Eventstudy

An event study is an empirical research method used to assess how a specific event affects the value of a firm, typically proxied by its stock price. The central idea is that new information released around the event should lead to price adjustments reflected in abnormal returns.

Methodology: identify the event date and define an estimation window before the event to estimate expected returns, then specify an event window around the date to observe price reactions. Expected returns are usually estimated with a market model or other risk-adjusted models; abnormal returns are the difference between observed and expected returns, and cumulative abnormal returns (CAR) summarize the effect over the window. Statistical tests (t-tests, nonparametric procedures, or cross-sectional tests) determine whether abnormal returns are statistically different from zero.

Data and scope: studies rely on daily or intraday price or return data. Common events include earnings announcements, mergers and acquisitions, stock splits, dividend changes, or regulatory actions. Extensions use multi-factor models such as the Fama-French or Carhart models, different estimation horizons, and cross-sectional designs to improve inference.

Limitations: results can be sensitive to event definition and window length, model risk, confounding events, data quality, and liquidity. Assumptions of market efficiency and independence across firms may not hold, and thinly traded stocks can yield biased estimates.

History: the event-study approach traces to foundational work in the late 1960s, notably Fama, Fisher, Jensen, and Roll on stock price adjustments to new information. Key methodological refinements were provided by Brown and Warner in the 1980s and by MacKinlay in later work.