Reporting Delays for Failed Firms
发现大量破产企业在最后一年延迟发布年报，甚至有些在报告发布前就申请破产，这会导致基于及时报告假设的破产模型高估预测准确性。
Corporate failure has been a subject of considerable academic interest since the mid 1960s. Major studies by Beaver [1966], Altman [1968], Edmister [1972], and Deakin [1972] have demonstrated that certain financial ratios of firms could be used to construct bankruptcy models for predicting corporate failures.1 Unfortunately, several of the models developed during the 1970s and more recently have assumed that the underlying corporate reports were available on a timely basis. For example, Altman and Brenner [1981] explicitly assumed the annual reports were available within three months of the companies' year-end. If this assumption does not hold, then their model and other research models constructed with data from Moody's Industrial Manuals and the Annual Compustat Tapes have a serious bias in their performance results. This note is an expansion of the work done by Ohlson [1980] on the financial reporting delays for firms filing for bankruptcy. His investigation was a milestone in the sense that it explicitly recognized that reporting lags could result in back-casting. This would occur in situations wherein information not disclosed until months after a firm petitioned for bankruptcy is included in the data set and used to predict the event. The result is to overstate the classification accuracy of the model. The results of this research are that a significant number of bankrupt firms incur delays in releasing their annual reports for the final year before bankruptcy, and that some even file for bankruptcy before the release of the report. More specifically, approximately 47 percent of the