Dear ,
Many thanks for your helpful comments regarding my patent data. Regarding your last comment, I guess in my enthusiasm I forgot to follow-up on your advice.
Besides the patent data, I am also working on a dataset based on accounting data, which already includes another innovation proxy: R&D expenditures.
While exploring the first steps in the analyses, I find myself wondering why, in the FE estimation as well in Pooled OLS, one extra indicator value is omitted in the regression due to collinearity, e.i., I specified the country base-level as NL, but also FR is omitted. The same goes for the year dummy: I set 2005 as base level and the regression omits 2012 additionally to 2005. Could you advice me on the reason for this omission?
I used:
Best regards,
Many thanks for your helpful comments regarding my patent data. Regarding your last comment, I guess in my enthusiasm I forgot to follow-up on your advice.
Besides the patent data, I am also working on a dataset based on accounting data, which already includes another innovation proxy: R&D expenditures.
Code:
* Example generated by -dataex-. To install: ssc install dataex clear input long ID double(Turnover RDexp TA empl) long sector float(survivor POST NL Post_NL) 1 5445000 0 5831000 16335 12 1 0 0 0 1 6083000 0 7300000 18180 12 1 0 0 0 1 6052000 0 7325000 17833 12 1 0 0 0 1 5982000 0 7782000 17371 12 1 0 0 0 1 6074000 0 7450000 16804 12 1 1 0 0 1 7035000 0 8511000 17371 12 1 1 0 0 1 6418000 0 8312000 15788 12 1 1 0 0 1 6458000 0 8243000 15859 12 1 1 0 0 2 . . . . 12 . 0 0 0 2 . . . . 12 . 0 0 0 2 . . . . 12 . 0 0 0 2 2254200 0 2668700 35313 12 0 0 0 0 2 2255900 0 2580000 34180 12 0 1 0 0 2 2295700 0 2473500 33616 12 0 1 0 0 2 2355700 0 2402900 32 12 0 1 0 0 2 2407300 0 2231200 29922 12 0 1 0 0 3 9133 0 1164018 75 11 1 0 0 0 3 13720 0 1327425 75 11 1 0 0 0 3 532497 0 1837167 3611 11 1 0 0 0 3 851452 0 1520790 4158 11 1 0 0 0 end label values ID ID_bvd label def ID_bvd 1 "BE0202239951", modify label def ID_bvd 2 "BE0214596464", modify label def ID_bvd 3 "BE0220324117", modify label values sector sector label def sector 11 "Other services", modify label def sector 12 "Post & telecommunications", modify
While exploring the first steps in the analyses, I find myself wondering why, in the FE estimation as well in Pooled OLS, one extra indicator value is omitted in the regression due to collinearity, e.i., I specified the country base-level as NL, but also FR is omitted. The same goes for the year dummy: I set 2005 as base level and the regression omits 2012 additionally to 2005. Could you advice me on the reason for this omission?
I used:
Code:
reg Turnover RDexp TA empl sector survivor POST NL Post_NL i.year ib(4).country i.sector, robust cluster(ID)
Best regards,