I have the following task. I want to estimate the effect of winning a particular auction on a firm's growth. For each auction, I observe the winner and second firm which I plan to use as a counterfactual. I build an example of how it may look. I have firmid which is unique for each firm, contractid which indicates a particular auction, year, rating which shows who was the first and second for an auction, artificial variable yprofit that shows firm's profit, d1 is a dummy = 1 if it is winner. Below is an example of a dataset.
I tried to implement it in this way:
bysort contractid: gen d1 = 1 if rating == 1
replace d1 = 0 if d1 ==.
replace d1 =. if rating ==.
xtset contractid
reg yprofit d1
I also tried in this way:
bysort contractid: reg yprofit d1
Actually, I have doubts about the correctness. Could you please suggest me the right way how to proceed with it?
firmid contractid year rating yprofit d1
1 | 111 | 2010 | 1 | 1000 | 1 |
2 | 111 | 2010 | 2 | 800 | 0 |
3 | 222 | 2011 | 1 | 1300 | 1 |
1 | 222 | 2011 | 2 | 1100 | 0 |
2 | 333 | 2012 | 2 | 1000 | 0 |
3 | 333 | 2012 | 1 | 1400 | 1 |
1 | 444 | 2013 | 2 | 1300 | 0 |
2 | 444 | 2013 | 1 | 1100 | 1 |
2 | . | 2011 | . | 900 | . |
3 | . | 2013 | . | 1500 | . |
3 | . | 2010 | . | 1200 | . |
1 | . | 2012 | . | 1200 | . |
0 Response to Treatment effect within a group. Implementation in Stata.
Post a Comment