Forecasting bankruptcy more accurately: A simple hazard model

I argue that hazard models are more appropriate than single-period models f or forecasting bankruptcy. Single-period models are inconsistent, while haz ard models produce consistent estimates. I describe a simple technique for estimating a discrete-time hazard model. I find that about half of the acco unting ratios that have been used in previous models are not statistically significant. Moreover, market size, past stock returns, and idiosyncratic r eturns variability are all strongly related to bankruptcy. I propose a mode l that uses both accounting ratios and market-driven variables to produce o ut-of-sample forecasts that are more accurate than those of alternative mod els.