A fine that became a price

Why did charging parents for collecting a child late make more of them arrive late?

A day-care centre has the problem every such place has: a handful of parents arrive after closing, and a member of staff waits with the last child. So the centre does the obvious thing and charges for it — a small fine, added to the monthly bill, for collecting a child late.

Lateness goes up. It roughly doubles, and stays doubled for as long as the fine is in force. Uri Gneezy and Aldo Rustichini ran exactly this as a field experiment across ten private day-care centres in Haifa over twenty weeks, introducing a fine of about ten shekels per late collection in six of them and leaving four as controls. They published it in 2000 under a title that is already the whole answer: A Fine is a Price.

The obvious reading is that the fine was too small to sting. Hold onto that, because it is half right and the half it gets wrong is the interesting half. Ask instead what a fine of zero shekels would have done. Nothing at all, on that reading — but zero was the situation before, and before, people were punctual.

Be precise about what a parent faced before the fine. They knew lateness was unwelcome. They did not know what would follow — an awkward look, a word at the door, eventually being asked to leave? The contract was silent, which is another way of saying the consequence was unbounded and unspecified. And notice what fills that silence: a judgement about what kind of situation this is. Someone is waiting past the end of her working day because of me — not a cost to be weighed but a wrong, and a wrong has no settlement price.

Now introduce ten shekels, and ask what a parent learns that they did not know an hour earlier. Two things, and both matter.

The first is the size of the sanction: the unspecified consequence has been named, and it turns out to be small — ten shekels against a monthly fee of well over a thousand. Any parent imagining something worse has been reassured. That is the "too small to sting" reading, and it is real.

The second is what the sanction implies about the transaction, and this is the part that does the work. A posted charge for a named behaviour is the form a price takes: this is a service the centre is willing to supply, here is the rate, and paying it settles the matter. Once something is a purchase, the ordinary logic of buying applies — you buy as much as you want, you owe nothing beyond the money, and nobody is entitled to an apology from a customer who paid. The centre had meant a rebuke. It issued a tariff.

So the payoff structure did not merely shift by ten shekels; the situation changed category. Before, the cost of lateness was a diffuse, unbounded social debt; after, a small definite dischargeable sum — and the second is cheaper for anyone the debt had genuinely constrained. Behaviour then followed the new payoff structure faithfully, which is what makes this an incentive story rather than a story about people becoming worse.

Which invites the sharper test. If the fine merely mispriced lateness, removing it should restore the original price of infinity, and punctuality with it. After roughly twelve weeks the researchers withdrew the fine. Lateness did not fall back; it stayed at the elevated level for the rest of the study, indistinguishable from the fined period.

Why should the change run one way and not the other? Because the two states are not symmetric in how they are established. A price is explicit, posted and unambiguous; it can be created by an announcement. The unpriced obligation was never announced — it was inferred from the absence of a price and from a shared reading of the relationship. Withdrawing the charge removes the charge; it does not un-tell the parents what kind of relationship they are in, and no announcement is available that would. The honest caveat is that the experiment cannot separate "the old inference was destroyed" from "it would return more slowly than twenty weeks allowed" — what it establishes is that reversal was not prompt, which for a manager weighing a fine is the operative fact either way.

Three refinements worth keeping.

First, this is not a claim that incentives fail. They worked exactly as designed, on a situation the designer had accidentally redefined. The parents optimised against the payoff structure in front of them; the error was upstream, in what that structure said the exchange was.

Second, the fix is not simply a bigger fine. A large enough charge would deter — but it deters as a price, in a relationship now commercial, and it keeps the property that paying settles the account. A centre that wants punctuality out of consideration cannot buy consideration; it can only buy compliance, at a higher rate than it expected.

Third, the boundary of the case: nothing here depends on children, on rewards, or on a taste for the activity itself — that is a related but separate mechanism. The move in Haifa is narrower and stranger. An unpriced obligation was converted into a purchasable service by the act of naming its price, and the conversion did not reverse.

• Incomplete contracts, and why leaving a consequence unspecified is sometimes deliberate.
• Whether non-monetary sanctions avoid the conversion, and what evidence exists either way.
• Gneezy and Rustichini's companion result that small payments for effort can produce worse performance than no payment at all.

Every term the collection defines is gathered in the glossary.