'Outlier Blindness': Efficient Coding Generates an Inability to Represent Extreme Values

Abstract

How do people record information about the economic outcomes they observe in their environment? Building on a well-established neuroscientific framework, we propose a model in which people are not attuned to making distinctions between realized outcomes that they seldom expect to encounter. We provide evidence for such `outlier blindness' in a series of controlled laboratory experiments and discuss important implications of our model for economic decision-making. In particular, we show how outlier blindness provides a microfoundation for well-established empirical features of risk perception such as tail risk neglect and context-dependent risk attitude.

Keywords: Tail risk, Instability, Imprecise perception, Efficient coding, Adaptation, Decision making under uncertainty, Behavioral finance, Neuroeconomics, Experiments

JEL Classification: C91, D87