Hello everyone,

I have been having problems with gathering all the information that I need for my study and also, some problems with fixed effects.

1) First of all, I am interested in comparing exports of Panama to certain pair countries (of Central America and some other). I am using panel data in stata (1994-2017) and variables of interest are:

1) Exportations (y)

2) Distance of the capitals

3) GDP (c_origin)

4) GDP (destination_c)

5) Population (c_origin)

6) Population (destination_c)

-->+ some dummy variables

7) common language (0,1)

8) borders between countries (0,1)

9) whether the destination country belongs to Central America (main variable of interest; 0, 1)

Is there any page where I could get most of the data? I used CEPII and could find data until 2015 for distance, language, population and GDP (but only the current one). ***I would need data for 2016 and 2017 and also a help of GDP deflator and exportations of Panama to those specific countries in a period of 1994-2017. How do I convert the GDP measured in current dollars with a help of GDP deflator and are there any pages where I already have those values? *** I first used GDP measured in current dollars and with the data I had, the outcome did not turn out okay.

2) Another problem I encountered when doing a hierarchical regression with my first "faulty" database was that, when including fixed effects of the countries, all of the variance was already explained by countries so there was no variance left for my variables of interest to explain. On the other hand, if I do not include it-the model will be biased. What is more, without any fixed effects, in some of the regression blocks I get counter intuitive correaltions/b coefficient values. Therefore, my questions are (apart from the *** sentence marked above):

a) How to solve this problem so my model is not biased and so my variables still explain significant variance of the y (exportations) variable?

b) Which fixed effects should I add to the model? Country, distance, population... which ones?

This is the first time I am using econometrics and fixed effects so your help would mean a lot!

Thank you in advance!

Similar questions and discussions