Abstract
Click Here:Access Full TextThe objective of this research study is to compare the performance of and test the reliability of four accounting-based models on a sample of 30 companies under the Insolvency and Bankruptcy Code in India and a matched sample of 30 solvent companies. The solvent companies are identified as those with highest and high safety, denoted with credit ratings ‘AAA’ and ‘AA’, respectively, and are matched on size (market capitalisation), year, and sector. The four models chosen are: Altman original Z-score, Altman Z-score for emerging markets, Ohlson model, and Zmijewski model. The scores are computed for T-1, T-2, and T-5 years, where T is the year the company filed under the IBC. The research findings reveal that all the models are able to predict default accurately for one and two years prior to bankruptcy, but fail to do so accurately for T-5 years. The predictive ability of the models is evaluated by ROC curves and the Gini coefficient. It is observed that the Ohlson model has the highest predictive ability, followed by the Zmijewski model and the Altman Z-score. In the next stage of our analysis, we compare the four accounting-based models with the Merton (market-based) model and the logit model to assess if the models, in their original form, are more accurate than these two models for the Indian markets. It is observed that the Merton model has the highest predictive accuracy, followed by Ohlson, Zmijewski and Altman Z-score. We conclude that neither the accounting nor the market models in themselves are sufficient; the models need to be reassessed based on specific regions and industries, and the scores need to be re-computed for better accuracy. We recommend that for accurate and early prediction, we need to look beyond financial statements to qualitative and other quantitative variables, so that defaults can be preempted and lenders be forewarned against impending distress.
Keywords: Accounting-Based, Bankruptcy, Predictive, Default, Accuracy
View PDF