MInPuD: Method Integration of Prediction Market and Delphi Study

• Applicant: Professor Dr. Christof Weinhardt
• Subject Area: Accounting and Finance
• Term: from 2014 to 2017
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 256547774

Prediction errors in the global financial and economic crises are but one example for the failure of traditional, time-series-based prediction methods in the light of abrupt changes. The objective of this project is studying whether the integration of the established feedback mechanisms Delphi study and prediction market can improve the quality of predictions in general -- and among other application scenarios the forecasting of economic indicators. It is well known that prediction markets are potent tools to forecast simple events. It is, however, not known how well they can handle complex interdependencies. In any case, they allow for measuring the prediction expertise of individual participants. These participants can then be invited and incentivized as experts in a Delphi study. Delphi studies are able to elicit more complex information settings and interdependencies than prediction markets. Thereby, they can qualitatively support market-based predictions. An integration of both methods alongside an intuitive design of the overall information system will result in a substantial increase in acceptance and activity and thereby improve prediction quality. To this end, we will develop a real-time Delphi tool and integrate it with the established EIX prediction market. The hypothesized improvement in prediction quality will be tested in laboratory and field experiments. Users' acceptance of the integrated system will be evaluated via extended Technology Acceptance Models.

• DFG Programme: Research Grants