---
title: Realized Tail-Risk Conditional Autoregressive Expectile Models: Accurate and Stable Forecasting Performance
url: https://www.ml-quant.com/papers/arxiv/2610.13446/
site: ML-Quant (https://www.ml-quant.com)
updated: 2026-10-16
license: Summaries CC BY 4.0; links go to the original sources
index: https://www.ml-quant.com/llms.txt
identifier: arXiv:2610.13446
source_url: https://arxiv.org/abs/2610.13446
featured: 2026-10-16
citations: unknown
topic: Econometrics & Forecasting
---


# Realized Tail-Risk Conditional Autoregressive Expectile Models: Accurate and Stable Forecasting Performance

A realized tail-risk model using high-frequency intraday data forecasts value-at-risk and expected shortfall with competitive performance across U.S. stock returns and multiple tail levels.

- Source: https://arxiv.org/abs/2610.13446
- Identifier: arXiv:2610.13446
- Released: 2026-10-12
- First featured: Quant Letter No. 135 (2026-10-16): https://www.ml-quant.com/issues/2026-10-16/
- Citations (Semantic Scholar): not tracked
- Published in: not yet
- Topic: Econometrics & Forecasting
- Authors: Keisuke Sasaki, Yasuhiro Omori

## Abstract (arXiv, CC0)

A realized tail-risk conditional autoregressive expectile (RTCARE) model is proposed for joint forecasting of value-at-risk (VaR) and expected shortfall (ES). Realized tail measures such as the realized quantile and the realized ES, constructed from intraday high-frequency returns, are incorporated into the dynamics and measurement equations of the conditional expectile, allowing information about lower-tail risk to enter the forecasts. The parameters are estimated by Markov chain Monte Carlo, and we evaluate the resulting VaR-ES forecasts in numerical examples and in an empirical application using U.S. stock-return data. Across the considered tail levels, the RTCARE models deliver competitive and often superior forecasting performance relative to existing alternatives, with stable results across Monte Carlo replications and rolling forecast evaluations.

## Related

- [Constrained Sampling with Primal-Dual Langevin Monte Carlo](https://www.ml-quant.com/papers/arxiv/2411.00568/): The study presents a PD-LMC algorithm that samples from a probability distribution while meeting statistical constraints, useful in Bayesian inference and prediction fairness.
- [An Itô-type Formula: High Frequency Estimation under Rough Volatility](https://www.ml-quant.com/papers/ssrn/7615661/): A new calculus framework for rough volatility below the semimartingale regime enables bias-corrected estimation of leverage effects and roughness with central limit theorems, rejecting semimartingale assumptions for recent S&P data.
- [Nonlinear Drivers of Macroeconomic Tail Risk: A Threshold Stochastic Volatility-in-Mean VAR with Regime-Dependent Leverage](https://www.ml-quant.com/papers/arxiv/2609.26994/): A regime-switching volatility-in-mean VAR reveals that the drivers of growth and inflation tails differ from median dynamics, with macroeconomic uncertainty playing a larger role in downside risk.
- [Network Realized GARCH--Itô Models: Volatility Spillovers with High-Frequency Identification](https://www.ml-quant.com/papers/arxiv/2609.29515/): Introduces a network realized GARCH-Itô model that identifies dynamic volatility transmission among assets using high-frequency data, outperforming recursive forecasts on sector ETFs.
- [Predicting Stock Market Crash with Bayesian Generalised Pareto Regression](https://www.ml-quant.com/papers/arxiv/2506.17549/): The study uses a Bayesian Generalised Pareto Regression model to predict extreme losses in India's Nifty 50 index, showing increased tail risk with higher market volatility and the model's effectiveness in forecasting tail risk in emerging markets.
- [Using quantile time series and historical simulation to forecast financial risk multiple steps ahead](https://www.ml-quant.com/papers/arxiv/2502.20978/): A new method for estimating Value-at-Risk (VaR) and Expected Shortfall (ES) models allows for tail risk forecasting and can be applied to any data or model where the VaR and ES relationship remains constant.