Dear all,
I'm currently writing my research on the effect of doing an acquisition on revenue in a sustainable sector vs a non-sustainable sector. I want to perform an event-study Difference-in-Difference model. My treatment group consists of companies that acquired in the timespan 2005-2020 and my control group consists of companies that didn't acquire in this timespan but are in the same sector.
To test the pre-trends assumption I want to perform a two-way fixed effects regression with panel data in this event study setting. My y-variable is Revenue and I want to add two fixed effects variables: Company_ID and Year.
I use relative years because the treatment occurs in different years for the treated companies (rel_year). rel_year_1, rel_year_2 and rel_year_3 show the years before the treatment occurs. The other years are the year of and after the treatment: rel_year_4, rel_year_5, rel_year_6 and rel_year_7. I want to use rel_year_3 as my reference year. I have a Post variable that shows 1 for the years after the treatment of the treated group and 0 before the treatment. This variable is 0 of all years of the control group. Treatment_group is 1 for the treated companies in all years and 0 for all control companies in all years.
I know how to perform the 'normal' DiD but I'm unsure of how to do it in an event-study setting with relative years and the Post variable always taking on the value 0 for the control group. I've tried different codes but all of them give 'omitted or empty' values for the interaction term in my regression.
My question is how to perform the two-way fixed effects regression in this event study Difference-in-Difference setting. Also, how do I check the pre-trends in this event study Difference-in-Different setting with relative years?
Any help would be of great help! Thank you in advance!
I'm currently writing my research on the effect of doing an acquisition on revenue in a sustainable sector vs a non-sustainable sector. I want to perform an event-study Difference-in-Difference model. My treatment group consists of companies that acquired in the timespan 2005-2020 and my control group consists of companies that didn't acquire in this timespan but are in the same sector.
To test the pre-trends assumption I want to perform a two-way fixed effects regression with panel data in this event study setting. My y-variable is Revenue and I want to add two fixed effects variables: Company_ID and Year.
I use relative years because the treatment occurs in different years for the treated companies (rel_year). rel_year_1, rel_year_2 and rel_year_3 show the years before the treatment occurs. The other years are the year of and after the treatment: rel_year_4, rel_year_5, rel_year_6 and rel_year_7. I want to use rel_year_3 as my reference year. I have a Post variable that shows 1 for the years after the treatment of the treated group and 0 before the treatment. This variable is 0 of all years of the control group. Treatment_group is 1 for the treated companies in all years and 0 for all control companies in all years.
I know how to perform the 'normal' DiD but I'm unsure of how to do it in an event-study setting with relative years and the Post variable always taking on the value 0 for the control group. I've tried different codes but all of them give 'omitted or empty' values for the interaction term in my regression.
My question is how to perform the two-way fixed effects regression in this event study Difference-in-Difference setting. Also, how do I check the pre-trends in this event study Difference-in-Different setting with relative years?
Any help would be of great help! Thank you in advance!
