This study examines the failure firm¡¦s stock risk character and return behavior prior announcements day. This study includes three parts. First, we examine the risk different between failure firms and control firms. Second, we analysis the returns character of failure firms, and third, we test whether the different judgment announcements will affect the failure firms¡¦ stock return. We examine the daily stock returns between 1997 and 1999, 26 failure firms and 50 control firms have been observed. The empirical results can be summarized as follows: 1. Systematic risk behaves quite differently for the failure and control groups as much as two years before failure. 2. Both the total variance and the firm-specific variance behave quite differently for the failure and control groups only two months before failure. 3. The market is effective before announcements and not effective after announcements. 4. The shareholders of failure firms lost between different judgment announcements are no different in a short time. 5. The magnitude of the price reaction to failure filings depend on the probability of bankruptcy and proxy for predisclosure information impounded in stock prices over the year prior to failure filing.