Modelling energy intensity with dynamic capital stock: a dual approach combining P-ARDL and MMQR

Abstract

This paper aims to analyse the behaviour of energy intensity in the context of European countries, with a particular focus on the role of dynamic capital stock. A panel dataset covering 32 years was constructed for a set of European countries. The application of preliminary diagnostic tests indicated the suitability of a Panel Autoregressive Distributed Lag (P-ARDL) model, specified in its Unrestricted Error Correction form, to capture both short- and long-term dynamics. To enhance the robustness of the findings and account for heterogeneity across the distribution of energy intensity, our approach was complemented with the Moments Quantile Regression (MMQR) approach. The methodology employed allowed for the model to capture both short- and long-run effects of covariates on energy intensity and uniquely explores how these effects vary across the distribution using MMQR. The results of our study indicate that when public capital is relatively scarce, private capital is more environmentally efficient, as evidenced by a positive relationship between capital stock and carbon intensity. Both energy intensity and ecological footprint are positively associated with higher carbon intensity, indicating their contribution to environmental degradation. Conversely, a higher share of renewable energy and increased trade are linked to lower carbon intensity, leading to reduced emissions in production and exports. To enhance energy efficiency, policymakers should promote public investment to boost private capital effectiveness, encourage renewable energy adoption and trade openness, and discourage high-energy-intensity activities while minimizing environmental footprint. Limitations of our paper stem from the availability of statistical data, which restricts the study to a historical snapshot. Given the novel approach linking capital structure to carbon intensity, the results should be viewed as preliminary and not yet generalized. Additionally, the evolving energy transition and rapid technological changes introduce uncertainty regarding the stability of the conclusions.