Definition of sunk cost

Sunk cost is an expression representing the unrecoverable amount of money that has already been placed into an ongoing investment or project. It is one of the simplest, yet most commonly misused financial measurements of a project. We’ll learn how to avoid the most common mistake in project (financial) management, and how to survive when our boss makes the mistake.

Obtuse?

In The Shawshank Redemption, Andy Dufresne asks the Warden if he’s obtuse. Andy gets a month in solitary confinement for his impertinence. We should remember this when our boss makes a statement like “We can’t cancel the project – we’ve already invested a million dollars in it!” If we want to avoid a month in the pokey, we need to restrain ourselves from telling our boss that he is irresponsible or ignorant.

Reinvestment decision example

Consider a project, with a projected return of $100,000 and a projected cost of $75,000. We previously determined that the project would be a good investment for us, so we funded it. Six months later, we’re reviewing the financials of the project, and discover that our project manager has already spent the $75,000, has not completed the software, and wants another $50,000 to complete the project. Assume that we believe the project manager, and it really will take an additional $50,000 to finish. If we don’t finish the project, the projected return is $0.

We’ve looked at methods for calculating the ROI of a proposed project. How do we make a decision about a project that is partially completed?

Financially incorrect reasoning

We now look at the project as having a total cost of $125,000 and a return of $100,000. We would lose $25,000 with this project – we should kill the project right now.

Financially sound reasoning

Unfortunately, the first $75,000 has already been “lost”, so we should not include it in our analysis. The previous calculation is wrong. Financial decisions are always made about future events, not past events. No decision we make now can recover the sunk costs. We should therefore ignore those costs in any decision making.
At the beginning of the project, the projected value of the project was $25,000 ($100,000 return minus $75,000 cost). [Note: This is similar to, but not the same as expected value]. The projected value of the additional investment is $50,000 ($100,000 return minus $50,000 additional cost). The first $75,000 we spent doesn’t matter any more – it’s already been spent and can not be recovered.

What we are faced with now is the decision to spend $50,000 to get a return of $100,000. There are two possible scenarios: invest nothing and get nothing in return, or invest $50,000 and get $100,000 in return.

The first $75,000 is a sunk cost in that we don’t get that money back just by choosing to not invest more money.

There’s usually more at stake here

In another example, imagine that we’ve invested a million dollars to get a return of 1.5 million. We’ve just discovered that we will have to invest another two million dollars to complete the project. Should we do it? Financially, no. This is where the pointy-haired boss might say “We can’t cancel the project – we’ve already invested a million dollars in it!”

However, that manager may be optimizing on his reputation, not the company’s finances. The reference to previously invested funds could be a red herring. What the boss is really saying is “We can’t cancel the project – the CEO knows that I’m in charge of making this happen!” It may also be that we don’t have all the information. There might be other hard to quantify benefits to completing the project (like goodwill or investor-perceptions) that would justify the added expense. In other words, the return exceeds 1.5 million dollars, we just don’t realize it, even if our boss does.

In cases like this – tread carefully. Those of you old enough to remember the SALT treaty will remember Reagan’s famous statement – “trust, but verify.” Take the same approach with your apparently illogical boss. And if you find out that you can’t trust your boss, find another boss.