Hybrid Input-Output tables in physical units and money value for E3 CGE model calibration: methodologies of construction and implications thereof for climate and energy policy analysis

Energy-Economy-Environment (E3) modelers have called to develop “hybrid” models that aim at reconciling bottom-up and top-down approaches (Hourcade et al. 2006) to overtake their respective limits: on the one hand, engineers focus on the fine representation of technologies but without taking into account the feedback loop of the changes in the energy system on the rest of the economy; on the other hand, traditional economic modeling proposes macroeconomic assessments that sacrify a precise description of technical systems. Hybrid CGE models have proved relevant for this challenge and even if one can find very different modeling approaches within this category, such models rely on similar benchmark databases. In the view to re-establish the conditions of control at the interface between economic systems and technical systems, hybrid CGE models are calibrated on accounting databases that are built with the description of economic flows both in monetary value and in “physical” volumes, notably for energy goods. Nonetheless, model documentations are usually very elusive on the methodology adopted to build such “hybrid” benchmark databases along with the compromises carried out. Moreover, the impacts of those compromises on modeling results are rarely if never discussed. Therefore, this paper aims at bridging two important gaps: - Analyzing the issue of data combination to build hybrid Input-Output tables and proposing a rationale and precise protocol to do it - Studying the impacts of benchmark database characteristics on CGE assessment of climate policies The paper contains four parts. We first go briefly over E3 models to justify the need to work on a database with the dual description of some economic flows both in monetary value and in “physical” quantities. The second part focuses on the issue of building such databases by combining data on prices, data from material balances and national accounts. We propose a rationale to classify the possible compromises to be done on the allocation of statistical “gaps” that emerge from the direct confrontation of the various databases source. We apply this rationale to characterize the existing hybrid databases. The third part then details a precise protocol to build an energy-economy hybrid IO table for a single country that makes it possible to overtake most caveats previously identified. We illustrate this protocol on the case of France (2006) and Brazil (2005). Finally, in the last part, we analyze with a two-sectors static CGE model, the consequences of the choice of the benchmark database (within three different databases) on the economic impact assessment of mitigation policies. The contrasted results on welfare costs of similar climate policies suggest that data combination to build the hybrid benchmark database is crucial for climate and energy policy analysis and that this aspect is usually mistakenly underestimated. See above See above