Knowledge Markets for Competitive Information

Abstract

Drawing on the research of Gans & Stern, Roth and Van Alstyne, we attempted to create an experimental knowledge market for information that is easily reproducible and highly value rivalrous, and is typically rarely shared. We tested the assumption that best price for information is free, by setting and comparing the performance of algorithmic traders in two different scenarios: (1) where they could exchange algorithms with each other for free and (2) where they could set an asking price for their algorithms and bid on each other's algorithms. While we did not find any statistically significant difference in performance or levels of information exchange between the two groups, we did find that those who elected to buy algorithms in group 2 were able to, on average, generate performance gains in excess of the cost of the algorithm.

Keywords: Market Design, Experimental Economics, Markets for Technology, Repeated Games, Knowledge Markets