Bankruptcy Classification Errors in the 1980s : An Empirical Analysis of Altman and Ohlson's Models

Bankruptcy prediction models developed by Altman (1968) and Ohlson (1980) are often used in empirical accounting research to indicate firms' relative likelihoods of bankruptcy. However, use of these models in their original forms on current datasets may introduce measurement error into the analysis. In this paper, we examine one aspect of this measurement error: the extent to which the original models misclassify firms when used on data from the 1980s. We hypothesize that the increased acceptability of the use of debt in the 1980s should result in increase Type II classification errors when the original models are used. We also hypothesize that the changes in bankruptcy laws that allowed for greater strategic use of bankruptcy in the 1980s should result in increased Type I classification errors when the original models are used. Our empirical results are consistent with these hypotheses. We also re-estimate both models using data from the 1980s. In the re-estimated models, the coefficients on the leverage variables decline, consistent with an increased acceptance of the use of debt. However, we find that re-estimation does not necessarily mitigate classification errors