportions had increased to 33 percent and 26 per- Whentheseentitiesissuesecurities,theissueswillbe 12. Estimatesofmonthlyasset-backedrepaymentflowssinceJune 11. Theeffectofa10percentappreciationintheU.S.dollarwillbe 2002areavailableatwww.treas.gov/tic/absprin.html. perceiveddifferentlydependingonwhethertheseholdingsareviewed 13. A special-purpose vehicle is a legal entity created in an offas U.S. liabilities (the U.S. perspective) or as foreign assets (the shore financial center (OFC) to engage in financial activities in a foreignperspective).Forexample,assumethatforeignresidentsown low-taxenvironment.AnonshorecorporationestablishesanSPVin $100inU.S.securities,80percentofwhicharedenominatedinU.S. anoffshorecentertoengageinaspecificactivity,suchastheissuance dollars and 20 percent of which are denominated in a foreign cur- ofasset-backedsecurities.Theonshorecorporationmayassignaset rency.Further,assumethattheU.S.dollarappreciates10percentwith ofassetstotheoffshoreSPV(forexample,aportfolioofmortgages, respecttothisforeigncurrency.FromtheU.S.perspective,theappre- loans, or credit card receivables). The SPV then offers to investors ciation will have the effect of reducing the value of U.S. liabilities avarietyofsecuritiesbasedontheunderlyingassets.TheSPV,and 2percent(the80percentdenominatedinU.S.dollarsisunaffected; hencetheonshoreparent,benefitfromthefavorabletaxtreatmentin the20percentdenominatedintheforeigncurrencydecreasesinvalue the OFC. Financial institutions also use SPVs to take advantage of 10percent;thetotaleffectisa2percentreductioninvalue).Fromthe less-restrictiveregulationsontheiractivities.Banks,inparticular,use foreignperspective,however,thevalueofforeignassetshasincreased themtoraiseTierIcapitalinthelowertaxenvironmentsofOFCs. 8 percent (the 20 percent denominated in foreign currency is unaf- Andnonbankfinancialinstitutionscreatethemtotakeadvantageof fected;the80percentdenominatedinU.S.dollarsincreasesinvalue more-liberalnettingrulesthanthosefacedinhomecountries—rules 10percent;thetotaleffectisan8percentincreaseinvalue). thatenablesuchinstitutionstoreducetheircapitalrequirements.

A66 Federal Reserve Bulletin 2006 attributed to the country of the offshore financial ters and held by U.S. investors, the U.S.-dollarcenterratherthantothecountryoftheonshoreparent denominated share of foreign long-term debt has corporation, even though the onshore parent corpo- increased in recent years, from 60 percent in 1997 to ration may be understood to be the ultimate obligor. 72percentbytheendof2004. Further, some companies have reincorporated from The increase in U.S. holdings of equity and debt their country of origin to offshore financial centers issued by offshore centers raises questions about the for tax purposes. Although the reincorporation prob- interpretation of such securities in the U.S. portfolio. ably has little or no effect on their locus of activity, Although these securities fit the definition of foreign securities issued by these companies will now be securities, U.S. investors may not regard them as attributedtothecountryofreincorporation. such, as they trade in U.S. dollars on U.S. exchanges U.S.holdingsofsecuritiesissuedinoffshorefinan- and are often issued by firms that conduct their cialcenters,especiallythoseintheCaribbean,posea marketactivitylargelyintheUnitedStatesandotherchallenge to measuring and interpreting U.S. inves- wise behave like U.S. firms. Likewise, when foreign tors’ portfolios. Equity issued in offshore centers, in investors acquire such securities, they may consider largepartreflectingtheequityofreincorporatedmul- themequivalenttoU.S.securities. tinationals and other entities controlled by onshore On the U.S. liabilities side, the acquisition of U.S. corporations, accounts for a growing percentage of securities by entities in offshore financial centers— the U.S. portfolio of foreign assets: In 1997, U.S. especially those in the Caribbean—poses additional holdings of equity issued by Caribbean offshore obstacles to interpreting cross-border financial activfinancial centers (Bahamas, Bermuda, British Virgin ity. Such activity partly reflects the importance of Islands, Cayman Islands, Netherlands Antilles, and these Caribbean countries as international financial Panama) amounted to $48 billion, or roughly 4 per- centers, and purchases and sales recorded against cent of all foreign equity held by U.S. investors. By these regions typically represent the first leg of a the end of 2004, these amounts had grown to seriesofinternationaltransactions.Itmayalsoreflect $277billion,ornearly11percentofallforeignequity the buying and selling of securities by the numerous held. investment funds that have been established in such ThegrowingshareoftheseholdingsinU.S.inves- offshorelocations.Moreover,becausemanyfinancial tors’ foreign equity portfolios affects the degree to institutions have affiliated banking and nonbanking which their portfolios are exposed to exchange rate offices in these offshore locations, analyzing securirisk. An increase in the share of foreign equity in an ties transactions through these centers can be diffiinvestor’s portfolio usually raises foreign currency cult without knowing whether offsetting transactions exposure. Equity issued through offshore financial are occurring through other parts of the financial centers, however, is typically either dollar denomi- accounts. For example, net financial outflows in the nated or denominated in currencies pegged to the form of net sales of U.S. securities through financial dollar, and so it carries a different exchange rate centers may be offset by equally sizable net inflows exposure. Although foreign equity still represents a reported in the TIC banking data from the same fairlysmallshareofthetotalequityportfolioheldby financialcenters. U.S. investors, the share has grown in recent years, Caribbean financial centers are an increasingly from about 9 percent in 1997 to about 14 percent in important location for cross-border transactions in 2004. But the increase in foreign currency exposure U.S. securities: Gross trading in long-term U.S. secuof U.S. equity portfolios has been more modest, as rities through these centers has grown from less than more than 1 percentage point of that 5 percentage 10percentoftotalcross-bordertradesinthefirsthalf pointincreasehasbeenattributabletotheacquisition of the 1990s to nearly 30 percent in the past five ofequityissuedinCaribbeanoffshorecenters. years. Net transactions through Caribbean financial Holdings of long-term debt issued in offshore centerscanalsobequitevolatile,astheymayrecord financial centers present a different challenge. Here large foreign net purchases of long-term securities U.S.holdingsconsistlargelyofdebtsecuritiesissued onemonthandlargeforeignnetsalesthenext. through SPVs, especially those established in the Cayman Islands. Such holdings of Cayman Island debt by U.S. investors amounted to about 2 percent Repurchase and Securities-Lending Agreements of U.S. investors’ holdings of foreign bonds in 1997 but had grown to nearly 12 percent by 2004. Partly Repurchase agreements, or repos, are arrangements because of the growth of both asset-backed and con- whereby the owner of a security sells it for cash ventional debt issued through offshore financial cen- and agrees to repurchase it at a future time (or under

Understanding U.S. Cross-Border Securities Data A67 other specified conditions) at an agreed-upon price. reportlargerforeignownershipofU.S.securitiesthan Although some market participants engage in repo it would if these agreements were recorded as outtransactionstogaincontrolofcertainsecurities,repos right sales, since there is both an economic owner areoftenstructuredascashloansfortraderswhouse andalegalownerofthesamesecurity.AlthoughTIC thecashreceivedintherepotransactionasalow-cost instructions specify that only the economic owner loan to fund their securities purchases, whereas the should be reported, TIC-reporting entities may lack lenders receive the securities as collateral against sufficient information to follow these instructions borrower default. The securities typically used as properly. Further, the legal owner has the right to collateralareU.S.Treasurysecuritiesand,toalesser resell a borrowed security, and the subsequent buyer extent, U.S. government agency and corporate debt ortheinstitutionreportingonbehalfofthebuyermay securities. have no knowledge that the security was originally Securities-lendingagreementsaresimilartorepur- borrowed. This situation can result in two different chase agreements in that the owner transfers title of foreign residents being reported on a liabilities surthe securities to a borrower who agrees to return a veyasholdingthesameU.S.security,oritcanresult like quantity of the same or similar securities at a in the same U.S. security being reported as having future date or under other agreed-upon conditions. been purchased twice by foreign residents with no Again,theborrowerprovidescollateral,butunlikein interveningsale. the case of repos, for which securities are used as collateral, the collateral for securities-lending agreements can be cash, other securities, or bank-issued Comprehensiveness of the Data letters of credit. Many market participants engage in securities-lending transactions to obtain securities In general, the data on U.S. liabilities are considered needed to meet delivery obligations; for example, to be reasonably comprehensive, as debt instruments brokersmayneedtocoverafailedtrade,orinvestors tendtobeissuedbyandboughtorsoldthroughlarge may want to create a ‘‘short’’ position.14 Both equity institutions that can be fairly readily identified and and debt securities are involved in securities-lending included in the data reporting network. U.S. foreign arrangements. assets held by or through large U.S. institutions Repurchaseandsecurities-lendingagreementspose should also be well recorded. However, for smaller a problem for the TIC system. Although both U.S. investors, directly purchasing foreign securities arrangements involve the outright sale of securities, abroad without using the services of a large, U.S.the TIC system does not treat them as such. Instead, residentinstitutionisincreasinglyeasy.Suchacquisiit treats them as collateralized loans, as the return of tions will not be captured in the U.S. recording systhe same or similar securities at a set price is agreed tem but will most likely be recorded as liabilities upon in advance and thus the economic risk of by the counterparty country’s measurement system. holding the securities continues to reside with the Because all countries face this problem, cross-border securities lender (the economic owner) even while assetsareprobablyundercountedworldwide. the lender does not own the securities. Repurchase andsecurities-lendingagreementsarenotrecordedas purchases or sales of securities in the monthly transactions data. Instead, funds loaned to or borrowed Stock Swaps from foreigners under such agreements are reported on the TIC B forms. For the benchmark surveys, An additional problem is that the TIC S data fail to lenders (or their custodians) are instructed to report captureU.S.acquisitionsofforeignstockandforeign thesecuritiesascontinuouslyheld,andborrowers(or acquisitionsofU.S.stockthatarisefromstockswaps their custodians) are instructed not to count them as associated with corporate mergers or takeovers. holdings. When a foreign company acquires a U.S. company As a result of treating securities-lending agree- andthedealisfinancedinpartthroughastockswap, ments as collateralized loans, the TIC system may U.S. residents who held stock in the target companybecomeholdersofforeignequity.Likewise,ifa U.S. company acquires a foreign company, a stock 14. Ashortsaleisthesaleofasecuritynotowned.Securitiesare borrowed—typically from a brokerage firm—and then sold in the swap can increase foreign holdings of U.S. equity. hope that the price of the security will fall. If the price drops, the Although missing from the TIC S data, stock swaps security can then be bought at the lower price and returned to are reported in the BEA’s quarterly balance of paythelenderataprofit.Conversely,ifthepriceofthesecurityrises,a lossisincurred. ments statistics, and monthly estimates of swaps

A68 Federal Reserve Bulletin 2006 based on these statistics are provided on the TIC consistentwiththesurveypositionstakenatannual— website.15 or, as in the past, at less frequent—intervals. The Although merger activity has tapered off in recent procedure is based on that described in a recent years, stock swaps previously were an important Federal Reserve Bulletin article and extended in a source of financing such activity. For example, in related research paper by Thomas, Warnock, and 2000, U.S. residents acquired $13 billion in foreign Wongswan (2004), although the numerical computaequity through net purchases but $80 billion through tion procedure has been simplified from that prestock swaps associated with foreign acquisitions of sentedinthepaper.16 U.S.companies. Basic Position Estimate Adjustments to the Transactions Data Webeginwithanillustrationofthebasicsituationfor Wehaveraisedanumberofcaveatsthatusersshould total U.S. long-term securities held by foreigners be aware of when using the TIC securities data. (figure 5). In the figure, the dots show the actual However, users can take some straightforward steps survey values for total foreign holdings of U.S. longthat will help them use the published data more term securities. Note that the length of time between effectively. Users can obtain more-comprehensive surveysvariesfromonetofiveyears.Theblacklines estimatesofcross-bordersecuritiesflowsbyincorpo- show ‘‘naive’’ position estimates obtained by sumrating the estimates described earlier for principal ming monthly net transactions from the date of repayment flows of asset-backed securities and for the previous survey. (As discussed above, we have stock swaps. Combining flows of short-term securi- reduced the naiveté of the estimates by adjusting the tieswiththetransactionsdataonlong-termsecurities net transactions for principal repayment flows of can also improve coverage. In addition, users should asset-backedsecuritiesandforstockswaps.) be aware of the problem of financial center transactions bias when attributing securities flows to individual countries. Because survey data on securities 16. Refer to William L. Griever, Gary A. Lee, and Francis E. Warnock (2001), ‘‘The U.S. System for Measuring Cross-Border holdings are believed to be more accurate than the Investment in Securities: A Primer with a Discussion of Recent higher-frequency transactions data, combining the Developments,’’FederalReserveBulletin,vol.87(October),pp.633– two sources can substantially improve one’s under- 50, www.federalreserve.gov/pubs/bulletin/2001/1001lead.pdf; and Charles P. Thomas, Francis E. Warnock, and Jon Wongswan, ‘‘The standing of both the magnitude and the country attri- Performance of International Portfolios,’’ International Finance Disbution of cross-border securities holdings. We cussionPapersSeries2004-817(Washington:BoardofGovernorsof explore these ideas in greater detail in the following theFederalReserveSystem,October),www.federalreserve.gov/pubs/ ifdp/2004/817/default.htm. sections. 5. Foreign holdings of U.S. long-term securities: ESTIMATING POSITIONS BY COUNTRY BETWEEN Comparison of survey values with monthly position SURVEYS estimates, 1984–2005 Billions of U.S. dollars Although the annual surveys give comprehensive measures of holdings by country at a point in time, 7,000 analysts often wish to have a time series of holdings 6,000 by country, as well as more-current measures of holdings. For short-term securities, the data from the 5,000 TICBformsprovidereasonablycomprehensivemea- 4,000 sures of current holdings. For long-term securities, 3,000 however,estimatesmustbeconstructed. 2,000 This section describes the construction of monthly estimates of asset and liabilities positions that are 1,000 0 15. DatafromSecurityDataCorporationonthefinancingofcorporate mergers and takeovers are used to distribute the quarterly 1985 1990 1995 2000 2005 statisticsofstockswapsbymonth.MonthlyestimatesfromJanuary NOTE: Dots represent positions reported in liabilities surveys. Lines 2000 through recent months are available at www.treas.gov/tic/ represent monthly position estimates obtained by adding net transactions to swapstk.html. the position of the previous survey.

Understanding U.S. Cross-Border Securities Data A69 The figure shows eight survey values. For the last 6. Foreign holdings of U.S. long-term securities: sevenofthese,wecancomparethesurveyresultwith Comparison of survey values with monthly position estimates adjusted for valuation changes, 1984–2005 the naive estimate made starting from the previous survey.Intwoyears(2000and2002),averysubstan- Billions of U.S. dollars tial ‘‘gap’’ separates the estimate and the actual surveyresult,whereasin1989and2004,amuchsmaller 7,000 gap exists, and in 1994, 2003, and 2005, virtually 6,000 no gap exists at all. (In 1984, of course, no gap is 5,000 defined, as this year is the starting point for our analysisbecauseofalackofusableearlierdata.) 4,000 3,000 2,000 Adjustment for Valuation Change 1,000 One way of improving the monthly position esti- 0 mates is to include an estimate of the valuation 1985 1990 1995 2000 2005 changefromonemonthtothenext.Startingfromthe previoussurvey,onecanapplythechangeinrelevant NOTE: Dots represent positions reported in liabilities surveys. Lines represent monthly position estimates obtained by adding net transactions to securities price indexes to the previous position to the position of the previous survey and by adjusting the result for valuation obtain an estimate of the capital gain or loss from changes. month to month. For foreign holdings of U.S. longterm securities, the composition of the security portfolio being held (for example, long-term U.S. Trea- 1990s now leads to an overestimate of actual foreign sury bonds) is known, and obtaining an appropriate holdings of U.S. securities. We also note that the price index for evaluating changes in Treasury bond adjustmentforvaluationchangesintroducesadegree prices is fairly easy. In contrast, U.S. holdings of of high-frequency variability into the estimated foreign securities comprise securities issued by positions. many different countries, greatly compounding the valuation estimation problem. To estimate valuation changes for foreign securities, we use individual Distribution of Gap Error country equity or bond price indexes for most countries, taking into account the currency composition The survey gaps represent known information that of U.S. holdings.17 However, in some cases—for can be used to estimate positions between surveys. example, for pricing the holdings of securities issued Because we believe the surveys are more likely than by offshore financial centers—the best estimate will themonthlypositionestimatestoaccuratelymeasure result from using information on the holdings of securities positions, we assume that the sum of the individual foreign securities derived from the U.S. net transactions, adjusted as described earlier, is in asset survey data to construct a customized index, as error by the amount of the gap. The gap represents nopublishedindexcurrentlyavailableisappropriate. errorsandomissionsinthemonthlytransactionsdata We present estimated monthly positions obtained of current S-form reporters, as well as transactions by adding net transactions to the previous survey conductedbyentitiesthathavenotyetbeenidentified benchmark, adjusting the results for valuation as prospective reporters. In addition, the gap may be changes (figure 6). In general, the gaps between the duetovariousmeasurementandapproximationerrors estimatedpositionsandthefollowingbenchmarksur- inconstructingthepricesusedtocalculatethevaluaveyaresmallerthaninfigure5,thoughthelargegap tionadjustment. in 2000 now switches sign: Adjusting for valuation Of course, we lack knowledge of how the errors changes during the stock market boom in the late reflected in the overall gap error are distributed over time; we know only how large the gap is and that it accumulated over the period from the previous 17. Forholdingsofbondsissuedbymostindustrialcountries,we survey to the present. To proceed, we must make an use local currency bond indexes to estimate valuation changes. For debtissuedbymostemerging-marketcountries,weuseJ.P.Morgan’s assumption about the distribution of the gap over Emerging Markets Bond Index Plus (EMBI+) indexes, which track the intersurvey period. We assume, as did Thomas, returns of debt denominated in external currencies. The EMBI+ Warnock, and Wongswan, that the gap is distributed indexesareappropriatebecausethemajorityofemerging-marketdebt heldbyU.S.investorsisdollardenominated. in proportion to the volume of transactions in each

A70 Federal Reserve Bulletin 2006 7. Foreign holdings of U.S. long-term securities: Estimated Positions Adjusted Comparison of survey values with monthly position for Transactions Bias estimates adjusted for valuation changes and gap error, 1984–2005 We present the same illustration (showing estimated monthly positions and forecast gaps) for U.S. securi- Billions of U.S. dollars ties held by investors in the United Kingdom and in 7,000 euro-area countries (figure 8). Note that the gaps for 6,000 the United Kingdom are consistently large and positive,anindicationthatthetransactions-basedmonthly 5,000 estimates consistently overstate the actual positions 4,000 eventually reported in the surveys. In contrast, the 3,000 gaps for the euro area are generally substantial and negative, a sign that the monthly estimates consis- 2,000 tentlyunderstatetheactualpositions. 1,000 The explanation for this result is the transactions 0 bias inherent in the monthly TIC S data reported for the United Kingdom: London is a major financial 1985 1990 1995 2000 2005 NOTE: Dots represent positions reported in liabilities surveys. Lines represent monthly position estimates obtained by adding net transactions to the position of the previous survey and by adjusting the result for valuation changes. Gap error is distributed in proportion to the volume of monthly 8. Holdings of U.S. long-term securities by the United transactions; the distribution allows for the effect of price changes. Dashed Kingdom and by euro-area countries: Comparison of line indicates forecast. survey values with adjusted monthly position estimates, 1994–2005 monthovertheperiodandthatthedistributionallows Billions of U.S. dollars fortheeffectofpricechanges. United Kingdom 1,800 Applying this procedure generates an estimated 1,600 monthly path for total foreign holdings of U.S. long- 1,400 term securities (figure 7). Note that the gaps have been eliminated, as the line passes through all the 1,200 dots, representing survey values. The line also 1,000 appears to exhibit attenuated versions of the line 800 variations in figure 6, illustrating that the estimation 600 procedure preserves the variability introduced by 400 valuationchanges. 200 Notealsothatadottedlineextendsbeyondthelast survey date. The extension represents an estimate of the future position based on observed transactions Euro-area countries 1,800 and prices and on a forecast of the survey gap. As 1,600 discussed in the next section, the estimates for many 1,400 countries seem to exhibit gaps that tend to be either 1,200 consistently positive or consistently negative. Accordingly, we forecast the next future survey gap 1,000 by using the simple average of the previous two 800 actualgaps,scaledbynominalpositionandbytime.18 600 400 200 18. Wefirstscalethegapsbynominalposition(theactualsurvey value)toconvertthenominalmagnitudeoftherawgapstoaunits- 1995 1997 1999 2001 2003 2005 freemeasure,allowingcomparisonsacrosscountriesandacrosssurveys.Wethenscalethegapsbytime(thenumberofmonthsbetween NOTE: Dots represent positions reported in liabilities surveys. In each panel, the line connecting all the dots represents monthly position estimates surveys), recognizing that errors in measuring the true position are adjusted for valuation changes and gap error as described in the note to additive from month to month. Applying this second normalization figure 7; the lines extending separately from the dots represent monthly significantlyreducesthemagnitudeofthegapsintheearlypartofthe position estimates adjusted only for valuation changes. Dashed lines indicate sample. forecasts.

Understanding U.S. Cross-Border Securities Data A71 center, and financial intermediaries in London often 9. U.S. holdings of foreign long-term debt securities: buy U.S. securities on behalf of customers in other Comparison of survey values with adjusted monthly position estimates, 2001–04 countries, many of whom are in the euro area. This pattern, which emerges quite clearly in the data, has Billions of U.S. dollars significant implications for forecasting the positions ofindividualcountries.Asshownbythedottedblack line in the first panel of figure 8, the projected U.K. 1,200 positionrisesbyabout$90billionbyyear-end2005, 1,000 reflecting the typical historical pattern, but this increase is less than what we would estimate on the 800 basis of flows and valuation adjustments alone (indicated by the line extending from the nearest dot). In 600 contrast, the projection for the euro area indicates an increase in holdings of about $200 billion, about 400 $130 billion more than would be implied by flows andvaluationadjustments. 2001 2002 2003 2004 NOTE: Dots represent positions reported in asset surveys. The line connecting all the dots represents monthly position estimates adjusted for Applications of Position Estimates valuation changes and gap error as described in the note to figure 7; the lines extending separately from the dots represent monthly position estimates adjusted only for valuation changes. Besides providing between-survey and more-recent estimates of both foreign holdings of U.S. securities United States, and indeed, further investigation of and U.S. holdings of foreign securities—by country S-form reporters indicated reporting errors in this and with an adjustment for transactions bias—our area. Although most respondents on the S form have estimates can also help identify potential problems not revised their reports for omissions in 2002 and withtheTICdatasystemandimprovereporting.Asa 2003,S-formreportingappearstocapturenewissues new survey becomes available, we can compare our offoreignbondsmorecompletelyforthedatabeginestimates,constructedforwardfromtheprevioussurning in 2004. The results from the December 2004 vey,withthereportedvaluesinthenewsurvey.Ifthe survey have been encouraging: Measured U.S. holdestimates are considerably different from what the ings of foreign debt came in much closer to the survey indicates, the difference suggests that errors estimatedpositions. exist in the transactions data, the valuation changes applied over the intersurvey period, the survey results,orsomecombinationofthesefactors. COMPARING U.S. LIABILITIES ESTIMATES The experience of the December 2003 survey of WITH CPIS ASSET POSITIONS U.S.holdingsofforeignsecuritiesillustrateshowthis approach can identify missed reporting and improve Although our estimated between-survey positions of the TIC reporting system. Estimating forward from foreign holders of U.S. securities are corrected for theDecember2001surveygeneratesavalueforU.S. transactions bias, they still suffer from the custodial holdings of about $650 billion in foreign long-term bias present in the liabilities surveys. However, we debt as of December 2003—slightly higher than the can perhaps gain a better understanding of the true roughly $600 billion measured at the end of Decemowners of U.S. securities—and of the possible holdber 2001—which resulted from small reported net ers of U.S. bearer bonds—by comparing our liabilisales of foreign debt over the two-year period plus ties positions with holdings of U.S. securities as positive valuation changes (figure 9). That estimate reported by other countries in the Coordinated Portwas significantly lower than the figure reported in folio Investment Surveys (CPIS), which are conthe December 2003 survey, which valued U.S. holdducted annually under the auspices of the Interings of foreign long-term debt securities at nearly national Monetary Fund (IMF).19 The CPIS asset $870 billion, about $220 billion more than expected. surveys represent a commitment to collect and pub- Anin-depthinvestigationofthesecuritiesreportedin the survey showed that U.S. investors held a sizable amount of newly issued foreign debt. These findings 19. BecausetheCPISassetsurveysareconductedannuallyasof December31,whereastheliabilitiessurveysareconductedannually suggestedthatanareaofmissedreportingmostlikely as of June 30, we compare the CPIS measures with our betweeninvolved new issues of foreign securities in the surveyestimatesofholdingsasofDecember31.

A72 Federal Reserve Bulletin 2006 lishcomprehensivedataonforeignportfoliosecurity 3. ForeignholdingsofU.S.long-termdebtsecurities: holdings. Approximately seventy countries partici- Estimatedforeignassets,estimatedU.S.liabilities,and thedifferencebetweentheestimates,forCPIS-reporting pate in the CPIS, measuring and reporting, by councountries,year-end2001–04 try, their domestic investors’ portfolio holdings of Billionsofdollars equity, long-term debt, and short-term debt; the U.S. asset survey is the United States’ contribution to this Marketvalueofholdings Item cross-countryeffort.20 2001 2002 2003 2004 Becausetheindividualcountryassetsurveysinthe Foreignassets ....................... 1,695 1,948 2,420 3,094 CPIS are expected to be accurate in terms of country CPISsurveys(nonreserves)........ 1,165 1,369 1,710 2,010 attribution,theirestimatesofholdingsofU.S.securi- Reserves1 ........................ 530 580 710 1,084 ties will be free of custodial bias. The absence of U.S.liabilities2...................... 1,899 2,164 2,532 3,131 suchbiasshouldpermitthecomparisonofeachCPIS Difference(liabilitieslessassets) ..... 204 216 112 37 country’s reported holdings of U.S. equity and long- Note: Coordinated Portfolio Investment Surveys (CPIS), also known as termdebtwithourreportedliabilitiesdata.Butsome CPISsurveys,areconductedannuallyasofDecember31,whereasU.S.liacomplications attend this comparison. First, the bilities surveys are conducted annually as of June 30. Liabilities estimates representbetween-surveyestimatesofholdingsasofDecember31. amounts of U.S. securities reported in the CPIS sur- 1. Estimated for each CPIS-reporting country from aggregate reported veysincludeonlynonreserveholdings,andthusthey reserve holdings of U.S. long-term debt securities in dollars as reported by the International Monetary Fund (IMF). For a description of the estimation will not be comparable to our liabilities measures, procedure,refertotextnote23. whichincludesizableforeignofficialholdings.21Sec- 2. Nonreserve and reserve asset holdings as well as a prorated share of unallocatedbearerbonds. ond,theCPISsurveyswillalsosufferfromtheprob- Source: For CPIS surveys, IMF, www.imf.org/external/np/sta/pi/cpis.htm; lem that asset surveys generally are not as compre- for reserves, IMF data (refer to text note 23); for U.S. liabilities, Treasury InternationalCapitalreportingsystem,www.treas.gov/tic/fpis.html. hensiveintermsoftheircoverageofholderswithina countryasareliabilitiessurveys.Finally,somecountriesthataremajorforeignholdersofU.S.securities, compare,countrybycountry,holdingsofU.S.securias identified in the U.S. liabilities surveys, do not ties as reported in the U.S. liabilities surveys with participateintheCPIS.22 thosereportedintheCPIS,weimputeanestimateof reserves in dollars for each CPIS-reporting country, using aggregate reported reserve holdings of U.S. Foreign Holdings of U.S. Long-Term Debt long-termdebtsecuritiesindollarsasreportedbythe Securities IMF.23 Because official reserves are rarely held in the form of equity, the omission of reserve holdings from the 23. The IMF-published reserve holdings consist of data on total CPIS surveys is more problematic for comparing reserveholdingsintheformofU.S.long-termsecuritiesasreported intheIMFSEFERsurvey,dataontotalreserveslessgoldasofthe holdings of U.S. debt securities than for comparing variousyear-enddatesforeachcountryasreportedinInternational holdings of U.S. equities. Nonreserve holdings of Financial Statistics, and data on the estimated fraction of reserves U.S. long-term debt securities as measured by CPIS- held in dollars as recorded in the IMF database known as COFER (CurrencyCompositionofForeignExchangeReserves).Westartby reporting countries were $1.2 trillion at year-end usingIMFCOFERdatatoconstructthefractionofreservesheldin 2001 and had grown to more than $2 trillion by dollars.Forindustrialcountries,theCOFERdataindicatethatroughly year-end2004(table3).Althoughwecandistinguish 76 percent of reserves were dollar denominated at year-end 2001, 72percentatyear-end2002,73percentatyear-end2003,and72perofficial holdings from private holdings in the U.S. cent at year-end 2004. We assume that all industrial countries held liabilities surveys (and in our year-end estimates these fractions of their nongold reserves in dollars. For all other based on these survey values), this distinction is countries, the COFER data indicate that, of the reserves for which currencywasidentified,70percentweredollardenominatedatyearconfidential at the country level. Because we want to end2001,65percentatyear-end2002,62percentatyear-end2003, and 61 percent at year-end 2004. We assume that all non-industrial countries held these fractions of their nongold reserves in dollars. 20. The IMF collects and publishes these data on its website at Thesecalculationsgiveanestimateofroughly$1.5trillioninaggrewww.imf.org/external/np/sta/pi/cpis.htm (select ‘‘CPIS Data’’). Also gatedollarreservesasoftheendof2003. availableisdescriptiveinformationabouteachcountry’ssurveytech- Because not all of these estimated dollar reserves are held in nique(select‘‘CPISMetadata’’),whichcanbeusedtohelpjudgethe long-termU.S.debtsecurities,wethenproratetheestimatedholdings qualityandcomprehensivenessofthecountrysurveys. of reserves in dollars per country by the proportion of SEFER- 21. ThetablesintheIMF’sSurveyofGeographicalDistributionof reported U.S. long-term debt securities in each period to estimated SecuritiesHeldasForeignExchangeReserves(SEFER)reportaggre- aggregatedollarreserves.FormoreinformationontheIMFdata,refer gatedataontotalU.S.securitiesheldasreservesbySEFER-reporting to the SEFER webpage, www.imf.org/external/np/sta/pi/sec.htm; countries;thedataarenotreportedbycountry. International Monetary Fund (2006), International Financial Statis- 22. Notably,theCPISexcludesmainlandChina,Taiwan,andmost tics, February; and the COFER webpage, www.imf.org/external/np/ MiddleEasternoil-exportingcountries. sta/cofer/eng/index.htm.

Understanding U.S. Cross-Border Securities Data A73 Adding an estimate of dollar reserves held in the countries’ combined holdings of U.S. long-term debt form of U.S. long-term debt securities increases the securitiesbyabout$660billion.24 CPIS-based estimates considerably. These estimates In contrast, our liabilities estimates for U.S. longare still somewhat lower than those based on the term debt securities are smaller than the CPIS-based liabilities surveys, although the gap has narrowed in estimates for Japan (by more than $120 billion in recent years. The estimates based on CPIS asset 2004) and for a number of other countries, most holdingsplusestimatedreserveholdingsofU.S.long- notably the Channel Islands, the United Kingdom, term debt securities give figures of $1.7 trillion in Bermuda, Hong Kong, Italy, and France. Taken suchsecuritiesasoftheendof2001and$3.1trillion together, the CPIS data suggest that our liabilities as of the end of 2004. In comparison, total holdings estimates understate investment in U.S. long-term of U.S. securities as reported in the U.S. liabilities debt securities in this group of countries by more surveys for the set of CPIS-reporting countries than $480 billion in 2004. These results indicate that (including a prorated share of the unallocated bearer atleastsomeofthecustodialbiasoverstatementmost bonds identified in the liabilities surveys) are evident for Belgium, Luxembourg, and the Cayman $1.9trillionattheendof2001and$3.1trillionatthe Islands likely reflects ultimate beneficial ownership end of 2004. Overall, the difference between the two by investors in the countries for which our estimates estimates of aggregate foreign holdings of U.S. secu- are understated, although a portion may also reflect rities is small enough to suggest that the country-by- ownership by investors in countries not included in country estimates are also reasonably accurate, and theCPISsurveys. indeed, we find for many countries that the CPISplus-reserve estimates are quite close to our liabilities figures. However, as would be expected, the two Foreign Holdings of U.S. Equities measures also diverge for many countries: Our reported holdings are considerably larger than the As noted earlier, the omission of reserve holdings in CPIS-based estimates in several instances, and they the CPIS surveys has little effect on our comparison aresmallerinothers. of liabilities estimates of foreign holdings of U.S. We illustrate the differences between the two estimates for 2004 (figure 10). Our liabilities estimates for 2004 are notably larger than the CPIS-based estimates for Belgium, the Cayman Islands, and 24. An additional problem for comparing the estimates for the Luxembourg, and they are larger by somewhat CaymanIslandsisthatthecoverageoftheCPISsurveyfortheislands isincomplete,asitisbasedonassetsasreportedbybanksonlyand smaller amounts for Switzerland and Germany, an thusexcludessecuritiesheldbythelargenumberofmutualfundsthat indication that custodial bias may overstate these operateintheCaymanIslands. 10. Differences between U.S. liabilities estimates and CPIS-based asset estimates of foreign holdings of U.S. long-term debt securities, for selected CPIS-reporting countries, December 2004 Billions of U.S. dollars Belgium Luxembourg 300 Cayman 250 Islands Switzerland Germany Korea 200 Canada Mexico 150 100 50 + _0 50 100 Brazil Ireland Denmark Norway U.K. Hong France Channel Kong Japan 150 Sweden Netherlands Australia Singapore Islands Bermuda Italy NOTE: Refer to notes to table 3. Difference for each country is calculated CPIS surveys and reserve holdings as estimated from data published by the by subtracting a CPIS-based asset estimate from a U.S. liabilities estimate. International Monetary Fund. CPIS-based asset estimate consists of nonreserve holdings as reported in the

A74 Federal Reserve Bulletin 2006 4. ForeignholdingsofU.S.equities:Reportedforeign countries for which our estimates are smaller (figassets,estimatedU.S.liabilities,andthedifference ure 11). Nonetheless, for several of the countries for betweenthesemeasures,forCPIS-reportingcountries, whichourliabilitiesestimatesforequityarelarger— year-end2001–04 the Cayman Islands, Switzerland, Germany, and Billionsofdollars Belgium—ourestimatesforlong-termdebtwerealso Marketvalueofholdings larger, providing further indications of custodial bias Item inourliabilitiesestimatesforthesecountries. 2001 2002 2003 2004 Foreignassets(CPISsurveys)1....... 997 902 1,267 1,470 U.S.liabilities2...................... 1,390 1,237 1,712 1,966 Foreign Holdings of All U.S. Long-Term Difference(liabilitieslessassets) ..... 393 335 445 497 Securities Note: Refertogeneralnotetotable3. 1. Nonreserveholdingsonly. Overall, the comparison of estimates from the U.S. 2. Nonreserveandreserveholdings.Forestimatedholdingsbyallforeign officialinstitutions,refertotextnote25. liabilities surveys with values reported in the CPIS Source: For foreign assets, International Monetary Fund, surveys supports our interpretation that the liabilities www.imf.org/external/np/sta/pi/cpis.htm; for U.S. liabilities, Treasury InternationalCapitalreportingsystem,www.treas.gov/tic/fpis.html. surveys overstate holdings in several large custodial centers, especially Belgium, Luxembourg, and the equities with CPIS-reported holdings.25 As of year- Cayman Islands. Correspondingly, the holdings end2001,CPIS-reportingcountriesheldinaggregate reportedintheU.S.liabilitiessurveysofseveralother a little more than $1 trillion in U.S. equity; this EuropeanaswellassomeAsiancountriesaresmaller amounthadincreasedtoalmost$1.5trillionbyyear- than the CPIS-based estimates, although some of end 2004 (table 4). Comparison with our liabilities the custodial center holdings as well as some bearer surveyssuggestsamoresignificantundercountinthe bonds no doubt represent holdings of countries that CPIS surveys for equities than for long-term debt, have not yet participated in the CPIS surveys. In as our liabilities-based estimates indicate holdings of general,ouranalysisprovidessupportforthecompa- U.S. equity by CPIS-reporting countries of $1.4 tril- rability of both the CPIS and the U.S. liabilities lion in 2001 and $2.0 trillion in 2004 (although in surveys,anditsuggeststhatbothtypesofsurveysare percentagetermsthatgaphasalsonarrowedinrecent capturingcomparablesecuritiesholdings. years). Consequently, we find more countries for which our liabilities-based estimates are larger than the CPIS-reported holdings, and we find fewer SUMMARY AND CONCLUSION In recent years, U.S. cross-border securities activity 25. Ourliabilitiesestimatesshowthatforeignofficialholdingsof has grown dramatically. As such activity attracts U.S.equitywere$94billionin2001,$87billionin2002,$125billion in2003,and$162billionin2004. greater attention, it is increasingly important to have 11. Differences between U.S. liabilities estimates and CPIS asset reports of foreign holdings of U.S. equities, for selected CPIS-reporting countries, December 2004 Billions of U.S. dollars Cayman Islands Canada Netherlands Bahamas 150 Switzerland Antilles Singapore Bermuda Germany Hong Belgium Norway Australia Kong 100 U.K. Denmark France Japan Luxembourg 50 + _0 50 Italy Sweden Ireland Channel Islands Netherlands NOTE: Refer to notes to table 4. Difference for each country is calculated by subtracting a CPIS-reported asset level from a U.S. liabilities estimate.

Understanding U.S. Cross-Border Securities Data A75 a data collection system that both accurately tracks rities transactions to provide more-timely estimates these positions and flows and is well understood by of cross-border securities holdings by country. This datausers. approach improves our ability to correct estimated Toproperlyinterpretthedataoncross-borderport- holdings for the bias inherent in the monthly transfolio activity, users should understand the strengths actions data and to adjust holdings for valuation andweaknessesofthesystemthatproducesthedata. changes. In this article, we outline a number of factors that Although our survey-based estimates of foreign influence the interpretation of the data collected in holdings of U.S. securities are considered to be comthe TIC system, focusing especially on the data on prehensiveintheircoverage,theircountryattribution cross-border securities. We note ways in which the isimperfectbecauseofthecustodialbiasinthedata. data on monthly transactions in long-term securities We illustrate how information on holdings of U.S. may provide an incomplete or misleading picture of securitiesasreportedinothercountries’assetsurveys cross-bordersecuritiesflows,andwedescribeseveral can help data users to better interpret the country adjustmentsthatcanimprovetheusefulnessandcom- attribution of data obtained from U.S. liabilities surprehensivenessofthepublisheddata. veys. As cross-border financial activity continues to We also discuss a procedure that combines infor- evolveandforeigndatareportingsystemscontinueto mation collected in the most comprehensive part improve,suchcomplementarysourcesofinformation of the TIC system, the annual surveys, with more- maybecomeincreasinglybeneficialinanalyzingU.S. currentbutlesscomprehensivedataonmonthlysecu- cross-bordersecuritiesdata.

A77 Profits and Balance Sheet Developments at U.S. Commercial Banks in 2005 Elizabeth C. Klee and Gretchen C. Weinbach, of the 1. Bank profitability, 1990–2005 Board’s Division of Monetary Affairs, prepared this article. Thomas C. Allard assisted in developing the Percent Percent database underlying much of the analysis. Arshia A. 18 1.8 Burneyprovidedresearchassistance. Return on equity 16 1.6 The profitability of the U.S. commercial banking 14 1.4 industry remained strong again in 2005, although it 12 1.2 was a bit below the levels of recent years. Asset Return on assets 10 1.0 quality was still sound, but pressure on net interest 8 .8 marginsloweredthereturnonassets,andanincrease in equity relative to assets—owing to an accumula- 6 .6 tion of goodwill from recent large mergers—pushed 4 .4 down the return on equity more substantially (figure1).Growthinindustryassetsremainedsolid. 1991 1993 1995 1997 1999 2001 2003 2005 These bank profit and balance sheet developments NOTE: The data are annual. were in large part attributable to the generally favorable financial and economic conditions of the U.S. ever, and thus the Treasury yield curve flattened economy in 2005. On the financial front, short- and considerably. Interest rates on fixed-rate homeintermediate-term interest rates increased as mone- mortgage loans were relatively low over most of tary policy tightening lifted the target federal funds 2005, although they rose somewhat late in the year. rate 2 percentage points during the year (figure 2). Corporateriskspreadsstayedquitenarrowbyhistori- Longer-term interest rates remained quite low, how- calstandards. The U.S. economy continued to expand at a solid Note:Thedatainthisarticlecoverinsureddomesticcommercial pace in 2005. In the household sector, consumer banks and nondeposit trust companies (hereafter, banks). Except spending remained vigorous despite higher energy whereotherwiseindicated,thedataarefromtheConsolidatedReports ofConditionandIncome(CallReport).TheCallReportconsistsof prices; it was supported by an improving labor martwo forms submitted by domestic banks to the Federal Financial ket and gains in household wealth that reflected fur- InstitutionsExaminationCouncil:FFIEC031(forthosewithdomestic ther substantial increases in house prices. The low andforeignoffices)andFFIEC041(forthosewithdomesticoffices only).Thedatathusconsolidateinformationfromforeignanddomes- level of residential mortgage interest rates last year ticoffices,andtheyhavebeenadjustedtotakeaccountofmergersand spurred sales of both new and existing homes to theeffectsofpush-downaccounting.Foradditionalinformationonthe recordlevels,althoughsalescooledsomewhatlatein adjustmentstothedata,seetheappendixinWilliamB.Englishand WilliamR.Nelson(1998),‘‘ProfitsandBalanceSheetDevelopments theyear.Mortgagerefinancingroseforatimeduring atU.S.CommercialBanksin1997,’’FederalReserveBulletin,vol.84 the first half of the year but dropped back in the sec- (June), p. 408. Size categories, based on assets at the start of each ond half. quarter,areasfollows:the10largestbanks,largebanks(thoseranked 11 through 100), medium-sized banks (those ranked 101 through In the corporate sector, investment spending 1,000),andsmallbanks.Atthestartofthefourthquarterof2005,the expandedatasolidpace,buoyedbyrobustgrowthin approximateassetsizesofthebanksinthosegroupswereasfollows: finalsales.Corporateprofitspostedstronggains,and thetenlargestbanks,morethan$87.6billion;largebanks,$7.3billion to$85.7billion;medium-sizedbanks,$457millionto$7.2billion;and firms maintained ample stocks of liquid assets. As a smallbanks,lessthan$457million. result, businesses financed much of their capital ex- Datashowninthisarticlemaynotmatchdatapublishedinearlier penditures out of internal funds. Nonetheless, elyearsbecauseofrevisionsandcorrections.Inthetables,components maynotsumtototalsbecauseofrounding.AppendixtableA.1,A–E, evated merger and acquisition activity and a considreports portfolio composition, income, and expense items, all as a erable rise in share buybacks contributed to a pickup percentageofoverallaveragenetconsolidatedassets,forallbanksand in business borrowing, particularly in the form of foreachofthefoursizecategories.AppendixtableA.2reportsincome statementdataforallbanks. commercialandindustrial(C&I)loans.Thegrowthin

A78 Federal Reserve Bulletin 2006 2. Selected interest rates, 2001–06 consumer loans.1 For both households and businesses, the combination of rising short-term market Percent interest rates and banks’ deposit pricing policies reduced the relative attractiveness of core deposits. 7 As a result, core deposits grew relatively slowly, Ten-year Treasury securities 6 while banks continued to issue managed liabilities at abriskpace.2 5 Other developments in 2005 also strongly influ- 4 encedbanks’profitability.Astheyieldcurveflattened considerably, net interest margins were squeezed at 3 large banks, but the general increase in the level of Target federal funds rate 2 short-term rates led to a widening of the net interest 1 margin at smaller banks, whose liabilities tend to reprice more slowly than those held by larger banks. Competition in business lending likely also contrib- 14 uted to narrower spreads of loan rates over banks’ High-yield bonds 13 cost of funds. At the same time, strong non-interest 12 income, especially trading revenue, and a significant 11 drop in non-interest expense helped shore up bank 10 profitability. Robust business balance sheets and co M rp o o o r d a y te ’s b B on aa ds 9 increases in household wealth kept overall credit 8 quality high. But changes to the bankruptcy law and 7 the devastation caused by last year’s hurricanes re- Thirty-year 6 duced earnings at some banks. Despite these chalfixed-rate mortgages 5 lenges, loss provisions as a share of assets remained 4 at about the level of 2004, and overall, delinquency 2001 2002 2003 2004 2005 2006 andcharge-offratesfellfortheyear. NOTE: The data are monthly and extend through March 2006. The number of new banks increased for the third SOURCE: For Treasury securities, mortgages, and Moody’s corporate yearinarow,butthenumberofconsolidationsofone bonds, Federal Reserve Board, Statistical Release H.15, “Selected Interest Rates” (www.federalreserve.gov/releases/h15); for federal funds, Federal bank charter into another also increased, pulling the Reserve Board (www.federalreserve.gov/fomc/fundsrate.htm); for high-yield number of banks at year-end down to 7,569 from bonds, Merrill Lynch Master II index. 7,677 at year-end 2004 (figure 3). Merger activity at the 10 largest institutions slowed from the pace in 2004, and the share of industry assets at these banks C&I loans was also fueled by increased availability; rose only 1.4 percentage points, to 49.4 percent, at banksreportedthattheyeasedstandardsandtermson year-end.Theshareheldatthe100largestbankswas C&I loans throughout the year. At the same time, steady at 76.9 percent. According to the Federal higher prices for commercial properties supported DepositInsuranceCorporation(FDIC),2005wasthe growthincommercialmortgages. firstyearwithoutanybankfailuressince1934,when These financial and economic conditions left an federaldepositinsurancebegan. imprint on banks’ balance sheets. The relatively low The number of mergers at the bank holding comlevel of fixed mortgage interest rates and the rapid pany level moderated a bit, as did the rate at which climbinthepricesofresidentialandcommercialreal bank holding companies formed. At year-end 2005, estate spurred the demand for bank loans secured by there were 5,155 bank holding companies, 6 more realestate,andtheshareofbankassetsattributableto than at year-end 2004 (for multitiered bank holding real estate loans rose to nearly one-third. Although companies, only the top-tier organization is counted the growth in residential mortgage lending by banks slowedsomewhattowardtheendoftheyear,thepace 1. Inthisarticle,consumerloansconsistofloanstohouseholdsthat ofcommercialrealestatelendingremainedvigorous. arenotsecuredbyrealestate,andtheyincludecreditcardloans. 2. In this article, core deposits consist of transaction deposits, C&I lending surged in 2005, and the share of such savings deposits (including money market deposit accounts), and loans in banks’ assets edged higher for the first time smalltimedeposits.Managedliabilitiesconsistoflargetimedeposits in several years. In contrast, consumer loan growth indomesticoffices,depositsbookedinforeignoffices,subordinated notes and debentures, federal funds purchased and securities sold was anemic, as households apparently continued to underrepurchaseagreements,FederalHomeLoanBankadvances,and favor mortgage financing over relatively high-cost otherborrowedmoney.

Profits and Balance Sheet Developments at U.S. Commercial Banks in 2005 A79 3. Number of banks, and share of assets at the largest BALANCE SHEET DEVELOPMENTS banks, 1990–2005 Total bank assets grew 7.7 percent in 2005, about Thousands 3 percentage points slower than in 2004 but in line Number 14 with the average pace over the preceding five years (table 1). The growth of assets last year was driven 12 mainly by lending secured by real estate. Continued 10 economic expansion and relatively low long-term 8 interest rates supported hefty increases in lending in 6 both residential and commercial real estate markets. A surge in C&I lending, fueled by favorable demand Percent and supply conditions, also supported the growth of Share of assets assets. Overall, total loans and leases expanded 80 10.4 percent; this ample loan volume led banks to 100 largest increasetheirsecuritiesholdingsonly2.4percent. 60 On the other side of the balance sheet, liabilities 40 10 largest expanded 7.7 percent on a year-end basis, a rate in 20 line with the gain in assets and nearly 2 percentage points lower than in 2004. Growth in core deposits 1991 1993 1995 1997 1999 2001 2003 2005 fell back a bit; given banks’ deposit pricing policies, NOTE: The data are as of year-end. For the definition of bank size, refer to rising short-term interest rates damped the relative the general note on the first page of the main text. attractiveness of liquid deposits for businesses and households. The expansion in managed liabilities, however,remainedbrisk. Banks continued to add to their capital positions; for this article). The share of bank holding company onanannualaveragebasis,bankcapitalasashareof assets at the 50 largest bank holding companies with average net consolidated assets edged up for the major commercial banking operations remained relasecond year in a row, to 10.1 percent. This ratio has tively steady at 74 percent.3 The number of financial risenoverthepastdecadeundertheinfluenceofthree holdingcompaniesfellmoderately,from639atyeartrends: Banks’ retained earnings have increased, end2004to625atyear-end2005.Mostofthelargest paid-in capital from parent holding companies has bank holding companies have elected to become moved higher on balance, and industry mergers and financialholdingcompanies.Consequently,attheend acquisitionshaveaugmentedthevalueofgoodwillon of 2005, about 84 percent of the assets of bank banks’ books. Regulatory capital ratios moved down holding companies were held by financial holding alittlein2005butremainedhigh. companies.4 Loans to Businesses 3. The50largestbankholdingcompaniesaredefinedhereasthe 50largest(asmeasuredbytotalconsolidatedassets)aftertheexclu- U.S. nonfinancial corporations needed only limited sion of a few institutions whose commercial banking operations accountforonlyasmallportionoftheirassetsandearnings.Thequar- recoursetoexternalfundslastyeargivenstrongcash terlyarticleintheFederalReserveBulletin,‘‘ReportontheCondi- positions and robust profits. Firms’ net financing gap tionoftheU.S.BankingIndustry,’’atwww.federalreserve.gov/pubs/ dropped into negative territory at the end of the year bulletin/default.htm,providesinformationonthe50largebankholding companies(the50largestasdefinedhere)andonthebankingindus- (figure 4).5 This decline also partly reflected tempotryfromtheperspectiveofbankholdingcompanies(includingfinan- rarytaxprovisionsthatencouragedtherepatriationof cialholdingcompanies)thatfilereportsFRY-9CandFRY-9LP;curprofits held at foreign subsidiaries. Net corporate rently,onlyabout2,300top-tierbankholdingcompaniesarerequired tofilethosereports(referto‘‘ReportontheCondition’’table1,last bond issuance was subdued in 2005, but C&I loan row,andnote1). growth surged to 12.5 percent. Apparently, a rise in 4. Financialholdingcompanystatisticsincludebothdomesticbank mergers and acquisitions and a considerable increase holding companies that have elected to become financial holding companiesandforeignbankingorganizationsoperatingintheUnited in share buybacks significantly lifted the demand for StatesasfinancialholdingcompaniesandsubjecttotheBankHolding bank financing. In addition, commercial real estate CompanyAct.Formoreinformation,refertoBoardofGovernorsofthe FederalReserveSystem(2003),ReporttotheCongressonFinancial HoldingCompaniesundertheGramm-Leach-BlileyAct(Washington: 5. Thenetfinancinggapisdefinedhereasthefour-quartermoving Board of Governors, November), at www.federalreserve.gov/pubs/ averageofthedifferencebetweencapitalexpendituresandU.S.funds reports_other.htm. generatedinternally.

A80 Federal Reserve Bulletin 2006 1. Changeinbalancesheetitems,allU.S.banks,1996–2005 Percent Memo Dec. 2005 Item 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 (billions of dollars) Assets .................................. 6.13 9.22 8.18 5.44 8.76 5.11 7.19 7.19 10.78 7.72 8,897 Interest-earningassets ................. 5.82 8.66 8.20 5.83 8.66 3.96 7.54 7.28 11.29 7.96 7,727 Loansandleases,net................ 8.17 5.32 8.76 8.03 9.24 1.82 5.90 6.52 11.20 10.38 5,227 Commercialandindustrial ........ 7.24 12.02 12.94 7.88 8.54 −6.73 −7.41 −4.56 4.35 12.54 1,012 Realestate ....................... 5.45 9.30 7.99 12.22 10.74 7.94 14.43 9.78 15.41 13.80 2,954 Bookedindomesticoffices ..... 5.51 9.53 7.97 12.36 11.02 8.02 14.85 9.68 15.08 13.92 2,902 One-tofour-family residential .............. 4.66 9.67 6.36 9.70 9.28 5.70 19.85 10.05 15.79 12.10 1,645 Other ....................... 6.75 9.32 10.29 16.06 13.31 10.95 8.81 9.20 14.14 16.41 1,257 Bookedinforeignoffices ....... 3.18 .34 8.79 6.28 −1.62 3.97 −7.41 15.74 35.59 7.19 52 Consumer........................ 5.12 −2.19 .34 −1.49 8.04 4.16 6.55 9.32 10.11 2.26 799 Otherloansandleases ............ 22.28 −7.91 13.95 6.71 7.01 −2.02 −.03 8.31 3.57 −.18 531 Loan-lossreservesand unearnedincome............. .04 −.45 3.11 2.34 7.99 13.15 5.73 −2.68 −4.19 −5.72 69 Securities .......................... .86 8.85 8.40 5.11 6.36 7.22 16.21 9.43 10.58 2.39 1,882 Investmentaccount ............... −1.10 8.66 12.06 6.68 2.86 8.88 13.54 8.70 6.15 1.18 1,528 U.S.Treasury .................. −14.28 −8.85 −25.17 −1.89 −32.72 −40.27 41.92 14.14 −15.87 −17.60 50 U.S.governmentagencyand corporationobligations..... 3.63 14.18 17.00 1.83 3.75 12.85 18.11 9.67 9.46 −1.85 970 Other.......................... 1.83 11.21 26.99 20.90 13.39 12.18 2.72 5.98 3.02 10.16 508 Tradingaccount .................. 14.44 10.00 −13.32 −6.93 37.16 −3.72 36.12 14.01 36.81 7.96 354 Other .............................. 1.06 38.54 3.79 −8.37 10.30 13.02 −2.92 6.83 14.28 5.83 617 Non-interest-earningassets ............ 8.29 13.03 8.10 2.90 9.45 12.79 5.10 6.61 7.60 6.18 1,170 Liabilities ............................... 5.99 9.11 8.06 5.58 8.59 4.45 7.13 7.25 9.55 7.74 8,003 Coredeposits ......................... 4.13 4.52 7.04 .23 7.53 10.55 7.58 7.30 8.25 6.50 4,233 Transactiondeposits ................ −3.44 −4.55 −1.41 −8.97 −1.31 10.21 −5.12 2.83 3.20 −1.07 736 Savingsdeposits(including MMDAs) ...................... 13.83 12.96 18.32 6.68 12.51 20.68 18.46 13.71 11.72 6.89 2,745 Smalltimedeposits ................. 2.25 4.18 .53 −.76 7.20 −7.23 −4.92 −6.77 1.58 13.47 751 Managedliabilities1 ................... 9.73 13.79 9.44 15.54 8.79 −2.73 5.34 6.97 12.06 12.11 3,263 Largetimedeposits ................. 21.17 20.15 9.09 14.19 19.37 −3.64 5.05 1.43 21.86 22.33 862 Depositsbookedinforeignoffices ... 4.27 11.13 8.71 14.60 7.84 −10.96 4.49 12.63 16.84 6.32 920 Subordinatednotesanddebentures ... 17.74 21.05 17.00 5.07 13.98 9.56 −.59 5.08 10.49 11.41 122 Grossfederalfundspurchased andRPs ....................... −2.20 30.51 4.35 1.56 6.49 5.72 12.75 −8.70 8.40 15.64 635 Othermanagedliabilities ............ 19.73 −4.04 15.65 35.27 1.80 −.28 .97 22.01 1.37 6.15 725 Revaluationlossesheldintrading accounts ......................... 8.89 36.94 3.44 −13.20 7.47 −17.06 33.44 14.02 −12.61 −17.96 135 Other................................. −1.30 14.82 12.75 −1.27 20.61 14.90 5.22 5.29 17.08 −1.70 373 Capitalaccount.......................... 7.77 10.44 9.53 3.89 10.65 12.30 7.83 6.63 23.14 7.60 893 Memo Commercialrealestateloans2 ............ 7.67 10.13 11.37 15.42 12.16 13.10 6.82 8.99 13.87 16.71 1,255 Mortgage-backedsecurities .............. 2.06 14.16 22.12 −3.34 3.29 29.05 15.56 10.10 13.45 2.06 882 FederalHomeLoanBankadvances....... n.a. n.a. n.a. n.a. n.a. n.a. 17.21 3.74 3.73 9.99 269 Note: Dataarefromyear-endtoyear-end. by multifamily residential properties; and loans to finance commercial real 1. Measuredasthesumoflargetimedepositsindomesticoffices,deposits estate,construction,andlanddevelopmentactivitiesnotsecuredbyrealestate. booked in foreign offices, subordinated notes and debentures, federal funds n.a. Notavailable. purchasedandsecuritiessoldunderrepurchaseagreements,FederalHomeLoan MMDAMoneymarketdepositaccount. Bankadvances,andotherborrowedmoney. RPRepurchaseagreement. 2. Measuredasthesumofconstructionandlanddevelopmentloanssecured byrealestate;realestateloanssecuredbynonfarmnonresidentialpropertiesor loans continued to expand rapidly, probably in part pointed to a pickup in mergers and acquisitions. becauseofsomeimprovedmarketfundamentals. Although the net percentage of banks reporting Unlike the pattern in preceding years, C&I loan strongerdemanddeclinedtowardtheendoflastyear, growth was rapid at banks of all sizes in 2005. thefractionremainedsubstantial. Responses to the Federal Reserve’s quarterly Senior BLPSrespondentsalsoindicatedthat,onnet,their LoanOfficerOpinionSurveyonBankLendingPrac- institutions had further eased credit standards and tices (BLPS) suggest that factors related to both terms on C&I lending last year, although the net demandandsupplylikelyplayedaroleinthegainsin percentageofbankssoreportingdeclinedtowardthe C&I lending. Throughout the year, most survey re- endoftheyear.Surveyrespondentstypicallyreported spondents reported greater demand for business that they had eased credit standards on C&I loans to financing, attributing it to a rising need to finance large and middle-market firms (figure 5, bottom inventories, accounts receivable, and investment in panel). Similarly, according to the Federal Reserve’s plantandequipment(figure5,toppanel).Asubstan- Survey of Terms of Business Lending, average tial fraction of respondents to some surveys also spreads of rates on C&I loans over those on

Profits and Balance Sheet Developments at U.S. Commercial Banks in 2005 A81 4. Financing gap at nonfarm nonfinancial corporations, 5. Changes in demand and supply conditions at selected 1990–2005 banks for C&I loans to large and middle-market firms, 1990–2006 Billions of dollars Percent 300 Net percentage of banks reporting stronger demand1 250 60 200 40 150 20 + 100 _0 50 + 20 _0 40 50 60 100 80 1991 1993 1995 1997 1999 2001 2003 2005 NOTE: The data are four-quarter moving averages. The financing gap is the difference between capital expenditures and internally generated funds. Net percentage of banks that tightened standards2 SOURCE: Federal Reserve Board, Statistical Release Z.1, “Flow of Funds 60 Accounts of the United States,” table F. 102 (www.federalreserve.gov/ releases/z1). 40 comparable-maturity securities declined notably on 20 balancein2005. Commercial real estate (CRE) loans grew rapidly + _0 lastyear;the16.7percentpacewasnearly3percentagepointsabovethebriskrateofexpansionin2004. 20 Unlike some preceding years, in which CRE loan growthwashighlyconcentratedatmedium-sizedand 1990 1992 1994 1996 1998 2000 2002 2004 2006 small banks, CRE lending in 2005 was quite brisk in allbanksizecategories,particularlyatthetenlargest NOTE: The data are drawn from a survey generally conducted four times per year; the last observation is for the January 2006 survey, which covers banks. 2005:Q4. Net percentage is the percentage of banks reporting an increase in Real-estate-securedloansforconstructionandland demand or a tightening of standards less, in each case, the percentage reporting the opposite. The definition for firm size suggested for, and developmentledthepickupinCRElendinglastyear generally used by, survey respondents is that large and middle-market firms and were likely fueled in part by the record-setting have sales of $50 million or more. 1. Series begins with the November 1991 survey. levels of new home construction. This category of 2. Series begins with the May 1990 survey. CRElendinggrew34.4percentlastyearand25.5per- SOURCE: Federal Reserve Board, Senior Loan Officer Opinion Survey on Bank Lending Practices (www.federalreserve.gov/boarddocs/snloansurvey). cent in 2004; it now accounts for nearly one-third of allCREloans,comparedwithaboutone-fourthatthe end of 2003 (figure 6). The largest category of CRE respondents also pointed to improvements in the lending,realestateloanssecuredbynonfarmnonresiconditionof,ortheoutlookfor,theCREsectorinthe dential structures, expanded 9.9 percent last year, a marketsinwhichtheyoperate. ratedownslightlyfrom2004. Respondents to the BLPS reported stronger demand for CRE loans throughout 2005, although the Loans to Households netpercentagedoingsodroppedmarkedlyattheend of the year (figure 7, top panel). Respondents also Gainsinpersonalincomeandemploymentaccompanoted that their institutions eased standards on CRE nied the solid economic expansion last year, while loans,onnet,overmostoftheyear(figure7,bottom interest rates on fixed-rate mortgages remained in a panel).Inaddition,responsestoaspecialquestionin relatively low range and house prices posted further the January 2006 BLPS about changes in various sizable increases. Accordingly, residential mortgage termsonCRElendingindicatedaconsiderableeasing lending, including both first- and second-lien loans over the past year. The main reason banks cited for secured by one- to four-family residential properties, such easing was more-aggressive competition from grew briskly again in 2005, at 12.1 percent, although other banks or nonbank lenders. In addition, some theadvancewasdownabitfromthepacein2004.

A82 Federal Reserve Bulletin 2006 6. Changes in the major components of commercial 8. Residential mortgage refinancing, 1990–2005 real estate loans, by purpose, 1990–2005 January 26, 1990 = 1 Percent 90 Construction and land 80 30 development 70 Multifamily residential 20 60 50 10 40 + _0 30 20 Nonfarm 10 nonresidential 10 + 20 _0 1991 1993 1995 1997 1999 2001 2003 2005 1991 1993 1995 1997 1999 2001 2003 2005 NOTE: The data are four-week moving averages. For definition of resi- NOTE: The data are annual. dential mortgages, refer to text. SOURCE: Mortgage Bankers Association. The pattern of residential mortgage lending during declines, and refinancing activity dropped back (figthe year was strongly influenced by changes in longure 8). The BLPS responses during the year were terminterestrates.Mortgageratesdroppedsomewhat consistent with these aggregate patterns: On net, over the first half of the year, and mortgagedemand for one- to four-family mortgages was rerefinancing activity picked up somewhat. But in the ported to have risen earlier in the year but to have second half, mortgage rates more than reversed their fallenoffatyear-end(figure9).6 To provide a longer-term perspective, a special 7. Changes in demand and supply conditions for commercial real estate loans at selected banks, question in the October 2005 BLPS asked domestic 1996–2006 banks how their terms on mortgage loans used to purchasehomeshadchangedoverthepasttwoyears. Percent A significant fraction of banks reported they had Net percentage of banks reporting stronger demand eased some terms on such loans. These easings 60 40 6. Inaskingbankshowdemandformortgagestopurchasehomes haschangedoverthepastthreemonths,theBLPSinstructsbanksto 20 consider only new originations as opposed to the refinancing of + existing mortgages. However, this distinction may be difficult for _0 bankstomakeinpractice. 20 9. Net percentage of selected banks reporting stronger 40 demand for residential mortgages, 1990–2006 60 Percent 60 Net percentage of banks that tightened standards 40 60 20 + 40 _0 20 20 40 + _0 60 80 20 1990 1992 1994 1996 1998 2000 2002 2004 2006 1996 1998 2000 2002 2004 2006 NOTE: Series begins with the October 1990 survey. For definition of residential mortgages, refer to text. Refer also to figure 5, general note and NOTE: Refer to figure 5, general note and source note. source note.

Profits and Balance Sheet Developments at U.S. Commercial Banks in 2005 A83 included increasing the maximum size of primary 10. Net percentage of selected banks reporting tightened mortgages they were willing to provide, increasing standards for consumer lending, 1996–2006 the maximum size of second mortgages, narrowing Percent spreadsofmortgageratesoveranappropriatemarket base rate, and increasing the maximum loan-to-value Consumer loans other than credit cards 25 ratioonsuchloans.Bycontrast,banksnotedthatthe 20 maximum length of extended interest-rate locks, minimumrequiredcreditscores,andloanorigination 15 feeswerelittlechanged. 10 Theslowergrowthofrevolvinghomeequityloans 5 in 2005 likely reflected the effects of higher interest + rates. Most revolving home equity loans carry vari- _0 able interest rates that are tied to short-term market 5 rates,whichrosesteadily.Asaresult,growthinthese 10 loans dropped steeply from the very brisk rate over the preceding several years, when short-term interest rates were particularly low. Still, these loans Credit card loans 50 expandedat8.1percentfortheyearasawhole,faster thantheoverallincreaseinbankassets. 40 A special set of questions in the July 2005 BLPS 30 queriedbanksabouttheircurrentholdingsandrecent 20 originationsofnontraditionalmortgageproducts.7Respondents generally reported that such loans ac- 10 counted for less than one-fourth of their residential + _0 mortgage originations and of the mortgages on their books. However, more than one-half of the respon- 10 dents to that survey noted that the share of mortgage originations attributable to nontraditional mortgage 1996 1998 2000 2002 2004 2006 productshadbeenhigheroverthepasttwelvemonths NOTE: Refer to figure 5, general note and source note. than over the previous twelve-month period.8 accordingtotheBLPS,butdidnoteasethemfurther Consumerloansonbanks’booksgrewjust2.3perin the fourth quarter (figure 10). BLPS respondents cent in 2005, less than half the pace of 2004 after generally kept terms on consumer loans about unadjusting the 2004 rate downward to account for the changed last year. Banks also reported in the BLPS effectofalargemerger.9Mortgage-relatedborrowing that demand for consumer loans had weakened over likely took the place of some consumer loans as the thesecondhalfoftheyear. rates on various types of consumer loans moved higher.Banksgenerallyeasedstandardsonconsumer lending over the first three quarters of last year, Other Loans and Leases 7. The July 2005 BLPS defined ‘‘nontraditional mortgage prod- The amount of other loans and leases reported on ucts’’toinclude,butnotbelimitedto,adjustable-ratemortgageswith banks’ balance sheets at year-end 2005 was essenmultiple payment options, interest-only mortgages, and so-called ‘‘Alt-A’’productssuchasmortgageswithlimitedincomeverification tiallyunchangedfromayearearlier.Loansandleases and mortgages secured by non-owner-occupied properties. Respon- to depository institutions grew 12.8 percent after dentstothatsurveywereinstructedtoexcludestandardadjustable-rate having been flat in 2004. Agricultural-related loans mortgagesandcommonhybridadjustable-ratemortgages—thoseon which the interest rate is initially fixed for a multiyear period and expanded moderately again after having contracted subsequentlyadjustsmorefrequently. over a period of several years, and estimates suggest 8. In December 2005, federal banking regulators proposed guidthat most categories of farm loans contributed to the ance to financial institutions on managing the risks associated with nontraditional mortgage products; the guidance also proposed con- pickup.10 Rising incomes, retail sales, and property sumer protection practices (‘‘Federal Financial Regulatory Agencies Propose Guidance on Nontraditional Mortgage Products,’’ press re- 10. Using its Survey of Terms of Bank Lending to Farmers, the lease, December 20, www.federalreserve.gov/boarddocs/press/bcreg/ FederalReserveestimatesnon-real-estatebankloansmadetofarmers 2005). by purpose of the loan, such as to obtain farm equipment and 9. Forinformationonthis2004adjustment,refertoElizabethC. machinery or to cover operating expenses. This information is pub- Klee and Fabio M. Natalucci (2005), ‘‘Profits and Balance Sheet lishedquarterlyinBoardofGovernorsoftheFederalReserveSystem, DevelopmentsatU.S.CommercialBanksin2004,’’FederalReserve StatisticalReleaseE.15,‘‘AgriculturalFinanceDatabook,’’sectionA Bulletin,vol.91(Spring),p.151,footnote6. (www.federalreserve.gov/releases/e15).

A84 Federal Reserve Bulletin 2006 11. Bank holdings of securities as a proportion of total 12. Selected domestic liabilities at banks as a proportion bank assets, 1990–2005 of their total domestic liabilities, 1990–2005 Percent Percent 26 40 Savings deposits 24 30 22 20 Small time deposits 20 Transaction deposits 10 18 1991 1993 1995 1997 1999 2001 2003 2005 1991 1993 1995 1997 1999 2001 2003 2005 NOTE: The data are quarterly. NOTE: The data are quarterly. Savings deposits include money market deposit accounts. values supported the continuing improvement in the up somewhat. However, as long-term interest rates fiscal positions of state and local governments. As a backed up over the second half of the year and consequence, growth in loans to this sector slowed a refinancing activity slowed, banks’ securities holdbit, to 11.7 percent, from 13.2 percent in 2004. By ingsdeclinedabit. contrast, the remaining categories of other loans and leases—whichinclude,amongothers,leasefinancing Liabilities receivables, loans for purchasing or carrying securities, and loans to foreign depository institutions, and The expansion of bank liabilities slowed from the which account for a bit more than half of the total— rapid pace posted in 2004. Even so, last year’s declinedlastyear. advance of 7.7 percent was still a bit above the averagerateoftheprevioustenyears.Thegrowthincore Securities deposits, at 6.5 percent, was the slowest in six years. Increases in short-term market rates typically make Withloansexpandingbrisklyin2005,bankstrimmed core deposits overall less attractive to deposit holdthegrowthintheirsecuritiesholdings.Atjust2.4per- ers because the rates banks pay on the majority of cent, the rate of increase was nearly 10 percentage these deposits tend to lag increases in market rates. points below the average over the previous three When short-term market rates rise, as they did in years.Accordingly,banks’holdingsofsecuritiesasa 2005, banks adjust rates on liquid deposits relatively shareoftotalassetsdeclined,althoughatyear-endthe slowly and by smaller margins than they do rates on sharewasonlyalittlebelowitsrecentelevatedrange small time deposits. As a result, transaction deposits (figure11). declined last year, and the rate of growth of savings Growth in securities held in both investment and accounts was only about half that posted in 2004, trading accounts ebbed significantly. In particular, whereas the issuance of small-denomination time mortgage-backed securities (MBS) held in invest- deposits surged. As a share of total liabilities, therement accounts grew only 2.1 percent after having fore, transaction accounts and savings deposits fell a expanded at a double-digit pace for several years. In bit, and small time deposits rose a little (figure 12). addition, banks continued to shed U.S. Treasury The growth of managed liabilities, at 12.1 percent, securities for the second consecutive year, and hold- was the same as in 2004. Large time deposits ings of non-mortgage-backed agency securities con- expanded quite briskly again, the growth of suborditractedforthefirsttimeinmorethanadecade.Banks nated debt continued at a low double-digit pace, accumulated securities at a fairly rapid clip on bal- advances from Federal Home Loan Banks rose a ance over the first half of 2005, as long-term interest rapid10.0percent,andthegrowthofdepositsbooked ratesdeclinedonnetandmortgagerefinancingpicked inforeignofficesdroppedbackto6.3percent.

Profits and Balance Sheet Developments at U.S. Commercial Banks in 2005 A85 13. Regulatory capital ratios, 1990–2005 14. Assets and regulatory capital at well-capitalized banks, 1990–2005 Percent Percent 14 Share of industry assets at well-capitalized banks Total (tier 1 + tier 2) 13 12 100 11 Tier 1 10 80 9 60 Leverage 8 7 40 6 20 1991 1993 1995 1997 1999 2001 2003 2005 NOTE: The data are as of year-end. For the components of the ratios, refer Percentage points to text notes 11 and 12. Average margin by which banks were well capitalized Capital 3.5 Banks’capitalpositionsremainedstrongin2005.The 3.0 7.6 percent rate of growth of capital accounts about matchedthatofbanks’assetsandliabilities.Retained 2.5 earnings of $48 billion accounted for about three- 2.0 fourths of the increase in banks’ capital accounts. Banksalsoboostedcapitalbyreceivingfundingfrom 1.5 theirparentholdingcompaniesandbyissuingshares tothepublic.Tier1capitalincreased9.9percent,and tier 2 capital advanced 7.5 percent.11 Unlike the 1991 1993 1995 1997 1999 2001 2003 2005 pattern of the preceding several years, risk-weighted NOTE: The data are annual. For the definitions of “well capitalized” and of the margin by which banks remain well capitalized, refer to text notes 13 assets grew more quickly than total assets in 2005. and 14. BanksreducedtheirholdingsofTreasuryandagency securities, which generally have a zero risk weight, 2000. By contrast, the leverage ratio, which is based while they modestly increased their agency-related ontangibleaverageassets,inchedup.12 MBS holdings, which have a 20 percent risk weight. The overall share of industry assets held by well- Meanwhile, assets that receive higher risk weights, capitalized banks in 2005 rose to 98 percent, from such as C&I loans, climbed rapidly. As a result, the 96percentattheendof2004(figure14,toppanel).13 ratiooftier1capitaltoallrisk-weightedassetsedged The estimated average margin by which welldown,tojustunder9.9percent(figure13).Thetier2 capitalized banks exceeded regulatory capital stanratiofellslightlyaswell,causingthetotalratio(tier1 dards fell slightly again in 2005 but remained in the plus tier 2) to move down for the second year in a middle of its range over the past decade (figure 14, row, although it remained above its recent low in bottompanel).14 12. Theleverageratioistheratiooftier1capitaltotangibleassets. 11. Tier1andtier2capitalareregulatorymeasures.Tier1capital Tangible assets are equal to total average consolidated assets less consistsprimarilyofcommonequity(excludingintangibleassetssuch assets excluded from common equity in the calculation of tier 1 asgoodwillandexcludingnetunrealizedgainsoninvestmentaccount capital. securities classified as available for sale) and certain perpetual pre- 13. Well-capitalizedbanksarethosewithatotalrisk-basedcapital ferred stock. Tier 2 capital consists primarily of subordinated debt, ratioof10percentorgreater,atier1risk-basedratioof6percentor preferredstocknotincludedintier1capital,andloan-lossreservesup greater, a leverage ratio of 5 percent or greater, and a composite toacapof1.25percentofrisk-weightedassets.Risk-weightedassets CAMELSratingof1or2.EachletterinCAMELSstandsforakey are calculated by multiplying the amount of assets and the credit- elementofbankfinancialcondition—Capitaladequacy,Assetquality, equivalent amount of off-balance-sheet items (an estimate of the Management,Earnings,Liquidity,andSensitivitytomarketrisks. potentialcreditexposureposedbytheitems)bytheriskweightfor 14. The estimated average margin by which banks were well eachcategory.Theriskweightsrisefrom0to1asthecreditriskofthe capitalizedwascomputedasfollows:Amongtheleverage,tier1,and assets increases. The tier 1 ratio is the ratio of tier 1 capital to total capital ratios of each well-capitalized bank, the institution’s risk-weightedassets;thetotalratioistheratioofthesumoftier1and ‘‘tightest’’capitalratioisdefinedastheoneclosesttotheregulatory tier2capitaltorisk-weightedassets. standardforbeingwellcapitalized.Thebank’smarginisthendefined

A86 Federal Reserve Bulletin 2006 2. Changeinnotionalvalueandfairvalueamountsofderivatives,allU.S.banks,2000–05 Percent Memo Dec. 2005 Item 2000 2001 2002 2003 2004 2005 (billions of dollars) Totalderivatives Notionalamount .......................... 16.87 11.47 24.14 26.54 23.70 15.42 101,886 Fairvalue Positive ................................ 20.80 26.42 85.41 .36 13.71 −6.47 1,262 Negative ............................... 21.86 20.82 89.18 1.00 13.75 −5.78 1,246 Interestratederivatives Notionalamount........................ 18.70 15.93 26.83 27.62 22.07 11.92 84,500 Fairvalue Positive.............................. 23.75 63.87 108.20 −5.95 13.14 −5.52 982 Negative ............................. 24.24 56.55 113.02 −5.07 12.94 −5.14 961 Exchangeratederivatives Notionalamount........................ 5.65 −7.00 7.34 18.81 21.03 7.61 9,712 Fairvalue Positive.............................. 25.85 −16.21 8.67 41.81 14.86 −35.87 147 Negative ............................. 27.61 −15.65 15.73 38.81 12.74 −37.41 144 Creditderivatives Notionalamount........................ 48.55 −1.20 52.47 55.98 134.52 148.09 5,822 Guarantor ............................ 34.40 21.84 38.57 61.82 139.07 137.87 2,681 Beneficiary........................... 60.03 −16.89 66.36 51.13 130.46 157.53 3,141 Fairvalue Guarantor ............................ n.a. n.a. n.a. 68.31 69.92 81.42 36 Positive ........................... n.a. n.a. n.a. 378.09 74.56 −5.63 17 Negative .......................... n.a. n.a. n.a. −68.87 38.37 827.99 19 Beneficiary........................... n.a. n.a. n.a. 19.85 51.28 83.50 41 Positive ........................... n.a. n.a. n.a. −63.13 2.64 505.51 22 Negative .......................... n.a. n.a. n.a. 295.74 66.36 2.78 19 Otherderivatives1 Notionalamount........................ 24.94 −12.06 6.70 3.77 33.15 32.25 1,852 Fairvalue .............................. Positive.............................. 1.40 −34.72 20.28 3.16 8.55 58.51 95 Negative ............................. 4.42 −42.63 24.62 −5.25 19.73 74.28 103 Note: Dataarefromyear-endtoyear-end. n.a. Notavailable. 1.Otherderivativesconsistofequityandcommodityderivativesandother contracts. Derivatives market value of those with negative values—was onlyabout$16billionattheendof2005,aboutone- Banks’ holdings of off-balance-sheet derivatives third lower than at the end of 2004.15 Derivates continued to expand in 2005 although not as rapidly holdings continued to be highly concentrated—the as they had over the preceding few years. The ten largest banks accounted for 98 percent of the notional principal amount of derivatives contracts notional value of all derivatives held by banks at the held by banks grew 15 percent last year, and it end of 2005, a ratio unchanged from 2004. totaled about $102 trillion at the end of the year The majority of banks’ derivatives holdings in (table 2). However, the fair market value of these notionaltermscontinuedtobeintheformofinterest holdings is typically much smaller than the notional rate swaps. Banks’ holdings of such swaps grew amount. In addition, because many of these holdings about 15 percent in 2005, and the share of these are linked to banks’ role as dealers, a considerable contracts relative to all derivatives contracts held portion of their derivatives positions are offsetting. steady at 64 percent in notional terms.16 Interest rate At the end of 2005, the aggregate fair market value swaps are typically used to hedge interest rate risk, of contracts with positive value was $1.26 trillion, includingthatrelatedtoholdingsofinterest-sensitive and the aggregate fair market value of contracts with assets such as mortgages, MBS, and assets related to negative value was at a similar level. As a result, the net fair value of all contracts—the total fair market 15. Thatthefairmarketvaluesofbanks’derivativescontractsare nearlyoffsettingdoesnotmeanthatbanks’aggregateexposuretothe value of contracts with positive values less the fair marketandcreditriskassociatedwiththecontractsarelikewisenearly offsettingbecause,forexample,thecounterpartiestobanks’positiveasthepercentagepointdifferencebetweenitstightestcapitalratioand andnegative-valuedcontractsmaydiffer. thecorrespondingregulatorystandard.Theaveragemarginamongall 16. Interestrateswapsareagreementsinwhichtwopartiescontract well-capitalizedbanks—themeasurereferredtoinfigure14—isthe toexchangetwopaymentstreams,onebasedonafloatinginterestrate weightedaverageofalltheindividualmargins;theweightsareeach and one based on a fixed interest rate; the payment streams are bank’sshareofthetotalassetsofwell-capitalizedbanks. calculatedonthebasisofsomenotionalprincipalamount.

Profits and Balance Sheet Developments at U.S. Commercial Banks in 2005 A87 mortgageservicing.Thegrowthinthenotionalvalue 15. Notional amounts of credit derivatives for which ofbanks’holdingsofinterestrateswapsin2005was banks were beneficiaries or guarantors, 2000–05 the slowest in more than a decade, perhaps because Trillions of dollars low interest rate volatility throughout the year damped demand for interest rate hedging by banks’ 3.0 customers. Thenotionalvalueofbanks’holdingsofallinterest 2.5 rate derivatives contracts, including interest rate fu- 2.0 tures, forwards, and options contracts, advanced 1.5 12 percent last year and accounted for 83 percent of thenotionalvalueofallderivativescontractsheldby 1.0 Beneficiary banks at year-end, down from about 86 percent for Guarantor .5 theprecedingthreeyears.Banks’notionalholdingsof + foreign exchange derivatives grew nearly 8 percent _0 lastyear,andtheirholdingsofequityandcommodity derivatives advanced briskly, at about 32 percent. 2000 2001 2002 2003 2004 2005 Commodityderivativesincludethoseonenergyprod- NOTE: The data are as of quarter-end. ucts,whosepriceswerequitevolatilelastyear.Banks positions as guarantors totaled nearly $2.7 trillion that are major derivatives dealers reportedly targeted (figure15). commodity derivatives for expansion in 2005.17 In notional terms, foreign exchange, equity, and commodityderivativescontractsaccountedforabitmore TRENDS IN PROFITABILITY than 11 percent of all derivatives held at year-end, downslightlyfromyear-end2004. Measures of banking industry profitability remained Banks’ holdings of credit derivatives continued to strong in 2005, although they were down a bit from surge last year.18 For the second year in a row, the recenthighlevels.Theeconomy’ssolidperformance notional amount of credit derivatives held by banks supportedtheindustry’sprofitabilitybykeepingloan more than doubled relative to the preceding year, lossprovisionslowandboostingthedemandforloans. reaching $5.8 trillion at the end of 2005. The fair- Returnonassets(ROA)tickeddown3basispoints,to value amount of credit derivatives on banks’ books 1.31percent,owinglargelytoanarrowednetinterest expanded 83 percent, to $77 billion at year-end 2005 margin,butitremainedintheupperhalfofitsrangeof from $42 billion at year-end 2004. Still, in terms of thepastdecade.Returnonequity(ROE)droppedmore notionalvalue,theseholdingsaccountforonlyabout than1percentagepoint,to13.01percent,reachingits 6 percent of all derivatives. Given their role as lowest level in more than ten years. The decline, dealers,banksarebothbuyers(beneficiariesofcredit however, is largely attributable to an expansion of protection)andsellers(protectionproviders,orguar- equityowingtoanaccumulationofgoodwillacquired antors) of these contracts, as they are for other in some recent large mergers. Excluding goodwill, derivatives.Bankswereagainnetrecipientsofcredit ROEwasnearthetopofitshistoricalrange.19Banks protection last year—the notional amount of banks’ operatinginhurricane-affectedareassufferedtovarypositions as beneficiaries totaled a bit more than ingdegreesinthethirdquarter;theirrevenuemaypick $3.1 trillion at the end of last year, while their up as rebuilding efforts continue, but the ultimate extentoflossesandthepossiblegainsfromrebuilding remain uncertain (for further discussion, refer to box ‘‘TheEffectsofHurricaneKatrina’’).Fortheindustry 17. Mostofthe32percentgrowthinbanks’holdingsofequityand commodity derivatives last year is attributable to the expansion of as a whole, the fraction of banks with negative net commodityderivativesatonelargebank.Accordingtoits2005annual income rose a bit from 2004, to 6.3 percent, but the bankholdingcompanyreportfiledwiththeSecuritiesandExchange shareofindustryassetsatbanksincurringlossesfellto Commission, this bank wanted to increase its commodities trading activities to provide a new source of trading revenue and to reduce about0.5percent. volatilityinitstradingresultsovertime. A further narrowing of the net interest margin at 18. Creditderivativesareover-the-counteragreementsinwhichthe the ten largest banks weighed on the profitability of risk of default of a certain reference entity is transferred from one party(thebeneficiary)toanother(theprotectionproviderorguaran- thoseinstitutions,butthenetinterestmarginwidened tor). For a general description of credit derivatives, including a discussion of how they are used by banks, refer to Roberto Perli, 19. Banks accumulated a significant amount of goodwill from a ‘‘Credit Derivatives,’’ in Klee and Natalucci, ‘‘Profits and Balance few large mergers in mid-2004. For more details, refer to Klee and SheetDevelopments,’’p.154. Natalucci,‘‘ProfitsandBalanceSheetDevelopments,’’p.152.

A88 Federal Reserve Bulletin 2006 TheEffectsofHurricaneKatrina Hurricane Katrina made its Gulf Coast landfall on In addition to the significant operational challenges it August29,2005;bythetimeitlefttheareaithadravaged created, the storm influenced the profitability of the sonearly 90,000 square miles in the Gulf Coast region and called hurricane-affected banks.2 The delinquency rates on displacedabout770,000people,accordingtodatafromthe consumer, real estate, and C&I loans at hurricane-affected Federal Emergency Management Agency (FEMA). Imme- banks spiked in the third quarter, but they declined somediately after the hurricane, the Federal Reserve took a what in the fourth quarter (figure A). To date, chargenumber of steps to help support the operation of financial off rates on loans at these banks have generally shown a institutions in the region. It adjusted its cash operations to smaller effect; the future effect is unknown because of alleviatetransportationburdensandmaintaincashdistribu- the uncertainty regarding ultimate repayment (figure B).3 tion.ItmoveditsNewOrleanscheckclearingoperationsto Nevertheless, hurricane-affected banks have begun to re- Atlantaandworkedwithnumerousfinancialinstitutionsto ducetheirearningsbyprovisioningforelevatedfutureloan address the special processing issues created by the hurri- losses. cane. In addition, the Federal Reserve Bank of Atlanta reminded depository institutions that the discount window 2. ‘‘Hurricane-affected banks’’ are defined as banks with more than 75 percent of their deposits in branches located in counties declared by was available to meet their liquidity needs. As part of FEMAtobe‘‘IndividualAssistance’’disasterareas.Thebanksfromthis ongoing efforts, the public website of the Federal Reserve grouparesmall–nationalandregionalbankstendtobediversifiedbeyond Boardhascontinuedtomaintaininformationclearinghouses theseFEMAcounties—andrepresentalittlemorethan1⁄2percentoftotal industryassets.Largernationalbankstookadditionalloanlossprovisionsbut thataddresshurricane-relatedconcerns.1 appeartobewelldiversifiedbeyondtheregion. 3. Regulatoryagencieshaveencouragedfinancialinstitutionstoconsider reasonableandprudentstepstoeaseburdensonaffectedindividualsandhave issued assurances to institutions that regulators will exercise prudence, discretion,andflexibilitywhenpossibleandappropriateinfulfillingsupervi- 1. Refertowww.federalreserve.gov/hurricanekatrina.htm. soryandregulatoryresponsibilities. A. Delinquency rates on loans at hurricane-affected B. Charge-off rates on loans at hurricane-affected banks, 2000–05 banks, 2000–05 Percent Percent 1.4 5 C&I Consumer 1.2 4 1.0 .8 Consumer 3 C&I .6 2 .4 Real estate Real estate .2 1 + _0 2000 2001 2002 2003 2004 2005 2000 2001 2002 2003 2004 2005 NOTE: The data are quarterly and seasonally adjusted. For definition of NOTE: The data are quarterly and seasonally adjusted. For definition of delinquencies, refer to the note for figure 24 of the main text. For definition net charge-offs, refer to the note for figure 24 of the main text. For of “hurricane-affected banks,” refer to footnote 2 in this box. definition of “hurricane-affected banks,” refer to footnote 2 in this box. abitforotherbanks.Increasesinnon-interestincome, theeffectofnarrowedinterestmargins.Assetquality ledbystrongtradingincome,buoyedindustryprofit- asawholeremainedstrong,andthedelinquencyrate ability. Results were also bolstered by a decline in on loans and leases fell, to 1.57 percent at the end of non-interestexpenseasashareofassets,animprove- the year. Loan loss provisioning as a share of assets mentlargelyreflectingthereversalofanearlierjump wasthesamein2005asin2004despiteprovisioning inchargesrelatedtolitigationandmergers. forlossesrelatedtothehurricanesandtheeffectsofa Solideconomicgrowthalongwithhealthybusiness changeinthebankruptcylawimplementedinthefall balance sheets and rising household wealth contrib- (forfurtherdiscussion,refertobox‘‘TheNewBankutedtoarobustcreditenvironmentthatpartiallyoffset ruptcyLawandItsEffectonCreditCardLoans’’).Net

Profits and Balance Sheet Developments at U.S. Commercial Banks in 2005 A89 TheNewBankruptcyLawandItsEffectonCreditCardLoans UndertheU.S.bankruptcycode,individualswhoareunable charge-offs on securitized credit card loans shot up almost tomeettheirobligationscanfileforbankruptcyeitherunder 130 basis points from September to October and moved chapter7orunderchapter13ofthecode.Underchapter7, up further through December, on net. Bankruptcy filings they can keep assets up to a state-defined exemption level dropped off considerably in mid-October, as many of the and their remaining unsecured debts are discharged; under pre-October17filingswouldapparentlyhavebeensubmitchapter13,theymustprovideaplanunderwhichtheywill tedlaterintheabsenceofthenewlaw.Bankruptcyfilings repayaportionoftheirdebts.OnApril20,2005,President have continued to run at low levels this year, and the BushsignedintolawtheBankruptcyAbuseandConsumer charge-offrateonsecuritizedcreditcardloanshasretreated Protection Act of 2005, which became effective on Octo- from its elevated levels of last fall. Effects of the new ber 17, 2005. The law aims to reduce the number of bankruptcy law on the charge-off rates for other types of bankruptcy filers and diminish creditors’ losses resulting consumer loans and residential mortgages appeared to be frombankruptcy.Beforethepassageofthenewlaw,alarge small. majority of debtors seeking bankruptcy protection filed Banks anticipated the growth in credit card charge-offs under chapter 7. The new law mandates that consumers and provisioned for these changes accordingly. At credit whose incomes are high enough and whose expenses are cardbanks,definedasthosebankshavingcreditcardassets lowenoughmustfileunderchapter13,thusmakingbank- that are greater than 40 percent of their total assets and ruptcyalessattractiveoption. ranked among the 1,000 largest in total assets, the ratio of Consumers rushed to file for bankruptcy in September provisionstoassetsclimbed20basispoints,to3.93percent, and early October as the implementation date of the new from the second to the third quarter and rose another 27 bankruptcylawapproached(figureA).Forthefourthquar- basispointsinthefourthquarter,to4.20percent.Theboost ter as a whole, bankruptcy filings ran at almost twice the inprovisioningpusheddownthereturnonassets(ROA)and pace of earlier in the year. Because credit card loans are return on equity (ROE) of credit card banks, especially in charged off by banks once borrowers are in bankruptcy, the fourth quarter: The ROA in that quarter fell about 50 basispoints,to2.35percent,itslowestlevelsince2000;and theROEdroppedabout230basispoints,to11.65percent, itslowestlevelofthepastdecade.Nonetheless,theannual A. Charge-off rate on credit card debt in securitized pools, and household bankruptcy filings, 1993–2006 returnsforcreditcardbanks—2.89fortheROAand14.18 for the ROE (figure B)—continued to exceed the corre- Percent Filings per 100,000 persons spondingaveragesforbanksasawhole. 900 8 Charge-off rate B. Credit card bank profitability, 1990–2005 on credit card loans 800 7 700 Percent Percent 6 600 30 Return on equity 7 5 500 6 25 400 4 Bankruptcy 5 filings 300 20 3 4 200 15 3 1994 1996 1998 2000 2002 2004 2006 10 2 NOTE: The data for bankruptcy filings are quarterly and extend through Return on assets 2006:Q1; the data for credit cards are monthly at an annual rate and extend 5 1 through March 2006. The charge-off rate is the proportion of total loans outstanding that have been written off as uncollectible. SOURCE: For bankruptcy filings, staff calculations based on data from 1991 1993 1995 1997 1999 2001 2003 2005 Lundquist Consulting, Inc.; for data on credit card debt in securitized pools, Moody’s Investors Service. NOTE: The data are annual. realized gains on investment account securities rela- performedtheS&P500.Forboththe50largestbank tivetoassetsdecreasedabitin2005. holding companies and the 225 largest, much of the As in 2004, dividends were a relatively low frac- gain was concentrated toward the end of the year tion of net income, and robust retained earnings (figure 16). The average spread of rates on banks’ boosted equity capital. Although, on net, bank hold- subordinateddebtoverthoseoncomparable-maturity ing company stocks made gains in 2005, they under- Treasurysecurities,whichwasaboutunchangedfrom

A90 Federal Reserve Bulletin 2006 16. Bank stock prices, by market value of bank, 18. Net interest margin, by size of bank, 1990–2005 and the S&P 500, 2001–06 Percent January 2005 = 100 All banks 4.50 120 S&P 500 4.25 100 4.00 80 3.75 50 largest banks 60 3.50 225 largest banks 40 2001 2002 2003 2004 2005 2006 4.75 NOTE: The data are monthly and extend through March 2006. Stock prices Medium Small are weighted by market value. 4.50 SOURCE: Standard & Poor’s and American Banker. Large 4.25 4.00 the very low levels of 2004 (figure 17), reflected a 3.75 relativelybenignriskoutlookforthebankingsector. 10 largest 3.50 Interest Income and Expense 3.25 3.00 As monetary policy tightened in 2005, the average 1991 1993 1995 1997 1999 2001 2003 2005 rateofreturnearnedonbanks’assetsandtheaverage rate of interest paid on banks’ liabilities moved NOTE: The data are annual. Net interest margin is net interest income divided by average interest-earning assets. For definition of bank size, refer to higher. The average rate earned rose less than the the general note on the first page of the main text. average rate paid, however, which caused the net interest margin to narrow for the third consecutive In line with the narrowing net interest margin, year, to 3.55 percent (figure 18). The rate of narrow- responses to the BLPS over the year indicated that ing was, however, somewhat slower in 2005 than it domesticbanksdecreasedthespreadofratesonC&I hadbeeninrecentyears. loans over their cost of funds (figure 19) and also reduced the costs of credit lines. In response to each 17. Average spread of rates on subordinated debt at 19. Net percentage of selected domestic banks reporting selected bank holding companies, 2002–06 increased spreads of rates on C&I loans over cost of funds, by size of borrower, 1990–2006 Basis points Percent 150 Large and middle-market 60 125 40 20 100 + _0 75 Small 20 40 50 60 80 2002 2003 2004 2005 2006 NOTE: The data are monthly and extend through March 2006. Spreads are 1990 1992 1994 1996 1998 2000 2002 2004 2006 over comparable-maturity Treasury securities. SOURCE: Merrill Lynch bond data. NOTE: Refer to figure 5, general note and source note.

Profits and Balance Sheet Developments at U.S. Commercial Banks in 2005 A91 of the four surveys covering 2005, the respondents divergence perhaps reflects greater competitive presindicated that the narrowing in spreads reflected suresonlargerinstitutionsfromnonbanks. more-aggressive competition from other banks or nonbanklenders.IntheApril2005andJanuary2006 Non-interest Income and Expense surveys,banksindicatedthattighterC&Iloanspreads stemmed from an increased tolerance for risk and Non-interest income at U.S. commercial banks grew improved liquidity in the secondary market for these 6.6 percent in 2005, a pickup from the sluggish loans. Spreads on new investment-grade syndicated 2.5 percent advance posted in 2004. As net interest business loans continued to edge lower over 2005, incomealsoincreased,non-interestincomeasashare while those on leveraged syndicated loans moved up of total revenue remained about flat, on balance, at abitbutremainedquitelowbyhistoricalstandards. about 43 percent (figure 20). The rise in non-interest The decline of the net interest margin for the incomewasfueledbystrongtradingrevenues,which banking sector as a whole was attributable to a drop surged44percentoverallandwhichareconcentrated of21basispointsatthetenlargestbanks;themargin at the ten largest banks. Income from interest rate atotherbankswidenedafewbasispoints.Attheten derivatives contracts and equity derivatives contracts largest banks, the average rate earned on interestgained considerably overall but fluctuated somewhat bearingassetsmovedupabout57basispoints,while duringtheyear. the average rate paid on interest-bearing liabilities Mostothercategoriesofnon-interestincomeposted leaped 100 basis points. Unlike other banks, the ten smallergains.Thegrowthoffiduciaryincomeslowed largest increased the share of their interest-earning to a 5 percent pace, and the level held steady as a assets attributable to securities, which at these banks proportion of revenue. Deposit fees grew at a slower had an average yield of 4.27 percent, a significantly ratethantotalrevenueandmoveddownasashareof loweraverageyieldthanloans,at6.16percent.They deposits for the third consecutive year (figure 21), a alsoincreasedtheshareofmanagedliabilities,which carry higher interest rates, on average, than other 20. Non-interest income and selected components as interest-bearing funding sources. In addition, larger a proportion of revenue, 1990–2005 banks rely more heavily on managed liabilities than smaller banks, and rates on managed liabilities rose Percent morerapidlythanthoseonotherliabilities. Total At large banks (those ranked 11 through 100), the net interest margin was about unchanged in 2005. 45 Likethe10largestbanks,thesebanksrelyheavilyon managed liabilities and saw the rates paid on those 40 liabilities rise significantly over the year. Unlike the 10 largest banks, however, large banks increased the 35 share of loans in their portfolios, which boosted interest earned on assets; and they raised the fraction 30 oftheirassetsfinancedbynon-interest-bearingliabilities,whichhelddowninterestexpense. In contrast to the situation at the ten largest banks, theaveragenetinterestmarginsatsmallandmedium- Selected components sized banks have widened a bit since the onset of 30 monetarypolicytighteningin2004andremainhigher than the average at larger banks. Small and medium- Other non-interest income 25 sizedbanksrelyrelativelymoreoncoredepositsthan 20 onmanagedliabilities,and,asnotedpreviously,core deposits have not repriced upward as much as man- 15 aged liabilities. In addition, real estate loans, which Fiduciary income plus trading income 10 earn relatively high yields, have expanded briskly at these institutions (especially commercial real estate Deposit fees 5 loans), whereas securities, which earn relatively low yields, have declined. Furthermore, rates on C&I 1991 1993 1995 1997 1999 2001 2003 2005 loans moved up significantly more rapidly at small NOTE: The data are annual. Revenue is calculated as the sum of nonand medium-sized banks than at larger banks; the interest income and net interest income.

A92 Federal Reserve Bulletin 2006 21. Deposit fee income as a proportion of total domestic 22. Non-interest expense and selected components as deposits, 1990–2005 a proportion of revenue, 1990–2005 Percent Percent Total .80 70 .75 68 .70 66 .65 64 .60 62 .55 60 .50 58 1991 1993 1995 1997 1999 2001 2003 2005 NOTE: The data are annual. Selected components 35 Other trend perhaps reflecting increased competition for 30 retail deposits, or perhaps reflecting higher spreads 25 Salaries and benefits nowthatbaserateshaverisen. 20 Items recorded under ‘‘other non-interest income’’ grew about 5.2 percent and maintained their share of 15 revenue at 27 percent. Revenues from investment Premises and fixed assets 10 banking fell slightly, while gains from sales of loans fell considerably, perhaps owing to a slowdown in 5 mortgageoriginations.Severaloftheremainingitems in other non-interest income—including check print- 1991 1993 1995 1997 1999 2001 2003 2005 ingfees,automatedtellermachinefees,andtherental NOTE: The data are annual. of safe deposit boxes—advanced solidly last year. However, trends for these items are difficult to infer Other expenses, which barely grew in 2005, fell because an item is reported only if it exceeds 1 perabout 1.5 percentage points as a share of total revcentoftotalincome. enue, to 24.7 percent. The decline in this ratio was The 4.1 percent growth of non-interest expense in greatest at the ten largest banks. Contributing to the 2005wasdownconsiderablyfromthe2004pace,and improved performance at the largest banks were the moderation led to a decline in the ratio of lower expenses related to merger activity and to non-interest expense to total revenue of about 1 perlitigationcharges,bothofwhichhadjumpedatafew centagepoint,to59percent(figure22).Expensesfor largebanksin2004. premises and fixed assets as a share of revenue have changed little over the past five years; according to datafromtheFDIC,thenumberofbranchesincreased Loan Performance and Loss Provisioning a bit. Growth in salary and employee benefits quickened more than 1 percentage point last year, to Indicatorsofcreditqualitygenerallyremainedrobust 7.6percent,buttheratioofsuchcoststototalrevenue last year. The interest-payment ratio of businesses onlyinchedup.Thenumberofbankemployeesgrew declined further from its level in 2004 and ended about2percentoverall.Employmentincreasedatthe 2005 at about its lowest level for the past decade ten largest institutions but declined somewhat at (figure23).Incontrast,thefinancialobligationsratio smaller institutions, a difference likely reflecting, in for households trended higher in 2005, as mortgage part, the effects of bank consolidation. Salaries and debt service increased as a share of income. On net, benefits per employee grew 5.7 percent last year, an loan-loss provisions were little changed as a share of increase reportedly attributable in part to higher in- assets from 2004; however, they increased modestly centivecompensation.Thegainwasanotablepickup in the second half of the year, in part owing to the from the rate in 2004 and faster than the 2005 rise in effects of the hurricanes and the increase in personal compensationperbusiness-sectoremployee. bankruptcies in advance of the implementation of

Profits and Balance Sheet Developments at U.S. Commercial Banks in 2005 A93 23. Interest-payment ratio for businesses, and financial 24. Delinquency and charge-off rates for loans to obligations ratio for households, 1990–2005 businesses, by type of loan, 1990–2005 Percent Percent Interest-payment ratio for nonfinancial corporations Delinquencies 22 12 20 Commercial real estate 9 18 16 6 14 C&I 3 12 + 10 _0 Financial obligations ratio for households Net charge-offs 3.0 19 2.5 Commercial real estate 2.0 18 1.5 17 C&I 1.0 .5 16 + _0 1991 1993 1995 1997 1999 2001 2003 2005 1991 1993 1995 1997 1999 2001 2003 2005 NOTE: The data are quarterly. The interest-payment ratio is calculated as NOTE: The data are quarterly and seasonally adjusted; the data for interest payments as a percentage of cash flow. The financial obligations ratio commercial real estate begin in 1991. Delinquent loans are loans that are not is an estimate of debt payments and recurring obligations as a percentage of accruing interest and those that are accruing interest but are more than thirty disposable personal income; debt payments and recurring obligations consist days past due. The delinquency rate is the end-of-period level of delinquent of required payments on outstanding mortgage debt, consumer debt, auto loans divided by the end-of-period level of outstanding loans. The net leases, rent, homeowner’s insurance, and property taxes. charge-off rate is the annualized amount of charge-offs over the period, net of SOURCE: For interest-payment ratio, national income and product accounts recoveries, divided by the average level of outstanding loans over the period. and Federal Reserve Board; for financial obligations ratio, Federal Reserve For the computation of these rates, commercial real estate loans exclude loans Board (www.federalreserve.gov/releases/housedebt). not secured by real estate (refer to table 1, note 2). new rules in October. As a share of all bank loans, The quality of bank C&I loans may slip a bit in delinquent loans remained low in 2005, and net 2006. Banks were asked in the January 2006 BLPS charge-offs declined a bit, to 0.54 percent of average abouttheoutlookforC&Iloanqualitythisyearunder loans. the assumption that economic activity progresses in line with consensus forecasts. On balance, the responses suggest that banks expect the quality of C&I loans to deteriorate somewhat in 2006 from C&I Loans recentrobustlevels. At the end of 2005, only 1.5 percent of C&I loans were delinquent, the lowest percentage in more than Commercial Real Estate Loans fifteenyears(figure24).Adeclineinthedelinquency rate of more than 40 basis points at the 100 largest The credit quality of CRE loans was supported by bankscontinuedthedownwardtrendindelinquencies improving market conditions in the sector and atthesebanksevidentsince2002.Thenetcharge-off remained strong last year. Vacancy rates in the office rate on C&I loans declined to 0.2 percent by year- sector and industrial sectors continued to trend down end,itslowestlevelsince1997;netcharge-offratesat modestlyfromtheirpeaksafewyearsearlier,andthe larger banks were lower than at medium-sized and vacancyrateintheretailsectorstayedrelativelylow. smallbanks. Rental rates and prices on commercial properties

A94 Federal Reserve Bulletin 2006 25. Delinquency and charge-off rates for loans 26. Credit card delinquency rate and household to households, by type of loan, 1990–2005 bankruptcy filings, 1993–2005 Percent Percent Per 100,000 persons Delinquencies Credit card delinquencies 900 6 5.0 800 Credit card 5 4.5 700 4 600 Other consumer 4.0 3 500 3.5 400 2 Residential real estate Bankruptcy filings 300 3.0 1 200 1993 1995 1997 1999 2001 2003 2005 Net charge-offs 7 NOTE: The data are quarterly and seasonally adjusted. The series shown for bankruptcy filings begins in 1995:Q1. For definition of delinquencies, refer to Credit card 6 the note for figure 24. SOURCE: For bankruptcy filings, staff calculations based on data from 5 Lundquist Consulting, Inc. 4 3 Association (GNMA) considerably boosted delin- 2 quent mortgages in the third and fourth quarters and Other consumer 1 accounted for much of the rise in the delinquency + _0 rate. In addition, broad measures of subprime mort- Residential real estate 1 gage delinquencies have moved up over the past few quarters. 1991 1993 1995 1997 1999 2001 2003 2005 Net charge-offs on residential real estate loans NOTE: The data are quarterly and seasonally adjusted; data for declined to a very low rate of 5 basis points by the delinquencies and for net charge-offs of residential real estate loans begin in fourth quarter as the economy remained strong and 1991. For definitions of delinquencies and net charge-offs, refer to the note for figure 24. house prices appreciated. Although the increases in housepricesmoderatedsomewhatinthelatterpartof the year, gains over the year were still robust and moved up. Against this backdrop, the delinquency boostedhomeowners’equity. rate on CRE loans fell a few basis points, to about Delinquenciesandcharge-offsoncreditcardloans 1.1 percent, its lowest level in more than a decade were significantly affected by a surge of bankruptcy (figure24).Thesedevelopmentshavebeenaccompafilings as households rushed to file before the rule nied in credit markets by relatively low spreads on changes in October (figure 25). As a result, the net commercial-mortgage-backed securities. The net charge-off rate on credit card loans jumped about charge-off rate on commercial real estate loans hov- 1 percentage point in the fourth quarter, to 5.7 pererednearzeroforthesecondyearinarow. cent.Thepatterninbankruptciesandcharge-offswas reflected in loan delinquencies, which had peaked at Loans to Households 3.9percentinthethirdquarterbutfellbacktoendthe yearlowerthanin2004(figure26).Responsestothe The delinquency rate on residential real estate loans January2006BLPSsurveyindicatedthatmuchofthe moved up 25 basis points, to 1.7 percent, over the riseinfourth-quartercreditcardcharge-offscouldbe four quarters of 2005 but was still well below its attributabletotheloansthatwouldhavebeenwritten historical average (figure 25). Delinquency rates off in later quarters if the bankruptcy law had not pickedupbothonone-tofour-familyresidentialreal been changed. Indeed, since October, bankruptcies estate loans and on home equity loans. A midyear havebeenverylow,andlikelyasaconsequence,the change in reporting instructions that requires seller- charge-off rate on securitized credit card loans fell orservicer-bankstorebookdelinquentmortgagesthat sharply in the first few months of 2006 (figure A of they had previously securitized and then sold as box ‘‘The New Bankruptcy Law and Its Effect on issuesbackedbytheGovernmentNationalMortgage CreditCardLoans’’).

Profits and Balance Sheet Developments at U.S. Commercial Banks in 2005 A95 The net charge-off rate on other consumer loans 27. Provisions for loan and lease losses as a rose to a quite elevated level in the third quarter, but proportion of total revenue, 1990–2005 the movement came primarily from a large bank’s Percent change in its accounting for loans extended at its European offices. For the year as a whole, the rate was little changed from 2004. The delinquency rate 20 on other consumer loans trended down in 2005 and ended the year at its lowest level in more than a 15 decade. 10 Securitized Loans 5 Delinquencyratesonmanytypesofsecuritizedloans continued to decline from the relatively low levels posted in 2004. The delinquency rate on securitized 1991 1993 1995 1997 1999 2001 2003 2005 creditcardreceivablesattheendofthefourthquarter, NOTE: The data are annual. about 3.3 percent, was down about 60 basis points fromayearearlier.Althoughthedelinquencyrateon cent. The dollar volume of U.S. banks’ exposures to securitized mortgages rose for a time, it remained India, Mexico, and Brazil, including lending and belowtheaveragelevelin2004andfinishedtheyear derivatives exposures for cross-border and localat just 3.7 percent. The delinquency rate on securioffice operations, rose somewhat, while exposures to tized home equity loans moved up a little and ended other countries remained about flat relative to tier 1 theyearatabout0.8percent,stillabitlowerthanthe capital (table 3). Improving economic prospects in averageleveloverthepastfewyears. IndiaandBrazilprobablycontributedtoariseinU.S. Loss Provisioning 28. Reserves for loan and lease losses, 1990–2005 Percent Banks’ provisioning for loan losses as a share of average net consolidated assets held steady from As a percentage of total loans and leases 3.0 2004, even with the increase in provisioning in the 2.5 third and fourth quarters caused by the surge in 2.0 personal bankruptcies and the hurricanes. Provisions for loan and lease losses as a share of total revenue 1.5 ticked up slightly, to 5.5 percent, but the measure 1.0 remains in the bottom part of its range seen over the pastdecade(figure27). As a percentage of delinquent loans However,charge-offsoutpacedprovisionsin2005, 100 and reserves for loan and lease losses fell about 80 5 percent. The ratio of reserves to total loans and leases fell for the third consecutive year, to 1.4 per- 60 cent, its lowest level in almost twenty years (fig- 40 ure28).Butbecauseofhealthycreditqualityoverall, the ratio of reserves to delinquent loans held about steady at 86 percent; the ratio of reserves to net As a percentage of net charge-offs charge-offs changed little and remained near the 500 middleofitsrangeofthepastseveralyears. 400 300 200 INTERNATIONAL OPERATIONS OF U.S. COMMERCIAL BANKS 100 1991 1993 1995 1997 1999 2001 2003 2005 In 2005, the share of bank assets booked in foreign NOTE: The data are annual. For definitions of delinquencies and net offices increased about 40 basis points, to 11.8 per- charge-offs, refer to the note for figure 24.

A96 Federal Reserve Bulletin 2006 3. ExposureofU.S.bankstoselectedcountriesatyear-endrelativetotier1capital,bybanksize,1997–2005 Percent EasternEurope Selected andRussia LatinAmerica Banksizeandyear Asian India Total countries1 All Russia All Mexico Argentina Brazil All 1997........................................ 16.11 1.48 3.47 1.80 29.67 5.48 5.84 9.74 50.73 1998........................................ 15.49 2.35 3.49 .43 42.93 9.88 9.66 11.27 64.26 1999........................................ 14.37 2.39 2.85 .37 39.00 9.50 9.40 10.49 58.61 2000........................................ 13.17 2.63 4.35 .49 37.88 9.08 8.41 11.15 58.03 2001........................................ 12.09 2.55 4.29 .60 54.06 25.97 6.61 2.99 72.99 2002........................................ 11.44 2.74 5.53 1.06 38.90 20.80 2.44 8.36 58.61 2003........................................ 11.15 3.86 5.44 1.48 32.85 17.95 1.73 6.77 53.30 2004........................................ 20.33 4.16 6.09 1.54 31.78 16.65 1.47 6.51 62.36 2005........................................ 17.77 4.92 5.87 1.95 31.84 17.36 1.34 6.94 60.40 Moneycenterandotherlargebanks 1997........................................ 26.87 2.58 6.12 3.16 48.37 8.40 10.01 16.13 83.94 1998........................................ 24.02 4.19 5.61 .68 64.20 14.10 15.19 17.04 98.02 1999........................................ 20.73 3.56 4.25 .55 53.90 12.62 13.63 14.53 82.44 2000........................................ 19.98 4.14 6.83 .77 54.98 12.69 12.68 16.40 85.93 2001........................................ 17.88 3.86 6.47 .91 79.08 34.54 9.79 18.74 107.29 2002........................................ 16.96 4.18 8.17 1.63 57.32 31.14 3.65 12.38 86.63 2003........................................ 16.98 5.93 8.41 2.29 49.19 27.13 2.64 10.02 80.51 2004........................................ 30.95 6.31 9.34 2.36 46.96 24.99 2.22 9.59 93.56 2005........................................ 27.16 7.52 8.93 2.96 47.39 26.16 2.04 10.35 91.00 Otherbanks 1997........................................ 2.34 .07 .08 .05 5.73 1.75 .51 1.56 8.22 1998........................................ 2.08 .05 .16 .00 9.51 3.24 .97 .00 11.80 1999........................................ 1.75 .07 .08 .01 9.41 3.31 1.01 2.47 11.31 2000........................................ 1.41 .03 .08 .00 8.35 2.84 1.04 2.08 9.87 2001........................................ 1.07 .06 .14 .00 6.45 2.04 .57 2.05 7.72 2002........................................ 1.03 .08 .65 .00 5.00 1.86 .02 .96 6.76 2003........................................ .90 .24 .21 .06 4.20 1.53 .13 1.05 5.55 2004........................................ .90 .21 .14 .04 4.00 1.39 .09 .85 5.25 2005........................................ .55 .16 .25 .10 3.37 1.24 .06 .69 1.06 Memo Totalexposure(billionsofdollars) 1997........................................ 55.24 5.07 11.91 6.16 101.73 18.80 20.03 33.40 173.96 1998........................................ 37.87 5.43 8.53 1.05 104.69 24.15 23.62 27.55 156.52 1999........................................ 37.45 6.23 7.43 .95 101.63 24.77 24.51 27.34 152.74 2000........................................ 37.30 7.46 12.33 1.39 107.31 25.71 23.82 31.59 164.40 2001........................................ 36.32 7.66 12.88 1.80 162.39 78.00 19.87 39.01 219.25 2002........................................ 36.32 8.70 17.55 3.37 123.53 66.15 7.75 26.55 186.10 2003........................................ 39.12 13.55 19.07 5.20 115.23 62.98 6.07 23.74 186.97 2004........................................ 79.57 16.27 23.85 6.02 124.39 65.17 5.75 25.46 244.07 2005........................................ 77.92 21.58 25.73 8.55 139.66 76.14 5.87 30.42 264.89 Note: Forthedefinitionoftier1capital,seetextnote11.Exposuresconsist lionintier1capital;theremainingfifty-nine(‘‘other’’)bankshad$154.9biloflendingandderivativesexposuresforcross-borderandlocal-officeopera- lionintier1capital.Theaverage‘‘other’’bankatyear-end2005had$33.8biltions.Respondentsmayfileinformationononebankoronthebankholding lioninassets. companyasawhole. 1. Indonesia,Korea,Malaysia,Philippines,andThailand. Theyear-end2005datacoversixty-sevenbankswithatotalof$438.6bil- Source: Federal Financial Institutions Examination Council (2006), Stalionintier1capital;oftheseinstitutions,fourweremoneycenterbanks,with tistical Release E.16, ‘‘Country Exposure Lending Survey’’ (March 31), $217.5billionintier1capital,andfourwereotherlargebanks,with$66.2bil- www.ffiec.gov/E16.htm. banks’ exposures to these countries; most of the ten-yearnominalTreasurysecuritiesincreasednearly increasewasatmoneycenterandotherlargebanks. 50basispoints. Weekly data published by the Federal Reserve indicates that asset growth at commercial banks DEVELOPMENTS IN EARLY 2006 remained rapid in the first quarter of 2006.20 Expansion in securities was supported by increased hold- The economy expanded briskly over the first three ings of Treasury and agency debt, and loan growth months of 2006. Growth in consumer spending and remained vigorous. C&I loans continued to advance businesspurchasesrebounded,andlevelsofresource briskly,and,althoughthegrowthinloanssecuredby utilization increased. The prices of crude oil and of real estate slowed, it remained near a double-digit some other commodities moved higher, and conpace. Robust increases in deposits continued to help sumer energy prices rose further, but core inflation fund the strong expansion in assets over the first remained relatively low. In these circumstances, the quarter. Federal Reserve firmed policy by raising the target federal funds rate 25 basis points at each of its first twomeetingsin2006.Longer-terminterestratesalso 20. StatisticalReleaseH.8,‘‘AssetsandLiabilitiesofCommercial rose considerably over the first quarter—yields on BanksintheUnitedStates’’(www.federalreserve.gov/releases/h8).

Profits and Balance Sheet Developments at U.S. Commercial Banks in 2005 A97 Bank profitability appeared to remain strong in Reflecting strong balance sheets and continuing early2006accordingtothestatementsonfirst-quarter profitabilityoverthefirstquarter,stockpricesoflarge earnings of several large bank holding companies. bank holding companies generally rose a bit more Although results varied across institutions, some thandidtheequitypricesoflargefirmsingeneral,as trimmedtheirloanlossprovisionsinthefirstquarter, measured by the S&P 500 stock index. However, the citing solid asset quality. The pace of personal bank- gains in bank stock prices were in line with those of ruptcyfilingsremainedwellbelowtheleveltheyhad the broader market, as measured by the Wilshire reached before the new bankruptcy law took effect 5000. Roughly the same number of mergers— last October. Non-interest income was supported by consolidations of one bank charter into another— gains in fee income and trading revenue. Despite occurred in the first quarter of 2006 as in the first continued increases in deposit rates, changes in net quarterof2005,butthemergersin2006havegenerinterestmarginsappeartohavebeenmixed.Andwith allyinvolvedsignificantlysmaller-sizedinstitutions. bank assets expanding notably, growth in net interest incomeappearstohavebeensolid.

A98 Federal Reserve Bulletin 2006 A.1. Portfoliocomposition,interestrates,andincomeandexpense,U.S.banks,1996–2005 A. Allbanks Item 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 Balancesheetitemsasapercentageofaveragenetconsolidatedassets Interest-earningassets ............................ 87.38 87.15 86.76 87.03 87.13 86.49 86.42 86.08 86.90 86.83 Loansandleases,net........................... 59.91 58.72 58.33 59.34 60.48 58.95 57.83 56.88 56.98 57.88 Commercialandindustrial ................... 15.59 15.77 16.36 17.07 17.16 16.08 14.07 12.18 11.06 11.17 U.S.addressees............................ 13.06 13.17 13.61 14.43 14.67 13.69 12.04 10.48 9.52 9.63 Foreignaddressees ........................ 2.53 2.60 2.75 2.64 2.49 2.39 2.04 1.70 1.54 1.53 Consumer ................................... 12.27 11.50 10.41 9.71 9.38 9.23 9.35 9.06 9.18 9.12 Creditcard................................ 4.93 4.62 4.02 3.51 3.52 3.63 3.78 3.55 3.86 4.05 Installmentandother ...................... 7.34 6.88 6.39 6.20 5.87 5.60 5.57 5.51 5.31 5.06 Realestate .................................. 25.04 25.00 24.85 25.44 27.04 27.10 28.39 29.91 30.78 32.40 Indomesticoffices ........................ 24.42 24.39 24.28 24.87 26.49 26.60 27.91 29.46 30.25 31.84 Constructionandlanddevelopment ...... 1.63 1.73 1.86 2.18 2.51 2.85 2.98 2.99 3.25 3.89 Farmland ............................... .56 .55 .55 .56 .56 .55 .56 .54 .54 .54 One-tofour-familyresidential........... 14.42 14.41 14.25 14.10 14.96 14.67 15.40 16.96 17.42 18.27 Homeequity ......................... 1.85 1.94 1.89 1.76 1.96 2.18 2.80 3.40 4.34 4.95 Other ................................ 12.57 12.47 12.37 12.34 13.00 12.49 12.60 13.57 13.09 13.32 Multifamilyresidential .................. .85 .83 .82 .88 .99 .97 1.02 1.05 1.06 1.08 Nonfarmnonresidential.................. 6.96 6.88 6.80 7.15 7.48 7.56 7.95 7.91 7.98 8.06 Inforeignoffices .......................... .63 .61 .57 .57 .54 .50 .48 .46 .53 .56 Todepositoryinstitutionsandacceptances ofotherbanks .......................... 2.33 1.93 1.91 1.96 1.87 1.83 1.87 1.98 2.11 1.73 Foreigngovernments ........................ .26 .18 .15 .16 .12 .10 .09 .08 .08 .06 Agriculturalproduction ...................... .92 .90 .89 .83 .78 .75 .70 .63 .59 .56 Otherloans.................................. 3.32 2.80 2.78 2.75 2.58 2.34 2.06 2.00 2.35 2.09 Lease-financingreceivables .................. 1.51 1.87 2.13 2.52 2.63 2.58 2.44 2.11 1.79 1.58 Less:Unearnedincomeonloans ............. −.12 −.09 −.07 −.06 −.05 −.04 −.05 −.04 −.04 −.03 Less:Lossreserves1 ......................... −1.21 −1.13 −1.07 −1.04 −1.02 −1.04 −1.11 −1.04 −.91 −.79 Securities...................................... 21.00 20.40 20.37 20.40 20.01 19.53 21.27 21.90 22.57 22.05 Investmentaccount .......................... 18.19 17.23 17.48 18.33 17.59 16.82 18.30 18.97 18.99 17.87 Debt...................................... 17.74 16.74 16.93 17.73 16.93 16.48 17.99 18.72 18.79 17.71 U.S.Treasury ........................... 4.19 3.38 2.71 2.14 1.66 .85 .78 .90 .89 .62 U.S.governmentagencyand corporationobligations ............. 9.74 9.73 10.28 10.85 10.31 10.08 11.46 12.26 12.37 11.51 Government-backedmortgagepools ... 4.80 4.93 5.16 5.24 4.75 5.13 6.09 6.75 7.13 6.78 Collateralizedmortgageobligations .... 2.11 1.93 2.12 2.15 1.92 1.95 2.35 2.34 2.01 1.80 Other ................................ 2.83 2.86 2.99 3.46 3.63 2.99 3.02 3.17 3.22 2.93 Stateandlocalgovernment .............. 1.68 1.59 1.57 1.62 1.52 1.49 1.49 1.48 1.41 1.36 Privatemortgage-backedsecurities ....... .61 .50 .67 .88 .95 1.09 1.25 1.30 1.41 1.76 Other................................... 1.51 1.54 1.70 2.24 2.48 2.98 3.01 2.78 2.72 2.47 Equity .................................... .45 .50 .55 .61 .66 .34 .31 .25 .20 .16 Tradingaccount ............................. 2.81 3.16 2.90 2.06 2.43 2.72 2.97 2.93 3.58 4.17 GrossfederalfundssoldandreverseRPs........ 3.81 5.18 5.37 4.61 4.12 5.11 4.81 4.85 4.58 4.75 Interest-bearingbalancesatdepositories ......... 2.66 2.86 2.69 2.68 2.52 2.90 2.52 2.45 2.76 2.15 Non-interest-earningassets........................ 12.62 12.85 13.24 12.97 12.87 13.51 13.58 13.92 13.10 13.17 Revaluationgainsheldintradingaccounts ...... 2.24 2.59 2.95 2.57 2.28 2.37 2.42 2.70 2.19 1.82 Other.......................................... 10.37 10.26 10.29 10.40 10.58 11.15 11.16 11.22 10.91 11.36 Liabilities ........................................ 91.73 91.57 91.51 91.52 91.58 91.25 90.85 90.96 90.57 89.91 Coredeposits .................................. 52.72 50.89 49.43 48.60 46.52 47.07 48.98 49.18 48.56 47.54 Transactiondeposits ......................... 17.57 15.76 14.10 12.58 11.07 10.36 10.06 9.73 9.10 8.46 Demanddeposits .......................... 12.81 12.15 10.99 9.78 8.61 8.00 7.67 7.26 6.58 6.16 Othercheckabledeposits .................. 4.75 3.61 3.11 2.81 2.46 2.36 2.39 2.47 2.52 2.30 Savingsdeposits(includingMMDAs)......... 19.07 19.76 20.87 22.47 22.43 24.53 28.13 30.12 31.19 30.85 Smalltimedeposits .......................... 16.08 15.37 14.46 13.55 13.01 12.18 10.80 9.33 8.27 8.24 Managedliabilities2............................ 32.77 34.13 34.97 36.59 38.83 37.42 35.05 34.61 35.69 36.22 Largetimedeposits .......................... 6.52 7.25 7.67 7.89 8.76 8.89 8.30 8.09 8.00 9.08 Depositsbookedinforeignoffices ............ 10.45 10.48 10.59 10.96 11.43 10.66 9.42 9.38 10.24 10.39 Subordinatednotesanddebentures............ 1.07 1.15 1.30 1.36 1.37 1.43 1.40 1.33 1.30 1.34 GrossfederalfundspurchasedandRPs ....... 7.18 8.13 7.98 7.97 7.83 7.95 7.77 7.75 7.24 7.05 Othermanagedliabilities..................... 7.54 7.13 7.43 8.41 9.44 8.49 8.16 8.06 8.91 8.37 Revaluationlossesheldintradingaccounts...... 2.14 2.64 2.97 2.52 2.29 2.21 2.09 2.30 1.95 1.67 Other.......................................... 4.10 3.91 4.14 3.81 3.94 4.54 4.73 4.87 4.36 4.47 Capitalaccount................................... 8.27 8.43 8.49 8.48 8.42 8.75 9.15 9.04 9.43 10.09 Memo Commercialrealestateloans3..................... 9.91 9.98 10.11 10.87 11.58 12.09 12.57 12.47 12.78 13.51 Otherrealestateowned4.......................... .14 .11 .08 .06 .05 .05 .06 .06 .06 .04 Mortgage-backedsecurities ....................... 7.53 7.37 7.96 8.27 7.63 8.17 9.69 10.39 10.56 10.33 FederalHomeLoanBankadvances ............... n.a. n.a. n.a. n.a. n.a. 2.89 3.17 3.19 3.07 3.04 Averagenetconsolidatedassets (billionsofdollars) .......................... 4,379 4,737 5,148 5,439 5,906 6,334 6,635 7,249 7,879 8,592

Profits and Balance Sheet Developments at U.S. Commercial Banks in 2005 A99 A.1.—Continued A. Allbanks—Continued Item 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 Effectiveinterestrate(percent)5 Ratesearned Interest-earningassets ......................... 8.16 8.17 8.02 7.71 8.20 7.37 6.11 5.30 5.10 5.71 Taxableequivalent........................ 8.22 8.23 8.07 7.76 8.26 7.42 6.15 5.34 5.14 5.75 Loansandleases,gross...................... 9.01 9.03 8.85 8.47 9.00 8.15 6.90 6.16 5.91 6.52 Netoflossprovisions .................. 8.56 8.50 8.30 7.97 8.33 7.15 5.85 5.48 5.48 6.10 Securities................................... 6.46 6.54 6.45 6.27 6.47 6.04 4.95 3.96 3.86 4.18 Taxableequivalent ..................... 6.66 6.73 6.63 6.46 6.65 6.22 5.10 4.10 3.99 4.30 Investmentaccount ....................... 6.39 6.50 6.38 6.25 6.45 6.05 5.04 4.00 3.96 4.29 U.S.TreasurysecuritiesandU.S. governmentagencyobligations (excludingMBS) .................. n.a. n.a. n.a. n.a. n.a. 5.76 4.42 3.29 3.11 3.46 Mortgage-backedsecurities ............. n.a. n.a. n.a. n.a. n.a. 6.45 5.44 4.24 4.38 4.60 Other .................................. n.a. n.a. n.a. n.a. n.a. 5.60 4.74 4.08 3.76 4.22 Tradingaccount .......................... 6.86 6.75 6.85 6.47 6.63 6.01 4.38 3.71 3.35 3.72 GrossfederalfundssoldandreverseRPs..... 5.21 5.45 5.29 4.78 5.56 3.86 1.93 1.43 1.43 2.65 Interest-bearingbalancesatdepositories ...... 6.20 6.23 6.32 5.95 6.48 4.01 2.79 2.09 1.98 3.70 Ratespaid Interest-bearingliabilities ...................... 4.82 4.92 4.88 4.47 5.17 4.15 2.54 1.87 1.77 2.68 Interest-bearingdeposits ..................... 4.34 4.39 4.31 3.87 4.45 3.61 2.12 1.48 1.37 2.05 Inforeignoffices ......................... 5.54 5.44 5.66 4.91 5.61 3.95 2.38 1.64 1.77 2.78 Indomesticoffices........................ 4.07 4.16 4.01 3.63 4.17 3.54 2.07 1.45 1.29 1.91 Othercheckabledeposits ............... 2.04 2.25 2.29 2.08 2.34 1.96 1.06 .75 .77 1.41 Savings(includingMMDAs)............ 3.00 2.93 2.79 2.49 2.86 2.19 1.13 .74 .72 1.24 Largetimedeposits6 ................... 5.39 5.45 5.22 4.92 5.78 5.04 3.38 2.59 2.35 3.19 Othertimedeposits6.................... 5.40 5.54 5.48 5.09 5.69 5.43 3.73 2.91 2.56 3.10 GrossfederalfundspurchasedandRPs....... 5.12 5.17 5.19 4.73 5.77 3.84 1.88 1.30 1.55 3.09 Otherinterest-bearingliabilities .............. 6.92 6.94 6.89 6.48 6.97 5.92 4.32 3.59 3.26 4.50 Incomeandexpenseasapercentageofaveragenetconsolidatedassets Grossinterestincome.......................... 7.16 7.15 6.99 6.73 7.18 6.38 5.28 4.55 4.44 4.98 Taxableequivalent........................ 7.22 7.21 7.04 6.78 7.22 6.43 5.32 4.59 4.48 5.02 Loans ...................................... 5.48 5.41 5.27 5.12 5.53 4.92 4.07 3.56 3.42 3.82 Securities................................... 1.16 1.11 1.10 1.14 1.15 1.00 .89 .74 .74 .77 GrossfederalfundssoldandreverseRPs..... .21 .29 .29 .23 .23 .20 .09 .07 .07 .13 Other....................................... .32 .35 .32 .24 .27 .27 .22 .18 .21 .26 Grossinterestexpense ......................... 3.43 3.48 3.46 3.22 3.76 2.98 1.79 1.30 1.26 1.89 Deposits.................................... 2.46 2.48 2.43 2.20 2.56 2.09 1.23 .87 .81 1.23 GrossfederalfundspurchasedandRPs....... .38 .43 .43 .39 .45 .31 .15 .10 .12 .23 Other....................................... .59 .57 .60 .63 .75 .58 .41 .33 .33 .44 Netinterestincome............................ 3.73 3.68 3.53 3.51 3.41 3.40 3.48 3.25 3.18 3.09 Taxableequivalent........................ 3.79 3.73 3.57 3.56 3.46 3.45 3.52 3.28 3.22 3.13 Lossprovisions7 .............................. .37 .41 .42 .39 .50 .68 .68 .45 .30 .30 Non-interestincome ........................... 2.18 2.23 2.40 2.66 2.59 2.54 2.54 2.53 2.39 2.33 Servicechargesondeposits.................. .39 .39 .38 .40 .40 .42 .45 .44 .42 .39 Fiduciaryactivities .......................... .33 .35 .37 .38 .38 .35 .32 .31 .32 .31 Tradingrevenue ............................ .17 .17 .15 .19 .21 .20 .16 .16 .13 .17 Interestrateexposures .................... .09 .08 .05 .07 .08 .09 .08 .06 .03 .05 Foreignexchangerateexposures .......... .06 .08 .09 .09 .08 .07 .07 .07 .07 .07 Othercommodityandequityexposures .... .02 * .01 .03 .04 .03 .01 .02 .03 .04 Other....................................... 1.29 1.32 1.49 1.69 1.61 1.57 1.61 1.63 1.52 1.47 Non-interestexpense .......................... 3.71 3.61 3.77 3.76 3.66 3.57 3.47 3.36 3.34 3.19 Salaries,wages,andemployeebenefits ....... 1.55 1.53 1.55 1.58 1.51 1.49 1.51 1.50 1.46 1.44 Occupancy ................................. .48 .47 .47 .48 .45 .44 .44 .43 .42 .41 Other....................................... 1.69 1.62 1.75 1.70 1.70 1.64 1.52 1.43 1.46 1.34 Netnon-interestexpense....................... 1.54 1.38 1.36 1.11 1.07 1.03 .93 .82 .95 .86 Gainsoninvestmentaccountsecurities ......... .03 .04 .06 * −.04 .07 .10 .08 .05 * Incomebeforetaxesandextraordinaryitems .... 1.85 1.92 1.81 2.02 1.81 1.77 1.97 2.05 1.97 1.93 Taxes ...................................... .65 .68 .62 .72 .63 .59 .65 .67 .64 .62 Extraordinaryitems,netofincometaxes ..... * * .01 * * −.01 * .01 * * Netincome ................................... 1.20 1.25 1.20 1.31 1.18 1.17 1.32 1.39 1.34 1.31 Cashdividendsdeclared ..................... .90 .90 .80 .96 .89 .87 1.01 1.07 .76 .75 Retainedincome ............................ .30 .35 .40 .35 .29 .31 .30 .31 .58 .56 Memo:Returnonequity....................... 14.51 14.83 14.08 15.39 13.97 13.40 14.40 15.33 14.16 13.01 1. Includesallocatedtransferriskreserve. 5. Whenpossible,basedontheaverageofquarterlybalancesheetdatare- 2. Measuredasthesumoflargetimedepositsindomesticoffices,deposits portedonscheduleRC-KofthequarterlyCallReport. bookedinforeignoffices,subordinatednotesanddebentures,federalfundspur- 6. Before1997,largetimedepositopenaccountswereincludedinothertime chasedandsecuritiessoldunderrepurchaseagreements,FederalHomeLoan deposits. Bankadvances,andotherborrowedmoney. 7. Includesprovisionsforallocatedtransferrisk. 3. Measuredasthesumofconstructionandlanddevelopmentloanssecured *Inabsolutevalue,lessthan0.005percent. byrealestate;realestateloanssecuredbynonfarmnonresidentialpropertiesor n.a. Notavailable. bymultifamilyresidentialproperties;andloanstofinancecommercialreales- MMDA Moneymarketdepositaccount. tate,construction,andlanddevelopmentactivitiesnotsecuredbyrealestate. RP Repurchaseagreement. 4. Otherrealestateownedisacomponentofothernon-interest-earningassets. MBS Mortgage-backedsecurities.

A100 Federal Reserve Bulletin 2006 A.1. Portfoliocomposition,interestrates,andincomeandexpense,U.S.banks,1996–2005 B. Tenlargestbanksbyassets Item 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 Balancesheetitemsasapercentageofaveragenetconsolidatedassets Interest-earningassets ............................ 80.12 81.84 81.25 81.49 82.23 81.74 81.68 81.39 83.54 83.96 Loansandleases,net........................... 53.51 50.91 50.76 53.37 55.22 53.86 53.61 52.20 51.29 51.35 Commercialandindustrial ................... 17.17 16.90 18.07 19.20 19.87 18.82 16.16 12.98 10.54 10.61 U.S.addressees............................ 9.59 10.24 11.76 13.14 13.95 13.42 11.69 9.40 7.49 7.74 Foreignaddressees ........................ 7.59 6.66 6.31 6.06 5.92 5.41 4.47 3.59 3.06 2.87 Consumer ................................... 6.22 6.40 6.04 5.94 5.43 6.17 7.82 7.96 8.49 8.80 Creditcard................................ 1.23 1.34 1.30 1.36 1.34 1.64 2.90 2.81 3.19 3.60 Installmentandother ...................... 4.99 5.06 4.74 4.58 4.09 4.53 4.92 5.15 5.30 5.21 Realestate .................................. 16.53 17.42 16.51 16.96 19.82 19.23 20.78 22.68 23.21 24.55 Indomesticoffices ........................ 14.44 15.69 15.08 15.55 18.48 18.05 19.70 21.74 22.21 23.52 Constructionandlanddevelopment ...... .51 .68 .77 .90 .98 1.27 1.42 1.36 1.40 1.70 Farmland ............................... .06 .09 .09 .10 .11 .11 .12 .10 .10 .10 One-tofour-familyresidential........... 10.43 11.02 10.33 10.77 13.37 12.41 13.51 16.03 16.71 17.73 Homeequity ......................... 1.53 1.70 1.72 1.54 1.61 1.78 2.35 2.96 4.04 5.22 Other ................................ 8.90 9.31 8.61 9.22 11.76 10.63 11.17 13.07 12.67 12.52 Multifamilyresidential .................. .38 .39 .38 .43 .60 .51 .55 .47 .45 .44 Nonfarmnonresidential.................. 3.05 3.52 3.51 3.35 3.42 3.76 4.09 3.78 3.55 3.55 Inforeignoffices .......................... 2.09 1.73 1.43 1.41 1.34 1.18 1.08 .94 1.00 1.03 Todepositoryinstitutionsandacceptances ofotherbanks .......................... 6.14 4.20 4.05 4.34 3.78 3.23 3.20 3.54 4.10 3.15 Foreigngovernments ........................ .69 .45 .35 .38 .28 .20 .20 .17 .16 .12 Agriculturalproduction ...................... .23 .31 .28 .26 .23 .28 .23 .19 .22 .20 Otherloans.................................. 6.34 4.15 3.74 3.96 3.75 3.51 2.94 2.87 3.32 2.81 Lease-financingreceivables .................. 1.59 2.24 2.81 3.40 3.07 3.43 3.44 2.87 2.08 1.78 Less:Unearnedincomeonloans ............. −.11 −.07 −.06 −.05 −.04 −.04 −.08 −.06 −.04 −.04 Less:Lossreserves1 ......................... −1.30 −1.08 −1.01 −1.03 −.97 −.97 −1.12 −1.02 −.80 −.65 Securities...................................... 19.83 20.00 19.72 18.34 18.98 17.81 20.54 21.22 22.95 23.37 Investmentaccount .......................... 10.60 10.97 12.12 13.08 13.71 12.14 14.36 15.31 15.99 15.59 Debt...................................... 10.22 10.55 11.64 12.57 13.03 11.88 14.13 15.11 15.83 15.44 U.S.Treasury ........................... 1.93 1.56 1.70 1.98 1.96 .68 .59 .82 .86 .56 U.S.governmentagencyand corporationobligations ............. 4.59 5.34 6.31 6.35 6.59 6.84 8.69 9.20 9.92 9.69 Government-backedmortgagepools ... 3.58 4.26 5.13 5.03 4.88 4.99 6.38 7.59 8.64 8.65 Collateralizedmortgageobligations .... .95 .93 .93 .79 .93 1.11 1.52 .91 .70 .54 Other ................................ .06 .15 .26 .52 .78 .74 .79 .70 .58 .50 Stateandlocalgovernment .............. .39 .51 .47 .45 .51 .55 .59 .59 .57 .58 Privatemortgage-backedsecurities ....... .30 .32 .60 .57 .51 .58 .92 1.10 .96 1.18 Other................................... 3.01 2.81 2.57 3.22 3.47 3.22 3.34 3.40 3.52 3.43 Equity .................................... .38 .42 .47 .51 .68 .26 .22 .20 .16 .14 Tradingaccount ............................. 9.23 9.03 7.60 5.25 5.26 5.67 6.18 5.91 6.96 7.79 GrossfederalfundssoldandreverseRPs........ 3.10 7.56 7.81 6.64 5.02 6.38 5.26 5.79 6.37 6.96 Interest-bearingbalancesatdepositories ......... 3.68 3.37 2.96 3.14 3.01 3.69 2.28 2.18 2.93 2.28 Non-interest-earningassets........................ 19.88 18.16 18.75 18.51 17.77 18.26 18.32 18.61 16.46 16.04 Revaluationgainsheldintradingaccounts ...... 7.63 7.36 7.62 6.66 5.66 5.48 5.40 5.79 4.45 3.50 Other.......................................... 12.25 10.80 11.13 11.85 12.11 12.78 12.93 12.83 12.01 12.54 Liabilities ........................................ 93.04 92.61 92.58 92.28 92.36 92.14 91.52 91.94 91.64 90.81 Coredeposits .................................. 29.12 31.66 32.94 33.76 33.28 36.38 40.61 41.07 42.02 40.18 Transactiondeposits ......................... 11.34 10.19 9.45 8.55 8.01 8.40 8.34 7.74 6.65 6.05 Demanddeposits .......................... 9.73 8.98 8.46 7.83 7.28 7.50 7.40 6.72 5.43 4.90 Othercheckabledeposits .................. 1.61 1.21 .99 .72 .74 .90 .95 1.02 1.22 1.15 Savingsdeposits(includingMMDAs)......... 12.93 15.32 17.07 18.94 19.24 22.21 26.82 28.99 31.54 30.11 Smalltimedeposits .......................... 4.85 6.15 6.42 6.26 6.03 5.77 5.44 4.34 3.83 4.02 Managedliabilities2............................ 47.39 46.02 44.42 45.49 46.84 43.41 38.89 38.60 39.33 40.83 Largetimedeposits .......................... 3.04 4.17 5.04 5.19 5.55 5.46 5.13 5.53 5.21 6.28 Depositsbookedinforeignoffices ............ 27.78 23.39 21.23 22.22 22.76 20.28 17.31 16.62 17.20 17.51 Subordinatednotesanddebentures............ 1.90 1.80 1.89 1.98 2.10 2.16 2.11 1.92 1.78 1.89 GrossfederalfundspurchasedandRPs ....... 5.88 10.26 9.78 8.84 8.89 9.04 8.83 8.62 7.79 8.39 Othermanagedliabilities..................... 8.79 6.40 6.49 7.27 7.55 6.47 5.53 5.90 7.35 6.76 Revaluationlossesheldintradingaccounts...... 7.27 7.53 7.67 6.51 5.69 5.10 4.63 4.88 3.95 3.21 Other.......................................... 9.26 7.39 7.55 6.52 6.55 7.26 7.39 7.40 6.34 6.60 Capitalaccount................................... 6.96 7.39 7.42 7.72 7.64 7.86 8.48 8.06 8.36 9.19 Memo Commercialrealestateloans3..................... 4.65 5.45 5.61 5.69 5.87 6.68 6.92 6.31 5.99 6.33 Otherrealestateowned4.......................... .18 .13 .09 .06 .04 .04 .03 .03 .03 .02 Mortgage-backedsecurities ....................... 4.83 5.52 6.65 6.40 6.32 6.68 8.82 9.60 10.30 10.36 FederalHomeLoanBankadvances ............... n.a. n.a. n.a. n.a. n.a. .82 .82 .84 .79 .63 Averagenetconsolidatedassets (billionsofdollars) .......................... 1,189 1,514 1,820 1,935 2,234 2,527 2,785 3,148 3,654 4,232

Profits and Balance Sheet Developments at U.S. Commercial Banks in 2005 A101 A.1.—Continued B. Tenlargestbanksbyassets—Continued Item 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 Effectiveinterestrate(percent)5 Ratesearned Interest-earningassets ......................... 7.72 7.57 7.55 7.37 7.76 6.83 5.83 5.01 4.72 5.29 Taxableequivalent........................ 7.74 7.60 7.57 7.39 7.78 6.86 5.86 5.03 4.74 5.31 Loansandleases,gross...................... 8.32 8.25 8.21 7.99 8.46 7.50 6.54 5.78 5.53 6.16 Netoflossprovisions .................. 8.31 8.10 7.77 7.65 7.92 6.55 5.32 5.21 5.30 5.84 Securities................................... 6.80 6.78 6.83 6.58 6.48 6.23 5.04 4.15 4.04 4.27 Taxableequivalent ..................... 6.85 6.85 6.89 6.65 6.55 6.31 5.11 4.21 4.10 4.33 Investmentaccount ....................... 6.70 6.76 6.78 6.59 6.40 6.23 5.30 4.26 4.37 4.63 U.S.TreasurysecuritiesandU.S. governmentagencyobligations (excludingMBS) .................. n.a. n.a. n.a. n.a. n.a. 5.01 3.74 2.62 2.92 3.29 Mortgage-backedsecurities ............. n.a. n.a. n.a. n.a. n.a. 6.42 5.55 4.51 4.83 4.93 Other .................................. n.a. n.a. n.a. n.a. n.a. 6.34 5.30 4.28 3.76 4.26 Tradingaccount .......................... 6.90 6.81 6.92 6.56 6.70 6.24 4.46 3.87 3.32 3.57 GrossfederalfundssoldandreverseRPs..... 4.92 5.45 5.20 4.52 4.93 3.86 2.20 1.66 1.47 2.45 Interest-bearingbalancesatdepositories ...... 6.71 6.91 7.16 7.22 7.43 3.73 3.40 2.49 1.80 4.06 Ratespaid Interest-bearingliabilities ...................... 5.44 5.41 5.29 4.79 5.37 4.09 2.55 1.86 1.80 2.80 Interest-bearingdeposits ..................... 4.57 4.54 4.40 3.82 4.40 3.27 1.95 1.36 1.30 2.00 Inforeignoffices ......................... 5.62 5.52 5.83 4.99 5.67 4.02 2.59 1.76 1.87 2.79 Indomesticoffices........................ 3.32 3.69 3.39 3.04 3.51 2.84 1.68 1.20 1.08 1.68 Othercheckabledeposits ............... 1.32 1.97 1.67 1.44 1.61 1.67 .93 .80 .97 2.27 Savings(includingMMDAs)............ 2.76 2.68 2.45 2.11 2.43 1.92 1.02 .73 .71 1.15 Largetimedeposits6 ................... 4.62 5.17 4.53 4.36 5.32 4.40 3.26 2.36 2.14 3.06 Othertimedeposits6.................... 4.58 5.45 5.21 4.95 5.53 5.11 3.55 2.86 2.61 3.23 GrossfederalfundspurchasedandRPs....... 4.93 5.02 5.18 4.53 5.47 3.81 2.02 1.39 1.71 3.16 Otherinterest-bearingliabilities .............. 8.86 9.13 8.85 8.61 8.15 7.01 5.39 4.26 3.69 5.25 Incomeandexpenseasapercentageofaveragenetconsolidatedassets Grossinterestincome.......................... 6.26 6.31 6.21 6.01 6.39 5.55 4.78 4.06 3.95 4.47 Taxableequivalent........................ 6.27 6.33 6.22 6.03 6.41 5.57 4.80 4.08 3.97 4.49 Loans ...................................... 4.48 4.31 4.27 4.35 4.74 4.13 3.58 3.05 2.86 3.19 Securities................................... .71 .73 .81 .85 .88 .72 .73 .63 .69 .72 GrossfederalfundssoldandreverseRPs..... .18 .45 .42 .30 .25 .25 .12 .10 .10 .18 Other....................................... .88 .82 .70 .51 .51 .44 .35 .28 .30 .38 Grossinterestexpense ......................... 3.52 3.55 3.48 3.16 3.60 2.69 1.65 1.20 1.22 1.89 Deposits.................................... 2.26 2.26 2.20 1.97 2.33 1.74 1.06 .75 .74 1.16 GrossfederalfundspurchasedandRPs....... .31 .54 .54 .40 .49 .35 .18 .13 .14 .27 Other....................................... .95 .75 .74 .79 .78 .59 .41 .33 .33 .45 Netinterestincome............................ 2.73 2.76 2.73 2.84 2.78 2.87 3.13 2.86 2.73 2.58 Taxableequivalent........................ 2.75 2.79 2.75 2.86 2.80 2.89 3.15 2.88 2.75 2.60 Lossprovisions7 .............. .............. .11 .16 .31 .26 .38 .59 .73 .35 .16 .20 Non-interestincome ........................... 2.34 2.12 2.15 2.55 2.54 2.26 2.32 2.31 2.21 2.37 Servicechargesondeposits.................. .28 .32 .33 .37 .40 .44 .48 .46 .45 .42 Fiduciaryactivities .......................... .31 .34 .32 .31 .27 .29 .25 .26 .24 .27 Tradingrevenue ............................ .52 .43 .33 .46 .48 .43 .32 .30 .22 .31 Interestrateexposures .................... .30 .23 .10 .17 .20 .20 .15 .12 .06 .11 Foreignexchangerateexposures .......... .17 .20 .20 .19 .18 .14 .14 .14 .12 .12 Othercommodityandequityexposures .... .05 * .03 .09 .11 .08 .03 .04 .04 .07 Other....................................... 1.23 1.04 1.17 1.41 1.39 1.10 1.26 1.29 1.30 1.38 Non-interestexpense .......................... 3.57 3.24 3.47 3.45 3.31 3.13 3.16 3.02 3.11 2.99 Salaries,wages,andemployeebenefits ....... 1.57 1.45 1.45 1.57 1.46 1.38 1.41 1.39 1.34 1.38 Occupancy ................................. .50 .47 .47 .50 .47 .45 .46 .45 .43 .43 Other....................................... 1.50 1.33 1.54 1.38 1.39 1.30 1.28 1.18 1.33 1.19 Netnon-interestexpense....................... 1.23 1.12 1.32 .90 .77 .87 .84 .71 .90 .62 Gainsoninvestmentaccountsecurities ......... .04 .08 .11 .03 −.03 .08 .13 .11 .08 * Incomebeforetaxesandextraordinaryitems .... 1.44 1.56 1.22 1.71 1.60 1.48 1.69 1.91 1.74 1.75 Taxes ...................................... .52 .58 .44 .66 .60 .49 .57 .62 .56 .57 Extraordinaryitems,netofincometaxes ..... * * * * * −.01 * * * * Netincome ................................... .92 .98 .78 1.05 1.00 .99 1.12 1.29 1.18 1.18 Cashdividendsdeclared ..................... .70 .82 .53 .79 .86 .66 1.05 .99 .65 .59 Retainedincome ............................ .21 .15 .25 .26 .13 .32 .07 .30 .53 .59 Memo:Returnonequity....................... 13.21 13.22 10.53 13.58 13.04 12.55 13.24 16.01 14.11 12.90 1. Includesallocatedtransferriskreserve. 5. Whenpossible,basedontheaverageofquarterlybalancesheetdatare- 2. Measuredasthesumoflargetimedepositsindomesticoffices,deposits portedonscheduleRC-KofthequarterlyCallReport. bookedinforeignoffices,subordinatednotesanddebentures,federalfundspur- 6. Before1997,largetimedepositopenaccountswereincludedinothertime chasedandsecuritiessoldunderrepurchaseagreements,FederalHomeLoan deposits. Bankadvances,andotherborrowedmoney. 7. Includesprovisionsforallocatedtransferrisk. 3. Measuredasthesumofconstructionandlanddevelopmentloanssecured *Inabsolutevalue,lessthan0.005percent. byrealestate;realestateloanssecuredbynonfarmnonresidentialpropertiesor n.a. Notavailable. bymultifamilyresidentialproperties;andloanstofinancecommercialreales- MMDA Moneymarketdepositaccount. tate,construction,andlanddevelopmentactivitiesnotsecuredbyrealestate. RP Repurchaseagreement. 4. Otherrealestateownedisacomponentofothernon-interest-earningassets. MBS Mortgage-backedsecurities.

A102 Federal Reserve Bulletin 2006 A.1. Portfoliocomposition,interestrates,andincomeandexpense,U.S.banks,1996–2005 C. Banksranked11through100byassets Item 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 Balancesheetitemsasapercentageofaveragenetconsolidatedassets Interest-earningassets ............................ 88.26 87.50 87.87 88.41 88.67 88.09 88.34 88.10 88.18 87.88 Loansandleases,net........................... 64.24 63.89 64.38 64.23 64.88 62.14 60.00 59.48 60.63 63.37 Commercialandindustrial ................... 18.95 19.01 18.92 19.40 18.19 15.84 13.27 11.96 11.90 12.18 U.S.addressees............................ 17.71 17.78 17.59 18.18 17.64 15.36 12.94 11.66 11.64 11.91 Foreignaddressees ........................ 1.24 1.22 1.33 1.22 .55 .48 .33 .30 .26 .27 Consumer ................................... 15.67 15.62 14.52 13.57 13.79 13.20 12.79 12.57 12.73 12.84 Creditcard................................ 8.26 8.50 7.67 6.78 6.97 6.97 6.56 6.35 6.90 7.44 Installmentandother ...................... 7.40 7.12 6.86 6.79 6.82 6.23 6.22 6.21 5.83 5.39 Realestate .................................. 23.26 22.99 24.59 24.80 26.21 27.29 28.94 30.67 32.16 34.89 Indomesticoffices ........................ 23.10 22.85 24.42 24.62 26.12 27.21 28.88 30.54 31.97 34.73 Constructionandlanddevelopment ...... 1.55 1.69 2.03 2.43 3.00 3.31 3.36 3.22 3.51 4.21 Farmland ............................... .13 .14 .17 .19 .22 .23 .22 .20 .19 .19 One-tofour-familyresidential........... 14.15 13.88 14.86 14.15 14.51 15.51 17.05 18.79 19.52 21.06 Homeequity ......................... 2.08 2.22 2.17 2.08 2.49 2.90 3.92 4.74 5.90 6.04 Other ................................ 12.07 11.65 12.69 12.07 12.02 12.60 13.13 14.05 13.62 15.02 Multifamilyresidential .................. .89 .93 1.00 1.02 1.11 1.16 1.20 1.32 1.34 1.45 Nonfarmnonresidential.................. 6.37 6.21 6.36 6.82 7.28 6.99 7.05 7.00 7.41 7.82 Inforeignoffices .......................... .16 .15 .18 .19 .09 .09 .06 .13 .20 .16 Todepositoryinstitutionsandacceptances ofotherbanks .......................... 1.53 1.30 1.09 .93 1.05 1.40 1.44 1.21 .54 .56 Foreigngovernments ........................ .20 .09 .06 .06 .03 .03 .02 .02 .01 .02 Agriculturalproduction ...................... .28 .29 .33 .33 .37 .32 .27 .23 .19 .19 Otherloans.................................. 3.27 3.18 3.35 2.99 2.57 2.03 1.80 1.59 1.87 1.62 Lease-financingreceivables .................. 2.41 2.70 2.72 3.29 3.82 3.18 2.65 2.35 2.30 2.07 Less:Unearnedincomeonloans ............. −.06 −.05 −.04 −.04 −.03 −.02 −.02 −.02 −.02 −.01 Less:Lossreserves1 ......................... −1.27 −1.24 −1.16 −1.11 −1.12 −1.13 −1.17 −1.10 −1.06 −.97 Securities...................................... 16.87 15.80 16.66 17.79 17.32 19.00 20.30 21.16 21.28 19.98 Investmentaccount .......................... 16.06 15.07 16.13 17.28 16.10 17.71 19.17 20.09 20.12 18.82 Debt...................................... 15.62 14.58 15.58 16.64 15.50 17.32 18.82 19.88 19.96 18.69 U.S.Treasury ........................... 3.34 2.81 2.25 1.70 1.12 .67 .74 .95 .89 .60 U.S.governmentagencyand corporationobligations ............. 9.12 8.98 9.93 10.57 9.70 10.09 11.45 12.99 12.80 11.62 Government-backedmortgagepools ... 5.42 5.17 4.98 5.12 4.31 5.19 6.00 6.08 5.74 4.83 Collateralizedmortgageobligations .... 2.16 2.13 2.83 2.89 2.55 2.42 2.79 3.72 3.42 3.39 Other ................................ 1.54 1.68 2.12 2.56 2.84 2.48 2.65 3.19 3.64 3.40 Stateandlocalgovernment .............. .99 .88 .92 .99 .96 .99 .97 .95 .96 .98 Privatemortgage-backedsecurities ....... .96 .73 .96 1.35 1.66 2.01 2.13 2.14 2.65 3.59 Other................................... 1.21 1.18 1.53 2.02 2.06 3.56 3.53 2.85 2.66 1.90 Equity .................................... .44 .49 .55 .65 .60 .39 .34 .21 .16 .13 Tradingaccount ............................. .80 .73 .54 .51 1.22 1.29 1.13 1.07 1.16 1.16 GrossfederalfundssoldandreverseRPs........ 4.26 4.38 3.57 3.34 3.76 4.06 4.71 4.20 2.98 2.30 Interest-bearingbalancesatdepositories ......... 2.89 3.43 3.24 3.06 2.71 2.88 3.33 3.26 3.29 2.24 Non-interest-earningassets........................ 11.74 12.50 12.13 11.59 11.33 11.91 11.66 11.90 11.82 12.12 Revaluationgainsheldintradingaccounts ...... .51 .69 .75 .56 .40 .55 .47 .60 .42 .33 Other.......................................... 11.23 11.81 11.38 11.03 10.92 11.37 11.19 11.30 11.40 11.79 Liabilities ........................................ 92.02 91.85 91.63 91.66 91.57 91.15 90.79 90.65 89.87 88.86 Coredeposits .................................. 52.96 51.51 49.89 48.35 46.28 46.28 47.07 47.93 46.55 48.24 Transactiondeposits ......................... 17.53 16.12 14.15 12.12 9.93 8.37 7.49 7.29 7.06 6.64 Demanddeposits .......................... 14.47 14.17 12.39 10.52 8.61 7.17 6.32 5.96 5.65 5.35 Othercheckabledeposits .................. 3.06 1.95 1.75 1.60 1.32 1.20 1.17 1.33 1.41 1.29 Savingsdeposits(includingMMDAs)......... 21.17 21.71 22.51 23.90 24.02 26.62 30.07 32.34 31.75 33.40 Smalltimedeposits .......................... 14.26 13.69 13.24 12.32 12.33 11.28 9.51 8.30 7.74 8.21 Managedliabilities2............................ 35.60 36.60 38.11 39.83 41.98 40.81 39.48 38.12 39.29 36.98 Largetimedeposits .......................... 6.54 7.37 7.83 8.17 9.54 9.72 8.99 8.20 8.76 10.04 Depositsbookedinforeignoffices ............ 7.73 8.08 8.37 8.19 7.56 7.05 6.28 6.54 7.21 6.02 Subordinatednotesanddebentures............ 1.41 1.48 1.66 1.71 1.54 1.53 1.44 1.38 1.39 1.31 GrossfederalfundspurchasedandRPs ....... 10.00 9.36 9.48 9.77 9.28 9.71 9.66 9.69 8.95 7.17 Othermanagedliabilities..................... 9.92 10.31 10.77 11.99 14.07 12.79 13.11 12.30 12.97 12.44 Revaluationlossesheldintradingaccounts...... .49 .68 .76 .58 .41 .52 .44 .56 .40 .33 Other.......................................... 2.97 3.05 2.87 2.90 2.91 3.54 3.80 4.05 3.64 3.30 Capitalaccount................................... 7.98 8.15 8.37 8.34 8.43 8.85 9.21 9.35 10.13 11.14 Memo Commercialrealestateloans3..................... 9.38 9.44 10.11 11.00 12.06 12.06 12.24 12.10 12.85 13.92 Otherrealestateowned4.......................... .08 .06 .04 .03 .03 .04 .05 .06 .05 .04 Mortgage-backedsecurities ....................... 8.54 8.03 8.76 9.36 8.52 9.63 10.93 11.93 11.81 11.81 FederalHomeLoanBankadvances ............... n.a. n.a. n.a. n.a. n.a. 4.07 4.85 4.75 4.65 5.19 Averagenetconsolidatedassets (billionsofdollars) .......................... 1,450 1,604 1,745 1,881 2,031 2,130 2,124 2,287 2,376 2,403

Profits and Balance Sheet Developments at U.S. Commercial Banks in 2005 A103 A.1.—Continued C. Banksranked11through100byassets—Continued Item 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 Effectiveinterestrate(percent)5 Ratesearned Interest-earningassets ......................... 8.18 8.33 8.13 7.84 8.44 7.54 6.03 5.30 5.26 6.05 Taxableequivalent........................ 8.23 8.36 8.17 7.88 8.48 7.58 6.07 5.33 5.29 6.08 Loansandleases,gross...................... 8.88 9.03 8.82 8.50 9.14 8.26 6.80 6.11 5.98 6.63 Netoflossprovisions .................. 8.21 8.27 8.15 7.80 8.25 6.96 5.59 5.11 5.19 5.90 Securities................................... 6.49 6.55 6.31 6.32 6.64 5.96 4.79 3.80 3.63 4.18 Taxableequivalent ..................... 6.66 6.70 6.46 6.46 6.77 6.08 4.91 3.90 3.73 4.29 Investmentaccount ....................... 6.49 6.57 6.33 6.34 6.66 6.04 4.86 3.87 3.64 4.11 U.S.TreasurysecuritiesandU.S. governmentagencyobligations (excludingMBS) .................. n.a. n.a. n.a. n.a. n.a. 5.83 4.28 3.17 2.94 3.47 Mortgage-backedsecurities ............. n.a. n.a. n.a. n.a. n.a. 6.60 5.34 4.20 4.02 4.34 Other .................................. n.a. n.a. n.a. n.a. n.a. 5.13 4.22 3.61 3.29 4.04 Tradingaccount .......................... 6.53 6.05 5.86 5.58 6.25 4.83 3.59 2.56 3.39 5.30 GrossfederalfundssoldandreverseRPs..... 5.31 5.45 5.46 5.12 6.06 3.86 1.68 1.14 1.25 3.24 Interest-bearingbalancesatdepositories ...... 5.82 5.76 5.67 4.81 5.49 4.38 2.46 1.93 2.27 3.20 Ratespaid Interest-bearingliabilities ...................... 4.70 4.79 4.77 4.38 5.22 4.16 2.41 1.80 1.71 2.68 Interest-bearingdeposits ..................... 4.15 4.22 4.15 3.76 4.42 3.60 1.96 1.35 1.29 2.03 Inforeignoffices ......................... 5.29 5.23 5.22 4.70 5.38 3.67 1.70 1.23 1.42 2.76 Indomesticoffices........................ 3.96 4.04 3.96 3.60 4.26 3.60 1.99 1.36 1.27 1.95 Othercheckabledeposits ............... 1.78 2.01 2.41 2.03 2.57 2.32 .94 .64 .72 1.28 Savings(includingMMDAs)............ 2.91 2.84 2.76 2.49 2.94 2.30 1.08 .66 .65 1.30 Largetimedeposits6 ................... 5.50 5.47 5.32 4.96 5.88 5.11 3.37 2.70 2.49 3.30 Othertimedeposits6.................... 5.26 5.43 5.35 5.03 5.73 5.42 3.68 2.95 2.58 3.04 GrossfederalfundspurchasedandRPs....... 5.19 5.29 5.22 4.87 6.02 3.86 1.73 1.20 1.37 3.04 Otherinterest-bearingliabilities .............. 5.95 5.85 5.81 5.41 6.36 5.30 3.54 3.02 2.76 3.87 Incomeandexpenseasapercentageofaveragenetconsolidatedassets Grossinterestincome.......................... 7.24 7.26 7.16 6.98 7.54 6.70 5.31 4.67 4.67 5.34 Taxableequivalent........................ 7.28 7.30 7.19 7.02 7.57 6.73 5.34 4.70 4.70 5.36 Loans ...................................... 5.80 5.87 5.79 5.56 6.05 5.28 4.15 3.72 3.72 4.27 Securities................................... 1.03 .98 1.00 1.10 1.09 1.06 .90 .75 .73 .77 GrossfederalfundssoldandreverseRPs..... .23 .22 .19 .18 .22 .15 .08 .04 .03 .06 Other....................................... .18 .19 .18 .14 .18 .21 .18 .15 .19 .22 Grossinterestexpense ......................... 3.39 3.41 3.45 3.26 3.96 3.14 1.77 1.30 1.26 1.94 Deposits.................................... 2.18 2.23 2.23 2.02 2.41 2.01 1.09 .77 .74 1.18 GrossfederalfundspurchasedandRPs....... .55 .51 .51 .51 .56 .38 .17 .12 .13 .23 Other....................................... .66 .68 .71 .74 .99 .75 .51 .41 .40 .53 Netinterestincome............................ 3.84 3.85 3.71 3.72 3.58 3.56 3.54 3.37 3.41 3.40 Taxableequivalent........................ 3.89 3.89 3.74 3.75 3.61 3.59 3.57 3.40 3.44 3.42 Lossprovisions7 .............................. .54 .60 .54 .55 .68 .91 .80 .67 .55 .52 Non-interestincome ........................... 2.61 2.76 3.07 3.36 3.18 3.35 3.30 3.29 3.05 2.75 Servicechargesondeposits.................. .44 .44 .42 .41 .42 .42 .42 .42 .40 .37 Fiduciaryactivities .......................... .43 .44 .49 .48 .52 .42 .42 .37 .42 .35 Tradingrevenue ............................ .08 .08 .09 .08 .07 .08 .08 .09 .07 .06 Interestrateexposures .................... .03 .02 .03 .02 .02 .04 .04 .04 −.01 −.01 Foreignexchangerateexposures .......... .04 .05 .06 .05 .04 .03 .04 .04 .05 .04 Othercommodityandequityexposures .... .01 * * * * * * .01 .03 .02 Other....................................... 1.67 1.79 2.07 2.39 2.18 2.43 2.37 2.41 2.16 1.98 Non-interestexpense .......................... 3.85 3.85 4.03 4.12 4.00 3.95 3.73 3.64 3.55 3.36 Salaries,wages,andemployeebenefits ....... 1.51 1.51 1.53 1.53 1.44 1.47 1.49 1.47 1.45 1.37 Occupancy ................................. .48 .46 .46 .45 .43 .42 .40 .41 .39 .37 Other....................................... 1.86 1.88 2.04 2.14 2.14 2.07 1.84 1.76 1.70 1.62 Netnon-interestexpense....................... 1.24 1.10 .96 .76 .82 .60 .43 .35 .50 .61 Gainsoninvestmentaccountsecurities ......... .02 .02 .03 −.01 −.05 .09 .10 .06 .03 * Incomebeforetaxesandextraordinaryitems .... 2.09 2.18 2.24 2.40 2.02 2.14 2.41 2.42 2.39 2.27 Taxes ...................................... .75 .77 .78 .86 .70 .74 .82 .82 .82 .77 Extraordinaryitems,netofincometaxes ..... * * * * * * * * * .01 Netincome ................................... 1.34 1.42 1.45 1.54 1.32 1.39 1.59 1.59 1.57 1.50 Cashdividendsdeclared ..................... 1.07 .93 .96 1.16 .94 .96 .99 1.05 .95 1.00 Retainedincome ............................ .26 .48 .50 .38 .38 .43 .60 .54 .62 .50 Memo:Returnonequity....................... 16.78 17.36 17.38 18.46 15.72 15.74 17.24 17.03 15.54 13.48 1. Includesallocatedtransferriskreserve. 5. Whenpossible,basedontheaverageofquarterlybalancesheetdatare- 2. Measuredasthesumoflargetimedepositsindomesticoffices,deposits portedonscheduleRC-KofthequarterlyCallReport. bookedinforeignoffices,subordinatednotesanddebentures,federalfundspur- 6. Before1997,largetimedepositopenaccountswereincludedinothertime chasedandsecuritiessoldunderrepurchaseagreements,FederalHomeLoan deposits. Bankadvances,andotherborrowedmoney. 7. Includesprovisionsforallocatedtransferrisk. 3. Measuredasthesumofconstructionandlanddevelopmentloanssecured *Inabsolutevalue,lessthan0.005percent. byrealestate;realestateloanssecuredbynonfarmnonresidentialpropertiesor n.a. Notavailable. bymultifamilyresidentialproperties;andloanstofinancecommercialreales- MMDA Moneymarketdepositaccount. tate,construction,andlanddevelopmentactivitiesnotsecuredbyrealestate. RP Repurchaseagreement. 4. Otherrealestateownedisacomponentofothernon-interest-earningassets. MBS Mortgage-backedsecurities.

A104 Federal Reserve Bulletin 2006 A.1. Portfoliocomposition,interestrates,andincomeandexpense,U.S.banks,1996–2005 D. Banksranked101through1,000byassets Item 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 Balancesheetitemsasapercentageofaveragenetconsolidatedassets Interest-earningassets ............................ 91.11 91.34 91.38 91.68 91.50 91.16 91.36 91.34 91.56 91.32 Loansandleases,net........................... 62.72 62.34 61.23 61.48 62.15 62.46 61.46 61.32 63.33 65.15 Commercialandindustrial ................... 12.76 12.38 12.45 12.64 12.95 13.03 12.38 11.51 11.52 11.78 U.S.addressees............................ 12.58 12.14 12.12 12.32 12.60 12.65 12.06 11.20 11.21 11.48 Foreignaddressees ........................ .18 .23 .32 .32 .36 .38 .31 .31 .31 .30 Consumer ................................... 16.11 14.36 12.56 10.79 10.19 9.76 8.13 6.79 6.33 5.42 Creditcard................................ 6.92 5.87 4.78 3.37 3.27 3.61 2.63 1.82 1.91 1.24 Installmentandother ...................... 9.19 8.49 7.78 7.41 6.92 6.15 5.50 4.97 4.42 4.18 Realestate .................................. 31.28 33.10 33.83 35.90 36.93 37.64 38.92 40.96 43.38 45.86 Indomesticoffices ........................ 31.26 33.08 33.81 35.87 36.91 37.62 38.89 40.91 43.32 45.78 Constructionandlanddevelopment ...... 2.38 2.68 2.87 3.48 4.15 4.90 5.40 5.89 6.98 8.81 Farmland ............................... .46 .52 .56 .58 .65 .66 .73 .80 .91 .99 One-tofour-familyresidential........... 17.29 18.08 18.14 18.26 17.17 16.18 15.39 15.71 15.36 15.22 Homeequity ......................... 2.30 2.29 2.14 1.99 2.10 2.21 2.51 2.92 3.46 3.61 Other ................................ 14.99 15.78 16.00 16.26 15.06 13.97 12.88 12.79 11.90 11.61 Multifamilyresidential .................. 1.28 1.28 1.25 1.44 1.58 1.69 1.83 2.00 2.24 2.37 Nonfarmnonresidential.................. 9.85 10.52 10.99 12.12 13.36 14.18 15.55 16.51 17.82 18.40 Inforeignoffices .......................... .02 .02 .02 .02 .02 .02 .03 .05 .06 .08 Todepositoryinstitutionsandacceptances ofotherbanks .......................... .50 .59 .52 .46 .37 .38 .37 .37 .25 .13 Foreigngovernments ........................ .02 .02 .03 .03 .03 .03 .02 .02 .01 * Agriculturalproduction ...................... .70 .73 .80 .78 .82 .85 .86 .83 .82 .81 Otherloans.................................. 1.67 1.47 1.30 1.25 1.22 1.22 1.18 1.25 1.32 1.36 Lease-financingreceivables .................. 1.00 .99 .99 .78 .75 .74 .75 .67 .75 .75 Less:Unearnedincomeonloans ............. −.10 −.10 −.09 −.08 −.08 −.07 −.06 −.06 −.06 −.06 Less:Lossreserves1 ......................... −1.23 −1.19 −1.15 −1.06 −1.04 −1.12 −1.10 −1.02 −.98 −.90 Securities...................................... 22.61 23.37 24.18 25.17 24.34 22.81 23.86 24.36 23.59 21.58 Investmentaccount .......................... 22.49 23.26 24.08 25.09 24.25 22.70 23.80 24.23 23.54 21.50 Debt...................................... 21.97 22.65 23.39 24.33 23.46 22.28 23.30 23.79 23.18 21.21 U.S.Treasury ........................... 5.59 4.94 3.91 2.53 1.81 1.32 1.22 1.00 1.02 .83 U.S.governmentagencyand corporationobligations ............. 12.62 13.91 15.08 16.29 15.56 14.70 15.85 16.96 16.70 15.05 Government-backedmortgagepools ... 5.67 6.20 6.45 6.72 6.22 6.27 6.55 7.03 6.80 5.73 Collateralizedmortgageobligations .... 3.11 3.00 3.21 3.52 3.04 3.08 3.69 3.69 3.41 3.16 Other ................................ 3.84 4.71 5.42 6.05 6.30 5.35 5.60 6.24 6.49 6.16 Stateandlocalgovernment .............. 2.23 2.43 2.69 2.91 2.91 2.90 2.89 2.95 2.92 2.78 Privatemortgage-backedsecurities ....... .76 .59 .65 1.00 .99 .94 .99 .87 1.08 1.17 Other................................... .76 .78 1.06 1.60 2.19 2.42 2.34 2.01 1.46 1.37 Equity .................................... .52 .61 .69 .77 .79 .43 .50 .43 .36 .29 Tradingaccount ............................. .12 .10 .11 .08 .09 .11 .06 .14 .05 .08 GrossfederalfundssoldandreverseRPs........ 3.86 3.59 4.16 3.35 3.40 4.20 4.15 3.85 2.95 2.83 Interest-bearingbalancesatdepositories ......... 1.93 2.05 1.80 1.68 1.60 1.68 1.89 1.81 1.69 1.76 Non-interest-earningassets........................ 8.89 8.66 8.62 8.32 8.50 8.84 8.64 8.66 8.44 8.68 Revaluationgainsheldintradingaccounts ...... .02 * * .01 .02 .01 .01 * * * Other.......................................... 8.86 8.66 8.62 8.31 8.49 8.84 8.64 8.66 8.44 8.68 Liabilities ........................................ 91.06 90.78 90.55 90.90 90.95 90.32 89.93 89.69 89.18 89.10 Coredeposits .................................. 64.28 64.06 63.87 62.48 60.80 60.33 61.26 61.31 60.40 59.08 Transactiondeposits ......................... 19.99 18.05 16.08 13.94 12.29 11.48 11.37 11.50 11.77 11.16 Demanddeposits .......................... 13.80 13.11 11.87 10.19 8.97 8.23 8.05 7.96 8.13 7.88 Othercheckabledeposits .................. 6.19 4.94 4.22 3.75 3.32 3.25 3.32 3.54 3.64 3.28 Savingsdeposits(includingMMDAs)......... 22.69 23.97 26.43 28.55 28.55 29.40 32.34 34.00 34.42 33.74 Smalltimedeposits .......................... 21.60 22.05 21.36 19.99 19.96 19.46 17.55 15.81 14.20 14.17 Managedliabilities2............................ 24.96 24.89 24.65 26.33 28.01 27.75 26.57 26.40 26.98 28.34 Largetimedeposits .......................... 8.34 9.68 10.09 10.30 11.98 12.60 12.17 11.92 12.12 13.60 Depositsbookedinforeignoffices ............ 1.34 1.23 1.31 1.20 1.28 1.24 .88 .64 .65 .57 Subordinatednotesanddebentures............ .36 .33 .37 .35 .30 .31 .34 .35 .35 .27 GrossfederalfundspurchasedandRPs ....... 8.17 7.06 6.15 6.90 6.30 5.77 5.27 5.35 5.52 5.54 Othermanagedliabilities..................... 6.74 6.59 6.73 7.57 8.15 7.84 7.90 8.13 8.34 8.35 Revaluationlossesheldintradingaccounts...... .02 .01 .01 .01 * .01 .01 * * * Other.......................................... 1.79 1.82 2.02 2.10 2.13 2.23 2.08 1.98 1.81 1.69 Capitalaccount................................... 8.94 9.22 9.45 9.10 9.05 9.68 10.07 10.31 10.82 10.90 Memo Commercialrealestateloans3..................... 13.80 14.72 15.33 17.28 19.32 21.03 23.05 24.62 27.25 29.79 Otherrealestateowned4.......................... .13 .11 .09 .08 .07 .08 .10 .11 .10 .08 Mortgage-backedsecurities ....................... 9.55 9.79 10.30 11.24 10.25 10.29 11.24 11.59 11.29 10.06 FederalHomeLoanBankadvances ............... n.a. n.a. n.a. n.a. n.a. 5.27 5.71 6.29 6.46 6.42 Averagenetconsolidatedassets (billionsofdollars) .......................... 1,078 971 938 972 986 1,002 1,022 1,072 1,080 1,152

Profits and Balance Sheet Developments at U.S. Commercial Banks in 2005 A105 A.1.—Continued D. Banksranked101through1,000byassets—Continued Item 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 Effectiveinterestrate(percent)5 Ratesearned Interest-earningassets ......................... 8.44 8.54 8.38 7.83 8.48 7.86 6.43 5.60 5.46 6.12 Taxableequivalent........................ 8.52 8.63 8.47 7.92 8.56 7.94 6.51 5.68 5.53 6.19 Loansandleases,gross...................... 9.41 9.53 9.42 8.74 9.42 8.76 7.32 6.57 6.25 6.90 Netoflossprovisions .................. 8.77 8.79 8.79 8.26 8.75 7.88 6.56 6.02 5.87 6.64 Securities................................... 6.34 6.43 6.31 6.03 6.45 5.97 4.95 3.81 3.79 4.03 Taxableequivalent ..................... 6.60 6.69 6.57 6.29 6.71 6.25 5.21 4.06 4.04 4.28 Investmentaccount ....................... 6.34 6.43 6.30 6.03 6.45 5.96 4.93 3.82 3.78 4.02 U.S.TreasurysecuritiesandU.S. governmentagencyobligations (excludingMBS) .................. n.a. n.a. n.a. n.a. n.a. 5.85 4.54 3.42 3.15 3.47 Mortgage-backedsecurities ............. n.a. n.a. n.a. n.a. n.a. 6.33 5.38 3.95 4.01 4.23 Other .................................. n.a. n.a. n.a. n.a. n.a. 5.40 4.51 4.07 4.21 4.42 Tradingaccount .......................... 5.94 6.37 6.84 7.33 9.30 6.60 14.05 3.07 10.30 5.82 GrossfederalfundssoldandreverseRPs..... 5.29 5.42 5.31 4.98 6.15 3.91 1.73 1.27 1.57 3.31 Interest-bearingbalancesatdepositories ...... 5.69 5.44 5.77 5.07 5.76 3.94 1.79 1.26 1.47 3.29 Ratespaid Interest-bearingliabilities ...................... 4.58 4.67 4.60 4.19 4.93 4.11 2.54 1.88 1.73 2.48 Interest-bearingdeposits .................... 4.27 4.34 4.28 3.84 4.46 3.82 2.28 1.61 1.44 2.09 Inforeignoffices ......................... 5.72 5.42 5.55 5.07 6.13 4.45 2.14 1.43 1.43 3.05 Indomesticoffices........................ 4.23 4.32 4.25 3.82 4.43 3.81 2.28 1.61 1.44 2.08 Othercheckabledeposits ............... 1.96 2.17 2.15 1.99 2.27 1.81 1.06 .74 .72 1.18 Savings(includingMMDAs)............ 3.11 3.08 2.96 2.65 3.07 2.22 1.17 .76 .74 1.27 Largetimedeposits6 ................... 5.48 5.56 5.51 5.17 6.00 5.27 3.34 2.58 2.33 3.21 Othertimedeposits6.................... 5.57 5.57 5.64 5.11 5.74 5.51 3.77 2.86 2.51 3.10 GrossfederalfundspurchasedandRPs....... 5.16 5.20 5.14 4.82 5.95 3.83 1.83 1.29 1.45 2.94 Otherinterest-bearingliabilities .............. 5.90 6.08 5.99 5.36 6.45 5.41 4.17 3.60 3.37 4.00 Incomeandexpenseasapercentageofaveragenetconsolidatedassets Grossinterestincome.......................... 7.70 7.79 7.66 7.19 7.79 7.16 5.85 5.08 4.99 5.57 Taxableequivalent........................ 7.78 7.87 7.74 7.27 7.86 7.24 5.93 5.16 5.06 5.64 Loans ...................................... 6.01 6.05 5.89 5.47 5.96 5.59 4.57 4.08 4.01 4.55 Securities................................... 1.42 1.49 1.50 1.51 1.58 1.33 1.15 .91 .88 .86 GrossfederalfundssoldandreverseRPs..... .20 .19 .22 .17 .21 .16 .07 .05 .05 .09 Other....................................... .06 .06 .06 .04 .04 .08 .06 .05 .05 .07 Grossinterestexpense ......................... 3.41 3.47 3.45 3.20 3.79 3.14 1.92 1.41 1.29 1.84 Deposits.................................... 2.57 2.69 2.70 2.44 2.87 2.48 1.49 1.04 .92 1.34 GrossfederalfundspurchasedandRPs....... .43 .37 .32 .34 .38 .22 .09 .07 .08 .16 Other....................................... .42 .42 .42 .42 .54 .44 .34 .30 .29 .34 Netinterestincome............................ 4.29 4.32 4.22 3.99 4.00 4.02 3.93 3.68 3.70 3.73 Taxableequivalent........................ 4.37 4.39 4.29 4.07 4.07 4.10 4.00 3.75 3.77 3.79 Lossprovisions7 .............................. .52 .58 .49 .39 .52 .65 .55 .40 .30 .24 Non-interestincome ........................... 1.88 2.07 2.26 2.31 2.35 2.37 2.37 2.31 2.26 2.02 Servicechargesondeposits.................. .41 .40 .39 .38 .36 .39 .41 .41 .39 .36 Fiduciaryactivities .......................... .29 .32 .37 .38 .44 .40 .35 .34 .37 .35 Tradingrevenue ............................ .02 .01 .02 .02 .01 * * .01 .01 .01 Interestrateexposures .................... .01 .01 .01 .01 .01 −.01 * .01 .01 .01 Foreignexchangerateexposures .......... .01 * * * * * * * * * Othercommodityandequityexposures .... * * * * * * * * * * Other....................................... 1.16 1.34 1.49 1.53 1.55 1.58 1.61 1.55 1.49 1.30 Non-interestexpense .......................... 3.69 3.73 3.86 3.70 3.84 3.88 3.73 3.60 3.54 3.37 Salaries,wages,andemployeebenefits ....... 1.44 1.50 1.56 1.56 1.59 1.61 1.64 1.64 1.64 1.61 Occupancy ................................. .45 .46 .47 .47 .47 .46 .45 .43 .43 .41 Other....................................... 1.80 1.77 1.83 1.68 1.78 1.81 1.64 1.53 1.47 1.35 Netnon-interestexpense....................... 1.81 1.66 1.60 1.39 1.48 1.52 1.36 1.29 1.28 1.35 Gainsoninvestmentaccountsecurities ......... .02 .02 .04 −.01 −.04 .05 .04 .05 .02 −.01 Incomebeforetaxesandextraordinaryitems .... 1.98 2.10 2.16 2.20 1.96 1.90 2.06 2.03 2.14 2.13 Taxes ...................................... .69 .73 .74 .74 .67 .66 .67 .66 .68 .68 Extraordinaryitems,netofincometaxes ..... * * .06 .01 * .01 * .03 * * Netincome ................................... 1.29 1.37 1.47 1.47 1.29 1.25 1.38 1.40 1.45 1.45 Cashdividendsdeclared ..................... 1.04 1.10 1.01 1.06 .92 1.33 1.19 1.64 .78 .87 Retainedincome ............................ .25 .28 .46 .40 .37 −.08 .19 −.25 .67 .58 Memo:Returnonequity........... ........... 14.42 14.89 15.60 16.11 14.21 12.93 13.75 13.54 13.43 13.34 1. Includesallocatedtransferriskreserve. 5. Whenpossible,basedontheaverageofquarterlybalancesheetdatare- 2. Measuredasthesumoflargetimedepositsindomesticoffices,deposits portedonscheduleRC-KofthequarterlyCallReport. bookedinforeignoffices,subordinatednotesanddebentures,federalfundspur- 6. Before1997,largetimedepositopenaccountswereincludedinothertime chasedandsecuritiessoldunderrepurchaseagreements,FederalHomeLoan deposits. Bankadvances,andotherborrowedmoney. 7. Includesprovisionsforallocatedtransferrisk. 3. Measuredasthesumofconstructionandlanddevelopmentloanssecured *Inabsolutevalue,lessthan0.005percent. byrealestate;realestateloanssecuredbynonfarmnonresidentialpropertiesor n.a. Notavailable. bymultifamilyresidentialproperties;andloanstofinancecommercialreales- MMDA Moneymarketdepositaccount. tate,construction,andlanddevelopmentactivitiesnotsecuredbyrealestate. RP Repurchaseagreement. 4. Otherrealestateownedisacomponentofothernon-interest-earningassets. MBS Mortgage-backedsecurities.

A106 Federal Reserve Bulletin 2006 A.1. Portfoliocomposition,interestrates,andincomeandexpense,U.S.banks,1996–2005 E. Banksnotrankedamongthe1,000largestbyassets Item 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 Balancesheetitemsasapercentageofaveragenetconsolidatedassets Interest-earningassets ............................ 92.45 92.45 92.64 92.55 92.52 92.26 92.22 92.14 92.34 92.31 Loansandleases,net........................... 57.38 58.76 59.11 59.76 62.31 62.67 62.72 62.32 63.81 65.45 Commercialandindustrial ................... 9.98 10.16 10.33 10.64 11.09 11.10 10.71 10.42 10.29 10.21 U.S.addressees............................ 9.91 10.08 10.25 10.55 11.02 11.02 10.64 10.37 10.25 10.15 Foreignaddressees ........................ .07 .08 .08 .08 .07 .08 .06 .05 .04 .06 Consumer ................................... 9.42 8.98 8.46 8.16 7.98 7.42 6.77 6.16 5.45 4.97 Creditcard................................ 1.04 .85 .70 .69 .59 .57 .49 .51 .40 .36 Installmentandother ...................... 8.39 8.14 7.76 7.47 7.39 6.85 6.28 5.64 5.05 4.61 Realestate .................................. 34.10 35.55 36.04 36.84 39.29 40.30 41.52 42.31 44.76 46.99 Indomesticoffices ........................ 34.10 35.55 36.04 36.83 39.29 40.30 41.52 42.31 44.76 46.99 Constructionandlanddevelopment ...... 2.61 2.82 3.02 3.28 3.70 4.23 4.51 4.99 6.01 7.46 Farmland ............................... 2.55 2.69 2.83 2.95 3.06 3.04 3.08 3.12 3.22 3.25 One-tofour-familyresidential........... 17.47 18.16 18.04 17.66 18.43 18.24 17.91 17.09 17.20 17.15 Homeequity ......................... 1.20 1.24 1.21 1.17 1.28 1.37 1.62 1.80 2.12 2.20 Other ................................ 16.28 16.92 16.83 16.49 17.15 16.87 16.29 15.30 15.08 14.95 Multifamilyresidential .................. .92 .95 .93 .98 1.04 1.06 1.16 1.28 1.41 1.48 Nonfarmnonresidential.................. 10.54 10.93 11.22 11.96 13.06 13.71 14.86 15.82 16.93 17.65 Inforeignoffices .......................... * * * * * * * * * * Todepositoryinstitutionsandacceptances ofotherbanks .......................... .21 .20 .14 .14 .12 .12 .10 .09 .07 .05 Foreigngovernments ........................ * * * .01 .01 * * * * * Agriculturalproduction ...................... 3.92 4.05 4.27 4.06 3.85 3.76 3.64 3.39 3.25 3.21 Otherloans.................................. .69 .67 .67 .67 .69 .67 .65 .66 .68 .70 Lease-financingreceivables .................. .23 .25 .24 .26 .27 .27 .31 .26 .25 .24 Less:Unearnedincomeonloans ............. −.27 −.24 −.20 −.15 −.11 −.09 −.07 −.06 −.06 −.05 Less:Lossreserves1 ......................... −.90 −.87 −.86 −.87 −.88 −.88 −.90 −.92 −.89 −.87 Securities...................................... 29.53 28.24 26.70 26.91 25.40 22.80 23.34 23.46 23.33 21.91 Investmentaccount .......................... 29.50 28.21 26.66 26.88 25.38 22.79 23.33 23.43 23.33 21.90 Debt...................................... 29.01 27.69 26.12 26.34 24.82 22.49 23.05 23.12 23.06 21.70 U.S.Treasury ........................... 7.85 6.70 5.05 3.34 2.12 1.33 1.04 .90 .81 .71 U.S.governmentagencyand corporationobligations ............. 15.67 15.58 15.43 16.89 16.95 15.27 16.07 16.22 16.57 15.63 Government-backedmortgagepools ... 4.21 4.01 3.90 3.95 3.47 3.78 4.54 4.84 4.75 4.23 Collateralizedmortgageobligations .... 2.46 2.19 2.02 2.00 1.70 1.94 2.30 2.20 1.96 1.70 Other ................................ 9.00 9.38 9.51 10.93 11.78 9.56 9.23 9.19 9.85 9.70 Stateandlocalgovernment .............. 4.62 4.60 4.80 4.96 4.64 4.51 4.56 4.73 4.67 4.49 Privatemortgage-backedsecurities ....... .18 .20 .16 .26 .23 .27 .26 .21 .19 .22 Other................................... .68 .61 .68 .89 .88 1.11 1.12 1.05 .83 .65 Equity .................................... .49 .52 .54 .53 .56 .30 .27 .31 .26 .20 Tradingaccount ............................. .03 .03 .04 .03 .02 .01 .01 .04 .01 .02 GrossfederalfundssoldandreverseRPs........ 4.04 3.95 5.12 4.17 3.22 5.01 4.26 4.26 3.33 3.24 Interest-bearingbalancesatdepositories ......... 1.51 1.49 1.72 1.71 1.59 1.78 1.90 2.08 1.86 1.70 Non-interest-earningassets........................ 7.55 7.55 7.36 7.45 7.48 7.74 7.78 7.86 7.66 7.69 Revaluationgainsheldintradingaccounts ...... * * * * * * * * * * Other.......................................... 7.55 7.55 7.36 7.45 7.48 7.74 7.78 7.86 7.66 7.69 Liabilities ........................................ 89.82 89.63 89.54 89.75 89.88 89.59 89.73 89.58 89.55 89.48 Coredeposits .................................. 75.76 74.58 73.75 72.74 70.87 69.92 70.04 69.96 69.24 67.68 Transactiondeposits ......................... 24.88 24.48 24.26 23.87 23.20 22.35 22.66 23.18 23.36 22.71 Demanddeposits .......................... 13.13 13.09 13.08 12.80 12.64 12.16 12.24 12.58 12.77 12.76 Othercheckabledeposits .................. 11.75 11.39 11.18 11.07 10.57 10.19 10.42 10.60 10.59 9.95 Savingsdeposits(includingMMDAs)......... 19.60 19.00 19.05 19.77 19.19 19.38 21.32 22.43 23.24 22.98 Smalltimedeposits .......................... 31.29 31.10 30.43 29.11 28.48 28.19 26.05 24.36 22.64 21.99 Managedliabilities2............................ 12.99 14.02 14.76 16.09 18.08 18.67 18.79 18.78 19.57 21.04 Largetimedeposits .......................... 9.77 10.51 11.11 11.52 12.51 13.55 13.21 13.07 13.15 14.53 Depositsbookedinforeignoffices ............ .11 .10 .07 .08 .05 .06 .07 .06 .07 .06 Subordinatednotesanddebentures............ .02 .01 .01 .01 .02 .02 .04 .03 .04 .03 GrossfederalfundspurchasedandRPs ....... 1.71 1.67 1.49 1.79 2.06 1.55 1.51 1.52 1.76 1.74 Othermanagedliabilities..................... 1.39 1.73 2.08 2.69 3.44 3.49 3.96 4.09 4.55 4.69 Revaluationlossesheldintradingaccounts...... * * * * * * * * * * Other.......................................... 1.06 1.02 1.03 .92 .93 1.00 .90 .84 .74 .77 Capitalaccount................................... 10.18 10.37 10.46 10.25 10.12 10.41 10.27 10.42 10.45 10.52 Memo Commercialrealestateloans3..................... 14.18 14.80 15.27 16.33 17.91 19.15 20.67 22.23 24.50 26.76 Otherrealestateowned4.......................... .20 .16 .13 .11 .11 .12 .14 .15 .14 .13 Mortgage-backedsecurities ....................... 6.85 6.39 6.07 6.22 5.39 5.99 7.10 7.24 6.90 6.15 FederalHomeLoanBankadvances ............... n.a. n.a. n.a. n.a. n.a. 3.34 3.71 3.87 4.33 4.47 Averagenetconsolidatedassets (billionsofdollars) .......................... 661 647 644 651 655 675 704 742 769 805

Profits and Balance Sheet Developments at U.S. Commercial Banks in 2005 A107 A.1.—Continued E. Banksnotrankedamongthe1,000largestbyassets—Continued Item 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 Effectiveinterestrate(percent)5 Ratesearned Interest-earningassets ......................... 8.37 8.50 8.35 8.05 8.44 7.94 6.79 5.94 5.73 6.23 Taxableequivalent........................ 8.50 8.63 8.48 8.18 8.56 8.05 6.91 6.05 5.84 6.33 Loansandleases,gross...................... 9.75 9.80 9.69 9.28 9.51 9.03 7.83 7.08 6.72 7.18 Netoflossprovisions .................. 9.47 9.49 9.34 8.89 9.14 8.59 7.39 6.72 6.45 6.95 Securities................................... 6.14 6.26 6.04 5.88 6.15 5.86 5.03 3.87 3.74 3.87 Taxableequivalent ..................... 6.52 6.65 6.46 6.29 6.54 6.28 5.43 4.26 4.11 4.24 Investmentaccount ....................... 6.14 6.26 6.04 5.89 6.15 5.86 5.02 3.87 3.73 3.87 U.S.TreasurysecuritiesandU.S. governmentagencyobligations (excludingMBS) .................. n.a. n.a. n.a. n.a. n.a. 5.97 4.80 3.74 3.39 3.53 Mortgage-backedsecurities ............. n.a. n.a. n.a. n.a. n.a. 6.20 5.47 3.58 3.90 4.18 Other .................................. n.a. n.a. n.a. n.a. n.a. 5.29 4.87 4.43 4.18 4.16 Tradingaccount .......................... 6.47 6.33 5.26 3.60 4.01 6.43 15.38 2.89 18.95 7.52 GrossfederalfundssoldandreverseRPs..... 5.34 5.51 5.36 4.96 6.25 3.83 1.63 1.08 1.32 3.21 Interest-bearingbalancesatdepositories ...... 5.63 5.62 5.67 5.69 6.38 4.56 2.68 1.96 2.03 3.21 Ratespaid Interest-bearingliabilities ...................... 4.49 4.61 4.60 4.28 4.80 4.40 2.92 2.13 1.87 2.43 Interest-bearingdeposits.......... .......... 4.44 4.54 4.53 4.22 4.67 4.32 2.78 2.02 1.75 2.29 Inforeignoffices ......................... 5.34 4.77 5.08 4.34 5.13 3.97 1.67 .85 1.04 2.86 Indomesticoffices........................ 4.44 4.53 4.53 4.22 4.67 4.32 2.79 2.02 1.75 2.29 Othercheckabledeposits ............... 2.41 2.46 2.44 2.28 2.47 1.97 1.16 .78 .69 .99 Savings(includingMMDAs)............ 3.26 3.36 3.39 3.21 3.56 2.81 1.72 1.13 1.04 1.53 Largetimedeposits6 ................... 5.48 5.53 5.53 5.21 5.89 5.53 3.62 2.78 2.47 3.22 Othertimedeposits6.................... 5.61 5.66 5.63 5.25 5.70 5.60 3.88 2.96 2.55 3.04 GrossfederalfundspurchasedandRPs....... 5.11 5.22 4.99 4.73 5.69 3.92 1.85 1.31 1.45 2.89 Otherinterest-bearingliabilities .............. 5.77 6.32 6.45 5.64 6.24 5.74 5.32 4.06 3.67 4.02 Incomeandexpenseasapercentageofaveragenetconsolidatedassets Grossinterestincome.......................... 7.77 7.90 7.75 7.48 7.83 7.35 6.31 5.46 5.32 5.78 Taxableequivalent........................ 7.89 8.02 7.87 7.60 7.95 7.45 6.41 5.56 5.42 5.88 Loans ...................................... 5.68 5.86 5.80 5.62 5.99 5.75 5.01 4.47 4.35 4.76 Securities................................... 1.80 1.76 1.59 1.58 1.57 1.32 1.16 .89 .87 .85 GrossfederalfundssoldandreverseRPs..... .24 .24 .29 .22 .21 .20 .07 .05 .05 .11 Other....................................... .04 .04 .06 .06 .05 .08 .06 .05 .05 .06 Grossinterestexpense ......................... 3.39 3.48 3.46 3.26 3.64 3.34 2.23 1.60 1.41 1.82 Deposits.................................... 3.22 3.28 3.25 3.02 3.30 3.08 1.98 1.41 1.22 1.58 GrossfederalfundspurchasedandRPs....... .08 .08 .07 .08 .12 .06 .03 .02 .02 .05 Other....................................... .08 .11 .13 .15 .21 .20 .21 .17 .17 .19 Netinterestincome............................ 4.38 4.42 4.28 4.22 4.20 4.01 4.08 3.86 3.91 3.96 Taxableequivalent........................ 4.50 4.54 4.41 4.35 4.31 4.12 4.19 3.96 4.00 4.06 Lossprovisions7 .............................. .25 .27 .29 .31 .32 .36 .35 .29 .23 .21 Non-interestincome ........................... 1.42 1.41 1.52 1.44 1.32 1.31 1.39 1.47 1.38 1.34 Servicechargesondeposits.................. .44 .44 .42 .42 .43 .44 .45 .43 .43 .40 Fiduciaryactivities .......................... .19 .20 .23 .26 .21 .25 .27 .28 .32 .33 Tradingrevenue ............................ * * * * .01 * * * * * Interestrateexposures .................... * * * * * * * * * * Foreignexchangerateexposures .......... * * * * * * * * * * Othercommodityandequityexposures .... * * * * * * * * * * Other....................................... .79 .77 .86 .75 .68 .62 .67 .76 .64 .61 Non-interestexpense .......................... 3.70 3.69 3.74 3.73 3.58 3.55 3.57 3.56 3.52 3.49 Salaries,wages,andemployeebenefits ....... 1.77 1.80 1.82 1.82 1.78 1.79 1.82 1.82 1.81 1.80 Occupancy ................................. .49 .49 .49 .49 .47 .47 .46 .45 .45 .44 Other....................................... 1.44 1.40 1.43 1.42 1.32 1.29 1.28 1.28 1.26 1.25 Netnon-interestexpense....................... 2.28 2.28 2.23 2.29 2.26 2.24 2.18 2.09 2.14 2.15 Gainsoninvestmentaccountsecurities ......... .01 .01 .02 * −.01 .04 .05 .04 .01 * Incomebeforetaxesandextraordinaryitems .... 1.85 1.89 1.79 1.62 1.61 1.45 1.60 1.53 1.55 1.60 Taxes ...................................... .59 .59 .53 .47 .45 .39 .41 .38 .37 .38 Extraordinaryitems,netofincometaxes ..... * * * * * * −.01 * * * Netincome ................................... 1.26 1.30 1.26 1.15 1.17 1.06 1.18 1.14 1.17 1.22 Cashdividendsdeclared ..................... .64 .74 .82 .70 .79 .64 .68 .67 .64 .67 Retainedincome ............................ .62 .56 .44 .45 .38 .42 .49 .47 .54 .55 Memo:Returnonequity....................... 12.37 12.53 12.02 11.26 11.52 10.17 11.46 10.96 11.24 11.58 1. Includesallocatedtransferriskreserve. 5. Whenpossible,basedontheaverageofquarterlybalancesheetdatare- 2. Measuredasthesumoflargetimedepositsindomesticoffices,deposits portedonscheduleRC-KofthequarterlyCallReport. bookedinforeignoffices,subordinatednotesanddebentures,federalfundspur- 6. Before1997,largetimedepositopenaccountswereincludedinothertime chasedandsecuritiessoldunderrepurchaseagreements,FederalHomeLoan deposits. Bankadvances,andotherborrowedmoney. 7. Includesprovisionsforallocatedtransferrisk. 3. Measuredasthesumofconstructionandlanddevelopmentloanssecured *Inabsolutevalue,lessthan0.005percent. byrealestate;realestateloanssecuredbynonfarmnonresidentialpropertiesor n.a. Notavailable. bymultifamilyresidentialproperties;andloanstofinancecommercialreales- MMDA Moneymarketdepositaccount. tate,construction,andlanddevelopmentactivitiesnotsecuredbyrealestate. RP Repurchaseagreement. 4. Otherrealestateownedisacomponentofothernon-interest-earningassets. MBS Mortgage-backedsecurities.

A108 Federal Reserve Bulletin 2006 A.2. Reportofincome,allU.S.banks,1996–2005 Millionsofdollars Item 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 Grossinterestincome..................... 313,696 338,865 359,675 366,137 423,839 404,406 350,040 329,757 350,028 428,064 Taxableequivalent................... 316,156 341,298 362,140 368,764 426,476 407,093 352,788 332,540 353,011 431,015 Loans ................................. 239,850 256,141 271,441 278,537 326,800 311,664 269,828 258,130 269,704 328,405 Securities.............................. 50,631 52,660 56,598 62,116 67,665 63,089 59,316 53,315 58,588 65,883 Grossfederalfundssoldandreverse repurchaseagreements ............. 9,272 13,658 14,999 12,330 13,546 12,649 6,223 5,122 5,245 11,012 Other.................................. 13,944 16,406 16,637 13,155 15,829 17,006 14,672 13,189 16,490 22,764 Grossinterestexpense .................... 150,249 164,692 178,161 174,946 222,159 188,799 118,920 94,471 99,261 162,488 Deposits............................... 107,512 117,350 125,217 119,665 151,145 132,352 81,899 62,753 64,001 105,731 Grossfederalfundspurchasedand repurchaseagreements ............. 16,780 20,439 22,182 21,130 26,860 19,590 9,920 7,590 9,203 19,345 Other.................................. 25,956 26,903 30,760 34,149 44,155 36,854 27,101 24,128 26,057 37,411 Netinterestincome....................... 163,447 174,173 181,514 191,191 201,680 215,607 231,120 235,286 250,767 265,576 Taxableequivalent................... 165,907 176,606 183,979 193,818 204,317 218,294 233,868 238,069 253,750 268,527 Lossprovisions .......................... 16,395 19,402 21,427 21,186 29,386 43,238 45,278 32,767 23,895 25,581 Non-interestincome ...................... 95,313 105,640 123,668 144,429 153,163 160,925 168,484 183,614 188,216 200,578 Servicechargesondeposits............. 17,050 18,558 19,769 21,497 23,719 26,873 29,631 31,693 33,459 33,880 Fiduciaryactivities ..................... 14,296 16,584 19,268 20,502 22,220 21,989 21,404 22,456 25,102 26,390 Tradingrevenue ....................... 7,525 8,018 7,693 10,429 12,235 12,382 10,735 11,446 9,954 14,355 Other.................................. 56,444 62,480 76,939 92,001 94,988 99,679 106,717 118,019 119,702 125,953 Non-interestexpense ..................... 162,581 171,060 193,833 204,632 216,432 226,027 230,292 243,306 263,364 274,150 Salaries,wages,andemployeebenefits .. 67,826 72,346 79,538 86,151 89,036 94,209 100,455 108,471 115,281 124,072 Occupancy ............................ 20,892 22,080 24,164 25,865 26,765 27,945 29,316 31,318 33,258 35,053 Other.................................. 73,865 76,634 90,129 92,616 100,631 103,875 100,520 103,516 114,825 115,025 Netnon-interestexpense.................. 67,268 65,420 70,165 60,203 63,269 65,102 61,808 59,692 75,148 73,572 Gainsoninvestmentaccount securities ............................ 1,123 1,825 3,090 250 −2,280 4,625 6,411 5,633 3,792 −220 Incomebeforetaxes ...................... 80,908 91,177 93,016 110,055 106,744 111,891 130,448 148,459 155,515 166,204 Taxes ................................. 28,447 32,001 31,965 39,211 37,250 37,284 42,956 48,456 50,340 53,652 Extraordinaryitems,netofincometaxes. 88 56 506 169 −31 −324 −78 427 59 240 Netincome .............................. 52,550 59,230 61,556 71,012 69,463 74,284 87,413 100,431 105,234 112,791 Cashdividendsdeclared ................ 39,419 42,801 41,205 52,101 52,547 54,844 67,230 77,757 59,587 64,617 Retainedincome ....................... 13,131 16,430 20,351 18,912 16,916 19,438 20,183 22,674 45,647 48,175

A109 Credit Card Disclosures, Solicitations, and Privacy Notices: Survey Results of Consumer Knowledge and Behavior ThomasA.Durkin,oftheBoard’sDivisionofResearch kets competitive, which benefits buyers of products and Statistics, prepared this article. Christine N. and services by minimizing the spread between Jonesprovidedresearchassistance. producers’ production costs and market price.1 The second advantage of information disclosure The mandatory dissemination of certain information over direct intervention through mandating specific by financial institutions is a key aspect of consumer product pricing or features is that the government protectionlaw.Itofferstwosignificantadvantagesfor need not know, or presume to know, the product consumer protection in the financial area over the feature preferences of all consumers. With effective alternative of direct government intervention into disclosures, consumers can decide what their preferproductpricingandcontent.First,informationdisclo- ences are in the tradeoff between price and product sure is compatible with competition, a significant features; the success of the disclosure approach to marketforcealreadyatworktoprotectconsumersby consumer protection does not depend on consumers’ keepingpricerisesincheck.Becauseofcompetition, preferences being the same. Disclosure requirements institutions already have incentives to make their may also be less costly for financial institutions to products known, to reveal favorable pricing and product features, and to treat consumers fairly by keeping them generally informed about what they 1.Researchershavepublishedasignificantbodyoftheoreticaland empiricalworkonthebenefitsofinformationanddisclosure.Among want and need to know. When a financial institution theimportantarticlesareGeorgeJ.Stigler(1961),‘‘TheEconomicsof employsthesestrategies,itgeneratesagoodbusiness Information,’’JournalofPoliticalEconomy,vol.69(June),pp.213– reputation that will produce referrals and repeat cus- 25; Phillip Nelson (1970), ‘‘Information and Consumer Behavior,’’ Journal of Political Economy, vol. 78 (March–April), pp. 311–20; tomers. Actions that firms use to accomplish these GeorgeAkerlof(1970),‘‘TheMarketforLemons:QualitativeUncergoals include advertising their prices and supplying taintyandtheMarketMechanism,’’QuarterlyJournalofEconomics, clients and potential customers with useful informa- vol. 84 (August), pp. 488–500; Michael A. Spence (1973), ‘‘Job MarketSignaling,’’QuarterlyJournalofEconomics,vol.87(August), tionaboutproductpricesandfeatures. pp.355–74;MichaelRothschild(1973),‘‘ModelsofMarketOrgani- The requirements for disclosures assist in the zation with Imperfect Information: A Survey,’’Journal of Political dissemination of financial information by standard- Economy, vol. 81 (November), pp. 1283–1308; Howard Beales, RichardCraswell,andStevenC.Salop(1981),‘‘TheEfficientReguizing concepts and terminology, such as the finance lation of Consumer Information,’’ Journal of Law and Economics, charge and annual percentage rate under theTruth in vol.24(December),pp.491–539;JosephE.Stiglitz(1985),‘‘Informa- Lending Act and the annual percentage yield under tionandEconomicAnalysis,’’TheEconomicJournal,vol.95,Supplement:ConferencePapers,1985(March),pp.21–41;andPaulineM. the Truth in Savings Act. Such standardization ad- Ippolito(1988),‘‘TheEconomicsofInformationinConsumerMarvances consumers’ knowledge about pricing and kets:WhatDoWeKnow?WhatDoWeNeedtoKnow?’’inE.Scott features of the financial products and institutions Maynes, ed., The Frontiers of Research in the Consumer Interest (Columbia,MO:AmericanCouncilonConsumerInterests),pp.235– and lowers consumers’ transactions costs by making 63.ImportantgovernmentreportsincludeFederalTradeCommission shopping easier. The standard format of required Staff(1979),ConsumerInformationRemediesPolicySession(Washdisclosures helps highlight the performance of the ington:FederalTradeCommission);andBoardofGovernorsofthe FederalReserveSystem(1987),AnnualPercentageRateDemonstrabest institutions and exposes the inadequacies of the tionProject(Washington:BoardofGovernorsoftheFederalReserve poorer ones. Well-informed shoppers help keep mar- System).

A110 Federal Reserve Bulletinh2006 implement and for the government to enforce than inLendingAct.4Asthesemandatorydisclosureshave consumer protection approaches that limit product taken their place in the financial marketplace and as features. consumer financial services have expanded and evolved, researchers and other observers have been interested in whether consumers use the disclosures TRUTH IN LENDING ACT AND DISCLOSURES they receive and, if so, how they use them. One way to examine consumer knowledge and use of the The Congress in May 1968 passed the Truth in disclosures is through surveys. For this reason, the LendingAct, the first in a series of federal consumer Federal Reserve Board has conducted and analyzed protection laws that addressed primarily financial targeted, nationally representative consumer surveys disclosures.2 This act was designed to protect conin this area since a before-and-after study of the sumersincredittransactionsbyrequiringcleardisclooriginalimplementationoftheTruthinLendingAct.5 sure of key terms of the credit arrangement and all Nationally representative surveys provide informacreditcosts.Thelawwasimplementedin1969bythe tionaboutconsumers’impressionsandexperiencesto Federal Reserve Board through Regulation Z, which supplementinstitutionalknowledgefrompubliccomprescribesuniformmethodsforcomputingthecostof credit, for disclosing credit terms, and for resolving errorsoncertaintypesofcreditaccounts.In1976,the 4. Various recent legislative and regulatory initiatives concerning creditcardshavecontinuedtounderscoreinterestinwhatconsumers Congressamendedtheacttocoverconsumerleasing, knowabouttheiraccounts,howtheyusedisclosureinformation,what and the Federal Reserve implemented Regulation M, theythinkisimportantinthedisclosures,andwhatinformationthey which covers the rules of all consumer leasing trans- wanttoreceive.Inthelegislativearea,theFairandAccurateCredit Transactions (FACT) Act of 2003, which amended the Fair Credit actionsandincludesdisclosureofleasingterms. ReportingActof1971,providedforadditionaldisclosurestoconsum- Credit cards are the most widely used method of ersabouthowtheycouldmoreeasilytaketheirnamesoffsolicitation generating consumer credit. The credit card industry listsforneworadditionalcreditcards,aremovalprocesssometimes also referred to as opting out. This act also required the Federal estimatesthatmorethan1billioncreditcardswerein ReserveBoardtoundertakeastudyofcreditcardsolicitations,which thehandsofcustomersintheUnitedStatesattheend the Board completed in December 2004. In April 2005, the Bankof 2004.3 When they use their cards, consumers ruptcyAbusePreventionandConsumerProtectionActmandatednew provisions for the Truth in Lending Act concerning open-end conreceive monthly account statements that contain dissumercredit,anditrequiredfurtherBoardstudiesofconsumersand closures about credit use, costs, and obligations for theircredit. payments. Elsewhere on the monthly statements, On the regulatory front, in December 2004 the Federal Reserve BoardbeganformalreviewandupdatingofRegulationZ,therulethat consumersreceivedisclosuresconcerningsuchitems implementsTruthinLending.Thefirststepinthereviewprocesswas asgraceperiods,membershipfees,minimumfinance anAdvanceNoticeofProposedRulemaking(ANPR),whichaskedfor charges,andproceduresforquestioningandresolving comments on a lengthy list of questions about open-end consumer credit.Thefullreviewprocesswilltakesometimeandwilllikelyraise billing errors. In addition, consumers frequently many additional questions about how well consumers understand receive mailed solicitations for new accounts, and credit products and how they use them, including credit cards. In these mailings carry disclosures. In recent years, October 2005, the Federal Reserve reopened the ANPR comment period, asking for public comment on issues raised by Truth in consumershavealsobegunreceivingprivacynotices Lending Act amendments in the bankruptcy reform legislation that from their financial institutions, and these notices year. Interagency initiatives to revise privacy notification rules for containdisclosuresoutliningtheinstitutions’privacy financialaccounts,includingcreditcardaccounts,arealsounderway. Ineachoftheseefforts,whatconsumersknowandwanttoknowhave policies and information on how customers can ‘‘opt beenimportantquestions. out’’ of certain kinds of information sharing among 5.RobertP.ShayandMiltonP.Schober(1973),ConsumerAwareinstitutions. nessofAnnualPercentageRatesofChargeinConsumerInstallment Credit:BeforeandAfterTruthinLendingBecameEffective,vol.1.: The content, format, and number of disclosures TechnicalStudiesoftheNationalCommissiononConsumerFinance haveevolvedandchangedsincepassageoftheTruth (Washington: Government Printing Office). For later survey results, refer toThomasA. Durkin and Gregory E. Elliehausen (1978), The 1977 Consumer Credit Survey (Washington: Board of Governors of the Federal Reserve System); Anthony W. Cyrnak and Glenn B. 2. Other statutes that focus on financial disclosure are the Real Canner(1986),‘‘ConsumerExperienceswithCreditInsurance:Some EstateSettlementProceduresAct(1974),theConsumerLeasingAct New Evidence,’’Federal Reserve Bank of San Francisco Economic (1976),andtheTruthinSavingsAct(1991).Themainintentofthese Review(Summer),pp.5–20;GregoryE.ElliehausenandBarbaraR. lawsistoprotectconsumersbythemandatorydisclosureofcertain Lowrey(1997),TheCostofImplementingConsumerFinancialReguinformation. Other consumer protection laws also contain important lations:AnAnalysisofExperiencewiththeTruthinSavingsAct,Staff requirementsfordisclosures,thoughtheyarenotprimarilydisclosure Study170(Washington:BoardofGovernorsoftheFederalReserve statutes;examplesincludetheFairCreditReportingAct(1971),the System); Thomas A. Durkin (2000), ‘‘Credit Cards: Use and Con- EqualCreditOpportunityAct(1974),andtheElectronicFundTrans- sumer Attitudes, 1970–2000,’’ Federal Reserve Bulletin, vol. 86 ferAct(1978). (September),pp.623–34;andThomasA.Durkin(2002),‘‘Consumers 3.ThomsonFinancialMedia(2005),CardIndustryDirectory,17th andCreditDisclosures:CreditCardsandCreditInsurance,’’Federal ed.(NewYork:ThomsonFinancialMedia),p.16. ReserveBulletin,vol.88(April),pp.201–13.

Credit Card Disclosures A111 ments generated through the regulatory process. Tar- 1. Examinationfrequencyofselecteddisclosure geted consumer surveys help reduce the need to rely informationamongholdersofbank-typecreditcards, 2005 unduly on opinions of interested parties or anecdotal Percentofcardholders reports for assessments of consumers’ disclosure use andtheirpreferences. Disclosureinformation In2004and2005,severalsurveyswereundertaken Item Annual Descriptive to assess consumers’ knowledge of, familiarity with, percentage material rate andattitudesaboutcreditcarddisclosures,creditcard solicitations, and privacy notices received from their Frequentexamination Everymonth ..................... 46.6 12.1 financial institutions.6 The targeted surveys supple- Everyothermonth ................ 5.6 6.2 Fourtofivetimesperyear ........ 11.3 15.5 mented the comprehensive Surveys of Consumer Subtotal........................ 61.5 33.8 Finances,whichareundertakeneverythreeyearsand Infrequentexamination whichprovidegeneralbenchmarksandgrowthtrends Lessoftenthanfourtofivetimes peryear ..................... 29.6 57.7 for consumer assets, debts, and use of financial Never(volunteered)............... 8.9 8.6 Subtotal........................ 38.5 66.3 services.7 In each case, the survey goal was to assess the frequency with which consumers examine or Total............................. 100.0 100.0 consult disclosures and their attitudes toward the Source: SurveysofConsumers,January2005. disclosuresreceived.Ifconsumerslookatthedisclosures frequently and are favorably inclined toward at a wide variety of outlets and which can generate their usefulness, then the benefits of informed credit revolving credit if the user chooses to pay less than use and of enhanced competition in the market for thefullstatementbalance. financialproductscanfollow. One line of questioning directly asked consumers with this type of card how often they examined the CREDIT CARD PERIODIC STATEMENTS AND pricing and other disclosures they received monthly as part of the periodic statements from the card- DISCLOSURES issuing bank. If the disclosures are examined fre- In January 2005, the Federal Reserve Board spon- quently, especially the pricing disclosures, then the sored a survey about the importance to consumers of Truth in LendingAct can have a favorable effect on various required disclosures for their credit card ‘‘the informed use of credit,’’ as the Congress inaccounts, consumers’ use of the disclosure informa- tended.8 tionprovided,andtheirnewaccountsandpaymentof Consumersgavethefullrangeofpossibleanswers fees. Slightly more than 73 percent of respondents tothequestiononhowoftentheyexaminetheannual reported holding one or more general-purpose credit percentagerates(APRs).Themajorityofcardholders cards with a revolving credit feature. The cards are (62 percent) said they looked at the APRs on their sometimescalledbank-typecreditcardsbecausethey card accounts at least four times or more per year, used to be issued only by banks; examples are cards timing characterized for discussion here as ‘‘frelikeDiscover,MasterCard,andVisa,whichareusable quently’’ (table 1). More than 40 percent said they lookedattheAPRsmonthly.Incontrast,themajority (66percent)saidtheylookedatthedescriptivemate- 6.ThesurveyscitedinthisarticlewereundertakenfortheFederal Reserve Board by the Survey Research Center of the University of rial, the information often found on the back of Michigan.Thecenterconducted500interviewsoncreditcardsolici- statements, fewer than four to five times a year, tationsandprivacynoticesinMay2004and494interviewsoncredit timingcharacterizedhereas‘‘infrequently.’’ cardperiodicstatementsinJanuary2005. 7.The1995,1998,2001,and2004SurveysofConsumerFinances Not surprisingly, the frequency of APR examinaare discussed, respectively, in Arthur B. Kennickell, Martha Starr- tioncorrelatesdirectlywiththeuseofcardsascredit McCluer,andAnnikaE.Sunden(1997),‘‘FamilyFinancesintheU.S.: generators rather than as convenient transaction de- Recent Evidence from the Survey of Consumer Finances,’’Federal ReserveBulletin,vol.83(January),pp.1–24;ArthurB.Kennickell, MarthaStarr-McCluer,andBrianJ.Surette(2000),‘‘RecentChanges inU.S.FamilyFinances:Resultsfromthe1998SurveyofConsumer 8.TheCongressarticulateditscentralgoalforTruthinLendingin Finances,’’FederalReserveBulletin,vol.86(January),pp.1–29;Ana section102oftheact:‘‘TheCongressfindsthateconomicstabilization M. Aizcorbe, Arthur B. Kennickell, and Kevin B. Moore (2003), wouldbeenhancedandthecompetitionamongthevariousfinancial ‘‘RecentChangesinU.S.FamilyFinances:Evidencefromthe1998 institutions and other firms engaged in the extension of consumer and2001SurveyofConsumerFinances,’’FederalReserveBulletin, credit would be strengthened by the informed use of credit. The vol.89(January),pp.1–32;andBrianK.Bucks,ArthurB.Kennickell, informeduseofcreditresultsfromanawarenessofthecostthereofby and Kevin B. Moore (2006), ‘‘Recent Changes in U.S. Family consumers. It is the purpose of this title to assure a meaningful Finances: Results from the 2001 and 2004 Survey of Consumer disclosureofcredittermssothattheconsumerwillbeabletocompare Finances,’’ Federal Reserve Bulletin, vol. 92, pp. A1–A38, morereadilythevariouscredittermsavailabletohimandavoidthe www.federalreserve.gov/pubs/bulletin/default.htm. uninformeduseofcredit.’’

A112 Federal Reserve Bulletinh2006 2. Examinationfrequencyofannualpercentagerate(APR) interesttomanyconsumersbecauseitismoregeneral andofdescriptivematerialbycardholdergroup,2005 andappearsmorelegalistic.Also,muchofitpertains Percentofgroup to more-limited circumstances, such as balance com- Cardholdergroup Response putation formulas and service calls for errors or for further information, which are less likely than price GroupsofconsumersmorelikelytoexamineAPRfrequently Accountstatus comparisontobetheobjectiveofconsumershopping. Havethreeormoregeneral-purposerevolvingcards ........ 66.5 The proportion of cardholders who reported frequent Openednewaccountinlastyear........................... 67.4 Balance examinationofthisinformationdidnotvaryasmuch Balanceafterlastpaymentpositivebutlessthan$1,500 .... 68.4 Balanceafterlastpaymentatleast$1,500butlessthan according to account usage asAPR examination, but $4,500 ............................................... 74.0 Balanceafterlastpayment$4,500ormore ................. 79.7 the pattern was generally the same. Those who used Revolvingbalance their credit cards as credit-generating devices exam- Payfullbalancehardlyever ............................... 79.2 Payfullbalancesometimes................................ 81.6 ined this material more frequently than those who GroupsofconsumerslesslikelytoexamineAPRfrequently used their cards as transactions devices.Almost two- Accountstatus Haveoneortwogeneral-purposerevolvingcards........... 57.4 fifthsoftherespondentswithbalancesreportedexam- Didnotopennewaccountinlastyear ..................... 59.8 ining this information frequently, but only about Balance Nobalanceoutstandingafterlastpayment.................. 40.8 one-quarter of those with no balance after their last Revolvingbalance Payfullbalancealmostalways ............................ 46.7 paymentdidso. Groupsofconsumersmorelikelytoexamine To ascertain what specific information on their descriptivematerialfrequently monthly statement consumers consider important, Accountstatus Openednewaccountinlastyear........................... 36.7 they were asked two open-ended questions: What Balance Balanceafterlastpaymentpositivebutlessthan$1,500 .... 38.8 information did they consider important enough to Balanceafterlastpaymentatleast$1,500butlessthan $4,500 ............................................... 37.0 lookateachmonth?Whatinformationismostimpor- Balanceafterlastpayment$4,500ormore ................. 39.7 tant?The interviewers recorded up to two replies for Revolvingbalance Payfullbalancehardlyever ............................... 39.9 eachquestion. Payfullbalancesometimes................................ 45.7 Respondents gave various answers to both ques- Groupsofconsumerslesslikelytoexamine descriptivematerialfrequently tions; however, their replies could be grouped into Accountstatus four categories: cost measures, correctness of credit Didnotopennewaccountinlastyear ..................... 32.9 Balance card statement, measures of personal finances, and Nobalanceoutstandingafterlastpayment.................. 23.5 Revolvingbalance miscellaneous (table 3). Seventeen percent replied Payfullbalancealmostalways ............................ 26.8 that cost measures were most important to them, Source: SurveysofConsumers,January2005. especially interest rates, compared with 29 percent who replied that aspects of statement correctness vices.Thesurveyfoundthatacreditcardholderwas likely to examine the rate more frequently as credit were most important to them. More than 60 percent use increased. Specifically, about 41 percent of card- repliedthatthedataontheperiodicstatementrelated holders reporting no balance outstanding said they toaspectsoftheiroverallpersonalfinancialcondition examined theAPR frequently; by comparison, about were the most important information, especially the 80 percent of those with an outstanding balance of balanceowed. $4,500ormoreexaminedtheAPRfrequently(table2). As with examination frequency of theAPR on the Ingeneral,thosewithsmalleroutstandingbalancesat periodicstatement,whichcategoryofinformationthe the time of the interview and those who said they respondents considered most important is correlated normallypaytheirbalancesinfullwerelesslikelyto withwhethertheyusetheircardsascredit-generating examine the APR frequently. These findings seem devices or mostly for transactions. Of those with a entirely reasonable because the cost of using credit credit card balance of more than $1,500, more than cards as credit devices would be more important to one-quarter said that cost measures were the most those using this kind of credit regularly, and so they important information for them; of those with no aremorelikelytoexamineAPRs.Creditcostisnotas balance outstanding, less than 10 percent said cost likely to be important to those who mostly use their measures were most important. In contrast, among cardsastransactiondevices. those who replied that correctness measures were Asnoted,thepercentageofthosewhoexaminethe most important, the percentage was higher among descriptivematerialfrequentlyismuchlowerthanthe thosewithnobalanceoutstandingthanthepercentage percentageofthosewhoexaminetheAPRfrequently. of those with revolving balances. Overall, however, The descriptive information tends to be denser than personalfinancialmeasures,especiallybalanceowed, the pricing disclosures, and arguably it is of less were mentioned most often as most important, and

Credit Card Disclosures A113 3. Importanceofselecteddisclosureinformationon 4. Responsetoquestionofwhetherannualpercentagerate periodicstatementsamongholdersofbank-typecredit (APR)orfinancechargeaffectedcardusedecisions cards,anddistributionofresponsesbycardholder amongholdersofbank-typecreditcards,and group,2005 distributionofresponsesbycardholdergroup,2005 Percent Percent Categoryofdisclosureinformation Mostimportant Important Groupcategory APR Finance Either charge Costs Anymentionofcosts.................... 17.0 40.2 Allrespondents ......................... 27.3 24.5 39.2 Interestrate........................... 13.5 30.1 Withincardholdergroups Specificfeesorfees................... 3.2 9.4 Payfullbalancealmostalways ........ 16.8 21.1 29.5 Financecharges....................... 1.1 5.0 Payfullbalancesometimes............ 50.0 29.5 57.4 Anymentionofcostswithin Payfullbalancehardlyever ........... 36.4 29.5 50.0 cardholdergroups Nobalanceoutstanding ............... 15.7 15.2 22.5 Payfullbalancealmostalways ........ 10.1 30.5 Balancepositivebutlessthan$1,500 .. 30.3 28.9 52.8 Payfullbalancesometimes............ 21.0 57.9 Balanceatleast$1,500butless Payfullbalancehardlyever ........... 25.1 49.5 than$4,500 ...................... 33.6 33.2 51.4 Nobalanceoutstanding ............... 7.2 24.0 Balance$4,500ormore ............... 41.7 31.7 51.0 Balancepositivebutlessthan$1,500 .. 15.5 41.2 Balanceatleast$1,500butless Source: SurveysofConsumers,January2005. than$4,500 ...................... 25.0 49.5 Balance$4,500ormore ............... 27.2 56.9 those with no balance outstanding after their last Correctnessofbillingstatement payment reviewed them monthly (first panel of Anymentionofcorrectness .............. 29.3 51.4 Purchases,transactions,charges, table 3). Again, measures of personal finances are returnsarecorrect ................ 28.2 46.2 Previouspaymentreceived ............ 1.8 9.1 mentioned most often as being consulted monthly, Accountnumber,name,andsoforth especiallytotalbalanceowed. arecorrect ....................... * 2.2 Anymentionofcorrectnesswithin The survey further explored the use of disclosures cardholdergroups Payfullbalancealmostalways ........ 35.3 53.9 aboutAPRs and finance charges through direct ques- Payfullbalancesometimes............ 31.0 56.3 Payfullbalancehardlyever ........... 19.8 47.5 tions. Toward the end of the interview, consumers Nobalanceoutstanding ............... 40.5 60.8 were asked specifically whether APRs and finance Balancepositivebutlessthan$1,500 .. 19.7 40.5 Balanceatleast$1,500butless charge disclosures had affected credit decisions. Not than$4,500 ...................... 18.3 48.1 Balance$4,500ormore ............... 26.6 50.3 surprisingly, the pattern of responses was similar to that discussed earlier. Those with frequently revolv- Personalfinances Anymentionofpersonalfinances ........ 61.0 83.3 ingbalancesandthosewithrelativelylargeoutstand- Balanceowed......................... 53.5 75.2 Duedate ............................. 5.0 19.5 ing balances were much more likely to respond that Itemization;wheremoneywasspent ... 2.5 11.6 Minimumpayment .................... 1.1 3.4 APRsandfinancechargeshadaffectedtheircarduse Creditlimit;availablecredit ........... .4 2.1 behavior(table4).Whentheresponsesofbothgroups Keepingtrackofdebt ................. * .3 Anymentionofpersonalfinances are combined, about 40 percent of respondents said withincardholdergroups Payfullbalancealmostalways ........ 61.6 84.7 thateithertheAPRorthefinancechargehadaffected Payfullbalancesometimes............ 53.7 73.2 theircardusedecisions. Payfullbalancehardlyever ........... 63.6 85.7 Nobalanceoutstanding ............... 58.5 82.2 The proportions differ sharply within cardholder Balancepositivebutlessthan$1,500 .. 70.4 86.3 Balanceatleast$1,500butless subgroups, however; the proportion of those with than$4,500 ...................... 66.4 85.1 Balance$4,500ormore ............... 55.2 83.4 revolving balances who stated that the cost information on their monthly statements had affected their Miscellaneous Anymentionofmiscellaneous ........... .9 4.6 card use decisions was much higher than the propor- Detailsofandchangesinpolicies ..... .9 2.4 Rebates,incentives,rewards ........... * 1.9 tionofconvenienceuserswhosaidso.Morethanhalf Billingdate ........................... * .3 of those reporting that they paid the full balance Graceperiod.......................... * * sometimes or hardly ever and those who reported a Note: Components do not sum to totals because some respondents gave morethanoneresponse. balance outstanding after making their last payment *Percentagetoosmalltobemeasured. indicated that theAPR, finance charge, or either had Source: SurveysofConsumers,January2005. affected their card use decisions. When asked how the percentage did not vary much with patterns of thisinformationhadaffectedtheirdecisions,themost credituse(thirdpaneloftable3). common responses were that it made them decide to The same patterns are visible among those vari- payofftheirbalancesmorequickly,encouragedthem ablesthatconsumersreportasbeingimportantenough to stop using a particular account or to use the that they look at them monthly, even if they are not account with the lowest rate, or encouraged them to the most important information. Notably, 57 percent limit card use altogether (data not in table). When ofthosewithrelativelylargeoutstandingbalanceson asked the rate on the general-purpose card they used their credit cards reported that they reviewed cost most often, only about 4 percent of those with measures at least monthly, whereas 24 percent of revolving balances said they did not know the rate,

A114 Federal Reserve Bulletinh2006 whereastheotherswithrevolvingbalancesresponded fees are really not so important to consumers as long with rates within reasonable ranges. These results asthechargesareclearandunderstandable. support the view that Truth in Lending has been The interviewers did not ask further questions important in providing useful information to users of about fees like annual fees, late fees, and over-limit creditcardsascredit-generatingdevices. fees that are not finance charges under Truth in Theintervieweraskedallrespondentswithgeneral- Lending.Buttheydidaskaboutexperiencewithfees purpose credit cards whether the distinction made on for cash advances and convenience checks that are credit card account statements between a finance considered finance charges under the regulation. charge and other kinds of charges and fees made a About15percentofthosewithbank-typecreditcards difference to them. All respondents to this question, reportedthattheyhadpaideitherorbothofthesefees whether answering yes or no, were then asked why inthepastyear(datanotintable). theyrespondedthewaytheydid. Thenumberofbank-typecardholderswhoreported Interestingly, 77 percent indicated that the distincpaying such finance charges is fairly small (15 pertion mattered.At first glance, it might seem that this cent),soadetailedanalysisofthefinancialcondition distinction likely would not matter because a fee is a of the consumers within this group is precluded. But fee regardless what it is called, but respondents it is worth noting that most of those reporting that indicatedthatthedistinctiondoesmattertothem.One they had paid these finance charges also said they possibility is that consumers like to feel comfortable knewaboutthechargesbeforethetransaction.When that someone else is keeping an eye on credit card asked, most indicated they had learned about the companies,aphenomenonnotedinanearlierBulletin chargesfromthedisclosuresgiventothem. article.9Thegovernmentprovidesthatmonitorshipof In contrast, relatively few of those reporting they creditcardcompaniesthroughtherequirementsofthe had paid these finance charges recalled seeing the Truth in Lending Act, and consumers can use the effect of the charges on theAPR on their statements disclosures to examine fees and how they are classithatmonth.TruthinLendingrequiresthatallfinance fiedundertheact. chargesbefactoredintoan‘‘effective’’APRthatisto This interest in monitoring fees is apparent in the be included on each periodic (monthly) statement. In answerstothequestionaboutwhytheyrespondedin a month in which a finance charge arises from a the way they did. For those who felt the distinction source other than application of the normal periodic wasimportant,themostcommonreasonsgivenwere rate to the balance (for example, finance charges for thattheywantedtoknowspecificallywhattheywere cash advances and for convenience checks), the paying for and that they wanted to avoid fees (data charge must be factored into the disclosedAPR that not in table).The specific fees most commonly menmonthastheeffectiveAPR(sometimesreferredtoas tioned were annual fees and late fees, neither of the‘‘historical’’APR). which is actually classified as a finance charge under Less than one-quarter of the 15 percent of card- Truth in Lending, and avoidance of ‘‘unexpected holders reporting these charges for cash advances or fees,’’ which was not further specified by type. convenience checks (about 3.5 percent of bank-type Absence in the open-ended response of a differentiation between fees by classification suggests that con- cardholders overall) said they noticed any change in sumersactuallydonothavemuchinterestinthelegal the APR in the month of the fee. Because every distinctions despite their expression to the contrary. cardholder paying one of these fees would find a Whatisapparent,however,isthattheywanttoknow higher APR on the statement that month, the low the amounts of fees so that they can fully understand response does not strongly support the efficacy and what is behind the payments requested on their usefulness of this disclosure for the majority of periodicbills. cardholders. For some, the change may have been so Amongthosewhosaidthatthedistinctionbetween smallinthemonthinquestionthatitcouldhavegone finance charges and other sorts of fees was not unnoticed, or the respondents may not have noticed important, the two most common reasons were that changes that occurred infrequently. Some may have theypaidofftheirbillsinfull,sosuchdistinctionsdid noticed the change at the time it happened but may not matter to them, or that all charges are the same. haveforgottenbythetimeoftheinterview. Thus, the responses of both groups suggest that Finally, those who responded affirmatively to a hair-splitting legal distinctions among categories of questionaboutopeninganewaccountinthepastyear wereaskedafewadditionalquestionsaboutaccountopening statement disclosures.This group was asked 9.Durkin,‘‘ConsumersandCreditDisclosures,’’p.208. only a few more questions because of the likelihood

Credit Card Disclosures A115 that the proportion with new accounts would be tions rather than on periodic statements. In recent relatively small; in a specialized interview, subdivid- years, prescreened mail solicitations, which employ ing the group in many interesting ways would not be creditexperienceinformationfromthefilesofcreditpossible because of its small size. Overall, about reportingagencies,havebecomeanimportantsource 22percentofrespondentswithgeneral-purposecredit of new accounts for card issuers; more than 6 billion cards said that they had opened a new credit card solicitations were mailed in 2005.10 Although many account within ‘‘the past year or so’’ (data not in of the surveyed consumers do not review carefully table). the mailings they receive, the requirement that the Each respondent with a new account was asked mailings contain pricing information means that a about the use and storage of the account-opening large volume of information on credit card pricing disclosure information. Most indicated that they had makes its way regularly into consumers’ mailboxes. read the information at least somewhat carefully and The specialized survey focused on consumers’ expehad kept it for later use if needed (data not in table). riencewiththereceiptofthesemailings.Someresults Some of the respondents may have been thinking of from this survey were included in a report to the the solicitation disclosures that are delineated in a Congress required in 2004 by the Fair and Accurate tabularformatratherthantheactualaccount-opening CreditTransactionsActof2003.11 disclosures, but the questioning did not make this The 2004 survey found that 77 percent of respondifferentiation. About three-quarters of those with dents had one or more credit cards at that time and new accounts said that the disclosures were useful, that 72 percent had one or more general-purpose mostlybecausetheyprovidedinformationaboutinter- creditcardswitharevolvingfeature(bank-typecredit est rates or reference information to catch mistakes cards).Ofthecardholdergroup,morethan96percent andmakecomparisons.Amongthelimitednumberof hadreceivedprescreenedmailsolicitationsforoneor respondents who said that the information was not more additional credit cards in the previous six useful,mostsaidthattherewastoomuchinformation months, and 81 percent of noncardholders had reorthatitwastooconfusingorlegalistic. ceived one or more solicitations. Most respondents In sum, the response to questions on the use of receivingsolicitationshadreceivedmorethanoneper periodicstatementsshowsthatmanyholdersofbank- month(table5).About50percentofcardholdersand type credit cards look at the disclosures fairly often. 30 percent of noncardholders had received six or The frequency with which they examine rates and more prescreened solicitations a month over the fees is correlated with whether or not they have a previous half year. Only a small proportion of either revolvingbalance.Thosepayingtheirbalancesinfull group had received only one prescreened solicitation each month also tend to look at the statements ornonepermonthduringthisperiod. frequently, but they do so to ensure accuracy of the Thesurveyaskedtherespondentswithcreditcards statements and to assess their personal financial con- who had received prescreened solicitations for more dition. The distinction between finance charges and information about their experiences with the mailotherfeesdoesnotseemespeciallyimportantdespite ings.12 One question concerned attitudes toward the proteststothecontrary.Mostofthosewithrevolving information in the mailings and asked if it was balances on their general-purpose credit cards seem helpful.About9percentindicatedtheydidnotknow, broadly aware of APRs, but the small number of so they probably did not pay much attention to the responsesabouttheeffectiveAPRsuggeststhatinfre- information (first panel of table 6). The rest of the quent changes do not receive much notice. However, respondents were almost evenly divided on whether the overall findings do indicate that the goal of the or not the information received was helpful: About informed use of credit by consumers is being addressed,astheCongressintended. 10.InformationPolicyInstitute(2003),TheFairCreditReporting Act:Access,Effıciency&Opportunity:TheEconomicImportanceof CREDIT CARD SOLICITATIONS AND Fair Credit Reauthorization, table 13, p. 57; and Synovate, Mail Monitor (2006), ‘‘Mail Monitor Reports Six Billion Card Offers DISCLOSURES Mailed in U.S. During 2005,’’ press release, June, http:// core.synovate.com. The general findings that reported usefulness of 11. Board of Governors of the Federal Reserve System (2004), ReporttotheCongressonFurtherRestrictionsonUnsolicitedWritten required disclosures on periodic statements is corre- OffersofCreditandInsurance(Washington:BoardofGovernorsof lated with measures of credit card usage are broadly theFederalReserveSystem,December). consistent with the results of another specialized 12. The survey did not ask those without credit cards many follow-upquestionsabouttheirexperienceswithsolicitationsbecause survey of credit card users undertaken in May 2004. thesmallsamplesizeofthisgroupwouldnotpermitfurtherclassifi- Thissurveyfocusedespeciallyoncreditcardsolicita- cationbreakdownoftheirexperiences.

A116 Federal Reserve Bulletinh2006 5. Cardholdersandnoncardholders:Creditcard 6. Creditcardsolicitations:Opinionsonhelpfulnessof solicitations,awarenessofopt-outlaw,andbehavior disclosureinformation,anddistributionofresponses response,2004 bycardholdergroup,2004 Percent Typeofinformation Response Consumersreceivingsolicitations Cardholders Noncardholders Informationingeneral Total .................................. 77.0 23.0 Veryhelpful .............................................. 7.2 Somewhathelpful ......................................... 35.2 Receivedsolicitationsforadditional Notveryhelpful .......................................... 17.6 cardsinprevioussixmonths ....... 96.1 80.8 Notatallhelpful .......................................... 31.3 Donotknow.............................................. 8.7 Approximatenumberofsolicitations receivedmonthly Informationthatishelpful Oneorfewer ........................ 11.8 18.9 Interestratesorannualpercentagerates(APRs) Twotofive .......................... 36.7 50.6 InterestratesorAPRs(notfurtherspecified) ............. 50.6 Sixormore ......................... 51.5 30.5 InterestratesorAPRs,introductory ...................... 11.4 Total ................................ 100.0 100.0 InterestratesorAPRs,standard.......................... 10.1 InterestratesorAPRsforbalancetransfers Awareofopt-outlaw ................... 20.8 16.9 orcashadvances ................................... 6.4 InterestratesorAPRs,fixedversusvariable .............. 6.7 Responsetoawareness AnymentionofinterestratesorAPRs1 .................. 68.5 Optedout ........................... 20.3 33.3 Thoughtaboutoptingout............. 38.2 13.9 Cardholdergroup:Anymentionofinterestrates Didnotconsideroptingout .......... 41.6 52.8 orAPRsashelpful Total ................................ 100.0 100.0 Nobalanceoutstanding ................................. 73.6 Balancelessthan$1,500 ................................ 49.6 Note: Componentsmaynotsumtototalsbecauseofrounding. Balanceatleast$1,500butlessthan$4,500.............. 69.4 Source: SurveysofConsumers,May2004. Balance$4,500ormore ................................. 79.3 Fees 40 percent said it was helpful, and slightly less than Fees(notfurtherspecified) .............................. 9.0 Annualfees,membershipfees ........................... 14.2 50 percent said it was not. Within the two negative- Balance-transferfees,transactionsfees,orboth........... 7.5 response subgroups, a much higher proportion took Latefees,penaltyfees .................................. 7.5 Anymentionoffees1 ................................... 35.4 the more extreme position that the information was Summary:Anymentionofinterestrates,APRs,orfees1 .... 80.0 very unhelpful, probably an indicator of frustration with receiving so much junk mail. Regardless, the Cardholdergroup:Anymentionofinterestrates, APRs,orfeesashelpful finding that a significant portion of consumers ap- Nobalanceoutstanding ................................. 84.4 Balancelessthan$1,500 ................................ 64.0 peared to be generally familiar with the kind of Balanceatleast$1,500butlessthan$4,500.............. 84.7 Balance$4,500ormore ................................. 86.4 informationintheprescreenedsolicitations—whether they stated it was helpful or not—is consistent with Othertypesofinformation Creditlimits,paymentpolicies,graceperiods ............ 13.3 the view that the prevalence of prescreened solicita- Benefits,rebates,rewards................................ 2.6 tions is useful in disseminating pricing information S G e e c n u e r r i a ty l , m p e ri n v t a io c n y s . ( . ‘‘ . t . e . rm .. s .. a . n . d .. c . o . n . d .. it . i . o . n . s . ,’ . ’ . ‘ . ‘ . p . r . e . - . a . p . p . r . o . v . a . l .. * and encouraging competitive conditions in markets qualifications,’’‘‘services,’’andsoforth) ............ 17.6 Anymentionofothertypesofinformation1 .............. 31.8 forcreditcardsgenerally,evenifonlyasmallminor- Donotknowresponses .................................... 3.3 ity of recipients actually responds to a given prescreenedsolicitationprogram. 1. Theresponsestothesecategoriesdonotsumtototalsbecauserespondentscouldgiveuptotworeplies. The respondents answering that the information *Lessthan0.5percent. washelpfulwereaskedwhatspecificinformationwas Source: SurveysofConsumers,May2004. helpful.About two-thirds mentioned interest rates or Anotherquestionaskedconsumerswhattheyactu- APRs(secondpaneloftable6).Somenotedthatthey ally do most often with the mailings they receive. found particular information on other rates helpful— Responsesindicatethatthemailingsarenotsolelyor for example, introductory rates or standard rates. always considered junk mail, even if they are so Aboutone-thirdmentionedspecificinformationabout regarded in many instances. Slightly more than half various fees, such as annual fees, balance-transfer ofrespondents(55percent)saidthattheythrowthem fees, and late fees. Again, the findings suggest that away,buttheotherssaidtheyatleastopenandlookat manyconsumersseemtoknowwhattheprescreened them—but not especially carefully (table 7). Memsolicitations contain, which is important for price bers of the group who said they usually opened the competition to work, even if they do not respond to, prescreened solicitations were asked a follow-up or even focus carefully on, the contents of any given questionaboutwhethertheylookedforanyparticular pieceofmailthattheyreceivefromcardissuers.13 formatandclarity(datanotintable).Thoseunfavorablyinclinedoften 13.Afollow-upquestionaskedboththosewhosaidtheinformation indicated either that they did know how the information could be washelpfulandthosewhosaiditwasnothowitcouldbemademore improvedorsaidthatfewermailingsshouldbesent.Thepossibilityis helpful.Respondentsgaveawidevarietyofanswers,butthosealready that the latter were not looking for any more credit cards and were favorablyinclinedtowardtheinformationoftensuggestedaspectsof frustratedwiththefrequencyofjunkmail.

Credit Card Disclosures A117 7. Dispositionofcreditcardsolicitations,anddistribution holders answered that they had heard of a federal ofbehaviorresponsebycardholdergroup,2004 law in this area. In turn, about 20 percent of card- Percent holders and 33 percent of noncardholders who said they knew of the federal law had placed their names Behavior Response on an opt-out list (lower panel of table 5). This Respondentsreceivingsolicitations response means that about 4 percent to 6 percent of Openandglanceatthem............................. 34.2 Openandlookmorecarefully........................ 10.0 respondents had placed their names on the opt-out Subtotal:Openandlookat ........................ 44.6 list.15 Because the survey found that only about Cardholdergroup:Thosewhoopenand lookatsolicitations 20 percent of consumers were aware of their right Withonegeneral-purposecreditcard ................. 41.7 under federal law to opt out from prescreened solici- Withtwogeneral-purposecreditcards ................ 40.7 Withthreeormoregeneral-purposecreditcards....... 48.7 tations, increased awareness may lead more consum- Withnobalanceoutstanding ......................... 40.9 Balancelessthan$1,500............................. 42.4 ers to do so. For consumers aware of the law, a Balanceatleast$1,500butlessthan$4,500 .......... 47.5 larger proportion of those with credit cards than Balance$4,500ormore ............................. 52.9 Receive1solicitationpermonth ..................... 72.2 those without (38 percent versus 14 percent) said Receive2to5solicitationspermonth ................ 46.8 Receive6ormoresolicitationspermonth ............ 37.1 that they had thought about placing their names on Throwthemaway ................................... 55.4 the opt-out list but had not yet done so. Consumers in the group aware of the federal law Total ............................................... 100.0 were also asked how they had heard of it. Most Source: SurveysofConsumers,May2004. mentioned their information was from the media, especially television, newspapers, and magazines information, and if so, what information.About two- (data not in table). Some consumers also mentioned thirds gave various answers that mostly focused on family, acquaintances, and other sources. Less than pricinginformation(datanotintable).Theremaining one-tenth of those with credit cards and aware of the one-third said that they looked for no particular law indicated that they learned of the right to opt out information. While many consider the mailings junk from the prescreened mail solicitation. With passage andtendtorapidlydisposeofthem,notallconsumers of additional time since the May 2004 survey on throwthemawaywithoutconsultingthem.Manyare solicitations and opting out, one possibility is that a clearly aware of the contents of the mailings they higher proportion of recipients would be aware of receive; in other words, they have direct access to opt-outrightstodaythanatthetimeofthesurvey. pricing and product information at a time when they Finally,afactorinanempiricalfindingfromearlier candecideaboutopeninganewcreditcardaccount.14 surveys may explain why the proportion of consum- As with attention to different aspects of periodic ers opting out was not higher in 2004 than it was, statements,behaviorwithrespecttosolicitationsvargiven that the number of those who knew of their ied by credit use. Those with more cards and those opt-outrightswasconsiderablyhigherthanthenumwith larger outstanding balances were more likely to ber of those who actually opted out. In particular, examinethemailingsthanthosewithfewercardsand earlier surveys found that consumers seem to mainthose with smaller balances. This finding suggests tainstrongfeelingsaboutwhatotherconsumersknow that active credit users are more likely to be looking or do not know and how they behave. This phenomfor cards that have more-favorable credit terms than enon,seeninearliersurveyresults,ischaracterizedas thoseoftheircurrentcardaccounts.Alsonoteworthy the ‘‘other guy effect,’’ whereby consumers indicate is that those who indicated receiving solicitations that they are better informed and likely to be more infrequently were also more likely to open and at least glance at the documents. This finding may responsiblethanunknown‘‘others.’’16Becausemany feel that their own private finances are under better indicatethatsomeofthemmayhavebeenpoorcredit controlthanthoseofconsumersingeneral,theymay risksinthepastandforthisreasonreceiveinfrequent feelthatoptingoutisunnecessary. solicitations,sotheonestheydoreceiveareofspecial interesttothem. The survey also asked respondents about removing their names from prescreened solicitation lists. 15. This figure is approximately the same proportion of opt-outs indicatedbyareviewofalargesampleofcredit-reportingagencyfiles About 20 percent of both cardholders and noncardatapproximatelythesametime.Formoreinformationonthatstudy, refer to Board of Governors, Report to the Congress on Further Restrictions on Unsolicited Written Offers of Credit and Insurance, 14.Responsestoafurtherquestioninthesurveyrevealedthatabout pp.17–27. 9percentofthosewithcreditcardsandreceivingsolicitationsinthe 16.Durkin,‘‘CreditCards:UseandConsumerAttitudes,’’pp.628–30 pastsixmonthshadrespondedtoasolicitationfromsomecardissuer and pp. 632–33; and Durkin, ‘‘Consumers and Credit Disclosures,’’ duringthattimeperiod(datanotintable). pp.204–6.

A118 Federal Reserve Bulletinh2006 8. Distributionofresponsesofcardholdersand most likely to receive prescreened solicitations for noncardholdersregardingopt-outlawandgovernment credit cards and are more likely than noncardholders prohibitionofcreditcardsolicitations,2004 to say that an opt-out list is a good idea, they largely Percent believe that the government should not prohibit such Attitudetoward solicitations. Presumably, this feeling is associated Cardholders Noncardholders All governmentintervention with the view that information about new products, Responsebygroup ...... 77.0 23.0 100.0 features,andpricingisworthwhile—evenifitisused Federalopt-outlaw only occasionally. Most consumers prefer the avail- Goodidea .............. 82.1 74.4 80.1 abilityofanopt-outlist,andtheyalsoprefertobethe Badidea ................ 16.7 20.6 17.7 Donotknow............ 1.3 4.9 2.1 onetochoosewhetherornottoplacetheirnameson Total ................... 100.0 100.0 100.0 thelist. Governmentshould prohibitsolicitations In sum, most consumers receive written offers of Yes ..................... 26.9 49.1 31.8 credit, and a significant portion appear to be at least No ..................... 70.8 49.1 65.2 Donotknow............ 2.3 1.9 2.1 somewhat familiar with the contents of the mailings, Total ................... 100.0 100.0 100.0 including a minority who were aware of the opt-out Note: Componentsmaynotsumtototalsbecauseofrounding. law. Among the more than 40 percent of respon- Source: SurveysofConsumers,May2004. dents who said that the information in credit card The May 2004 survey also found evidence of the solicitations is helpful, a large majority cited the ‘‘otherguyeffect.’’Whenaskeddirectlywhetherthey pricing information as helpful. Only a relatively think that pre-approved offers of credit cards cause small proportion had actually acted on the opt-out otherpeople,ingeneral,tousetoomuchcredit,about information and had their names placed on the 85percentsaidyes.Whenaskedalternativelywhether opt-out list maintained by credit-reporting agencies, pre-approved offers have led them to use too much popularly known as credit bureaus. Only a small credit, only 15 percent agreed. But most respondents percentage of consumers (15 percent) acknowledged also did not want the government to take specific that pre-approved offers of credit cards had led them actions to rectify the difficulty with excessive credit to overuse credit, but a large majority of consumers use that they perceived others might be facing.After (85 percent) believed the solicitations caused other questions about voluntarily placing their name on an consumers to overuse credit. Nonetheless, the majoropt-outlist,allrespondents(cardholdersandnoncard- ity of respondents indicated they did not want the holders)wereaskedfortheirviewsontheopt-outlaw government to restrict prescreened solicitations, preand on government intervention regarding pre- sumably because they did not want to restrict their screenedsolicitations.17 own opportunities to receive future offers, even if Theirresponsestothesequestionsexhibitadistinct they mostly responded by disposing of them. preference for an opt-out law, even if most do not personallyemployit,andfornogovernmentinterven- PRIVACY NOTICES AND DISCLOSURES tion in their opt-out decisions.About 80 percent said that a federal opt-out law is a good idea; the propor- Also in May 2004, consumers were surveyed about tion was somewhat higher among cardholders than theirknowledge,attitudes,andbehaviorregardingthe noncardholders (table 8). But almost two-thirds said privacy policies of their financial institutions. The they prefer that the government not prohibit pre- survey was designed to obtain a benchmark indicascreened solicitations, even though a majority said tion of consumers’ responses to these notices and a they do not open and peruse the ones they receive. basis for comparison should the format of such Again, this proportion is higher for cardholders than noticeschangemeasurablyinthefuture. noncardholders(70percentversus49percent). As might be expected, consumers responded over- Thesefindingssuggestthattheinconvenienceasso- whelmingly that privacy policies were important to ciated with receiving the mailings is overall not too them, but they gave various reasons why. They great for consumers, even though many apparently indicated that they did not often use privacy notices consider the mailings junk.Although cardholders are for direct comparisons between policies of institutions. This latter finding is not surprising, given the receipt frequency of these notices and their some- 17.Specifically,thefirstquestionwas‘‘Doyouthinkitisagood ideaorabadideathatthereisafederallawthatpermitsyoutoput timesdenseappearance. yournameonalistandthencreditcardcompaniescannotsendyou Survey results indicated that consumers are generthese offers?’’This question was followed immediately by a related allyawareofprivacypoliciesatfinancialinstitutions, question:‘‘Doyouthinkthegovernmentshouldprohibitcreditcard companiesfromsendingpre-approvedoffersforcreditcards?’’ a finding that is not surprising because those with

Credit Card Disclosures A119 9. Responsesregardingprivacynoticesoffinancial thought that privacy notices were ‘‘very useful’’ institutions,2004 (second panel of table 10), and 21 percent reported Percent thattheywere‘‘veryconfident’’thattheyunderstood Response thepolicies.Whenaskedinafollow-upquestionwhy Question Total they thought that the statements were useful or not, DoNot Yes No Know two-thirds of those who said the notices were useful gavevariousreasons,buttheresponsesweregrouped Awarenessoffederalprivacylaw forfinancialinstitutions ......... 73 27 * 100 aroundasmallernumberofthemes(datanotintable). Receiptofprivacynoticesfrom financialinstitutions ............ 78 22 * 100 About12percentcitedspecificfeaturesorusesofthe Knowledgethatmaindepository notices: explaining rights, enabling them to opt out, institutionhasanopt-outpolicy orhelpingthemevaluateinstitutions.Another16per- Allrespondents ..................... 67 7 26 100 Thoserecallingreceiptofprivacy cent were more vague, noting how the information notices ......................... 76 5 19 100 kept them up to date or provided useful general *Lessthan0.5percent. informationordetails.Another30percentmentioned Source: SurveysofConsumers,May2004. consumerprotections,suchaspreventionfrominforaccounts at financial institutions have received them, mation misuse by the institution or from identity possibly many. About three consumers in four were theft.About one-quarter mentioned general customer aware of a federal law in this area and had received awareness, peace of mind, and usefulness for any privacy notices from their financial institutions future problems. About 4 percent said they did not (table 9).About two in three recalled that their main knowwhytheyfeltthenoticeswereuseful;theother depository financial institution has an opt-out policy 10 percent gave other various answers.Among those concerning information sharing with other compa- whoreportedthenoticeswerenotuseful,themajority nies.Among those in this group who recalled receiv- gave reasons such as too much information, inundaing notices, this proportion was three in four. While tion,junkmail,andunhappinesswith‘‘legalese.’’ the response may well contain many guesses, it does Concerningbehaviorwithrespecttoprivacynotices indicate that the awareness that such a law exists is received, most consumers who recalled receiving widespread. them reported that they generally open and at least Concerning attitudes toward financial privacy and glance at them (table 11). Only about 14 percent privacy policies, consumers appear to regard the reported that they threw away the notices without privacy policies of financial institutions as generally consulting them or filed them for possible later use. importanttothem(table10).Alargemajorityofusers Most of those who threw away the notices without of financial institutions regard the protection of pri- even glancing at them did so because they felt the vacy as very important, and they generally believe noticeswereprettymuchthesameorbecausetheyjust that their own institutions protect their privacy well. didnothavetimeorinterestinthem(datanotintable). Asomewhatsmallerproportionofconsumersreplied Those who at least open and glance at the notices thattheprivacypolicyisakeycriterionforchoosing were asked whether they found any particular inforfinancialinstitutions. mationimportant.About3percentdeclinedtoanswer Among those respondents who recalled receiving the question, and another 4 percent said they did not privacynoticesfromfinancialinstitutions,24percent know. A majority (57 percent) said no information 10. Attitudesregardingprivacynoticesoffinancialinstitutions,2004 Percent Topic Very Somewhat Notvery Notatall Donotknow Total Importancethatprimaryfinancialinstitutionprotects personalinformationaboutaccounts ................. 88 9 2 * * 100 Likelihoodoftransferringinstitutionsifprimary institutiondidnotprotectpersonalfinancial informationadequately.............................. 70 17 6 6 1 100 Adequacyofcurrentinstitution’sprotectionofpersonal financialinformation ................................ 60 32 2 1 5 100 Importanceofinstitution’sprivacypolicycompared withotherreasonsforchoosingfinancialinstitutions . 48 36 12 3 1 100 Amongrespondentswhorecalledreceivingprivacynotices Usefulnessofprivacynotices ............................ 24 43 19 14 1 100 Confidenceinunderstandingprivacypolicies ............. 21 52 20 6 1 100 Note: Componentsmaynotaddtototalsbecauseofrounding. Source: SurveysofConsumers,May2004. *Lessthan0.5percent.

A120 Federal Reserve Bulletinh2006 11. Behaviorresponsetoreceiptofprivacynoticesandto 12. Accuracyofdefinitionsofselectedtermsinprivacy questionregardinguseofprivacynoticestocompare noticesandconfidencelevelsinunderstandingprivacy institutions,2004 noticesbylevelofeducation,2004 Percent Percent Behavior Response Affıliateterm:Definitionandconfidencelevel Response Openandlookcarefullyatnotices ................... 29.0 Definitionreferringtoaspectsofjointownership1 ............ 43.4 Openandglanceatnotices........................... 57.0 Byeducationlevel Throwthemaway ................................... 10.0 High-schooleducationorless.............................. 34.3 Other(fileforlaterreference) ........................ 4.0 Somecollege ............................................. 38.4 Donotknow ........................................ * Collegegraduateormore.................................. 50.7 Total ............................................... 100.0 Byconfidencelevelinunderstandingprivacynotices Veryconfident ............................................ 55.2 Privacynoticesusedtomakecomparisons Somewhatconfident....................................... 40.5 Yes ................................................. 5.0 Notveryconfident ........................................ 39.1 No.................................................. 95.0 Notatallconfident ....................................... 31.8 Total ............................................... 100.0 Definitionreferringtoothersortsofformalizedrelationships2 . 18.7 *Lessthan0.5percent. Byeducationlevel Source: SurveysofConsumers,May2004. High-schooleducationorless.............................. 12.4 Somecollege ............................................. 23.6 Collegegraduateormore.................................. 20.4 was particularly important, and the rest offered a Byconfidencelevelinunderstandingprivacynotices Veryconfident ............................................ 15.3 variety of answers that mostly focused simply on Somewhatconfident....................................... 19.0 Notveryconfident ........................................ 21.1 knowing that the institution would limit access to Notatallconfident ....................................... 23.9 personalinformation. Definitionreferringtoothersortsoforganizations3 ........... 27.7 Onlyaboutoneconsumerintwentyreportedusing Byeducationlevel High-schooleducationorless.............................. 34.6 thenoticetomakespecificcomparisonsamonginsti- Somecollege ............................................. 33.0 tutions, however (second panel of table 11). Further- Collegegraduateormore.................................. 21.3 Byconfidencelevelinunderstandingprivacynotices more,lessthan10percentwhosaidtheyhadnotused Veryconfident ............................................ 17.9 Somewhatconfident....................................... 30.6 thenoticesforcomparisonssaidtheyplannedtodoso Notveryconfident ........................................ 29.9 Notatallconfident ....................................... 34.1 in the future (data not in table). Because most consumers are generally satisfied with the privacy poli- Donotknowornoanswer .................................. 10.2 Byeducationlevel cies of their current institutions as they understand High-schooleducationorless.............................. 18.8 Somecollege ............................................. 5.0 them,infrequentuseforcomparisonsisnotinconsis- Collegegraduateormore.................................. 7.6 tent with stated importance of privacy policies. If Byconfidencelevelinunderstandingofprivacypolicies Veryconfident ............................................ 11.6 moreconsumerswereconcernedaboutthepoliciesof Somewhatconfident....................................... 10.0 Notveryconfident ........................................ 9.9 their own institutions, presumably more of them Notatallconfident ....................................... 10.2 would use the notices to seek replacement institutions. The findings indicate that consumers do not or less gave generalized answers that did not fit review closely the notices they receive, but that lack common definitions of affıliate, but only 21 percent of attention does not stem from a perception of their ofcollegegraduatesdidso.‘‘Donotknow’’responses inabilitytounderstandthem. weremuchmorecommonamongthosewithlessthan Finally,thesurveyaskedrespondentsthemeanings ahigh-schooleducation. of the terms affıliate and opt out, which are some- Littlerelationexistedbetweenanswerstothequestimes used in privacy notices or in discussions about tion about the meaning of affıliate and responses to financialprivacy.Whatwasevidentfromtheresponses the question about the usefulness of privacy notices. to these questions was that consumers assign various Cross-tabulation of the affıliate definition with the meanings to these two terms, especially to the term response on notice usefulness showed little variation affıliate. Slightly more than 40 percent thought the among groups of consumers (data not in table). termmeantaformalrelationshipthroughsomesortof Cross-tabulation of correct definition of the term jointownership(table12).Onepossibilityisthatthey affıliate with confidence in understanding the notices acquired this meaning by examining privacy notices, (last line of table 10) revealed a correlation, one butgeneralknowledgeandexperienceisalsoalikely slightlyhigherthantheonefromthecross-tabulation source. Educational level is also associated with of the term and response on usefulness. But the correctly defining the term. Slightly more than one- correlation was not as strong as the one between third of respondents with a high-school education or correct definition and education level. Thus, underless attached the meaning of a joint or combined standingoftheterminology,atleastatthetimeofthe ownershiprelationshiptoaffıliate,butmorethanhalf interview, was more dependent on educational level of college graduates did so. In contrast, more than than on specific attitudes expressed toward experione-third of those with only a high-school education enceswiththenotices.

Credit Card Disclosures A121 12. Accuracyofdefinitionsofselectedtermsinprivacy Itisnotpossibletodeterminefromtheanswersthe noticesandconfidencelevelsinunderstandingprivacy respondents gave whether their understanding of opt noticesbylevelofeducation,2004—Continued out was influenced in some way by the privacy notices they had received. Nonetheless, reasonable Optoutterm:Definitionandconfidencelevel Response understandingofthemeaningofthetermseemsfairly Definitionreferringtorequestingnosharingofinformation4 .. 75.2 widespread at the time of the interview. Here again, Byeducationlevel High-schooleducationorless.............................. 62.7 the correlation with education was greater than with Somecollege ............................................. 74.8 Collegegraduateormore.................................. 82.7 specific measures of attitudes toward usefulness of Otherdefinitions5 ........................................... 12.6 the notices. In future studies, another measurement Byeducationlevel may discover whether privacy notices and general High-schooleducationorless.............................. 12.6 Somecollege ............................................. 11.6 usageofthetermamongthepublic,includingitsuse Collegegraduateormore.................................. 13.2 in other areas like credit card solicitations, has had a Donotknowornoanswer .................................. 12.1 Byeducationlevel long-termeffectonunderstandingoftheterm. High-schooleducationorless.............................. 24.6 Somecollege ............................................. 13.5 Collegegraduateormore.................................. 4.1 Note: Affıliatemeansanycompanythatcontrols,iscontrolledby,orisun- CONCLUSION dercommoncontrolwithanothercompany.Optoutmeansadirectionbythe consumerthatyounotdisclosenonpublicpersonalinformationaboutthatcon- Surveys of consumers regarding their knowledge of, sumertoanonaffiliatedthirdpartyotherthanaspermitted.Definitionscome fromtheBoardofGovernorsoftheFederalReserveSystem(2000),‘‘Regula- attitudes toward, and use of various required disclotionP:PrivacyofConsumerFinancialInformation,’’FederalReserveRegulasures about consumer financial services indicate that toryService,6−2253and6−2300respectively(November). 1. Examplesareemployees;peoplewithinthecompany;theirownbanking many consumers are aware of the disclosures, have group;partoftheirfamilyofcompanies;partsoftheircompany;companiesin generally favorable attitudes toward them, and often thecorporation;anotherentityownedbytheparentcompany;intheirnetwork; same ownership; same corporation; branches; satellite banks; joint ventures; usethemforthepurposesenvisionedbytheiroriginal sister banks; brother banks; parent company; holding company; subsidiaries; companies they might own; off-shoot companies; organizations connected sponsors. That is not to say that the responses were with,relatedto,involvedwith,undertheumbrellaofthebank;andinstitutions not diverse, however; the responses probably always affiliated to or working with the bank to offer services (for example, loan department, mortgage company, credit card companies owned by the bank, will be as long as individuals have diverse backinsurancecompany,andsubcontractors). grounds, educations, experiences, and needs. Con- 2. Examplesarepeople,companies,organizations,banks,institutionsthey workwithordealwith,‘‘anybodytheydobusinesswith,’’investmentcompa- sumers who use their cards to generate credit tend to nies,stockbrokers,stockmarkets,andreciprocalmarketagreements. reviewmorefrequentlythedisclosuresofcreditcosts 3. Examplesareotherbanks,creditunions,lendinginstitutions,otherinstitutions,someoneintheirindustry,counterparts,creditcardcompanies,outside than those who mainly use their cards as convenient vendors, third-party companies, other organizations, other companies, other payment devices. This pattern also appears to hold businesses,competitors,marketers,companiesorbusinessestheywouldgive orsellyourinformationto,peopletheyselltheirliststo,institutionstheyswap true for the disclosures in credit card solicitations. loanswith,governmentagencies,anybusinessoftheirchoosing,and‘‘who- Thoseusingthecardsascredit-generatingdevicesare evertheywant.’’ 4. Exampleswererequestsinwritingthattheinstitutionnotdiscloseinfor- morelikelytoreviewinformationonannualpercentmationtoaffiliates,choicetoparticipate,optiontoparticipateornot,anddo age rates than those using cards primarily to make notwantthemtoshareinformationwithaffiliates. 5. Exampleswerehavethemnotdisclosefinancialstanding,geton‘‘donot payments. Consumers also indicate that institutional call’’list,getonoroffalist(notfurtherspecified),theycouldofferselected privacy policies are important to them, but they tend informationforrelease,youcangetoutofgivingtheminformation,choosenot bepartofthataffiliation,andbeleftoutordropped. nottoexaminethenoticestheyreceivecloselyforthe Source: SurveysofConsumers,May2004. mostpart.Thosewithmoreeducationaremorelikely Concerning the meaning of opt out in the context to understand the terms affıliate and opt out in the of privacy notices, about three-quarters of respon- context of a privacy notice. Overall, the survey dentsansweredcorrectlythatthetermmeantarequest responses suggest that the disclosures contribute to that information about them not be shared. As with theinformeduseofcreditbyconsumersandenhance affıliate, correctness of the response also varied by the competitiveness of consumer credit markets, as education. envisionedbythesponsorsofthedisclosurelaws.

A123 Higher-Priced Home Lending and the 2005 HMDA Data (Table 8 revised September 18, 2006) Robert B. Avery, Kenneth P. Brevoort, and Glenn B. mation has prompted widespread public discussion Canner, of the Division of Research and Statistics, about the fairness of mortgage lending decisions, as prepared this article. Caitlin G. Coslett and Sean M. the disclosures revealed wide disparities in the rates Wallaceprovidedresearchassistance. of approval of loan applications across racial and ethnic lines. The disclosures triggered debate about Since 1975, the Home Mortgage Disclosure Act the proper interpretation of the data and about the (HMDA)hasrequiredmostmortgagelendinginstitu- meaningfulnessofthedisparitiesinthedispositionof tions with offices in metropolitan areas to disclose to loanapplicationsandinlendingpatterns.1Thedisclothe public information about the geographic location sures also led many lenders to strengthen their fair and other characteristics of the home loans they lending compliance programs and to expand their originate or purchase during each calendar year. outreachtounderservedcommunities. Disclosure of home-lending activity is intended to Periodically, the Federal Reserve Board reviews help the public determine whether institutions are each of the regulations that it promulgates, including adequately serving their communities’ housing fi- RegulationC,whichimplementsHMDA.2Asaresult oftheBoard’smostrecentreviewofRegulationC,a nance needs, to facilitate enforcement of the nation’s number of important changes were made to the fair lending laws, and to guide public- and privatereporting requirements, changes that substantially sector investment activities. Although the act is increase the types and the amount of information intended to help achieve important public policy madeavailableabouthomelending(fordetails,refer goals, the law itself does not include mandates or totheappendix).3TheBoardstatedthattherevisions restrictions on lending—that is, it does not direct were intended to keep the regulation in step with lenders to make loans to particular areas or persons, recent developments in mortgage markets and with nordoesitdirectthemtomakecertainkindsofloans the revised standards of classification for the collecortorefrainfromcertainloantermsorpractices. tion of information on race and ethnicity as estab- Taken together, the nearly 8,850 lenders currently lished by the Office of Management and Budget covered by the law account for an estimated 80 per- (OMB).4 cent of home lending nationwide. Consequently, The2004HMDAdata,thefirsttoreflecttherecent HMDAdatalikelyprovidearepresentativepictureof revisionstoRegulationC,werereleasedtothepublic mosthomelendingintheUnitedStates.Theinformaby individual lending institutions in the spring of tion thus provided is rich, but it is limited: The data 2005.InSeptember2005,theFederalFinancialInstirevealagreatdealaboutwhatthelendingpatternsare tutions Examination Council (FFIEC) made publicly but relatively little about what causes the patterns. available various summary reports (statistical tables) Nonetheless, by drawing attention to these patterns, pertainingtoeachlenderandlendingactivityineach the data promote further analysis and discussion that candeepenunderstandingoftheircausesandencouragemarketplaceefficiencybyfosteringcompetition. 1.Refer,forexample,toJohnGoeringandRonWienk,eds.(1996), The Congress has amended HMDA on several MortgageLending,RacialDiscrimination,andFederalPolicy(Washington:UrbanInstitutePress). occasions to extend the reach of the law to more 2.RefertoHomeMortgageDisclosureAct(12U.S.C.§§2801-11), institutions and to expand the types of information RegulationC(12C.F.R.pt.203),andthestaffcommentaryaccompathatmustbedisclosed.Themostsweepinglegislative nyingRegulationC(12C.F.R.pt.203,Supp.I). 3.ThefinalrevisionstoRegulationCwereissuedonFebruary15, amendments occurred in 1989; they required the 2002,andJune27,2002. disclosure of application and loan-level information 4.Since2003,HMDAdatahaveusedthenewlyestablishedOMB for home loans, including the disposition of applica- standardsfordefiningmetropolitanandmicropolitanstatisticalareas. Refer to OMB (2000), ‘‘Standards for Defining Metropolitan and tionsandtheincome,sex,andraceorethnicityofthe MicropolitanStatisticalAreas,’’noticeofdecision,FederalRegister, individualsapplyingforcredit.Analysisofthisinfor- vol.65(December27),pp.82228–38.

A124 Federal Reserve Bulletinh2006 metropolitanstatisticalarea,alongwithacomprehen- (LTV) ratios, debt-to-income (DTI) ratios, and measive data file that included all the reported informa- suresofborrowercredithistory(forexample,acredit tion (except the dates of loan applications and of history score)—are not included in the HMDA data creditdecisions).5Atthattime,thestaffoftheFederal and, consequently, cannot be accounted for in an Reserve Board prepared the first comprehensive analysisofpricingdifferencesthatreliesonthesedata assessment of the expanded data, which was pub- alone. lishedasanarticleintheFederalReserveBulletin.6 Differencesinloan-pricingoutcomes,suchasthose The most important change made to Regulation C revealed in the HMDAdata, have increased concern is the requirement that lenders disclose pricing inforaboutthefairnessofthelendingprocess.Lendersare mationforloanswithpricesabovedesignatedthreshresponsibleforensuringcompliancewithfairlending olds;suchloansarereferredtohereas‘‘higher-priced laws,andtheexpandedHMDAdatamaybothencourloans.’’ The new pricing data allow a better underage and facilitate improved compliance efforts. The standing of lending activity in the higher-priced segregulatory agencies charged with enforcement of the ment of the mortgage market, a market segment that fair lending laws also use the expanded data to has grown substantially over the past decade or so in facilitateenforcementactivities. response to improvements in information processing This article reviews the 2005 HMDA data, which technology and in the ability of lenders to measure havejustbeenreleasedtothepublic.The2004article andpriceforcreditrisk. covered a wide range of topics, including ways in Greater understanding of the market and an improved ability to monitor the activities of individual which the expanded data might be used to aid fair lenders in the higher-priced market segment are lending enforcement, but this article is more limited: important because the expansion of such lending, Thefocushereisprimarilyontheloan-pricingaspects though affording some consumers greater access to ofthedata,includingthosethatpermitanassessment credit, has been accompanied by a variety of con- oftheeffectsofthechanginginterestratesituationin cerns. The concerns relate to the appropriateness of 2004 and 2005 on the disclosure of higher-priced loan terms and lending practices, constraints on con- lending.Toidentifytheeffectsonlendingpatternsof sumer shopping and on access to the full range of changing interest rates, the analysis presented here creditopportunities,thecompetitivenessofthehigher- uses adjusted sets of the 2004 and 2005 HMDAdata pricedmarket,andthepotentialforunequaltreatment in an attempt to distinguish the loans that exceeded of borrowers on the basis of race, ethnicity, or some the pricing thresholds solely because of a changed othercharacteristicprotectedbylaw. interest rate situation from other higher-priced loans. Areviewofthe2004HMDAdatafoundthat,inthe Thissectionoftheanalysisreliesonmonthlysurveys aggregate, less than one-fifth of borrowers took out of loan terms and pricing conducted by Freddie Mac higher-priced loans. However, the data also showed andtheFederalHousingFinanceBoardtohelpgauge that the incidence (measured as the proportion of theeffectsofchanginginterestratesovertheperiod. borrowers) of higher-priced lending varied substan- The analysis indicates that the substantial narrowtially across racial and ethnic lines: Blacks and His- ing of the difference between short- and long-term panicwhitesweremorelikely,andAsianslesslikely, interest rates in 2005 compared with 2004 not only to have received higher-priced loans than non- increased the overall share of reported loans that Hispanic whites. Information included in the HMDA exceededthepricingthresholdsestablishedbyRegudata on borrower or loan characteristics, such as lation C but also affected to some degree the gap in income and amount borrowed, was insufficient to loan-pricing outcomes among groups of borrowers account fully for the variation in loan pricing across sortedbytheirraceorethnicity. groups. Many factors routinely used by lenders to Theanalysisfurtherrevealsthatchangesininterest underwrite and price loans—including loan-to-value rates substantially affected the types and the proportions of loans that exceeded the price-reporting thresholds. Because of a combination of (1) the 5.IndividuallenderscoveredbyHMDAarerequiredtomaketheir owndataavailabletothepublicbeginningonMarch31oftheyear procedure specified in Regulation C for determining after the calendar year for which the data apply. However, the data which loans are higher priced and (2) the rules madeavailableatthattimehavenotbeensystematicallycheckedby governing how annual percentage rates (APRs) are thesupervisoryagenciesforerrorsoromissions,ashavetheHMDA datareleasedbytheFFIECinSeptembereachyear. calculated for adjustable-rate loans, adjustable-rate 6.RefertoRobertB.Avery,GlennB.Canner,andRobertE.Cook loans were much more likely than fixed-rate loans (2005),‘‘NewInformationReportedunderHMDAandItsApplication with similar risk profiles to be below the HMDA in Fair Lending Enforcement,’’ Federal Reserve Bulletin, vol. 91 (Summer),pp.344–94. price-reporting thresholds in 2004 but were about as

Higher-Priced Home Lending and the 2005 HMDA Data A125 likelyasfixed-rateloanstobeabovethethresholdby ‘‘near prime.’’ Individuals in the subprime category the end of 2005. One consequence of this changed typically pay the highest prices because they pose relationshipisthatcertainpopulations—suchasthose greater credit or prepayment risk or are otherwise residinginthewesternpartofthecountry—thatused more costly to serve. In practice, the dividing line adjustable-rateloansrelativelymoreoftenthanfixed- betweenthesetwo‘‘nonprime’’marketsegmentscan rateloanslikelywitnessedarelativelylargerincrease besomewhatamorphous,ascanthelinebetweenthe inreportedhigher-pricedlendingin2005. prime and nonprime markets. Moreover, the thresholds that separate these market segments can change asmarketinterestratesmove,aslenders’appetitesfor LOAN PRICING IN THE MORTGAGE MARKET interest rate or credit risk change, and as technologi- Over the past decade or so, the mortgage market has calimprovementsallowformore-preciseriskassesschanged markedly. Before that, mortgage lenders ment. offered consumers a relatively limited array of loan Estimates of the annual volume of nonprime lendproductsatprices(interestrates,points,andfees)that ing vary, but all sources agree that the nonprime variednotbythecreditworthinessoftheborrowerbut market segment has grown substantially in recent by loan type (for example, conventional or years. One source estimates that from 1994 to 2005, government-backed), loan characteristic (for ex- the dollar volume of subprime loans increased from ample, amount borrowed, term to maturity, or LTV about $35 billion to more than $600 billion. Further, ratio), type of structure securing the loan (for ex- subprime lending is no longer a minor portion of the ample, traditional ‘‘site built’’ home or factory- mortgage market. Subprime loans are estimated to manufactured unit), and ownership status (owner- have accounted for 20 percent of all mortgage origioccupied or nonowner-occupied). Effectively, nationsin2005,upfromlessthan5percentin1994.9 borrowerseitherdidordidnotmeettheunderwriting criteria for a particular product. Those who met the Concerns about Loan Pricing criteria paid about the same price; those who did not weredeniedcredit. As price flexibility has emerged in the mortgage Advances in technology, better access to informa- market,sohaveconcernsaboutthefairnessofpricing tion on the credit histories of individuals, increased outcomes. Such considerations generally fall into competition,andthematurationofarobustsecondary three broad categories: In the first category are conmarket for loans representing the full spectrum of cernsaboutpossiblediscriminationbasedontherace credit risks have helped spur remarkable changes in or ethnicity of the borrower. These concerns are themortgagemarket.Mostprominenthasbeencredit heightened because, for some loans, prices are deterpricingbasedexplicitlyonrisk.Today,muchmoreso mined on an individual basis and not strictly accordthaninthepast,differencesinthecreditworthinessof ing to credit risk, cost factors, or competitive condidifferentborrowerscanleadtodifferentpricesforthe tions. same product.7Applicants who are less creditworthy Inthesecondcategoryareconcernsaboutwhether or who are unwilling or unable to document their borrowers in the higher-priced segment of the loan creditworthiness or income are increasingly less market have sufficient resources (for example, time, likely to be turned down for a loan; rather, they are information, and financial experience) to shop effecoffered credit at higher prices.8 Explicit risk-based tively for the loan terms most appropriate to their pricing has expanded opportunities for homeowner- circumstances. These concerns relate to both borshipandhasallowedindividuals,includingthosewho rower and lender behavior. For example, some borare otherwise credit constrained, to more readily rowers may not shop or negotiate for the best availpurchase homes or to borrow against the equity they able rates and terms because they need funds haveaccumulatedintheirhomes. immediatelyandarefocusedprimarilyontheamount Borrowers in the higher-priced market generally theycanborrowandthesizeofthemonthlypayment, fallintooneoftwomarketsegments—‘‘subprime’’or not on the interest rate, fees, or other loan features. 7. Refer to Souphala Chomsisengphet and Anthony Pennington- 9. Estimates pertain to mortgages backed by one- to four-family Cross (2006), ‘‘The Evolution of the Subprime Mortgage Market,’’ homes. Estimates are based on information from Inside Mortgage Federal Reserve Bank of St. Louis, Review, vol. 88 (January/ Finance Publications (2005 and earlier years), Mortgage Market February),pp.31–56. Statistical Annual (Bethesda, Md.: IMFP), www.imfpubs.com; and 8.Refer,forexample,toDarrylE.Getter(2006),‘‘ConsumerCredit on information from LoanPerformance, www.loanperformance.com, Risk and Pricing,’’Journal of ConsumerAffairs, vol. 40 (Summer), a subsidiary of First American Real Estate Solutions, pp.41–63. www.firstamres.com/jsp/index.jsp.

A126 Federal Reserve Bulletinh2006 And some lenders may engage in aggressive ‘‘push’’ and the rate on Treasury securities of comparable marketingthatmayconfuseborrowersaboutthecost maturity. The thresholds for reporting differ by lien andtermsofloans. status: 3 percentage points for first liens and 5 per- Finally, concerns have been raised about whether centage points for junior, or subordinate, liens.12The competition is adequate to ensure that borrowers in different thresholds for first and junior liens are the higher-priced segment of the loan market are intended to reflect differences in the credit risk and provided with the full range of credit opportunities. other features of the loans in these two different Some believe that prime-market lenders are not markets.Tobetterinterpretthereportedpricinginforpresent, or do not offer or promote their prime mation, the Board has also required institutions to products sufficiently, in certain geographic markets, reportthelienstatusforeachloan. including neighborhoods that have larger minority In limiting the reporting of price information to populations. In this view, limited access to prime only the higher-priced segment of the market, the lenders and the products they offer diminishes the Board weighed the costs and benefits of moreopportunities for borrowers in affected communities expansive data collection and reporting and deterto obtain lower-priced loans. These concerns are minednottoadoptmore-expansivereportingrequireextraordinarilycomplexandbeyondthescopeofthis ments. The Board also chose to refer to loans with article. The Federal Reserve Board’s recent hearings pricesthatexceedthereportingthresholdas‘‘higheronhomeequitylendingsoughttocollectmoreinfor- priced loans’’ rather than as ‘‘subprime loans.’’ The mation about these and other concerns raised by the correspondence between subprime loans and loans rapid growth of the higher-priced segment of the with prices exceeding the threshold is not precise. market.10 TheBoard’sregulationsetstheprice-reportingthresholds in such a way that the number or proportion of Determining What Pricing Information Is loansreportedashigherpricedcanvaryfromyearto Reported year even if the size and the share of the subprime markethavestayedthesame. In2002,theFederalReserveBoardamendedRegulationCtorequirethedisclosureofpricinginformation Reasons for Loan-Price Variation for higher-priced loans. In establishing the loanpricing disclosure rule, the Board sought to select Mortgage pricing is complex and reflects a wide thresholds that would limit regulatory burdens by range of factors. Many of these factors are easily focusing data reporting on only those loans in the quantifiable and objectively measured. Some, howhigher-pricedsegmentofthemarket.11 ever, are less readily quantified—for example, the Specifically, for loans with spreads above desig- extent of negotiations, if any, between lender and nated thresholds, revised Regulation C requires the borrower. The expanded HMDAdata include few of reporting of the spread between the APR on a loan the factors that may help explain variations in the prices of reported loans. Even if all of the readily quantifiable factors were included in the data, they 10. For more information about the hearings, refer to Board of would not necessarily fully explain loan pricing Governors of the Federal Reserve System (2006), ‘‘Board to Hold Four Public Hearings on the Home Equity Lending Market,’’press becausesomefactorsaredifficulttomeasure. release,May1,www.federalreserve.gov/boarddocs/press/bcreg/2006. Important factors not included in the HMDAdata 11.WhentheBoardamendedHMDAtoexpanddatareporting,it include the costs of raising the funds to be lent; alsoestablishedtransitionrulesforcompliancewithRegulationC.The transition rules provide that for loans with application dates before considerations related to credit risk, such as those January 1, 2004, lenders need not report pricing information.As a reflected in the borrower’s credit history, LTV ratio, consequenceofthetransitionrules,someindeterminateproportionsof orDTIratio;prepaymentrisk(theriskthataloanwill higher-pricedloansarereportedwiththesamecodeasloansthatdid not meet the threshold requirements. The inability to distinguish higher-pricedloansfromothersthatwereoriginatedin2004and2005 butwithapplicationdatesbeforeJanuary1,2004,meansthatusersof 12.Incalculatingtheratespread,thelenderusestheTreasuryyield thedataneedtotakethislimitationintoaccountwhenassessingthe for securities of a comparable maturity as of the fifteenth day of a data.Theeffectsofthetransitionruleweresignificantforassessments givenmonthdependingonwhentheinterestratewassetontheloan. ofthe2004databutareofmuchlessimportanceforanalysisofthe Forsuchacalculation,theruledirectslenderstousethefifteenthday 2005 data. Nonetheless, to identify which applications had dates ofagivenmonthforanyloanonwhichtheinterestratewassetonor before January 1, 2004, the FFIEC added a flag to the 2005 ‘‘loan/ after that day through the fourteenth day of the next month. The applicationregister’’(LAR)dataitmakesavailabletothepublic.The relevant date to use is the date the interest rate on the loan was LARisaregisterthatispreparedannuallybyeachlendercoveredby determined,whichisoften,butnotalways,setpursuanttoalock-in HMDAand that includes data on each of the items reported under agreementbetweentheborrowerandthelender.TheAPRusedinthe HMDA.Fortheanalysisofloanpricingthatfollowshere,weexclude calculationsistheonecalculatedanddisclosedtotheconsumerunder allloanswithapplicationdatesbeforeJanuary1,2004. section226.18ofRegulationZ(12C.F.R.pt.226).

Higher-Priced Home Lending and the 2005 HMDA Data A127 be prepaid before the term of the loan); overhead lowerflow-throughrate—thatis,thenumberofappliexpenses, such as those related to providing offices cations processed to successfully extend a loan may and to compensating staff for finding prospective be higher for such lenders than for lenders that deal borrowers and underwriting loans; loan-servicing primarilywithborrowerswithfewcreditproblemsor costs;andpossiblytheextentofnegotiationsbetween withtheabilitytomakelargedownpayments. creditorandborrower.Marketconditionsandcompetition also bear on pricing, as local economic Discretionary, or Flexible, Pricing conditions—including, importantly, those of local housingmarkets—caninfluencethedemandandsup- Some creditors provide their loan officers and agents ply of credit.13 Finally, the legal situation in a state, working on their behalf (for example, mortgage broincluding foreclosure rules, may affect loan pricing kers or loan correspondents) with rate sheets that by constraining to a greater or lesser degree the indicatethecreditors’baselinepricesbyloanproduct ability of lenders to recover and dispose of the (for example, conventional loans of various types), collateralusedtobackloansthatareindefault. owner-occupancy status, loan characteristic (for ex- Mortgages are typically priced at a spread above ample, amount of loan, prepayment penalty option, the yields on Treasury securities or on other, similar term to maturity, or LTV ratio), and borrower creditinstruments or indexes of funding costs that corre- worthiness (as reflected in, for example, a credit spondtothetimealoanisexpectedtobeoutstanding. historyscoreorDTIratio). Each of the factors noted earlier may influence the Rate sheets vary across lenders. For some lenders, magnitudeofthespread.Elevatedcreditriskforloans therateonthesheetisa‘‘sticker’’price;forothers,it in the higher-priced mortgage market results in sub- istheminimumacceptedprice;andforstillothers,it stantially higher default and foreclosure rates and is the actual target price. Some lenders have a single costs and, consequently, in higher price levels. Pre- rate sheet for the entire organization (for each loan payment risk is also greater for higher-priced loans product);othershavedifferentratesheetsfordifferent not only because borrowers in the higher-priced geographicmarketsthatreflectlocalmarketcompetimarket have an incentive to refinance when interest tion and costs. Rate sheets can change daily with ratesfall(asdoborrowersinthelower-pricedmarket changesinbasiceconomicconditions,suchasmarket segment) but also because they have an incentive to interestrates. prepaywhentheircredithistoryimprovestothepoint Loanratespaidbyborrowerscandeviatefromthe that they qualify for lower-priced credit.14 Because interest rates shown on sheets for many reasons. For creditandprepaymentrisksarehigherforloansinthe example, the rates on the sheets may not reflect higher-priced segment of the market, such risks tend differences in loan origination costs. Also, in some tovarymoreinthismarketsegment. cases,loanofficersandbrokersareallowedtodeviate Lenders active in the higher-priced market may from prices on rate sheets as market conditions, also face a cost structure different from that faced by includingtheextentofcompetition,warrantorallow. lenders focused on the lower-priced segment of the Deviations may also occur because of negotiated market.Lendersfocusedonthehigher-pricedmarket outcomes. Loan officers or brokers may benefit from segment may face steeper funding costs, may incur pricing flexibility through higher compensation by higher marketing expenses, and may have a much obtainingapriceabovetheratestatedonaratesheet (orabovepricesobtainedbyothers). Borrowers differ in their propensity to negotiate— 13. For example, in areas that have experienced sustained rapid increases in home prices, more prospective borrowers may rely on for example, borrowers with less experience in the mortgage products intended to minimize initial monthly payment mortgage market, such as first-time homebuyers, burdens,suchasadjustable-rateloans.Also,differencesinprepayment may be less likely than experienced borrowers to propensitiesmayresultinpricingdifferencesacrossstates. 14. Refer, for example, to Office of Thrift Supervision, Office of negotiate. These differences in negotiating propensi- ResearchandAnalysis(2000),‘‘WhataboutSubprimeMortgages?’’ ties may be correlated with race, ethnicity, or sex. Mortgage Market Trends, vol. 4 (June), pp.1–13. Borrowers with For example, minorities are disproportionately firsthigher-pricedloansmayalsoprepaymorefrequentlythanborrowers with other loans if they have a greater propensity to extract equity time homebuyers. throughacash-outrefinance.Suchmaybethecaseifborrowerswith Discretionary, or flexible, pricing may be a legitihigher-pricedloanshavefeweralternativesourcesoffundstoaddress mate business practice. Properly developed, monipressingfinancialproblems.Also,borrowerswithhigher-pricedloans may prepay more often if, over time, they become more aware of tored, and administered, discretionary pricing proless-expensive credit opportunities. Refer to Anthony Pennington- grams may help to ensure that markets allocate Cross (2003), ‘‘Credit History and the Performance of Prime and resources in an efficient way. However, when loan Nonprime Mortgages,’’ Journal of Real Estate Finance, vol. 27 (November),pp.279–301. officers have latitude in deviating from rate sheets or

A128 Federal Reserve Bulletinh2006 in determining which rate sheet applies to each toryinstitutionmayofferabroaderrangeofmortgage borrower, the lender runs the risk that differential products through its retail branch network. If morttreatmentonabasisprohibitedbylawmayarise.For gagebrokersorloancorrespondentsthatfocusonthe this reason, the Interagency Fair Lending Examina- subprime market tend to work disproportionately tion Procedures provide that discretionary pricing with borrowers from minority neighborhoods, then should be considered an examination ‘‘risk factor’’ the depository institution’s overall pricing pattern whenalender’sriskforengaginginpricingdiscrimi- mayshowahigherincidenceofhigher-pricedlending nationisevaluated.15 forminoritiesthanforwhites. Variations in Loan-Processing Channels GENERAL FINDINGS FROM THE 2005 HMDA DATA The delivery channels through which borrowers obtain loans vary across lenders, and such variation For 2005, lenders covered by HMDAreported informay affect loan pricing. On the one hand, underwrit- mation on roughly 30.2 million home-loan applicaing and pricing may be centrally controlled even tions—11.7 million for purchasing one- to fourthough applications for credit may begin through family homes, 15.9 million for refinancing existing differentchannels,suchastheInternet,themail,ora home loans, 2.5 million for improving one- to fourvisit to a bank office. On the other hand, in complex familydwellings,andthebalanceforloansonmultifinancialorganizationswithnumerousbankbranches, familydwellingsforfiveormorefamilies(table1).17 multiple affiliates (both bank and nonbank), decen- Theseapplicationsresultedinsome15.6millionloan tralized loan production offices, and third-party bro- extensions. Lenders also reported information on kerageoperations,eachapplicationmaybesubjectto about 5.9 million loans they purchased from other a different underwriting and pricing regime depend- institutions and on some 397,000 requests for preingonitspointofinitiation. approvals of home-purchase loans that either were The 2004 HMDA pricing data suggested that the turned down by the lender at the time the predelivery channel through which a borrower obtains a approvalwassoughtorweregrantedbutnotactedon loan may matter. For example, the incidence of by the applicant (data not shown in table). The total higher-priced lending was significantly higher for number of reported applications and purchased loans borrowers who lived outside the assessment areas of increased about 2.8 million, or 7 percent, from 2004; lenderscoveredbytheCommunityReinvestmentAct most of the increase was for applications for homeof 1977 (CRA) than for those who lived inside these purchaseloans.Thenumberofapplicationsforloans areas.16The HMDAdata do not provide a reason for to refinance an existing loan fell about 1 percent, this pattern, but several explanations that warrant likelybecauseofanincreaseininterestratesin2005. furtherresearcharepossible.Forexample,thediffer- From the 2005 HMDA data, the FFIEC prepared ence may be due, at least in part, to a reliance on disclosure statements for 8,848 HMDA-reporting different delivery channels for loans within and out- institutions—3,904 commercial banks, 974 savings sidetheselenders’assessmentareas. institutions, 2,047 credit unions, and 1,923 mortgage Differences in the incidence of higher-priced lend- companies (table 2). Of the mortgage companies, ingacrossgroupsmayalsoariseifdifferentchannels 70percentwereindependententities—thatis,institutendtoservedifferentcustomergroups.Forexample, tions that were neither subsidiaries of depository mortgage brokers or loan correspondents that origi- institutions nor affiliates of bank holding companies nateloansonbehalfofadepositoryinstitution(com- (data derived from table). The disclosure statements mercial bank, savings association, or credit union) consistedof78,193distinctreports,eachcoveringthe mayfocusonthesubprimemarket,whilethedeposi- lending activity of a particular institution in each metropolitan statistical area (MSA) in which it had a homeorbranchoffice(table1,lastcolumn).Thetotal 15.Refertowww.ffiec.gov/PDF/fairlend.pdf. 16.The assessment areas of lenders covered by the CRAinclude principallythelocalesinwhichalenderhasitsmainorbranchoffices 17. In recent years, many lending institutions have developed anditsdeposit-takingautomatedtellermachines.Foramorecomplete programs to respond to prospective homebuyers’ need to provide definition of CRA assessment areas, refer to the Federal Reserve sellerswithevidencethattheyarelikelytoqualifyforfinancingoncea Board’sRegulationBB,section228.41.AlsorefertoRobertB.Avery, contractforsalehasbeensigned.Suchprogramsreviewrequestsfor Glenn B. Canner, Shannon C. Mok, and Dan S. Sokolov (2005), pre-approvalsofhome-purchaseloansandtypicallyprovideaprospec- ‘‘Community Banks and Rural Development: Research Relating to tive homebuyer with a binding written commitment to finance a ProposalstoRevisetheRegulationsThatImplementtheCommunity purchase(subjecttocertainconditions).Theapplicationcountsshown Reinvestment Act,’’ Federal Reserve Bulletin, vol. 91 (Spring), intable1excludeinformationreportedonpre-approvalsthatdidnot pp.202–35. resultinaloan.

Higher-Priced Home Lending and the 2005 HMDA Data A129 1. HomeloanandreportingactivityofhomelenderscoveredunderHMDA,1990–2005 Number Applicationsreceivedforhomeloansonone-tofour-family properties,andhomeloanspurchasedfromotherlenders(millions) Disclosure Year Applications Reporters Loans reports2 Total1 Home Home purchased Refinance Total1 purchase improvement 1990............ 3.27 1.07 1.16 5.51 1.15 6.66 9,332 24,041 1991............ 3.26 2.11 1.18 6.55 1.36 7.91 9,358 25,934 1992............ 3.54 5.24 1.23 10.01 1.98 12.00 9,073 28,782 1993............ 4.52 7.72 1.40 13.64 1.80 15.44 9,650 35,976 1994............ 5.20 3.80 1.69 10.69 1.48 12.17 9,858 38,750 1995............ 5.51 2.70 1.75 9.96 1.28 11.24 9,539 36,611 1996............ 6.33 4.54 2.14 13.01 1.82 14.83 9,328 42,946 1997............ 6.75 5.39 2.16 14.30 2.08 16.38 7,925 47,416 1998............ 7.96 11.42 2.04 21.43 3.23 24.65 7,836 57,294 1999............ 8.43 9.37 2.05 19.85 3.01 22.86 7,832 56,966 2000............ 8.28 6.54 1.99 16.81 2.40 19.21 7,713 52,776 2001............ 7.69 14.29 1.85 23.83 3.77 27.59 7,631 53,066 2002............ 7.40 17.48 1.53 26.41 4.83 31.24 7,771 56,506 2003............ 8.15 24.60 1.51 34.26 7.23 41.49 8,121 65,808 2004............ 9.79 16.10 2.20 28.13 5.14 33.27 8,853 72,246 2005............ 11.67 15.90 2.54 30.17 5.87 36.04 8,848 78,193 Note: Hereandinallsubsequenttablesexcepttables3and5,for2004and 2. Areportcoversthemortgagelendingactivityofalenderinasinglemetro- 2005,applicationsexcluderequestsforpre-approvalthatweredeniedbythe politanstatisticalareainwhichithadanofficeduringtheyear. lenderorwereacceptedbythelenderbutnotacteduponbytheborrower.In Source: Hereandinsubsequenttablesandfiguresexceptasnoted,Federal thisarticle,applicationsaredefinedasbeingforaloanonaspecificproperty; Financial Institutions Examination Council, data reported under the Home theyarethusdistinctfromrequestsforpre-approval,whicharenotrelatedtoa MortgageDisclosureAct(www.ffiec.gov/hmda). specificproperty. 1. Applicationsformultifamilyhomesareincludedonlyinthe‘‘total’’columns;for2005,theseapplicationsnumberednearly57,700. number of reporting institutions was little changed tended to be very active lenders: The 576 mortgage from2004,aswasthedistributionofreportersbytype company affiliates processed 24 percent of the appliofinstitution. cations in 2005. Different types of lending institutions tend to Lender Specialization specialize in different types of home loans, although less so than in the past. The most notable change Mortgage companies, as distinct from depository has been the diminished role that mortgage compainstitutions, received more than 60 percent of all the nies play in originating government-backed loans. home-loan applications reported in the 2005 HMDA In 2005, mortgage companies accounted for nearly data, although such companies accounted for only 64 percent of government-backed originations. As about one-fifth of the reporting institutions (table 3). recently as 2002, their share of originations of this Among mortgage companies, those affiliated (either type had been 83 percent. Depository institutions directly or indirectly) with a depository institution extended 71 percent of reported home-improvement loans and about 88 percent of multifamily loans 2. DistributionofhomelenderscoveredbyHMDA, (data not shown in tables). Commercial banks acbytypeofinstitution,2005 counted for about half the loans for manufactured homes in 2005. Type Number Percent Depositoryinstitution Activity and Size of Lender Commercialbank ........... 3,904 44.1 Savingsinstitution .......... 974 11.0 Creditunion ................ 2,047 23.1 Although the number of lending institutions covered All ......................... 6,925 78.2 byHMDAislarge,mostoftheseinstitutions,whether Mortgagecompany Independent ................ 1,347 15.2 measuredbyassetsizeorbysomemeasureoflending Affiliated1 .................. 576 6.5 All ......................... 1,923 21.7 activity (such as the number of reported applications orloans),aresmall(table3).For2005,60percentof Allinstitutions ............... 8,848 100 the reporting institutions each provided information 1. Subsidiary of a depository institution or an affiliate of a bank holding company. onfewerthan250loansorapplications,accountingfor

A130 Federal Reserve Bulletinh2006 3. DistributionofhomelenderscoveredbyHMDA,bytypeoflenderandthenumberofapplicationstheyreceive,2005 Numberofapplications Typeoflender, andsubcategory 1–99 100–249 250–999 (assetsizeinmillionsof dollars,oraffiliation) Percentof Percentof Percentof Percentof Percent Percentof type1 subcategory2 type1 subcategory2 oftype1 subcategory2 Depositoryinstitution Commercialbank Lessthan250 .......................... 78.8 58.7 66.0 30.4 27.4 10.1 250–999 ............................... 17.9 25.0 29.8 25.8 60.6 41.7 1,000ormore .......................... 3.3 12.6 4.1 9.6 12.1 22.4 All..................................... 100 41.8 100 26.6 100 21.2 Savingsinstitution Lessthan250 .......................... 84.6 40.4 70.3 38.0 25.2 18.4 250–999 ............................... 11.8 7.2 27.3 18.9 65.1 61.3 1,000ormore .......................... 3.6 5.2 2.4 3.9 9.8 21.4 All..................................... 100 22.6 100 25.6 100 34.7 Creditunion Lessthan250 .......................... 96.1 63.6 84.4 26.7 34.8 9.5 250–999 ............................... 3.8 9.1 14.6 16.7 58.7 58.0 1,000ormore .......................... .1 .9 1.0 4.7 6.5 25.2 All..................................... 100 49.2 100 23.5 100 20.2 Alldepositoryinstitutions Lessthan250 .......................... 85.2 58.5 71.6 29.9 28.8 10.8 250–999 ............................... 12.6 18.4 25.3 22.6 61.0 48.7 1,000ormore .......................... 2.2 9.2 3.1 7.6 10.1 22.6 All..................................... 100 42.0 100 25.6 100 22.8 Mortgagecompany Independent .............................. 41.2 11.4 73.3 13.1 79.4 28.4 Affiliated ................................. 58.8 38.2 26.7 11.1 20.6 17.2 All ....................................... 100 19.5 100 12.5 100 25.0 Allinstitutions .............................. ... 37.1 ... 22.7 ... 23.3 Memo Allapplications,bynumberreported bylender............................... ... .5 ... 1.1 ... 3.4 Note: Refertotable2,note1,andgeneralnotetotable1. 2. Distributionsumshorizontally. 1. Distributionsumsvertically. ... Notapplicable. just 1.6 percent of all the reported data.At the other DISPOSITION OF APPLICATIONS, SELECTED endofthespectrum,6percentofreportinginstitutions CATEGORIES OF LOAN PRODUCTS, AND eachprovidedinformationon5,000ormoreloansor THE SECONDARY MARKET applications, but these few highly active lenders ac- The HMDAdata provide opportunities to categorize countedfor88percentofallthereporteddata. applicationsandloansinawidevarietyofways.For Asset size is available only for depository instituthe analysis here, applications were grouped into tions. Asset size and lending activity are highly twenty-five product categories based on loan and correlated. For example, the 707 depository instituproperty type, purpose of the loan, and lien and tions with assets of $1 billion or more reported owner-occupancystatus.18Foreachproductcategory, 86 percent of all applications reported by depositoinformation is provided on the number of total and ries, whereas the 4,236 HMDA-reporting depository pre-approval applications, application denials, origiinstitutions with assets of less than $250 million nated loans, loans with prices above the thresholds, accountedforonlyabout5percentoftheapplications loans covered by the Home Ownership and Equity (percentagesderivedfromtable3). Protection Act of 1994, and the mean and median Many HMDA reporters are affiliated with each APR spreads for loans priced above certain threshother.IfindividualHMDAreportersareaggregatedto olds(table4). theirhighestlevelofcorporateorganization(suchasa Becausethetransitionrulesregardingthereporting bank holding company), the concentration of mortof data create problems for assessing some of the gage lending nationwide is evident. The twenty-five organizationsreportingthelargestnumberofapplications and loans accounted for 54 percent of the 2005 18.Applicationsinwhichthelenderreportedthattherace,ethnicity, or sex of the applicant or co-applicant was ‘‘not applicable’’were data, a proportion essentially unchanged from 2004 assumed to have been made by businesses (including trusts) rather (datanotshownintables). thanbyindividuals.

Higher-Priced Home Lending and the 2005 HMDA Data A131 3.DistributionofhomelenderscoveredbyHMDA,bytypeoflenderandthenumberofapplicationstheyreceive,2005—Continued Numberofapplications Memo Typeoflender, andsubcategory 1,000–4,999 5,000ormore Any (assetsizeinmillionsof Numberof Percentof dollars,oraffiliation) Percentof Percentof Percentof Percentof Percentof Percentof lenders applications type1 subcategory2 type1 subcategory2 type1 subcategory Depositoryinstitution Commercialbank Lessthan250 .......................... 6.2 .8 1.2 0 57.7 100 2,254 1.0 250–999 ............................... 32.4 7.4 2.5 .2 30.8 100 1,204 1.6 1,000ormore .......................... 61.5 37.9 96.3 17.5 11.4 100 446 19.0 All..................................... 100 7.0 100 2.1 100 100 3,904 21.6 Savingsinstitution Lessthan250 .......................... 12.0 2.8 3.4 .4 47.3 100 461 .5 250–999 ............................... 38.9 11.7 5.1 .8 36.9 100 359 .7 1,000ormore .......................... 49.1 34.4 91.5 35.1 15.8 100 154 11.8 All..................................... 100 11.1 100 6.1 100 100 974 12.9 Creditunion Lessthan250 .......................... 3.0 .3 0 0 74.3 100 1,521 .5 250–999 ............................... 50.4 16.2 0 0 20.5 100 419 .8 1,000ormore .......................... 46.7 58.9 100 10.3 5.2 100 107 1.1 All..................................... 100 6.6 100 .5 100 100 2,047 2.5 Alldepositoryinstitutions Lessthan250 .......................... 6.6 .8 2.0 .1 61.2 100 4,236 2.0 250–999 ............................... 38.4 10.0 3.3 .3 28.6 100 1,982 3.1 1,000ormore .......................... 55.0 40.3 94.7 20.2 10.2 100 707 31.9 All..................................... 100 7.5 100 2.2 100 100 6,925 37.0 Mortgagecompany Independent .............................. 82.5 30.0 68.3 17.2 70.1 100 1,347 39.5 Affiliated ................................. 17.6 14.9 31.7 18.6 30.0 100 576 23.5 All ....................................... 100 25.5 100 17.6 100 100 1,923 63.0 Allinstitutions .............................. ... 11.4 ... 5.5 ... 100 8,848 100 Memo Allapplications,bynumberreported bylender............................... ... 7.1 ... 88.0 ... 100 ... 100 2004 and 2005 data on loan pricing, as they do for prospectiveborrowersgothroughbeforepurchasinga assessing the data on manufactured homes and pre- home(table4).Denialratesarelowerforgovernmentapprovals,theanalysisthatfollowsexcludes‘‘transi- backed loans than for conventional loans and are tion’’ applications—that is, those submitted before especially high for loans to purchase manufactured January1,2004(dataontheseapplicationsareshown homes. Requests for pre-approval are denied at a asmemoitemsintables4and5).Otherwise,informa- higher rate than applications initiated through a pretion is given on all applications reported under approvalprogram(table5). HMDA. Comparedwithdenialratesin2004,thosein2005 are slightly higher for conventional home-purchase Disposition of Applications and refinance loans and are either unchanged or slightly lower for other loan products. Overall, the HMDAdataaretheonlypubliclyavailablesourceof denial rate for all loans in 2005 was 27.1 percent, information on the disposition of individual applicacomparedwith26.5percentin2004. tions for home loans. Because the data include informationontherace,ethnicity,andsexofapplicantsas Conventional and Government-Backed Loans well as the type and purpose of the loan and the location of the property, the disposition of applica- As in 2004, most applications (about 95 percent in tionscanbeassessedalongmanydimensions. 2005) for loans to purchase owner-occupied one- to The HMDA data for 2005 indicate that lenders four-familyhomes(eithersite-builtormanufactured) approve most of the applications they receive, al- were for conventional loans—that is, nonthoughtheproportionapprovedordeniedvariessome- government-backed loans (table 4). The remainder whatbyloanpurposeandproductandbylienstatus.In were for government-backed forms of credit, mostly general, denial rates are notably higher for refinanc- those involving the Federal Housing Administration ingsandforhome-improvementloansthanforhome- (FHA). purchase loans, perhaps because of the prequalifica- The share of all HMDA-reported loans backed by tion and financial counseling activities that many theFHAhasbeendecliningoverthepastseveralyears,

A132 Federal Reserve Bulletinh2006 4. Dispositionofapplicationsforhomeloans,andoriginationandpricingofloans,bytypeofhomeandtypeofloan,2005 Applications Loansoriginated Loanswithannualpercentagerate(APR) spreadabovethethreshold1 Acteduponbylender Distribution, Typeofhomeandloan bypercentagepoints Number Number ofAPRspread submitted Number Percent Number Percent Number 3–3.99 4–4.99 denied denied ONE-TOFOUR-FAMILY Nonbusinessrelated3 Owneroccupied Sitebuilt Homepurchase Conventional Firstlien ...................... 6,838,946 5,922,478 969,271 16.4 4,399,445 1,080,344 24.6 27.0 35.4 Juniorlien .................... 1,930,805 1,701,237 304,874 17.9 1,215,902 604,924 49.8 ... ... Governmentbacked Firstlien ...................... 554,607 494,785 61,859 12.5 408,618 3,654 .9 76.3 13.0 Juniorlien .................... 1,157 941 106 11.3 789 29 3.7 ... ... Refinance Conventional Firstlien ...................... 12,752,498 9,637,488 3,176,225 33.0 5,518,481 1,418,459 25.7 27.4 31.7 Juniorlien .................... 1,449,919 1,205,491 359,090 29.8 720,380 217,570 30.2 ... ... Governmentbacked Firstlien ...................... 247,768 212,745 42,752 20.1 150,000 1,349 .9 42.8 41.2 Juniorlien .................... 433 331 50 15.1 257 24 9.3 ... ... Homeimprovement Conventional Firstlien ...................... 932,159 712,434 252,675 35.5 399,723 104,930 26.3 34.6 29.2 Juniorlien .................... 1,090,972 954,402 400,022 41.9 461,296 82,013 17.8 ... ... Governmentbacked Firstlien ...................... 3,547 3,082 768 24.9 2,003 110 5.5 52.7 13.6 Juniorlien .................... 3,440 2,972 753 25.3 1,867 1,116 59.8 ... ... Conventionalorgovernmentbacked,unsecured ........... 325,391 315,102 149,744 47.5 143,716 ... ... ... ... Manufactured Conventional,firstlien Homepurchase .................. 386,286 367,166 193,285 52.6 99,964 58,304 58.3 26.8 24.7 Refinance........................ 233,159 190,832 103,360 54.2 69,807 38,482 55.1 30.0 30.0 Other .............................. 131,221 119,064 48,584 40.8 60,264 12,957 21.5 17.2 18.0 Nonowneroccupied4 Conventional,firstlien Homepurchase .................. 1,548,496 1,361,256 241,699 17.8 1,010,518 205,020 20.3 41.5 27.5 Refinance........................ 1,053,842 888,321 249,826 28.1 557,262 125,333 22.5 30.7 29.6 Other .............................. 440,842 386,483 118,046 30.5 235,844 112,909 47.9 3.6 2.4 Businessrelated3 Conventional,firstlien Homepurchase .................. 72,619 62,161 4,377 7.0 52,601 6,194 11.8 53.5 23.4 Refinance........................ 59,831 48,215 4,913 10.2 38,694 5,366 13.9 36.6 24.3 Other .............................. 31,417 25,969 3,645 14.0 19,277 4,235 22.0 3.1 .9 MULTIFAMILY5 Conventional,firstlien Homepurchase .................. 27,132 24,867 2,354 9.5 21,526 1,283 6.0 43.8 25.1 Refinance........................ 24,262 21,840 2,192 10.0 18,872 1,198 6.3 47.5 24.6 Other .............................. 6,144 5,403 598 11.1 4,605 230 5.0 22.6 10.0 Total .............................. 30,146,893 24,665,065 6,691,068 27.1 15,611,711 4,086,033 26.2 21.6 24.3 Note:Excludestransition-periodapplications(thosesubmittedbefore2004) 3. Business-related applications and loans are those for which the lender andtransition-periodloans(thoseforwhichtheapplicationwassubmittedbe- reportedthattherace,ethnicity,andsexoftheapplicantorco-applicantare fore2004). ‘‘notapplicable’’;allotherapplicationsandloansarenonbusinessrelated. 1. APRspreadisthedifferencebetweentheAPRontheloanandtheyield 4. Includesapplicationsandloansforwhichoccupancystatuswasmissing. onacomparable-maturityTreasurysecurity.Thethresholdforfirst-lienloansis 5. Includesbusiness-relatedandnonbusiness-relatedapplicationsandloans aspreadof3percentagepoints;forjunior-lienloans,itisaspreadof5per- forowner-occupiedandnonowner-occupiedproperties. centagepoints. ... Notapplicable. 2. Loans covered by the Home Ownership and Equity ProtectionAct of 1994,whichdoesnotapplytohome-purchaseloans.

Higher-Priced Home Lending and the 2005 HMDA Data A133 4. Dispositionofapplicationsforhomeloans,andoriginationandpricingofloans,bytypeofhomeandtypeofloan,2005—Continued Loansoriginated Memo Loanswithannualpercentagerate(APR) Transition-periodapplications(thosesubmittedbefore2004) spreadabovethethreshold1 Distribution, APRspread bypercentagepoints Loansoriginated (percentagepoints) ofAPRspread Numberof Numberof HOEPA- Number Number Percent HOEPAcovered submitted denied denied Percentwith covered 5–6.99 7–8.99 9ormore Mean Median loans2 Number APRspread loans2 above threshold 34.0 3.4 .2 4.8 4.7 ... 9,178 718 9.5 5,367 2.6 ... 72.7 25.9 1.4 6.5 6.3 ... 449 36 10.9 222 9.0 ... 8.8 1.4 .5 3.8 3.3 ... 972 124 21.1 302 .3 ... 51.7 34.5 13.8 7.1 7.0 ... 2 0 0 0 0 ... 35.9 4.4 .6 4.8 4.7 15,602 4,382 630 21.8 1,447 9.7 1 59.9 30.9 9.2 7.0 6.7 7,225 206 19 14.7 80 10.0 1 11.9 3.9 .1 4.3 4.5 19 332 66 34.2 49 2.0 0 54.2 41.7 4.2 6.9 6.8 1 0 0 0 0 0 0 30.3 5.0 .9 4.7 4.5 1,873 92 7 9.3 54 11.1 0 43.8 31.4 24.8 7.7 7.4 5,726 31 10 71.4 4 0 0 24.5 9.1 ... 4.5 3.9 1 1 0 0 0 0 0 42.7 27.8 29.6 8.1 7.4 472 0 0 0 0 0 0 ... ... ... ... ... ... 5 1 50.0 0 0 ... 31.3 12.4 4.7 5.4 4.9 ... 89 10 11.9 51 11.8 ... 30.6 7.0 2.4 5.0 4.7 1,760 87 14 21.2 24 20.8 0 27.5 21.9 15.4 6.5 5.9 1,059 85 14 20.9 30 6.7 0 27.7 2.8 .5 4.5 4.3 ... 1,599 159 12.1 903 4.8 ... 34.9 4.2 .5 4.8 4.7 1,534 634 90 21.4 251 15.5 0 48.7 32.6 12.7 7.0 6.8 470 77 14 23.0 36 30.6 0 16.7 4.2 2.1 4.4 3.9 ... 1,778 123 8.0 1,084 1.6 ... 32.1 5.4 1.7 4.8 4.6 134 641 80 17.3 167 1.8 1 60.0 29.3 6.7 6.7 6.5 92 361 62 23.2 73 6.8 1 29.0 1.8 .3 4.5 4.2 ... 59 3 5.7 46 0 ... 24.6 3.2 .1 4.4 4.1 5 62 3 5.3 34 2.9 0 56.1 8.7 2.6 5.4 5.3 7 9 0 0 8 12.5 0 41.8 10.2 2.0 5.3 5.1 35,980 21,131 2,183 13.5 10,232 4.4 4 fromabout16percentin2000tolessthan3percentin homepriceshavediminishedborrowerinterestinthe 2005 (data not shown in tables). Of all first-lien FHA program as FHA insurance limits have fallen home-purchaseloansreportedin2005,theFHAshare behindincreasesinlocalhomevalues.Insomepartsof was 5 percent. New, more flexibly underwritten con- thecountry,FHA-insuredproductsaccountforanegventionalloanproductsareattractingborrowerswho, ligible share of the market. In the metropolitan diviinthepast,mightotherwisehavesoughtFHAbacking, sion that includes San Francisco, for example, only particularly those borrowers seeking loans with high two of the roughly 23,000 first-lien home-purchase LTVratios.Also,insomeareas,highandrapidlyrising loanswereFHA-insuredin2005.

A134 Federal Reserve Bulletinh2006 5. Home-purchaselendingthatbeganwitharequestforpre-approval:Dispositionandpricing,bytypeofhome,2005 Applicationsprecededby Loanoriginationswhoseapplicationswere Requestsforpre-approval requestsforpre-approval1 precededbyrequestsforpre-approval Loanswithannualpercentage Acteduponbylender rate(APR)spread abovethethreshold2 Typeofhome Number Number Percent Number actedupon Number denied denied submitted bylender Number Number Number Percent denied ONE-TOFOUR-FAMILY Nonbusinessrelated3 Owneroccupied Sitebuilt Conventional Firstlien ........................ 834,824 205,707 24.6 548,224 484,423 38,343 409,856 62,189 15.2 Juniorlien ....................... 137,063 25,952 18.9 100,161 90,799 5,991 77,428 22,986 29.7 Governmentbacked Firstlien ........................ 94,105 28,830 30.6 64,370 57,719 4,948 48,774 902 1.8 Juniorlien ....................... 186 35 18.8 156 130 17 111 4 3.6 Manufactured Conventional,firstlien ............. 43,042 22,200 51.6 40,178 34,042 19,715 8,980 6,363 70.9 Other .............................. 4,958 1,837 37.1 3,375 3,027 564 2,181 163 7.5 Nonowneroccupied4 Conventional,firstlien ............. 121,816 21,453 17.6 86,844 75,387 7,917 61,782 10,355 16.8 Other .............................. 16,600 2,322 14.0 14,375 12,009 1,131 9,659 5,830 60.4 Businessrelated3 Conventional,firstlien ............. 5,197 1,619 31.2 3,784 2,619 263 2,239 420 18.8 Other .............................. 1,107 91 8.2 1,061 810 63 705 272 38.6 MULTIFAMILY5 Conventional,firstlien ............. 420 43 10.2 402 299 33 248 29 11.7 Other .............................. 79 7 8.9 77 57 5 45 14 31.1 Total .............................. 1,259,397 310,096 24.6 863,007 761,321 78,990 622,008 109,527 17.6 Note: Excludestransition-periodrequestsforpre-approval(thosesubmitted reportedthattherace,ethnicity,andsexoftheapplicantorco-applicantare before2004).Refertogeneralnotetotable1. ‘‘notapplicable’’;allotherapplicationsandloansarenonbusinessrelated. 1. These applications are included in the total of 30,146,893 reported in 4. Includesapplicationsandloansforwhichoccupancystatuswasmissing. table4. 5. Includesbusiness-relatedandnonbusiness-relatedapplicationsandloans 2. Refertotable4,note1. forowner-occupiedandnonowner-occupiedproperties. 3. Business-related applications and loans are those for which the lender ... Notapplicable. Owner-Occupancy Status buyers is that a portion of these buyers do not use mortgages;rather,theypaycashforthepropertiesor Some believe that part of the strength in housing take out commercial loans. (Of course, some ownermarkets over the past several years is due to a occupants also purchase homes solely with cash.) In growingnumberandshareofhomesalestoinvestors 2005,lenderscoveredbyHMDAreportedonroughly or individuals purchasing second homes, as distinct 3millionapplicationsfornonowner-occupiedproperfrom buyers who intend to make the units being ties (data derived from table 4). About half of these purchased their primary residence. HMDA data can applications were conventional first liens for home beusedtodocumenttheroleofinvestorsandsecond- purchase. home buyers in the housing market because the data TheHMDAdataindicatethattheshareofreported indicatewhetherthepropertytowhichanapplication lending for nonowner-occupied purposes remained orloanrelatesisintendedastheborrower’sprincipal steadyfrom1990throughthemid-1990s,primarilyin dwelling (that is, as an owner-occupied unit).19 A the range of 4.5 percent to 6.0 percent (whether limitation to using mortgage lending information to measured in number of loans or dollar amount of gauge the activity of investors and second-home loans), and then began rising (table 6). In 2005, the nonowner-occupied share of the home-purchase loan market in terms of number of loans was about 19.Aninvestmentpropertyisanonowner-occupieddwellingthatis intendedtobecontinuouslyrented.Somenonowner-occupiedunits— 17percentandintermsofdollaramountofloanswas vacation homes and second homes—are for the primary use of the roughly 16 percent. Both figures rose from 2004, ownerandwouldthusnotbeconsideredinvestmentproperties.The when the shares were 15 percent and 13 percent HMDAdatadonot,however,distinguishbetweenthesetwotypesof nonowner-occupieddwellings. respectively.

Higher-Priced Home Lending and the 2005 HMDA Data A135 5. Home-purchaselendingthatbeganwitharequestforpre-approval:Dispositionandpricing,bytypeofhome,2005—Continued Loanoriginationswhoseapplicationswere precededbyrequestsforpre-approval Memo Applicationswithtransition-periodrequestsforpre-approval (requestsubmittedbefore2004) Loanswithannualpercentagerate(APR)spreadabovethethreshold2 Distribution, APRspread Loansoriginated bypercentagepointsofAPRspread (percentagepoints) Number Number Percent Percent submitted denied denied withAPR Mean Median 3–3.99 4–4.99 5–6.99 7–8.99 9ormore Number spread spread spread above threshold 30.3 25.3 36.0 7.8 .6 4.9 4.8 435 14 4.6 207 6.3 ... ... 66.8 30.1 3.1 6.6 6.4 28 0 0 16 6.3 57.4 33.1 8.0 1.1 .3 4.1 3.8 133 7 9.6 57 0 ... ... 75.0 25.0 0 6.2 5.6 0 0 0 0 14.4 20.3 32.8 24.1 8.4 6.2 5.8 3 0 0 0 0 28.8 .6 61.3 9.2 0 5.4 6.0 1 0 0 1 0 54.5 22.6 17.3 4.7 .9 4.3 3.9 90 6 9.1 37 10.8 .1 0 39.8 41.4 18.7 7.6 7.5 5 4 80.0 1 100 20.5 11.4 25.2 16.2 26.7 6.8 6.4 41 0 0 23 0 2.2 0 36.4 20.2 41.2 8.4 8.0 7 0 0 2 0 27.6 27.6 41.4 3.4 0 4.7 4.7 0 0 0 0 0 0 0 57.1 28.6 14.3 7.7 6.1 0 0 0 0 0 23.8 18.0 40.5 14.9 2.8 5.5 5.3 743 31 6.2 344 5.5 Theextentoflendingfornonowner-occupiedprop- significantsecond-homemarkets.Eachofthesestates erties varies considerably by geography (figure 1). has also experienced elevated shares of lending for Someofthestateswiththehighestincidenceofsuch nonowner-occupied properties for the past several lending in 2005 included Florida, Nevada, Hawaii, years. South Carolina, and Vermont, all of which have Piggyback Lending 6. Home-purchaseloansonnonowner-occupiedsite-built homesasashareofallfirst-lienhome-purchaseloans TheexpandedHMDAdataprovideanopportunityto onone-tofour-familyhomes,bynumberanddollar measuretheextenttowhichhomebuyersaresimultaamountofloans,1990–2005 neously obtaining first- and junior-lien loans. Such Percent simultaneous borrowing has been a feature of the Year Number Dollaramount conventional mortgage marketplace for some time 1990 .......................... 6.6 5.9 but has grown in importance in recent years as 1991 .......................... 5.6 4.5 lenders have marketed products intended to offer 1992 .......................... 5.2 4.0 1993 .......................... 5.1 3.8 consumers an alternative to private mortgage insur- 1994 .......................... 5.7 4.3 ance(PMI)or,insomecases,alineofcreditthatmay 1995 .......................... 6.4 5.0 be used for a variety of purposes. Simultaneous 1996 .......................... 6.4 5.1 1997 .......................... 7.0 5.8 borrowing of this type is often referred to as a 1998 .......................... 7.1 6.0 1999 .......................... 7.4 6.4 ‘‘piggyback’’loanoran‘‘80-10-10’’loan. 2000 .......................... 8.0 7.2 Manyfirst-timehomebuyershavefewassetsavail- 2001 .......................... 8.6 7.6 abletosatisfydown-paymentandclosing-costrequire- 2002 .......................... 10.5 9.2 2003 .......................... 11.9 10.6 ments, and thus they can ordinarily qualify for a 2004 .......................... 14.9 13.1 mortgage only with a high LTV ratio and some type 2005 .......................... 17.3 15.7 of mortgage backing that protects the lender in case

,semoh tliub-etis ,ylimaf-ruof ot -eno no snaol esahcrup-emoh neil-tsrif rof gnidnel deipucco-renwonon fo ecnedicnI .1 5002 ,elitniuq yb dna etats yb tnecreP (cid:19)(cid:27)(cid:17)(cid:28)(cid:177)(cid:19)(cid:19)(cid:17)(cid:26)(cid:3)(cid:3) (cid:19)(cid:27)(cid:17)(cid:20)(cid:20)(cid:177)(cid:20)(cid:27)(cid:17)(cid:28)(cid:3)(cid:3) (cid:19)(cid:21)(cid:17)(cid:23)(cid:20)(cid:177)(cid:20)(cid:27)(cid:17)(cid:20)(cid:20) (cid:19)(cid:20)(cid:17)(cid:21)(cid:21)(cid:177)(cid:20)(cid:21)(cid:17)(cid:23)(cid:20) (cid:19)(cid:22)(cid:17)(cid:23)(cid:22)(cid:177)(cid:20)(cid:20)(cid:17)(cid:21)(cid:21) A136 Federal Reserve Bulletinh2006

Higher-Priced Home Lending and the 2005 HMDA Data A137 of default. Other borrowers have the financial capac- families involved a junior-lien or piggyback loan itytomakealargedownpaymentbutprefernottodo reported by the same lender, up from nearly 14 perso. Traditionally, lenders and secondary-market pur- centin2004(table7). chasers have sought protection in case of borrower Piggyback lending varies by borrower income defaultforloanswithhighLTVratios.PMIreducesa and race or ethnicity as well as by geography and lender’s credit risk by insuring against losses associ- loan characteristic.21 Minority borrowers, borrowers ated with borrower default up to a contractually with middle or upper incomes, and borrowers who established percentage of the claim amount. PMI purchased homes in lower-income census tracts are premiums are paid by the borrower, usually as an more likely to use piggyback loans to purchase add-ontothemonthlymortgagepayment. homes than non-Hispanic whites or lower-income Typically, PMI is required on conventional loans borrowers.22The apparent inconsistency between the with LTVratios above 80 percent. Over the past few results for borrower income and those for censusyears, lenders have become more active in self- tract income appears to be driven by the relatively insuring by waiving PMI requirements if a borrower high incidence of the use of piggyback loans by simultaneouslytakesoutafirst-lienloanwithanLTV middle- and upper-income borrowers purchasing ratioof80percentormoreandajunior-lienloanata homes in lower-income areas. Piggyback lending is higher price to cover the remaining portion of the also related to the amount borrowed, as larger firstloan.The combined loans are often competitive on a lien loans are more likely to be associated with pricebasiswithasingleloaninvolvingPMIandoffer piggyback lending than are smaller loans. Regionthe borrower a tax advantage because the interest ally, piggyback lending is most common in the payments on the junior-lien loan are generally tax- western region of the country and is particularly deductible,whereasthePMIpremiumsarenot. frequent in California, Nevada, and Colorado. Piggyback loans are not identified as such in the Piggybacklendingiscloselyrelatedtothelocation HMDA data. However, the data provide a basis for ofapropertyrelativetothelender’sassessmentareas identifying piggyback loans if one assumes that two asdefinedbytheCRA.Borrowerswhoareobtaining conventional home-purchase loans involving proper- loanstopurchasehomesintheCRAassessmentareas ties in the same census tract, from the same lender, oftheirlendersaremuchlesslikelytousepiggyback with identical time of application and closing, and loansthanareborrowerspurchasinghomesoutsideof with the same owner-occupancy status, borrower their lenders’ assessment areas or borrowers obtainincome, race or ethnicity, and sex involved the same ing loans from lenders not covered by the CRA borrowerandthesamehome.Since2004,theidenti- (independent mortgage companies and credit ficationprocesshasbeenimprovedbytheadditionof unions).23 Although the HMDA data do not provide lienstatus,whichearliercouldonlybeapproximated bycomparingthesizeofloansthatwerematched.For 2005, we estimate that about 85 percent of the 21.Onlyloanswithcompleteinformationoncensus-tractcharacteristicsareincludedintheanalysis. junior-lien home-purchase loans for owner-occupied 22. The income category of a borrower is relative to the median properties can be matched to a first-lien loan by this familyincomeofthearea(MSAorstatewidenon-MSA)inwhichthe process.20 property being purchased is located, and the income category of a censustractisthemedianfamilyincomeofthetractrelativetothatof The expanded HMDA data document the importhearea(MSAorstatewidenon-MSA)inwhichthetractislocated: tance of the junior-lien home-purchase loan market. ‘‘Low’’islessthan50percentofthemedian;‘‘moderate’’is50percent For 2005, lenders reported on a total of 1.37 million to79percent(inthisarticle,‘‘lowerincome’’encompassesthelowand moderate categories); ‘‘middle’’ is 80 percent to 119 percent; and junior-lien loans used for the purpose of home pur- ‘‘upper’’ is 120 percent or more. For loans with two or more chase, up 74 percent from 2004 (data not shown in applicants,HMDA-coveredlendersreportdataononlytwo.Income tables). The vast majority of junior-lien loans are fortwoapplicantsisreportedjointly. For tables 7 and 12, minority means that the applicant or coconventionalloans:Onlyaverysmallnumber(fewer applicantisHispanicorhasgivenatleastonenonwhiterace.Forother than1,000nationwide)ofthejunior-lienloansissued tables,weuseadifferenttaxonomy.Applicantsareplacedunderonly in 2005 involved government-backed forms of credit onecategoryforraceandethnicity,generallyaccordingtotheraceand ethnicityofthepersonlistedfirstontheapplication.However,under (table 4). Overall, for 2005, we estimate that 22 perrace,theapplicationisdesignatedasjointifoneapplicantreportedthe centofthereportedfirst-lienhome-purchaseloanson single designation of white and the other reported one or more owner-occupied site-built homes for one to four minorityraces.Iftheapplicationisnotjointbutmorethanoneraceis reported,thefollowingdesignationsaremade:Ifatleasttwominority races are reported, the application is designated as two or more minorityraces;ifthefirstpersonlistedonanapplicationreportstwo 20.DateinformationcollectedunderHMDA,whichiscriticaltothe races, and one is white, the application is categorized under the accuracyofthematchingprocess,isnotmadeavailabletothepublic minorityrace. butisavailabletotheagenciesthatoverseeHMDAreporting(includ- 23.Largercommercialbanksandsavingsassociationscoveredby ingtheFederalReserveBoard). the CRA (generally those with assets of $1 billion or more) are

A138 Federal Reserve Bulletinh2006 an explanation for this finding, one possibility is the 7. Incidenceofpiggybacklendingforhome-purchase availability of special low-down-payment lending loansonowner-occupied,one-tofour-family,site-built homes,andtheincidenceofsuchlendingthatinvolved programs for homebuyers purchasing homes in lendahigher-pricedfirst-lienloan,bycharacteristicof ers’CRAassessmentareas,programsthatwouldtend borrowerandofcensustractandbyamountofloan, todiminishtheneedforajunior-lienloantoprovidea typeoflender,andlocationofproperty,2004and2005 sourceofdownpaymentwhenpurchasingahome. Percent The incidence of piggyback lending varies across Shareof neighborhoods according to the distribution of credit Shareofloans piggybackloans thatare involving scores among those with outstanding mortgages, the Characteristicandstatus piggyback higher-priced distributionofeducationalattainmentlevelsofneigh- firstliens borhood residents, and the proportion of minority 2004 2005 2004 2005 residents in the neighborhood.24 The incidence of Borrower piggyback lending is higher in areas that have larger Incomeratio proportions of mortgage borrowers with low credit (percentofareamedian) Lessthan80 ....................... 11.9 18.9 25.6 61.9 scores and that have larger minority populations and 80–100 ............................ 15.9 24.6 21.7 56.4 100ormore ....................... 14.3 21.9 16.1 50.9 is smaller in areas that have larger proportions of Notreported1 ...................... 8.3 19.4 4.7 20.9 residents with more than a high-school education. Total .......................... 13.9 21.8 19.6 53.6 Thesethreerelationshipsgenerallyholdregardlessof Minoritystatus Minority ........................... 20.7 32.6 26.8 69.7 the level of census-tract income (data not shown in Non-Hispanicwhite ................ 11.4 17.7 15.2 41.2 table). Missing2........................... 13.3 18.2 16.9 51.4 Total .......................... 13.9 21.8 19.6 53.6 Sex Loans for Manufactured Homes Female ............................ 15.1 24.8 24.3 59.6 Male .............................. 16.2 25.9 22.7 58.6 Joint3.............................. 11.5 16.8 13.3 42.4 Until the release of the 2004 data, users of HMDA Total4......................... 13.9 21.8 19.6 53.6 datahadnocertainwaytoidentifywhichapplications AmountofLoan and loans involved manufactured homes.25 To help (ThousandsofDollars) overcome this limitation, the Department of Housing Lessthan100 ...................... 10.7 16.7 33.3 65.6 100–250 ........................... 15.1 23.6 18.9 51.8 andUrbanDevelopment(HUD)producedannuallya 250ormore ....................... 13.9 21.6 13.6 51.5 Total .......................... 13.9 21.8 19.6 53.6 list of reporting institutions (typically about twenty) that it believed were primarily in the business of TypeofLender, extending such credit.26 Users of the HMDA data byPropertyLocation oftenreliedontheHUDlisttoidentify,albeitimper- Depositorywithinassessmentarea5 . 6.2 9.8 5.0 15.0 Depositoryoutsideofassessment fectly,loansandapplicationsrelatedtomanufactured area........................... 12.1 19.5 23.0 56.0 LendernotcoveredbyCRA6 ....... 22.2 32.2 21.0 60.3 homes.Thispracticehaditsownlimitations:Itcould Total4......................... 13.9 21.8 19.6 53.6 not be used to identify applications and loans related LocationofProperty, tomanufacturedhomesreportedbylendersnotonthe byFreddieMacRegion7 HUDlist,anddatausersoftenassumedthatallloans Northeast .......................... 10.3 18.6 18.5 47.6 Southeast .......................... 11.2 19.8 23.0 54.9 by lenders on the list were for manufactured homes NorthCentral ...................... 9.0 16.4 25.9 53.5 when some were not. The expanded HMDA data Southwest.......................... 15.6 24.0 21.3 47.6 West............................... 21.9 28.8 16.0 59.1 resolve this problem by including a code to identify Total4......................... 13.9 21.8 19.6 53.6 applicationsandloansformanufacturedhomes. CensusTractofProperty Incomeratio (percentofareamedian) requiredtoidentifythecensustractsintheirCRAassessmentareasas Lessthan80 ....................... 18.7 29.4 27.3 67.6 of the end of each calendar year. That information was used to 80–119 ............................ 14.1 22.2 20.7 55.2 120ormore ....................... 11.8 18.0 13.3 41.1 determinewhichloansintheHMDAdatawereforpropertieswithin Total4......................... 13.9 21.8 19.6 53.6 thelenders’CRAassessmentareas.Whenlenderswerepartofabank orthriftholdingcompany,thecombinedassessmentareasofallbanks Racialorethniccomposition intheholdingcompanywereusedfortheanalysis. (minoritiesaspercentage ofpopulation) 24. The distribution of credit scores for mortgage borrowers by Lessthan10 ....................... 9.0 15.0 17.1 42.0 census tract relates to all individuals with an outstanding mortgage 10–50 ............................. 14.5 22.4 17.6 50.3 loanasoftheendof2004.Nonetheless,webelieveitislikelytobe 50–100 ............................ 22.3 33.6 25.9 70.3 Total4......................... 13.9 21.8 19.6 53.6 representativeofthecredit-scoredistributionof2005borrowers.The data were provided by one of the three national credit-reporting Location agencies. Centralcity ........................ 13.7 21.7 19.2 52.7 25.Asdistinctfromsite-builthomes,mostmanufacturedhomesare Noncentralcity..................... 15.5 23.8 19.9 54.9 Ruraloronlystateknown .......... 8.0 13.4 21.7 51.5 assembledinfactoriesandshippedtoahomesite. Total4......................... 13.9 21.8 19.6 53.6 26.Refertowww.huduser.org/datasets/manu.html.

Higher-Priced Home Lending and the 2005 HMDA Data A139 7. Incidenceofpiggybacklendingforhome-purchaseloans countedfor29percentoftheloans,andthetoptwenty onowner-occupied,one-tofour-family,site-builthomes, accountedfor39percent(datanotshownintables). andtheincidenceofsuchlendingthatinvolvedahigher- Loans for manufactured homes entail more credit pricedfirst-lienloan,bycharacteristicofborrowerandof risk than do most other forms of secured credit censustractandbyamountofloan,typeoflender,and locationofproperty,2004and2005—Continued extendedtoconsumers.Lendercautionisreflectedin Percent the very high denial rates on applications for loans backed by manufactured homes. As noted, past Shareof Shareofloans piggybackloans HMDA data did not distinguish applications for thatare involving Characteristicandstatus piggyback higher-priced manufacturedhomesfromthoseforsite-builtproperfirstliens ties.Analysis of the HUD list of manufactured-home 2004 2005 2004 2005 loan specialists suggested that such lenders had very Creditscoreofborrowers high denial rates and that, for lenders offering both (percentofmortgageborrowers manufactured-home loans and other home loans, a withscoresbelow600)8 20ormore......................... 16.1 27.4 35.8 70.6 distorted picture of their propensity to deny credit 10–19 ............................. 15.6 24.4 22.1 57.7 Lessthan10 ....................... 12.4 18.7 13.2 43.7 could easily be drawn. The 2005 data document the Total4......................... 13.9 21.8 19.6 53.6 importance of distinguishing applications for manu- Educationalattainmentofresidents factured homes from those for site-built properties. (percentofadultswithhigh-school educationorless) Forexample,denialratesforapplicationsforconven- 30orless .......................... 12.2 18.2 11.5 36.8 31–60 ............................. 14.4 22.6 20.2 54.7 tional first-lien home-purchase loans on manufac- Morethan60 ...................... 15.3 24.6 28.8 68.7 tured homes were 52.6 percent in 2005, compared Total4......................... 13.9 21.8 19.6 53.6 with 16.4 percent for such applications related to the Realpriceappreciation ofrealestate9 purchase of one- to four-family site-built homes Lessthanzero ..................... 15.7 24.1 20.2 58.2 0–20............................... 12.4 20.1 20.3 51.4 (table4). Morethan20 ...................... 12.4 19.6 17.9 46.7 Manufactured housing also differs from site-built Total4......................... 13.9 21.8 19.6 53.6 homes in that it serves relatively more lower-income Note: Fordefinitionsofpiggybacklendingandhigher-pricedloan,referto households but fewer minorities. Of those obtaining text. Excludestransition-periodloans(thoseforwhichtheapplicationwassub- loans to purchase manufactured homes, 38 percent mittedbefore2004).Fordefinitionofincomecategoriesforborrowerandcenwereoflowerincome,whereasofthoseborrowingto sustract,refertotextnote22.Censustractisforthepropertysecuringthe loan.ThetermminoritymeansHispanicorLatinoethnicityoranyraceother purchasesite-builthomes,about20percenthadlower thanwhiteforeithertheborrowerorthecoborrower.Census-tractdatareflect incomes (table 8). On average, minority borrowers the2000decennialcensus;theyalsoreflectdefinitionsformetropolitanstatisticalareasestablishedbytheOfficeofManagementandBudgetinJune2003 have lower incomes than do non-Hispanic white andusedinHMDAforthefirsttimeinthe2004data. borrowers,butonlyabout15percentofmanufactured- 1. Informationforincomewasnotreported. 2. Information for the characteristic was missing on the application. home purchasers were members of a racial or ethnic 3. Ontheapplicationsfortheseloans,oneapplicantreported‘‘male,’’and minoritygroup,whereasabout28percentofpurchasthe other reported ‘‘female.’’For female and for male, only sole applicants wereconsidered. ersofsite-builthomeswereminorities. 4. Excludesloansforwhichtheinformationforthecharacteristicwasmissingontheapplication. 5. IncludeslendingbynonbankaffiliatesintheCRAassessmentareaofthe Secondary-Market Activity depositoryinstitution. 6. Includes credit unions and mortgage companies not affiliated with a depositoryinstitutionorwithabankorthriftholdingcompany. HMDAdatadocumenttheimportanceofthesecond- 7. Freddie Mac defines its regions as follows: Northeast: N.Y., N.J., Pa., arymarketforhomeloans.Ofthe21.5millionhome Del., Md., D.C., Va., W.V., P.R., Maine, N.H., Vt., Mass., R.I., Conn., V.I.; Southeast:N.C.,S.C.,Tenn.,Ky.,Ga.,Ala.,Fla.,Miss.;NorthCentral:Ohio, loans originated or purchased in 2005 by lenders Ind.,Ill.,Mich.,Wis.,Minn.,Iowa,N.D.,S.D.;Southwest:Texas,La.,N.M., covered by HMDA, 14.9 million, or nearly 70 per- Okla., Ark., Mo., Kan., Colo., Neb., Wyo.; West: Calif., Ariz., Nev., Ore., Wash.,Utah,Idaho,Mont.,Hawaii,Alaska,Guam. cent,weresoldin2005(datanotshownintables).27 8. Includes all borrowers with an outstanding mortgage regardless of the Prominentinthesecondarymarketaregovernmentyearinwhichtheloanwastakenout. 9. Basedonthechangeinmedianhomevaluesforaconstant2000-defined sponsored enterprises (GSEs)—in particular, Fannie geography. Mae and Freddie Mac. For the most part, the pur- Source: For Freddie Mac data, Primary Mortgage Market Survey; for census-tractcharacteristics,the1990and2000decennialcensuses;forcredit- chases of Fannie Mae and Freddie Mac in 2005 scoredata,oneofthethreenationalcredit-reportingagencies. The 2005 HMDAdata indicate that roughly 4,400 27.TheHMDAdatatendtoundercountsomewhatthevolumeof lenders extended more than 256,000 manufactured- secondary-marketsales.Onereasonisthat,forexample,someloans home loans, a loan volume up slightly from that in originated in 2005 will be sold to a secondary-market institution in 2006orlaterandthuswillneverbereportedasasale.Anotheristhat, 2004.Amongtheselenders,thetenthatextendedthe aswithotherHMDAdata,about20percentofhomeloansoriginated largest number of manufactured-home loans ac- in2005wereextendedbylendersnotcoveredbyHMDA.

A140 Federal Reserve Bulletinh2006 8. Distributionofhome-purchaseloansforone-tofour-familyowner-occupiedhomes,bycharacteristicofborrower andofcensustractandbytypeofhome,2005 Note:DatarevisedonSept.18,2006,tocorrectcomputationalerrors. Site-built Manufactured Total Characteristic Memo andstatus Percentof Percentof Percentof Percentof Percentof Percentof Number characteristic1 status2 characteristic1 status2 characteristic1 status Borrower3 Incomeratio(percentofareamedian) Lessthan50 .................................... 3.7 93.0 10.4 7.0 3.9 100 181,818 50–79 .......................................... 15.9 95.6 27.4 4.4 16.2 100 765,134 80–119 ......................................... 27.1 97.1 30.7 2.9 27.2 100 1,281,742 120ormore .................................... 53.3 98.5 31.5 1.6 52.7 100 2,483,787 Total4...................................... 100 97.4 100 2.6 100 100 4,712,481 Race AmericanIndianorAlaskaNative ............... .7 95.8 1.2 4.2 .7 100 36,064 Asian ........................................... 5.1 99.8 .4 .2 4.9 100 244,143 BlackorAfricanAmerican ...................... 7.7 98.4 4.6 1.6 7.6 100 375,188 NativeHawaiianorotherPacificIslander......... .5 98.2 .4 1.8 .5 100 26,045 White........................................... 75.1 97.1 85.9 2.9 75.3 100 3,730,468 Twoormoreminorityraces ..................... .1 97.0 .1 3.0 .1 100 2,453 Joint............................................ 1.3 97.9 1.0 2.1 1.3 100 61,723 Missing5........................................ 9.7 98.3 6.4 1.7 9.6 100 475,141 Total ....................................... 100 97.4 100 2.6 100 100 4,951,225 Ethnicity HispanicorLatino .............................. 12.3 98.4 7.6 1.6 12.2 100 602,774 NotHispanicorLatino .......................... 76.5 97.2 84.1 2.8 76.7 100 3,798,888 Joint6........................................... 1.3 97.8 1.1 2.2 1.3 100 64,609 Missing5........................................ 9.9 98.1 7.3 1.9 9.8 100 484,954 Total ....................................... 100 97.4 100 2.6 100 100 4,951,225 Minoritystatus Minority ........................................ 27.8 98.6 15.4 1.4 27.5 100 1,360,100 Non-Hispanicwhite ............................. 61.8 96.8 77.0 3.2 62.2 100 3,080,720 Missing5........................................ 10.4 98.1 7.6 1.9 10.3 100 510,405 Total ....................................... 100 97.4 100 2.6 100 100 4,951,225 CensusTractofProperty Incomeratio(percentofareamedian) Lessthan50 .................................... 1.7 99.4 .4 .6 1.7 100 81,222 50–79 .......................................... 13.6 97.0 16.1 3.0 13.7 100 668,547 80–119 ......................................... 49.7 96.4 72.2 3.6 50.3 100 2,461,940 120ormore .................................... 35.1 99.2 11.2 .8 34.5 100 1,687,639 Total4...................................... 100 97.5 100 2.5 100 100 4,899,348 Racialorethniccomposition (minoritiesaspercentageofpopulation) Lessthan10 .................................... 32.1 96.5 45.3 3.5 32.4 100 1,589,295 10–19 .......................................... 22.8 97.7 21.0 2.3 22.8 100 1,114,804 20–49 .......................................... 28.0 97.8 24.9 2.2 27.9 100 1,366,972 50–79 .......................................... 10.4 98.3 6.8 1.7 10.3 100 505,574 80–100 ......................................... 6.7 99.2 2.0 .8 6.6 100 324,229 Total4...................................... 100 97.5 100 2.5 100 100 4,900,874 Location Centralcity ..................................... 38.6 98.9 16.3 1.1 38.0 100 1,866,761 Noncentralcity.................................. 52.3 97.7 48.4 2.3 52.2 100 2,562,936 Ruraloronlystateknown ....................... 9.1 90.9 35.3 9.1 9.8 100 479,951 Total4...................................... 100 97.5 100 2.5 100 100 4,909,648 Note: Excludestransition-periodloans(thoseforwhichtheapplicationwas 1. Distributionsumsvertically. submittedbefore2004).Fordefinitionofincomecategoriesforborrowerand 2. Distributionsumshorizontally. censustract,refertotextnote22.Censustractisforthepropertysecuringthe 3. Fordetailsontheidentificationofborrowerincome,race,andethnicity, loan.Categoriesforraceandethnicityreflecttherevisedstandardsestablished refertotextnote22. in1997bytheOfficeofManagementandBudget(OMB).Thetermminority 4. ExcludesloansforwhichtheinformationforthecharacteristicwasmissmeansHispanicorLatinoethnicityoranyraceotherthanwhiteforboththe ingontheapplication. borrowerandthecoborrower.Census-tractdatareflectthe2000decennialcen- 5. Informationforthecharacteristicwasmissingontheapplication. sus;theyalsoreflectdefinitionsformetropolitanstatisticalareasestablishedby 6. Ontheapplicationsfortheseloans,oneapplicantreported‘‘Hispanicor theOMBinJune2003andusedinHMDAforthefirsttimeinthe2004data. Latino,’’andtheotherreported‘‘notHispanicorLatino.’’ consistedofconventionalfirst-lienloansoriginatedto by the Federal Housing Finance Board.Among such purchase homes or to refinance existing loans.These loans,thesetwoGSEsaccountedforabout32percent two GSEs accounted for 28 percent of all loans of the purchased conventional conforming loans.28 purchased by all secondary-market institutions as reportedintheHMDAdata.FannieMaeandFreddie 28.Conformingloansareloansthatarewithintheloan-sizelimits Mac,however,focusonthepurchaseofconventional determinedbytheFederalHousingFinanceBoardandthatmeetother home loans within size limits established each year requirements used by Freddie Mac and Fannie Mae to determine

Higher-Priced Home Lending and the 2005 HMDA Data A141 Moreover, mortgage loans purchased by Fannie Mae not a direct correspondence between higher-priced and Freddie Mac are largely resold in the form of and subprime lending. mortgage-backedsecurities. Three factors may lead to changes in the reporting Othertypesofpurchasinginstitutionsactiveinthe ofhigher-pricedlending.Thefirstislenders’business secondarymarketincludeprivatesecuritizationpools practices, particularly lenders’ willingness or ability (12 percent of all loans sold); mortgage, finance, and to bear credit risk. For example, an increase in insurancecompanies(13percent);depositoryinstitu- competitive conditions in the higher-credit-risk portions(6percent);GinnieMae(3percent);affiliatesof tion of the market has driven down margins and institutions covered by HMDA (16 percent); and encouragedlenderstoofferawiderrangeofproducts. ‘‘other’’purchasers(22percent).29 The second factor that may affect the reporting of higher-priced lending is consumers’ borrowing prac- THE 2005 HMDA DATA ON LOAN PRICING tices or credit-risk profiles. Changes in borrower credit-riskprofilescanaltertheincidenceofsubprime The expanded HMDA data allow analysis of loan lendingevenwhentheinterestratesituationisstable. pricing along a number of dimensions, including by Such changes reflect real fluctuations in economic loan product, across lenders and markets, and by behavior or conditions rather than an artifact of the financial and personal characteristics of borrowers. HMDA reporting criteria. The credit-risk profile of The results of this analysis have implications for fair the population of borrowers can vary as changes in lending enforcement and CRA supervision activities generaleconomicconditionsencourageonegroupor and for consumer financial education efforts. The another to be relatively more active in the homebuyrelease of the 2005 HMDAdata adds a time dimen- ing or refinancing markets or to alter the types of sion to the analysis that can be undertaken because refinancings that are undertaken (for example, the data users now have two years of loan-pricing infor- share that involves cashing out equity). The creditmation at their disposal. However, caution is war- risk profiles of borrowers may also be affected by ranted,asthedifferentinterestratesituationsin2004 local economic conditions. For example, when local and2005affectedthereportedpricingdatainimpor- house prices are high relative to incomes or rise tantways. rapidly, more borrowers may have to stretch financiallytoqualifyforloans,andtheresultisanincrease The Interest Rate Situation and the Reporting inthepoolofborrowerswithhighDTIorLTVratios, of Higher-Priced Loans bothofwhicharerelatedtoelevatedcreditrisk. The third factor is the interest rate situation— Year-to-year changes in the number or proportion of specifically,therelationshipbetweenshort-andlongloans with prices that exceed the thresholds for term interest rates. Generally, interest rate changes reporting price information under HMDA must be can significantly affect whether loans are reported as interpreted with great care. It is tempting to assume higher priced but are less likely to affect the creditthat a change in the incidence of higher-priced risk component of loan pricing. The credit-risk comlending from one year to the next simply reflects ponent can be affected if interest rate movements changes in the volume of subprime lending activity. influencetheloan-productmixthatborrowersuse:In This simple interpretation ignores a number of facsome years, for example, adjustable-rate loans may tors that may influence the incidence of reported berelativelymoreattractivethanfixed-rateloans. higher-priced lending.An important consideration is the difference between the criteria used to distin- The Interest Rate Situation and the Yield Curve guish loans that are reportable under HMDA as higher priced and the factors that truly reflect the The yield curve displays how the yield on financial elevated credit risks or costs associated with instruments, such as U.S. Treasury securities, varies subprime lending. The difference means that there is with maturity and, therefore, reflects the relationship betweenshort-andlong-terminterestrates.Theyield whichloanstheymaypurchase.Loan-sizelimitsfor2005,byproperty curveistypicallyupwardsloping—thatis,short-term size, were as follows: one-family unit, $359,650; two-family unit, rates are typically lower than long-term rates. It $460,400;three-familyunit,$556,500;andfour-familyunit,$691,600. usually has such a slope because longer-term invest- Limitsare50percenthigherinAlaska,Hawaii,theVirginIslands,and Guam. ments ordinarily involve greater risk (credit risk, 29. The ‘‘other’’category includes depository institution holding market interest rate risk, and inflation premium), and companiesandsubsidiariesofdepositoryinstitutionsthatareneither consequently investors require a higher return to be depositoryinstitutionsthemselvesnoraffiliatesofmortgageorfinance companies. willingtoinvesttheirfundsforlongerperiods.

A142 Federal Reserve Bulletinh2006 2. Spread between interest rates on thirty-year and short- and long-term rates can cause a change in the five-year Treasury bonds, 1977–2006 proportionofloansthatarereportedashigherpriced, allotherthingsbeingequal. Percentage points For example, if short-term rates rise relative to 2.0 long-term rates, then the number and proportion of loans reported as higher priced will increase even if 1.5 all other factors that may influence the number and 1.0 proportion of higher-priced loans, such as the busi- .5 ness practices of lenders and the credit-risk profiles + _0 and borrowing practices of borrowers, remain constant. Conversely, if short-term rates fall relative to .5 long-term rates, then the number and proportion of 1.0 loans reported as higher priced will fall even if all 1.5 otherpossiblyinfluentialfactorsremainconstant. 1978 1982 1986 1990 1994 1998 2002 2006 Changes in the Yield Curve from 2004 to 2005 NOTE: After March 2002, the spread is between twenty-year and five-year Treasury bonds. The yield curve at the start of 2004 (the first year SOURCE: Federal Financial Institutions Examination Council, “FFIEC Rate lendersweresubjecttothepricedisclosureprovisions Spread Calculator,” www.ffiec.gov/ratespread/default.aspx. of HMDA) was upward sloping: In mid-January, for Over the past twenty years, longer-term rates (for example, the yield on five-year Treasuries was example, the average annual yield on thirty-year 2.97 percent, and the yield on thirty-year Treasuries Treasury securities) have almost always exceeded was 4.87 percent. Over the course of the year, the shorter-term rates (for example, the average annual difference narrowed as shorter-term rates rose and yieldonfive-yearTreasuries),apatternillustratedby longer-termratesfellslightly.ByearlyJanuary2005, the positive difference in these rates over time (fig- theyieldonfive-yearTreasurieshadrisento3.71perure 2). Sometimes, however, the yield curve is rela- cent,andtheyieldonthirty-yearTreasurieshadfallen tively flat—that is, short-term rates are close to to 4.72 percent. Shorter-term interest rates continued long-term rates; occasionally, the yield curve inverts, to rise through 2005 (4.33 percent at the end of and short-term rates rise above long-term rates. A December), while longer-term rates were essentially review of the rate spreads between five-year and unchanged (4.75 percent). Thus, although at the thirty-year Treasury securities over the past two beginning of 2004 short-term rates were well below decadesindicatesthat2003and2004weresomewhat long-term rates, by the end of 2005 short- and longunusual years by historical standards because the termratesweremuchcloser. yield curve was particularly steep during this time, Becauseofthechangesintherelationshipbetween andconsequentlythegapbetweenlonger-andshorter- short- and long-term interest rates, the gap between termrateswasparticularlylarge. the effective interest rate (measured by the APR on Changes in the shape of the yield curve affect the the loan) on most mortgages and the HMDAthreshreporting of higher-priced loans under HMDA. Be- oldforreportinghigher-pricedloansnarrowedmarkcause most mortgages prepay in a relatively short edly between 2004 and 2005. For example, for loans period (well before the stated term of the loan is priced during the week of January 15, 2004, the reached), lenders use relatively short-term interest average APR on conventional first-lien fixed-rate ratestosetmortgagerates.Forexample,lendersoften thirty-year prime loans reported by Freddie Mac was pricethirty-yearmortgagesaccordingtointerestrates 5.72.30 As a result, a gap of 215 basis points, or 2.15 percentage points, separated the APR of the on maturities of fewer than ten years, and they frequently price certain loan products, such as adjustable-rate mortgages, on the basis of much 30.DataarefromFreddieMac’sPrimaryMortgageMarketSurvey shorter terms than those for fixed-rate loans. But for (PMMS). We calculated the effective rate (orAPR) on the basis of most loans, Regulation C requires lenders to use interest rates and points reported in the survey for conventional first-lienfixed-ratethirty-yearprimeloans.SinceApril1971,Freddie longer-termrates(fortermsoftwentyyearsormore) Machassurveyedlendersweeklytodeterminetheaveragethirty-year todeterminewhethertoreportaloanashigherpriced fixed rate offered to prime consumers during the Tuesday of the becausethestatedmaturityofmostloans,particularly surveyedweek.Currently,125lendersaresurveyedeachweek,and the mix of lender types—thrifts, commercial banks, and mortgage first-lienloans,exceedstwentyyears.Thus,achange lendingcompanies—isroughlyproportionaltothelevelofmortgage fromoneyeartothenextintherelationshipbetween businessthateachtypecommandsnationwide.Overtime,thePMMS

Higher-Priced Home Lending and the 2005 HMDA Data A143 typical prime loan priced that week and the HMDA 3. HMDA price-reporting threshold, interest rates for fixed- and adjustable-rate loans, and spreads between reporting threshold. By December 15, 2005, the gap the threshold and such rates, 2004–05 betweenthecalculatedAPRandtheHMDAthreshold had narrowed to 140 basis points. Although factors Percentage points other than interest rate changes may also have influenced the proportion of higher-priced loans reported APR, HMDA threshold 8.0 underHMDA,thisexampleclearlydemonstratesthat APR, 30-year fixed APR, 1-year adjustable 7.0 even if such factors (including business practices or 6.0 consumercredit-riskprofiles)hadremainedthesame, the proportion of higher-priced loans reported under APR, 5-year adjustable 5.0 HMDAwouldhaveincreasedin2005. Spread, 1-year adjustable 4.0 Although the year is not complete, the yield curve 3.0 for 2006 has experienced further flattening and, if Spread, 5-year adjustable 2.0 otherconditionsremainthesame,willlikelyresultin Spread, 30-year fixed anevengreaterincidenceofhigher-pricedlendingas 1.0 definedbyRegulationC.Throughmid-July2006,the 2004 2005 gap between the calculated APR on conventional first-lien fixed-rate thirty-year prime loans reported NOTE: For explanation of HMDA price-reporting threshold, refer to text. Threshold and annual percentage rates (APRs) are for conventional first-lien by Freddie Mac and the HMDA threshold had de- thirty-year prime loans. creasedtoabout120basispoints. SOURCE: APRs are estimated from Freddie Mac, Primary Mortgage Market Survey. The Interest Rate Situation and the Relative Thelikelyresultoftheflatteningoftheyieldcurve APRs of Fixed- and Adjustable-Rate Loans was an increase in the proportion of adjustable-rate loansthatexceededtheHMDAprice-reportingthresh- The Federal Reserve Board’s Regulation Z requires olds. The increase occurred because many relatively that,incalculatingtheAPRforadjustable-rateloans, high-rate adjustable-rate loans that would not have lenders assume that the interest rate situation at the been reported as higher priced in 2004 because of time of origination will continue for the term of the comparatively low APRs were reported that way in loan. When the yield curve is steep, it suggests that 2005. themarketexpectsshort-terminterestratestorise,yet To illustrate this effect, we show the APRs of the theAPRcalculationforadjustable-rateloansassumes primethirty-yearfixed-rateloans,theprimeone-year thatinterestrateswillstaythesame.31Becauseofthis adjustable-rate loans, and the prime five-year regulatory construct, when the yield curve is posiadjustable-rate loans reported in the Freddie Mac tivelysloped,theAPRsforadjustable-rateloanstend mortgage interest rate survey for 2004–05 (figto be lower than those for fixed-rate loans of similar ure3).33Thebottomthreelinesofthefigurerepresent termandcreditrisk. the differences (gaps) between the effective rates Thus, the flattening of the yield curve over the (APRs) reported by Freddie Mac and the HMDA 2004–05periodhadtwoeffects.First,asnotedearlier, reporting threshold. As noted earlier, the reporting itnarrowedthegapbetweenthelonger-termratesused gap between the typical prime thirty-year fixed-rate for the HMDA reporting threshold and the shorterloan and the reporting threshold narrowed from 215 termratesusedforpricingloans.Second,theflattenbasis points at the beginning of 2004 to 140 basis ing narrowed the APR gap between adjustable- and pointsattheendof2005.Forone-yearadjustable-rate fixed-rate loans because, as short-term interest rates loans, the gap narrowed much more, from 404 basis increased, it reduced the effect of the comparatively pointsatthebeginningof2004toonly75basispoints lowAPRcalculationsforadjustable-rateloans.32 attheendof2005. Although the differences between the APRs on hasexpandedtoincludeothertypesofloans.Formoreinformation, fixed- and adjustable-rate loans and the reporting refertowww.freddiemac.com/pmms/pmms_archives.html. threshold decreased for both types of loans, the 31. Under Regulation Z, borrowers are provided a variety of disclosures explaining the possibility of a rise in loan rates, the possiblesizeoftheincrease,andthecircumstancesunderwhichan 33. The Freddie Mac series for five-year adjustable rates did not increasemightoccur. beginuntilJanuary1,2005.For2004,weshowestimatesforfive-year 32.Theflatteningoftheyieldcurveactuallyhadathirdeffect:It adjustable rates based on a statistical model using the one-year alsocausedageneralincreaseintheinterestratesonadjustable-rate adjustableratesandthirty-yearfixedratesreportedinFreddieMac’s mortgages.Thisriseinrealratesforadjustable-rateloansmayhave PrimaryMortgageMarketSurveyandtheone-andfive-yearratesfor affectedborrowerbehavior. Treasurysecurities.

A144 Federal Reserve Bulletinh2006 decrease for adjustable-rate loans was much larger. have differed geographically, as rates of home-price Thus, the gap between the APRs on fixed- and appreciation and the levels of home prices varied adjustable-rate loans, which was substantial at the across the country.Analysis of HMDAdata provides beginning of 2004, had been virtually eliminated by support for this conjecture, as it shows a positive the beginning of 2005. This finding suggests that, as correlation between the rate of house-price appreciaanartifactofregulation,geographicareaswithdiffer- tion in a state and the loan-to-income ratio of homeent percentages of fixed-rate versus adjustable-rate buyers.34 loans might have shown different incidences of Industrysourcesprovidesomesupportfortheview higher-rateloansin2004.Thatis,in2004,areaswith that the incidence of higher-priced lending experilargersharesofadjustable-rateloanslikelyhadfewer encedarealincreasefrom2004to2005.Mostofthe higher-priced loans than areas with larger shares of increaseseemstohavetakenplaceinthenear-prime, fixed-rate loans. This effect should have been much or ‘‘alt-A,’’ market. For example, Inside Mortgage smallerin2005(andinthefirsthalfof2006)because Finance Publications reports that from 2004 to 2005, interestratesonadjustable-andfixed-rateloanswere the subprime share of the overall market rose someclosertogether. what,from18.5percentto20percent.35Butoverthe same period, the near-prime portion of the market rosesubstantiallymore,from7percentto13percent. The Interest Rate Situation and Junior-Lien Loans Incidence of Higher-Priced Lending The effects of the changing yield curve are reflected primarilyamongfirst-lienloans,whichtypicallyhave Mostloansreportedin2005werenothigherpricedas longtermstomaturity.Theeffectonjunior-lienloans definedunderRegulationC,althoughtheincidenceof is likely much less, as these loans typically have higher-priced lending was significantly greater in maturitiesconsiderablyshorterthanthoseoffirst-lien 2005 than in 2004. For 2005, 26.2 percent of all loans and are priced accordingly. Also, the HMDA reportedloans(excludingloanswithapplicationdates price-reporting threshold for junior-lien loans is set before 2004) were higher priced (table 4). This per- 2 percentage points higher than that for first-lien centage represents an increase of nearly 70 percent loans, a fact that may make the price reporting for overthe15.5percentratein2004. junior-lienloanslesssensitivetochangesintheyield Theincidenceofhigher-pricedlendingvariesconcurve. siderably across loan products. First, in almost all cases,government-backedloans—insuredbytheFederalHousingAdministration(FHA)orguaranteedby Real Incidence of Higher-Priced Lending theVeteransAdministration(VA)—havemuchlower Changes in the incidence of higher-priced lending incidences of higher-priced lending than do compacausedbytheyieldcurveeffectsdescribedearlierare rable conventional loan products. For example, in to a large extent an artifact of the way Regulation C 2005, among first-lien home-purchase loans for sitedefines a higher-priced loan. That is, they reflect built homes, 24.6 percent of conventional loans had changes in the way the threshold andAPRs (particu- APRsabovethepricingthresholdversusonly0.9perlarly for adjustable-rate loans) are computed and not cent of government-backed loans. Second, with few necessarilychangesinthebusinesspracticesoflend- exceptions, first-lien loans have a lower incidence of ers or in the credit-risk profiles or preferences of higher-pricedlendingthandojunior-lienloansforthe consumers. It is difficult to speculate on the impor- samepurposes.Forexample,in2005theincidenceof tanceofthelattertwofactorsinexplainingchangesin higher-priced lending for conventional first-lien refithe ‘‘real’’ incidence of higher-priced lending over nance loans was 25.7 percent, whereas for compatime. rable junior-lien loans it was 30.2 percent. Third, The 2004–05 period was characterized by a rela- manufactured-home loans exhibit the greatest incitively robust housing market without equity declines dence of higher pricing across all loan products, a or economic downturns. However, rapidly rising homepricesinseveralareasofthecountrymayhave 34.Dataonhouse-priceappreciationarefromtheOfficeofFederal HousingEnterpriseOversight(OFHEO).OFHEOestimatesandmakes put upward pressure on LTVand DTI ratios, particupublicly available a quarterly house-price index for single-family larly for first-time homebuyers, many of whom detachedhomes.TheindexusesdatafromFannieMaeandFreddie stretched financially to buy homes. These changes Mac on conventional loan transactions. For details, refer to www.ofheo.gov/hpiabout.asp. may have increased the proportion of homebuyers 35.EstimatesarederivedfromInsideMortgageFinancePublicawho obtained higher-priced loans. The effects may tions,MortgageMarketStatisticalAnnual2006.

Higher-Priced Home Lending and the 2005 HMDA Data A145 result consistent with the elevated credit risk associ- meanspreads.Aswiththepricingofjunior-lienloans, atedwithsuchlending.For2005,nearly60percentof prices on loans for manufactured homes were little the conventional first-lien loans used to purchase changed from 2004, an indication that most of these manufacturedhomeswerehigherpriced. loans have shorter terms to maturity than do most In the secondary market, the vast majority of the first-lienloans. purchases by Fannie Mae and Freddie Mac involved Although the changes in the means and the mediloans with prices below the thresholds for reporting ans are consistent with an upward shift in the distriprice information under HMDA (data not shown in bution of reported interest rates from 2004 to 2005, tables). In total, institutions reporting under HMDA the changes in the distribution of spreads for higherindicatedthat3percentoftheirloansalestothesetwo priced loans are somewhat puzzling. In 2004, for GSEshadinvolvedhigher-pricedloansandthatFan- conventional first-lien products, almost 60 percent of nieMaehadpurchasedthebulkoftheloans.36Other thehigher-pricedloansfellwithin1percentagepoint secondary-market purchasers were active in buying of the reporting threshold, and the percentage dehigher-priced loans, which accounted for more than clined in each subsequent pricing segment (refer to half the sales of private securitization pools; about segment ranges—such as 3–3.99, 4–4.99, and so one-thirdthesalesofinsurancecompanies,mortgage on—in table 4). The pattern was similar to the bankers, finance companies, and credit unions; and truncated upper tail of a normal (bell-shaped) about one-third the sales of ‘‘other’’ purchasers. distribution—that is, the distribution was monotonically declining. For 2005, the pattern was quite Rate Spreads for Higher-Priced Loans different. Only about 27 percent of the higher-priced loans fell within 1 percentage point of the reporting There is considerable variation across loan products threshold, and the percentage increased in the next in the incidence of higher-priced lending, but variatwopricingsegmentsbeforedecliningthereafter.This tionacrossproductsinmeanandmedianAPRspreads nonmonotonicpatternisnotwhatonewouldexpectif as reported in the HMDA data is much smaller. For the changes in interest rates in 2004–05 uniformly example, for 2005, the mean APR spreads reported shifted the distribution of loan rates. The pattern is for higher-priced conventional first-lien loans for the purchase or refinancing of an owner-occupied site- not a consequence of reporting by any one (highly built home were both about 4.8 percent (table 4). active) lender or for any one loan product or area of Reflecting, at least in part, the changing interest rate the country. Interestingly, this pattern does not hold situation, the levels of the average spreads for these for junior liens, which exhibited the same declining two large loan product categories were both about segment share of a truncated normal curve for 2005 70basispointshigherin2005thanin2004. astheydidfor2004. Becausethethresholdforreportingissethigherfor As in 2004, only a very small proportion of the junior liens than for first liens, higher-priced junior- higher-priced first-lien loans reported in 2005 had lien products have higher mean and median spreads spreads that exceeded 7 percentage points. Similarly, thandohigher-pricedfirst-lienloans.However,unlike only a small proportion of most types of junior-lien the average spreads for first-lien loans, those for loanshadspreadsof9percentagepointsormore.For junior liens rose little between 2004 and 2005. As example,amongthehigher-pricedconventionalfirstnotedearlier,thetypicaljunior-lienloanhasatermto lienloansusedtopurchaseowner-occupiedsite-built maturity that is much shorter than that for first-lien homes, only 3.6 percent had spreads that exceeded loans,andsoitsfundingcosttypicallydependsmore 7percentagepoints(in2004,theshareofloansofthis on shorter-term sources of funds; consequently, the type with rate spreads exceeding 7 percentage points flattening of the yield curve had much less effect on was1.4percent).Amongtheconventionaljunior-lien price reporting for junior-lien products. In fact, the loans, only those for home improvement had large mean spreads reported for the refinancing of junior- proportions (about 25 percent) with rate spreads lienloanswereactuallysomewhatlowerin2005than above9percentagepoints. in2004. Loans for manufactured homes differ from other loan products in that they generally have the highest Pre-Approval Programs and Loan Pricing Since 2004, the HMDAdata have included informa- 36.TheroleofFannieMaeandFreddieMacinthehigher-priced tion about certain types of requests for pre-approval portion of the loan market is incompletely measured in the HMDA data,asthedatareflectonlytheirpurchasesofloans. of home-purchase loans. But for purposes of report-

A146 Federal Reserve Bulletinh2006 ing under Regulation C, pre-approval programs per- nearly all their loans to near-prime or subprime tain only to requests for home-purchase loans, and borrowers.However,manyinstitutionsserveabroader consequently the data do not include pre-approval market, including borrowers from the prime and information for applications involving a refinance or nonprime market segments. If one considers a lender home-improvementloan. that devotes 60 percent or more of its business to As with the 2004 data, the data for 2005 indicate higher-priced lending a ‘‘specialist’’ in this business that the incidence of higher-priced lending is notably segment, then among the roughly 1,120 lenders lowerforconventionalloansforsite-builthomesthat reporting at least 100 higher-priced conventional were initiated through a pre-approval program than homeloans,346,or4percentofallreportinginstitufor all such loans. For example, for conventional tions,canbecharacterizedasspecialists.Itshouldbe loans secured by a first lien on a site-built home, the kept in mind that the HMDA data can be used to incidence of higher-priced lending for loans initiated gaugealender’sbusinessfocusonlyroughly,assome through a pre-approval program was 15.2 percent prime loans will exceed the HMDA price-reporting (table 5), whereas the rate for all similar first-lien thresholdandsomesubprimeloansmaynotreachthe conventional loans was 24.6 percent (table 4). The threshold. pattern differs for conventional loans to purchase manufactured homes: Loans initiated through a pre- Loans Covered by HOEPA approval program were more likely to be higher priced. Perhaps those who seek pre-approvals for Under the Home Ownership and Equity Protection manufactured homes are more likely to be stretching Act of 1994 (HOEPA), certain types of mortgage financially and feel a need to provide prospective loans that have rates or fees above specified levels sellers with some assurance that they will qualify for require additional disclosures to consumers and are credit. subject to certain restrictions on loan terms.37 Under For borrowers who received higher-priced loans the 2002 revisions to Regulation C, the HMDAdata for site-built homes, the data do not suggest any indicatewhetheraloanissubjecttotheprotectionsof meaningful differences in actual prices paid, as the HOEPA. meanandmedianspreadswerequitesimilarwhether Coverage under HOEPA is determined by a twoor not a borrower went through a pre-approval pro- part test that considers both the APR and the dollar gram.Forthoseobtainingloanstobuymanufactured amount of points and fees. The APR portion of the homes,themeanandmedianspreadswereabout100 coverage test is similar to the method used to deterbasis points higher for loans initiated through pre- mine which loans are higher priced under HMDA. approvalprograms. The difference relates to the rules for choosing the specific Treasury security to use for determining coverage under the two regulations. In the case of Differences among Lenders in the Propensity HMDA, determining which loans are higher priced to Make Higher-Priced Loans requires using the Treasury security of comparable Asin2004,mostofthenearly8,500lenderscovered maturity for the fifteenth day of the month preceding by HMDA reported extending few if any higher- thedateonwhichtheloanratewasset.ForHOEPA, priced loans in 2005: Nearly 3,200 lenders made no the APR portion of the coverage test requires using such loans, and an additional 2,000 reported only the Treasury security of comparable maturity for the between one and nine higher-priced loans (data not fifteenth day of the month preceding the month in shown in tables). Toward the other end of the spec- which the application was received. Another differtrum, about 1,120 lenders reported making at least enceisthattheAPRspreadsfordeterminingHOEPA 100 higher-priced loans; these more-active lenders coverage are 8 percent and 10 percent for first- and accountedfor98percentofallreportedhigher-priced junior-lienloansrespectively. loans. Moreover, the ten lenders with the largest Before the release of the 2004 data, little informavolume of higher-priced loans extended 59 percent tion was publicly available about the extent of of all such loans, a share that had increased from HOEPA-related lending or the number or type of 38percentin2004. institutions involved in this activity. Although the Lenders extending large numbers of higher-priced loans can be quite different from other lenders in 37.HOEPA,whichisimplementedbytheFederalReserveBoard’s business orientation. Some lenders focus on the RegulationZ,appliestohome-refinanceloansandothernonpurchase higher-priced segment of the market and extend loanssecuredbyaconsumer’sprincipaldwelling.

Higher-Priced Home Lending and the 2005 HMDA Data A147 expanded HMDA data provide important new infor- involved conventional first-lien loans (of these, more mation, the data fail to capture all HOEPA-related than 80 percent were for refinancings), and about lending.SomeHOEPAloansareextendedbyinstitu- 40 percent involved conventional junior-lien loans tionsnotcoveredbyHMDA,andsomeHOEPAloans (morethanhalfofthesewereforrefinancings). madebyHMDA-coveredinstitutionsarenotreported ReportedHOEPAlendingvariesamongborrowers underRegulationC,whichimplementsHMDA.Most sorted by borrower income, race, and ethnicity and notably, if the proceeds of a home-secured loan are among census tracts sorted by census-tract income, not used to refinance an existing home loan or to population, and location. However, the data do not finance home improvement, then the loan may be indicate that HMDA-reportable HOEPA lending is coveredbyHOEPAbutisnotreportableunderRegu- focused primarily on lower-income or minority indilation C.38 The extent of HOEPA-related lending not viduals or on those residing in lower-income neighreportedunderHMDAisunknown. borhoods or neighborhoods with high concentrations of minority individuals. For example, although re- Incidence of HOEPA-Related Lending ported HOEPA loans were extended to borrowers in allincomegroups,nearlytwo-thirdswereextendedto For 2005, more than 1,300 lenders reported nearly middle-andupper-incomeborrowers(datanotshown 36,000 loans covered by HOEPA, an increase of in tables). Similarly, more than 70 percent of the 53 percent from 2004 (table 4). As in 2004, most reported HOEPA loans were extended to nonlendersdidnotreportextendinganyHOEPAloansin Hispanic white borrowers. Most of the homes secur- 2005. For 2005, HOEPA-related lending appears to ing HOEPA loans were in middle- or upper-income have been quite concentrated: The ten lenders that areas, and a large proportion were in areas where the reported the largest number of HOEPA originations minority population was less than 20 percent of the accountedfor70percentofallreportedHOEPAloans census-tractpopulation. (data not shown in tables).At the other extreme, 730 institutionsreportedmakingonlyoneortwoHOEPA loans. PRICING ANALYSIS USING ADJUSTED 2004 Although the incidence of HOEPA-related lending AND 2005 DATA was up significantly over that reported in 2004, such lendingstillaccountedforaverysmallproportionof As discussed earlier, the flattening of the yield curve the market. HOEPA loans accounted for less than over the 2004–05 period affected the proportion of one-half of 1 percent of all the originations of home- loansreportedashigherpricedbecauseofthewaythe secured refinance or home-improvement loans re- price-reportingthresholdandtheadjustable-rateAPR ported for 2005 (data derived from table 4). The are determined. The size of these effects cannot be volumeofHOEPA-relatedlending,likethatofhigher- quantified precisely with the limited information priced lending, was affected by the flattening of the available in the HMDAdata. However, we can comyieldcurvefrom2004to2005.However,itisimpos- pute rough estimates of the magnitude of the yield sible to determine precisely how much of the in- curve effects on the incidence of higher-priced lendcreased volume of HOEPA-related lending was due ing, although our estimates likely understate the to changes in interest rates because, as noted earlier, effects. HOEPAcoverageisbasednotonlyonAPRlevelsbut alsoonthedollaramountofloanpointsandfees. Effects on Loan Pricing of the Method for Setting the HMDA Price-Reporting Threshold Characteristics of HOEPA-Related Lending To estimate the effect on loan pricing of the way the For2005,thevastmajorityofHOEPAloansinvolved HMDA price-reporting threshold is determined, we conventional loan products: Only a very small per- useanadjustedsetofthe2004and2005HMDAdata centage of such loans were government backed. that enables us to identify those loans that exceeded About 60 percent of the reported HOEPA loans the pricing thresholds solely because of a change in the interest rate situation. We separate all reported higher-priced loans into two groups: (1) those that 38.Forexample,ifahomeownertakesoutaHOEPA-coveredloan topayoffoutstandingcreditcarddebtorsomeothertypeofconsumer would have been reported under any interest rate credit and the loan does not involve the refinancing of an existing situation that prevailed during the 2004–05 period home loan or home improvement, then the loan is not covered by and (2) those that were reported only because of the Regulation C and is thus not required to be part of an institution’s HMDAreporting. interestratesituationthatexistedatthetimetheloan

A148 Federal Reserve Bulletinh2006 was made. In separating the higher-priced loans, we and the applicable HMDAthreshold in effect on the assume that a nonprime borrower would receive a datetheloanwasestimatedtohavelocked. loan rate that is no less than a constant markup over We estimate that a loan with an adjusted spread of the rate on a ‘‘prime mortgage’’and that this markup 228 or more basis points above prime would have (discussed below) is independent of interest rates. been reported as higher priced regardless of the date Thus, our exercise is to determine how much above of origination during 2004–05—that is, 228 basis the interest rate for a prime mortgage a cutoff would pointsistheminimumspreadforaloantohavebeen needtobesetsuchthataloanpricedabovethecutoff reported as higher priced during this period. Loans would have been reported under any interest rate with adjusted spreads between 140 basis points and situationprevailingin2004–05.Toconducttheexer- 228 basis points would have been reported as higher cise, we must determine the prime rate that would priced if originated on some days during the period applytoeachloan.Inreality,theprimeratecanvary butnotonothers.Loanswithadjustedspreadsbelow fromdaytodayandfromproducttoproduct,depend- 140 basis points would not have been reported under ingonthetermofthemortgage,thetypeofrate(fixed anycircumstancesduringthistimeframe. or adjustable), the date the loan price was set, the WecomputeincidencesandAPRspreadsfor2004 geographiclocationinwhichtheloanwasmade,and and 2005 that have been ‘‘spread adjusted’’ for otherfactors. changes in the yield curve. These figures are com- Theonlyportionofthisinformationthatisexplicputed in exactly the same way as the overall inciitlyincludedintheHMDAdataislocation.Nevertheless, the necessary information can be approximated. dences and meanAPR spreads, as shown in tables 4 Almost80percentoffirst-lienprimemortgageshave and 5, except that those loans with adjusted spreads a fixed rate of interest, according to LoanPerfor- below 228 basis points are deemed not to be higher mance, and most of these have a thirty-year term to priced. And the adjusted spreads are spreads above maturity.39 The date the loan price was set (the lock the markup over the rate on a prime mortgage rather date) is not reported in the HMDA data, but the than spreads above the yield on a comparableapplication and origination dates are recorded. We maturity Treasury security. By construction, the adapproximate the loan terms and the lock dates for all justed spreads for higher-priced loans have a miniconventional first-lien mortgages by assuming that mumof228basispointsinsteadof300. they are all thirty-year fixed-rate mortgages and that Overall, the incidence of higher-priced lending for the day on which the mortgage pricing was set is conventional home-purchase loans on ownerhalfway between the date of application and the date occupied site-built homes was 11.5 percent in 2004 the loan was originated.40 Because terms vary so and24.6percentin2005,anincreaseof13.1percentmuch for junior-lien loans, we conduct this exercise agepoints(table9).Thespread-adjustedestimatesfor onlyforfirst-lienloans. the same period were 10.4 percent and 21.5 percent To estimate the prime rate, we use the weekly respectively, an increase of 11.1 percentage points. Freddie Mac Primary Mortgage Market Survey. The Thiscomparisonsuggeststhat2percentagepoints,or survey reports the average contract rates and points roughly15percent,ofthetotaldifferenceinreported for all loans and the margin for adjustable-rate higher-priced lending for this product can be attribloans.41 We use this information to estimate the uted solely to the flattening of the yield curve. For average APR for adjustable- and fixed-rate loans refinancingsforsimilarproperties,about2percentage prevailing each week. We calculate the ‘‘adjusted points of the 10.2 percent increase in higher-priced spread’’ for each loan in the HMDA data as the lendingforrefinanceloanscanbeattributedsolelyto difference between the estimated prime fixed APR theyieldcurve. Estimated mean APR spreads are also lower after 39.DatafromLoanPerformancesuggestthatabout90percentofthe spread adjustment for the two conventional loan first-lienloansextendedin2004and2005hadatermofthirtyyears. 40.WithintheHMDAdatafor2004,themediantimebetweenthe products. The mean APR spreads for conventional dateofapplicationandthedateofloanoriginationforconventional first-lien home-purchase loans, and for conventional first-lienhome-purchaseloanswasthirtydays;for2005,thecomparable figure was twenty-eight days. For refinancings, the median first-lien refinance loans, on owner-occupied sitenumbersofdaysfor2004and2005weretwenty-sevenandtwenty-six built homes were both about 4.8 percentage points respectively. Less than 10 percent of home-purchase loans had a before spread adjustment and 3.7 percentage points differenceindatesofapplicationandoriginationofmorethanninety days.Forrefinancings,lessthan10percenthaddifferencesindatesof and 3.8 percentage points respectively using the applicationandoriginationofmorethansixtydays. spread-adjusted 2005 data. The unadjusted spreads 41.Themarginforanadjustable-rateloanisthemarkupabovethe increase about 70 basis points, and the adjusted interestrateestablishedbytheindex(suchastherateforaTreasury security)usedtosetthebaseratefortheloan. spreadsincreaseabout40basispoints.

Higher-Priced Home Lending and the 2005 HMDA Data A149 9. Incidenceofhigher-pricedlendingforfirst-lienloans,andthemeanandmedianAPRspreadsforsuchloans, unadjustedandadjustedforchangesininterestrates,bytypeofhomeandtypeofloan,2004and2005 Percentagepointsexceptasnoted Loanswithannualpercentagerate(APR)spreadabovethethreshold1 2004 2005 Typeofhomeandloan Spreadunadjusted Spreadadjusted Spreadunadjusted Spreadadjusted APRspread APRspread APRspread APRspread Incidence Incidence Incidence Incidence (percent) (percent) (percent) (percent) Mean Median Mean Median Mean Median Mean Median ONE-TO FOUR-FAMILY Nonbusinessrelated2 Owneroccupied Sitebuilt Homepurchase Conventional Firstlien ........ 11.5 4.1 3.8 10.4 3.3 3.1 24.6 4.8 4.7 21.5 3.7 3.6 Governmentbacked Firstlien ........ 1.3 4.2 3.9 1.2 3.5 3.1 .9 3.8 3.3 .4 3.7 3.0 Refinance Conventional Firstlien ........ 15.5 4.2 3.9 14.2 3.4 3.1 25.7 4.8 4.7 22.4 3.8 3.7 Governmentbacked Firstlien ........ 1.5 3.9 3.6 1.4 3.2 2.8 .9 4.3 4.5 .6 3.7 3.4 Homeimprovement Conventional Firstlien ........ 21.9 4.4 4.0 20.3 3.7 3.3 26.3 4.7 4.5 21.5 3.8 3.6 Governmentbacked Firstlien ........ 3.8 4.7 4.0 3.7 3.9 3.2 5.5 4.5 3.9 3.9 3.9 3.5 Manufactured Conventional,firstlien Homepurchase .... 57.1 5.7 5.2 55.5 4.9 4.5 58.3 5.4 4.9 50.4 4.5 4.0 Refinance.......... 47.8 5.0 4.6 45.8 4.2 3.8 55.1 5.0 4.7 46.7 4.1 3.7 Nonowneroccupied3 Conventional,firstlien Homepurchase .... 12.2 4.1 3.8 11.1 3.4 3.0 20.3 4.5 4.3 15.3 3.7 3.5 Refinance.......... 14.0 4.2 3.9 12.9 3.5 3.2 22.5 4.8 4.7 18.8 3.8 3.7 Businessrelated2 Conventional,firstlien Homepurchase .... 9.4 4.4 4.0 8.5 3.7 3.3 11.8 4.4 3.9 8.3 3.7 3.2 Refinance.......... 10.3 4.4 4.1 9.3 3.7 3.3 13.9 4.8 4.6 10.9 3.9 3.7 MULTIFAMILY4 Conventional,firstlien Homepurchase .... 4.5 4.1 3.7 4.1 3.4 3.0 6.0 4.5 4.2 4.3 3.7 3.7 Refinance.......... 4.8 4.1 3.8 4.4 3.4 3.0 6.3 4.4 4.1 4.6 3.7 3.5 Total5 ............... 15.5 4.8 4.3 14.5 4.3 3.6 26.2 5.3 5.1 23.4 4.6 4.2 Note: For definition of higher-priced lending and explanation of spread reportedthattherace,ethnicity,andsexoftheapplicantorco-applicantare adjustment,refertotext. ‘‘notapplicable’’;allotherapplicationsandloansarenonbusinessrelated. 1. APRspreadisthedifferencebetweentheAPRontheloanandtheyield 3. Includesapplicationsandloansforwhichoccupancystatuswasmissing. onacomparable-maturityTreasurysecurity.Thethresholdforfirst-lienloansis 4. Includesbusiness-relatedandnonbusiness-relatedapplicationsandloans aspreadof3percentagepoints. forowner-occupiedandnonowner-occupiedproperties. 2. Business-related applications and loans are those for which the lender 5. Totalisforallsecuredloans,includingjuniorliensnotshownintable. Effects of APR Calculations on Pricing of havebeenifallfirst-lienloansreportedintheHMDA Adjustable-Rate Loans datahadbeenthirty-yearfixed-ratemortgages.However, many loans included in the HMDA data are The spread adjustments just described address only adjustable-rateloans,and,asnotedearlier,theflattentheeffectthattheflatteningoftheyieldcurvehadon ing of the yield curve also affected the gap between the gap between the HMDAreporting threshold and thecalculatedAPRsonadjustable-ratemortgagesand the interest rate at which long-term mortgages are the calculated APRs on fixed-rate mortgages. At the typically priced, approximated by the Freddie Mac beginning of 2004, a one-year adjustable-rate loan primeAPRforthirty-yearfixed-rateloans.Ourspread would have been treated comparably for purposes of adjustment reflects what the yield curve effect would HMDA price reporting to a thirty-year fixed-rate

A150 Federal Reserve Bulletinh2006 10. Incidenceofhigher-pricedlendingforstatesgroupedbytheshareofloansoriginatedthathadanadjustablerate, andthechangeintheincidenceofsuchlending,unadjustedandadjustedforchangesininterestrates, bytypeofloanandbyquintileorstate,2005 Percentexceptasnoted Homepurchase Refinance Quintile1 Change,2004–05 Change,2004–05 or 2005 (percentagepoints) 2005 (percentagepoints) state Spread Spread Spread Spread Spread Spread Spread Spread unadjusted adjusted unadjusted adjusted unadjusted adjusted unadjusted adjusted Lowest ...................... 23.4 19.7 7.8 5.2 33.1 28.0 7.8 4.3 Secondlowest ............... 25.0 21.5 10.0 7.6 30.0 25.7 9.5 6.5 Middle ...................... 21.3 18.8 9.8 8.2 28.2 24.4 10.7 8.1 Secondhighest .............. 23.5 20.2 12.1 10.0 26.5 22.9 11.2 8.9 Highest...................... 24.8 22.0 14.0 12.3 26.2 22.9 10.4 8.5 California2 .................. 31.4 28.7 23.5 22.0 19.0 17.3 10.3 9.6 Total........................ 24.8 21.8 13.2 11.2 26.0 22.6 10.3 8.2 Note: For definition of higher-priced lending and explanation of spread Source: For share of adjustable-rate loans originated, LoanPerformance adjustment,refertotext. (www.loanperformance.com). 1. Basedonshareofloansoriginatedin2005thathadanadjustablerate. 2. Californiaisshownseparatelybecauseitaccountsforalargenumberof loansandhasahighincidenceofadjustable-ratelending. mortgage (of the same term to maturity) with an ‘‘Change, 2004–05’’ in home-purchase section of adjusted spread 200 basis points lower (refer to fig- table), patterns that would have been predicted as ure 3). By the beginning of 2005, this gap had been resulting from the narrowing of the adjustable-fixed virtuallyeliminated.Theimplicationofthisnarrowing APR gap. It is noteworthy that the average spreadof the gap is that, relative to fixed-rate loans, fewer unadjustedincidenceofhigher-pricedhome-purchase adjustable-rate loans would have met the HMDA lending for each of the five quintiles for 2005 was price-reporting thresholds in 2004 than in 2005. almost the same, an indication that the distortions Fullyquantifyingthiseffectwouldbedifficulteven causedbythedifferenceinAPRsbetweenadjustableif the HMDA data distinguished fixed- from and fixed-rate loans had been virtually eliminated adjustable-rate loans.As shown in figure 3, applying during2005.42 the same method to one-year adjustable-rate loans California shows the same pattern as other states thatweemployedforfixed-rateloanswouldnecessi- with a high percentage of adjustable-rate hometate using an adjusted threshold of about 400 basis purchase loans, as it witnessed a significant increase points above the APR on the Freddie Mac prime in the spread-unadjusted incidence of higher-priced one-year adjustable-rate loan. This approach would lending from 2004 to 2005. California’s spreadpotentially exclude a large share of the higher-priced unadjustedincidenceofhigher-pricedlendingin2005 adjustable-rate loans reported under HMDA and (31.4 percent) is substantially higher than that of would reflect only changes at the higher end of the other states with a large proportion of adjustable-rate subprimemarket. loans, but this finding may be due not just to the To provide some rough approximations as to what flattening of the yield curve but also to the effects of theeffectmighthavebeen,weuseinformationonthe borrowersstretchingfinanciallybecauseofhighhome mix of adjustable- and fixed-rate loans for each state prices in California. The pattern for refinancing is asderivedfromtheLoanPerformancedatabase.States different in California because large increases in are arrayed into quintiles based on the percentage of home values are likely to benefit rather than hurt loans originated in 2005 that had an adjustable rate refinancers. (table10).California,whichwouldhavebeenplaced in the quintile with the highest percentage of Effects of Yield Curve Changes on Pricing of adjustable-rate mortgages, is treated as a special Adjustable- and Fixed-Rate Loans categoryandshownseparately.Foreachquintile,we To a limited extent, we can assess the differential calculate the average incidence (spread unadjusted effect of the flattening of the yield curve on the and spread adjusted) of higher-priced lending for 2004 and 2005. For home-purchase loans, the analysisindicatesthatstateswithhighlevelsofadjustable- 42.Thepatternsarenotaspronouncedforrefinanceloans,perhaps rate lending had both relatively low levels of higher- because other factors are more relevant. The rank order of the five quintilesisaspredicted,asisthenarrowingofdifferencesfrom2004 priced lending in 2004 and larger increases in such to 2005; but unlike the results for home-purchase lending, some lending from 2004 to 2005 (refer to data under differencesremainin2005.

Higher-Priced Home Lending and the 2005 HMDA Data A151 incidenceofhigher-pricedlendingamongadjustable- 1 percent to almost 4 percent, during 2005 (the and fixed-rate loans using individual loan data. The thirty-year fixed-rate loans were not pruned to ex- Monthly Interest Rate Survey (MIRS) of the Federal cludenear-primeloans). HousingFinanceBoardisamonthlysurveyofmajor The patterns for the thirty-year fixed-rate and the lenders that collects detailed information on each one-year adjustable-rate loans in the MIRS data are conventionalsingle-familynonfarmloanusedtopur- consistentwithwhatonewouldexpectfromtheyield chaseahomeclosedduringthelastfivebusinessdays curve changes shown in figure 3. We emphasize that of each month.43 The survey includes enough infor- becausetheMIRSdatadonotincludeafullsampling mation to calculate an APR for each loan, to deter- of near-prime and subprime loans, the incidence of mine whether it is an adjustable- or fixed-rate loan, loansinthesamplethatwouldhavebeenreportedas and, among the adjustable-rate loans, to identify the higher priced under HMDA is not representative of typeofloan. all the loans included in the HMDA data. Neverthe- The focus of the survey is on conventional prime less, the analysis here suggests that the flattening of rateloans.For2004,thedataforone-yearadjustabletheyieldcurvehadasignificantlylargereffectonthe rate loans included near-prime loans. But starting reporting of adjustable-rate loans as higher priced with the data for February 2005, most of the nearthan on the reporting of fixed-rate loans as higher primeloanswereexcludedfromthesample.Thus,we priced.Therefore,ourearlierestimateoftheeffectof limittheanalysisoftheone-yearadjustable-rateloans the flattening of the yield curve is likely understated, tothosemadebeforeFebruary2005. perhapssubstantially. Forthefirstfewmonthsof2004,thepercentageof one-year adjustable-rate loans included in the MIRS Differences in Pricing across Geographies data that we estimate would have been reported as higherpricedunderHMDAwasonlyabout1percent. The HMDA data allow analysis of higher-priced This percentage began to rise substantially in the lending along geographic lines. The analysis can be middleof2004,andbytheendoftheyeartheportion conducted by region of the country, metropolitan had risen to more than 18 percent. It appears that, at area,orcensustract. least within the MIRS data, the increase in the incidence of loans that would have been reported as higher priced under HMDA was driven almost en- Region of the Country tirely by the narrowing of the gap between the Interest rates on prime home loans vary across calculated APRs on adjustable- and fixed-rate loans. regions. For example, for 2005, there is a difference The distribution of rates of one-year adjustable-rate loans in the MIRS data relative to the APRs of the of12basispointsbetweentheaverageprimeratefor Freddie Mac prime one-year adjustable-rate loans the region with the highest rates (the North Central remained unchanged—that is, their movements mir- region) and that for the region with the lowest rates roredeachother. (the Southeast) as reported by Freddie Mac in its Thepercentageofthirty-yearfixed-rateloansinthe survey of interest rates (table 11). This variation MIRSdatathatweestimatewouldhavebeenreported likely reflects differences across regions in such facas higher priced under HMDA was little changed tors as prepayment rates, foreclosure laws, originaduring the course of 2004 but rose, from about tioncosts,ordegreeofcompetition. The variation in the incidence of higher-priced lending across regions in both the 2004 and 2005 HMDA data is much larger than might be expected 43. Information collected includes the contract interest rate, fees, loan terms (for example, the LTV ratio and the term to maturity), given the difference in prime rate variation and does propertyvalue,propertytype(newlyconstructedorpreviouslyoccunot show the same rank ordering. For example, the pied unit), loan type (fixed or adjustable rate), and type of lender (savings association, mortgage company, or commercial bank). The Northeast region shows the lowest incidence of dataalsoincludeanestimatedeffectiveinterestrate.Foradjustable- higher-pricedlendingforhome-purchaseloansinthe rate loans, the survey includes information on the annual limit (the 2005 HMDA data and the West region the highest. ‘‘cap’’)onhowmuchtheinterestratemayincrease,themargin,and theindexusedtosetthecontractinterestrate.Thesurveyexcludes Further, the variation in the incidence of higher- FHA-insuredandVA-guaranteedloans,multifamilyloans,andmobilepricedlendingacrossregionsis9percentagepoints,a homeloansandislimitedtohome-purchaseloans.RefertoFederal HousingFinanceBoard,www.fhfb.gov. sizable difference. The large differences in the inci- Thedatainthesurveyreflectthesharesoflendingbylendersize dences of higher-priced lending across regions sugandlendertypeasreportedintheHMDAdata.Althoughthescopeof gest that the regions differ considerably in terms of the survey varies from month to month, it typically covers about 20,000loansandabout100lenders. borrowercredit-riskcharacteristicsorotherfactors.

A152 Federal Reserve Bulletinh2006 11. Interestratesforthirty-yearhome-purchaseloans, above the APR thresholds established by Regulaandthespread-unadjustedincidenceofhigher-priced tionC. lendingforsuchloans,byregionofthecountry, The great variation in the incidence of higher- 2004and2005 priced lending across MSAs is seen in a simple Percent comparison.Ifonefocusesontheincidenceofhigher- 30-yearAPR2 Spread-unadjusted priced lending among conventional first-lien homeincidence Region1 purchase loans on owner-occupied, one- to four- 2004 2005 2004 2005 family, site-built homes, the MSA in the continental Northeast ....... 5.90 5.94 9.3 19.1 United States with the lowest incidence of higher- Southeast ....... 5.86 5.87 14.0 26.1 pricedlendingforthisproductisIthaca,NewYork,at NorthCentral ... 5.96 5.99 12.7 23.9 Southwest....... 5.89 5.92 15.8 26.9 4 percent; the MSA with the highest incidence is West............ 5.89 5.89 8.7 28.2 McAllen-Edinburg-Pharr,Texas,at53percent. Total ........... 5.90 5.92 11.7 24.8 Assessmentofthereasonsforthewidevariationin Note: For definition of higher-priced lending and explanation of spread the incidence of higher-priced lending for homeadjustment,refertotext. 1. DefinedbyFreddieMacasfollows:Northeast:N.Y.,N.J.,Pa.,Del.,Md., purchaseloansacrossMSAsfindsacloseassociation D.C., Va., W.V., P.R., Maine, N.H., Vt., Mass., R.I., Conn., V.I.; Southeast: between the proportion of individuals in an MSA N.C.,S.C.,Tenn.,Ky.,Ga.,Ala.,Fla.,Miss.;NorthCentral:Ohio,Ind.,Ill., Mich., Wis., Minn., Iowa, N.D., S.D.; Southwest: Texas, La., N.M., Okla., county with low credit scores and the incidence of Ark., Mo., Kan., Colo., Neb., Wyo.; West: Calif., Ariz., Nev., Ore., Wash., higher-priced lending in that area.45 Other factors Utah,Idaho,Mont.,Hawaii,Alaska,Guam. 2. Annualpercentagerate(APR)istheaveragefortheyear. positively related to a greater incidence of higher- Source: APRsareestimatedfromFreddieMac,PrimaryMortgageMarket pricedlendingacrossMSAsincludethepercentageof Survey. the MSA’s adult population with less than a high- Metropolitan Area school education, rates of unemployment, and the racial or ethnic makeup of the MSA. Areas with Analysis of loan-pricing patterns for 2004 revealed higherunemploymentratesandlargerminoritypoputhattheincidenceofhigher-pricedlendingforhome- lations are more likely to have higher incidences of purchase loans varied widely across regions of the higher-pricedlending. country and MSAs.44 The 2005 data reveal a similar The geographic pattern in the incidence of higherpattern, with some notable differences. A review of priced lending for refinancings is similar to the pathome-purchase lending at the level of the MSA tern for home-purchase loans, although the proporindicates that, as compared with the 2004 pricing tions of loans with prices above the reporting patterns, areas of the country with the highest inci- thresholds are generally higher for refinancings. The dence of higher-priced lending are not primarily in findingsforMSAsalongthePacificCoast,andinthe state of California in particular, are noteworthy thesouthernandsouthwesternregionsofthecountry because they differ from the general pattern: The but also include a number of MSAs in California incidences of higher-priced lending for refinancings (figure 4). The presence of several MSAs in Califorin the MSAs in this area are typically much lower niaonthelistofareaswitharelativelyhighincidence thanthoseforhome-purchaseloansandforrefinance of higher-priced lending may reflect the effects of loansinotherareas.Asnotedearlier,thispatternmay rapid house-price increases, which result in more reflect,atleastinpart,theneedformanyhomebuyers borrowers in these areas stretching financially to in this region, which has experienced rapid housepurchase homes. The fact that California MSAs price appreciation, to stretch financially to purchase tended to have relatively high proportions of higherhomes, while those refinancing have generally benpriced lending in 2005 but did not in 2004 may also efited from increased home equity as a result of be due to the relatively more frequent use in these home-price appreciation and consequently tend to areas of adjustable-rate loan products that are priced poselesscreditrisk. off of extremely short-term sources of funds, a practice that would, because of the flattening of the yield Census Tract curve, tend to result in large increases from 2004 to 2005 in the number of loans with prices reported Theincidenceofhigher-pricedlendingvariesconsiderably across census tracts, with similar patterns for home-purchase loans (table 12.A.) and refinance 44.Reportinginstitutionsarerequiredtoreportalltheirlendingin MSAsaswellasinthenonmetropolitanportionsofstates.However, becauseinstitutionsoperatingexclusivelyinnonmetropolitanareasare notcoveredbyHMDA,loansinsuchareasareunderrepresentedinthe 45.Thedistributionofcreditscoresbygeographyisalsoconsidered data. For this reason, the geographic analysis here is focused on inMattFellowes(2006),‘‘CreditScores,Reports,andGettingAhead MSAs. inAmerica,’’SurveySeries(Washington:BrookingsInstitution,May).

,ylimaf-ruof ot -eno ,deipucco-renwo no snaol esahcrup-emoh neil-tsrif lanoitnevnoc rof gnidnel decirp-rehgih fo ecnedicnI .4 5002 ,elitniuq yb dna aera lacitsitats natiloportem yb ,semoh tliub-etis tnecreP (cid:28)(cid:24)(cid:17)(cid:25)(cid:20)(cid:177)(cid:26)(cid:27)(cid:17)(cid:22)(cid:3)(cid:3) (cid:24)(cid:27)(cid:17)(cid:28)(cid:20)(cid:177)(cid:19)(cid:25)(cid:17)(cid:25)(cid:20) (cid:21)(cid:23)(cid:17)(cid:22)(cid:21)(cid:177)(cid:25)(cid:27)(cid:17)(cid:28)(cid:20) (cid:26)(cid:24)(cid:17)(cid:27)(cid:21)(cid:177)(cid:22)(cid:23)(cid:17)(cid:22)(cid:21) (cid:23)(cid:25)(cid:17)(cid:21)(cid:24)(cid:177)(cid:27)(cid:24)(cid:17)(cid:27)(cid:21) Higher-Priced Home Lending and the 2005 HMDA Data A153

A154 Federal Reserve Bulletinh2006 12. Incidenceofhigher-pricedlending,unadjustedandadjustedforchangesininterestrates,forloansonone-tofourfamilyhomes,andthechangeinsuchincidence,bycharacteristicofborrower,loan,andcensustractandbytypeof lenderandlocationofproperty,2005 A. Homepurchase,owner-occupiedsite-builthome Percentexceptasnoted Conventional,firstlien Numberofloans Incidence Characteristicandstatus Change,2004–05 2005 Percentage (percentagepoints) 2005 change, 2004–05 Spread Spread Spread Spread unadjusted adjusted unadjusted adjusted Borrower Incomeratio(percentofareamedian) Lessthan80 ................................................ 988,156 5.6 31.9 28.1 14.9 12.5 80–100 ..................................................... 1,079,629 12.1 28.1 24.9 14.9 13.0 100ormore ................................................ 2,014,029 26.9 20.3 18.1 12.4 11.1 Notreported1 ............................................... 176,941 25.8 17.4 9.6 8.1 1.4 Total ................................................... 4,258,755 17.4 24.8 21.8 13.2 11.2 Minoritystatus Minority .................................................... 1,162,532 28.8 40.6 36.3 21.3 19.0 Non-Hispanicwhite ......................................... 2,726,119 12.7 17.3 14.7 8.5 6.8 Missing2.................................................... 370,104 21.5 31.0 28.0 18.7 16.9 Total ................................................... 4,258,755 17.4 24.8 21.8 13.2 11.2 Sex Female ..................................................... 1,001,191 20.1 30.9 27.4 15.5 13.4 Male ....................................................... 1,452,993 23.6 31.9 28.1 16.7 14.3 Joint3....................................................... 1,804,571 11.6 15.8 13.6 8.6 7.1 Total4.................................................. 4,258,755 17.4 24.8 21.8 13.2 11.2 Loan Amount(thousandsofdollars) Lessthan100 ............................................... 731,341 3.6 37.2 32.7 15.2 12.2 100–250 .................................................... 2,169,777 12.3 24.1 21.0 13.3 11.3 250ormore ................................................ 1,357,637 37.4 19.4 17.2 13.4 12.0 Total ................................................... 4,258,755 17.4 24.8 21.8 13.2 11.2 Status Piggyback5 ................................................. 927,451 84.2 53.6 48.5 33.9 31.8 Notpiggyback .............................................. 3,331,304 6.7 16.8 14.3 6.4 4.8 Total4.................................................. 4,258,755 17.4 24.8 21.8 13.2 11.2 TypeofLender,byPropertyLocation Depositorywithinassessmentarea6 .......................... 1,108,769 8.5 7.0 5.1 2.6 1.2 Depositoryoutsideofassessmentarea ........................ 1,535,824 11.7 23.5 20.6 12.8 10.7 LendernotcoveredbyCRA7 ................................ 1,614,162 31.3 38.4 34.3 19.5 17.6 Total4.................................................. 4,258,755 17.4 24.8 21.8 13.2 11.2 LocationofProperty,byFreddieMacRegion8 Northeast ................................................... 915,573 13.6 19.1 16.6 9.8 8.2 Southeast ................................................... 907,695 23.3 26.1 23.0 12.1 10.1 NorthCentral ............................................... 753,081 12.8 23.9 20.3 11.2 8.7 Southwest................................................... 635,200 20.2 26.9 23.3 11.1 8.6 West........................................................ 1,047,206 17.9 28.2 25.5 19.5 18.0 Total4.................................................. 4,258,755 17.4 24.8 21.8 13.2 11.2 loans (table 12.B.). Higher-priced lending is most (table 12.A.). Almost 40 percent of borrowers in common in census tracts with lower incomes, high census tracts with a low percentage of adults with percentages of minorities, depreciating real home schooling beyond high school had higher-priced values,loweducationalattainment,lowcreditscores, loans,comparedwith13percentintractswithahigh and high application denial rates. The variations percentageofresidentswithmorethanahigh-school acrossmostofthesecategoriesarequiteconsiderable. education. For example, in 2005, the incidence of higher-priced Theserelationshipsappeartoberobustandpersist lendingforhome-purchaseloansaveraged47percent evenwhenwecontrolforotherfactors.Forexample, for census tracts in the category with the largest ifacomparisonismadeforcensustractswithsimilar percentage of mortgage borrowers with low credit income levels and minority percentages but with scores, compared with an incidence of only 16 per- varying educational attainment and credit scores, cent for census tracts with a low percentage of wide differences in the incidence of higher-priced mortgage borrowers with low credit scores lendingpersist.Thefactthattherelationshipbetween

Higher-Priced Home Lending and the 2005 HMDA Data A155 12. Incidenceofhigher-pricedlending,unadjustedandadjustedforchangesininterestrates,forloansonone-tofour-family homes,andthechangeinsuchincidence,bycharacteristicofborrower,loan,andcensustractandbytypeoflenderand locationofproperty,2005—Continued A. Homepurchase,owner-occupiedsite-builthome—Continued Percentexceptasnoted Conventional,firstlien Numberofloans Incidence Characteristicandstatus Change,2004–05 2005 Percentage (percentagepoints) 2005 change, 2004–05 Spread Spread Spread Spread unadjusted adjusted unadjusted adjusted CensusTractofProperty Incomeratio(percentofareamedian) Lessthan80 ................................................ 640,985 24.3 41.6 37.2 20.4 17.9 80–119 ..................................................... 2,099,773 19.2 26.6 23.3 13.8 11.7 120ormore ................................................ 1,517,997 12.6 15.3 13.2 8.7 7.3 Total4.................................................. 4,258,755 17.4 24.8 21.8 13.2 11.2 Racialorethniccomposition (minoritiesaspercentageofpopulation) Lessthan10 ................................................ 1,395,467 14.5 17.5 14.9 8.3 6.6 10–50 ...................................................... 2,159,244 17.9 23.2 20.3 12.7 10.8 50–100 ..................................................... 704,044 22.0 44.4 39.9 23.5 21.2 Total4.................................................. 4,258,755 17.4 24.8 21.8 13.2 11.2 Location Centralcity ................................................. 2,263,107 16.6 23.3 20.4 12.7 10.9 Noncentralcity.............................................. 1,622,306 17.8 26.8 23.7 14.3 12.5 Ruraloronlystateknown ................................... 373,342 21.4 25.8 21.8 10.3 7.4 Total4.................................................. 4,258,755 17.4 24.8 21.8 13.2 11.2 Creditscoreofborrowers (percentofmortgageborrowerswithscoresbelow600)9 20ormore.................................................. 567,150 28.5 47.1 42.0 18.1 15.1 10–19 ...................................................... 1,453,580 22.7 29.6 26.0 15.2 13.0 Lessthan10 ................................................ 2,238,025 11.9 16.1 13.9 9.8 8.4 Total4.................................................. 4,258,755 17.4 24.8 21.8 13.2 11.2 Educationalattainmentofresidents (percentofadultswithhigh-schooleducationorless) 30orless ................................................... 1,117,002 9.3 13.0 11.2 7.8 6.6 31–60 ...................................................... 2,388,381 19.1 25.6 22.5 13.7 11.7 Morethan60 ............................................... 753,372 25.9 39.9 35.3 18.0 15.2 Total4.................................................. 4,258,755 17.4 24.8 21.8 13.2 11.2 Denialrateforloantype(percentofapplicantsdeniedcredit) 10orless ................................................... 1,178,823 −13.4 12.6 10.7 6.1 4.9 11–20 ...................................................... 2,161,489 25.5 23.9 20.8 12.0 10.2 Morethan20 ............................................... 918,443 69.4 42.9 38.4 18.5 16.0 Total4.................................................. 4,258,755 17.4 24.8 21.8 13.2 11.2 Realpriceappreciationofrealestate10 Lessthanzero .............................................. 1,913,712 16.5 27.3 24.2 15.3 13.5 0–20........................................................ 1,089,984 18.1 23.9 20.7 11.7 9.7 Morethan20 ............................................... 1,255,059 18.4 21.9 19.0 11.0 9.1 Total4.................................................. 4,258,755 17.4 24.8 21.8 13.2 11.2 Note: Fordefinitionofhigher-pricedloansandexplanationofspreadad- 6. Includes lending by nonbank affiliates in the CRAassessment area of justment,refertotext. depositoryinstitution. Excludestransition-periodloans(thoseforwhichtheapplicationwassub- 7. Includes credit unions and mortgage companies not affiliated with a mittedbefore2004).Fordefinitionofincomecategoriesforborrowerandcen- depositoryinstitutionorbankorthriftholdingcompany. sustract,refertotextnote22.Censustractisforthepropertysecuringthe 8. Freddie Mac defines its regions as follows: Northeast: N.Y., N.J., Pa., loan.ThetermminoritymeansHispanicorLatinoethnicityoranyraceother Del., Md., D.C., Va., W.V., P.R., Maine, N.H., Vt., Mass., R.I., Conn., V.I.; thanwhiteforeithertheborrowerorthecoborrower.Census-tractdatareflect Southeast:N.C.,S.C.,Tenn.,Ky.,Ga.,Ala.,Fla.,Miss.;NorthCentral:Ohio, the2000decennialcensus;theyalsoreflectdefinitionsformetropolitanstatisti- Ind.,Ill.,Mich.,Wis.,Minn.,Iowa,N.D.,S.D.;Southwest:Texas,La.,N.M., calareasestablishedbytheOfficeofManagementandBudgetinJune2003 Okla., Ark., Mo., Kan., Colo., Neb., Wyo.; West: Calif., Ariz., Nev., Ore., andusedinHMDAforthefirsttimeinthe2004data. Wash.,Utah,Idaho,Mont.,Hawaii,Alaska,Guam. 1. Informationforincomewasnotreportedontheapplication. 9. Includes all borrowers with an outstanding mortgage regardless of the 2. Informationforthecharacteristicwasmissingontheapplication. yearinwhichtheloanwastakenout. 3. Ontheapplicationsfortheseloans,oneapplicantreported‘‘male,’’and 10. Basedonthechangeinmedianhomevaluesforaconstant2000-defined the other reported ‘‘female.’’For female and for male, only sole applicants geography. wereconsidered. Source: For Freddie Mac data, Primary Mortgage Market Survey; for 4. Excludesloansforwhichtheinformationforthecharacteristicwasmiss- census-tractcharacteristics,the1990and2000decennialcensuses;forcreditingontheapplication. scoredata,oneofthethreenationalcredit-reportingagencies. 5. Fordefinitionofpiggyback,refertotext. the denial rate and the incidence of higher-priced high correlation between the incidence of higherlending persists after credit scores of tracts are con- pricedlendingandelevateddenialratesisduetoboth trolled for is noteworthy. Certainly, a portion of the factors being related to borrower indicators of ele-

A156 Federal Reserve Bulletinh2006 12. Incidenceofhigher-pricedlending,unadjustedandadjustedforchangesininterestrates,forloansonone-tofourfamilyhomes,andthechangeinsuchincidence,bycharacteristicofborrower,loan,andcensustractandbytypeof lenderandlocationofproperty,2005—Continued B. Refinance,owner-occupiedsite-builthome Percentexceptasnoted Conventional,firstlien Numberofloans Incidence Characteristicandstatus Change,2004–05 2005 Percentage (percentagepoints) 2005 change, 2004–05 Spread Spread Spread Spread unadjusted adjusted unadjusted adjusted Borrower Incomeratio(percentofareamedian) Lessthan80 ................................................ 1,410,566 −7.2 35.2 30.8 12.0 9.3 80–100 ..................................................... 1,481,181 −3.1 29.4 25.7 11.8 9.6 100ormore ................................................ 2,226,956 −.1 20.1 17.5 9.4 7.7 Notreported1 ............................................... 275,362 −6.3 7.3 4.8 4.3 2.1 Total ................................................... 5,394,065 −3.2 26.0 22.6 10.3 8.2 Minoritystatus Minority .................................................... 1,291,246 4.5 35.5 31.5 14.4 12.2 Non-Hispanicwhite ......................................... 3,433,488 −5.5 21.1 18.0 8.1 6.1 Missing2.................................................... 669,331 −4.9 32.5 28.9 12.6 10.4 Total ................................................... 5,394,065 −3.2 26.0 22.6 10.3 8.2 Sex Female ..................................................... 1,213,093 4.8 31.2 27.5 11.3 9.1 Male ....................................................... 1,539,606 4.5 30.3 26.6 11.9 9.8 Joint3....................................................... 2,641,366 −10.2 21.1 18.0 8.5 6.5 Total4.................................................. 5,394,065 −3.2 26.0 22.6 10.3 8.2 Loan Amount(thousandsofdollars) Lessthan100 ............................................... 1,101,140 −23.9 33.4 28.7 9.2 5.9 100–250 .................................................... 2,709,682 −5.6 26.8 23.3 12.2 10.0 250ormore ................................................ 1,583,243 26.4 19.3 17.1 11.3 10.1 Total ................................................... 5,394,065 −3.2 26.0 22.6 10.3 8.2 TypeofLender,byPropertyLocation Depositorywithinassessmentarea6 .......................... 1,433,289 −14.8 9.2 6.6 3.5 1.4 Depositoryoutsideofassessmentarea ........................ 1,920,562 −1.8 24.8 21.2 9.9 7.4 LendernotcoveredbyCRA7 ................................ 2,040,214 5.6 38.9 35.1 13.9 12.3 Total4.................................................. 5,394,065 −3.2 26.0 22.6 10.3 8.2 LocationofProperty,byFreddieMacRegion8 Northeast ................................................... 1,318,512 .6 25.1 21.9 10.5 8.4 Southeast ................................................... 910,382 7.2 31.5 27.3 9.1 6.3 NorthCentral ............................................... 967,734 −13.1 28.2 24.0 11.7 8.7 Southwest................................................... 559,113 −16.4 31.5 27.1 10.3 7.3 West........................................................ 1,638,324 .5 20.3 18.2 10.4 9.4 Total4.................................................. 5,394,065 −3.2 26.0 22.6 10.3 8.2 vated credit risk. However, the persistence of this characteristics (tables 12.A. and 12.B.). Lowerpatternwhencreditscoresarecontrolledforsuggests income borrowers, borrowers with loan amounts that high denial rates may also increase the per-loan below$100,000,borrowersusingpiggybackloansor costs of loan origination and thus may influence loans from lenders not covered by the CRA, and pricingacrosscensustracts.46 minorityborrowersallshowelevatedlevelsofhigherpricedlendinginthe2005HMDAdata.Thesefactors reflect more than census-tract characteristics, as they Differences in Pricing by Characteristic of show the same variation within census tracts of Borrower, Loan, and Lender similarincomes.Forexample,theincidenceofhigherpriced lending for home-purchase loans averaged There is considerable variation in the incidence of about 37 percent for loans below $100,000 but only higher-priced lending by borrower, loan, and lender 19 percent for loans of at least $250,000. Although HMDAdata do not contain sufficient information to 46.TheHMDAdataindicatethatsucharelationshipmayhold,as explainthelatterpattern,onehypothesisisthatsome lenderswhoservemainlyborrowersinthehigher-pricedmarkethave ofthevariationmaybeduetothefactthatindividuals originationratesthatareabout20percentlowerthanthoseoflenders who borrow small amounts may be more likely to whoserveprimarilyborrowersreceivingloanswithratesbelowthe HMDAprice-reportingthresholds. have riskier credit attributes, such as lower credit

Higher-Priced Home Lending and the 2005 HMDA Data A157 12. Incidenceofhigher-pricedlending,unadjustedandadjustedforchangesininterestrates,forloansonone-tofour-family homes,andthechangeinsuchincidence,bycharacteristicofborrower,loan,andcensustractandbytypeoflenderand locationofproperty,2005—Continued B. Refinance,owner-occupiedsite-builthome—Continued Percentexceptasnoted Conventional,firstlien Numberofloans Incidence Characteristicandstatus Change,2004–05 2005 Percentage (percentagepoints) 2005 change, 2004–05 Spread Spread Spread Spread unadjusted adjusted unadjusted adjusted CensusTractofProperty Incomeratio(percentofareamedian) Lessthan80 ................................................ 896,846 5.0 39.6 35.1 13.2 10.6 80–119 ..................................................... 2,798,392 −.6 27.4 23.7 10.3 8.0 120ormore ................................................ 1,698,827 −10.7 16.4 14.1 7.8 6.3 Total4.................................................. 5,394,065 −3.2 26.0 22.6 10.3 8.2 Racialorethniccomposition (minoritiesaspercentageofpopulation) Lessthan10 ................................................ 1,770,723 −7.6 22.4 19.0 8.6 6.3 10–50 ...................................................... 2,516,983 −2.7 23.6 20.5 9.7 7.7 50–100 ..................................................... 1,106,359 3.6 37.1 33.1 14.0 11.9 Total4.................................................. 5,394,065 −3.2 26.0 22.6 10.3 8.2 Location Centralcity ................................................. 2,947,231 −2.7 24.3 21.1 10.3 8.3 Noncentralcity.............................................. 1,911,159 −4.2 27.1 23.8 11.0 8.9 Ruraloronlystateknown ................................... 535,675 −1.9 30.9 26.4 8.4 5.3 Total4.................................................. 5,394,065 −3.2 26.0 22.6 10.3 8.2 Creditscoreofborrowers (percentofmortgageborrowerswithscoresbelow600)9 20ormore.................................................. 758,118 5.0 47.8 42.2 12.8 9.4 10–19 ...................................................... 1,835,281 2.6 30.8 26.8 11.4 8.9 Lessthan10 ................................................ 2,800,666 −8.5 16.8 14.5 8.0 6.6 Total4.................................................. 5,394,065 −3.2 26.0 22.6 10.3 8.2 Educationalattainmentofresidents (percentofadultswithhigh-schooleducationorless) 30orless ................................................... 1,220,761 −14.3 13.4 11.6 6.8 5.6 31–60 ...................................................... 3,037,703 −.1 26.2 22.8 10.4 8.3 Morethan60 ............................................... 1,135,601 2.7 38.7 33.8 11.9 9.0 Total4.................................................. 5,394,065 −3.2 26.0 22.6 10.3 8.2 Denialrateforloantype(percentofapplicantsdeniedcredit) 20orless ................................................... 915,057 −39.1 11.6 10.1 5.7 4.7 21–40 ...................................................... 3,441,463 5.3 25.2 21.9 9.3 7.4 Morethan40 ............................................... 1,037,545 29.9 41.3 35.9 8.5 5.2 Total4.................................................. 5,394,065 −3.2 26.0 22.6 10.3 8.2 Realpriceappreciationofrealestate10 Lessthanzero .............................................. 2,679,049 .4 26.2 23.1 10.9 9.1 0–20........................................................ 1,279,137 −4.5 27.1 23.3 10.1 7.6 Morethan20 ............................................... 1,435,879 −8.1 24.5 21.0 9.5 7.2 Total4.................................................. 5,394,065 −3.2 26.0 22.6 10.3 8.2 Note: Refertonotestotable12.A. scores.47 Another hypothesis is that small loans are lenders that are covered by the CRAand that lend in moreexpensiveonaper-dollarbasistooriginateand their assessment areas than is shown by the same thusaremorelikelytobehigherpriced. lenderswhentheymakeloansoutsideoftheirassess- As in 2004, the data for 2005 continue to show a ment areas.Although the HMDAdata do not contain much lower incidence of higher-priced lending by sufficient information to enable us to determine the causes of this pattern, several hypotheses are possible.Asnotedearlier,onepossibleexplanationforat 47.The hypothesized relationship between credit scores and loan least part of the assessment-area effect may be that amountsisborneoutinthecreditrecordsofanationallyrepresentative sampleofcreditfilesobtainedbytheFederalReserveBoardfromone the channel through which loans are originated matof the three national credit-reporting agencies. Refer to Robert B. ters. Loans extended to borrowers outside an institu- Avery,PaulS.Calem,andGlennB.Canner(2004),‘‘CreditReport tion’s assessment area may be more likely to come Accuracy andAccess to Credit,’’Federal Reserve Bulletin, vol. 90 (Summer),pp.297–322. throughmortgagebrokers,whomaypricedifferently

A158 Federal Reserve Bulletinh2006 or who operate in areas with different market condi- Forexample,accountingforborrowerincomeandfor tions than those faced by institutions that originate loan amount is a measure of the financial burden loans directly. Another possible factor is that these associated with the loan payments, as larger loans brokers serve markets or individuals who are more relative to income imply higher monthly payment costly to serve or whose credit profiles are weaker, burdens(ifweassumethathousingvaluesareproporandthebrokerspriceaccordingly. tionate to income, higher loan-to-income ratios may alsoreflecthigherLTVs).Becausewearefocusingon specific loan products, we are already controlling in Differences in Pricing by Race, Ethnicity, broad terms for loan type and purpose, type of and Sex of Borrower property securing the loan, lien status, and owner- Analysisofthe2004HMDAdatarevealedsubstantial occupancystatus. disparities in the incidence of higher-priced lending In comparing lending outcomes across racial and acrossracialandethniclinesandfurthershowedthat ethnic groups, one can match for the sex of the such differences could not be fully explained by applicant and co-applicant.Accounting for sex in the factors included in the HMDA data. The 2005 data analysisisintendedtobetterdistinguishpricingissues showsimilarpatterns. relatedpurelytotheraceorethnicityoftheborrower Becauseofitsimportance,welookattheincidence from those that may be related to sex. In assessing ofhigher-pricedlendingbyrace,ethnicity,andsexin lendingoutcomesbysex,onecanmatchforraceand a more detailed way than in previous sections. The ethnicity. analysis is more detailed in three respects. First, we The analysis focuses on both the incidence of examine pricing patterns for specific racial, ethnic, higher-pricedlendingandthemeanAPRspreadspaid and gender groups. Second, we examine the inci- by borrowers with higher-priced loans, and we comdence of higher-priced lending (both spread unad- pare these outcomes across eleven groups—nine justed and spread adjusted for changes in the yield racial or ethnic groups and the two sexes. Comparicurve) and the APR spreads (also spread unadjusted sons of average outcomes for each group are made and spread adjusted) paid by those receiving higher- bothbeforeandaftermodifyingtheresultsfordifferpriced loans.Third, and most important, we examine ences in the borrower-related factors cited earlier whether these patterns persist when other factors (income;loanamount;location—MSA—ofthepropincluded in the HMDA data are accounted for. We erty; presence of a co-applicant; and, in the comparirestrict our analysis to conventional first-lien home- sonsbyraceandethnicity,sex)andfordifferencesin purchase and refinance loans on owner-occupied, borrower-related factors plus the specific lending one- to four-family, site-built homes, as these are by institution used by the borrower. Excluded from the far the largest two loan product categories in the pricing analysis are applicants residing outside the HMDAdata.48In2005,home-purchaseandrefinance fifty states and the District of Columbia, applications loan products involved roughly 4.4 million and deemed to be business related, and applications filed 5.5millionloansrespectively. during the transition period. Otherwise, the sample TheHMDAdatadonotincludemanyofthefactors includes all 2005 HMDA loans for the two loan considered in credit underwriting and pricing. How- product categories we examine. Our method of conever, our analysis can include some variables likely trolling for these factors is to group borrowers into related to the loan-pricing process. Specifically, the cells,aswedidinour2005articleassessingthe2004 HMDA data allow an accounting for property loca- HMDAdata.49 tion (for example, same metropolitan area), income Comparisons for lending outcomes across groups relied on for underwriting, loan amount, and time of arediscussedinthefollowingsections.Thecompariyear the loan was made as well as presence of a sons are of three types: unmodified (or ‘‘gross’’), co-applicant.TotheextentthatsomeoftheseHMDA modified for borrower-related factors (or ‘‘borrower factors are not used directly in loan underwriting or modified’’),andmodifiedforborrower-relatedfactors pricing, they are included in the analysis as proxies plus lender (or ‘‘lender modified’’). For purposes of for at least some of the factors that are considered. presentation, the borrower- and lender-modified outcomesshowninthetablesarenormalizedsothat,for the base comparison group (non-Hispanic whites in the case of comparison by race and ethnicity, and 48. In the analysis of the 2004 HMDAdata, we assessed pricing patternsforabroadergroupofloanproductsthanispresentedhere. malesinthecaseofcomparisonbysex),themeanat Wealsoexaminedpatternsforborrowersgroupedbyincome,censustractincome,typeoflender,anddispositionofloan.Wedonotpresent thecorrespondinganalysisfor2005becausethepatternsarelargely 49.Foradescriptionofourapproach,refertoAvery,Canner,and unchangedfrom2004. Cook,‘‘NewInformationReportedunderHMDA,’’pp.387–88.

Higher-Priced Home Lending and the 2005 HMDA Data A159 eachmodificationlevelisthesameasthegrossmean. For 2005, for refinancings, the gross difference Consequently,theborrower-andlender-modifiedout- between blacks and non-Hispanic whites is 28.3 percomes for any other group represent the expected centage points; the difference is reduced to 6.2 peraverageoutcomeifthemembersofthatgrouphadthe centage points after controlling for borrower-related samedistributionofcontrolfactorsasthatofthebase factors plus lender; most of the reduction in differcomparisongroup. ences comes from the addition of the control for lender (table 13.B.). By comparison, in 2004, the Incidence of Higher-Priced Lending by Race and gross difference between blacks and non-Hispanic Ethnicity whiteswas21.7percentagepoints;thedifferencewas reduced to 4.7 percentage points after controlling for The2005HMDAdata,likethe2004data,indicatethat borrower-related factors plus lender, and about twoblack and Hispanic borrowers are more likely, and thirdsofthatreductioncamefromtheadditionofthe Asians borrowers less likely, to obtain loans with controlforlender. prices above the pricing thresholds than are non- The picture forAsians differs greatly from that for Hispanic white borrowers. These relationships hold blacks or Hispanic whites: Compared with nonforbothloanproductsandpersistwhentheincidence Hispanic whites, Asians have a lower gross mean isspreadadjustedfortheeffectsoftheflatteningofthe incidenceofhigher-pricedlendingforhome-purchase yield curve (table 13, sections labeled ‘‘Spread adand refinance loans. The gap is affected some by justed’’).Grossdifferencesintheincidenceofhighercontrolling for borrower-related factors plus lender; priced lending between non-Hispanic whites, on the for home-purchase loans, the incidence of higheronehand,andblacksorHispanicwhites,ontheother, priced lending remains lower for Asians than for are large, but these differences are substantially renon-Hispanic whites; for refinancings, the gap is duced after controlling for borrower-related factors essentially eliminated. Hispanic whites show a patpluslender.Mostofthereductioninthedifferencein tern similar to that of blacks but with smaller differthe incidence across groups comes from adding the encesrelativetonon-Hispanicwhites. control for lender to the control for borrower-related One of the more notable pricing patterns that factors, an indication that the pricing differences emerges is much narrower gaps across racial and within a given lender are typically smaller than the ethnic groups for refinancings as compared with differencesamongloansacrosslenders.50 home-purchase lending. This pattern occurs despite For 2005, for conventional home-purchase loans, the fact that the gross incidence of higher-priced the gross mean incidence of higher-priced lending lending is higher for refinancings for at least some was 54.7 percent for blacks and 17.2 percent for groups,includingnon-Hispanicwhites.Also,thegap non-Hispanic whites, a difference of 37.5 percentage between blacks and non-Hispanic whites is notably points (table 13.A.). Borrower-related factors inlargerthanthatforotherminoritygroups. cluded in the HMDA data accounted for about onefifthofthedifference.Addingtothismodificationthe control for lender reduces the remaining gap to Rate Spreads by Race and Ethnicity 10 percentage points. By comparison, in 2004, the unmodified mean incidence of higher-priced lending The 2005 data, like the data for 2004, indicate that for conventional first-lien home-purchase loans was among borrowers with higher-priced loans, the gross 32.4 percent for blacks and 8.7 percent for non- mean prices paid by black and Hispanic white bor- Hispanic whites, a difference of 23.7 percentage rowers are about the same as those paid by nonpoints. Borrower-related factors accounted for about Hispanicwhiteborrowers(table14).Asianborrowers one-fourthofthedifference.Addingtothismodifica- with higher-priced loans also paid about the same tionthecontrolforlenderreducedtheremaininggap price,onaverage,asnon-Hispanicwhiteswithhigherto7percentagepoints. priced loans. These relationships are consistent for bothtypesofloans. 50. Racial and ethnic differences in higher-priced lending vary substantiallyacrossloanproductcategories(datanotshownintables). Pricing Differences by Sex Forgovernment-backedloanproducts,smallproportionsofborrowers havehigher-pricedloans,andnomeaningfuldifferencesappearacross The 2005 HMDA data reveal little difference in racial and ethnic groups. At the other extreme, the majority of borrowersformanufacturedhomeshavehigher-pricedloans,andfor pricing when borrowers are distinguished by sex. this product significant differences appear across racial and ethnic For example, sole female borrowers generally have groups (although the differences are smaller than for some other a slightly lower incidence of higher-priced lending products).These relationships persist after controlling for borrowerrelatedfactorsandforborrower-relatedfactorspluslender. than sole male borrowers for home-purchase loans

A160 Federal Reserve Bulletinh2006 13. Incidenceofhigher-pricedlending,unadjustedandadjustedforchangesininterestrates,andunmodifiedandmodified forborrower-andlender-relatedfactors,forloansonone-tofour-familyhomes,bytypeofloanandbyrace,ethnicity, andsexofborrower,2004and2005 A. Homepurchase,owner-occupiedsite-builthome Percentexceptasnoted Conventional,firstlien 2005 2004 Race,ethnicity,andsex Modifiedincidence, Modifiedincidence, bymodificationfactor bymodificationfactor Number Unmodified Number Unmodified ofloans incidence Borrower- ofloans incidence Borrower- Borrower- Borrowerrelated related related related pluslender pluslender Spreadunadjusted Race AmericanIndianorAlaskaNative ... 28,107 18.1 17.2 11.8 27,766 35.3 29.5 21.8 Asian ............................... 199,359 5.9 7.4 8.1 237,383 16.6 15.8 16.6 BlackorAfricanAmerican .......... 232,688 32.4 26.7 15.7 312,451 54.7 47.0 27.2 NativeHawaiianorother PacificIslander ................. 20,293 15.7 16.3 11.1 23,450 34.8 30.4 21.0 Twoormoreminorityraces ......... 2,613 22.9 22.2 12.2 2,112 30.4 28.7 20.8 Joint................................ 47,299 6.9 10.8 9.4 51,881 18.2 23.0 19.0 Notavailable........................ 390,136 13.4 16.8 11.1 431,159 32.4 33.6 21.6 Ethnicity Hispanicwhite ...................... 301,915 20.3 16.6 11.6 464,634 46.1 34.2 21.9 Non-Hispanicwhite ................. 2,476,255 8.7 8.7 8.7 2,789,265 17.2 17.2 17.2 Sex Onemale ........................... 1,129,781 15.3 15.3 15.3 1,392,947 31.7 31.7 31.7 Onefemale ......................... 850,213 15.3 14.4 15.0 1,021,006 30.8 29.8 30.8 Twomales .......................... 38,170 9.5 9.5 9.5 44,278 23.1 23.1 23.1 Twofemales ........................ 31,083 10.4 9.0 9.8 36,140 24.7 22.4 23.9 Spreadadjusted Race AmericanIndianorAlaskaNative ... 28,107 16.4 15.6 10.8 27,766 31.2 25.6 18.3 Asian ............................... 199,359 5.1 6.7 7.4 237,383 14.5 13.4 14.1 BlackorAfricanAmerican .......... 232,688 30.0 24.6 14.4 312,451 50.1 42.6 23.3 NativeHawaiianorother PacificIslander ................. 20,293 13.9 14.9 10.3 23,450 30.9 26.9 17.6 Twoormoreminorityraces ......... 2,613 20.2 19.2 10.5 2,112 27.7 25.9 17.7 Joint................................ 47,299 6.1 9.7 8.5 51,881 15.9 20.0 16.2 Notavailable........................ 390,136 12.1 15.4 10.1 431,159 29.3 30.1 18.5 Ethnicity Hispanicwhite ...................... 301,915 17.8 14.8 10.3 464,634 40.8 29.6 18.2 Non-Hispanicwhite ................. 2,476,255 7.9 7.9 7.9 2,789,265 14.6 14.6 14.6 Sex Onemale ........................... 1,129,781 13.8 13.8 13.8 1,392,947 27.9 27.9 27.9 Onefemale ......................... 850,213 13.9 13.0 13.5 1,021,006 27.2 26.4 27.3 Twomales .......................... 38,170 8.5 8.5 8.5 44,278 19.7 19.7 19.7 Twofemales ........................ 31,083 9.5 8.2 8.9 36,140 21.8 19.6 20.6 Note: Excludestransition-periodloans(thoseforwhichtheapplicationwas coborrower.Formethodofallocationintoracialandethniccategoriesanddefisubmittedbefore2004).Fordefinitionofhigher-pricedlendingandexplana- nitionsofcategories,refertotextnote22.Loanstakenoutjointlybyamale tionsofspreadadjustmentandofmodificationfactors,refertotext.Categories andfemalearenottabulatedherebecausetheywouldnotbedirectlycompaforraceandethnicityreflecttherevisedstandardsestablishedin1997bythe rablewithloanstakenoutbyoneborrowerorbytwoborrowersofthesame Office of Management and Budget. The term minority means Hispanic or sex. Latino ethnicity or any race other than white for both the borrower and the after accounting for borrower-related factors plus yield curve had a different effect across racial and lender but a slightly higher incidence for refinanc- ethnicgroupsandconsequentlyaffectedtheobserved ings (table 13). Similarly, few if any differences are gaps in loan pricing from 2004 to 2005. Evidence revealed in the average prices (mean APR spreads) suggeststhatsuchdifferentialyieldcurveeffectsexist paid by those receiving higher-priced loans but were likely not large. For example, for conven- (table 14). tional home-purchase lending, the borrower- and lender-modifiedgapbetweenblackandnon-Hispanic whiteswas7.0percentagepointsin2004and10.0per- Effects of the Yield Curve on Pricing Differences centagepointsin2005(table13.A.).Thecomparable across Racial and Ethnic Groups spread-adjusted gaps are 6.5 percentage points for Animportantquestioniswhethertheflatteningofthe 2004and8.7percentagepointsfor2005.Thefactthat

Higher-Priced Home Lending and the 2005 HMDA Data A161 13. Incidenceofhigher-pricedlending,unadjustedandadjustedforchangesininterestrates,andunmodifiedandmodifiedfor borrower-andlender-relatedfactors,forloansonone-tofour-familyhomes,bytypeofloanandbyrace,ethnicity,and sexofborrower,2004and2005—Continued B. Refinance,owner-occupiedsite-builthome Percentexceptasnoted B.Percentexceptasnoted Conventional,firstlien 2005 2004 Race,ethnicity,andsex Modifiedincidence, Modifiedincidence, bymodificationfactor bymodificationfactor Number Unmodified Number Unmodified ofloans incidence Borrower- ofloans incidence Borrower- Borrower- Borrowerrelated related related related pluslender pluslender Spreadunadjusted Race AmericanIndianorAlaskaNative ... 44,503 20.2 21.0 14.7 37,213 28.9 32.1 24.1 Asian ............................... 207,114 5.9 9.7 12.1 165,011 15.2 18.9 21.1 BlackorAfricanAmerican .......... 391,524 34.6 29.5 17.6 441,299 49.3 45.0 27.2 NativeHawaiianorother PacificIslander ................. 31,381 16.4 18.6 14.5 31,453 28.4 32.2 24.3 Twoormoreminorityraces ......... 5,089 21.1 22.4 15.0 3,650 28.6 29.5 24.2 Joint................................ 67,199 10.4 14.7 13.5 61,200 19.3 26.2 22.4 Notavailable........................ 827,590 19.3 25.4 15.3 752,573 32.2 38.0 24.5 Ethnicity Hispanicwhite ...................... 378,826 19.3 18.5 14.3 478,381 33.8 31.5 23.6 Non-Hispanicwhite ................. 3,698,309 12.9 12.9 12.9 3,496,425 21.0 21.0 21.0 Sex Onemale ........................... 1,360,350 18.6 18.6 18.6 1,424,721 30.3 30.3 30.3 Onefemale ......................... 1,173,835 19.8 18.5 18.7 1,229,138 31.1 30.0 30.4 Twomales .......................... 40,012 12.1 12.1 12.1 37,442 21.2 21.2 21.2 Twofemales ........................ 43,208 17.3 14.5 13.4 41,572 27.0 23.5 22.5 Spreadadjusted Race AmericanIndianorAlaskaNative ... 44,503 18.3 19.3 13.3 37,213 25.1 27.9 20.6 Asian ............................... 207,114 5.2 8.9 11.1 165,011 13.4 16.1 18.1 BlackorAfricanAmerican .......... 391,524 32.3 27.5 16.3 441,299 43.9 39.9 23.3 NativeHawaiianorother PacificIslander ................. 31,381 14.8 17.0 13.2 31,453 25.3 28.5 20.9 Twoormoreminorityraces ......... 5,089 18.7 20.4 13.8 3,650 25.5 25.7 20.2 Joint................................ 67,199 9.4 13.5 12.3 61,200 16.6 22.6 19.2 Notavailable........................ 827,590 17.8 23.6 14.1 752,573 28.6 33.8 20.9 Ethnicity Hispanicwhite ...................... 378,826 17.4 16.8 13.0 478,381 30.1 27.4 20.2 Non-Hispanicwhite ................. 3,698,309 11.8 11.8 11.8 3,496,425 17.9 17.9 17.9 Sex Onemale ........................... 1,360,350 17.0 17.0 17.0 1,424,721 26.8 26.8 26.8 Onefemale ......................... 1,173,835 18.2 16.9 17.2 1,229,138 27.4 26.4 26.8 Twomales .......................... 40,012 11.0 11.0 11.0 37,442 18.2 18.2 18.2 Twofemales ........................ 43,208 16.0 13.3 12.2 41,572 23.4 20.2 19.4 Note: Refertonotetotable13.A. both spread-adjusted gaps are lower than the compa- blacks, and Hispanic whites had higher denial rates rable unadjusted figures suggests that to the extent than non-Hispanic whites; blacks generally had the thattheyieldcurvechangesaffectedthemeasurement highest rates, and Hispanic whites had rates about ofracialandethnicpricingdifferences,theytendedto halfway between those for blacks and those for widengapsratherthannarrowthem. non-Hispanic whites.The pattern was less consistent for Asians, who had higher denial rates than non- Hispanic whites for some loan products but lower Denial Rates by Race, Ethnicity, and Sex ratesforothers. These patterns reflect gross differences in lending Analyses of the HMDA data from different years outcomes but do not account for differences in consistently find that denial rates vary across appli- economic or financial circumstances that may vary cantsgroupedbyraceorethnicity(table15).Foreach across groups. To account for the subset of these loan product category in 2005, American Indians, factors included in the HMDA data, we conducted

A162 Federal Reserve Bulletinh2006 14. MeanAPRspreads,unadjustedandadjustedforchangesininterestrates,andunmodifiedandmodifiedforborrowerandlender-relatedfactors,forhigher-pricedloansonone-tofour-familyhomes,bytypeofloanandbyrace,ethnicity, andsexofborrower,2004and2005 A. Homepurchase,owner-occupiedsite-builthome Percentagepointsexceptasnoted Conventional,firstlien 2004 2005 Race,ethnicity,andsex Modifiedmeanspread, Modifiedmeanspread, bymodificationfactor bymodificationfactor Numberof Numberof Unmodified Unmodified higher-priced higher-priced meanspread Borrower- meanspread Borrowerloans Borrower- loans Borrowerrelated related related related pluslender pluslender Spreadunadjusted Race AmericanIndianorAlaskaNative ... 5,101 4.0 4.1 4.1 9,799 4.6 4.8 4.8 Asian ............................... 11,771 3.8 4.0 4.0 39,471 4.6 4.7 4.7 BlackorAfricanAmerican .......... 75,427 4.2 4.2 4.2 171,009 5.0 4.9 4.9 NativeHawaiianorother PacificIslander ................. 3,186 4.0 4.1 4.1 8,162 4.6 4.8 4.8 Twoormoreminorityraces ......... 598 4.1 4.3 4.1 641 4.8 4.9 4.8 Joint................................ 3,242 4.0 4.1 4.1 9,468 4.6 4.8 4.8 Notavailable........................ 52,094 4.1 4.1 4.1 139,740 4.9 4.9 4.8 Ethnicity Hispanicwhite ...................... 61,248 3.9 4.0 4.1 214,415 4.6 4.7 4.8 Non-Hispanicwhite ................. 216,409 4.1 4.1 4.1 479,338 4.7 4.7 4.7 Sex Onemale ........................... 173,166 4.0 4.0 4.0 441,919 4.8 4.8 4.8 Onefemale ......................... 130,250 4.1 4.0 4.0 313,959 4.8 4.8 4.8 Twomales .......................... 3,632 4.1 4.1 4.1 10,213 4.5 4.5 4.5 Twofemales ........................ 3,246 4.1 4.0 4.1 8,943 4.7 4.6 4.5 Spreadadjusted Race AmericanIndianorAlaskaNative ... 4,603 3.2 3.3 3.3 8,658 3.6 3.8 3.8 Asian ............................... 10,222 3.1 3.3 3.3 34,340 3.5 3.7 3.7 BlackorAfricanAmerican .......... 69,867 3.4 3.4 3.4 156,504 3.9 3.9 3.8 NativeHawaiianorother PacificIslander ................. 2,824 3.2 3.4 3.4 7,243 3.5 3.8 3.8 Twoormoreminorityraces ......... 528 3.3 3.5 3.4 586 3.7 3.8 3.8 Joint................................ 2,885 3.3 3.4 3.3 8,247 3.6 3.7 3.8 Notavailable........................ 47,071 3.3 3.4 3.4 126,398 3.8 3.9 3.8 Ethnicity Hispanicwhite ...................... 53,750 3.2 3.3 3.3 189,768 3.6 3.7 3.8 Non-Hispanicwhite ................. 195,778 3.3 3.3 3.3 408,297 3.7 3.7 3.7 Sex Onemale ........................... 156,146 3.3 3.3 3.3 388,632 3.7 3.7 3.7 Onefemale ......................... 117,763 3.3 3.3 3.3 277,536 3.8 3.7 3.7 Twomales .......................... 3,246 3.3 3.3 3.3 8,706 3.5 3.5 3.5 Twofemales ........................ 2,944 3.3 3.3 3.4 7,874 3.6 3.5 3.5 Note: Spread-unadjustedAPR is the difference between theAPR on the ethnicity reflect the revised standards established in 1997 by the Office of loanandtheyieldonacomparable-maturityTreasurysecurity.Spread-adjusted ManagementandBudget.ThetermminoritymeansHispanicorLatinoethnic- APRisthedifferencebetweentheAPRontheloanandtheestimatedAPR ityoranyraceotherthanwhiteforboththeborrowerandthecoborrower.For reportedbyFreddieMacforathirty-yearfixed-rateloaninitsPrimaryMort- methodofallocationintoracialandethniccategoriesanddefinitionsofcategogage Market Survey. Excludes transition-period loans (those for which the ries,refertotextnote22.Loanstakenoutjointlybyamaleandfemalearenot applicationwassubmittedbefore2004).Fordefinitionofhigher-pricedlending tabulatedherebecausetheywouldnotbedirectlycomparablewithloanstaken andexplanationofmodificationfactors,refertotext.Categoriesforraceand outbyoneborrowerorbytwoborrowersofthesamesex. an analysis analogous to that undertaken in the the specific lender used by the applicant almost pricing discussion.51 always reduces differences further, although large With few exceptions, controlling for borrower- differences remain between non-Hispanic whites and related factors in the HMDAdata reduces the differ- most of the other racial and ethnic groups. For encesamongracialandethnicgroups.Accountingfor example, for conventional first-lien home-purchase loans,thegrossmeandenialratewas27.5percentfor blacks and 12.3 percent for non-Hispanic whites, a 51.Thesamplerulesusedforthedenialrateanalysisareidenticalto difference of 15.2 percentage points.Accounting for those used for the pricing analysis except that transition-period applicationswerenotexcluded. income, loan amount, and other borrower-related

Higher-Priced Home Lending and the 2005 HMDA Data A163 14. MeanAPRspreads,unadjustedandadjustedforchangesininterestrates,andunmodifiedandmodifiedforborrower-and lender-relatedfactors,forhigher-pricedloansonone-tofour-familyhomes,bytypeofloanandbyrace,ethnicity,and sexofborrower,2004and2005—Continued B. Refinance,owner-occupiedsite-builthome Percentagepointsexceptasnoted Conventional,firstlien 2004 2005 Race,ethnicity,andsex Modifiedmeanspread, Modifiedmeanspread, bymodificationfactor bymodificationfactor Numberof Numberof Unmodified Unmodified higher-priced higher-priced meanspread Borrower- meanspread Borrowerloans Borrower- loans Borrowerrelated related related related pluslender pluslender Spreadunadjusted Race AmericanIndianorAlaskaNative ... 8,977 4.1 4.2 4.1 10,770 4.8 4.8 4.8 Asian ............................... 12,250 3.9 4.1 4.1 25,119 4.7 4.8 4.8 BlackorAfricanAmerican .......... 135,467 4.3 4.3 4.3 217,351 5.0 5.0 4.9 NativeHawaiianorother PacificIslander ................. 5,153 4.1 4.2 4.2 8,945 4.8 4.8 4.8 Twoormoreminorityraces ......... 1,072 4.0 4.1 4.1 1,043 4.9 4.9 4.8 Joint................................ 6,973 4.1 4.2 4.2 11,815 4.7 4.8 4.8 Notavailable........................ 159,741 4.2 4.2 4.2 242,666 5.0 5.0 4.8 Ethnicity Hispanicwhite ...................... 73,181 4.0 4.1 4.2 161,713 4.8 4.8 4.8 Non-Hispanicwhite ................. 476,034 4.2 4.2 4.2 733,290 4.8 4.8 4.8 Sex Onemale ........................... 252,618 4.1 4.1 4.1 432,386 4.9 4.9 4.9 Onefemale ......................... 232,583 4.2 4.2 4.1 382,071 4.9 4.9 4.9 Twomales .......................... 4,833 4.2 4.2 4.2 7,937 4.8 4.8 4.8 Twofemales ........................ 7,479 4.3 4.2 4.2 11,208 4.8 4.8 4.8 Spreadadjusted Race AmericanIndianorAlaskaNative ... 8,160 3.4 3.4 3.4 9,354 3.8 3.8 3.8 Asian ............................... 10,867 3.2 3.4 3.4 22,074 3.6 3.8 3.8 BlackorAfricanAmerican .......... 126,314 3.5 3.5 3.5 193,660 3.9 3.9 3.9 NativeHawaiianorother PacificIslander ................. 4,630 3.3 3.4 3.4 7,943 3.7 3.8 3.8 Twoormoreminorityraces ......... 951 3.2 3.3 3.4 929 3.9 3.9 3.9 Joint................................ 6,343 3.3 3.4 3.4 10,139 3.7 3.8 3.8 Notavailable........................ 147,619 3.5 3.5 3.4 215,508 4.0 4.0 3.8 Ethnicity Hispanicwhite ...................... 65,733 3.3 3.4 3.4 143,893 3.7 3.8 3.8 Non-Hispanicwhite ................. 436,611 3.4 3.4 3.4 625,890 3.8 3.8 3.8 Sex Onemale ........................... 231,756 3.4 3.4 3.4 381,119 3.9 3.9 3.9 Onefemale ......................... 214,180 3.4 3.4 3.4 336,179 3.9 3.9 3.9 Twomales .......................... 4,402 3.5 3.5 3.5 6,821 3.8 3.8 3.8 Twofemales ........................ 6,897 3.5 3.5 3.4 9,713 3.8 3.8 3.8 Note: Refertonotetotable14.A. factors in the HMDA data reduces the difference factors plus lender generally has only a small effect 3.1 percentage points. Controlling for borrower- ondifferencesindenialrates. related factors plus lender further reduces the gap to 7.0percentagepoints.Thereductionforconventional Limitations of the HMDA Data in Accounting first-lien refinance loans is similar. The gross differ- for Differences across Groups ence between denial rates for blacks and those for non-Hispanic whites is 15.9 percentage points, a Like the 2004 data, the data for 2005 show large difference cut in half when modified for borrower- differences in the incidence of higher-priced lending relatedfactorspluslender. between minorities and non-Hispanic whites.Analy- With regard to the sex of applicants, sole male sis indicates that the information in the HMDA applicants typically have higher denial rates than data—that is, the data modified for borrower-related femalesdo,butingeneral,thesizesofthedifferences factors plus lender—is insufficient to account fully by sex are small. Controlling for borrower-related for racial or ethnic differences in the incidence of

A164 Federal Reserve Bulletinh2006 15. Denialratesonapplications,unmodifiedandmodifiedforborrower-andlender-relatedfactors,forloansonowneroccupied,one-tofour-family,site-builthomes,bytypeofloanandbyrace,ethnicity,andsexofapplicant,2005 Percentexceptasnoted Conventional,firstlien Homepurchase Refinance Race,ethnicity,andsex Modifieddenialrate, Modifieddenialrate, Numberof bymodificationfactor Numberof bymodificationfactor applications Unmodified applications Unmodified actedupon denialrate Borrower- actedupon denialrate Borrower- Borrower- Borrowerbylender related bylender related related related pluslender pluslender Race AmericanIndianorAlaskaNative .. 41,081 22.4 21.5 17.4 76,922 39.9 40.8 34.1 Asian .............................. 325,881 15.8 14.4 14.3 257,577 23.6 30.0 31.2 BlackorAfricanAmerican ......... 512,130 27.5 24.4 19.3 953,323 43.1 42.5 35.2 NativeHawaiianorother PacificIslander ................ 33,931 19.6 17.3 16.2 54,290 30.9 37.1 32.8 Twoormoreminorityraces ........ 3,052 20.1 19.2 15.7 6,782 36.6 39.0 33.5 Joint............................... 65,752 12.5 15.3 13.6 96,179 27.6 34.8 30.5 Notavailable....................... 672,062 22.5 22.2 17.1 1,824,626 47.7 49.3 34.8 Ethnicity Hispanicwhite ..................... 669,703 21.3 18.0 16.0 807,409 30.5 33.7 31.6 Non-Hispanicwhite ................ 3,490,403 12.3 12.3 12.3 5,482,979 27.2 27.2 27.2 Sex Onemale .......................... 1,944,385 18.9 18.9 18.9 2,671,069 36.2 36.2 36.2 Onefemale ........................ 1,410,239 18.3 17.6 18.0 2,187,420 34.1 33.1 34.3 Twomales ......................... 59,548 17.5 17.5 17.5 63,351 31.9 31.9 31.9 Twofemales ....................... 48,745 17.5 16.1 16.1 71,160 33.0 30.3 30.7 Note: Includes transition-period applications (those submitted before methodofallocationintoracialandethniccategoriesanddefinitionsofcatego- 2004).Forexplanationofmodificationfactors,refertotext.Categoriesforrace ries,refertotextnote22.Applicationsmadejointlybyamaleandfemaleare andethnicityreflecttherevisedstandardsestablishedin1997bytheOfficeof nottabulatedherebecausetheywouldnotbedirectlycomparablewithapplica- ManagementandBudget.ThetermminoritymeansHispanicorLatinoethnic- tionsmadebyoneapplicantorbytwoapplicantsofthesamesex. ityoranyraceotherthanwhiteforboththeborrowerandthecoborrower.For higher-priced lending; significant differences remain uted to this increased incidence of higher-priced unexplained. Similar patterns are shown in racial or lending. ethnicdifferencesindenialrates. The first effect was driven by the flattening of the Theunexplaineddifferencesmaystemfromcredit- yield curve and its relationship to fixed-rate loans. related factors not available in the HMDAdata, such The gap between the APRs on thirty-year fixed-rate as measures of credit history, LTV and DTI ratios, mortgages and the yield on the thirty-year Treasury anddifferencesinloanproducts.Differentialcostsof security used to compute the threshold for higherloan origination may also bear on the differences in priced loan reporting under HMDA narrowed over pricing. Differences in pricing and underwriting outthe 2004–05 period. This narrowing was primarily comes may also reflect discriminatory treatment of driven by rising mortgage rates, though the yield on minority groups. Further research is needed to assess the thirty-year Treasury security did fall slightly the extent to which credit- or cost-related factors during the period. This increase in mortgage rates account for the unexplained differences in loan pricaffectedallmortgageborrowers. inganddenialrates. ThesecondeffectwasacombinationoftheflatteningoftheyieldcurveandanartifactofthewayAPRs CONCLUDING THOUGHTS on adjustable-rate loans are determined. The APRs used to determine whether adjustable-rate loans met Much of the attention paid to the 2005 HMDA data the threshold for being reported as higher priced will likely focus on loan pricing and, in particular, underHMDAwereartificiallylowin2004becauseof on the significant increase in the reported incidence the nature of the formula used to constructAPRs for of higher-priced lending relative to that reported in such loans and the interest rate situation that pre- 2004. For example, the incidence of higher-priced vailedduringtheyear.Bythebeginningof2005,this lending for conventional first-lien home-purchase effect had been largely eliminated because of the loans on owner-occupied, one- to four-family, site- flattening yield curve. For the same credit-risk charbuilt homes rose from 11.5 percent in 2004 to acteristics, adjustable-rate loans would have had 24.6 percent in 2005. At least three effects contrib- higherAPRs in 2005 than in 2004, and consequently

Higher-Priced Home Lending and the 2005 HMDA Data A165 some of them would have surpassed the HMDA HMDA. However, if a lender is required to report, it threshold in 2005, whereas a loan with the same risk must report information on all of its applications and characteristicswouldnothavebeenreportedashigher loans in all locations, including nonmetropolitan pricedin2004. areas. Thethirdfactorinfluencingtheincidenceofhigher- TheFederalReserveBoard’sRegulationCrequires priced lending was borrower- or lender-specific and lenderstoreportthefollowinginformationonhomereflected changes in the risk characteristics of lend- purchase and home-improvement loans and on the ing.Evidenceindicatesthatchangesinriskcharacterrefinancingofsuchloans: istics varied across geographic regions, largely becauseofsubstantialhouse-priceappreciationinsome Foreachapplicationorloan locales, and likely caused more borrowers to stretch (cid:129) applicationdateandthedateanactionwastakenon financially to obtain loans. The substantial growth in theapplication piggyback lending from 2004 to 2005—more than (cid:129) actiontakenontheapplication 84 percent—is consistent with financial stretching. —approvedandoriginated Indeed, the increase in the number of higher-priced —approvedbutnotacceptedbytheapplicant piggybackloansin2005accountedformorethanhalf —denied (with the reasons for denial—voluntary of the increase in the number of all higher-priced forsomelenders) loans. —withdrawnbytheapplicant Allocating the increase in the incidence of higher- —fileclosedforincompleteness priced lending across these three effects is difficult. (cid:129) pre-approval program used (for home-purchase We estimate that 2 percentage points of the 13.1 perloansonly) centage point increase in the incidence of higher- (cid:129) loanamount priced lending for conventional first-lien home- (cid:129) borrowerincomereliedoninloanunderwriting purchase loans on owner-occupied, one- to four- (cid:129) loantype family, site-built homes can be attributed to the first —conventional effect, the narrowing of the gap between mortgage —insuredbytheFederalHousingAdministration rates for fixed-rate loans and the HMDA pricereporting threshold. Although we are unable to esti- —guaranteedbytheVeteransAdministration matetheshareoftheincreasedincidenceattributable —backed by the Farm Service Agency or Rural totheothertwoeffects,ourcomparisonofchangesin HousingService the incidence of higher-priced lending in areas with (cid:129) pre-approvalstatus different mixes of adjustable- and fixed-rate mort- (cid:129) lienstatus gagessuggeststhatthesecondeffect,thereductionin —firstlien the distortion in the APR calculation for adjustable- —juniorlien rateloans,couldbesubstantial. —unsecured (cid:129) loanpurpose APPENDIX: REQUIREMENTS OF —homepurchase REGULATION C —refinance —homeimprovement Under the Home Mortgage DisclosureAct (HMDA), (cid:129) type of purchaser (if the lender subsequently sold lenders use a ‘‘loan/application register’’ (HMDA/ theloan) LAR) to report information annually to their federal supervisory agencies for each application and loan Foreachapplicantorco-applicant actedonduringthecalendaryear.Lendersmustmake (cid:129) race their HMDA/LARs available to the public by March (cid:129) ethnicity 31 following the year to which the data relate, and (cid:129) sex they must remove the two date-related fields to help preserveapplicants’privacy.52 Foreachproperty Only lenders that have offices (or, for nondeposi- (cid:129) location,bystate,county,andcensustract tory institutions, are deemed to have offices) in (cid:129) type metropolitan areas are required to report under —one-tofour-familydwelling —manufacturedhome —multifamilyproperty(dwellingwithfiveormore 52.Lendersmustmaketheirdate-modifiedregisteravailabletothe publicforaperiodofthreeyears. units)

A166 Federal Reserve Bulletinh2006 (cid:129) occupancy status (owner-occupied or nonowner- 2002 revisions to Regulation C, additional items occupied) became subject to reporting beginning with the data collectedfor2004. Informationisalsoreportedonhomeloanspurchased by an institution during the calendar year. Under the

A167 Financial Services Used by Small Businesses: Evidence from the 2003 Survey of Small Business Finances Traci L. Mach and John D. Wolken, of the Board’s changes reported in this article are for the period Division of Research and Statistics, prepared this betweenthe1998and2003surveys.2 article. Courtney M. Carter, John A. Holmes, and The latest survey gathered data from 4,240 firms LieuN.Hazelwoodprovidedresearchassistance. selected to be representative of small businesses operatingintheUnitedStatesattheendof2003.3As Small businesses—nonfarm entities with fewer than in previous surveys, the data show that most busi- 500 employees—are an integral part of the U.S. nesses were very small and were located in urban economy. They account for about half of private- areas.Also as in previous surveys, the percentage of sectoroutput,employmorethanhalfofprivate-sector firms involved in the provision of business and proworkers,andhavegenerated60percentto80percent fessional services increased somewhat, whereas the ofnetnewjobsannuallyoverthepastdecade.1Given percentagesoffirmsengagedinmanufacturingandin thesignificantroleofsmallbusinessesinthenational retailandwholesaletradedeclined.Amongfirmsthat economy, understanding trends in the types and were corporations, those organized under subchapter sources of financing they use is important for eco- SoftheU.S.InternalRevenueCode(Scorporations) nomicresearchandpolicymaking,especiallybecause continued to grow as a proportion of all small busismall businesses typically finance their operations nessesrelativetothoseorganizedundersubchapterC quite differently than large corporations do. For (Ccorporations).4 example,asmallbusinessoftenreliesonthepersonal resources and credit history of the firm’s owners to accesscredit. 2.The1987and1993surveyswerecalledtheNationalSurveyof SmallBusinessFinances.Forsummariesoftheearliersurveys,refer Newly available data from the Federal Reserve tothefollowingarticles:GregoryE.ElliehausenandJohnD.Wolken Board’s 2003 Survey of Small Business Finances (1990),‘‘BankingMarketsandtheUseofFinancialServicesbySmall (SSBF) provide detailed information on the use of and Medium-Sized Businesses,’’ Federal Reserve Bulletin, vol. 76 (October), pp. 801–17; RebelA. Cole and John D. Wolken (1995), creditandotherfinancialservicesbythesefirms.The ‘‘Financial Services Used by Small Businesses: Evidence from the SSBF is the most comprehensive source of informa- 1993NationalSurveyofSmallBusinessFinances,’’FederalReserve tion available on the characteristics of small busi- Bulletin,vol.81(July),pp.629–67;andMarianneP.Bitler,AliciaM. Robb,andJohnD.Wolken(2001),‘‘FinancialServicesUsedbySmall nessesandtheirowners;noothersurveyprovidesthe Businesses: Evidence from the 1998 Survey of Small Business breadth and detail of information for a nationally Finances,’’Federal Reserve Bulletin, vol. 87 (April), pp. 183–205. representative sample of such firms. Moreover, poli- InformationontheavailabilityoftheSSBFdataaswellastechnical information, data from previous surveys, and a bibliography of cymakers and researchers can compare the newest researchusingtheSSBFareavailableonthewebsiteoftheFederal SSBF data with results from the previous surveys, ReserveBoard,atwww.federalreserve.gov/pubs/oss/oss3/nssbftoc.htm. which cover 1987, 1993, and 1998. Most of the 3. Interviewing began in mid-2004 and for the most part was completedbyyear-end.Firmswereaskedtoreportbalancesheetand incomedataforthefirm’sfiscalyearthatendedbetweenMay1,2003, and April 30, 2004; other data were reported as of the date of the 1.Theseproportionsarerelativetothenonfarmsectorandarefrom interview.Resultsfromthe2003surveyarereferredtointhisarticleas ‘‘Frequently Asked Questions’’ at the website of the U.S. Small 2003 data. Further information on the survey’s methodology is in Business Administration (SBA), Office of Advocacy, http:// appendixA. app1.sba.gov/faqs/faqIndexAll.cfm?areaid=24.Forresearchpurposes, 4.Theorganizationalformshavedifferentrulesaboutliabilityand theSBAOfficeofAdvocacydefinesasmallbusinessasanindepen- taxes. In sole proprietorships (hereafter, proprietorships) the owners dentfirmhavingfewerthan500employees.Forpurposesofcontract- receivealltheincomefromthebusinessandbearfullliabilityforits ingwiththefederalgovernment,smallbusinessesaredefinedbythe obligations. Partnerships must have more than one owner. As in SBAOfficeofSizeStandards,andthedefinitionvariesbyindustryand proprietorships,thepartnersreceivealltheincomefromthebusiness insomecasesisbydollarvalueofsalesorofassetsratherthannumber and, in general, are fully liable for its obligations. Corporations are of employees. This article will in some instances also use sales or separate legal entities, and the owners’ liability is limited to the assetstodefinesubsetsofsmallbusinesses. amount of their original equity investment. The primary difference

A168 Federal Reserve Bulletinh 2006 Thefinancialaffairsofsmallbusinessin2003were thefindingsmayrequireamore-detailedandin-depth conducted in a financial marketplace whose analysis than is possible in this article. To facilitate elements—including regulations, technology, and or- additional research, a micro-level data set for public ganizationalstructures—havechangedmarkedlysince use will be released shortly after the publication of the Federal Reserve Board’s first small business this article.6These data will permit a rigorous analysurvey. For example, state and federal restrictions on sis that takes into account characteristics of the interstate branching and banking have been relaxed, businesses, their owners, and local banking markets. certain financial institutions are now permitted to Researchers will be able to study many aspects of offer a wider range of financial services, lenders small business finance, including, for example, how employ complex credit-scoring models to evaluate the proximity of financial institutions affects the mix would-be borrowers, and mergers and acquisitions offinancialproductsthefirmuses,whichcharacterishave produced a financial industry with fewer but tics of firms and owners affect the ability of small largerorganizations. businesses to obtain credit, and how lending patterns In this changing financial marketplace, small busi- varywiththesecharacteristics. nesseshavebeendiversifyingtheirprovidersoffinancialservices.Nondepositoryinstitutionshavebecome ECONOMIC AND FINANCIAL SERVICES increasinglyimportantsourcesoffinancialservicesto ENVIRONMENT small businesses; more than half reported using non- In 1998 the economy was in its seventh year of depository sources in 2003, compared with about sustainedeconomicexpansion.Theannualunemploy- 40 percent in 1998. Among these sources, finance ment rate had fallen to 4.5 percent; the consumer companies and leasing companies were important price index rose 1.6 percent, gross domestic product suppliers of credit and financial management sergrew 4.4 percent, and productivity in the nonfarm vices, especially for the largest small businesses, and businesssectorincreased2.7percent. brokerage firms were important suppliers of broker- In 2003 the economic climate for small busiage and trust and pension services. Nonetheless, nesses was quite different than in 1998. A recession commercialbankscontinuedtobe,byawidemargin, in 2001 was followed by a sluggish recovery. By thesuppliermostcommonlyusedbysmallbusinesses the end of 2003, the pace of economic activity was for checking and savings accounts, for loans other picking up, although many small businesses were than leases and vehicle loans, and for financial manlikely still feeling some effects from the subpar agement services other than brokerage and trust and performance in the preceding few years. Many small pension services. They were the second most combusinesses had failed, and those that had weathered monly reported provider of vehicle loans and trust the period were probably facing declining revenues. andpensionservices. Health-carecostshadincreasedsharply,venturecapi- The types of credit used by small businesses have tal opportunities had declined, and banks had instialso been changing.The percentage of firms that had tuted new fees and raised existing fees and balance outstanding vehicle loans and credit lines increased requirements. At the same time, interest rates in betweenthe1998and2003surveys;theuseofcapital 2003 were lower than they had been in decades; leases declined somewhat; and the use of equipment these low rates made relatively low cost new loans loans,mortgages,andotherloansremainedaboutthe available and provided opportunities for substantial same. The use of personal credit cards for business savings from refinancing. purposesremainedroughlyconstant,whereastheuse These differences in the overall economy between ofbusinesscreditcardsincreasedsubstantially. the two most recent surveys are reflected in the This article focuses on some of the major results problems reported by firms (table 1). In 2003, poor from the 2003 SSBF for broad subgroups of small businesses.5 Understanding and explaining many of sales topped the list, particularly among the smallest firms. In 1998, firms reported that their most importantproblemswerecompetitionfromotherfirmsand between the two types of corporations is how they are taxed: S the quality of labor. The quality of labor remained a corporations are not subject to corporate income tax, whereas C commonlyreportedconcernin2003,especiallyamong corporationsare.Scorporationsarelegallyconstrainedtohavefewer firms with ten or more employees. The 2003 survey thanseventy-fiveshareholders,arerestrictedtooneclassofstock,and mustpassallfirmincometotheownersattheendofeachfiscalyear. also recorded a marked increase in the percentage of 5.Spacelimitationspreventtheinclusionofstandarderrorsforall thestatisticspresentedhere.Althoughwedonotdirectlyaddressthe statisticalsignificanceoftheresults,thearticlehighlightsfindingsthat 6. The data will be available on the Federal Reserve Board’s aresignificantorareinterestinginabroadercontext. website,atwww.federalreserve.gov/pubs/oss/oss3/nssbftoc.htm

Financial Services Used by Small Businesses: Evidence from the 2003 Survey of Small Business Finances A169 1. Mostimportantproblemreportedbysmallbusinesses,bynumberofemployeesinfirm,1998and2003surveys Percent 1998 2003 Problem 9orfewer 10–49 50–99 100–499 9orfewer 10–49 50–99 100–499 Taxes ............................... 7.2 5.5 2.8 4.0 5.4 6.1 1.5 2.9 Inflation............................. .4 .4 .3 .3 1.9 3.7 2.9 2.2 Poorsales........................... 7.6 6.5 7.6 7.1 20.9 13.7 15.4 12.3 Financingandinterestrates .......... 6.9 6.7 4.2 3.2 6.5 3.5 3.0 1.5 Costoflabor........................ 3.2 6.2 14.8 9.7 1.2 1.9 4.1 2.6 Governmentregulationsandredtape .. 6.9 6.0 7.4 8.9 4.1 4.9 8.2 5.0 Competitionfromlargerfirms........ 11.0 12.6 9.7 15.6 3.3 4.5 5.1 4.0 Qualityoflabor ..................... 10.2 23.7 25.6 24.4 4.8 15.3 13.8 15.1 Costandavailabilityofinsurance .... 2.3 2.5 .3 .1 8.5 14.9 13.7 14.6 Other ............................... 44.3 29.9 27.4 26.7 43.4 31.5 32.3 39.8 Total ............................. 100 100 100 100 100 100 100 100 Note: Inthisarticle,numberofemployeesconsistsoffull-andpart-time data reflect the population of small businesses rather than sample measures workers,countedequally,andownerswhoworkinthefirm;itexcludeswork- (moreinformationisavailableintheappendix).Alsohereandinsubsequent ersemployedtemporarilyorunderothernonstandardworkarrangements(refer tables, components may not sum to totals because of rounding, or because to table 5). Here and in subsequent tables except as noted, percentages are somefirmsdidnotansweroranswered“Donotknow,”or,inafewcases,beweightedtoadjustfordifferencesinsamplingandresponserates;theweighted causevaluesforsomevariablesaremissing. firms reporting the cost and availability of insurance tion Act of 1996, which liberalized the rules for astheirmostimportantproblem. subchapterSqualification.8 Service industries (both business and professional CHARACTERISTICS OF SMALL BUSINESSES services) accounted for the largest fraction—46 percent—of small businesses’ primary activities, and Likeitspredecessors,the2003SSBFcollectedawide 18 percent of all firms were primarily in retail trade. varietyofinformationaboutfirmsandowners,includ- Thisdistributionissimilartothatin1998;betweenthe ing the firm’s size, primary industry, and organiza- two surveys, small increases were observed in busitional structure and the owners’ race, ethnicity, sex, nessandprofessionalservicesandsmalldecreasesin andextentofparticipationinthefirm(table2). manufacturingandretailandwholesaletrade. Thecompositionofsmallbusinesseshasremained The geographic distribution of the firms correlargely unchanged between the 1998 and 2003 sur- sponded closely to the distribution of the population: veys. The large majority continued to be very small 35 percent in the South, 24 percent in the West, and owner-managed. More than 80 percent of firms 21 percent in the Midwest, and 20 percent in the employed fewer than ten workers, and less than Northeast.9 About 79 percent of firms had their 3percentemployedfiftyormore.7Morethan70per- headquarters or main office in an urban area, and the cent of firms had annual sales of less than $500,000, remaining 21 percent were in rural areas. The vast and more than 80 percent had assets of less than majorityofthefirms(95percent)conductedbusiness $500,000. Finally, more than 85 percent conducted primarilywithintheUnitedStates,andtheremaining businessoutofasinglelocation,andthevastmajority 5percentoperatedinternationally. ofowners(94percent)managedday-to-dayactivities Number and Ownership Shares themselves. of Small Business Owners In 2003, 47 percent of all small businesses were corporations (31 percent were S corporations and Information on the owners of the firm was collected 16 percent were C corporations), 45 percent were differently for the 2003 survey than it had been proprietorships, and the remaining 9 percent were partnerships. The proportion of S corporations rela- 8. The act increased the number of permitted shareholders from tive to C corporations has grown since 1993, when thirty-fivetoseventy-five;allowedan‘‘electingsmallbusinesstrust’’ theyaccountedfor20percentand30percent,respec- withmultiplebeneficiariestoqualifyasanScorporationshareholder; allowed charitable organizations and qualified retirement plans (but tively, of all small businesses.Aportion of this shift notindividualretirementaccounts)tobeScorporationshareholders; maybeattributabletotheSmallBusinessJobProtec- andallowedcorporationswithsubsidiariestobecomeScorporations (andprovidedthatwhollyownedsubsidiariescouldbeconsideredpart oftheScorporationforfederalincometaxpurposes). 7.Thenumberofemployeesincludespaidandunpaidworkersand 9.In2003,36percentofthepopulationwasintheSouth,23percent ownerswhoworkinthefirm;part-timeandfull-timeworkersareeach intheWest,22percentintheMidwestand19percentintheNortheast. countedasone.Forexample,ifatotaloftwopart-timeworkers,one Geographic areas of the United States cited in this article are as full-timeworker,andanownerworkinthefirm,thefirmisconsidered defined by the U.S. Census Bureau (www.census.gov/geo/www/ tohavefouremployees. us_regdiv.pdf).

A170 Federal Reserve Bulletinh 2006 2. Numberandpopulationproportionofsmallbusinesses 2.—Continued insurveysample,byselectedcharacteristicsoffirms, 2003survey Memo Memo Number Percentage 1998 Number Percentage 1998 Characteristic insample1 ofpopulation percentageof Characteristic insample1 ofpopulation percentageof population2 population2 Allfirms ................... 4,240 100 100 Censusareaofmainoffıce Northeast ................... 756 19.8 18.9 Numberofemployees NewEngland ............. 247 6.0 5.2 0–1......................... 640 20.6 21.9 MiddleAtlantic ........... 509 13.8 13.7 2–4......................... 1,167 40.0 41.5 5–9......................... 632 20.2 19.6 Midwest .................... 1,015 21.1 21.8 10–19....................... 389 10.6 8.8 EastNorthCentral ........ 652 14.2 14.6 20–49....................... 566 6.0 5.6 WestNorthCentral........ 363 6.9 7.2 50–99....................... 444 1.7 1.6 South ....................... 1,386 34.7 32.7 100–499 .................... 402 1.0 1.2 SouthAtlantic ............ 747 18.9 16.9 EastSouthCentral ........ 231 5.3 5.5 Sales(thousandsofdollars) WestSouthCentral........ 408 10.5 10.4 Lessthan25 ................ 430 14.6 16.3 25–49....................... 289 9.9 9.4 West........................ 1,083 24.4 26.6 50–99....................... 350 11.6 14.3 Mountain ................. 344 7.6 6.6 100–249 .................... 598 19.8 21.9 Pacific.................... 739 16.8 20.0 250–499 .................... 459 14.3 13.3 Urbanizationatmainoffıce 500–999 .................... 441 12.2 10.2 Urban....................... 3,350 79.4 79.9 1,000–2,499................. 532 10.0 8.1 Rural ....................... 890 20.6 20.1 2,500–4,999................. 338 3.6 3.3 5,000–9,999................. 319 2.3 1.6 Numberofoffıces 10,000ormore.............. 483 1.7 1.7 One......................... 3,235 86.0 87.8 Two ........................ 474 9.4 8.6 Assets(thousandsofdollars) Threeormore............... 531 4.6 3.6 Lessthan25 ................ 934 31.3 34.8 25–49....................... 372 12.5 12.8 Salesarea 50–99....................... 447 13.5 14.2 Primarilywithinthe 100–249 .................... 573 15.9 15.7 UnitedStates ........... 3,995 95.4 95.5 250–499 .................... 401 10.0 9.0 Internationalorglobal3 ...... 245 4.6 4.5 500–999 .................... 361 7.1 6.1 1,000–2,499................. 439 5.8 4.3 Management 2,500–4,999................. 279 1.9 1.7 Byowner ................... 3,794 94.3 92.5 5,000ormore............... 434 2.0 1.4 Hired ....................... 387 5.8 7.5 Organizationalform Race,ethnicity,andsexof majorityowners Proprietorship ............... 1,347 44.5 49.4 Partnership.................. 344 8.7 7.0 NonwhiteorHispanic ....... 484 13.1 14.6 Scorporation................ 1,548 31.0 23.9 Non-Hispanicwhite ......... 3,697 86.6 85.4 Ccorporation ............... 1,001 15.8 19.8 White....................... 3,853 91.0 90.7 Black ....................... 119 3.7 4.1 StandardIndustrial Asian,NativeHawaiian,or Classification otherPacificIslander.... 170 4.2 4.4 Constructionandmining..... 440 11.8 11.9 AmericanIndianorAlaska Manufacturing............... 499 7.1 8.3 Native.................. 58 1.3 .8 Transportation............... 171 3.8 3.7 Wholesaletrade ............. 288 5.9 7.2 Hispanic .................... 149 4.2 5.6 Retailtrade ................. 821 18.4 19.0 Non-Hispanic ............... 4,032 95.8 94.4 Insuranceandrealestate..... 262 7.2 6.5 Female...................... 783 22.4 24.3 Businessservices............ 934 25.1 24.8 Male........................ 2,923 64.8 72.0 Professionalservices......... 823 20.7 18.5 Ownershipdividedequally bysex.................. 475 12.8 3.7 Yearsundercurrentownership 0–4......................... 686 20.6 22.4 5–9......................... 822 22.1 22.8 1.Unweighted. 10–14....................... 666 16.0 19.2 2.Thepercentagesreportedherearefinalandmaydifferslightlyfromthe 15–19....................... 596 12.6 12.9 preliminary data reported in Bitler, Robb, and Wolken (2001), “Financial 20–24....................... 512 10.9 8.9 ServicesUsedbySmallBusinesses.” 25ormore.................. 957 17.9 14.0 3.InternationalreferstosalesareasoutsidetheUnitedStates;globalrefersto combinedU.S.andinternationalsalesareas. previously.Inthepast,characteristicsofownerswere ferred to in this article as the first owner); if that collected only for the owner with the largest share, individual did not have a controlling interest in the and respondents were asked whether a majority of company (an ownership share of at least 51 percent), firmownerswereHispanic,nonwhite,orfemale.The information was also collected on up to two addi- 2003 survey followed the lead taken by the U.S. tionalowners. CensusBureauinitsSurveyofBusinessOwnersand This new method confirmed the implicit assumpcollected demographic information on up to three tion under which previous information was collected owners.10 Respondents were asked to report first on aboutfirmowners:Smallbusinessesareveryclosely the individual with the largest ownership share (re- held.Theaveragefirmhadonlythreeowners,andthe owner with the largest share held an 81.5 percent interestinthefirm(table3).Thelargestdifferencesin 10.FurtherinformationontheCensusBureausurveyisavailableat www.census.gov/sbo/index.html. ownership dispersion of the firms can be seen across

Financial Services Used by Small Businesses: Evidence from the 2003 Survey of Small Business Finances A171 3. Numberofownersandaverageownershipsharesof characteristic was 51 percent or more, the firm was principalowners,bynumberofemployeesand determinedtobeofthatgroup.12 organizationalformofthefirm,2003survey In 2003, 13.1 percent of firms were owned by nonwhite or Hispanic individuals (table 2); the share Ownershipshare Numberofowners (percent) is statistically lower than that recorded by the 1998 Characteristic First Second survey(14.6percent).Thesharesfornonwhitegroups Median Average owner owner alone did not change by a statistically significant Allfirms ............... 1 3.0 81.5 44.3 amount:TheshareforblacksandtheshareforAsians Numberofemployees eachheldatroughly4percent;13theshareforAmeri- 0–1..................... 1 1.2 94.6 48.3 2–4..................... 1 1.6 82.5 46.5 can Indians and Alaska Natives held at roughly 5–19.................... 2 2.4 75.2 44.0 1 percent. However, the share of Hispanic-owned 20–49................... 2 8.7 69.7 38.6 50–99................... 2 43.8 66.2 35.9 firms fell a statistically significant amount, from 100–499 ................ 3 13.0 62.9 30.2 5.6 percent to 4.2 percent (refer to appendix B for a Organizationalform Proprietorship ........... 1 1.2 93.6 50.0 discussion of changes in the estimated rates of non- Partnership.............. 2 2.9 52.3 42.9 whiteandHispanicownership). Scorporation............ 2 2.0 76.6 44.5 Ccorporation ........... 2 10.2 73.0 41.0 The largest change in ownership composition in 2003 was among firms owned equally by males and Note:Thesurveydesignatestheownerwiththelargestshareasthefirst owner,andtheownerwiththesecond-largestshareasthesecondowner. females. The proportion of such firms rose sharply, from 3.7 percent in 1998 to 12.8 percent in 2003, organizational type and firm size. Among partneralthoughpartofthisincreasemaystemfromchanges ships, the average firm had 2.9 owners, and the in how the question was asked.14 This increase is partner with the largest share controlled 52.3 percent reflectedinthedeclineinthepercentageoffirmsthat of the firm. Compared with partnerships, C corporawereownedbymales,from72.0percentto64.8pertions had more owners (10.2 on average), but the cent; the percentage of firms owned by females also largestownerheldalargershareofthefirm(73.0perdeclined between the two surveys, but much less— cent). S corporations had 2 owners on average, with from24.3percentto22.4percent. thelargestshareholdercontrolling76.6percentofthe Firms owned by females, nonwhites, or Hispanics firm. differed in several ways from firms owned by males, The average number of owners increased with the whites, or non-Hispanics (table A.1). As seen in number of employees: The smallest firms (0–1 employees)hadanaverageof1.2owners;intermediatesized firms (5–19 employees and 20–49 employees) 12.Forexample,considerafirminwhichthetotalreportedshares had 2.4 and 8.7 owners respectively; and the largest summedto75percent.Thelargestownerheld40percentofthefirm andidentifiedhimselfasbothwhiteandnonblack,asHispanic,andas firms(100–499employees)had13owners.Theownmale; the second owner held 25 percent of the firm and identified ership share of the first owner decreased as the herselfasbothwhiteandblack,asnon-Hispanic,andasfemale;and number of owners increased, from 94.6 percent thethirdownerheld10percentofthefirmandidentifiedhimselfas bothnonwhiteandblack,asnon-Hispanic,andasmale.Theownershipamong the smallest firms to 62.9 percent among the weightedcharacteristicsoftheownerswere87percentwhite,46perlargest. centblack,53percentHispanic,and67percentmale.Thefirmwould bedeemedtobewhite,Hispanic,andmale(butnotblackbecausethe Race, Ethnicity, and Sex of Small Business ownership-weighted percentage of black ownership was less than 51percent).Hereishowthefirm’sshareofHispanicownershipwould Owners becalculatedinthisexample: [(1x.40)+(0x.25)+(0x.10)]x(1/.75)x100=53percent. The race, ethnicity, and sex of the ownership of a Thefirstterm,1x.40,istheproductofanindicatorofthefirstowner’s Hispanicstatus(1,indicatingHispanic)timesthefirstowner’sshare small business in the survey were defined by the (40 percent). The second term is the product of an indicator of the weighted sum of the characteristics of the firms’ secondowner’sHispanicstatus(0,indicatingnon-Hispanic)timesthe owners.11 Unlike in previous years, owners were secondowner’sshare,andthethirdtermistheproductofanindicator ofthethirdowner’sHispanicstatustimesthethirdowner’sshare.The allowedtoidentifythemselvesasbeingofmorethan sumofthethreetermsarethenmultipliedbythereciprocalofthetotal one race, and therefore firms could be classified as sharesreported(.40+.25+.10)=.75sothatthetotalshareswillsum being of more than one race. For firms in which less to100percent.Last,theadjustedsumismultipliedby100toconvert thevaluetopercent. than 100 percent of the ownership was reported, 13. Throughout this article, the term ‘‘Asian’’is used for convecharacteristics were scaled up by a factor that made niencetorefertoindividualswhocharacterizedthemselvesasbeingin the reported ownership equal 100 percent. If the thecategory‘‘Asian,NativeHawaiian,orotherPacificIslander.’’ 14.Thisincreaseislikelytobeatleastpartiallyattributabletothe fact that equal ownership was a ‘‘volunteered’’response rather than 11. Characteristics of each owner were weighted by the owner’s specifically asked about in 1998. In 2003 this statistic was derived shareinthebusiness. fromthereportedownershipshares.

A172 Federal Reserve Bulletinh 2006 4. Useofcomputersbysmallbusinessesandtypeofuse,bynumberofemployees,industry,andageoffirm,1998and 2003surveys Percent Percent 1998 2003 Typeofuseamongusers Typeofuseamongusers Characteristic Uses Applica- Uses Applicacomputers Online Internet tionsfor Adminis- computers Online Internet tionsfor Adminis- Other Other banking access loansor tration banking access loansor tration credit credit Allfirms ............... 76.2 15.0 75.3 5.3 95.8 45.6 85.9 46.8 90.9 12.9 96.6 77.0 Numberofemployees 0–1..................... 63.2 17.4 74.6 5.8 95.1 46.2 80.3 42.4 88.7 12.1 95.0 73.0 2–4..................... 72.3 12.2 74.6 4.9 94.4 43.6 82.1 42.3 89.9 12.4 95.2 74.4 5–19 ................... 86.5 14.1 73.5 5.8 97.0 46.3 91.2 49.3 91.9 13.5 98.4 80.2 20–49 .................. 94.1 19.1 81.8 4.5 97.9 51.5 97.6 61.6 94.5 15.3 99.2 83.3 50–99 .................. 94.2 34.6 89.4 2.8 98.8 48.2 98.3 65.1 97.3 12.4 99.4 83.9 100–499................ 96.5 36.9 91.0 3.6 99.5 47.3 100.0 77.7 98.9 13.4 99.2 90.1 StandardIndustrial Classification Constructionandmining.. 67.9 11.9 72.3 3.0 97.4 29.5 85.3 42.2 87.0 9.4 97.4 65.3 Manufacturing .......... 87.6 16.5 75.7 3.9 92.7 51.1 90.7 54.1 99.1 14.0 96.4 84.3 Transportation .......... 85.1 25.8 75.1 5.2 97.9 46.4 82.1 43.1 81.4 14.0 98.8 69.1 Wholesaletrade......... 84.6 17.4 72.6 5.3 98.0 50.9 90.9 48.9 88.9 12.5 96.5 70.2 Retailtrade ............. 61.4 10.3 66.7 5.7 95.3 45.3 73.2 47.4 87.0 13.4 98.2 75.1 Insuranceandrealestate. 90.1 18.2 83.5 5.4 96.1 38.3 94.7 47.3 96.1 14.6 98.6 80.2 Businessservices ....... 70.4 16.3 73.2 5.9 96.2 51.9 83.0 50.0 90.7 13.5 93.1 81.0 Professionalservices .... 89.5 13.9 82.7 6.0 94.8 45.1 95.7 42.7 93.6 12.9 97.9 79.6 Yearsundercurrent ownership 0–4..................... 78.3 14.1 77.0 7.5 94.5 51.6 87.9 56.2 92.8 15.8 97.1 80.0 5–9..................... 78.3 15.7 78.9 5.6 96.1 48.3 88.5 49.7 91.0 12.7 97.3 76.5 10–14 .................. 77.5 14.6 74.4 4.0 94.9 45.5 84.8 44.9 93.7 13.7 96.3 78.3 15–19 .................. 76.6 16.1 77.7 4.4 96.1 42.1 85.9 46.1 93.8 14.4 96.1 79.8 20–24 .................. 76.6 15.2 72.1 3.3 97.0 40.8 88.1 40.4 86.3 8.9 96.5 74.9 25ormore ............. 65.6 14.8 65.3 4.7 97.8 36.0 79.2 36.9 86.6 10.6 95.8 71.4 previous surveys, the female-owned firms tended to that used computers to apply for credit or loans also be younger and smaller in terms of employment, rosesubstantially,from5.3percentto12.9percent. sales, and assets than those owned by males. They Withtheriseinprevalenceofcomputerusecamea were also more likely to organize as proprietorships rise in incidence and a narrowing in the variation of and less likely to organize as S corporations than incidence across firm age and size. For example male-owned firms. Female-owned firms were more although the incidence of use still varied with the likely to be engaged in professional and business numberofemployees,therangein2003—80percent services than male-owned firms and less likely to be to100percent—washigherandnarrowerthanthatin engagedinconstruction,mining,andmanufacturing. 1998—63 percent to 97 percent. And although inci- Relativetowhite,non-Hispanicfirms,nonwhiteor denceofusevariesinverselywithfirmage,therange Hispanic firms were younger and smaller in employ- of incidence by age also rose and narrowed between ment,sales,andassetsandweremoreoftenorganized thetwosurveys. as proprietorships. Similarly, nonwhite or Hispanic Nonstandard Work Arrangements firmswerealsomorelikelytobeengagedinbusiness servicesandlesslikelytobeengagedinconstruction The use of nonstandard work arrangements has been andminingandinsuranceandrealestatebusinesses. ontherisesinceatleastthemid-1990s.Forexample, Computer Use within the Firm estimatesfromtheFebruary1995CurrentPopulation Surveyindicatethat12.1millionworkers(or9.8per- Useofacomputerwithinafirmisoneindicatorofthe cent of the total) were independent contractors, extent of the firm’s adoption of technological ad- on-callworkers,temporaryagencyworkers,orworkvances.Inthe1998survey,whichwasthefirstSSBF ers provided through contract firms.15 By 2005, estito ask firms about their use of computers, 76.2 per- mates indicate that 14.8 million workers, or 10.7 cent reported using them (table 4). By 2003, the proportion had increased to 85.9 percent.Among the 15. Anne E. Polivka (1996), ‘‘Contingent and Alternative Work firms using computers, the proportion that used them Arrangements, Defined,’’Monthly Labor Review, October, pp. 3–9. The Current Population Survey, conducted monthly by the U.S. foronlinebankingrosebetweenthetwosurveysfrom CensusBureaufortheU.S.BureauofLaborStatistics,coversabout 15.0percentto46.8percent;likewise,theproportion 50,000households(www.bls.census.gov/cps/cpsmain.htm).

Financial Services Used by Small Businesses: Evidence from the 2003 Survey of Small Business Finances A173 5. Useofnonstandardworkarrangementsbysmallbusinessesduringatypicalpayperiod,andtypesofarrangements, bynumberofemployeesandindustry,2003survey Percentexceptasnoted Averagenumber Anynonstandard Temporaryagency Contractorsor Characteristic Paiddaylaborers Leasedemployees ofnonstandard arrangement employees consultants employees Allfirms ................... 45.7 9.8 7.6 2.9 40.2 4.9 Numberofemployees 0–1......................... 33.6 6.9 2.5 1.5 30.0 4.1 2–4......................... 43.8 10.9 4.2 1.8 39.0 3.3 5–19........................ 52.3 10.4 11.0 4.3 45.7 4.4 20–49....................... 57.0 10.5 18.3 5.6 48.8 7.6 50–99....................... 62.0 7.3 28.4 6.8 52.4 23.1 100–499 .................... 74.6 10.3 38.6 8.9 59.1 21.9 StandardIndustrial Classification Constructionandmining..... 69.4 19.4 10.4 4.4 62.1 5.1 Manufacturing............... 50.7 7.8 14.2 2.1 42.6 4.5 Transportation............... 56.1 8.3 7.1 6.1 52.3 5.3 Wholesaletrade ............. 43.3 9.2 12.1 2.7 32.6 3.6 Retailtrade ................. 30.8 9.2 4.6 2.2 25.4 4.7 Insuranceandrealestate..... 56.0 9.2 6.3 2.8 52.2 6.0 Businessservices............ 41.7 10.5 6.1 2.9 37.0 5.9 Professionalservices......... 44.0 5.6 7.5 2.3 40.0 3.5 percent of total employment, fell into one of these portionoffirmsthatusedcontractorsandconsultants groups.16 rangedfrom25percentinretailtradeto62percentin For the 2003 survey, the SSBF asked respondents constructionandmining. for the first time about the use of nonstandard work arrangements.Afteraseriesofquestionsontheuseof TYPES OF FINANCIAL SERVICES USED standard employees (both paid and unpaid), respon- BY SMALL BUSINESSES dentswereaskedwhether,duringatypicalpayperiod, they used any paid day laborers, temporary agency Firmswereaskedwhichoffourteenfinancialservices employees, workers from an employee-leasing firm, theyusedatuptotwentyinstitutions.17Thefinancial or contractors.About half of the firms reported using services can be grouped into three broad categories: atleastoneofthesearrangements. (1) liquid asset accounts, which are checking and The use of nonstandard work arrangements varied savings-type accounts, (2) credit lines, loans, and substantially by firm size. In general, the larger the capital leases, which are lines of credit, mortgages firm,themorelikelyitwastohaveemployedatleast used for business purposes, motor vehicle loans, one worker in each of the nonstandard arrangements equipmentloans,capitalleases,andmiscellaneousor (table 5). Across all firm sizes, contractors and con- ‘‘other’’ loans, and (3) financial management sersultantswerethemostcommontypesofnonstandard vices, which are transaction services, credit card and workers reported, a result consistent with statistics debit card processing services, cash management calculatedfromtheemployeeside(refertotextnotes services, credit-related services, brokerage services, 15 and 16). About 30 percent of the smallest firms and trust and pension services. Loans from owners, used contractors and consultants, and about 59 per- creditcards,andtradecreditarediscussedseparately cent of the very largest firms did so. Among firms and are not included in the tabulations for ‘‘any usinganynonstandardworkers,thenumbergenerally financial service’’ because no information was colincreasedwithfirmsize. lectedabouttheprovidersoftheseservices. The use of nonstandard work arrangements varied Nearlyallsmallbusinesses(about96percent)used substantiallybyfirmindustry.Attheextremes,nearly at least one financial service in 2003, a finding 70 percent of firms involved in construction and mining reported some type of nonstandard arrange- 17.Forthisarticle,useofafinancialservicewasmeasuredbythe ment, whereas only 31 percent of retail trade firms percentageofsmallbusinessesusingthatservice.Dataonusethatare basedondollaramountsornumbersofaccountswillbeavailableata reported doing so. Across all industries, contractors later date. However, previous analysis has shown that conclusions andconsultantsagainwerethemostcommontypesof based on dollar amounts or on number of accounts are usually nonstandard workers reported; by industry, the pro- qualitatively very similar to conclusions based on percentages of firms.FurtherdiscussionisinRebelA.Cole,JohnD.Wolken,andR. LouiseWoodburn(1996),‘‘BankandNonbankCompetitionforSmall 16. U.S. Bureau of Labor Statistics (2005), ‘‘Contingent and BusinessCredit:Evidencefromthe1987and1993NationalSurveys AlternativeEmploymentArrangements,February2005,’’pressrelease, of Small Business Finances,’’ Federal Reserve Bulletin, vol. 83 July27,www.bls.gov/news.release/conemp.nr0.htm. (November),pp.983–95.

A174 Federal Reserve Bulletinh 2006 6. Useofselectedfinancialservicesbysmallbusinesses,byselectedcharacteristicsoffirms,2003survey A.Anyservice;liquidassetaccounts;creditlines,loans,andcapitalleases Percent Creditline,loan,orcapitallease Liquidassetaccount Any Characteristic service1 Credit Loan Capital Any Other line lease Any Checking Savings Mortgage Vehicle Equipment Allfirms 2003 ........................ 96.4 95.0 94.6 22.1 60.4 34.3 13.3 25.5 10.3 8.7 10.1 1998 ........................ 96.1 94.4 94.0 22.2 55.0 27.7 13.2 20.5 9.9 10.6 9.8 Numberofemployees 0–1 ......................... 88.9 85.9 85.0 15.4 42.1 19.4 5.6 17.4 4.3 4.0 7.1 2–4 ......................... 96.9 95.1 94.8 18.6 53.9 27.2 12.6 22.0 5.2 6.6 7.1 5–9 ......................... 99.8 99.4 99.2 25.7 72.7 43.1 15.8 30.8 13.6 11.6 13.3 10–19....................... 100.0 99.8 99.8 30.9 77.4 50.2 19.2 35.9 21.1 12.1 16.5 20–49....................... 100.0 99.9 99.9 33.3 82.7 57.5 21.4 36.2 26.3 16.0 15.7 50–99....................... 100.0 98.7 98.5 41.1 87.4 68.0 18.6 36.5 27.6 22.9 16.5 100–499..................... 100.0 100.0 100.0 36.5 93.8 82.3 28.0 35.9 32.6 27.9 18.6 Sales(thousandsofdollars) Lessthan25................. 81.4 76.7 75.3 10.8 29.4 12.3 5.3 10.2 * * 6.0 25–49....................... 97.2 94.1 93.7 19.6 45.6 14.1 8.9 17.3 * 5.8 8.9 50–99....................... 97.6 96.7 96.2 16.7 49.5 24.2 10.1 23.0 7.4 3.6 5.3 100–249..................... 98.9 98.3 98.3 19.1 59.9 29.2 13.3 21.9 8.0 9.1 9.1 250–499..................... 99.6 99.2 98.9 22.9 70.7 39.8 19.1 27.9 10.0 10.4 12.5 500–999..................... 100.0 99.4 99.4 27.4 80.0 47.8 15.6 38.7 16.7 12.9 14.4 1,000–2,499 ................. 100.0 99.0 99.0 37.2 76.4 56.6 18.9 35.5 21.4 13.1 12.5 2,500–4,999 ................. 100.0 99.2 99.1 30.3 79.9 65.2 14.6 37.2 19.1 15.1 15.0 5,000–9,999 ................. 100.0 100.0 100.0 34.7 90.4 65.6 22.3 49.4 28.4 16.3 14.1 10,000ormore .............. 100.0 99.8 99.8 39.1 90.7 83.8 16.3 33.3 23.2 20.9 18.1 Assets(thousandsofdollars) Lessthan25................. 90.0 86.8 86.2 13.3 39.0 16.2 4.3 16.2 3.3 5.2 6.1 25–49....................... 98.2 97.9 97.7 17.4 57.2 31.0 6.8 23.5 6.5 8.3 5.5 50–99....................... 99.8 98.5 98.0 22.9 66.2 32.8 13.8 26.9 9.1 9.4 13.4 100–249..................... 99.1 98.4 98.4 25.9 67.2 37.1 16.3 31.4 12.5 7.9 10.3 250–499..................... 99.7 99.1 98.6 29.7 78.1 48.5 22.4 32.3 17.6 12.3 12.9 500–999..................... 100.0 99.7 99.7 27.5 79.0 55.7 23.8 30.5 16.5 12.0 19.7 1,000–2,499 ................. 100.0 100.0 100.0 38.7 82.1 62.2 27.3 39.5 24.6 12.5 12.2 2,500–4,999 ................. 100.0 96.0 95.9 41.8 88.1 64.8 27.6 34.5 26.7 20.6 21.4 5,000ormore ............... 100.0 100.0 100.0 32.8 80.3 66.6 28.8 28.0 17.9 15.9 14.0 Organizationalform Proprietorship ............... 93.0 90.5 89.9 17.2 52.4 24.2 11.1 21.5 6.9 5.5 9.5 Partnership .................. 97.3 95.7 95.7 22.1 57.1 28.7 20.5 20.4 10.4 8.9 6.2 Scorporation ................ 99.5 98.9 98.8 24.1 70.0 43.8 15.4 31.3 11.9 11.7 12.5 Ccorporation................ 99.6 99.4 99.1 32.1 66.2 47.2 11.3 28.2 16.6 12.0 9.7 StandardIndustrial Classification Constructionandmining ..... 96.5 95.0 95.0 23.9 70.8 44.6 13.6 43.9 16.5 6.3 7.5 Manufacturing............... 97.7 97.2 96.9 27.3 70.0 47.8 18.0 27.3 17.6 10.9 10.5 Transportation ............... 99.5 98.3 98.3 31.3 79.1 36.5 9.5 42.9 16.0 9.3 20.0 Wholesaletrade ............. 98.8 96.7 96.3 22.7 62.6 49.5 13.1 30.1 7.1 7.8 9.6 Retailtrade.................. 97.4 96.3 96.3 17.6 58.9 32.8 14.7 19.7 9.2 7.3 14.0 Insuranceandrealestate ..... 96.3 95.5 95.1 28.6 59.2 28.8 23.1 21.1 * 5.3 5.4 Businessservices ............ 94.9 93.2 92.4 19.9 56.4 28.4 11.2 25.0 7.9 7.4 10.1 Professionalservices ......... 95.9 93.9 93.5 22.0 54.1 29.4 10.1 17.3 10.2 13.5 8.1 Yearsundercurrentownership 0–4 ......................... 96.1 94.3 93.8 16.8 61.8 30.2 13.3 21.2 7.0 11.0 15.4 5–9 ......................... 95.8 94.4 93.9 20.0 60.0 30.2 11.7 24.9 11.1 7.1 12.8 10–14....................... 98.1 96.5 96.1 25.7 60.4 35.1 14.6 28.0 10.8 10.4 7.7 15–19....................... 94.7 92.5 92.4 20.3 61.4 40.0 12.5 28.8 11.4 9.4 7.5 20–24....................... 96.0 94.6 94.6 22.8 58.1 33.9 13.7 25.2 12.6 8.4 5.4 25ormore .................. 97.8 97.0 96.8 28.6 60.2 39.3 14.5 26.9 10.4 6.3 7.7 essentiallythesameasin1998(table6.A).Ingeneral, Proprietorships were less likely than corporations useincreasedwithfirmsize,andnearlyallfirmswith or partnerships to use any financial service. The at least five employees, or with sales of at least difference may be due to the tendency of many $250,000, or with assets of at least $50,000 used proprietorships to commingle business and personal somefinancialservice.About11percentoffirmswith finances; for example, the owners may use personal oneworkerusednofinancialservicein2003.18 across all financial suppliers for each firm. The average number of services used by small businesses in 2003 was 3.9. One-half of the 18.Analternativetoincidenceasameasureofintensityofuseisthe firmsused3orfewerservices,one-thirdused2orfewer,one-fourth numberofdistinctservices(whichrangefrom0to14foreachsource) used5ormore,and2.5percentused10ormore.

Financial Services Used by Small Businesses: Evidence from the 2003 Survey of Small Business Finances A175 6.—Continued A.Anyservice;liquidassetaccounts;creditlines,loans,andcapitalleases Percent Creditline,loan,orcapitallease Liquidassetaccount Any Characteristic service1 Credit Loan Capital Any Other line lease Any Checking Savings Mortgage Vehicle Equipment Censusareaofmainoffıce Northeast.................... 95.5 94.9 94.9 19.2 57.8 33.8 11.5 21.4 9.1 7.6 9.6 NewEngland ............. 97.9 97.5 97.5 18.3 62.3 35.5 13.6 28.2 10.8 6.9 10.7 MiddleAtlantic............ 94.4 93.8 93.8 19.7 55.9 33.0 10.6 18.5 8.3 7.9 9.2 Midwest..................... 98.4 96.5 96.1 26.8 67.0 41.1 17.1 26.5 12.3 9.6 9.7 EastNorthCentral......... 98.0 95.8 95.3 26.3 67.1 39.4 19.2 25.7 10.6 10.1 11.0 WestNorthCentral ........ 99.3 98.1 97.8 27.8 66.7 44.4 12.9 28.2 15.9 8.5 7.0 South ....................... 96.5 94.7 94.1 17.7 61.8 31.9 14.1 28.9 10.4 7.8 10.5 SouthAtlantic............. 97.5 95.9 94.8 17.6 63.9 31.3 13.9 32.0 10.5 8.5 9.1 EastSouthCentral......... 94.9 93.1 93.1 20.9 57.0 31.0 9.2 21.3 9.5 6.4 14.9 WestSouthCentral ........ 95.5 93.3 93.3 16.2 60.5 33.3 16.8 27.1 10.7 7.4 10.9 West ........................ 95.4 94.1 93.8 26.8 54.8 32.2 10.2 23.1 9.4 10.0 10.3 Mountain ................. 95.6 94.2 93.8 23.4 62.4 36.4 14.5 28.9 11.9 9.2 14.1 Pacific .................... 95.4 94.0 93.8 28.4 51.4 30.3 8.3 20.5 8.2 10.4 8.6 Urbanizationatmainoffıce Urban ....................... 96.5 95.0 94.6 22.4 60.2 33.8 12.2 24.9 9.4 9.4 10.1 Rural........................ 96.2 95.1 94.9 21.3 61.5 36.2 17.6 27.9 13.7 6.0 10.4 Numberofoffıces One......................... 95.9 94.4 94.0 21.0 58.1 31.8 12.9 24.9 9.7 7.7 9.0 Two......................... 99.2 98.4 98.1 27.5 70.4 45.5 13.2 27.4 11.5 14.1 18.5 Threeormore ............... 100.0 99.2 99.2 32.7 82.7 56.8 21.0 33.9 18.4 17.2 14.8 Salesarea Primarilywithinthe UnitedStates ........... 96.4 95.0 94.6 22.1 60.8 34.4 13.5 25.7 10.4 8.7 10.1 Internationalorglobal ....... 97.1 94.3 94.3 23.8 51.8 32.2 8.4 22.6 8.6 8.7 11.5 Management Byowner ................... 96.2 94.7 94.3 21.5 59.9 34.1 12.7 25.5 10.1 8.4 10.0 Hired ....................... 99.9 99.9 99.9 31.8 68.3 35.1 21.5 26.5 13.7 12.6 11.6 Race,ethnicity,andsexof majorityowners NonwhiteorHispanic........ 95.5 92.8 92.0 17.8 54.8 25.6 11.6 25.7 5.5 5.9 12.5 Non-Hispanicwhite.......... 96.5 95.3 95.0 22.8 60.9 35.5 13.3 25.4 11.0 9.0 9.6 White ....................... 96.5 95.1 94.8 22.5 61.0 35.4 13.1 25.7 10.7 8.9 9.6 Black ....................... 91.7 89.8 88.0 19.5 47.5 12.1 10.6 25.4 * 8.9 11.9 Asian,NativeHawaiian,or otherPacificIslander.... 99.1 97.4 97.4 17.6 52.4 28.1 12.8 19.5 6.3 * 14.9 AmericanIndianorAlaska Native.................. 92.7 90.9 88.5 * 63.1 35.5 * 28.2 * * * Hispanic .................... 94.9 90.4 90.4 17.5 61.1 31.6 13.6 33.3 5.8 6.7 11.9 Non-Hispanic................ 96.5 95.2 94.8 22.3 60.3 34.3 13.2 25.2 10.5 8.8 10.0 Female...................... 92.3 89.6 89.0 19.9 47.6 23.4 10.7 19.6 8.4 7.2 7.3 Male ........................ 97.4 96.3 96.0 22.0 64.4 38.1 13.0 27.7 11.0 9.1 11.1 Ownershipdividedequally bysex.................. 98.6 97.8 97.6 26.3 62.3 32.9 18.8 25.1 10.1 9.0 10.1 Note:Fordefinitionsofservices,refertotext;fordefinitionofsalesareas,refertotable2,note3. 1.Memoitemsintable6.Bareexcludedfromthesedata. *Fewerthanfifteenobservations. savings and checking accounts for business pur- Liquid Asset Services poses.19 Also, firms whose ownership was black, American Indian or Alaska Native, Hispanic, or Most small businesses (95 percent) had a checking female were less likely to use any financial service account in 2003, the same percentage as used any than were firms whose ownership was white, non- liquid asset account, that is, a checking or savings Hispanic,ormale. account (table 6.A).20 Because a checking account 19. Respondents were asked to count as a business service any 20. Checking accounts were defined as accounts with unlimited personal account that was used at least 50 percent of the time for check-writing privileges and included those in credit unions (share businesspurposes.Mostofthefirmsthatreportedusingnofinancial draft accounts). Money market accounts, including money market serviceswereextremelysmall;itispossiblethattheownersofthose depositaccounts,wereconsideredtobecheckingaccountsonlyifthey firmsusedpersonalaccountsforbusinesspurposesbutdidsolessthan offered unlimited check-writing privileges. Savings accounts were 50percentofthetime. definedaspassbooksavings,creditunionshareaccounts,certificates

A176 Federal Reserve Bulletinh 2006 6. Useofselectedfinancialservicesbysmallbusinesses,byselectedcharacteristicsoffirms,2003survey B.Financialmanagementservices Percent Financialmanagementservice Memo Credit Nontraditionalcredit Characteristic and Cash Trust Traditional Trans- Credit- andnon- Any action pro d c c e a e b r s d i s t ing m m an e a n g t e- related Brokerage pe a n n s d ion o L f w r o o n a m e n r Perso C n r a e l dit B ca u r s d iness T cr r e a d d i e t tra c d r i e ti d o i n t al Allfirms 2003.................... 64.7 38.9 37.2 6.7 5.0 5.6 17.2 30.3 46.7 48.1 60.1 92.9 1998.................... 49.8 41.1 n.a. 5.2 3.1 4.3 12.6 28.1 46.0 34.1 61.9 89.5 Numberofemployees 0–1..................... 41.5 25.2 16.1 2.7 * 2.6 8.0 25.7 48.6 32.0 35.7 83.8 2–4..................... 59.6 35.0 33.7 3.8 4.1 5.0 10.3 27.0 48.1 45.7 55.9 93.3 5–9..................... 77.4 45.1 47.9 8.3 6.9 7.9 20.5 33.3 47.8 56.8 71.6 96.6 10–19................... 84.4 49.9 57.3 8.1 4.9 7.0 29.2 31.3 45.6 59.7 80.4 97.3 20–49................... 90.1 61.0 56.3 15.6 9.0 7.1 41.9 36.0 34.4 61.8 85.0 99.8 50–99................... 88.3 61.3 51.5 37.8 15.8 12.8 59.5 32.9 34.6 63.5 88.5 98.3 100–499 ................ 96.1 70.3 50.2 50.1 26.7 13.3 68.4 28.4 32.2 71.5 85.4 98.9 Sales(thousands ofdollars) Lessthan25 ............ 36.3 23.2 15.4 * * * * 22.3 48.1 25.7 27.7 79.2 25–49................... 49.7 30.5 24.5 * * * 7.5 33.5 51.7 34.1 40.2 89.2 50–99................... 49.6 29.1 25.6 * * * 10.5 34.9 46.5 41.0 46.3 91.4 100–249 ................ 62.6 34.3 37.5 3.7 4.6 4.4 10.1 28.1 49.9 48.2 61.0 94.5 250–499 ................ 74.7 43.9 48.3 7.4 4.2 6.7 15.6 27.7 49.4 54.9 70.0 96.3 500–999 ................ 81.2 49.4 53.9 8.4 7.6 7.1 20.7 27.7 44.1 62.5 82.2 98.4 1,000–2,499............. 87.2 51.8 48.8 9.2 5.9 9.4 40.5 36.3 40.9 63.6 79.3 98.7 2,500–4,999............. 92.2 58.4 56.7 17.2 15.1 12.6 47.6 38.7 39.7 61.9 87.6 100.0 5,000–9,999............. 91.7 65.2 39.2 19.7 14.1 15.7 52.6 26.9 30.8 63.3 83.6 99.5 10,000ormore.......... 94.9 75.2 50.4 50.5 27.5 17.9 68.4 30.0 35.8 68.9 89.8 99.9 Assets(thousands ofdollars) Lessthan25 ............ 43.2 25.1 20.8 2.7 * 2.4 6.5 23.8 47.3 35.9 38.4 85.5 25–49................... 60.3 37.0 32.3 * * * 11.6 28.1 47.4 48.1 57.6 93.5 50–99................... 70.5 42.6 43.5 4.2 3.5 5.3 14.8 28.1 51.4 47.5 64.8 95.4 100–249 ................ 75.5 42.7 48.9 5.7 5.5 7.4 18.7 31.5 43.8 53.4 68.1 96.0 250–499 ................ 77.4 43.7 48.6 8.5 6.4 9.2 23.3 34.1 43.8 61.4 76.2 98.9 500–999 ................ 84.9 52.1 52.2 7.5 11.8 10.1 30.0 37.8 47.7 53.6 76.2 96.7 1,000–2,499............. 83.0 53.2 48.8 16.4 13.5 9.7 37.4 31.8 50.5 64.7 84.5 99.0 2,500–4,999............. 88.2 62.7 49.3 22.2 19.6 12.2 48.1 30.3 35.5 61.9 86.0 99.1 5,000ormore........... 95.5 78.6 36.3 51.5 25.3 16.1 56.7 34.6 33.6 55.8 85.2 98.5 Organizationalform Proprietorship ........... 52.3 30.4 28.3 3.1 2.1 3.5 10.1 ... 52.3 35.1 46.9 88.3 Partnership.............. 70.4 44.4 34.7 7.6 6.6 9.5 13.1 25.2 41.7 46.9 58.5 93.9 Scorporation............ 74.9 45.6 47.1 8.5 6.8 5.5 22.4 31.3 43.6 61.6 71.2 97.4 Ccorporation ........... 76.8 46.9 44.5 12.6 8.6 9.7 29.0 31.1 39.6 58.9 76.1 96.8 StandardIndustrial Classification Constructionandmining. 45.5 28.4 13.9 4.2 6.1 3.1 15.2 28.3 44.7 52.1 80.5 95.0 Manufacturing........... 67.2 45.1 36.5 10.0 11.9 5.1 19.7 39.8 47.1 54.8 76.0 97.0 Transportation........... 56.7 33.5 37.5 8.4 13.2 * 14.0 28.3 41.4 51.8 60.7 96.9 Wholesaletrade ......... 73.6 44.6 48.2 7.5 8.2 5.9 22.5 34.1 46.7 54.4 72.2 94.1 Retailtrade ............. 82.4 46.8 72.1 4.9 3.4 2.4 11.5 38.2 47.9 45.0 67.4 91.4 Insuranceandrealestate. 63.4 40.9 9.3 15.3 6.9 13.6 19.9 17.7 47.4 43.0 45.9 89.8 Businessservices........ 62.3 35.7 38.5 6.1 3.4 3.9 11.6 27.4 45.2 47.0 52.5 91.5 Professionalservices..... 61.7 38.6 24.8 5.6 2.1 9.9 27.4 26.2 48.9 47.1 46.8 93.5 Yearsundercurrent ownership 0–4..................... 59.4 38.3 37.9 4.5 4.3 3.1 8.5 34.4 45.2 46.6 47.8 91.1 5–9..................... 68.9 42.8 40.7 5.8 5.5 2.9 14.3 28.9 44.6 49.7 61.0 93.1 10–14................... 67.8 39.7 39.4 7.0 5.3 7.5 20.4 30.5 49.8 48.7 63.0 94.3 15–19................... 67.0 38.7 38.0 9.0 4.0 6.8 21.1 32.1 48.5 52.3 62.3 92.6 20–24................... 63.0 37.3 33.9 6.8 3.2 7.8 23.7 28.1 46.9 50.7 62.8 92.1 25ormore.............. 62.5 35.4 31.7 8.1 6.6 8.1 20.9 27.2 46.6 43.1 67.0 94.5 (including a share draft account at a credit union) is a vehicle for paying suppliers and depositing sales of deposit, and other time deposits; also considered to be savings receipts, it is not surprising that the reported use of accountsweremoneymarketaccountsthatwerelimitedineitherthe numberortheamountofchecksthatcouldbewritten.Incomparison withsmallbusinesses,91.4percentofhouseholdsin2004hadsome Bucks,Arthur B. Kennickell, and Kevin B. Moore (2006), ‘‘Recent type of transaction account (checking account, savings account, ChangesinU.S.FamilyFinances:Evidencefromthe2001and2004 money market deposit account, money market mutual fund, or call SurveyofConsumerFinances,’’FederalReserveBulletin,vol.92,pp. account at a brokerage). More information is available in Brian K. A1–A38.

Financial Services Used by Small Businesses: Evidence from the 2003 Survey of Small Business Finances A177 6.—Continued B.Financialmanagementservices Percent Financialmanagementservice Memo Credit Nontraditionalcredit Characteristic and Cash Trust Traditional Trans- Credit- andnon- Any action pro d c c e a e b r s d i s t ing m m an e a n g t e- related Brokerage pe a n n s d ion o L f w r o o n a m e n r Perso C n r a e l dit B ca u r s d iness T cr r e a d d i e t tra c d r i e ti d o i n t al Censusarea ofmainoffıce Northeast ............... 65.2 39.6 33.6 4.9 3.4 4.5 18.9 34.9 45.1 44.7 59.3 92.6 NewEngland ......... 67.2 42.3 35.5 3.7 3.1 * 21.5 41.7 53.1 48.0 69.7 98.5 MiddleAtlantic ....... 64.3 38.4 32.7 5.4 3.5 5.1 17.7 32.2 41.7 43.2 54.9 90.1 Midwest ................ 67.2 38.2 38.4 10.0 6.3 6.8 22.1 28.1 46.3 50.9 65.8 94.6 EastNorthCentral .... 64.7 37.6 36.7 9.7 6.4 8.6 21.8 30.0 40.8 53.2 63.8 94.6 WestNorthCentral.... 72.4 39.5 41.9 10.6 6.1 3.0 22.9 24.5 57.5 46.1 70.0 94.6 South ................... 65.0 40.0 38.8 6.5 5.7 5.0 14.3 29.4 45.0 47.0 59.6 92.5 SouthAtlantic ........ 68.9 41.5 43.3 6.7 5.6 4.5 15.9 29.5 44.9 53.9 61.3 94.3 EastSouthCentral .... 58.0 34.4 33.8 7.0 5.4 5.2 13.7 26.6 37.1 41.4 57.4 88.1 WestSouthCentral.... 61.7 40.0 33.2 5.8 6.0 5.7 11.6 30.3 49.0 37.3 57.5 91.4 West.................... 61.9 37.5 37.0 5.4 4.1 6.4 15.6 30.1 50.7 50.2 56.4 92.4 Mountain ............. 64.2 43.3 38.3 7.2 5.5 5.1 11.8 25.4 50.8 53.4 61.2 94.3 Pacific................ 60.8 34.9 36.4 4.6 3.4 7.0 17.3 32.8 50.7 48.7 54.1 91.5 Urbanization atmainoffıce Urban................... 65.1 39.5 37.2 6.8 4.7 6.0 17.7 31.9 46.4 49.6 58.4 93.0 Rural ................... 63.3 36.6 37.4 6.3 6.1 4.4 14.9 23.0 47.8 42.6 66.6 92.8 Numberofoffıces One..................... 62.2 36.4 35.9 5.7 4.2 5.3 15.6 30.3 46.9 46.4 58.0 92.1 Two .................... 78.2 50.4 44.2 7.0 9.3 6.2 21.8 29.0 44.6 56.9 71.8 97.8 Threeormore........... 85.6 61.9 48.0 24.7 11.3 9.9 36.7 32.9 45.9 62.9 74.9 99.4 Salesarea Primarilywithinthe UnitedStates ....... 64.1 38.2 36.9 6.7 4.8 5.7 16.7 29.7 46.3 48.2 60.4 92.9 Internationalorglobal ... 77.2 53.1 44.2 5.0 8.4 5.2 27.4 40.7 53.6 46.0 53.3 92.9 Management Byowner ............... 63.7 38.2 36.4 6.1 4.7 5.4 16.3 30.3 47.2 48.0 59.2 92.6 Hired ................... 80.4 50.7 49.1 14.7 9.2 10.2 29.2 30.7 39.4 49.7 72.3 99.0 Race,ethnicity,andsex ofmajorityowners NonwhiteorHispanic ... 62.3 35.7 38.1 3.9 4.8 4.7 10.5 36.6 39.9 46.7 48.6 90.6 Non-Hispanicwhite ..... 64.9 39.6 36.9 7.1 5.0 5.6 18.1 29.8 47.9 48.3 61.9 93.1 White................... 64.8 39.5 36.9 6.9 4.9 5.7 17.9 30.2 47.4 48.4 61.4 93.2 Black ................... 51.8 31.7 29.2 * * * * 35.2 30.5 36.1 34.8 86.1 Asian,NativeHawaiian, orotherPacific Islander ............ 71.2 37.7 51.2 5.0 7.7 3.9 10.5 31.6 48.0 53.4 56.4 92.5 AmericanIndianor AlaskaNative ...... 51.3 28.4 28.2 * * * * * 51.4 39.7 51.4 87.4 Hispanic ................ 64.7 39.1 34.8 * * 8.0 14.6 39.0 35.3 50.5 49.1 92.3 Non-Hispanic ........... 64.6 38.9 37.2 6.9 5.1 5.5 17.2 30.0 47.2 48.0 60.5 93.0 Female.................. 61.6 38.6 36.0 6.3 2.7 3.6 10.2 35.9 49.2 42.9 46.9 88.2 Male.................... 64.6 38.4 36.2 6.9 6.0 6.6 19.9 29.8 45.1 49.8 64.0 93.8 Ownershipdivided equallybysex...... 69.9 41.8 43.7 5.7 3.6 4.6 14.7 26.6 50.6 48.5 62.9 96.9 Note:Refertonotestotable6.A. n.a.Notavailable. ...Notapplicable. ‘‘any service’’ (96 percent) nearly matches the re- organizational form, corporations were the most ported use of ‘‘any liquid asset account.’’ likely type of firm to have a savings account. Firms For business savings accounts, firm size seems to owned by non-Hispanic whites were more likely play a significant role in usage, with the smallest than nonwhite or Hispanic firms to have a savings firms the least likely to have savings accounts. Less account, and male-owned businesses or businesses than one-fifth of firms with four or fewer workers owned equally by males and females were more had savings accounts, compared with more than likely than were female-owned firms to have a one-third of firms with twenty or more workers. By business savings account.

A178 Federal Reserve Bulletinh 2006 Credit Lines, Loans, and Capital Leases depository institutions typically require borrowers to have several years of financial history to qualify for More than 60 percent of small businesses reported credit.The youngest firms (those under current ownoutstandingcreditintheformofacreditline,aloan, ershipfewerthanfiveyears)reportednearlythesame oracapitallease(table6.A).Overall,theincidenceof incidence of borrowing as did more mature firms. credit lines, loans, and capital leases increased from However,asomewhatdifferentpictureemergeswhen 55percenttomorethan60percentbetween1998and specific types of credit are examined. The youngest 2003. firms were least likely to have outstanding lines, Asin1998,themostwidelyusedtypesofcreditin vehicle loans, and equipment loans. They were also 2003 were credit lines and vehicle loans.21 The themostlikelytohavehadleasesand‘‘other’’loans. importance of these two credit types seems to be It may be that the hypothesis regarding young firms growing.Theshareoffirmswithlinesincreasedfrom and depository institutions does not apply equally to 28 percent to 34 percent, and the share with vehicle all loan types, or that other factors, such as personal loans rose from 21 percent to 26 percent. The incidences of mortgages, equipment loans, and ‘‘other’’ relationshipswithfinancialinstitutions,mayoffsetto loansweresimilartotheir1998levels;leasesdeclined some degree the lack of information available for somewhat.22 The increase in the percentage of firms youngerfirms. with lines may be due in part to commercial banks’ Theincidenceoflines,loans,andleasesalsovaried increaseduseofcredit-scoringmodelsforthattypeof somewhat with owner characteristics, such as race, loan.23Alternatively,theincreasemayhavebeendue ethnicity, and sex. In 2003, 61 percent of whiteto differences in the economic environment. For ownedbusinesseshadoutstandingcredit.Incontrast, example, 2002 and 2003 were years of historically about 48 percent of either black-owned or femalelowinterestrates.Thelowratesmayhavestimulated owned businesses and 52 percent of Asian-owned increasedloandemandamongsmallbusinesses. firmshadoutstandingcredit.Firmsownedequallyby The incidence of lines, loans, and leases increased men and women appear to be most similar to malewith firm size. More than 90 percent of the largest ownedbusinessesintheirreporteduseoflines,loans, firms (100–499 employees or at least $5 million in andleases. sales) had one of these types of credit in 2003, Incidence also varied by credit type for these firm compared with fewer than 50 percent of the firms types.Foreachcredittype,female-ownedbusinesses with 1 employee or with less than $100,000 in sales. were less apt to have credit than were male-owned Corporations were more likely than other types of firms.Black-ownedfirmswerelesslikelythanwhitefirmstohavehadoutstandingloansin2003.Firmsin owned firms to have had lines of credit, mortgages, services (business or professional) had fewer out- and equipment loans. Black-owned firms had about standing loans than those in other industries, perhaps thesameincidenceofvehicleloansandcapitalleases becausetheyrequirelessinventoryandequipment. butreportedahigherincidenceof‘‘other’’loans. Credit incidence did not appear to vary systemati- Non-Hispanic-owned firms and Hispanic-owned cally with the age of the business, even though firms were equally likely to have outstanding credit, although the incidence varied by type of credit. Non-Hispanic firms were more likely than Hispanic- 21.Inthisarticle,useofacreditlinereferstotheavailabilityofa owned firms to have equipment loans and leases, creditlineandnotnecessarilytotheborrowingoffundsfromtheline. Surveyinformationonoutstandingcredit-linebalanceswillbeavail- whereas Hispanic-owned firms were more likely to ableatalaterdate. have vehicle and ‘‘other’’loans. Some of the differ- 22. The majority of ‘‘other’’loans were loans that could not be ences by owner race, ethnicity, and sex may be classifiedascreditlines,mortgages,equipmentloans,vehicleloans,or capital leases. Such loans were most likely term loans, and roughly attributabletodifferencesinfirmcharacteristics,such 70percentofthemwereunsecured. assize.24 23.Althoughstatisticsontheuseofcreditscoringbycommercial banks are somewhat dated (W. Scott Frame,Aruna Srinavasan, and LynnWoosley,2001,‘‘TheEffectofCreditScoringonSmallBusiness Lending,’’Journal of Money, Credit, and Banking, vol. 33,August, 24.Researchonthistopicusingmultivariateanalysisisavailablein, pp.813–25),someindirectevidenceisavailablefromdatagathered for example, Ken S. Cavalluzzo, Linda C. Cavalluzzo, and John D. under the Community Reinvestment Act. These data indicate that Wolken (2002), ‘‘Competition, Small Business Financing, and Disbetween1998and2003,thegrowthinthenumberofsmallbusiness crimination: Evidence from a New Survey,’’ Journal of Business, loanshasbeenfargreaterthangrowthinthedollaramountofsmall vol. 75 (October), pp. 641–79; and Ken S. Cavalluzzo and John D. business loans, which suggests that much of the growth in business Wolken(2005),‘‘SmallBusinessLoanTurndowns,PersonalWealth, loanshasbeeninsmallerloans.Smalllinesofcreditarelikelytobe and Discrimination,’’ Journal of Business, vol. 78 (November), oneofthetypesofcreditmostamenabletocreditscoring. pp.2153–77.

Financial Services Used by Small Businesses: Evidence from the 2003 Survey of Small Business Finances A179 Financial Management Services (39percent)followedcloselybycreditanddebitcard processing (37 percent). Trust and pension services In the 2003 survey, financial management services wereusedbynearly20percentoffirms,whereascash covered six broad categories of service, one more management, credit-related, and brokerage services than in previous surveys. The 2003 categories were wereeachusedbyroughly5percentoffirms.Aswas transaction services, credit card and debit card pro- true for liquid asset services and for lines, loans, and cessing, cash management, credit-related services, leases, the use of financial management services brokerage services, and trust and pension services.25 increased with firm size. For the smallest firms (as Nearly two-thirds of firms used at least one financial measured by employment) just over 40 percent of managementservicein2003,comparedwithone-half firmsusedanyfinancialmanagementservice,25perof firms in 1998 (table 6.B). Part of the increase is cent used transaction services, 16 percent used card likely due to the difference between the two surveys processing, and only a very small portion used other in the wording of the questions. In 1998, transaction financial management services. In contrast, 96 perservices included credit card (but not debit card) centofthelargestfirms(100–499employees)usedat processing. In 2003, credit card processing was least one financial management service; the most removed from transaction services and added to a common was transaction services (70 percent), folnewservicecategorythatconsistedofcreditcardand lowed closely by trust and pension services (68 perdebit card processing. Consequently, direct compari- cent).Cashmanagementandcardprocessingservices sons of ‘‘any’’ financial management service or of wereusedby50percentofthelargestfirms. transactionserviceswithearliersurveysisdifficult.26 The proportion of proprietorships that used finan- In particular, the decline in the use of transaction cial management services (52 percent) was smaller services, from 41 percent to 39 percent, probably than that of firms with other organizational forms reflects,atleastinpart,thechangeinthedefinitionof (70–77 percent); proprietorships may not need these thatcategory. services as much because they tend to be small and more likely than other types of firms to commingle Overall Results personalandbusinessaccounts. Firmsdifferedintheiruseoffinancialmanagement The evidence generally points to a growing imporservices by race, with black-owned firms and firms tanceforfinancialmanagementservices.Anincrease ownedbyAmericanIndiansorAlaskaNativessomein incidence was recorded for each of the financial what less likely to have used financial management management services whose definitions were conservices than white non-Hispanic-owned firms. stant across surveys: for cash management services, Female-owned firms were somewhat less likely to from 5 percent to 7 percent; for credit-related serhave used credit-related, brokerage, and trust and vices, from 3 percent to 5 percent; for brokerage pension services than male-owned firms. Hispanicservices,from4percentto6percent;andfortrustand owned and Asian-owned firms used one or more pensionservices,from13percentto17percent. financial-management services with about the same The most widely used financial management serfrequency as firms owned by non-Hispanic whites. vice in 2003 continued to be transaction services However,differencesinuseexistamongthesegroups for specific financial management categories. For example,Asian-ownedfirmsweremorelikelytohave 25. These categories cover the following specific services— Transactionservices:theprovisionofpapermoneyandcoins,depos- used card processing and credit-related services than iting or clearing checks or drafts from business customers, the werewhite-ownedfirms.Hispanic-ownedfirmswere collectionofnightdeposits,andwiretransfers.Creditcardanddebit less likely to have used trust and pension services card processing services: the processing of credit card receipts, of signature-based debit (check-card) transactions, and of PIN-based than were non-Hispanic-owned firms. These differdebit transactions (credit card processing was previously combined ences could be related to firm size and industry withtransactionservicesbutwasaskedaboutseparatelyinthe2003 classification. survey).Cashmanagementservices:theprovisionofsweepaccounts, zero-balanceaccounts,lockboxservices,andotherservicesdesigned toautomaticallyinvestliquidfundsinliquid,interest-bearingassets. Transaction and Card Processing Services Credit-relatedservices:theprovisionofbanker’sacceptances,letters ofcredit,andfactoring.Trustandpensionservices:theprovisionof 401(k)plans,pensionfunds,andbusinesstrusts.Brokerageservices: Althoughthecollectionofinformationontransaction brokeringthepurchaseandsaleofstocks,bonds,andothersecurities. services and card processing changed in 2003, the 26.Acomparisonoftheservicesinthe1998transactionsservices data exist to measure whether a firm used a transaccategorytosimilarservicesfromthe2003surveyisdiscussedlater,in thesection‘‘TransactionandCardProcessingServices.’’ tion service or a card processor. In the 2003 survey,

A180 Federal Reserve Bulletinh 2006 7. Useoftransactionservicesandcreditcardprocessing Among small businesses, 37 percent used a card bysmallbusinesses,bynumberofemployees,1998and processing service (table 8). The majority of these 2003surveys businesses used a credit card processor (96 percent). Percent The usage pattern varied across firm industry. For 2003 example, 72 percent of retail trade firms used some 1998 typeofcardprocessor,butonly9percentofinsurance Numberofemployees Transaction transaction Transaction Creditcard serviceand service agents and real estate firms did so. The pattern by service processing creditcard processing typeofprocessorused,however,wasthesameforall typesoffirms;creditcardprocessorswerealwaysthe Allfirms ........... 38.9 35.6 56.2 41.1 most common, followed by processors for signature 0–1................. 25.2 14.7 34.3 28.1 2–4................. 34.9 32.1 51.9 35.7 debitcardandPINdebitcardtransactions.Usagealso 5–19................ 46.8 49.0 69.9 53.3 20–49............... 61.0 55.4 80.8 59.2 increased with the size of the firm: 16 percent of the 50–99............... 61.3 50.8 76.2 56.8 100–499 ............ 70.3 49.4 79.8 70.3 very smallest firms (0–1 employees) used credit or debitcardservicesorboth,comparedwith50percent Note:Inthe2003survey,creditcardprocessingwasseparatefromtransaction services; in the 1998 survey, transaction services included credit card ofthelargestfirms(100–499employees). processing.Forfurtherdetails,refertotext. Owner Loans, Credit Cards, and Trade Credit respondentswhoreportedusingacreditordebitcard processor were asked to identify which services they In addition to using credit lines, loans, and capital received. The responses make it possible to separate leases, many small businesses obtain financing by firmsthatusedcreditcardprocessingfromthosethat borrowing from the firm’s owners (owner loans), used only debit card processing. The data indicate borrowingviacreditcards,orborrowingfromsupplithatifthe2003questionwereaskedexactlyasithad ersofgoodsandservices(tradecredit). been in 1998, growth in transaction services would Thesealternativeformsofcreditaredifferentfrom have been observed. In 2003, 56 percent of all firms lines, loans, and leases in a number of ways. For reported using transaction services or credit card example, owner loans are clearly not arm’s-length processing (a category of service comparable to that transactions. In the case of credit cards, the interest asked in 1998), up from the 41 percent of firms who rates charged may exceed the interest rates for other reported doing so in 1998 (table 7). The use of types of loans; moreover, credit cards, unlike typical transaction services increased with firm size, with loans,provideaconvenientmeansofpayingbillsand more than three-fourths of the largest firms using tracking expenses. Trade credit is generally used in themin2003. connection with the purchase of goods and services 8. Useofcreditanddebitcardprocessingservicesbysmallbusinessesandspecifictypeandnumberofservicesused,by numberofemployeesandindustry,2003survey Percentexceptasnoted Typeandmediannumberofservicesusedamongusers ofcardprocessingservices Anycard Characteristic processing Debitcardprocessing, service Creditcard bytypeoftransaction Median numberused processing Signature PIN1 (ofthree) Allfirms ............................... 37.2 95.6 48.8 26.6 2 Numberofemployees 0–1..................................... 16.1 91.1 51.7 17.1 1 2–4..................................... 33.7 95.3 49.4 25.8 2 5–19.................................... 51.1 95.8 48.9 30.8 2 20–49................................... 56.3 98.4 49.2 22.3 2 50–99................................... 51.5 98.7 33.1 21.9 1 100–499 ................................ 50.1 98.6 34.4 19.0 1 StandardIndustrialClassification Constructionandmining................. 13.9 86.5 24.9 18.8 1 Manufacturing........................... 36.5 94.8 31.6 16.8 1 Transportation........................... 37.4 81.8 43.2 20.7 1 Wholesaletrade ......................... 48.2 100.0 26.2 19.3 1 Retailtrade ............................. 72.1 96.8 61.8 33.7 2 Insuranceandrealestate................. 9.3 81.9 42.5 22.1 1 Businessservices........................ 38.5 98.2 47.3 28.3 2 Professionalservices..................... 24.8 93.7 49.1 18.8 2 1.PINPersonalidentificationnumber.

Financial Services Used by Small Businesses: Evidence from the 2003 Survey of Small Business Finances A181 from a specific supplier, whereas funds from lines, cards did not appear to be related to age. Proprietorloans, and leases are often available for general ships were the organizational form most likely to purposes and are not restricted to purchases from a haveusedpersonalcreditcardsandtheleastlikelyto single supplier. Also, when outstanding trade credit haveusedbusinesscards,butproprietorshipsarealso balancesarenotrepaidinarelativelyshortperiod,the generally smaller than other organizational forms. financechargesmayexceedthoseonotherloans. Use of credit cards did not vary much by industry— transportation firms were the least likely to use personalcreditcards,andinsuranceandrealestatefirms Loans from Owners and retail trade firms the least likely to use business Of the small businesses that could have received credit cards. By owner characteristic, non-Hispanic loansfromowners(thatis,thosethatwereorganized white-owned businesses were more likely to use as corporations or partnerships), the proportion with personal credit cards and just about as likely to use such loans rose slightly between the 1998 and 2003 businesscreditcardsasHispanic-ownedornonwhitesurveys,from28percentto30percent(table6.B). owned businesses. Relative to male-owned firms, Because they generally have fewer credit options, female-ownedfirmsweremorelikelytousepersonal smaller firms might seem more likely than larger credit cards and less likely to use business credit firms to borrow from their owners. This was not the cards. case in 2003. The incidence of owner loans differed As indicated earlier, some firms may use credit acrosssizegroupswithnospecificpatternexceptthat cardssimplyfortheconvenienceofmakingpayments the smallest size groups (0–1 employee, less than and tracking expenses. In 2003, 73 percent of per- $25,000insales,orlessthan$25,000inassets)were sonal credit card users and 71 percent of business the least likely to have reported owner loans. Only credit card users reported that they paid their stateaboutone-fourthofthesmallestfirmsreportedowner ment balance in full by the payment-due date. For loans,versus30percentforallsmallbusinesses. these businesses, credit cards appear to be used primarilyfortransactionalconvenience. Credit Cards Trade Credit Smallbusinessesweresomewhatmorelikelytohave used credit cards in 2003 than in 1998. The percent- Afirmreceivestradecreditwhenitssupplierscollect age that used personal credit cards (47 percent in paymentafter,ratherthanatthetimeof,thereceiptof 2003) remained about the same, but the percentage goods or services. Most trade credit is extended to thatusedbusinesscreditcardsincreasedfrom34per- firmsforaveryshortperiod(thirtyorsixtydays)and centto48percent. is always granted in connection with specific pur- Credit cards are a convenient means of making chases.Businessesusetradecredittoreducetransacpaymentsandtrackingexpenses.Anecdotalevidence tion costs and sometimes as a form of financing. suggeststhatmanysmallerandnewerbusinessesalso Allowing available funds to be used for other puruse credit cards as a source of credit, even though posesisonewaythattradecreditreducesthetransacthey may have higher interest rates than other forms tion cost that businesses would incur if they had to ofcredit.Lenderssometimesrationcredittohigh-risk make payment at the time of receipt. If the firm does firms. Thus, firms just starting out and those having not pay the bill for the goods or services on time, littlecredithistorymaybeperceivedashighriskand tradecreditbecomesaformoffinancing.Becausethe may therefore rely on credit cards as a substitute for interest rates charged on overdue balances can be other types of loans. The descriptive statistics on the quite high, it is reasonable to expect that the firms useofcreditcardsareonlysomewhatconsistentwith using trade credit for longer-term financing purposes this hypothesis. Personal credit card use was highest wouldhavehaddifficultyobtainingcreditfromother amongthesmallestfirms,averagingabout50percent. sources.27 But even among the largest firms, about 33 percent Trade credit was used by 60 percent of small reported using personal credit cards for business businesses in 2003, an incidence of use that exceeds expenses.Incontrast,theuseofbusinesscreditcards thatforallotherfinancialservicesexceptchecking.In generallyincreasedwithfirmsize—roughlyone-third ofthesmallestfirmsusedthem,comparedwithabout three-fifthsofthelargerfirms. 27.Firmswereaskedtoreportonthemonthlypenaltytheywould bechargediftheypaidaftertheduedate.Theaverageratereported The use of personal credit cards did decline somewasalittlemorethan1percent.Themedian,75thpercentile,and90th what with firm age, but the use of business credit percentilewere1percent,1.5percent,and2percentrespectively.

A182 Federal Reserve Bulletinh 2006 9. Useofselectedsuppliersoffinancialservicesbysmallbusinesses,byselectedcharacteristicsoffirms,2003survey Percent Nondepository Depository Primarilyfinancial Othernondepository Characteristic su A pp n l y ier Com- Thrift Any Finance Broker- Leasing a I n n c s e ur o - r C c a h r e d ck or Family Any mercial institu- Credit Any com- ageor com- mort- Any pro- Govern- andin- Other3 bank tion1 union panyor pension pany gage cessing ment dividfactor firm com- firm2 uals pany Allfirms 2003................. 96.4 95.9 86.5 13.8 8.1 54.1 40.3 25.2 14.9 4.5 5.4 24.3 13.4 1.2 6.6 5.5 1998................. 96.1 95.2 87.3 9.2 5.9 39.9 31.0 15.5 11.3 7.0 3.7 15.6 4.0 1.0 6.2 5.6 Numberofemployees 0–1.................. 88.9 87.2 71.6 14.2 9.6 35.6 25.7 16.4 8.1 * 3.0 14.7 6.8 * 4.0 3.1 2–4.................. 96.9 96.4 86.2 13.9 9.2 47.1 33.4 20.8 10.6 3.0 4.2 21.2 12.5 * 5.5 4.3 5–9.................. 99.8 99.5 92.8 13.5 6.6 66.9 48.5 30.8 18.2 6.2 5.3 33.2 18.5 * 9.2 7.0 10–19................ 100.0 100.0 95.1 14.2 7.1 71.4 57.3 34.5 25.2 7.5 6.1 32.3 18.2 * 10.2 8.6 20–49................ 100.0 100.0 97.8 13.0 4.4 75.9 62.6 40.3 27.5 7.0 14.9 30.5 16.3 3.4 7.7 7.2 50–99................ 100.0 99.9 98.7 9.2 * 86.4 78.0 46.0 36.4 13.3 18.6 33.0 14.2 * 4.6 14.8 100–499 ............. 100.0 100.0 96.8 17.0 * 82.6 75.6 42.1 40.9 14.1 16.7 30.7 11.6 4.8 10.9 11.7 Sales(thousands ofdollars) Lessthan25 ......... 81.4 80.0 65.5 12.3 9.0 24.7 13.7 9.2 4.4 * * 13.5 6.7 * 5.2 * 25–49................ 97.2 96.1 78.0 19.3 14.0 42.7 28.4 17.5 9.9 * * 19.8 10.4 * 6.4 * 50–99................ 97.6 96.0 84.9 13.2 9.6 43.8 28.7 17.7 10.3 * * 18.6 9.4 * 5.3 * 100–249 ............. 98.9 98.9 89.8 13.0 8.1 49.4 34.9 22.1 9.0 4.3 4.1 22.8 13.4 * 6.7 3.4 250–499 ............. 99.6 99.4 91.8 14.7 7.0 63.8 45.8 26.8 16.1 5.5 7.1 29.1 16.9 * 8.4 6.2 500–999 ............. 100.0 99.8 94.6 12.4 6.4 72.0 57.0 38.4 17.4 7.4 5.9 36.2 21.0 * 8.6 8.2 1,000–2,499.......... 100.0 99.6 94.4 15.4 5.4 71.8 60.7 34.5 29.7 8.5 7.4 29.7 18.3 * 5.0 9.9 2,500–4,999.......... 100.0 100.0 98.4 10.4 * 77.3 68.3 46.9 31.3 5.0 14.4 27.3 11.8 4.5 7.5 8.6 5,000–9,999.......... 100.0 99.9 99.4 10.7 * 89.0 79.9 55.5 41.2 8.1 13.4 27.5 9.0 * 6.1 13.8 10,000ormore....... 100.0 100.0 96.3 11.8 * 84.1 79.7 42.6 48.8 14.6 17.2 30.8 13.0 3.0 7.2 12.9 Assets(thousands ofdollars) Lessthan25 ......... 90.0 89.0 73.0 13.7 8.9 35.7 23.7 15.5 6.5 2.2 2.2 16.9 8.0 * 5.1 3.6 25–49................ 98.2 97.7 89.6 15.9 6.9 43.5 30.0 20.3 10.1 * * 18.3 12.3 * * * 50–99................ 99.8 99.4 90.5 11.0 12.7 59.8 41.9 27.1 14.3 3.8 6.1 26.8 15.6 * 9.5 4.4 100–249 ............. 99.1 98.7 92.6 11.4 6.6 62.6 44.8 26.2 17.2 5.3 5.5 29.8 19.9 * 5.0 6.3 250–499 ............. 99.7 99.3 94.2 15.2 6.2 71.0 55.0 32.2 19.9 10.2 5.3 34.5 19.8 * 9.3 8.2 500–999 ............. 100.0 100.0 95.5 16.4 9.1 66.3 56.6 34.5 22.3 5.4 8.9 32.4 15.4 * 12.1 8.4 1,000–2,499.......... 100.0 100.0 95.2 15.4 5.3 79.3 69.4 44.3 33.7 8.6 13.1 27.8 11.0 * 7.8 10.4 2,500–4,999.......... 100.0 100.0 94.9 14.0 * 79.6 74.1 48.2 32.9 10.1 20.4 30.0 10.6 * 6.1 11.8 5,000ormore........ 100.0 98.5 97.8 19.1 * 74.9 71.1 39.3 40.3 7.8 16.5 21.7 8.5 2.4 5.7 9.6 Organizationalform Proprietorship ........ 93.0 92.0 78.5 15.9 11.1 44.3 31.3 20.1 9.8 2.2 4.4 19.8 10.3 * 6.8 3.8 Partnership........... 97.3 96.9 89.6 14.0 9.1 49.8 35.1 20.3 13.1 4.4 4.9 22.4 11.9 * * 7.1 Scorporation......... 99.5 99.2 93.2 12.5 5.0 64.0 48.9 30.8 18.3 7.3 5.5 29.9 17.6 2.3 7.4 6.5 Ccorporation ........ 99.6 99.4 94.1 10.4 5.3 64.3 51.7 30.9 23.6 5.4 8.2 27.1 14.6 1.4 6.1 7.1 StandardIndustrial Classification Constructionand mining .......... 96.5 96.2 86.7 13.3 9.1 51.4 45.0 34.3 10.6 4.8 5.7 14.8 5.2 * 3.4 4.4 Manufacturing........ 97.7 97.7 90.6 11.6 8.1 59.2 42.6 24.5 16.8 6.1 4.7 30.6 13.5 2.4 10.6 7.8 Transportation........ 99.5 99.5 90.9 10.0 15.7 61.9 47.5 38.4 6.3 6.6 5.8 27.1 11.0 * 13.1 6.7 Wholesaletrade ...... 98.8 97.7 92.0 10.5 * 62.7 47.8 31.0 21.7 3.2 5.8 27.7 17.8 * 5.2 6.8 Retailtrade .......... 97.4 97.1 88.4 15.7 5.6 59.4 31.9 22.3 8.7 4.3 4.9 39.8 28.3 1.9 7.8 6.1 Insuranceand realestate ....... 96.3 96.0 90.2 18.3 7.8 46.4 43.3 19.3 21.2 * 12.2 8.3 * * * * Businessservices..... 94.9 94.0 82.0 14.2 9.6 49.9 35.3 23.4 11.3 3.4 4.4 24.3 13.9 .8 7.2 4.9 Professionalservices.. 95.9 95.1 85.0 12.7 7.4 53.0 46.0 22.9 24.3 5.7 4.5 17.8 7.5 * 5.4 5.7 Yearsundercurrent ownership 0–4.................. 96.1 95.0 83.5 14.2 9.3 51.1 31.8 23.3 5.5 4.3 3.3 29.1 15.7 * 9.5 5.3 5–9.................. 95.8 95.4 86.5 12.2 7.7 56.2 39.6 25.9 12.4 4.6 4.7 27.7 15.0 * 8.0 6.8 10–14................ 98.1 97.7 88.5 12.4 10.5 58.4 45.8 26.5 20.9 5.9 4.4 22.6 12.6 * 4.3 5.9 15–19................ 94.7 94.1 83.6 15.7 5.3 54.7 44.8 27.1 18.8 3.4 6.5 23.3 15.4 * 5.3 4.9 20–24................ 96.0 95.3 85.1 15.7 6.6 55.4 45.1 25.3 18.8 4.8 8.2 21.4 12.5 * 5.0 4.6 25ormore........... 97.8 97.3 90.9 14.0 8.0 49.7 40.0 23.8 18.5 3.8 7.1 18.5 8.6 1.2 5.5 4.6 1998, 62 percent of small businesses reported using firmsandothergroupsaresimilartothedifferencesin trade credit. Use generally increased with firm size, usebetweensmallerandlargerfirms. rising from about one-third of the smallest firms to The use of trade credit was most common among morethan85percentofthelargestfirms.Youngfirms, firms in construction, manufacturing, and wholesale proprietorships, and firms owned by nonwhites, His- and retail trade—industries for which nonlabor expanics, or females were less likely than others to use penses, such those for equipment and inventory, are this service.The differences between these groups of large relative to labor costs. Among industries for

Financial Services Used by Small Businesses: Evidence from the 2003 Survey of Small Business Finances A183 9.—Continued Percent Nondepository Depository Primarilyfinancial Othernondepository Characteristic su A pp n l y ier Com- Thrift Any Finance Broker- Leasing a I n n c s e ur o - r C c a h r e d ck or Family Any mercial institu- Credit Any com- ageor com- mort- Any pro- Govern- andin- Other3 bank tion1 union panyor pension pany gage cessing ment dividfactor firm com- firm2 uals pany Censusarea ofmainoffıce Northeast ............. 95.5 94.6 77.4 28.4 6.3 54.9 40.2 26.1 17.6 3.1 4.6 25.2 14.8 * 6.9 4.7 NewEngland....... 97.9 97.5 63.9 51.7 * 53.2 40.7 28.8 16.6 4.0 6.1 23.9 16.5 * * * MiddleAtlantic..... 94.4 93.4 83.2 18.3 6.2 55.6 40.0 25.0 18.1 2.8 3.9 25.8 14.1 * 7.1 5.7 Midwest.............. 98.4 98.1 92.4 7.8 9.5 55.2 41.9 22.9 17.8 5.2 5.8 24.1 13.0 2.6 4.7 6.2 EastNorthCentral.. 98.0 97.6 91.2 9.6 9.8 55.8 43.1 21.5 19.0 5.9 5.4 24.8 12.2 * 5.3 6.3 WestNorthCentral . 99.3 99.3 94.7 4.0 8.9 54.0 39.3 25.8 15.4 3.7 6.7 22.6 14.8 * 3.5 6.0 South................. 96.5 95.8 89.0 9.8 7.0 55.7 40.8 27.4 11.6 5.3 5.9 24.8 13.9 .7 7.0 5.3 SouthAtlantic ...... 97.5 96.7 88.8 12.3 5.6 60.0 44.9 30.1 12.5 6.0 6.4 26.6 15.9 * 6.3 5.7 EastSouthCentral.. 94.9 94.9 89.5 * * 45.7 33.2 21.6 11.0 4.3 4.7 19.6 9.5 * * 4.4 WestSouthCentral . 95.5 94.7 89.1 7.7 7.1 52.9 37.1 25.7 10.2 4.6 5.5 24.0 12.4 * 8.3 5.1 West ................. 95.4 94.9 85.2 12.8 10.0 50.1 38.3 23.1 14.9 3.7 4.9 23.0 11.9 1.3 7.5 5.7 Mountain........... 95.6 94.8 90.0 7.1 9.5 52.1 37.9 26.5 10.3 2.1 4.6 24.7 11.3 * 9.8 5.8 Pacific ............. 95.4 95.0 83.0 15.4 10.2 49.2 38.5 21.5 17.0 4.5 5.1 22.3 12.2 .6 6.5 5.7 Urbanization atmainoffıce Urban ................ 96.5 95.9 86.2 14.2 7.7 55.4 41.9 25.7 15.8 4.8 5.5 25.1 13.7 1.3 6.7 6.0 Rural................. 96.2 95.9 87.5 12.2 9.9 48.7 34.3 23.0 11.7 3.3 4.8 21.3 12.5 .8 6.4 3.3 Numberofoffıces One .................. 95.9 95.4 85.5 14.0 7.8 52.0 38.4 24.6 13.7 3.6 5.2 23.1 13.3 .8 6.0 5.1 Two .................. 99.2 98.7 90.8 14.3 10.0 68.9 50.6 28.7 19.0 9.2 4.8 34.6 16.7 * 11.5 7.0 Threeormore ........ 100.0 99.2 96.3 9.0 11.0 62.7 54.0 28.8 29.0 10.5 10.0 25.5 9.3 2.4 7.4 8.8 Salesarea Primarilywithinthe UnitedStates..... 96.4 95.9 86.7 13.8 8.1 53.8 40.1 25.0 14.7 4.4 5.3 24.0 13.2 1.2 6.5 5.5 Internationalor global............ 97.1 95.9 82.8 14.0 7.8 60.1 45.5 28.5 19.2 4.8 7.1 29.6 17.7 * 9.3 4.6 Management Byowner............. 96.2 95.6 86.0 13.6 8.2 53.0 39.3 24.6 14.3 4.4 5.0 24.0 13.2 1.2 6.7 5.2 Hired................. 99.9 99.9 93.4 16.7 7.6 69.3 56.2 34.5 24.7 5.3 10.5 28.9 16.9 * 5.3 8.9 Race,ethnicity,and sexofmajorityowners NonwhiteorHispanic.. 95.5 94.9 85.8 12.0 7.4 53.3 35.3 25.7 10.9 2.1 4.9 27.5 15.0 * 7.8 5.1 Non-Hispanicwhite ... 96.5 95.9 86.5 14.0 8.2 53.7 40.6 24.7 15.3 4.8 5.3 23.6 13.1 1.2 6.4 5.5 White ................ 96.5 95.9 86.3 13.9 8.4 53.9 40.6 25.0 15.3 4.6 5.4 23.7 13.2 1.2 6.4 5.5 Black................. 91.7 91.7 81.0 13.7 * 46.2 36.0 28.9 * * * 20.6 * * * * Asian,NativeHawaiian, orotherPacific Islander........... 99.1 97.3 93.6 13.0 * 59.8 33.9 22.9 12.3 * * 36.2 23.8 * * * AmericanIndianor AlaskaNative .... 92.7 92.7 81.5 * * 45.0 22.7 18.7 * * * * * * * * Hispanic.............. 94.9 94.9 82.4 * * 54.7 40.0 28.9 15.1 * 7.6 25.4 14.1 * * * Non-Hispanic ......... 96.5 95.9 86.6 13.9 8.0 53.9 40.3 25.0 14.9 4.6 5.2 24.2 13.4 1.3 6.6 5.5 Female ............... 92.3 91.4 79.8 13.9 8.2 45.0 30.5 20.3 10.5 3.4 2.7 23.9 14.2 1.1 6.5 3.7 Male ................. 97.4 96.8 88.2 13.2 8.1 57.1 44.0 26.8 17.0 4.8 6.1 24.4 12.9 1.2 6.9 6.1 Ownershipdivided equallybysex.... 98.6 98.6 89.1 16.8 8.2 53.8 38.6 25.1 12.1 4.6 6.4 24.3 14.4 * 5.3 5.1 Note:Fordefinitionofsalesareas,refertotable2,note3. 3.Includesbusinessfirms,suppliers,andventurecapitalfirms. 1.Savingsbankorsavingsassociation. *Fewerthanfifteenobservations. 2.Includescreditanddebitcardprocessingfirms. whichlabor’sshareofcostsishigh,suchasbusiness pay on time paid late only about 30 percent of the and professional services, the use of trade credit was time, which suggests that, even for these firms, the somewhatlesscommon. majoruseoftradecreditwasfortransactionpurposes. As indicated earlier, trade credit can be used for transaction purposes and for financing. Some of the use patterns (for example, variation by industry) are SUPPLIERS OF FINANCIAL SERVICES USED more consistent with the transaction hypothesis. BY SMALL BUSINESSES Among the firms that reported using trade credit in 2003, 59 percent reported that they always paid the Thesuppliersoffinancialservicestosmallbusinesses debt by the due date. The firms that did not always consistofdepositoryinstitutions—commercialbanks,

A184 Federal Reserve Bulletinh 2006 thriftinstitutions(savingsbanksandsavingsassocia- (8 percent). In general, the percentage of firms using tions), and credit unions—and nondepository institu- commercial banks increased with firm size, and once tions. Nondepositories consist of primarily financial a certain size threshold was crossed (for example, at nondepositories—financecompaniesandfactors,bro- least twenty employees or at least $2.5 million in kerage and pension firms, leasing companies, and sales),virtuallyallfirms(97percentormore)usedat insurance and mortgage companies—and other least one commercial bank. In contrast, the use of nondepositories—credit card and check processing credit unions declined with size, and the use of thrift firms, government sources, family and individuals, institutions did not vary systematically with size. business firms, suppliers, venture capital firms, and Proprietorships, which are generally the smallest miscellaneous types. The survey collected informa- firms,weresomewhatlesslikelythanotherfirmtypes tiononthesourcesofcheckingandsavingsaccounts; tousecommercialbanksandmorelikelytousecredit linesofcredit,loans,andcapitalleases;andfinancial unionsandthriftinstitutions. managementservices.28 The use of thrift institutions increased from 9 per- In 2003, depository institutions were used by centtonearly14percentbetween1998and2003,and 96 percent of all firms, roughly the same percentage the use of credit unions increased from 6 percent to as in 1998 (table 9). Among depository institutions, 8 percent. The increased use of thrift institutions the use of commercial banks remained about the reversesdeclinesobservedinearliersurveys,declines same, whereas the use of thrift institutions and credit thathadbeenattributedtothedecreaseinthenumber unionsincreased. of thrifts during the 1990s. The increased use of In contrast, the proportion of firms using nonde- thriftsandcreditunionssuggeststhatthederegulation pository institutions increased from 40 percent in of business lending by those institutions and the 1998 to 54 percent in 2003, with increases recorded expansion in potential credit union membership perboth for primarily financial nondepository sources mittedbytherelaxationinthedefinitionof‘‘common and for other nondepository sources. Among the bonds’’by the National Credit UnionAdministration primarily financial nondepository sources, only leas- inrecentyearsmayhaveenabledtheseinstitutionsto ingcompanieswereusedsomewhatlessin2003than better meet the financial service demands of small in 1998. Among other nondepository institutions, businesses. credit card processors logged the largest increase As was true in earlier surveys, small businesses in between 1998 and 2003.29These changes are consis- New England were much more likely to use thrift tent with the finding that the percentage of small institutions(52percent)thanwerefirmsinotherparts businesses that used credit card processing services ofthecountry.Thisfindingisconsistentwiththefact increasedovertheperiodbetweensurveys. that thrifts account for a larger proportion of depositories in New England than in other areas of the Depository Financial Institutions country. Moreover, thrifts in New England tend to look more like commercial banks in terms of their Depositoryinstitutionsprovidedatleastonefinancial business lending than do thrifts outside of New Enservice to about 96 percent of small businesses in gland.30 Credit unions were most likely to have been 2003 (table 9), roughly the same percentage of busi- usedbyfirmslocatedinthePacificpartoftheWest. nessesthathadacheckingorsavingsaccountin2003 Black-ownedandHispanic-ownedbusinesseswere (table 6.A). Commercial banks continued to be used less likely than non-Hispanic-owned or white-owned by a far larger proportion of firms (87 percent) than firms to use commercial banks. Asian-owned firms were thrift institutions (14 percent) or credit unions weremorelikelytousecommercialbanksthanother ownership groups. Also, female-owned firms were less likely than male-owned firms to use commercial 28.Noinformationonthesourcesofownerloans,creditcards,and tradecreditwascollected. banks.In1998,thegroupsmostlikelytousecommer- 29.As noted earlier, an additional service—credit card and debit cial banks were also the least likely to use thrift card processing—was added to the list of financial management institutions and credit unions. In 2003, this apparent services in the 2003 survey. This additional service may have been partlyresponsiblefortheobservedincreaseintheuseofnondeposi- substitution across institution types altered sometorysources—especiallyofcreditcardprocessingsources.However, what. The use of thrifts and credit unions was often thedatasuggestthatpartoftheincreaseintheuseofnondepository greatest for those firm types that were also the most sourceswasindependentoftheadditionofthecreditcardprocessing service. For most types of nondepository sources, including finance companies, brokerages, insurance and mortgage companies, government, and family and individuals, the incidence increased between 30.StevenJ.PilloffandRobinA.Prager(1998),‘‘ThriftInvolve- 1998and2003.Theonlynondepositorysourcesforwhichincidence mentinCommercialandIndustrialLending,’’FederalReserveBulledecreasedwereleasingcompaniesand‘‘other’’types. tin,vol.84(December),pp.1025–37.

Financial Services Used by Small Businesses: Evidence from the 2003 Survey of Small Business Finances A185 likelytousecommercialbanks.Forexample,theuse family and individuals, other businesses, supplier of thrifts was highest for white-owned firms and for businesses,andventurecapitalfirms),upfrom16perfirmsownedequallybymalesandfemales;theuseof cent in 1998. Most of this increase is due to a rise in credit unions was highest for white-owned firms and the use of card and check processors (from 4 percent formale-ownedfirms. in 1998 to 13 percent in 2003), which may, in turn, partly reflect increased use and acceptance of credit Nondepository Sources and debit cards by small businesses.31 The use of other subgroups was largely unchanged from 1998. Small businesses obtained their financial services About 1 percent used government, 7 percent used from a variety of sources besides depository institufamily and individuals, and 6 percent used other tions. Between 1998 and 2003, the incidence of sources,includingotherbusinessfirms.32 nondepository use by small firms increased from The use of other nondepository sources did not 40 percent to 54 percent (table 9). Both primarily consistently increase with firm size, although it was financial nondepositories and other nondepositories least likely for the smallest firms. For example, the wereusedmorefrequentlyin2003thanin1998. use of other nondepository sources increased from 15percentoffirmswithfewerthantwoemployeesto Primarily Financial Nondepositories 33percentforfirmswithfivetonineemployees.For largerfirms,thepercentageusingsuchsourcesgener- Primarily financial nondepositories were a source of allyremainedatthehigherlevels.Proprietorshipsand financial services for about 40 percent of all small partnerships were less likely than were corporations firms, a sizable increase over the roughly 30 percent tousethesesources. incidencereportedin1998.Andasin1998,themost Younger firms were more likely to use other noncommonly used financial nondepositories were fidepositories, including family and individuals, than nancecompanies,followedbybrokeragecompanies. were older firms. Younger firms sometimes have The use of financial nondepositories (and of each difficulty borrowing from financial institutions, in subgroup of suppliers within that group) is strongly part because financial institutions often require that relatedtofirmsize,increasingfromaboutone-fourth prospectiveborrowersprovideseveralyearsoffinanof the smallest firms to about three-fourths of the cialstatementswiththeirloanapplications.Nonfinanlargestfirms.Forexample,16percentofthesmallest cialsources,suchasindividualswhomaybefamiliar firms (those with fewer than 2 employees) used with the prospective borrowers, may be better posifinance companies, and 8 percent used brokerages, tionedtoevaluatecreditworthinessandtomonitorthe whereas about 42 percent of firms with 100 or more financial condition of younger firms, or alternatively employees used finance companies and 41 percent they may have different credit standards than finanused brokerage companies. Use of financial nondecialsources.Thesurveyresultsshowthatin2003the positories also differed by organizational form, ranguse of family and individuals was most common ing from 31 percent of proprietorships to 52 percent amongyoungerfirms. of C corporations. Proprietorships and partnerships were about half as likely as corporations to use brokerages and about two-thirds as likely to use USE OF FINANCIAL SERVICES SUPPLIERS, financecompanies. BY SERVICE The use of financial nondepositories also varied The data reviewed thus far have examined variations with the race, ethnicity, and sex of the business in the use of financial services by firms according to owners. White-owned and male-owned firms used characteristicsofthefirmandvariationsinthesource financialnondepositories,financecompanies,brokerof financial services used by firms according to the ages, and leasing companies more often than did type of firm. This section reports on the types of other types of firms. The differences were largest among groups using brokerage firms: For example, 15 percent of white-owned firms used brokerages, 31. As indicated earlier, some of the observed increase is likely compared with 11 percent of nonwhite or Hispanic- causedbyexplicitlyaskingaboutcreditcardprocessingservicesinthe 2003survey.Before2003,creditcardprocessingwasincludedinthe ownedfirms. questionabouttransactionservices. 32.The1percentfigurelikelyunderstatesthetrueroleofgovern- Other Nondepositories ment in providing financial services to small businesses. Many governmententities,suchastheU.S.SmallBusinessAdministration, providecreditguarantees,whichensurerepaymentofsmallbusiness In 2003, 24 percent of firms used other nondepositoloans made by institutional lenders such as commercial banks and ries(suchascardandcheckprocessors,government, thriftinstitutions.

A186 Federal Reserve Bulletinh 2006 10. Useofselectedsuppliersoffinancialservicesbysmallbusinesses,byselectedservice,2003survey Percent Nondepository Depository Primarilyfinancial Othernondepository Service su A pp n l y ier Com- Thrift Any Finance Broker- Leasing a I n n c s e ur o - r C c a h r e d ck or Family Credit com- ageor mort- Govern- andin- Any mercial institu- Any com- Any pro- Other union panyor pension gage ment dividbank tion pany cessing factor firm com- uals firm pany Any................... 96.4 95.9 86.5 13.8 8.1 54.1 40.3 25.2 14.9 4.5 5.4 24.3 13.4 1.2 6.6 5.5 Liquidassetaccount ... 95.0 94.5 83.2 10.9 4.9 4.5 4.2 .4 3.8 0 .1 .3 .1 0 0 .2 Checking ........... 94.6 94.1 82.7 10.5 4.4 1.8 1.7 .1 1.6 0 0 .2 .1 0 0 .1 Savings ............. 22.1 20.0 16.2 2.4 2.3 3.1 3.0 .4 2.6 0 .1 .1 .0 0 0 .1 Creditline,loan, orcapitallease.... 60.4 46.4 41.1 5.5 3.9 33.3 26.6 22.2 .8 4.3 2.3 10.4 .3 1.1 6.5 2.9 Creditline .......... 34.3 32.4 29.5 2.8 .9 3.6 2.9 2.2 .4 .1 .4 .8 0 .1 .1 .6 Mortgage ........... 13.3 10.8 9.1 1.9 .3 3.4 1.9 .6 .1 0 1.2 1.5 0 .4 .9 .2 Vehicleloan......... 25.5 13.2 10.0 1.1 2.5 15.2 14.8 14.5 0 .3 .1 .4 0 0 .2 .2 Equipmentloan ..... 10.3 5.1 4.3 .5 .4 5.8 4.4 3.4 .2 1.0 .1 1.5 .1 .1 .5 .8 Capitallease ........ 8.7 1.1 1.0 .1 0 7.8 6.9 4.6 0 2.9 0 1.1 .1 0 .1 1.0 Other ............... 10.1 3.4 3.1 .2 .2 7.2 1.4 .6 .1 .2 .6 5.9 0 .5 4.9 .5 Financialmanagement . 64.8 53.2 48.2 5.6 2.2 33.1 20.1 5.1 13.3 .3 3.2 16.3 13.3 .1 .3 2.8 Transaction ......... 39.0 38.3 34.1 4.0 1.9 3.0 2.0 .8 1.0 0 .2 1.1 .6 0 .2 .4 Creditanddebitcard processing ...... 37.2 22.5 20.7 1.7 .4 16.7 3.1 2.8 .1 .1 .2 14.2 13.1 0 0 1.0 Cashmanagement ... 6.7 6.2 5.8 .4 .1 .8 .7 .2 .5 0 0 .1 0 .1 0 0 Credit-related ....... 5.0 4.1 3.8 .4 .2 1.4 1.0 .8 .1 .1 0 .4 .1 0 0 .3 Brokerage........... 5.6 .8 .8 0 .1 5.0 4.9 .3 4.5 0 .3 .2 0 0 0 .2 Trustandpension ... 17.2 3.6 3.2 .4 .2 14.3 13.2 .8 10.1 .1 2.7 1.2 .1 0 .1 1.0 Note:Refertonumberednotestotable9. financial services provided to small businesses by were also the dominant supplier of savings accounts, eachtypeoffinancialservicesupplier. faroutpacingthenextmostcommonproviders(thrifts, Not all financial service suppliers provide the full creditunions,andbrokerages). range of financial services (table 10). Depository institutionswereusedforthefullrangeofservices— Credit Lines, Loans, and Capital Leases checking and savings (95 percent of firms); lines, loans, and leases (46 percent); and financial manage- In 2003, commercial banks were the most common ment services (53 percent). In fact, a much larger supplier of credit lines, loans, and capital leases, percentage of firms used depository institutions for providing about 41 percent of firms with such sereach of these services than used nondepository vices, slightly up from 39 percent in 1998 (table 10). sources. One in four firms obtained lines, loans, or Primarily financial nondepositories and family and leases from primarily financial nondepositories, and individuals were also important suppliers. In 2003, one in five firms obtained financial management one-fourth of businesses obtained lines from, or had services from primarily financial nondepositories. outstanding balances on loans or leases with, prima- Among other nondepositories, used by about one in rily financial nondepositories (specifically, finance four firms, credit card processors were important and leasing companies); 7 percent had loans from sourcesoffinancialmanagementservices,andfamily familyandindividuals.Thesepercentagesweresomeandindividualsweremostoftenusedforcredit. what lower in 1998. Although suppliers other than commercial banks were important sources of lines, Checking and Savings Accounts loans, and leases, commercial banks in 2003 were about two times more likely than finance companies As was the case in previous surveys, commercial to have been the source of these services for small banks dominated the provision of checking accounts businesses, six times more likely than family and tosmallbusinessesin2003,supplyingthemto83per- individuals, and ten times more likely than leasing centofallsmallfirms(table10).Firmsalsoobtained companies. checking accounts from thrift institutions (11 per- Creditlines,usedbyaboutone-thirdofbusinesses, cent),creditunions(4percent),andbrokerages(2per- were the most commonly used form of credit. They cent).Noothertypeofsupplierprovidedmorethana were supplied primarily by commercial banks, thrift trivialshareofcheckingaccounts.Commercialbanks institutions, and finance companies. Commercial

Financial Services Used by Small Businesses: Evidence from the 2003 Survey of Small Business Finances A187 banks were the most important source and were ten istics of small businesses and on the types and times more likely than thrifts and fifteen times more sourcesofcreditandotherfinancialservicestheyuse. likelythanfinancecompaniestohavebeenthesource Theprecedingsurveyscoveredtheyears1987,1993, of this service. Most mortgages used for business and 1998. Although the discussion in this article is purposeswereobtainedfromcommercialbanks(9per- based on descriptive statistics, the data suggest intercent of firms), thrifts (2 percent), mortgage compa- esting behavior patterns and differences in the use of nies (1 percent), and family and individuals (1 per- creditbysmallbusinesses. cent). Vehicle loans were obtained primarily from Straightforwardcomparisonsrevealsomesimilarifinancecompanies(15percentoffirms)andcommerties in the findings from the earlier surveys. In cial banks (10 percent). Equipment loans were also particular,commercialbankshaveremainedthedomiobtained mainly from these sources, with finance nant supplier of financial services for small busicompanies used somewhat less than commercial nesses; these services include checking and savings banks.Leaseswereobtainedmainlyfromfinanceand accounts, most forms of credit other than leases, and leasing companies. However, less than 9 percent of most financial management services other than brobusinesses reported a capital lease in 2003. Finally, kerageservices. family and individuals were the source of ‘‘other’’ Comparisons also reveal some changes between loans for 5 percent of firms, and commercial banks surveys. The share of small businesses that are S were the source of these loans for about 3 percent of corporations has risen at the expense of C corporafirms. Insum,commercialbankswerethedominantsource tions and proprietorships. Computer use, especially oflines,loans,andleases.Bycredittype,commercial for Internet banking and online loan applications, banks were the dominant source of lines, mortgages, increased markedly between 1998 and 2003. The and equipment loans; and they were the second most payment of business expenses with credit cards, importantsourceforvehicleandotherloans.Theonly especially business credit cards, has grown substancredit type for which commercial banks were not an tially between surveys. The incidence of credit lines importantsourcewascapitalleases.33 and vehicle loans has increased, whereas the incidence of capital leases declined somewhat. Since the Financial Management Services 1987survey,smallbusinesseshaveincreasinglyused nondepository institutions to obtain some of their Commercial banks were the dominant supplier of financialservices.However,despitethegrowthinthe financialmanagementservices,providingalmosthalf use of nondepository sources—from 25 percent of ofsmallbusinesseswiththoseservicesin2003(table firms in 1987 to 54 percent in 2003—commercial 10).Brokeragesandcreditcardprocessingfirms,tied banks remained the dominant supplier of most finanas the second most common source of financial cialservices. management services, were each used by 13 percent The2003surveyalsoprovidessomenewinformaof firms. By individual service, commercial banks tionunavailableinprevioussurveys.Inparticular,the were the dominant supplier of transaction services (used by 34 percent of firms), credit card processing survey collected demographic characteristics for up (21 percent), cash services (6 percent), and credit to three individual owners and thereby helped to services(4percent).Brokerageswerethemostwidely refine the firm-level measurement of characteristics used source of trust and pension services (used by suchasrace,ethnicity,sex,andownershipconcentra- 10 percent of firms) and brokerage services (5 per- tion. In addition, the survey collected information on cent). Credit card processing firms, used by 13 per- the use of nonstandard work arrangements by firms cent of firms for credit and debit card processing, and about their use of credit and debit card processwere the second most common supplier of these ing. services. Explainingthedifferencesamongfirmsintheiruse of financing and, more fundamentally, understanding SUMMARY thefactorsthataffectsmallbusinessfinancingrequire The 2003 Survey of Small Business Finances, the arigorousanalyticalframeworkthataccountsforthe fourth in a series sponsored by the Federal Reserve financialcharacteristicsofborrowersandthemarkets Board,providesdetailedinformationonthecharacter- in which they operate. Although the use of such a framework is beyond the scope of this article, the finalsurveydatawillprovideconsiderableopportuni- 33. Of the six types of credit, capital leases were used by the smallestpercentageoffirms(8.7percent). tiesformoreformalandcompleteanalyses.

A188 Federal Reserve Bulletinh 2006 A.1. Smallbusinessesgroupedbyselectedcharacteristicsanddistributedbyselectedcharacteristicsoffirms,2003survey Percent Yearsunder Urbanization Organizational Majorityowners Numberofemployees current atmainoffice form ownership All Characteristic firms Non- Non- Maleor Prowhite His- divided 10or Female 0–1 2–4 5–19 20–49 50–499 0–9 Urban Rural prietor- Other orHis- panic equally more ship panic white bysex Allfirms ............... 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 Numberofemployees 0–1..................... 20.6 19.7 21.0 18.1 29.5 100 ... ... ... ... 22.9 18.9 20.7 20.1 35.2 8.9 2–4..................... 40.0 46.1 39.3 40.2 39.7 ... 100 ... ... ... 41.1 39.2 39.9 40.4 46.7 34.7 5–9..................... 20.2 21.1 19.8 20.6 18.9 ... ... 65.6 ... ... 20.2 20.1 20.3 19.6 13.4 25.6 10–19................... 10.6 5.7 11.2 11.7 6.6 ... ... 34.4 ... ... 9.6 11.3 10.3 11.5 3.7 16.1 20–49................... 6.0 5.1 6.1 6.5 4.0 ... ... ... 100 ... 4.9 6.8 5.9 6.4 .9 10.1 50–99................... 1.7 1.4 1.7 1.8 .9 ... ... ... ... 63.1 1.0 2.2 1.8 1.3 * 2.9 100–499 ................ 1.0 .8 .9 1.1 .4 ... ... ... ... 37.0 .4 1.4 1.1 .8 * 1.8 Sales(thousands ofdollars) Lessthan25 ............ 14.6 20.0 14.0 11.2 26.6 37.3 15.5 2.2 * * 18.4 11.8 14.3 15.6 24.9 6.3 25–49................... 9.9 12.6 9.8 8.6 14.7 22.5 11.3 2.3 * * 11.4 8.8 9.8 10.5 15.8 5.2 50–99................... 11.6 16.4 11.0 10.7 14.9 17.5 16.8 4.1 * * 12.2 11.1 11.1 13.3 18.5 6.1 100–249 ................ 19.8 14.9 20.4 20.5 17.2 14.4 29.5 16.0 * * 19.7 19.9 19.7 20.3 21.7 18.3 250–499 ................ 14.3 14.1 14.5 15.5 10.3 4.5 16.5 21.1 4.5 * 14.0 14.5 14.5 13.5 9.8 17.8 500–999 ................ 12.2 8.8 12.4 13.5 7.6 * 6.1 27.1 14.6 * 11.4 12.7 12.9 9.2 5.8 17.2 1,000–2,499............. 10.0 7.5 10.3 11.4 5.6 * 3.0 19.6 37.6 12.6 8.6 11.1 10.0 10.3 2.5 16.1 2,500–4,999............. 3.6 2.8 3.8 4.1 2.0 * 1.1 4.9 20.7 14.5 2.4 4.5 3.6 3.6 * 6.3 5,000–9,999............. 2.3 1.8 2.1 2.7 .6 * * 2.6 12.4 27.1 1.3 3.0 2.3 2.2 * 3.7 10,000ormore.......... 1.7 1.2 1.7 2.0 .5 * * * 5.9 43.4 .7 2.5 1.8 1.5 * 3.0 Assets(thousands ofdollars) Lessthan25 ............ 31.3 37.9 30.6 27.0 46.9 63.1 37.5 10.4 * * 37.0 27.1 32.7 26.1 45.7 19.8 25–49................... 12.5 14.4 12.4 12.6 12.4 13.9 16.3 9.2 * * 13.4 11.9 12.4 13.0 14.0 11.3 50–99................... 13.5 12.6 13.6 13.5 13.6 10.5 16.1 14.6 5.3 * 13.3 13.6 13.4 13.6 15.2 12.1 100–249 ................ 15.9 15.0 15.8 17.3 11.0 6.5 15.8 23.5 15.4 3.6 15.3 16.3 15.4 18.0 13.2 18.1 250–499 ................ 10.0 5.5 10.7 11.0 6.8 3.1 7.4 17.6 14.3 6.8 8.5 11.2 10.2 9.6 6.1 13.2 500–999 ................ 7.1 5.0 7.4 7.7 4.8 * 3.8 12.8 17.7 8.4 5.8 8.0 6.3 9.8 2.9 10.4 1,000–2,499............. 5.8 6.3 5.7 6.6 2.8 * 1.8 9.2 25.8 20.7 3.9 7.2 5.6 6.5 2.3 8.6 2,500–4,999............. 1.9 2.8 1.7 2.1 1.0 * * 1.5 9.9 22.7 1.2 2.4 2.0 1.3 * 3.0 5,000ormore........... 2.0 .6 2.2 2.3 .7 * * * 4.8 35.4 1.6 2.4 2.0 2.1 * 3.5 Organizationalform Proprietorship ........... 44.5 47.8 44.6 41.3 56.3 76.0 51.9 24.7 6.9 3.9 42.2 46.2 42.7 51.6 100 ... Partnership.............. 8.7 7.7 8.6 9.0 7.5 * 10.9 9.9 9.7 7.2 12.3 6.0 8.5 9.6 ... 15.7 Scorporation............ 31.0 25.5 31.9 32.9 24.8 15.5 25.9 43.2 49.0 47.3 34.5 28.4 32.8 24.1 ... 55.9 Ccorporation ........... 15.8 19.0 15.0 16.9 11.4 5.8 11.3 22.3 34.4 41.7 11.0 19.3 16.0 14.8 ... 28.4 StandardIndustrial Classification Constructionandmining.. 11.8 5.7 12.8 13.7 5.2 13.1 10.5 12.1 15.0 10.9 10.6 12.7 10.4 17.3 12.0 11.7 Manufacturing........... 7.1 6.0 7.5 7.9 4.4 4.8 6.7 7.6 9.7 20.3 7.3 7.0 7.1 7.1 4.7 9.0 Transportation........... 3.8 2.6 4.0 4.3 1.8 4.5 3.2 4.0 2.6 7.4 3.8 3.8 3.3 5.3 3.6 3.9 Wholesaletrade ......... 5.9 6.8 5.7 6.2 4.7 5.1 6.0 6.0 6.4 8.6 4.8 6.7 6.2 4.8 3.0 8.2 Retailtrade ............. 18.4 21.1 18.1 18.1 19.5 10.7 16.9 23.2 30.9 19.4 21.8 15.9 18.1 19.7 18.0 18.7 Insuranceandrealestate. 7.2 4.4 7.5 7.6 5.8 5.6 9.3 6.7 3.0 3.5 6.0 8.1 7.1 7.5 6.5 7.8 Businessservices........ 25.1 31.6 24.2 23.4 31.3 31.1 26.2 22.5 15.0 15.4 27.7 23.2 24.9 25.7 29.9 21.3 Professionalservices..... 20.7 21.8 20.2 18.7 27.3 25.0 21.3 18.1 17.3 14.5 18.1 22.6 22.8 12.5 22.3 19.3 Yearsundercurrent ownership 0–4..................... 20.6 33.4 18.5 20.0 23.1 26.7 21.5 18.0 13.4 6.6 48.3 ... 22.1 15.0 20.6 20.6 5–9..................... 22.1 23.2 21.9 20.7 26.9 20.7 22.4 23.3 21.5 14.8 51.7 ... 21.8 23.1 19.9 23.8 10–14................... 16.0 13.2 16.5 15.6 17.1 17.3 16.2 14.9 15.9 15.9 ... 27.9 15.8 16.7 16.7 15.4 15–19................... 12.6 10.1 13.0 12.7 12.3 11.5 11.6 14.3 12.7 16.4 ... 22.0 12.5 12.9 12.6 12.6 20–24................... 10.9 9.1 11.1 11.4 8.9 11.0 10.1 11.6 9.9 14.2 ... 18.9 11.2 9.5 12.2 9.7 25ormore.............. 17.9 11.0 18.9 19.6 11.7 12.7 18.3 18.0 26.5 32.1 ... 31.2 16.6 22.8 18.0 17.9 tural, nongovernmental businesses with fewer than APPENDIX A: SURVEY METHODS 500 employees that were in operation both at year- The 2003 Survey of Small Business Finances was end 2003 and at the time of the interview. Most conducted in 2004 and 2005 for the Board of Gover- interviewstookplacebetweenJune2004andJanuary nors of the Federal Reserve System by the National 2005.34 OpinionResearchCenter(NORC),aresearchorganization at the University of Chicago. The survey 34.FurtherdetailsareinNORC(2005),‘‘The2003SurveyofSmall covered a nationally representative sample of U.S. BusinessFinancesMethodologyReport,’’June,www.federalreserve. for-profit, nonfinancial, nonsubsidiary, nonagricul- gov/pubs/oss/oss3/nssbftoc.htm.

Financial Services Used by Small Businesses: Evidence from the 2003 Survey of Small Business Finances A189 A.1.—Continued Percent Yearsunder Urbanization Organizational Majorityowners Numberofemployees current atmainoffice form ownership All Characteristic firms Non- Non- Maleor Prowhite His- divided 10or Female 0–1 2–4 5–19 20–49 50–499 0–9 Urban Rural prietor- Other orHis- panic equally more ship panic white bysex Censusarea ofmainoffıce Northeast ............... 19.8 14.9 20.5 19.1 22.1 24.9 19.1 17.9 18.3 15.5 18.3 20.8 21.5 13.2 20.5 19.2 NewEngland ......... 6.0 * 6.5 5.6 7.2 7.6 5.0 6.1 6.6 4.4 4.8 6.8 6.1 5.6 6.7 5.4 MiddleAtlantic ....... 13.8 12.7 14.0 13.5 14.9 17.3 14.1 11.8 11.7 11.1 13.5 14.0 15.4 7.7 13.9 13.7 Midwest ................ 21.1 10.8 22.5 21.8 18.3 16.6 21.4 22.1 27.8 24.9 19.1 22.6 18.7 30.3 19.9 22.1 EastNorthCentral .... 14.2 8.4 15.0 14.8 12.2 12.2 14.8 13.4 20.4 15.7 13.6 14.7 14.0 15.1 13.8 14.5 WestNorthCentral.... 6.9 2.5 7.5 7.1 6.1 4.4 6.6 8.7 7.4 9.3 5.5 7.9 4.7 15.2 6.0 7.6 South ................... 34.7 44.7 33.6 34.5 35.4 31.6 35.7 36.9 29.7 30.6 39.0 31.6 34.6 35.1 32.6 36.5 SouthAtlantic ........ 18.9 26.2 17.8 19.4 17.3 15.9 18.6 22.3 15.0 16.6 20.7 17.6 19.9 15.2 15.0 22.1 EastSouthCentral .... 5.3 6.2 5.3 5.1 5.9 5.4 6.0 4.3 5.7 4.7 6.1 4.7 3.9 10.6 6.3 4.5 WestSouthCentral.... 10.5 12.4 10.5 10.1 12.1 10.3 11.1 10.3 9.0 9.3 12.1 9.3 10.8 9.4 11.3 9.9 West.................... 24.4 29.6 23.5 24.5 24.3 26.9 23.9 23.1 24.3 28.9 23.6 25.0 25.2 21.4 27.0 22.3 Mountain ............. 7.6 5.6 8.0 7.9 6.7 9.1 6.3 8.2 9.2 5.1 7.5 7.7 6.5 11.9 6.9 8.2 Pacific................ 16.8 24.0 15.5 16.6 17.6 17.8 17.6 14.9 15.0 23.8 16.1 17.3 18.7 9.5 20.1 14.1 Urbanization atmainoffıce Urban................... 79.4 89.6 77.6 79.2 80.2 79.9 79.2 79.1 77.8 84.5 81.5 77.8 100 ... 76.1 82.0 Rural ................... 20.6 10.4 22.4 20.8 19.8 20.1 20.8 20.9 22.2 15.6 18.5 22.2 ... 100 23.9 18.0 Numberofoffıces One..................... 86.0 86.6 86.0 85.2 88.9 93.5 89.7 82.8 71.1 41.0 85.3 86.5 85.6 87.3 93.0 80.3 Two .................... 9.4 9.5 9.5 9.8 8.1 5.2 8.3 11.8 15.1 18.0 10.8 8.4 9.9 7.7 5.8 12.4 Threeormore........... 4.6 3.9 4.6 5.0 3.0 * 2.0 5.4 13.8 41.1 4.0 5.1 4.5 5.0 1.2 7.4 Salesarea Primarilywithinthe UnitedStates ....... 95.4 95.0 95.6 95.2 95.9 94.6 95.4 96.1 95.7 92.0 95.0 95.6 94.7 97.9 96.3 94.6 Internationalorglobal ... 4.6 5.0 4.4 4.8 4.1 5.4 4.6 4.0 4.3 8.0 5.0 4.4 5.3 2.1 3.7 5.4 Management Byowner ............... 94.3 92.8 94.6 94.6 93.0 99.1 96.2 91.5 85.9 76.9 94.7 93.9 94.0 95.2 97.4 91.7 Hired ................... 5.8 7.2 5.4 5.4 7.0 * 3.8 8.5 14.1 23.1 5.3 6.1 6.0 4.8 2.6 8.3 Race,ethnicity,and sexofmajorityowners NonwhiteorHispanic ... 13.1 100.0 1.0 12.3 15.7 12.5 15.0 11.4 11.3 11.5 17.3 9.9 14.7 6.6 14.0 12.3 Non-Hispanicwhite ..... 86.6 6.3 100.0 86.9 85.8 87.9 85.0 87.3 89.0 88.8 82.0 90.1 84.6 94.3 86.4 86.8 White................... 91.0 35.0 100.0 91.4 89.8 91.9 90.0 90.9 94.4 92.2 87.9 93.3 89.6 96.3 90.7 91.2 Black ................... 3.7 28.1 * 3.2 5.5 3.5 4.6 3.3 * * 4.8 2.8 4.3 * 5.1 2.5 Asian,NativeHawaiian, orotherPacific Islander ............ 4.2 32.4 * 4.2 4.3 4.0 4.4 4.0 4.3 5.4 5.8 3.1 4.9 1.6 3.3 5.0 AmericanIndianor AlaskaNative ...... 1.3 10.3 .7 1.1 2.2 * 1.6 * * * 1.7 1.1 1.2 1.9 1.7 1.1 Hispanic ................ 4.2 31.8 ... 4.1 4.3 4.0 4.8 3.3 5.6 3.5 5.2 3.4 4.7 2.2 4.5 3.9 Non-Hispanic ........... 95.8 68.2 100.0 95.9 95.7 96.1 95.3 96.7 94.5 96.5 94.8 96.6 95.3 97.8 95.5 96.1 Female.................. 22.4 26.8 22.2 ... 100 32.0 22.1 18.5 15.0 11.5 26.2 19.5 22.6 21.5 28.2 17.7 Male.................... 64.8 65.1 65.3 83.5 ... 62.2 62.6 66.3 75.4 80.5 60.7 67.9 65.3 62.9 62.4 66.8 Ownershipdivided equallybysex...... 12.8 8.1 12.6 16.5 ... 5.9 15.3 15.2 9.6 8.0 13.1 12.6 12.1 15.6 9.4 15.5 Note:Fordefinitionofsalesareas,refertotable2,note3. *Fewerthanfifteenobservations. ...Notapplicable. ThesamplewasdrawnfromtheDun&Bradstreet Dun & Bradstreet database covers approximately Market Identifier (DMI) file.35 The DMI file is 93percentoffull-timebusinessactivity.36 broadlyrepresentativeofallbusinessesintheUnited Entitiesknowntobeineligibleforthesurvey(such States (though it may underrepresent the newest and as firms with 500 or more employees, branches, smallest businesses). It has been estimated that the subsidiaries, and firms in certain industries) were removed from the DMI file, and then the DMI file 36.BruceD.PhillipsandBruceA.Kirchhoff,‘‘Formation,Growth, and Survival: Small Firm Dynamics in the U.S. Economy,’’ Small 35.Dun’sMarketingService,Dun&Bradstreet,Inc. BusinessEconomics,vol.1(March),pp.65–74.

A190 Federal Reserve Bulletinh 2006 was sampled according to a stratified systematic interviewingwasfinished,4,240eligiblefirms,repredesign. The design consisted of seventy-two strata senting 6.3 million businesses, had completed defined by the cross-classification of firm size by interviews—acompletionrateof52.4percentforthe number of employees (less than 20, 20–49, 50–99, main interview. The weighted response rate for and 100–499), Census division, and urban or rural screeningandmaininterviewscombinedwas32.4perstatus.EachstratumwassortedbyStandardIndustrial cent.40 Classification(SIC)codetohelpensureproportionate The actual main telephone interview took an averindustry coverage. Larger small businesses (those age of about fifty-five minutes, and the total time with 20 or more employees), which account for a (whichincludedestablishingcontact,settingappointsmall proportion of the target population, were over- ments,andsoon)averagedmorethanthreehoursper sampled to ensure a large enough sample to permit completed case. Typically, respondents that started comparisonswithsmallersmallbusinesses. the main interview and got through the first few A sample of 37,600 firms (representing nearly questionsendedupcompletingtheinterview. 9.7 million enterprises) was initially selected for a The following categories of information were colbrief,computer-assistedtelephoneinterview.Thepur- lectedfromeachfirm: poseofthisscreeninginterviewwastoverifythatthe (cid:129) thedemographicsofthefirmandoftheownerswith firm was eligible for inclusion in the sample, to thelargestshares—uptothreeownersperfirm confirm contact information, and to secure coopera- (cid:129) the firm’s use of financial services and the sources tionforthemaininterview.Inadvanceofthescreenprovidingtheservices ing interview, selected firms received a brochure (cid:129) themostrecentapplicationsforcreditbythefirmin describing the survey. The average screening interthepastthreeyears view itself took less than eleven minutes, and the (cid:129) thefirm’sbalancesheetandincomedata average total time per screening interview (which (cid:129) therecentcredithistoryofthefirmanditsowners included establishing contact with firm owners, setting appointments, and calling back to complete the With the exception of the income statement, balance screening interview) was about one hour. Of these sheet, and most recent credit applications, the data 37,600 firms, 23,798 were selected for the screening werecollectedasofthedateoftheinterview.41 interview, 14,061 were actually screened, and 9,687 A public-use version of the data set and a user’s weredeterminedtobeeligible.Theweightedresponse manualwillbepostedontheFederalReserveBoard’s rateonthescreeninginterviewwas62percent.37 website, at www.federalreserve.gov/pubs/oss/oss3/ In the second stage, the main survey, also a nsbftoc.htm. computer-assistedtelephoneinterview,wasattempted with all 9,687 firms determined to be eligible during the screening stage.38 Within a couple of weeks of APPENDIX B: RACE, ETHNICITY, AND completing the screening interview, firms received a BUSINESS OWNERSHIP customized worksheet to help the owners collect and In contrast to earlier years, the most recent SSBF organize their records in preparation for the main survey and the most recent data released by the U.S. interview.Theworksheetrequestedfinancialdatafor Census Bureau differ in the estimated share of all thefirm,informationaboutthefinancialservicesused by the firm, and the sources of those services. The worksheets were customized according to the firm’s 40.Theresponserateforthe1998SSBFwas33percent.Oneofthe legalorganizationalformanddirectedrespondentsto goalsofthe2003surveywastoimproveresponserates.Tothisend, the appropriate lines on their tax forms.39 When the2003surveyofferedincentivesforcompletingthemaininterview, shortened the time between the screening interview and the main interview to an average of two weeks or less, sent worksheets to potential respondents using next-day delivery, added an automated 37. Details on response rates are in NORC (2005), ‘‘The 2003 procedure for looking up institutions during the main interview to SurveyofSmallBusinessFinancesMethodologyReport.’’ reduce the amount of time needed to identify branch locations and 38.Duringthemaininterview,afewadditionalfirmsweredeter- obtainbranchaddresses,andcarefullyreviewedandstreamlinedthe minedtobeineligible.Inmostofthesecases,screeninginterviewshad questionnaire.Theseeffortsarebelievedtohavepreventedthe2003 beencompletedbysomeoneotherthananowner. responseratefromdecliningfromits1998level. 39.At the end of the main interview, respondents were asked to 41.Toeasereportingburdenandtoensuresomeconsistencyacross returntheirworksheetsorotherrecords(taxfiles,financialreports)in firms,balancesheetandincomeitemsweretiedtothetaxformsthat aself-addressedstampedenvelope.Aboutone-thirdoftheparticipat- mostfirmsarerequiredtofileeachfiscalyear.Becausenotallfiscal ingfirmsreturnedcompletedworksheetsorotherrecords.Theprepa- yearsendonDecember31,andbecauseittakestimeforbusinessesto rationoftheworksheetshelpedrespondentspreparefortheinterview, organize records and prepare tax forms, firms were asked to report and the returned worksheets often helped resolve discrepancies or balance sheet and income data for the firm’s fiscal year that ended supplyitemsmissingafterthemaininterview. betweenMay1,2003,andApril30,2004.

Financial Services Used by Small Businesses: Evidence from the 2003 Survey of Small Business Finances A191 businessesthatwereownedbyAsians,byblacks,and B.1. Ratesofbusinessownership,byselectedraceand by Hispanics.42 In the Census Bureau’s Survey of ethnicityofowners,1997and2002SBOand1998 and2003SSBF BusinessOwners(SBO),theownershipratesofthese Percent minority groups rose between the 1997 and 2002 surveys. However, the SSBF estimates indicate that, SBO SSBF Raceorethnicity between 1998 and 2003, rates of ownership among 1997 2002 1998 2003 Asians and blacks were essentially unchanged, Asian1 .......... 4.4 5.0 4.4 4.2 whereas rates of Hispanic ownership declined (table (.3) (.4) B.1).43 Black ........... 4.0 5.2 4.1 3.7 (.3) (.4) Thesedifferencescanlargelybeexplainedbythree Hispanic ........ 5.8 6.9 5.6 4.2 factors: (.4) (.4) (cid:129) First, the lists of businesses from which the two Note:Numbersinparenthesesarestandarderrors. 1.Fordefinitionofthistermhereandinthefollowingtables,refertotext surveys draw their samples (the sample frames) note42. differ in ways that are sensitive to any dispropor- Source:Hereandinthefollowingtables,datafortheSurveyofBusiness Owners(SBO)arefromtotalnumbersofbusinesses,1997and2002surveys, tionate change in minority ownership rates. Such www.census.gov/csd/sbo. disproportionatechangeswereseeninthe2002–03 tion of survey structure with changes specific to the periodbutnotinthe1997–98period. periods in question. Such complex interactions high- (cid:129) Second, to improve the uniformity of its financial lighttheneedtotreatcomparisonsofresultsfromthe data,the2003SSBFlengthenedthetimethatafirm SSBFandSBO—indeed,betweenanytwosurveys— would have to be in business to be eligible for the withcare. survey.Thischangewouldtendtocauseadecrease in the observed rate of minority ownership in the Differences in Universes SSBF relative to the SBO because minority-owned firms tend to have shorter life spans than non- Perhaps the most important reason that the estimates minority-ownedfirms. of minority ownership rates from the most recent (cid:129) Third, unlike the 1998 SSBF and the two SBO SSBFandSBOmayhavedivergedisthedifferencein surveys, the 2003 SSBF did not oversample the promptness with which the surveys’ sample minority-owned businesses. Although that is a frames pick up new and small businesses, together notable change in methodology, the specific effects withsomeparticulardynamicsofbusinessownership ofthechangeareunknown. intheperiodsbetween1997and2003. For both the 1997 and 2002 surveys, the SBO Takentogether,thesethreefactorssuggestthatthe samplewasdrawnfromeconomiccensusreportsand difference in minority ownership rates between the fromalistoffirmscompiledfromfederaltaxreturns 2003SSBFandthe2002SBOstemfromtheinteracfiled by businesses.44 Because the list used by the Census Bureau is drawn in part from restricted 42. Rates of ownership byAmerican Indians andAlaska Natives sources and is not available to the Federal Reserve werecloseinthetwosurveysandarethusnottreatedinthisappendix. Board,theSSBFsamplemustbedrawnfromanother In the 1997 SBO, the Native Hawaiian category did not appear source. The Dun & Bradstreet Market Identifier File separately; in that survey, the category was ‘‘Asian and Pacific Islander.’’Asintherestofthisarticle,‘‘Asian’’usedinreferencetothe (DMI), thought to be the best publicly available SSBFreferstoindividualswhoareAsian,NativeHawaiian,orPacific listing of current businesses in the United States, is Islander. the source for the 1998 and 2003 SSBF samples.45 43. The 1998 SSBF covered 3,561 firms, of which 214 were Asian-owned,273wereblack-owned,and260wereHispanic-owned. The DMI file is updated using information from new Thecorrespondingownershipnumbersforthe4,240firmsinthe2003 SSBFwere170,119,and149respectively.Reportedpercentagesare weighted to adjust for sample design and nonresponse. Because the 44.TheCensusBureauobtainselectronicversionsofthefollowing samplesineachsurveyweredrawnfromdifferentpopulationsandthe formssubmittedbybusinessestotheInternalRevenueService:Form CensusBureaudoesnotprovidestandarderrors,thepointestimates 1040,ScheduleC(individualproprietorshiporself-employedperson); arenotdirectlycomparableacrosssurveys.Forthisappendix,estimate Form 1065 (partnership); all Form 1120 corporation tax forms; and BisconsideredstatisticallydifferentfromestimateAifitdoesnotlie Form941(Employer’sQuarterlyFederalTaxReturn). within the 95 percent confidence interval of A. For example, in the 45.The Small BusinessAdministration (The State of Small Busi- 2003SSBF,thepercentageoffirmsthatwereownedbyblackswas ness:AReportofthePresident,U.S.SmallBusinessAdministration, estimatedtobe3.7percent.The95percentconfidenceintervalforthat 1988)estimatesthattheDMIframerepresentedapproximately93perestimateisfrom2.9percentto4.5percent—calculatedas3.7±(1.96x centofprivateemploymentin1987.Noupdatedestimateoftheextent 0.4)(seetableB.1).Becausethe2002SBOestimateof5.2percent of DMI coverage is available, but its coverage is likely to have fallsoutsidethisrange,thedifferencebetweenthesetwoestimatesis improved(oratleastheldsteady)withimprovedinformationtechnolstatisticallysignificant. ogy.Forexample,intheearly1990s,theDMIframebeganincluding

A192 Federal Reserve Bulletinh 2006 B.2. Ratesofbusinessownership,byselectedraceand B.3. Ratesofownershipofbusinesseswithpaid ethnicityofownersandpresenceofpaidemployees, employees,byselectedraceandethnicityofowners 1997and2002SBO andnumberofpaidemployees,1997and2002SBO Percent Percent Nopaidemployees Paidemployees Asian Black Hispanic Numberof Raceorethnicity paidemployees 1997 2002 1997 2002 1997 2002 1997 2002 1997 2002 Asian ........... 4.0 4.6 5.5 5.9 1–4............. 5.7 6.2 1.9 1.8 4.3 3.8 Black ........... 4.7 6.3 1.8 1.7 5–9............. 5.6 5.8 1.5 1.5 3.9 3.3 10–19........... 4.5 5.2 1.3 1.3 3.5 3.2 Hispanic ........ 6.4 7.9 4.0 3.6 20–49........... 4.3 4.1 1.2 1.2 2.6 2.5 50–99........... 3.5 2.9 1.3 1.1 2.7 2.4 100–499 ........ 2.5 2.2 1.0 1.1 1.3 1.6 credit applications; in-person and telephone inter- Memo views;county,state,andfederalgovernmentsources; Allfirms........ 5.5 5.9 1.8 1.7 4.0 3.6 business trade-tape exchange programs; and thirdparty sources such as listings in the Yellow Pages. Asian ownership rates were higher for firms with Giventhatthenewestandsmallestfirmsoftendonot paid employees than for those without paid employapplyfornewcredit,advertiseintheYellowPages,or ees, and Asian ownership rates grew somewhat for incorporate,itislikelythattheDMIframeunderrep- both these types of firms. However,Asian ownership resentsthesmallestandnewestfirms.46 rates were highest for the smallest firms; indeed, the Iftheratesofminorityownershiparehigherforthe growthbetweenthe1997and2002SBOwaslimited firmsmissingfromtheDMIfilethanforthoselisted, tofirmswithfewerthantwentypaidemployees(table we would expect to see lower rates of minority B.3).47 ownership in the SSBF than in the SBO. The SBO doesnotprovideinformationontheownershipofthe Composition of Self-Employment from the CPS firm by firm age, so we are unable to say anything Data from the March Current Population Survey about the part of the DMI file that is likely to be (CPS) provide further details on the changes in missing because the firm is too new. However, the business ownership composition in the periods be- SBO does provide information according to whether tweenthe1997and2002SBOandthe1998and2003 thefirmhaspaidemployeesand,ifitdoes,according SSBF (table B.4). Although the CPS data are not tothenumberofemployees.Non-employerfirmsand intended to measure business ownership, they do firms with few paid employees are less likely to be provide information on self-employment, with a represented on the DMI file than the employer firms, breakdown by business incorporation status.48 We especially employer firms with a large number of would expect new unincorporated businesses to be employees. picked up less quickly in the DMI file than incorporatedbusinessesforthereasonsnotedabove.Changes Business Ownership Rates from the SBO in the make-up of self-employed individuals in unin- Among the firms represented by the 2002 SBO corporated businesses are thus likely to be observed sample,lessthanone-fourthhadanypaidemployees, only with a lag in the DMI file. However, if the a fraction consistent with the 1997 SBO. The SBO change is proportionate by race and ethnicity, the lag estimatesindicatethatownershipratesforblacksand should not affect estimates of minority ownership Hispanics were much lower among firms with paid rates. employees(tableB.2).Theyalsoindicatethatownership rates among blacks and Hispanics grew only 47.Minorityownershipratesbythenumberofemployeesarenot among firms with no paid employees. Among firms availableforfirmswithoutpaidemployees,asthenumbersofunpaid with paid employees, the rate of ownership among employeesarenotcollected. 48.Estimatesofbusinessownershipratesbasedonself-employment blacksremainedroughlyconstant,andtherateamong will not be exactly comparable to actual business ownership rates Hispanicsshowedaslightdecline. becausetheunitofobservationistheindividualratherthanthefirm. Countingthiswaywillmissallindividualswhodonotactivelywork intheirfirmsandovercountfirmsthatarejointlyowned;noinformaallfirmsinthebusinesslistingsoftelephonedirectories,suchasthe tion is available on the effect this is likely to have on estimates of YellowPages. minority ownership. Estimates are based on the ‘‘Class ofWorker’’ 46. An estimate of the coverage of the DMI file is in David A. variablesforthemainandsecondjob;respondentswerecountedas MarkerandW.ShermanEdwards(1997),‘‘QualityoftheDMIFileas self-employediftheyreportedself-employmentforeitherjob.From aBusinessSamplingFrame,’’inproceedingsoftheSurveyResearch 1994 to 2003, between 10 percent and 12 percent of the employed Methods Section, American Statistical Association, pp. 21–30, populationreportedbeingself-employedineithertheirmainorsecond www.amstat.org/sections/SRMS/proceedings. job.

Financial Services Used by Small Businesses: Evidence from the 2003 Survey of Small Business Finances A193 B.4. Shareofself-employedpopulationthatisofselectedraceandethnicity,bystatusofbusinessincorporation,Current PopulationSurvey,1994–2003 Percent Allbusinesses Incorporatedbusinesses Unincorporatedbusinesses Year Asian Black Hispanic Asian Black Hispanic Asian Black Hispanic 1994.... 2.8 4.4 5.3 3.6 2.9 4.0 2.4 4.9 5.8 (.2) (.3) (.3) (.4) (.4) (.5) (.2) (.3) (.4) 1995.... 2.5 4.1 4.7 3.4 2.6 4.0 2.1 4.7 5.0 (.2) (.3) (.3) (.4) (.4) (.5) (.2) (.3) (.4) 1996.... 3.3 4.4 5.5 4.5 2.6 3.2 2.8 5.0 6.3 (.2) (.3) (.3) (.6) (.5) (.5) (.3) (.4) (.4) 1997.... 4.0 4.3 5.6 4.6 2.5 3.1 3.8 5.0 6.6 (.3) (.3) (.3) (.5) (.4) (.5) (.3) (.4) (.4) 1998.... 3.9 4.4 5.5 4.3 2.3 3.3 3.7 5.2 6.4 (.3) (.3) (.3) (.5) (.4) (.5) (.3) (.4) (.4) 1999.... 4.3 5.1 6.4 4.9 3.9 3.4 4.1 5.6 7.5 (.3) (.3) (.4) (.5) (.5) (.5) (.3) (.4) (.5) 2000.... 3.8 5.8 5.9 4.4 4.3 4.7 3.5 6.5 6.4 (.3) (.4) (.3) (.5) (.6) (.5) (.3) (.4) (.4) 2001.... 4.4 5.2 6.3 4.6 4.0 4.8 4.3 5.7 6.8 (.3) (.3) (.3) (.5) (.5) (.5) (.4) (.4) (.4) 2002.... 3.5 5.6 6.7 4.0 3.4 3.5 3.3 6.6 8.2 (.3) (.3) (.4) (.5) (.4) (.4) (.3) (.4) (.5) 2003.... 4.5 5.4 7.7 5.4 4.2 4.0 4.1 6.0 9.4 (.3) (.3) (.4) (.5) (.5) (.5) (.4) (.4) (.5) Note:Numbersinparenthesesarestandarderrors. Source:CalculatedfromCurrentPopulationSurveyforMarchofeachyear,www.bls.census.gov/cps/cpsbasic.htm. The data do indicate disproportionate changes in minority ownership relative to surveys, such as the thecompositionoftheself-employedinthe2002–03 SBO, that pick up changes in the population of period. Between 1997 and 1998, the proportions of unincorporated businesses more promptly. Although the self-employed population that wereAsian, black, the 2003 SSBF rates are somewhat lower than the and Hispanic were essentially unchanged. In particu- overallself-employmentproportionsinthe2003CPS, lar,theproportionsofself-employedpersonsinunin- they are not statistically different from the 2003 CPS corporated businesses who were Asian, black, or proportionsofpersonsself-employedinincorporated Hispanic remained constant. This constancy would businesses. imply that, despite the lag with which the smallest Thisanalysisprovidesevidencethatminorityownnew businesses are likely to appear in the DMI file, ershipratesarehighestamongthesmallestfirms,the the estimated rates of minority ownership from a firmsmostlikelytobemissingfromtheDMIfile.In sample drawn from the DMI file during that period addition, the analysis shows differential trends for should be very close to the actual rates. Indeed, the minorityself-employment.Takentogether,thesefindratesofminorityownershipinthe1998SSBFandthe ingsprovidesomeinsightintothedissimilarityofthe 1997SBOwerebothveryclosetothe1998estimated comparisons between the 1997 SBO and the 1998 overall self-employment proportions calculated from SSBF on the one hand, and the 2002 SBO and the theCPS. 2003SSBFontheother. Between 2002 and 2003, however, the proportion of self-employed individuals who were Asian rose about 1 percentage point, the proportion that was Differences in Coverage Hispanic also rose about 1 percentage point, and the share that was black changed little. Moreover, the Both the SSBF and the SBO limit their samples to proportions of self-employed individuals in unincor- firmsinbusinessasofcertaindefinedperiods,butthe porated businesses who wereAsian or Hispanic also SSBF changed its population definition between the rosebetweentheseyears.Thus,unlikeinthe1997–98 1998 and 2003 surveys.The SBO, focused on owner period, the 2002–03 period had disproportionate characteristics, restricts its coverage to businesses in changesinthesharesoftheself-employedwhowere operation at any point during the calendar year; the AsianorHispanic.Ashiftofthissortbetweenthetwo 1997 survey covered businesses in operation during periods should cause a sample drawn from the DMI 1997, and the 2002 survey covered businesses in file for 2003 (the SSBF) to understate these rates of operationduring2002.

A194 Federal Reserve Bulletinh 2006 The SSBF, on the other hand, attempts to collect a B.5. Numberofstarts,closures,andbankruptciesof largeamountoffinancialinformationaboutfirmsthat businesseswithpaidemployees,2000–04 were in business as of December 31 of the survey Percent year. In 2003, an additional in-business constraint Category 2000 2001 2002 2003 2004 was imposed on firms so that account information Starts........ 574,300 585,140 569,750 612,296 642,600e could be captured as of the date of the interview and Closures..... 542,831 553,291 586,890 540,658 544,300e balancesheetinformationcouldbecollectedasofthe Bankruptcies . 35,472 40,099 38,540 35,037 34,317 end of the fiscal year (December 31, 2003, for the eEstimated. majority of firms).49 Thus, whereas the 1998 SSBF Source:SBAOfficeofAdvocacy,http://app1.sba.gov/faqs/ collected information on firms that were in business faqIndexAll.cfm?areaid=24. as of December 31, 1998, the 2003 SSBF collected applied only to the 2003 SSBF, we should expect to information on firms that were in business as of observe differences in comparisons of estimates of December 31, 2003, and at the date of the interview rates of minority ownership between the 1997 SBO (between June 2004 and January 2005). The 2003 and the 1998 SSBF and between the 2002 SBO and procedurethuseffectivelyimposedalongerlongevity the2003SSBF. constraintonthepopulationofinterest. Because there is a significant amount of ‘‘churn- Differences in Methodology ing’’inthebusinesspopulationeachyear(tableB.5), the added longevity constraint implies that fewer Another area that might have contributed to the firmswouldbeeligibleunderthe2003SSBFeligibil- observed differences in minority ownership rates ity rules than under the 1998 SSBF rules. Research betweenthelatestSSBFandSBOcanbefoundinthe indicates that minority-owned businesses are less different ways the two surveys drew their samples. likelytoremaininbusinessthannon-minority-owned For the 1997 and 2002 SBO, predicted race and businesses.50 Among the non-minority-owned busi- ethnicitycategorieswereusedassamplingstrata,and nesses established in 1997, 72.6 percent were still in minorities were oversampled. The 1998 SSBF also businessasof2001.AmongAsian-ownedbusinesses, oversampled minorities using information collected thecorrespondingproportionwas72.1percent;among during the screening interview. The 2003 SSBF did Hispanic-ownedbusinesses,68.6percent;andamong not, however, oversample minorities. The effects of black-owned businesses, 61.0 percent. Some evi- this divergence between the surveys are unclear, but dence also indicates that the longer a firm is in themethodologicaldifferenceitselfisworthnoting. business, and the larger a firm is at a given time, the morelikelyitistosurviveoversomefiniteperiod.51 Summary Given (1) the significant churning of businesses, (2)thefactthatmanyminority-ownedbusinessesare In contrast to earlier years, the 2003 SSBF reported quite small, and (3) the higher business closure rates lower rates of business ownership by Asians, by byminorities,weshouldexpecttoseefewerminority blacks,andbyHispanicsthandidthe2002SBO.This businesses in the SSBF sample than in the SBO appendix looked at differences in the universes, covsampleevenifthesampleweredrawnfromthesame erage,andmethodologyacrossthetwosurveys.Each list. Furthermore, given that the longevity constraint of these three factors, together with the changing dynamics of the small business population, could be expected to contribute to the differential ownership 49. The additional in-operation restriction was imposed so that rates observed between the most recent surveys by respondents could provide information on accounts and financial theSSBFandSBO. service providers as of the date of the interview.Although previous interviewsaskedrespondentstoprovidethisinformationasoftheend ofthelastfiscalyear,interviewerreportsindicatedthatmanyrespon- (cid:129) TheSSBFdrawsitssamplefromalistthatislikely dentsgaveaccountinformationasoftheinterviewdate.Tomakethe to be slower than the SBO list in picking up the reportedinformationconsistentacrossallrespondents,thequestionnewestandsmallestfirms.Between1997and1998, nairewaschangedtoaskaboutaccountsasofthedateoftheinterview. 50.YingLowrey(2005),‘‘DynamicsofMinority-OwnedEmployer the fluctuations in the self-employment population Establishments,1997–2001:AnAnalysisofEmployerDatafromthe were relatively stable within racial and ethnic cat- SurveyofMinority-OwnedBusinessEstablishments,’’ResearchStudy, egories. Between 2002 and 2003, however, there U.S.SmallBusinessAdministration,OfficeofAdvocacy(February), www.sba.gov/advo/research/chron.html. were sizable increases in the proportion of self- 51.JoelPopkinandCompany(1992),‘‘BusinessSurvivalRatesby employedindividualswhowereAsianorHispanic, Age Cohort of Business,’’ Research Study, U.S. Small Business andtheSSBFlikelyunderestimatedthatrecentrise Administration,OfficeofAdvocacy;summaryatwww.sba.gov/advo/ research/chron.html. becauseitssourcelistwasslowtorecordit.

Financial Services Used by Small Businesses: Evidence from the 2003 Survey of Small Business Finances A195 (cid:129) The2003SSBFextendedthetimethatafirmhadto Although counterfactuals—such as the percentage of beinbusinesssoastocapturecertainfinancialdata firmsthatwentoutofbusinessthatareincludedinthe moreaccurately.Thatchangewouldtend,however, SBO sample or the newly formed firms that are not tocausefewerminority-ownedbusinessestoqualify included in the DMI frame—are not available to for the survey relative to the SBO and to the 1998 definitively identify the cause of the differences, and SSBFbecausesmallandminority-ownedfirmstend although all of the factors are likely to have played a toclosemorequicklythanothers. role,thedifferencesinthelistfromwhichthesamples (cid:129) Finally, of the four surveys at issue (the two most weredrawnappeartohavebeenthedominantfactor. recentreleasesoftheSBOandthetwomostrecent oftheSSBF),the2003SSBFwastheonlyonethat did not oversample minority-owned businesses, a differencewithspecificeffectsthatarenotknown.