Abnormal returns using WRDS event study program

[Glen Young]

Hello,

Has anybody used WRDS’ event study application and SAS program at the following link?

https://wrds-web.wharton.upenn.edu/wrds/research/applications/event/run/

I’m using it to look at stock performance around a particular event (shareholder activist announcement). The output I’m primarily interested in are its four measures of abnormal returns, which are essentially a firm’s return minus some estimation of “normal” returns. Each of these four abnormal return measures is then aggregated over the event window using two different methods to create cumulative abnormal returns (CARs) and buy-and-hold abnormal returns (BHARs). With 4 measures of abnormal returns and 2 aggregation methods, I’m ultimately interested in 8 measures.

My problem is that I get weird results for a few of the BHAR measures, as summarized in the picture at the following link (note, I didn't embed the image directly it will only embed at the top of my message): http://imgur.com/4QqH71A

Basically, for three of the BHAR measures, returns start to plunge dramatically over time, especially with longer horizons. Here are BHAR returns measured from days [-10, 200]: http://imgur.com/kzOfUWJ

Here’s the same graph for days [-100, 200]: http://imgur.com/jP9Jgzj

Obviously, a firm can have, at most, a -100% return, so the fact that the average return from day -100 to 200 is about -500 to -600% tells me that there’s something wrong with the code. Also, a similar pattern is observed if I use an entirely different type of event (restatement). As the graphs above show, the market-adjusted return (bhar0) behaves as expected. So do the four CAR measures: http://imgur.com/VDZ0Jqw

I've been unable to figure out what's going on for days now, so any light that anyone could shed would be a lifesaver. Thanks so much!

Glen