A test for comparative income effects in an ultimatum bargaining experiment

We test Bolton's (1991) comparative bargaining model by conducting an ultimatum experiment with two primary treatments distinguished only by their payoff rules. In the first treatment subjects play a series of basic ultimatum games. Because responders can increase comparative income by rejecting offers deemed too low, outcomes are expected to diverge from extreme divisions in the direction of equality. In the second treatment subjects play a series of ultimatum tournaments. Because rejection of low offers can only decrease comparative income, extreme splits are expected. Hence, the comparative model predicts that offers will be lower and near zero in the second treatment. The results of our experiment do not support the comparative model. Mean offers in the ultimatum tournaments are not extreme, nor are they significantly lower than those in the basic ultimatum games.