Eventstudy2 Command (User-Written): Missing Values in Result Files

[Leon Gehlmann]

Deal all,

I am using the user-written eventstudy2 command to calculate announcement returns. The program works excellent, however I still have two question regarding the result files.

When I open the "arfile", I notice that in some rows a value for the column with the abnormal return is missing. In addition, these rows (with the missing abnormal returns) also show missing values for the security returns ("rit") and the market returns ("rmkt"), which are however all present in my "security_returns_file" and "marketfile". Is it correct that the corresponding values in the "arfile" are set to missing anyway?

As a result (according to my impression), the CARs for some events and eventwindows are not calculated either. What can I do so that all ARs and subsequently all CARs are properly calculated?

Attached you can find the corresponding code I used.

eventstudy2 gvkey datadate using trading_file, returns(rit) car1LB(-1) car1UB(1) model(FM) marketfile(market_file) marketreturns(rmkt)  evwlb(-10) evwub(10) eswlb(-250) eswub(-50)    ///
aarfile(aar_v1) carfile(car_v1) arfile(ar_v1) crossfile(car_crosssection_v1) diagnosticsfile(diagonistics_v1) graphfile(graph_v1) delweekend datelinethreshold(0.2) shift(3) replace

I would appreciate a brief answer to my question.

Many thanks in advance.

Best wishes,
Leon

[Leon Gehlmann]

Please find attached a screenshot of my security_returns_file and one of my arfile:

security_returns_file:

arfile:

[Thomas Kaspereit]

Dear Loen,

Thank you. Please have a look at your return data. You did not choose the right decimals.

For instance, when you tell eventstudy2 that a return is -1.065, as you do for 18Aug2000 for XYZ Inc, you are effectively saying that the return on that day was -106.5%.

A return of less than 100% is impossible, which is why eventstudy2 sets it to missing.

Please divide all your return data by 100 and re-run.

Best
Thomas

[Leon Gehlmann]

Dear Thomas,

Thank you very much for the fast reply. Your suggestion has solved my problem.

Best wishes,
Leon