Abstract. The aim of this study is trying to identify a group of variables that can be
used to predict firm bankruptcy. The examined variables consist of Net Income to Total
Assets (NIMTA), Total Liabilities to Market Value of Total Assets (TLMTA), Cash to
Market Value of Total Assets (CASHMTA), Relative Size (RSIZE), Excess Return
(EXRET), Volatility of Return (SIGMA), Stock Price (PRICE) and Market to Book
Equity (MB). By using Hazard Model as a modelling basis, the result of this study
found there were six variables that could be used as a predictor of firm bankruptcy,
including TLMTA, TLMTAsq2, TLMTAsq3, EXRET, SIGMA dan PRICE. The
evaluation of the model showed that it has a good accuracy. In accordance with model
accuracy approaches, the level of accuracy of the model showed a range between
89.36-96.51 percent; Area Under Curves (AUC) of ROC Curves reached 0.8476; and
the Brier Score showed a very low number which was 0.0309.