Research

We calculate the social cost of carbon (SCC) under stochastic climate volatility resulting from uncertainty about future climate risk regimes where weather extremes are becoming more frequent and intense. Using a stochastic dynamic integrated climate-economy model where representative agents are endowed with Duffie-Epstein recursive preferences, we find that climate volatility risks substantially increase the SCC both in the business-as-usual and optimal abatement policy scenario. We also show that switching to a regime with more intense disasters increases the SCC more than a switch to a regime with more frequent disasters for equal expected value. Overall we show that stochastic volatility has a major impact on the SCC.

We study the impact of tipping in the Greenland Ice Sheet (GIS) and permafrost, as well as their interactions, on the Social Cost of Carbon (SCC), optimal emission abatement policy, temperature and climate damage to global GDP. By incorporating reduced-form models of both tipping points into a stochastic dynamic integrated climate-economy model, we find that both tipping points can significantly affect future climate conditions when emission abatement and climate mitigation policies are absent. In the short term, thawing permafrost leads to larger climate damages and SCC estimations, largely through the permafrost methane emissions. Melting GIS alone does not increase the SCC much, but its impact can be significantly amplified jointly with the thawing permafrost. With stringent abatement policies, the estimations of future climate damages and the SCC will be greatly reduced. Accounting for thawing permafrost in the policy analysis will lead to more stringent abatement policies than the melting GIS.

We construct a behavioral SEIRS (BSEIRS) epidemic model to simulate the COVID-19 pandemic, and evaluate the effectiveness of different lockdown policies under uncertainty about vaccines. Our BSEIRS model has three new features: (i) a micro-founded theory of people's behavioral response to fear of contagion, (ii) Bayesian learning about the COVID-19 mortality rate, and (iii) the possibility of reinfection, all contributing to a better fit of the pandemic data. We highlight that lockdowns confer a real option value with uncertain exercise timing in the presence of uncertainty about vaccine effectiveness: Before an effective vaccine becomes available, lockdown is the only way to decelerate the spread of virus which in turn will give more people access to potentially benefit from a future vaccine. Stricter lockdown policies grant more people this real option of getting vaccinated before being infected once the vaccine arrives, and thus lead to a lower mortality rate. The real option value makes up a large proportion of the total health benefit from lockdowns, but has not been extensively discussed in the literature.

• Jan Brouwer Thesis Prize (2021) by Royal Holland Society of Sciences and Humanities (KHMW)