We study the problem of simultaneous and coherent assessment the probability of a firm's bankruptcy at various time horizons in future. In contrast with usual (one-period) formulations of the problem, such multi-period formulation better matches the nature of bankruptcy process (bankruptcy occurs in time) and allows an easier and more natural incorporation of bankruptcy (default) prognoses in valuation of risky debt and equity, optimization of corporate capital structure etc. The study uses a new mathematical apparatus - multi-alternative decision rules of statistical decision theory. We investigate a new type of predictive variables that can be extracted from the maturity schedule of a firm's long-term debt. The study develops Bayesian-type forecasting rules that use both maturity schedule factors and traditional financial ratios. These rules noticeably enhance bankruptcy prediction (compared with the familiar one-period Z-score rules of Altman) for bankruptcy within the first 1, 2 or 3 years. Predictive factors derived from schedule information enhance bankruptcy prediction at distant time horizons.
- Bayesian decision rules
- Forecast efficiency
- Multi-period bankruptcy prediction
- Schedule of paying off long-term debt
- Time to bankruptcy