Answers are grounded in this paper and cite their sources.
Purpose: This study evaluates the predictive accuracy of various econometric models in forecasting stock market returns within the Nigerian Exchange Group (NGX). Given the high volatility and structural breaks inherent in emerging African markets, the research investigates whether linear (ARIMA), volatility-sensitive (GARCH), or hybrid rolling-window frameworks provide the most robust tools for financial forecasting. Methodology: Employing a quantitative design, the study analyses daily All-Share Index (ASI) closing prices (sourced from the NGX official database) from January 1, 2021, to August 30, 2025. The approach utilizes Box-Jenkins formalization for ARIMA parameters and GARCH (1,1) to account for volatility clustering, optimized via a 180-day rolling window. Stationarity was verified via ADF tests, with selection guided by the Akaike Information Criterion (AIC). Results and conclusion: Empirical findings indicate that hybrid ARIMA-GARCH models significantly outperform standalone ARIMA and GARCH models in forecasting the NGX ASI. While linear models fail to account for significant ARCH effects, the hybrid configuration markedly reduces forecast errors, as confirmed by superior MAPE and RMSE metrics. The NGX exhibits high volatility persistence, rendering traditional linear models insufficient. Findings suggest the NGX deviates from the Weak-Form Efficient Market Hypothesis, as historical patterns retain predictive value. findings: These results advise institutional investors to prioritize volatility-aware models for asset pricing and Value-at-Risk (VaR) estimations. Furthermore, they underscore the need for policymakers to implement stabilizing interventions during periods of high turbulence.
No takes yet. Share an insight, caveat, or question.
Ekpete et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: