2T-POT Hawkes model for left- and right-tail conditional quantile forecasts of financial log returns: Out-of-sample comparison of conditional EVT models

Conditional extreme value theory (EVT) methods promise enhanced forecasting of the tail events that often dominate systemic risk. We present an improved two-tailed peaks-over-threshold (2T-POT) Hawkes model is adapted for conditional quantile in both left and right tails a univariate time series. This applied to daily log returns six large-cap indices. also take unique step fitting at multiple exceedance thresholds (from 1.25% 25.00% mirrored quantiles). Quantitatively similar asymmetries parameters are found across all indices, adding further empirical support temporal leverage effect financial price series which impact losses not only larger but more immediate. Out-of-sample backtests find our 2T-POT reliably accurate than GARCH-EVT when (mirrored) risk expected shortfall 5% coverage level below. suggests asymmetric Hawkes-type arrival dynamics better approximation true generating process GARCH-type volatility. Our therefore presents better-performing alternative modelling.