We describe a model where agents can access two independent signals and can use no, one or both signals. We assume that while the first signal is costly and informative, i.e., it directly affects the return generation process, the second signal is pure misinformation
and free. We find that large swaths of the agents’ population do not use information in their trading process. This holds especially in parametric configurations that are roughly calibrated to realistic situations. Traders often (learn to) select prudent strategies which, in practice, appear to produce a relatively small demand for equity. As a consequence, the model offers some explanation for the well‐known equity premium puzzle, the general
tendency to hold fewer stocks than normal risk‐aversion measures would suggest.

Paolo Pellizzari is Professor of Mathematics for Economics and Social Sciences at Ca‘ Foscari University in Venice, Italy. His research interests include computational economics, agent‐ based models and complex systems. He is the local coordinator of the ITN project EPOC „Economic Policy in Complex Environments“ (https://epoc‐itn.eu/) and is currently the president of the Economic Library at Ca‘ Foscari.

Wednesday, April 19th 2023 | 3pm | N.2.35