Credit risk modelling for private firms under distressed economic and financial conditions: evidence from Zimbabwe.

Abstract

Since the outburst of the recent 2007 - 2008 global financial and economic crisis, modelling of credit risk for private non-financial firms under economic and financial stress has been receiving a lot of regulatory and scientific attention the world over. Nevertheless, the quandary is that there seems to be no well-defined estimation procedures and industry consensus on how to incorporate economic downturn conditions in private firm credit risk models, which have led to the introduction of diverse default probability, exposure at default and rate of recovery prediction methodologies. Moreover, there is no consensus on which predictor variables have the most significant impact on private firm credit risk under downturn conditions. This study strives to design forecasting models in order to estimate key credit risk components (default probability, recovery rate and exposure at default) for private nonfinancial firms under downturn conditions in a developing economy. The main aim of the thesis is to identify and interpret the drivers of probability of default, recovery rate and credit conversion factor. In the first part, the study reviews literature using a scoping review framework in order to identify the reasons and motives for research, emerging trends and research gaps in modelling bankruptcy risk for private nonfinancial corporations in developing economies. The second part of the thesis creates stepwise logit models to detect the default probability for privately-owned non-financial corporates under downturn conditions in a developing country. In the third section of the study, stepwise logit models are designed to separately forecast probability of default for audited and unaudited privately-traded non-financial corporations under downturn conditions in a developing economy. The fourth part of the thesis develops stepwise Ordinary Least Squares regression models to predict workout recovery rates for defaulted bank loans for private non-financial corporates under downturn conditions in a developing market. In the fifth section of the study, stepwise Ordinary Least Squares regression models are developed to estimate the credit conversion factor to precisely predict, at the account level, the exposure at default for defaulted private nonfinancial corporations having credit lines under downturn conditions in a developing economy. To fit the models, the study adopts unique real-world data sets pooled from an anonymised major Zimbabwean commercial bank. This study finds that the forecasting of probability of bankruptcy for private non-financial corporates in developing economies is an appropriate discipline that has not been properly studied and has some distinctive and unexplored zones due to its complexity and the diverse business ethos of private firms. The thesis discovers that accounting information is imperative in predicting the default probability, rate of recovery and exposure at default for Zimbabwean private non-financial corporations under downturn conditions. Further, the study reveals evidence indicating that the forecasting results of the designed credit risk models are improved by incorporating macroeconomic variables. The incorporation of macroeconomic factors is vital since it enables stress testing and provides a way of modelling the default probability, recovery rate and exposure at default under downturn conditions. In light of these findings, it is recommended that firm and/or loan features, accounting information and macroeconomic factors should be adopted when predicting credit risk parameters for private non-financial corporates under downturn conditions in a developing country.