Using Machine Learning Techniques to Test the Load Capacity Curve Hypothesis: A Classifier-Lasso Application on Global Panel Data

This study investigates the Load Capacity Curve hypothesis to analyze the relationship between GDP per capita and the Load Capacity Factor using panel data from 147 countries over the period 1995–2018. The study employs the classifier-Lasso method, developed by Su, Shi, and Phillips (2016) to address slope heterogeneity. This approach groups countries with similar economic and environmental characteristics, allowing for the estimation of group-specific coefficients. The findings reveal that developed countries exhibit a U-shaped relationship between GDP per capita and environmental outcomes. In the early stages of economic growth, environmental degradation occurs due to industrialization and resource-intensive activities. However, once a certain GDP per capita threshold is surpassed, environmental quality improves as these nations adopt cleaner technologies and implement stricter environmental regulations. In contrast, developing countries display an inverted U-shaped relationship. As GDP per capita increases beyond a specific turning point, industrial expansion and inadequate environmental management lead to significant environmental degradation. These results highlight the complex relationship between economic growth and environmental sustainability, offering valuable insights for policymakers. For developed countries, the study emphasizes the need to uphold stringent environmental standards, foster technological innovation, and take the lead in global sustainability initiatives. For developing countries, it recommends building adaptive capacity, integrating sustainability into economic development strategies, and securing international support to implement effective environmental policies. Graphical Abstract