3

Sunk costs are costs that can not be recovered when new a planning decision is made. Taking them into account is therefore characterized as a fallacy. But are the examples below really fallacies?

[ Source: ] [1.] Imagine you go see a movie which costs $10 for a ticket. When you open your wallet or purse you realize you’ve lost a $10 bill. Would you still buy a ticket? You probably would. Only 12 percent of subjects said they wouldn’t.

[2.] Now, imagine you go to see the movie and pay $10 for a ticket, but right before you hand it over to get inside you realize you’ve lost it. Would you go back and buy another ticket? Maybe, but it would hurt a lot more. In the experiment, 54 percent of people said they would not.

The situation is the exact same. You lose $10 and then must pay $10 to see the movie, but the second scenario feels different. It seems as if the money was assigned to a specific purpose and then lost, and loss sucks. [...]

Source: p 276, Choices, Values, and Frames by Daniel Kahneman, Amos Tversky

Example 4. A family pays $40 for tickets to a basketball game to be played 60 miles from their home. On the day of the game there is a snowstorm. They decide to go anyway, but note in passing that had the tickets been given to them, they would have stayed home.

Example 5. A man joins a tennis club and pays a $300 yearly membership fee. After two weeks of playing he develops a tennis elbow. He continues to play (in pain) saying, "I don't want to waste the $300!"

For hypotheticals 1-5 above, if aversion to paying for the lost amount is the only reason for persisting (with the action that caused Sunk Costs), then I accept that 1-5 do exemplify the Sunk Cost Fallacy.

But did Profs Kahneman and Tversky consider the following possibility 6? Does 6 disprove the hypotheticals above as any Fallacy at all (a fortiori examples of the Sunk Cost Fallacy)?

  1. Suppose that in 1-5, loss of the money spent instigated so much agony or distress in the victim that the victim can no longer enjoy, and so abandons, the planned activity (e.g. the victim is too shocked by his carelessness or incompetence to care about the activity anymore).
2

None of these are examples of the "Sunk Cost fallacy", except #5.

How the "Sunk Cost fallacy" works: You decided that doing X is a good idea. You start doing X, and spend $Y on it. Then you figure out that doing X wasn't actually a good idea. Not only was it a bad idea to start it, but it is so bad that it is even a bad idea to continue. The "Sunk Cost fallacy" is to notice that the $Y will be wasted if we don't continue, and therefore continue even if it is a bad idea.

Quite often the argument is: If we stop, then it is obvious to everyone that we were idiots. If we continue, then we have a case to argue. We know we would be idiots, but it would be harder to prove.

(When is it not a fallacy: Say you figure that doing X is worth 10 million, and costs 8 million, so you start doing it. After spending 15 million you figure out it was a bad idea to start it, but you need to spend another 5 million to finish X and get the 10 million benefit. In that case, it is correct to continue).

(Number 5 wouldn't have been a fallacy if the man decided that he enjoys tennis a lot less than he thought, and wouldn't have paid $300 if he knew what he knows now, but he still enjoys it a little bit and continues playing).

2

Sunk cost fallacy refers to a particular context of idealized "rational agents" of classical economics. In that context basing current decision on costs that have already occurred and can not be recovered regardless of what is done now, is indeed irrational. Your scenario does not work as stated because you explicitly presuppose irrational motivations like agony, distress, shock, and so on. One is free to take those things into account but then one is not talking about rational agents. How well this idealization applies to real life is a different question entirely, Keynes called such economic impulses "animal spirits", and explicitly took them into account in The General Theory of Employment, Interest and Money.

However, McAfee, Mialon and Mialon argue in the paper Do Sunk Costs Matter? that taking sunk costs into account may be justified even for a rational agent within a broader context:"Agents may rationally react to sunk costs because of informational content, reputational concerns, or financial and time constraints". For example, the expected additional investment may be correlated with the sunk investment, a commitment to finishing projects that appear unprofitable may have reputational value nonetheless, and the sunk funds may constrain ability to make future expenditures. This last version may work for your scenario if we replace emotions by rational reflection on financial constraints as the motive for abandoning the planned activity.

2

[1.] May not be a fallacy.

The amount of wealth that you should be spending depends on the total wealth you have.

You do not have an unlimited budget. If you lost 10$, and after it you buy a ticket, then that day you lost 20$.

If you are millionaire, then it doesn't matters, but if you are homeless, then it may cost you your next meal, or make you fall in debt.

For example: On finance, there is taught Kelly's criterion. It advises what percentage of your of your money you should risk on an investment. A bet may be favorable to you, so it may be a good idea to gamble. But you should not bet all of your money, because there is a chance you will lose all of it, and if you repeat the bet enough times, you will lose all your capital.

There is always uncertainty and risks in life, so it is natural and reasonable to become more conservative in your spending after unfortunate financial events.

Kelly criterion says that the amount of money a person should risk, depends on the percentage of his wealth that the person is risking, and also the probabilities of the outcomes.

On the first case, people got poorer before buying the ticket, but not everybody got equally poorer as percentage of to his wealth, so it is rational for some people to avoid spending more, and other people to not mind it. On the second case, people got poorer, but there is also an empirical risk of losing tickets, so the investment is riskier, and that explains rationally why more people should avoid buying a second ticket.

It does not means that people answered after evaluating the probabilities, but it provides a rational explanation for why people should make those choices, so it is not necessarily a valid example of the Sunk Cost Fallacy.

  • @jovexeni So your deduction is that ten dollar tickets in general are a bad idea because they are easier to lose than ten dollar bills. But this is just not true. 'Learning' from the specific example is illogical. It lacks adequate consideration of the rest of the world. At some point you just have to admit that we are basically illogical and need to refer to logic to correct for that. – jobermark Jan 7 at 19:21
  • If you have unexpainable and unpredictable evidence that you may lose the ticket, then it is rational to not risk when you have a choice. In [1] the lost was independent of the choice. In [2] the risk is not independent of the choice. – jovexeni Jan 7 at 19:24
  • @jovexeni Repeating yourself is not arguing. Do you deduce from this occurrance that tickets are easier to lose than bills? Because that is not true. Local deductions that are disproved by general experience are not learning opportunities, they are illogical choices. – jobermark Jan 7 at 19:25
  • @jobermark NO. I deduce that in that PARTICULAR EXAMPLE, there is risk of losing a ticket. In general there is no risk of losing a ticket. You are making an unjustified generalization from that particular example. – jovexeni Jan 7 at 19:26
  • In general is not normal "to magically lose a ticket", so you cannot generalize. But if the assumption is that a ticket was lost, for unexplained reasons, in that PARTICULAR EXAMPLE, that risk should be accounted. – jovexeni Jan 7 at 19:28

Your Answer

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Not the answer you're looking for? Browse other questions tagged or ask your own question.