L'énergie et les fonctions de production agrégées : perspectives historique et méthodologique

From a historical and methodological perspective, this paper focuses on empirical work on energy based on the aggregate production function, from the early 1970s to the late 2000s. It starts with the standard neoclassical approach, and in particular the controversy over the substitutability between capital and energy. Then it tackles the thermodynamic approach, which focuses on the explanation of the long-term growth. It shows continuity in the methodological issues raised by this work. At the theoretical level, the aggregate production function offers little conceptual insight into the physical aspects of the production process. At the empirical level, the results of estimates of energy production functions raise questions. In the neoclassical framework, the estimation is done indirectly through the cost function, so that the result is overdetermined by the marginal productivity pricing assumption. The thermodynamic approach proceeds in the opposite direction to a direct estimate, which encounters statistical problems no less important. If these difficulties relate more generally to the aggregate production function, energy issues reveal them in a very striking way.