To Subsidize Or Not to Subsidize: A Comparison of Market Scoring Rules and Continuous Double Auctions for Price Discovery

Abstract

We investigate which of the two most common prediction market mechanisms – market scoring rules and continuous double auctions – leads to better price discovery. The relative contribution of a particular prediction market to price discovery also depends on the total number of trades observed in that market. We use real-world prediction market price data to estimate price discovery shares of each market and their relationship to the number of trades. We find that when the number of trades is low, prediction markets that use market scoring rules have a higher speed of incorporating information into prices. When the number of trade is high, however, the continuous double auctions have a higher price discovery share. As market scoring rules require a variable cost – a subsidy – to operate, and continuous double auctions are costless; our results provide important implications for the design of prediction markets. By combining the methods for measuring price discovery and information aggregation, we introduce a new data-driven approach that can be used by researchers and practitioners to gain further insight into the exact values of the number of trades favoring the use of market scoring rules instead of continuous double auctions or vice versa.