Confronting collinearity in environmental regression models: evidence from world data

摘要: Despite the evidence, correlation between environmental impact factors has mostly been neglected in econometric models or treated with traditional methodologies such as ridge regression, which are recommended when goal is prediction and estimated parameters not interpreted causal effects. This paper addresses existing collinearity alternative methodologies, only to mitigate problem mechanically, but also isolate effects of main objective designing better policies for countries. The applied analyze CO $$_2$$ emissions 114 countries covering thirteen most recent years available data, results from empirical methodological perspectives compared. treatment residualization raise regression procedures allows researcher obtain a global vision relationship different affecting emissions, thus reaching conclusions those methodologies.