RDP 2016-09: Why Do Companies Fail? Appendix A: Regression Results

Rose Kenney, Gianni La Cava and David Rodgers

November 2016

Table A1: Complementary Log-log Model Estimates of Failure
Benchmark model
  Estimated coefficient Exponentiated coefficient
Structural characteristics
Public 1.13***
(0.00)		 3.09***
(0.00)
Public * log of age −0.24*
(0.08)		 0.79*
(0.08)
Public listed −1.25***
(0.00)		 0.29***
(0.00)
Public listed * log of age 0.57***
(0.00)		 1.76***
(0.00)
Cyclical characteristics
Log of age −0.34***
(0.00)		 0.71***
(0.00)
Log of assets −0.09***
(0.00)		 0.91***
(0.00)
Debt-to-assets ratio 0.76***
(0.00)		 2.13***
(0.00)
Trade credit-to-assets ratio 0.73***
(0.00)		 2.08***
(0.00)
Cash-to-assets ratio −1.88***
(0.00)		 0.15***
(0.00)
Return on assets −0.07***
(0.00)		 −0.93***
(0.00)
Time fixed effects Yes Yes
Industry fixed effects Yes Yes
Company random effects No No
Wald test (p-value) 0.00 0.00
Area under ROC curve 0.73 0.73
Observations 90,729 90,729
Companies 23,326 23,326
Failure observations 532 532

Notes: Standard errors are clustered by company; p-values are reported in parentheses; ***, ** and * denote statistical significance at the 1, 5 and 10 per cent level, respectively