Dear FernandoRios Jeff Wooldridge
Currently I am trying to perform an analysis of the impact of credit incentives on loans' terms (interest rates, etc.). My dataset is at the loan level, and therefore each row is a loan that appears only once. The treatment is staggered in the way that some banks adopted the program in 2020Q3, other ones in 2021Q2, and so on. My dataset goes from 2017 through 2023. At the end, there are 6 banks in the program and 18 in the control. So, I think that this is a repeated cross section (am I wrong?).
With that in mind, I am running the following specification:
where term_w is the loan term in months weighted by credit, numeric_quarter is the time period and numeric_quarter_accreditation is the quarter when banks joined the program.
My questions are:
1. Is my specification well? I am not including ivar since this is not a panel, but I am not totally sure if my dataset can be classified as "repeated cross section".
2. I want to add Bank and Quarter FE, how can I do that?
3. When running the above specification I lose observations (from 97,578 to 39,349), why is this happening? is something related to the fact that numeric_quarter_accreditation is missing for those banks that are never treated? Should I recode that?
Thanks in advance for your help!!
Currently I am trying to perform an analysis of the impact of credit incentives on loans' terms (interest rates, etc.). My dataset is at the loan level, and therefore each row is a loan that appears only once. The treatment is staggered in the way that some banks adopted the program in 2020Q3, other ones in 2021Q2, and so on. My dataset goes from 2017 through 2023. At the end, there are 6 banks in the program and 18 in the control. So, I think that this is a repeated cross section (am I wrong?).
With that in mind, I am running the following specification:
Code:
csdid term_w, time(numeric_quarter) gvar(numeric_quarter_accreditation) notyet rseed(654252) agg(simple) wboot(rseed(6589652)) cluster(loanid)
My questions are:
1. Is my specification well? I am not including ivar since this is not a panel, but I am not totally sure if my dataset can be classified as "repeated cross section".
2. I want to add Bank and Quarter FE, how can I do that?
3. When running the above specification I lose observations (from 97,578 to 39,349), why is this happening? is something related to the fact that numeric_quarter_accreditation is missing for those banks that are never treated? Should I recode that?
Thanks in advance for your help!!
