How do prior outcomes affect risk attitude?

Modelling choice in multi-period asset pricing requires assumptions about how prior outcomes affect risk attitude. We present an experimental study of the influence of prior outcomes on risky choice. We document a strong framing effect. By manipulating the presentation format of the decision problem we can induce increased risk taking following a gain, i.e. the house-money effect (Thaler and Johnson 1990) or, alternatively, increased risk taking following a loss, i.e. escalation of commitment (Staw 1976). Maximization of a value function from prospect theory can explain some of our results. Escalation of commitment in our experiment does not appear to be driven by a need to justify or rationalize the initial decision