Jose Apesteguia

Your Morals Depend on Language

Should you sacrifice one man to save five? Whatever your answer, it should not depend on whether you were asked the question in your native language or a foreign tongue so long as you understood the problem. And yet here we report evidence that people using a foreign language make substantially more utilitarian decisions when faced with such moral dilemmas. We argue that this stems from the reduced emotional response elicited by the foreign language, consequently reducing the impact of intuitive emotional concerns. In general, we suggest that the increased psychological distance of using a foreign language induces utilitarianism. This shows that moral judgments can be heavily affected by an orthogonal property to moral principles, and importantly, one that is relevant to hundreds of millions of individuals on a daily basis.

The Impact of Gender Composition on Team Performance and Decision Making: Evidence from the Field

We investigate whether the gender composition of teams affects their economic performance. We study a large business game, played in groups of three, in which each group takes the role of a general manager. There are two parallel competitions, one involving undergraduates and the other involving MBA students. Our analysis shows that teams formed by three women are significantly outperformed by all other gender combinations, both at the undergraduate and MBA levels. Looking across the performance distribution, we find that for undergraduates, three-women teams are outperformed throughout, but by as much as 0.47 of a standard deviation of the mean at the bottom and by only 0.09 at the top. For MBA students, at the top, the best performing group is two men and one woman. The differences in performance are explained by differences in decision making. We observe that three-women teams are less aggressive in their pricing strategies, invest less in research and development, and invest more in social sustainability initiatives than does any other gender combination. This paper was accepted by Brad Barber, Teck Ho, and Terrance Odean, special issue editors.

''Piensa'' twice: On the foreign language effect in decision making

In this article, we assess to what extent decision making is affected by the language in which a given problem is presented (native vs. foreign). In particular, we aim to ask whether the impact of various heuristic biases in decision making is diminished when the problems are presented in a foreign language. To this end, we report four main studies in which more than 700 participants were tested on different types of individual decision making problems. In the first study, we replicated Keysar et al.s (2012) recent observation regarding the foreign language effect on framing effects related to loss aversion. In the second section, we assessed whether the foreign language effect is present in other types of framing problems that involve psychological accounting biases rather than gain/loss dichotomies. In the third section, we studied the foreign language effect in several key aspects of the theory of decision making under risk and uncertainty. In the fourth study, we assessed the presence of a foreign language effect in the cognitive reflection test, a test that includes logical problems that do not carry emotional connotations. The absence of such an effect in this test suggests that foreign language leads to a reduction of heuristic biases in decision making across a range of decision making situations and provide also some evidence about the boundaries of the phenomenon. We explore several potential factors that may underlie the foreign language effect in decision making.

Choice By Sequential Procedures

We propose a rule of decision-making, the sequential procedure guided by routes, and show that three influential boundedly rational choice models can be equivalently understood as special cases of this rule. In addition, the sequential procedure guided by routes is instrumental in showing that the three models are intimately related. We show that choice with a status quo bias is a refinement of rationalizability by game trees, which, in turn, is also a refinement of sequential rationalizability. Thus, we provide a sharp taxonomy of these choice models, and show that they all can be understood as choice by sequential procedures.

A theory of reference dependent behavior

There is extensive field and experimental evidence in a wide variety of environments showing that behavior depends on a reference point. This paper provides an axiomatic characterization for such behavior. Our approach is dual, we study choice behavior and preference relations. We proceed by gradually imposing more structure on behavior, requiring higher levels of rationality, that free the decision-maker from certain types of manipulations. Depending on the phenomena one wants to model, one degree of behavioral structure will be appropriate or another. We provide two applications of the theory: one to model the status-quo bias, and another to model addictive behavior.

A measure of rationality and welfare

Evidence showing that individual behavior often deviates from the classical principle of preference maximization has raised at least two important questions: (1) How serious are the deviations? (2) What is the best way to analyze choice behavior in order to extract information for the purpose of welfare analysis? This paper addresses these questions by proposing a new way to identify the preference relation that is closest, in terms of welfare loss, to the revealed choice.

Imitation and the evolution of Walrasian behavior: Theoretically fragile but behaviorally robust☆

A well-known result by Vega-Redondo (1997) implies that in symmetric Cournot oligopoly, imitation leads to the Walrasian outcome where price equals marginal cost. In this paper, we show that this result is not robust to the slightest asymmetry in fixed costs. Instead of obtaining the Walrasian outcome as unique prediction, every outcome where agents choose identical actions will be played some fraction of the time in the long run. We then conduct experiments to check this fragility. We obtain that, contrary to the theoretical prediction, the Walrasian outcome is still a good predictor of behavior.

A Foundation for Strategic Agenda Voting

We offer complete characterizations of the equilibrium outcomes of two prominent agenda voting institutions that are widely used in the democratic world: the amendment, also known as the Anglo-American procedure, and the successive, or equivalently the Euro-Latin procedure. Our axiomatic approach provides a proper understanding of these voting institutions, and allows comparisons between them, and with other voting procedures.

Minimal Books of Rationales

Kalai, Rubinstein, and Spiegler (2002) propose the rationalization of choice functions that violate the “independence of irrelevant alternatives” axiom through a collection (book) of linear orders (rationales). In this paper we present an algorithm which, for any choice function, gives (i) the minimal number of rationales that rationalizes the choice function, (ii) the composition of such rationales, and (iii) information on how choice problems are related to rationales. As in the classical case, this renders the information given by a choice function completely equivalent to that given by a minimal book of rationales. We also study the structure of several choice procedures that are prominent in the literature.

Monotone stochastic choice models: The case of risk and time preferences

Suppose that, when evaluating two alternatives x and y by means of a parametric utility function, low values of the parameter indicate a preference for x and high values indicate a preference for y. We say that a stochastic choice model is monotone whenever the probability of choosing x is decreasing in the preference parameter. We show that the standard use of random utility models in the context of risk and time preferences may sharply violate this monotonicity property and argue that their use in preference estimation may be problematic. They may pose identification problems and could yield biased estimations. We then establish that the alternative random parameter models are always monotone.