Modeling Saudi Stock Index Returns and volatility: A Dual Approach using GARCH and Neural Networks

SUKAINAH ALBESHER^*Dania Al-Najjar^*

• King Faisal University, Al-Ahsa, Saudi Arabia

Financial markets are one of the supports of economic growth that helps in the mobilization of local savings, attraction of international capital, and efficient allocation of resources. Risk reduction and optimal investment decisions depend on market returns and volatility dynamics as the main factor which requires more and more sophisticated quantitative models. The economy of Saudi Arabia, which is a member of the GCC, unified economy, is characterized by large and fast-growing economies that are dominated by oil production, which is why the energy aspects play an essential role in the economic and financial performance. Proper prediction of the stock market is therefore vital to both the policy makers and investors. The Saudi Stock Exchange (Tadawul) plays one of the independent roles in the region, as it is supported by a strong economy, a highly developed banking system, and structural changes as part of Vision 2030. However, the market is very unpredictable because of internal and external factors such as fluctuation of oil prices and the economic state of the world, and predicting its rebound is complicated. Although the Saudi market volatility has been studied using econometric models like the GARCH models, the studies conducted have already employed univariate models, having ignored the key conventional volatility drivers, including oil and gold prices. In addition, literature gap in the comparison of classical econometric methods with the contemporary machine learning methods exists, which is capable of capturing nonlinear correlations and intricate time variations. This research fills these gaps through simulations and prediction of both the returns and volatility of the TASI with traditional GARCH models and an LSTM neural network. The timeframe will be January 1, 2000, to December 31, 2022, and the explanatory variables will be the oil and gold prices. Comparing the models is possible to compare their forecasting capabilities and determine how oil price changes affect the Saudi stock market economically. The results offer theoretical and practical information, which will enhance better policymaking and investment plans.