Research

Abstract: Adding a portfolio choice problem to the consumption-saving model of Deaton (1991) or Carroll (1992) shows that the optimal portfolio share of equity declines with wealth, even in the presence of precautionary saving motives. Glaringly, the poor invest all their savings in equity. Using an infinite-horizon consumption-saving problem, I examine the effects of a positive covariance between permanent income growth and returns on equity (Viceira 2001) on optimal portfolio allocation. A positive covariance between the two decreases the optimal portfolio share of equity at all wealth levels (Kimball's (1991) temperance motive). If the precautionary saving motive is a no-borrowing constraint, with a positive covariance between permanent income growth and returns on equity beyond a threshold value that depends only on the equity premium percentage and risk aversion (i) the optimal portfolio share of equity increases with wealth, (ii) the poor invest all their wealth in the risk-free asset, and (iii) the optimal portfolio share is lower than the Merton-Samuelson (no labor income) share. If the precautionary saving motive is a positive probability of zero labor income, however, (i) and (ii) do not hold true, though (iii) does. When calibrated to U.S. data, neither precautionary saving motive can fully explain observed portfolio shares of equity even with a very high covariance between permanent income growth and returns on equity. The calibrated temperance motive may, however, explain the observed equity portfolio share in conjunction with other popular explanations documented in the literature.

Conference Presentations: DSE Winter School 2023, BREW-ESA 2023, ISI ACEGD 2023, Ahmedabad University 2024, AMES-CSW 2024

Abstract: We introduce an interactive model of belief formation and transmission within echo chambers. Individuals in our model have subjective core beliefs, but these are not the same as the beliefs underlying their behavior. An individual's behavioral beliefs incorporate her core beliefs but are also influenced by the behavioral beliefs of others within her echo chamber. Therefore, echo chambers feature interactive behavioral beliefs with any individual's beliefs both being influenced by as well as influencing the beliefs of others in her echo chamber. The echo chamber representation of a profile of such behavioral beliefs that we propose captures the steady state of this process of interaction and influence. We show that the model is falsifiable by characterizing it based on two axioms. The first emphasizes the need for conformity when it comes to assessments about certain events, while the second highlights the possibility of differing behavioral beliefs about uncertain events and the potential for everyone to exercise influence. We also analyze when the model permits exact identification, i.e., when can an analyst draw on the profile of behavioral beliefs to uniquely identify the composition of the echo chambers in society, the core beliefs of different individuals, and the degree to which each of them is immune to influence. Further, we provide an observational learning foundation for our theory by incorporating sequential arrival of private information into the DeGroot (1974) model. We use this foundation to provide a novel theory of interactive belief updating for non-Bayesian decision-makers.

Abstract: I propose a model of decision making under uncertainty that make ambiguity attitudes susceptible to menu effects. The starting point for this exercise is a choice rule defined over finite sets of Anscombe-Aumann acts. Each act is evaluated, in the context of a menu, through (i) a non-comparative component: a vector expected utility evaluation of the act with standard menu-independent parameters (ii) a comparative component: a cost assigned in relation to the other acts in the menu. I introduce a notion of comparative ambiguity, which measures the excess ambiguity to which an act is exposed, against the benchmark of the least ambiguous comparable act in the menu. The decision maker then makes an adjustment in their utility evaluation for this excess. The choice rule is determined by the maximization of this menu-dependent utility. Under minimal assumptions, my model lends to sharp experimental identification of how ambiguity attitudes are susceptible to menu effects. The paper details the underpinnings of the model through an extensive literature review and benchmark its predictions against AI generated responses using GPT-4.