Compustat Segments and SFAS 131 - How to eliminate duplicate records?
Posted: 16 January 2013 09:01 AM   [ Ignore ]
Total Posts:  17
Joined  2013-01-12

Dear Forum,

I am using Compustat Segments Data from the period 1978-2010. However, in 1997 the SFAS131 replaced the SFAS14 and firms changed their segment reporting.

I want to account for this and, besides others, want to make sure that I only use the non-restated Segment data. My idea was to match the variables “datadate” and “srcdate” and eliminate all records for which These two variables do not have the same value, e.g. srcdate lies after the datadate because a company restated segment data on year t=1 in t=2.

What are your thoughts on that issue? Thx

Posted: 17 January 2013 08:56 AM   [ Ignore ]   [ # 1 ]
Total Posts:  901
Joined  2011-09-19

hi Chris,

What is most commonly done, is use an indicator variable to identify which firm-years are under SFAS 14 vs 131. You can run the same analysis for the two separate samples to see how much it affects the coefficients and then group the samples and include the indicator variable, and possible interactions of that variable with other variables of interest.

The distinction between datadate and scrdate is relevant regardless the standard. When I use the data, I require these to be the same, i.e., I use the info as it was disclosed in the year of the annual report. (That way, revisions in later years are ignored). I would think this would work for you as well (you wouldn’t have to worry about disclosures in 1998 with restated 1997 numbers.

Hope this helps,



To reply/post new questions: Please use the group WRDS/SAS on Google Groups!