Queen, Maggie and Richard Roll, "Firm Mortality: Using Market Indicators to Predict Survival", Financial Analysts Journal, Vol. 43, No. 3, (May/June 1987), pp. 9-26.
Abstract: A firm may disappear from the public marketplace for several reasons. It may go bankrupt, be suspended from trading or delisted, in which case its shareholders are likely to experience unfavorable results. Or a firm may merge with or be acquired by another firm, in which case the results for shareholders may be quite favorable. An examination of five readily available market indicators -- firm size, price, return, volatility and beta -- reveals that all but beta can be of use in predicting favorable and unfavorable firm mortality.
Size seems to be the best predictor of both favorable and unfavorable mortality over both the long and short terms. The smallest firms have about even odds of disappearing, for favorable or unfavorable reasons, within a decade. The largest firms have a mortality rate of about 20% over the two decades.
The effect of market price varies depending on whether price is used as the sole predictor or in conjunction with the other variables. As the sole predictor, price shows a monotonic negative relation to unfavorable mortality, but both high and low-priced firms tend to have lower favorable mortality rates than mid-priced firms. When used with the other predictors, however, price has a strong positive relation to favorable mortality and no relation to unfavorable mortality.
Total return and total volatility of return both appear to have strong predictive powers. As return increases, the likelihood of unfavorable mortality declines and the likelihood of favorable mortality increases, while high total volatility increases the rates of both types of mortality. When one skips a year between calculating these variables and observing mortality, however, return and volatility retain their predictive abilities only for unfavorable mortality. The predictive powers of these variables appear to be at least partly ascribable to the effects of mortality events that are announced in the year during which the variables are calculated but not consummated until the following year.
Download paper (793K PDF) 18 pages
Related reading: Fitch Equity Implied Rating and Probability of Default Model