Application of OLS Regression and VAR Models to Analyse the Economies of Varying Political Regimes

Abstract

The GDP per capita is a popular method of measuring the economic success of a country. This paper uses regression analysis to predict the GDP per capita (GDPPC) of countries using different independent variables. We applied Ordinary Linear Squares Regression and Vector Autoregression to check for a correlation between the chosen independent variables (Corruption Perception Index, Political Rights score, Civil Liberties score, Gender Inequality Index, Consumer Price Index, Population Density, and the percentage of people using the Internet) and the GDPPC. Using empirical evidence, we determine which model might be more accurate to attain this goal. Four countries of varying political regimes are studied - USA and Canada are categorised as democratic countries and China and Russia are non-democratic countries. Our results show trends in the correlations between the independent and dependent variables, and we can draw a distinction between the political regimes.  We found that Corruption Perception Index and Population Density negatively correlates with the GDPPC of all 4 countries. We also noticed that the percentage of people using the internet and Gender Inequality Index correlates negatively with the GDPPC for non-democratic countries and in democratic countries the Consumer Price Index negatively influences the economy.