Asset integration and risk-taking in the laboratory

We report on a laboratory experiment designed to assess risk preferences in a decision environment where real losses can occur. Specifically, we utilize an asset integration protocol designed to ensure that cash provided to treatment group participants by the experimenter is fully integrated into each individual's wealth. This cash is placed at stake in an incentivized risk-preference elicitation task based on the well-known Holt and Laury (2002, 2005) methodology. Our experimental design allows us to distinguish between the predictions of expected utility and prospect theory. We find that features consistent with expected utility theory, constant relative risk aversion and rank dependent expected utility functions, are insufficient to explain our experimental results. However, preference functions based on prospect theory, accounting specifically for loss aversion, capture the observed behaviour of participants in the experiment.