Internal Capital Markets and Investment: Do the Cash Flow Constraints Really Bind?

Internal Capital Markets andInvestment: Do the Cash Flow Constraints Really Bind? Calvin Schnure” July 1997 ‘$EconomistF,ederalReserveBoard.IthankJamesBohn,GeorgeFenn,KevinHassett,SpencerKrane, DeanMaki,SteveSharpeandKarlWhelanforhelpfulcomments.Theviewsexpressedinthispaperarethoseof theauthoranddonotreflecttheviewsoftheBoardofGovernorsorthestaffoftheFederalReserveSystem. Email:mlcdsOl@frb.gov.

Internal Capital Markets and Investment: Do the Cash Flow Constraints Really Bind? < Abstract Lament (1997)claimsto find evidenceof creditmarketimperfectionsthatdistort financingandinvestmentdecisionsof asampleof oil-dependentfirms, asinvestmentby nonoil unitsfellwhen oil cashflow dropped. However, asimpletestrevealsthatfew of these firmsbehavedin afashionconsistentwith binding cashflow constraints. In addition,most werecashrich. The dataprovide strongevidenceagainstthehypothesisthatinvestment decisionsby nonoil unitswere significantlyaffectedby oil cashflow, or thatcreditmarket imperfectionsareanimportantfactor for thissetof firms.

Internal Capital Markets and Investment: Do the Cash Flow Constraints Really Bind? A lively debatehastakenplaceintheliteratureon the relevanceof cashflow constraintsfor businessinvestment(seeFazzari,Hubbard andPetersen(1988);Hoshi, KashyapandScharfstein(1991)). The existenceof sucheffectscould havewide-ranging implications,from macroeconomic issueslikethebehavior of investmentoverthebusiness cycle, to theefficiency of financialmarkets,to distortionsin firm behaviordueto agency problems. While some claimthatsuchconstraintsarelimitedto smaller,less-well-known firms (seeGilchristandHimmelberg (1995);Whited (1992)),othershavearguedthateven largefirmswith accessto public bond marketsmayfacesuchconstraints. Takingthelatterview isLament (1997),who findsacorrelation betweencashflow from oil subsidiariesandinvestmentby nonoil unitsof 26largediversifiedcompanies. When oil revenuesfellfollowing thecollapseof oil pricesin 1986,investmentby subsidiariesin industriesunrelatedto the oil businessalsodeclined. This, he argues,isevidenceof imperfectionsin creditmarketsthatdistortfinancingandinvestmentdecisionsby thesefirms. The criticalassumptioninthispieceisthattotalcashflow of thefirm fellsharply, prompting managersto cutinvestmentby nonoil segmentsby more thanthey would havein the absenceof the oil price decline, astheconstraintimposed by totalcashflow became binding. Thus, itwould be well worth examiningdirectly iftotalinvestmentof oil andnonoil segmentswasrestrictedby totalcashflow (Lamentdoesnot); firmsthatdo not meetthistest cannotprovide evidencein supportof the mainhypothesis. Interestingly,only aminority of thefirmsin hissamplemeetthisbasic--andnecessary--condition. Furthermore,many of thefirmsthatdo meetthecashflow condition actedin other waysthatstronglysuggestthat,farfrom beingconstrained,they were actuallyflushwith cash in 1986. Examplesof suchcharacteristicsinclude balancesheetsso cash-richthatmanyfirms heldmore thantwice theamountof cashandsecuritiesrelativeto totalassetsasdidthe medianfirm in similarindustries,firmsthatrepurchasedhundredsof millions of dollars worth of common stock, andlargeincreasesin common dividendspaid. Thesefindingscast \ seriousdoubt on thevalidityof Lament’s results. The nextsectionreviewsthe “internalcapitalmarket”argumentandexaminesthe 1

condition thatLament’s hypothesisimposeson therelationshipbetweentotalcashflow and investmentby the entirefirm. Then, Ipresentdatashowing thatamajority of thefirmsinhis samplefailthisbasiccondition. Iproceed with otherindicatorsdemonstratingfew firmsin thissamplecould havebeenfinanciallyconstrained,aswell assome otherweaknessesof the study,before concluding. -.. I. Cash flow and investment. Lament’s argumentisasfollows: diversifiedfirmssubsidizetheinvestmentof nonoil unitswith cashflow generatedby othersegmentsthrough the “internalcapitalmarket”. They do so becausecreditmarketimperfectionsdrive awedgebetweenthecostof internalfunds andexternalfunds. As aresult,thedifferentcorporate segmentsarefinancially interdependent,so thatfinancialshocksto one segmentimpactthecostof fundsin other segments. According to thisargument,when oil pricesfellin 1986,thedeclinein cashflow generatedfrom the oil segmentsof thisgroup of firmswasso sharpthatthey hadto curtail capitalspending. Facedwith theneedto balancetheircashinflows andexpenditures, companiesstoppedsubsidizingnonoil unitsandcutcapitalexpendituresacrosstheboard, including investmentby nonoil segments. Beforeturningto theinvestmentbehaviorof nonoil units,itisworthwhile verifyingif cashflow constraintsdid bind for thefirm (asopposed to industrysegmentswithin thefirm).l The overallfirm isthe appropriateunitof analysis,asinternalcapitalmarketsoperateby linkingsourcesandusesof cashwithintheentirefirm. Supposethereisadesiredinvestment level1’,,determinedby thenetpresentvalueof theexpectedfuturecashflows of thefirm’s investmentopportunities. The firm mayfacecreditmarketimperfectionslikethosedescribed by Lament. Actual investment,1,,isgivenby ~ = min{I*,,CF,} (1) where CF~iscashflow inperiodt. Therearetwo periods (1985and 1986);theconstraintmay 2

bind in eitheror both periods, or not atall. Next considerthechangein investment,dI,, betweenthetwo periods. There arefour possibleoutcomes: A. Constraint binds in both periods. 1(.l= CF,l < 1*,1,and I,= CF, < I*,. Inthis case,the changein investmentequalsthechangeincashflow: dI, = CF,-CF,.l= dCF1 (2) B. Constraint binds att-1 only. 1,-1= cFtl < I“t-l,andCF, > It= I’to Thus,the changein investmentislessthanthechangeincashflow: dI, = 1*,-CF,.l< CF,-CF,.l d~ < dCF, (3) c. Constraint binds att only. CF,.l > 1,.l= 1’,.l, and 1,= CF1< I*,. Investmentfalls by lessthancashflow: dI,= CF,-l*,.l> CF,-CF,.l d~ > dCF, (4) D. Constraint never binds. CF,.l > ~.l = 1’,.l,CF, >1, = I*,. Inthisinstancethesign of therelationshipbetweenthechangein investmentandthechangein cashflow is indeterminate. dIt= 1“,-l’,.l? CF,-CF,-l (5) Lament’s hypothesiscorrespondsto thethirdoutcome, wherethecashflow constraint wasnot binding in 1985butbecamebinding in 1986. Cashflow exceededdesiredinvestment 3

in 1985,so actualinvestmentequaledI*. However, in 1986,cashflow fallsshortof desired investment. As aresult,actualinvestmentequalscashflow. For theconstraintto have become binding, cashflow hadto havefallenmore sharplythandid desiredinvestment,so dI, > dCF,. Otherwise,theconstraint(by definition)could not havebecome binding. Table 1showsthechangesin investmentandcashflow for the26firms in Lament’s --study. Most of thesefirmsfailedthisbasictestof cashconstraint,thatcashflow fallmore sharplythaninvestment: l Lament states: “thisstudyexaminesa...groupof firmsthatexperienceacash shortfall.” (p.85) However, cashflow actually rose for six firms. DuPont’s cash flow roseby more thanthree-quartersof abillion dollarsin 1986,yettotalcapital expendituresfellby $3o9million. Mobil Oil’s cashflow increasedby more thanhalfa billion dollars,yetitcut investmentexpendituresby $329million. Four otherfirms hadhighercashflow but cutinvestment. This group’s declinein investmentwasnot a responseto fallingcashflow. l One firm increasedcapital expenditures in 1986. Litton spent$33million more on totalcapitalexpendituresin 1986thanitdidin 1985,inspiteof lower cashflow. l Seven firms cut capital expenditures by far more than the drop in cashflow. Unocal’s capitalexpendituresfellby more thanfive timesthedrop in totalcashflow, asdid Kerr-McGee’s; Tenneco’s fellby nearlyfour timesasmuch, andAmoco’s by nearlytwice asmuch. Suchoutsizeddecreasesviolatethecondition implied by Lament’s cashflow constrainthypothesis,butareconsistentwith adrop indesired investmentin responseto lessfavorableinvestmentopportunities. l Twelve firms cut investment by lessthan the fall in cash flow. Thesedecreases would be consistentwith constrainedbehavior;however, giventhatmore thanhalfthe sampleevidentlycutcapitalspendingin responseto adrop in desiredinvestment,we cannot ruleout thepossibility thattheseotherfirmsweremerely reactingto less profitable investmentprojectsaswell. This comparison assumesthatothersourcesandusesof fundsareconstant. However, examiningtheimpactof otheractivitieswe find that,farfrom causinganyconstraint,they on netprovided additionalfundsfor most firmsinthissample. Table 2liststhe changesin 4

sourcesotherthancashflow andusesotherthancapitalexpenditures.2DuPont hadalarge jump in otherusesdueto sixacquisitionstotaling$1.2billion in 1986. Other usesfellby $2.6 billion atAtlantic Richfield Co., asthefirm reduceditssharerepurchases.While other sourcesfellfor allof thefirmswhose cashflow increasedin 1986,thedeclineswere smalleror roughly equalto theincreasesin cashflow, leavingtotalsourcesfor most of thisgroup higher --or aboutunchangedfrom 1985. Two firms,Royal Dutch ShellGroup andAtlantic Richfield Co., hadlargedeclinesintotalsourcesasthey scaledbacktheirborrowing in 1986. Similarly, Litton’s andUnocal’s drop in othersourcesisalmostentirelyattributableto slower debt issuance,andTenneco’s otherusesof fundsjumped in 1986dueto areduction in long term debtoutstanding. There islittleevidencethatotheractivitiesdrainedcashfrom investment for thesefirms. Forthefirmsinthefinalgroup, whose investmentfellby lessthanthe changein cash flow--thatis,consistentwith thehypothesisof bindingcashconstraintsin 1986-most had increasesinfundsfrom othersources,some of themsubstantial.The largestcatego~ of other sourcesof fundsisfrom increasesin long-termdebt. As will bediscussedbelow, many of thesefirmshadheavybond issuancein 1986. Overall,dataon othersourcesandusesof funds do not suggestthatfinancialconstraintstightenedin 1986. One could arguethatitisnot justcashflow thatmatters,butcashstocksaswell. However, consideringcashstocksfor thissetof firmsin 1986arguesagainstthe existenceof financialconstraints. For example,afirmthatdecreaseditscashholdingswould havehad additionalresourcesto spendon investmentin 1986,furtherloosening anyconstraints. On theotherhand,anincreasein cashstockswould provide evidencethatthefirm wasnot short on cash,especiallyconsidering (asisdiscussedbelow) thatmostof thesefirmshadrelatively largecashholdings comparedto firmsof similarsizeandindustry. II. Other possible indicators of financially constrained firms. Of course,one shouldnot look atinvestmentandcashflow in isolation. Do other characteristicsof thisgroup of firmssuggestthey werefinanciallyconstrainedin 1986? No. Infact,anumber of othermeasuressuggestthatthesefirmswerecashrich: 5

l High cash holdings relative to industry norms. Holdings of cashandliquid securitieson thesefirms’ balancesheetsin 1986roseby more than25percent,to $31.5 billion--acashhoardfor thisgroup exceededduringtheentire 1982-1994period only once, by the $31.9billion holdings in 1987(Table3). Over thisspanof time,in only , ,, two yearsdid cashandsecuritiesholdingsfor thegroupexceed80percentof total capitalexpenditures: 1986,124percent,and1987,113percent. Most firms’ cash holdingswere abovethe medianfor similarfirms,3 and10were inthetop quartileof .-. comparablefirms;only threefirmswerein thebottom quartile(Table4). Most of thesefirmshadampleresourcesto maintaininvestmentplansinsmallersubsidiaries. l Increasesin common dividends paid. Firmswith cashresourcesexceedingtheir investmentopportunitiesmay returnsome cashto shareholdersthrough higher dividends. In 1985,before the sharpdrop in oil revenues,the averagecommon dividendspaidby thefirmsin thisgroup roseby $9million ~able 5). Ifthesefirms becameconstrainedin 1986one mightexpectsmallerincreasesindividends. However, the 1986increasewasmore thanthreetimesaslargeasin 1985,averaging$30million. Moreover, firmscontinued to raisedividendsinsubsequentyears. Total dividends paidby thegroup soaredby 30percentover thenextfew years, from $8.2billion in 1985to $10.7billion in 1988. The oil shock of 1986seemsto havehadlittlelong-term impacton thefinancialresourcesof thesefirms. l Stock repurchases. Firmsthatareshorton cashmightbe expectedto avoid any discretionaryusesof fundslike buying backsharesof common stock. However, more thanhalfthefirmsinthissamplehadsufficientfundsto carry out sharerepurchasesin 1986(Table6). Four firmsthathadnot repurchasedsharesin 1985did so in 1986, including abuyback of nearly$600million by W R Grace. Another tenfirms repurchasedsharesin both years. While repurchaseamountsin 1986weresmallerfor mostfirms (butnot all–SchlumbergerandUnion Pacificsteppedup theirrepurchases), four of thesefirmshadrepurchasesin excessof aquarterof abillion dollarsin 1986. l New debt issuesin 1986. Did low collateralvaluemakeitdifficult for firmsto borrow in 1986?4Evidently not-12 firmsinthissampleissuedpublic debtin 1986, andmany mademore thanone offering (Table7).5 Total proceedsof theseissues exceeded$8billion. Of course, not allof thismoney representsnew debt,but rather refinancingof existingissues. However, firmswith multi-billion dollar cashstockpiles mighteasilyhavescaleddown thesizeof thebond offeringsif agencyCOSKof borrowing were amajor impediment. Many, infact,increaseddebtduring 1986. l Early retirement of debt. Inyet anotherdisplayof thecomplete lackof financial constraintsfacedby thesefirms,on September9, 1986,PhillipsPetroleumretiredan entireissueof $200million of guaranteednotesthatwerenot dueuntil 1989. 6

Table 8summarizesthedatafrom theprevioussections. A “V” incolumns 1-3 indicatesthatthefirm violatestestsof beingfinanciallyconstraine,d,,eithedrueto the relationshipbetweencashflow andinvestment,by increasingdividendpayout or by repurchasingshares,respectively. Column 4liststhequartileof theratioof cashplus securitiesto totalassetsrelativeto agroup of COMPUSTAT firmsof comparablesize,3-digit SICindustry,andbond ratings(including259firmswith publicly-rateddebtand 1291firms without rateddebt). Ido not redisplaydataon bond issueshere. Most firmsfailone or more tests;DuPont failsallthree(andissuedover halfabillion dollarsof bonds in 1986),andhasaratioof cashholdingsto assetsnearthatof themedian firm. Only threefirmsmeetallthe restrictions: Chevron, FinaandZapata. Chevron, however, heldliquid assetsequalto 9percentof totalassets,almostthreetimesthemedian sharefor comparablefirms,andraisednearly$1billion in thebond marketduring 1986. One cansafelyruleout financialconstraintsasafactorfor Chevron. ZapataCorp, on the otherhand,wasclearlyfinanciallyconstrainedin 1986. Having justdefaultedon two of itssubordinateddebtissues,itwasoperatingunderanagreement negotiatedwith itsbanklenderswhile itpursuedstepsto restructureitsdebt. The direct influencetheselendershadon investmentdecisionslikely contributedto the 10.3percent decreasein investmentrelativeto salesby itsnonoil segment,thesecond largestpercentage drop inthissample. Moreover, Zapatawasableto obtain anew loanto financethe completion of thecompany’s Wisdom gasfield,subjectto liensplacedon theproject by the bank. Thatis,evenfirmsin defaultmaybe ableto financedesirableprojects. Finaalsowasexperiencingdifficultiesin 1986. Itscashflow andliquid assetsboth plummetedinthatyear. Furthermore,while Finapaid$25million in common dividendsin 1985,iteliminateditsdividend in 1986. Curiously, thechangeininvestmentrelativeto sales by itschemicalsegmentwas-0.95percent–themedianfor thesegmentslistedin Larnont’s Table III,andlessthantheaveragecutof -1.46percent. That is,inspiteof severeconstraints ,onFina,theinvestmentbehaviorof itsnonoil unitwasfairlytypicalcomparedto thefirms thatwere not constrained. 7

Of thefirmsin Lament’s samplethatwerenot in default,only Finaappearsto have beenconstrainedin 1986. However, FinaandZapatawere likely constrainedin 1985aswell, in violation of Lament’s assumptionof,being “awashin cash”inthatyear. Both cuttotal investmentby farmore in 1985thanthey did in 1986. In addition,Zapata’scashflow began deterioratingin 1983,anddecreasedby $166million from 1983to 1985,alargertotaldrop thanthe 1986declineof $81million. Itisimportantto underscorethatthedatado not point to pervasivedistortionsin financingandinvestmentdecisionsof major corporations, asLament asserts.Rather,these firmswereeasilyableto cushion anydeteriorationof oil cashflow with other resources. III. Other pitfalls of the paper. A. Misinterpretation of coefficients. The crux of Lament’s argumentisthathigh cashflow from oil operationssubsidizedinvestmentby inefficientnon-oil industrysegments of diversifiedoil companiesin 1985. In 1986,when oil pricesfellandoil cashflow withered, investmentby non-oil segmentsof thefirm alsodeclined. To demonstratethisrelationship,in Table XI he reportsthe resultsof regressinginvestmentby thenon-oil segmenton the segment’sown cashflow, aswell asoil cashflow of thefirm (allvariablesarescaledby total firm sales).Insupportof thepaper’scentralthesis,usingdatafor 1985only, the coefficient on oil cashflow ispositive andstatisticallydifferentfrom zero: firmswith agreatershareof total cashflow from oil hadhigherinvestmentin nonoil segments,perhapsindicatingthat“internal capitalmarkets”aresubsidizinginvestmentby thenon-oil segment. In 1986thisresultdisappears:the coefficient on oil cashflow is0.00. Lament interpretsthisas “In 1985,oil companieswere awashin cash,andsubsidizedunderperforming nonoil businesses.In 1986,theparentcompaniesstoppedsubsidizingtheir nonoil segments;thesesegmentsthereforereliedonly on theirown cashflow to financeinvestment.”(p.103) This isakinto assertingthatthewealtheffecton consumption holds when the stock market rises,butdisappearswhenthestock marketfalls. Infact,Lament’s hypothesisabout 8

subsidizationoi nonoil investmentwith oil cashflow implieswe shouldalsoobservea positivecoefficient on oil cashflow in 1986. Firmswhose oil shareof cashflow waslowestin 1986shou~dalsohavelessinvestment. A coefficient of zero indicatestheopposite occurred: a lower shareof oil cashflow wasnot associatedwith below-averageinvestmentin 1986. B. Nonoil segments ashigh investment in 1985. Lament characterizesthe nonoil segmentsas“overinvesting”because5of 39units(I3 percent)hadinvestment exceedingpretaxincome. However, for firmswith profitable investmentopportunitiesitis quitecommon for thisto occur. For example,of the 877firmson the COMPUSTAT tapein 1985inthesame3-digitSICindustriesasthenonoil segmentsin Lament’s sample,601firms, or 69percent,hadinvestmentgreaterthanpretaxincome. This suggeststhat,if income or cashflow isthecorrect metricby which to judge if investmentisexcessive,bthenthe 13 percent“high investors”in 1985for Lament’s samplewasactuallyquitelow by industry standards.Most of thesefirmswereunderinvestingintheirnonoil segmentby Lament’s argument. The other 34unitswere,presumably,subsidizingoil operationswith thecashflow from nonoil segmentsin 1985. IV. ~ Conclusion A majority of thefirmsin Lament’ssamplefailthe mostbasictestof whetheradrop of investmentby nonoil segmentsin 1986could havebeencausedby newly-binding cashflow constraintsrelatedto sinkingoil prices: cashflow did not fallmore sharplythandid investmentor, in many cases,actuallyrose. Moreover, mostof therestof thefirmsappearto havebeenflushwith cash–acondition inconsistentwith financialconstraintsbeingimportant for thesefirms. The dataon firm-levelcashholdings andcashflow provide strongevidence againstthe hypothesisthatinvestmentdecisionsby nonoil unitsweresignificantlyaffectedby oil cashflow, or thatcreditmarketimperfectionsareanimportantfactorfor thissetof firms. On thecontrary, theselargepublic corporations havereadyaccessto credit,andchangesin cashflow do not impose binding constraintson theirinvestment. Theseresultsarein accordwith otherstudiesthatcallinto questiontheimportanceof liquidity constraintsandthecausallinkbetweencashflow andinvestment(Kaplanand 9

Zingales(1997)),suggestingthatone mustexercisegreatcautionin interpretinginvestmentcashflow sensitivitiesasevidenceof liquidityconstraints. 10

References Fazzari,Steven,R. Glenn Hubbard, andBrucePetersen,1988,Financingconstraintsand corporate investment,BYookingsPapersonEconomicActivity, 141-195. Gilchrist,Simon, andCharlesP.Himmelberg, 1995,Evidenceon the role of cashflow for investment,JournalofMonetaryEconomics36,541-572. Hoshi, Takeo, Anil Kashyap,andDavid Scharfstein,1991,Corporate structure,liquidity, andinvestment: Evidencefrom Japaneseindustrialgroups, QuarterlyJournalof Economics56,33-60. Jensen,Michael, 1986,Agency costsof freecashflow, corporate finance,andtakeovers, AmericanEconomicReview 76,323-329. Kaplan,StevenN., andLuigiZingales,1997,Do investment-cashflow sensitivitiesprovide usefulmeasuresof financingconstraints?QuarterlyJournalofEconomics,168-215. Lamont, Owen, 1997,Cashflow andinvestment: Evidencefrom internalcapitalmarkets, JournalofFinance52,83-109. Whited, Toni M., 1992,Debt, liquidity constraints,andcorporate investment: Evidence from paneldata,JournalofFinance47, 1425-1470. 11

Notes 1. This analysisassumes,asdid Lament, thatcashflow istheappropriatemeasureof financialconstraint. However, broader measuresof sourcesandusesof fundsalsosupportthe argumentmadeinthispaper,thatfinancialconstraintsdid not pinch for most firmsinthis samplein 1986. Furthermore,most of thesefirmshadamplestocksof cashandsecuritiesto maintaininvestmentexpendituresandweatherashortfallin cashflow, andmany actually addedto theirstock of liquidfinancialassetsin 1986. 2. Total sourcesof fundsexceededtotalusesof fundsfor 14of the 26firmsinthesample, up from 10firmsin 1985,prior to the sharpdrop in oil prices. 3. Relativeto acontrol group comprised of COMPUSTAT firmsof comparablesize,3digitSICindustry,andbond ratings(including 259firmswith publicly-rateddebtand 1291 firmswithout rateddebt). 4. Lament states:“thevalueof thepetroleum-relatedcollateralowned by thecompany alsofell,so externalfinancemay havebeenmore difficult to obtain” (p. 86). 5. At leastone otherfirm issuedprivateplacementsin 1986. 6. Note thatthisisnot themeasurethatwould be suggestedby economic theory; rather, theexpectedNPV of cashflows (or Tobin’s q) drivesinvestment. However, cashflow or pretaxincome isthemeasurethatLament haschosen, so Ihavepresentedthisasanindustv comparison. 12

Table I Change in investment and cash flow from 1985to 1986. Lament’s hypothesisthatfirms’ investmentwasconstrainedby cashflow impliesthat, . cashflow fell by m-o~ethaninvestment,or (dI/dCF) < 1. Comt)anv d dCF dI/dCF ($ million) ($million) Cash flow rose: Du pent deNemours -309 818 * Mobil Corp -329 520 * Royal Dutch/Shell Group -2269 406 * Atlantic Richfield Co -1773 133 * Placer Dome Inc -20 31 * Southdown Inc -25 16 * Investment rose: Litton IndustriesInc 33 -228 * Investment fell by more than cash flow: HomestakeMining -38 -3 12.01 Unocal Corp -665 -128 5.20 Kerr-McGee Corp -225 -45 5.00 Tenneco Inc -1756 -474 3.70 Imperial Oil Ltd -381 -148 2.57 Amoco Corp -1625 -850 1.91 Nova Corporation -40 -32 1.25 Investment fell by less than cash flow: 1.00 CanadianPacific Ltd -57 -58 Burlington Northern RR Co. -299 -330 .91 Chevron Corp -845 -960 .88 Fina Inc -48 -60 .79 ZapataCorp -57 -81 .70 Phillips Petroleum Co -269 -430 .63 Dekalb Energy Co -42 -73 .59 Grace (W R) &Co -321 -567 .57 Occidental Petroleum.Corp -128 -267 .48 Union Pacific Corp -291 -917 .32 Schlumberger Ltd -340 -2120 .16 USX Corp-Consolidated -222 -1641 .14 * Violates assumption of dCF < 0 or dI < 0. 1COMPUSTAT doesnot reportdepreciationin 1985;however, income fellby $18million.

Table II Change in sources of funds other than cash flow and usesof funds other than investment, fro~ 1985to 1986. Thesesourcesandusesrepresentadditionalresourcesthatfirmshadavailableto easefinancial constraints. Change in Change in Net change in -.. Com~anv Other Sources Other Uses other funds Cash flow rose: Du pent deNemours -236 1372 -1608 Mobil Corp -438 -204 -234 Royal Dutch/Shell Group -2353 -850 -1503 Atlantic Richfield Co -1558 -2596 1038 Placer Dome Inc -89 -152 63 Southdown Inc -28 -6 -23 Investment rose: Litton Industries Inc -1423 -1439 16 Investment fell by more than cash flow: Homestake Mining -1 60 -61 Unocal Corp -3894 -3287 -607 Kerr-McGee Corp 49 -5 54 Tenneco Inc 777 2324 -1547 Imperial Oil Ltd -51 224 -275 Amoco Corp -619 -1250 631 Nova Corporation 61 64 -3 Investment fell by less than cash flow: Canadian Pacific Ltd -387 -922 535 Burlington Northern RR Co. 218 370 -153 Chevron Corp -3527 -5734 2207 Fina Inc 226 176 49 ZapataCorp 8 -20 28 Phillips Petroleum Co -6581 -6500 -81 Dekalb Energy Co -18 -34 16 Grace (W R) &Co 1137 972 165 Occidental Petroleum Corp 10514 9914 600 Union Pacific Corp 1873 1575 298 Schlumberger Ltd 1044 835 209 USX Corp-Consolidated 5401 659 4742

Table III Cash andliquid securities holdings. Total holdings of cashandsecuritiesfor firmsin thesample,andcashandsecurities holdings asapercentof capitalexpenditures. Year Total Cash Holdin~s Cash/Investment ($ billions) (%) 1982 21.2 42 1983 26.7 79 1984 25.3 56 1985 25.0 66 1986 31.5 124 1987 31.9 113 1988 23.5 58 1989 21.5 53 1990 24.2 57 1991 22.1 51 1992 21.6 61 1993 24.8 73 1994 25.8 74

Table IV Holdings of cash and securities, by firm. * Stocksof cashandliquid securitiesatendof year,andpercentchangefrom 1985to 1986. Cashratioisholdingsof cashandsecuritiesdividedby totalassetsin 1986,expressedas apercentage;H indicateshigh cashratio:intop quartileof firmsin similarindustries,size .. g~oupsandbond ratings in i986; L indicates low-cashratio, in bottom quartile. Cash and Securities Cash ratio Comvanv 1985 1986 change ($ millions) (0/0) (0/0) Amoco Corp 991 441 -56 1.9 Atlantic Richfield Co 1083 2397 121 11.1 H BurlingtonNorthern RR Co 563 652 16 9.7 H CanadianPacificLtd 234 126 -46 1.0 L Chevron Corp 3168 3131 -1 9.1 H Dekalb Energy Co 198 102 -48 15.1 Du pent deNemours 583 584 0 2.2 FinaInc 30 9 -69 0.5 L Grace (W R) &Co 152 93 -39 2.3 HomestakeMining 90 104 16 14.9 H ImperialOil Ltd 371 532 43 8.5 Kerr-McGee Corp 212 165 -22 5.3 Litton IndustriesInc 1564 1515 -3 33.2 H Mobil Corp 1546 1582 2 4.0 Nova Corporation 145 135 -6 3.9 Occidental PetroleumCorp 726 655 -10 3.7 PhillipsPetroleum Co 676 1141 69 9.2 H PlacerDome Inc 30 155 415 17.8 H Royal Dutch/Shell Group 6949 9400 35 12.3 H SchlumbergerLtd 4590 3810 -17 47.6 H Southdown Inc 7 15 122 2.7 Tenneco Inc 81 111 37 0.6 L USX Corp-Consolidated 289 3915 1255 17.9 H Union PacificCorp 453 300 -34 2.8 Unocal Corp 242 370 53 3.7 ZapataCorp 41 29 -29 3.4

Table V Change in common dividendspaid. Average changein common dividendspaidby firms inthesample,in millions of dollarsandpercent. Average change Percent change in common dividends common dividends b ($ millions) (0/0) 1982 14.4 6.3 1983 5.4 1.9 1984 21.8 7.7 1985 9.3 3.0 1986 29.7 9.4 1987 30.5 8.8 1988 37.0 9.8 1989 28.0 6.8 1990 44.0 10.0 1991 -17.4 -3.6 1992 -0.2 -0.0 1993 -5.8 -1.3 1994 36.1 7.2

Table VI Stock repurchases Firms’ expendituresto repurchasecommon stock in 1985and 1986. Re~urchase amount ($ million Firm 1985 1986 Repurchaseisn1986butnot1985: Du Pent 0 162 W R Grace 0 598 HomesteadMining 0 28 Mobil Corp 0 20 Repurchases in both years: Amoco 806 202 Atlantic Richfield 3525 260 Dekalb Energy 23 1 Kerr-McGee 97 39 Litton Industries 1320 15 Nova Corporation 12 12 Occidental Petroleum 1389 381 PhillipsPetroleum 4972 15 Schlumberger 184 474 Union PacificCorp 153 273 Repurchases in 1985but not 1986: CanadianPacific 8 0 Tenneco 51 0 USX Corp 335 0 Unocal 4178 0 Repurchases in neither year: BurlingtonNorthern 0 0 Chevron 0 0 Fina 0 0 ImperialOil 0 0 PlacerDome 0 0 Royal Dutch Shell 0 0 Southdown 0 0 Zapata 0 0

Table VII Firms with new bond issuesin 1986 Number of s&P Total Com~anv issues Rating Amount m W ($million) Amoco Corp 4 AAA 1, 050 Atlantic Richfield Co 41 AA- A 1,252 BurlingtonNorthern RR Co. 2 AA AA 500 CanadianPacificLtd 1 AA AA 10 billion yen Chevron Corp 3 AA AA- 900 Dekalb Energy Co 0 BB+ Du pent deNemours 2 AA AA 550 FinaInc 0 Grace (W R) &Co 1 BBB+ BBB-2 250 HomestakeMining 0 ImperialOil Ltd 0 AA+ AA+ Kerr-McGee Corp 1 A A- 100 Litton IndustriesInc 0 A- A- Mobil Corp 3 AA- AA- 295 Nova Corporation 0 Occidental PetroleumCorp 4 BBB+ BBB 2,500 PhillipsPetroleum Co 0 BBB BB- PlacerDome Inc 0 Royal Dutch/Shell Group 0 3 3 . SchlumbergerLtd 0 Southdown Inc 0 BBB BBB Tenneco Inc 0 A- BBB+ USX Corp-Consolidated 1 BBB- BBB- 400 Union Pacific Corp 0 AA A+ Unocal Corp 34 BBB BBB- 420 ZapataCorp 0 D5 Total 29 $8,217 + 10 B yen 1Atlantic Richfield includes$2OOmillion of eurobonds, $102million of euroyen bonds anda $200million MTN program. Actual amountoutstandingundertheMTN program isnot available. zw R Grace downgradedto Ba2by Moody ’s. was JTheoperatingsubsidiariesof Royal Dutch ShellGroup hadinvestment-gradedebt ratings. 4Unocal includes$2OOmillion of eurodollarnotes,$110million of SwissFrancbonds and $110of DeutscheMarkbonds. ‘D = In default.

Table VIII Summary of tests of financial constraint in 1986. “V” indicatesthatafirm violatesone of thefollowing testsof theexistenceof fi~l~ncial constraint:Column 1,cashflow fallsby more thaninvestmentin 1986;column 2,no increase individendsin 1986;column 3,firm did not repurchaseshares.Column 4 liststhequartileof .. theratioof cashplussecuritiesto totalassets,relativeto agroup of COMPUSTAT firmsof comparablesize,3-digitSICindustry,andbond ratings(including259firmswith publiclyrateddebtand 1291firmswithout rateddebt),with 1beinglowestcashholdings,4highest. Stock Cash Firm dCF < dI w Repurchase auartile Amoco Corp v v ‘2 Atlantic Richfield Co v v 4 Burlington Northern Rr Co v 4 Canadian Pacific Ltd v 1 Chevron Corp 4 v Dekalb Energy Co -Cl B 3 v v Du Pent (E I) De Nemours v 2 Fina Inc -Cl a 1 v Grace (W R) &Co 2 v Homestake Mining v 4 v Imperial Oil Ltd 3 v v Kerr-Mcgee Corp 3 v v Litton Industries Inc 4 v v Mobil Corp 3 v v Nova Corporation 3 v v Occidental Petroleum Corp 3 v Phillips Petroleum Co 4 v v Placer Dome Inc 4 v v Royal Dutch/Shell Grp Comb 4 v Schlumberger Ltd 4 v Southdown Inc 2 v v Tenneco Inc 1 v Usx Corp-Consolidated 4 v Union Pacific Corp 2 v Unocal Corp 3 ZapataCorp 2