Artificial Intelligence for Risk Mitigation in the Financial Industry

9Determinants of Financial Distress in Select Indian Asset Reconstruction Companies Using Artificial Neural Networks

Shashank Sharma* and Ajay Kumar Kansal

School of Management, Gautam Buddha University, Greater Noida, India

Abstract

An analysis has been performed to examine the financial health of select asset reconstruction companies (ARCs) registered with the Reserve Bank of India (RBI). For the analysis, 10 key firmlevel financial variables have been assessed for the period 2011–2012 to 2019–2020 and analysed through multilayer perceptron artificial neural networks (MLP-ANNs) and popular Altman’s Zscore models. In addition, an impact analysis of Insolvency and Bankruptcy Code (IBC) on the capital structure of ARCs has been presented as an additional section. Rationale behind the analysis has been the concern that any contingent financial distress situation in ARCs may trigger a triplebalancesheet problem, which may affect non-financial corporates, the banking sector, and ARCs altogether. Findings indicate a gradual rise of financial distress in select ARCs over the last decade and among the selected variables—return on equity, debt to total assets, and return on capital employed surfaced as the most important indicators. Similarly, the MLP-ANN model has returned an average root mean square error of 0.15 with an overall average classification accuracy of 95.4% during the training and testing phases of the model. Furthermore, introduction of IBC may have a significant positive ...