The classical event-study pipeline, modernized.

How an event study works

An event study isolates the market's reaction to a specific event by comparing actual returns against what would have been expected absent the event. The difference is the abnormal return. Aggregating abnormal returns across firms or across the event window, and testing for statistical significance, gives you a publication-grade quantification of impact.

The five-step recipe

1. Define the event. Pick a ticker, a date, and an event window.
2. Fit the expected-return model. Market model, CAPM, Fama-French 3/5-factor, Carhart, or constant mean.
3. Compute abnormal returns. AR_i,t = R_i,t − E(R_i,t).
4. Test significance. Parametric (t-test, BMP, Patell) and non-parametric (Corrado rank, generalized sign).
5. Publish. Export to LaTeX, CSV, or Stata-ready formats.

Where to go next

• Researcher new to the topic → start with Introduction
• Need formulas → Significance tests
• Picking a model → Expected-return models
• Building a study → Application blueprint