Home » Research



Working Papers

  • An Evolutionary Perspective on Updating Risk and Ambiguity Preferences (with Philipp Sadowski)
    December 2019, Updated June 2021  [PDF] [SSRN]

    Evolution of preferences refers to the notion that natural selection not only can influence physical traits, but can also shape preferences. Using this approach, we develop a foundation for non-expected-utility and ambiguity-averse preferences and study updating of such preferences in response to information. Evolutionarily optimal preferences will be dynamically consistent, but may violate consequentialism.

  • Risk Attitude Optimization and Heterogeneous Stock Market Participation
    October 2017  [PDF] [SSRN]

    This paper uses a special case of mixture-averse preferences (Sarver, 2018) to study equilibrium portfolio choice and asset returns. We solve for the dynamic stochastic general equilibrium of a calibrated economy and show that the model can generate endogenous heterogeneity in equilibrium stock market participation, even for a population of identical consumers.



  • Naivete about Temptation and Self-Control: Foundations for Recursive Naive Quasi-Hyperbolic Discounting (with David Ahn and Ryota Iijima)
    Journal of Economic Theory, September 2020  [Published Version] [Preprint PDF]

    Behavioral definitions of naivete that can be applied to the Gul-Pesendorfer model of temptation and self-control. Unlike the Strotzian model, the self-control representation is continuous and can be applied recursively. We use our approach to axiomatically characterize recursive naive quasi-hyperbolic discounting.

  • Behavioral Characterizations of Naivete for Time-Inconsistent Preferences (with David Ahn, Ryota Iijima, Yves Le Yaouanq)
    Review of Economic Studies, November 2019  [Published Version] [Supplementary Appendix]

    Behavioral definitions of naivete and sophistication, and comparative measures of naivete. Think Arrow-Pratt, but for naivete instead of risk aversion. We show that our definitions yield tight parametric restrictions for the random Strotz model of time inconsistency, which includes hyperbolic, quasi-hyperbolic, and random quasi-hyperbolic discounting as special cases.

  • Dynamic Mixture-Averse Preferences
    Econometrica, July 2018  [PDF] [Supplementary Appendix]

    A new recursive utility representation where an individual optimally adjusts her risk attitude in response to the uncertainty that she faces. The key axiom is a dynamic form of aversion to mixtures of lotteries. The model permits the disutility from marginal increases in risk to decrease with exposure, and we demonstrate the usefulness of this novel feature in applications to insurance demand, portfolio choice, and the Rabin paradox.

  • Hidden Actions and Preferences for Timing of Resolution of Uncertainty (with Haluk Ergin)
    Theoretical Economics, May 2015  [PDF]

    Extending the analysis of intrinsic preference for timing of information from Kreps-Porteus to include preference for flexibility, we show that this broad class of dynamic preferences can be represented as if the decision maker takes an interim action that is not observable to the analyst. Novel special cases are considered, including unobserved information acquisition.

  • Preference for Flexibility and Random Choice (with David Ahn)
    Econometrica, January 2013  [PDF] [Supplementary Appendix]

    Preferences over option sets and random choice from option sets can both be represented using random utility models. Our model simultaneously considers both pieces of choice data, provides an axiom that ensures consistency of behavior across the two domains, and shows that this combined data can be used to identify both utility and beliefs.

  • The Unique Minimal Dual Representation of a Convex Function (with Haluk Ergin)
    Journal of Mathematical Analysis and Applications, October 2010  [Published Version] [Preprint PDF]

    Fenchel-Moreau duality allows a convex function to be expressed as the supremum of a set of linear functions minus the conjugate (the “intercept”). Under suitable assumptions on the domain, we show that any locally Lipschitz continuous function has a unique minimal closed set of linear functions for which this duality holds.

  • A Unique Costly Contemplation Representation (with Haluk Ergin)
    Econometrica, July 2010  [PDF] [Supplementary Appendix]

    Individuals are often uncertain of their tastes (utility) over available alternatives when making choices, but may be able to engage in contemplation to improve their decisions. We model such behavior and analyze it axiomatically. Our key axiom, aversion to contingent planning, captures the idea that since contemplation is costly, individuals will prefer to economize on cognitive resources by avoiding unnecessary planning.

  • Anticipating Regret: Why Fewer Options May Be Better
    Econometrica, March 2008 (Lead Article)  [PDF]

    Individuals may experience regret if their choices end up being inferior to other available alternatives. Larger choice sets therefore come with both benefits and costs: Adding an option that is certain to be better than anything currently available will clearly benefit an individual. However, adding options that are dominated in expectation at the time of choice, but that may ultimately prove superior ex post, opens the individual up to regret.

  • Representing Preferences with a Unique Subjective State Space: A Corrigendum (with Eddie Dekel, Barton L. Lipman, and Aldo Rustichini)
    Econometrica, March 2007  [PDF] [Supplementary Appendix]

    Correction of Dekel, Lipman, and Rustichini (2001). Check out the supplementary appendix for a complete proof of DLR’s additive expected-utility representation theorem.

  • Correlated Equilibrium in Evolutionary Models with Subpopulations (with Justin Lenzo)
    Games and Economic Behavior, August 2006  [Published Version] [Preprint PDF]

    Classic results in evolutionary game theory draw a connection between limit states in the replicator dynamics and Nash equilibrium. In a setting where individuals are divided into subpopulations, with correlation in how these subpopulations are matched for play, we provide a similar foundation to correlated equilibrium.


Email: todd.sarver@duke.edu

Office: 232 Social Sciences

Mail: Department of Economics
213 Social Sciences
Box 90097
Durham, NC 27708