Retentions ban means quality work, installed, certified and paid

The Commercial Payments Bill will end cash retentions in UK construction. After commencement, expected no earlier than 2027, contracts undergo a short transition, and then comes the day the Bill’s drafters call the last retention day: every retention clause in every construction contract stops working, including those already signed.

Failure to release money owed after that will result in a penalty that is the greater of £40 or half the retention debt, with statutory interest on top.

Most of the industry’s reaction so far has been about the money, and no wonder. By the government’s own assessment, around £4.5bn sits in retentions in England in any year, and when a business above you fails, your share tends to disappear with it.

But the most telling moment in the whole debate came at committee stage, when the minister ruled out exemptions and any return of retentions in another form, and restated the government’s view that retentions are ineffective as a mechanism for assuring quality. Not unfair, but ineffective. The cash never fixed a defect – it sat in someone else’s account while the defect was covered up by the next trade.

That verdict is the opportunity. The retentions ban is a lever to drive exactly the behaviour the Building Safety Act is already targeting: quality work, installed and certified. The Act’s golden thread requires higher-risk buildings to carry an evidence trail from design through occupation, kept digitally: what was built, what changed and how it complies. The ban quietly extends the same logic to money. If the proof exists, the cash hold is redundant. If the proof does not exist, the cash hold was never protecting anyone anyway.

So what does pulling the lever the right way look like? Pay on proof: work installed, inspected, certified, then paid. Three disciplines make that real on a live job.

Progress means signed off, not claimed. The programme moves when the quality record moves. A job run this way cannot quietly build over an unresolved defect, because the unresolved defect is what is holding the programme.

Independent checks before cover-up. The moment of truth for most defects is the moment the next trade hides them. Inspect before that, with a name on the record, and the latent defect problem that retentions notionally covered largely stops existing.

An evidence trail that survives handover. Not a room of boxes assembled in the last month, but a record built as the job runs, the way the golden thread already demands.

None of this is aspiration any more. The tools arriving on sites make it systematic. Field quality apps put certification at the workface, with photographs, location and a name attached at the moment of installation.

Digital sign-offs can drive the programme directly, so reported progress is whatever the certified record says it is, not what was claimed in a meeting. And a payment application drawn from that record changes the incentive completely: the cash motive lands purely on delivering quality work on time, because that is the only thing that gets certified and the only thing that gets paid.

My sincere hope is that we do not spend the transition years doing what this industry has sometimes done best, which is finding a new legal loophole to hold a partner’s money and call it something else. The Bill anticipates us: the door to exceptions, and to retentions under another name, has already been closed. If our energy goes into inventing the next hold, we will have kept the worst of the old system (the cash pressure on the supply chain) and thrown away the chance the ban offers.

The contractors who use the lever positively will see performance improve twice over. Quality improves first, because defects get found while they are still cheap, before they are covered up. But the bigger prize is programme. A job that cannot build over its own defects does not accumulate the hidden rework that surfaces in the final months and wrecks the completion date.

A job that is not rushed at the end is a safer job and usually a more profitable one. In an industry where the average margin is 2.4% and the direct cost of avoidable error runs at around 5% of project value – roughly twice the profit – health and safety and margin both improve as by-products of simply not having to rush.

The window is real, but it is not long. Commencement is expected no earlier than 2027 and the last retention day is around three years after that. That is the time a business has to make its quality provable. Start with one project: count how much of this month’s reported progress is certified rather than claimed, and how many inspections happened before work was covered up rather than after. The answer is usually uncomfortable, and it is the baseline everything else builds from.

The businesses that arrive at the last retention day with a certified record instead of a promise will find their bonds priced as a formality rather than a risk. The ones still hunting for somewhere to hold the cash will have missed the point twice.

Andy Pritchard is a business improvement consultant, Construction Manager of the Year Gold Medal winner and author of Going for Gold: Constructing Project Managers.