The sunk costs fallacy: an unlikely friend

Roland Allen


April 17th, 2016

Sunk costs are one of the first things any economics student learns about. For the uninitiated, a sunk cost is defined as ‘a cost that is already incurred and cannot be recovered’. Specifically, it is taught that any rational decision maker should not take sunk costs into account when making decisions. This might seem like […]

Sunk costs are one of the first things any economics student learns about. For the uninitiated, a sunk cost is defined as ‘a cost that is already incurred and cannot be recovered’. Specifically, it is taught that any rational decision maker should not take sunk costs into account when making decisions. This might seem like common sense; as the old adage states, ‘it’s no use crying over spilt milk’. But taking sunk costs into consideration is something that people tend to do a lot; so much so that this even has its own (unsurprising) name: the sunk costs fallacy. You might have encountered the situation at a restaurant where your hunger is perfectly sated and you can’t fathom eating another bite, yet you haven’t quite finished your meal. A friend might insist that you finish and ‘get your money’s worth’, or else the food will be wasted. But of course, you’ve already paid and the food has already been prepared, so these are costs are sunk; if you don’t want to eat the rest (if you would derive no benefit from doing so) then you shouldn’t. Often, the sunk costs fallacy can spur us to continue doing something that we should have abandoned.

Another, completely separate way in which we are irrational is in our tendency to value short-term reward at the expense of a greater long-term reward. This can be considered irrational because it doesn’t maximise our overall satisfaction (or ‘utility’) when taking the long-term into account. We’re constantly short-sighted, and even when we’re not myopic and understand the long-term effects of our actions our willpower is finite. We can see this manifest in countless instances in our own lives and those of others. People eating junk food despite the long-term health effects (but it’s so tasty!), uni students procrastinating and under-dedicating time to assignments and studying (who can be bothered?), partygoers drinking excessively with no thought given to the crippling hangover they will be greeted with the following morning (another one!), so on and so forth.

What might these two seemingly unrelated biases have to do with each other? Here’s an example.
Kathy is studying for her upcoming maths exam. Any given amount of time she spends studying provides her some perceived benefit, but also some perceived opportunity cost (the foregone benefits of the next best thing she could be doing with her time, like watching Netflix). The ‘net benefit’ (NB) can subsequently be defined as the perceived benefits minus the perceived costs for the amount of hours dedicated. This gives us an indication of how many hours she will study. We can represent this graphically:

Screen Shot 2016-04-17 at 5.02.32 PM

Each additional hour she works gives her less benefit than the previous one (the slope of the NB curve decreases as the hours worked increases), meaning that the total net benefit she feels she will receive from extra study increases up until a certain point (h), but from there on the costs outweigh the benefits; she will do no more or no less than h hours of study for her exam in an attempt to maximise her net benefit and utility.
But note the importance of the word ‘perceived’ here. What if Kathy’s perceived net benefits don’t reflect the actual net benefits from studying? She might underestimate the benefits of studying beyond a certain point (perhaps a result of overconfidence), or fail to fully comprehend the consequences of her exam performance. Alternatively, she could overestimate the opportunity cost. Catching up on the last season of House of Cards seems like a great idea to her now, but it may not be that beneficial in the long run, compared to study. As explained before, people tend to value benefit in the short-term more than the long-term. So Kathy’s perceived benefits from spur-of-the-moment activities might be overvalued. This makes intuitive sense; if spending time watching TV and going on Facebook is more attractive than it should be, then any activity that would result in foregoing time spent doing these things will be less attractive than it should be. This would result in her actual net benefit curve looking a little bit different:

Screen Shot 2016-04-17 at 5.02.40 PM

Unsurprisingly, she really should be doing more study if she wants to maximise her utility (the optimal hours worked, h* exceeds h).

Now we can see how the sunk costs fallacy might come into play. Imagine, right before she finished all of the studying she was going to do (all h hours of it), she thought to herself: “I’ve come this far, I’ve dedicated all these hours, it would be a waste not to continue and put more time into the subject!” This sort of logic mirrors that of our nagging friend at the restaurant, only this time it might accidentally be useful. While Kathy’s rationale is based on an emotional feeling of investment in her studies and NOT a calculated future cost/benefits analysis (she would otherwise only complete h hours), in this case it may motivate her to study further, pushing her closer to the actual optimal amount of hours worked (to hnew) due to a shift in her perceived net benefits:

Screen Shot 2016-04-17 at 5.02.46 PM

If she fully understood the totality of the net benefits associated with studying, this thinking would cause her to overwork; but as it so happens, she was going to study a less than desirable amount otherwise. The two irrationalities ‘cancel out’ to an extent.
The sunk costs fallacy tends to encourage people to over-dedicate resources to their pursuits, while people’s ‘short-termism’ tends to lead people to under-dedicate resources to pursuits with long-term gains. Thus, when pursuing long-term goals, our tendency to factor in past information that ‘should’ be irrelevant into our decision making can counteract our tendency for inaction.
If we look closely, we can see this phenomenon all around us. Unsurprisingly, it can be applied to economic activity. Let’s analyse the apocryphal tale of Pete, an uncomfortably portly man who is currently making the switch from eating high amounts of junk food to healthier alternatives (a choice that we will assume is rational, in that it maximises his long-run utility); junk food is, as we touched on briefly before, a classic example of a good whose benefits are enjoyed exclusively in the short-term and whose adverse consequences are almost entirely felt in the long-term. A few weeks into his switch he is feeling his willpower deteriorating and despite knowing that sticking with consuming healthy food will make him better off, he is worried that his cravings could start to get the better of him. But as he begins contemplating falling off the wagon and back into the comforting embrace of fast food burgers, it strikes him that doing so would waste all of his previous effort and sacrifice; would all that spinach and broccoli be for nothing? And every time that he would consider indulging his vice, a version of this thought pattern dominates his thinking and he continues on the straight and narrow. He becomes a healthier and happier individual as a result. It appears that two wrongs can make a right, after all.
Certain patterns of thinking which, on their own, would cause us to act against our best interests can surprisingly be an ally in fending off the effects of other biases. There are likely countless other interactions that could be observed; we have just looked at one. Economics is predicated on the decision-making of consumers; the more we understand about them, the clearer the picture gets on how the world really works.

The views expressed within this article are those of the author and do not represent the views of the ESSA Committee or the Society's sponsors. Use of any content from this article should clearly attribute the work to the author and not to ESSA or its sponsors.

Founding sponsors




Gold sponsors



Silver sponsors