Modeling High-Frequency Volatility with Volume-Driven Intraday Effects

Modeling and forecasting volatility is critically important for portfolio construction and risk management. Numerous volatility models exist, and one established line of research incorporates trading volume, motivated by the idea that volume reflects the rate of information flow into prices and is therefore positively related to volatility.

Reference [1] extends this line of research by proposing a model in which volume enters volatility dynamics in a lagged, rather than contemporaneous, manner. Specifically, the paper develops a high-frequency stochastic volatility model that integrates a persistent component, intraday periodicity, and volume-driven time-of-day effects.

The key novelty lies in modeling these volume-driven intraday effects as time-varying through lagged trading activity. The authors pointed out,

This paper develops a framework for modeling high-frequency volatility that jointly accounts for persistent behavior, static time-of-day adjustments, and volume-driven time-of-days effects. By conditioning intraday periodicity on past trading volume, our model captures timely and economically relevant deviations from static volatility patterns, providing a more accurate and microstructure consistent description of intraday risk dynamics.

Across equity indexes, currency, and commodity futures, the volume-driven component explains a large share of intraday volatility variation, highlighting the importance of conditioning on trading activity. The model not only reproduces well-known U-shaped volatility patterns but also uncovers features associated with global trading activity. It indicates that the deviations from the time-of-day average volume are deeply tied to volatility. Our results complement the Sequential Information Arrival Hypothesis (SIAH) literature by showing that the lead–lag relationship between trading volume and volatility also holds in a high-frequency setting…

From an economic perspective, these statistical improvements translate into superior portfolio performance. When applied to volatility-managed investment strategies, our forecasts consistently improve Sharpe ratios relative to unmanaged benchmarks and competing volatility models. The benefits are most striking in equity index futures, where dynamic exposure based on our specification more than triples the Sharpe ratio compared to GARCH or standard stochastic volatility alternatives. Overall, our results suggest that lagged trading volume provides a critical dimension for understanding and forecasting intraday volatility.

In short, the paper develops a high-frequency volatility framework that integrates persistent dynamics, time-of-day effects, and lagged trading volume. The results demonstrate that this volume-driven specification improves volatility forecasts and translates into higher risk-adjusted performance in volatility-managed portfolios, particularly for equity index futures.

This contribution is meaningful. We note that the instruments examined in the paper belong to relatively mature markets. It would be interesting to investigate whether similar effects hold in new, developing markets, such as cryptocurrencies.

Let us know what you think in the comments below or in the discussion forum.

References

[1] Igor Ferreira Batista Martins, Audronè Virbickaitè, Hoang Nguyen and Hedibert Freitas Lopes, Volume-driven time-of-day effects in intraday volatility models, Örebro University School of Business, 2025

Originally Published Here: Modeling High-Frequency Volatility with Volume-Driven Intraday Effects

source https://harbourfronts.com/modeling-high-frequency-volatility-volume-driven-intraday-effects/