The Role of Risk‐Neutral Moments in Forecasting Future Realised Variance: An International Perspective
ABSTRACT This paper estimates the predictability of future realised variance for eight stock markets using US‐listed proxy ETFs, based on their risk‐neutral variance, skewness, and excess kurtosis. Our evidence shows that risk‐neutral variance and skewness have a positive relationship with future realised variance, while excess kurtosis has a negative relationship. Results from various statistical tests for model comparison indicate that incorporating second and higher‐order risk‐neutral moments from specific market proxies, as well as aggregated moments across market proxies, into the HAR‐RV model enhances the predictability of future realised variance. The inclusion of these moments provides additional information, significantly improving forecasting power and reducing extreme underestimation. Moreover, portfolio allocation exercises demonstrate that these improvements can translate into economically meaningful certainty‐equivalent return gains.
Authors
- Jin E. Zhang (ORCID: https://orcid.org/0000-0002-8556-474X)
- Junyu Zhang (ORCID: https://orcid.org/0000-0003-3178-1295)
- Xinfeng Ruan (ORCID: https://orcid.org/0000-0002-1447-6603)
Institutions
- Queen's University Belfast (GB)
- Xi’an Jiaotong-Liverpool University (CN)
- University of Otago (NZ)
Publication Details
- Journal
- International Journal of Finance & Economics
- Published
- 2026-11-15
- DOI
- https://doi.org/10.1002/ijfe.70328
- Primary Topic
- Financial Risk and Volatility Modeling
- Type
- article
- Field-Weighted Citation Impact
- 0.00