Herfindahl-Hirschman Index

[TANVEER Khan]

Hello all,
I am working on a panel data of 200 firms from 1995-2019. Data is for a single industry. When I calculate the HHI from market shares, I get a time series of HHI for a 25 years and not for all firms to be used in panel. I tried to use the value of HHI for a particular year same for all firms. Is this the right approach? Is there any method by which I can get HHI for each firm so that I can run a meaningful analysis.
with regards

[ashar awan]

Dear HHI is used for measuring market power, and you got correct HHI for 25 years, I hope you will find this article useful

An institutional framework for the development of sustainable and competitive power market in Pakistan

[TANVEER Khan]

Thank you sir for your response, I went through this paper. Alas! It didn't help. Let me put it here with an example.what my concern is.
Firm year ms
1 1995 .34
1 1996 .29
1 1997 .14
2 1995 .4
2 1996 .3
2 1997 .15

Now I want to calculate HHI which is equal to sum of squares of market shares(ms). When I calculate it in excel, I summate the market shares of all firms in a particular year to get HHI. Doing same thing for all years, finally I get a series of HHI values for each year. So in a way I converted my panel data of ms into a time series data of HHI. However I need a panel of HHI to run the panel analysis. I have tried to use HHI value for a particular year same for all firms in that year in order to get panel data back for HHI. Can it be done?
I hope you got my worry

[ashar awan]

I am also interested in the answer now after understanding, if you get the answer, please share it here on SATAlist forum too.

[Nick Cox]

I can't see a data example where market shares add to 1 across the relevant subset.

[Phil Bromiley]

If you search Nick Cox's responses in the last week or so, you will find a very elegant way to calculate what I think is the equivalent of your HHI. It uses by conditions with egen and you can readily add years to the by conditions.

[Anuradha Saikia]

TANVEER Khan this thread might be useful for you. though am very late.
https://www.statalist.org/forums/for...irschman-index