Analysis of World Economic Growth Using Panel Data

In this paper, I consider panel data analysis of world economies to identify the relationship between GDP growth and the independent variables, and the nature of effect that may exist. Annual data for 161 countries from 1990 to 2020 sourced from the World Bank and UN are used. GDP is the independent variable while the independent variables are population, gross value added, total natural resources rent, and labor force. Firstly, a poolability test is performed to test for the joint significance of the fixed effects. The null hypothesis is rejected as there exist significant individual effects. Secondly, Huesman’s test is performed to test for consistency of fixed effect and random effect. The null hypothesis is rejected implying significant random effects exist. Thirdly, Bruesch-Pagan test for heteroscedasticity is performed, which shows the existence of heteroscedasticity. Finally, White’s covariance matrix estimator for random effect is performed which results to consistent and efficient parameter estimate in the presence of heteroscedasticity. The random effect model is statistically significant at the 5% level, with 68 % of the variation being explained by the model. All the explanatory variables are statistically significant at a 5 % level. They all have a positive effect on GDP with the labor force and population having the highest effect. The random effect is large and significant with a variance of 2.25±1.5 and 94.1 % share of the error component. Countries that have had consistent increases in either labor force or population growth or both, has over the period experienced consistent economic growth after controlling for the other variables and the random effect.