Central Clearing and Systemic Liquidity Risk

Finance and Economics Discussion Series Federal Reserve Board, Washington, D.C. ISSN 1936-2854 (Print) ISSN 2767-3898 (Online) Central Clearing and Systemic Liquidity Risk Thomas B. King, Travis D. Nesmith, Anna Paulson, and Todd Prono 2020-009 Please cite this paper as: King, Thomas B., Travis D. Nesmith, Anna Paulson, and Todd Prono (2022). “Central Clearing and Systemic Liquidity Risk,” Finance and Economics Discussion Series 2020-009r1. Washington: Board of Governors of the Federal Reserve System, https://doi.org/10.17016/FEDS.2020.009r1. NOTE: Staff working papers in the Finance and Economics Discussion Series (FEDS) are preliminary materials circulated to stimulate discussion and critical comment. The analysis and conclusions set forth are those of the authors and do not indicate concurrence by other members of the research staff or the Board of Governors. References in publications to the Finance and Economics Discussion Series (other than acknowledgement) should be cleared with the author(s) to protect the tentative character of these papers.

CENTRAL CLEARING AND SYSTEMIC LIQUIDITY RISK THOMASB.KING,TRAVISD.NESMITH,ANNAPAULSON,ANDTODDPRONO Abstract. Bysteppingbetweenbilateralcounterparties,centralcounterparties(CCPs) transformcreditexposure,therebyimprovingfinancialstability.But,largeCCPsare concentratedandinterconnectedwithmajorglobalbanks.Moreover,althoughtheymitigate creditrisk,CCPscreateliquidityrisks,becausetheyrequireparticipantstoprovidecash.Such requirementsincreasewithmarketvolatility;consequently,CCPliquidityneedsareinherently procyclical.ThisprocyclicalitymakesitmorechallengingtoassessCCPs’resilienceintherare eventthatoneormorelargefinancialinstitutionsdefault.Liquidity-focusedmacroprudential stresstestscouldhelptoassessandmanagethissystemicliquidityrisk. Keywords:Financialsystems;Centralcounterparties;CCPs;margin;liquidityrisk;systemic risk;financialstability;procyclicality JEL:g23;g21;g28;e58;n22 1. Introduction Centralcounterparties(CCPs)playacriticalroleinthefinancialsystem. Byinsertingitself betweenthetwoparticipantsinasecuritiesorderivativestransaction,acentralcounterparty guaranteespaymentsthatcouldotherwisebejeopardizedbythedefaultofeitherparticipant. CCPs’importancehasgrownsignificantlysincethe2008financialcrisis,inpartasaresultof newregulatorymandatesrequiringcentralclearingofover-the-counter(OTC)derivatives. By the first half of 2018, over 75% of the notional value of interest rate swaps (IRS) and creditdefaultswaps(CDS)—thetwolargestcategoriesofOTCderivativesaffectedbyclearing WethankIvanaRuffiniforhercommentsandforprovidingsubstantialinsightsintothedata.Wealsothank KelseyBurr,MartaChaffee,RoyCheruvelil,JoshGallin,RichardHaynes,MikeKiley,BethKlee,DavidLi,Jennifer Lucier,DavidMurphy,SamSchulhofer-WohlandJonathanRoseforhelpfulcommentsandfeedback. Wealso wishtothankthethreerefereeswhosecarefulreviewandcommentsimprovedthepaper.Mostofall,wewantto recognizethedebtweoweourformercolleagueRobertCox;henotonlyinfluencedthispaperwithhiscomments, butmoregenerallyinfluencedourunderstandingofthetopic.DavidKelley,TimLarachandJohnSpenceprovided excellentresearchassistance.Theanalysisandconclusionssetfortharethoseoftheauthorsanddonotindicate concurrencebyothermembersoftheresearchstaff,theFederalReserveBankofChicago,theBoardofGovernors, ortheFederalReserveSystem. 1

2 THOMASB.KING,TRAVISD.NESMITH,ANNAPAULSON,ANDTODDPRONO mandates—wascentrallycleared.1 TherisksmanagedandposedbyCCPsareasubjectof intenseinterestforfinancialregulatorsandpolicymakers,particularlyascertainlargeCCPs havebeendesignatedtobesystemicallyimportant. We review the central clearing landscape and highlight some financial- stability issues, focusingonthelargeCCPsthatarethe mostcriticalfortheUSfinancialsystem. A robust literaturediscussesCCPs,andwesynthesizesomeofthisworkinouroverview. Wealsomake threenovelcontributions. Firstandmostimportantly,wehighlightapotentialriskofcentral clearing: CCPsmayplacesignificantliquiditystrainsonthebankingsystempreciselyinthe momentswhenthebankingsystemisleastabletobearsuchstrains. Thisriskstemmingfrom CCPs’procyclicalneedforliquiditywasarguablyunder-appreciatedwhenanearlierversion of this paper circulated, but became more widely recognized following the severe market volatilityresultingfromtheCOVID-19pandemic.2 Theincreasedattentionhasfocusedonthe procyclicalityofmargin—forexample,seeFIA(2020)andISDAClearingMemberCommittee (2021)forindustryresponsesandFSB(2021)andBCBS-CPMI-IOSCO(2021)fortheglobal regulatory response—which is only part of the potential risk. CCPs are connected to and dependentupongloballysystemicallyimportantbanks(GSIBs)throughavarietyofchannels inadditiontodirectclearingrelationships. Theseconnectionsinclude: banks’obligationsto postmargintoCCPs;theirrequiredcontributionstoCCPdefaultfunds;and,theirprovision oflinesofcredittoCCPs. TheliquidresourcesthatCCPsdemandthrougheachoftheseand other channels are inherently procyclical with respect to market conditions. The materiality ofthisriskmayhavebeenunder-appreciatedpartlyduetoafocusintheliteratureonOTC derivatives clearing in isolation. Despite the impact of COVID-19 on market volatility, the full scope of the potential liquidity risk has not received as much attention because there were no defaults requiring CCPs to use liquidity resources beyond margin. Second, given thisprocyclicalliquidityrisk,wediscussthepotentialformacro-prudentialliquiditystress teststoimprovethepreparednessofCCPs,GSIBs,andotheractorsbylookingatpotential liquidityneedsacrossthesystem. Third,weexploit“quantitativedisclosure”dataonCCPsin combinationwithmarketvolatilitydata. Thequarterlydisclosuredata,whichmostlargeCCPs beganreportingpubliclyin2015,presentarichsourceofpotentialinformationforresearchers 1Clearing percentages calculated from Bank for International Settlements data at https://www.bis.org/ statistics/derstats.htm. 2Kingetal.(2020)waspublishedjustpriortothemarketvolatilityinearly2020.

CENTRALCLEARINGANDSYSTEMICLIQUIDITYRISK 3 andpolicymakers.3 Althoughthelengthofthedatalimitsanalysis,itislongenoughtoatleast indicaterelationshipsbetweenclearingandmarketvolatility. Attheoutset,itisimportanttorecognizethatinmostcircumstancesCCPsimprovefinancial stabilityrelativetoaworldinwhichtradesarebilateral.4 Thereareavarietyofwaysinwhich theydoso,including: byinsulatingmembersfromeachothers’defaults;bysimplifyingand reducing members’ gross exposures through the netting of positions; by centralizing risk management within a small number of relatively transparent entities; by pooling financial resourcestoaddressextremetailrisks;andbyincreasingthetransparencyandpredictability ofmarketoperations. DeBandtandHartmann(2019)createatypologyofsystemicriskand identifythefirstmechanismthatcanleadtowidespreadinstabilitywithinthefinancialsystem ascontagion—thepotentialtransferofdistress“horizontally”fromonefinancialintermediary ormarkettoanother.5 CCPsareexactlytargetedatmitigatingsuchcontagion,byservingasa “firebreak”tocontaindefaults. Indeed,severalrecentstudies—DuffieandZhu(2011)and Aminietal.(2016),amongothers—findthatCCPsreducecontagionstemmingfrompotential counterpartydefaults. Arguably,CCPsperformedwellduringtheCOVID-19crisis(CCP12, 2020). However,DeBandtandHartmann(2019)alsoidentifyinterconnectednessasapotentialdriver offinancialinstability. LargeCCPsare,virtuallybydefinition,highlyconcentratedandinterconnected. ItisthusworthexploringchannelsthroughwhichCCPscouldamplifysystemicrisk throughinterconnectedness,evenastheydampencontagionrisk.6 Somepreviouspapers(e.g. Faruquietal.,2018)havenotedtheinterconnectednessriskassociatedwithCCPsbeingunable tomeettheirobligationsfollowingmemberdefault. Wehighlightadifferentproblem: the abilityofCCPstofulfilltheirobligationsinstressfulperiodsinvolvestheiraccessingcontingent liquidityfrommembers;suchprocyclicalliquidityriskcanbeasourceofsystemicriskthat couldlimithoweffectiveCCPsareatreducingsystemicriskoverall. Ourdiscussionrecalls Bernanke(1990),who,inthewakeofthe1987stockmarketcrash,expressedconcernsboth 3SeeAlfransederetal.(2018)forfurtherdiscussionofthisdataanditsusefulnessformonitoringCCPs.Weaccess thedatathroughClarusCCPView. 4Forexample,Kiffetal.(2010)arguethatwell-runCCPsreducesystemicrisk,inprinciple,andFSB(2017b,page1) concludesthatreformshavemitigatedsystemicriskinOTCderivativeslargelybecauseofincreasedcentralclearing. 5DeBandtandHartmann(2019)dividesystemicriskintohorizontalrisksthatremaininthefinancialsystemand verticalriskswherefinancialsystemriskscreaterisksintherealeconomy. 6FSB(2017a,2018a)detailsthehighinterconnectednessbetweenthelargestCCPsandeleventotwentylarge financialinstitutions.Barkeretal.(2017)arguethatbanksneedtomodeltheirinterconnectedexposurestoCCPs andthatthemodelingisextremelycomplex.

4 THOMASB.KING,TRAVISD.NESMITH,ANNAPAULSON,ANDTODDPRONO aboutthelimitstoCCPs’abilitytoaddresssystematicrisksontheirownandthepotentialfor CCPstobeasourceofrisk. The potential for CCPs to produce procyclical liquidity strains was recently illustrated bythelargeincreasesinmarginrequirementsduringtheextraordinarymarketvolatilityin early2020broughtonbytheCOVID-19pandemic,BCBS-CPMI-IOSCO(2021)reportsthat dailyvariationmargincalls,whichcoveralreadyrealizedlossesonclearedportfolios,increased around$115billionfromearly2020tothepeakinaggregate. Atthesametime,initialmargin requirements,whichcoverCCPsagainstpotentiallossesifaclearingmemberdefaults,increased $300billioninaggregate. Theserequirementsweremet,butthepotentialstresscouldhave beenworse,asitwasattenuatedbythestrongliquiditypositionofbanks. Critically,therewere nomajorconcurrentdefaults,likeLehmanduringthe2009financialcrisis. Defaultscancreate additionalliquiditydraw,aswasseenintheSeptember2018defaultofalargeclearingmember attheSwedishCCP,NasdaqClearing. Thedefaultoriginatedduethemember’sinabilityto meettheCCPs’liquiditydemands,whichwereinturngeneratedbyextremelylargemarket movements(inthiscase,inthespreadbetweenNordicandGermanPowerfutures). Losses stemming from the default quickly ate through much of the CCP’s prefunded buffer, and survivingmembersfoundthemselvesrequiredtoreplenish€107millionofthatbufferwithina fewdays.7 Whilesomeelementsofthiseventwereidiosyncratic—anditfortunatelyoccurred amidotherwisebenignmarketconditions—thedifficultiesinliquidatingtheportfolioandthe resultingimpactonotherclearingmembersshowthepotentialforliquidityproblemsataCCP toamplifysystemicrisk. TheimplicationisthattheobservedliquiditystrainsatCCPsduring COVID-19didnotfullyreflectthesystemicriskthatcouldhavematerializediftherehadbeen asignificantdefault. ProcyclicalityofmarginrequirementshasbeenstudiedbyMurphyetal.(2014)andGlassermanandWu(2018),amongothers. Regulatorybodieshavealsohighlightedmarginprocyclicality as a concern prior toMarch 2020 (e.g. CPMI-IOSCO, 2012, 2016). But this issue has notbeenexplicitlytiedtosystemicrisk,andwearenotawareofanystudiesthatfocusonthe impactofthisprocyclicalityonbankcondition.8 7For details, see https://www.nasdaq.com/solutions/updates-on-the-nasdaq-clearing-memberdefaultaccessedAugust4,2020. 8Duffie(2014)studyhowcollateraldemandsshiftwiththeintroductionofmandatorymarginandcentralclearing. Similarly,HellerandVause(2012)showthatthecurrentsetofrulescanplacesignificantliquidityburdenson clearingmembers,potentiallycontributingtotheirfailure,andGibsonandMurawski(2013)studybanks’trading behaviorandwelfareunderdifferentmarginregimes.However,theanalysesinthesepapersarestatic;theydo

CENTRALCLEARINGANDSYSTEMICLIQUIDITYRISK 5 Moreover,weemphasizetheadditionaldemandsthatCCPsmayplaceonbanksduring stressfulperiods,beyondmarginrequirements. Mostlargebanks,particularlyGSIBs,have multipletypesofobligationstowardmultipleCCPssimultaneously. Someoftheseobligations aremorelikelytocomedueortobelargerduringtimesofheightenedmarketvolatility. These obligationshavereceivedlittleattentionintheliteratureeventhough,duringseverecrises, theycouldbesubstantial.9 LiquidityobligationstoCCPscouldconcentrateatbanksduringan alreadystressfulperiodandcouldstrainabank’sabilitytomanagetheirliquidity,eitherby makingitchallengingtomaintainregulatoryliquidityrequirementsorinextremisbyexceedinga bank’sfundingcapacity.10 Ineithercase,CCPs’liquiditydemandscouldincreasesystemicrisk, eitherbydecreasingtheliquidityavailabletosupportothermarketfunctionsorbyincreasing thelikelihoodthatpaymentsaremissedoreventhatinstitutionsdefault. Tobeclear,wedonotarguethatthepotentialforprocyclicalliquiditydemandoutweighsthe systemicbenefitsprovidedbyCCPs,orthattheworldwouldbebetteroffifCCPsdidnotmake suchdemands. BothprudenceandregulatorydirectivesleadCCPstobuildstrongerdefenses duringperiodsofgreaterrealizedandanticipatedstress,andsomeamountofprocyclicalityis anunavoidableconsequenceofthedynamicriskmanagementrequiredofCCPs. Nevertheless, CCPsaremostcriticalpreciselywhensystemicstressoccurs,heighteningconcernsoverwhether CCPs’dynamicriskmanagementsufficientlyanticipatessystemicstress. Althoughindividual CCPsregularlystresstestthesufficiencyoftheirownliquidresources,that“micro”approach isnotdesignedtoestimatethepotentialliquidityimpactsacrossmultipleCCPsatatimeof financialstress.11 Thisobservationmotivatesthecallformacroprudentialliquiditystresstests thatlookacrossmultipleinstitutionssimultaneouslywithafocusonthesystemicimpact. (See CPMI-IOSCO(2017)andAndersonetal.(2020).) Suchmacroprudentialstresstestscould provide insights into potential aggregate liquidity demands in response to an extreme but plausiblemarketevent. Attheendofthispaper,wediscusssomeproposalsforincorporating notconsiderhowliquiditydemandsvaryovertimewithfinancial-marketconditions.SidaniusandZikes(2012) estimatethatcollateraldemandfrommarginrequirements(bothinandoutofCCPs)wouldbeapproximately twiceaslargeunder“stressed”marketconditionsasitisunder“normal”conditions. 9MaruyamaandCerezetti(2019)alsostudiestheinherentprocyclicalityofCCPriskmanagement,includingsome oftheadditionaldemandsthatbankscouldface,butdoesnotfocusasmuchonfinancialstability. 10Underextrememarketstress,regulatoryrequirementsmayberelaxed,increasingtheamountofavailability liquidity,butnotcompletelyforestallingthepossibilitythatliquiditydemandscouldexceedcapacity. 11InadditiontoCCPs’ownliquiditystresstests,theU.S.CommodityFuturesTradingCommission(CFTC)andthe EuropeanSecuritiesandMarketsAuthority(ESMA)bothhaveconductedliquiditystresstestingusingtheirown scenariosappliedjointlytosetsofCCPs,buttheobjectivehasremainedfocusedonthemicroobjectiveofassessing individualCCPs’resilience(seeCFTC(2017)andESMA(2018a)respectively).

6 THOMASB.KING,TRAVISD.NESMITH,ANNAPAULSON,ANDTODDPRONO CCPliquidityintoamacroprudentialstress-testingframework. Insightsfromsuchtestscould becrucialinmoreaccuratelyassessingwhetherCCPsandthebroaderfinancialsystemcould weatheranyprocyclicaldemandsforliquidityemanatingfromcentralclearing. Inaddition, most bank stress-testing implementations (e.g., those surveyed in Borio et al. (2014) and Schuermann(2014))havenotincorporatedliquidityexposurestoCCPs. AfterCCPmacroprudentialstresstestingmatures,integratingbankandCCPmacro-prudentialstress-testing shouldbeconsidered. OurdiscussionoftheprocyclicalityofCCPdemandsfitsintoabroaderliteratureshowing how synchronized liquidity needs in the financial sector can be destabilizing and lead to systemic risk.12 Adrian and Shin (2008, 2010) show that intermediaries broadly decrease leverageduringperiodsoffinancialstress,andtheyarguethatthisbehaviorisindicativeof procyclicalliquidityinthefinancialsystemasawhole. Allenetal.(2009),Acharyaetal.(2010), andHeideretal.(2015)presentmodelsinwhichbankdemandforliquidityisinefficientlyhigh duringcrises,contributingtoinstabilityoftheinterbankmarket. Ashcraftetal.(2011),Acharya andMerrouche(2012),andBerrospide(2013)documentthatbanksdidindeedhoardliquidity duringthe2008financialcrisis.13 Theseshiftsinthedemandforsafeandliquidassetscancause financialinstitutionstofire-salesecurities(ShleiferandVishny,2011,Greenwoodetal.,2015) andreducelending(Cornettetal.,2011,Iyeretal.,2014). Theincreasingcontingentobligations ofbankstoCCPsconstituteanadditionalsourceofliquiditypressure—likelycorrelatedwith bothassetandliabilitydrawsduringstressfulperiods—addingtothefundingpressuresfaced bybanksduringstress.14 Section2ofthispaperprovidesabriefoverviewoftheU.S.CCPlandscapeanddiscusses howCCPsfitintothebroaderfinancialsystem. Section3reviewsthebasicsofCCPoperation 12ThecontingentobligationsofbankstoCCPsalsohaveimplicationsforsystem-wideliquiditymeasurement.For example,thoughpotentiallylarge,thesecommitmentsdonotgenerallyfactorintomeasuresofaggregateliquidity conditionsthathavebeenproposedintheacademicliterature(BergerandBouwman,2009,2017,Baietal.,2018). 13Cornettetal.(2011)andIppolitoetal.(2016)emphasizethatbankexposuretoliquidityriskduringstressful periodscancomebothfromcreditordemandsandfromoff-balance-sheetcommitments,suchaslinesofcredit. Whiletheabilityofbankstoprovideliquidityonbothsidesoftheirbalancesheetshastraditionallybeenseen asastabilizingforce(Kashyapetal.,2002),thisisonlytruetotheextentthatliquidityoutflowsareimperfectly correlated.IvashinaandScharfstein(2010)showthatdrawsonlinesofcreditoccurredduringthecrisisatprecisely thetimethatbanksalsofaceddifficultyrollingovertheirliabilities. 14Similareffectsextendoutsideofthecommercialbankingsector. Inparticular,liquidityshortagesmayaffect broker-dealers’willingnesstoprovidesecuritiesfinancingtoclients. BreachandKing(2018)demonstratethis relationshipempirically.InBrunnermeier(2009),increasesinthecollateraldemandedbydealersforsecurities financingleadstoasset-pricevolatilitythatfeedsbackintofundingconditionsinadestabilizingspiral. Tothe extentthatdealersexperienceoranticipateextraordinaryliquiditydemandsfromCCPs,theymayrespondby makingfundingconditionsmorerestrictive,increasingthelikelihoodofsuchaspiral.

CENTRALCLEARINGANDSYSTEMICLIQUIDITYRISK 7 andriskmanagement. Section4discussesvariouswaysthatCCPriskmanagementcanextract liquidityfromtherestofthefinancialsystemduringtimesoffinancialstress. Section5describes theroleofliquiditystresstestsatboththeCCPandmacro-level. Section6concludes. 2. TheUSCCPLandscape AscatalogedbyKroszner(1999,2000),CCPsdevelopedintheUStoaddresscounterparty risksinresponsetofinancialcrises. Indeed,theessentialfunctionofaCCPistransforming counterpartyexposure. ACCPstepsintoeachtradebetweenitsclearingmembers. Bydesign, aCCPisaveryrestrictedcounterpartyasitdoesnottakeanypositionsonitsownbehalf. By beingthe“buyertoeverysellerandsellertoeverybuyer,”aCCPalwaysmaintainsaperfectly matched book and therefore takes no direct market risk in the markets it clears. However, counterpartyriskisconcentratedattheCCP,andtheCCPisexposedtocontingentmarketrisk uponthedefaultofaclearingmember,becauseitacquiresthedefaultingmember’spositions. ACCPthenhastotakestepstoreturnitselftoamatchedbookandflatmarket-riskposition. (WediscussCCPriskmanagementinthenextsection.) CCPsfaredwell,andservedtheirroleasabufferagainstdefaults,duringthe2008financial crisis. Forexample,fromValukas(2010),CMEclosedouttheclearedderivativesportfolio of Lehman Brothers, which had a net value around $21billion in May 2008, within a few days;Lehman’smarginwassufficienttocoverauction-relatedlossesamountingto$1.2billion. Similarly,atLCHSwapClear,Lehman’sinterestrateswapportfolio,withanotionalvalueof $9trillion spread over 66390 trades across five currencies, was closed out by early October (LCH.Clearnet,2008). AsdiscussedinFlemingandSarkar(2014)andWigginsandMetrick (2019),centralclearingdidfacesomechallengesandconsequentcriticism. Yet,despitethis stress,asshowninFigure1onpage30,thepercentageofclearedIRStrades(theblueline) jumpedfromlessthan25%in2007tonearly50%by2009,astraderssoughtthesafetyofCCPs duringthecrisis. In2009,theG-20leaderscommittedtoclearingallstandardizedOTCderivativescontractsthroughCCPs. ThegoalwastoaddressrisksinthebilateralOTCderivatives’ markets,includinglargeconcentrationsofcounterpartyexposures,inconsistentriskmanagement,ascarcityofprefundedresourcestocoverrealizedlosses,alackoftransparency,and adversefeedbackloops(e.g.,marginspirals). Subsequently,asshowninthefigure,clearingof IRSroseto75%.

8 THOMASB.KING,TRAVISD.NESMITH,ANNAPAULSON,ANDTODDPRONO Intheyearssincethe2008crisis,avarietyoffactorshaveledtofurtherincreasesinclearing volumes for both exchange-traded and OTC derivatives. The most direct impetus was the establishment of central clearing mandates for OTC derivatives noted above.15 In Figure 1, for example, the jump in cleared interest-rate swaps from about 50% in 2012 to near 75% by2014iscoincidentwithmandatescomingintoforce. Similarly,CDSclearing,whichbegan atICC(formerlyICETrust)inMarch2009andatICEClearEurope(ICEU)inJuly2009,has steadilyclimbedtonearly40%by2017accordingtoBISdata(notshowninthefigure).16 This increasecamedespitedelaysinimplementingclearingmandates. ForbothIRSandCDS,over 80% of new dollar- denominated trades are now centrally cleared. Some of the growth in centralclearinglikelyreflectsanincreasedappreciationfortherisk-mitigatingfunctionofCCPs followingthecrisis. Italsolikelyreflectsmarketmovesawayfrommoreexoticandbespoke tradestomorestandardizedones. Theincreasedclearingratemeansthatclearingvolumeshave heldroughlysteadyevenasnotionalamountsfellbyathirdinthelastfewyears. Morerecent supportforcentralclearinghasbeenprovidedbybilateralmarginrequirements;FSB(2018b) foundthatOTCderivativeclearingdramaticallyacceleratedintermsofnotionalclearedupon theimplementationofbilateralmarginrequirements,evenfortradeswherecentralclearing wasnotrequired. Thesedevelopmentshelpedtocontributetoamoreresilientfinancialsystem thatwasabletoweathertheCOVID-19marketstress(FSB,2021). In the U.S. six CCPs clear the most important financial markets; these are listed in Figure 2. Five of these CCPs are designated as systemically important in the US. The sixth, LondonClearingHouseSwapClear(LCH),isaUKCCPbutclearsasubstantialamountof USdollar-denominatedinterestrateswaps. ThesesixCCPscanbeclassifiedasclearingsecuritiesandderivatives,respectively. ThesecuritiesCCPsincludetheFixedIncomeClearing Corporation(FICC),whichoperatestwoseparateclearingservices: oneforUSTreasuriesand repurchaseagreements(repos),andoneformortgage-backedsecurities(MBS).TheNational Securities Clearing Corporation (NSCC) also clears securities; it is the primary clearer for USequitymarkets,andalsoclearscorporateandmunicipalbonds. ThesesecuritiesCCPsare allpartoftheDepositoryTrustClearingCorporation. DerivativesCCPsinclude: theOptions ClearingCorporation(OCC),whichistheprimaryclearerforUSequityoptions;theChicago 15Culp(2010)reviewstheregulatoryhistoryofOTCderivatives’clearingthroughtheDodd-Franklegislation.FSB (2018b)analyzedtheincentivestocentrallyclearcreatedbyvariousreformsbeyondjustclearingmandates. 16CMEalsostartedclearingCDSin2009,butterminatedtheserviceinearly2018.

CENTRALCLEARINGANDSYSTEMICLIQUIDITYRISK 9 MercantileExchangeInc.(CME),whichoperatestwoseparateclearingservices,oneforfutures andoptions(Base),andoneforinterestrateswaps(IRS)andswaptions;ICEClearCredit(ICC), whichclearscreditdefaultswap(CDS)indexesandsinglenames;and,LCHSwapClear,which alsoclearsIRS.DerivativesCCPscanfurtherbedividedbywhethertheclearedderivativesare exchange-tradedorOTC.OCCandCME’sBaseserviceclearexchange-tradedderivatives. The remainder—CME’sIRSservice,ICC,andLCHSwapClear—clearOTCderivatives.17 Measuring the size of CCPs is not entirely straightforward, but Figure 2 provides some context. Thefirstnumericcolumnpresentsapproximatetotalprefundedresourcesin2019; althoughsubstantial,theresourcesalsojumpedsignificantly—forNSCCnearlytripling—inthe firstquarterof2020asshowninthesecondnumericcolumn. Thelasttwocolumnspresents themaximumdailycallformarginsince2015priorto2020andduringthefirstquarterof 2020respectively. Again,thepre-2020peakswerelarge,butweredwarfedbythecallsduring theCOVID-19marketstress. Notethatthemeasuresofsizeusedherelikelyunderstatethe economicimpactofsecuritiesclearinghouses,becausetheydonotcapturethetotalcashthat flowsthroughsecuritiesCCPs,whichissubstantialduetothesettlementdemandsforcash. Despitemeasurementchallenges,clearly,alltheCCPshavesignificantresourcesavailable,but alsoallhavemadesignificantcallsforadditionalfundsfromthemarket,bothbeforethemarket shockinearly2020andevenmoresoduringit. Thispatternisconsistentwiththeanalysisin BCBS-CPMI-IOSCO(2021). 3. CCPFunctionandRiskManagement TheconcentrationofcounterpartyexposureataCCPmeansthesuccess,orfailure,ofitsrisk managementcontrolscouldhaveprofoundimplicationsforthemarketsitclears. Successmeans defaultsarenotamplifiedwithpotentialsystemicimplications; failure, althoughdesigned tobeanextremelyremotepossibility,potentiallycreatescontagioneffectsbetweenclearing membersthatwouldnototherwisehavebeenexposedtothedefault. CCPsmanagetheircredit 17UndertheDodd-Franklegislation,theSecuritiesandExchangeCommission(SEC)istheSupervisoryAgencyfor FICC,NSCC,andOCCwhiletheCFTCistheSupervisoryAgencyforCMEandICC.BecauseICCisregistered withtheSECasaclearingagency,theSECalsohasauthorityoverallofitsclearingservices.TheFederalReserve hasauthoritytoparticipateinthedesignatedsupervisor’sexaminationsandreviewsofmaterialchangesatthe designatedCCPs.LCHSwapClearisregisteredwith,andsupervisedby,theCFTCfortheclearingofUSdollar denominatedpositions;TheFederalReservealsoparticipatesininternationaloversightofLCHSwapClearunder theauspicesoftheBankofEngland.

10 THOMASB.KING,TRAVISD.NESMITH,ANNAPAULSON,ANDTODDPRONO andliquidityrisksdifferentlythanbanks, partlyreflectingthatCCPsdonotactivelytrade financialcontracts,butrathermanagetherisksgeneratedbyparticipants’tradingactivity.18 CCPriskmanagementgenerallyincludesmembershipeligibilityrequirements,thenetting ofexposures,marginrequirements,mutualizedfinancialresources(usuallyinadefaultorguarantyfund),anddefault-managementprocedures. SeeMurphy(2013)forafullerdiscussion. CCPsmayalsohaveadditionaltoolstoemployinrecoveryscenarios,suchasassessmentpowers,variationmarginhaircutting,contracttear-ups,orlossallocationpowers(CPMI-IOSCO, 2014,Sec. 4). Inadditiontothemanagementofcreditrisk,CCPsalsomanageliquidityrisk, andhaveavarietyoftoolstoaddressliquidityneeds,suchasaccesstocollateralmarketsand pre-arrangedrepolines. Withaneyetowardsthesubsequentdiscussionofprocyclicality,we reviewtwoofthemostimportantelementsofCCPriskmanagement: marginandmutualized financialresources,whichrepresentpre-fundedresourcesheldbytheCCP.Bothresources arecriticaltomanagingandmitigatingaCCP’scontingentriskexposuresandaretwomajor driversofprocyclicalityinCCPs’resourcedemands. MarginisacriticalcomponentofCCPriskmanagement. Itcanbedividedintovariation margin(VM)andinitialmargin(IM);Figure3illustrateshowVMextinguishesaCCP’scurrent exposureandIMcoverspotentialfutureexposure. Wediscusseachtypeofmargininturn. Variationmargincoverstherealizedchangeinaclearedposition. Regulationsgenerally require CCPs to collect and pay out in cash the daily change in value that each member’s portfolioexperiencesasVM.19InthestylizedexampleinFigure3,thechangeinvalueofthe portfolio at the end of the day is negative. The clearing member must pay this amount to theCCP.ThetotalvalueofthepaymentsofVMexactlyequalstheamountowedtoclearing memberswhoseportfoliosgainedinvalue. Bymarkingeveryportfoliotomarketdailyand exchangingcashtocoverthechanges,theCCPresetsitscurrentexposuretozeroeverydayand preventsexposuresfromaccumulating.20 TherequirementtopayVM,however,alsocreatesa pointoffailure;ifaclearingmemberfailstomakeaVMpaymentforitselforitsclientsinthe 18Intheeventofaclearingmemberdefault,aCCPmaytradeinordertohedgeandliquidatethedefaulter’sportfolio, butusuallyneedstosecondtradersfromitsclearingmemberstodoso.Forfurtherdiscussionofthedifferences betweenbanksandCCPs,particularlythedifferentrolesplayedbycapitalandcollateralintheirrespectiverisk managementapproaches,seeManningandHughes(2016)andCoxandSteigerwald(2018) 19Thereareexceptions.Inparticular,FICC’sclearingstructureformortgage-backedsecuritiescollectsVM,butdoes notpassitthrough.OCCcollectslittleVM,becausethecoveredoptionspositionsitclearsaregenerallyhedgedby theunderlyingsecurities. 20Inaddition,VMissometimesexchangedintra-dailytomoreactivelylimitthebuildupofcurrentexposuresthat mightoccurmorerapidly,forexample,inmorevolatilemarkets.

CENTRALCLEARINGANDSYSTEMICLIQUIDITYRISK 11 timerequired,theCCPcan,andlikelywill,declaretheclearingmembertobeindefault.21 As emphasizedbyMaruyamaandCerezetti(2019),VMisthemaindriverofCCPmargincalls. TherequirementtomeetVMeffectivelytestseachclearingmember’sperformanceatleastdaily, whichinsomesensetransformsthecreditexposureitfacedwithitsoriginalcounterpartytoa liquidityexposuretotheCCP(Cont,2017). Thoughnotwithoutitsownrisks(aswediscuss indetailbelow),theliquidityexposurehasthevirtueoftransparency,sinceitdependsdirectly onobservablemarketmovements,ratherthanondifficulttoobservecounterpartyactions. Initial marginis designed to cover the potential future exposure of a clearing member’s portfoliofromthetimeofthelastVMpaymentuntiltheportfoliocouldbeliquidatedina default. Becausethelossthatcouldberealizedonaportfolioisuncertain,IMtargetsahigh quantile—atleastthe99thpercentile—ofmarketmovesoverthespecifiedclose-outperiod(also sometimescalledthemarginperiodofrisk). InFigure3,IMcoversboththeVMthatwasnot paidduringadefaultplusadditionallossesonthepositionthatcouldberealizedduringthe close-out. Attheircore,CCPs’IMestimatesforaportfoliobearsimilaritiestostandardmarket riskcalculations,likevalue-at-riskorexpectedshortfall. However,regulatoryrequirements alsospecifythatIMcoversotherexposuresthataremoredifficulttoquantify,likeestimatesof portfoliotransactioncosts,jump-to-defaultexposures,andconcentrationrisks. Heckingeretal. (2017)reviewbothhistoricalandcurrentIMpractices. LookingacrosscertainderivativesCCPs forfuturesandswaps,asshowninFigure4thesizeofIMrequirementsmorethandoubled fromlessthan$194billionattheendof2013toover$425billioninAugust,2019;thatsteady increasewasmatchedbyroughlyanadditional$200billioninMarch2020aloneandattheend of2021IMismorethantriplethatheldin2013attheseCCPs.22 ThisgrowthinIMimplies growthintheamountofriskCCPsaremanagingalthoughnotindirectlockstep. Tocoverpotentiallossesbeyondmarginintheeventofadefault, CCPsrequireclearing memberstoprovidemutualizedfinancialresourcessufficienttomeetaspecificcoveragetarget inawiderangeofextremebutplausiblestressscenarios. Themutualizedresourcesaresized sothattheCCPhassufficientfinancialresourcestocoverthedefaultofeitheranysingleorany 21ArmakollaandLaurent(2017)showstheimportanceofmembers’abilitiestomeettheirobligationsforthe resilienceofaCCP. 22ThedataincludesrequirementsfrommorethanthesystemicallyimportantCCPs.BesidesICEU,whichclears energyfuturesandoptionsinadditiontoCDS,andotherpartsofLCHLtd.besidesSwapClear,requirementsfrom ICEUS,whichclearsavarietyoffuturesandoptions,andLCHSA,whichistheEuropeansitedcounterpartto LCHLtd.intheUKareincluded.

12 THOMASB.KING,TRAVISD.NESMITH,ANNAPAULSON,ANDTODDPRONO pairofclearingmembersinstressedmarketconditions.23 AtmostCCPs,mutualizedresources aremaintainedinaseparatefund,calledadefaultorguarantyfund. TheUSsecuritiesCCPs insteadmutualizeallIM.Ineithercase,theadequacyoftheseresourcesistesteddailybythe CCPthroughitsstress-testingprogram,whichisrequiredtoapplyavarietyofextremebut plausiblescenarios: bothhistoricallyobservedmarketstressesandhypotheticalscenarios. Tocoveranylossesfromadefault,aCCPwoulduseprefundedresourcesinaprescribed sequence,oftencalledtheCCP“waterfall.”Thedefaulter’smarginabsorbslossesfirst. Ifthe defaulter’smarginisexhausted,anyadditionalresourcesprovidedbythedefaultingparty,such asitscontributionstoadefaultfund,areapplied. Ifthedefaulter’sresourcesareexhausted, beforemutualizinglossestosurvivingclearingmembers,CCPstypicallyapplyaportionof their own capital—known as “skin-in-the-game”—to cover remaining losses. Mutualized resourceswouldbeappliedtoanyremaininglosses. CCPsarerequiredtohaveexplicitrules and procedures that address how potentially uncovered credit losses would be allocated. Theymayalsohaveadditionalprescribedassessmentpowerstoseekfurtherresourcesfrom non-defaultingmembers. ThestrongandconsistentriskmanagementpracticesadoptedbyCCPshavebeenenhanced bystrongerregulatoryexpectationspromulgatedinthewakeofthe2008financialcrisis. This strengtheningwasfacilitatedbythedevelopmentofinternationally-agreedrisk-management standardssetoutinthe”PrinciplesforFinancialMarketInfrastructures”(PFMI).24TheprinciplesinthePFMIsetseveralexpectationsforCCPs,includingeffectivelymanagingalldimensionsofitscreditandliquidityrisks,employingarobustmarginingsystem,andmaintaininga minimumleveloffinancialresourcestocoverpotentiallossesandhonorpaymentobligations, bothinextrememarketconditions. TheprinciplesaredesignedtoensurethatCCPswillhalt contagionamongtheirmembersandmitigatesystemicriskacrosstheinterconnections. 4. PotentialProcyclicalResourceDemand AnyreductioninsystemicriskaffordedbycentralclearingdependsonCCPsperforming asdesignedinstressedmarkets. Asnotedabove,CCPshavegrownbothinsizeandinthe scopeoftheproductstheyclear. Furthermore,individualsystemicallyimportantbankstendto 23Specifically,intheUS,FICC,NSCCandOCCarerequiredtocoverthesinglelargestdefaultloss,whileCMEand ICCarerequiredtocoverthelargesttwodefaultlosses,asisLCHSwapClear. 24SeeCPMI-IOSCO(2012).

CENTRALCLEARINGANDSYSTEMICLIQUIDITYRISK 13 participateinmultipleCCPs,inordertoaccessdifferentmarketsacrossdifferentjurisdictions. Forexample,therecentanalysisinFSB(2018a)showsthatoutofasetof26majorglobalCCPs, the11largestclearingmembersintermsofaggregateresourcesaremembersofatleast16of the26withamedianparticipationof22. Manyofthesesamelargefinancialinstitutionsalso provideotherservicestoCCPs,suchassettlementandinvestmentservicesandlinesofcredit. Thisinterconnectedness,togetherwithCCPs’criticalfunctionssupportingmarkets,naturally raisestheneedtomonitortheirimpactonfinancialstability. CCPs’risk-managementcontrolscanaffecttheotherfinancialinstitutionsexpectedtomeet oneormoreresourcedemands,particularlyifCCPresourcedemandscomeattimesofheightenedfinancialstress. Thesedemandsinclude: (1)variationmargin;(2)initialmargin;(3)settlementrequirements;(4)defaultfundcontributionsandassessments;(5)linesofcreditand otherliquidityarrangements;and(6)capitalandliquiditytoabsorbthepositionsofmembers and their clients in the event of default. In this section, we review each of these potential resourcedemandsinturn. Fromasystemic-riskperspective,thereisatrade-offbetweenrigorouslymanagingtherisk of positions cleared through CCPs and minimizing interconnected liquidity strains on the system. ThefailureofalargeCCPwouldbeadevastatingeventfortheglobalfinancialsystem; itwouldlikelybeaccompaniedbymassivedisruptionsinfinancialmarketsandthesevere distressorfailureofmanyothersystemicallyimportantinstitutions. Fromafinancial-stability perspective,itisimportantforCCPstomaintainfinancialresourcessizedtocoverpotential lossesinawiderangeofstressscenarios. Onehypotheticalwaytoachievethiswouldbefor CCPstoprefundallofthesefinancialresourcesexantebyholdingverylargequantitiesofsafe andliquidassetsatalltimes. Suchanapproachwouldtieupliquidityandcollateralinaccounts that,mostofthetime,wouldbewellinexcessofanyplausiblelossthatCCPsmightexperience. Moreover,therequirementtokeeplargeamountsofassetsparkedatCCPscontinuouslywould imposecostsonclearingmembers;thosecostscouldwellinducebanksandtheirclientsto moveactivityintounclearedpositions(eliminatingthestabilitybenefitsofcentralclearing) ortoexitpositionsaltogether,whichcouldresultinlesshedgingactivityandareductionin overallmarketliquidity. Ratherthanholdinganextremelylargebufferoffinancialresourcesall thetime,CCPsholdsmaller(thoughstillsignificant)amountsofliquidresourcesduringcalm periodsandincreasetheseresourcesduringtimesofhighfinancialmarketvolatility,when

14 THOMASB.KING,TRAVISD.NESMITH,ANNAPAULSON,ANDTODDPRONO theirmarketandcounterpartyrisksincrease. Theincreasemustbemetbyawebofbanks andotherfinancialinstitutions,whichimpliesanincreaseintheriskassociatedwiththeCCPs interconnectedness. Importantly,themostcommonresourcedrawsthatcomefromCCPs,margincalls,arenot uniquetoclearingrelationships. PriortoimplementationoftheDodd-FrankActandsimilar regulatorychangesabroadmanybankarrangementswithnon-bankcounterpartiesincluded provisionsforIM,andformostderivativeexposurestherewassomeperiodicexchangeofVM. However,thesearrangementsweresomewhatadhocandflexible. Withcentralclearing,margin requirementsareuniversalforclearedcontractsand,toalargedegree,standardizeddueto consistentregulatoryexpectations.25 Nevertheless,becauseclearingtakesplacethroughasmallnumberofCCPs,themodels andrulesusedbyanyoneCCPcanhavesignificantimplicationsforalargenumberofmarket participants. Inaddition,muchoftheactivityatCCPsisconcentratedinafewlargemembers. TheconcentrationisillustratedinFigure5,whichshowsthatthefivelargestclearingmembers alsoaccountforthemajorityofIMrequirementsfacingclientsforbothdollar-denominated clearing of both futures and swaps; in fact, at one point the five largest clearing members accounted for nearly three-quarters of clients’ swap IM requirements, before declining to around60%. TheconcentrationassociatedwithCCPsandtheirlargemembersimplythat margincallsarelikelytocomeinamorecoordinatedmannerinaworldwithcentralclearing, reflectingstrongerinterconnectedness. Furthermore,asdiscussedbelow,CCPsalsoimpose othertypesofobligationsontheirclearingmembers—obligationsthathavenocounterpartsin bilateraltrading—andthesearemorelikelytobetriggeredinthesamestatesoftheworldin whichmargincallsarelarge. 4.1. Variationmargin. Asnotedearlier,VMistheamountthataclearingmembermustpost totheCCPtocovermarked-to-marketchangesinthevalueofitsportfolio. Maruyamaand Cerezetti(2019)illustratehowVMisthemargincomponentthatismostsensitivetomarket volatility. Becausechangesaregreatestondayswhenmarketpricesmovemost,theamount ofVMpaidtoCCPsnecessarilyincreasesduringtimesofhighrealizedvolatilityinfinancial markets. Figure6illustratesthisrelationshipbyshowingthepeakamountofVMcollectedin 25Marginrequirements,bothforVMandIM,havealsobeenestablishedforbilateraltrades(BCBS-IOSCO,2015); foranalysisoftheirimpactontheincentivestocentrallyclear,seeFSB(2018b).

CENTRALCLEARINGANDSYSTEMICLIQUIDITYRISK 15 eachquarterbytheCCPsthroughwhichU.S.banksclearmostoftheirtrades. Thismarginis plottedagainstthepeakintherealizedvolatilityofthestockmarketineachquarter. (Although thedatadonottellusforcertain,itislikelythatthepeaksinthetwoseriesoccuronthesame daysofeachquarter.) ThiscomparisoniscrudebecausesomeCCPsdonotclearproductsthat aredirectlylinkedtoequitymarkets. Nevertheless,thetendencyofvolatilityinmostmarkets tomovetogethermakesthecomparisoninformative,andthepatternisclear. Notsurprisingly, thepeakinbothisduringthefirstquarterof2020. TomeetVMcalls,clearingmembersmustmakepaymentstotheCCPinashortamountof time,ofteninaslittleasonehour,andthesepaymentsgenerallyaremadeincash. Furthermore, although clearing members who clear on behalf of clients pass through VM calls on client positions,thispass-throughsometimesdoesnotoccuruntilthefollowingday,andtheclearing memberisitselfresponsibleforensuringthattheproperamountofmarginispostedtocover their clients’ losses. Clearing members and their clients expect to make (or receive) VM daily;however,becausetheyreflectchangesinassetprices,thesizeofVMcallsareessentially unpredictable. Toaddressthisrisk,clearingmembersmaintainreservesofcashorhaveaccess tofundingmarketsthatcanbedrawnonwithveryshortnotice. Evenintimeswhenfinancialinstitutionsarenotunderseverepressure,VMcallsdueto spikesinvolatilitycanbeburdensome. Eventssurroundingthe“Brexit”referenduminthe U.K.inJune2016provideanexample. Thereferendum,theresultofwhichsurprisedmany, causedaspateofvolatilityinfinancialmarkets,particularlythoseinvolvingtheBritishpound, theeuro,andassociatedfixed-incomeproducts. LCH,whichclearsmanysuchinstruments, calledforlargeamountsofVMinthewakeofthisepisode. Figure7showsthatonthepeakday duringthesecondquarterof2016,VMpaymentstoLCHtotaled$16billion,incontrasttothe dailyaverageofabout$3billion. ThesecallscamesimultaneouslywithsmallerVMcallsfrom otherCCPs;CFTC(2016)foundthatLCHtogetherwithfourotherCCPscalledfor$27billion overthetwodaysfollowingthereferendum.26 ThepeakdemandcausedbyBrexitwasmuch higherthantheelevatedaveragedemandduringCOVID-19,eventhoughmuchlargercalls weremadeatLCHinthelaterperiod. 26LCHwasalsoslowertomakepaymentstomemberswithpositiongainsthanitwastocollectfrommembers withlosses,creatingfurtherliquiditypressure;somemarketparticipantsseemedtofindtheobligationtopostthis marginonerous,despitethefactthatthefinancialsystemitselfdidnotappeartobeunderaparticularlyhighlevel ofpressure(Madiganetal.,2016).

16 THOMASB.KING,TRAVISD.NESMITH,ANNAPAULSON,ANDTODDPRONO Undermostcircumstances,VMpaymentstoCCPsdonotreduceaggregateliquidresources availabletomarketparticipants,becauseCCPsrunmatchedbooks,sothateverymarked-tomarketlosstheyfaceisaccompaniedbyanequalmarked-to-marketgain. Thus,VMoughtto bepassedthrough,dollarfordollar,fromonesetofclearingmembers(thosewithnetlosses) toanother(thosewithnetgains).27 Evenso,thetransferoffundscancausestressforthose withunanticipatedliquidityoutflows. Furthermore,particularlyintimesofstress,disruptions anddelaysinpaymentchainscanoccur,potentiallypreventingtheliquidassetsfromflowing totheirrecipientsinatimelymanner. CCPsandsettlementbanksthemselvesmaydeliberately slowthetransferofliquidityoutofafearthattheirownliquidresourcesmayproveinadequate. Thestock-marketcrashof1987,forexample,wasaccompaniedbyanumberofdisruptionsin thepaymentssystem,includingthereluctanceofsettlementbankstoprovidelargeamountsof intradaycredit. Asaresult,VMpaymentstoclearingmembersatCMEandOCCweredelayed, addingtothestressofmarketparticipantsonanalreadyvolatileday(PresidentialTaskForce onMarketMechanisms,1988). DuringtheCOVID-19stress,someclientsdidreportneedingto conductrepotransactionsandassetssalestomeetmargincalls(BCBS-CPMI-IOSCO,2021), whichmayhaveaddedadditionalstrainonfundingmarkets. Finally,extremecircumstancesmayalsopreventclearingmembersfromreceivingthefull VMduetothem. Inparticular,ifaCCP’sresourcesshouldbeexhaustedbyclearingmember defaults,oneoptionsomeCCPshavetomanageliquidityisto“haircut”theVMitpaysout. Inthiscase, theCCPwouldabsorb, atleasttemporarily, partoftheVMitreceivedonloss positions. As another option in such situations, some CCPs may pay out VM in securities collateral,ratherthancash,puttingtheburdenofliquiditytransformationontomembers. 4.2. Initialmargin. InitialmarginispostedtoCCPstoaccountforpossibledeteriorationinthe valueofclearingmemberpositionsthatmightoccurbetweenthetimeofadefaultandthetime thedefaultingmember’spositionscanbeliquidated. LikeVM,IMgenerallyincreasesduring timesoffinancial-marketvolatility. AsdiscussedinMurphyetal.(2014),therequirements to cover a quantile of market moves requires IM models to be risk-sensitive, and therefore procyclical,toatleastsomedegree. However,whilethepass-throughofvolatilitytoVMis largelymechanicalandoutoftheCCP’scontrol,thefactorsdeterminingIMaremorecomplex 27Therearesomeexceptionsforintradayvariationmargincalls.

CENTRALCLEARINGANDSYSTEMICLIQUIDITYRISK 17 anddependonthespecificpracticestheCCPfollows. Thesepracticesarelargelyspelledoutin CCPrulebooks,buttheyalsoinvolveelementsofdiscretion. As a baseline, CCPs maintain models of the value that positions could lose with some confidence: typically,99%or99.5%overaspecifiedliquidationperiod(typicallyafewdays). Thesetaillossesarealmostalwaysgreaterwhenmarketvolatilityisexpectedtobehigher, andconsequentlyIMtendstoincreaseinperiodsofhighrealizedandimpliedvolatility. As anexample,Figure8plotstheamountofIMthatCMEhasrequiredonanS&P500futures contract,asapercentageofthecontractvalue,withtheVIXindexofimpliedvolatilityonthe S&P.Itisclearthatthesetwoseriesmovecloselytogetherparticularlyduringlargespikesin impliedvolatility. ThestrengthofthisrelationshipisfurthershowninFigure9,whichplots thesamedataagainsteachother;thesignificantpositiverelationshipcanbeseenbytheslope oftheregressionline.28 ThereddotsshowthebehaviorduringMarch2020astheIMlevel increasedasVIXspiked,butcontinuedtodosoevenasVIXfellback. AsanotherexampleoftheconnectionbetweenvolatilityandIM,Figure10showspeakIM calls at the OCC, which primarily clears equity options; the jumps in equity volatility that occurredinFebruary2018,December2018,andMarch2020areclearlyidentifiedascausing correspondingjumpsinIMduringthosequarters. Althoughitisrelatedtovolatility,thetotalamountofIMrequiredofaclearingmember is,inmostcases,morecomplicatedthanthedirectrelationshipssuggestedbyFigures8to10. ThenetpositionsthatclearingmembersandtheirclientsmaintainwithCCPsconsistoflarge arraysoffinancialinstrumentsthatareexposedtodifferenttypesofrisk. CCPmarginmodels attempttoaccountforthecomovementofdifferentassetsandemployavarietyofmodeling techniques to capture the tails of the joint distributions of returns. In addition, the overall levelofmarginchargedoftenincludescertain“add-ons,”whichtendtobelesssensitive(to varyingdegrees)tomarketvolatility(CPMI-IOSCO,2016). And,ofcourse,thetotalamount ofIMrequiredtendstoriseandfallwiththesizeofparticipants’positions. Figure11shows quarterlychangesinthetotalquantityofIMrequiredbythemajorUSCCPs,plottedagainst contemporaneous changes in the VIX. Although the figure suggests a correlation between volatilityandIM,thatcorrelationappearsmutedrelativetotheonebetweenvolatilityandVM. Potentialexplanationsforthisdampenedrelationshipincludetheanti-procyclicalitytoolsthat 28ThelineiscalculatedbyanM-typeestimatorwithHuber’sTnorm(Huber,1981)tomaketheregressionrobust againstthelargepositiveoutliers.

18 THOMASB.KING,TRAVISD.NESMITH,ANNAPAULSON,ANDTODDPRONO havebeenintroduced;ESMA(2018b)codifiessomeanti-procyclicalitytools,whichhavealso beenappliedtoUSCCPsduetotheirglobalactivity. Examplesoftoolsincludeabufferthat allowsforaproportionofmargintobetemporarilyexhaustedfollowingasignificantincrease inmarginrequirements,whichlessensthesizeofanyIMcallsthatresultfromsuddenspikes involatility,andafloorthatlimitstheamountthatmarginrequirementscandeclineduring tranquilperiods,therebymutingincreasesthatresultwhentranquilitygiveswaytoturbulence. Despitethesecaveats,BCBS-CPMI-IOSCO(2021)findsthatmarketvolatilitywasthemajor driverofIMrequirementsduringthesevereCOVID-19stress. AnotherreasonthatCCPIMrequirementsmaynotbedirectfunctionsofmarketvolatility isthatCCPsmayattachmarginsurchargestoparticularclearingmemberstocompensatefor concentrationorcounterpartyrisk. Thoughcounterparty-risksurchargesarerare,theycan resultinsuddenlargeincreasesintheamountofIMmarginrequiredfromfinancialinstitutions thatarealreadyindistress. Forexample,asdiscussedbyHeckinger(2014),amidconcerns abouttheriskassociatedwithitsrepopositions,thefuturescommissionmerchantMFGlobal wasdowngradedbyMoody’sandStandard&PoorsontheweekofOctober24,2011. Over the following few days, LCH (through which these repo positions were cleared) imposed counterparty-risksurchargesthatapproximatelydoubledtheIMthatMFGlobalwasrequired tohold. Theresultingmargincalls, whichtotaledover$500million, exceededMFGlobal’s resources, which led directly to its failure. This sort of example illustrates how margin requirementscancontainaprocyclicalelement,eveniftheyarenotexplicitlytiedtoassetprice volatility.29 Suchunusualchargesalsoreceivedsignificantattentionduringthe“Memestock” episodeinU.S.equitymarketsinJanuary2021(Mourselas,2021). Finally,itisnotonlytheamountofIMrequiredthatislikelytoriseduringtimesofmarket stress. Accordingtothequantitativedisclosuredata,approximatelyhalfofIMisintheform of securities, rather than cash, and those securities are subject to haircuts. Lewandowska andGlaser(2017)findthatonemajorCCPdoesnotsignificantlyincreaseitshaircutsduring timesofmarketstress. Nevertheless,inprinciple,risinghaircutsanddeclinesinmarketvalue thatcouldbeassociatedwithepisodesofmarketvolatilitycouldleadtochangesinhaircuts 29StandardsrequireIMmodelstoincorporateanti-procyclicalitymeasures(CPMI-IOSCO,2012),butultimatelyIM willrespondtochangingriskexposures.HoullierandMurphy(2017),WongandPei(2017),andGlassermanand Wu(2018)studytheproblemofreducingmarginprocyclicality.O’NeillandVause(2016)studyhowatime-varying “macroprudentialbuffer”addedtoIMcanaddressfire-saleexternalities.Raykov(2012)examinesthetrade-offs inherentinreducingtheprocyclicalityofIMandfindsthatreducingitdoesnotnecessarilyreducesystemicrisk.

CENTRALCLEARINGANDSYSTEMICLIQUIDITYRISK 19 andwouldrequirethepostingofadditionalcollateralevenifthemarginrequirementwere unchanged. 4.3. Settlementrequirements. IndiscussinghowVMandIMcanvarywithmarketvolatility, thereisatendencytofocusonthechangeinmarketprices. However,marketvolatilityusually isalsoassociatedwithincreasedtradingactivity. Asaresult,portfoliovaluationscanchange becausebothpricesandpositionsfluctuate. Importantly,increasesintradingactivitydirectly createliquidityneedsatsecuritiesCCPsinordertosettlethetrades. Becausethecashneeded tosettlesecuritiescorrespondsdirectlytothefullvalueofthesecuritiestrades,theresulting liquidityneedscanbesimultaneouslymaterialandprocyclical. Intermsofmateriality,from thelatestquantitativedisclosures,thesumofthepeakpaymentobligationsinthepriortwelve months was just under $194billion. For derivatives, most trades do not require up-front paymentstosettle,sotheeffectthereisatmostminor.30 Fromasystemicview,theliquidity needsgeneratedbythesettlementofcentrally-clearedsecuritiesneedtobecombinedwiththe liquidityneedsfromVMandIMinassessingtheresiliencyoftheclearingsystemoverall. 4.4. Default fund contributions and assessments. While VM and IM calls are the most commonwaysinwhichCCPsdrawinresourcesfromtheirclearingmembers,severalother contingentrelationshipscanalsocomeintoplay,particularlyintimesoffinancialstress. One important way in which CCPs can require resources from their members is by calling for contributionstothedefaultfund. Figure12showstheaveragequarterlychangesindefault fundsizesperclearingmembersplitintothefivelargestversustherest.31 Since2015,changes in average changes in default fund contributions have been small on average, however the averageofaveragesmasksvolatility,particularlyforlargerclearingmembers,whereeventhe quarterlyaverageacrosstheCCPshasbeenover$100million. Thelargestcallatanindividual CCPwasover$540million,andevenclearingmembersnotamongthefivelargestfacedacall ofnearly$160million. Althoughlikelylesssensitivethanmargin,defaultfundscanchange due to changing portfolios, market risk, and risk management practices. The fact that the changeindefaultfundassessmentswasnotmorepronouncedduringthefirstquarterof2020 30Derivativescontractscanhavesettlementrequirements,suchasifanoptionisexercisedorafuturematures,but suchrequirementslikelyarenotstronglyprocyclical.CDSpotentiallycangenerateprocyclicalpaymentsduetothe defaultsofreferenceentities,butsuchpaymentsaredifficulttopredictorregularlyobserve. 31ThisfigureonlyincludesthederivativesCCPsasthesecuritiesCCPsdonothaveclearingfundsthataredistinct fromIM.

20 THOMASB.KING,TRAVISD.NESMITH,ANNAPAULSON,ANDTODDPRONO likelyreflectsalaginupdatingstressscenarios;suchalagacrossmultipleCCPsmayindicatea correlatedweaknessinCCPstresstestingandthereforeahigherlevelofsystemicrisk. TherearethreedistinctcircumstancesunderwhichCCPsmaycallonclearingmembersto addresourcestothedefaultfund. Thefirstisintra-monthcallsforcontributionsinresponseto changesinriskthattheCCPperceives. Typically,CCPslevydefault-fundassessmentsmonthly toreflectchangesinmarketconditionsandconcentrationthatoccuroverthecourseofany givenmonth. Intimesofmarketstress,however,CCPsmaynotwaitfortheendofthemonth. Intra-monthdefaultfundcallsmaybeissuedfollowingsuddenchangesinmarketvolatility.32 The second situation in which unscheduled payments to the default fund may occur is whenaclearingmemberdefaultsandthatmember’sIMandowndefault-fundcontribution is insufficient to cover the liquidation value of its position. An example occurred in 2013 when a topping-up of the default fund was required at the Korean CCP KRX following a tradingerrorthatresultedinaclearingmemberlosing$45millionbeforeamargincallcould beissued. Amorerecentexampleisthepreviouslymentioned2018defaultofalargeclearing member at Nasdaq Clearing, which ultimately consumed approximately two-thirds of the mutualizeddefaultfundresources. Althoughbothofthesedefaultswereidiosyncraticand occurredduringtimesofrelativemarketcalm,theobligationtoreplenishthedefaultfundis asourceofprocyclicality,becauseboththelikelihoodofdefaultsandthepotentialimpactof arealizeddefaultincreaseduringtimesofmarketstress,whenthenecessarycapitalmaybe scarce. BecauseclearingmemberdefaultsthatarelargeenoughtobreachaCCP’smutualized defaultresourcesarequiteuncommonandfortunatelydidnotoccurduringthepandemic, thereisalsoariskthatmarketparticipantsmaynotbeattunedtoorpreparedfortheresource demandssuchdefaultscouldgenerate. Inaddition,andfromtheperspectiveofagivenclearing member,thereisadistinctionbetweentherisksposedbyVMandIMcallsrelativetocalls foradditionalmutualizeddefaultfundresources,regardlessofeitherthefrequencyorsizeof theassociatedcalls,sincetheformertworelateentirelytotheportfoliothatagivenclearing member(CM)bringstotheclearinghouse,whilethelatter,toadegree,dependsuponthe portfoliosbroughtbyotherCMs,whichareoutsideofthegivenCM’simmediatecontrol. Finally,CCPshavepowersofassessmentintheeventthattheprefundedportionofthedefault fundisexhausted. Figure13showstheunfundedcommitmentsthatclearingmemberswould 32Increaseddefaultfundneedscouldalternativelybeaddressedbyissuingmargincalls.

CENTRALCLEARINGANDSYSTEMICLIQUIDITYRISK 21 havetopayCCPsinsuchcircumstancesasapercentageoftotaldefaultresourcesexcludingIM. TheexercisebyaCCPofitsunfundedcommitmentsisaveryraresituation,sinceprefunded resourcesarecalibratedtocoverthelossesofaCCP’ssingleortwolargestclearingmembers. Thus,inprinciple,CCPsshouldonlyneedtodrawonunfundedcommitmentsincasesinwhich threeormoremembersdefaultnearlysimultaneouslyandhaveinsufficientmargintomakeup forthelossassociatedwiththeirpositions. Thecaveat“inprinciple”reflectsthatoneortwo memberdefaultsalsocouldexhaustprefundedresourcesiftheCCP’smodelsareinaccurateor ifliquidationofdefaultedpositionsprovesunexpectedlychallenging. Clearly,anenvironment inwhichthosedefaultsoccurredwouldbeassociatedwithveryhighlevelsofmarketstressand liquiditydemandandcouldstrainclearingmembers’abilitytomeettheirpaymentobligations. Notably,theincreaseinpre-fundedresourcesheldfollowingMarch2020, hasreducedthe unfundedcommitmentsaspercentage. However,thisindicatesthatassessmentpowershave notnecessarilyexpandedevenasCCPsseemtoviewriskasbeinghigherasreflectedintheir increasedholdingofpre-fundedresources. 4.5. Absorbingdefaultingmemberpositions. Intheeventofaclearingmemberdefault,the remainingclearingmembersmayalsohaveotherobligations. Inparticular,theymayacquire someor allof thedefaultingmember’spositions andmayalsobecome responsibleforthe positionsofthedefaultingmember’sclient. Aspartofitsdefault-managementprocedures, anOTCderivativesCCPwouldtypicallyauctionpartorallofthedefaultingmember’shouse portfolio. DependingontherulesoftheCCP,survivingclearingmembersmaybeobligatedto participateintheauction. Inaddition,thepositionsofthedefaultingmember’sclientsmust be transferred to remaining clearing members or be liquidated. Many market participants whoclearindirectlyhaveestablishedbackuprelationshipswithoneormoredirectclearing membersthattheycouldactivateintheeventofthedefaultoftheirprimaryclearer. Absorbingthisadditionalbusinessplacesanaddedburdenonclearingmembers’resources. Inparticular,clearingmembersthatarebanksorbrokerdealersarerequiredtoholdcapital againstboththeirownpositionsandthoseoftheirclients. Althoughinstitution-leveldataon housepositionsarenotavailable, Figure14showsthesizeofthelargestclientpositionsat anyclearingmember,relativetotheexcesscapitalattheremainingclearingmembers. (The dataarefromtheCFTCandonlycoverthederivativesCCPs.) “Excesscapital”represents theamountofcapitalthatclearingmembershaveavailabletosupportadditionalpositions.

22 THOMASB.KING,TRAVISD.NESMITH,ANNAPAULSON,ANDTODDPRONO Thus,theratiointhegraphisameasureofmembers’abilitytoabsorbtheclientbusinessof anothermember. TheratiowassubduedintheyearsleadinguptotheCOVID-19crisis,partly reflectingthehighlevelsofcapitalinthebankingsector,butitrosequicklywhenvolatility spiked,whichisalsowhentheprobabilityofadefaultandthepossibleneedtoportclient positionslikelyincreased. Wenotethat,unlikemanyoftheotherprocyclicalresourcedemandsnotedinthispaper, theneedforbankstoabsorbhouseandclientpositionsofdefaultingmembersislargelyan issueofcapital. However,duringstressfulperiods,capitalandliquidityadequacycanbecome intertwined. Moreover,theabsorptionofnewpositionsalsorequiresbankliquidity,because itrequirespostingadditionalcontributionstoCCPdefaultfundsandreservingadditional liquidity buffers to meet future margin calls. Again, given that at least one member has defaulted,thesecommitmentsarelikelytooccurinatimewhenthecapitalandliquidityof othermembersarealreadystretchedthin. Furthermore,addingnewpositionsandposting newcollateralcanimpactregulatoryliquidityratiorequirements. Evenifnotbinding,abank’s willingnesstoabsorbpositionscouldbeinfluencedbyconcernsovernegativelyimpactingits liquidityratio,whichcouldbereflectedinitsvaluationofthedefaultedportfolio.33 4.6. Liquidity provision. Finally, CCPs maintain liquidity arrangements with many large banks. These arrangements include committed lines of credit, repo facilities, and foreignexchangeswapagreements. Ideally,thecounterpartiesontheothersideofthesecontracts are liquidity providers that do not face the CCP in other types of transactions. In practice, mostoftheinstitutionsthatareinapositiontocommittoprovidingsignificantliquidityare largebanksthatalsoparticipatedirectlyinCCPs,clearinghighvolumesofderivativesand securitiestransactions. Consequently,sincelargebankstendtobemembersofmultipleCCPs andofferliquidityservicestoeach,thesupplyoftheseservicestotheoverallsystemofCCPs tendstobeconcentrated. FSB(2018a)reportedthat27percentofclearingmemberssurveyed across26CCPsalsoprovidecreditlinesthatprovideliquiditytotheCCP.Manyoftheseclearing membersprovidesuchlinestomultipleCCPs.34 33Theseconcernsmotivatedinparttherecentannouncementofarevisiontothetreatmentofclientmarginin calculatingtheliquiditycoverageratio(BCBS,2019). 34Intheeventofdefault,liquiditydemandsmaybeparticularlylargeatthetwosecuritiesCCPs,becausesuch CCPsguaranteesettlementofthefullpurchasepriceofsecurities.(Thesettlementvalueofderivativescontracts istypicallyasmallfractionoftheirnotionalvalue.) Forexample,inthefewdaysfollowingthebankruptcyof LehmanBrothersin2008,NSCCandFICCsettled(withoutloss)over$300billionofsecuritiestransactionsthat

CENTRALCLEARINGANDSYSTEMICLIQUIDITYRISK 23 Figure15showstotalcommittedlinesofcreditandrepoarrangementsatCCPsasafraction oftheholdingsofcashandcashplusliquidsecuritiesatlargeUSbanks.35 Althoughnotall oftheCCPliquidityprovidersarelargeUSbanks,thecomparisonnonethelessshowsthat, if these facilities were suddenly and significantly drawn upon, the cash demand could be significant. Interestingly starting in 2018, the significance rose particularly relative to cash holdings,bothbecauseCCPsworkedtoimprovetheiraccesstoliquidityandbanksreduced cashholdings. Fortunately,CCPsdidnothavetodrawontheseresourcesduringtheCOVID-19 crisis,andthesizeexposureshavedeclinedsomewhatasbankshavemaintainedhighercash holdings. Butifneeded,evenifbankswereabletomeetthisdemand,doingsocouldwellput pressureonshort-termfundingmarkets,includingtheinterbankmarketandtherepomarket forhigh-qualitycollateral. Again, theCCPsareonlylikelytoneedtoaccesstheirliquidity arrangements following the default of a very large clearing member, so it is probable that fundingmarketswouldalreadybeundersomepressure. Forexample,CCPswouldalsomost likelybewithdrawinglargesumsfromtheircashdepositaccountsatbanksinsuchasituation. 4.7. ThescaleofCCPresources. Figures16-19showhowthemainresourcesabsorbedand potentiallydemandedbyCCPscomparetovariousmeasuresofthefinancialsector’scapacity to handle those demands. In particular, we scale CCP resource draws by two measures of intermediarycapitalandfourmeasuresofintermediaryliquidity. Themeasuresofcapital are the total Tier-1 capital at commercial banks and the capital in excess of required at the clearingmembersofU.S.futurescommissionmerchants. Theliquiditymeasuresaretheliquid assets(cashandTreasurysecurities)ofU.S.broker-dealers,thecashheldbylargedomestically charteredcommercialbanks,thetotalamountofhigh-qualityliquidassets(HQLA)heldby thesixlargestbankholdingcompanies,andtheamountofHQLAinexcessoftheliquidity coverageratiorequirement.36 Figure16comparesthestockofresourcesheldbyCCPsasofQ12021tothelevelsofthe balance-sheetmeasuresatintermediariesatthattime. ThesixCCPscollectivelyheld$700 Lehmanhadexecutedwithitscustomersandothercounterparties. Incontrast,themarketvalueofLehman’s obligationsunderderivativecontractswasabout$45billion(Valukas,2010). 35Largebanksincludesthosewithover$250billioninassets. 36TotalbankcapitalcomesfromtheFDICQuarterlyBankingProfile. Clearingmembercapitalcomesfromthe monthlyCFTCreportonFinancialDataforFCMs.Broker-dealerliquidassetscomesfromtheFinancialAccounts oftheUnitedStates. Large-bankliquidassetscomesfromtheFederalReserveH.8Report. Informationabout HQLAandLCRrequirementsistakenfromtheindividualregulatoryfilingsofJPMorganChase,BankofAmerica, Citibank,WellsFargo,GoldmanSachs,andMorganStanley.

24 THOMASB.KING,TRAVISD.NESMITH,ANNAPAULSON,ANDTODDPRONO billioninIManddefaultfunds,anamountequaltoseveraltimestheexcesscapitalofclearing members’ordealers’holdingsofliquidassets. Figures 17 and 18 consider the situation in terms of flows, rather than stocks. Figure 17 reports the largest daily variation and initial margin calls reported at any of the six CCPs overthesampleperiod,relativetoestimatesofthestandarddeviationofdailychangesinthe balance-sheetcategories.37 Thesemargincallsrepresentedchangesofseveraldailystandard deviationsineachofourcapitalandliquiditymeasures. Notethatthesearecallsatindividual CCPs. Thedatadonotallowustomeasurethemaximumaggregatedailymargincalls,but theywerenecessarilylargerandlikelysignificantlyso. Figure18showsthemaximumquarterlychangesininitialmarginanddefaultfunds,comparingtothequarterlystandarddeviationofchangesinthebalance-sheetmeasures. Again, themaximumaggregateIMcallinoursamplewasequaltoachangeofseveralstandarddeviationsinintermediarycashandliquiditypositionsatthequarterlyfrequency. Themaximum default-fundcalloverthissamplewassmaller,butitstillrepresentedalargeflowrelativeto typicalquarterlychangesinintermediarycapital. Finally,Figure19reportstheexplicitclaimsthatCCPshadonfinancialinstitutionsasof Q1 2021. These include lines of credit, deposits (both the CCPs’ own house accounts and the portion of initial margin held as bank deposits), and unfunded commitments that the CCPscancallfrommembersintheeventofadepletionofthedefaultfund. Becausethese itemsallrepresentpotentialflows,weagainscalethembythequarterlystandarddeviationof intermediarybalance-sheetchanges. Again,itisclearthatthesepotentialresourcedrawscould belargerelativetothetypicalfluctuationsbanksanddealersseeintheirbalancesheets. Totake theextremecase,ifalloftheseCCPcommitmentsweredrawnatonce,theresultcouldbeequal toa7-standard-deviationmoveinexcessHQLAatlargebanksanda32-standard-deviation moveinCMexcesscapital. ThiswouldbeontopoftheVM,IM,anddefaultfundcallsthat wouldlikelyoccuratthesametime. TheupshotofthisdiscussionisthatthedemandbyCCPsforliquidresourcesthathasbeen observedoverthelastseveralyearshasbeenquantitativelylargerelativetomarketparticipants’ capitalandliquiditycushions,andithasthepotentialtobeevenlargerinfutureperiodsof 37Theestimatesareconstructedbycalculatingthevarianceofquarterlybalance-sheetchangesandassumingindependenceofchangesatthebusiness-dayfrequencywithinquarters.

CENTRALCLEARINGANDSYSTEMICLIQUIDITYRISK 25 stress. Becausesuchperiodsarealsoassociatedwithotherpressuresonfinancialinstitutions’ balancesheets,CCPprocyclicalitycouldbematerialininducingsystemicliquidityshortfalls. 4.8. TheimportanceofCCPincentives. Althoughtheirproceduresaregovernedbydetailed rulebooks,CCPsretainsomediscretionintheextentandtimingoftheircallsforresources. Twoexampleswillillustratethispoint. BignonandVuillemey(2018)describeoneoftheveryfewfailuresofaCCP,theCaissede LiquidationdesAffairesenMarchandisesinParis. In1974,followingacollapseofsugarprices,the CCPfacedthedefaultofaclearingmemberthathadclearedalargevolumeofsugarfutures. BecausetheCCP’smarginingpracticeswereinadequate, thedefaultwouldhavecreateda lossfortheCCP’smembers,butwouldhaveallowedittocontinueoperating. ButtheCCP engagedinrisk-shifting: itdelayeddeclaringtheunderwatermemberindefault,apparently inthehopesthatitspositionwouldrightitself. Instead, furtherlossesaccrued, eventually resultinginashortfallsolargethattheCCPwasforcedtoshutdown. Morerecently,asHeckinger(2014)describes,priortothefailureofMFGlobalin2011,at leasttwoCCPs(ICEClearUSandFICC)refusedtoreleasemargin,totalingabout$100million, thatwasduetotheteeteringbroker-dealer. TheCCPsmayhavefearedthatMFGlobalmight subsequentlytakelossesonpositionsclearedthroughthemandnotbeabletomakeupthe shortfall. AlthoughtheseactionsaffordedtheCCPsanadditionallayerofprotection, they contributedtoMFGlobal’sliquidityshortfallsandultimatedemise. As illustrated by these episodes and others noted above, CCPs can bend or circumvent rules,atleastforatime,whenitisintheirinteresttodoso. Moreover,thereisnoguarantee thattheinterestsofCCPownerswillalignwiththeadvancementoffinancialstability. Some CCPsareownedbytheirmembersorexchanges,whileothersareownedbypubliclytraded companies. Theabilityoftheseownerstoexercisediscretionmakesitimportanttoconsider theincentivesthatCCPsmightfaceinstresssituations. TheCCPsdiscussedhereinallhave theirowncapitalexposedintheirdefaultwaterfalls. BeingexposedtolossprovidesCCPs anincentivetomanagerisks,butcouldalsoincentivizedefensivemoves—likechoosingto withholdMFGlobal’smargin—duringastressfuldefault. ForCCPsthataremember-owned, thepotentialthatclearingmembersmightneedtoreplenishlostcapitalcanbeanadditional procyclicalliquidityriskthatcouldaffectincentivesevenifalossisnotactuallyrealized. The broaderpointisthat,whenfacingthepossibilityofasevereliquidityorsolvencythreat,CCPs

26 THOMASB.KING,TRAVISD.NESMITH,ANNAPAULSON,ANDTODDPRONO mayhavetheabilityandincentivetohoardevenmoreresourcesthantheirnormalpractice suggests.38 5. LiquidityStressTests: IndividualandMacroprudential OneofthemainwaysthatthePFMIraisedstandardsforCCPswastheexplicitrequirements aroundmanagingtheirliquidityrisk. ACCPmustbeabletomakeallofitspaymentobligations ontimeinallrelevantcurrencieswithahighdegreeofconfidence. ThePFMIestablishedthe expectationthatCCPsestablisharobustframeworktoidentify,measure,monitor,andmanage liquidityriskfromparticipants,settlementandcustodianbanks,liquidityproviders,andany otherrelevantentities. ThePFMIspecifiedthatCCPshadtobeabletomakeitspayments underthedefaultoftheclearingmemberthatwouldgeneratethelargestaggregateliquidity obligationfortheCCPinextremebutplausiblemarketconditions. Furthermore,thePFMI definedwhatresourcesshouldqualifyforthepurposesofmeetingsuchrequirements. The abilityoftheCCPtomeetitspotentialliquiditydemandswithitsliquidresourcesmustbe testeddailythroughrigorousstresstestingsimilartohowthetotalfinancialresourcesarestress testeddaily.39 Therequirementtostresstestliquidity,inadditiontobeingnewer,addsnewcomplexities. First,thenumberofrelevantpartiesgoesbeyondjusttheclearingmemberfunctionandincludes thoseentitieswhogeneratealiquidityexposure. Second,multiplerolesplayedbyaclearing member,forexampleiftheywerealsoaliquidityprovider,needtobeconsidered. Third,the abilitytomakepayments,oftenincash,indifferentcurrenciesmustbetestedandmet. Fourth, thetimedimensionmatters,becauseitisnotenoughtohavesufficientcashorotherliquid resourcesinaggregateoveraclose-outperiod,butrathertheCCPneedstobeabletomake paymentsontimewhendue. ThePFMIsignificantlyenhancedexpectationsforliquidityriskmanagementatCCPsand advancedthestresstestingofeachCCP’sparticularliquidityneeds. Bytheirnature,however, 38CCPdiscretioncanalsogotheotherway,providingmarketparticipantsmoretimetomeetpaymentrequirements, orevenoverridingmarginrequirements. Suchexercisesofdiscretioncanresultinsmallerburdensonclearing membersthanwouldotherwisehaveoccurred,butpotentiallyexposetheCCPtohigherriskduringaperiodof marketstress.Moregenerally,CCPdiscretionmeansthatitmaybedifficulttopredictCCPactions. 39InadditiontoCCPs’ownliquiditystresstests,theU.S.CommodityFuturesTradingCommission(CFTC)and theEuropeanSecuritiesandMarketsAuthority(ESMA)bothhaveconductedliquiditystresstestingusingtheir ownscenariosappliedjointlytosetsofCCPs(seeCFTC(2017)andESMA(2018a)respectively).

CENTRALCLEARINGANDSYSTEMICLIQUIDITYRISK 27 thesemicro-levelstresstestscannotmeasureliquiditydemandsthatmayariseacrossthefinancialsysteminamarket-stressevent. Whileparticularpaymentobligationsareisolatedwithin individualCCPs,theresourcesavailabletomakesuchpaymentsextendbeyondtheCCP’s boundaries. Theresultinginterdependenciesaredifficultifnotimpossibleforanindividual CCPtodisentangle,evaluate,andstresstest. ThesimplestrelationshipillustratingdependenciesacrossCCPsisthecommittedlinesof creditextendedtoCCPsbyliquidityproviders. Asnotedabove,afewofthelargestclearing membersprovidesuchlinesofcredit,oftentomultipleCCPs. AsshowninFSB(Figure11 2018a,onpage21),apartfromafewoutliers,thereisapositivecorrelationbetweentheamount ofpre-fundedresourcesaclearingmemberprovidesinaggregateandtheamountofaggregate resourcesitprovidesasaliquiditybackstop. TheimplicationisthatlargerCCPmembersalso providesignificantliquidresourcestothesameCCPs. IndividualCCPsdonotnecessarily have a view to the obligations its clearing members have to other CCPs. The default of a largeclearingmemberlikelywouldentailasimultaneousdefaultacrossmultipleCCPs,and thereforemultipleCCPsactivatingliquidityrelationships. StresstestingatindividualCCPs cannotcapturethisinterdependency. Acoordinated,macroprudentialsupervisorystresstestacrossmultipleCCPscancomplement the micro-oriented stress tests conducted by individual CCPs. The goal would be to evaluatethecollectiveimpacttheparticipatingCCPshaveonthebroaderfinancialsystem during such a default. The results would inform regulators and participants on potential liquiditydemandsandperhapsavoidtheliquidityhoardingthatincreasedsystemicriskduring the 2008 crisis (Ashcraft et al., 2011, Acharya and Merrouche, 2012, Berrospide, 2013). ThisobjectivestandsincontrasttotheCCPs’ownstresstests,whichlookattheirindividual resiliency. Suchatestwouldprovideasenseofthesizeofthesystemicliquidityriskthata centralbankcouldfaceinextremecircumstanceseitherinitsroleasthelenderoflastresortto thebankingsystemorpotentiallythroughdirectlendingtoCCPs.40 Suchatestwouldbedesignedinamannerconsistentwiththeinternationalframeworkfor supervisorystresstestingpublishedbyCPMI-IOSCOin2017.41 But,wearemorefocusedthan 40Centralbanks’abilityandwillingnesstolenddirectlytoCCPsvarieswidely.TheFederalReservehasaconstrained abilitytolenddirectlytoCCPsthathavebeendesignatedtobesystemicallyimportant,butonlyinunusualor exigentcircumstancesandifotherprovisionsaremet(Baker,2012).Incontrast,theBankofEnglandexplicitly includesCCPsinitsregularlenderoflastresortfunction. 41SeeCPMI-IOSCO(2017).Andersonetal.(2020)discusssuchmacroprudentialCCPstresstestsfurther,including discussingtheirrationaleandobjectives;seealsoTompaidis(2012).

28 THOMASB.KING,TRAVISD.NESMITH,ANNAPAULSON,ANDTODDPRONO theframeworkinexplicitlycallingforstressingliquidityneedsandfocusingontheoverall impactonthesystem. Suchtestswouldgreatlyenhancemarketparticipants’andregulators understandingofhowliquidityriskscouldariseintheCCPnetworkandpotentiallyaffect therestofthefinancialsystem,goingbeyondprioranalysis,suchasHeathetal.(2016). Such a tests may require iterative defaults to trace the potential for stress to spread through the networkofliquidityneedandprovision. Suchiterationwouldbedifferentthansupervisory stresstestingforbanks,whichisafirmlyestablishedregulatorytoolpostthe2008crisis,and generallyisviewedaseffective.42 Testingtheresiliencyoftheclearingsystemlikelywould similarlyimproveourunderstandingofthefeedbackloopsdiscussedinFaruquietal.(2018). It would expand on the recent supervisory stress tests conducted by the CFTC on CCPs it regulates(CFTC,2016,2017),particularlybyincludingintheanalysissecuritiesCCPs,which, asnotedabove,requirelargeamountsofliquiditytoeffectsettlement. AlthoughmacroprudentialstresstestingofCCPs’liquidityriskwouldbeanimportantstep,it islikelynotsufficienttofullyidentifythescopeofrisksassociatedwiththeirinterconnectedness. To do so, regulators should work to integrate macroprudential stress testing of CCPs with moreestablishedbankstresstesting,whichhasgenerallyignoredbanks’largeexposuresto CCPs(Borioetal.,2014,Schuermann,2014). Suchanintegratedapproachgoesbeyondthe CPMI-IOSCOframeworkforCCPs. 6. Conclusion Bytakingbothsidesofderivativesandsecuritiestrades,CCPsabsorbcounterpartycredit risk. Insodoing,theygenerallyimprovefinancialstabilitythroughmultilateralnetting,risk mutualization,andmargining. FollowingthesuccessofCCPsinmanagingriskandpreventing contagionduringthe2008financialcrisis,regulatoryreformshavemovedevenmoreactivityto centralclearing,inparticularthroughclearingmandatesforthemostcommonOTCderivatives. Despitetheirrolesinpromotingfinancialstability,andparticularlyinreducingcontagion, large CCPs are concentrated and interconnected and pose risks of their own. While attentionhasgenerallyfocusedonthepotentiallydisastrousconsequencesofafailureorsevere 42SeeDrehmann(2008),PetrellaandResti(2013),Borioetal.(2014),Schuermann(2014),Fernandesetal.(2017), andFlanneryetal.(2017).AlthoughthegenerationofsuchtestscouldcreateincentivesforCCPstoadjusttheir riskandliquiditymanagement,theirmacroprudentialnatureandthefactthatCCPsdonottradeintotheirown positionslikelyeliminatesthe‘window-dressing’effectobservedbyCornettetal.(2018)inbankstresstestingby supervisors.

CENTRALCLEARINGANDSYSTEMICLIQUIDITYRISK 29 disruptionofaCCP,wehavehighlightedadifficultythatmayoccurinamuchlessremote state of the world. Namely, to protect themselves, CCPs necessarily require liquidity from largebanksand othermarketparticipants. Fromtheperspectiveof clearingmembers, the counterpartyriskthatismitigatedwithcentralclearingis,inasense,replacedwithliquidity risk. ThedemandforresourcescantaketheformofVMandIMcalls,defaultfundassessments, drawsonliquiditylines,thetransferofpositionsassociatedwithdefaultingmembers,and other obligations. This liquidity risk spans banks and funding markets, creating risks due tointerconnectedness. Furthermore,thestrengthoftheinterconnectednessincreasesunder marketstress. Fromasystemicperspective,thetrade-offbetweenreducingcounterpartycredit exposurewhilepotentiallyincreasinginterconnectedliquidityneedsisexpectedtoreduce systemicrisk. Nevertheless,becauseresourcesaremostlikelytobecalledforbyaCCPattimes whenbankliquiditypositionsarealreadyunderstress,theyareinherentlyprocyclicalwith respecttomarketconditions. Theprocyclicalityofliquidityriskmustbemanaged. Although CCPs,banksandotherfinancialinstitutionssuccessfullymanagedtheunprecedentedmargin calls during the market stress driven by COVID-19, not all the potential liquidity risk was realizedasthefewdefaultsthatoccurredwererelativelynon-impactful. Consequently,the systemicriskmayyetremainunder-estimated. Theexpansionofcentralclearingatwell-managedCCPsstrengthensmarketfunctionand resiliency. But, as more activity becomes concentrated in CCPs, the possibility that CCP liquidity demands could strain banks and other market participants looms larger. On the policyfront,aswehavediscussed,liquidity-focusedmacroprudentialstresstestscouldhelpto assesstheimpactofshockstoCCPsonsystemwideliquidity. Suchtestsalsowouldbeastep towardsintegratingsupervisorystresstestingofbanksandCCPs. Fromaresearchperspective, moreworkisneededtounderstandhowtheliquidityrisksposedbyCCPsfitintothebroader contextofdemandforsafeandliquidassets. Theanswertothatquestionhasimplicationsfor themeasurementofsystem-wideliquidity,forthemodelingoffundingmarkets,andforour understandingofthepropagationoffinancialcrises.

30 THOMASB.KING,TRAVISD.NESMITH,ANNAPAULSON,ANDTODDPRONO Figure1. Interestrateswapsoutstanding 80% 70% 60% 50% 40% 30% 20% 10% 0% 2000 2005 2010 2015 2020 Semi-Annual deraelCtnecreP 600 500 400 300 200 100 0 )lirt$(gnidnatstuO Thebluebars,associatedwiththeleft-handscale,representIRSnotional outstanding;Theredline,associatedwiththeright-handscale,shows thepercentagecentrallycleared. Source:BIS Figure2. SixmajorCCPsclearingUSmarkets Max. dailymargin Approx. prefunded callsince2015 resources($bil.) ($bil. est.) CCP MainProducts 2019 Q1-2020 pre-2020 Q1-2020 Securities FICC 39.7 66.6 16.2 34.2 GSD USTreasuriesandrepos MBSD Mortgage-backedsecurities NSCC USequities,corps. &munis 12.5 36.7 8.6 24.5 Derivatives OCC USequityoptions&futures 57.2 103.0 17.3 31.8 CME 139.4 239.8 16.0 21.3 Base Commodityandfinancial futures&options IRS Interestrateswaps andswaptions ICC Creditdefaultswaps 36.7 53.3 2.4 9.1 LCHSwapClear Interestrateswaps 170.7 206.9 24.9 32.9 MaximumdailymargincallestimatedbyaddingthepeakVMandIMcallsduringeachquarter. Source:Quant.disclosuresviaClarusCCPView.DatathroughQ3-2021

CENTRALCLEARINGANDSYSTEMICLIQUIDITYRISK 31 Figure3. VariationandInitialMargin 1. Variation margin is owed daily Portfolio value of clearing • Equal to daily shortfall in value member • Failure to pay is a default Potential • Prevents current exposures from default accumulating Close point out Time 2. Initial margin is collected to protect CCP from defaulted positions losing value before they are closed • Equal to estimated potential loss during close-out at 99th percentile of assumed period( e.g. 99th percentile of a 5-day loss) Figure4. IMforIRS,Futures,andCDS 600 500 400 300 200 100 0 2014 2015 2016 2017 2018 2019 2020 2021 2022 EndofMonth )lib$(MIderiuqeR IRS CDS Futures Totalrequirementsheldbyclearinghousesfromclearingmembers,includingadd-ons. IRSdataincludeCME&LCHLtd. Futuresdatainclude CME,ICEU,&ICEUS.CDSdataincludeCME,ICC,ICEU,&LCHSA. Dataaremonth-endthroughDec. 2021. Source:CFTC

32 THOMASB.KING,TRAVISD.NESMITH,ANNAPAULSON,ANDTODDPRONO Figure5. PercentageofclientIMrequirementsatfivelargestparentfirms 80% 70% 60% 50% 40% 30% 20% 10% 0% 2014 2015 2016 2017 2018 2019 2020 2021 2022 EndofMonth egatnecreP Swaps Futures FuturesdataincludeCME,ICEU,&ICEUS.Swapsdataincludesboth IRSandCDSrequirements. IRSdataincludeCME&LCHLtd. CDSdata includeCME,ICC,ICEU,&LCHSA.Dataaremonth-endthroughDec. 2021. Identityoftop5firmscanvaryfrommonth-to-month. Source:CFTC Figure6. DailypeakVMpaidversusrealizedvolatility 12 10 8 6 4 2 0 Q3Q4Q1Q2Q3Q4Q1Q2Q3Q4Q1Q2Q3Q4Q1Q2Q3Q4Q1Q2Q3Q4Q1Q2Q3 2016 2017 2018 2019 2020 2021 Quarters )lib$(nigraMnoitairaVkaeP 100% 80% 60% 40% 20% ytilitaloVdezilaeR Peak VM, represented by the blue bars associated with the left-hand scale,istheindustryaverageofthemaximumpaidby/toeachCCPin thequarter. Realizedvolatilityisannualized. Source:Quant.disclosuresaccessedviaClarusCCPView,andHeberetal.(2009).Data throughQ3-2021.

CENTRALCLEARINGANDSYSTEMICLIQUIDITYRISK 33 Figure7. VMatLCHSwapclear 25 20 15 10 5 Q3Q4 Q1 Q2 Q3Q4 Q1 Q2 Q3Q4 Q1 Q2 Q3Q4 Q1 Q2 Q3Q4 Q1 Q2 Q3Q4 Q1 Q2 Q3 2016 2017 2018 2019 2020 2021 Quarters )lib$(nigraMnoitairaV Peak Average PeakVM,thehigherredline,isthemaximumpaidby/toLCHSwapclear inthequarter. Thebluelineshowsthecorrespondingquarterlyaverage. Source:Quant.disclosuresaccessedviaClarusCCPView.DatathroughQ3-2021. Figure8. CMEFuturesIMrequirementsandVIXindex 40 35 30 25 20 15 10 5 0 2004 2008 2012 2016 2020 egatnecrePnigraMlaitinI 100 CMEFutures Margin 80 60 40 20 0 sleveLXIV VIX Theblueline,associatedwiththeleft-handscale,showsCME’sIMrequirementsonanS&P500futurescontractasapercentageofthecontractvalue. Theredline,associatedwiththerighthandscale,istheChicagoBoard OptionsExchange’sVolatilityIndex(VIX),whichthatreflectsmarketexpectationsof30-dayforward-lookingimpliedvolatility. Itiscalculated fromS&P500indexoptions. Source:Dailydata3/1/2001–1/12/2022fromChicagoBoardOptionsExchange,CME, andHaverAnalytics.

34 THOMASB.KING,TRAVISD.NESMITH,ANNAPAULSON,ANDTODDPRONO Figure9. CMEFuturesIMrequirementversusVIXindex 80 70 60 50 40 30 20 10 4 6 8 10 12 14 CMEFuturesMargin leveLXIV 𝛽=3.58 March2020 The joint plot shows CME’s IM requirements on an S&P 500 futures contract as a percentage of the contract value versus VIX; the strong positiverelationshipisreflectedintheupwardslopingregressionline; thisregression,whichishighly-significantasshownbythebarelyvisible 99thpercentbootstrappedconfidenceinterval,isestimatedrobustlyto reducetheinfluenceofoutliersthatwouldincreasetheslopefurther. The marginaldistributions,plottedaboveandtotherightashistograms,show thatbothserieshavealongpositivetail,althoughpresumablymarginas apercentiscapped. TheredpointshighlightvaluesfromMarch2020 duringtheCOVID-19inducedmarketshock. Source: Dailydata3/1/2001–1/12/2022fromChicagoBoardOptionsExchange,CME andHaverAnalytics.

CENTRALCLEARINGANDSYSTEMICLIQUIDITYRISK 35 Figure10. OCCquarterlypeakIMcalls 30 25 20 15 10 5 Q3Q4 Q1 Q2Q3Q4 Q1 Q2Q3Q4 Q1 Q2Q3Q4 Q1 Q2Q3Q4 Q1 Q2Q3Q4 Q1 Q2Q3 2016 2017 2018 2019 2020 2021 Quarters )lib$(llaCMIfotnuomA Source:Quant.disclosuresaccessedviaClarusCCPView.DatathroughQ3-2021. Figure11. QuarterlychangesinrequiredIMandVIX 60 40 20 0 −20 −40 Q4Q1Q2Q3Q4Q1Q2Q3Q4Q1Q2Q3Q4Q1Q2Q3Q4Q1Q2Q3Q4Q1Q2Q3 2016 2017 2018 2019 2020 2021 Quarters )lib$(MIderiuqernisegnahC IM 60 40 20 0 −20 −40 )stniop.loV(XIVnisegnahC ViX The blue bars, associated with the left-hand scale, represent quarterly changesinthesumofthelargestcallsinrequiredIMacrosstheseven CCPs;Theredline,associatedwiththeright-handscale,showsthechange inthemaximumVIXlevelineachquarter. Source:Quant.disclosuresaccessedviaClarusCCPView,andChicagoBoardOptions Exchange.DatathroughQ3-2021.

36 THOMASB.KING,TRAVISD.NESMITH,ANNAPAULSON,ANDTODDPRONO Figure12. ChangeinaverageDefaultFundassessments 100 50 0 −50 −100 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 2016 2017 2018 2019 2020 2021 Quarters )lim$(FDΔ.gvA Top5CMs OtherCMs The thicker red line represents average change at the six CCPs in the defaultfundassessmentsofeachCCP’sfivelargestclearingmembers. Thethinnerbluelineshowsthesameaveragechangefortheremaining CMsacrosstheCCPs. Source:Quant.disclosuresaccessedviaClarusCCPView.DatathroughQ3-2021. Figure13. UnfundedCMcommitmentsasapercentoftotaldefaultresources excl. IM 35% 30% 25% 20% 15% Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 2016 2017 2018 2019 2020 2021 Quarters stnemtimmoCdednufnU Source:Quant.disclosuresaccessedviaClarusCCPView.DatathroughQ3-2021.

CENTRALCLEARINGANDSYSTEMICLIQUIDITYRISK 37 Figure14. LargesttotalclientpositionofCMasapercentageofotherCMs’ excesscapital 60% 50% 40% 30% 20% 10% 0% 2004 2009 2014 2019 EndofMonth egatnecreP Source:EndofmonthdataMar.2002–Nov.2021fromCFTC. Figure15. Committedlinesofcreditasapercentoflargebankliquidassets 25% 20% 15% 10% 5% 0% Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 2016 2017 2018 2019 2020 2021 Quarters tnecreP CashandTreasuries Cashonly Large banks are domestic banks with more than $250billion in assets. Liquidassetsincludescashandotherhighlyliquidassets. Source: Quant. disclosuresaccessedviaClarusCCPView, andFederalReserveH.8 release.DatathroughQ3-2021.

38 THOMASB.KING,TRAVISD.NESMITH,ANNAPAULSON,ANDTODDPRONO Figure16. StockofResourcesAbsorbedbyCCPs Relativetofinancialintermediarybalance-sheetquantities,3/31/21 HQLA above Totalvalue Tier1 Excess Liquid Cash HQLA min. LCR Resources 3/31/21 capital capital assets (large (large (large at6CCPs ($bil) (allbanks) (FCMs) (dealers) banks) banks) banks) InitialMargin $572 30% 426% 247% 30% 22% 141% Defaultfund $128 7% 95% 55% 7% 5% 32% Total $700 36% 521% 302% 37% 27% 173% Source:Quant.disclosuresviaClarusCCPView;FDICQuarterlyBankingProfile;CFTCFinancialDataforFCMs; FinancialAccountsoftheUnitedStates;FederalReserveH.8Report;regulatoryfilingsofJPMorganChase,Bankof America,Citibank,WellsFargo,GoldmanSachs,andMorganStanley.DataasofQ1-2022 Figure17. FlowofresourcestoCCPs–maximumdaily Relativetodailystd.dev.ofintermediarybalance-sheetquantities,2015-21 HQLA above Resources Totalvalue Tier1 Excess Liquid Cash HQLA min. LCR demanded 3/31/21 capital capital assets (large (large (large by6CCPs ($bil) (allbanks) (FCMs) (dealers) banks) banks) banks) Maxdaily VMcallata singleCCP $26 1644% 2070% 431% 328% 227% 440% Maxdaily IMcallata singleCCP $32 1984% 2498% 520% 396% 274% 532% Source:Quant.disclosuresviaClarusCCPView;FDICQuarterlyBankingProfile;CFTCFinancialDataforFCMs; FinancialAccountsoftheUnitedStates;FederalReserveH.8Report;regulatoryfilingsofJPMorganChase,Bankof America,Citibank,WellsFargo,GoldmanSachs,andMorganStanley.DatathroughQ3-2021.

CENTRALCLEARINGANDSYSTEMICLIQUIDITYRISK 39 Figure18. FlowofResourcestoCCPs-maximumquarterly Relativetoquarterlystd.dev.ofintermediarybalance-sheetquantities,2015-21 HQLA above Resources Totalvalue Tier1 Excess Liquid Cash HQLA min. LCR demanded 3/31/21 capital capital assets (large (large (large by6CCPs ($bil) (allbanks) (FCMs) (dealers) banks) banks) banks) Maxqtrly IMcall $199 1538% 1938% 403% 307% 213% 412% Maxqtrly def. fund call $22 170% 214% 45% 34% 24% 46% Source:Quant.disclosuresviaClarusCCPView;FDICQuarterlyBankingProfile;CFTCFinancialDataforFCMs; FinancialAccountsoftheUnitedStates;FederalReserveH.8Report;regulatoryfilingsofJPMorganChase,Bankof America,Citibank,WellsFargo,GoldmanSachs,andMorganStanley.DataasofQ1-2022 Figure19. ClaimsonIntermediaries,asof3/31/21 Relativetoquarterlystd.dev.ofintermediarybalance-sheetquantities,2015-21 HQLA above Resources Totalvalue Tier1 Excess Liquid Cash HQLA min. LCR dueto 3/31/21 capital capital assets (large (large (large 6CCPs ($bil) (allbanks) (FCMs) (dealers) banks) banks) banks) Linesof credit $232 1797% 2263% 471% 359% 248% 482% Bank deposits (CCPacct +IMheld) $70 539% 678% 141% 108% 4% 144% Unfunded commitments $30 232% 292% 61% 46% 32% 62% Source:Quant.disclosuresviaClarusCCPView;FDICQuarterlyBankingProfile;CFTCFinancialDataforFCMs; FinancialAccountsoftheUnitedStates;FederalReserveH.8Report;regulatoryfilingsofJPMorganChase,Bankof America,Citibank,WellsFargo,GoldmanSachs,andMorganStanley.DataasofQ1-2022

40 THOMASB.KING,TRAVISD.NESMITH,ANNAPAULSON,ANDTODDPRONO References Acharya, Viral V., and Ouarda Merrouche. 2012, “Precautionary Hoarding of Liquidity and Interbank Markets: Evidence from the Subprime Crisis.” Rev.Finan., 17(1): 107–160. doi:10.1093/rof/rfs022. Acharya,ViralV.,HyunSongShin,andTanjuYorulmazer.2010,“CrisisResolutionandBank Liquidity.”Rev.Finan.Stud.,24(6): 2166–2205.doi:10.1093/rfs/hhq073. Adrian,Tobias,andHyunSongShin.2008,“Liquidity,MonetaryPolicy,andFinancialCycles.”Curr.IssuesEcon.Financ.,14(1): 1–7,URLhttps://www.newyorkfed.org/research/ current_issues/ci14-1.html. Adrian, Tobias, and Hyun Song Shin. 2010, “Liquidity and Leverage.” J. Finan. Intermediation,19(3): 418–437.doi:10.1016/j.jfi.2008.12.002.RiskTransferMechanismsandFinancial Stability. Alfranseder, Emanuel, Paweł Fiedor, Sarah Lapschies, Lucia Orszaghova, and Paweł Sobolewski. 2018, “Indicators for the Monitoring of Central Counterparties in the EU.” TechnicalReport14,EuropeanSystemicRiskBoard,EuropeanSystemofFinancialSupervision,URLhttps://www.esrb.europa.eu/pub/pdf/occasional/esrb.op14.en.pdf. OccasionalPaperSeries. Allen,Franklin,ElenaCarletti,andDouglasGale.2009,“InterbankMarketLiquidityand CentralBankIntervention.”J.Monet.Econ.,56(5): 639–652.doi:10.1016/j.jmoneco.2009.04.003. Carnegie-RochesterConferenceSeriesonPublicPolicy: DistressinCreditMarkets: Theory, Empirics,andPolicyNovember14-15,2008. Amini,Hamed,RamaCont,andAndreeaMinca.2016,“ResiliencetoContagioninFinancial Networks.”Math.Finance,26(2): 329–365.doi:10.1111/mafi.12051. Anderson,Edward,FernandoCerezetti,andMarkManning.2020,“SupervisoryStressTesting forCCPs: AMacro-prudential,Two-tierApproach.”J.Finan.MarketInfrastructures,8(1): 1–25. doi:10.21314/JFMI.2019.115. Armakolla,Angela,andJeanPaulLaurent.2017,“CCPResilienceandClearingMembership.” doi:10.2139/ssrn.2625579.AvailableatSSRN. Ashcraft,Adam,JamesMcandrews,andDavidSkeie.2011,“PrecautionaryReservesandthe InterbankMarket.”J.MoneyCreditBank.,43(s2): 311–348.doi:10.1111/j.1538-4616.2011.00438.x. Bai,Jennie,ArvindKrishnamurthy,andCharles-HenriWeymuller.2018,“MeasuringLiquidityMismatchintheBankingSector.”J.Finance,73(1): 51–93.doi:10.1111/jofi.12591. Baker,Colleen.2012,“TheFederalReserveasLastResort.”U.Mich.J.L.Reform,46(1): 69–133, URLhttps://repository.law.umich.edu/mjlr/vol46/iss1/2.Fall. Barker,Russell,AndrewDickinson,AlexLipton,andRajeevVirmani.2017,“SystemicRisks inCCPNetworks.”Risk, 91–97, URLhttp://www.risk.net/cutting-edge/banking/ 2479766/systemic-risks-in-ccp-networks. BCBS. 2019, Leverage Ratio Treatment of Client Cleared Derivatives. URL https://www.bis.org/ bcbs/publ/d467.htm. BCBS-CPMI-IOSCO.2021,ReviewofMarginingPractices.URLhttps://www.bis.org/bcbs/ publ/d526.htm.ConsultativeReport. BCBS-IOSCO. 2015, Margin Requirements for Non-centrally Cleared Derivatives. URL https:// www.bis.org/bcbs/publ/d317.pdf. Berger,AllenN.,andChristaH.S.Bouwman.2009,“BankLiquidityCreation.”Rev.Finan. Stud.,22(9): 3779–3837.doi:10.1093/rfs/hhn104. Berger, Allen N., and Christa H.S. Bouwman. 2017, “Bank Liquidity Creation, Monetary Policy,andFinancialCrises.”J.Finan.Stab.,30: 139–155.doi:10.1016/j.jfs.2017.05.001. Bernanke, Ben S. 1990, “Clearing and Settlement during the Crash.” Rev.Finan.Stud., 3(1): 133–151,URLhttp://www.jstor.org/stable/2961962.

CENTRALCLEARINGANDSYSTEMICLIQUIDITYRISK 41 Berrospide, Jose. 2013, “Bank Liquidity Hoarding and the Financial Crisis: An Empirical Evaluation.” Technical Report 3, Federal Reserve Board, URL https:// www.federalreserve.gov/pubs/feds/2013/201303/201303abs.html. Finance and EconomicsDiscussionSeries. Bignon,Vincent,andGuillaumeVuillemey.2018,“TheFailureofaClearinghouse: Empirical Evidence.”Rev.Finance,1–30.doi:10.1093/rof/rfy039.Forthcoming. Borio,Claudio,MathiasDrehmann,andKostasTsatsaronis.2014,“Stress-testingMacroStress Testing: DoesItLiveuptoExpectations?” J.Finan.Stab.,12: 3–15.doi:10.1016/j.jfs.2013.06.001. Reformingfinance. Breach, Tomas, and Thomas King. 2018, “Securities Financing and Asset Markets: New Evidence.” Working Paper 2018-22, Federal Reserve Bank of Chicago, Chicago, IL, URL https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3301242. Brunnermeier,MarkusK.2009,“DecipheringtheLiquidityandCreditCrunch2007–2008.”J. Econ.Perspect.,23(1): 77–100.doi:10.1257/jep.23.1.77. CCP12.2020,“CCPsAgainDemonstrateStrongResilienceinTimesofCrisis.”Technicalreport, URLhttps://ccp12.org/papers/. CFTC. 2016, Supervisory Stress Test of Clearinghouses. URL https://www.cftc.gov/ sites/default/files/idc/groups/public/@newsroom/documents/file/ dcr_ecl1017.pdf. CFTC. 2017, Evaluation of Clearinghouse Liquidity. URL https://www.cftc.gov/ sites/default/files/idc/groups/public/@newsroom/documents/file/ dcr_ecl1017.pdf. ChicagoBoardOptionsExchange,“CBOEVolatilityIndex: VIX[VIXCLS].”retrievedfrom FRED,FederalReserveBankofSt.Louis,URLhttps://fred.stlouisfed.org/series/ VIXCLS. ClarusCCPView.ClarusFinancialTechnologyLtd.,URLhttps://ccpview.clarusft.com. Cont, Rama. 2017, “Central Clearing and Risk Transformation.” Working Paper 3, Norges Bank, URL https://www.norges-bank.no/en/news-events/news-publications/ Papers/Working-Papers/2017/32017/.NorgesBankResearch. Cornett,MarciaMillon,JamieJohnMcNutt,PhilipE.Strahan,andHassanTehranian.2011, “LiquidityRiskManagementandCreditSupplyintheFinancialCrisis.”J.Finan.Econ.,101(2): 297–312.doi:10.1016/j.jfineco.2011.03.001. Cornett,MarciaMillon,KristinaMinnick,PatrickJ.Schorno,andHassanTehranian.2018, “AnExaminationofBankBehavioraroundFederalReserveStressTests.”J.Finan.Intermediation. doi:10.1016/j.jfi.2018.05.001.Inpress. Cox,RobertT.,andRobertS.Steigerwald.2018,“ACCPisaCCPisaCCP.”J.Finan.Market Infrastructures,6(4): 1–18.doi:10.21314/JFMI.2018.085. CPMI-IOSCO. 2012, PrinciplesforFinancialMarketInfrastructures. URL http://www.bis.org/ cpmi/publ/d101.htm. CPMI-IOSCO. 2014, Recovery of Financial Market Infrastructures. URL https://www.bis.org/ cpmi/publ/d121.htm. CPMI-IOSCO.2016,ResilienceandRecoveryofCentralCounterparties(CCPs): FurtherGuidanceonthe PFMI.URLhttp://www.bis.org/cpmi/publ/d101.htm. CPMI-IOSCO. 2017, FrameworkforSupervisoryStressTestingofCentralCounterparties(CCPs). URL http://www.bis.org/cpmi/publ/d161.htm. Culp,ChristopherL.2010,“OTC-ClearedDerivatives: Benefits,Costs,andImplicationsofthe ’Dodd-FrankWallStreetReformandConsumerProtectionAct’.”J.Appl.Finance,20(2): 1–27, URLhttps://ssrn.com/abstract=2693059. DeBandt,Olivier,andPhilippHartmann.2019,“SystemicRiskinBankingAftertheGreat FinancialCrisis.”InOxfordHandbookofBanking,eds.AllenN.Berger,PhilipMolyneux,and

42 THOMASB.KING,TRAVISD.NESMITH,ANNAPAULSON,ANDTODDPRONO JohnO.S.Wilson,chapter26,847–889,OxfordUniversityPress,3rdedition.doi:10.1093/oxfordhb/9780198824633.013.27. Drehmann, Mathias. 2008, “Stress Tests: Objectives, Challenges And Modelling Choices.” RiksbankEcon.Rev., (2): 60–92, URL http://archive.riksbank.se/Upload/ Dokument_riksbank/Kat_publicerat/PoV_sve/eng/2008/er2008_2_ny.pdf. Duffie,Darrell.2014,“FinancialMarketInfrastructure: TooImportanttoFail.”InAcrossthe GreatDivide: NewPerspectivesontheFinancialCrisis, eds. John B. Taylor and Martin Neil Baily, 251–257, Hoover Institution Press, URL https://www.hoover.org/sites/default/ files/across-the-great-divide-ch11.pdf. Duffie, Darrell, and Haoxiang Zhu. 2011, “Does a Central Clearing Counterparty Reduce CounterpartyRisk?” Rev.AssetPric.Stud.,1(1): 74–95.doi:10.1093/rapstu/rar001. ESMA. 2018a, EU-wide CCP Stress Test 2017. URL http://firds.esma.europa.eu/webst/ ESMA70-151-1154%20EU-wide%20CCP%20Stress%20Test%202017%20Report.pdf. ESMA. 2018b, Guidelines on EMIR Anti-Procyclicality Margin Measures for Central Counterparties. URL https://www.esma.europa.eu/sites/default/files/library/esma70-151- 1293_final_report_on_guidelines_on_ccp_apc_margin_measures.pdf. Faruqui,Umar,WenqianHuang,andElődTakáts.2018,“ClearingRisksinOTCandDerivativesMarkets: TheCCP-BankNexus.”BISQuart.Rev.,73–90,URLhttps://www.bis.org/ publ/qtrpdf/r_qt1812h.htm. Fernandes, Marcelo, Deniz Igan, and Marcelo Pinheiro. 2017, “March Madness in Wall Street: (What)DoestheMarketLearnfromStressTests?” J.Bank.Financ.doi:10.1016/j.jbankfin.2017.11.005.Inpress. FIA.2020,“RevisitingProcyclicality: TheImpactoftheCOVIDCrisisonCCPMarginRequirements.”Technicalreport,URLhttps://www.fia.org/resources/fia-issues-whitepaper-impact-pandemic-volatility-ccp-margin-requirements. Flannery,Mark,BeverlyHirtle,andAnnaKovner.2017,“EvaluatingtheInformationinthe FederalReserveStressTests.”J.Finan.Intermediation,29: 1–18.doi:10.1016/j.jfi.2016.08.001. Fleming,MichaelJ.,andAsaniSarkar.2014,“TheFailureResolutionofLehmanBrothers.”Fed. ReserveBankNewYorkEcon.Pol.Rev.,20(2): 175–206,URLhttps://www.newyorkfed.org/ research/epr/2014/1412flem.html. FSB. 2017a, Analysis of Central Clearing Interdependencies. URL http://www.fsb.org/wpcontent/uploads/P050717-2.pdf. FSB.2017b,ReviewofOTCDerivativesMarketReforms:EffectivenessandBroaderEffectsoftheReforms.URL http://www.fsb.org/wp-content/uploads/P290617-1.pdf. FSB. 2018a, Analysis of Central Clearing Interdependencies. URL http://www.fsb.org/wpcontent/uploads/P090818.pdf. FSB.2018b,IncentivestoCentrallyClearOver-the-Counter(OTC)Derivatives:APost-ImplementationEvaluationoftheEffectsoftheG20FinancialRegulatoryReforms.URLhttp://www.fsb.org/2018/11/ incentives-to-centrally-clear-over-the-counter-otc-derivatives-2/. FSB. 2021, Lessons Learnt from the COVID-19 Pandemic from a Financial Stability Perspective: Final Report.URLhttps://www.fsb.org/2021/10/lessons-learnt-from-the-covid-19pandemic-from-a-financial-stability-perspective-final-report/. Gibson,Rajna,andCarstenMurawski.2013,“MargininginDerivativesMarketsandtheStabilityoftheBankingSector.”J.Bank.Financ.,37(4): 1119–1132.doi:10.1016/j.jbankfin.2012.10.005. Glasserman,Paul,andQiWu.2018,“PersistenceandProcyclicalityinMarginRequirements.” Manage.Sci.,64(12): 5705–5724.doi:10.1287/mnsc.2017.2915. Greenwood,Robin,AugustinLandier,andDavidThesmar.2015,“VulnerableBanks.”J.Finan. Econ.,115(3): 471–485.doi:10.1016/j.jfineco.2014.11.006. HaverAnalytics.HaverAnalytics,URLhttp://www.haver.com/our_data.html.

CENTRALCLEARINGANDSYSTEMICLIQUIDITYRISK 43 Heath,Alexandra,GerardKelly,MarkManning,SheriMarkose,andAliRaisShaghaghi. 2016,“CCPsandNetworkStabilityinOTCDerivativesMarkets.”J.Finan.Stab.,27: 217–233. doi:10.1016/j.jfs.2015.12.004. Heber, Gerd, Asger Lunde, Neil Shephard, and Kevin K. Sheppard, “Oxford-Man Institute’s realized library.” University of Oxford, URL https://realized.oxfordman.ox.ac.uk/.Versionno.0.3. Heckinger, Richard. 2014, “MF Global: A Case Study of Liquidity Risks.” J. Finan. Market Infrastructures,3(2): 79–96.doi:10.21314/JFMI.2014.037. Heckinger, Richard, Robert T. Cox, and David Marshall. 2017, “Cleared Margin Setting at Selected Central Counterparties.” J. Finan. Market Infrastructures, 5(4): 1–21. doi:10.21314/JFMI.2017.075. Heider,Florian,MarieHoerova,andCorneliaHolthausen.2015,“LiquidityHoardingand Interbank Market Rates: The Role of Counterparty Risk.” J.Finan.Econ., 118(2): 336–354. doi:10.1016/j.jfineco.2015.07.002. Heller,Daniel,andNicholasVause.2012,“CollateralRequirementsforMandatoryCentral ClearingofOver-the-counterDerivatives.”TechnicalReport373,BISWorkingPapers,URL http://www.bis.org/publ/work373.pdf. Houllier,Melanie,andDavidMurphy.2017,“InitialMarginModelSensitivityAnalysisand VolatilityEstimation.”J.Finan.MarketInfrastructures,5(4): 77–103.doi:10.21314/JFMI.2017.078. Huber,PeterJ.1981,RobustStatistics.WileySeriesinProbabilityandMathematicalStatistics, JohnWiley&Sons,Hoboken,NJ. Ippolito, Filippo, José-Luis Peydró, Andrea Polo, and Enrico Sette. 2016, “Double Bank Runs and Liquidity Risk Management.” J. Finan. Econ., 122(1): 135–154. doi:10.1016/j.jfineco.2015.11.004. ISDAClearingMemberCommittee.2021,“COVID-19andCCPRiskManagementFrameworks.” Technical report, URL https://www.isda.org/2021/01/06/covid-19-andccp-risk-management-frameworks/. Ivashina,Victoria,andDavidScharfstein.2010,“BankLendingduringtheFinancialCrisisof 2008.”J.Finan.Econ.,97(3): 319–338.doi:10.1016/j.jfineco.2009.12.001.The2007-8financial crisis: Lessonsfromcorporatefinance. Iyer, Rajkamal, José-Luis Peydró, Samuel da Rocha-Lopes, and Antoinette Schoar. 2014, “InterbankLiquidityCrunchandtheFirmCreditCrunch: Evidencefromthe2007–2009 Crisis.”Rev.Finan.Stud.,27(1): 347–372.doi:10.1093/rfs/hht056. Kashyap,AnilK.,RaghuramRajan,andJeremyC.Stein.2002,“BanksAsLiquidityProviders: AnExplanationfortheCoexistenceofLendingandDeposit-taking.”J.Finance,57(1): 33–73. doi:10.1111/1540-6261.00415. Kiff,John,RandallDodd,AlessandroGullo,EliasKazarian,IsaacLustgarten,Christine Sampic,andManmohanSingh.2010,“MakingOver-the-CounterDerivativesSafer: The Role of Central Counterparties.” In Global Financial Stability Report: Meeting New Challenges to StabilityandBuildingaSaferSystem,chapter3,1–27,InternationalMonetaryFund,URLhttps: //www.imf.org/external/pubs/ft/gfsr/2010/01/pdf/chap3.pdf. King,Thomas,TravisD.Nesmith,AnnaPaulson,andToddProno.2020,“CentralClearing andSystemicLiquidityRisk.”FinanceandEconomicsDiscussionSe-ries2020-009,Washington,DC.doi:0.17016/FEDS.2020.009. Kroszner,RandallS.1999,“CantheFinancialMarketsPrivatelyRegulateRisk? TheDevelopmentofDerivativesClearinghousesandRecentOver-The-CounterInnovations.”J.Money CreditBank.,31(3): 596–618.doi:10.2307/2601077. Kroszner,RandallS.2000,“LessonsfromFinancialCrises: TheRoleofClearinghouses.”J. Finan.ServicesRes.,18(2): 157–171.doi:10.1023/A:1026534619637.

44 THOMASB.KING,TRAVISD.NESMITH,ANNAPAULSON,ANDTODDPRONO LCH.Clearnet, “$9 Trillion Lehman OTC Interest Rate Swap Default Successfully Resolved.”PressRelease,URLhttp://secure-area.lchclearnet.com/media_centre/ press_releases/2008-10-08.asp. Lewandowska,Olga,andFlorianGlaser.2017,“TheRecentCrisesandCentralCounterparty RiskPracticesintheLightofProcyclicality: EmpiricalEvidence.”J.Finan.MarketInfrastructures, 5(3): 1–24.doi:10.21314/JFMI.2017.072. Madigan,Peter,DuncanWood,andLukasBecker.2016,“LCHunderscrutinyafteroutsized Brexit margin calls.” Risk, URL https://www.risk.net/risk-management/2474560/ lch-under-scrutiny-after-outsized-brexit-margin-calls. Manning,MarkJozsef,andDavidHughes.2016,“CentralCounterpartiesandBanks: ViveLa Difference.”J.Finan.MarketInfrastructures,4(3): 1–24.doi:10.21314/JFMI.2016.058. Maruyama, Atsushi, and Fernando Cerezetti. 2019, “Central Counterparty Antiprocyclicality Tools: A Closer Assessment.” J Finan Market Infrastructures, 7(4): 1–25. doi:10.21314/JFMI.2018.110. Mourselas, Costas. 2021, “Hero or villain? NSCC draws fire for Robinhood margin waiver.” Risk.net, URL https://www.risk.net/risk-management/7817001/hero-orvillain-nscc-draws-fire-for-robinhood-margin-waiver. Murphy,David.2013,OTCDerivatives:BilateralTradingandCentralClearing.GlobalFinancialMarkets, PalgraveMacmillan.doi:10.1057/9781137293862. Murphy, David, Michalis Vasios, and Nick Vause. 2014, “An Investigation into the ProcyclicalityofRisk-BasedInitialMarginModels.”Technicalreport,BankofEngland,URL https://www.bankofengland.co.uk/-/media/boe/files/financial-stabilitypaper/2014/an-investigation-into-the-procyclicality-of-risk-basedinitial-margin-models.pdf.FinancialStabilityPaperNo.29. O’Neill, Cian, and Nick Vause. 2016, “Macroprudential Margins: A New Countercyclical Tool?” Technicalreport, BankofEngland, URLhttps://www.bankofengland.co.uk/ working-paper/2018/macroprudential-margins-a-new-countercyclical-tool. WorkingPaperNo.765. Petrella,Giovanni,andAndreaResti.2013,“SupervisorsAsInformationProducers: DoStress Tests Reduce Bank Opaqueness?” J.Bank.Financ., 37(12): 5406–5420. doi:10.1016/j.jbankfin.2013.01.005. PresidentialTaskForceonMarketMechanisms.1988,ReportofthePresidentialTaskForceonMarket Mechanisms.NicholasBrady(Chairman). Raykov,Radoslav.2012,“ReducingMarginProcyclicalityatCentralCounterparties.”J.Finan. MarketInfrastructures,7(2): 43–59.doi:10.21314/JFMI.2018.106. Schuermann,Til.2014,“StressTestingBanks.”Int.J.Forecasting,30(3): 717–728.doi:10.1016/j.ijforecast.2013.10.003. Shleifer,Andrei,andRobertW.Vishny.2011,“FireSalesinFinanceandMacroeconomics.”J. Econ.Perspect.,25(1): 29–48.doi:10.1257/jep.25.1.29. Sidanius, Che, and Filip Zikes. 2012, “OTC Derivatives Reform and Collateral Demand Impact.”Technicalreport,BankofEngland,URLhttps://www.bankofengland.co.uk/ financial-stability-paper/2012/otc-derivatives-reform-and-collateraldemand-impact.FinancialStabilityPaperNo.18. Tompaidis,Stathis.2012,“MeasuringSystem-wideResilienceofCentralCounterparties.”J. Finan.MarketInfrastructures,6(4): 41–54.doi:10.21314/JFMI.2018.098. Valukas,AntonR.,“ReportofExaminer,UnitedStatesBankruptcyCourt,SouthernDistrictof NewYork,inre.LehmanBrothersHoldings,Inc.,etal.,Chapter11caseno.08-13555.”Jenner&BlockLLP,URLhttps://web.stanford.edu/~jbulow/Lehmandocs/menu.html. (Examiner’sReport).

CENTRALCLEARINGANDSYSTEMICLIQUIDITYRISK 45 Wiggins, Rosalind Z., and Andrew Metrick. 2019, “The Lehman Brothers Bankruptcy G: The Special Case of Derivatives.” J. Finan. Crises, 1(1): 151–171, URL https:// elischolar.library.yale.edu/journal-of-financial-crises/vol1/iss1/8. Wong,Max,andPatrickGePei.2017,“PerformanceTestingofMarginModelsUsingTime SeriesSimilarity.”J.Finan.MarketInfrastructures,5(4): 51–75.doi:10.21314/JFMI.2017.076. (ThomasKingandAnnaPaulson)FederalReserveBankofChicago 230S.LaSalleSt.,ChicagoIL60604 (TravisD.NesmithandToddProno)BoardofGovernorsoftheFederalReserveSystem 20th&CSts.NW,Wash.DC20551 Emailaddress,ThomasKing:Thomas.King@chi.frb.org Emailaddress,TravisD.Nesmith:Travis.D.Nesmith@frb.gov Emailaddress,AnnaPaulson:Anna.Paulson@chi.frb.org Emailaddress,ToddProno:Todd.Prono@frb.gov