Sunk costs in the NBA: why draft order affects playing time and survival in professional basketball - National Basketball Association - includes appendix

Administrative Science Quarterly, Sept, 1995 by Barry M. Staw, Ha Hoang

Common sense tells us that people try to avoid losing courses of action. They move away from lines of behavior that have not been rewarded and hesitate to follow strategies that are not likely to yield future benefits. Yet some behavioral research has challenged this logic. Coming under the rubric of escalation of commitment, a number of studies have shown that people can become stuck in losing courses of action, sometimes to the point of "throwing good money after bad."

Evidence of this escalation effect was initially provided by three independent lines of research. Staw (1976) used a simulated business case to show that people responsible for a losing course of action will invest further than those not responsible for prior losses. Tegar (1980) took advantage of an unusual competitive bidding game (Shubik, 1971) to demonstrate that people can become so committed to a position that they will pay more for a monetary reward than it is worth. Finally, in several related studies, Brockner and Rubin (1985) showed that people are likely to expend substantial amounts of time and money in efforts to reach a receding or elusive goal. These initial investigations have been followed by a wide range of studies on conditions likely to foster persistence in a course of action, along with a set of theories accounting for these effects (see Staw and Ross, 1987, 1989; Brockner, 1992, for reviews).

Though the escalation literature has grown dramatically over the past two decades, it has continued to suffer from some serious problems. One issue is that escalation researchers have borrowed heavily from other research areas, such as cognitive and social psychology, without strict guidelines for selecting those variables most parallel to the conditions or events present in escalation situations. A second problem is that much of the escalation literature, despite its intent to explain nonrational sources of commitment, has not directly challenged the assumptions of economic decision making. By and large, the escalation literature has demonstrated that psychological and social factors can influence resource allocation decisions, not that the rational assumptions of decision making are in error. A third weakness is that almost all the escalation literature is laboratory based. Aside from a few recent qualitative case studies (e.g., Ross and Staw, 1986, 1993), escalation predictions have not been confirmed or falsified in real organizational settings, using data that are generated in their natural context. Therefore, despite the size of the escalation literature, it is still uncertain if escalation effects can be generalized from the laboratory to the field.

This paper presents one of the first quantitative field studies in the escalation literature. The study does not resolve all the problems of the escalation area, but it was designed with these deficiencies in mind. Because escalation situations involve the expenditure of resources over time, it is important to know whether the amount one initially spends on a course of action can affect subsequent commitment. Therefore, the study of sunk costs (past and irreversible expenditures) is central to the escalation question. Research on sunk costs is also a form of inquiry that confronts directly the assumptions of rational economic decision making. Economists universally caution against the use of sunk (rather than incremental) costs in decisions to invest further time, money, or energy in a course of action (Samuelson and Nordhaus, 1985; Frank, 1991). Therefore, any demonstration that sunk costs influence subsequent investment decisions calls into question the description of people as economically rational decision makers. Finally, and perhaps most importantly, by constructing a test of sunk costs using real organizational data, a large void in the escalation literature can be filled. If sunk-cost effects can be demonstrated in the field, then we may have greater confidence that escalation hypotheses can be generalized to situations devoid of the props, scenarios, and student samples generally used by laboratory researchers.

Research on Sunk Costs

Probably the most important set of sunk-cost studies is a series of ten experiments conducted by Arkes and Blumer (1985). Their most well-known study used a "radar-blank plane" scenario. Students were asked to imagine they were the president of an aircraft company deciding whether to invest $1 million in research on an airplane not detectable by conventional radar. These students were also told that the radar-blank plane was not an economically promising project because another firm already had a superior product. As one might expect, only 16.7 percent chose to commit funds to the project when funding was characterized as being used to start the unpromising venture. But, as predicted, over 85 percent chose to fund the venture when it was described as already 90 percent completed.

Follow-up studies by Garland and his colleagues replicated the sunk-cost effect yet posed questions about its interpretation. Using variations of Arkes and Blumer's (1985) radar-blank plane scenario, Garland (1990) demonstrated that sunk costs influenced investment decisions across several combinations of prior expenditures and degrees of project completion. When ConIon and Garland (1993) independently manipulated the level of prior expenditures and degree of project completion, however, they found only effects for degree of completion. Garland, Sandefur, and Rogers (1990) found a similar absence of sunk-cost effects in an experiment using an oil-drilling scenario. Prior expenditures on dry wells were not associated with continued drilling, perhaps because dry wells were so clearly seen as reducing rather than increasing the likelihood of future oil production. Thus it appears that sunk costs may only be influential on project decisions when they are linked to the perception (if not the reality) of progress on a course of action.