Comparison of cross sectional entities in panel data

[Vijay Singh]

Hello
I am new to stata and stuck at one point. I am writing a paper on entrepreneurship wherein i have taken a panel of BRICS nations for the time period 2005 - 2017 (Brazil, Russia, India, China, and South Africa). I want to make a comparison among the countries, ie. which country is more positive in terms of entrepreneurship. The dependent variables are business density score and independent variables are corruption and economic freedom. When doing OLS in stata it gives overall result of the panel but not for individual country. i want to see if the entrepreneurial activities in india are more than as compared to other countries. or to rank countries from the panel.

Is it possible to do this analysis in stata.

Please help

Thanks
Vijay

[Carlo Lazzaro]

Vijay:
welcome to this forum.
You should take a look at -xtreg,re-.
If you're not familiar with panel data econometrics, reading a textbook on this topic is strongly recommended.
Stata listers dealing with this stuff are usually referred to https://www.stata.com/bookstore/micr...metrics-stata/.

[Vijay Singh]

Thanks for the reply

I have used that command but it doesn't give country specific results. it just gives overall result of the panel

Can i create dummy variables for each country (i have already given an ID to respective country) and then perform and LSDV

What do you think on that????

Regards
Vijay

[Carlo Lazzaro]

Vijay:
please act on one of the most helpful FAQ recommendations: post what you typed and what Stata gave you back (you can even top off this first recommendation with sharing an excerpt/example of your dataset via -dataex-).Thanks.

[Vijay Singh]

@ Lazzaro Sir

I have gone through the book Microeconometrics Using Stata by Cameron and Trivedi (2009), Chapter 8, Page No 270
Here the respected authors have mentioned Separate Regressions and have advised for using command: statsby
so if i do the same on my dataset which contains 5 countries it gives regression output for each country seperatly. but now the question is that are the results valid as time span is only 13 years for each country. I haven't checked for stationarity of the dataset. does that makes results unreliable.

Kindly give your suggestions

Thanks & regards
Vijay Singh

[Carlo Lazzaro]

Vijay:
since you're seemigly dealing with a short N, long T panel dataset, why not considering -xtgls- and adding -i-country- in the right-hand side of your regression equation?