Survey_FT_2011

Abstract

Our experimental design differs in a number of important ways from former experi- ments. Most importantly, in all previous experiments, subjects had symmetric information. Obviously, the degree of divergence of opinion cannot be varied in a symmetric information setting. Another important difference is that the experiments in three of the four previous studies featured a long-lived asset with a fundamental value that declined through the course of the experiment. Endowments were not re-initialized. Consequently, a subject exhausting her short selling capacity in one period was unable to sell in the next period. In contrast, our expe- riments consisted of stationary replications of a one-period economy. We conducted 18 experimental sessions in which a total of 180 subjects participated.10 Participants were recruited among economics students at the University of Bonn using the on- line recruiting system ORSEE (Greiner 2004). Ten subjects were assigned to one cohort. Each cohort participated in one experimental session that consists of three distinct parts. To allow the subjects to get acquainted with the computerized trading system, the sessions started with three training periods that have been excluded from the analysis. Subsequently, there were 20 trading periods. In 10 of these periods, short selling was prohibited, and in ten periods, it was allowed. We thus chose a within-subjects design, i.e., each cohort faces both the short selling condition and the no short selling condition. To control for order effects, half of the cohorts encountered the short selling condition first, and the other half faced the no short selling condi- tion first. Subjects received a 20 € show-up fee for participation. In addition, at the end of the ses- sion, two periods (one period of the no short selling condition and one period of the short sell- ing condition) were determined randomly. The profit of these periods were converted into Eu- ros at a rate of 20 ECU11 = 1 € and added to (or subtracted from) the show-up fee.