Essays in Economic TheoryPublic Deposited
The main theme of this dissertation are departures from standard assumptions in economic theory, specifically, departures from the model of subjective expected utility in decision theory. Part 1 axiomatizes the robust control criterion of multiplier preferences introduced by Hansen and Sargent (2001). The axiomatization shows that the class of multiplier preferences is precisely the intersection of the class of variational preferences, of Maccheroni, Marinacci and Rustichini (2006), and the class of second order expected utility preferences, of Ergin and Gul (2004) and Neilson (1993). The main contributions of Part 2 are an axiomatization of dynamic multiplier preferences and a characterization of preference for earlier resolution of uncertainty in the class of variational preferences. The latter result says that in the class of variational preferences the only preferences that satisfy indifference to timing are maxmin expected utility preferences of Gilboa and Schmeilder (1989). These questions are studied in a recursive setting, where time is infinite. Part 3 studies the same questions in a different setting--that of Maccheroni, Marinacci, and Rustichini (2006b) and Epstein and Schneider (2003b). In this setting time is finite, but there is more flexibility to model the state space, which does not have to be an infinite product of identical sets, as is the case in the setting of Part 2. The main result of Part 4 shows that probabilistic sophistication implies expected utility under an assumption that there exists a nontrivial unambiguous event. This means that although variational preferences are an excellent tool for studying behavior exemplified by the Ellsberg paradox, their ability to account for the Allais paradox is limited. Part 5 studies a definition of subjective beliefs for general ambiguity averse preferences. This definition leads to a characterization of the efficiency of ex ante trade: when aggregate uncertainty is absent, full insurance is efficient if and only if agents share some common subjective beliefs. This part of my dissertation was written jointly with Luca Rigotti and Chris Shannon and is forthcoming in Econometrica.