Managing the Health Impacts of Climate Change through Sector Projects 25 Cost–Benefit Analysis of Adaptation: Accounting for Risk and Uncertainty Conducting any cost–benefit analysis implies looking into the future and asking how the “universe” of interest may look without the project, and with the project (the impacts of the project being the difference between these two scenarios). The exercise (looking into the future) is lacking complete information, and fraught with risk and uncertainty. This is true of all cost–benefit analyses, whether they be related to climate change or not. To this extent, the risk and/or uncertainty associated with projected climate change impacts is no different, and the analytical tools available to account for risk and uncertainty in the conduct of a project cost–benefit analysis are of relevance in the context of assessing the costs and benefits of climate change adaptation options. While climate change brings a new source of uncertainty, methodological tools exist to handle such uncertainty in the context of cost–benefit analysis. Two approaches may be applied to explicitly account for risk and uncertainty within the framework of the cost–benefit analysis. Each is briefly discussed below.17 Approach 1: Sensitivity Analysis The technique most widely applied to account for risk and uncertainty is sensitivity analysis (or sensitivity testing). In the context of conducting the cost–benefit analysis of an adaptation option, this type of analysis essentially involves changing the value of one or more variables which affect the adaptation option’s costs or benefits (for example, assuming that the cost of the adaptation option could be 20% higher than estimated; or assuming that the storm return period could be 1 in 30 years instead of the estimated 1 in 50 years), and for each such changes to re-compute the net present value of the option. This exercise may be repeated as much as may be deemed necessary. Switching values are often computed in the context of sensitivity testing. Switching value is the value of a specific variable that makes the net present value switch from positive to negative, or conversely. The purpose of such sensitivity testing is to raise the level of confidence in the outcome of the analysis. A key advantage of sensitivity testing is the simplicity, though it has severe limitations, including: (i) (ii) Sensitivity testing is highly subjective, due to the fact that no specific reason justifies the direction (smaller or larger) or the extent by which the value of a specific variable may be assumed to change; More importantly, sensitivity testing does not take into account the probability that the value of any specific variable may differ from the value originally estimated. As a result of this serious limitation, while sensitivity analysis allows for computing a range of net present values, within which the actual net present value of the adaptation option may fall, it does not allow for computing the expected net present value of the adaptation option. This last shortcoming explains the second approach used to account for risk and uncertainty in the cost–benefit analysis. Approach 2: Probabilistic Analysis Conducting a “probabilistic cost–benefit analysis” involves attaching a probability distribution for the possible value of any given specific cost or benefit component of the project, instead of attaching a single deterministic value. Such probability distributions may be constructed using records of historical data. Probabilistic analysis allows selecting multiple variables that can all be varied simultaneously according to the specific probability distribution attached to each variable. This process, known as a Monte Carlo simulation analysis, requires that a specific value of each individual variable (cost component or benefit component) be generated randomly according to the specific probability distribution attached to each 17 For more details, see for example ADB (2002) and Rayner et al. (2002).