Empirical Studies of the Securitization Market Credit Rating Risks Vivian Marit van Breemen
1 Credit Rating Risks Empirical Studies of the Securitization Market Vivian M. van Breemen
©NoVpivairatnoMf t. hviasntBhreeseismmenay, 2b0e23re. pArllordiughcetsd,resstoerrevdedo. r transmitted in any form or by any mpreeavniosuwslyithpouubtlisphreiodrchpaeprtmerisss. ion of the author, or the copyright-owning journal for TU hn ei v earcsai dt ye. mT hi ce rreesseeaarrcchh hs taus dbi eese ni ns utphpi so rbt eodo kb ywDe er eN pe de refrol ar mn desdc haet BNaynekn (r Do dNeB )Bauns di n tehs es European Central Bank (ECB). This thesis should not be reported as representing the views of DNB or the ECB. The views expressed are those of the authors and do not necessarily reflect those of DNB or the ECB. DPreisnitgend&byla:yR-iodudte: rMparainikt.enlDisco, proefschriftopmaak.nl ISBN: 9789089801715
NYENRODE BUSINESS UNIVERSITEIT PHD THESIS w i t ha tr Ne gyaerndr ot od et hBeudsoi nc et os sr aUt en/i vPehrDs i dt eei gt r e e on authority of the Rector Magnificus, Prof. dr. Koen Becking in accordance with the Doctorate Committee. T h eF pr iudbalyi cSde epfteenms be et ra k2e2s, 2p 0l a2c3e o n at exactly 4 o’clock by Vivian Marit van Breemen borninoHn eMeamys2k6er, k1993 the Netherlands Credit Rating Risks Empirical Studies of the Securitization Market
Examination Committee Supervisor Prof dr. Dennis Vink - Nyenrode Business Universiteit Other members Prof dr. Arnoud W.A. Boot - University of Amsterdam Prof dr. Jakob de Haan - Rijksuniversiteit Groningen Prof dr. Ivo Arnold - Nyenrode Business Universiteit Prof dr. Ruud G.A. Vergoossen RA - Nyenrode Business Universiteit
To my beloved parents – Nico and Marjan van Breemen
Preface The creation of this dissertation has been an amazing journey. It has allowed me to meet incredibly inspiring people and travel to beautiful places. My first academic paper (Chapter 3) was created in collaboration with Professor Vink and dr. Mike Nawas. The paper also benefitted from feedback of Professor Jakob de Haan aBna dn kP r( oDfNe sBs)oWr Forrakni nk gJ .PFaapbeorzSz ei . rAi ef st e(rWpPu Sb)l ,i sI hhi na vg et hper epsaepnet er di nt hDe ep Na pe edre ralta snedvsecrhael international conferences. For instance, with thanks to Professor Duc Khuong NS hgourytel yn , awf t ee rw, Ie rweaisn vf oi tret du naat tteh ee nPoaur igshFti on af un rc ti ah le Mr taankaeg epma ret ni tn Cmo na nf eyr ei nntceel lienc 2t u0a1l l9y. s( At i FmAu) l aCtoi nngf e rdeins cc ue s2s 0i o2n2s i no nS aonuDr i epgaop.eSr p ae tc i at hl et h aAnmkes rtioc aPnr oFf iensas no cr eJ oAhsns oGcriaaht iaomn faonrdi nd vr.i tMi ni kg emNea owvaesr haal ls tghree awt layy ht eo l pS ae nd mD iee gt oo . eTvhe en tsuuapl leyr vp ius bi ol ins ho ft hPer opfaeps se or ri nV ti nh ke Journal of International Financial Markets, Institutions & Money. Not long after that, my second academic paper (Chapter 2) was published in the Journal of Financial Services Research, again with the wonderful guidance of three highly regarded experts in my field; Professor Frank J. Fabozzi, Professor Dennis Vink and dr. Mike Nawas. Ib yw Pa sr oqf ue si tseo rp rMo aurdk uwshBe nr u, nf onre mr my etiheirr da nadc aPdreomf eiscs op ra pLearu r( aC hSat aprt ke sr t4o) , pI r we saesnitnavgi taei nd aC te nt ht rea Al BF Aa nCko (nEf eCrBe)n Wc e Pi Sn, 2f o0r2w2 .hMi c yh tI hwi rodupl da pl iekre wt oa st hpaunbkl i tshhee dE di ni t ot hr iea lEBu or oa pr de aonf tghuei dEaCnBc eWoPf SP rf oo rf etshseoi rr Dr eevni ne wi s Va ni ndki na nv idt aPt ri oo nf e tsos opru Fbrl iasnhk i nJ . Ft ha eb oEzCzBi , Wt h Pe Sp. aWp ei trh wt ha es published in the Journal Financial Markets, Institutions & Instruments.
My fourth academic paper (Chapter 5) is created in collaboration with Professor Dennis Vink, Professor Frank J. Fabozzi and dr. Mike Nawas. The paper benefitted f2r0o2m2 iAn np nu ut aolf Ct ho en fEe Cr eBn cr ee s oe fa rEcShC sBe mR ei ns ea ar rpc ha r Ct ilcui ps taenr t s3 . a In dw owual sd pl irkees et no t tehda antk t dh re. DL iisabnoanBwohnef irremI wa na sd ads kr.eDd etno npirse sReeni nt hoaurrdfti nfdoirn gi nsv. i t i n g u s to t h e c o n fe re n c e i n For my fifth and final academic paper (Chapter 6), I am very grateful to be working together with dr. Claudia Schwarz and Professor Dennis Vink. The paper also benefitted from feedback of dr. Klaus Düllmann and the participants of the EB Ca nBkRoefs Ne aerwc hYoS er km) i, nPaart. rIi cwi ao Mu l ods lsiekre (tCoo tl uh ma nbki aL Si nI PdAa )G, Lo ol dr bi aenr ag P( Fe el idz ez or anl (RLeesi ebrnvi ez Iannsdt i tCuEt Be Rf oAr )F fi on ra nt hc ieai lr Ri ensvei taar tciho nS At Fo Ep) raensde nRta po uh ra epl aSpc he ro ednul rei (nBg r at hned eCi es nUt nr ai vl eBr as int ky Re xecsiet ea dr ctho Ah as vs oe cmi aat ni oynm(oCrEeBf Rr uAi )t f uAlnrneus ea al rMc he edtiisncgu s2s0i o2n3s iann dN teowc oYnotri kn.u Ie awmo r vk ei nr yg on the paper until its publication. Ic own inl le cf ot iroenvse ar nbde e txhpaenr ki ef nu cl ef so rt htahtet hgirseda its sneurmt abt ieorn ohf a ps rporfoevs isdi oe nd aml ea .n d p e r s o n a l
8 Table of Contents
9 Chapter 1 General Introduction 11 Chapter 2 HE mo wp i mr i cuaclhEdv oi dIennvcees tf or or smRtehl ey Uo nS Ca nr eddEi tURCaLt iOn g s : primary market 25 Chapter 3 Security Design and Credit Rating Risk in the CLO market 67 Chapter 4 Intensified Competition and the Impact on Credit Ratings in the RMBS market 125 Chapter 5 TUhS eS It ma t pe sa cotnotfhCer Re dMi tBoSr mP raortkeectt i o n i n 181 Chapter 6 RS ei sc ku rRi teitzeant ti oi onn Mi na rt hk ee t E: Su kr oi mp ema end b y Skin-in-the-Game Methods? 223 Chapter 7 General Discussion & Conclusion 277 Chapter 8 Summary (English) 293 Samenvatting (Dutch summary) 300 Publications 307 About the Author 313 Acknowledgement 317
10
11 General Introduction Chapter 1
12 General Introduction Over the years, regulators, policymakers and supervisors have put significant eq fuf oe sr tt iso ni n r termy iani gn st, oh oi mwpervoevr,e htohwe feuf fneccttiiovne i nt hge os ef tehf feo rstesc u( or fi tt ei znattiimo ne smi na rtkheet . f oTrhme oi nf trhuel ems aarnkde tr. eIgnu al ant iaotnt es m) aprtet oa ns dh ehdo lwi g hi nt voens ttohriss pr eelractei vi veel y t choemr pi slke sx el amn bd es cdadpeed, tt hh ies edvioslsuetri toant i oo fn ma ii tmi gsa ttioo np raot tvei md ep tas bb ey t tt ehre ug no vdeerrns ti na ng dbi no dg i eosf . tThhoes eg orai sl kosf at hnids di ni sdsi reer ct at ltyi o ni n vi so l tvheedn i tno tphreo vsiedceu rui st iezf autl i onne wm ianr skiegth, t sp af rotri caullla rplayr tfioers dr ei rgeucl tal tyo rosr, policymakers, supervisors and investors. This dissertation focuses on answering the following question: To what extent do factors beyond credit ratings affect securities’ credit quality, and to what extent do investors rely upon these ratings? Tmhairskient t raondduccot ov reyr sc thhaep tme ra ienx cpol ani cnesptths ea nk edyt hc he oa rr ai ecst etrhi as tt i cf os romf tthhee sf oe uc unrdi at itzi oa tni oonf trhe si se adricshs e rqtuaet isot ino. nT ha er e d it fhf ee rne ndte sr ce rs iebaerdc ,h f soul lbo-wq eu de s tbi yo nas uf unrdt he relry i no ug t ol i un re oovf e rt ha lel remaining part of this dissertation. 1.1 Securitization design and market overview Sa ne cdusrtirt ui zcat tui or endi si nt thoe tpr ar odceeasbsl ei ns we chuirciht i ve as .r Ti ohuess teyspeecsuor fi tlioeas ncsa na rbeebsuonl dd lteodi nt ovgeesttho er sr wu nhdoe, rilny i nt ugr np ,o roel coefi vl eo atnhse. Ti nht ee rseesct uarni tdi z aptriionnc i pp raol cpeasysms teanrttss gwe int ehr at ht eedo frri og imn atthoer
13 (i.e., a bank or other financial institution) which removes the loans from its own balance sheet by selling the asset pool to an issuer1. The issuer divides the asset pool into what is known as tranches or note classes, that each have a different of risk and return profile. The level of seniority of each tranche defines tl ohses ea sl l. oTc ahtei osne noifo irni tvye sl et mv eel ni ts ruest uu ar nl l ys (di i. ev .i,dperdi ni nc itpoa jl uanni do r,i nmt eerzezsat npi na ye ma ne nd t ss )e nainodr tranches. The most senior tranches have the lowest risk as they are the first taol s roe cc oe mi v ee sa wr ei tthu rt nh eo lf ocwa epsi tt ael xapnedc ttehde rleatsut rtno mb ea kai lnl og ctaht ee md lloosws e- rsi. s Hk ol owwe v- reerw, tahrids ab ses ea tl lso. cTaht eedml oossst ejsu. nHi oorwtervaenrc, hi ne sr ec at ur rr yn tt hh ee yh ibgehneesfti tr if sr ko ma st ht heehy i gahr ee stth ee xfpi resctt et do roef tpurrinn cmi paakli nagn dt hienmt e rhei sgth -prai sykmheingths- raentdu rtnh ea saslel ot sc.aTt ihoen roufl el os sfsoers t ha reedri es tf er irbr ue dt i ot no ac os rtrhees pwo antdesr f taol l tphaey mr i sekn tl esvt reul cot fu rt eh.e Etarcahn cthr aen(ceh. ge . , r es ce eniivoers t ar acnrcehdei t breaitni ng gAtAhAa t- r( Ca tReAds) ). , Tshuec hc raesdMi t oroadt iyn’ sg, sS at ar ne dparrodv &i d ePdo obry’ s c(oSm& Pm)e, racni adl Fcirt ec hd i It nrvaetsi nt ogr sa gSeenr vc ii ce es (oFbilticgha )t .i oCnRsA. sT hp er ocvrieddei tf oqruwa al i rt dy ,- ol oroski imn gp lpy epr us pt , etchtei vceaspoanb itlhi tey carne dd ipt rqeupaal ri teyd on fe ds se bo tf the debtor to complete payments on its debt obligations, is reflected in the credit rwaot iunl gd. oTthhee ri wn ti rsoe dhuacvt ei opne rocf eCi vReAds t ha lel soewseedc ui sr si tui ee rs sa st oo rpeaaqcuhe ma na dn yc oi nmvpe lsetxo r( sD at hl eayt eatndalp., e2n0s2io0n). fTuhnedsse(ucusuriatlilzyathioolndiinngvetshteorsatfyepsitcatrlalyncrhanesg)es from banks, insurers 2 to hedge funds, openet hned re ids kf uy nt rdasn, ca hs se es t) . mA asnt ya lgi ze er sd aenx da mo pt hl ee ro fi nt ht eer ns eact ui orni tai zl ai nt ivoens tdoerssi g( nu si us apl rl yo vhi od leddi ni ng Figure 1.1. 1 This entity is a bankruptcy remove legal entity known as special purpose vehicle or special purpose entity. 2 Other market participants are for example legal advisors, auditors, servicers, trustees and liquidity providers. Chapter 1 - General Introduction
14 Ts ehceu ur int idzeart liyoinnsg ba as cs ke tesd obf yt hae ps oe coul roi ft i rz ea at il oens tcaatne bme oorft vg aa rg ieosu as rtey pc ea sl l. eFdo rmeoxrat mg apglee-, bb ya cckoerdp soer ca ut er i ltoi easn (sM, aBnSd) . aCsos lel ta-tbear ca kl iez de ds el oc aunr i ot ibe lsi g(aAt Bi oSn) s a(rCeLcOo)ma pr er i tsheods eo fbaascskeetds sa us sceht apsosot ul di ne na tsleocaunrsi ,t ci zraetdi oi tnc ac radn rveacreyi ,vba ub ltetshaensde cc aurr iltoi az antsi.oHnednecsei, gt nh er eumn daei nr lsy ti nh ge ss ae mc uer .i tIinz at thi iosnddi sessei rg tna, t iroant h, eI ra mt h amn a si no ll ye l iyn tset ur edsyt iendg i tnh teh pe oi sosl uoef s l os ua rnrso uu nn dd ienr gl y ti nh ge tinhteesrecchuarnigtye.aTbhleertehfroorueg, Ihuosuet soeucrusrittuidzaytions with different underlying asset pools 3. The added benefit is that it allows us to study a larger and more diversified sample of the securitization market. Securitization is a financial innovation that was first introduced by US government ao gf etnh cei erse scirdeeant et ida lb ryetahl ee Us tSa tCeo mn gorretsgsa gi ne tmh ea r1k9e7t 0(sJ ot ob sfta, c2i l0i t0a8t e) . t hOen ldye ivne l tohpemleant et 1990s, European markets followed. The issuance volumes of securitization were 3 Only in some cases we include specific asset-related variables, such as the average house price for analyses on residential mortgage-backed securities (RMBS). Figure 1.1: Stylized example of securitization design.
15 rapidly increasing in the run up to the 2007-2009 Global Financial Crisis (GFC). Following the GFC, the securitization markets experienced a significant decline isni z ies, siuna2n0c 2e 1b, ut ht eh aUvSe ms ianrckee tb roeupnrceesde nbtascrko tuog hu lnyp4r 1e c%e daenndt et hd ehEeUi g hmt as .r kI ne tt er or mu gsh ol yf 7r o%u gohf l yt h€e2g3l 3o .b1abl isl el icounr ai tni zda €t i 3o n. 8 9m1abr ki lel ito(nS, &r ePs, p2e0c2t 1i v)e, lwy, ist ehenFeiwg uirses u1 e. 2 v. o l u m e s o f 4 Isno at hs i tshde isses emr at artki eotnm, Iofdoeclus sa op np etahre t oU Sb ea nqdu i Et eUs isme ci luarri t(iez .agt. i, ocno mmpaarrkeedt st oo nt hl yo. sIe di no tt ha regCe ht ei nd ebs ye mr e ag rukl ae tt i)oann tdh be er ec aa fut seer. tThhe isse cmr eaartkeest sa wu ne ri qe uhei t sbeyt t ti nh ge Gt oF Cc oamn dp ahreea av ni l dy cdoenvetrlaosptmbeonthtsmbaerfokreets, daunrdiningvaensdtigaaftteertthheeimcrpisaisc.t of regulation and other market 4 mTahrek eCt hi inn e2 s0e2 1s e. cTuhrei tri ze amt iaoi nn i mn ga mr kaert kreatps i wd liyt hi nac rr ee laastei vde il ny st hmea ll la spto dr tei coandien, crl eupdree Jsaepnat ni n, gA ur os ut rgahl il ay , 4C 0a %n a doaf , tahnedgLl oabt ianl America (S&P, 2021). Figure 1.2: Total amount of new issue volumes in the US and EU securitization market (in € billions). Source: Association for Financial Markets in Europe (AFME) and Securities Industry and Financial Markets Association (SIFMA). Chapter 1 - General Introduction
16 1.2 Securitizations’ rise to fame Sq eucaur rt ei tri zoaft i o2 n0 0p7l .a yMe adnay kme ya rrkoelte oi nb stehrev eGrFs Ca rt hg aute stthaar tt etdh ei n dtehpet hs eacno dn dl eonrg tt hh i rodf t( hr eef ecrrri se ids two aa ss “t shue brpersi uml te ob fo ra rno wo veerr”e) xt theant swi oenr eosf emc uorrittgi zaegde sa nt od swoel da kt obionrvreoswt oerrss. Investors overpriced these securities as the credit ratings attached to the tranches rr ae tf li encgtse dw ear emtuocoh omp toi mr ei spt oi cs iat ni vde doi ud t nl ooot kr et fhl ea cnt wt haes aacct tuuaal l cl yr etdhiet rciasske o. fTah et r ac rnecdhiet (also known as inflated ratings). When mortgage rates started to rise, debtors, ep sapy emc ei anl tl ys hhooc mk . eTohwi sn reer ss u lwt eidt hi n ahdi jguhs tdaebf laeu- lrta rt ea t ems oa rntdg acgoenss, e qf auceendt l ya h img ho rl ot gsas ge es froerl i ba no tche i on nv e cs rt oe dr si ti nr at ht ien sges nai onrdt rt ahne ci rh easpapnedt i tj ue ntioo ri nt rvaensct hiens .s Ne caut ruirtai zl layt, iionnvse swt oerrse’ significantly reduced after the crisis5. It is alleged that this chain of events was predominantly caused by the CRAs assigning inflated credit ratings as a result of cmoomdpeel ti si t isvter upcrteusrseudr e( sseien, et h. ge. , rDaet i nHga amna&r kAemt atne ndbtrhien kw, a2y0 1i n1 )w. h i c h t h e i r b u s i n e s s 1.3 Issuer pays revenue model Twhaey ipna rwa hmi cohu nr te vreenaus oe ni s wg he yn eCr aRtAe ds ba sys iCgRnA tsh; et shee i‘ insfsl ua teerdp ac ryes d’ irte vr ea nt iuneg smios dtehl e 6. . In this model, a conflict of interest might arise as the issuer is the key client of the CRA. Issuers are likely to select only the best, most optimistic rating to ot hpetiirm iczlei e nt ht se’ i r( i p. er. ,o fiist ss u. eCrosn) s we qi us heenst l yb, yC RpAr os v imd iing hg t mboe r ei n coepnt ti mi v iizs et idc t(oi n fpl al et ea ds e) 5c oEmx pa lnetxe ,niantvuerset oorf ss we ceurrei tfioz ua nt i do nt ot hs ai gt nmi fai ckaens t il ty mr eol yr eodni ftfhi ceucl tr ef od ri t i rnavtei ns tgo irns tt oh emi rarkies kt haesisreos swmn erni st ,kma as si ne lsys md ueen tt o( stehee, e.g., Furfine, 2014). 6 See, e.g., Flynn & Ghent (2018), Griffin et al. (2013), He et al. (2016) and Zhou et al. (2017).
17 ratings, rather than those that reflect the securities true credit quality7. The conflict of interest between the issuer and CRA is broadly described in lpi ht eernaot umr ee nboyn t. wT oh ed or amt ii nn ga nsthtohpepoi rni ge st: ht ehoer “yr as tt iant ge ss ht hoaptp ii snsgu”ear ns da r“ er a at ibnl ge ct ao t es roilni cgi ”t preliminary credit ratings from numerous CRAs and only select those of their liking. Only the selected CRAs benefit from this process as they receive the full pc oany mt r aecntt i no f rteht ue rins.s Fuoe lrlso, wwi nhgi l et htihset hoet oh reyr,sorneec emi vi eg hot nal rygau emt ihnaot rt hf eeei sf os ur ebrrceaanc hp uo tf pr artei snsgusr)e, roant hCeRrAt sh taon aas sriagtni nrga ttihnagts stohlaetl ys artei fslfeyc tt hs et hwei suhnedseor fl yt hi negi scsrue ed ri t( qi . eu. a, il ni t fyl aot fe da sc er ec du irti trya. t Ti nhge s rt aa nt i dn ag r cdast be raisnegd tohne oc or ymbpue itlidt isv eu pp or ens st hu irse aosf tChReAi rspme ei gr sh. tI na dpjaurstti ctuhleairr, they might try to match their standards (i.e., by inflating their ratings) to those of their competitors to reduce the chance of rating shopping behavior by issuers. The unfolding of the global financial crisis has revealed the misfunctioning of the ss ek ci lul sr oi t fi zCaRt Ai osnh amv ae rbkeeet n. Twhi de esl yc rqeueensi tni go nceadp tahbei lriet iaef st e or . fC oo rni sgei nq auteonr tsl ya, pn od l itchyemraakt ei nr gs aTnhde Dreogdudla-FtorarsnkhaAvcet created more and stricter rules and regulations thereafter. 8 was implemented in the US market and, similarly, a variety of regulations have been proposed in the EU9. These rules and regulations have, aI nmtohne gUs St omt ha er kr et th, itnhgi ss , ws oa us gdhotnteo brye di nuccree at hs ien ag bt irl ai tnys op fa ri sesnuceyr so ft oi nsf oh romp af ot iro rnaitni ntghse. rating market. In the EU, issuers are now required by regulation to disclose at 7 a Tn hd eWd oo ludbCtos mr e, gCaRr dAisn wg tehr ee reexl ti ar ebmi l iet yl y olfact re eidni tardaj tui ns tgi sn gd at theesi rb ac cr ek dt iot br ae tf oi nr ge twh he eGnF tCh. eF ocro emx pa amnpi el es , ei nx pt he er i ceanscee od f sEenvreor ne difficulties (see e.g., Boot, 2006). 8 SuDbosdedct-iFornan1k5GA. ct, 2010, Dodd-Frank Wall Street Reform and Consumer Protection Act of 2010, Section 941 9 (ERCe)guNloat1io0n60(E/U20) 0N9oo4n62C/R2A0s1. 3 of the European Parliament and of the Council of 21 May 2013 amending Regulation Chapter 1 - General Introduction
18 laenads tottwh eorcrrue ldeist ar antdi nrgesg uo lf awt iho inc sh hoanvee sahl soou ltdr i iedde taol l ys tbi me iuslsaut ee dt hbey eanst mr a anlcl eC oRfAs. mT ha lel se er or newer CRAs to reduce the duopoly of Moody’s and S&P in the rating market. 1.4 Transformative banking Tt oh- eh op lrdo cme os sd oe fl st oe caunr iot irzi ag ti ni oant eh- at os -tdr ai snt sr if bo rumt eemd tohdee rl .oTl er aodf ibt ai onnkasl lf yr ,obma na kn so ar icgt i an sa taen- it nh teei rr me xepdoi as ruyr ebse ut wn teiel nmbaot ur rroi twy . eIrns tahni sd odrei pg ionsai tt oe r- tso, -whho el dr emboadnekls, rhi os kl dma na nd amg eomn iet on rt is achieved mainly by constructing a well-diversified portfolio. The securitization tt er ac dh en ai qbul ee , l ihqouwi de vs ee rc, u or if tf ei erss abnadn kr se mt hoev eo pl opaonrst uf nr oi tmy tt ho e ti ru rbna l ial nl i cqeu isdh eaest s e– t sa gi notaol sboe ul ogwh .t I bt yt hbe ar enbkys rbeemc aouvsees po af rht i og hf t hc aepci rt ea dl irteeqxupi roesmu reens t fsr of omr tshuec hb al no ak n’ ssbaasl aenxcpel as ihneeedt at on dt rfarne es fseur pc cr ea dp ii tt aal nt od i spsouret fpo ol i toe nr itsi akl on fe wb alno ka sn st .oI nc aa pd idt ai tli omn a, irtkpertos v(i sdeees, ae m. g .e, cAhyadni ni s m& Altunbas, 2016). The securitization process thus intertwines banks with financial markets, creating more risk sharing between them but also make banks vulnerable to volatility in financial markets (e.g., Boot & Marinč, 2010). Hi mopwl iecvaet ri o, ntsh et h aot r iwg ienraet er-et ov -edailsetdr i bduutrei n gm tohdee l G FaCl s. oB ahnakss st eonmdee d s teor i or ue ds u caed vtehresier st hc reeseonl ei npguarnpdo sme ot on isteocrui nr igt isztea. nI tdaalrsdose, np ha ratni ccue lda trhl ye froi rs kt haopspeeltoi taenos f tbhaant kwsearse tihs es yu ekdn ef owr ti mh apt l itchaet i roi ns ks so fwt hi l el bo er i gt ri an na st ef e- tror -eddi st tor i tbhui rt ed mp aordt ei el ,sr ea gn uy lwa taoyr. sF ho al l voewi innt gr otdh ue cs ee dntehgea rt ii vs ke retention requirements after the GFC. The risk retention rule requires the originator or sponsor to retain a significant portion (5%) of the securitization on their balance sf rhoeme t dt hi frf oe ruegnhto ruet gtuhlea tl oi freyomf ae tt hr aondssa ct ot i or ne t. aTi hn et ht rea npcohret i or ent aoi fn et hr ei ssaelcl ou wr i tei dz attoi ocnh. oTohs ee
19 purpose of this rule is to incentivize banks to make good loans by ensuring that they have some ‘skin-in-the-game’ (see, e.g., Daley et al., 2019). 1.5 Research sub-questions To address the overarching research question, ‘To what extent do factors beyond credit ratings affect securities credit quality, and to what extent do investors rely upon these ratings?’, I have established five sub-questions. The first question fcorceudsitersaotinngtsh:e reliance of investors on credit ratings and the factors beyond 1. Hc roewd i td roa ti innvgess ti on r ds erteelrymoi nn i nc rge tdhi te rf au tni nd gi ns g acnods to, tahnedr hf aocwt o dr so bUeSy oa nn dd EU market differences impact these relationships? The second and third sub-questions relate to the implications of the ‘issuer pays’ business model and competition between CRAs: 2. Tr aot iwn gh asth eoxptpeinntg i sa nt dh er actoi mn gp lceaxtietryi nogf ba esheacvuiroirt ys ,’ sa nd de s hi gonwr edloaet es dt ht oe GFC impact these relationships? 3. Hquoawlitdyoaensdcormatipnegtisttioanndbaertdwse?en large and small CRAs impact rating Tthheeirfocruerdthit qrauteinstgiso:n touches upon the consistency between CRAs in providing 4. Tr aot i nwghsa tg i veexnt e nt ht e ddoi f fCeRr eAnst pl ervoevlisd eo f ccornesdi isttoern tp raontde c trieolni a ba lcer ocsrse dUiSt states? Chapter 1 - General Introduction
20 In the fifth and final sub-question, we investigate if differences exist between the various regulatory risk retention methods that are currently in place in the EU market: 5. Td oi f f ewr ehnatt r ee xg tuel na tt o rdyo r iisnkv ersettoernst i oann dm eCtRhAo ds s di ne v pi artiec i nbge tawn ede nr a t ti nh ge securitization tranches? The results derived from all the five sub-questions will eventually allow us to answer the overarching research question. 1.6 Dissertation outline Chapters 2 to 6 are based on five stand-alone papers that each answer a separate st iumbe- qpueersi toido ,nm. Fairgkuertea1n. d3 svei scuuarliitzi zeas tti ho en ot yupt leo uo skeodf itnh iesadc hi s ss et ur tdayt. i o n , i n c l u d i n g t h e In Chapter 2, we seek to investigate the degree to which investors rely on credit ratings in pricing CLO tranches that were originated and sold between 1997 and Figure 1.3: Dissertation outline.
21 2015 (first sub-question). Furthermore, we compare the behavior of investors in the US and EU primary CLO market. In Chapter 3, we study the extent to which tcha et ecroi nmgpbl eexhiat yv ioofrai ns etchuer pi t ryi’ms da er ys i gC nL Oi s mr eal ar kt eedt t( os eccroe nd di t sr ua tbi -nqgusehsot ipopni)n. gWaen da nr aa ltyi nz ge bd oi f tf he rtehnec Ue sS tahnadt EmUi gmh at rekxei st tf reoxm- a n1 t9e9 a6 nt do 2e x0-1p3o sa tn tdh feo cs ut as r st poef c ti hf i ec agl ll yo boanl pf iontae nn ct ii aa ll crisis. In Chapter 4 we explore our third sub-question by analyzing if, and if so hr aotwi n, gcso amnpde ct irtei od ni t br aettiwn ge se ns t as mn daal lr dasn. dWl ea rugsee CpRr iAms airmy pma ca trsk et ht ed aqt ua aol if t rye os ifd cernetdi ai tl mthoerUtgSagaen-dbaEcUkemdasrekceutrbiteietws e(RenM2B0S1) 7traanndch2e0s 2w0h. iIcnh are originated and sold in Chapter 5, we restrict our sa anmd pr el el i at ob lteh ec rUe Sd iot nr layt, i tnog isngvievsetni gtaht ee dt hi fef eerxetne tn ltetvoe wl s hoi fc hc rCe Rd iAt os rp pr or ov it de ce t ci oonn saics rt oe ns st Ut hSa st twa teerse(of orui gritnhast ue db -aqnude sstoi ol dn )b. eWt we eset und2y0R1M7 BaSn dt r a2n0c2h0e. sI na t t h e t i m e o f i s s u a n c e Chapter 6, we focus o2 n0 2t h1 e, ttoo tsat luEduyr tohpee ai mn ps ea cc ut roi ft itzhaet i do ni f fme raernkte rt e( gAuBl Sa ,t oC rLyO r, iMs kB Sr e) ,t ienn tt hi oenpme rei ot hdo2d0s 1o1n- the pricing and rating differences of securitization tranches (fifth sub-question). Ts ehvee rqaul aonrtdi ti antai vr ye rl ee as se ta rs cqhu as rt ue sd i(eOs Li Sn ) Cahnadp t(eorrsd 2e r teod )6 laorgei tc roengdrue cs tsei od nb my ao pd pe ll ys i on ng pooled cross-sectional data (tranche-level). Finally, Chapter 7 discusses the main findings of each chapter and how they, cr eo smu bl tisn, ewd ,e asnpsewc ief ry tt hh ee coovne trrai rbcuhtiinogn raensde arreccho mq umeesnt idoant.i oBnus i lodfi nogu ru ps tound yt haens de cfuotnucrleudreesebayrcphr.oviding limitations of our research and suggested avenues for Chapter 1 - General Introduction
22 References Association for Financial Markets in Europe (2022, 24 October). AFME securitisation dDaettaa isl ns /aApFs hMoEt :- SQe3c u3r0i t3i 3s a. thi totnp-sD: /a/twa -wS nwa.pa sf mh oet. e- Qu 3/ P- 2u0b2l i2c a t i o n s / D a t a - R e s e a r c h / Badoer, D. C., Demiroglu, C., & James, C. M. (2019). Ratings quality and borrowing choice. Journal of Finance, 74(5), 2619-2665. https://doi.org/10.1111/jofi.12820 Boot, A. W. A. (2006). De toegevoegde waarde van credit ratings. Maandblad voor Accountancy en Bedrijfseconomie, 80(3), 108-117. Boot, A. W. A., & Marinč, M. (2010). Financial innovation: Economic growth versus instability in bank-based versus financial market driven economies. International Journal of Business and Commerce, 2(1), 1-32. Daley, B., Green, B., & Vanasco, V. (2020). Securitization, ratings and credit supply. Journal of Finance, 75(2), 1037-1082. https://doi.org/10.1111/jofi.12866 Flynn, S., & Ghent, A. (2018). Competition and credit ratings after the fall. Management Science, 64(4), 1477–1973. https://doi.org/10.1287/mnsc.2016.2604 Furfine, C. H. (2014). Complexity and loan performance: Evidence from the securitization of commercial mortgages. The Review of Corporate Finance Studies, 2(2), 154– 187. https://doi.org/10.1093/rcfs/cft008 Griffin, J. M., Nickerson, J., & Tang, D. Y. (2013). Rating shopping or catering? An examination of the response to competitive pressure for CDO credit ratings. Review of Financial Studies, 26(9), 2270–2310. https://doi.org/10.1093/rfs/ hht036 Haan, J. de, & Amtenbrink, F. (2011). Credit rating agencies [Working paper]. DNB Working Paper No. 278. https://doi.org/10.2139/ssrn.1760951 He, J. J., Qian, J. Q. J., & Strahan, P. E. (2016). Does the market understand rating shopping? Predicting MBS losses with initial yields. Review of Financial Studies, 29(2), 457–485. https://doi.org/10.1093/rfs/hhv067 Jobst, A. (2008, September). What is securitization? International Monetary Fund (faIMndFd)/F2in0a0n8c/e09&/Dpdevf/ebloapsmicse.nptd.fhttps://www.imf.org/external/pubs/ft/ Securities Industry and Financial Markets Association. (2008). ESF securitization data rseepcuorritt.ihsatttpiosn:/-/dwatwa-wre.spifomrta-.2o0r0g/8w-qp4-.cpodnftent/uploads/2017/05/afme-esfS&P Global Ratings. (2021). Global structured finance 2021 outlook. https://www. spglobal.com/_assets/documents/ratings/research/100048329.pdf Zhuo, X., Xu, G., & Wang, Y. (2017). The issuer-pays business model and competitive rating market: Rating network structure. The Journal of Real Estate Finance and Economics, 55(2), 216–241. https://doi.org/10.1007/s11146-016-9563-2
23 Chapter 1 - General Introduction
24
25 Journal of Financial Services Research, 2022, DOI: 10.1007/s10693-021-00372-x Frank J. Fabozzi Vivian M. van Breemen DM ei kn en iNs aVwi na ks Austin Gengos How much do Investors Rely on Credit Ratings: Empirical Evidence from the US and EU CLO primary market Chapter 2
26 Abstract Wf a cet oi nr sv ebset yi goant de ct hr eed ei tx treant itn gt os iwn hdi ceht e irnmv iensitnogr st hree l fyu no dn i nc gr ecdoi ts t r faot ri ncgosl l aa nt edr aol ti zheedr loan obligations (CLOs) tranches in the period 1997-2015. We find significant differences between the United States (US) and European Union (EU) markets. In the US., we find a much higher and more consistent degree of reliance on credit ratings and other factors in pricing CLOs over time compared to the EU market. Finally, we find that investors in both markets reduce funding costs when rating standards loosened. The regulatory implications are discussed. Keywords: credit ratings, collateralized loan obligations, regulations, structured finance. JEL Classifications: G12, G24, G28, G32, L11.
27 2.1 Introduction Ta nhde Fci rt cehd. iTt hreast ei nt gh rienedcur se tdriyt riast i dn og ma gi ennact ieeds (bCyR AMso) ohda yv e’ s r, oSut ga hnldy a9r1d%&o fPtohoerm’ s a(rSk&e tPi )n, Europe and 95% of the market in the US (European Securities and Markets Authorities (CEoSl lMa tAe r) ,a l2i z0e2d0L; oSaenc uOrbi tl ii eg sa t ai onnds (ECxLc Oh as n) gi sea Cs oe gmmmeins ts ioof nt h(eS sEtCr u) , c 2t u0r2e0d)f. i nTahnec emsaerckuerti t fi oe sr market10 and in assigning the credit ratings of CLOs the dominance of only two of CRAs b2 e0 i0n8g, mC Ro Ar es pwr eorneo aucnccuesde, dMoof oadsys ’isg na ni ndg Sb&i aPs. eI nd trhaet i wn gask teoosf t trhuec tgul or ebda l f if ni naannccei asleccrui rs ii tsi eosf snuoct ha pa ps rCoLpOr si a(tee. lgy. , rGe rf li ef fci nt tehtearl i. ,s2k 0s 1a 3s s) oocri, amt eodr ewgi tehnCe Lr aOl ,s t(os eh ea ,vee. ga. s, Fc raibboezdz ri a&t iVnignskt, h2a0t1d0o; Flynn & Ghent; 2018; He et al., 2016; Zhou et al., 2017). Due to the complexity of CLO structures, investors are exposed to the risk that the assigned credit rating does not fully or precisely reflect actual credit risk (Vink et al., 2021). Concerned that investors may rely too heavily on potentially biased or inflated rCaRt Ai nsg isn, at tht ei t ui nddeus st torwy ahradss ct hh ae nrgoel ed obf rCoRa Ad lsyaannddt, hi en ds oo mm ien ca anscees o, fc rt hy se ttahl irzeeedl aartgtehset regulatory level. The Dodd-Frank Act11 in the United States (US) and regulations12 in the European Union (EU) have sought to reduce the reliance on credit ratings, especially for structured finance securities. The stated goal is for the market to mc aol lv ef oarwaany fermo mp i rrieclai al ni cnev eosnt i cgraet di oi tn r aa ti mi n eg ds . Ta th eb Ue tSt earn du nEdUe rr es tgaunl ad ti on rgy trhees peoxnt es en st to which investors rely on credit ratings in the CLO market. This improved 10p eIrni o2d0 1i n8 t, hCeL OE Ui s smuaarnkceet iint wt haesUaSpmp raorxki emt aatme loyu€n2t e8dbtiol l iroonu g( Bhlloy o$m1 2b5e rbgi,l l2i o0n1 9( S) &. P G l o b a l , 2 0 1 8 ) a n d a t t h e s a m e t i m e 11SubDsoedcdti-oFnra1n5kGA. ct, 2010, Dodd-Frank Wall Street Reform and Consumer Protection Act of 2010, Section 941 12(ERCe)gNuola1ti0o6n0(/E2U0)0N9oo4n6C2R/A20s.13 of the European Parliament and of the Council of 21 May 2013 amending Regulation Chapter 2 - How much do Investors Rely on Credit Ratings
28 uefnfedcetrisvteannedsisnogf lceoguisldlatiinont.urn inform policymakers seeking to improve the Wc r ee di ni tv reas tt ii ng ga tse atshsei gdneegdr ebeyt oC RwAhsi cihn i tnhteh ep rUi cSi na ng dt hEeU CmL aOr ka et tt ihnev et ismt oer so rf ei lsys uo anntchee. Using data of CLO tranches that are originated and sold between 1997 and 2015, we first test the extent to which investors rely on CLO credit ratings by Moody’s aq nu do /t eodr mS &a Pr g, ians aetviisdseunacnecde . bWy et hael s roe el axtai omni snhei pi n bv ee st wt oer erne lci ar endc ei t orna t oi nt hg se ra fnadc t tohr es beyond credit ratings that are specific to the CLO market, which we refer to as “i ns ev ce us troi trys dper si ci ge nt ”h feai rc tionrvse. sI tnmt ehni st swi anyC, wL Oe sgoa ni n t ah ne ub an sdi es ros ft acnr eddi ni tg raast i tnog st hoer doeng tr he ee basis of other factors that influence their investment appetite. W“ s tee at dh ye -nb ocoomn s-ibduesrt -irfe cs ouvc ehr yl e” vpeel sr i oodf sroe bl i saenrcvee dc hi na nt hg ee dC LoOv emr atri mk eet ,, at nh dr owu ghhe t ht heer there are further differences to be observed in that regard between the US and tt hh ee bE uUsmi naersks ectysc, lae c(oBnasri-dI searaa ct i &o nSphraopmi r po ,t e2d0 1b3y ; rDa ti li lnyg&mMo ädhe ll ms baeninn,g2i0n1f l6u)e. nWc ee dt ebsyt wa nhde ft rheeqr ut ehne t pvaetrt seur ns si nof br es qe ruveendt ii sns tuheer sUiSn at nh de steh emEa Ur kde itfsf ,egr i vf oern l sa ur gg eg evsetri sounss si nmtahl el , literature that investors may vary their reliance on credit ratings, depending on tsht reutcyt pu er eodf fiisnsaunecret hr aetyi nagrse, ds teuadl iinegs wh ai vt he (f oHuenedt tahl .a, t2C0R1 A2 )s. mO ua tys ihdaev eo fvtahrei efdi etl hdeoi rf raantai nl ygs isst, awn ed at er ds ts wo hv ee rt ht ei mr ien v( sees et oer.sg .t, aAklep i, n2t0o 1a3c )c. oIunntthceh laans gt epsairnt roaf t oi nugr setma npdi raircdasl in pricing CLOs in the US and EU markets. Os uubrs traenstui al tl sl y sghr eoawt e rt heax tt e ni nt otnh ec r eUdSi t mr aat irnkgest , a so nt h eayv edreat ge er mi innvee tsht oe rfsu nrde il ny g tcoo sat
29 ot hf i Cs LgOr etartaenr cehxetse nt ht aonf ii nn vv ee ss tt oo rr sr edloi ainnc tehoe nE cUr emd ai tr kr ae tt i. nAglss oi s, omuor rree cs uo nl t ss i ss theonwt ot vh ea rt tainmde EtUh ai nn vi ens tt oh res EcUa nmbaer keextp. lNa ienxetd, wb ey ci )o bn us isdi ne re si sf tchyecsl ees di ii f)f et hr eenicme sp abcett wo feiesns uUeSr sf iinz ae nocri a il i ci )r ics hi sa, nt hg ee sr ei lni arnactei nogn sctraenddi ta rr da tsi.nWg ser ef imn da i nt headt mt hoer ye odrol. eFs si rsstta, balfet ei rn tt hh ee US but decreased significantly in the EU market. Second, funding cost required bi ny t ihnev eUsSt odros ni no tt mh ea Ek eU smu cahr kdei tf f ae rr ee ndtiifaf et iroenn. tTbhai sr de d, oounr irsessuuel tr sssi zheo, ww thhi al et ii nn vv ee ss tt oo rr ss dmeomreansod ianlothweeUrSfutnhdaninignctohset EfoUr mCLaOrksewt.hen CRAs loosen their rating standards, To the best of our knowledge, this paper is the first to compare investors in the Ur aSt i anng ds Ei nU pCrLi cOi nmg aCr kLeOts waitt hi srseusep. e Oc tutro stthued ye x ct eonntt rtiob uwt ehsi c ht ot hae yr erceel ny t o bn ocdrye doi tf l(istee er aet.ug r. , eMoanr qcur ee ds i&t rPaitni nt og, s2a0n2d0 ;cYr ae nd gi t estparle. ,a2d0s 2i 0n ) t, ha en ds ttroutcht ue rl ei tde rfai nt ua rnec eo nmcarrekdei tt rating standards (see e.g. Alp, 2013; Cafarelli, 2020). Our results are relevant t(od ipveorl igciynmg )alkeegriss liant itvheef rUaSmaenwdo rEkUs soeneckri endgi tt oraitmi npgrso. ve t h e e f fe c t ive n e s s o f t h e i r Tt hhee rreelsetv oa nf tt hl ei t ecrhaatputreer ri es l oa tregda ntioz ecdr eadsi tf orlal ot iwn gs .s .S eS cetcitoi on n2 .22. 3c odnetsaci rnisb ea sr eovuire wC LoOf t2r. a5 npcrhoev iddaet sa aa nd di s cSue sc st ii oo nn , 2c .o4npc rl ue ss ieonnt sa nt hde s reet ss uol ut st op fo ol i uc yr iemmppl ii rc iactai ol nt ess. t s . S e c t i o n 2.2 Literature Review The academic literature on structured products has benefited greatly from amarked increase both in theoretical and empirical studies of market reliance on Chapter 2 - How much do Investors Rely on Credit Ratings
30 cr er eg da ri tdrl ea st isn og fs . t Ihne f pa cr to, dt uh cet mt yapj oer, iftoycoufs sst ou md i ee ws ph ea rt ,f oi fr mn oetd porni ms tarrui lcyt, uorne dt hper ordo ul ec tosf, the CRAs13.This is likely due to the widespread belief that CRAs assign favorable ratings, especially to structured finance securities. The issuer-paid business model in place for the entire CRA industry gives CRAs an incentive to cater rHaet ientgas l .t, o2 0i s1s6u;eZr sh’odueemt aanl . ,d2(0s1e7e ), . eC. gR. ,AFs l ay rnen a&c c uGsheedn to, f 2h0a1v8i n; gGcroi fnf itnr i be tu tael d. , 2t o0 t1h3e; ds terputcht uarnedd lf ei nnagnt hc eosfe tchuer igt il eo sb (asl ef ei n, ea .ngc. , i Fa ll ycnr ni s &i s Gbhy eansts, i2g0n1i n8g; Hf aev eotr aa bl . l2e 0r1a6t i; nZghsotuo, ei nt vaels. ,t o2r0s1r7e)l i, eedi tthoeor hdeuaev i ltyo opno corre dr ai tt irnagt i ns tgasnidnaervdasl ui na t ti nh ge iar sas ne tasl.y s i s o r b e c a u s e In addition, CRAs are found to more likely issue less-accurate ratings during bMoäohml mpaenrni o, d2s0 1( s6e; eHee. g e. , tBaalr. ,- I2s 0a 1a c2 )&. BS ah ra-pI si raoa, c2a0n1d3 ; SBhoa lpt oi rno e(t2 a0l1. , 32)0, 1f o2r; De xi lal ymapnl ed, gc oi vme ma enr cuima l bme ro toi vf ees xopfl aCnRaAtsi otnosmf oa xr i mt hiizse br ue tsui nr ensssa nc ydctl he eedf fee-cetm: rpehpausti sa toi onnc rrei ds ki t, mbooonmit)o.ring by CRAs in periods of low default probability (i.e., in an economic Another factor that is found to influence rating quality is the size of the issuer. He et al. (2012) examine the role of the CRAs in the rating of private-label residential mortgage-backed securities. They find that larger issuers experience he ai grhl ieerr fsut unddyi nogncCo Rs tAs st ha na nd smmoar ltlgeargies-sbuaecrkse. dT hs eeci ru rrietsi eusl ttshaart es uc og gnes si st tseinnta wc ciut hr aac ny of ratings related to issuer size (He et al., 2011). 13p r oTdhuecr tes .aFr oe r set ux admi e ps l et h, aDte kf ouc ue st aoln. ( 2o t0h1e9r) ,f aucstionrgs abseasmi dpe lse corfe 4d ,i2t 0r1a tEi nugr otpheaat nd oe trei gr imn ai nt ee dt hMeBpS rti rcai nngc hoefs ssthr uo wc t utrheadt tfihneaqnuciaallitcyriosfisth. e trustee has an impact on the pricing of structured finance securities during the most recent global
31 In the rating of other credit products such as corporate bonds, there is an eBxetceknesri v&e Mb oi ldbyooufr lni t, e2r0a 1t u1r; eCoanf atrheel l qi , u2a0l i2t 0y )o. fBrlautmi neg es tt aanl .d(a1r 9d 9s 8( s) eseh eo. wg . ,, uA sl pi n, g2 0S1&3P; bd oe cnldi nrea ti ni n gc rse, dt hi ta tq ut hael i tnyuomf bceorr poof rcarteeddi te br at ,t ibnugt droawt hnegrr abdy eCs Ri sA sn oatp pc al yui snegdmboyr ae sAtlrpi n( g2 e0n1t3 r) apt ri no gv i sdteasn md aorrdes ei vni dt he en cUeSt hma at rokveetr. At hger epeeirni og dw1i t9h8 5B -l 2u 0m0e2 ectr ae ld. i (t 1r9a 9t i 8n )g, sgtraanddearradtsinvgasr.ied, with divergent patterns for investment-grade and speculative- These findings suggest that there are factors outside of the bond structure or cc oo lnl as itsetreanl t i tws ei tl fh t Fhaabt oazf zf ei cat ncdr eVdiint kr a(t2i n0 g1s0 )a nadn dt hMe aprrqi uc iensg aonfds ePci nu troi t i(e2s0, 2w0h)i cwh hios found that investors look beyond the credit rating in determining the funding cost of structured finance securities. Our assessment of the literature is that rating quality can be impacted by (1) business cycles (see e.g., Bar-Isaac & S2 h0 a1p2i)r, oa, n2d0 1( 33); Dc hi lal yn g&e sMi änhrl amt ianngn s, t2a0n1d6a)r d( 2s )( si seseu ee .rg .s, i Az el p(, s2e 0e 1e3. g; . C, Ha fea reet l al i l, . ,22002101) ., Twhhies t hper or vi ni dveess tfourrst ht ae kr ems oe tciuvraittiyo nd efsoirg no uf ar c st ot ur sd yi ,n ti no awchc oi cuhn twwe hseene kp rt ioc ienxga CmLi On es adti ftf he reetni mt i aet eo fi ni s stuhaen pc er .i cWi neg boufi l dC LuOpso nb et ht we seee ns t ubduisei sn easnsd ci yncvleesst, i gi sast ue ei rf i snivzees taonrds wi n hoeut hr ea rn ar al yt si ni sg bs eatrwe eiemn pCaLc Ot esdi sbsyu ec dh ai nn gt eh se iUnS raant idn gE Us ttaon sdtaurddys .dWi f fee rdeinf fceerse ni nt i taht ee ut ensdt et rhl ey i dn eg gfraecet ot ros wwhhi icchh ihnavvees ttohres grreel ya toe ns t tihme praact ti nogns tahses ipgrni ce idn gb yo fCCRLAOs saat nt di mt oe of issuance. Chapter 2 - How much do Investors Rely on Credit Ratings
32 2.3 Data and Methods We begin the process by manually collecting data obtained from Bloomberg, w$ 1h. 0i c5h tpr irlol ivoind,etshaa tc owme rpel ei tses uuendi vaenr ds es oo fl d8 ,i3n2t4h eC LUOS torra nEcUh ems awr ki teht saf rt oo tma l 1v9a9l u6e uopf tcoh a2r0a1c 5t e. rFi os tri cesa, c ph rCi cLeOddaet ea ,l , tthhee 3d-amt aosne tt hp rboevnicdhems da reka/l raenf de rterna cnec hi ne tne ar ems et sr, ai tses uf oe rr the floating-rate tranches, credit ratings, balance and primary issuance spread.14 ches are rated by either Moody’s or S&P, or both. There are an insufficient number oT fh eCrLeOf osr er ,a ti end obuyr Fd iat tcahs eot r woet hue sr e s mo nallyl e trr aCnRcAhse st ot heant a obbl et asi nt aetdi s tai c raal t iannga l fyrsoems . Moody’s and/or S&P, consistent with the dataset used by Griffin et al. (2013). We apply several filters to our dataset and remove tranches with incomplete it nh feo rnmu ma t bi oenr . oBf eccraeudsiet wr aet i an rges , i nwt ee roe ns tl ye di ni nc l ut hdee ei nf f eocut ro fs tCuLdOy s CdLeOa sl ct or amn pc hl eexsi t wy iot nh af rto lme a8s ,t3 o2n4 et oc r7e, 9d 1i t0r. aWt ien gf u dr ti hs celro ds ei sdc aartdi sasl luter. aTn hc hi se sr ewdiut hc ems ios us irn go rtirgai nn saal cstai omnpol er tranche size (14 tranches) and missing information on the funding cost at issue (305 tranches). This filtering resulted in a full sample of 7,591 CLO tranches, of wt hhe i cEhU 5m, 9a3r5k et tr.aPnac nh ee lss aAr et oi sDs uoefdTianb tl eh e2 .U1Sr me paorrktes t saunmdm1a, 6r 5y 6s ttar tains tcihc es sf oi srstuheed Ui nS and EU market, respectively. 14 If fixed-rate tranches were to be included in our study, then it would be necessary to determine the appropriate benchmark yield curve for each tranche in the sample in order to obtain primary issuance spreads that could be consistently compared across the sample. By restricting the tranches in our sample to 3-month floating-rate tranches where the reference rate is the same interest rate benchmark, we avoid this problem. Furthermore, in constructing the final sample, we had to eliminate some tranches due to errant data or metrics that represented v co as s t t l . y atypical observations. For our analysis, we want to have a consistent benchmark for assessing the funding
33 2.3.1 Empirical Model Wr a et i ni ngvseasst si gi gant ee dt hbeydCeRg Ar esei nt ot hwehpi crhi cEi nUg aonf dC LUOS amt at rhkeetti mi nev eosf tiosrssu raenlcyeo. nTot heex acmr eidniet ti shsi us ,a wn cee l oi no kt haets et h tewi om pmaacrt koe ft st hues icnrge doi rt dri antai nr yg loefa Cs tL Os qs uoanr etsh e( Of Lu Sn)d irnegg rceosssti oant ai nn aS leycstiiso, nw2h.i2c ,hwi se caolns os i setxeanmt iwn iet hi nHvee settoarl .r(e2l i0a1n6c )e. oBna soet dh eornf oa cutrolri st ebr ae yt uornedr ec rvei edwi t ratings that are specific to the CLO market (i.e., security design factors). We are primarily interested in the following for each market: (1) the size of the credit rvaatl iuneg ocfotehf ef i ccireendti tc or anttirnogl l ceodeffof irc ti ei mn te aasnmd ei sassuuerre df i xbeyd t ehfef eacdt sj u, s( 2t e) dt h e e x p l a n a t o r y R², and (3) the si ne cduertiet rymd iensiinggn tfha ec tporrisc et haatt i isns vuees. tToor sa ct ah ki eevien tt ho i as ,c wc oeupnet rbf oe yr mo nsde tvheer aclr reedgi rt ersast ii on ng ss that are generally based on the following model: Spreadijt = β0 + β1 Credit Ratingijt + β2 Tranche Countijt + β3 Capital Allocationijt + β4 Log Tranche Sizeijt + β5 Log Transaction Valueijt + β 6 Rating Discrepancyijt + Issuer and Market Controlsijt + ε ijt The data vary by year (t), deal (i) and security (j). We control for security-design characteristics, issuer-fixed effects and time-fixed effects. The specification used is an OLS regression with primary issuance spread as the dependent variable, Credit Rating as the independent variable and the other variables shown in thheet emr oogdeenl eai tbyo v ei n a socuorn t er os tl i vmaar ti ai obnl e, s .wBee c auus se e tah eheert reor ro st ke er md ass thi ac ivtey -scyosnt es ims taetni ct covariance matrix as suggested by White (1980). Due to the possibility of issuer- as tnadn dtai mr de -efri xr oe rds eofff ec oc tesf,f i cwi ehni ct hs , wweo ut hl de nl er audn tohuera nOaLlSy s irse st ruel tast i nt og euancdhesraems tpi ml eaat es panel data. We achieve this by including the issuance-year effects and we doubleChapter 2 - How much do Investors Rely on Credit Ratings
34 cbluuisl dt errofbour satl ls ttar anndcahr de se sr or ol dr sb, ya st hr ee csoammme ei snsdueedr bayn dP ei nt e trhs ee ns a( 2m0e0y9e) a. r i n o r d e r t o Next, in order to investigate whether changing credit rating standards over time ha navdeLai un ai mn dpWa cat nogn (t2h0e1f9u)nadni ndgecsot ismt sa toef tahCe LfoOl laotwi si ns ug amnoc ed,ewl se: fo l l o w A l p ( 2 0 1 3 ) Where Rit denotes the credit rating of security i in issuance year t. αt is the intercept for year t, β is the vector of slope coefficients, and Zit is a latent variable that relates to Rit in the ranges between different partition points µi. Rit ranges from 1 to 21 as we have 21 rating categories in our sample. The matrix Xit denotes columns with explanatory variables. The variable definitions are given in Section 2.3.2. It nh eor redf oe rree dn ol ot gei ct omn oo md ei lcsa, l cl yo emf fei ca ineinntg fvuall, usei sn caer et hi en yuenai rt si nodf i ac al taot re nc to evfaf ri ci ai ebnl et a n d at is not in the same unit as Zit. Therefore, consistent with Alp (2013) and Liu and Wang (2019), we convert at into a rating notch, that is the average distance
35 between the partition points. The average rating notch length is calculated as (µ20 - µ1m o) /d1e9l .dDe ifvi ni de idn gi nt hE eq uy ea at iro innsd(i 2c a. 2t o) r- c( o2 e. 4f f)i, cbi ey nt thse, craaltci un lga nt eodt cuhs il ne gn gt ht he, owr ed ec rree da tleo ga int icnr de di ciat tsoprr feoardrsa, twi negusstea nt hdiasridnsd. iIcna toorrd ienr mt ootdees lt 1t h, ew ihme rpea c t o f r a t i n g s t a n d a r d s o n Rating Standards denote the year indicator coefficients divided by the rating notch length in year t. 2.3.2 Variable Construction and Summary Statistics 2.3.2.1 Dependent variable The dependent variable of our study is the specific funding cost of CLO tranches. Wwhei cmheeaqsuuaretetshtios tbhye tqhueopterdi mmaarrygiisnsfuoarntchee st prarne ac hde, s(i m p ly re fe r re d to a s s p re a d , Spread). For a given tranche, tt hh ee fruenf edri en ng cceo sr ta ft oe r rtehper iessseunetrs i st ht eh eproerf tei roenn co ef rt ah tee fpul nu ds itnhge cqousot t et hd amt ai sr gai nm, wa rhkeerte- wide benchmark and the quoted margin equates to the portion of the funding cost that is tranche-specific. This latter tranche-specific portion of the funding cboy s pt ui sr cthhaes iandgd ti thi ao tn apla rptei cr ual na rn ut rma nccohme , pwe nh si caht i omne faonrs t thhea tr itshke f qa cueodt ebdy mi navr eg si nt o irss, fCoLrOoturra npcuhrep. oI sne os ,u trhset uadpyp, rwo pe rui as tee omn leyafsluo raet i no gf -trhaet estpreacnicf ihce fsuins ds ui negd caot spt aor f t ht haet wE Ue rCeLbOe tnrcahnmc haer sk ei nd oouf fr tsht ue dEyuar no pd eUa Sn di no tl learrb La on nk doof nf e irne tde rrba at en k( EoUf fRe IrBe dO Rr a) t feo (rUtShDe LIBOR) for the US CLO tranches in our study.15 For securities issued at par, the Spread at issue – the dependent variable in model (1) – equals the quoted margin 15u nEsUe cRuI Br eOdRbraesf il se. cUt sS tDh eL Ii Bn tOeRr erset f rl ea ct et sa tt hwe hi ni cthe rhei sgth rl ya tcer ea dt iwt rhai tcehdhbi ga hn lkys ccraend bi torrartoewd , bi na ne ku sr ocsa, nf rbo omr root wh e, ri nb Ua nS kdsool lna ra sn, fbraosmis ofothr earvbaarinektsy oonf manatuunristeiecsu.red basis. EURIBOR and USD LIBOR are determined and communicated on a daily Chapter 2 - How much do Investors Rely on Credit Ratings
36 bo fe tt wh ee ei nni tt ihael ybieenl dc h, mm ea ar ks urraetde iang rbeaesdi s uppooi nn tas t (tbhpes d) . aItses uoaf npcrei csi pn rgeaa nd di st ha emc eoauspuorne otrfatnhcehreisskispsrueemd iautma pdreimceadnidffeedrebnytifnrvoemstoprasr.when issued at par. We do not use any 2.3.2.2 Independent variables The independent variable of the model, Credit Rating, is defined as the credit rating of Moody’s and/or S&P provided for each tranche at issuance. We measure Credit Rating via a numerical scale to convert credit ratings of Moody’s (and, in pr easrpe enct thi ev se el ys , 1S &f oPr ) At oa an u( AmAeAr )i c, a2l fsoc ro rAeas 1c o( Ar rAe+s p) , o3n df oi nr gAtao2 t h( Ae Ar )a, t i4n gf onr oAt ca h3e (sAwA i–t )h, aC n) da ns od oEnU. Tma ba lrek e2t. 1( Pr ea pn oe rl st s Bs uamn dm Da r) y. Fs it rast ti s, twi ces foobrs tehr ev eU Si nmPaarnkeel ts ( CP aannedl s DA tahnadt wt hiat nh o5u, 9r3E5Ut rma na rckheets doaut ra sUeSt , mw ahri ckhe tc do au tnatsse1t , h6 a5s6 ad ag trae aptoeirn nt su. mS ebceornodf, wd aet ao bpsoei rnvt es that in the US there are more dual-rated tranches than single-rated tranches, whereas in the EU the opposite is the case. Specifically, in the US market, for 42% (528, %5 0 8( 3t,r4a2n7c ht er as n) cohf et hs )e at r adnucahl ersa tai ns gi n. gSllei grhattliyn gmwoar es doifs ct hl oes es di nagtl ei s- rsautaendc et raanndc hf oe sr rt he ce eEi vUe dmaa rr ka et itn, g4 4b %y S (&7P3 2( 1t, r4a9n2c ht reasn) c ohfe st h) et htar na nbcyh eMs owo de yr e’ s d( u1 a, 0l -1r6a tterda nacnhde s5) 6. %I n (n9u2m4bterra nr ac theeds )b yo fMt ho eo dt rya’ sn (c6h4e 7s wt r ea rnec hs ei ns g) lteh- ar ant eb dy , So&f Pw(h2i 7c h7 wt r ea nocbhseesr)v. e a h i g h e r 2.3.2.3 Control variables WB ei nr eTpa bo lret 2t h. 1e . dWe es cirni pc ltui vdee ssteavt iesrtai cl sc oa nn tdr ovla vr iaarbi al eb lde iss tt roi bcua tpi ot unrse i sne Pc ua rniet yl s dAe sai ng nd characteristics of the underlying tranche: the number of tranches the CLO deal of
37 wo fhtihceh Ct hL eO tdr ae na lc ho ef wi shai cpha trht ;ectarpaint ac hl ae l li os caapt iaornt ; orfatthi negt rdai ns ccrheep; at nr acny,c ihf ea nv ay,l ubee;t wv ael eune Moody’s and S&P, per tranche; and, finally, the year of issuance of the tranche. Invariably a CLO deal is made up of a number of tranches. Tranche Count equals the total number of tranches in a corresponding CLO deal. In our total sample, the tranche count 16 per CLO deal ranges from 1 to 23 with acomnsetarnucotft7h.e6 for the US market sample and 6.7 for the EU market sample. We Capital Allocation measure per tranche as the percent of protection fi nr oomu rl ossasme sp lf eo ri se 2a c3h%t ri anntchheeUi nS mt haerckaept iat anlds2t r5u%c t ui nr et .hTe hEeUmme aa rnkceat p. Ti thailsailnl odci ac at itoens the percent of cushioning in the capital structure of a CLO deal, against credit losses that a specific tranche could suffer. The cushioning is provided by other tBrlaonocmh eb se ri gn tdhoee ss anmoet Cr eLaOd di l ye arl etphoa rt ta vr ea lsuuebs ofrodri ncaatpeidt atloatlhl oec tartai onnc haenidn tqhueersetfioorne. wcaelchualadtitoon cwalacsulacotendituscvteadluefofroreaecahchtrtarannchcheeomn aanudaelalyl.-bTyh-idseraal thbearsisla. borious Capital allocation aligns with credit ratings in those tranches with higher levels of shui gbhoerrd icnr ae tdei dt r ca at ipni gt a, ls oc ue svhe ino nt hi no gu gthh ewme laagbaei lntsht e ctrreadni ct hleoss sbeys c ur es du iat l lrya tri ne cge, itvhee i ar credit rating also reflects the subordination structure of cash flows in the entire deal of which the tranche is a part. Wfaceefvuarltuheerofcaontrtaronlchfoerattriasnscuhaencseiz(e, measured as the natural logarithm of the Log Tranche Size). The mean size of tranches ims seuaend t irna nt hc he eU sSi zme aor kf e$t2i5s 0$ 9m8i lml i oi lnl i. oTnh. iFso rd itfhf eer Ee nUc me ai sr kseutb, swt aenot bi asle, ravneda ohni lgyh ei nr pUaSr mt da ur ke e ttot ht ha ne rien bt he ienEg U, ams amr ek ne tt :i odneeadl s ai nb ot vhee, EoUn wa veerrea og ne amvoe rr ea gter a3n9c. h2 e%s li anr gt he er 16 We excluded one outlier with 29 tranches in one deal. Chapter 2 - How much do Investors Rely on Credit Ratings
www.ridderprint.nlRkJQdWJsaXNoZXIy MTk4NDMw