I am analyzing an unbalanced data set. I have to do propensity matching to see the treatment effect and I need to control for firm fixed effects and industry-year fixed cohort effects to allow for variation within an industry to vary with time. Anyone know?
Maria Aguilera , you can manually create fixed effects, as fixed effects are only (a matrix of) dummy variables. So you can create a binary matrix of firms and a binary matrix of industry-year, and then add these matrices to your model. Be careful about multicollinearity.
Thank you very much!!! hough I have no idea how to do that in rstudio but i am sure i can find a way!Thank you
Maria Aguilera , it is quite easy indeed. Please see below:
https://cran.r-project.org/web/packages/fastDummies/vignettes/making-dummy-variables.html
Thank you so much!! The problem is I need to put industry--by year fixed effects and firm fixed effects...I don't understand how to do that.. three dummy variables?
You might create an interaction between the industry and year
https://stackoverflow.com/questions/56710380/equivalent-for-statas-egen-group-function
Then, you create a dummy for each of the industry-year categories.
You would have a set of dummy variables for each industry-year (nº of industries*nº of years) pair and a set of dummy variables for firms (one for each firm)
- Badges
- Science topic
- Similar topics
- Biological Science
- Histology
- Fixatives