Volume Event Study

What a volume event study is

A volume event study examines whether the trading volume of an asset shows statistically significant abnormalities around a defined event. Instead of asking how prices moved, it asks how heavily the security traded relative to its normal activity. Two settings have made the approach popular. The first is the investigation of possible insider trading ahead of merger and acquisition announcements, where unusual volume can appear before the news becomes public. The second is securities fraud litigation, where abnormal volume is increasingly used as supporting evidence of how the market absorbed information.

How it differs from a return study

A return event study measures abnormal returns: the gap between the return a security actually earned and the return that would have been expected without the event. A volume event study keeps the same event-window and estimation-window logic but swaps the outcome variable. The metric of interest becomes abnormal trading volume rather than abnormal return. Because raw share counts vary widely across firms, volume studies typically work with a relative, log-transformed measure of volume per firm so that securities of very different sizes can be compared on a common scale.

Why volume reacts to events

Under strict market efficiency, trading volume should carry no predictive power over future returns. Relaxing those assumptions, several arguments link abnormal volume to information. Volume and returns can show autocorrelation, abnormal trading can reflect shifts in liquidity, and unusual volume can signal undisclosed information. When insider trading is present, the abnormal volume roughly corresponds to the trading generated by insiders. This signal tends to be asymmetric: short-selling restrictions make it harder to act on negative private information, so the informational content of abnormal volume is often stronger on the positive side. Early work by Epps (1975) and Karpoff (1985) helped establish the volume-return relationship.

How to measure abnormal volume

Abnormal volume is the difference between observed volume and a benchmark of normal volume. Two benchmarks are common. The mean-adjusted model treats the average volume over the estimation period as normal and measures deviations from it. The market model instead regresses a firm's volume on overall market volume and measures the residual, which controls for marketwide swings in activity but requires more data. An estimated generalized least squares variant of the market model is also available for more demanding analyses. The foundational methodology for measuring abnormal daily trading volume is set out by Campbell and Wasley (1996).

Running the analysis

You can run a volume event study with the Abnormal Volume Calculator. It applies these benchmark models to your event and estimation windows and returns abnormal volume measures together with statistical tests, so you can judge whether the volume reaction around your event is significant.