Who inflates the bubble? Forecasters and traders in experimental asset markets (Q2291436)

The authors use a laboratory experiment to study how forecasting contributes to mispricing. The experiment design is based on a market where participants interact in groups of 8 companies across 20 rounds. Each company consists of an analyst, who is supplied with information on the asset's future dividends, and a trader, who has an initial endowment of cash and assets and decides on sales and purchases. Companies are not informed about the total amount of cash in the market. Each of the assets earns a dividend \(dt\) that varies across rounds. Cash earns a fixed interest rate \(r.\) In each round, traders earn dividends and interest rate payments, depending on the number of shares and cash they are holding. The asset is traded between companies in a continuous double auction market. The average price of the last round is used as the closing price.\newline In the Baseline treatment, both the task of forecasting and the task of trading are assigned to the same subject. In treatment SamePay, these tasks are separated and assigned to two different subjects, who share the profits from trade. In treatment Accuracy, forecasters receive a payment according to the accuracy of their forecasts. \newline The authors find that the separation of tasks induces some mispricing and paying for accuracy reduces attention towards the fundamental value and generates major and persistent mispricing as well as trend extrapolation. The findings are informative for tracing the sources of mispricing as well as for enhancing financial stability.