<![CDATA[This study focuses on designing a carbon tax policy based on spatial clustering and machine learning to identify optimal jurisdictions based on environmental-economic performance. The research aims to identify spatial patterns of environmental-economic performance across 38 countries, cluster countries based on similarity profiles using data-driven clustering methods, model the relationship between carbon prices/taxes, economic indicators, and environmental indicators, and recommend optimal carbon tax ranges for each jurisdictional cluster. Adopting a quantitative approach, this study utilizes secondary data from 38 countries, encompassing variables such as carbon prices/taxes, GDP, carbon emissions, energy consumption, industrial contribution to GDP, Environmental Performance Index (EPI), and climate change scores. The analysis employs Ward's hierarchical clustering method and evaluates silhouette coefficients to assess clustering validity. The results classify countries into five distinct clusters with varying environmental-economic characteristics. Developed nations with high environmental performance (e.g., Sweden, Norway, Denmark) are recommended to implement high carbon taxes (USD 100–150 per ton CO₂), while developing countries with high emission intensity (e.g., Indonesia, Kazakhstan) are advised to adopt low initial rates (<USD 15 per ton CO₂). Transitional economies are suggested to implement intermediate rates (USD 20–60 per ton CO₂). This study underscores the necessity of carbon tax policy differentiation based on economic capacity and environmental performance, as well as the importance of international cooperation in technology transfer and energy transition financing. The theoretical contribution lies in integrating Pigouvian Tax frameworks, spatial approaches, and machine learning to develop a more adaptive environmental fiscal policy design.]]>
Copyrights © 2025
