How Different Are We? Identifying the Degree of Revealed Preference Heterogeneity

From unobserved to observed preference heterogeneity: a revealed preference methodology

“…number of subsets V (⋅) ) of a given WARP partition. In general, WARP constitutes a necessary condition for utility-maximizing behaviour (see Varian 1982). However, WARP is also a sufficient condition in a two-goods setting (see Rose 1958).…”

“…20 The literature has proposed alternative procedures to relax the 'sharp' binary WARP conditions. A most popular procedure is to make use of Afriat's CCEI (Afriat 1973;Varian 1990). As extending our method to include these relaxed rationality conditions is fairly straightforward, we will not consider this explicitly in the current paper.…”

“…making use of non-parametric revealed preference techniques in the tradition of Samuelson (1938Samuelson ( , 1948, Houthakker (1950), Afriat (1967), Diewert (1973) and Varian (1982). We present a novel and easy-to-implement method to identify the observable consumer characteristics that drive the preference heterogeneity underlying the observed variation in consumers' choice behaviour.…”

“…Before entering our analysis, it is useful to contrast the approach that we develop in the current paper with a frequent practice in the revealed preference literature that relates the degree of consumer rationality-usually measured through Afriat's Critical Cost Efficiency Index (CCEI) (Afriat 1973;Varian 1990)-to observable consumer characteristics; see, for example, Choi et al (2014). Both approaches make use of non-parametric revealed preference methods to relate observed choice behaviour to consumer characteristics.…”

“…11 In follow‐up work, Surana (2022) proposes a revealed preference method to quantify the degree of unobserved preference heterogeneity between different individuals in a panel data setting. This method makes use of partitioning techniques that are closely similar to those used by Cosaert (2019).…”