Acquisition of costly information:: an experimental study

I study decision making in a situation where subjects can buy information in order to uncover the initially unknown deterministic structure. The majority of subjects fail to find the most economical way of choosing between two possible prospects. Many subjects underestimate the value of information and, as a result, make avoidable forecasting errors. Other subjects learn correctly which action is most successful under any given condition but fail to implement this knowledge efficiently. They either accept errors for the sake of economizing on small amounts of information costs or they invest in redundant information. The measured losses are substantial. However, in certain settings the fact that half of the economist-subjects and one-fourth of non-economist-subjects learn to behave optimally may be sufficient for efficient market functioning. The present findings thus help to explain the (near-)competitive outcomes observed in many experimental and actual markets. (C) 2001 Elsevier Science B.V. All rights reserved.