Realized Tail-Risk Conditional Autoregressive Expectile Models Improve Stability of VaR and ES Forecasting
ORIGINAL / Realized Tail-Risk Conditional Autoregressive Expectile Models: Accurate and Stable Forecasting Performance
This study proposes the RTCARE model, which incorporates realized quantile and realized ES constructed from intraday high-frequency returns into the dynamics and measurement equations of conditional expectiles for joint VaR and ES forecasting. Its novelty lies in combining low-frequency tail-risk forecasting with high-frequency realized tail information; it has potential practical value for quantitative risk management and backtesting.
01 ABSTRACT
The paper proposes a realized tail-risk conditional autoregressive expectile (RTCARE) model that introduces realized tail measures such as realized quantile and realized ES from intraday high-frequency returns into the dynamics and measurement equations of conditional expectiles to jointly forecast VaR and ES. Parameters are estimated via Markov chain Monte Carlo, and the resulting forecasts are evaluated in numerical examples and an empirical application using U.S. stock-return data. The authors report that RTCARE delivers competitive and often superior forecasting performance relative to existing alternatives across tail levels, with stable results in Monte Carlo replications and rolling forecast evaluations.
02 KEY FINDINGS
- RTCARE incorporates realized quantile and realized ES constructed from intraday high-frequency returns into the dynamics and measurement equations of conditional expectiles.
- Parameters are estimated using Markov chain Monte Carlo, targeting joint forecasting of VaR and ES.
- In numerical examples and U.S. stock-return empirical analysis, the authors report that RTCARE has competitive and often superior forecasting performance across multiple tail levels.
AI GENERATED SUMMARY / DISCOVERED BY ARXIV Q-FIN