Predicting Corporate Failure in Malaysia During Financial Crisis Using Logistic Model Approach
Thai, Siew Bee (2003) Predicting Corporate Failure in Malaysia During Financial Crisis Using Logistic Model Approach. Masters thesis, Universiti Putra Malaysia.
Since the financial crisis of 1997, more and more companies are facing financial difficulties. As of 31 March 2003, a total of 376 companies had successfully seek protection under Section 176 (10) of the Companies' Act 1965 with restraining orders. Prior to that, the distressed firms were gazetted under Practice Note No.4 (PN4) of the KLSE listing requirement. As firms' financial health signs were transmitted prior to actual financial reporting, we addressed the following issue: Can financial distress be predicted? According to Zulkarnain (2000), corporate failure was not a sudden even and it developed gradually over many years. Some of the symptoms that lead to firms failure are declining in profits, working capital, liquidity and asset quality. In light of this, this study attempted to develop a model that can discriminate between failed firms and non- failed firms using logistic regression analysis. This study also investigated whether cash flow ratios apart from the usual accrual ratios were more significant in predicting corporate failures compared to previous studies which were based on accrual ratios as major predictors. A total of 42 companies from 8 different sectors that were gazetted under Practice Note No.4/2001 (PN4) were included in the sample. Failed firms were matched to non - failed firms on a one to one basis for the period of analysis stretching from 1999 to 2001. Using Pearson Correlation test, highly correlated variables were removed. In addition, the enter method in logistic regression was employed to detect any outliers. A total of 14 variables for 1 year before failure and 13 variables for 2 years and 3 years respectively before failure were entered into the analysis to find the most parsimonious model in predicting corporate failures. The result showed that the overall of correct classification for all of the 3 models were well above 90% accurate rate with 1 year prior to failure showed a predictive accuracy of 97.62%, 2 years prior to failure showed a 91.67% accuracy rate and 3 years prior to failure showed a 90.48% accuracy rate. From the result of this study, we concluded that the prediction model with the combination of cash flow ratios as the major ratios along with accrual ratios outperformed the prediction model based solely on accrual ratios as documented in previous studies. Furthermore, this study also suggested that cash flow ratios were particularly useful in predicting bankruptcy and financial distress.
Repository Staff Only: Edit item detail