Quantcast
Channel: Statalist
Viewing all articles
Browse latest Browse all 73258

Fixed effects

$
0
0
Hi guys!

I have the following problem. I would like to replicate the following xtreg , fe:

Code:
xtreg tasso_crescita_salespr L.log_sales L.dummy_2 L2.dummy_2 L3.dummy_2 mean_gr_rate_atc2 recalls_sales ageprodcat1 ageprodcat2 ageprodcat3 ageprodcat4 newmolfirm newmolmarket i.Year, fe vce(cluster idpr)
with demeaned OLS. So first of all I have demeaned the variables like that:

Code:
bysort idpr (Year) : center tasso_crescita_sales_prod, gen (tasso_crescita_salespr_demeaned) // equivalente a fare: egen tasso_crescita_salespr_demeaned = mean(tasso_crescita_s), by(idpr)
bysort idpr (Year) : center log_sales, gen (log_sales_demeaned)  
bysort idpr (Year) : center mean_gr_rate_atc2, gen (mean_gr_rate_atc2_demeaned) 
bysort idpr (Year) : center dummy_2, gen (dummy_2_demeaned) 
bysort idpr (Year) : center newmolfirm, gen (newmolfirm_demeaned)
bysort idpr (Year) : center newmolmarket, gen (newmolmarket_demeaned)
bysort idpr (Year) : center recalls_sales, gen (recalls_sales_demeaned)

*N.B. ageprodcat1 ageprodcat2 ageprodcat3 ageprodcat4 sono già dummizzate (non sono categoriche) quindi posso già farci il demeaning:
bysort idpr (Year) : center ageprodcat1, gen (ageprodcat1_demeaned) 
bysort idpr (Year) : center ageprodcat2, gen (ageprodcat2_demeaned) 
bysort idpr (Year) : center ageprodcat3, gen (ageprodcat3_demeaned) 
bysort idpr (Year) : center ageprodcat4, gen (ageprodcat4_demeaned)
The DB looks like this (just reported one category):
Code:
* Example generated by -dataex-. To install: ssc install dataex
clear
input float dummy_2 byte ageprodcat1 float(newmolmarket newmolfirm newmolfirm_demeaned newmolmarket_demeaned tasso_crescita_sales_prod)
0 1 1 .         .         .5         .
1 1 0 1         0        -.5         .
0 0 1 .         .       .875         .
0 0 0 0         0      -.125 3.1807585
0 0 0 0         0      -.125         .
0 0 0 0         0      -.125 12.165902
0 0 0 0         0      -.125         .
1 0 0 0         0      -.125         .
0 0 0 0         0      -.125         .
0 0 0 0         0      -.125         .
1 1 1 .         .   .9166667         .
1 1 0 0         0 -.08333334 18.725441
0 1 0 0         0 -.08333334 18.340553
0 0 0 0         0 -.08333334 18.435717
0 0 0 0         0 -.08333334  18.86105
0 0 0 0         0 -.08333334  19.39682
0 0 0 0         0 -.08333334 18.886513
0 0 0 0         0 -.08333334 18.683449
0 0 0 0         0 -.08333334 18.159758
1 0 0 0         0 -.08333334 19.067064
0 0 0 0         0 -.08333334  19.46246
0 0 0 0         0 -.08333334         .
0 1 0 1  .9090909          0         .
0 1 0 0 -.0909091          0 16.825748
0 1 0 0 -.0909091          0 17.012917
0 1 0 0 -.0909091          0 14.926245
1 1 0 0 -.0909091          0  15.29373
0 0 0 0 -.0909091          0  15.47137
0 0 0 0 -.0909091          0         .
0 0 0 0 -.0909091          0 13.862047
0 0 0 0 -.0909091          0 17.476244
0 0 0 0 -.0909091          0 17.279263
0 0 0 0 -.0909091          0   15.7814
0 0 1 .         .   .9166667         .
0 0 0 0         0 -.08333334         .
0 0 0 0         0 -.08333334         .
0 0 0 0         0 -.08333334         .
0 0 0 0         0 -.08333334         .
1 0 0 0         0 -.08333334         .
0 0 0 0         0 -.08333334         .
0 0 0 0         0 -.08333334         .
0 0 0 0         0 -.08333334         .
0 0 0 0         0 -.08333334         .
0 0 0 0         0 -.08333334         .
0 0 0 0         0 -.08333334 11.292548
0 0 1 .         .   .8571429         .
0 0 0 0         0 -.14285715         .
0 0 0 0         0 -.14285715         .
0 0 0 0         0 -.14285715         .
0 0 0 0         0 -.14285715         .
1 0 0 0         0 -.14285715         .
0 0 0 0         0 -.14285715         .
0 0 1 .         .   .9166667         .
0 0 0 0         0 -.08333334         .
0 0 0 0         0 -.08333334         .
0 0 0 0         0 -.08333334         .
0 0 0 0         0 -.08333334         .
0 0 0 0         0 -.08333334 14.164447
0 1 0 0         0 -.08333334         .
1 1 0 0         0 -.08333334         .
0 1 0 0         0 -.08333334         .
0 1 0 0         0 -.08333334         .
0 0 0 0         0 -.08333334         .
0 0 0 0         0 -.08333334  10.16614
0 0 1 .         .   .9166667         .
0 0 0 0         0 -.08333334         .
0 0 0 0         0 -.08333334  13.10945
0 0 0 0         0 -.08333334         .
0 0 0 0         0 -.08333334         .
1 0 0 0         0 -.08333334         .
1 0 0 0         0 -.08333334         .
0 0 0 0         0 -.08333334         .
0 0 0 0         0 -.08333334         .
0 0 0 0         0 -.08333334         .
0 0 0 0         0 -.08333334         .
0 0 0 0         0 -.08333334         .
0 1 1 .         .   .9166667         .
0 1 0 0         0 -.08333334 13.415682
0 1 0 0         0 -.08333334 13.316372
0 1 0 0         0 -.08333334   13.7403
0 1 0 0         0 -.08333334 14.663067
0 0 0 0         0 -.08333334 14.747814
1 0 0 0         0 -.08333334 14.361357
1 0 0 0         0 -.08333334 14.021648
0 0 0 0         0 -.08333334         .
0 0 0 0         0 -.08333334         .
0 0 0 0         0 -.08333334         .
0 0 0 0         0 -.08333334         .
0 1 1 .         .   .9166667         .
0 1 0 0         0 -.08333334         .
0 1 0 0         0 -.08333334  9.691856
0 1 0 0         0 -.08333334 12.495008
0 1 0 0         0 -.08333334 10.372454
0 0 0 0         0 -.08333334 12.004423
0 0 0 0         0 -.08333334         .
0 0 0 0         0 -.08333334  10.91284
0 0 0 0         0 -.08333334 11.242186
0 0 0 0         0 -.08333334         .
0 0 0 0         0 -.08333334         .
1 0 0 0         0 -.08333334         .
end

At this point however I have some doubts:
  1. my original fixed effect model contains lags. How should I treat them in the demeaned case?
  2. the dummies (dummy_2, newmolfirm, newmolmarket, ageprodcat*) represent: dummy_2 = recall of product I at time t; newmolfirm = 1 the first year that the product isappears in the firm's portfolio, newmolmarket = 1 if the molecule related to the product appeared for the very first time in the sample; ageprodcat = age of the product. Since they vary in time I have to demean them also right?
  3. Finally recalls_sales is an interaction between sales (continuous variable) and recalls (dummy). Do I have to demean also this?
  4. Are fixed effects automatically thrown away?
Thanks a lot!

Viewing all articles
Browse latest Browse all 73258

Trending Articles



<script src="https://jsc.adskeeper.com/r/s/rssing.com.1596347.js" async> </script>