Using Virtual Stock Markets as a Research Tool: Insights from Marketing and Management Science

Background

VSM are similar to regular stock markets; however, the listed stocks do not represent company shares. Rather, shares are tied to the outcomes of future events, e.g., market shares of a given product or shares of votes in an election. Every stock has a fixed lifetime after which the actual outcome of the predicted event can be measured. The final value of the stock is determined accordingly. During the lifetime of the stock market, traders compare the current market prices with their individual expectations of the outcome and make trades accordingly. Supply and demand determine the prices of stocks.² Following the logic of the Hayek hypothesis (von Hayek, 1945) and information efficiency hypothesis (Fama, 1970), the resulting market prices reflect the traders’ aggregate expectations of the future events to which the stocks are tied. According to von Hayek (1945), this information aggregating mechanism works efficiently even in the extreme case of all market participants holding diverging information.

VSM allow participants in a market to bet on specific outcomes of an event. For instance, a participant can buy a contract that pays $1 if the next president is a Republican. This contract can be traded among participants within the VSM. Until the election is decided, the stock will trade for an amount less than or equal to the maximum payment of the contract ($1). If the contract trades at the hypothetical value of $0.70, this would mean that the group assesses the probability that the next President is a Republican at 70% (Borison & Hamm, 2010). It is important to note that values above 0.50 do not imply that an event will definitely occur. Instead, a value of 0.70 implies that if the event (in this case the election) were repeated 100 times, the election of a Republican would be estimated to occur 70 times. The first application of VSM was the Iowa Presidential Stock Market (Forsythe et al., 1992). In this example, virtual stocks representing the vote shares of the candidates in the 1988 Presidential election were traded. Actual outcomes could be predicted very precisely. Forecasts based on VSM outperformed every pollster’s forecast in terms of prediction accuracy and low fluctuation levels in forecasts prior to the election. Currently, this principle is being applied to a variety of situations, including the success of Hollywood movies (Karniouchina, 2011), outcomes of sporting events (Borghesi, 2012), and success of corporate projects (Chen, Gao, Goldstein, & Kash, 2011).

VSM have been able to repeat this remarkable performance in subsequent implementations (e.g., Berg et al., 2008), sparking academic interest and laying the basis for different applications in related fields, most importantly in business research and practice.