97 icsa ct eorni nt rga r(yH t1oAt)h. eHgoewn ee vr ae lr, i di ne at hoef rpaet irni ogds ha of tpepr i nt hge( Hc r1i s) i bs ,u wt ceo on bs isset revnet wa idt hr arma tai tni gc change and find evidence that supports rating shopping. In this period, we see thhi ga ht eCrLcOa pt ri taanl cahl leosc at ht iaotnalreev el al sr ga er er ma nodr et hl iokseel ywt oh ehraev et haes iAnAg Al e rcar teeddi t trraatni nc hg er sa thhaedr than dual credit ratings. A possible explanation for these findings is that before ti shseu ec rr si s ti os pt huet pr ar et isnsgu reenov ni r oCnRmA se nt ot fmo ra t mc hoer ea c choomt hpel erx’ s dr ae tailnsg m, aandde t hi ta te aasf ti ee rr tfhoer crisis, even though issuers still sold complex CLOs, they had fewer opportunities to influence the credit ratings quality because of stricter quality controls within the CRAs themselves, i.e., a reduced likelihood of rating catering. 3.4.2 Number of ratings disclosed at issuance, variations between Moody’s and S&P Wt o ea snsoews s sthhief t eoxut er natt tteonwt i ho inc ht ot heeayc hmoafy t hh ae vCeRcAa st e(rMe do ot hdeyi’ rs raantdi nSg&t oP )ms ae tpcahr at ht eeliyr, competitor for complex CLOs, before and after the crisis. In Panel A of Table 3.3, we take a finer approach in our logit model by including Moody’s Exclusively and S&P Exclusively, as opposed to just Single Credit Rating. We first look at the effects oi sfs ud ae na lc ec oi nmspt el eaxdi toyf oa ns itnhgel ei scsrueedri ’ts rpa trienf ge rbeyn coen ltyo Md iosocdl oys’ se (t w o c r e d i t r a t i n g s a t Moody’s Exclusively). To do so, in Panel A we exclude all single ratings from S&P (S&P Exclusively). In Pc oa mn epl l Be x, wi t ye or ne pt re aa nt ct hh ee sa tnhaal yt sa irse dr iastceuds bs ey db ao bt ho va eg ,ebnuc ti ensocwo mwpe aarne adl yt oz et rtahnec ehfef se ct ht oa tf received only a single rating by S&P (i.e., Moody’s Exclusively are excluded). The mbyotdheel of Panel B explains a substantial higher proportion of variation, denoted R2 (48%), compared to the model in Panel A (24%). Chapter 3 - Security Design and Credit Rating Risk
RkJQdWJsaXNoZXIy MTk4NDMw