Working Paper: NBER ID: w25326
Authors: Christine L. Exley; Judd B. Kessler
Abstract: Distributional decisions regularly involve multiple payoff components. In a series of experiments involving over 3,300 subjects and 81,000 decisions, we find that—even when payoff components can be easily aggregated—many subjects exhibit narrow equity concerns, applying fairness preferences to a single component of payoffs. This behavior leads to preference reversals; subjects make different choices depending on which payoff component is used to denominate their decision. In our simplest setting, in which the two payoff components are small and large tokens, displaying narrow equity concerns is 63%–83% as prevalent as achieving equity in total payoffs and just as prevalent as applying the well-documented norm of a 50/50-split. Subjects also exhibit narrowly equity concerns over payoffs of time and money.
Keywords: narrow equity; distributional decisions; fairness preferences
JEL Codes: C91; D63; H2
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| narrow equity concerns (D63) | preference reversals (D11) |
| token denomination used in decision-making (E42) | preference reversals (D11) |
| narrow equity concerns (D63) | achieving narrow equity (D63) |
| narrow equity concerns (D63) | achieving overall equity (D63) |
| narrow equity concerns (D63) | decision-making process (D70) |
| narrow equity concerns (D63) | robustness across different experimental designs (C90) |