The energy module of FRIDA v2.1: endogenous investment and the timing of climate policy in the global energy system
Abstract. Integrated assessment models that derive energy-transition pathways from cost optimization or prescribed scenarios often diverge from observed investment behaviour. Here we present the energy module of FRIDA v2.1, a feedback-based, fully coupled global integrated assessment model built using System Dynamics, in which energy investment is endogenous. Total energy investment is set by a proportional–integral–derivative controller that responds to the relative mismatch between energy supply and demand, and is allocated across coal, oil, gas, solar, wind, hydropower, biofuel and nuclear technologies according to marginal costs that evolve with learning-by-doing, stepping-on-toes effects, resource scarcity, integration costs, climate damages and policy. Energy demand depends on population, GDP per person, energy prices, and climate-driven changes in heating and cooling degree days. Calibrated against 53 historical time series (1980–2023), the module reproduces observed trends in global energy supply, investment and emissions. Using 100 000-member ensembles, we compare endogenous model behaviour with four stylised policy scenarios that differ in the timing of carbon taxes and renewable subsidies. Without additional policy, warming exceeds 3 °C by 2100. All policy scenarios accelerate decarbonization, mainly through supply-side substitution rather than reduced demand, yet median warming remains above 2 °C; implementing both instruments early lowers end-of-century warming by roughly a third of a degree compared with delayed implementation. Abrupt policy introduction carries a substantial risk of persistent macroeconomic disruption, showing that successful decarbonization depends not only on policy strength but also on its timing and sequencing.