Understanding the Sunk Cost Effect: An Experimental Approach

The purpose of this project is to examine the relevance of sunk costs in decision-making. A “sunk” cost is an unrecoverable monetary or time commitment made to an endeavor. The prevailing economic paradigm suggests that, in general, rational consumers and producers will ignore sunk costs when making decisions. Thaler (1980) and Arkes and Blumer (1985) are two early and influential attempts to show that decision-makers may not always ignore sunk costs (and, in failing to do so, demonstrate a “sunk cost effect”). Two general approaches have emerged for empirically examining the relevance of sunk costs in decision-making: surveys and experiments. Explicitly or implicitly, surveys and experiments on sunk costs necessarily share (at least) three common elements: (i) the sunk cost, (ii) an initial endowment (or budget), and (iii) the marginal cost and benefit of escalation (i.e. continuing with the sunk cost activity). Within an experimental framework Meyer (1993) finds that the sunk cost effect becomes more pronounced as the size of the sunk cost increases. Experiments have subjects act as decision-makers who face a cost to enter into a risky investment. In our initial experiments in this area we look at two factors that may influence the likelihood that participants will display the sunk cost effect. The first is the size of the incremental, or marginal cost of continuing with a project. It is suggested that as the marginal cost of continuing an investment increases the likelihood that participants will display the sunk cost effect will fall. The second factor is the level of complexity associated with the decision to proceed with a sunk cost project, holding the expected pay out and variance constant. As the complexity of the continuation decision increases, participants are less able to assess the true expected value of continuing, and may instead give more weight to the highest possible pay out promised. Understanding The Sunk Cost Effect: An Experimental Approach