psmatch2

[Faiza Zafar]

Hi everyone,

I am trying to run psmatch2 for my PhD paper and I am stuck at a few points. Would really appreciate any help I can get

The problem:

I want to match my treated firms based on ROA, ROE and industry from t-1 where t= reference period of policy change.

The current command i am using is

 psmatch2 treated ROA ROE MTB INDUSTRY if year==2019, out(TF)   mahalanobis(ROA ROE MTB INDUSTRY) caliper(.01) logit

The results that i get reduces my number of treatment firms from 20 firms to 7 firms.

I am unable to understand why is it doing that.

MANY thanks in advance!

[Øyvind Snilsberg]

for 13 of the treated firms, there are no control firms with pscore within the caliper. if you want to keep all of the treated firms, you need to increase the caliper or take out the option.

[Faiza Zafar]

yes that works! Thank you

Just to check if i have written the command correct:

I am trying to do psmatch2 on treated based on control variables ROA , ROE , MTB and industry from t-1 ( in my case 2019) .

I know i have specified mahalanobis (ROA ROE MTB INDUSTRY) but not specified year 2019 in the brackets .

Is it only matching based on ROA ROE MTB INDUSTRY and not the year?

Note: I dont think stata allows to specify year in the brackets for the mahalanobis

[Øyvind Snilsberg]

you are matching firms based on the 2019 values of the variables roa, roe, mtb and industry