Incorporating the square of GDP in the EKC allows for this non-linear relationship to be captured more accurately.
Multicollinearity, a term used to describe the presence of high correlation among the independent variables, has been a longstanding issue in econometric modeling. However, despite its adverse effects on regression results, the use of GDP and GDP squared in the Environmental Kuznets Curve (EKC) has remained a popular practice.
We see alot of published papers that have high VIF/multicollinearity since the model include both GDP per capita and GDP per capita squred terms.
See further useful discussion
https://agris.fao.org/agris-search/search.do?recordID=CH2018103372
https://dergipark.org.tr/en/download/article-file/1792935
Article On the Estimation of the CO2 Emission, Economic Growth and E...
Article Handling multicollinearity in quantile regression through th...
My ultimate question will be; is it justified to use both GDP and GDP square in one regression?
Is it ever justified to use both x and. X**2 in the same regression.the answer is obviously yes. You have to take more than very simple ideas into account. David Booth
- Badges
- Science topic