課課2017長(zhǎng)2017長(zhǎng)線班內(nèi)容：【前導(dǎo)班+基礎(chǔ)班+強(qiáng)化班+沖刺班+百題班班+TotalTLandLiabilitiesandMarginTotalTLandNext,the uses50%borrowedLiabilitiesandMarginTotalTLandThus,theleverageratiohasincreasedto1.5(i.e.,Answer:BCVA=EPE×Creditspreadofthecounterparty–ENE×CreditspreadofForLocal,Creditspreadofthecounterpartyincreasemorethanthecreditspreadofitself.Therefore,LocalshouldrequestareductionintheCVAchargeitpays.GlobalbankrequestsareductionintheCVAchargeitreceives.Answer:StatementIisincorrect.AMAisthemostflexiblemethod.StatementIIiscorrect.ThemostpopularmethodtosatisfytheAMAisthelossdistributionapproach(LDA).StatementIIIiscorrect.TheAMAmethodallowsabanktobuilditsownoperationalriskmodelandmeasurementsystemthatiscomparabletomarketriskstandards.StatementIVisincorrect.LDAiswidelyusedininsuranceandactuarialscience.Answer:InterestrateriskisthebroadmarketvariablecapturedbyVaRwhilecreditriskisthespecificvariablecapturedbythespecificriskcharge(SRC).Answer:StatementsAandDarenotnecessarilytrueastherelativeamountsofseniorandsubordinateddebtcanvarydramatically.StatementBisfalse.Subordinateddebthaspriorityafterseniordebtbutaheadofequity.Therefore,thesubordinatedclaimisbetweenseniordebt,F,andtotaldebt,F+U.ItfollowsthatthesubordinateddebtcanbemodeledbyalongcalloptionwithstrikepriceFandashortcallwithstrikepriceF+U(notU).Statementciscorrect.Duringfinancialdistress,theequityvalueisrelativelysmall,andthesubordinateddebtclaimbehavesmorelikeequityasitisclosertoreceiving“residual”cashflows.Similarly,whenthefirmvalueisrelativelyhigh,thesubordinateddebtclaimbehavesmoreliketraditionaldebt.Answer:ThisQ-Qplothassteeperslopesatthetailsoftheplot,whichindicatefattailsinthedistribution.AnormaldistributionwouldresultinalinearQQplot.AdistributionwiththintailswouldproduceaQQplotwithlesssteepslopesatthetailsoftheplotthanalinearrelationship,whilethisoneissteeperatthetails.Itisnotanegativelyskeweddistribution,astheQ-Qplotissymmetric.Answer:TheprojectshouldberejectedbecausetheARAROCof2.6%islessthanthe Answer:r[2,0]=r(0)+2×lambda×dt–2×sigma×SQRT(dt)=3.0%+2×0.50%×1/12–2×2.0%×SQRT(1/12)=1.929%Answer:Defaultsensitivitiesarecomputedbyshockingvarioushypotheticaldefaultprobabilities.Defaultsensitivitiesarealwayspositivealthoughtheywillconvergetozeroathighdefaultratesforalltranches.Defaultsensitivitiesarelargestatvaluesthatcreatelossesclosetotheattaentpoints.Answer:Theseveritydistributionwillbeheavy-tailedandskewedtotheGirling:"Themostcommonandleastcomplexapproachtomodelingseverityistousealognormaldistribution,althoughlowfrequencylossesmayfitbettertootheroptionssuchasGeneralizedG ,TransformedBeta,GeneralizedPareto,orWeibull.Regulatorstakeakeeninterestinhowwelltheselecteddistributiondemonstrates'goodnessoffit'--orinotherwords,howcertainareyouthatthesamplecomesfromthepopulationwiththeclaimeddistribution.Whenselectingwhichapproachtouse,theAMAguidelinesalsoprovidethefollowingguidance:TheselectionofprobabilitydistributionsshouldbeconsistentwithallelementsoftheAMAmodel.Inadditiontostatisticalgoodnessoffit,DuttaandPerry(2007)haveproposedthefollowingcriteriaforassessingamodel'ssuitability:Realistic(e.g.,itgeneratesalossdistributionwitharealisticcapitalrequirementsestimate,withouttheneedtoimplement“correctiveadjustments”suchascaps),Wellspecified(e.g.,thecharacteristicsofthefitteddataaresimilartothelossdataandlogicallyconsistent),Flexible(e.g.,themethodisabletoreasonably modateawidevarietyofempiricaldata)andSimple(e.g.,itiseasytoimplementanditiseasytogeneraterandomnumbersforthepurposeoflosssimulation).Answer:VaRmeasureswillvaryaccordingtotheapproach(delta-normal,historicalsimulation,MonteCarlosimulation).Thevariationinthesevaluesdoesnotsuggestbiggerproblemswithdataorprogramming/implementationnoristhereanyreasontoendogenousmodelrisk(e.g.,tradersgamingthesystemtolowerriskvalues).Answer:Themeanreversionrate,a,indicatesthespeedofthechangeorreversionbacktothemean.Ifthemeanreversionrateis0.4andthedifferencebetweenthelastvariableandlong-runmeanis10(=40–30),theexpectedchangeforthenextperiodis4(i.e.,0.4×10=4).Answer:Abreakclause(liquidityput)wouldallowJorgenstoterminatetransactionsatprespecifiedfuturedates.WalkawayclauseswouldallowJorgenstowalkawayfromitsliabilitiesifitscounterpartydefaultedwhilemaintainingtheabilitytomakeaclaimiftheexposurewaspositive.AdditionalterminationeventswouldallowJorgenstoclose-outspecifictransactionsifaspecifiedadditionalterminationevent(e.g.,ratingsdowngrade,netassetvaluechange,managementchange)occurred.Answer:Theleft-handsideofJensen’sinequalityistheexpectedpriceinoneyearusingthe1-yearspotratesof6%and4%as:$10.5 $10.5 0.50.943400.50.96154E1r Theexpectedpriceinoneyearusinganexpectedrateof6%computestheright-handsideofinequality $10.51.060.5 Next,divideeachsideoftheequationby1.05todiscounttheexpected1-yearzerocouponbondpriceforonemoreyearat5%.Thepriceofthe2-yearzerocouponbondequals$0.90711(calculatedas$0.95247/1.05),whichisgreaterthan$0.90703orthepriceofatwo-yearzero-couponbonddiscountedfortwoyearsattheexpectedrateof5%:1

Thus,Jensen’sinequalityrevealsthat$0.90711>Answer:Dowd:“Wehaveassumedsofarthatthestochasticprocessdrivingourdataisiid,butmostfinancialreturnsexhibitsomeformoftimedependency(orpatternovertime).Thistimedependencyusuallytakestheformofclustering,wherehigh/lowobservationsareclusteredtogether.Clusteringmattersforanumberofreasons,including:Itviolatesanimportantpremiseonwhichtheearlierresultsdepend,andthestatisticalimplicationsofclusteringarenotwellunderstood.Therearetwosimplemethodsofdealingwithtimedependencyinourdata.Perhapsthemostcommon(andcertainlytheeasiest)isjusttoapplyGEVdistributionstoblock a.Thisisthesimplestandmostwidelyusedapproach.Itexploitsthepointthat aareusuallylessclusteredthantheunderlyingdatafromwhichtheyaredrawn,and eevenlessclusteredastheperiodsoftimefromwhichtheyaredrawngetlonger.”Answer:AnobjectiveratingsystemproducesjudgmentsbasedonconsiderationstiedcocreditAnswer:EVproblemsareintrinsicallydifficult,becausebydefinitionwealwayshaverelativelyfewextreme-valueobservationstoworkwith.ThismeansthatanyEVestimateswillnecessarilybeveryuncertain,relativetoanyestimateswemightmakeofmorecentrallesorprobabilities.EVestimateswillthereforehaverelativelywideconfidenceintervalsattachedtothem.ThisuncertaintyisnotafaultofEVTassuch,butaninevitableconsequenceofourpaucityofdata.Answer:Thedecreaseinprobabilityofdefaultwouldincreasethevalueoftheequitytranche.Also,adefaultoftheequitytranchewouldincreasetheprobabilityofdefaultofthemezzaninetranche,duetoincreasedcorrelation,reducingitsvalue.Thus,itisbettertogolongtheequitytrancheandshortthemezzaninetranche.Answer:ofX(i)Assigned(sameRank-13--22-314554Concordantpairs:(1,3),(4,5);(1,3),(5,4);(3,1),(5,4);(3,1),(4,5),4Discordantpairs:(1,3),(3,1);(4,5),(5,4),2tau

4

Answer:ThisisexactlytheapproachtakenbyBasel'sinternalratings-based(IRB)approachwhichisessentiallyanoriginalone-factorGaussianCopula(OFGC)model.Answer:Thelonger-termfundingratioisequaltotheavailableamountofstablefundingdividedbytherequiredamountofstablefunding.UnderBaselIII,thisratiomustequalorexceed100%.SmallBank’snetstablefundingratio=$255/$250=102%.Answer:Oneofthekeyfrictionsintheprocessofsecuritizationinvolvesaninformationproblembetweentheoriginatorandarranger.Inparticular,theoriginatorhasaninformationadvantageoverthearrangerwithregardtothequalityoftheborrower.Withoutadequatesafeguardsince,anoriginatorcanhavetheincentivetocollaboratewithaborrowerinordertomakesignificantmisrepresentationsontheloanapplication.Dependingonthesituation,thiscouldbeeitherconstruedaspredatorylending(wherethelenderconvincestheborrowertoborrowtoolargeofasumgiventheborrower’sfinancialsituation)orpredatoryborrowing(theborrowerconvincesthelendertolendtoolargeasum).Themajorratingagenciesarenotpaidbytheinvestors.Escrowaccountscan allbutnoteliminatetheriskofforeclosure.Answer:Derivativesmarketparticipantsdonotusetreasuryratesasrisk-rates.Thisisbecausetreasuryratesaregenerallyconsideredtobeartificiallylow.Followingthecreditcrisis,mostbankshavechangedtheirrisk-discountratesforcollalizedtransactionsfromLIBORtowhatareknownasovernightindexedswap(OIS)rates.Butfornon-collalizedtransactionstheycontinuetouseLIBOR,oranevenhigherdiscountrate.Anovernightindexedswap(OIS)isaswapwhereafixedrateforaperiod(e.g.,1monthormonths)isexchangedforthegeometricaverageoftheovernightratesduringtheperiod.OISrateisgenerallylowerbecauseovernightlendersbearmuchlessrisk.LIBORisshortforLondonInterbankOfferedRate.Itisanunsecuredshort-termborrowingratebetweenbanks.LIBORrateshavetraditionallybeencalculatedeachbusinessdayfor10currenciesand15borrowingperiods.Theborrowingperiodsrangefromonedaytooneyear.LIBORratesareusedasreferenceratesforhundredsoftrillionsofdollarsoftransactionsthroughouttheworld.Answer:Mapernmentbondspayingregularcouponsontozerocouponernmentbondsisanadequateprocess,becausebothcategoriesofbondsareernmentissuedandthereforehaveaverysimilarsensitivitytoriskfactors.However,thisisnotaperfectmapsincethesensitivityofbothclassesofbondstospecificriskfactor(i.e.changesininterestrates)maydiffer.Answer:Thethreecommon“l(fā)inesofdefense”suggestedbytheBaselCommitteeonBankingSupervisionandemployedbyfirmstocontroloperationalrisksare:(1)businesslinemanagement,(2)anindependentoperationalriskmanagementfunction,and(3)independentreviewsofoperationalrisksandriskmanagement.Answer:UsingaVasicekmodel,theupperandlowernodesfortime1arecomputedas0.413.2% uppernode6.2% 0.413.2% lowernode6.2% Answer:Iftheswapspecifiesphysicaldelivery,thebuyeroftheswapwilldeliverthereferenceobligationtothesellerandreceivetheparvalueoftheobligation.Answer:LVaRVaR4520,0000.38%1.6450.54520,0000.1Answer:Thefollowingmodelrisksarerepresentedbychoicesathroughc:incorrectmodelspecification,implementationrisk,andcalibrationerror.ChoiceDiscorrect.Answer:Iftheimpliedvolatilitiesforactualcurrencyoptionsaregreaterforaway-from-the-moneyoptionsthanat-the-moneyoptions,thencurrencytradersmustthinkthereisagreaterchanceofextremepricemovementsthanpredictedbyalognormaldistribution.EmpiricalevidencesupportsthisAnswer:Answer:Ifthenumberofexceptionsismorethan2andlessthan12,wewouldnotrejectthemodelbecausethecalculatedLRislessthan3.84.Ifwedonotrejectthemodel,wemaycommitaTypeIIerror.ATypeIIerrorisdefinedasacceptinganinaccuratemodel.Answer:HQLAL1L21haircut430.82TotalNetCashOuflows8010%min10,0.758010%2LCR6.463.232Answer:ThevalueofequityusingtheMertonmodelrequiresthefollowing ESNdKertNd 20%ln100100e10%20%20%20% d11.341641,d2Nd10.9101E1000.9101100e10%50.8144Therefore,thevalueofequityofthefirmis$41.61million.Sincethevalueofdebtmustbethedifferencebetweenthefirm’svalueof$100millionandthevalueofequity,thevalueofdebtis$58.39Answer:λ=8/12=0.6667defaultperP(nextdefaultwithinonemonth)=1–exp(-0.6667×2)=Answer:Theseniortranchewillgainvalueifthedefaultcorrelationdecreases.Highcorrelationthatisonenamedefaults,alargenumberofothernamesintheCDOwillalsodefault.Lowcorrelationimpliesthatifonenamedefaults,therewouldbelittleimpactonthedefaultprobabilityoftheothernames.Therefore,asthecorrelationdecreases,thecumulativeprobabilityofenoughdefaultsoccurringtoexceedthecreditenhancementontheseniortranchewillalsodecrease.Hencetheinvestorwhohassoldprotectionontheseniortranchewillseeagain.Answer:ThescenariodescribedbyWaliarelatestoriskshifting.Riskshiftingoccurswhenrisksandrewardsaretransferredfromonegroupofmarketparticipantstoanothergroupholdingdifferentpositionsinthefirm’scapitalstructure.Inthiscase,risktothefirm’sassetswouldbenefitequityholderstothedetrimentofdebtholderswhohavefixedreturnsbutincreasedriskofloss.Answer:Increasingcollalwouldeffectivelyreducecurrentcreditexposuredependingonthecontractparameters,mainlyminimumtransferamountandthreshold.Answer:Forinvestors,therewaslittleexpectationofportfoliolossesinseniortranches.However,whenrepaymentissuessurfacedintheriskiesttranches,alackofconfidencespreadtoholdersofmoreseniortranches.Thiscausedpanicamonginvestorsandaflightintosaferassets,resultinginafiresaleofsecuritizeddebt.Answer:Theliquidity-adjustedVaRisthesumoftwocomponents.ThefirstcomponentistheVaR,whichisthestockpricetimesthez-scoretimesthestockpricestandarddeviation:$200×0.03×1.65=$9.90.Thesecondcomponentadjustsforliquidityriskandishalfthestockpricetimesabid-askspreadcomponent:0.5×[$200×(0.01+1.96×0.005)]=$1.98.Notethatweuseacriticalz-scoreof1.96whencalculatingtheliquidityriskcomponent.Theliquidity-adjustedVaRisthus:$9.90+$1.98=Answer:Inyears(1)to(3),theloaninterestexceedsthe$5.6750totalbondcouponobligation,suchthatpositiveexcessspreadremains.Onlyinyear(4)isthereaninterestshortfall,butthe65,000isdrawnfromtheOCaccounttopaythebondinterest.Inthisway,thetotalcouponpaymentsof$5.6750arefullyfundedthroughyearThedeficitinyear(5)is$86.356-100.675million=-$14.3186million.Whiletheobligationconsistsof$85principal+$4.675interest=$89.675toseniorbondholders;and$10principal+$1interest=$11tojuniorbondholders.Sothereisnotenoughtobuffertheseniorbondholders:withan$11.0millionmezzanineshortfall,theseniorbondholdersstillexperiencea$3.318millionAnswer:TheadjustedorsecondgenerationRAROC=0.15-1.5(0.1-0.03)TheprojectshouldbeacceptedbecausetheARAROCisgreaterthanthe Answer:AtitsmeetinginMarch2008,theBaselCommitteereviewedcommentsreceivedanddecidedtoexpandthescopeofthecapitalcharge(i.e.,expandtheincrementaldefaultriskchargetotheincrementalriskcharge,IRC).Thedecisionwastakeninlightoftherecentcreditmarketturmoilwhereanumberofmajorrankingorganizationshaveexperiencedlargelosses,mostofwhichweresustainedinbanks’tradingbooks.Mostofthoselosseswerenotcapturedinthe99%/10-dayVaR.Sincethelosseshavenotarisenfromactualdefaultsbutratherfromcreditmigrationscombinedwithwideningofcreditspreadsandthelossofliquidity,applyinganincrementalriskchargecoveringdefaultriskonlywouldnotappearadequate.Forexample,anumberofglobalfinancialinstitutionscommentedthatsinglingoutjustdefaultriskwasinconsistentwiththeirinternalpracticesandcouldbepotentiallyburdensome.Answer:Aslongasthegiverofcollalisnotindefaultthentheyremaintheownerfromaneconomicpointofview.Hence,thereceiverofcollalmustpassoncouponpayments,dividendsandanyothercashflows.Answer:Theconstantprepaymentrate(CPR)andthePublicSecuritiesAssociation(PSA)methodarecommonmethodologiesusedtoestimateprepaymentsforstudentloansandmortgages.Answer:Subordination,excessspread,andshiftinginterestprovideprotectionforseniortranches.Overcollalizationalsoprovidesprotectionforseniortranches.Timingoflossesimpactsexcessspreads.Prepaymentscanaccelerateordeceleratethecashflowstoseniortranches.Answer:Ifafirmisinfinancialdistress,thesubordinateddebtbehavesmorelikeequityandacallItwillincreaseinvalueastimetomaturityincreases,volatilityincreases,andinterestratesincrease.Theseniordebtwillhavenegativeexposurestothesefactors.Ifthefirmisnotindistress,boththeseniordebtandsubordinateddebthavenegativeexposurestothesefactorsbecausethesubordinateddebtbehavesmorelikeseniordebtthanequity.Inthiscase,choiceDwouldbecorrect.Answer:542,750,000562,250,000581,750,000605,150,000636,350,000677,300,0006Answer:ComponentVaRB=26.6/100×1.480×50=Answer:BusinessLinesBetaCorporateFinance=Tradingandsales=Retailbanking=Commercialbanking=Paymentandsettlement=Agencyservices=Assetmanagement=Retailbrokerage=Answer:ESurplus6008%6006%Surplus60018%260014%220.660018%60014%88.51SurplusatRisk121.64588.51133.60Answer:50%1,00050%1,0001.07 1.1 1.04 50%1,00050%1,0001.07 1.049 Answer:PD=(0.1765-Answer:Iistrue.Usingthecopulaapproach,wecancalculatethestructuresofcorrelationbetweenvariablesseparayfromthemarginaldistributions.IVisalsotrue.Correlationisagoodmeasureofdependencewhenthemeasuredvariablesaredistributedasmultivariateelliptical.IIisfalse.Thecorrelationbetweentransformedvariableswillnotalwaysbethesameasthecorrelationbetweenthosesamevariablesbeforetransformation.Datatransformationcansometimesalterthecorrelationestimate.IIIisalsofalse.Correlationisnotdefinedunlessvariancesarefinite.Answer:Whatisneededisaliquidityadjustmentthatreflectstheresponseofthemarkettoapossibletrade.TheformulatouseistheratioofLVaRtoVaR:LVaR1P1EN10.730%1.21 Answer:BecausebankswereabletouseawiderangeofmodelsforcalculatingtheAMA,therewasmoreflexibilitytotheseapproachesthanunderthenewSMA.TSAandASAwereolderapproachesratherthanvariationsoftheSMA.AMAdidaccountforinternallosses.TheSMAisnon-model-based,whereastheAMAdidincorporatebank-specificmodels.Answer:Moody’sKMVmodelevaluatesthehistoricalfrequencyofdefaultforfirmswithsimilardistancestodefaultandusesthisastheprobabilityofdefault.Answer:APoissondistributionisfrequentlyassumedforthedistributionofoperationalriskeventfrequency(i.e.,numberoflossesperyear).Incontrast,aWeibulldistributionisusedwhenmodelingtheseverityofoperationalrisklosses.Therefore,thePoissondistributionisthemoreappropriateonetousetomodelthedistributionoffrequency.AFrechetdistributionhas“heavy”tails,whichsuggeststhatthereisagreaterlikelihoodofanextremeeventoccurring.Incontrast,aGumbeldistributionhas“l(fā)ight”tails,whichsuggeststhatcomparedtotheFrechetdistribution,thereisalesserlikelihoodofanextremeeventoccurring.Gumbeldistributionsaresimilartonormalandlognormaldistributionswherethereisalesserlikelihoodofanextremeeventoccurring.theFrechetdistributionisthemoreappropriateonetousetomodelthedistributionofAnswer:Bothstatementsarecorrect.InthecontextofusinganERMframeworktodecentralizetherisk-rewardtradeoffina ,statementsIandIIarebothcorrect.Answer:EL=1,000,000×0.02=VaR=WCL–EL=3×1,000,000/50–20,000=60,000–20,000=Answer:Therisk-weightfunctiondoesindeedincludeaneffectivematurityadjustmentthatisequaltoageneric2.5yearsinFIRBandwhichisdefinedforeachfacilityinAIRB.Answer:Currencyoptionsexhibitvolatilitysbecausetheat-the-moneyoptionshavelowerimpliedvolatilitythanaway-from-the-moneyoptions.Equitytradersbelievethattheprobabilityoflargepricedecreasesisgreaterthantheprobabilityoflargepriceincreases.Currencytraders’beliefsaboutvolatilityaremoresymmetricasthereisnolargeskewinthedistributionofexpectedcurrencyvalues.(i.e.,thereisagreaterchanceoflargepricemovementsineitherdirection).Answer:Ifthesolidbluelineistheexpectedexposure,thenthepotentialfutureexposure(PFE)willbegreaterandthe umPFEwillbethehighestPFEvalueovertheinterval(butthePFEprofileisnotshown).Answer:Statement1iscorrectasvariabilityinriskmeasures,includinglackofuniformityintheuseofconfidenceintervalsandtimehorizons,canleadtovariabilityinVaRestimates.Statements2isincorrectasotherfactorscanalsocausevariability,includinglengthofthetimeseriesunderysis,waysofestimatingmoments,maptechniques,decayfactors,andnumberofAnswer:ThechoiceofdistributionrelatedtoseverityiscriticaltoAdvancedMeasurementApproachmodels.Medium/heavy-taileddistributionsarefrequentlyused.Thechoiceofthefrequencydistributionis