Risk Aversion and Incentive Effects: New Data without Order Effects

Holt and Laury (2002) used a menu of ordered lottery choices to make inferences about risk aversion under various payment conditions. The main results of that paper were: (a) subjects are risk averse, even for relatively small payments of less than 5; (b) risk aversion increases sharply with large increases in the scale of cash payoffs; and (c) there is no significant effect from increasing the scale of hypothetical payment. With a few exceptions noted in the paper, all treatments began with a low-payment choice, followed by a choice with hypothetical payments that had been scaled up (by 20, 50, or 90), followed by a real-cash decision with the same high payment scale (20, 50, or 90), followed by a single, final, low (1) real payment choice. Those in the 90 treatment could earn amounts ranging from 9. 00 to 346. 50 in this task. As Glenn W. Harrison et al. (2004) correctly note, this design confounds order and treatment effects since the high real payment choice was always completed after the low real and high hypothetical payment tasks. In a new experiment reported below, we first seek to replicate Harrison et al. ’s finding that the order effect (participating in a low-payment choice before making a high-payment choice) magnifies the scale effect. In a second treatment, each subject completes the menu of lottery choices under just one payment condition (1 or 20, real or hypothetical), thereby eliminating any order effects. I. New Data