In almost all walks of life, predicting uncertain future events plays an essential role in decision-making processes. However, information related to future events frequently exists only as dispersed opinions, insights, and intuitions of individuals. Each individual only knows a little, but aggregating the dispersed information together may make considerable contribution to decision making. This is typical in many domains including business, politics, and entertainment. Therefore, how to aggregate such dispersed information for useful decision support is a crucial task.
Markets have shown great potential as one of the most effective mechanisms for gathering distributed information and generating accurate forecasts, often surpassing many existing methods in practice. This research studies information markets, markets that are specially designed for information aggregation and forecasting, from four different perspectives: theoretical examination, experimental evaluation, empirical analysis, and design.
With the ultimate goal of better understanding information markets as a forecasting device, this thesis makes four contributions to the field of information markets. The first contribution is a theoretical model of information markets that generalizes an existing model to situations with aggregate uncertainty, which is ubiquitous in the real world. It helps answering the question of why information markets work, by modeling how information flows from traders to the market and back again, and characterizing convergence properties of information markets.
The second contribution is an experimental evaluation of several theoretical models of information markets. Because theoretical models often have to make simplified assumptions about human behavior for tractableness, we use human subject experiments to test them, while still maintaining close parallel settings with the theoretical models. Results of this part demonstrate whether and to what extent theoretical models are supported in a more realistic environment and point out important areas to be improved by theoretical models.
The third contribution is an initial attempt to compare the prediction accuracy of information markets and opinion pools using real-world market data. The results provide insights into the predictive performance of information markets, and the relative merits of selecting among various opinion pooling methods.
The last contribution of the thesis is a generic framework of information market development. Although evidence has shown that information markets can make accurate predictions, there are certainly cases that markets fail. How to design an information market for accurate predictions in practice remains an open question. To facilitate the development process, the proposed framework illustrates the life cycle of information market development and explains issues to be considered at each stage.