Event studies are better suited to studying events that occur on different days for different companies. This reduces the probability of “event contamination.”

For example, let’s presume that you are interested in the effect of a low inflation announcement on September 12, 2005. Your evidence shows that stock prices went up on this day. Therefore, you might be tempted to conclude that the inflation announcement had a positive stock price influence. However, this overlooks an important problem. On this day, a million other things may have happened: the President coughed, the Congress squabbled, the Fed grumbled, the FDA changed its mind on genetic engineering, investors grew colder on mining stocks and hotter on game stocks, OPEC met, the Europeans demonstrated against U.S.

policy, and so on. Are you really sure that it was the inflation announcement that made stocks go up and none of the other events? In contrast, if the event day is different for every firm, sometimes these other events will positively influence the market, sometimes negatively. Net in net, this other-events contamination is more likely different on different days and thus it will more likely wash out.

Of course, if your event is on different days but still always on firms’ annual meetings, then you have the different problem that there could be a lot of other value-relevant news that is being disclosed simultaneously.

In this case, you are likely to have more noise, uncertainty, and contamination to deal with than in the case where event days occurred randomly for different firms.