42 TABLE 2.2: REGRESSION CREDIT FACTORS ON FUNDING COST FOR CLOS (1) (2) (3) (4) (5) (6) Credit Rating 27.86*** 27.64*** 26.87*** 27.46*** 27.89*** 26.30*** (73.26) (99.46) (98.42) (66.70) (67.86) (53.57) Tranche Count 0.94 -1.40* * 2.511*** (1.53) (-2.02) (5.70) Capital Allocation 30.32*** 44.94*** 19.25*** (5.12) (6.71) (3.23) Log Tranche Size -8.10 ** -13.17*** -22.25*** (-2.09) (-3.46) (-9.20) Log Transaction Value 0.22 4.53* ** -3.58*** (0.25) (4.50) (-3.16) Rating Discrepancy 9.19 *** 63.47*** -26.31*** (3.35) (23.90) (-10.88) Year dummy N N Y Y N Y Issuer dummy N Y Y N Y Y Observations 7,591 7,591 7,591 7,591 7,591 7,591 Adjusted R-squared 0.567 0.762 0.835 0.836 0.787 0.788 This table reports OLS regression of the underlying credit factors on the yield spread (at issuance) of CLOs for the full sample. The sample is based on securities that received a rating from Moody’s and/or S&P as reported on Bloomberg between 1997 and 2015. The dependent variable is the ‘Spread at issue’, measuring the quoted margin between the benchmark rate and the coupon of the initial yield, in basis points. ‘Credit Rating’ are a set of dummy variables to indicate the credit rating of a security at issuance by Moody’s and/or S&P, after we convert the ratings into a numerical value by setting 1 for Aaa (AAA), 2 for Aa1 (AA+), 3 for Aa2 (Aa), and so on. ‘Tranche Count’ stands for the total number of tranches in the CLO of which the security is a part. ‘Capital Allocation’ is the level of internal credit enhancement supporting such a security within a CLO, measured as the ratio of subordinated tranches and percent of protection from losses in the capital structure. ‘Log Tranche Size’ is the natural logarithm of the face value of the security at issuance. ‘Log Transaction Value’ is the natural logarithm of the transaction value of the security at issuance. ‘Rating Discrepancy’ that equals 1 if, at issuance, a security had two different credit ratings and 0 if only one rating or no rating differences. ‘Years Controls’ indicate specifications when year fixed effects were used for years Q1/1997 – Q1/2015. White (1980) heteroskedasticity-adjusted t-statistics in parentheses and (*), (**), (***) denote significance levels of 10%, 5% and 1%, respectively. Table 2.2: Regression credit factors on funding cost for CLOs.
RkJQdWJsaXNoZXIy MTk4NDMw