Schedule - Parallel Session 1 - Other-Regarding Preferences

IMC Room 246 - 11:00 - 12:30

Asymmetric Attention to Opportunity Costs Hinders Generosity

Christopher Olivola; John Han


Altruistic decisions are often thought to result from a salient internal conflict between self-serving and other-regarding motives. We show that such decisions are also governed by default “attentional” factors that operate independently of any internal motivational conflicts. In our studies, participants make a series of choices between receiving money themselves (e.g., “Option A: You receive $4.30”) versus having a larger sum of money go to another party (e.g., “Option B: A homeless person receives $7.50”), and we measure their overall generosity (how frequently they select the option that benefits the other party). We show that adding an “other-nothing” wording to the selfish option which highlights the fact that choosing to receive the money entails a forgone gain for the other party (e.g., “Option A: You receive $4.30 and a homeless person receives nothing”) increases generosity. Critically, however, this effect is not symmetric: adding a “self-nothing” wording to the altruistic option which highlights the fact that being generous entails a forgone gain for self (e.g., “Option B: You receive nothing and a homeless person receives $7.50”) does not decrease (nor increase) generosity. In other words, we demonstrate an asymmetric effect of highlighting forgone gains for self vs. other(s), such that subtle reminders of others’ forgone payoffs increase generosity, whereas equivalent reminders of one’s own forgone payoffs have no effect. We argue, and empirically show, that this asymmetric effect occurs because people are naturally attentive to their own (forgone) gains, but less so to others’ (forgone) gains. Consequently, reminding them of their potential forgone gains has no effect (since these considerations are already highly salient to them), whereas reminding them of others’ potential forgone gains increases generosity by increasing the (otherwise limited) attention given to other people’s needs. We demonstrate this “other-nothing” effect across a variety of contexts, formats, and payoff sizes. We also test the hypothesis that it operates by counterbalancing people’s natural asymmetric attention to their own (over others’) payoffs. For example, we show that the “other-nothing” wording decreases satisfaction with the selfish outcome to a greater extent than the “self-nothing” wording decreases satisfaction with the altruistic outcome. We also provide process evidence that the effect operates by “putting a break” on the default tendency to focus on one’s own outcomes. Finally, we repeatedly rule out experimenter demand as a driver of the effect. In sum, these studies provide converging evidence for a subtle, but consequential, default asymmetry in how people attend to their own vs. others’ outcomes, and they highlight a novel semantic “nudge” that increases generosity by counteracting this tendency. We conclude by discussing the implications of our results for ongoing debates about whether altruism is intuitive vs. reflective.

Christopher Olivola

Assistant Professor, Carnegie Mellon University

Skewness and Preferences for Non-Instrumental Information

Collin Raymond; Yusufcan Masatlioglu; Yesim Orhun


Psychologists have long recognized that the desire to regulate anticipatory emotions regarding an uncertain outcome in the future, such as hope, anxiety and suspense, may influence what people want to learn, and when they want to learn it, even in the absence of their ability to condition their actions on that information. Anecdotal evidence, such as from medical testing, also suggests that try to avoid situations that make them feel anxious about a future event and cultivate instances of hope at the face of uncertainty, even when doing so is costly. We present experimental results from a broad investigation of preferences for non-instrumental information. Most earlier work investigated whether people have preferences that would lead them to obtain non-instrumental information. However, more generally, individuals may care not only about whether they observe information or not, but also about what kind of information they observe. For example, even if they reveal the same amount of information overall, certain information structures eliminate more uncertainty about the undesired outcome conditional on generating a bad signal, but are unlikely to generate a bad signal (i.e., they are negatively skewed). Yet others eliminate more uncertainty about the desired outcome conditional on generating a good signal, but are unlikely to generate a good signal (i.e., they are positively skewed). This paper, in addition to studying whether people prefer more information to less when information is non-instrumental, specifically explores i) whether people prefer negatively skewed information or positively skewed information, and ii) how individual preferences over the skewness and the degree of information relate to one another. Exploring a wider range of non-instrumental information preferences allows us to assess existing models in this domain including Kreps and Porteus (1978), Grant, Kajii and Polak (1998), Caplin and Leahy (2001), Brunnermeier and Parker (2005), Koszegi and Rabin (2009}, Dillenberger (2010), Dillenberger and Segal (2014) and dynamic extensions of Gul (1991) and Quiggin (1982). Although existing evidence, both empirical and experimental, cannot distinguish between predictions of these models, these models make different predictions regarding preferences for skewness. Therefore, we can assess the extent to which existing models can accommodate the data, and identify ways in which they can be modified to capture these behavioral patterns. We find that individuals exhibit a strong preference for positively skewed information structures as well as Blackwell more informative information structures. Moreover, we find that preference for Blackwell more informative information structures dominate preferences for right-skewed signals. These results provide the strongest support for the class of preferences introduced by Kreps and Porteus (1978).

Collin Raymond

Assistant Professor, Amherst College

The Rich, the Poor, and Social Capital

Stefan Trautmann

Stefan Trautmann

Associate Professor, Tilburg School of Economics and Management

Reciprocity Towards Groups and Individuals: A Comparison

David Hugh-Jones; Roi Zultan


Conflicts are frequently driven by cycles of intergroup revenge, in which innocent bystanders are targeted in revenge for the supposed actions of others in the group. We test for the existence of reciprocity towards groups, using an incentivized trust game followed by a sequence of money allocations. We compare the effect of a subject’s partner’s behaviour in the trust game on his or her allocations to three kinds of people: (a) the partner themselves; (b) other people in the partner’s group; and (c) members of a neutral group. Trustors who were betrayed in the trust game display group reciprocity, by allocating less money towards members of the partner’s group. Trustees do not display group reciprocity.

David Hugh-Jones

Senior Lecturer, University of East Anglia