Premium
The prediction of default for high yield bond issues
Author(s) -
Huffman Stephen P.,
Ward David J.
Publication year - 1996
Publication title -
review of financial economics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.347
H-Index - 41
eISSN - 1873-5924
pISSN - 1058-3300
DOI - 10.1016/s1058-3300(96)90007-5
Subject(s) - bond , default , econometrics , bankruptcy , logistic regression , economics , multivariate statistics , yield (engineering) , bankruptcy prediction , financial ratio , actuarial science , default risk , asset (computer security) , financial economics , credit risk , statistics , finance , mathematics , computer science , materials science , computer security , metallurgy
Bondholders and financial analysts have long sought models which will predict financial distress in corporations. Prior research has produced a number of useful models to predict bankruptcy in the short term. This paper looks at four models which predict default based upon public information at the time of issuance of high yield bonds. Multivariate results using logistic regression analysis indicate that high yield issues that default are characterized by having higher asset growth rates, lower operating profit margins, larger levels of collateralizable assets, and larger changes in net working capital. Models using Altman (1968) variables have lower likelihood ratio indexes than models employing alternative explanatory variables. Predictive ability tests of models excluding the traditional variables on a holdout sample are able to correctly predict 73.3 percent of the defaulted bonds and 68.6 percent of the nondefaulted bonds.