Papers
Published and Accepted Papers
A Model of Non-Belief in the Law of Large Numbers (with Daniel Benjamin and Matthew Rabin)
Journal of European Economic Association, 2015
People believe that, even in very large samples, proportions of binary signals might depart significantly from the population mean. We model this “nonbelief in the Law of Large Numbers” by assuming that a person believes that proportions in any given sample might be determined by a rate different than the true rate. In prediction, a nonbeliever expects the distribution of signals will have fat tails. In inference, a nonbeliever remains uncertain and influenced by priors even after observing an arbitrarily large sample. We explore implications for beliefs and behavior in a variety of economic settings.
A Behavioral Analysis of Stochastic Reference Dependence (with Yusufcan Masatlioglu)
American Economic Review, 2016
We examine the reference-dependent risk preferences of Koszegi and Rabin (2007), focusing on their choice-acclimating personal equilibria concept. Although their model has only a trivial intersection (expected utility) with other reference dependent models, it has very strong connections with models that rely on different psychological intuitions. We prove that the intersection of rank-dependent utility and quadratic utility, two well-known generalizations of expected utility, is exactly monotone linear gain-loss choice-acclimating personal equilibria. We use these relationships to identify parameters of the model, discuss loss and risk aversion, and demonstrate new applications.
Preferences for Truth-Telling (with Johannes Abeler and Daniele Nosenzo)
Econometrica, 2019
Private information is at the heart of many economic activities. For decades, economists have assumed that individuals are willing to misreport private information if this maximizes their material payoff. We combine data from 72 experimental studies in economics, psychology and sociology, and show that, in fact, people lie surprisingly little. We then formalize a wide range of potential explanations for the observed behavior, identify testable predictions that can distinguish between the models and conduct new experiments to do so. None of the most popular explanations suggested in the literature can explain the data. We show that only combining a preference for being honest with a preference for being seen as honest can organize the empirical evidence.
One in a Million: Fields Experiment on Belief Formation and Pivotal Voting (with Alan Gerber, Mitchell Hoffman and John Morgan)
AEJ: Applied, 2020
A common feature of many models of voter turnout is that increasing the perceived closeness of the election should increase voter turnout. However, cleanly testing this prediction is difficult and little is known about voter beliefs regarding the closeness of a given race. We conduct a field experiment during the 2010 US gubernatorial elections where we elicit voter beliefs about the closeness of the election before and after showing different polls, which, depending on treatment, indicate a close race or a not close race. We find that subjects update their beliefs in response to new information, but systematically overestimate the probability of a very close election. However, the decision to vote is unaffected by beliefs about the closeness of the election. A follow-up field experiment, conducted during the 2014 gubernatorial elections but at much larger scale, also points to little relationship between poll information about closeness and voter turnout.
On the Consensus Effect (with David Dillenberger)
Journal of Economic Theory, 2019
It is well documented that individuals make different choices in the context of group decisions, such as elections, from choices made in isolation. In particular, individuals tend to conform to the decisions of others — a property we call the consensus effect — which in turn implies phenomena such as group polarization and the bandwagon effect. We show that the consensus effect is equivalent to a well-known violation of expected utility, namely strict quasi-convexity of preferences. Our results qualify and extend those of Eliaz, Ray and Razin (2006), who focus on choice-shifts in group when one option is safe (i.e., a degenerate lottery). In contrast to the equilibrium outcome when individuals are expected utility maximizers, the consensus effect implies that group decisions may fail to properly aggregate preferences in strategic contexts and strictly Pareto-dominated equilibria may arise. Moreover, these problems become more severe as the size of the group grows.
Tell all the Truth, but Tell it Slant: Testing Models of Media Bias (with Sarah Taylor)
Journal of Economic Behavior and Organization, 2021
Media outlets often appear to bias news reports, but it is difficult to identify bias clearly. In addition, it is often unclear what motivates media bias and whether this bias improves the welfare of news consumers. We develop a model of demand-driven media bias, and we test for the existence of bias using a novel data set. We consider how daily weather predictions by The New York Times in the late 19th century differ across days when the New York Giants (the local baseball team) played a home game and days when they did not. The historical context provides a clean natural experiment for identifying bias: on game days the Times' weather reports were relatively more accurate at predicting sunny weather and less accurate at predicting rainy weather. We provide evidence that a model of demand-driven bias is a more plausible explanation for the patterns observed in the data than an analogous model of supply-driven bias. This has important implications for welfare, as news reports which cater to a demand-side preference for biased news can be beneficial to consumers
Persistent Overconfidence and Biased Memory: Evidence from Managers (with David Huffman and Julia Shvets)
American Economic Review, 2022
A long-standing puzzle is whether and how overconfidence can persist in field settings characterized by repeated feedback. This paper studies managers who participate repeatedly in a high-powered tournament incentive system, learning relative performance each time. Using both reduced form and structural methods we find that: (i) managers make overconfident predictions about future performance; (ii) managers have overly-positive memories of past performance; (iii) the two phenomenon are linked at an individual level in a way consistent with models of motivated beliefs.
Intrinsic Information Preferences and Skewness (with Yusufcan Masatlioglu and Yesim Orhun)
Forthcoming, American Economic Review
We present experimental results from a broad investigation of intrinsic preferences for information. We examine whether people prefer negatively skewed or positively skewed information structures when they are equally informative, whether people prefer Blackwell more informative information structures, and how individual preferences over the skewness and the degree of information relate to one another. The wide scope of our investigation not only reveals new insights regarding intrinsic preferences for information, but as we show, also allows for testing of existing models in this domain. We find that models based on the framework of Kreps and Porteus (1978) and Caplin and Leahy (2001), are the most consistent with the data we observe.
Working Papers
Weighted Linear Discrete Choice (with Chris Chambers, Yusufcan Masatlioglu and Paulo Natenzon)
Revise and Resubmit, American Economic Review
We introduce a new model of stochastic choice. The model modifies the Luce model by adding one additional parameter which reflects salience or other economic frictions involved in choosing. The model is consistent with many classical approaches including random utility as well as preference maximization as in Machina (1985). We characterize our model behaviorally and investigate its comparative statics properties. Finally, we demonstrate the applicability of our model in equilibrium settings where firms can choose price, quality, and advertising. The model generates intuitive closed form solutions in these situations, produces results consistent with data which other models cannot accommodate, and allows for preference parameter identification.
Incentive Complexity, Bounded Rationality and Effort Provision (with Johannes Abeler and David Huffman)
Revise and Resubmit, American Economic Review
Using field and laboratory experiments, we demonstrate that the complexity of incentive schemes and worker bounded rationality can affect effort provision, by shrouding attributes of the incentives. In our setting, complexity leads workers to over-provide effort relative to a fully rational benchmark, and improves efficiency. We identify contract features, and facets of worker cognitive ability, that matter for shrouding. We find that even relatively small degrees of shrouding can cause large shifts in behavior. Our results illustrate important implications of complexity for designing and regulating workplace incentive contracts.
Motivated Optimism and Workplace Risk (with Alain Cohn and Yesim Orhun)
Revise and Resubmit, Economic Journal
We provide field evidence that individuals engage in motivated optimism in the face of impending risk, and that their belief distortions are time- and stake-dependent. Our study leverages exogenous variation in when people are required to return to their workplaces during the COVID-19 pandemic. Among workers currently staying at home, individuals who are temporally closer to returning to their workplace are relatively more optimistic about the increase in infection risk associated with going back. Temporal belief differences are larger for people who are more likely to get severely ill if infected.
Additive-Belief-Based Preferences (with David Dillenberger)
We introduce a new class of preferences - which we call additive-belief-based (ABB) utility - that captures a general yet tractable approach to belief-based utility, and that encompasses many popular models in the behavioral literature. We show that the general class of ABB preferences and two prominent special cases, which allow utility to depend on the level of each period's beliefs but not on changes in beliefs across periods, are fully characterized by suitable relaxations of the standard Independence Axiom. We also identify the intersection of ABB preferences with the class of recursive preferences and characterize attitudes towards the timing of resolution of uncertainty for ABB preferences.
Work In Progress
Girls will be Boys: Gender, Beliefs and Selection in the Field (with David Huffman and Julia Shvets)
The Role of Privilege in Statistical Discrimination (with Itai Ashalgi and Faidra Monachou)
Emotional Inattention (with Lukas Bolte)
Complexity, Communication and Misperception (with Junya Zhou)