Choice of model for panel data analysis

[Karoline Hoelgaard]

I have a question on panel data and two way fixed effects models:
I have a panel dataset on 267 danish municipalities in the time period 1978-2006, and I have formatted the data "long".
I am trying to run a regression where my dependent variable is an index including three different variables on the municipality's spending and the independent variable is the percentage of each municipality who voted for a right wing party. I want to investigate if the preferences that the citizens express through their voting behavior is listened to by the municipalities - i.e. is the municipalities responsive to their citizens?

I have been told that a two way fixed effects might be the right fit for this kind of data, but I wanted to ask if that is correct? Right now my commands look like this:
xtset muni year
xi: reg index bluevote1 i.muni i.year

I hope you can help!

Best
Karoline

[Carlo Lazzaro]

Karoline:
welcome to this forum.
Please note:
1) you probably meant:

xtset muni year
xtreg index bluevote i.year, vce(cluster muni)

or (right, but less advisable if you have evidence of a panel-wise effect):
2)

reg index bluevote1 i.muni i.year, vce(cluster muni)

In addition:
a) -regress- does not need previous -xset-;
b) -vce(cluster panelid) standard error is mandatory if you go -regress- (that is not conceived for panel data regression); otherwise, Stata will consider all your observations as independent, which is not the case due to the panel structure of your dataset;
c) with 267 -muni-, -vce(cluster panelid)- standard error is highly recommended with -xtreg,fe-, too;
d) -xi:- prefix is now redundant for Stata built-in commands.