從承運交易到貿(mào)易信貸 由非金融公司進(jìn)行金融中介 From Carry Trades to Trade Credit - Financial Intermediation by Non－Financial Corporations

NBERWORKINGPAPERSERIES

FROMCARRYTRADESTOTRADECREDIT:

FINANCIALINTERMEDIATIONBYNON-FINANCIALCORPORATIONS

BryanHardy

FelipeSaffie

WorkingPaper31183

/papers/w31183

NATIONALBUREAUOFECONOMICRESEARCH

1050MassachusettsAvenue

Cambridge,MA02138

April2023

WebenefitedfromhelpfulcommentsbyBoraganAruoba,StefanAvdjiev,PierreDeLeo,KindaHachem,SebnemKalemli-Ozcan,WilliamMullins,HyunSongShin,InaSimonovska,PabloSlutzky,ColinWard,andFrankWarnock,aswellasseminarandconferenceparticipantsattheBIS,IMF,CBC-IMF-IMFERemergingmarketsconference,theWashingtonAreaSymposiumonInternationalFinance,andespeciallyattheNBERconferenceonInternationalFragmentation,SupplyChains,andFinancialFrictions.WearealsothankfultoVictoriaNuguer,ValentinaBruno,andTimSchmidt-Eisenlohrforexcellentdiscussionsthatgreatlyimprovedthismanuscript.Allerrorsareourown.TheviewsexpressedherearethoseoftheauthorsandnotnecessarilythoseoftheBankforInternationalSettlements.TheviewsexpressedhereinarethoseoftheauthorsanddonotnecessarilyreflecttheviewsoftheNationalBureauofEconomic

Research.

NBERworkingpapersarecirculatedfordiscussionandcommentpurposes.Theyhavenotbeenpeer-reviewedorbeensubjecttothereviewbytheNBERBoardofDirectorsthataccompaniesofficialNBERpublications.

?2023byBryanHardyandFelipeSaffie.Allrightsreserved.Shortsectionsoftext,nottoexceedtwoparagraphs,maybequotedwithoutexplicitpermissionprovidedthatfullcredit,including?notice,isgiventothesource.

FromCarryTradestoTradeCredit:FinancialIntermediationbyNon-FinancialCorporationsBryanHardyandFelipeSaffie

NBERWorkingPaperNo.31183

April2023

JELNo.E44,F30,G15,G30

ABSTRACT

Weuseuniquefirm-leveldatafromMexicotodocumentthatnon-financialcorporationsengageincarrytradesbyborrowinginforeigncurrency(FX)andlendingindomesticcurrency,largelyintheformoftradecredit,accumulatingcurrencyriskintheprocess.WeshowataquarterlyfrequencythatthepracticeofborrowinginFXandextendingtradecreditismoreprevalentwhenforeigncurrencyborrowingisrelativelycheaperthanlocalcurrencyborrowing,anditisassociatedwithexpansionsinbothgrosstradecreditandsales.Firmsthatweremoreactiveincarry-trades,accumulatingcurrencyrisk,experiencedlargerreductionsininvestmentandprofitsfollowingalargedepreciationevent.Nevertheless,theirextensionoftradecreditremainedstable,insulatingtheirtradingpartnersfromtheshock.Afirm-levelpanelfor20emergingcountriesprovideexternalvalidityforthelinkbetweencarrytradesandtradecredit.

BryanHardy

BankforInternationalSettlements

Centralbahnplatz2

4051Basel

Switzerland

bryan.hardy@

FelipeSaffie

DardenBusinessSchool

UniversityofVirginia

SaffieF@

1

1Introduction

Non-?nancial?rmsareanimportantproviderof?nancialresourcestotheeconomy,includingtheprovisionofcredittocustomersandsuppliers.Forexample,when?rmAbuysintermediategoodsfrom?rmBandtheagreementinvolvesapaymentafterthedeliveryofthegoods,?rmArecordsaliabilityunderaccountspayableand?rmBrecordsanassetunderaccountsreceivable.Inorderto?nancetheprovisionoftradecredit,?rmBhastouseeitherinternal(e.g.retainedearnings)orexternalfunds(e.g.newloans).Foremergingmarket?rms,lowcostforeigncurrency(FX)creditisoftenusedasasourceofexternalfunds,particularlyforlarge?rmswithbetteraccesstoit.

1

FinancingtradecreditwithFXfundscouldgenerate?nancialstabilityconcerns.Carrytradeincentives(cheapFXcreditrelativetolocalcurrencycredit)canfosterbalancesheetmismatches,as?rmsmightuseFXcredittofundlocalcurrencyassets.

2

Ifusedtosupportvaluechainswithtradecredit,FXriskcouldpropagatetoother?rmsviathosetradecreditlinks,generatingsystemicriskintheeconomy.

3

Despitetheserisks,regulationandprudentialsupervisiontendtofocusprimarilyonbanksandother?nancialinstitutions,whilenon-?nancial?rmsarelessregulatedintheir?nancialintermediationactivitiesandcurrencyriskexposure.

4

1Throughoutthepaper,weusetheterm“tradecredit”togenerallyrefertointer-?rmcredit.Tradecreditcanbeextended(tradecreditassets:accountsreceivable)orreceived(tradecreditliabilities:accountspayable).Tradecreditisdifferentfromtheterm“trade?nance”,whichreferstobank-based?nanceusedtofacilitatecross-bordertrade.

FinkelsteinShapiro,Gonz

lezGo′mez,Nuguer,andRoldn-Pea(

2018

)showthattradecreditprovidesover50%oftheexternalfundsusedforworkingcapitalonaverage,and28%ofoftheexternalfundsusedforinvestmentfor?rmsin13emergingmarkets.Thedataby

Chui,Kuruc,and

Turner

(

2016

)showthatFXdebtaccountsfor31%ofdebtonaverageacrossthesecountries.

2

Klapper,Laeven,andRajan

(

2012

)documentthattheeffectiveinterestrateassociatedtotradecreditis54%onaverageand31%forthemedian.Wheretheeffectiveinterestrateisde?nedastheimpliedinterestrateif?rmspayattheduedateinsteadofpayingearlyandgettingadiscount.Therefore,tradecreditisapro?tableuseforcheapFXfunding.

3Seeforinstance

Altinoglu

(

2021

),whichpresentsatheoreticalmodelshowingshockspropagatingviatradecreditnetworks,and

Alfaro,Garcia-Santana,andMoral-Benito

(

2019

),whichshowsempiricalev-idencethatcreditsupplyshockspropagatetodownstream?rms,potentiallyviatradecredit.Similarly,

JacobsonandVonSchedvin

(

2015

)showsthattradecreditdefaultcangenerateriskalongthesupplychain.

4Recentregulatoryeffortshavestartedtoaccountfor?rmFXrisk,thoughthistypicallyisdonethroughregulationonbanks.

2

Weusedetailedquarterly?nancialdataforMexicannon-?nancialcorporationstostudytheinterplaybetweencarrytradebehaviorandtradecreditprovisionanddocumentfourkeyrelation-ships.First,wedirectlyshowthatFXliabilitiesareusedtofundshort-termpesoassets.Second,themaindestinationofFXliabilitiesthatfundshort-termassetsisthe?nancingofaccountre-ceivables.Third,thislinkbetweenFXborrowingandtradecreditextensionisparticularlypromi-nentwhencarrytradeincentivesarehigh(FXborrowingisrelativelycheap).Speci?cally,duringfavorablecarrytradetimes,?rmsaccumulatecurrencymismatchbyborrowinginFXandaccu-mulatingshort-termpesoassets,includingaccountsreceivables,exposingtheirbalancesheettocurrencyrisk.Fourth,quiteinterestingly,duringthe2008Pesodepreciation,inter-?rmlendingbe-havedmoreasabufferthanamagni?erforsystemicriskas?rmsadjustedmoretheirinvestmentdecisionsthantheirtradecreditprovisiontorelatedparties.

Emergingmarketshavesigni?cantforeigncurrencydebtintheircorporatesector,aswellasprominentuseofinter-?rmcredit.Ouruniquedatasetprovidesanidealopportunitytostudytheserelationships.Itprovides?veadvantagesovertheexistingliteraturestudyingcarrytradebehaviorofnon-?nancialcorporations.First,webuildapaneldatabaseataquarterlyfrequency,whichcapturestheshort-maturitydynamicsofcarrytradeandtradecreditactivities,thesedy-namicsaretypicallymissedbystudiesrelyingonannualdata.Second,ourdatasetincludesde-tailedinformationofthecurrencycompositionofbothliabilitiesandassets,allowingustodirectlyconnectFXliabilitiestotheirfundingofpesoassets,abehavioronlyimpliedorindirectlyobservedinpreviousstudies.Third,weindependentlycaptureallsourcesofFXandPesoborrowing(e.g.bonds,loans,etc.)andtheinternalcash?owgeneratedbythe?rmeachquarter,allowingustodoacompleteaccountingdecompositionofallsourcesandusesoffundsbythe?rm.

5

Fourth,

5Thisstandsincontrasttotheexistingliteraturewhichexaminesonlyonesourceoffunds(e.g.bonds)andoneuseoffunds(e.g.cash),ouranalysiscanprovideanaccountingdecompositionofallsourcesandusesoffundsbythe?rm.Examiningonlyonesourcetacitlyassumesthatitsuseandallocationareindependentofothersourcesanduses.Theonlysourceofexternalfundingthatwedonotobserveisstockissuance.Thislong-termandinfrequent?nancingstrategyisunlikelytoplayaroleincarrytradesortradecreditbehavior.

3

thedataalsoincludeadetailedbreakdownofshort-termassetsbyinstrument,allowingustoex-aminehow?rmsadjusttheirextensionoftradecreditandbenchmarkitagainstotherassets(e.g.inventories).The?naladvantageistheinclusionofrealoutcomessuchassales,investment,andemployment,makingitpossibletoconnectcarrytradeand?nancialactivitiesofthe?rmtoitsrealactivities.Thisdimensionallowsustostudytherealeffectsofacurrencydepreciationon?rmsthataccumulatedcurrencymismatchduetotheircarrytradeactivities.

Usingthisuniquedatabasewedocumentthattradecreditisakeycomponentofa?rm’sbalancesheets.Infact,accountsreceivableconstitute40%ofshort-termassetsonaverage,whileaccountspayabledenominatedinFX(Peso)makeup32%(23%)oftotalFX(Peso)liabilitiesonaverage.Bycomparison,accountspayableinanycurrencyaremoreimportantthanbondsandasimportantasloansfortheMexican?rmsinoursample.

Havingestablishedtheimportanceoftradecreditin?rmbalancesheets,weusearegressionframeworktoperformanaccountingdecompositionlinkingshort-termassetchangestochangesininternalandexternalfunds.ThisrevealsthateachdollarofFXliabilitiesbacks38centsofshorttermassets,17centsofwhichareusedtodirectlyaccumulateshort-termpesoassets.WethusprovidedirectevidenceofcurrencymismatchassociatedwithFXfunding.Moreover,whendecomposingthose38centsbyinstrument,wedocumentthatthemaindestination(40%)ofthesefundsisaccountsreceivables.Therefore,?rmsuseFXfundsto?nancetradecreditandintheprocesstheygeneratecurrencyexposureontheirbalancesheet.

Wethusdocumentthatnon-?nancial?rmsactakinto?nancialintermediaries,withapositiveco-movementbetween?nancialassetsandliabilities-fundingpesoassetswithFXliabilities-butfurthershowthemaindimensionalongwhichtheyactasintermediariesisbyextendingtradecredittoother?rms.Inordertoshowthatthispatternisrelatedtocarrytrademotives,weexaminethemaindriverofcarrytradedynamicsintheliterature:interestratedifferentials.Tothisend,weleveragetheloanlevelcomponentofourdatabasetobuilda?rmlevelmeasure

4

oftheinterestdifferentialbetweenFXandPesoloans.

6

Usingthismeasure,wedocumentthat?rms’tendencytofundpesoassetswithFXliabilitiesiscorrelatedwithcarrytradeincentives.Whenforeigncurrencyborrowingratesarerelativelylower,?rms’short-termFXliabilitiesandshort-termpesoassets,includingaccountsreceivable,arerelativelyhigher.

Carrytradeincentivesareassociatedwithanexpansionoftradecreditnetworksandincreasedsales.

7

Becauseinvestmenttypicallyexhibitsalagwhenaffectingproduction,ouruseofquarterlydatarevealsthatFXcreditfundinginvestmentdoesnotdrivethesalesandtradecreditdynamics.Thusourevidencedirectlylinkscarrytradebehaviortoproductionandsalesviatradecreditlinkages.

Thegranularityofourpanelallowsustoprovidedirectevidencelinkingcarrytradebehaviortointer?rmlending,andtheassociatedbuildupexposuretocurrencyrisk.Whataretherealeffectsforthesecarry-trade?rmswhenthecurrencyriskmaterializes?Dotheytransmittheirriskalongtheirvaluechainsorabsorbmostoftheimpact,providinginsurancetotheirtradingpartners?Weusetherealvariablesofour?rmlevelpaneltoanswerthesequestionsandthe2008MexicanpesodepreciationtriggeredbytheLehmanBrothers’defaultasaquasi-naturalexperi-ment.Largedepreciationepisodeswreakhavocon?rmbalancesheetsandthemacro-economygenerally,andtheendof2008featuredadepreciationofthepesoof33%againsttheUSdollar.Priortothatshock,oursamplefeaturesaperiodofhighcarrytradeincentivesover2005-2008,witharelativelystableexchangerateandalargeandincreasinginterestratedifferential.

We?ndevidencethattheFXriskdoesnotpropagatethroughtradecreditnetworkswhenashockarises.Investmentandemploymentfallafterthedepreciationforall?rms,asdotradecreditandsales,re?ectingthegeneralimpactoftheshock.Moreover,?rmsthataccumulated

6Ourbaselineanalysiscomputestheinterestratedifferentialfromaverageinterestratesacross?rmstocapturegeneralborrowingconditionsfornon-?nancial?rmsineachperiod.Wealsoshowthatourresultsarerobusttousinga?rmspeci?cmeasureforthesubsampleof?rmsthatborrowsimultaneouslyinbothcurrencies,andsaturatingthatregressionwithtime?xedeffects.

7Firmsdonotchangetheamountofeachsalesoldoncredit.Rather,theyappeartopassthecostsavingsfromthecheaperFXborrowingontotheircustomers.

5

moreshort-termFXexposureoverthecarrytradeperiodperformedpoorlyfollowingthedepre-ciation,withlowerinvestmentgrowththansimilar?rmsthatdidnotincreasetheirexposureandlowerpro?ts.Interestingly,tradecredit(borrowingorlending)forthesecarrytrade?rmsisnotcomparativelydifferentfollowingthedepreciationshock.Therefore,investmentandpro?tsfallaftertheexogenousexchangerateshockfor?rmsthatbuiltupFXexposurethroughcarrytrade,buttheirtradecreditremainsrobust.Thisasymmetrysuggeststhat?rmsplaceahighvalueontheirinter-?rmcreditandrelationships,astheyprefertodecreasephysicalinvestmentortodrawfromother?nancialassetsinordernottocutcredittorelatedpartners.Interestingly,whiletheliteraturetypicallyviewsbanksaspropagatorsofshocks,pullingbackcreditwhentheyarehit(e.g.thebanklendingchannelofmonetarypolicy),inter-?rmlendingbehavesmorelikeabuffer. Therichnessofourdatacomesatthecostofonelimitation:ourresultsarebasedonasampleofMexican?rms.Toaddressconcernsaboutexternalvalidity,weusetheCapitalIQcapitalstruc-turedatafor?rmsin20emergingmarketstodocumentthelinkbetweencarrytradeandtradecredit.Wegenerateaquarterly?rm-levelpanelwithalmost8000?rms.Onelimitationofthisdata-setisthatthereisnoinformationaboutthecurrencyoftheassets,onlythedebtliabilities.Therefore,itcannotbeusedtodirectlymeasurecurrencyexposure.ConsistentwithourMex-icanresults,ouranalysisshowsthat1)themaindestinationamongshort-termassetsof?rm’sborrowingistradecredit?nance;2)cash,potentiallyinlocalcurrency(

Bruno&Shin

,

2017

),andaccountreceivablesare?nancedbyFXdebt;and3)carrytradeincentivesincreaseFXborrowing,

decreaselocalcurrencyborrowing,andexpandtradecreditandsales.Therefore,wecon?rmthelinkbetweentradecreditandcarrytradebeyondthecaseofMexico.

RelatedLiterature

Evidenceofcarrytradebehaviorinnon-?nancial?rmshasbeenshownintheliteratureinthecaseofemergingmarket?rms,borrowingviaUSDbondsandholdingcashwiththeproceeds.Usinga

6

cross-countryannualpanelof?rms,

BrunoandShin

(

2017

)showthatemergingmarketeconomy(EME)?rmsissueUSDbondswhenthecarrytradeisfavorable.These?rmsusetheproceedstodisproportionatelyaccumulatemorecash,suggestingacarrytrademotive.ByassumingthatcashisdenominatedinlocalcurrencyandthatFXbondissuanceisorthogonaltoeveryothersourceofexternal?nance,

BrunoandShin

(

2017

)providegranularevidenceofnon-?nancial?rmsengagingincarrytrade.

Caballero,Panizza,andPowell

(

2014

)complementthese?ndingsforasampleofemergingeconomiesandstudytheroleofcapitalcontrolsinpromotingthisbehavior.Similarly,

AcharyaandVij

(

2017

)usesyearlydataforIndiaandstudytheeffectsofmacro-prudentialpolicyon?rm’sFXborrowing.InsteadofholdingcashafterissuingFXbonds,adominantstrategywouldbetoinvesttheproceedsinhighreturnlocalcurrencyassets.Inthisvein,andclosertoourstudy,

Huang,Panizza,andPortes

(

2018

)performacountryspeci?canalysisforChinausingyearlybonddata,theyshowthat?rmsrespondtocarry-tradeincentivesusingFXbondproceedstofundaccountsreceivable.

Wecontributetothisliteraturebystudyingforthe?rsttimeallsourcesofexternalfundinganddirectlyobservingthecurrencycompositionofthewholebalancesheetatquarterlyfrequency.Becausewecandecomposeassetsbyinstrumentandcurrency,wecanrelaxtheassumptionthatallcashholdingisdenominatedinlocalcurrencyandthatfundingsourcesareuncorrelatedbydirectlyshowingthat?rmsusecarrytradeproceeds(i.e.FXborrowing)tofundshort-termassetsinpesos,takingintoaccountallsourcesoffunding.Interestingly,althoughwecanreplicatetheFXbond-cashcorrelationdocumentedbytheliterature(

Bruno&Shin

,

2017

),themoreimportantdestinationforFXliabilitiesarenon-cashassets(especiallyaccountsreceivable),drivenbyFXbankloanandFXaccountspayableinsteadofFXbonds.Becausecarrytradestrategiesareshortmaturity,itisnaturalforshort-term?nancialinstruments(whosedynamicsarebettercapturedataquarterlyfrequency)tobethekeydriversofthisbehavior.Thus,these?rmsnotonlyborrowatlowFXrates,buttheyusetradecredit,ahighyieldingasset(

Klapperetal

.,

2012

),tobene?tfromtheinterestratedifferential.

7

Ourresultsprovideimportantevidenceforhowcreditconditionscanaffectproductionviasupplychainsandproductionnetworks.Whenproductionschainsarelong,creditshockscanamplifyrecessionsbydisruptingthetradecreditlinkagesthatsustainthechain(

Alfaroetal.

,

2019

;

Kalemli-O

¨zcan,Kim,Shin,S?rensen,&Yes?iltas?,

2014

).

8

BrunoandShin

(

2018

)showthatwithastrongerdollar,creditconditionstightenandleadstoareductionininternationalsupply

chains.Thus,FXcreditconditionsmaysynchronizetradecreditbyincreasingthe?owofcreditthroughthenetworkof?rms.WebuildabridgebetweenthevaluechainliteratureandthestudiesonexchangeraterelatedbalancesheetshocksbyshowingthatcarrytradeincentivescanincreaseboththeFXexposureandtradecreditnetworkof?rms.

9

Uncoveredinterestrateparity(UIP)conditionsareoftenviolatedinemergingmarkets,biasingborrowingtowardsforeigncurrency(

Burnside,Eichenbaum,&Rebelo

,

2007

;

Gilmore&Hayashi

,

2011

;

Hassan

,

2013

;

Niepmann&Schmidt-Eisenlohr

,

2022

)andexposingtheireconomiestocur-rencycrises.

10

BrunoandShin

(

2020a

)showthat?rmsthatexploitcarrytradeopportunitiesseetheirsharepriceaffectedduringcurrencydepreciationepisodes.Consistentwiththeirresults,weshowthat?rmsinvolvedincarrytradeseedecreasedinvestmentandpro?ts.However,ourre-sultssuggestthatinter-?rmtradecreditnetworksarevaluabletothe?rm,astheyaremaintaineddespitedeclinesininvestmentandotherresourcesintheeventofanadverseshocktothe?rm.

8

Garcia-Marin,Justel,andSchmidt-Eisenlohr

(

2022

)providesamodelandevidenceshowingthat?rmswithhighermarkupsbene?tmorefromextendingtradecredit,asbankcreditisonlyneededtocoverproductioncostsratherthanthefullsaleprice.Weprovideacomplementaryviewthatcheaperbankcreditcanmakeiteasierfora?rmtoextendmoretradecredittotheirtradingpartners.

9FXborrowingandbalancesheetexposuregenerallyresultinlowerinvestmentfollowingadepreciation(

Aguiar

,

2005

;

Cowan,Hansen,&O

′scarHerrera,

2005

;

Hardy

,

2018

;

Kalemli-O

¨zcan,Kamil,&Villegas-

Sanchez

,

2016

;

SerenaGarralda&Sousa

,

2017

).

10See

diGiovanni,Kalemli-O

¨zcan,Ulu,andBaskaya(

2018

);

SalomaoandVarela

(

2018

)formorerecentevidenceonUIPdeviations.Monetarypolicyofthelocalorforeigncurrencycanaffecttheinterestratedifferentialandthus,theincentivestoborrowandlendineachcurrency(

Avdjiev,Koch,McGuire,&von

Peter

,

2018

;

Ongena,Schindele,&Vonnak

,

2016

).Capitalcontrolscanalsoin?uencetheFXborrowingof?rms(

Keller

,

2018

).

BocolaandLorenzoni

(

2018

);

GabaixandMaggiori

(

2015

);

GopinathandStein

(

2018

)providemodelswhichmicrofounddeviationsfromUIPandprovideframeworkstounderstandriskofcurrencyexposure.Ourresultssuggestthatinter-?rmlendingisanimportantelementyettobeincludedinthesemodels.

8

Thisaccordswithevidenceonthevalueoftraderelationshipsinthecontextofinternationaltrade,which?ndsthatdisruptionstotraderelationshipsiscostlytoboththe?rmandthemacroecon-omy(

Monarch&Schmidt-Eisenlohr

,

2020

).Thus,whileFXcreditconditionsmayin?uencethevolumeoftradecreditandproductioninsupplychains,shockstoindividual?rmstendtobeabsorbedratherthantransmittedviatradecreditlinks.

11

Theremainderofthepaperproceedsasfollows:inSection

2

,wedescribeourdataandsample;Section

3

examinestheborrowingandsavingof?rmsbycurrencyandinstrument;Section

4

providesevidenceofcarrytradeactivityin?rmshort-termFXpositions;therealconsequencesfor?rmsofthatexhibitcarrytradebehaviorisexploredinSection

5

;Section

6

providesevidenceofexternalvaliditylinkingcarrytradeincentivestotradecreditusing?rmleveldatafor20emergingmarkets;andSection

7

concludes.

2DataandSample

Weuseanoveldatasetoflistednon-?nancial?rmsinMexicothatincludesdetailedinformation

onbothassetandliabilityFXexposure.Thisdatasetisderivedfromquarterly?nancialstatementsmadebycompanieslistedontheMexicanStockExchange(BMV).

12

Thisisaquarterly?rm-level

datasetof183?rms(unbalanced)over2005q1-2015q2.Table

1

summarizestheavailablebreak-

downsoftheFXliabilitiesandassetsinthedata.

11Tradecreditmayinvolvenon-?nancialmotives(

Klapperetal.

,

2012

)orbeusedtosmoothcustomerprices(

FinkelsteinShapiroetal

.,

2018

).Suchtradecreditisespeciallyimportantto?rmswithoutaccesstobankcredit(

Minetti,Murro,Rotondi,&Zhu

,

inpress

).

12See

Hardy

(

2018

)formoredetailonthedataset.

9

Table1:CurrencyCompositionData

FXandPesoLiabilitiesFXandPesoAssets

Total

byMat-urity

byIns-trument

byInst.&Mat.

Total

byMat-urity

2005q1-2007q42008q1-2011q42012q1-2015q2

Thedataismore?exibleontheliabilitythanontheassetside.Infact,wecandirectlyob-servea?rm’sliabilitybyinstrument,currency,andmaturity(e.g.,shorttermbondsinFX),butwecannotsimultaneouslydecomposeassetsbycurrencyandinstrument(e.g,totalcashortotalFXassets).

13

Theinstrumentbreakdownontheliabilitysideincludesbankcredit,marketcredit(bonds),accountspayable(tradecreditreceived),andother.Theassetscanalsobesplitbyin-strument,withshorttermassetssplitintocash,?nancialassets,inventories,accountsreceivable(tradecreditextended),andother.Thisdetailinthebalancesheetdataisuniqueinthelitera-tureandmakesitpossibletodirectlyexaminehow