Time Series Characteristics of Four Primary Accounting Variables for Predicting Business Failure of Firms

摘要

This paper investigates time-series characteristics of four primary accounting variables that are commonly used for predicting business failure of firms in Korea. Although four accounting ratios including Return on Assets, Liability Ratio, Working Capital Ratio, and Current Ratio have frequently been used in the prediction models of business failure, little studies have been done for understanding, the time-series characteristics of the ratios. By classifying firms into non-failed firms and failed firms, we investigate time-series characteristics of four variables in terms of two groups. In particular, we measure four primary accounting ratios with net income to total assets, total liability to total assets, working capital to total assets, and current assets to current liability respectively. Using 117 matched samples from 1980 to 2006, we find that whereas all financial ratios of failed firms show random walk patterns of times series, the ratios of non-failed firms show in part stationary pattern. We also find that the accounting ratios showing random walk become stationary through the linear combination with other accounting ratios showing random walk. We finally find that the ratios of non-failed firms and the stationeries ratios of failed firms converge into the level of the corresponding industry, but the ratios of non-failed firms take longer time to adjust to the level of an industry.