The prediction of bankruptcy: an investigation into the time and industry sensitivity of predictive models
Date
2021
Authors
Govender, Tamara
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
The overall objective of this research is to evaluate the predictive ability of the Altman (1968), Ohlson (1980) and Zmijewski (1984) models when evaluating firms for bankruptcy. An industry-adjusted and re-estimated version of each of the three models will be assessed to determine if the re-estimation of the models, using a data from South African firms from 1990 to 2020, results in an increased predictive ability as compared to the original models. The results suggest two main challenges faced when using predictive models in industry: the issue of time sensitivity, as well as industry sensitivity. The issue of time sensitivity was resolved post re-estimating each of the original models using the new sample. The re-estimation of each of the three models resulted in significant improvements in predictive ability across industries. The challenge of industry sensitivity was addressed byre-estimating the models using industry-specific samples of data. The findings showcase near perfect predictive accuracy up to five years prior to bankruptcy. The intended contribution of this research is the practical application of the methods and findings which would serve as a guide for risk assessment by lending institutions, and performance benchmarking for firms