My dissertation consists of three chapters that examine the role of behavioral biases in both expert updating and asset pricing. It provides empirical evidence for confirmatory bias and Bayesian reassessment in expert updating, and utilizes a unique, regression-discontinuity approach for identifying confirmatory bias, using insights from a new model of confirmatory bias. Additionally, I propose a rational explanation for the home underdog bias, which has been found in many sports betting markets. I use evidence from a set of binary choice models to propose that betting houses intentionally leave betting on home underdogs open for profitable betting in order to eliminate the behavioral strategy of betting on hot teams.
Chapter 1 considers how experts use Bayes' rule, the common benchmark for expert updating. Experts are regularly relied upon to provide opinions in financial markets, but the expert's incentives to truthfully reveal his opinions depend on the environment, making it difficult to identify forms of expert bias. I consider how experts report their beliefs in an environment devoid of incentives. Using data from an incentive-free environment-the Associated Press Top 25 Poll for college football-I find that experts reassess data selectively, depending on their prior hypotheses over time. I also find that experts do not fully incorporate all of the information contained in a signal for a given time, as they "reassess" the quality of information in light of new information. Furthermore, I also find that experts react more strongly to imprecise signals than would be expected if the experts were truly Bayesian. These reactions cause persistence in beliefs over time, since prior beliefs are weighed in current belief calculations. Overall, the manner by which experts respond mirrors confirmatory bias, the idea that individuals interpret ambiguous evidence as being consistent with their prior hypotheses. I conclude that the evidence that experts may be biased even without incentives suggests that incentives may exacerbate biases.
Chapter 2 considers market efficiency in betting markets for amateur sports. In this essay, I test for betting market efficiency in an amateur sport, American college football, using data from over 11,000 games from 1985 to 2003, and find that the market is inefficient. In particular, favorites are systematically overpriced. I show that the magnitude of this bias is large enough to generate both economic and statistical inefficiency in this betting market. More importantly, I provide suggestive evidence for the cause of this inefficiency: betting houses deliberately inflate the betting lines for favorites in order to counteract bettors' "hot hand" beliefs. While eliminating the "hot hand" bias is efficient for a betting house, tempering the "hot hand" results in consistently profitable simple betting strategies.
Chapter 3 proposes a way to identify confirmatory bias in empirical settings. Confirmatory bias is pervasive in many settings, including principal/agent relationships, investing decisions, and expert opinions. Empirical evidence of confirmatory bias and its consequences is largely limited to laboratory experiments. I present empirical evidence of confirmatory bias in individual beliefs from a unique source- the Associated Press Top 25 College Football Poll. Individuals in the poll are tasked with assessing team quality, and their beliefs are randomly treated from week-to-week with game results that serve as new signals about individual teams' qualities. I use a regression-discontinuity approach to exploit the unique variation in signal treatments to identify confirmatory bias in this real-world setting. I find that individuals with positive hypotheses about a team's quality react significantly to weakly positive signals about a team's quality, but do not react to weakly negative signals about a team's quality. The estimates suggest that nearly one-third of the magnitude of belief changes in the AP College Football Poll is due to confirmatory bias, indicating that announced beliefs may be highly sensitive to prior beliefs.