Data for: To see is to believe: Common expectations in experimental asset markets

Published: 21 November 2016| Version 1 | DOI: 10.17632/w6kdfhrwpm.1
Stephen L. Cheung, Morten Hedegaard, Stefan Palan


Abstract of associated article: We experimentally manipulate agents' information regarding the rationality of others in a setting in which previous studies have found irrationality to be present, namely the asset market experiments introduced by Smith et al. (1988). Recent studies suggest that mispricing in such markets may be an artefact of confusion, which can be reduced by training subjects to understand the diminishing fundamental value. We reconsider this view, and propose that when it is made public knowledge that training has occurred, this may also reduce uncertainty over the behavior of others and facilitate the formation of common expectations. Our design disentangles the direct effect of training from the indirect effect of its public knowledge, and our results demonstrate a distinct and statistically significant effect of public knowledge over and above that of training alone.


Steps to reproduce

To run these analyses, it should only be necessary to change the file path in the first line, and to run “net install st0134.pkg” if the permtest package is not already installed. This final version of the analysis code was prepared using STATA version 13.1. Please note that some permutation test statistics may differ marginally from the published paper due to numerical approximation.