Benchmarking the utility theory: a data envelopment approach

The purpose of this paper is to benchmark alternatives of decision problems that include risk and uncertainty considering different risk attitudes via a new data envelopment analysis (DEA) decision model.

A new utility function of strict bounds is applied in a data envelopment model to evaluate all possible stochastic alternatives (i.e. gambles). The amount of risk in the alternatives is measured by a newly introduced risk ratio (RR). Each alternative is considered as a decision making unit (DMU). The alternatives efficiency frontier is found via linear optimization of the DEA model.

In contrast to literature studies of binary decision alternatives, here, benchmarking is conducted to evaluate multiple decision alternatives with unbounded utility of the payoffs along with a new DEA decision model. Different surveys and studies have been used to validate the model. DEA could demonstrate the ability to uncover relationships that remain hidden for other methodologies. The resulting rankings remarkably conform to those elicited by subjects.

Individuals of different wealth backgrounds evaluate risky decision problems differently.

The paper contributes to the existing research by benchmarking multiple alternatives as compared to the literature research which usually assesses binary problems. Instead of using explicit utilities, the model implements the efficiencies along with a new utility function and a new RR. The introduction of DEA to such a decision field is found to be successful in benchmarking numerous alternatives under different risk attitudes.