关闭

关闭

关闭

封号提示

内容

首页 What role did piggyback lending play in the hou…

What role did piggyback lending play in the housing bubble.pdf

What role did piggyback lending…

上传者: fengqiyaoye 2013-10-10 评分 0 0 0 0 0 0 暂无简介 简介 举报

简介:本文档为《What role did piggyback lending play in the housing bubblepdf》,可适用于高等教育领域,主题内容包含yiWeullertoomonastatelevelandZipcodeleveldataovertheperiod–,wefindthatthefr符等。

yiWeullertoomonastatelevelandZipcodeleveldataovertheperiod–,wefindthatthefractionofpiggybackoriginationsisrelatedtohigherforeclosureanddefaultratesinsubsequentyears,andthisrelationisstrongestfornonowneroccupiedpropertiesThepattern,howothers)OneparticulartypeoflendingthatgrewrapidlyduringtherecenthousingboomispiggybacklendingPiggybackloans,moretechnicallyreferredtoassimultaneouscloseerstakingoutapiggybackloanin(StateofNewYorkCity’sHousingandNeighborhoodsReport,)Similarly,aboutofCaliforniaborrowersalsousedpiggybackloanstofinancehomepurchasein(Fishbien,)Theriseofpiggybacklendingduring–mayalsohavecontributedtotheriseindefaultandforeclosureratesSomehavearguedthatpiggybacklendingenableshouseholdstotakeontoomuchdebtviathepurchaseof$seefrontmatterPublishedbyElsevierIncCorrespondingauthorEmailaddresses:mlacourlittlefullertonedu(MLaCourLittle),charlescalhouncoxnet(CACalhoun),weiyucsupomonaedu(WYu)JournalofHousingEconomics()–ContentslistsavailableatScienceDirectJournalofHousiweldoi:jjheThecurrentfinancialcrisishaditsoriginsinashousepricesbegantofallandthemortgagemarketexperiencedasharpincreaseinsubprimemortgagedefaultsandforeclosuresNumerouspapershavestudiedthefactorsthatcontributedtotheunprecedentedincreaseindefaultandforeclosureratesIrrationalexpectationsregardingfuturehousepricegrowth,aproliferationofnonagencymortgagesecuritization,laxunderwriting,andchangingeconomicconditionsareamongthecitedfactors(Bajarietal,Domsetal,Keysetal,MianandSufi,Colemanetal,,amongchaseThesearegenerallyusedbyhomebuyerstofinancemorethanofthehousevaluewithoutpayingprivatemortgageinsurance,atleastifthefirstlienisGSEfinancedPiggybacklendingplayedanimportantroleinhomesales,especiallyfromto,andwasinvolvedinaboutoftheonetofourfamilyowneroccupiedhomepurchasesin(Averyetal,b)ItisparticularlypopularinhighcosthousingareasForexample,betweenand,thenumberofpiggybackloansissuedinNewYorkcitymorethantripled,resultinginmorethanofthehomepurchaseborrowJELClassification:GGRKeywords:MortgageForeclosureHousingPiggybacklendingIntroductionever,appearstobelimitedtotheuseofsubprimepiggybacks,ratherthanamoregeneralphenomenonUsingatopologybasedhousingsupplyelasticitymeasureasaninstrumentforhousepricegrowth,wefurtherconfirmthatthestrongassociationofsubprimepiggybackoriginationwithworseloanperformancewasnotdrivenbytheendogeneityofhousepriceappreciationPublishedbyElsevierIncseconds,arejuniorlienmortgageloanstakenoutconcurrentlywiththefirstmortgagetofinancethehomepurArticlehistory:ReceivedAprilAvailableonlineFebruaryWeexaminetheuseofsimultaneousclosejuniorlienlending(‘‘piggybacks’’)overthecourseoftherecenthousingbubbleandsubsequentmortgagemarketcollapseUsingbothWhatroledidpiggybacklendingplaandmortgagecollapseMichaelLaCourLittlea,,CharlesACalhounb,aCollegeofBusinessandEconomics,CaliforniaStateUniversityatFullerton,FbCalhounConsultingLLC,Annandale,VA,USAcCollegeofBusinessAdministration,CaliforniaStatePolytechnicUniversity,Particleinfoabstractjournalhomepage:wwnthehousingbubbleiYucn,CA,USA,CA,USAngEconomicsseviercomlocatejhecinflatedassets(WSJ,),andthereforehelpedtofurtherinflatethehousingbubbleOncethebubbleburst,itmadehighlyleveragedhouseholdsatgreaterriskofnegativenonowneroccupiedpiggybacksandhighLTVpiggybacksthatwehavebeenabletoidentifythatdirectlyaddressesthedeterminantsofhomeequityborrowingisSalandroasinmostdataonhomeequitylending,theirdatadoesMLaCourLittleetalJournalofHousingEconomics()–havestrongerimpactonmortgageperformanceOfparticularinterest,wedistinguishamongthreeformsofpiggybacklending:()primefirstlienandprimesecondlien,()primefirstlienandsubprimesecondlien,and()subprimefirstlienandsubprimesecondlien,andfurtherexplorethedifferenceinthesethreepiggybacklendingpatternsinexplainingstatelevelandZipcodelevelmortgageperformanceTheplanofthepaperisasfollowsInthenextsection,wereviewthelimitedresearchrelatedtothistopicInthethirdsection,wedescribeourdataandempiricalmethodology,includingourmethodforidentifyingpiggybackloansfromHMDAdataInthefourthsectionwepresentresultsofourempiricalresultsThefinalsectionconcludesLiteraturereviewComparedtootherresearchonmortgagemarkets,juniorlienlendingisarelativelyunexploredarenaThestillnarrowertopicofpiggybacklendinghasreceivedevenlessrigorousresearchBeginningwiththebroaderresearchonjuniorliendebt,Canneretal()describetheearlystagesandgrowthofthehomeequitylendingsegment,followingpassageofthetaxlawchangeswhicharegenerallyacknowledgedtohavespurredthissegmentofconsumerlendingCannerandLuckett()andCanneretal()updatethosefindings,includingSurveyofConsumerFinance(SCF)datashowinghomeequitybalancesoutstandingreached$billionbyWeicher()reviewsthegrowthofthehomeequitylendingindustryduringthes,describingitasbusinessbasedonrecapitalizingborrowerswithsubstantialhousingequity,butimpairedcreditTheonlypaperPriortomostinterestonconsumerdebtwastaxdeductibleaftertaxlawchanges,onlydebtsecuredbyresidentialmortgagedebtremainedgenerallydeductibleforthosewhoitemizedeductionsequityandmorevulnerabletodefaultTheuseofpiggybackloanshasbeenshowntobeimportantinexplainingthemagnitudeofnegativeequity(LaCourLittleetal,)Piggybackloansalsomaketheloanmodificationprocessmorecomplicatedbecausefirstlienandjuniorlienloansarepackagedandsoldtodifferentportfoliosecuritizations(Rosengren,)Moreover,juniorlienlenders,ifdifferentfromfirstlienlenders,usuallyhavelittleincentiveinmodifyingtheloantoavoidforeclosureifthereisnoequityprotectingthem(TheWashingtonPost,)Inthispaper,westudytherelationshipbetweenthemortgageperformance(delinquency,foreclosure,anddefaultrates)andhomeownerpiggybackborrowingpatternsatboththestatelevelandtheZipcodelevelWeaskwhetherstatesandZipcodeswithhigherproportionsofpiggybackloansissuedduring–areassociatedwithworsemortgageperformanceinlaterperiods,especiallyduringandWealsoexaminewhethernotcontaininformationabouttheunderlyingfirstmortgageloans,sincefirstandjuniordebtisoftenheldbydifferentlenders,apatternthatappliestopiggybacklendingaswellLaCourLittleetal()reportthatroughlyofSouthernCaliforniaborrowersfacingforeclosureduring–hadatleastonejuniorlienoutstanding,thoughinformationontheloansthemselvesislimitedCalhoun(b)developsamethodforidentifyingpiggybackloansfromHMDAdataCalhoun(a)arguesthatsimultaneouscloseor‘‘piggyback’’transactionssystematicallyraiseriskthroughoutthemortgagefinancesystem,yetpresentednoloanperformancedataBernstein(),aswellasothersmentionedintheintroduction,documenttheincreaseintheuseofpiggybacklendingovertheperiodwestudyUsingAmericanHousingSurveydata,Bernstein()reportsthatmultiplemortgagefinancingpackagesasapercentofnewlyoriginatedmortgagesincreasedfrominsurveyyeartoinsurveyyear,corroboratingthegrowthinthiscategorydocumentedbyothersClearlyconsiderableadditionalresearchisnecessarytomorecompletelyunderstandthisnewmarketphenomenon,itscauses,anditseffectsOurefforthereaddressesthisgapintheliteratureDataandempiricalmethodologyDataTocalculatetheproportionofpiggybackloanstototalhomepurchaseloanoriginationsforeachstateandZipcode,weuseHomeMortgageDisclosureAct(HMDA)datafromtoWefirstidentifypiggybackloansusingeachofthemethodsproposedbyAveryetal(a)andCalhoun(b)Tocalculatetheproportionofpiggybackoriginationsatthestatelevelforagivenyear,weaggregatethenumberofpiggybacksbystateandyearanddivideitbythetotalnumberoffirstlienhomepurchaseloanoriginationsForZipcodelevelpiggybackoriginations,wecalculatethepiggybackoriginationsatthecensustractlevelfirstandDetailsoftheidentificationmethodareaddressedlaterandHarrison(),whousedandSCFdata,wellbeforethedramaticincreaseinhomeequitylendingoccurredInmorerecentwork,LaCourLittle()arguesthataborrower’spostoriginationhomeequityborrowingdilutestheirequity,increasingtheriskofdefaultontheseniordebtAgarwaletal(a)showsthatpatternsofhomeequitylineusearealsorelatedtoborrowercreditquality,asmeasuredbytheirFICOscoreExtendingthatanalysisfurther,Agarwaletal(b)examinetheperformanceofhomeequitylinesandloans,findingconsiderabledifferenceintermsofdefaultandprepaymentriskMianandSufi(forthcoming)examinehomeequitybasedborrowingfromtoandfindthatitisassociatedwithhighdefaultratesfromtoUnfortunately,thenaggregateittotheZipcodeusingadatabasethatmatchescensustractnumberswithZipcodesfromMissouriCensusDataCenterBecauseagivencensustractcancorrespondtomorethanoneZipcodes,wecreateaweightvariablebasedontheshareofhousingunitsineachcensustractthatliewithinagivenZipcodesUsingthisweightvariable,wecancalculatetheweightedaveragepiggybackoriginationsattheZipcodelevelForstatelevelloanperformancemeasures,weusethestatelevelpercentageofmortgageforeclosureinventoryfromto,obtainedfromtheMortgageBanker’sAssociation(MBA),asourproxyforstateforeclosureratesSincetheMBAdataareonlyavailableatthestatelevel,weusetheZipcodeproportionofnoncurrent(delinquentanddefaulted)mortgagesandtheproportionofmortgagesindefaultfromto,obtainedfromEquifax,asourproxyforZipcodedelinquencyanddefaultratesWesupplementthedatawithadditionaleconomicvariablesthatmayalsoaffectloanperformanceFirst,ourstatelevelhousepricedatacomefromFederalHousingFinanceAgency(FHFA,formerlyOfficeofFederalHousingMLaCourLittleetalJournalofHousingEconomics()–EnterpriseOversight,orOFHEO)purchaseonlyHousePriceIndexesfromtoWeusethisdatatocalculateannualhousepriceappreciationratesAtZipcodelevel,weusetheCoreLogicZipcodeSingleFamilyDetachedHousePriceIndextocalculatehousepriceappreciationratesTheCoreLogicHousePriceIndexcoversaboutZipCodesintheUnitedStatesSecond,weusebothstateandMSAlevelpercapitaincomedatafromtheBureauofEconomicAnalysisThestatepercapitaincomeisusedinourstatelevelanalysis,andtheMSApercapitaincomeisusedasaproxyforZipcodeincomeThird,weobtainedstateunemploymentdatafromtheDepartmentofLaborStatisticsWeusetheMSAdataasaproxyfortheZipcodeincomeandunemploymentrateFourth,wesupplementtheZipcodeleveldatawithZipcodecreditriskdataobtainedfromEquifaxLastly,subprimeloansareidentifiedintwoways:ForloansinHMDAfromandlateryearswetreathighWerecognizethattheHMDAdatausesthecensustractdefinitionsbeforeandthecensustractdefinitionsstartingTherefore,weuseadatabasethatmatchescensustractnumberswithZipcodesfordatabeforeandanotherdatabasethatmatchescensustractnumberswithZipcodesfordataonandafterBothdatabasesareavailablefromMissouriCensusDataCenterForexample,censustracthashousingunitsandofthehousingunitsarewithinZipcodeACensustracthashousingunitsandofthehousingunitsarewithinZipcodeACensustracthaspiggybackoriginationsandcensustracthaspiggybackoriginationsForsimplicity,ZipcodeAaresolelycomposedofcensustractandTocalculatethepiggybackoriginationforZipcodeA,wefirstcreateaweightvariableforcensustractandbasedonthenumberofhousingunitsofeachcensustractthatliewithinZipcodeATheweightforcensustractis(())andtheweightforcensustractis(())WethencalculatetheweightedaveragepiggybackoriginationsforZipcodeAas=AmortgageiscodedasdelinquentifitismorethandaysandlessthandayspastdueAmortgageiscodedasindefaultifitismorethandayspastdue,inbankruptcystatusorinseverederogatorystatusCoreLogicwasspunoffinanIPOfromTheFirstAmericanCorporationonJune,costorspreadreportableloansassubprimeloansForyearspriortoweusetheDepartmentofHousingandUrbanDevelopment’s(HUD)listofsubprimelenderstoidentifysubprimeloansbasedonthelenderIDsinHMDAWethenaggregatethesubprimeloansbystateandZipcodeandcalculatetheproportionofsubprimeloanoriginationstototalloanoriginationsinagivenyearBothmeasuresofsubprimeloansareimperfectUseoftheHUDlistisasomewhatcrudeproxybecauseallloansmadebyasubprimelenderwillbeclassifiedassubprimeloansandallloansmadebyaprimelenderwillbeclassifiedasprimeloansRelianceonthespreadreportablethresholdinHMDAmayunderoroverestimatetheprevalenceofsubprimeloansdependingoncurrentinterestrateconditions,andspecificallytherelationshipbetweenmortgageratesandcomparablematurityTreasuryratesDetaileddefinitionsofthekeyvariablesusedinstatelevelandZipcodelevelregressionsandtheirsourcesarelistedinTableIdentificationofpiggybackloansWeusetwomethodstoidentifypiggybackloansintheHMDAdataThefirstmethodisbasedonAveryetal(a)Sinceinformationaboutlienstatusisonlyavailablestartingin,theidentificationprocessisalittledifferentbeforeandafterBefore,wesorthomepurchaseloanseachyearbystate,county,censustractnumber,lenderID,owneroccupancystatus,borrowerincome,race,andsexIfwefindduplicateloanrecordsaccordingtothissetofmatchingfactors,thentheonewiththesmallerloanamountisidentifiedasapiggybackloanThebasicassumptionunderlyingthismethodisthatiftwohomepurchaseloansinvolveapropertyinthesamecensustractandsameowneroccupancystatus,borrowerswithidenticalincome,raceandsex,andwasissuedbythesamelender,thenmostlikelythesetwoloansareusedforthepurchaseofthesamehomeAfter,withtheadditionoflienstatusinHMDA,weseparatethehomepurchaseloansintotwosamplesThefirstsampleincludesalljuniorlienpurchaseloansandthesecondsampleincludesallfirstlienpurchaseloansWethenmatchthesecondsampletothefirstsamplebycensustract,lenderID,owneroccupancystatus,borrowerincome,raceandethnicity,andsexIfthereisamatch,thenthematchedjuniorlienloanisidentifiedaspiggybackloansOnelimitationoftheAverymethodisthatitmayunderestimatethenumberofpiggybackloansbecauseitcannotidentifypiggybackloansthatareissuedbylendersdifferentfromthefirstlienlendersThesecondmethodfollowsCalhoun(b)TheCalhounapproachissimilartothatofAvery,butrecognizesthefollowingtwopotentialproblems:First,piggybackloansmaybeissuedbyadifferentlenderfromthefirstlienlenderSecond,thelienstatusisusuallymissingforloansthatarereportedassecondarymarketpurchasesinHMDACalhounproposedatwostepmatchingprocedureThefirststepfollowsAvery,byfindingduplicatesaccordingtoasetofmatchingfactorsInthesecondstep,ifthelienstatusisnotmissing,thenthedataaresortedbystate,county,censustract,borrowerincome,lienstatus,andSourceMortgageBanker’sAssociationlienpurchaseloanoriginationsHMDAlforeclosureproceduresandyearlagofpiggybackHMDABureauofEconomicnationrtgageerofmlienlforecMLaCourLittleetalJournalofHousingEconomics()–TableVariabledefinitionsVariableDefinitionPanelA:StatelevelregressionForeclosureStateforeclosureratePiggybackProportionofpiggybackloanoriginationstototalfirstPiggylagTwoyearlagofpiggybackforstateswithnonjudiciaforstateswithjudicialforeclosureproceduresLnincomeLogofstatepercapitaincomeUnemployStateunemploymentratePctsubprimelagTwoyearlagoftheproportionofsubprimeloanorigiHpigrowthHousepriceappreciationrateHpigrowthlagOneyearlagofhousepriceappreciationratePanelB:ZipcodelevelregressionNoncurrentProportionofnoncurrent(delinquentanddefault)mogivenZipcodeDefaultProportionofdefaultedmortgageloanstototalnumbPiggybackProportionofpiggybackloanoriginationstototalfirstPiggylagTwoyearlagofpiggybackforstateswithnonjudiciaforstateswithjudicialforeclosureproceduresaLnincomeLogofMSApercapitaincomeUnemployMSAunemploymentratebloanamountAdjacentfirstandsecondlienswithidenticalvaluesofborrowerincomeareidentifiedaspiggybackcombinationsandremovedfromthedataIfthelienstatusismissing,thentheremainingdataaresortedbystate,county,censustract,borrowerincome,andloanamountandadjacentloanswithidenticalvaluesofborrowerincomearematchedInthesecondstepofthistwostepmatching,thematchingfactorsdonotincludelenderIDTherelaxationofthesamelenderassumptionhelpstoidentifyadditionalpiggybackcombinationsthatareissuedbydifferentlendersTofurtherconfirmapiggybackmatch,CalhounthencalculatedtheratioofthelowerloanamounttohigherloanamountfortheduplicatesThepiggybackcombinationsarefurtherconfirmediftheratioofthesmallerloanamounttothelargerloanamountfallsintocertainrangesthatareconsistentwiththepiggybackloanstructures(suchasstructure,structure,etc)Toexaminewhethernonowneroccupiedandhighlyleveragedpiggybackshaveastrongereffectondefaultandforeclosurerates,wealsoidentifypiggybackloansusedfornonowneroccupiedhousepurchasesandpiggybackloanswithcombinedloantovalue(CLTV)ratiosaboveSinceinformationonthehousevalueisnotprovidedbyHMDAdata,weassumethefirstlienisapproximatelyofthehousevalueThecombinedPctsubprimelagTwoyearlagoftheproportionofsubprimeloanoriginationHpigrowthZipcodehousepriceappreciationrateHpigrowthlagOneyearlagofho

职业精品

用户评论

0/200
    暂无评论

精彩专题

上传我的资料

热门资料

资料评价:

/20
1下载券 下载 加入VIP, 送下载券

意见
反馈

返回
顶部

Q