Confusion with entropy measure
Posted: 15 July 2015 12:39 PM   [ Ignore ]
Total Posts:  1
Joined  2015-07-15

Hi All,

I am trying to calculate entropy measure using Palepu, Krishna (1985), “Diversification Strategy , Profit Performance and the Entropy Measure,” 6 (3), 239–55.

What’s confusing me is that how do I treat two segments with the same 4 digit codes. For example, my sample has a company with a segment named “Health Management” and another named “Professional Diagnostics”, with both having SIC code as 2835. In my calculations, do I treat the segments separately (i.e. to take pi*ln(1/pi) for both the segments) or do I add the segment sales together and treat sum of both as one segment i.e. 2835. From Palepu, what I understood is that Group is on 2 digit level while segment is on 4 digit level.

I hope the question makes sense. Looking forward to hearing from you guys!


Posted: 15 July 2015 03:42 PM   [ Ignore ]   [ # 1 ]
Total Posts:  901
Joined  2011-09-19

hi Mo,

As you point out, you could aggregate identical SIC industries and treat them as 1, or treat them as separate industries.If aggregating, you need to decide at what level (2 digit, 4 digit, etc). If I am not mistaken, the names for these versions are ‘related’ and ‘unrelated’ entropy measure. I would expect the paper to provide details as to how it is computed (if aggregating, then how, etc)

Hope this helps,



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