Regression on Gdp Per Capita – Cross-Sectional Data Analysis Paper
Statistics and Econometric refresher course
Cross-Sectional Data Analysis Paper
Eva Sloff
12-09-2012
1 Introduction
The purpose of this paper is to describe a model concerning the GDP per capita (Gross Domestic Product per capita) and the possible relation between three other economic variables that are available of different countries. I am looking at the GDP per capita because it measures the national output divided by the population and therefore it is a relevant measurement for comparing one country to another and it shows the relevant performance among different countries. Furthermore I am looking at three different random economic variables, which I have chosen because I expect that they have an influence on the GDP per capita. The three variables are respectively; Total Unemployment, Total Expenditures on Education and Total Expenditures on Health. The aim is to see whether there is a significant relationship between one or more of the explanatory variables and the dependent variable. We are going to look at a simple way of establishing the relation between the 3 explanatory variables and the dependent variable through a linear regression of the GDP per capita. After that we are going to check the CLRM assumptions, to see whether they are correct. If some of the assumptions are not correctly made we may exclude one of the explanatory variables to solve this problem.
2 Data
I am using cross-sectional data to examine the impact of unemployment, expenditures on health and expenditures on education, on the GDP per capita. All the data I am using is from the World Bank database. I am looking at year 2007 because that specific year had the most recent and corresponding data. The dataset contains 73 countries.
Figure 1: Scatterplots
In this figure you see three scatterplots which show the relationship between the dependent variable GDP per capita and the three independent variables.
The first scatterplot shows a relationship between the GDP per capita and the total unemployment. The relationship is negative as I expected, because when the total unemployment goes down, the GDP per capita logically goes up. The R2 explains how well the regression model actually fits the data. The Total Unemployment variable explains 13 percent variations in the dependent variable about its mean value. The second scatterplot shows a small positive relationship (as I expected) between the GDP per capita and Expenditures on Education and it explains 14 percent variations in the dependent variable. The positive relationship explains