rob engle china research luncheon february 5, 2013pages.stern.nyu.edu/~jcarpen0/chinaluncheon/engle...

Post on 30-Sep-2020

3 Views

Category:

Documents

0 Downloads

Preview:

Click to see full reader

TRANSCRIPT

RobEngleChinaResearchLuncheonFebruary5,2013

  Whenthefailureofoneormorefinancialinstitutionstomeetitsobligationshassubstantialnegativeimpactsontherealeconomy,thenthesearesystemicinstitutions.

  WesawthatLehmanbankruptcybegantheworstepisodeofthefinancialcrisis.ButitcoincidedwithdeepdistressatFANNYandFREDDIE,Citi,BankofAmerica,MerrillLynch,Goldman,MorganStanley,AIG,WAMU,Wachovia,andmanymore.

  Howmuchcapitalwouldafinancialinstitutionneedtoraiseinordertofunctionnormallyifwehaveanotherfinancialcrisis?

  Wemeasurethiseconometricallybasedonmarketdataonequitiesandbalancesheetdataonliabilities.WeupdateweeklyonV‐LABforUSandGlobalfinancialfirms.WecallthisSRISK.

  Principleinvestigators:ViralAcharya,MattRichardsonandmeattheVolatilityInstituteatNYU’sSternSchool.CollaborationwithHECLausanneandtheInstituteforGlobalFinanceatUniversityofNewSouthWales.ContributionsbyChristianBrownlees,RobCapellini,DianePerriet,EmilSiriwardane.

  Regulatorsmeasurethisbasedonsupervisorydataandstressscenarios.

  ManyotherrelatedmeasuresarebeingdevelopedorareinusebyregulatorsinEuropeandtheUS.

  SomemeasuresarefirmspecificsuchasCoVaR,andnetworkmodelsthattracelinkages.Othersarefinancialindustryqualitymeasuressuchasvolatility.

  RecentsurveysbyBrunnermeierandOehmkeandbyBisias,Flood,LoandValvaniscovermanymeasures.

2/5/13 VOLATILITYINSTITUTE 4

  Thedominoornetworkmodelsaysthatonefirmcanbesoimportantorinterconnectedthatitsfailurewillprecipitatethefailureofitscounterpartiesandthentheywillbringdowntheircounterpartiesuntilthesectorandtherealeconomyfails.  Question:whataretheotherdominosdoingwhilethefirstarefalling?Whatareinvestorsdoing?

  Question:whatisthedifferencebetweeninterconnectednessandrisksharing?

  Thetsunamimodelisbasedonmanyinstitutionshavinglargerisksonthesameevents.Iftheseeventsoccur,thenalltheinstitutionswillbestressedatthesametimeandthesectorandeconomywillcollapse.

  SRISKiscomputedfrom:

  Wherekisaprudentiallevelofequityrelativetoassetstakentobe8%(and5.5%forIFRSfirms)andLRMESisthedeclineinequityvaluestobeexpectedifthereisanotherfinancialcrisis.

  SRISKdependsuponsize,leverageandrisk.

  BankofAmericahasamarketcapof$114billion.Itsaccountingliabilitiesare$1.9trillionforaleverageratioof17.9

  Ifwehaveanotherfinancialcrisiswhichisassumedtobeafallof40%inbroadUSequitiesoversixmonths,thenweestimatesharesinBACwillfallby60%.

  ThisreflectsaDynamicConditionalBetaof1.7todaythatwillmoveinthefutureduetomeanreversioninvolatilitiesandcorrelationsandalsowillrisewithdownsidereturns.

  SRISK=$112billion.  Itisundercapitalizedsomewhattodayandthiswillbemoresevere

underthestressofanequitydecline.

  CreditAgricolehasamarketcapof$19billion  Ithasliabilitiesof$2.1trillionforaleverageratioof124

  Anyfluctuationinassetorliabilityvaluationscaneasilymovethefirmintobankruptcy.

  Mostofthecapitalshortfallisneededtobringtheleveragedownnow.Theriskisonlyasmallpartofthecapitalshortfallcalculation.

  Mostlikely,CreditAgricoleisnolongermakingloansexceptpossiblythemostsecure.

  Ifwehaveafinancialcrisis,thenallfirmswithpositiveSRISKwilltrysimultaneouslytoraisecapitalandtheonlysourceislikelytobetaxpayers.ThebiggerSRISK,themoreseriousthethreattofinancialstability.

  SRISKisestimatedconditionalonanendogenousvariable–astresstestdoesnotindicatecausality.

  Buthowdoesthishappen?

  IfanyfirmshavehighSRISK,theywillrecognizetheirvulnerabilityandwillbegintodeleverandderisk,therebyimpactingtherealeconomy.IfonlyafewfirmshavehighSRISK,theremainingfirmscantakeuptheslack.

  Asthemacroeconomyslows,stockpriceswillfall,volatilitywillrise,andSRISKwillgoupmore.

  Firmsmaydeleverandderiskbyattemptingtosellilliquidassetsandhoardingcashleadingtofurtherdeclinesinrealandfinancialsectors.

  Investorsrecognizefinancialinstitutionweaknessandlowervaluations,increasingSRISK

  Forwardlookinginvestorscouldmakethishappeninonestep.

  Bankruptciesandotherfailureswilloccuruntileventually,thereturntocapitalishighenoughtobringnewcapitaltotheindustry.

  Thespiralcanbearrestedbeforethebottom.

  However,thiswillerodemarketdisciplineandmayimposehugeregulatorycostsonthefinancialsectorgoingforward.

  Thusregulationisneededinadvance.Ideallyitwouldbecountercyclical.

  Externalities–ifonlyonefirmhashighSRISK,thereisnospiral.

  ImplicitandExplicitgovernmentguaranteessuchasdepositinsuranceor“toobigtofail”

  Regulatoryincentives–themeasure:“riskweightedassets”ignorescorrelationandhenceleadstonon‐diversifiedassetmix

  Riskweightsmaybepoormeasuresofrisk.

 Miscalculation:useshortrunriskmeasurestochooseleverageratherthanlongrunrisk.

 Miscalculation:valuingexoticsecuritiessuchasCDOswithoutrecognizingalltherisks.

 Miscalculation:housingpricescangodown  Agencyproblems–wallstreetbigshots. ……..Toomanypossibilities

2/5/13 VOLATILITYINSTITUTE 15

2/5/13 VOLATILITYINSTITUTE 16

2/5/13 VOLATILITYINSTITUTE 17

2/5/13 VOLATILITYINSTITUTE 18

2/5/13 VOLATILITYINSTITUTE 19

  RegulatorsmightrequirethatfirmsholdsufficientcapitalsothattheirSRISKiszero.Thustheywouldnothavetoraisecapitalinafuturecrisis.

  ThusfirmswouldberequiredtoreduceSRISKwhichcanbedoneby  Deleveraging  Demerging  Derisking  Decliningtofollowtheherdwithidenticalbets.

  Itisbestifcapitalrequirementscanbeincreasedingoodtimessincethebankscaneasilyraisecapitalandincreasetheirbuffer.

  Inbadtimes,itisnaturaltoreducerequirementsbecausenewcapitalisveryhardandexpensivetoraiseatthattimeandbecausedraconiancutswillhurttherestoftheeconomy.

DYNAMICCONDITIONALBETA

2/5/13 VOLATILITYINSTITUTE 22

  Conditionont‐2

  Theequation

  ButucanbeanMA(1)andGARCH.Infact,itmusthaveMA(1)ifRiistobeaMartingaledifference.

  Combiningtheconstantbetaanddynamicconditionalbetaintooneregression:

 WhereuwillbeanMA(1)GARCH

  For1200globalfinancialinstitutionsweupdateweeklyestimatesofSRISK.ThesenowuseNestedDynamicConditionalBetawithMA(1)andGARCH.

  http://vlab.stern.nyu.edu  ImayalsoshowyouresultscorrectingfordifferencesbetweenGAAPandIFRSaccountingthatarenotyetonthewebsite.

  ThechangeinSRISKfromonetimeperiodtoanothercanbeattributedtochangesindebt,equityorrisk.

  Inearly2008LocalGovernmentDebtinChinawas1.7trillionyuan.Attheendof2010itwas10.7trillionor27%ofGDP(NationalAuditOfficeJune2011)

  LocalGovernmentscannotborrowdirectlyandestablished6,576specialfinancingvehicles.

  8.5trillionyuanarefrombankloansdue2011‐2013

  530billionyuanofirregularitiesinlocalgovernmentdebt  46billioninirregularcreditguarantees  73billionsecuredwithirregularcollateral  35billionspentonstocks,housingandpollutingplants

  132billionnotmadebyapproveddeadline  244billionoffraudulentunderpaymentofregisteredcapitalinfinancingvehicles

  Orderstocorrecttheseirregularitieshaveonlybeenpartiallysuccessfulthusfar.

  Localgovernmentsarenotpermittedtoborroworissuebonds

  Localgovernmenttaxrevenuesgrewmoreslowlythanexpenditures

  Stimuluswasallocatedtolocalexpendituresandexacerbatedtheproblem.

  4trillionyuan($618billion)allocatedto10areasincludingincludinglow‐incomehousing,ruralinfrastructure,water,electricity,transportation,theenvironment,technologicalinnovation,andreconstructionfromseveraldisasters.

  Financedby1.2trillionfromcentralgovernmentandtherestbylocalgovernments.Theysetupfinancingvehiclesandborrowedmassivelyfromthebanksascreditstandardswereeasedandregulatorsshiftedflowstolocalgovernments.

  Netinterestmarginsshrunkbecauseofmacroeconomicpolicy;banksexpandedloanstomakeuptherevenue.

  Chiefrevenuesourceforlocalgovernmentsislandsaleswhichareslowing.Inflationfearswillleadtomacroeconomictighteningwhichwillfurtherslowthisgrowth.

  Projectshavelonglivesandcannotcoverinterestpaymentsintheshortrun.MorethanhalfofChina’sGDPgrowthoverthelastcoupleofyearsisbasedonfixedinvestment.Butsuchstimulusisnotnecessarilycommerciallyviableonitsown.

  Muchofthisborrowingisoffbalancesheetusinggovernmentlandorassetsascollateral.

  RecentlyWSJreportedthatallmunicipalloanswererolledoverattheendof2012.

  Banklendingtostateownedenterprisesgrewrapidlyinthisperiodpresumablyastheseweremoresecure.

  GrowthofSOEsattheexpenseofprivatefirmsislikelytosloweconomicgrowth

  SOEsreportedlyinvestedinrealestateratherthancorebusinesses.

  Debtforgivenessorrolloveraretemporaryfixes.

 Municipalitiesneedgreateraccesstofunds.Needtaxreformtogivemoresources.

 Municipalitiesneedtobeabletoissuebondswhichwouldmakefinancingmoretransparent.Thisisnowbeingexperimentallytried.

  “TOOBIGTOFAIL”guaranteeleadseachinstitutiontotakemoreriskthansociallyorprivatelyoptimal.

  “LEVERAGEEXTERNALITY”leadseachbanktotakemoreleveragethanissociallyoptimal

  “RISKMYOPIA”leadsinstitutionstotakelongrunpositionsbasedonshortrunrisks.

  “HERDING”leadseachinstitutiontochoosesameinvestmentportfolio.Regulationmayalsodothis.

  “REGULATION”mustoffsettheseincentives.

  Theyarestate‐ownedandwillsurelyberescuedshouldtheyneedit.

  Thesovereignhas$3trillionofreservessorescueshouldbeeasy.However,sellingthese$assetswillmakeithardtomanageexchangerates.

  TotalSRISKisonly$200billiontoday.  OffbalancesheetitemsarelikelytobeimportantsoSRISKisprobablyunderstated.

  FinancialequitymaybemorevaluablebecauseoftheguaranteesandhenceleadtolowerSRISK.

  ThecausesofsystemicfinancialriskareallveryapparentintheChinesesettingthustherisktothefinancialsectorislikelytogrow.

  Regulationmayoffsettheseincentivesifitiseffective.

  FSBlistsonlyBankofChinaasGSIFI. MyGuess–probablynotnowbutpayattention!

top related