An empirical researcher who wants the price effect of one corporate event — here, a quarterly earnings announcement — has to name the event before any market-model abnormal return exists. The announcement date, the rule that sorts the surprise into good news, no news, or bad news, and the listing screen that builds the sample all change the object the cumulative abnormal return later claims to measure.
This note fixes that object. Window lengths and the normal-return regression sit in the next two lessons.
Learning objectives. By the end of this note you should be able to:
- state the seven steps of an event study and say which step the event definition is;
- write the selection screen for a sample and name the bias that screen can introduce;
- classify an earnings release with the 2.5 percent surprise rule used in the Dow example;
- separate the event date from the economic hypothesis the price reaction is asked to test.
In this lesson
- The seven steps
- Selection and the surprise sort
- Dating, screens, and other events
- What the definition does not deliver
- 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.