You are testing the effect of merger announcements on stock prices.

(This is an event study.) Your procedure goes through the following steps:

Step 1: You choose the 20 biggest mergers of the year.

Step 2: You isolate the date the merger became effective as the key day around which you will examine the data.

Step 3: You look at the returns for the five days after the effective merger date.

By looking at these returns (0.13%) you conclude that you could not have made money on merger announcements. Are there any flaws that you can detect in this test? How would you correct for them? Can you devise a stronger test?