The Predominant Role of Signal Precision in Experimental Beauty Contests

Abstract

The weight assigned to public information in Keynesian beauty contests depends on both the precision of signals and the degree of strategic complementarities. This experimental study shows that the response of subjects to changes in signal precision and the degree of strategic complementarities is qualitatively consistent with theoretical predictions, though quantitatively weaker. The weaker response of subjects to changes in the precision of signals, however, mainly drives the weight observed in the experiment, qualifying the role of strategic complementarities and overreaction in experimental beauty contests.

Experiments on Macroeconomics: Methods and Applications

“…The main result of these experiments is that subjects attribute more weight to public than to private signals, but the weight on public signals is, in general, smaller than predicted by the equilibrium and too small to generate welfare detrimental overreactions. 26 Baeriswyl and Cornand (2016) qualify the previous results in terms of overreaction: They find that the weights that subjects attribute to the more precise signal is smaller than the Bayesian weight. If private signals are more precise than public signals, subjects' behavior gets closer to the equilibrium prediction with its potential negative welfare implications.…”

“…For reasons of accountability and fairness, they argue that central banks should favor limited transparency when they want to avoid overreactions. However, in the light of the results by Baeriswyl and Cornand (2016), the issue only arises, if a central bank publishes information that is less precise than private information. It is remarkable that negative welfare effects of imprecise public information can be avoided if the precision is deliberately reduced by adding idiosyncratic noise.…”

Clarity Trumps Content: An Experiment on Information Acquisition in Beauty Contests

“…Prior literature on beauty-contest games have used level-k models to explain subjects' nonequilibrium behavior (Cornand and Heinemann, 2013;Baeriswyl and Cornand, 2016). Unlike these studies, we have an additional stage of information acquisition.…”

“…Like Nagel (1995) and Stahl and Wilson (1994), we assume that an level-0 action has a uniform distribution over the action space, which, in our case, is the reals. Following Cornand and Heinemann (2013); Baeriswyl and Cornand (2016), we define a level-1 action to be the first order expectation of the state conditional on the player's information set:…”

“…public nature of information in beauty contests. With exogenous information, prior research finds that over-weighting of public information in laboratory experiments is less severe than theoretical predictions (e.g., Cornand and Heinemann, 2014;Baeriswyl and Cornand, 2016). While theory imposes the assumption of infinite level of reasoning, experimental evidence indicates that subjects' behavior is more consistent with level-2 reasoning.…”