258 at the 1% level. This indicates that CRAs are more likely to report ratings that anreeg at ht iev es aamn de fhoi rg ht hl ye sVi gEnSi fmi c ea tnht ocdo et hf fai cni ef on rt tf oh re FF LLTE , mwei tt hh oodd. dS si moifl a-r1l .y0, 5w (ez -f isnt adt =a -3.43). While for the OBS method, we find no significant results at all, column (5). Oo rui rg irneastuolrt ss ’ asri ze er vo abrui sa tb wl e hi ne nc owl ue me xncsl u( 1d )e toou(r4 s) e. Tc uhrei st ye rdeessui gl tns cs hu ag rgaecsttes rti hs tai tc sC Ra nAds have less rating disagreement for tranches with the VES and FLE method, relative to our base method FLT. So apparently, CRAs find it more difficult to evaluate the c( 2r e) doi ft Tr iasbkl eo fI ,t rAapnpcehneds iwx iItIh, wt hee rFeLpTe amt et ht heoadn. aI nl y soius rorfoTbaubslten e6s. 6s abnuat l ny soiws , ci nocl ul umdne the following additional control variables; STS Compliant, Single Originator, GDP Growth Rate and Country of Risk controls. We show that our results of Table 6.6 remain robust when including several additional controls to our model. Ibne tTwa ebel en 6t.r7a, nwc he easg at hi na tr eapr ee aot rtihgei naant eadl y sbi ys of rf eTqaubel ne t6o. 6r,i gbiunta twoer s(pi . el i.t, oour irg si na amt oprl es aa mn do nt hgotshee wt ohpo a1r0e%o rmi gei ansaut er eddb by yi nnf ur emqbueern to fo rtirgainncaht oe rs s) , (ctohleu mr enms a(i n1 i)nagn9d0(%2 )) ,, columns (3) and (4). We do so to analyze if CRAs are sensitive to the size and experience of the originator by assigning split ratings for specific risk retention methods. Interestingly, we show that CRAs indeed deviate between the size of originators ww hh ee nn wa sesliog onki nagt tshpel i ftr er qa tui en ng ts ofroirg i snpaet oc irfsi c, wr ei sfki nrde nt eonht iiog nh l ymsei gt hnoi fdi cs a. nRterme saur lkt as bf ol yr, Vc oEeSf fai cnide nOtB, wS ii tnhcoodl udms rnast i(o1s) oafn-d0 .(828) .(Wz - shtial et =f o- 2r .F2L4E) ,, awl be ef ii tn adt at hnee g5a%t i vl eevseilg, nc oi fliuc ma nnt (2). However, when we remove our set of independent variables, in column (1), our results turn insignificant. Henceforth, when a tranche is originated by a frequent (more experienced) originator, CRAs are less likely to have rating
RkJQdWJsaXNoZXIy MTk4NDMw