Vivian van Breemen

91 c( Ao rAr+e)s, p3o nf odri nAga t2o (tAhAe )r, a4t i fnogr nAoat3c h( eAsAw- )i, t ahnrde sspoeoc nt i .vAe lsy s1h of owr nA ai na P( Aa nAeAl )A, 2a fnodr AC ao1f Table 3.1, the mean credit rating28 of the whole sample is between A1(A+) and A2 (r Aa t)e, df osrutbh-es es ti nt gh lee -mr aetaend csruebd-i st erta it ti ni sg bi se tAw1e(eAn+A) , 2m( eAa) nainndg At h3a (t Ad-u) aal n- rda tf eo dr tt hr ae ndcuhaels- are, on average, higher rated. We qeu ca ol sn ot rnoel fi fo rt ht eh ei sms uaerrk ei st as hmaor ne go tf ht he et oi sps u1 e0r%. Woef ii ns scul uedr se ma de au smu mr eyd vuasriinagb lgel ot hb aa tl Co fL TOasbml ea3r k. 1e ts hs oh wa rse ,t ha antd 2z, 9e r3o8 itf rtahnec hi sessu ae rr ei si sas mu eodn gb yt ht oe pr e1m0 %a i ni isns gu e9r0s %a n. Pd a5n, 9e l9 B3 tt rhaenscahme sp laer ea ri es si us seudebdybtyh teo rpe 1m0a%i n ii snsgu9e 0r s%. I ni s ms uoedr se .l A( 3b)o, uwt e3a3l %s o oc fo tnht er ot lr faonrc ihs es su ei nr fixed effects (Petersen, 2009). We further control for market factors by including Euro Market, a dummy vi na rtihaeb lUeSs me t atrok eotn. eI ni f t thhee f us el lc suarmi t yp lies, i3s 8s u%e do fi nt hteh es eEcUu rmi t iaersk (e3t , a3n6d3 zt rearno cihf ei ss )s uaerde ibsysuaedddiinngthteheEUcoanntdro6l2v%ariianbtlhee US (5,568 tranches). Finally, we control for time Year of Issuance, which equals a dummy of one tohthatercworirsees. ponds to the year of issuance (ranging from 1996 to 2013) and zero 3.4 Complexity of a CLO’s Design and the Choice for Multiple Ratings It nr atnhci hs es se ci tsi or ne l, awt eede xt oa mt hi ne enwu mh ebt ehre ro cf ocmr epdliet xri at yt i on fg tshde i ss ec lcousrei tdy adte issi sguneo. fCtohme pC lLeOx CLO tranches with one rating are more likely to have been shopped than those 28 Table 3.1 reports the numerical mean credit rating scores corresponding to the rating notches. Chapter 3 - Security Design and Credit Rating Risk

RkJQdWJsaXNoZXIy MTk4NDMw