Thomas Graeber

Photo Thomas Graeber I am an Assistant Professor at Harvard Business School.
My research is in behavioral and experimental economics.

I teach Negotiation in the MBA curriculum.

[cv]     [Harvard faculty page]     [contact: tgraeber@hbs.edu]

Office location: Baker Library | Bloomberg Center 440



Current Research

Complexity and Time
with Benjamin Enke and Ryan Oprea
[abstract]   [pdf]   [latest version: March 10, 2023]

We provide experimental evidence that core intertemporal choice anomalies – including extreme short-run impatience, structural estimates of present bias, hyperbolicity and transitivity violations – are driven by complexity rather than time or risk preferences. First, all anomalies also arise in structurally similar atemporal decision problems involving valuation of iteratively discounted (but immediately paid) rewards. These computational errors are strongly predictive of intertemporal decisions. Second, intertemporal choice anomalies are highly correlated with indices of complexity responses including cognitive uncertainty and choice inconsistency. We show that model misspecification resulting from ignoring behavioral responses to complexity severely inflates structural estimates of present bias.

Stories, Statistics, and Memory
Revise & Resubmit at The Quarterly Journal of Economics
with Chris Roth and Florian Zimmermann
[abstract]   [pdf]   [latest version: April 7, 2023]

For most decisions, we encounter relevant information over the course of days, months or years. We consume such information in various forms, including collections of data shown in numbers – statistics – and anecdotes about individual instances – stories. This paper proposes that the information type – story versus statistic – shapes selective memory. In controlled experiments, we document a pronounced story-statistic gap in memory: the average impact of stories on beliefs fades by 33% over the course of a day, but by 73% for statistics. Consistent with a model of similarity and interference in memory, prompting contextual associations with statistics improves recall. A set of mechanism experiments reveals that lower similarity of stories to interfering information is a key force behind the story-statistic gap.

The Complexity of Economic Decisions
with Xavier Gabaix
[abstract]   [pdf]   [latest version: July 10, 2023]

We propose a theory of the complexity of economic decisions. Leveraging a macroeconomic framework of production functions, we conceptualize the mind as a cognitive economy, where a task’s complexity is determined by its composition of cognitive operations. Complexity emerges as the inverse of the total factor productivity of thinking about a task. It increases in the number of importance-weighted components and decreases in the degree to which the effect of one or a few components on the optimal action dominates. Higher complexity generates larger decision errors and behavioral attenuation to variation in problem parameters. The model applies both to continuous and discrete choice. We develop a theory-guided experimental methodology for measuring subjective perceptions of complexity that is simple and portable. A series of experiments test and confirm the central predictions of our model for perceptions of complexity, behavioral attenuation, and decision errors. We apply our framework to core economic decision domains, including the complexity of static consumption choice with one or several interacting goods, consumption over time, the tax system, forecasting, and discrete choice between goods or lotteries. These applications demonstrate how our approach to complexity can be used in empirical and theoretical work.

Cognitive Uncertainty  
Forthcoming in The Quarterly Journal of Economics
with Benjamin Enke
[abstract]   [pdf]   [latest version: March 14, 2023]

This paper documents the economic relevance of measuring cognitive uncertainty: people’s subjective uncertainty over their ex-ante utility-maximizing decision. In a series of experiments on choice under risk, the formation of beliefs and forecasts of economic variables, we show that cognitive uncertainty predicts various systematic biases in economic decisions. When people are cognitively uncertain – either endogenously or because the problem is designed to be complex – their decisions are heavily attenuated functions of objective probabilities, which gives rise to average behavior that is regressive to an intermediate option. This insight ties together a wide range of empirical regularities in behavioral economics that are typically viewed as distinct phenomena or even as reflecting preferences, including the probability weighting function in choice under risk; base rate insensitivity, conservatism and sample size effects in belief updating; and predictable overoptimism and -pessimism in forecasts of economic variables. Our results offer a blueprint for how a simple measurement of cognitive uncertainty generates novel insights about what people find complex and how they respond to it.

Intertemporal Altruism
Forthcoming in The American Economic Journal: Microeconomics
with Felix Chopra and Armin Falk
[abstract]   [pdf]   [latest version: August 13, 2022]

Most prosocial decisions involve intertemporal tradeoffs. Yet, the timing of prosocial utility flows is ambiguous and bypassed by most models of other-regarding preferences. We study the behavioral implications of the time structure of prosocial utility, leveraging a conceptual distinction between consequence-dated and choice-dated utility flows. We conduct a high-stakes donation experiment that comprehensively characterizes discounting behavior in self-other tradeoffs and allows us to identify different prosocial motives from their distinct time profiles. Our data can only be explained by a combination of choice- and consequence- dated prosocial utility. Both motives are pervasive and negatively correlated at the individual level.

Heterogeneity of Gain-Loss Attitudes and Expectations-Based Reference Points
with Pol Campos-Mercade, Lorenz Goette, Alex Kellogg and Charles Sprenger
[abstract]   [pdf]   [latest version: August 4, 2022]

Existing tests of reference-dependent preferences assume universal loss aversion. This paper examines heterogeneity in gain-loss attitudes, and explores its implications for identifying models of the reference point. In two experimental settings we measure gain-loss attitudes, and then study a canonical treatment effect which distinguishes different models of the reference point. Accounting for measurement error, we document substantial heterogeneity in gain-loss attitudes, with approximately three-quarters loss-averse subjects. We then document heterogeneous treatment effects over gain-loss attitudes consistent with formulations of expectations-based reference points. Our findings provide foundational support for reference points derived from expectations, and explain inconsistencies across prior exercises.


Publications

Confidence, Self-Selection and Bias in the Aggregate  
with Benjamin Enke and Ryan Oprea
(The American Economic Review, 2023, vol. 113 (7), pp. 1933-1966)
[abstract]   [pdf]

The influence of behavioral biases on aggregate outcomes depends in part on self-selection: whether rational people opt more strongly into aggregate interactions than biased individuals. In betting market, auction and committee experiments, we document that some errors are strongly reduced through self-selection, while others are not affected at all or even amplified. A large part of this variation is explained by differences in the relationship between confidence and performance. In some tasks, they are positively correlated, such that self-selection attenuates errors. In other tasks, rational and biased people are equally confident, such that self-selection has no effects on aggregate quantities.

Inattentive Inference  
(Journal of the European Economic Association, 2023, vol. 21 (2), pp. 560–592)
[abstract]   [pdf]

This paper studies how people infer a state of the world from information structures that include additional, payoff-irrelevant states. For example, learning from a customer review about a product’s quality requires accounting for the reviewer’s otherwise-irrelevant taste. This creates an attribution problem common to all information structures with multiple causes. We report controlled experimental evidence for pervasive overinference about states that affect utility – a form of “omitted variable bias” in belief updating –, providing an explanation for various misattribution patterns. In studying why systematic misattribution arises, we consistently find that errors are not due to deliberate effort avoidance or a lack of cognitive capacity. Instead, people behave as if they form incomplete mental models of the information structure and fail to notice the need to account for alternative causes. These mental models are not stable but context-dependent: misattribution responds to a variety of attentional manipulations, but not to changes in the costs of inattention.

Measuring the Scientific Effectiveness of Contact Tracing: Evidence from a Natural Experiment
with Thiemo Fetzer
(Proceedings of the National Academy of Sciences, 2021, vol. 118 (33))
[abstract]   [pdf]

Contact tracing has for decades been a cornerstone of the public health approach to epidemics, including Ebola, severe acute respiratory syndrome, and now COVID-19. It has not yet been possible, however, to causally assess the method’s effectiveness using a randomized controlled trial of the sort familiar throughout other areas of science. This study provides evidence that comes close to that ideal. It exploits a large-scale natural experiment that occurred by accident in England in late September 2020. Because of a coding error involving spreadsheet data used by the health authorities, a total of 15,841 COVID-19 cases (around 20% of all cases) failed to have timely contact tracing. By chance, some areas of England were much more severely affected than others. This study finds that the random breakdown of contact tracing led to more illness and death. Conservative causal estimates imply that, relative to cases that were initially missed by the contact tracing system, cases subject to proper contact tracing were associated with a reduction in subsequent new infections of 63% and a reduction insubsequent COVID-19–related deaths of 66% across the 6 wk following the data glitch.

Bayesian signatures of confidence and central tendency in perceptual judgment  
with Yang Xiang, Benjamin Enke and Samuel Gershman  
(Attention, Perception & Psychophysics, 2021)
[abstract]   [pdf]

This paper theoretically and empirically investigates the role of Bayesian noisy cognition in perceptual judgment, focusing on the central tendency effect: the well-known empirical regularity that perceptual judgments are biased towards the center of the stimulus distribution. Based on a formal Bayesian framework, we show that measures of subjective confidence can be used to explain the central tendency effects and response variability through a Bayesian lens. Specifically, our model clarifies that lower subjective confidence as a measure of posterior uncertainty about a judgment should predict (i) a lower sensitivity of magnitude estimates to objective stimuli; (ii) a higher sensitivity to the mean of the stimulus distribution; (iii) a stronger central tendency effect at higher stimulus magnitudes; and (iv) higher response variability. To test these predictions, we collect a tailored large-scale experimental data set and additionally re-analyze perceptual judgment data from several previous experiments. Across data sets, subjective confidence is strongly predictive of the central tendency effect and response variability, both correlationally and when we exogenously manipulate the magnitude of sensory noise. Our results lend support to Bayesian explanations of both confidence and the central tendency effect.

Delayed Negative Effects of Prosocial Spending on Happiness
with Armin Falk
(Proceedings of the National Academy of Sciences, 2020, vol. 117 (12), pp. 6463-6468)
[abstract]   [pdf]  
Does prosocial behavior promote happiness? We test this longstanding hypothesis in a behavioral experiment that extends the scope of previous research. In our Saving a Life paradigm, every participant either saved one human life in expectation by triggering a targeted donation of 350 euros or received an amount of 100 euros. Using a choice paradigm between two binary lotteries with different chances of saving a life, we observed subjects’ intentions at the same time as creating random variation in prosocial outcomes. We repeatedly measured happiness at various delays. Our data weakly replicate the positive effect identified in previous research but only for the very short run. One month later, the sign of the effect reversed, and prosocial behavior led to significantly lower happiness than obtaining the money. Notably, even those subjects who chose prosocially were ultimately happier if they ended up getting the money for themselves. Our findings revealed a more nuanced causal relationship than previously suggested, providing an explanation for the apparent absence of universal prosocial behavior.