The Dow earnings sample is 30 firms and 600 announcements: each name contributes about twenty quarterly dates.
Those dates can be separated in calendar time, so the
overlap problem need
not bind, and the CARs can still be dependent because they belong to the same firm.
This note treats that repeated-event structure, the cross-sectional regression that links the abnormal return to
the size of the surprise, and the power arithmetic that says whether twenty names can detect a half-percent move.
It uses the core
regression
tools for the cross section and does not re-teach them.
Learning objectives. By the end of this note you should be able to:
- count events and firms separately in the Dow example;
- write the cross-sectional regression of the abnormal return on event characteristics;
- state why White standard errors are the default in that regression;
- explain the anticipation problem that can hide a valuation slope;
- read MacKinlay's power figures at for a percent and a percent abnormal return.
In this lesson
- The same firm, many quarters
- Regression on the surprise
- Power when each firm is one draw
- Audit on the running announcement
Before you start
The full lesson, with worked examples and exercise solutions, is for members.
Read the full lessonNew to SelfMadeQuant.com? See membership options or browse the Event Studies and Causal Measurement in Finance syllabus.