Aggregation mechanisms for crowd predictions

• DOI: 10.1007/s10683-019-09631-0
• Publication Date: 2019-11-09T05:15:50Z
• Authors: Stefan Palan; Jürgen Huber; Larissa Senninger
• Keywords: Information aggregation; Value (mathematics)

When the information of many individuals is pooled, resulting aggregate often a good predictor unknown quantities or facts. This frequently outperforms forecasts experts even best individual forecast included in aggregation process ("wisdom crowds"). However, an appropriate mechanism considered crucial to reaping benefits "wise crowd". Of possible ways forecasts, we compare (uncensored and censored) arithmetic geometric mean median, continuous double auction market prices sealed bid-offer call controlled experiment. We use asymmetric structure, where participants know different sub-sets total needed exactly calculate asset value be estimated. find that from markets clearly outperform all alternative approaches for aggregating dispersed lets only best-informed generate excess returns.