Logit on panel data

[Drobi]

Hi everyone,

I am would like to run a logit model on my panel data (attached) with a limited variable as dependent one (here the Gini_Index) and several independent variables (e.g. GDP growth, schooling years, inflation and FC dummy). However, the Gini_index is not a discrete variables, it's values can lie between 0 and 1. Since it is a limited variable I thought the Logit model is more appropriate as a standard LS regression.
When I try to run the binary method it always shows the error that there are not enough oberservations (to many missing or non 0-1).

Any help or advice is appreciated.

Mtahias

[startz]

Logit refers to a model where the lhs variable is zero or one, so it's not what you want.

[Drobi]

thank you startz,

However, it was the only method I could came up with when considering that I have a limited dependent variable. Do you have any other suggestion how to account for such variables?

[startz]

You might try using

Code: Select all

log(Gini_Index/(1-Gini_Index))

as your dependent variable