Prediction and Indirect Costs of Bankruptcy: A Multi-Stage Approach to Financial Distress

Abstract

Prediction and Indirect Costs of Bankruptcy: A Multi-Stage Approach to Financial Distress This dissertation investigates corporate failure from three perspectives using a sample of non-financial firms listed at Pakistan Stock Exchange. First, the dynamic behaviour of different stages of financial distress were explored. Second, a prediction model of multistage financial distress was developed. Third, impact size and determinants of indirect cost of financial distress were studied. To investigate these three perspectives, the data of 321 on going and 91 delisted non-financial firms is extracted from annual publications of State Bank of Pakistan during 2002-2015. This research starts with the proposed dynamic framework of multistage financial distress showing multiple adverse heterogeneous events leading a healthy firm closer to bankruptcy progressively. It is found that initially healthy firms face profitability problems or mild liquidity issues. While continuity of both the problems leads to severe liquidity that results in bankruptcy. This research also developed a machine learning model to forecast such profitability problem, mild liquidity and severe liquidity. In doing so, criticisms on prior prediction models particularly regarding sampling, feature selection and model selection are explored using Systematic Literature Review. This research contributes by applying recommended solutions of such criticisms to obtain more accurate multistage financial distress prediction model. Results showed that the proposed model predicted multistage financial distress with 84.06% accuracy. This accuracy increased to 89.57% when relevant cut-off values were applied. Furthermore, the indirect cost of financial distress was studied using the sample of on going firms that were healthy in the previous year and documented positive gross profit. To measure indirect cost, appropriate measure of distress, receivable and inventory management are used based on the recommendations of Systematic Literature Review. Results showed that healthy firms do not lose their market share when faced with profitability problems or mild liquidity issues. Conversely, leverage showed a nonlinear relation with indirect cost. It was also found that healthy firms that remain healthy should increase their receivables and inventory to capture more market share. However, results revealed that firms facing profitability problems should follow industry averages for receivables management and hold more inventory xi to recover from profitability problems. Conversely, healthy firms that face mild liquidity issues can decrease their receivables and inventory to enhance their liquidity position without affecting market share. This dissertation provides useful practical implications for managers to respond during financial distress. The results will help stakeholders to recognize intensity of financial problems earlier in order to respond accordingly. Moreover, this dissertation provides useful insight that how managers can minimize the adverse effects of indirect cost of financial distress in terms of loss of opportunistic market share. Keywords: Bankruptcy Prediction; Financial Distress; Liquidity; Machine Learning; Indirect Cost; Opportunity Loss; Pakistan JEL Classification: C53; G17; G32; G33