Specification tests of parametric dynamic conditional quantiles

This article proposes omnibus specification tests of parametric dynamic quantile models. Contrary to the existing procedures, we allow for a flexible specification, where a possibly continuum of quantiles are simultaneously specified under fairly weak conditions on the serial dependence in the underlying data generating process. Since the null limit distribution of tests is not pivotal, we propose a subsampling approximation of the asymptotic critical values. A Monte Carlo study shows that the asymptotic results provide good approximations for small sample sizes. Finally, an application suggests that our methodology is a powerful alternative to standard backtesting procedures in evaluating market risk.