Reading a reference to the sink cost fallacy in another thread, I got to wondering, are there formulae to help one calculate when one should bail? To have a reasonably useful measurement to say, “okay, this much and no more”?
I would imagine that the time is… whenever the expected value (discounting past expenditures) is less than zero.
Ideally, your formula is X+Y*Z, where X is the likely amount that you gain from continuing, Y is the amount you already spent (in time or money) and Z is zero. In other words, completely forget all the time you spent learning which Pokemon was which and all the money you spent on your coin collection - if watching the next episode of Pokemon or collecting the next coin will give you pleasure or money, do it, and if not, don’t do it.
Next week - the “opportunity cost” issue!
If you would not buy it today at the current price, why would you continue to hold it today? Ignoring tax, spread, and other transaction costs, if you wouldn’t buy it today, you should sell it today. Assuming you have some other investment you’d be willing to buy today. And even if, in the extreme, you don’t.
The “sunk cost fallacy” is the idea that somehow people care about what they’ve already lost more than they care about what they’re about to lose. So they’re willing to lose more to be able to pretend for longer that they haven’t already lost what they already really have already lost.
A “fallacy” means “you’re thinking about it wrong.” The sunk cost fallacy is thinking that what you have invested already has any relevance to your decision-making now. It doesn’t. Or rather, it shouldn’t.
Which is what a lot of people seemed to have failed to understand when it comes to something like the Overwatch 2 debacle. Blizzard poured a lot into that, so to just scrap it seems totally awful, even if it’s the case that they currently feel they cannot produce the game they intended to produce and they’re better off stopping the bleeding now than try to work on it any further. There are other factors involved in this particular case, but big companies have to cut loose major projects that they invested a lot of money into every so often, and it generally comes down to whether there’s anything salvageable or not.
That also potentially leads to a reverse fallacy though, where someone looks at a project that has.spent so much money, that the bosses can’t see it possibly recouping all of it, and fail to allow what has been already worked on being used in something smaller scale at a cost for less of what that smaller scale development would cost from this point on. I’m reminded of how Annie Hall started out as a completely different movie, and after it was filmed, they decided to throw most of it out and focus on just one small bit of what was filmed, which became a classic. If they had looked at the whole mess of what they had filmed and thought that it would never worked as originally planned and completely scrapped it, they would have missed making one of the best movies of the era.
“This time she’s really gonna love me!”
No. No rational calculus.
It boils down to (obviously)
Profit = Revenue - Costs
And when considering whether to carry on with a project, Profit is negative then you cancel as it will lose you money. Nothing controversial there.
But the critical part is:
Costs = cost of completion + amount already spent
This was actually pretty controversial when it was first proposed. I mean you aren’t getting that money back, whether you choose to cancel the project or not, so why should you cancel a project you’ve already spent a million dollars just because the extra 100 bucks needed to finish it would take you into the red?
So I’d say it’s the opposite. The sunk cost fallacy is saying you shouldn’t consider what you’ve already spent, when you consider the cost of a project. So if you have spent a million dollars and only have half a million left to complete the project, but you expect the project to the earn a million bucks, then its incorrect to say "I need to spend half a million dollars and I’ll earn a million, I should carry on ". The total cost is 1.5 million not 0.5 million even if cancellation won’t get you that million back.
Agree with this completely.
But here’s a nuance … For the example of a $1M project lacking just the last $100 = 0.01% of spending, we ought to add uncertainty into our calcs. How certain are we of the total revenue? For a fixed price construction project the answer might be “very”. For e.g. a computer game, the expected future revenue is an estimate with some error bars.
Likewise, often our “costs” have significant error bars introduced by artifacts of our cost accounting procedures and habits. Lots of dumb business decisions have been made in both directions (wrongfully cancelling or wrongfully proceeding) because cost accounting did not reflect the true direct and indirect costs attributable to the project.
I think we’re saying the same thing from a different reference frame. Or perhaps there’s really two variants of “sunk cost fallacy”, one in an accounting sense and one in psychological / organizational behavior sense.
-
The true cost of a completed project is what you have spent plus what you will spend to completion. Measuring expected profitability against only the latter spend is incorrect. Doing so is a fallacy. As you rightly say.
-
There is a psychological desire to “protect your investment” by continuing even a hopeless project. The belief persists that the future (small) spend will leverage the past (large) spend to create a good result. In your example, that’s “spend a hundred to save a million”. It’s a arithmetically correct but accounting nonsense equation using real inputs to generate a nonsense result.
But it is seductive, whether the decider is an executive with personal prestige on the line, or you or I thinking about selling our shares of some losing stock. The fallacy is a form of delayed recognition. As long as I haven’t cancelled / sold, some miracle may occur and salvage the situation for me. Once I have cancelled / sold, the loss becomes real, not notional. Again incorrect and therefore fallacious thinking. The loss is real whether you choose to recognize it or not.
What I meant by the snip you quoted (snipped a bit too short to reflect my actual point) is more in the second sense. Spending e.g. 500K to “preserve the value” of the e.g. 500K already spent on a losing project means thinking that the first half is somehow more meaningful, more important, than the second half.
In a stock price example, if you bought at $100 and it’s now trading at $50, you’re more worried about, more protective of, the $50 you already lost than the $50 you’re going to lose when the company finishes tanking. Much less the opportunity cost of what other profitable thing you’re not buying with the same money while waiting for the recovery that won’t ever come.
Emotionally overvaluing the known “paper losses” already sustained primes you to continue losing in the usually vain hope of winning it all back later. cf. Gambler’s fallacies of all sorts.
I’m sorry, but I think you guys have it completely wrong (mostly you, griffin1977, sorry).
The sunk cost fallacy is when you have already lost a lot on a project (the amount already spent is high) but you keep on going as you become convinced that surely if you finish the project, whatever the cost, (dumping whatever it takes into the cost of completion) then surely you will overcome your losses with a profit, even though you would be better off at that moment to abandon the project rather than continue. Because in your one you have to profit.
It’s not spending a million dollars and finding that if you spend half a million more, you’ll get three quarters of a million in revenue (for a net loss of three quarters of a million dollars). It’s spending a million dollars and being convinced that if you just spend another half million surely you will be better off, maybe even profit without either (a) actually doing the math as in the preceding example where the math shows losses will indeed be reduced by continuing or (b) doing the math but refusing to believe it.
It makes perfect sense to continue a project if continuing will minimize losses already incurred. That is not the sink cost fallacy.
This is correct.
In simple terms the sunk cost fallacy is “throwing good money after bad.” It is the emotional attachment to an investment that causes a person to invest even more, even though that additional investment is not rational. It’s like having an old car, you pay $2000 for a new transmission, then when it needs a new engine you say, “Well, I’ve already got $2000 invested in the transmission so I need to spend another $3000 on a new engine” even though the car is only worth $2500.
Sometimes it’s that simple. The trouble is distinguishing between sound and fallacious reasoning when it’s more complicated. And life usually is more complicated.
Right now I have a car that could need as much as $4K invested to get looking good and mechanically sound. I might be able to sell it for $4K right now, but likely not for $8K after the work is done. I cannot buy an equivalent requiring no extra investment for either $4K, and probably not even $8K. So my choices are:
- No car and $4K profit from selling it
- This car looking fine, running, sound, and $4K spent on it
- A $4K ugly crappy car that maybe works for a while
- A $8K slightly less ugly less crappy car that works for a longer time
Since I have another car I can choose option 1 if I want, but it’s a lot like option 3 or 4 without spending any money. Trouble is that cars are really consumables. They’re terrible investments financially outside of their use in making money, although that would include commuting to a job. Every cent you put in a car for any other reason is wasted money. In a case like this it doesn’t come down to a dollars and cents calculation. The solution is my satisfaction with the result. The same reason we all pay more for clothes, cars, houses, furniture than the absolute minimum because we aren’t Anabaptists.
I mean, in that scenario you should actually spend the half million. Even though you still lose money overall, you lose less money than you would if you just scrapped the project (1 mil vs 0.5 mil).
Is there not also a predictive analytical (and/or gambling) component as well? I think that it’s been alluded to upthread but if you’re smarter and better at reading the tea leaves than someone else your decision on when or whether to bail or continue may be different and more accurate than someone else’s.
I think I agree with you.
Let’s talk specifics: On January 1, someone comes to you with a proposal - spend $20 million so you’ll have a factory on December 31st that will make you $200 million dollars thereafter (ie. the expected value of the whole factory for its entire lifespan is $200 million in today’s dollars - you could sell it for that amount on Dec 31 with no problem). Good deal! You start the project.
On March 31, you realize that you’ve already spent $15 million and the building’s not half done yet. Your best estimate is that it will be another $10 million to get the project done. Still a good deal for $200 million on Dec 31.
On June 30th, the project is going pretty well - you’ve spent another $5 million and the project and you think that you can actually spend only $3 million to get it completed. Then you hear bad news. Technology changes make your planned 8 track tape factory nearly valueless. If you finish it, on December 31st, you could sell it to a streaming company for them to refurbish into a spa - but you’d only get $1 million; if you sell the land and the half finished factory to someone they’ll give you 0.5 million to tear it down, so they can build a parking lot.
So you’ve spent $20 million so far, it’s almost done, and it will cost $3 million to finish it, and there is a strong tendency to not want to “throw away” the 20 million you’ve already spent. But in fact, that money is gone. You have a choice between pulling the plug to get a half million, or spending 3 million more to have a factory worth 1 million. The sunk cost fallacy would have you throw the good money (3 million) after the bad (20 million) - to avoid the fallacy, make the decision you would make if you were magically given the company on June 30th, and had no vested interest in the part-built factory.
This is fascinating! Thanks to everyone who contributes!
I think @Andy_L has the best real world example of how this happens.
It starts out as an obviously winning plan. Over time, the costs go up, the schedule stretches out, and the expected benefit withers. Eventually those curves of past + future expense versus future expected benefit cross and now you’re holding a money-loser, not a money-winner.
So now you’re faced with spending X more to finish something with some completed value, but not nearly the value you expected, or stopping right now to have only something you can sell for salvage value. If that.
Said another way, at that awkward point sorting out the least-bad exit strategy is now the sensible correct goal. Pretending you’re not already screwed and pressing ahead blindly is falling for the fallacy.
In the ideal world the project management folks would be watching these two curves approaching each other over time and warn top management even earlier that “we’re showing profitable today but if the unplanned adverse surprises keep coming at the current rate we’ll transition to a loss before it’s done.”
Big exogenous shocks of course wreck even those sorts of pessimistic projections. You can be cruising along fine when suddenly somebody invades Iraq or the stock market crashes or an oil embargo hits and “all bets are off” as they say. Really meaning that “all prior projections are invalidated.”