Remy Levin

PhD Candidate – Department of Economics

Research Summary

I am an applied behavioral economist with interests in development economics and microeconomic theory. My research agenda focuses on exploring how the economic preferences of individuals and groups adapt over time to changes in their environments. My research employs a diverse set of empirical tools in close concert with economic theory, and draws on ideas from anthropology and psychology.

Research Statement

Job Market Paper

Adaptive Risk-Taking: Theory and Evidence from Developing Countries

(with Daniela Vidart)

Developing countries exhibit high rates of macroeconomic volatility and low rates of individual risk-taking. In this paper we present evidence for a new channel linking these phenomena: lifetime experiences of macroeconomic volatility directly change individual risk attitudes. We combine two panel data sets from Indonesia and Mexico (n=15,000), containing elicited measures of risk aversion for the same subjects years apart, with state-level real GDP growth time series capturing their lifetime macroeconomic experiences. We find that living through periods of increasing volatility increases measured risk aversion. This effect is robust to controlling for changes in income, wealth, savings, and exposure to violence and natural disasters. Increases in measured risk aversion correlate with decreases in risky behaviors like smoking, migration, and the planting of cash crops. To better understand the mechanisms at play we develop a model of risk-taking adaptation over the life course. In our model a Bayesian expected utility maximizer who is exposed to exogenous background risk learns about the mean and variance of that risk from its realizations over their lifetime, in turn affecting their endogenous risky choice. Our model generates a novel testable prediction: large negative shocks will increase measured risk aversion more than commensurate positive shocks. We find significant evidence for this predicted asymmetry in our Indonesian sample.


Richard Karlsson Linner, Pietro Biroli, Edward Kong,..., Remy Z. Levin,..., Daniel J. Benjamin, Philipp D Koellinger, Jonathan P. Beauchamp. Genome-wide association analyses of risk tolerance and risky behaviors in over 1 million individuals identify hundreds of loci and shared genetic influences. Nature Genetics, 2019.

Humans vary substantially in their willingness to take risks. In a combined sample of over 1 million individuals, we conducted genome-wide association studies (GWAS) of general risk tolerance, adventurousness, and risky behaviors in the driving, drinking, smoking, and sexual domains. Across all GWAS, we identified hundreds of associated loci, including 99 loci associated with general risk tolerance. We report evidence of substantial shared genetic influences across risk tolerance and the risky behaviors: 46 of the 99 general risk tolerance loci contain a lead SNP for at least one of our other GWAS, and general risk tolerance is genetically correlated (|𝑟̂_g| ~ 0.25 to 0.50) with a range of risky behaviors. Bioinformatics analyses imply that genes near SNPs associated with general risk tolerance are highly expressed in brain tissues and point to a role for glutamatergic and GABAergic neurotransmission. We found no evidence of enrichment for genes previously hypothesized to relate to risk tolerance.

Levin, Remy Z., and Chen, Paul. Rethinking the Constitution-Treaty Relationship. International Journal of Constitutional Law, 10(1), 242-260, 2012.

Work in Progress

The Dynamic Decision-Making of Utilitarian Firms

(with Alejandro Nakab)

How do the preferences of individuals who make up the firm affect the firm's dynamic decision-making? We study this question by examining how employees' measured time preferences influence the firm's trade-off between long-term growth and short-term profits. Using longitudinal, linked data on the universe of firms and their employees in Brazil, we show that firms with a higher variance of measured time preferences among managers (1) grow less overall, especially in the long term, and (2) extract more profits for managers, in the form of wages, in the short- and medium-terms. These findings are robust to the inclusion of firm and time fixed effects and numerous controls for time-varying firm characteristics. Our results are consistent with a model of the firm as a utilitarian aggregator of its employees' time preferences, where present-biased behavior of the collective is driven by heterogeneity among rational actors at the individual level.

Climate Change and Risk-Taking Adaptation

(with Wesley Howden)

One of the primary effects of climate change is to expose individuals, particularly those in developing countries, to new sources of environmental risk to which they must adapt. In this paper we study how lifetime experiences of climate change affect individual propensities to take risks. We link a large panel data set from Indonesia containing experimental measures of risk aversion for the same subjects years apart with data on subjects' lifetime experiences of heat, and examine the effects of changes in experienced heat on risk attitudes at different temporal frequencies. We find that individuals who live through periods of increasing temperature mean become measurably less risk averse, while those who live through increases in temperature volatility become measurably more risk averse.

The Physiognomy of Cooperation: Facial Communication in Strategic Interaction

Human cooperation underpins the organization of every level of society. Lab studies routinely observe 40\%-60\% rates of cooperative behavior in the one-shot prisoner's dilemma, in contrast to theoretical predictions of fully selfish play. This observed behavioral heterogeneity suggests that at least in the short-run, first impressions matter. One important way people form first impressions is by observing the expression and structure of others' faces. I design a series of experiments to study how facial communication shapes cooperation. Using a combination of online data collection, machine learning, and lab experiments I will (1) establish whether subjects have consistent facial perceptions of cooperativeness in the one-shot PD; (2) uncover the correlation structure between these and facial perceptions of economic preferences, psychological measures, and observed demographics; (3) test the ability of these perceptions to predict actual behavior and preferences; and (4) generate clean causal estimates for the effects of one- and bi-directional facial communication on cooperation in the lab. To achieve this last goal, I will train a Generative Adversarial Network (GAN) on the labeled data from an online survey. Using the GAN I will create photo-realistic simulated facial images which are perceived to be more cooperative without varying along other facial perception dimensions, for use as treatment stimuli.