Retention money

Why would a builder agree to let a client hold back part of his payment?

A builder does the work, the work is certified, and a slice of the money — commonly around five per cent under standard forms, though the figure is negotiated and varies — is simply not paid. Half of it comes back at practical completion; the rest waits out a defects period of six or twelve months. He finances the gap himself, and he may have to chase it.

Every instinct says he should refuse. He has the labour cost, he has the materials cost, and he is lending the client his own margin at no interest. Yet builders sign these terms routinely, and the ones who tender without accepting them often do not get the work. What is he being paid for, if not in money?

Start by asking what the client is afraid of, and when the fear bites. Not during the job — during the job the client has all the leverage there is, because the next payment has not been made and the builder wants it. The fear is about afterwards: the screed that lifts in month four, the flashing that leaks in the first winter, the sockets that were never quite finished.

Now ask what the builder's incentives look like the day after final payment. He has the money. Coming back costs him a van, two men and a week, and produces no revenue. The client's remedy is to sue — slow, expensive, and worth less than the defect in most cases. So the builder's rational move, at that moment, is not to come back promptly.

Here is the trap that follows, and it is the heart of it. The builder may be entirely honest and fully intend to return. It does not help him. The client cannot tell an honest builder from a plausible one at tender stage, because both make the same promise at the same cost — which is nothing. A promise that costs nothing carries no information, so the client must price for the possibility that this builder is the other sort. The good builder is taxed for the bad one's behaviour, and has no way to say otherwise, because saying is free.

So the question turns around. It is not "why does the builder tolerate retention?" but "how can a builder who will come back make that believable?" The answer has to be an action that is cheap for him and expensive for the sort of builder he is not. Merely asserting it fails that test. Leaving money in the client's hands passes it: the builder who intends to return loses only the use of the sum for a year, while the builder who intends to disappear loses the sum outright. Same term, wildly different cost, depending on which type you are.

That is what makes retention a commitment rather than a fee. It changes the payoffs facing his own future self. On the day the client rings about the leaking flashing, the builder is no longer choosing between a week's work and nothing; he is choosing between a week's work and forfeiting a sum that may exceed it. The remedy stops depending on his goodwill or on a court, because the money is already sitting on the client's side of the table.

Then ask why it is deducted continuously through the job rather than paid as a lump at the end. Because it has to accumulate as the exposure does. Halfway through, half the defect risk exists and roughly half the retention is held. And the staged release maps onto the two distinct risks: the first half returns at practical completion, when the risk of non-completion has gone, and the balance at the end of the defects liability period, when the risk of latent faults has largely gone.

Three refinements.

First, the builder is not doing this for nothing. He is buying the job: where a client can hold retention, tenders that refuse it are either excluded or priced as riskier. The compensation is diffuse but real — he is paid in the currency of being awarded work.

Second, retention is one instrument in a family, and the family shares a single logic — put something of the promisor's at stake so the promise stops being free. A performance bond does the same job through a third party who pays out on default; a parent-company guarantee does it by widening who can be pursued; a retention bond substitutes an insurer's undertaking for the cash, which is strictly better for the builder because the money stays in his account. That these instruments substitute for one another in negotiation is good evidence that what is being bought is credibility rather than a cash cushion.

Third — the honest complaint — the mechanism is regularly abused. Retention held long past the defects period, released only after chasing, or lost outright when a main contractor becomes insolvent partway up the chain is a well-documented problem, and it falls hardest on small subcontractors furthest from the client. Several jurisdictions have moved to hold retentions in trust or restrict them. That abuse does not refute the mechanism; it shows what happens when a security given to one party is itself held with no security.

What would refute the account? If retention were principally a cash-flow device, clients would drop it where the builder posts a bond instead — and by and large they do accept the substitution, which points at credibility rather than cash.

• Why retention bonds and project bank accounts are pushed as reforms, and what each fixes about who holds the money.
• How the same commitment logic appears elsewhere in the trades — deposits, sale-or-return, and holding back the last stage payment on a domestic job.

Every term the collection defines is gathered in the glossary.