Two Experiments to Test a Model of Herd Behaviour

We carry out two experiments to test a model of herd behaviour based on the work of Banerjee (Quarterly Journal of Economics, CVII, 797–817, 1992). He shows that herding occurs as a result of people observing the actions of others and using this information in their own decision rule. In our experiments herding does not occur as frequently as Banerjee predicts. Contrary to his results, the subjects' behaviour appears to depend on the probabilities of receiving a signal and of this signal being correct. Furthermore, Banerjee finds that the pattern of decision making over a number of rounds of the game is volatile whereas we find that decision making is volatile within rounds.