242 We also assess whether specific attributes of the originator influence the choice f“oerxpaepriaerntciceu”lainr sreisckurriettizeinntgio, nwhmicehthiosdp. rOonxieedatbtryibute is the originator’s size or ‘Frequent Originator’. This is a mi s aanmuoanl lgys ct at lhceu ltaotpe d1 0d%u ml amr gye vs ta roiraibg li en ai nt odrisc,amt i ne ag sounr ee di fb tyhne utmr abnecrhoe f’ st roarni gc ihneast oi nr the total number of securities issued in the EU (2011-2021), and zero otherwise. In addition, we include the variable ‘Single Originator’, a dummy that is 1 if tmh oe r de etahla ins oonr ieg ionrai gt ei nda bt oyr. aF suirnt ghleer mo roi gr ei n, awt oe rt, easnt da t0t r ii fb ut ht ee s doefa tl hi es do er iagl isn, antaemd ebl yy ‘No. of Tranches’, the total number of tranches in the securitization of which the security is part of, and Security Type, the different types of underlying assets in the securitization, ranging from ABS, RMBS, CMBS, to CLO. ‘Credit Rating’ is a proxy for the riskiness of the underlying assets and represents the average credit rating provided by Moody’s, S&P, Fitch, DBRS and KBRA. We have converted the r o a n t . ings into a numerical value by setting 1 for Aaa, 2 for Aa2, 3 for Aa2, and so 76 The specification of our first model is: Nd eesxitg, nt oc fhuarrtahcetre roibs tsiecrsv ea rter etnhdes si anmo eu rf odra tsae, cwu er i itni zvaetsi ot ing attrea wn chheet sh ewr itthhe dsief fceurrei nt yt risk retention methods. We do so by applying t-tests, using our full tranche-level 76 We use Moody’s credit rating scale as an example, but we have converted the different credit rating scales of all CRAs in our sample (Moody’s, S&P, Fitch, DBRS and KBRA).
RkJQdWJsaXNoZXIy MTk4NDMw