GREEN GROWTH OR EMISSIONS LOCK-IN? A HYBRID ECONOMETRIC–AI FORECAST OF CHINA’S INDUSTRIAL CO₂ PATHWAYS

Abstract

This paper examines whether China’s industrial sector is advancing toward green growth or facing an emissions lock-in by integrating econometric analysis with artificial intelligence–based forecasting. Using panel data from 30 provinces between 2000 and 2023, the study applies a System GMM model alongside machine learning algorithms—Random Forest, Gradient Boosting, and Support Vector Regression—to identify the key drivers and predict future CO₂ trajectories. The results reveal that energy intensity and industrial output significantly raise emissions, while renewable-energy deployment and R&D investment exert strong mitigating effects. Among AI models, Random Forest achieves the highest predictive accuracy (R² = 0.86), validating its effectiveness for nonlinear environmental systems. Scenario forecasts for 2025–2030 indicate that moderate efficiency gains and renewable expansion could lower emissions by nearly 10 percent compared with the baseline path. Overall, the findings underscore that improving energy efficiency, stimulating innovation, and addressing regional disparities are essential to prevent long-term carbon lock-in and to align industrial development with China’s 2060 carbon-neutrality goal.