What it means
Share prices move every day for many reasons, so it is hard to tell whether a particular announcement mattered. An event study isolates the effect by first estimating a normal pattern, based on how the stock usually behaves relative to the overall market.
Any movement beyond that pattern during the event window is attributed to the news. The analyst picks an estimation window, often a few months of data before the event, to calibrate the normal relationship.
An event window, typically a few days around the announcement, is then examined. Summing the abnormal returns across the window gives the cumulative abnormal return (CAR).
Event studies are widely used in corporate finance and legal disputes. Companies use them to judge whether acquisitions, dividend changes or leadership appointments created value, and courts use them in securities cases to estimate how much a misleading statement affected the price.
Results depend on good design. If other news arrives at the same time, the effect is contaminated, and if the market anticipated the event, the price reaction may have occurred before the window opens.
The result is an estimate of average market reaction, not proof of cause. The method assumes that markets are reasonably efficient, meaning prices adjust quickly to new information.
That is why the event window is kept short, since the longer the window, the more unrelated noise is captured. Practitioners often test whether an abnormal return is statistically significant rather than random.
They compare it with the typical day-to-day variation of the stock outside the event window, expressed as a standard deviation. An abnormal return that is large relative to normal variation is more likely to reflect the news itself.
In practice
Real-world examples
Example 1
A bank studies 40 takeovers it has completed and finds the acquirers' shares fell an average of 2% around the announcement. The board uses this to question its deal strategy. The pattern across many deals is more informative than any single result.
Example 2
A lawyer in a shareholder dispute uses an event study to estimate that a false statement inflated a company's share price by about $3 per share before the truth emerged. The estimate then forms part of the damages calculation presented to the court.
Example 3
A consumer brand measures how its share price reacted to the unveiling of a new chief executive and compares it with the reaction to rivals' appointments. This gives the board a benchmark for judging how the market views leadership changes.
Think of it
Imagine tracking a runner's speed on a normal day. If you suddenly hand them a heavy backpack, an event study helps you measure exactly how much slower they ran compared to their usual pace, filtering out the effect of a sudden headwind.
Formula
Calculation
Abnormal return = Actual return - Expected return
Expected return = Alpha + Beta x Market return
Cumulative abnormal return = Sum of daily abnormal returns across the event window
Suppose Halden Pharma, a fictional company, announces a successful drug trial. On the announcement day the market returns 1.0%, and the stock has a beta of 1.2 and an alpha of zero, so the expected return is 1.2 x 1.0% = 1.2%. The actual return is 5.2%, giving an abnormal return of 5.2% - 1.2% = 4.0%. The next two days show abnormal returns of 1.0% and -0.5%, so the CAR is 4.0% + 1.0% - 0.5% = 4.5%. On a market value of $400,000,000, that equals $400,000,000 x 4.5% = $18,000,000 of value attributed to the news.Case study
Seen in the real world
Lakeshore Energy is an illustrative, fictional utility that announced a surprise dividend increase. Executives believed it would please investors, but the CFO asked an analyst to test that belief.
The analyst built an event study with a 120-day estimation window and a three-day event window. The stock's abnormal return was a modest 0.8%, and it was not distinguishable from normal random movement.
In the illustrative conclusion, the team realised that investors had already expected the increase, so little new information was released. The company decided to focus future communication on genuine surprises, such as new investment plans, rather than on changes the market had priced in. The CFO then asked the analyst to run the same test on past announcements of other kinds, to guide how Lakeshore would frame future ones.
Watch out
Common mistakes
- Reading the raw share price change as the effect of the event, without removing the normal market movement.
- Using a long event window that captures many unrelated news items.
- Assuming an abnormal return proves the event caused the move, when other simultaneous news may be responsible.
Questions
People also ask
What is an abnormal return?
It is the difference between the stock's actual return and the return expected from its usual relationship with the market.
How long should an event window be?
It is commonly one to a few days around the announcement, because short windows reduce noise.
Can event studies be used on private companies?
Not directly, because they need frequently traded prices, so analysts usually study listed peers instead.