SIC is the industry classification. The definition of SIC from Bigelli, 2006 is

Due to the low numbers of firms in some industries, we use the most detailed industry (SIC4) only if the number of companies in the industry was greater or equal to 3, otherwise we use the less detailed industry (SIC3) or the least detailed (SIC2).

In my sample, I want to use industry*year fixed effect. What I should do to justify that SIC2 or SIC3 is more suitable in my fixed effect ( I am working on an international study and my sample lasting 20 years)? Because in SIC2, is it too general? And in SIC3, maybe we face too many missing variables? I saw a couple of studies using SIC3 fixed effect but SIC2 (e.g. Dasgupta, 2019)

From table IA5 of this paper, they did the same thing with me and the least number of observation for their SIC3 is 36 ( but they did not say that clearly in their manuscript)



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.