Climate Change 2001:
Working Group II: Impacts, Adaptation and Vulnerability
Other reports in this collection

2.5.4. The Cost of Uncertainty

This section reviews the primary methods for incorporating uncertainty into analyses of climate impacts. Here we look at how to judge the cost associated with uncertainty. Cost and valuation depend, in general, on the entire distribution of the range of outcomes. Insurance and the Cost of Uncertainty

Risk-averse individuals who face uncertainty try to buy insurance to protect themselves from the associated risk (e.g., different incomes next year or over the distant future, depending on the state of nature that actually occurs). How much? Assuming the availability of "actuarially fair" coverage (i.e., coverage available from an insurance provider for which the expected cost of claims over a specified period of time equals the expected income from selling coverage), individuals try to insure themselves fully so that the uncertainty would be eliminated. How? By purchasing an amount of insurance that is equal to the difference between the expected monetary value of all possible outcomes and the certainty-equivalent outcome that insurance would guarantee—the income for which utility equals the expected utility of all possible outcomes.

For a risk-averse person, the certainty-equivalent income is less than the expected income, so the difference can be regarded as WTP to avoid risk. In a real sense, therefore, willingly paid insurance premiums represent a measure of the cost of uncertainty. Therefore, they can represent society's WTP for the assurance that nondiversifiable uncertainty would disappear (if that were possible). Thus, this is a precise, utility-based measure of economic cost. The cost of uncertainty would be zero if the objective utility function were risk-neutral; indeed, the WTP to avoid risk is positive only if the marginal utility of economic activity declines as income increases. Moreover, different agents could approach the same uncertain circumstance with different subjective views of the relative likelihoods of each outcome and/or different utility functions. The amount of insurance that they would be willing to purchase would be different in either case. Application of this approach to society therefore must be interpreted as the result of contemplating risk from the perspective of a representative individual. Yohe et al. (2000), for example, apply these structures to and offer interpretations for the distributional international impact of Kyoto-style climate policy.

Other reports in this collection