Here you have an example of recente research in this area;

Article Inequality and development: is the Kuznets curve in effect today?

Basically, people regress the Gini index on the natural logarithm of GDP per capita.

Older research (among may others)

http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.194.8608&rep=rep1&type=pdf

https://www.econstor.eu/bitstream/10419/109529/1/wp-011.pdf

Hi Kostas Simoglou, this is a kinda broad question. There are so many papers on this topic, you can find many different approaches. I will give you a few notes. I will assume that your aim is to examine the impact of inequality on economic growth (GDP).

1) For the independent variable you can use the Gini index or income share (https://wid.world/). There is even possible to try different income groups and test which group affect economic growth most.

2) For the dependent variable: GDP per capita (constant 2010 US$) - WorldBank database. You can also use growth variables, but be cautious about results interpretation.

3) Don't forget to use some control variables (unemployment, etc.)

4) Try different models with different approaches and analyze the difference in variables significance. (Robustness tests), try lagged (t-1) variables to mitigate endogeneity issue etc.

I am really grateful for your replies, as I am really stressed right now. Both the answers so far helped me clear some things in my head.

Another issue that I have is that in the beginning I wanted to test the relationship between regional inequality (across 50 states) and economic growth. The thing is that I can't find state-specific data (as most of them are only short-term) and I decided to go for national data. Do you think it's decent enough for a bachelor thesis?

Also, @Miroslav Navrátil, for the control variables I am thinking of using Average Income , education, human capital and crime rates, as I see them a lot in other papers. Do you have any suggestions for them?

It is hard to offer you the best answer, depends on your econometric techniques level. Environmental Science and Pollution Research is a good place for your ref.

I don't think such endogenous question can be solved by single regression. Maybe

structural equations can do this better.

Kostas Simoglou

To your question about appropriate control variables. It is common practice to support the choice of control variables through studies (papers). So find some papers which analyze determinants of economic growth. And the most significant variables from these research papers you can use in your bachelor thesis (with citations). It is also suitable to mention in your thesis conclusion that your variables act as other studies point out.

I have to mention that the use of a control variable can lead to lower significance of your main variable (inequality) which is OK, you need to test it and from it draw a conclusion.

What research question are you trying to answer:

1.how does income inequality effect Economic growth or

2. how does Economic growth effect income inequality?

From your original post its 1, from some of the answers its 2. These are 2 totally different research questions.

Then, whatever the right research question is, when you know, you should have a clear knowledge of what is your dependant and independant variable.

then you right that Y(dependant) = funcion (x independant) + error

so if 1. you have that

Y(inequality)= function (Economic growth-x)+ error

You would add to the right side (add to the model) gdp per capita, and variables like maybe immigration, population growth, school level attainment, training of workforce, or others variables that from literature or theory the

at are maybe linked to inequality besides Econ growth.

- you would try to describe how when taking everything else into account Econ growth influences inequality and the result of the regression maybe will be: that if econ growth goes up by 1%, inequality goes maybe up/or down by 0,5%, with everything else beiing fixed (from literature: ceteris paribus)- taken into account

If research question i s 2.

Y(Economic growth)= function (inequality) + error

Thus stated it is clear that in this equation has more "truth" describing Economic growth in error then in independant variable (inequality) so you would add domestic consumption, employment, foreign trade surplus, ...things that make up Econ growth, and say with everything else being fixed (from literature: ceteris paribus)- taken into account 1% increase in inequality leads maybe to 1.5% less econ growth.

Good luck!

I think the real GDP per capita woulld be the best choice! In this case you can determine the trend of the changing of inequality for the anylized period of time.