Dear Researchers,

I am using a panel data model for my study on firms in financial services in India (totally there are 24 firms, 4 of them are in government category and 20 in private category), My panel data set is for the period 2005-2020.

Since I am using the entire 24 companies my sample equals population.

Theoretically it is said that if your observation cannot be described as a random sample you must use Fixed effects model.

My questions in the regard are:

a) Am I logically correct to apply Fixed effect model because I study the entire population of companies i.e. choosing all 24 companies together.

b) What if I study them separately by categorizing them in Government and Private sector, then also the sampling won't be random and I will have to choose the Fixed effect model.

Request you to kindly share your views.

P.S. - The government firm panel data is balanced whereas the private firms is unbalanced.