st: Calculations in panel data

[Andreas Dall Frøseth <Andreas.Froseth@stud.nhh.no>]

Dear Statalisters.

I'm new to stata, and apologize if these questions appear to be way to basic.

I'm currently working with several databases (one for each year from 1992-2010) each containing company and accounting number (balance sheet and income statement) data for approximately 200 000 companies (observations differ from year to year).

In short terms, my objective is to combine these sets, and categorize each company in to groups, based on calculations for growth and ROA (return on assets). Ie. a company with growth and ROA above industry average would be categorized as a "star"-company. And then use these to follow the transitions of companies trough-out the periode, in an transition-matrix kind of way. The main objective is to test a hypotesis that a company categorized as ie a "value" company one year is more likely to become a "star" company the consecutive year, than a growth company.

My question is therefore:

Because my calculations would require both sales-numbers, as well as asset-numbers, for two consecutive years (sales growth would be calculated by the formula: (sales2010-sales2009)/sales2009). Will the best solution be to merge the sets for 2009 and 2010 or append the sets? I've been looking into the xt-command for panel data in STATA, and was hoping to use these during my analysis. If I append the sets, how will i then be able to do the calculations (lets say i'm defining xtset orgid year), and create new variables with the results? I need to do the calculations both for each company each year, as well as on an industry level (to be able to compare a company to it's industry mean).

Best regards
Andreas

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/