Modeling Volatility Asymmetry in Government-Owned Stocks: Evidence from Value at Risk Estimation in Indonesia
This study evaluates the comparative effectiveness of four volatility models—EWMA, GARCH(1,1), EGARCH, and TGARCH—in estimating daily Value at Risk (VaR) for a portfolio of Indonesian state-owned enterprise (SOE) stocks over the period 2019–2024. Motivated by the rapid growth of retail investor participation and increasing exposure to market risk in Indonesia’s emerging capital market, the research addresses a critical gap in empirical risk modeling for government-owned equities. A key contribution of this study lies in the integration of asymmetric GARCH-family models with Student-t innovations into a VaR estimation framework, tailored specifically to SOE stocks—an approach seldom explored in the Southeast Asian context. The analysis uses daily return data from ten liquid, sectorally diverse SOEs. Volatility is estimated via parametric methods, assuming a normal distribution for EWMA and Student-t distributions for GARCH-type models. Model accuracy is evaluated through in-sample and out-of-sample backtesting, employing MAE, RMSE, and the Kupiec and Christoffersen tests. The findings indicate that GARCH(1,1) performs most reliably at the 95% confidence level, while TGARCH demonstrates superior performance at the 99% level, particularly in capturing downside risk. EGARCH tends to produce conservative estimates, whereas EWMA underestimates tail risk. The results support a dual-model strategy for operational monitoring and capital risk management.