The Statistical Discrepancy in the U. S. International Transactions Accounts: Sources and Suggested Remedies

Board of Governors of the Federal Reserve System International Finance Discussion Papers

Number 404

July 1991

THE STATISTICAL DISCREPANCY IN THE U.S. INTERNATIONAL TRANSACTIONS ACCOUNTS: SOURCES AND SUGGESTED REMEDIES

Lois E. Stekler

NOTE: International Finance Discussion Papers are preliminary materials circulated to stimulate discussion and critical comment. References in publications to International Finance Discussion Papers (other than an acknowledgment that the writer has had access to unpublished material) should be cleared with the author or authors.

ABSTRACT

The statistical discrepancy in the U.S. international transactions accounts has tended to be both large and positive over the last decade and a half. In 1990 the statistical discrepancy rose by $45 billion to a record $64 billion and brought the cumulative discrepancy since 1960 to. almost $250 billion. The size and persistence of this discrepancy has called into question the accuracy of the data on the U.S. current and capital accounts.

This paper attempts to find clues to the sources of the statistical discrepancy by 1) reviewing past history, 2) examining the data sources for each major component of the U.S. international transactions accounts, and 3) using regression analysis. The paper concludes with a list of recommendations for data improvements.

While inadequacies are evident in the data for a wide variety of international transactions, both current and capital account, the search for sources of the big increase in the discrepancy between 1989 and 1990 probably can be narrowed largely to the capital account. It seems unlikely that net exports of goods, services, or investment income increased by an additional $45 billion in 1990. On the capital account side, increases in foreign holdings of U.S. currency probably played a significant role, but the bulk of the increase in the statistical

discrepancy in 1990 remains a mystery.

THE STATISTICAL DISCREPANCY IN THE U.S. INTERNATIONAL TRANSACTIONS ACCOUNTS : SOURGES AND SUGGESTED REMEDIES

Lois Steklert I. INTRODUCTION

The statistical discrepancy in the U.S. international transactions accounts has tended to be both large and positive over the last decade and a half. In 1990 the statistical discrepancy reached a record $64 billion and the cumulative discrepancy since the beginning of 1960 summed to almost $250 billion. The size and persistence of this discrepancy has called into question the accuracy of the data on the U.S. current and capital accounts.

In principle, the sum of all transactions in the U.S. balance of payments accounts, a double-entry bookkeeping system, should equal zero; for each transaction there should be two equal entries of opposite sign. In practice, the recorded accounts never sum exactly to zero because the data that would reflect the debit and credit counterparts of each single transaction generally are obtained from different sources. The statistical discrepancy is the net of errors and omissions in all the components of the international transactions accounts. A positive

statistical discrepancy represents some combination of net unrecorded

1. The author is a staff economist in the Division of International Finance. This paper reflects the views of the author and should not be interpreted as reflecting the views of the Board of Governors of the Federal Reserve System or other members of its staff. The author would like to acknowledge many helpful conversations with staff members at BEA, Treasury, the Federal Reserve Bank of New York, and the IMF Working Party on the Measurement of International Capital Flows. The author would also like to thank Allen Frankel, Bill Helkie, Sarah Holden, Peter Hooper, Russ Krueger, Walther Lederer, Kathy Morisse, and Larry Promisel for helpful comments on an earlier draft and Vladimir Gutin for his assistance.

“eqep smoTF Teatdeo “S'n 84a Fo AdeANODe aYyQ ANoge suoT3Zsenb snotazes sester (ueder pue ‘SIq ‘aaresey Te19peq) Seotnos JeyIO WorAF Beep YITM sSquNODDe suOTIORSURIR TeuoTIeUASAUT “S$'Mn ey WOAJ eIep Fo uostTzeduog ‘uoTRIeTTOO eIep 02 a20AeP 02 SBdANOSeA TeUOTITpPpe 2e3 09 A[NOTJFTIp ASA useq sey AT usyM aut. @ Qe ‘Q[NOTFJIp orzow AezZ smoTZ Tegtdeo Teuotjaeurzequt jo ButTyoes3 oy. spew sary squeudoTeAep esey, ‘AND00 suOoTQOeSURPAQ YOTYM Ysnory3 STeuueyd ay} pue sjeyreU UsTeT0Z uT squedtotjaed Jo azequmu ay yI0q SuTATdtQ {nu ‘sjoeyAeM TeTOUeUTF JO uoTAeZTTeEGoTS 9yQ 03 pejnqtazjUOD eAeYy asueyo [POTZoTOUYIeR pur UOCTIRAOUUT ‘apts AuUNODOe [eQTdeo 9YyQ UO *prorge sZejysuerzz Quenbesqns ATeyQ Ao squeASTWUIT Meu JO SQesse oy} UO UOTReUWAOJUT eTAIITT sey se7ejs peqatug eyQ ‘uOoTITppe UT ‘pe_eqszepun AT TeTIUeQSQns eq ASU sWOOUT QUeUQseAUT OTTOZQA0d “S'g 2eYUQA AseBBns (SIgq) squeUe T3325 TeuoTIeUAeAU] AOZ yUeG ey 02 seTAQUNOD AeyQO Aq peqrzoder eIep YAIM e3ep ‘§'n Jo suostaedwoo f‘uanjexr Jo seaqei pue suotjtsod jueuqseautT Anoqe suotqjdumsse Ayeys uo peseq soqeUTqSe ete sWOoUT QUeUqZseAUT oTTOFQI10d uo B2ep ‘S'Qn ‘paepeseu st sserZ0id AaeyAANZ 3anq ‘szeek Queder UT peAorzdut Aqje013 useeq eaey ‘(AAQUNOD 9YyQ sAPeT AO ABQUSe AT[POTSAYd ABdYQ se perz0jtuow eq Jouuros Aey. esneoeq) spooS uo ejep urYyA AdeT[OS 02 ATNOTJITpP etow AT[PAVUEeZ ee YOTYM ‘suOTQOBSUeAQ SedTAAZEeS uo BeRIeq ‘sqiodxe ueyQ AT[NjJereo eAoW pezojtTuow ATTereUes ere sqzoduT uo ejep swojsno ‘epts quNnODDe qUueTAND ey. UQ ‘SUOTROBSURAA QUNODDE [RQITdeo pure QUezTINSO YyIOqG jo Zutpz00er ey UT SUOTSSTWO pue szozz9 ‘ATpejqnopun ‘ere ere], ‘peoiqe worzy smoTjJuT [eatdeo pejzazodeizun

qeu pure sWOOUT QUeUTQSseAUT pue ‘seoTAZes ‘spoo3 jo szeusTez10F 03 sqazodxe

Inadequacies in the data on U.S. international transactions have produced both increased volatility in the statistical discrepancy in recent years and a tendency towards large positive values. The sources for these trends need not be the same. While timing problems and errors and omissions in the reporting of capital flows are more likely sources of the increases in short-term volatility, errors and omissions in the recording of current account transactions also have probably contributed to the upward trend in the discrepancy.

Part II of this paper reviews the past history of the statistical discrepancy for clues about its sources. Part III examines the data sources for each major component of the U.S. international accounts, points to problem areas, and suggests certain improvements. Part IV uses statistical tests to try to shed light on the sources of the statistical discrepancy and Part V summarizes the conclusions and suggestions for improving the data.

II. CLUES FROM PAST HISTORY

The growth of the statistical discrepancy in the U.S. international transactions accounts is a relatively recent development. In both the 1950's and the 1960's the statistical discrepancy was close to zero. (See chart 1.) In contrast, during the early 1970's there were substantial net unrecorded outflows or payments. Since 1974 the statistical discrepancy has tended to be both large and positive, indicating net unrecorded receipts or inflows. This increase in magnitude is not just the result of the inflation of nominal values. The ratio of the statistical discrepancy to the value of trade (the average

of recorded exports and imports of goods, services, and investment

($ billions)

70.0

60.0

50.0

40.0

30.0

20.0

10.0

1955

1960

Statistical Discrepancy

1965 1970 1975

1980

1985

1990

- e¢ =

-4-

income) rose from a mean absolute value of .02 in the 1950's and 1960's to .05 in the 1970's and 1980's.”

The statistical discrepancy has increased not only in size, but also in volatility. Increases in the statistical discrepancy in one quarter, tend to be followed by decreases in the next quarter,” The average absolute size of the quarterly changes in the statistical discrepancy, scaled by the value of trade, has risen from .04 in the 1960's, to .06 in the 1970's, and .09 in the 1980's.

In the early 1980's, it was assumed that the sudden increase in the positive discrepancy was largely accounted for by unrecorded capital flows. The wide quarterly swings in the size of the statistical discrepancy also supported that conclusion. It was recognized that errors and omissions occurred in the reporting of trade transactions, but there seemed little reason to assume that these errors would suddenly increase or that they would vary widely from quarter to quarter.

Previous periods of relatively large positive statistical discrepancies had coincided with unsettled political and economic conditions abroad. The ratio of the value of the statistical discrepancy to trade was about as high as or higher than the 1979-1980 levels (.08) in certain Depression years (1934, 1935, and 1937), the early years of World War II (1939-41), and 1948. It seemed reasonable to assume that these earlier episodes were associated with the flight of capital to a safe haven in the United States in forms that were not fully reported,

either because these investors wanted to remain anonymous or because the

2. The peak values for this ratio in the postwar period were .14 in 1971, .10 in 1982, and .09 in 1990.

3. A regression between changes in the statistical discrepancy in one quarter and changes in the previous quarter (1960Q2-1990Q4) produces an R2.o0f .28 and a statistically significant coefficient of -.53.

“L861 tequeqdes ‘oq uoq8utysem ‘AWT ‘Sooueleg JuNocoy jUeaany plTaom ut ouedetosTd TRoTastieis ey uo Ajieg BUTYIOM oy jJO jxodoey Teula ‘AWT ey)

pe qsezejuT eyL “ejep SutTATAepun ey AoJZ seoanos azeyjo Auew uo seT[er

qanq ‘sjunoooe suoTjoRsUeTA [TeUOCTIeUEQUT ‘$*p aya seYsTTqnd yaq

SWHLSAS ONILYOdaa JO MAIATA “III “Pore SsTy UT Jstsaed [[T3s sweTqorid qnq ‘eTqeTTeae arom elep pesTAer ueym sieek 103 AouederosTp [eot4stjeqs eaTATsod eyQ Butonpez ‘pequeweTdut ueeq eaey AueW ‘squoueaoirdut pepesou AJTAUepT 02 sAA0FTJe Tetqueqjsqns 02 pet seotArzes Jo sqiodut pue sqiodxe jo Butjaode1z ayy uy seToenbepeut pue Awouoose *g*p ay UT 1020eS soTAIZeS ey. Fo soueqrodut ey? JO UoTITUZODeA BuTMOAS ‘uOoTIATppe UT * pordtesex SUOOUT QUdUIQSeAUT OTTOFIA0d AToeyA But ewTAserepun AT TeoTWeWeASKks e1eM SeTIQUNOD KUeU ey? peqse#8ns ATZuo0zAs eqep qunooose jueszino [eqoTZ ut SoeToueqSTsuooUuT Jo Apnqs gWI ue ‘teaocezom ‘“aYySITI Teatdeo Joj ,Sueaey aejyes, [etTjuejod ZayIO FO UOTANTOAS ey eATdsep pue sUOCTITpUOD AoyAeU TeTOUeUTJ Jo AQeTIeA eB Jepun eAT{Tsod peuteuer sey Aourdet0sTp [eoTAsTIeIS ey ‘QUNODOe quezTIND ey FO squsucduiod uo osTe uoTIUeIIe pesnooz sey SOZST PUI

guyanp setouedetostp [eoTasTqeqs eu eaTyTsod jo eoueqstsied ou “O861 PUP 6/61 UT SMoTzUT TeQtdeo peprzooeizun ey. 02 siojVOeF But AnqTAzquoo [eTQuejod pereptsuoos [Te erem ‘sqjesse ueTUPAT JO eZe07JF *S'n EY pur spueY DFdO UT YATeeM Fo uoTAeTNUMDOF ptdez yATM peuTquiod ‘STSTI9 [TO puodes ey pure g/6T UT UPI UT UOTIANTOAEZ eyL ‘reek 2eYyI UT SMOTJINO TeATdeo pepAoderUN 0} peT pey squTeIAseA qaodxe [Teqtdeo uteqieo sn{d uotTjeloerdep Aze[[op Jo suotjeqoedxe 2ey eumsse 0} eTqTSne Td pouwless QI “T/6T sem ‘aaTQeZou ysnoyaTe ‘ere T ATeaTzeTar sem AouederosTp [eoTastqzeqs eyQ USeYyM J] eM p[z0M eoUTS APek

Zeyjo ATuo ey, ‘“sroqjseauT [Teus Auew yo ed 02 peTtey wejsks Butqiodez

reader is referred to The Balance of Payments of the United States: Concepts, Data Sources, and Estimating Procedures, published by the

Department of Commerce, Bureau of Economic Analysis, May 1990, for a detailed description of the accounts. This paper provides only an impressionistic survey, focusing on problem areas and particularly on capital flows and investment income, which have not been the focus of other recent studies of U.S. data adequacy .> It should be noted that

the statistical discrepancy is the net of all errors and omissions in the accounts; correction of some data inadequacies would add to the net statistical discrepancy rather than reduce it.

Before proceeding to discuss each of the components of the U.S. international transactions accounts, it would be useful to touch on certain general problems. One pervasive problem in the last decade has been the scarcity of resources devoted to the collection and improvement of data. Generally, staffs at Treasury, Commerce, and the Federal Reserve Bank of New York (FRBNY) are well aware of inadequacies in the international transactions data they collect or estimate, but resources have not been available to make substantial improvements.

Preoccupation with reducing the Federal budget deficit and getting the government "off the backs" of the private sector has outweighed concerns about data adequacy and the potential for faulty policies based upon inadequate data. In some areas, automation has reduced the costs of data

collection; data have been improved despite constant or even reduced

5. The Panel on Foreign Trade Statistics of the National Academy of Sciences expects to release its report assessing the data on trade in goods and services during 1991.

6. In some cases planning is currently underway to correct some of the inadequacies discussed in this paper and requests for additional funding are included in current budget proposals.

‘eoremMIOD pue AAinseer], UseMjeq peqeoOTTe eq PTNoYs uoTIeWAOJUT sTYyI

BuqyutTejqo JO sjsoo ey. Moy st uoTISenb uedo zeyQouy ‘seQeUTISe yzeuyoueq

epTAoid 03 sZUTp[OYy UO UOTIEWAOJUT peTTeIep YATM UOTIeUTQUIOD UT pesn

2q PTNOD saoinos AseyAO Asy}eYM JO FT esaT Aeaaing oy ut pted spueptatp

pue 3Sse79jUT UO sUOT ZSeNb epnToUT 09 ST UOoTIeWAOJUT STYQ UTeIGo 07 ABM

380q 94} TeYyIOYM oul OF APEeTO Jou sT 3 ‘AeAANS YAewYoOUSg 6861 O42 YITM UoTJO9UUOD UT UoOTIeUAOJUT sTYyI YIIM WAG eptTaoad 09 uetd seop Aanseery “/

* (peares

St esodaind Arostazedns ou ueym ‘st 3ey.) sesodind squeuked jo aoueTeq 105 AT@TOS aWOOUT aOTATES UZTEeIOF UO UOTIeWIOJUT epTaord ptnom 2ey. syueq jo perztnbez sqziodea ey 03 suotTqsenb ppe 03 [ToUN0D UoTIeUTWeXY sUOTIANITISUT TeToueuly [erlepey eyQ ZuTouTAUOD. uT [nJsseoons useq sAeMTe Jou sey Vad ‘“1eAC<IOW “II APM PTA0M eOUTS peQonpuoD useq sey prorge Aue seAUT OTTOJR3I40d *S'qg Jo AdaAANS YAeUYoUSeG ou ‘UOTITIPpe UT j WeewaseauT OTTOF340d uZTez0Z uo pted omoout uo uoTjeUWAOJUT YITM WAG eptaoid Jou ptp Aanseoayl ‘seqeqs peqatun ey. UT QuewjseaUT oT[OFI}A0d usTeI05 Jo ASAANS YAeUYyoUSed HEE ey2 Jo aed se ‘TeAemoy ‘eWlODUT JUeUTSeAUT OTTOFIA0d Fo saqeutqyse s,yqG A0J YAeuyoueq e eptaoad sButpjtoy jo Aanseesal &q sKaains yxreWyoueq oTpotied ‘uoTatTppe ut ‘smoTz Teatdeo otTosj1z0d uo BIep FO voAnos 9y2 e1e sqaoderz (OIL) Teatdep [Teuotzeurequy Aansveazy eyy ‘oT duexe Joy ‘umop syverq ATTBUOCTSeDD0 UoTIeAedooOD pue AezZJIp ATQUenbeazz SeT3TsoTad ‘wey. ezATeue pue esn oymM esoyQ pue ejep ey. OeTTOO OYM esoyy useMjeq severe oulos UT AT[Tds 9YyQ Aq poeqeqrzeoexe ATQUenberz ore peddoap aie suoTjsenb AeyjO sseTun sAsarns 02 peppe eq Jouuro suoTjsenb mou

qey eoUeqSTsSUT pue ZuTpunz eQenbepeut Aq pesneo smetqoad ou *AWOUODS “S$* 2BYQ FO UOTReZT[eUOCTJeUAEeQUT BUTSPeTOUT eya pue suoTzeAOUUT [eTOUeUTZ YATM voed Adey Jou sAeY UOTIDeTTOO eIep 03 paqOASp sedANOser eyQ ‘owooUT JUeUQSeAUT pUe sMoT}F [TeIaTdeso

Ajaep[noyqaed ‘seorze AeyQO UT ‘TeAeMOH ‘UOTIOVT TOO 03 peOASep seoANnoser

Access to data across agencies also has been a problem. Confidentiality requirements preclude the wide sharing of micro data, but problems frequently arise even when data are supposed to be shared. For example, it was agreed in 1983 (after a decade of negotiations between the Federal Reserve Board and the Treasury) that the Federal Reserve Board was entitled to have access to banks’ TIC reports. While information is currently supplied on a special request basis, we are still working to establish on-line access for the Federal Reserve Board to the micro data (which is stored on the FRBNY’s computer) .° Another example of problems with data access across agencies was dealt with by Congressional legislation in 1990. This legislation allows BEA access to the data provided to Census by the establishments owned by foreign direct “investors in the United States. Implementation of this legislation will require both the expenditure of resources and time.

Other problems are caused by the lack of interaction between those who report the data, those who collect the data, and those who analyze them. In rapidly changing areas such as international financial markets, innovations may require frequent adaptations of reporting systems. Moreover, those who focus narrowly on data collection may not spot implausible reports without years of experience. Automated editchecks are useful under these circumstances, but are not a complete answer to this problem.

In past decades, an Office of Federal Statistical Policy and Standards played an active role in coordinating statistical efforts

across government agencies. However, in the 1980's this office atrophied

8. The last remaining hurdles are acquiring and installing certain transmitting equipment that meets government security standards and training personnel to use the necessary software.

‘(peppe en[ eA AoOGeT Jeu) sedTATeS se paqeetQ are suOoT IOPSUPAQ

esey. BIep UPOTXeW eyQ UT Aang ‘sqzodwut pue sqiodxe esTpueyoreu

UT pepNpToUT st 3T BIeP “S*pn aU UT ‘feperq eAOpeT[InbeW Jo quoewjere7, eyQ

UT seoueTessTIp Aq poeqeot{Tdwoos st ooTxe_ YIM UOTIeT[TOUCDeY ‘SsoTAsTqQeIsS

eperz2d peustT{qnd Jo sjes omQ ay eTToUODeA 09 BuTqdweqqe QuewUr2A0?

UBOTXeW e4Q YATM Apnys e& UT peATOAUT sT ATQUEZ"INO QuewqQIredeq soreWMI0D sUL

‘weTqoid e oq ostTe Aew ooTxe_ 02 sqaodxe Jo ZutTqaodezzepun ‘psr0jTuoW

AT@eSOTO eTOW ST OOTXeW UITM Aept0g ‘*S‘n 9UuQ YsnoUurATY ‘seqeIaIS peatug

aya Aq peqoeT[oo eqep jarzodxe Joy eqep jarzodutT uetpeueg ZuTanqtasqns Aq

wetTqoid sTyq 2o0F sz0e7100 (snsueDg ey Jo nesing ‘uOoTSTATG epery usTer0g)

quowqzedeq eoremiog ouL ‘wetTqoid aze[noTqied e st epeurg pue “sn eI usvenjeq sZutTssoz19 2zepz0oq uedo Auew eyq Ae sqiodxe jo e8e1ea09

‘squowewze Jo squowdtys ,yooq ySTu, PeTT#O OS se Yyons suOoTQeTOTA

@SUZ0DTT QAlodxe eTqIssod aeATOAUT 3YZTW QeYQ Spoos 02 peqATWTT st Sutuserzos

quesutaqas {AT [NJereo perzoRTUOW You st squowerzTnbexr Zut{Ty yszodxe

yITM souet{Tduos ‘sjzodut exT[UM ‘eqenbepe sseT AeyMeuios eq 09 ATeATT

ST ‘Zeaemoy ‘sqiodxe Jo ederzaa0g «6‘sqazodwt 3nip se yons ‘spoo3 paTssnus

toy ydaoxe ‘aqetdwuoo ATqeuoseez oq 02 ATeATT Sst.seqeas peatun ey teque Aey. se spooZ jo sqiodwt jo aoTarzesg swoqsng ey Aq eBereA0g

| Spoop Jo Ssjaxoduy pue sjxzodxy “Vv

"SQZOJJO UOTIOST[OO eQep AeYyAO se [TEM

se ‘SUOT]JOBSUPIQ [PUOTIEUATEQUT *S' UO BQep ey. BuTAOCAdWT ut deASs AsATF

[njesn e& eq p[Nom seTouese [eLepey ssoitoe UOTIOETI[OO BIep BuTJeUTpI00D

pue ‘setatzotad Sutystiqeqse ‘soTqastqeqys Terepey go AqtyTenb eyQ Zutsorzdut

0} UOTIPOTPeper sqT pue ‘eo0TFjJo sTyQ JO [TeATASeY “JoVOeS oQeaTAd

aya 10jJ uepaing Zutj1odez BuTonper uo ATeATSNTOKe JsouTe pesnooj pue

- 10 -

Efforts to reconcile U.S. trade data with other major trading partners (to improve the basis on which to judge the reliability of U.S. statistics) is also being undertaken with the European Community and Korea, although the coverage problem is unlikely to be as large in the absence of a long land border. A study was completed in March 1991 that reconciled the U.S. and Japanese merchandise trade data for 1989; when estimated adjustments for differences in coverage and definitions were made, the difference between U.S. export and Japanese import data was only $1.3 billion out of total U.S. exports to Japan of $44.5 billion.

Audits of a sample of exports through four major airports in 1988 also suggested that coverage at airports of exports to countries other than Canada is substantially better than coverage of exports to Canada. It should be noted, however, that even if overall coverage of exports were fairly complete (e.g., 95 percent), the dollar value of omitted exports would still be substantial (about $20 billion in 1990). Improvement in the coverage of exports through air, sea, or overland ports would require additional manpower and efforts devoted by the Commerce Department and the Customs Service to verify documents. In addition, it would be useful to clarify the legal authority of the Commerce Department to investigate exporters’ books and to penalize them for failure to file timely or accurate declarations.

While coverage of U.S. trade in goods is probably fairly good, the accuracy of valuations is probably a more serious problem. Understandably, the Customs Service monitors declared values of dutiable

imports more carefully than other imports or exports.” At the same

9, For this reason free trade agreements, such as the one recently signed between the U.S. and Canada, may result in a deterioration in the trade data.

‘saoTAzes AZepun sjazodxe pue sqizodut AzeqtT{qtw SOPN[OUT suoT_oORSUReAQ TeUOTQeUAEQUT “S$'n Jo uoTQequeserd S,VAd “OT

("eTQIZTTZeu You ereM SsTsTAO J[NH uesirzsog

aud UaTA peqetoosse sqiodwit z0 sqazodxe AzeqtT{[qTw jo Butjizodez eyQ UT

sueTqoid Jey eTqeaToouod st 4]) ‘AouederzosTp TeoT 3sTqeqIs eyQ UT s3urms

zofew 20F QUNODDe 09 YSnous eBzeT Jou ere sqaodxe AzeAT[IW ‘seouRQSUNMOATO

Qsow tepun ‘xzeaemoy =‘ AOUedeTOSTp [eoTAsTIeISs 9YyQ UT sSuUTMS pue sweTqoad

Bulut} 02 eAnqTzQUOD ‘uoTseoDD0 uo ‘Aew pue ‘ATYQOOUS un sAeMTe JOU

Soop Vad 09 WeyQ BuTAQTWsUReIQ pue QUoUWQAedeg esueyeq eyQ JO soeyoret APF

eyQ WorlF eQep QUBAETOA 9YQ BUTQeTNUMDOe Jo ssaoo0ad ayL or Roza asn oF

peseyoand ao ‘peoarqe se{Ttdyoo0qs worjz pros ‘seseq AxzeAT[ IM worzy ATIOOATp

peddtyus jueudtnbe Azeqt{qwm jo Butqiodexz ay AzA0jJ Quewjzedeg esuezeq ayy uo ATor 4ysnu Wad 2eUuI ST eIep sperzq eYyQ UT wWeTqord [eUOCTITppe uy

"eqep eperzq aya

jo Aodeainooe ay} ut Queweaoadut oq Aeu sqz0jjye SYI esey. Fo Yonpoad-Aq V¥

“SOTQTLTTGETT XBQ ewooUT “S*n ATEeYQ eonper oj suoTIeAOdZ0D TeUOTIeUTITNU

kq Sutotad azeysuezqy Fo asn oy2 YIIM Teep 02 SY 243 FO siromod

ay? pepuedxe ATQUeDer sey ssetZu0D *sqi0dxe jo enteaA ey ZuTQeIsSAepun

pue sqjiodut Jo enqtea ey BZuTqeqszeao Aq peorige suoTjoTpstTanf xeqQ AeMmoT

02 SOTITTIGeTT xeq eouey pue sqtjzoad aztys 09 seotaid zezsuerzq esn Aeu

‘sqzodut pue sjiodxe ‘s‘pn yQ0q Jo e8equeorzed QueoTjsTusTS eB OZ JuNoOdoe

yorym ‘setuedwoo [euoTeUTQ[NU ‘UOTITppe UT “sToazquOD TeqTdeo YITM

seTzqunod wory sjiodxe [eqtTdeo estn3stp 09 Azepszo ut sjaodwt 10 sqaodxe

Jo enyTeaA ey Aueseaderstu Aew steperq ‘squeuureAc’ AseYyQO JO “§'n eY4? 0F

pted setjnp swojsno ozTWwTuTW 02 Sqz0jJJe sepTseg ©‘sentTea qQueserderstu 03

@ATQUGOUT BU0IAS e SAOperQ.2ATZ ABW sreTazeq eperq Jo soueserd oy ‘out

The net effect of inadequacies in the trade data on the statistical discrepancy is not clear. Omitted drug imports may very well be on about the same scale as omitted or understated exports. As for contributing to wide swings in the statistical discrepancy, it seems unlikely that errors and omissions in the reporting of trade in goods would vary sharply from period to period.

B. Exports and Imports of Services

Over the last decade, the Office of the U.S. Trade Representative has lobbied hard and effectively for improved data on international trade in services for use in the current GATT negotiations. BEA has responded by adding estimates for a long list of services to the accounts (many based on new surveys), improving estimates for services already included in the accounts, and providing separate estimates for certain services that previously had been aggregated with other items in the accounts.+! Table 1, reproduced from Ascher and Whichara!?, provides an overview of these improvements. The net impact of the introduction of these improvements was to increase recorded net exports of services and to reduce the positive statistical discrepancy in the U.S. international transactions accounts by about $10 billion per year in 1987 and 1988.

Much has been accomplished, but certain problems remain. Responses to the survey of travelers’ expenditures are very low and

estimates of receipts and payments are probably subject to a substantial

11. See Anthony J. DiLullo and Obie Whichard, "U.S. International Sales

and Purchases of Services", Survey of Current Business, Volume 70 number 9 (September 1990). pp. 37-72.

12. Bernard Ascher and Obie Whichard, "Developing a Data System for International Sales of Services: Progress, Problems, and Prospects", in P. Hooper and J. Richardson (eds.) International Economic Transactions:

Issues in Measurement and Empirical Research. Chicago, Chicago University Press, forthcoming.

UL papsodaa se sysodul pue sqsodxa paadxa aiay uMOYS saanbt4y saseysind pue sales 34} yo aoueteg “S*f} ayy ut uMOYS ate Saanbt} jeu ayy asnesag

aud yey} OS *(SaTezS Paztun ayy ut uBtauo} pue peouge

“syunoase asoy} *syunoaze sjuauxed

“pausaasip aq jOU PLNOD apesz jO MOL 4 Aem-OM} "S°M) JUaWySaAUuL zDeAIP JO adh} yoea 410} pazjau 340M

sjuawked pue sqdtaoas yng ‘6/61 UL pa}Ia, [09 asam Siaublasoy payer !yse YIM SBDLALaS UL apes} UO OYE “T

stsf| Puy JtwoU0dz yO Neaing ‘adsauMO) jO Juawziedag “S$ °N_+994N0S

Ov 08 rach

“eu

el 86 819

SOV

6E

dot 9LI le

698 LLs vo60 19S SS9 9S9 evs cel 660 Sys SL9 2e6 860 980

Tel

Bee 6IT Lye vee S02 6€S

988

LYO

saseyound

v8 Tel 9€2

ys 2g9e

le

I a7 *2

°2 "I

“L “TT “el ‘L "€€

“T

"EL

"€ "€ ‘2 *T

“€

‘9

“08

sn

geet

682

7a)

961 SSe vst Le

9s £02 69T 00S 699 1e8 evl SSé 028 y9e 6e6 TZZ St6 22s

efe

iev SSY 288 820 BE2 992

Br

ogy

‘T I “T °S °2 “1

T ‘€

v

wal ‘Ss "eI 8 “92

°2 “Ul ‘9 ‘8 ‘vI "€

‘€

“al

°S6

sales

“oA

€9 “e'u “eo

c

090 0 «UD cece

© Cc

00 0 9D eccce

i)

bee’ Ses "etu “etu ‘evu “e°u Tes yg9'e 102°9 90601 2:3 eo €lp'6 60€

Ce ee ee on 0 n 0 9 ececcece

“eu

saseyound “sf

bee -p- -p-

ec

090 0 0 ceec

yS0'l

og 0 9 cececce

Toll Ble “evu -evu ‘evu "er u 6bt

06€° 2c" 126° 9st" Typ" v02°

= ONOMNM WO

00 00 DD cececcceece

“at qe, Leae ON

49410

sadtasas Asozesoge, pue ‘butysa} LeLosammo0d *juawdo] aaap pue YydsPasay Lecce nee ee eee eee een eneee saotAsas | P9Lpay “+ suotzetas otyqnd pue *Burz_nsuod * yuawabeueq re $991 Aas , e637

a ee ee quauwdi nba yo a.edau pue aoueuajulew ‘uo! ze, | e}Suy

i you "Bulut pue *uoryonaysuod *| 24nzdaz1Yyoue *Butuaaut bug S9DLAs|S UOL}EWUOJUL 19YyZO Puke aseq eG tots rss s sess Hurssac0ud eyep pue tayndwo) a ee Butst Jaapy Burdaayx00q pue ‘Huryipne ‘Bur zunodsy

-e°

spequas wpty

- sad_auas Leotuyoay pue ‘| euolssajoid ‘ssauisng

jou ‘aduesnsulay yau ‘aoueunsut Asewlid

43430 SAIL ABS 440g

a a cs qybias4

SadtAsas 439410 teres sete sess ess saa} asuadt, pue saz, esoy

SdD1AuaS 4138439 cece eee eee eee eeeeeee saa} asuadi, pue saiyyecoy

SUOL TEDL UNUWODA| aL

aoueinsu] sadlAuas [eyoueuty uolyeonpy

uotyeysodsuesy 139430 sauey sabuassed [aaest saa} asuadi, pue sat}, eAoy

sazerptjyge ublasos YIM

a ee sjuaied ub,as0} UutlA

(sae_Lop 30 suot{ lth)

pue G£6T “S0{Asag ul suolzoesuUeI] 419ps0g-Sso1) “S'N--'T ALGeL

- eT

u

suaublasoy pazelltsseun YIM

margin of error. Estimates of financial services provided to foreigners such as swaps and foreign exchange transactions are incomplete. The survey of business, professional, and technical services probably is more adequate on the export side than the import side; providers of these services to foreigners are more likely to be easily identified than purchasers from foreigners. Although recent improvements in the data on services have reduced net omitted receipts, it is not clear that further improvements would necessarily operate in the same direction. Moreover, continued efforts are required to keep up with a rapidly changing world.

C. Unilateral Transfers

Unilateral transfers include a wide range of both government and private transfers. Among the more problematic items are estimates of immigrants’ transfers. The United States receives large numbers of immigrants each year; not all are penniless, but no estimate of their wealth is included in the accounts. This omission contributes to a positive statistical discrepancy in the accounts to the extent that reported holdings of foreign residents decline and those of U.S. residents increase as a result of immigration. !3 Lack of information on immigrants’ assets abroad would also lead to a continuing underestimation of U.S. investment income. On the other hand, immigrants frequently make remittances to relatives left behind; coverage of these remittances is

incomplete, and does not include any currency sent. /* It is not clear

13. BEA has occasionally made adjustments to unilateral transfers of this sort. For example, an adjustment of about $1 billion was made when John Paul Getty died, because, although he was a U.K. resident, his estate was a U.S. resident.

14. Evidence gathered in connection with the recent amnesty granted to illegal immigrants indicates that the total of these remittances sent abroad has been substantially understated.

io seseyoind rey,0 YATM peqeser88e) peqioder ere saSueyoxe uo pepern suoTjedjtotqaed ATuo ‘sotqoerd ut ‘sdtysazeujaed peqtwt[ uo eqep epnqout ATQuerino sqiodei-g (OIL) Teatdeg [euotqeurzequy Aanseerzy ayy ‘Ar00yQ ul ‘(estadieque pe,ertodiooutun ue ut 4setequT queTeatnbe ey, 10 Aueduoo e ut AjiInbe SutqoA eyQ Jo queozed OT Jo [oAQUOD se peUTjJep sT AuewQseAUT Q0eITp eouTs) sdtysrzeujqied peatwtyT ysnorzyA quewqseauT e3e3Se [eer Zo osn Teuosied oF oQeISe [eer [eTIUepTseI uO perTnber ere sjiaodez juewQseAUT Qo0eAITp ou :squewertTnbex SutTqiodez snotizea Jo syoerIo sy usemjeq TleF 93e9Se [PAT *S'h UT squeuwqseAUT USTeTOF Ysow ‘AoAoeZOW ‘uOoMMIOD are SqUeUJSeAUT [[eUS esnedeq eete wetqoid ATeyxTT & ST eQeqse TROY *suOTSTADZ TeTQUeQSqns uT S3uTA[Nser ‘aTqeTtTeae evie sieqiodez mou worZ eRep eATOJeq Anooo Aew s3eT ‘uoTIIppe ut (‘setTueduos *s'q peperzq AToOTTqnd Jo AAtInbe ey Jo siow Zo queoied ¢ Jo eseyoind ey} Jo uoTIeAASTZeX eATNbeA suoTyAe[N3ea 9yS) *suoT}oPsurIA ojeATId To squeujseauT [][euUSs Jo sseo aeyQ UT ATeXTT ero ST STUL ‘S02e92S peqaTUQ oY. UT SquUeUQSeAUT UZTeT0J sWoSs oTe se ‘pessTU ale peorge squeuqQseAUT QoerTp B3uTyew setueduos ‘s‘qg euos ATpeqqnopup *szeqiodei mou Jo soeuweu ay. AOF uoTIeWAOFUT eTqeTTeae AToOTTqGnd ZuTIOQTUOW OR SedTNOSeT [eTQUReASQnS seqQoAep WAG ‘eQeTdWoo AT qeuosPer eq 09 A2YUZnoyA st e8et9A0D 6‘sqiodez perztnbex A{TeBeT Jo seTres saTsueqxe ue yZnorzyu2 wad Aq peqQoeT[Oo ere squeWjseAUT oSeyQ UO peuree sUODUT eya SB [eM se sMoTJQNO pue sMOTFUT TeQTdeo QuoeuwqseauT 3001TG ‘aaTQeZeu Jo saTatTsod eq prnom ATTe198Ued

Slojsuezq [ele_eTTuUN UT sUOTSSTWO pure SiOIZe 9YyQ Fo Jou 9YyQ A2eYyQeYmA

sales of stocks) .1> Perhaps some thought should be given to changing the way data are collected on the international transactions that are associated with limited partnerships or with various forms of cooperation such as research partnerships, technology sharing, or contractual production.

Direct investment reports for balance of payments purposes include information on income (dividends, interest, and reinvested earnings) and capital flows (changes in equity capital, intercompany debt, and reinvested earnings). Since these data are required by firms for a variety of purposes, accurate reporting should be routine. BEA does monitor the data carefully, but is necessarily dependent on reporters to devote sufficient resources to ensure accuracy; BEA does not conduct audits of firms’ books. Late reports are an increasing problem. In addition, foreign investors in the United States are sometimes unfamiliar with reporting requirements and recently acquired firms, heavily debt-burdened and stripped of "fat", may not devote sufficient resources to ensure accurate reporting.

In addition, it is uncertain whether wholly-owned U.S. affiliates of foreigners report as instructed using U.S. generally accepted accounting practices rather than U.S. tax accounting or foreign accounting practices. Differences between economic and tax depreciation charges could have substantial impacts on reported income and reinvested earnings. Since reinvested earnings are included in capital flows as

well, the net impact on the statistical discrepancy of the use of these

15. Clarification of the TIC-S reporting instructions in July 1990 may help some, but since many real estate partnerships are formed without the participation of current S-form reporters, coverage is likely to continue to be inadequate.

"QueuwqseAUT JOeATp jo MOTJUT O32eT Queder 9yQ wWorz BuTATNseI ueping ButotTazes eangnjz ATOXTT eya SuTssesse UT [eTonz9 sT uot Asenb STYQ 09 AeMsSUe oUL ‘“eANQeU sQqUew QsoAUT SB est uinjer Jo seqei peqioder rzeyjeymM se yons suotTjsenb sseippe jouues ‘sqzodez1 Auedwoo [enpyTaTput ey 03 sseo0e worjZ perzeq ‘sizeptsino ‘LT *sesodind swoqsno Joy ereToep AeyQ se sqtjoid aqe[noteo

07 eNTeA sles ay esn Asnu sWATJ ‘seTAnp suojsno 09 Aoelqns spoo3

toy ‘Jey peTna ATAuenbesqns syI eu, ‘sesodind swo jsno Joy pezeloep Aeyq uey. satzoad eqeT[TJZJe “S'y Burqe[noTeo ut sieo peqiodwt z0z voTad zeyustTy ATTeTQuejsqns e pesn AodyQ ey JOR eyQ UO peseq ‘uessTN pue ejoMo]

Wory sexe Yorq [eTIUBASGNS peqoeTI[OO SYI eyL ‘peqIeasteAo eq pTnom ATqeqoid sqiodut jo enzea eyQ anq ‘pejzeqjsiepun eq p[nom squewted ouoout JUSUIZSeAUT JoOoTTp uey ‘sieved pejAodUT AOZ eAeT[TJJe eperzq eTeseToyM

"S°N S32} peSzeyoreao Aeanjorjnuew oqne ssoueder e JT ‘aTdwexe t0g ‘OL

JO esed oY} UT ST QeedTJoU eq p[nod AouederosTp styl, ‘eQeAnooe st weqsds Azque eTqnop ey ut Azque ButTpuodseizzoo ay jo Sutq10de1 ey jo Butwty ey 3ey. Butumsse AouedertosTp ey, uT s8uTMs 03 eqnqTtqUOD ptnos pepnypout ere Avy. YOTYA UT TeqaTenb 9yQ Aen0d ATQOeXe JOU Op BIep eyQ jeu qoeF SUL "GT tequieoeq 03 CT ‘Ideg WorAF UNA p[NomM HH UT pepntoUT eIep ‘GT Aequieseq uo spue azeef [eOSTF S,WITF ey FT ‘oTdwexe toq ‘saeeX [eosTy azATeEUI JO STseq ey uo Qio0de1 02 pemoT{e ere switqg ‘Aouederostp [eotastjeqs ayy UT sZuTMs 03 e4nqT4AQUOD p[nod Jey suOoTIOPSUeIA QUeUSeAUT

Joe1Tp Jo ButTqazodez ey} ut swetqord ButTwtQ [eTjUejod ere ereUZ ry Bet os zvedde sejeqgs peqtun ey. uT QuemqseaUT 2001Tp USTetOZ UO uinjet Jo seqer ey. AYyM But IeStqAseauT uo AqTAoTad ySty eoetd p[noys vag gt swunowe ewes ey Aq pejeqssTU e1eM OSTe SqUNODDe seodTAXeS pue opera ey. UT pepntouT ejep eyQ JT AouederzosTp TeoTastqjeqas ey uo AoedWT Jeu ou aaey p[nom Zutotad tezsuezq Jo esn sTyQ ‘JeaemoyH ‘peoraqe suotjoTpstan[ XB] JEMOT 02 eWOoUT eTqQexe} AFTYS 09 seTuedwoo [TeUOCTIeUTIATNU USTetOF pue “s'n yjoq Aq Sutotad rzezsuetq Jo esn oyQ oq Aew JOROeZ ASYQoOUYy °s02BIS peaTuN ey}. UT SAoQSaAUT JO8ATp usTer0Z Aq peqjaodeaz uanjez jo sojei moyT AeA ayQ BuTuTeTdxe ut eTor queqiodwt ue Aeq{d Aew saotjoeid

BuTj{UNODDe ‘TeAeMOH ‘aTQTST[Zeu oq Aew spilepueqs But Aunoooe Burress Ip

large acquisitions or sharp shifts in financing patterns. However, it would be an unreasonable imposition of burden on firms to insist that they keep a separate set of books on a calendar year basis for their reports to BEA.

Another potential problem in the direct investment accounts that could contribute to the statistical discrepancy is caused by the fact that accounts receivable and payable between affiliates and parents may be denominated in foreign currency as well as dollars. Changes in their dollar value, which are used to estimate capital flows, could reflect exchange rate changes as well. If these accounts are largely dollar denominated or if payables and receivables in each foreign currency largely balance, then only small errors would be introduced by assuming that changes in the dollar value reflect capital flows. However, no information on currency denomination is currently collected; an occasional question, perhaps added to the periodic benchmark surveys, would be useful in determining whether the practice of denominating accounts receivable or payable in foreign currencies is widespread.

E. Currency

BEA does not include estimates of increases in foreign holdings of U.S. currency in the U.S. international transactions accounts. Conceptually, they belong in the accounts; foreign assets in the United States include non-interest bearing obligations of the U.S. government (currency) as well as interest bearing obligations such as Treasury securities. Estimates of U.S. currency in the hands of foreigners are necessarily imprecise. Piecing together information from surveys of households’ currency holdings, assumptions about businesses’ holdings,

IRS estimates of the volume of illegal domestic transactions, and

‘Sutpueqsqno Aouezaino ‘s*p aya Jo FTeY Anoge poy sszeUsTer0J

ey. Yses3ns ‘puewep uste10z Aq poutetdxe ere AzQunod oy Jo ser

aya pue JOTAASTG sATesSOYy [eLepey YOK MON OYA UseMQeq AQTOOTeA ADUeAAND UT SsouUsrTSeTITP Aey. uot jdumsse sy. uo peseq ‘seqeWTQSse QUeDerT BIO “gT

B ANoyATA ‘(MOTJUT peprooer &) SASUBTEeAOF 09 SeTAT[IqeT[T yueq ‘s'n UT esPverTOUT Ue MOUS PTNOM squnoooe oyQ ‘syueq eYyQ YsnoryQ ATTeZeT seqeas PeITUN SY OQUT YOeq MOTT pue ‘peorqe perepuNnel, ere sBnaip 10x squewtked yseo ey. JT ‘teAemoH ‘sSqunowe [enbe Aq pejeqjsrspun ere (AouezIno "S'Q JO sZuTpyToy uster0y peseezouT) [eqtdeo Jo moTJuT ou eyQ pue ATOTJep Qunosoe quezAnd ey YyAOG ‘st AeYyL ‘peqzodezun 03 wey A0F sAzsonpolzd uZTet0z 02 squeuked oyq pue s8nap jo sjiodwt ey. YyI0q aouTs ‘pejoezje A[30eATp Qou st AouederzOsTp TeoOTISTIeIS Jou 9yQ eseo STYQ UL ‘epRIa Snip ey Jo A[nser e se spuey ustet0z ut dn pue ATqeqoad Aouezano ‘s'n JO squnowe QUeOTJTUSTS ‘septs YyQOq Uo peqaTwo ATqeqoad eae Aoueazano ‘s‘n JO SZUIPTOY UZSTeTOJ pesverTOUT UT Q[Nser Jey sUOTIOPSUeIQ AUP ‘peqqtwo o1e uoTQOeSURIQ 9sYQ Fo septs yjoq JT AouedetOSTp [ROTIASTQeIS JOU 9YyQ OF Ppe YOU [TM UOTIOPSURIQ e ‘Sutdeeyxx00q ArQUe eTqnop jo wejsks e UL ‘SUOTSSTWO pue siOtzZe sATIeZOU pue eATatsod z93zeT, yonu Fo qeu eyQ st AouedezOsTIp [eoTAsT_eAs ey. AeUI putw ut 1eeq p[noys Aepeer ey} ‘AeAemoH *(UOTTTTG OGZ$ Jnoqe) sArveK OF ased ay. AZeAO QuNODDe sUOTQOBSURPAQ TeUOTQJeUTEQUT ‘S‘pQ 2eY4Q UT AodUedeZOSTp TBOTASTIeAS SATIETNUMS 9YyQ FO 9ZTS 9YyQ OF SATQIVTOA QUeOTJTUSTS aze AoUeTAND ‘S*g JO sBuTplToYy ustTez0F uo sTejo0Q esau “UOTTITG ST§ ueyQ orzom sdeyrzed ‘O66 UT e3ATeT ATABTNOTIIAed 910M AouezInd “s'g Fo sZuIp[oYy ustet0Z uy seseetouT ey. seqQeoTpUT seodAnos jo Aqoetiea e& WOTF UOTIeUWIOFUT gt UOFITEA OLI$ pue ogg useeMQeq- -sTEeUsTeIOF jo spuey eyq ut eq Aew Butpueqsqno AouezANd *s‘q eyQ FO SpzTYyy-o0M}Q

pue suo Udsemjeq QeYyQ sqses3ns pekorzAsep Ao Aso, AoUeTAND Jo se jeUtqAss

- 8ST -

corresponding entry of the opposite sign (an unrecorded outflow). To the extent that currency payments abroad flow back into U.S. banks, the net impact of the drug trade on the statistical discrepancy would be negative.

In addition to cash payments to cover drug imports, U.S. currency ends up in foreign hands through a variety of channels. Banks (including the Federal Reserve Bank of New York) both make outgoing shipments and receive incoming shipments. U.S. tourists spend currency abroad and foreign tourists bring cash to the United States with them. U.S. residents send cash to relatives abroad and migrant workers take cash home with them.

Only limited data are available on these currency shipments. Persons taking more than $10,000 in currency into or out of the United States are generally required to file a Currency and Monetary Instruments Report with Customs (CMIR). However, Customs’ emphasis is on finding evidence of illegal transactions, not collecting accurate aggregate data. There are certain exemptions from the reporting requirement (which Treasury currently is considering eliminating), and compliance and enforcement are spotty, particularly on the outflow side.

Efforts should be made to close these gaps in the Treasury CMIR reports and the data should be incorporated in the international transactions accounts. Alternatively, a regular survey of banks involved in incoming and outgoing shipments could be initiated. Admittedly, the accounts would still lack data on illegal shipments of currency, currency inflows and outflows associated with most tourists, and currency mailed by U.S. residents to relatives abroad. One would guess that the currency

outflows through these other channels are likely to exceed the currency

‘sqe8pnq umo Ateya Aq peteaod ere eQep OIL e842 BUTQJOeTTOO Jo syueg eATesey Telepegy oy 02 sqsoo eyy ‘AAnseary ey AoF squese se ejep ey. AoeTToo ATTenQjoe syUeG sATESOY TeAEPeT ou] *6T

‘“suoTQovsuRrAQ [BUOTITperq-uou Uo ATAe[NOTAAed ‘uoTIeWAOJUT [NJZutTueou ZutTptaorad Jo yseq ey 02 sqQenbaepeut A[BurTseerout ‘yazomyojed e auwooeq sey weqsks Quezand sy, ‘weqasAs UOT ADeTTOD e2ep smoTZ Teqatdes eToym oyQ YUTYQeA 03 sQa0FJe OTpotazed seatnbez SMOTJ 9SeY4Q UO eIep aejenbepe AoseT[oOo oOo], ‘smoTF Teatdeo [TeuoTIeureqUT JO AING ey SAeAOD AVsBuOT ou suoTAeAodiod. pue seTAPTpowWreQUT TeToueuTy aZieT Fo Asequmu peATWTT e wWoTF uoTAeWAOJUT ‘squedtoTqaed meu pue squoumMAQSUT MoU BuTATOAUT ‘pedoTeaep esaey SsMOoTJ TeqtTdeos AJoJZ sTeuueyo men ‘sqeyxAeU [eBTOUeUTJ JO UOTAeSZT[eUOCTIeUASQUT BuTSeeTDUT 2eYyQ 09 paegngqTazqUOD IlB® eaey sTorzquos Teqtdeo jo uoTjeUTWITe pue ‘sJeyxAeW [Te TOURPUTF

JO UoTAe[NZerep ‘esueyo TeoTuyoeQ ‘uOoTIeAOUUT ‘sepeoep qUeder UT qt SuOTa Fpuocos qoyAeW [eToueUTF BuTZSueyo ATptdea JO 90RF 9YyQ UT BRep ey Jo AQTTenb [[eAea0 VY UTeQUTeW 03 aeqenbepe useq Qou eaey (YAOR Mon ATAe[NoOTAAed) syuegG eAXEeSeYy TeAepegq ouQ pue AAanseerL aya 32 SMOTF Teatdeos ofjToFaz0d esayQ AOQTUOW 09 BT qeTTeAe saodAnoser ey ‘UuOoTIeWOINe UT seoURPApe aqTdseq ‘990 ‘sqTpeto epeizq ‘suoTjoORsUeIy yueg-1equT ‘sueo,T ‘sqtsodep yueq ‘squeunizqsuTt eTqetTj,oZeu w1i9eq-j10Ys ‘seTAatinoss w1ieq-3uo, Jo eTes pue sseyoind ey se yons suoT AOesURIQ yo AQOTACA OPTM & VpNToOUT eseu, ‘smoTZ [eatdeo ofToJja0d uo ejep AoF

aoanos oy eae sqiodeaz (OIL) Teatdeg [euotqeuzequy Aanseery eu ‘sTeuueyo Tey,0 YB8noryA smo[J3no AduezAno qou jo sejewtyse Aq pequewetddns st 31 sseTun ‘peorqe prey Aouerzano “S‘Q UT eseetTDUT ey JO seqjeUTAserepun sonpord of ATOEATT st (sjaodea

yueq Jo weqsks e 10) e3ep squeudtys qeu AAnseery ey. JO asn ‘smoTzutT

In addition to a systematic review of the entire data collection system, it would also probably be useful to change the instructions to reporters on identification of foreigners. Currently, foreign residents are identified by their address; however, some foreigners may arrange to get their mail through a U.S. affiliate or relative or use a U.S. agent. More accurate data would probably be obtained if reporters were instructed to use their customers’ tax identification information (which financial institutions are legally required to collect) instead, 2°

1) Bank claims and liabilities

The data provided by banks on their own claims and liabilities vis-a-vis foreigners are generally assumed to be fairly accurate. The data are reported to regional Federal Reserve Banks and aggregated by the Federal Reserve Bank of New York, as agents for the Treasury. Given Federal Reserve powers over the banks, they tend to be cooperative, but errors do occur: respondents enter data in the wrong columns or on the wrong lines, they report in dollars rather than thousands of dollars, or bugs appear when changes are made to reporters’ automated systems or when their new employees are inadequately trained.

As a user of these data, I have uncovered numerous errors in the process of investigating large changes that would not be expected on the

basis of known financial market developments. Undoubtedly, many less

glaring errors go undetected. The FRBNY always questions reporters about

20. Use of W-8 tax filings to identify foreigners would need to be supplemented by instructions to classify certain tax exempt foreigners (e.g., IBRD, official monetary authorities, etc.) correctly. Treasury, in conducting the 1984 Benchmark Survey of Foreign Portfolio Investment in the United States compiled the necessary list. One drawback of using the W-8 rather than address as the criterion for identifying foreigners is that certain U.S. citizens resident abroad may still be subject to U.S. taxes, and therefore would be incorrectly identified as U.S. residents.

‘pezeaooun oq 02 ATeATT

ero ete STOATTe ATey ‘syUueq UO UOT IeUIOFUT Jo seodINOs sATIeUTEITe sey eATesey [Plepey ey. esneoeq ‘AeAemMOH ‘sUOTANATIASUT JeYyQO Urey sieqrodexz se ejeinooe sseT Aue oie syueq ey. eumsse 03 Uoseer OU sT eeu ad *Su0IM ST Tey.O eYyA AO AazodeA suo AeYA IeEeTO

ATyTensn st aT ‘UOTTITG OT$ se 3Tq se ore SeTousqjsTsUuOoUT UusymM ‘soyourrq usTe1OF SAT Jo sqeeys souel[eq 9y. Ao Aeoys eourTeq s,yueq oyq UT sesueyo anoqe UOT IeUITOFUT ASYIO YATM QueqjstTsuoouT sT eSueyo peqaodea e yAeYQ

epew eq ued ueayo squewspnf[ ‘reaemoy ‘aeTqtssodut suostaeduoo estoead oxeu YoTYM UOTITUTJep UT SeouerezJTp ATQuenberzy ere er9yQ ‘ATpeaatwpy “[z

sZuUIMS QUeOTJTUSTS 02 eqnqtazqUOD ptnoo Aaya ‘azeqaenb oues ey. UT squNodso0e ey ut dn moys jou op Aey. JT Anqg ‘suotjoesuerzq eseyA Jo epts JeyAO aya uo eTqeTTeae st uoTjeuzojsutT oN *(Atequeqdes pue yoreW-pue ATAzeTNoTAAed) seqep Sutqiodea uteqie0 je syueq peseq-eseueder Aq Zutsseap moputm useq sey (10Q0eF3 ATUO oyA Jou ATQueredde ynq) 1z0320eF BuTanqtaquos sud “Tensnun jou ete UOCTTTTA QZ$-OT$ JO sButmg ‘“yQuouw yaxau eyQ jo Aep jasaty ey pue syquou uTeqz90 Jo Aep JseT 9y. useemjeq ATAP[NoOTIAAed ‘saoTFTO UBTeAOJF UMO ATEYI STA-e-STA SOTIT[IGeT]T pue swteyo ,syueq ut AQTTTIeTOA Ajtep esaey AXA soeqeoTpuTt (squnoooe suotqoRsuelz TeuoTzeureqQUT “S'N 84a UT pesn Jou ‘T[¢6Z PUe OSG6Z YA) SUOTAOeSUeAL ADUeATAINDOIANy

utTeIzaD Jo qriodey sAtesoy [eAApey ey. WoAF o[Qe[Teae uoTIeWAOJUT “sZ0l1e BSutTrsaooun JoJ suvew ATUO 9y3 eptTaoad ATQuenberz sdetteao ssey} ey. UOT ITUZODeA ey Aq peredweq eq p[Tnoys sqiodea uo uot jeuxtozut jo uotTaeot {dnp [Te @AISUTWTTS OQ SsqAosTY tz Sttodes ssojoe AdUaSTSUOD AOF syoeyo Aeqndwoo JO epeu eq p[Nod asn JejveIs ‘XeAemMoH ‘sqtodea1 AeyQO UO pagazodel ejep jAsutese syxoeyo-atTpe Jo ANGUA ea Aq epew ATQUeTANO sT asn aos “sqiodert earesey [elepey T9syAO pur OIL ATeYyR UseMmjeq seTOUeASTSUODUT YyaTM pequoazjuoo are AeyQ sseTun uoTQeSTAseauT Yyonw AnoyATA eqeanooe ete sqiodexz eyA 2eYyQ seoueinsse eptaord oj pusQ Teuuosied

,syxueg ey. ‘TeAeMOY {SQTWTT soUeTETOQ UTeQIeD UPYQ AeSAzeT seBuryo

in the statistical discrepancy .?? Similarly, timing discrepancies between crediting and debiting of accounts in other transactions could also contribute to the swings in the statistical discrepancy.

Another potential problem is that capital flows for banks are calculated from changes in asset positions, but other factors can also produce changes in asset positions. Write-offs of loans are not systematically reported and may produce apparent capital flows where none actually occurred. First differencing of positions may also be inaccurate for foreign-currency denominated claims and liabilities; changes in the dollar value of these claims and liabilities would be the result of exchange rate changes as well as capital flows. However, if currency positions are about matched for each currency, the net error introduced may be small. There is no information collected on the composition of foreign currency claims and liabilities; these claims and liabilities were each about $60 billion in March 1990, up from only about $10 billion as recently as 1984.

2) Securities purchases and sales

Purchases and sales of foreign securities (stocks and bonds or notes with maturities longer than one year) by U.S. residents are reported by securities dealers in the United States. Communications advances in recent years have made it more likely that ordinary U.S. residents (not only tax evaders or criminals) may conduct securities transactions with intermediaries outside the United States (e.g., the London office of Merrill Lynch), undermining the coverage of the reports.

The problem of incomplete coverage may also be aggravated by recent SEC

23. This problem is compounded by the fact that, while banks are required to report their books as of the last business day of the year, there is some latitude for other reporting dates in the TIC reports.

jo seseyoind [Te epn{our ejep eseueder ey asneoeq etqeredwoo ATI9eATP Jou ere ejep vsey} ‘qOoPF UL ‘seTATANDes AInseeTy ‘Sn Jo seseyoand esoueder uo ejep esoueder pue ‘g*p ueemqeq setTouederzostp querzedde

aya 09 sseid [eToueuTJF oyj UT peqOoAep Useq sey UOTQUEeIIe YOnW *sleqazodex ze[NZe1 Jou ere yeyA swatz Aq squewsoe{d sqeatad jo aseo ay. UT eqjeTduoouTt eq 09 ATOXTI ST spuoq yons jo e8erzea0g * (spuoqoiny ‘*3°9) szeuZTezt0Z 09 ATQ9OEATP Spuog [Tes AeU sUOTAeAOdI09 WorZ sqrzodez 4q pequowetddns ere sze[eep seTatTaAnoes Jo sqzodez sul ‘*sy00Qs pue ‘spuoq 2tey0 pue ejerodio0os ‘satoue3e pezosuods A{[elepez pue suotjerzodi0o quowureac3 *s‘g Fo spuoqg ‘sejou pue spuog Arnseery *S‘n :paqazodez aze sotz0Zeqed ANO_ ‘se7eIS peqtTuQ ey UT sAeTRep seTATANoes Aq pejrzoderz

osTe@ e1e sxz0usTez0zy Aq SeTATANDes ‘S'g Jo sales pue seseyoAng ‘eATQIeSeU 9q MOU PTNOM sseuQ ‘spoujeu BuTQeUWTAISse JUezAND Zutsn ‘esneoeq syo0qjs esourder jo s8uTptoy ‘s‘g uo eqep Butysttqnd peddojs sey waq ‘oe Tdwexe At0q ‘uoTReWAOJUT eTqeTTeAe ATQUeAIND uo eaordUT ATQSPA PTNom ‘ejeTduoouT ys8noyaATe ‘suoTAnqTAsuT [TeTouPUTJ pue ‘siojseAUT [euoT3AnzTAsuT ‘suotjerzodizoo azofew ATuo Butzeaoo AeAAns e WorAF BIeG ‘AvgaIns yAeWYyoUEQ eB YoNs ZuTyeqrzepun Butrzeptsuos ATeaTjoe ATQuUeAANO ST Aanseerl ey. ‘AeAeMOH ‘SAETYOR 07 A[NOTJJTIp eAtow AeJZ oq P[Nom sZer2eao0d aqeTduos A[qeuosee1 pue ‘syAeUYOUEeq PABMUT 9YyQ UY SQUepuOodser osAOU TeF jzo AeAANS & OATNbeA pTNom yAAewYyoueq przeAqno saTSuseyertduoo e ey} penzie S@M IT {suoTjerepTsuod 4sood Jo ssneoeq AsAANS YAeWYoOUSq MoU B syAeITepun 03 Quejonter sem Aanseeay ‘ATQUSDeT [TIUQ ‘II ABM PTA0M BuTAnp pe Zonpuos SM peoige jUeUjseAUT oOT[OJRAI0d ‘s'g Jo Avarns yAewYoueq AAnseei]

ASeT eyuL ‘UTeQAeOUN sT WeTqord e8ezeA0cd. Vy Jo QueQKXe su] *spuoqg

perzeqsTZeiun ut suotqnqtasut Aq ButTperzq Arepuooes BuTAeqi[Torz ssut[n1z

U.S. Treasuries worldwide while the U.S. data include only securities sold directly from the United States to Japan. The development of active markets for U.S. Treasury securities in London and the Far East have made such bilateral data comparisons impossible; the U.S. reporting system cannot identify the ultimate purchaser of a U.S. Treasury security in the London market. However, the cumulative total of Treasury securities sold to all foreigners according to the U.S. data appears too low to be consistent with the Japanese estimates of their holdings. This issue may be settled by the results of the Treasury benchmark survey of foreign portfolio investment in the United States in 1989 that is currently underway.

As a result of the growth of markets abroad where U.S. securities are actively traded, the vast geographic detail collected on the TIC-S form is of limited use, for it yields no insight into the ultimate purchaser of U.S. debt instruments. Eurobonds account for most U.S. corporate bonds sold to foreigners; they are generally distributed through underwriters in London and are therefore reported as sales to the United Kingdom. U.S. Treasury securities are also actively traded in London. The geographic detail on U.S. purchases and sales of foreign securities is of limited use as well. All foreign securities are aggregated so it is impossible to distinguish the U.S. sale of a Canadian bond to a Japanese resident from the sale of a Japanese bond to a Japanese resident. The problem with equities is currently less acute, but likely to become serious as well, as the listing of equities on foreign exchanges becomes more common.

The vast geographic detail is of limited use in facilitating

bilateral comparisons of data as well. For example, the Japanese report

“seTiqunoo patyz jo squeptsez Aq pensst spuoqg

SPNTOUT p[Nnom eIep ,SeTAQUNOD YoRe uT ATOBZeA"D puog UBTeTOF sy. esnedeq peryeseri3Zestp spuoq esoueder 10 ‘s'n 10J epew eq p[noo uostTaeduios sstoeird ou ‘ZeAeMoH ‘“SQUepTsez ‘S‘q 02 esourder wory spuog [[e Jo seseyoand io seTes uo ejep osourder oy ITM peredwod eq p[Nod squeptsez eseoueder worz spuoq [[® Jo seTes to seseyoind ‘s‘q uo eIep ‘S‘q eyQ ‘eTdwexe Jog ‘cz "JeTTes AO Aeseyoand ey? Jo uoTQePdOT 9YQ

ZO ‘peasTT st Aqtanoes oyQ erayM YoyAeW sy Jo uoTRPDOT 9YA ‘AOAGep sayy yo AqTTeuoTAeu yA 02 Zutpioooe peqazodez oq Aew suotjoePsuerQ ‘pexTu st seT{Tinoes ejeiodi0o *s'g uT suoTjoesueiq jo ButTqaodexz esourder oy ‘Zz

cz TeAeT a7e3ei38e ayQ Qe suosTiedwoo [Tereqe[Iq BuTAatwAed ‘z0R0ePSUPIQ ey. Jo AQTTeuoTAeU 9YyQ Fo stTseq 9yA Uo pejroder suoTjoORSURIQ SeTATANDES eaey pue SOUTTEpTNS JWI eu eSueyo 02 eq PTNOM sATQeUASATe AeYQOUY *(suoTaTUuTjJep eTqerzeduoo qadope pue seuTTeptn3’ JWI ey MOTTOF 02 BuTT TIM eq p[nom seTzqunoo aofew [Te Butumsse) seTatTAnoes ‘s*q ButAnq st oun ssesse 03 seTiquNOd Aeyjo Aq peptaoid ejep osn 03 pue ‘seTATAnoes ‘sg Jo seseyoind u8tTez0jJ uo [TeAep oTyderZ0eZ ou AoeTTOO 09 9q PTNOM sATIeUIEITe aug *SeTAQUNOD UBTeZ0F Aq peqQoeT[OO eIep sey 07 eTqeazedwod uoTIeWAOJUT eptaoid p[noo szeqzodex wi0F S-DI]L 242 AemM ou ST orayQ ‘rzeseyoand eaqeWTA[N ey Jo AATQUepT eYyQ ButTMouy Fo Aem ou saPY Seq7e4S pe Itun 2yQ UT Steqioder ‘setqatanoes ‘S'n AOJ ‘AeAeMOY ‘“ASTTeS To Aahnq 9YyQ Jo UOoT_POOT eya ueyQ A9yuQeA ‘AOQAGep eYyQ Jo AATTeUCTAeU 9YQ eReOTpPUT PTNOD srzeqirodez ‘seTAtinoess ustTez0z Jo saseyoind ‘s‘gq AoF suotjzoONAQAsuT ButTjazodez

S-OIL 242 eBueYyo pTNod sejeqs peAtuQ ey], ‘peyooq ere suoTJoOesUrAQ eTSyM Jo uotjsenb oyy UT st eteyQ UrYA SeTATANOeS ‘S‘q Fo seseyoand ssoueder

JO OZTS OY} UT QSeT9QUT oTOW ABZ ST oT9YQ AeYQ esues 9YyQ UT eqetTadordde st yoroidde sty] ‘Joaqep eyQ Jo soueptsez ey Jo stseq ey} uo seseyoand SOTATANDes BuTAJTSSeTO yseBBns soutTeptn3 qWI ‘“wuopuoy ut setatTanoes uSTetoJ UT suoTIOeSURAQ UZTeTOF uO eJep JOeTTOO OU ssOop WopsuTY peIAtun euL yz UOPUCT ut AleTpewzequyT [etToureuTZ e Aq peyoog st uoTIoesUeIAQ 2eYI

JT uervs ‘seqeqsg peatup ey. wory seseyoind se set ATAnoes Aanseerl *S'n 1?

Periodic benchmark surveys of holdings would be necessary to provide information on the basis of the nationality of the debtor.

3) Options, warrants, and futures

Derivative instruments such as options, warrants, and futures are currently only partially covered in the U.S. international transactions accounts. Some data on these transactions are available; however, thorny theoretical and practical problems make it difficult to separate capital flows from service or investment income and capital gains in connection with many of these transactions.

In theory, warrants and options are included in the TIC-S reports when the underlying security is a stock or long-term bond . Data on purchases and sales of these instruments are aggregated with purchases and sales of U.S. or foreign bonds; since they are not available separately, judgments about the adequacy of coverage are impossible. All other options or warrants are omitted. In addition, BEA estimates the margin accounts and profits and losses on futures trading, based on information on foreign transactions on U.S. futures exchanges, and includes these estimates in liabilities to unaffiliated foreigners (TIC- C). U.S. residents’ participation in foreign futures transactions are omitted entirely. Cooperative efforts with various regulatory agencies such as the CFTC and SEC might lead to data improvements in this area.

4) Limited Partnerships

Only those investments in limited partnerships that are listed on exchanges are covered; they are aggregated with foreign purchases of U.S. stocks in the TIC-S reports. The IRS is currently automating responses on tax returns, and as a by-product they soon will be able to

produce a list of limited partnerships in which foreigners have some

*szeqazodez w1r0; D-OIL TPuctitppe mez ZutprTetA ‘peqonpuoo sem swaty goog jo Aeaans y ‘97

pue [eToueuTy eaey 07 satuedwos peumo-'s‘p ueyQ ATeAT[ etou o1e AjTqeqoad setueduos uSter0z Jo sejel{TjJjy ‘pewoozysnu sey seqeqs peatun ay UT QUdUQZSeAUT JOeATpP UBTerTOF ‘aut, 2eYUQ eOdUTS gz 8461 uy pejIonpuos. SWM stejytoder [eTJuejod Jo seauro oTQewWeqsks AseT ey, ‘weTqord Aue qsuoOD @ ST sleusTe10F STA-e-STA SOTAIT[TQeT[ IO swTeToO eTqeqzodexz yQATM esoyur TTF BZuTyorer pue ‘e8re{ st saeqizodex wiogz-9 TeTQueqod JO Zequmu sy ‘osTe Ppeteptsuos eq AYystTu sTeuinof eperq pue sazededsmeu ut (ja0dex 03 eanq{tTez 1OJZ SeTqQTeued pue) squewerztnbexr Butjaodez ButsTjreapy ‘“[nyJesn eq pTNom SOTISTIeIS ejJeANODe AOF pasu sy} uo ‘10R90es eAetZodi09 ay AT Ae[NoTAAed pue ‘ot[qnd ey} ejeonpe 03 sqiz0zze ponutquog ‘sqaodezx atey, jo Aoeznooe ay. Jnoge pouoTjsenb ueym (eTTIsoy seulTjJewos pure) eaATAeTedood ssaT eq 03 pueq Aey pue eAresey [eAepey eyQ Jo AATA0yAne AroOReTNZeA ay 07 Qoefqns Jou ete steqiodert 9-911 ‘[ezeue8 ut ‘weqsks OIL ey UT YUTT Aseyeem ayy Atqeqoid eie sqioder esoyuy ‘sxeusTeTOZ STA-e-STA SOTATTIQeT[ pue suteyo TeToreuulod pue [etToueuTy Jo AqeTzeA e& apntToUT sjazoder 9-91 euL SleusTeIOF peqetT[TFJeun 09 sotAt[Iqet], pue uo suteto yuequon (¢ “petpnas eq prnoys sdyyszeuqied peqywt,T ‘s'n ut uotqedtorqaed u8tezoy uo ejep Buy Qoe Too JO poyjew Juezino sy 0 saATQeUATEeQqTY °*(perzAnooso AayA ueYyM Jou YsnoyITe) sdtysieuqied peqtut[ uy squewqseauy uSyper0F Jo uoTsstwo ey. jo Queqxe ey. Mouy [TTA Arnseery eyQ ‘eqeTdwoo sq AAoFZo uoTAewo4ne Syl ey. usYyM *sdtyszeuqied peqIWy, Jo wiof ay uT seA AT osneodeq squnoooe squouted JO sour, eq ey. WOTF peq ATWO Useq sey a4e4Se [PAT *S'n UT QueUjseAUT

usTet0F yonu yey peqoedsns st AL ‘AeT[Aee pessnostTp sy ‘eHeIS

- 29 -

commercial ties with unaffiliated foreigners. They are also more likely to be unfamiliar with U.S. reporting requirements. Rather than conduct another general canvas, it might be more cost-effective to focus on these affiliates; Treasury could check its list of C-form reporters against a list of U.S. affiliates of foreign companies provided by BEA and canvas those who are not current reporters.2/

In some sense, the C reports are a residual report, catching transactions that are not reported by someone else. They exclude direct investment transactions, securities, and custody items reported by banks, brokers, and dealers. There is frequently room for confusion about reporting responsibilities. One serious problem, the reporting of loans to U.S. residents booked at banks outside the United States, was resolved (at least partially) by the introduction of the BL3 report in mid-1986. Banks in the United States, with any knowledge of loans to U.S. residents booked at their offshore offices, are required to report these loans as custody liabilities or file a BL3 form notifying the borrower (with a copy to FRBNY) of their responsibility to report the loan on the C-forn. Banks have uniformly chosen the first alternative. As a result of this clarification of reporting responsibility, custody liabilities reported by banks increased by $18 billion, while financial liabilities reported on the C form remained virtually unchanged.

However, comparisons of the data reported on custody liabilities

and data reported to the BIS suggest that reporting of bank loans from

27. If focusing only on U.S. affiliates of foreign companies was viewed as discriminatory, U.S. companies with foreign affiliates might also be canvassed.

‘satsodep uey. rey QeI

sZuUTMOII0q yUeq peteptTsuos ere Aeyq ‘satsodep se satsodep yueq 2YusTUIeAO AJTSSBTO 09 Jou pejzonajsuT ere szejaodez ‘aTduexe tog ‘saseo uTeq190

UT ae8esn uoummo0d worz 2Aedep ATQUeAedde osTe suOoTIATUTJeEp DIL eUyL "OE ‘aoueISTSeT eTqerepTsuod yATM Jou seqeqs

pe3tun x4} epTsqno peyooq sjueptset *Ss'n YATM suOoTjORSURIQ ,syUeq poseq -uSTe10F FO (¢Eg6T) APAAINS oArESEY [eLEpegd V °(ZOGZ Ud) SQUepTsez ‘s'n uo sute[o SuTpnTouy ‘setat{tqet][ pue sqesse ATey, uo syueq peseq-‘s‘n

JO S90TJJO USTer0J ayQ worZ sqaodez ATYQuoW sje03 eAAesSeYy [eAEpeg YL °6Z “sqoustezJIp elep oy. AOJ Aunoooe AT[NF uoTAluTjep ut sesouereFsJIp

esoeyq AoyeYAM AePeTO Jou st AIT ‘seoueqdeoor sxzeHUeq pue AzAeded [TeTorewMOD se yons sjueumaqsUT eTqeTJOZeu pue seTATAnoses Jo sZurpl[oy epntout Ajjuenbery swytey[oO uo e_ep sig eya ‘aze[noTqaed uy ‘suotsnzouo; PATITUTJEp pue suostieduod estoeid epn[oeiad suoTjtutjep ut sesouerzesstTq °9z

Apojsno se (71g 9yQ UO T2eYQe80Q peQe3ei33e) pejtoderz ere pue

suoTtqnqT3sut [eToueuTy *s‘q Aq Apoqsno ut pyey ATQuenberzz ere srAsusTer0J

4q peseyoind (syq) seourqdeooe szeyueq pue ‘(sq9) 3}Tsodep jo seqeoTytqj1e9 ‘zeded [eyTolomos se Yyons squewMAQSUT eTqeTIJOZeu w1eq-AAOCUS ‘S'n SqueUMAASUT eTqeTIOZeu w1zeq- AAz0EYS (9

*SuTQUNOD-eTqnop pToae

02 suOoTIONAASUT OIL ey. UT seZBueyo erzTnbez p[nom seoAnos ejep sxsaTIeUIEQ TE

Fo osn oyQ ‘uTesy og P22838z2ePpun A[TsnoTies ere SAeP OIL 34a 2ey.? 4seBBns

BIep sATOSeY [ePAepey pue STG yaTM sTeI0R OIL Jo uostaeduoos ‘sqrodez

JIL ey. UT pepntouT ATQUezTINS squnowe sy TOF peqAnyatasqns eq prnood sqtsodep ,squeptsez ‘sg Jo peorige syueq worz sqazodex ‘ATAeTTWITS

*ZutTjunoo.

-2TQnop 4ueraerzd 09 suotTjonaqsuT 3uTAzodez oI eyQ ut se8ueyo erztnbez

OST® P[NOM sqUepTSer *S$‘pM 02 sUPOT YyUeq UO eRep BuTUTeAGO TOF weqsds

quezind oy. AOF soeqeqgs peAtugQ ey. epTsqno peqeFd0T seotsyo SuTyueq wo1rs

eqep Jo uot3nAzTIAsqns 6z 923838 peatug ey. eptsqno seoTFjFo worzz squeptse1

"S'Q 02 BuTpueT, ATeYyR UO syUeq UZTezT0F wWorAF eQep Bututeqggo jo AATTITqQtTssod

au. ST syueq TRIQUeD ASsYQIO pur STq e42 YITM peroTdxe oq pnoys

2eYyQ UOCTINTOS 9UuCQ gz 22eTduoout 2q IITas Aew squeptsert ‘s'n 03 peorge

- O€ -

liabilities. + Separate information is also provided on CDs held in custody. In addition, U.S. firms that issue commercial paper or short term instruments (with maturities of 1 year or less) directly in the Euromarkets are supposed to report these as financial liabilities to foreigners on the TIC-C forms.

Since 1978, when the TIC system was last redesigned, the U.S. commercial paper market has grown to about six times its earlier size; borrowers have found it advantageous at times to issue their own shortterm negotiable instruments rather than turn to the banks for funds. It would be useful to have more information on the participation of foreigners in the U.S. commercial paper market. At minimum, these shortterm negotiable instruments should be covered in the Treasury's periodic benchmark surveys of foreign portfolio investment in the United States.

U.S. holdings of foreign commercial paper, CDs, or BAs are also frequently held in custody by financial institutions and reported as custody claims (TIC-BQ1). Instruments not held in custody are supposed to be reported directly as financial claims on the TIC-C reports. Coverage is clearly inadequate. For example, at the end of June 1990, $27.5 billion in foreign negotiable instruments were reported as held in custody for U.S. residents by TIC-B form reporters. Another $17.5 billion in financial claims on foreigners was included in the TIC-C reports, summing to $45 billion. However, the FRBNY survey of commercial paper outstanding in the United States indicates that foreigners had $65

billion in commercial paper alone outstanding in the United States at the

31. Some of the CDs included in these reports have maturities of longer than 1 year.

‘QueoTjTusts Jou ere g9-oang Jo sZutpToy ‘s‘n 3euy

pue squeptsez usterz0j Aq seqeqs pe ytug eyq ut pensst zoded [eTorOuMI0D

JO sieyeQ eSreT Qou ere sA9usTerAoF QeyQ soumsse saTAeUAEITe STUL “€€ *szeusTez0j Jo soqet[ Tse

"s'n ey Aq sonssTt eyQ uste10f se ButAjJtssej[o ATQoOeAAOOUT ere s19aqi0de1 eWOS QeYA STQFATSD.UOD OSTe ST AI ‘eTPOS STYQ UO ere saseyoAnd ATeYy. yeUQ ATeXTTUN swees AT Aanq ‘seqeqIg peATuQ eYyQ UT sieUusTeA0zZ Aq poensst rzeded TeToremmos ey Jo ewos aseyornd op sAeusTet0Z ey aeTqeATeoU0D ST AI ‘“ZE

QuowjseauT oTToOJRjA0d peqiodey ‘seeze wetTqord snotaes ysow ayQ Jo auo se

SWOOUT JUSUIQSeAUT OTTOJRA0d peTjJTjuepT sey ‘suotjeu [[e AOF eAep AuNODOe

Quezand eTTouodez 02 sqdueqje zeqze ‘qWI eu, ‘“Suoseez [eAVsAeS AO0F qoedsns e1e sMOdUT QUeUqseAUT oTTOJQI0d *S'p UO soeQeUTASS oY

‘uinjer Jo seaqeX poumsse pue s3uTpjToy

uo ejep jo stseq ay uo ATTedtoutad yyq Aq pejeutqse ere squewted

pue sqdteoer QueuqseauT oT[Tofzqaizod ‘yxA10R meN Jo YUeG sAASSOY [eAepay our

22 pley seTAtanoess Aanseery uo seTAtTAoyqne AAeQOUOU TeTOTFFO 09 YsoreQUT

jo squowked pue sqdteo0er AserequT quoewureAocs wory Aaedy ‘peqazodez uryy AeyAeA peqeutyse ATesrzeT ore squouwXked pue sqdtooer AserzequT ‘sn

cg POtePESUcO aq ptnoys (sqzo0dex 9 pue

9q ey worzy azeded [eToATeMMIOD epnToKe 0} sUOTIONAASUT OIL ey. uT seZueyo

yQTM BuoTe) seqeqIg peatun ey ut reded [eToreuMIO0D Jo senssT usTer0J

uo e1ep ANGYA FO uoTAnATAsqns ‘uoTATppe UL ‘“squeUMAQSUT WI0eQ QAOYS

Zeyjo pue zeded [eToLemMIOD epn[oUT pl[noys perorqe juewjseAUT oT[OFII0d

‘s‘n jo Aeaans yrewyoueq Aansvery Auy ‘sjueumazqsuT eTqeTjJoseu ustTer0z

jo sZurIptToy ,squeptTsel *S'g eqQeqsizepun ATssoi3 eQep OIL aud 3eyQ savedde

3I ‘syueq ‘S'n A0F Apoqjsno ut pTey ST UOTTTTq [zZ$ AeyQoue pue ‘syUequoU

"S$°Q z0z Apojsnd ut sqo UT UOTTTTq ZZ§ Inoge poy wopsuTy peatun

aur ur syueq 2eYQ e2eOTPUT eQep pueTZUq Fo yueG “sqoQ UsTetT0Z Fo sButTpToy

TeTqueqsqns sary squepTser *s'g ‘uOTITppe ur ze 0661 eunr jo pues

income payments world-wide far exceed reported portfolio investment income receipts. Moreover, the U.S. estimates of portfolio investment income receipts and payments are based on shaky data on holdings. In addition, there are problems in estimating certain returns.

The accuracy of the data on holdings depends on the TIC data discussed above. As noted earlier, there are serious problems with the data on U.S. residents’ holdings of bank deposits abroad. Comparison with Federal Reserve, BIS, and IMF data suggest that they may be underestimated by more than $100 billion, leading to a substantial underestimation of interest income, 34 Holdings of foreign commercial paper also appear to be substantially underestimated. On the other hand, BIS data suggest that U.S. borrowing from banks outside the United States may also be understated, resulting in underestimation of interest payments to foreigners. In addition, the statistical discrepancy in the U.S. international transactions accounts has cumulated to about $250 billion in recent decades; if a substantial fraction of these errors and omissions were unrecorded capital inflows, then payments of investment income to foreigners would be underestimated. >>

Another source of errors in BEA’s estimates of portfolio investment income is assumed returns. BEA is aware of inadequacies in this area and is actively exploring possible improvements. Because of resource shortages, BEA’s assumptions about appropriate rates of return are sometimes out of date. For example, no interest is assumed paid or

earned on foreign currency deposits because, many years ago, these were

34. However, acceptance of these higher estimates for U.S. assets would imply that there were omitted capital outflows in earlier years.

35. There would be no payments to foreigners on holdings of currency or residential real estate for personal use, however.

ey. FO uoTIeUTISe eYyQ ‘szTPeA QUedeT UT ‘sSATeUZTeTOZ yATM sdems AouerzAnsd AO 93e1 ASOeTEeRqUT JO eased 9Yy UT AO soeqou aqer BuTAeOTJ AOF JueQsuOD peumsse oq qQouueos seqer ‘AeAeMOH “peATIeA AO PTOS spuoq Jo eZequtTa ey. pue peseyoind spuoq Jo saeTATAnjew 9y. Anoqe oepew eq Asnu ssoen3 ouos ‘eSueyo Jou op spuoq BZuTpurqsjno uo seqer YsereqUT 2eYI peumsse oq pTnood aT JT ueaAqg “SpuepTATp 4909S AOZ poesn poyjew oy. of snoZojTeue st spuoqg wi0}-ZuoT uo s8uTuree 4serequT BuTQewtqse zA0F AZoTopoyjeu s,yag “II 22M P[z0mM Butanp peqonpuos sem seesieAo QueuwqseautT ofjTozazod -s'q Fo AeAAns yAeWMYoUSq AANSPeATL ASPT 9 edoUTS Aozie Jo uTZ1eu epta AeA e& 09 Qoefqns ere syd0RSs UZTeTOF Jo sButppoy "S'O JO seqeUTjse ‘TeAeMOH ‘“SQeYyTeW YO0IS UZTe10F Aoflew uo spTeTtA puepTATp e3ezeae uo uoTIeWAOJUT Butsn ‘sqdteover puepTATp BuTjewtq3se AoOF poyjew sat ut eZueyo snoZo,eue ue suetd yaq ‘seotad yo09s uT seZuryo jo seqeutqise pue (S-D9IL) seseyoind yeu jo sqzodez1 quenbesqns yATmM peuTquos ‘saqeqg peatuy ey. UT 4UeUWQSeAUT OTT[OFQI0d uUstTeT0ZF Jo skaarns yAeWMYyoUSq Aanseeiy otTpotied eyq uodn peseq ere setqtnbe ‘s'p jo sButptoy uster0z JO saqewtqase eyy “syx909s QOS UO pTeTA eqTsodwod s,i00g pue piepuejs uo UOoTIeWAOJFUT YATM pouTquioos seTAInbe ‘s*q Jo sButTpjToy ustTero0F Jo saqeutqse SOSn poy ell MOU SUL “8G 02 AOR BIeP pesTAer pure O661 eUNL Jo se AZoTopoyjeu BuTAewtyAse sql pesTael wad ‘epts squewded oyQ ug ‘o8e sireek O€ PIP 21 se Aepoj PUePTATP sues ey UAve OF poeUMSsSe ST Q96T UT perztnbor yooras e {s8uTpToy 09 suoTATppe uo spuepTtatp sntd potied Ase, se owes 24a ST potied sty ewooUuT QeYyQ soumsse poyjow eu ‘ejenbepeut ost{e st seTaInbe uo ewoouTtT Jo sqdtevexr ButTjewtyse A0z AZoTopoyjeu s,Vvag ‘ezqTs sTyq 09 umorZ eAeY soeouRPTeq BuTAesuedwoo jeYyd AToeATTUN suses AI :UOTTITG g9$ 09 pequnowe seTAT{TqeT{ Aouerino usTe10Z

Sued ‘6861 FO pue oy. Aq ‘AeAemoH ‘SeoURTeq BuTAesueduoo ATeBI1eT

appropriate interest rate has been complicated by the expansion of foreign currency bond markets, the use of interest rate and currency swaps, and the issuance of zero coupons, bonds with warrants attached, and other innovative instruments. The TIC data does not disaggregate these instruments. In this environment, periodic benchmark surveys of U.S. portfolio investment abroad and foreign portfolio investment in the United States are crucial; without periodic benchmarks to check against, any estimating methodology could, over time, produce large errors. In addition, information is needed on currency and interest rate swaps.

Finally, more needs to be done to automate BEA’s estimating procedures. PC spreadsheets are used, but much data is entered by hand, rather than transferred directly from BEA’s mainframe TIC data base. IV. STATISTICAL EVIDENCE

The review of the data sources and recent history in the preceding sections suggest that problems exist in the reporting of both current and capital account data. This section turns to statistical techniques, particularly regressions, to try to explore the importance of various factors in explaining the statistical discrepancy in the accounts. An obvious approach would be to regress the statistical discrepancy against various current and capital account components of the U.S. international transactions accounts (or the underlying variables that might explain movements in these components) and attempt to compare their contributions to explaining movements in the statistical discrepancy. Unfortunately, the insights that can be obtained from such

regressions are limited.

pue queuoduod petpnjs ay. UT suOoTSsTWO pure si0i1ze Jo oedUT Jou eyQ

JFOSTJOA TLTTA YUSTOTZJ909 pajeutjse eyuy, ‘petpnqas potiaed oyQ AeAo squNoDDe

ey. UT perinodso saeZueyo Aey.O AeYyM UO BuTpuedep Area TTTM squUNoD0e

SUOTIOBPSUPIQ [RPUOTIFUZSRUT 9YyQ Jo AueuoduOD eT3uTs Aue pue AouederosIp TeOTasTIeAs: eyq UseMqeq UOoTSsserZe1 eTdUITs eB Jo sq[nser aU

‘AouedsiosTp TeotjasTqeqs

ey. UT UOTID.eATp eQATsoddo ayQA uT eZueyo e Aq ATVOeXe peyoQeU oq AsnuU

Qunooor quezIno eyQ uT e8ueyo Aue ‘Queqsuod sMmoTJ Teqtdeo you BuTptoy

a a C) Cc) 4A- WO- = At WO = dS

‘[- sewTQ Qunoooe [eqtdeo pue qQuezaAno

eya Fo ums oy 02 Tenbe st AouederOSTp TeoT|AsTAeAS 9UyQ ‘suz9Q BuTsuer1e9sy "x + oe + °yo + VO = 0

‘orezZ Tenbe yAsnu (x) eIep SMOTJF TeQTdeo eyQ UT suUOTSsTWO pue

si0ii9 pue Cy SMOTJUT [eqtTdeo you pepz0;ez eyQ sntd (yo) eqep qunoooe

quezinod vy} UT suOoTSSTWO pue sioize pue (7Wo) Qunoooe 4UuezAIND pepsz0s;er

aya FO ums oy ‘eQPAASNTTT OL ‘eBueyo oyQ soutQ [- Tenbe ATTrAesssoeUu

4snu ‘juejsuod sjuseuUoduOs reyjO BuTpToYy ‘squnoodoe ey jo jueuoduoo

Aue ult o8ueyo e WorF A[NSeA pTNom yeyQ AouedsrzOsTp TRoOTASTQeIS ey UT

asueyo eyQ ‘o1eZ 09 ums Qsnu ‘uoTQTUTJep Aq ‘YoTYM weqsks AAQUe eTqnop eB ere squNoodoe suOTIORSURITA TRUOTIBUAEQUT “S§'h eYQ eOUTS ‘AeAOCaTOW

*peqeutqjsestu

Zo pejiodeizun sem queuoduos squewked Jo sourlTeq ey Jo uoTIOPAZ eT qQeqQsS

eB Jt ATUo y8TYy eq P[NOM UoOTReTeAZ0D ey, ~“QUsUOdMOD AeYQ Jo UOTIeUITISe

Zo ZutTj10dei UT suoTssTwo pue si0719 [eTQURQSqnsS OU 91e eTZYA WeYI

aaoid jou seop sjunooor squewked jo sourTeq ey} Jo Queuoduoo aze[noTqaed e pue Aouedez0SsTp [ePOTASTQeIS |9YQ UsEeMQEq UOTIeTeII090 Jo YyoeyT

SuUOTJe}IUT] Vv

errors and omissions in the balancing components. The estimated coefficient could easily be insignificant, despite systematic errors and omissions in the reporting of the component studied, because there were also systematic errors and omissions opposite in sign in the reporting of the balancing changes in the accounts. Alternatively, the estimated coefficient could be significant even if there were no systematic errors and omissions in a particular component studied if there were systematic errors and omissions in the balancing component. °° In addition, since the composition of these balancing changes may vary depending on the cause of the initial change, the estimated coefficients are likely to be unstable, and very sensitive to the addition or subtraction of observations. B. Regression results

Given these problems, it is not surprising that regressions between the statistical discrepancy and various components of the U.S. international transactions accounts yield little insight into the sources of the statistical discrepancy. The R25 generally are very low and the size and sign of the coefficients vary depending on whether the sample ends in 1989Q4 or 1990Q4 and whether the data are first differenced or used in level form.

However, several other explanatory variables do shed some light on probable sources of the statistical discrepancy in the U.S. international transactions accounts. These variables include 1)

variables that are frequently used to explain international capital

36. For example, there appears to be a systematic negative relationship between official capital inflows and the statistical discrepancy; the most likely explanation is that part of the decrease in private assets in the United States that occurs when official holders buy dollar assets net is unrecorded.

‘quZTTJ Teatdeo Jo serznseow 09 pejetTer eq prnom Aouederostp [eoT Ast eqs "S'n aya ‘seQe2g peaTUQ 9YyQ OCJUT SMoTjJUT [eqTdeo peprooerzun ut peat[nsezt si0j3qep OQT peTqnoszy AT[eToueUTZ wors AYBTTZ [eatdeo FT SOUT WorF TUSTTF TeITaeS Jo SInSeSH VY (Z “SMOTJ} PeptOdetUN pue sjesse peqeUTWOUEP-IeT [Op Jo sseueaAT }0F1939e SATISTEA eyQ usemjeq dtysuotje,er eTdwts e jou st ezeYy AeYQ ase83ns op AeyQ ‘Aouederostp TeotastTaeqgs eyQ 09 ATTeTQUeASqns eqAnqgTzqUO0D SMOT} Teatdeo peprzooszun yey AATTTqQTssod eyA Ano eTnA Aou Op saATNser eseud eTTYUM ‘O1ezZ WoT QUeAEZFJTp ATAUBOTJTUBTS JUEeTOTFJZOI00 TeYyATeE ST zou ‘({[ uoTQenbe) AouederzosTp [eOTASTIeIAS eYyQ UT UOTIeTAeA |YyW Fo Yonu ute,Tdxe ou op seTqetizea esey. ‘Z eTqeqa UT uMOYS sy ‘(peqoedxe ATIoOeI1I00 ezem seZueyo ajel eBueyoxe [enjQoe AeUQ peumsse ST 4T ezeym) seBueyo aqei eSueyoxe pojoedxe pue soqer ysexeQuT wreq-Buo0T,T usTetoy esereae peqyustem pue ‘S$'pn useemqQeq TeTIUeTeTJTIP eyQ epnpouT AYySTW seTqeTrzeaA yong ‘Aouedezostp [eoTAsTIeIs ey UT squeweAow uTeTdxe osTe AYysTW SMOT} Yons BZutTute,Tdxe ut queaeTer ATTPOTJeAT0eYyR ere AeYI SeTqeTreA ‘smoTZ Teatdeo pepzooerzun ATesrzeT etem AdUedezOSTp [POTASTIeIAS 9YyA FI SHOTJ [EITAeO TEUSTISUTSIUT UPETAKS JeyT SSTASTIEA “1 ‘eaaTqses3ns ore Avy esneoeq pejsoder ere sqz[nser aya ‘(squnoooe ese [Ie UT suOTSSTWO pure stOIIe Fo YoU aYyA VZOeTJeI T1tm Aouedez0stp [eoTasT_eIs eYyQ pue eTeyMesTe seZueyo Aq peourTeq aq Asnm sqUNODDe sUOTIOBSURIQ [TeUOTIeUTEQUT “S'p ey Fo suo ut eBueryo eB sesneo 2ey oTqetTszea Aue) eAoge pessnosTp suOoTIeITWTT ey 02 eUuNnuMT Jou ete suOoTsserZe1 eseyQ eTTUM ‘epetQ eSTpUReYyoTeM “S'yn BuTUTeTdxe [epow e UT si0ize (9 pue ‘AouederosTp [eoT3sTqeqs ey. UT seBueyo pesse,t (cg ‘seer YserequT Jo TeAeT ey. (¥ ‘BuTpueqsqno AoueraZnd “sn

ult se8ueyo (¢ ‘sto03qgep OGT wory AYSTTJ Teqatdeo Jo eanseow e (7 ‘smoTF

- 38a -

Table 2 Regression Results

Dependent variable: Level of the statistical discrepancy Sample period: 1972Q1-1990Q4

Explanatory variables Coefficient T. stat. R” Constant 4.0 4.18 .03 U.S.-foreign interest rate differential 2/ -47 72 Exchange rate changes 3/ 41 1.70 Dependent variable: Level of the statistical discrepancy

Sample period: 1974-1989 (annual data) 2 Explanatory variable Coefficient I. stat. R Constant 4.7 1.06 .29 Capital flight 4/ -96 2.65 Dependent variable: Level of the statistical discrepancy

Sample period: 1970Q1-1990Q4 2 Explanatory variable Coefficient T. stat. R Constant -1.2 . 83 15 Change in U.S. currency outstanding 5/ 1.86 3.89 Dependent variable: Level of the statistical discrepancy

Sample period: 1970Q1-1990Q4 2 Explanatory variable Coefficient T. stat. R Constant -2.1 -.79 04 U.S. Treasury bill rate 6/ 73 2.18 Dependent variable: Change in the statistical discrepancy

Sample period: 1970Q3-1990Q4

Explanatory variable Coefficient T. stat. R- Constant 2 .32 -36 Lagged change in the statistical discrepancy ~.38 -3.65

RHO ~.38 -3.62 Dependent variable: Level of the statistical discrepancy

Sample period: 1973Q1-1988Q4 2 Explanatory variable Coefficient TI. stat. R Constant 3.1 3.43 -.01 Trade model residual -.27 ~,56

Data are in billions of dollars. The quarterly data on the statistical discrepancy exclude the seasonal adjustment discrepancy. All regressions were OLS except equation 5, where adjustment for serial correlation was necessary. Interest rate on 10 year U.S. treasury bonds minus the trade weighted average of rates on 10 year government bonds for the G-10 countries.

Percent change in the Federal Reserve trade weighted index of the value of the dollar against G-10 currencies ((I - Tp > x 100.

Capital flight from 10 Latin American countries and the Philippines. Equal to the gross external debt plus the inflow of net foreign direct investment minus the current account deficit, minus the change in external assets of the central banks and the commercial banks.

Source: Flow of Funds accounts.

U.S. Treasury bill rate - 3 month, secondary market.

“@ATSNTOUT 003 Aez ‘ATpoeqaTupe ‘sT eAnseow eAoge eyy, ‘3Seq Ae Azey ST AYSTTJ Teatdeo pue uoT_eoTFISAeATp oTToJAA10d wseemjeq SUTT eYyL “/€

UT J[Nset QeYyQ suOTJoOeSUrAQ JO UOTSSTWQ ‘“SQUNODDe [eUOTIeUTeQUT eY4Q worz AdUeTAIND ‘sn jo sSuTprToy u8tez0jJ ut seseerout sjtwo vag KOWSTIND (¢

‘peqqtwo o1e

0661 OJ BIeP JT QueoTFTUSTS Qou st dtysuotaejerl styQ ‘Aeaemoy ‘{squnoooe

suOoTIOBPSUBIQ TeUCTJeUAEQUT “S‘Q 9YyQ UT AoUedeTDSTp TeOTIASTIeAS eA

pue PoOTTeUy UTIYT UT UOTIeTJUT Fo eQer oZereAe 9yQ usemjeq dtTysuotAeyjesz

qUBOTJTUSTS eB ST eT0UT *“AODUeTAND TeooT AOF AoUSeTANSD ‘S'p Jo uoT{ANITIASqns

eyd 02 speeT YOTYM ‘uOTIeTJUT BuTQereTeo0R Aey_eA Anq ‘TeTONIAO

ST 2eUQ ,IUSTTZ Teqatdeo, ou st at sdeyxzeg *peqeyoorzAys AouederoOstp

Te°TasT3eas eyQ eTTYM peuTToOep ,AUSTITJF Teqatdeo, Aeyq eze suoTAeoOTpUT

AzeutwtTerd ‘aTqe[tTeae OU ST Q66T UT ,IUSTTF [eatdeo, uo eqep

aqeTduos ysnoyATe ‘AeAceTOW “SMOTJUT TeITdeo *s'q peproodezun ut dn pepue

SeTAQUNOD 9seyQ WOT ,AUZTTJF Teatdeo, Jo Ae[T Top Arevae Jo sjued ¢GE Anoge

2eyq ButAtdwt ‘y8ty ATqtsne,Tdut swees queToTgZyeoo ay ‘AeAeMOH *“sSqUNODDR

SUOTQOPSUPIQ TBUOTIFUAEQUT *S’'q eYyQ UT AoUedeZDSTp [POTIASTIeIS eY4Q pure

seTijunod esey Worz ,IAUSTTJF Teqatdeo, usemqeq dtysuotqej[er queoTyTusts

AT Teotast3e3s eB 3q 09 zeedde seop erzeyy, ‘“ATAeqazenb ueyA AzsyQeI

[enuue ore eleEp oy. esnedeq suUOTSserZeA TeyQO eYyQ UT esoyQ 09 eTqeareduods

qou st uotTqjenbe sty ut za aul pe Sse [eToreumuios eyj pure yueq [eAQUeD

ay. JO sjosse [eUAEQKe UT eZueYyo eyQ snuTW ATOTJep Qunoosse Quezanod eyQ

SNUTW JUSUQSeAUT 2OeATp USTeA0F Qou Jo moTJuT ey sn{d seTaqunod eseyQ Fo

Aqep [eUreQxe ssoiZ ey. 02 TeNbe se pernseow ATepnazaod st ,IYsTTF Teqatdeo,

‘seutddt{ttud eyQ pue eotAouy ute] Wory AYSTTZ Teatdeo Jo eAanseeu suo

0q Aouedezostp [eotastjzeqs eyQ seqetea (Z uoTqenbe) uotTsseiZer Qxou sul,

- 4O -

increased foreign holdings of U.S. currency would contribute to the net statistical discrepancy if the other side of the transaction were recorded in the accounts. For example, net shipments of currency abroad by banks would contribute to a positive discrepancy because the currency shipments are omitted from the accounts, but the payments for the currency would be recorded. Transactions that are omitted on both sides would not contribute to the net discrepancy (e.g., cash payments for drug imports).

A sharp increase in net currency shipments abroad by banks appears to explain part of the very large statistical discrepancy in 1990, 38 Economic disruption in Eastern Europe, the Soviet Union, Latin America (particularly run-away inflation in Argentina), and the crisis in the Persian Gulf all probably contributed to foreign demand for U.S. currency. However, since the total increase in U.S. currency outstanding (excluding bank vault cash) in 1990 was only about $22 billion, factors other than increased foreign holdings of U.S. currency must also have played a significant role in explaining the growth of the statistical discrepancy to $64 billion.

Comprehensive historical data on currency shipments abroad by banks are, unfortunately, unavailable. It is therefore not possible to assess directly the contribution of such currency shipments to the statistical discrepancy over a longer period of time. An indirect approach, using data on total U.S. currency outstanding, was tried

instead.

38. This information is based on informal discussions between FRBNY and certain banks.

"asn 09 eIep eqeTidoadde oy o1e eQep peasn[pe AT[euUosveS 9Y3 usey. ‘puewep “s'p UT sUoTQeNQONTJ sqQoeTzor AoueATAND ut uzreqQ qed [euosees ey. ey peumsse sT AT JI ‘_Y eyQ seonper oOsTe pue T MOTeq AQUeTOTJJe0o0

au. jo ezTs a3 soofiper SuTpueqsqno AouezAno Az0F ejep pogqsn[pe

AT [euosees-uou BuTSH ‘QUueOTFTUSTS os[Te pue AeZ1eT UsAe ST JUETOTJJI00 AouezAINd 9yQ ‘eqQep psouereFJIp-AsATJ Butsn una are suotsseiZe1

aud JI ‘QueoTzTUSTS ATTeOTAISTIeIAS TITIs pue G‘[T ST JUeTOTFJI00

ay. 4nq ‘pooS se jou st ATF oyy pepny{oxe 21P OSGI AOF SREP USM ‘GE

uoTjenbs) saqer 4seteQUT *‘S$'n JO TeAeT 9yQ pue AouedeTDSTp [eOTASTIeIS eyQ ueemjeq dtysuotjelter queotTsTusts e aq 03 Aaevedde ssop ere, ‘eSOL SOQPT JSeAeQUT SB eSeeTOUT 02 pUeaq PTNOM PpaeATOAUT squUNOUe ayy ‘peqeutysetzepun or9M peorqe sjesse “S$'p V8Snedeq sQUNODOe [eUOCTIeUASQUT "S$' 9UQ UT peqeUTqsezepun Zuteq e1eM sUWIODUT JSeATeQUT JI ‘ewooUT queuqseAUT OTTOFAA0d ATeYyA eQeWTAsetepuN seTAQUNOD AUeW QeUQ peqse3Zns Aouedeito0sTp Aunoooe AuerzAnd TeqoT3 sya Fo Apnas AWI eUuL Tease, oFerT FsorsUT “4 ¢ Aouedezostp Teotastqeqas eyQ 09 pegnqTaquos aAeY AZUSTW ZeYR SA0QDPJ ASYIO YATAM pepToUuTOoO eAeYy BuUTpUeASqno AoUerINO "S$'h UT SquUemweAow QeyuQ BSuTAseB88ns (suo ueYyQ 39eQve78) YSTY ATqTsne,dwt sU9e9S JUSeTOTJJIOO OYA ‘AeAeMOH ‘BSuTpueqsqno AouezAINO ‘S‘n UT sesueyo pue Aoueder0stp TeoTAsTIeIAS |YyQ UseMjeq dTysuOTAeTeI QUeOTJTUSTS e eq 03 azeedde svsop ezsyQ ‘¢ uoTjenbe Aq poeqeotput sy ZAouedexzosTp [ePoOTIASTIeIS ayy FO TeASeT SU YITM pe_geTeti0o. BZuTpuejasqno AoduerAno ut sesueyo aie ‘puooseg ‘AouezAINO ‘S'p AOJZ puewep usTetT0F uT seseero0uT azoyz Axoad poo8 e eq 07 avedde jou svop AouederzostTp [eoTasTqeqs eyy_ ‘“AerTNoTQAed UT STITG OOTS$ pue [ereues ut AoUeTANO AOoZ puewep 9syQ ZuTuTeTdxe sTepou wolz STenptTsez 9yQ pue AouedetosTp [eoTAsTIeIS BY} USeMQeq UOTIeTeAI0D queotTjTusts ou oq 03 saeedde orey, ypuewep utTetdxe djtey Aouezano ‘*s'n 1Oj pueuep Zututetdxe ~Tepow e 09 AouederosTp [eoTIsTIeIs ey BuTppe seop

‘asaTq :seaTqoedsied om3 worzz peyororidde eq ued uoTjsenb styl

4). These regressions results appear to support the contention that BEA

. . . og : 40 is underestimating portfolio investment income.

5) Lagged changes in the statistical discrepancy

The statistical discrepancy is highly volatile from quarter to quarter. As indicated by equation 5, changes in the statistical discrepancy in one direction tend to be followed by changes in the opposite direction in the next quarter. Timing problems (i.e., the recording of one side of a transaction in one quarter and the other side in the next quarter) probably contributed to this result.

6) Residuals in a trade model

If there were substantial quarterly swings in the errors and omissions in trade transactions, there would be no reason to assume that such swings would be mirrored in trade model estimates, and the model residuals would reflect these swings. As indicated by equation 6, there does not appear to be any correlation between the residuals of the model used by the U.S. International Transactions Section of the Federal Reserve Board to project the partial trade balance and the statistical discrepancy in the U.S. international transactions accounts, These

results suggest that the wide quarterly swings observed in the

40. The estimated coefficient implies that if U.S. short term interest rates were 100 basis point higher for a year, the statistical discrepancy would be about $3 billion higher for the year. Another inference that can be drawn from these results is that BEA is underestimating U.S. net interest bearing assets abroad by about $300 billion and if interest rates averaged about 7 percent for the year, the contribution of underestimated interest receipts to the statistical discrepancy would be about $20 billion. Given the standard error of the coefficient, these estimates are undoubtedly subject to a very large margin of error, but they do suggest the importance of improving the estimates of U.S. portfolio assets abroad and the income earned on these assets.

41. The model residual equals the model estimate minus the recorded

partial trade balance (the value of nonagricultural exports minus nonoil imports).

"0661 UT PpezZT[eerl sem uy. ATOTJep eperzq azeBaey, ATQueoTJTUsTS e peqotperd szeas qyd ey Aq pouTequTeU [epou syq ‘ORF UT “Ev ‘AouederosTp [eoTAsTAeyAs 9yq 03

aqnqTiqUOD ATperD [TeToOAemMI0D Jo uoTsUeyxe oy) 10 spood ssoyQ AF Quewked

jo BuTpioo.er |yy UT sweTqorid 3utwtT, ey suTewer AQT[IqISssod aul “ZH

4 OFPOTPUT eIep peprzooer sy ueYyQ AesA7eT, Aue sem souRTeq eprAq ay UT AuoueaordmT peqoedxe ey qeyQ yseZ3ns jou op sTepou pejeUtqse uo peseq eoUueTeq epely esTpurYyoreM ay UT QuoeweaorduT Jo suotqjoeford ‘peqexyoorkhys Aouedetostp [eoTAsTqaeqs syQ uUsyM Q66T UT pedunf{ ATueppns

spoo3 Jo sjiodxe Jou peqqTwo QeyQ eumsse 0} UoSeeX AUP 919 ST ION ‘zeeToun st Aouedei0sTp TeoOTASTAeAs ay UO QUNODOe epelzy ssTpuFYyoreU ey uy SUOTSSTWO pUe S1OIZe AeYyQO pue sseYy. Fo AoeTJe Jou ey, ‘“pesPetOUT sey apelq se OWT JeAO eZTS oAnTOsqe UT uMOIZS ATqeqord sey AT Anq ‘weTqoad Mou B OU ST Ssqtodxe Jou Fo UoTIeUTJSerepuN ‘sqioduT uey A per0jTUOM Ay[Nyerzeo sseT ere sqiodxe ‘puey reyjOo eyQ uQ “TeTIUeASqns ATpejqnopunN ‘aze pue AouedetosTp [eoOTASTIeIS 9YQ OQ UOTANQTAQUOD aeATQeZOU Pe

ayeul sooueqsqns [eseT[T Jo sqizodwt ‘epts qunoooe Querzino eyQ uO ‘Aourdet0sTp [eoOTAsTAeAS Jou sYyQ UeYQ JesreT yonu e1e squnoooe ey} UT SUOTSSTWO puke Siorte ssoiZ QeYyA ATEXTT OsTe ST AI ‘squnoooe [eqIdeo pue quezANo eYyW YIoq UT sUOTSSTWO puke Ssi0I1e

Jo A[Nsez ey st AouederOSTp [eoTAsTAeIAs YSTY ey. Jey ATeXTT swees AT

‘sqseq [eBoOTISTIeIS SNoTAeA puUe sqUNODOe suUOTIOFSUPAQ [PUOTIeUAeIUT “Sf ayq 10F seoAnos eqep sy UT seToenbepeuT Jo MeTAeT eB UO paseg *saeo0k

queded ut eTTae{oA ATYSTY pue ere, ATeA uVveq sey SqQUNODOe suUOTIOPSUPIA

[euoTjeureQUT *s'n ey ut AouederoSTp [eOTASTIeIS SUL SNOISMTONOD “A 4 SP008 Jo sqaodxe ato sqazodut jo Zutjiaodez ey uT suoTssTuo

pue sioize Jo A[Nseaz eyQ you ATqeqoad ore AouederosTp TeOTISTIeIS

Trade in services is much more difficult to monitor than trade in goods, which can be observed as goods enter or leave the United States. BEA has made substantial advances in improving its estimates of U.S. receipts and payments for services in the past decade, but holes remain. It is not clear, however, whether improving these data would add more to receipts or to payments.

In contrast, both alternative data sources on U.S. assets abroad and the statistical tests of the relationship between the statistical discrepancy and U.S. interest rate levels suggest that net portfolio investment income is substantially understated in the U.S. international transactions accounts. However, because the level of U.S. interest rates did not rise sharply in 1990 on average, omitted portfolio investment income could not have been an important contributing factor to the rise in the statistical discrepancy between 1989 and 1990, "4

The sources of the big increase in the statistical discrepancy in 1990 should be sought in the capital flows data, and perhaps in unilateral transfers. One obvious source of the discrepancy is the omission of estimates of increases in foreign holdings of U.S. currency. In 1990, as in other periods of economic and political turmoil abroad, foreigners turned to U.S. currency for transactions and as a store of wealth. Inflation in Argentina, economic disintegration in the Soviet Union and parts of Eastern Europe, and the crisis in the Persian Gulf all

contributed to the demand for dollars. However, given the total increase

in U.S. currency outstanding in 1990, increases in foreign demand for

44. Interest rates in some foreign countries did rise, but since only a small part of U.S. claims on foreigners is denominated in foreign currencies, only a small impact on omitted interest receipts would be expected.

sqjiodeit 9-DIL eTTJ 03 sA0QSeAUT QOeATP USTeT0J Jo saAeT[IzZze Yuequou

"S°A FO sANT Tey eyQ useq saeY AYSTU Q66T UT AoOUedeTOSTp TeOTASTIAeAS 9A ut esta daeys ayq ut etTor e pedketd sary Aew AeYQ AOROPJ ASYQOUYV

"0661 ut AouedetoOSTp [eoTasTjaeas “S‘n ea

Sututetdxe ut qu3sT[TF Teatdeo so soueqiodwit ayq Jo jueuwssesse aatjeqtTjuUeNnb

@ BATS 09 Q[NOTFJIP AABA 9q P[NOM AI ‘pepzAOdeAUN st sAezTsueIR [erEQeTTUN

uo MOTJUT BSuToueTeq eyQ eTTYM sreUBTeAOF uo squepTset ‘S$’ JO swTeTO UT

BSPOTOUT Ue pur SqUepTSserT UBTeAOJ 09 SOTATTIGeI[ pepzooer UT uoTIAONper eB

YATM peqetoosse oq 09 ATOAXTT ST AT esneoeq AouederzOstp [eotqastTjzeqas 3yQ oF

@ANGTAQUOD PTNom see js peaTUNQ eyy OF sTeNpTaTpuT AyATeeM Jo uoTIeASTUMUT

‘pezanooco eaey ostTe Aew Teqtdeo AtTeyA se [Tem se suoszed AyATeOmM

JO AUZTI™ ‘peseezouT ovary Aew Buoy Buoy Jo eAngAny oyQ Anoqe AqQutTeqrzeoun

Zo ‘edoing uzeqseq uT Aserzun ‘sTsTi9 JIN) uetsizeg ayQ Aq perzands

SMOTJ ‘Q661 UT eseeTDUT Ue MoUs Jou [ITM ATqeqoid AeTTAee pequesead

seutddt[tyg eyQ pue eotrowy uTIeT] worAF ,IYSTTJ [Teatdeo, Jo eianseouw

eyd eTTyYM ‘weqsks SuTQAiodez ayQ Aq pessTW erOFerTeyA pue ‘sjuepTsel

‘S$’ 212 OYM spueTAZ AO soeaTAeTEeT YBnorzyQA spew oq AeW sQUEUIQZSeAUT

queoltgtugts ‘ATeaTQeUAZeAaTY ‘saqaodei moTZ Teqrdeo eyQ UT pereaod

ATeqenbepeul st 4eyQ vere ue ‘eqeQSe TeeT UT peqseAUT aq 03 pus squUNoOUe

queotytusts sdeyzeg ‘eseo ayQ useq oaey 03 savedde qeyQ ATT POTIOIASTY

qnq ‘smoTz [Teatdes ‘s'g Az0F weqsks Butqazodexz eyQ edeose p[nom Aeyq suat0j

ayeq ATTAessvoeu pTNom AYyZTTF Teqatdeo AyM AeeTO Jou st AI ‘seQeIS peaTug

ey. UL UsaeYy eFeS 09 SWAOZ ATOYyQO UT TeATdeo Jo AYSTTF ey. YAITM poeQetoosse osjte ATqeqoid st peorzqe [TowAznq TeoTIATTod pue oTwouo0Dy

“TTEms

SP WUPOTJTUZTS useq ovPY Asnu sZOQOPZ AeYyQO ‘AouedetOSTp TeOTISTIeIS

ey JO UOTTTTG OZ$ O29 STS ‘3Sow Qe ‘uTeTdxe prtnoo ATqeqoad Aduezanso “s'h

- cv -

- 46 -

indicating their liabilities to unaffiliated foreigners. These reports are probably the weakest link in the TIC reports system; the number of reporters is small and the attention devoted by’companies to the accuracy of these reports is minimal. Capital inflows from foreign direct investors to their U.S. affiliates dropped from $71 billion in 1989 to $37 billion in 1990; however information available on acquisitions of U.S. companies by foreigners does not indicate a precipitous decline, raising the question of how these takeovers were financed.*> While many acquisitions were funded by borrowing by affiliates in the United States, borrowing from financial institutions outside the United States also may have played a role. Much of such borrowing is, in principle, reportable on the TIC-C forms, but if omitted would contribute to the statistical discrepancy. Given the widening spread between prime and libor at the end of 1990, increased borrowing in the Eurodollar market would not be surprising.

The wide swings in the statistical discrepancy from quarter to quarter seem more likely to be the result of errors and omissions in the recording of capital flows as well. Timing problems in the recording of direct investment flows or the counterpart to highly volatile bank flows probably contributed to these swings.

A. Recommendations

1. Renewed and sustained commitment to improving the data on

U.S. international transactions is needed. In many cases, this would

involve increasing the budgets of the agencies responsible for data

45. Starting with data for 1991, BEA’s annual survey of U.S. business enterprises acquired or established by foreign direct investors (BE-13) will provide more information on sources of funding.

46. If the U.S. office of a bank played some role in the loan, the reporting responsibility would rest with the bank and not the borrower.

*sqiodez OIL eyQ UT sweqzTtT 1AsyAO0 YATA peqeseisZe ere Aeyq ‘ieded [TetozemMMI0D se Yyons ‘squeumAqsuT w10Q AeqAoOYS UO UOTIeWAOJUT epnyoUT p[noys sAeAAIns yxAewMYoUSeq eYyQ ‘12eACaTOW ‘QWOoUT JuUeUIQZSeAUT OTTOJFAA0d Fo sejeUTAse s,V_yq OF yAeuyoueq e epTAoad 0 peustsep eq ptnoys sAeains yAeUOUeq pAreMQNO pue pAeMUT eYQ ‘sBUTpPTOY uo uoTjeWAOJUT ButTptaoird 03 uocTITppe UL ‘“peorge QueuyjseauT oT[oF QI10d "$'g Fo AoAANS YyAeWYOUSG eB aXxeIAepuN p[Noys AAnseer] sul ‘+ *suzoz eseyQ uo pejriodexz ATQUeTAINS UOTIeWAOJUT eWlos ATOZ (eIeP SI_ JO sATESOYy [elepeg ‘°3°9) seornos ejep eATIeUTEQTe eqanqTasqns 09 sqZATOzJe 10/pue Quewertnbexr Butqiodex eyq yaTM soueT{duos eaoriduT 02 QA0FFe SnoTies eB sertnberz (sqaodez 9-9IL) sreuSTeATOJ peqet[tTjyeun yIIM syuequou “s'n Jo suoTjoRsuety oy Jo BuTq10dex1 eqenbepeut ‘ze[noT_aed ul “¢ ‘sqZeyreU TeTOUPUTJ BUTATOAS USsATS paejoeTTOO 2q seq P[noo e2ep yons Moy ssesse oj pue stsATeue AotTTod pue sesodind squowfhed jo soue,teq toz Azessodeu are BRep Fo SpUuTy WeYM AepTsuodeT 02 pssUu 3M ‘szezk[eue pue sz0VOeTIOO eQep QuewureAOD’ se [Tem se squedytotqzed QexreEW [eTouRUT] pue soTWepeoe Jo sMeTA ayy UO ZUTMeIP ‘pepseu sT smoTZ [eITdeO oT[ToFz10d sozF weqsAs UoTAoOeTTOO BIep ey} FO MeTAeA OTSeq V “Z ‘ATQueotjtTusts dytey prnom seToueSe [elepey ssoioe uoTOeTTOO BQep ZuTIeUTpI009 pue ‘seTqTs0TAd ZuTUSTTqQe4se ‘soTasTqeqs [erepez Jo AjtTenb oy Butaorduy 02 uoT eoTpeper SI pue ‘spiepueqs pue AdT[od TeOTASTIeIS TerepeT JO eoTFFJO 942 FO [BATASA oy ‘UCTATppe UL (‘Vad TOF JseeT Je ‘uoTIOeATP STY? UT ssei30id aWOS UT pejt[nser sey yAoder ee AQTMMOD UTYSOG SeYyL) “UOTIPOTFTASA BQepP OF seoanosei ejenbepe erow ejoAep pue ‘spoyjew Sut ewT4se eaocaduy ‘s19qI10de1

Zutsstw AJTQUepT ‘eqep Qquezino ut sdeB [TTF uero Aeya os uoTQVeT{[OD

5. Obvious gaps in the international accounts should be closed. Estimates of increases in foreign holdings of U.S. currency should be included, based on improved Treasury CMIR data and/or a survey of banks involved in incomng and outgoing currency shipments. Problems in the coverage of real estate transactions should be tackled.

6. Efforts by Commerce and Customs to improve the merchandise trade data should continue. Periodic sample audits of shipments through specific ports are useful to identify problem areas, as are efforts to reconcile trade data with other countries.

7. Miscellaneous recommendations to BEA would include a) improving the method used to estimate dividend and interest income on holdings of securities, b) estimating interest receipts and payments on foreign-currency denominated bank deposits, and c) periodically collecting information on the currency composition of accounts receivable and payable between direct investment affiliates and their parents.

8. Miscellaneous recommendations to Treasury would include a) instructing TIC reporters to use tax identification information rather than address to identify foreigners, b) collecting information periodically on the currency composition of banks’ foreign currency denominated claims and liabilities, c) improve reporting of banks’ writeoffs of loans, and d) disaggregate reporting of foreign investments in limited partnerships from purchases of equities and study alternative

ways of collecting data on limited partnerships.

"0661 ‘Og u0QZuTYseM ‘“Seanpesorg SUTJEUTISY pue “Sooanog ejeq Sjdoou0,) .Sojeq9 ped}Tuy oy. JO SqUSWAEY

Jo souerleg oyL ‘stsAyTeuy oTwouocg Jo neeing ‘aozemiog Jo quewjiredeq “s'n

‘ZuTwooyj10F ‘sseig AATSAeATUQ OZeoTYD ‘oZeoTYD

‘YOaeesoy [eolTAtduy pue jUdoANnSeoy Ul Sonss] .SUOTJOeSUET] OTWOUODF JeuoTIeUAS JU, (‘Sspe) uOospzreyoTY ‘f pue aedooy ‘gq ut ‘,eQ2eq JuUeUQseAUT q0e1Tq °§*n jo Aoenbepy oul, ‘sueaeqas *5°a Any pue “gq sTOT ‘19THxOIS

"6861 Azenaged ‘WW ‘e8ptaquey ‘gz ‘ou zedeg Butyz0m YAGN ‘,UOTIJeUTPI00D AdT[Og AOF eRe uoTATSOg pue suOoTjJORSUBAY [ePUOCTJeUASQUI FO Aoenboepy, ‘'q SToyT ‘29T4e9S

"L861 ‘od uoqZuTysem ‘(WI ‘SoeoueTeg quNnoooYy JueTAND pptzomM UT Aoueddaostqd Leorastaeds

ayy uo Ajieg SUpTAAOM SY) JO JaAOdeyg [euTY ‘puny AreQouow TeuoTJeUurEqUL

‘od uo jZuTYysem *(066T 72equeades) 6 Atequmu o/ eunto, ‘SseuTsng quezing FO AsAANEg ‘,SeoTAreg Jo seseyoand

pue seTes [euoTJeUTeqUI “S'Ny ‘PAeYyoOTYM eTqGO pue “f Auoy UY ‘oOTINTTC ‘Zutwooyji0y ‘sserg AqTsaAeaTug o8eoTyD ‘oZeoTyQ ‘Poressey TedTArduy pue JUsWsinseoy UT Senss] -SudTJoOeSUeAL OTWOUOCY TeuoT eure {UT

(‘spe) uosprzeyoty “c¢ pue rzedooy *g uT ‘,SeoTATeg Jo seTeg [euoTIeureqUT toy weqsks ejeq e BuTdojeneq, ‘preyoTymM eTtqo pue preureg ‘1eyoOsSy

Sodouer0 Fog

IFDP NUMBER

404

403

402

401

400

399

398

397

396

395

394

393

392

391

390

- 50 -

International Finance Discussion Papers

TITLES 1991

The Statistical Discrepancy in the U.S. International Transactions Accounts:

Sources and Suggested Remedies

In Search of the Liquidity Effect

Exchange Rate Rules in Support of Disinflation Programs in Developing Countries The Adequacy of U.S. Direct Investment Data Determining Foreign Exchange Risk and Bank Capital Requirements

Precautionary Money Balances with Aggregate Uncertainty

Using External Sustainability to Forecast the Dollar

Terms of Trade, The Trade Balance, and Stability: The Role of Savings Behavior

The Econometrics of Elasticities or the Elasticity of Econometrics: An Empirical Analysis of the Behavior of U.S. Imports

Expected and Predicted Realignments: The FF/DM Exchange Rate during the EMS Market Segmentation and 1992: Toward a

Theory of Trade in Financial Services

1990

Post Econometric Policy Evaluation A Critique

Mercantilism as Strategic Trade Policy: The Anglo-Dutch Rivalry for the East India Trade

Free Trade at Risk? Perspective

An Historical

Why Has Trade Grown Faster Than Income?

_

Please address requests for copies to International

Papers, Division of International Finance, Federal Reserve System, Washington, D.C.

20551.

AUTHOR(s)

Lois E. Stekler

Eric M. Leeper David B. Gordon

Steven B. Kamin

Lois E. Stekler

Guy V.G. Stevens Michael P. Leahy Wilbur John Coleman II Ellen E. Meade

Charles P. Thomas

Michael Gavin

Jaime Marquez

Andrew K. Rose Lars E. 0. Svensson

John D. Montgomery

Beth Ingram Eric M. Leeper

Douglas A. Irwin

Douglas A. Irwin

Andrew K. Rose

Finance Discussion

Stop 24, Board of Governors of the

uueWe “J suTASYAeD

eUuTAeT ssoy

qedee] “HW oTag

uueW “7 euTreyje9 uossoTag “Y TION zenbiaey outer epeoW “A ueT Ty

uouse5 ‘gq ydesor Aepuexely “Ss sTMey

uosiepuey “mM eTeg uosTpad “f£ TIPH

SUTAST ssoy

aeddog ueTey uOoSSOTIG “Y [TEN Azpuey “gq ptTaeg uTMAI “Y seT3noqg key ‘pf Aeups

aeddog ueTeay

AOTTOWN “Y WETTTIM uoseTYuIeW “Cf prTeuog

esoy “y Merpuy pooTg “gq 31eqoy

te euy “W TeeyoTH uousey ‘q ydesor

(5) aOHINV

asuey) AoT[og snsiaa sesueyg [Teanjonaqs :9oUuRTeg [eure XY ‘S “UT Jusuleaoidwy peuteqsng z0z sqoedsozg

AOTTog pue ‘yamory ‘sqeyzeW yooIs

Tepow OTAseyoors e UT seTny xel pue a3eYy 4sereqU] Jo soqweudg eu

303995

SuTinqoejnuew ey. pue ADT [Og OTWoucse01 eR TOF SOTOY PUL :seTwouccy BuTztTerazqsnpuy ATMOeN JO uoTRIeA9UeD QxoN ey. sprzeMol

STepoW JUNoooYy-epeazy jo eourUAOFIeg SATIOTperg eyQ ButTjenqteag

SeoTig SUTTTe@y eya Jo asey PUL -3TOTJed epeazy ey. pue s1zeqnduo)

uoTIeOTJTUQ UPWIZED JO suoTIeoTT dW] oTWOUCDY TeqoTD aul

AydeaZotTqtg :suotqeiedg Aoduerzing ustez04

pereqjouuy uy queudoTereq oTwouosq pue einqonaqg [eroueuty ueder pue ‘Aueuzey

*S028I$ peqITUy ey Jo sqoyITeW eroysuO

eyi UT sejeYy AserequI Jo eanjonazqg warez sul

$97e7§ peItun ey. pue wopZuTyY peqtup eu ut AeuoW MorIeN OJ pueuag ey. BUT [epoy

yororaddy ofzQeU0ey ‘Teng y *STOpow wUNTAqTTInby Ter9eueDg 10QOey-eerUL

soTIoeIg pue Artosy]L ut AdTTog

“S'N é9TGPTA TIT3§ Juowjeezy TeuotqeN sj sdemg Aouezang wie]-3uoc] wosz souepTaqg ‘:AITTTGOW Teatdeng [euotjeuzequy

AATTTGQeIS

pue 9injoni3g = :Yysnoiryj-sseg But eWTIS|

SeuoZ 3e31e], eeYy eSueYyoxy Jo uoTqer0Tdxy Teotatdug uy gxTq Aoezreg ou SHA eu sT essTpueyoreN [eqIo]L pue

SeTTqowoqny uo ejeq Teueg WorzZ doUEPTAY

:eperzAL TeucTeUAEQUI UT JeyAeW 09 BuTOTIg

O66T SHILIL

siedeq WOTSsnosTq SoueurTy TeuoT eudejUuy

tLe VLE

SLE

OLE

LLE

BLE

6LE

O8E

T8€

C8E

€8E

VBE

S8E

98€

L8E

88E

68€

Yad WAN dda