Computing HHI with Compustat annual/quarterly

[dodo]

Hi. I am trying to compute annual time series of HHI based on sales in each 4-digit SIC industries, but I am having some difficulties in choosing which database/method would be best for my purpose. Specifically, I was wondering which databases I have to choose from: Compustat annual, Compustat quarterly, Compustat-CCM annual, and Compustat-CCM quarterly.

About Compustat annual and Compustat CCM annual, I think that it would be problematic to compute HHI based on fiscal year, because of the differences in fiscal year end, right? For example, suppose that I compute HHI for fiscal year 2015 with two firms 1 and 2 whose fiscal years end June and May respectively. Then the firm 1's sales occur between July 2014 and June 2015 while the firm 2's sales occur between June 2015 and April 2016, a totally different periods. Is this concern valid? If it is a valid concern, do you think that it is OK to use Compustat/Compustat-CCM quarterly to extract sales and adjust inflation, then use the quarterly time series to compute the annual real sales?

Also, I have some difficulties in deciding between Compustat and Compustat-CCM. What's the main difference between those two besides that some Canadian/Bahaman countries not listed in US markets are not included in Compustat-CCM? My main focus is US market, and I will subsequently use HHI to analyze CRSP. Should I just use Compustat CCM?

[dodo]

Also, I was wondering why there are significant differences in SIC code found in CRSP and CRSP-CCM. 4-digit SIC in CRSP (especially in the past) seems much more detailed than CRSP-CCM. Were there changes in SIC reporting or something?