Energy efficiency and economic growth in China: The effects of technological progress and policy intervention

China reduced the energy intensity of its output by roughly 45% between 2005 and 2022 while its economy expanded more than threefold in real terms. This paper asks how energy efficiency relates to that growth, and how technological progress and policy intervention shape efficiency itself. We assemble a balanced panel of 30 provinces for 2005 to 2022. We measure total-factor energy efficiency with a slacks-based data envelopment model that treats carbon dioxide as an undesirable output. We then follow a second-generation panel strategy. Cross-sectional dependence tests, the CIPS unit root test, slope homogeneity tests, and the Kao cointegration test guide the choice of estimators. We estimate the growth equation with feasible generalized least squares, Prais-Winsten regression, and fixed effects with Driscoll-Kraay standard errors. We identify the policy effect through a staggered difference-in-differences design built around the national low-carbon pilot programme. The results show a stable long-run link between efficiency and income. A 0.1 point rise in total-factor energy efficiency is associated with about 4.0% higher real GDP per capita. Research and development intensity raises both income and efficiency. Pilot designation lifts efficiency by about 0.028 points, and the effect builds over the first four years. Short-run causality runs from income and research effort to efficiency rather than the reverse. The findings suggest that China's efficiency gains are largely induced by growth and directed by policy, and that further gains depend on sustained research investment and a lower coal share.