Degree Name

Doctor of Philosophy


School of Accounting and Finance


This dissertation investigates whether trained professionals can be adversely influenced by sunk cost information when making financial decisions. The rational decisionmaking model posits that sunk costs are irrelevant to choices between alternative courses of action. However, a review of the literature reveals that, contrary to the assumptions of normative theory, people do not always act in a rational manner when making financial decisions. Moreover, sunk costs are taken into consideration and induce economically irrational behaviour, such as committing more funds to doomed projects which can be described as "throwing good money after bad". Such theoretically irrational behaviour is the subject of behavioural (positive) theories of decision making developed in psychology. Research in cognitive psychology has recognised that people often use heuristics to simplify a decision task.

To test the predicted decision outcomes of expected utility theory a model was developed based on prospect theory and image theory. In the context of prospect theory, sunk cost and framing of the decision task were manipulated. Image theory was tested by inclusion of problem space items and image compatibility questions. The model provided a suitable conceptual framework for empirically testing the "sunk cost effect". Testing for the existence of the sunk cost effect was focused on the decision behaviour of particular professional groups and evidence was gathered using survey instruments adapted from previous research. The professions surveyed were managers, accountants and financial planners. A hypothetical outsourcing task was the basis of the survey of managers and accountants. A hypothetical investment task was used to survey the financial planners and was also included in the survey of the accountants.

The results of the investment task are consistent with image theory, the level of responsibility was found to be positively correlated with a higher level of funding. In the investment task the perceived level of responsibility for making the initial decision was found to be a contributing factor to escalation of commitment. A n unexpected result was that both the low and high image compatibility were found to be significant predictors of the level of additional funding in the investment task. However, the statistical significance of the high image compatibility was smaller compared to that of low image compatibility. The sunk cost effect was found to be significant in both the outsourcing task and in the negative framed version of the investment task. The sunk cost effect was not statistically significant in the positive framed version of the investment task. A n unexpected result was that framing produced risk behaviour that was contradictory to the prediction of prospect theory. Negative framing, which should have produced risk-taking behaviour resulted in risk-avoidance behaviour and positive framing which should have produced risk-avoidance behaviour resulted in risk-taking behaviour. The sunk cost effect was found to be nearly ubiquitous in the three professional groups. This effect was observed across a range of tasks. The strength of the effect depended on the relative magnitude of the sunk cost and other contextual factors allowed for in the survey instruments, but remained constant regardless of these influences to an observable degree.