Estimating preference parameters from strictly concave budget restrictions

We propose an easy-to-use method for estimating preference parameters experimentally: choices from strictly concave budget restrictions (SCBRs). SCBRs generalize the popular method of analyzing choices from linear budget restrictions (LBRs). SCBRs promise (i) to improve the informational content of individual choices by reducing the number of corner allocations and (ii) to increase the range of identifiable behavioral types. Two online studies on risk and time preferences confirm the benefits of SCBRs vis-à-vis LBRs: (i) They reduce corner allocations drastically and make more participants estimable individually. (ii) They elicit a richer distribution of preference parameters, specifically, distinguishing linear from convex utility.