1. Motivation and Context

The critical economic dichotomy lies between structural change and technological progress. Is the observed decline in energy intensity merely a composition effect, resulting from a shift in economic activity away from energy-intensive manufacturing toward services? Or does it represent genuine technological advancement in energy utilization? This distinction is not merely academic; if the decline is driven by technology, the path is replicable for developing economies. If it is driven solely by structural shifts (outsourcing dirty production), it is not.

This contribution seeks to disentangle these effects using a comprehensive dataset to "peel the onion" of European energy intensity between 1995 and 2009.

2. Data and Methodology

To rigorously assess these dynamics, we utilize the World Input-Output Database (WIOD). This harmonized dataset provides time-series input-output tables, environmental satellite accounts, and socioeconomic information across 27 EU countries and 34 sectors from 1995 to 2009.

The methodological framework employs a two-stage approach:

• Index Decomposition Analysis (IDA): We apply the Logarithmic Mean Divisia Index (LMDI-II) method. This approach is preferred as it is zero-value robust and yields a "perfect decomposition" with no unexplained residuals. We decompose changes in energy intensity into a within-country structural effect, a between-country structural effect, and a technology effect.
• Econometric Panel Analysis: We extend the descriptive IDA results by constructing a panel to identify causal drivers, controlling for variables such as Total Factor Productivity (TFP), trade openness, and capital-labor ratios.

3. Empirical Findings

The analysis reveals substantial heterogeneity across Member States, yet a clear aggregate trend emerges. For the EU27 as a whole, energy intensity declined by 27.4% over the period. The decomposition analysis indicates that this clean-up was driven predominantly by the technology effect.

We classify the Member States into four clusters based on their performance regarding Gross Output growth relative to Energy Use growth:

• Best Performers: Primarily Eastern European nations (e.g., Poland, Estonia, Czech Republic) which experienced massive decoupling—above-average output growth with below-average energy use growth.
• Mature Economies: Countries like Germany and the UK showed moderate improvements, driven largely by industrial technological efficiency rather than structural shifts.

4. Theoretical Implications and Econometric Evidence

The econometric analysis provides insight into the determinants of these trends. We find statistical evidence for an inverted U-shaped relationship between income and energy intensity, supporting the Environmental Kuznets Curve hypothesis.

Furthermore, the results highlight the role of production factors:

• Trade Openness: An increase in openness lowers energy intensity, likely due to the import of improved technology or the outsourcing of energy-intensive intermediate production.
• Capital-Labor Substitution: We find that capital and energy act as complements in the short run; an increasing capital-to-labor ratio leads to an increase in the total energy intensity effect.
• Total Factor Productivity (TFP): There is a robust negative correlation between TFP growth and energy intensity, confirming that general productivity advancements spill over into energy efficiency.

5. Conclusion

The empirical evidence suggests that the decline in European energy intensity is not merely a statistical artifact of deindustrialization or structural change. Rather, technological progress is the key driving force.

This conclusion offers a significant positive implication for global climate policy. If the European clean-up were driven solely by structural change, developing nations would be unable to replicate this trajectory without sacrificing industrial growth. However, because the driver is technological, this development path is replicable for emerging economies such as China and India.