[PDF] courbe zero coupon
[PDF] calcul propositionnel cours
[PDF] logique propositionnelle exercice corrigé
[PDF] calcul propositionnel exercices corrigés
[PDF] logique propositionnelle table de vérité
[PDF] calcul propositionnel formel
[PDF] calcul des propositions exercices corrigés
[PDF] le resultat d'une multiplication est
[PDF] comment calculer le taux de possession du stock
[PDF] cout de stockage annuel
[PDF] calcul du stock moyen
[PDF] exercice cout de passation et possession
[PDF] calcul tangente formule
[PDF] calcul tangente fonction
[PDF] calcul tangente en ligne
AnExpla nationofNegativeSwapSpreads:
DemandforDuration fromUnderf undedPensionPlans
SVENKLINGLE RandSURESHSUNDARESAN
ABSTRACT
The30-ye arU.S.swapspreadshaveb eennegat ivesinceSept ember2008.Weo eranov el explanationforthispersistentanomal y. Throughanillustrat ivemodel,weshowthatu n- derfundedpensionplansoptim allyuseswapsfordurati onhedging.Combi nedwithdealer banks'balancesheetc onstraints,thisde mandcandrivesw apspreadstobecomenegative. Empirically,weconstructameasureoftheaggr egatefundi ngstatusofDefinedBenefit pen- sionplan sandshowthatthi smeasurei sasignificantex planatoryvar iableof30-yearswap spreads.Wefindasimilarli nkbetw eenpensi onf unds'underfundingands wapspreadsfo r twoothe rregions. Klingler(correspondingau thor)isfromtheDepartmentofFinance,BINorwegi anBusinessSchool and SundaresanisfromColumbiaBus inessS chool.Wear egratefultoStefanNagel(the Editor),theAssociate
Editor,twoanonymousre ferees,Dar rellDu
e,Robin Greenwood,W eiJiang,TomasKokholm,David Lando,HarryMama ysky,Scot tMcDermott,StephenSchaefe r,PedroSerran o,MortenSørensen,Hyun Shin,SavitarSun daresan,andDimitriVa yanosforhelpfulcomments.Klingle racknow ledgessupportfrom theCenter forFinancialFrictions (FRIC) ,grantno.DNRF102.Sundaresanack nowledgeswiththanksthe productivesabbaticalspentatBISwhe rethepaperreceiveduse fulsuggestionsandc omme nts.Wealso thankparticip antsintheseminarsattheSwissFinanc eInsti tuteinGe rzenseeandattheIndianSc hool of Businessfortheircomments. Theauth orshavenoconflicts ofinterest,asidentifiedint heDisclos urePolicy. InSep tember2008,shortlyafterthe defaultofLehman Brothers,thedi!erencebetweenthe swaprate( whichisthe fixed-rateintheswap )ofa30-yeari nterestrateswap (IRS)and theyield ofaTreas urybondwiththe samematurit y,commonlyreferredto asswapspre ad, droppedsharplyandbeca menegative.Asweexplai ninmorede taillater,thisisa theoretical arbitrageopportunit yandanassetpricinganomaly. Incontrast toothercrises phenomena, the30-ye arnegativeswapspre adisverypers istentandstillataround -40basispointsas ofDecemb er2015.Inthispaper,we examinethep ersistentnegative30-y earswap spread ando eranew per spectiveon thepossiblereasonsbehindthis anomal y.Ourhypo thesisis thatdemandf ordurationhedgingb yunderfu ndedpensionplanscoupledw ithbalancesheet constraintsfacedbyswapdealersput spressureonlong -terms wapfixedrate sandultimately turnedthe30-year swap spreadtobecome negative. Negativeswapspreadsare apricingano malyandpresentachall engetovie wst hathave beenheldpr iortothefinan cialcrisisthat sugges tedthatsw apspreadsareindicatorso f marketuncertainty,w hichincreaseintimes offinancialdistres s.T hisisbecaus ethefixed paymentinanIRSisexchang edagains tafloating payment,w hichis typicallybasedon Libor,andentailsc reditrisk. Hence,eventhoughIRSar ecollateralizedand viewedasfree of counterpartycreditrisk,theswaprate shouldbeabovet he(theoretical)ri sk-freeratebecau se ofthe creditris kthatisimplicitin Libor.Theref ore,swapspre adssho uldincreaseintimes ofelevat edbankcreditrisk(seeCollin-DufresneandSolnik,2001, foratreatmen tofthis andrelat edissues).Inadditionto that,treasuries(whichar etheb enchmarksagainstwhich swapspreadsar ecomputed)h aveasta tusas"safehaven",i.e.,assetst hatinvestorsvalue fort heirsafetyandliqui dity.In timesoffi nancialdistress,investorsvaluetheconvenience ofholdingsafe andliquidassets evenmore, whichdecreases thetreasuryyield andmak es themtradeata liquiditypre miumorc onve nienceyield(s ee,for instance ,Longsta ,2004, 1 KrishnamurthyandVissing-Jorgensen,2012, orFeldh¨ut terandLando,2008).Insummary, theseargumentsshow thatthe30-yearswaps preadshouldhaveinc rease daroundthe default ofLehman Brothers. Weo erade man d-drivenexplanationfornegativeswapspre ads.Inparticular,wede- velopamodelinwh ich underfundedpens ionpla ns'demandfordu rationhedgingleadsthem tooptimallyre ceiv ethefixedrateinIRSwithlongmaturities.P ensio nfundshav elong-term liabilitiesintheform ofunf undedp ensionclaimsand inve stinaportfolioof assets,suchas stocks,aswellasinoth erl ong-termassets,lik egov ernment bonds.Theycanbalancet heir asset-liabilitydurationbyin vestingin long-termbondsorby receivingfixedinanIRSwith longmaturit y.Ourtheorypredictsthat,ifp ension fundsareunderfunde d,theypreferto hedgetheirdurat ionriskwithIRS ratherthanbuyingTreasuri es,whichmayb enotfeasibl e giventheirfundingstatus.The preferencefor IRStohedge durationris karisesb ecausethe swaprequi resonlymodestinvestmentt ocovermargins, whereasbuyingagov ernmentbond tomatch durationrequiresoutright inve stment. 1
Thisdemand, whencoupledwithdealer
balancesheetconstrain tsresultsinneg ativeswapspreads. GreenwoodandVayanos(2010)show thatpen sionfunds'demandfordurat ionhedgin g inthe U.K.cana ectthete rmstructu reofBritish giltsbyloweringlong-termra tes.In this sense,ourpaperbea rsacloser elationshiptot heirwork.Howeve r,o urapproachdi ersfrom theirssincew efocusonunderfundedpensionfunds'optimalpr eferenceforthe useofIRS fordu rationhedging.Themodelthat wedevelopshowsthatthe demandfor IRSincreases asthe fundbecom esmoreunderf unded. Weprov idenon-parametricevi dencesuggestingthattheswapspreadstendtoben egative inperi odswhenDBplansareund erfunded.Wet husill ustrateanewchan nelthatma ybe atwor kindrivinglo ng-ter mswapspreadsdown.Usin gdatafromthefinancialaccoun ts 2 oftheUn itedSt ates(formerflow offundstable)fromt heFederalReserve,wec onstructa measureoftheaggregate fundi ngstatuso fDBplans(bothprivateandpubli c)intheUnited States.Wethenuse thismeasureto testtherelations hipbetw een theunderfundedratio (UFR)ofDBp lansan dlong- termswapspread sinaregre ssionsetting.Evenaftercont rolling foroth ercommondriv ersofswapspreads,recogn izedintheliterature,such ast hespread betweenLiborandrepora tes,Debt-to-GD Pratio,de aler-banks' financialconstraint s,market volatility,andlevelaswellasthes lopeofthe yieldcurve,wefind thatUFRisas ignificant variableinexplaining3 0-year swapspreads.Inlinewithournarr ative,wealsoshowthat swapspreads ofshortermaturiti esare nota!ectedby changesinUFR. Wecond uctanumberofrobustne sst ests.Onepotentialc oncern aboutusingUFRasan explanatoryvariableforswapspreadsi sthatthesamefactorsthatha veb eenshownto a ect swapspreads canalsoa ectpensio nfunds.Forexample,ade creaseinthelev eloftheyi eld curvecana ectswapsp readsandals oincreasesthelevelofp ensionfu nds'underfund ing.To addressthisconcern,we usestockretu rnsasaninstrumentalvari ablein atwo-s tageleast squaresetting.The ideahereisthatstockretur nsdirectl yaquotesdbs_dbs2.pdfusesText_3
Derivative Securities: Lecture 7