Estimating value-at-risk models for non-conventional equity market index

In this study, we evaluate Value-at-Risk (VaR) forecasts for non-conventional (i.e., Islamic) equity markets using various time-varying volatility models. Recent evidence suggests that volatility shifts in returns cause non-normality by significantly increasing kurtosis. Consequently, we endogenously detect significant shifts in the volatility of our Islamic equity market index returns and incorporate this information into selected models to estimate the downside risk. Our results show that the best forecast is generated by the dynamic quantile regression (DQR) model that does not make any specific assumptions about the underlying returns distribution. We also show the economic implications of our findings by calculating daily capital charges under Basel II Accord.