The Design and Testing of Information Aggregation Mechanisms: A Two-Stage Parimutuel IAM

Abstract

The research reported here is focused on the design of new information aggregation mechanisms. These are competitive processes designed for collecting and aggregating dispersed information held in the form of impression and belief, that might otherwise be impossible to get. The research explores alternative institutional forms of IAMs and how they work. That specially designed markets can aggregate information is well documented in the literature as are the problems encountered when parimutuel-type systems are employed in an information aggregation capacity. This research is focused on new mechanisms, unlike any found evolving naturally, which mitigate these problems. These new mechanisms speed the process through which information is revealed, reduce deceptive behavior and reduce the instances of substantially incorrect aggregation (i.e., bubbles). The paper finds that a special, "two-stage" parimutuel mechanism is an improvement over previously studied parimutuel mechanisms. The two-stage parimutuel, on average, makes a prediction closer to that predicted using all available information. The mechanism suffers from fewer mirages (bubbles) than do previous parimutuel structures and it produces indicators for assessing the reliability of the information produced.

Files

Additional details