Tail risks and forecastability of stock returns of advanced economies: evidence from centuries of data*

This study examines the out-of-sample predictability of market risks measured as tail for stock returns eight advanced countries using a long-range monthly data over century. We follow Conditional Autoregressive Value at Risk (CAViaR) Engle and Manganelli (2004) to measure consequently, we produce results both 1% 5% VaRs across four variants (Adaptive, Symmetric absolute value, Asymmetric slope Indirect GARCH) CAViaR. Thereafter, use “best” fit in return selected markets. For forecasting exercise, construct three predictive models (one-predictor, two-predictor three-predictor models) examine their forecast performance contrast with driftless random walk model. Three findings are discernible from empirical analysis. First, find that choice VaR matters when determining CAViaR model each series outcome seems differ between VaRs. Second, incorporates risk oil produces better outcomes than one own indicating significance domestic global countries.