Dear all,
I'm trying to run a DiD model, to estimate the effect of a policy that raises retirement age
The data is perfectly balanced from 2007-2018 on aggregated agegroups, age 30-59 and age 60 and over.
The outcome variabel im using, is days away from work (due to personal illness). The variabel is measured as the mean value within each agegroup for each year.
The impact of the policy starts 1. January 2014. Before the policy, the retirement age was 60, whereas the reform raises the age to 60,5 in 2014, 61 in 2015, 61,5 in 2015 and so on until 2018.
Agegroups:
agegrp=1 if age>=60
agegrp=2 if age<60
Outcome variabel: daysoff
To estimate the effects of the policy on "days off" among the elderly (age 60 and older), my first intuition was to set up the model like this:
[xtset agegroup year]
[gen logDO = log(daysoff)]
[gen post=0]
[replace post=1 if year>=2014]
[gen treat=0]
[replace treat=1 if agegrp=1]
[xtreg logDO treat##post i.year, r ]
My first question: Is this correctly done?
Secondly, can I in any ways estimate how the treatment effect develops after the reform, due to the increasing retirement age?
etc. have multiple treatment periods? or include leads and lags?
Many thanks in advance!
I'm trying to run a DiD model, to estimate the effect of a policy that raises retirement age
The data is perfectly balanced from 2007-2018 on aggregated agegroups, age 30-59 and age 60 and over.
The outcome variabel im using, is days away from work (due to personal illness). The variabel is measured as the mean value within each agegroup for each year.
The impact of the policy starts 1. January 2014. Before the policy, the retirement age was 60, whereas the reform raises the age to 60,5 in 2014, 61 in 2015, 61,5 in 2015 and so on until 2018.
Agegroups:
agegrp=1 if age>=60
agegrp=2 if age<60
Outcome variabel: daysoff
To estimate the effects of the policy on "days off" among the elderly (age 60 and older), my first intuition was to set up the model like this:
[xtset agegroup year]
[gen logDO = log(daysoff)]
[gen post=0]
[replace post=1 if year>=2014]
[gen treat=0]
[replace treat=1 if agegrp=1]
[xtreg logDO treat##post i.year, r ]
My first question: Is this correctly done?
Secondly, can I in any ways estimate how the treatment effect develops after the reform, due to the increasing retirement age?
etc. have multiple treatment periods? or include leads and lags?
Many thanks in advance!
