Volatility dynamics and regime switching in Ghanaian gold returns: Evidence from GARCH, GJR-GARCH and time-varying-transition-probability models

Authors

DOI:

https://doi.org/10.18488/62.v13i1.5370

Keywords:

Asymmetric volatility, GARCH models, Ghanaian gold returns, Markov-switching models, Volatility forecasting.

Abstract

This study examines the volatility dynamics of Ghanaian gold returns using GARCH, EGARCH, GJR-GARCH, and a time-varying-transition-probability Markov-switching GJR-GARCH model. The analysis uses 807 daily Gold/GHS log returns derived from 808 price observations spanning 3 January 2022 to 30 December 2025. The return series is stationary, non-normal, negatively skewed, leptokurtic, and conditionally heteroskedastic. Student-t and generalized-error innovations generally improve the fit of the single-regime models relative to the normal distribution. The Markov-switching specification identifies a persistent low-volatility regime and a shorter high-volatility regime, with average conditional variances of 2.7150 and 9.5231, respectively. Asymmetry is concentrated in the high-volatility regime, where the estimated coefficient is 1.0427. Out-of-sample comparisons and the trading-rule evaluation show that model rankings depend on the evaluation criterion and forecast horizon. Overall, Ghanaian gold-return volatility is persistent, heavy-tailed, asymmetric during stressed periods, and regime-dependent. The findings provide practical implications for volatility forecasting, market-risk management, hedging, and investment decisions in commodity-dependent economies.

Downloads

Download data is not yet available.

Downloads

Published

2026-11-19