Hello ,
I run a model in the form of a DID using reghdfe .
I look at the impact of a crime shock on my education outcome.
I have seen an earlier paper that uses a lagged dependent shock variable if there is a dynamic relationship that is expected (https://www.journals.uchicago.edu/do...ournalCode=jpe)
(1)I have included my lagged dependent variable as a control variable.My time period is 8 years.Is it okay to use this model?
I cannot run a differenced GMM in this context as my data is not longitudinal at the student level, I can only run GMM at the district level.
(2)So for comparability, I should aggregate the results of my fixed effects model at the district level and then present them with my GMM?
I run a model in the form of a DID using reghdfe .
I look at the impact of a crime shock on my education outcome.
I have seen an earlier paper that uses a lagged dependent shock variable if there is a dynamic relationship that is expected (https://www.journals.uchicago.edu/do...ournalCode=jpe)
(1)I have included my lagged dependent variable as a control variable.My time period is 8 years.Is it okay to use this model?
I cannot run a differenced GMM in this context as my data is not longitudinal at the student level, I can only run GMM at the district level.
(2)So for comparability, I should aggregate the results of my fixed effects model at the district level and then present them with my GMM?
