Event study
Measure a market reaction around a dated event
An event study estimates how security prices change around a clearly timed announcement relative to a model of normal returns. Abnormal return is observed minus expected return; cumulative abnormal return sums these deviations over a specified window. Short-window announcement studies differ from long-horizon performance studies, where benchmark and compounding choices become especially consequential.
Choose this for questions about market reactions to announcements when event dates, adjusted returns and a defensible benchmark are available. It works best when narrow windows limit competing news.
Strengths
- Makes event timing and the return benchmark explicit
- Can aggregate reactions across events and examine heterogeneous responses
Limitations
- Confounding news can contaminate the window
- Results depend on the normal-return model and dependence across events
Know the boundary
CAR is a model-relative price reaction. An observational announcement study does not automatically identify the causal effect of the underlying corporate decision.