From Léon Walras to Lance Taylor: Model Closures and Theoretical Uncertainty in E3 Modelling
The Energy-Economy-Environment (E3) modelling literature has so far left aside the question of theoretical uncertainty in the building of macroeconomic models or modules. As a matter of fact, most of the latter are based on entrenched neoclassical assumptions, such as marginal pricing, full employment or causality going from savings to investment (Lavoie, 2015). Plus, if exceptions do exist (Bibas et al. (2010), SEURECO (2018)), the relative importance of their departure from classical hypotheses is yet to be assessed. This state of fact is all the more surprising since the Computable General Equilibrium (CGE) community in more traditional branches of economics have long been concerned with such uncertainties. From Sen’s (1963) seminal article, the question of the effects of models “closures”, i.e, ways to write the causal structure of a given CGE, has indeed elicited intense discussions. The aim of this paper is to put these two literatures together in order to sketch a preliminary picture of the effect of such “closures” on the behaviour of E3 models. For this purpose, we implement a stylised, neoclassical-inspired KLEM framework (Hudson & Jorgenson, 1974) and test a whole spectrum of targets of reduction of energy consumption using different causality structures. Our results show that model outcomes depend highly on the considered closure. Our conclusion is that relying quasi-exclusively on neoclassical assumptions can lead to a misleading and artificially unified picture of transition possibilities. These results are also an invitation to redefine the role of economists as “honest brokers” in facilitating policy debates, through the highlighting of the theoretical and epistemological implications of choosing one closure or the other on modelling results, but also on the understanding of economic phenomena at large.