PREDICTION OF BUSINESS FAILURE USING ACCOUNTING DATA

The original point of departure for this effort was William Beaver's now classic study.2 He presented empirical evidence that certain financial ratios, most notably cash flow/total debt, gave statistically significant signals well before actual business failure. The test design used here is, with minor differences, a replication of his original design, which employed pairs of matched failed and nonfailed firms of similar size and industry characteristics. In earlier writings, I presented a simple theoretical framework utilizing a stochastic rather than deterministic model which plausibly explained Beaver's result.3 The present study tests the predictive usefulness of a statistic derived from a binomial process with an absorbing state. The evidence indicates a substantial improvement over Beaver's ratios through at least four years before bankruptcy.4