On the Latin America Evidence for the Environmental Kuznets Curve: A Two-Stage Approach using K-means Clustering and Polynomial Regression

Abstract

This paper uses a two-stage methodological approach to analyze the existence and shape of the Environmental Kuznets Curve (EKC) hypothesis across 19 Latin American countries from 2000 to 2020. First, K-means clustering is used to categorize countries based on their economic development, environmental efficiency, energy sustainability and technical efficiency characteristics. This clustering reveals five distinct groups: transitioning economies, resource-dependent economies, industrial economies, early transition economies, and advanced economies. Subsequently, polynomial regressions are estimated within each cluster to analyze the specific EKC relationships. The results demonstrate significant heterogeneity in the growth- environmental relationship across clusters, with coefficients of determinations ranging from 0.20 to 0.94. While some countries exhibit traditional inverted U-shaped relationships, others show different patterns, including U-shaped and linear relationships. The turning points at which environmental degradation (CO2 emissions) begins to decrease vary substantially, from 1,722 USD in Panama to 241,024 USD in Venezuela, in per capita terms, highlighting diverse development trajectories. The empirical findings obtained suggest that the EKC relationship in Latin America is more complex than traditionally assumed and varies significantly based on countries’ development stages and structural characteristics. These results have important implications for environmental policy design, suggesting the need for cluster-specific approaches rather than one-size-fits-all solutions.