Hello Statalist team, may I ask you the following qustion?
I would like to use the DID with propensity score matching in the following settings:
1) Pre treatment period = 2017, After = 2018
2) I have the panel datasets both in 2017 and 2018 for bank level.
3) I would like to major the average treatment effect of a policy occurred in 2018 on a bank deposit.
I use the following STATA code:
"teffects psmatch (Delta_Deposit) (treatment var1 var2 time) if fiscal_year == 2018"
treatment = 2018*treated dummy (bank level)
In this case, do I need to add the regression period of 2017 in the code?
So the correct code could be:
"teffects psmatch (Delta_Deposit) (treatment var1 var2 time) if fiscal_year == 2018 | fiscal_year == 2017"
In fact, I tried both but there is not a big difference between "if fiscal_year == 2018" and "if fiscal_year == 2018 | fiscal_year == 2017".
If I use only the year of 2018, matched sample should be selected based on the year of 2018?
But if I use the both years of 2017 and 2018, the matched sample should be selected based on the years of 2017 and 2018?
Thus in theory, propensity score matching should select the matched sample based on the treated sample meaning only to use 2018 could be the right answer.
Is this understanding correct?
Thank you for your advice in advance.
Best regards;
0 Response to Difference in Difference with propensity score matching
Post a Comment