I am estimating the following DiD model: log(Credit)=B1 BankExposureMeasure*Post+B2 BankSize*Post+Firm-Time FE+Firm-Bank FE
Here Bank Exposure Measure is a time-invariant, continuous treatment measure and I control for Bank Size by using a time invariant measure (pre-event) measure of bank size and interacting it with post which is a dummy for the post treatment years.
I want to run a dynamic DiD for this specification to check for pre-trends. Is it okay if I just interact the treatment measure with a year dummy and continue interacting Bank Size with the Post indicator or do I need to interact the bank size also with the year dummy?
Basically, is running this specification wrong: log(Credit)=B1 BankExposureMeasure*i.Year+B2 BankSize*Post+Firm-Time FE+Firm-Bank FE when I want to estimate a Dynamic DiD to check for pre-trends. If it is wrong, can anyone explain why?
Should I be estimating log(Credit)=B1 BankExposureMeasure*i.Year+B2 BankSize*i.Year+Firm-Time FE+Firm-Bank FE instead?
Thank you.
Here Bank Exposure Measure is a time-invariant, continuous treatment measure and I control for Bank Size by using a time invariant measure (pre-event) measure of bank size and interacting it with post which is a dummy for the post treatment years.
I want to run a dynamic DiD for this specification to check for pre-trends. Is it okay if I just interact the treatment measure with a year dummy and continue interacting Bank Size with the Post indicator or do I need to interact the bank size also with the year dummy?
Basically, is running this specification wrong: log(Credit)=B1 BankExposureMeasure*i.Year+B2 BankSize*Post+Firm-Time FE+Firm-Bank FE when I want to estimate a Dynamic DiD to check for pre-trends. If it is wrong, can anyone explain why?
Should I be estimating log(Credit)=B1 BankExposureMeasure*i.Year+B2 BankSize*i.Year+Firm-Time FE+Firm-Bank FE instead?
Thank you.
