STATISTICAL VS. IDENTIFIED LIVES INBENEFIT-COST ANALYSIS

Economic evaluation of projects involving changes in mortality risk conventionally assumesthat lives are statistical, i.e., that risks and policy-induced changes in risk are small andsimilar among a population. In reality, baseline mortality risks and policy-induced changes inrisk often differ among individuals although these differences are imperfectly known. Weexamine the effects of information about heterogeneity of risk on economic evaluation.Although social welfare (defined as aggregate expected utility) is unaffected by informationabout risk heterogeneity, the economic valuation of changes in risk (the sum of individualcompensating or equivalent variations) is sensitive to this information. The effect ofinformation on economic valuation and hence the outcome of a benefit-cost analysis (BCA)depends on: i) whether information is about heterogeneity of the baseline and/or change inrisk, ii) whether risk is valued using willingness to pay (WTP) or willingness to accept(WTA) measures, iii) the status quo policy, and iv) whether individuals are risk-averse orrisk-neutral in wealth. We show that BCA does not systematically favor identified overstatistical lives and suggest some political factors that may explain the apparent publicdecisionbias toward protecting identified lives.