Free-riding, Free-driving, and Moral Hazard of SRM in a Simple Climate-Economy Model

CO2 emission reduction is hampered by free-riding problems: Countries not reducing their emissions still benefit from other countries effort, so countries are tempted to let others do the effort. By contrast, solar geoengineering is considered cheap enough that individual countries or small coalitions could implement it, requiring no collaboration (free-driving). While many fear that the prospect of Geoengineering reduces incentives to decarbonise, others argue that the threat of unilateral Geoengineering might actually trigger decarbonization, if a country that fear climate change threatens to perform Geoengineering unless decarbonization efforts are made (Fabre and Wagner, 2020). We investigate this hypothesis using a simple climate-economy model based on Nordhaus' famous DICE model. We first add SRM as a policy option (Helwegen et al., 2019). Next, we rewrite the energy module of DICE. DICE unrealistically assumes that the cost of reducing emissions in any year is independent of previous abatement efforts (Grubb et al., 2021), while in fact green investments have a much longer lifetime. We replace DICEs abatement cost by a stylized energy sector wherein the decision maker can invest in fuel-based or green energy plants. Green plants get cheaper through learning-by-doing. Decarbonising therefore inflicts significant costs while the green transition takes place, but not perpetual costs as in DICE. This makes rapid decarbonisation more attractive relative to SAI, and adding the policy tool of SAI leads only to minimal delays in decarbonisation. If decarbonisation is transition-like, a rational actor will not delay it significantly even if SAI is available to generate additional cooling. Finally, we expand the model to allow for two or more regions with independent decision makers (Brede and de Vries, 2013). These can either be altruistic (optimising global welfare) or competitive (optimising their own welfare). For two identical regions, decarbonisation is significantly delayed if they act competitively rather than altruistically (free-riding), while the presence or absence of SAI has again hardly affects the timing of the transition. We explore the parameter space for unequal regions to identify conditions under which the option of SAI may accelerate or seriously delay the green transition.