Factors Affecting Banking Financial Stability in Vietnam's Emerging Economy: A Lasso, Ridge, and Elasticnet Regression Analysis

Abstract

Background: The COVID-19 pandemic has significantly impacted the stability of the global economy and the banking sector in developing nations such as Vietnam.

Objective: This study examines the variables impacting the Z-score index and uses them to evaluate the stability of commercial banks in Vietnam.

Methodology: The research uses ElasticNet, Ridge, and Lasso regression methods to tackle the problems caused by a large number of independent variables and reduce the likelihood of issues like autocorrelation and multicollinearity. These approaches help adjust the model coefficients within the appropriate thresholds or select important variables to minimise the risk of multicollinearity or autocorrelation as well as in other emerging markets, for the period from 2010 to 2023 based on the regression methods Ridge, Lasso, and ElasticNet. Model performance is evaluated using MSE and R² scores.

Result: The results showed that Ridge performs best with an R² score of 88.75% and an MSE of 0.002607; ElasticNet performs better than Lasso, though the difference is negligible.

Conclusion: Empirical data demonstrate that several parameters, such as the previous year's Z-score, the ratio of loans to total mobilised capital, total asset growth rate, ROE, competitiveness, market power, or inflation rate, significantly influence the Z-score. In contrast, the Z-score is less affected by the risk provision ratio, income diversification, or bank size.

Unique Contribution: The study uses the Z-score index to measure banking stability and provides empirical information on its influences. This contribution is especially relevant for emerging economies, where the drivers of financial stability are underexplored.

Key Recommendation: The authors suggest several implications that could improve both experimental and operational efficiency for bank managers and maintain financial stability.