Skip to content
Event Studies and Causal Measurement in Finance / Abnormal-return measurement

Overlapping Events

The variance of an average cumulative abnormal return sets the covariance between securities to zero. That step is justified when event windows do not share calendar days.

Earnings announcements of different Dow names sometimes fall on the same morning, and an index inclusion hits many names on one date by construction. The covariances are then not zero, and the usual standard error is the wrong one.

This note is about that calendar overlap: how to see it in the sample, and the two repairs MacKinlay records. Inference that stays with firm-level residuals and estimates their cross-sectional variance is the next topic, in cross-sectional dependence.
Repeated quarters for one firm are multiple events, a different dependence.
Learning objectives. By the end of this note you should be able to:
  • mark two event windows as overlapping or disjoint in calendar time;
  • state what the zero-covariance formula does when they overlap;
  • describe the event-time portfolio repair and the multivariate dummy-variable repair;
  • say which repair can accommodate a positive reaction at one firm and a negative reaction at another.

In this lesson

  1. What overlap does to the variance
  2. A two-firm variance fixture
  3. Two repairs
  4. Audit on the running announcement

Before you start

The full lesson, with worked examples and exercise solutions, is for members.

Read the full lesson

New to SelfMadeQuant.com? See membership options or browse the Event Studies and Causal Measurement in Finance syllabus.