Exploring the elicitation of value life, health and consumer goods

In this thesis, we set out to explore some of the issues that arise when we elicit the societal value of life and health. Many governments use societal preferences to inform their policy and budget allocation decisions. However, there is plenty of evidence contesting the reliability and validity of the elicited preference estimates. A possible driver for this may be the fact that the economic theory underlying elicitation methods and value metrics lacks descriptive validity. First, expected utility theory (EUT) may not provide a valid description of people’s behaviour under risk. Second, the account that people exclusively care about maximising their own utility, disregarding that of others, may not be accurate. Using the wrong value estimates could lead to sub-optimally allocating enormous amounts of money, with the consequent loss of welfare; hence the relevance of this research. The first part of this thesis focuses around violations of EUT, in particular the reduction of compound lotteries axiom (RCLA) and the independence axiom. In Chapter 2, the disparities between the utility estimates elicited directly and indirectly using the standard gamble method are studied. They could be caused by a failure of the RCLA, but also by the context in which the estimates were elicited. We found that after having removed the effect of affect and the lack of incentives, and when choice complexity was the lowest, the disparities, and hence the violation of the RCLA, persisted. This suggests that context may play a role, but it is not sufficient to explain the disparities. In Chapter 3, the common ratio effect was examined as the complexity of choices was manipulated by having either money or objects as outcomes. As prior studies did, we found the independence axiom was violated with money, but we did not find this with objects. The higher level of consistency with objects should not be taken as evidence of the validity of EUT; rather, it is likely to stem from participants resorting to heuristics (such as “always choose the prospect with the best payoff”) when choices are too complex and hence difficult to make. Both of these chapters provide evidence that decisions may be domain-specific; this should be taken into account when testing models of decision making in the lab. The second part of the thesis focuses on social preferences in the context of fatality and physical risk reductions. In Chapter 4, the effect of self-interest on the efficiency-equity trade-off was studied. Efficiency was found to be the most important concern, closely followed by equity. Self-interest significantly modified this trade-off, but the magnitude of this effect was not large enough to overturn the general preference for efficiency over equity. On the other hand, stakes (whether the risk under consideration was fatal or non-fatal) did not change these preferences