Saudi Green Banks and Stock Return Volatility: GLE Algorithm and Neural Network Models

Abstract

This study investigates the effects of ESG factors on stock return volatility from 2012 to 2020 using linear regression, GLE algorithm, and neural network models. This paper used the ESG factors and main control variables (ROA, EPS, and year) as independent variables. The regression model results showed that both year and E scores significantly positively affected Saudi banks’ stock return volatility. However, the S score and ROA significantly negatively impacted the volatility. The results indicated that the prediction models were more efficient in analysing the volatility and building an accurate prediction model using all independent variables. The results of the GLE algorithm model showed that the level of importance of the variables was sorted from highest to least significant as follows: S score, ROA, E score, and then G score. While the result of the neural network was sorted as ROA, ROE, and EPS, then the E score, S score, and G score factors all had the same minor importance in predicting the stock return volatility. Linear regression and prediction models indicated that the S score was the most crucial variable in predicting stock return volatility. Both policymakers and investors can benefit from our findings.

The Influence of Environmental, Social, Governance, and Corporates Scandals on the Stock Volatility

“…Laokulrach (2022) also found that companies that implement ESG in 7 out of 8 industries have higher stock volatility than the industry average. Several other studies also attempted to examine further the influence of ESG on stock volatility based on each of its components, namely E, S, G and found that there was a positive, negative, or insignificant influences of each component on stock volatility (Assous, 2022;Meher et al, 2020;Yoo et al, 2021).…”