Régulation et Environnement

Du 19/02/2018 de 12:00 à 14:00

Salle R1-13, Campus Jourdan, 48 boulevard Jourdan, 75014 Paris

CREPIN Anne-Sophie ()

Inertia in risk; improving economic models of catastrophes

We provide a new way to model endogenous catastrophic risk termed inertia risk, which accounts for dynamic lags between physical variables and the hazard rate—a characteristic which is often observed in real life problems. For example the stock of some particular species in an ecosystem might influence the level of stress (or resilience) in that system, which is a slowly changing variable compared to the species dynamics. When the level of stress passes a critical threshold the ecosystem undertakes a regime shift, a rapid, substantial and persistent change in the system. In such context the inertia hazard rate defines the risk that the system will shift in the next period of time, which depends on both variables. We show that the added realism in our risk model has intuitive appeal and significantly impacts optimal solutions. With inertia risk, the probability that a catastrophe will ever occur may span the entire interval [0, 1]. This as opposed to the standard approach where this probability is either zero or one. We also show how inertia risk may generate path dependency as the hazard rate depends on learning about how risk is distributed in state space. We illustrate the implications for policy in a simple model ­­of climate change. The optimal solution with inertia depends on parameters, such as damage and discount rates in a qualitatively different way compared to the standard approach. Hence for problems with lagged effects, where inertia risk is a more realistic way to represent risk, using the standard models of catastrophic risk that discard these lagged effects could generate substantially flawed policy recommendations.