TY - JOUR TI - Three essays in the theory of preferences DO - https://doi.org/doi:10.7282/t3-t00x-by36 PY - 2018 AB - This dissertation consists of three chapters. The first chapter addresses the classical questions of utility representation and maximization. It relaxes the notion of weak upper continuity(Campbell and Walker) to obtain a property called partial weak upper continuity and shows that both maximization of preferences and representation by a utility function can be achieved under this new property. The rest of this dissertation focuses on extending revealed preference theory to accommodate behavioral anomalies observed in the experimental data. In particular, I offer a framework to expand the theory of revealed preferences to the case where a DM's choice is not completely identified with a single preferences. In Chapter 2, I use a divide and conquer procedure in order to expand the revealed preference theory to accommodate behavioral anomalies such as attraction effect, compromise effect, and reverse dominance effect. These effects are induced when a third alternative is used as a "decoy" to change the relative ranking from a pair. Therefore, they could be rationalized using the notion of referenced preferences: that is, when the pairwise preference between a pair of alternatives is referenced when a third alternative. I model such behavior as a partially rational inductive divide and conquer procedure where the deviation from WARP only take place on tripletons where references operate. Keeping the rational choice axioms on sets with higher cardinalities retains the predictive power in the classical theory to the extent possible. In order to do this, I assume a DM only drops reference elements when facing more than one such elements in a set. I show that under this particular division rule the choice is characterized via simple majority rule over the collection of referenced preference. In Chapter 3, I consider a variation of the results in Chapter 2 where in dividing a set of alternative in an inductive manner, the DM considers all possible first-diminished subsets. I show that such division rule results in a more sophisticated behavior than the simple majority rule. Here the DM uses here referenced preferences over the pairwise preference in a consecutive manner to squeeze the choice problem. This, indeed is an extension of the rational short-list method in Manzini and Mariotti (2007) with three preference. The distinction, however, is that the method is enogenized in this set up. KW - Economics KW - Revealed preference theory LA - eng ER -