Eventus performs event studies using data read directly from CRSP stock databases. This means you can only do an event study with American stocks. If you want to do an event study with non-American stocks you have to use data from Datastream. For this data the Datastream Event Study Tool has been developed.
In this manual we will guide you through the steps you have to take in Eventus. However, you have to make most choices yourself. For example: how long are the estimation and event windows, what statistical tests do you want to use, which model? The EDSC can’t help you answering these questions; you have to ask your supervisor or coach for help and check the academic literature – what choices have previous researchers made?
Under the link Manuals and Overviews at the top of the page you find the Eventus User Guide, which provides a lot of background information.
Access to Eventus
Eventus is part of Wharton Research Data Service (WRDS). You find WRDS in the list of financial databases on the website of the EDSC. To get access to WRDS (and Eventus) you need a VPN-connection. WRDS also asks for your e-mail-address. Enter your ERNA-e-mail-address. WRDS sends a link to the entered e-mail-address; this link is valid for 14 days. After these 14 days you can request a new link.
Within Eventus you can choose different types of event studies. Most students use the Basic Daily Event Study: an event study with daily data centered on a single date for each company. You can make this selection in the column at the left side of the screen.
Step 1: What identifier and request file will you use?
You have to upload a text file (.txt) consisting of 8-digit Cusips or PERMNO’s and the event dates in the format yyyymmdd. Make sure you mark the identifier that’s in your text file.
Under Manuals and Overviews you find the Eventus Validation Tool: here you can upload your txt-file – the tool will convert the dates in the file into the required Eventus format.
Step 2: Which Market Indices and Benchmark Options would you like?
Here you select the market portfolio you want to compare the return of the stock with and the model you want to use. The Market Model is selected by default.
In the Eventus User Guide you can find more information about the benchmark options in Appendix A.1.
Step 3: Which Estimation Options would you like?
The estimation period is the period before the event, used to calculate the ‘normal’ return.
Autodate is useful when you have event dates on non-trading dates: by default NONE is marked. This means that event dates on which the market is closed (for example weekends, holidays) are dropped from the analysis.
When you select YES non-trading dates are converted to the next trading day (Sunday is converted in Monday). When you select BACK non-trading dates are converted to the previous trading day (Sunday is converted into Friday).
Step 4: Which event windows would you like to search?
The event period is the period around the event. When you enter alternative windows abnormal returns will also be calculated for these windows. Make sure that these alternative windows are within the Event period.
Step 5: Which tests would you like to search?
Some background information of the statistical tests can be found in the Eventus User Guide, from page 29.
You can find the meanings of this parameters in the Eventus User Guide in the section B.5.
Most asked for is:
DETAIL=FULL – when you mark this option, you will get the individual daily abnormal returns and standardized abnormal returns for each security-event.
When you have entered all information, you can enter your e-mail address (WRDS will send an e-mail with a link to the output file to this address, so make sure you enter an existing e-mail address – the link will be valid for 48 hours).
Click Submit Query. A new webpage will appear, which will be refreshed every 10 seconds, until your output appears.
The output of a Basic Daily Event Study is a text-file. You can read it more easily when you copy the text and paste it into Word (use in landscape and the largest format possible), but even then it’s not very user friendly. (Tips are welcome!)
The first part of the output consists of an overview of problems detected in the sample (for example the data is outside the range of data for the security). Here you can see if and why events were deleted.
Then you get the Parameter Estimates and Estimation Period Statistics, for each event and the mean and median.
If you marked Detail=Fulll you will also get data on the security-event level. This data is grouped: first you get the Abnormal Day Returns, then the Cumulative Abnormal Returns of the windows entered at Step 4. When you have marked different models at Step 2 you will first see the results of the Market Model for all security-events, and then the results for the other model(s).
Cumulative Abnormal Returns
If you are (only) interested in the Cumulative Abnormal Returns, there is an alternative: the Cross-Sectional Analysis – Daily event study in Eventus (this option can be found in the column at the left side of the screen in Eventus). The output of this study is a SAS-file, but you can open it in SPSS as well (Choose in SPSS ‘Open an existing data source’ and select behind Files of Type SAS). From SPSS you can export the file to Excel.
When you don’t have SAS or SPSS you can send the file to the Datateam, we can export it from SAS to Excel. In that case, make sure there are no spaces and dots in the file name.
The output looks like this (you can click on the image to enlarge it):
Please note 1: As you can see, this output has little additional information: you have to remember the windows entered at step 4 and you can’t see if and why events have been deleted. for this information you still need to run the Basic Daily Event Study.
Please note 2: When you open the output in SPSS, you’ll see less decimals for the CARs. You can adjust the number of decimals in the Variable View of the file in SPSS.