Empirically calibrating damage functions and considering stochasticity when integrated assessment models are used as decision tools

Abstract

Benefit-cost integrated assessment models (IAMs), though developed originally for exploratory research, are now being applied as decision-making tools. This application places new demands on model calibration and capabilities. We suggest two directions for increasing the policy applicability of IAMs. First, employ recent work in the impacts community on empirical impact functions, grounded in the observed response of human systems to climate variability, to parameterize and calibrate IAM damage functions. Empirical damage functions can supplement and, in some cases, replace the often-dated damage estimates in IAMs with alternatives that can be directly compared to contemporary observations. Second, explicitly model the interactions between changes in mean climate and stochasticity in natural and human systems (e.g., weather, business cycles). Explicit stochasticity enables consideration of risk aversion with respect to episodic factors, such as extreme weather, thereby providing a natural way to examine the benefits of consumption-smoothing adaptive measures, such as insurance.