NSF Award
2416935
Economic decision makers face an important constraint: while they need to know consumer preferences to identify welfare improvements, they can only observe choices made. Hence, welfare economics and its resulting implications rest heavily on the idea that choices are driven by one’s preferences. Under the classic model of rationality, a consumer has a stable preference over alternatives and systematically uses it to choose the best affordable option. Because neither decision process nor preference are observable, economists have had to content themselves with choices that, to the naked eye, are consistent with rational behavior under some preference. Choices look “as if they are rational” when, taken altogether, they do not lead to contradictory inferences about preferences (for an example of such a contradiction, imagine selecting option A when option B is affordable under some prices, but switching to option B when A remains affordable under some other prices). Comfortingly, a rich experimental literature establishes the prevalence of as-if rationality in a variety of economic contexts. It may seem natural to presume that as-if rationality is tightly coupled with the use of a rational decision process. The current research challenges this presumption. The researchers use a novel experimental design to show that even when experimental participants fail to maximize their welfare, they tend to use coherent decision rules, or choice procedures, that mimic properties of rational behavior. The ability to disentangle as-if rationality (which may be driven by rules of thumb or heuristics) from truly (i.e., substantively) rational behavior rationality has many important implications that this research investigates. For instance, given prior work establishing that as-if rationality correlates with wealth, is it substantive rationality or heuristic behavior that drives the correlation? Does heuristic behavior disappear given feedback and time to learn? Is the prevalence of substantive rationality responsive to the complexity of a decision problem? This research will also investigate multiple important questions arising in a society with some seemingly-rational consumers. These include how to aggregate choices, elect leaders, and minimize price distortions in markets.<br/><br/>The experimental literature has documented, using various metrics of compliance with the General Axiom of Revealed Preference (GARP), that subjects tend to be highly as-if rational. Yet what drives this GARP compliance is unknown: there is an indeterminate relationship between the subject's welfare and the utility function(s) which seem to capture choices in these experiments. The researchers introduce a novel experimental design that pairs traditional experiments measuring subjects' natural preferences (i.e., their true preferences for risk, time, etc.) with isomorphic experiments that instead induce canonical preferences. Presuming only that money is desirable, this approach removes scope for subjects to rationally express unobserved ``home grown'' preferences: the utility-maximizing choice is the optimal demand under the induced preference. The initial finding, which is robust to multiple metrics of assessing substantive rationality, is that the majority of as-if rational subjects use suboptimal simplifying heuristics which only mimic rational demands. The researchers will extend and deploy this new experimental approach to address multiple questions on the nature and extent of this distinction. Do people become more procedurally rational in complex problems? Does procedural rationality correlate with classic behavioral biases? Which forms of rationality vary with experience? How do different forms of rationality correlate with socio-economic variables? The researchers will also experimentally investigate key questions relating to group rationality and the aggregation of choices. Is the pivotal agent in group decisions substantively rational? Are prices distorted in the presence of procedurally-rational agents, thereby impacting everyone’s welfare (including those who are substantively rational)? Does the fraction of substantively-rational choices decrease in strategic voting if being pivotal is unlikely? Do voters recognize and value candidates’ degrees of rationality? What becomes of the Coase theorem when bargainers fail to be substantively rational? The results have the potential to profoundly change how economists envision welfare economics, and to make significant contributions in applications to general equilibrium, strategic voting, and bargaining.<br/><br/>This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.
