Propensity score matching by year and industry

[Hoyoung Kim]

Hi STATA Masters,

I wondering if there is any way to select control samples by year and industry when running psmatch2?
I need to select control samples based on the year and industry of treatment samples.

My current codes are like this....

qui logit dummy x1 x2 x3, robust cluster(ind)
predict pscore if e(sample)

psmatch2 dummy, pscore(pscore) neighbor(3) caliper(0.10) common

However, this codes result in 1:3 match based on pscore of x1, x2, and x3 regardless of same industry and year.
Could you help me add additional conditions for same industry and year for control sample selection in psmatch2 procedure?

Best,

[Phil Bromiley]

You could loop over years and industries and set the pscore to missing if it isn't in the right industry/year. That will be slow but would probably work.
There are many matching tools in the treatment documentation for recent Stata versions. You might find they are more flexible than psmatch2.