Essays On Optimal Portfolio Choice PDF Download

This Thesis, consisting of three chapters, studies the effects of uncertainty on decision-making with portfolio choice applications. Chapter 1 studies how experimental subjects report subjective probability distributions in the presence of ambiguity characterized by uncertainty over a fixed set of possible probability distributions generating future outcomes. The level of distribution uncertainty varies according to the observed outcomes and the rules used by the subjects to update the distribution uncertainty. This chapter introduces several reporting and updating rules and our empirical analysis focuses on estimating the sample distribution of these rules. Two dominant reporting rules emerge from our analysis: we find that 65% of subjects report distributions by properly weighting the possible distributions using their expressed uncertainty, while 22% of subjects report distributions close to the distribution they perceive as most likely. Further, we find significant heterogeneity in how subjects update their expressed uncertainty. On average, subjects tend to overweight the importance of their prior uncertainty relative to new information, leading to ambiguity that is substantially more persistent than would be predicted using Bayes' rule. Counterfactual simulations suggest that this persistence will likely hold in settings not covered by our experiment. Uncertainty in financial markets is a natural consequence of investors being unaware of objective probabilities of asset returns. Chapter 2 highlights that ambiguity and loss aversion have opposite effects on financial markets and can coexist in the presence of uncertainty. This chapter addresses the normative question of the optimal portfolio evaluation frequency for an investor in order to minimize the effect of myopia, but to learn about the investment opportunities in the market. Towards this end, we present a new experimental design in which investors are asked to make repeated portfolio choices facing initial ambiguity concerning the distribution of returns of one of the available assets. We exploit exogenous variations in evaluation frequency along with time variation of probabilistic beliefs over the possible return distributions to jointly identify ambiguity, loss, and risk aversion along with rules investors use to update their ambiguity. Estimates from a structural model suggest seven different classes of investors. Investor class membership depends on loss aversion, ambiguity aversion as well as risk aversion preferences. Further, we find that at the aggregated level, investors are loss averse, ambiguity averse and they display risk aversion over gains and risk seeking over losses. We conclude our analysis by using our model estimates to predict the distribution of optimal evaluation periods for our sample. Our predictions suggest that approximatively 70% of investors prefer the highest possible evaluation period frequency. Finally, Chapter 3 investigates whether or not the discount factor of the elderly affects their portfolio choices. We estimate time preferences using inter-temporal choice data from a hypothetical experiment in a representative sample of American elders and a structural model of decision-making accounting for lifetime uncertainty. Our results indicate considerable heterogeneity in the elderly population. Moreover, we find that older people who display a higher discount factor are more likely to own retirement accounts and risky assets. These older people also tend to decrease the share of financial wealth held in safe assets and increase the share of financial wealth held in risky assets. These findings suggest that time preferences affect investment choices from safe assets toward other financial assets, all else being equal.