What Factors Drive Systemic Risks In Asian Emerging Markets

Abstract

Determinants of systemic risk have been extensively investigated in developed markets and using the Ordinary Least Squares (OLS) regression models. However, several studies in the top-tier journals in finance and banking about the prediction of stock returns and default risks show that the prediction of the OLS regression models is not as good as that of machine learning models, namely the Ridge, the LASSO, and the Elastic Net. In this study, I employ machine learning techniques to predict systemic risk in Emerging Asian Markets to fill out the gaps of the shortage of research of systemic risk in emerging markets and the lack of the application of machine learning methods in predicting systemic risk. I find that the Elastic Net is the best among the four regression models: Stepwise Regression, Ridge Regression, LASSO Regression, and Elastic Net Regression and that non-performance loans are the best predicted factor and is negatively related to systemic risk as well as its two components. Thus, too low high-quality non-performance loans give the signals or early warning about systemic risk. Thereby, increasing high-quality non-performance loans, in that case, may help prevent the damage that systemic risk can bring.