International Conference on Economics, Finance & Business, Rome

MODELING THE BEHAVIOUR OF CONDITIONAL VOLATILITY IN EMERGING AFRICAN STOCK MARKETS

SIBANISEZWE ALWYN KHUMALO

Abstract:

Stock market volatility remains a crucial element in economic investigations as it is a leading indicator of a country’s financial performance. Volatility, defined as the variation in asset prices over time, can be classified into unconditional volatility—characterised by a constant variance without additional information—and conditional volatility, which adjusts based on new information. The continuous price discovery process in stock markets, driven by the release of new information, creates dynamic volatility patterns that necessitate robust modelling techniques for risk management and portfolio diversification. This study focuses on modelling the conditional volatility of stock markets in emerging African economies. African markets, often exhibiting low or negative correlations with developed economies, present unique opportunities and challenges as they are less influenced by external news yet display characteristics such as persistence, long memory, and asymmetry in volatility. Understanding these features is critical for investors and policymakers who rely on accurate volatility measurements to mitigate risk and make informed decisions. Drawing on Asset Pricing Theory, the Efficient Market Hypothesis, and Behavioural Finance, this research employs a quantitative, deductive approach under a positivist paradigm. The study analyses daily stock returns derived from MSCI U.S. dollar national equity indices for ten emerging African markets across Northern, Eastern, Southern, and Western Africa from January 2006 to September 2024. This period encompasses significant economic shocks, including the 2007/09 Global Financial Crisis and the COVID-19 pandemic, providing a robust data set for analysis. The research evaluates a comprehensive family of Generalised Autoregressive Conditional Heteroscedasticity (GARCH) models—specifically, GARCH, P-GARCH, GARCH-M, E-GARCH, T-GARCH, FIGARCH, and FIEGARCH—to determine the optimal model for capturing the conditional volatility of each market. Each model's ability to capture key volatility characteristics such as persistence, long memory, asymmetry, and the risk-return relationship is assessed through stationarity, non-negativity, and significance tests. Goodness-of-fit measures, including the Schwarz Information Criterion, Log Likelihood, Akaike Information Criterion, and News Impact Curves, are then applied to select the best-fitting model. By bridging the research gap in multi-country studies of emerging African markets, this study aims to provide a holistic view of volatility risk. Its findings will benefit stock exchanges, regulators, policymakers, and investors, contributing significantly to the literature on stock volatility.

Keywords: Conditional volatility, GARCH, Emerging African Markets, Asset prices