The Cost of a Thing Is the Other Thing

There is a shop on a street near you that has been open for eleven years. The owner works six days a week, employs nobody, and clears twenty thousand pounds a year after every bill is paid. By the ordinary meaning of the word, the shop is profitable.

She could earn thirty-five thousand doing the same work for a chain half a mile away.

Economics says the shop is losing fifteen thousand pounds a year, and this is not a rhetorical trick. Her labour is an input to the business. It has a market price. The fact that she does not write herself an invoice does not make it free, any more than a factory using its own coal is using free coal. Economic profit subtracts the value of every input at what it could have earned elsewhere, and by that measure a great many small businesses that look modestly successful are quietly consuming their owner's wages.

The concept doing the work is opportunity cost: the value of the best alternative you gave up. Not the sum of all the alternatives (you could only have done one other thing with the afternoon) and not the money that left your account, which is a different measure with a different use.

It is worth being precise about why the accounting measure is not enough. Accounting cost answers the question "what did we pay?", which is what a tax authority and a creditor need to know. Opportunity cost answers "what did this resource stop us doing?", which is the only question relevant to a decision. Those come apart whenever a resource is not bought fresh for the purpose: your own time, a building you already own, a machine standing idle, the three years you spent on a qualification.

And they come apart most sharply where the price is zero. A free ticket to a concert three hours away is not free; it costs the evening. Publicly owned land in a city centre costs a council nothing and has an opportunity cost running to millions, which is precisely why "it costs us nothing" is among the least reliable sentences in public spending. A council that lets a building to a charity rent-free is making a subsidy exactly equal to the forgone rent. That may be an excellent use of the money. It is money, and describing it as costless means nobody ever asks whether a grant of the same size would have done more good somewhere else.

The other half of the idea runs in the opposite direction and gets used even more often: costs already incurred are irrelevant.

The only thing any decision can change is what happens next, so the question is always what the available options are worth from here. Money already spent appears identically under every option, which means it cannot distinguish between them, which means it cannot bear on the choice. A sunk cost is not a factor to be weighed carefully. It is not a factor.

This is easy to state and famously hard to feel. People finish films they are not enjoying because they paid for the ticket, stay in projects that have already consumed two years, and keep repairing a car on the grounds of what they have spent on it so far. The pull is strong because walking away feels like wasting the earlier expenditure, when in truth the waste happened when it was spent and no future action can retrieve it. The alternative, spending more on something that will not repay it, wastes the earlier money and some new money too.

Behind both halves sits the habit that most distinguishes economic thinking, which is reasoning at the margin. Classical economists tied themselves in knots over the paradox of value: water is essential and cheap, diamonds are useless and dear. The knot comes undone the moment you notice that nobody ever chooses between all the water and all the diamonds. They choose whether to buy one more bottle, and the marginal bottle of water is abundant and nearly worthless while the marginal diamond is scarce and expensive. Total usefulness and marginal value are simply different quantities.

Once you have the habit, a whole class of confused arguments dissolves. An airline's average cost per seat tells you nothing about whether to sell one more ticket on a plane that is departing anyway, because the plane, the fuel and the crew are already committed and the marginal cost of one extra passenger is a sandwich. That is why the last seats are cheap, and why the same airline would be destroyed if it sold every seat at that price. The average answers a question about the business; the margin answers a question about the decision. Nearly every bad argument about cost is one of those two standing in for the other.