Observation Period

The observation period in an Event Study is the post-event window in which abnormal returns are computed. The baseline model fit during the Estimation Period is projected into this window; any deviation from the model's predictions is attributed to the event. The window length depends on how quickly you expect effects to unfold: instantly for financial news, over years for policy interventions with behavioural adjustment.

When to use

PS_3's observation window is Nov 20 1989 – Dec 31 1993 (~4.1 years post-HNC) — long enough to detect delayed effects from behavioural responses (the second-car-purchase substitution Davis 2008 documents takes time). A too-short window misses delayed responses; a too-long window risks contaminating with other concurrent events. State the window choice and its justification explicitly in any event-study writeup.