Education

Postdoctoral Fellow (2019 – current)
Harvard Kennedy School, Harvard University

Visiting Scholar (2017 – 2018)
The Wharton School, University of Pennsylvania

PhD in Behavioral Decision Research (2012 – 2018)
Department of Social and Decision Sciences, Carnegie Mellon University

About

David Hagmann is a Postdoctoral Fellow at the Harvard Kennedy School at Harvard University. He received his PhD in Behavioral Decision Research from the Department of Social and Decision Sciences at Carnegie Mellon University. David’s research looks at the desire to avoid information that may be painful to learn yet improve decision-making, persuasion on politically charged topics, and behavioral policy interventions (nudges).

Research Publications

[*] Denotes alphabetical ordering of authors, as is customary in economics

Measuring Information Preferences
Management Science (in press)
Emily Ho, David Hagmann, and George Loewenstein
( Abstract ) (PDF) (OSF)
(Select Media: Scientific American)

Nudging Out Support For a Carbon Tax
Nature Climate Change (2019)
David Hagmann, Emily H. Ho, and George Loewenstein
( Abstract ) (PDF) (OSF)
(Select Media: ArsTechnica, Grist)

A Behavioral Blueprint For Improving Health Care Policy 
Behavioral Science & Policy (2017)
George Loewenstein, David Hagmann, Janet Schwartz, Keith Ericson, Judd B. Kessler, Saurabh Bhargava, Jennifer Blumenthal-Barby, Thomas D’Aunno, Ben Handel, Jonathan Kolstad, David Nussbaum, Victoria Shaffer, Jonathan Skinner, Peter Ubel, & Brian J. Zikmund-Fisher
( Abstract ) (PDF)

Information Avoidance
Journal of Economic Literature (2017)
[*] Russell Golman, David Hagmann, and George Loewenstein
( Abstract ) (PDF)
(Select Media: American Economic Association, YANSS Podcast)

Polya’s Bees: A Model of Decentralized Decision Making
Science Advances (2015)
[*] Russell Golman, David Hagmann, and John H. Miller
( Abstract ) (PDF)

Warning: You Are About to be Nudged 
Behavioral Science & Policy (2015)
George Loewenstein, Cindy Bryce, David Hagmann, and Sachin Rajpal
( Abstract ) (PDF)

Endogenous Movement and Equilibrium Selection in Spatial Coordination Games
Computational Economics (2014)
David Hagmann, and Troy Tassier
( Abstract ) (PDF)

Working Papers

Take My Word For It: Personal Narratives and the Suspension of Disbelief
David Hagmann, Julia A. Minson, and Catherine Tinsley
( Abstract )

The Agent-Selection Dilemma in Ultimatum Bargaining
[*] Daniel Feiler and David Hagmann
( Abstract ) (PDF) (OSF)

Tax Information Avoidance
[*] David Hagmann, Alex Rees-Jones, and Jeffrey Yang
( Abstract )

Persuasion With Motivated Beliefs
David Hagmann and George Loewenstein
( Abstract )

Wait, Wait… Don’t Tell Me: Learning More From Less Feedback
David Hagmann, Jason L. Harman, and Cleotilde Gonzalez
( Abstract )

Fear and Promise of the Unknown: Explore-Exploit Decisions in the Presence of Losses
[*] Alycia Chin, David Hagmann, and George Loewenstein
( Abstract )

Book Chapters

The Deep Structure of Deliberate Ignorance: Mapping the Terrain
MIT Press (in press)
Barry Schwartz, Peter J. Richerson, Benjamin E. Berkman, Jens Frankenreiter, David Hagmann, Derek M. Isaacowitz, Thorsten Pachur, Lael J. Schooler, and Peter Wehling
In: Deliberate Ignorance: Choosing Not to Know, edited by R. Hertwig and C. Engel. Strüngmann Forum Reports, vol. 29, J. R. Lupp, series editor. Cambridge, MA.
(PDF) (Book)

Measuring Information Preferences

Advances in medical testing and widespread access to the internet have made it easier than ever to obtain information. Yet, when it comes to some of the most important decisions in life, people often choose to remain ignorant for a variety of psychological and economic reasons. We design and validate an information preferences scale to measure an individual’s desire to obtain or avoid information that may be unpleasant but could improve future decisions. The scale measures information preferences in three domains that are psychologically and materially consequential: consumer finance, personal characteristics, and health. In three studies incorporating responses from over 2,300 individuals, we present tests of the scale’s reliability and validity. We show that the scale predicts a real decision to obtain (or avoid) information in each of the domains as well as decisions from out-of-sample, unrelated domains. Across settings, many respondents prefer to remain in a state of active ignorance even when information is freely available. Moreover, we find that information preferences are a stable trait but that an individual’s preference for information can differ across domains.

Nudging Out Support For a Carbon Tax

A carbon tax is widely accepted as the most effective policy for curbing carbon emissions but is controversial because it imposes costs on consumers. An alternative, ‘nudge,’ approach promises smaller benefits but with much lower costs. However, nudges aimed at reducing carbon emissions could have a pernicious indirect effect if they offer the promise of a ‘quick fix’ and thereby undermine support for policies of greater impact. Across six experiments, including one conducted with individuals involved in policymaking, we show that introducing a green energy default nudge diminishes support for a carbon tax. We propose that nudges decrease support for substantive policies by providing false hope that problems can be tackled without imposing considerable costs. Consistent with this account, we show that by minimizing the perceived economic cost of the tax and disclosing the small impact of the nudge, eliminates crowding-out without diminishing support for the nudge.

A Behavioral Blueprint For Improving Health Care Policy

Behavioral policy to improve health and health care often relies on interventions, such as nudges, which target individual behaviors. But the most promising applications of behavioral insights in this area involve more far-reaching and systemic interventions. In this article, we propose a series of policies inspired by behavioral research that we believe offer the greatest potential for success. These include interventions to improve health-related behaviors, health insurance access, decisions about insurance plans, end-of-life care, and rates of medical (for example, organ and blood) donation. We conclude with a discussion of new technologies, such as electronic medical records and web- or mobile-based decision apps, which can enhance doctor and patient adherence to best medical practices. These technologies, however, also pose new challenges that can undermine the effectiveness of medical care delivery.

Information Avoidance

We commonly think of information as a means to an end. However, a growing theoretical and experimental literature suggests that information may directly enter the agent’s utility function. This can create an incentive to avoid information, even when it is useful, free, and independent of strategic considerations. We review research documenting the occurrence of information avoidance, as well as theoretical and empirical research on reasons why people avoid information, drawing from economics, psychology, and other disciplines. The review concludes with a discussion of some of the diverse (and often costly) individual and societal consequences of information avoidance.

Polya’s Bees: A Model of Decentralized Decision-Making

How do social systems make decisions with no single individual in control? We observe that a variety of natural systems, including colonies of ants and bees and perhaps even neurons in the human brain, make decentralized decisions using common processes involving information search with positive feedback and consensus choice through quorum sensing. We model this process with an urn scheme that runs until hitting a threshold, and we characterize an inherent tradeoff between the speed and the accuracy of a decision. The proposed common mechanism provides a robust and effective means by which a decentralized system can navigate the speed-accuracy tradeoff and make reasonably good, quick decisions in a variety of environments. Additionally, consensus choice exhibits systemic risk aversion even while individuals are idiosyncratically risk-neutral. This too is adaptive. The model illustrates how natural systems make decentralized decisions, illuminating a mechanism that engineers of social and artificial systems could imitate.

Warning: You are About to be Nudged

Presenting a default option is known to influence important decisions. That includes decisions regarding advance medical directives, documents people prepare to convey which medical treatments they favor in the event that they are too ill to make their wishes clear. Some observers have argued that defaults are unethical because people are typically unaware that they are being nudged toward a decision. We informed people of the presence of default options before they completed a hypothetical advance directive, or after, then gave them the opportunity to revise their decisions. The effect of the defaults persisted, despite the disclosure, suggesting that their effectiveness may not depend on deceit. These findings may help address concerns that behavioral interventions are necessarily duplicitous or manipulative.

Endogenous Movement and Equilibrium Selection in Spatial Coordination Games

We study the effects of agent movement on equilibrium selection in network based spatial coordination games with Pareto dominant and risk dominant Nash equilibria. Our primary interest is in understanding how endogenous partner selection on networks influences equilibrium selection in games with multiple equilibria. We use agent based models and best response behaviors of agents to study our questions of interest. In general, we find that allowing agents to move and choose new game play partners greatly increases the probability of attaining the Pareto dominant Nash equilibrium in coordination games. We also find that agent diversity increases the ability of agents to attain larger payoffs on average.

Establishing Trust Despite Disagreement Using Personal Narratives

Working with those who hold different political views can be challenging, as they are often perceived as less capable and less trustworthy than those who share one's own views. Across six experiments, we find that those who share personal stories about their subjective experience are rated as more trustworthy than those who convey objective information based on data. This holds in naturalistic communication (Study 1) as well as when we closely control the informational content of the messages. In two studies, we find that participants view those sharing personal narratives as more trustworthy than those sharing either data-driven or anecdotal accounts (Study 2A), and that trustworthiness is greater when the narrative reveals hardship of the sender (Study 2B). Trust remains high when the narrative is combined with objective data (Study 2C) and the extent to which the speaker reveals vulnerability partially mediates the effect of narrative on trustworthiness (Study 3). Finally, we show that people are more willing to work with someone who shared a personal narrative on a trust task, but with someone who shared a data-driven argument when the task involves quantitative abilities (Study 4). While data-driven arguments have the potential to be more informative, narratives establish trust, which in many settings may be a more important goal.

The Agent-Selection Dilemma

Many disputes and negotiations end with the parties failing to reach a mutually acceptable agreement. Such costly impasses may arise because both sides interpret information in a way that is favorable to them, creating a disagreement about what a fair resolution is. In high stakes negotiations, principals frequently retain agents to act on their behalf. These agents may be less biased than the principals themselves, giving hope that the principals' self-serving bias may not be as consequential as believed. We present evidence from two large (n = 1,288), pre-registered experiments showing the opposite: although agents are indeed less biased than the principals on average, principals do not hire agents randomly: instead, they choose those who provide them with a favorable estimate of the negotiation outcome. We introduce a novel “ultimatum negotiation game” that allows us to establish the counter-factual outcome that would have been achieved, had the principal retained a different agent. We use empirical data from the agents to show there is a unique Nash Equilibrium strategy for the principals. Moreover, we show that principals systematically choose agents who are more favorable to them than the equilibrium choice and that both parties could improve their outcomes by deviating unilaterally.

The Agent-Selection Dilemma

Many disputes and negotiations end with the parties failing to reach a mutually acceptable agreement. Such costly impasses may arise because both sides interpret information in a way that is favorable to them, creating a disagreement about what a fair resolution is. In high stakes negotiations, principals frequently retain agents to act on their behalf. These agents may be less biased than the principals themselves, giving hope that the principals' self-serving bias may not be as consequential as believed. We present evidence from two large (n = 1,288), pre-registered experiments showing the opposite: although agents are indeed less biased than the principals on average, principals do not hire agents randomly: instead, they choose those who provide them with a favorable estimate of the negotiation outcome. We introduce a novel “ultimatum negotiation game” that allows us to establish the counter-factual outcome that would have been achieved, had the principal retained a different agent. We use empirical data from the agents to show there is a unique Nash Equilibrium strategy for the principals. Moreover, we show that principals systematically choose agents who are more favorable to them than the equilibrium choice and that both parties could improve their outcomes by deviating unilaterally.

Persuasion with Motivated Beliefs

Considerable research finds that people derive utility not only from consumption, but also from their beliefs about themselves and the world. Rather than dispassionately updating their views in response to new information, such belief-based utility leads people to avoid information and use other strategies to protect their existing beliefs. We present a two-stage model of persuasion in the presence of belief-protecting strategies and test it in an incentive compatible persuasion experiment. Persuaders seek to shift receivers’ numeric estimates related to emotionally charged topics, such as abortion and racial discrimination. We manipulate whether the persuader first acknowledges her own lack of certainty and whether she first has an opportunity to build rapport with the receiver, which our theory predicts should enhance persuasiveness, but should be irrelevant or may even go in the opposite direction under the standard account.

Wait, Wait... Don't Tell Me: Learning More From Less Feedback

People frequently learn about the options available to them through personal experience, receiving feedback about the outcomes of their decisions. Previous research suggests that providing immediate feedback faciliates learning, steering people quickly away from bad options. We propose that immediate feedback may lead to worse choices when decision-makers can determine when to terminate the learning process. With uncertain outcomes, early negative experiences may discourage people from exploring further and realizing the option's expected value. To outside observers, this would look like risk aversion, but it would in fact not be driven by preferences, but by a mistaken belief about what the option is. In such cases, withholding feedback can promote learning, lead to more accurate beliefs about the options, and ultimately to choices consistent with those made with full information. In three experiments, we present participants with a choice between an option with a guaranteed payoff and a binary lottery with a higher expected value. We find that with immediate feedback, people are less likely to choose the risky option when they have to learn from experience than when they are told what the options are in advance. When feedback is delayed, however, this gap narrows (Study 1) or closes entirely (Studies 2 and 3). Moreover, the benefit of withholding feedback is greatest when the lottery returns a high outcome with low probability, that is when a decision-maker may have to persist despite multiple adverse outcomes to learn about the option's high potenial outcome.

Fear and Promise of the Unknown: Explore-Exploit Decisions in the Presence of Losses

In “explore-exploit” situations, decision makers choose between taking a risk on an unknown option (exploration) and taking advantage of a known option (exploitation). We study how explore-exploit behavior varies in the presence of losses, predicting that people will be (a) less likely to exploit when incurring losses and (b) less likely to explore when they anticipate incurring losses. To examine these predictions, we created a task in which participants explored a one-dimensional environment, choosing in each period to either exploit a known location or move to a neighboring location. To study the influence of losses, we randomly assigned participants to encounter, via a framing manipulation, only positive payoffs or both positive and negative payoffs. In two studies, we find support for our predictions and demonstrate that loss aversion can be adaptive, leading participants in certain environments to higher total earnings.