A liquidity risk stresstesting framework with interaction between Market and Credit Risk.docx
《A liquidity risk stresstesting framework with interaction between Market and Credit Risk.docx》由会员分享,可在线阅读,更多相关《A liquidity risk stresstesting framework with interaction between Market and Credit Risk.docx(73页珍藏版)》请在冰豆网上搜索。
AliquidityriskstresstestingframeworkwithinteractionbetweenMarketandCreditRisk
1
Aliquidityriskstress-testingframeworkwithinteractionbetween
marketandcreditrisks
EricWong*+
andCho-HoiHui*
Thisversion:
April11,2011
Abstract
Thisstudydevelopsaframeworkforstresstestingbanks’liquidityrisk,where
liquidityanddefaultriskscanstemfrommarketriskarisingfromassetpriceshocks.
Therisksareassumedtobetransmittedthroughthreechannels.First,banks’
mark-to-marketlossesonassetsincreasetheirdefaultriskandthusinducedeposit
outflows.Second,assetpricedeclinesreducebanks’assetliquidity.Third,banksalso
facehighercontingentliquidityrisk,asthelikelihoodofdrawdownsonirrevocable
commitmentsbycustomersincreasesinthestressscenario.Contagionriskisalso
incorporatedintheframeworkthroughbanks’linkagesintheinterbankandcapital
markets.Theframeworkquantifiesliquidityriskbyestimatingtheexpectedcash
shortagetimeandtheexpecteddefaulttimeofbanks.Stresstestingresultsfromthis
frameworksuggestthatliquidityriskintheHongKongbankingsectorwouldbe
containedevenifsuchstressscenarioweretooccur.
JELclassifications:
C60,G21,G28
Keywords:
Liquidityrisk,stresstesting,defaultrisk,marketrisk,banks,HongKong
*ResearchDepartment,HongKongMonetaryAuthority,55/F,TwoInternationalFinanceCentre,8,
FinanceStreet,Central,HongKong,China.Email:
etcwong@hkma.gov.hk,andchhui@hkma.gov.hk
Phone:
(852)28788735Fax:
(852)28781891
+Correspondence.
TheauthorswouldliketothankNaohikoBaba,KlausDuellmann,HansGenberg,MartinGrieder,
participantsatthethirdannualworkshopoftheAsianResearchNetworkson25March2010
co-organisedbytheBankforInternationalSettlementsandBankofJapan;theresearchworkshop
“ChallengesinBankingResearch”on28-29May2009organisedbytheResearchTaskForceofthe
BaselCommitteeonBankingSupervisionandsponsoredtheBankofSpain,and“theThirdStress
TestingExpertForum:
AdvancedTechniquesinStressTesting”on19-20May2009organisedbythe
IMFandDeutscheBundesbankfortheirusefulsuggestionsandcomments.Theviewsandanalysis
expressedinthepaperarethoseoftheauthors,anddonotnecessarilyrepresenttheviewsoftheHong
KongMonetaryAuthority.1
1.Introduction
AsillustratedbyrecentdevelopmentsintheUSandEuropean
bankingsystems,liquidityanddefaultrisksofbankscanstemfromthecrystallisation
ofmarketriskandsuchinteractionofriskscouldleadtosystemiccrises,e.g.the
sub-primecrisisemergedin2007.1
Whilethebankingsystemsinmostother
economieshaveshownrelativelyresilient,theyarenotimmunetosimilarcrises
becauseofthreecommonfeaturesrunningthroughallbankingsystems.First,banks’
balancesheetsareinevitablyexposedtocommonmarketriskfactors,asthey
generallyholdsimilarfinancialassets.Thus,significantassetpricedeclines,even
onlyinasinglemarket,couldexposealargenumberofbankstosubstantialmarket
risklosses.Secondly,thecapitalavailableforbankstobufferagainstsuchmarket
risklossesislimited,asbanksusuallyoperatewithahighleveloffinancialleverage.
Thissuggeststhatallbankingsystemsareessentiallyvulnerabletomultipledefault
riskduringseveremarketriskshocks.Thirdly,interbankmarketsareextremely
sensitivetodefaultrisk.Anyshockthatincreasesthedefaultriskofbankscould
resultintightenedinterbankmarkets,creatingsystemicliquidityshortages.
Forbankingstabilityitis,therefore,importanttoassesstheextenttowhich
abankingsystemisexposedtosuchaninteractionofrisks.However,intheliterature,
stress-testingframeworkscapturingtheinteractionofrisksarerelativelyscant.Tofill
thisgap,thisstudydevelopsanewstress-testingframeworktoassesstheliquidity
1
Thedevelopmentofthesub-primecrisiscanbesummarisedasfollows:
Intheearlystage(around
thesecondhalfof2007),therewerecontinualannouncementsofsignificantmark-to-market
write-downsofsub-primemortgage-relatedsecuritiesbyfinancialinstitutionsintheUSandEuropeas
aresultofdeteriorationintheassetqualityofsub-primemortgages.Suchcrystallisationofmarketrisk
triggeredconcernsofbanks’defaultriskquickly,asevidencedbyincreasesincreditdefaultswap
spreadsofbankssincethethirdquarterof2007.Defaultriskofbankswasamplifiedfurtherfollowing
thedebacleofsomelargefinancialinstitutions,includingBearStearns,FreddieMacandFannieMae.
Asdefaultriskheightened,banksbecameincreasinglyreluctanttolendamongthemselves,resultingin
thesystemicliquidityproblemsintheglobalinterbankmarketsinmid-September2008followingthe
collapseofLehmanBrothers,despitetheunprecedentedactionsandmeasurestakenbeforebyvarious
centralbanksandgovernmentstoinjectliquidityintheglobalbankingsystem.2
riskofbanksinthiscontext.
Intheframework,weassumethatthereisaprolongedperiod(i.e.oneyear)
ofnegativeexogenousassetpriceshocksinsomemajorfinancialmarkets,which
affectbanks’liquidityriskthroughthreechannels:
(i)increasesinbanks’defaultrisk
anddepositoutflows;(ii)reductioninbanks’liquiditygenerationcapability;and(iii)
increasesincontingentdrawdowns.Defaultriskofbanksisendogenously
determinedusingaMerton-typemodelintheframework.2
Contagiousdefaultriskis
incorporatedthroughbanks’linkageoninterbankandcapitalmarketsthatis
consistentwiththetheoriesintheliterature.
Withthisframework,dailycashoutflowsofbankscanbesimulatedgiven
exogenousassetpriceshocks.UsingtheMonteCarlomethod,theframework
quantifiestheliquidityriskofindividualbanksbyestimatingtheprobabilityofcash
shortageandtheprobabilityofdefaultduetoliquidityproblems.Inaddition,
conditionalonoccurrencesofcashshortageanddefaultinthesimulations,thefirst
cashshortagetimeandthedefaulttimecanbeestimatedrespectively.The
correspondingprobabilityofmultipledefaultsofbanksinabankingsystemcanbe
alsoestimated,whichisanimportantmeasureforassessingthesystemicriskinthe
bankingsystem.Theframeworkwithtwostressscenariosisappliedtoassessthe
liquidityriskofagroupof12listedbanksinHongKongwithpubliclyavailable
data.
Thisstudydrawsontheliteraturethatrelatestotheimpactofassetprice
2
UsingMerton-typemodelstoendogenisedefaultriskofbanksinsystemicriskassessment
frameworksisalsoadoptedbyAspachsetal.(2006),whichisbasedonthetheoreticalframeworkby
Goodhartetal.(2006).3
declinesonbanks’defaultrisk.Cifuentesetal.(2005)andAdrianandShin(2008a,
2008b)provideatheoreticalfoundationonhowasmallassetpriceshockcanbe
amplifiedbyitsmark-to-market(MTM)effectsonbanks’balancesheets,andthus
leadstoadownwardspiralinassetpricesandcontagiousdefaultsofbanksthrough
interbanklinkages.3
Theimportanceofthelinkagebetweenassetpricesanddefault
riskofbankshasrecentlybeenrecognisedandstudiedempiricallybyBossetal.
(2006)andAlessandrietal.(2007)).However,theimplicationsforbanks’liquidity
risk,whicharecrucialforpolicymarkersinviewofthesub-primecrisis,arenotthe
mainfocusofthesestudies.vandenEnd(2008)incorporatestheinteractionbetween
marketandfundingliquidityandpotentialfeedbackonbanksintoaframeworkto
assessliquidityriskofDutchbanksbyestimatingthedistributionsofliquiditybuffers
andprobabilityofliquidityshortfallofbanks.However,theeffectofmarketrisk
shocksondefaultriskofbanksandinturntheirdepositoutflows(andliquidityrisk)
arenotexplicitlymodelledintheframework.Bycontrast,theproposedframeworkin
thispapertriestoestablishthelinkagesbetweenliquidityrisk,marketriskandbanks’
defaultrisk.ItisalsorelatedtothetheoreticalframeworkbyDiamondandRajan
(2005)whichshowsthatanaggregateliquidityshortagecanbecausedbybankruns
perceivedbybankinsolvency.
Thisstudycontributestotheliteratureintwoaspects.First,thisisamong
fewempiricalstudiestoincorporateinteractionofrisksinaliquidityrisk
stress-testingframework.Giventhatthesub-primecrisisishighlyrelevanttosuch
interactionofrisks,theframeworkcouldbeusefulforpolicymakerstoassesshow
resilientabankingsectorisunderliquidityshockssimilartoorevensevererthan
3
SeealsoAllenandGale(1994,1998and2000a),AcharyaandYorulmazer(2007),andNieretal
(2007).InadditiontotheMTMeffects,assetpricecollapsecouldcausebankfailuresbecauseof
widespreaddefaultsofbanks’lendingtoinvestorsforacquiringriskyassets(seeAllenandGale
(2000b)andGoetzVonPeter(2004)).4
thoseoccurredinthesub-primecrisis.Secondly,theframeworkcouldserveasa
complementarytooltothebottom-upapproachforliquidityriskstresstesting.Thisis
particularlysoinviewofthedifficultytoincorporatecontagiousdefaultriskunder
thebottom-upapproach.4
Bycontrast,defaultriskofbanksisendogenisedinthis
frameworkandcontagiousdefaultriskisthuspossiblethroughinterbankandcapital
markets.Theproposedframeworkcanbereadilyappliedtootherbankingsystemas
therequiredinputdataarepubliclyavailable.
Thereminderofthepaperisorganisedasfollows.Thestress-testing
frameworkisoutlinedinthefollowingsection.Sections