The complexity of the climate-economy nexus: agent-based modelling and policy evaluation

While the consensus on the urgency of climate actions has grown in the last decades, what is the pathway to be followed to translate proposal into actions is still argument of debates in the climate change economics literature. Most economists believe that carbon pricing is the main and the most efficient option to reduce GHGs emissions, however a growing number of works point out that this result is highly dependent on the type of model used, claiming the superiority of a policy mix when a more realistic representation of the economy is used. My research work deals with the study of different climate policies with a complex system science approach, in particular, using the Eurace macroeconomic agent-based model. This work has two main objectives: first, to test the common belief that the carbon tax policy is the main and powerful instrument we have to induce the desired climate transition; second, to study the policy mix problem within the Eurace model economy, in particular, a mix of a carbon tax and a feed-in tariff policy. I enriched the Eurace model with a new agent, the climate module, to account for the climate-economy feedback. The economy affects the climate through greenhouse gas emissions from fossil fuels use for the energy production while the climate affects the economy damaging physical capital, with damages dependent on the temperature anomaly. Moreover, I introduced heterogeneity in the capital good sector, in order to include energy intensity improvements as a factor of technological change. In order to establish a relation between real world and model quantities, I followed an initialization procedure based on imposing physical constraints on model's quantities. I have developed an extended multi-criteria analysis method to evaluate policies performance accounting for both multiple objectives and variability of the outcomes of computational experiments. To pursue the research objectives I performed a set of computational experiments with the Eurace model, in which I analyzed a carbon tax policy, a feed-in tariff policy, and a mix of the two policies. Results of computational experiments show that the carbon tax is not the best performing climate policy when analyzed with the Eurace model, both the feed-in tariff and the policy mix perform better. This result is independent from the presence of climate damages. In absence of climate damages the PM performs better than its components, however, climate damages reduce the positive effects of the interaction between the components leading to higher economic costs for the same emission reduction obtained. According to the extended multi-criteria analysis, in presence of climate damages, the feed-in tariff policy is almost always preferred to the policy mix.