To What Extent Are People Willing to Avoid Uncertainty?

For the first edition of the Jabarian Research Letter, I would like to share the findings of my job market paper, “A Two-Ball Ellsberg Paradox: An Experiment”, co-authored with Ph.D. student Simon Lazarus (Princeton University).

Background

Previous research in economics, specifically Daniel Ellsberg's seminal work, has shown that individuals tend to exhibit ambiguity aversion, i.e., avoid choosing ambiguous options — options for which individuals cannot access precise probabilities of possible outcomes. The literature standardly assumes that this ambiguity aversion stems from the belief that ambiguous options may result in worse outcomes than risky or certain options. Hence, ambiguity aversion is conceptualized as a choice-outcome phenomenon.

New Experimental

Evidence

My job market paper presents experimental evidence from a US representative randomized experiment that individuals prefer dominated risky options over ambiguous ones, even when the latter always yields better outcomes.

Through various experimental tests and treatments, we rule out the possibility that this behavior is merely due to irrationality; we demonstrate that individuals are inherently averse to ambiguous situations, regardless of their associated outcomes. Hence, experimental evidence establishes the relevance of ambiguity aversion as a choice-processing phenomenon.

→Implications

for Science & Policy

Ambiguity aversion is relevant to economists and policymakers since, in most real-world situations, individuals cannot access precise probabilities of possible outcomes. Since Ellsberg’s seminal contribution, standard models have been developed to deal with ambiguity aversion as a choice-outcome phenomenon.

Relying on these models, researchers have explored the important implications of ambiguity in diverse policy settings: for instance, in environmental economics, health economics, and macro-finance. All these standard models can accommodate typical Ellsberg preferences, but they cannot explain the results of our experiment. Hence, the use of such models in the policymaking context is challenged.

Our results suggest the need for new models and applications considering ambiguity aversion as a choice-processing phenomenon for better policy recommendations.

twitter thread

research paper