| Summary: | 碩士 === 國立東華大學 === 應用數學系 === 99 === In this paper, the frailty correlated default model is suggested to predict the probabilities of financial distresses for Taiwan companies. It is built by applying the frailty factor to the intensity default model. The frailty factor is used to explain the latent macroeconomic dynamic factor. To build the frailty correlated default model for Taiwan corporations, we collect time-varying predictor values for each set of Altman’s predictors (Altman 1968), Campbell’s predictors (Campbell et al. 2008), Duffie’s predictors (Duffie et al. 2007), and Shumway’s predictors (Shumway 2001). The empirical results indicate that, for each panel dataset, the frailty correlated default model is better than the intensity default model, in the sense of yielding more information about financial distress, more accurate out-of-sample predicted number of distresses and predictive intervals. Also, the frailty correlated default model using Campbell’s predictors has the best out-of-sample performance.
|
|---|
