Modeling and Forecasting Stock Market Volatility of CPEC Founding Countries: Using Nonlinear Time Series and Machine Learning Models

Abstract

The highly sensitive, nonlinear, and unpredictable stock market behaviours are always challenging for researchers. Stock markets of Pakistan and China, i.e., KSE-100 and SSE-100, respectively, are the two most attractive stock markets after the official announcement of CPEC. Thus, the daily closing price of KSE-100 and SSE-100 Stock returns are used to evaluate the volatility forecast performance of the machine learning technique, GARCH family and the nonlinear regime-switching models. The findings of this study revealed that the standard GARCH model is the best-fitted model based on Akaike’s Information Criteria (AIC) and Bayesian Information Criteria (BIC). Furthermore, the forecast performance of the machine learning LSTM model outperforms other models based on RMSE for SSE-100. In contrast, the forecast performance of CGARCH for SSE-100 and the Markov-regime-switching model for KSE-100 outperforms other models based on MAE, MAPE, and SMAPE evaluation criteria. It is also revealed that the predictive power of the machine learning model is very close to CGARCH and MRS model; therefore, the LSTM model can be used as an alternative to GARCH and regime-switching models for stock market volatility. These findings will help national and international investors, policy-makers, geographical economists, and industrialists to use thebestforecastmodeltomakebetterpoliciesandgaintremendousprofit.