Essays in stochastic modeling with applications to economics

This dissertation examines how stochastic dynamic systems shape economic outcomes and policy design in two domains marked by uncertainty: the COVID-19 pandemic and climate change. Using stochastic differential equation-based frameworks, it highlights how randomness in key epidemiological, climatic, and economic processes fundamentally alters welfare assessments and optimal interventions.
Chapter 2 develops a stochastic behavioral SEIRS model to study lockdown policies during COVID-19. The model incorporates behavioral adaptation and learning about mortality risk, improving the ability to replicate observed infection dynamics. It shows that lockdowns generate substantial “real option value” by delaying infections until vaccines arrive, accounting for nearly 70% of the value of statistical lives saved under realistic policy scenarios. This mechanism, absent in deterministic models, underscores the importance of uncertainty for evaluating public health interventions.
Chapters 3 and 4 extend stochastic modeling to climate economics using Dynamic Stochastic Integrated Assessment Models (DSICE). Chapter 3 analyzes stochastic volatility in climate damages, showing that higher-order uncertainty about the frequency and intensity of extreme events can raise the Social Cost of Carbon (SCC) by more than 30%. Chapter 4 examines tipping risks associated with permafrost thaw and Greenland Ice Sheet melting. Incorporating these dynamic tipping elements significantly increases the SCC and supports more aggressive abatement, with combined tipping risks reducing optimal warming by 0.7°C and climate damages by over 30% relative to business-as-usual.
Overall, the dissertation demonstrates that uncertainty is not a nuisance but a defining feature of complex economic-environmental systems. Explicit stochastic modeling provides clearer guidance for optimal policy under deep uncertainty.