82 There are an insufficient number of CLOs rated by Fitch, the third of the three globally dominant CRAs, to enable statistical analyses on Fitch ratings. We apply several filters to our dataset and remove tranches with incomplete information. Bo fe cc ar euds ei t wr aet ianrges ,i nwt ee r eo sntleydi ni nc l tuhdee ei fnf eoc ut ro fs tCuLdOy s CdLeOa sl ct or amn pc hl eexsi t wy iot hn taht el enaus mt obne er c9r, 1e 1d 2i t. rWa tei nfgu rdtihs ec lro ds ei sdc aa rt di s as ul l et. rTa hn icshreesdwu ci tehs mo ui rs soi nr igg iins as ul seadmapt al e (f1r5o 4m t1r a0n, 4c0h0e st)o, ttrraannscahcetsi.on or tranche size (27 tranches), resulting in a full sample of 8,931 CLO 3.3.2 Empirical Model We conduct three sets of tests. First, we use a univariate dichotomous (logit) model to study how CLO deal complexity influences the number of credit ratings dt hi secr leo sies d aa t r ieslsaut ieo. nS sehc iopn db, ewt we eeemn p cl ooymoprlde ex ri teyd al ongdi t trhe eg r edsesgi roenes toof t de si st cwr ehpeat nh ce yr oo fr driant ai nr yg sl eaassst i sg qnue adr et os (tOh eL Ss)atme set ss teoc ui nr vi tei se tsi gbayt eSt&hPe iamn pd a Mc t ooof dt hy ’es .c rTehdi irtdr, awt i ne guos ne tvhaeluyeiealsdmfoerasseucruerditbieys that received one or two credit ratings and the explanatory R2, consistent with He et al. (2016). Be xaps leadn aotno royu fra cl ittoe rr sa toufr et hree vsieecwu riint y Sde ec st ii og nn 3t h. 2a.t3 ,mwa ey iddeetne trimf yi nt eh rtehee kCeLyOu’ sn idqeuael c ( omplexity: the natural logarithm of the face value of the security at issuance Log Tranche Size), the capital allocation (Capital Allocation) measured as the po fe rt rcaennct hoef spirnotthe cet ci oonr rfersopmo nl odsi ns egsCiLnOt hoef wc ahpi ictha lt hs ter us ec ct uurrei t, ya insdi nt hc leu dt oetda l( n u m b e r Tranche Count).
RkJQdWJsaXNoZXIy MTk4NDMw