Prediction markets: an experimental approach to forecasting cattle on feed
Abstract
Purpose
The purpose of this paper is to use prediction markets to forecast an agricultural event: United States Department of Agriculture's number of cattle on feed (COF). Prediction markets are increasingly popular forecast tools due to their flexibility and proven accuracy to forecast a diverse array of events.
Design/methodology/approach
During spring 2008, a market was constructed comprised of student traders in which they bought and sold contracts whose value was contingent on the number of COF to be reported on April 18, 2008. During a nine‐week period, students were presented three types of contracts to forecast the number of COF. To estimate forecasts a uniform price sealed bid auction mechanism was used.
Findings
The results showed that prediction markets forecasted 11.5 million head on feed, which was about 1.6 percent lower than the actual number of COF (11.684 million). The prediction market also fared slightly worse than analysts' predictions, which on average suggested there would be about 11.795 million head (an over‐estimate of about 1 percent).
Originality/value
The contribution of this study was not to provide conclusive evidence on the efficacy of using prediction markets to forecast COF, but rather to present an empirical example that will spark interest among agricultural economists on the promises and pitfalls of a research method that has been relatively underutilized in the agricultural economics literature.
Keywords
Citation
Karina Gallardo, R., Wade Brorsen, B. and Lusk, J. (2010), "Prediction markets: an experimental approach to forecasting cattle on feed", Agricultural Finance Review, Vol. 70 No. 3, pp. 414-426. https://doi.org/10.1108/00021461011088521
Publisher
:Emerald Group Publishing Limited
Copyright © 2010, Emerald Group Publishing Limited