How do prior outcomes affect risk attitude? Comparing escalation of commitment and the house-money effect

When making sequential decisions, do prior gains induce more or less risk taking than prior losses? Prior studies have found evidence for a house-money effect, where risk taking increases following gains. These studies also found seemingly contradictory evidence for escalation of commitment, where risk taking increases following losses. We present an experimental study investigating this apparent contradiction. By manipulating the presentation format of a decision problem we can induce both types of behavior. When the problem is presented as a portfolio decision, risk taking is greater following losses than following gains, consistent with escalation of commitment. When the problem is presented as a two-stage betting game, risk taking is greater following gains than following losses, consistent with the house-money effect. Some of these results can be accounted for using the prospect-theory value function. Escalation of commitment in our experiment does not appear to be driven by a need to justify or rationalize the initial decision. These results are potentially relevant to modeling risk preference in multiperiod investment decisions.