Assessing the Probability of Bankruptcy

Abstract

This paper assesses whether two popular accounting-based measures, Altman's (1968) Z-Score and an O-Score derived from Ohlson (1980), effectively summarize publicly-available information about the probability of bankruptcy (PB). According to option-pricing theories (Black and Scholes, 1973, Merton, 1974), a market-based measure, which we call BSM-PB, should reflect all available information about PB. These theories imply that not only will BSM-PB contain relatively more information than the Score variables, but that the accounting measures will not be incrementally informative to BSM-PB. We test the validity of these implications using a large sample consisting of 65,960 firm-year observations including 516 bankruptcies during the 1979-1997 period. Our statistical methodology utilizes a discrete hazard model that incorporates up to nineteen years of data per firm and produces unbiased coefficient estimates. Our main conclusion is that the traditional reliance on accounting-based measures of bankruptcy risk is inadequate. We find that BSM-PB has relatively more explanatory power than either of the two Scores, even when the Scores are decomposed to reflect industry differences or annual changes. However, the Scores contain significant, incremental information, and thus, BSM-PB is not a sufficient statistic for PB. Finally, we find that BSM-PB does not reflect all available market-based information regarding PB. Specifically, excess returns and relative market size provide incremental information.