A Bootstrap Approach for Generalized Autocontour Testing. Implications for VIX Forecast Densities

We propose an extension of the Generalized Autocontour (G-ACR) tests for dynamic specification of in-sample conditional densities and for evaluation of out-of-sample forecast densities. The new tests are based on probability integral transforms (PITs) computed from bootstrap conditional densities that incorporate parameter uncertainty. Then, the parametric specification of the conditional moments can be tested without relying on any parametric error distribution yet exploiting distributional properties of the variable of interest. We show that the finite sample distribution of the bootstrapped G-ACR (BG-ACR) tests are well approximated using standard asymptotic distributions. Furthermore, the proposed tests are easy to implement and are accompanied by graphical tools that provide information about the potential sources of misspecification. We apply the BG-ACR tests to the Heterogeneous Autoregressive (HAR) model and the Multiplicative Error Model (MEM) of the U.S. volatility index VIX. We find strong evidence against the parametric assumptions of the conditional densities, i.e. normality in the HAR model and semi non-parametric Gamma (GSNP) in the MEM. In both cases, the true conditional density seems to be more skewed to the right and more peaked than either normal or GSNP densities, with location, variance and skewness changing over time. The preferred specification is the heteroscedastic HAR model with bootstrap conditional densities of the log-VIX.