What if? Revisiting the MacroeconomicImpact of the Energy Crisis withPeak-load Electricity Balthazar de Vaulchier, Lionel Fontagné & Yu Zheng Highlights We introduce a base--peak load structure to electricity modeling in large-scale CGE models to better capturemerit order dispatch, capacity constraints and renewable intermittency. A transparent toy model demonstrates how ignoring peak load constraints systematically underestimatesthe macroeconomic impacts of energy shocks. We argue that second-order approximation are not well-suited for this kind of exercise, due to the large sizeof the shock compared to the value of the elasticity of substitution. Wetreatbaseloadandpeakloadelectricityasnon-substitutableLeontiefcomplements,reflectingoperationalconstraints in electricity systems. Applyingtherefinedmodeltothe2022RussiangasshockrevealslargerGDPandwelfarelossesinGermanyandtheEUthanstandardelectricityrepresentationspredict.Applyingtherefinedmodeltothe Abstract Electricity generation presents distinctive modeling challenges due to the absence of storage, instantaneous demand-supply balancing requirements, and heterogeneous generation technologies with different cost structures. Thispaper addresses these challenges by incorporating a base--peak load structure into large-scale computable generalequilibrium (CGE) models, offering a middle ground between detailed energy system models and multisectoral globaleconomicframeworks.Wefirstdevelopatransparenttoymodelinspiredby\textcite{bachmann2022}todemonstratethatfirst-orderapproximationsofcascadingeffects,followingHulten'stheorem,areinadequatewhenshocksarelarge and elasticities of substitution are low. Building on the theoretical insights, we embed a base--peak structure intotheMIRAGECGEmodel,treatingelectricityasaLeontiefproductionfunctionbetweenbaseloadgeneration(coal,nuclear,hydro,andpartofrenewables)andpeakloadgeneration(gas,oil,andpeakrenewables).Thisrefinementcaptures the merit order dispatch mechanism and bottleneck effects when peak generation is constrained. We applytheenhancedmodeltoassessthe2022RussiangasshockinGermanyandtheEuropeanUnion.Ourresultsdemonstrate that the base--peak structure more accurately reproduces observed macroeconomic impacts compared Keywords ElectricityModeling,BaseLoadandPeakLoad,ComputableGeneralEquilibrium,EnergyShocks,RussianGasCrisis,MIRAGEModel. JEL C68,D58,Q41,Q43. Working Paper What if? Revisiting the macroeconomic impact of theenergy crisis with peak-load electricity Lionel Fontagné1, Balthazar de Vaulchier2, and Yu Zheng1Paris School of Economics, 48 boulevard Jourdan, 75014, Paris, France2CEPII, 20 avenue de Ségur, 75007, Paris, France 1Introduction With growing policy concerns around energy security, climate change, and their interactionswith international trade, large-scale general equilibrium models have been regainingattention among economists as valuable tools for energy and climate policy analysis (see,e.g., Bekkers and Cariola 2022; Bellora and Fontagné 2023; Clora and Yu 2022). However,there is a tension between tractability of the model and detailed representation of the Electricity generation presents modeling challenges due to several distinctive features.The first one is the (quasi) absence of storage, coupled to the necessity of equalizing supplyand demand at all time if blackouts are to be avoided. The second one is that, contrary towhat we would think of an electron identical to another one, electricity is differentiated:electricity produced by a nuclear plant is not the same, economically speaking, as electricitycoming from a gas plant. Indeed, different technologies exist to produce electricity, each This design has two implications for modeling strategies.First, a standardcesproduction function between the different electricity sectors would not be suited torepresent the mechanism of electricity generation. Indeed, due to the merit order curve andthe fixed capacities of the existing plants at short term, substitution between the different These issues lead us to hypothesize that electricity should be split between base loadand peak load electricity in the modeling. Although not fully realistic, such approach hasthe merit of being quite simple for the macroeconomic modeling, while still capturing part of electricity specificities. Because of the merit order curve on the supply side, and the verypoor elasticity of electricity demand,1one can treat electricity generation as a Leontief Against this background, this paper aims at illustrating how adopting a more refinedmodeling of electricity helps reconsider the economic consequences of an energy shock ascaptured by a large scale model of the global economy. We first introduce a transparenttoy model inspired by Bachmann et al. (2024) – the reference paper on the consequences ofthe energy crisis in Germany. We then embed the same base-peak structure intoMirage(Bouët, Fontagné, et al. 2026), a computable general equilibrium model. We apply the