GDP ENERGY EFFICIENCY IN THE ENERGY SECTOR: AN INTEGRATED LMDI AND OLS APPROACH

This study systematises and compares analytical methods for macroeconomic assessment of GDP energy efficiency, and applies two complementary approaches — the LMDI additive decomposition (Ang & Liu, 2001) and log-linear OLS regression — to Uzbekistan's energy sector for 2011–2023. The LMDI decomposition quantifies the separate contributions of primary energy consumption growth (activity effect, D_EC) and GDP growth (intensity effect, D_GDP) to the total observed change in energy intensity. Over 2011–2023, the cumulative D_GDP effect amounted to −0.4526 kgoe/USD (driving intensity down), while D_EC contributed +0.1692 kgoe/USD (partially offsetting improvement). The OLS model confirms a GDP elasticity of −0.975 and an energy consumption elasticity of 0.873, with R² = 0.997 and MAPE = 0.18% — substantially outperforming both naïve (MAPE = 12.71%) and simple trend (MAPE = 3.79%) benchmarks. A programme-target KPI framework for monitoring energy efficiency to 2030 is constructed from both methodologies. The integrated approach addresses limitations inherent in any single method and yields actionable guidance for Uzbekistan's energy and development policy.