THAT human beings often continue to pour money into bad projects because they have already invested in them and cannot bring themselves to lose that investment is well known. Indeed the sunk-cost fallacy, as this phenomenon is called, is frequently cited as an example of people failing to behave in the “rational” way that classical economics suggests they should.
Though the exact psychological underpinning of the sunk-cost fallacy is debated, it might reasonably be expected to apply only when the person displaying it also made the original investment. However a study published recently in Psychological Science by Christopher Olivola of Carnegie Melon University suggests this is not true. In making decisions, people may also take into account the sunk costs of others.
Dr Olivola was led into his investigation by a thought experiment of the sort sometimes conducted by physicists. His imagined experimental subject had just received, as a present from a well-intentioned aunt, a gaudy and uncomfortable jumper. He asked himself whether the putative subject would be more likely to wear the jumper if he also knew that his aunt had made significant sacrifices to buy it, and he suspected that the answer would be “yes”.
Having experimented reflectively on himself, he decided to try something like it on other people. He recruited volunteers and posed them similar hypothetical questions, though not involving aunts.
In his first experiment he asked 602 people to imagine that they had obtained a front-row ticket to a basketball game but that a terrible storm on the day of the game meant travelling to watch it would be cold, slow and potentially hazardous. Participants were also told that it was too late to exchange the ticket or to give it to someone else. They were then asked to imagine either that they had obtained the ticket for themselves or that a friend had obtained it, but because of an unexpected work-related trip could not attend and had therefore given it to them. They were also asked to imagine either that they or their friend had obtained the ticket free, or had paid $200 for it. Armed with all this information they were then asked whether they would go to see the game live or stay at home and watch it on television.
As sunk-cost theory predicts, those told they had paid for the ticket themselves opted to attend the match, rather than watch it on TV, more often than those told they had obtained it free. Intriguingly, though, this was also true of those told they had been given the ticket, if they were told as well that the ticket had originally cost money rather than being a freebie. Moreover, similar results obtained in other experiments Dr Olivola conducted, involving imaginary tennis-club memberships, movie-watching and chocolate cake.
Oooo! It’s lovely!
A possible explanation for these results, and also for Dr Olivola’s own intuitive response to the aunt problem, is that social signalling is involved. In all cases the gift was supposed to have come from a close social connection (either a friend or a relative), so part of the act of using it was to show appreciation for its receipt. The costlier the gift, the more appreciation a donor might expect to be demonstrated, which was consistent with what he found.
To double-check the role of social connection, however, he decided to conduct one final round of experiments. In these the putative gift was supposed to have come not from a bosom buddy but rather from a casual acquaintance or a stranger. To his surprise, the effect was often stronger with these people than it was with friends and relatives.
What is going on here is obscure. Perhaps exaggerated gratitude towards acquaintances and strangers is a way of turning them into friends. All told, however, Dr Olivola believes he has demonstrated that the sunk-cost phenomenon shapes human behaviour much more broadly than was previously thought. Yet more evidence, then, that Homo sapiens and Homo economicus are different species.
Click to get a 90% discount on Udemy.