Are non-monetary payment modes more uncertain for stated preference elicitation in developing countries?

An increasing number of stated preference studies adopt both monetary and non-monetary payment modes to elicit preferences for goods and services in developing and transition countries. The extent to which these alternative payment modes approximate the underlying human preferences for these goods and services is poorly understood. The circumstances under which monetary and non-monetary welfare measures can be combined for efficient estimation of welfare measures and to guide public resource allocation also remain unclear. In a split-sample design, we present a choice experiment on the purchase of flood insurance in which insurance premiums are paid in money, labour time and harvests. We use an integrated choice-modeling framework to test for differences in relative scale parameters among these three alternative payment modes. We find that the relative scale parameters for non-monetary payment modes are lower than the relative scale for monetary payment mode. We argue that the two non-monetary payment modes exhibit higher degrees of uncertainties in the choice experiment. We discuss possible causes and the implications of these results for the design of stated preference studies and the use of resulting welfare measures in cost-benefit analyses.