The UK retentions ban, explained for contractors (2026): what replaces the five percent?

Long read, August 2026. Every figure sourced inline; sources listed in full at the end. Last checked against the bill on 11 August 2026.

Cash retentions in UK construction are being banned. Not phased down, not reformed. Void, on a date already written into a bill. The Commercial Payments Bill, introduced to the House of Lords on 19 May 2026, prohibits the deduction and withholding of retention sums under construction contracts (House of Lords Library, LLN-2026-0028, June 2026). The mechanics, in one line: commencement no earlier than 2027, retention clauses still agreeable for two years after that, then a hard stop three years from commencement that the bill calls the last retention day, when every retention clause becomes void, including in contracts signed before the ban (Macfarlanes; Beale & Co; Osborne Clarke; CMS, 2026). What replaces the five percent is partly financial: retention bonds, escrow and insurance will each take a share of the job cash used to do badly. But the real replacement is not a financial instrument at all. It is proof. When nobody is allowed to hold your money against the possibility that your work is wrong, the question becomes whether you can show it is right, at the moment it is built, before it is covered up. The contractors who can will find every alternative instrument cheap. The contractors who cannot will find the new world prices them accordingly.

That is the short answer. The rest of this page is the long one: what the bill actually says, when it bites, why Parliament is doing it, what the five percent was really securing, and what to build in its place.

What has the government actually announced?

The government confirmed its intention on 24 March 2026, in its response to the late payment consultation, describing the package as the most ambitious late payment legislation in over 25 years (Construction News, 24 March 2026; Osborne Clarke, March 2026). The Commercial Payments Bill [HL] entered the House of Lords on 19 May 2026 as HL Bill 4 of the 2026-27 session. It passed second reading on 9 June 2026, completed committee stage in July with the retentions policy unchanged, and has been reprinted as amended (HL Bill 45), with report stage to be announced (bills.parliament.uk/bills/4128, checked 11 August 2026). This is a bill moving at pace, not a consultation that might drift. And it is not softening: at committee stage the Department for Business and Trade minister Lord Leong confirmed the government does not intend to introduce exceptions to the ban or to permit retentions to continue in another form, and no committee amendment introduced a carve-out (Burges Salmon, 2026). The measures:

• A retentions ban with a hard stop. The bill prohibits the deduction and withholding of retention sums under a construction contract (House of Lords Library, June 2026). Retention clauses can still be agreed or varied for two years from commencement. At the three-year mark, the day the bill calls the last retention day, every retention clause becomes void, including in contracts signed before the ban, and retained sums must be released regardless of what the contract says (Beale & Co; Osborne Clarke; CMS, 2026). Not phased down. Void.
• A penalty aimed squarely at retentions. Fail to release a retention sum within the required timescale after the last retention day and the payee is entitled to the greater of £40 or 50 percent of the retention sum, on top, plus statutory interest (Beale & Co; Osborne Clarke, 2026). Half the retention, for holding it too long.
• A 60-day cap on payment terms. Large firms must pay smaller suppliers within a maximum of 60 days, reducing to 30 days where the purchaser is a public authority (GOV.UK, 19 May 2026; Macfarlanes, 2026).
• Statutory interest that cannot be signed away. The rate is not new: 8 percent above the Bank of England base rate is the existing Late Payment Act rate, which is 11.75 percent all-in at the time of writing (CMS, 2026; Bank of England base rate 3.75 percent, August 2026). What changes is that contractual terms excluding it become void (Macfarlanes, 2026). The rate stays. The escape route goes.
• Enforcement with teeth. The Small Business Commissioner gains powers to adjudicate payment disputes and investigate persistent poor payers, with financial penalties of up to 1 percent of a business’s annual UK turnover for persistent late payment (Macfarlanes, 2026), alongside penalties for breach of the bill’s publication requirements (House of Lords Library, June 2026).

The government’s own framing: late payments cost the UK economy £11 billion a year and close 38 businesses every day (GOV.UK press release, 19 May 2026).

When will retentions actually end?

On that arithmetic, the last cash retention in UK construction will be held around the end of the decade. But the practical deadline is much closer than that, because retentions run years behind the work. A contract signed this year with a twelve-month defects period already straddles the transition. The businesses that treat the last retention day as the deadline will discover their cash flow model changed while they were not looking. The ones that treat today as the start date get a running start measured in years, and every one of those years makes the alternatives below cheaper.

Why is Parliament banning retentions?

Because the numbers stopped being defensible. The government’s own evidence base, assembled when BEIS consulted on retentions, found:

• £4.5 billion of retentions held over the course of a year in England alone, with a range of £3.2 billion to £5.9 billion (BEIS, Retentions Consultation Impact Assessment, 2017, in 2015 prices).
• £229 million a year of that money never comes back, lost to upstream insolvency: the firm holding the retention goes under before the money is released (BEIS, 2017, in 2015 prices).
• 4.8 percent average retention withheld, against the 3 to 5 percent the industry tells itself (Pye Tait review for BEIS, 2017).
• Carillion made the abstract concrete: its collapse in January 2018 took hundreds of millions of pounds of retained supply chain money with it, and forced smaller firms into insolvency for work they had already done correctly (Lord Aberdare, House of Lords retentions debates, via The Construction Index).

Set those numbers side by side with what the money is notionally for, and the case collapses. As I wrote when the ban was first signalled: retentions were never a good defect management tool in the first place. They are a blunt financial hold, not a quality assurance mechanism (Constructing Culture, The Construction Retentions Ban UK and Supply Chain Trust, 2026).

What was the five percent actually securing?

In theory, the retention was the client’s security that defects would be put right. In practice it secured two other things entirely: free working capital for whoever held it, and a signal, priced into every tender, about who carries the risk when trust is absent.

What it never secured was quality. Money held in an account does not stop a wall being built wrong, and the industry has an expensive proof of that. From Going for Gold: “The inquiries into the Edinburgh school failures found signed-off paperwork for walls nobody had actually inspected. Paper QA is not assurance.” Every one of those jobs will have carried a retention. The retention did not catch the walls, because a financial hold placed at the end of a payment chain cannot see inside one. It prices failure after the fact. It prevents nothing.

That distinction, pricing failure versus preventing it, is the key to choosing what replaces the five percent. It is also now the government’s own position: ministers defended the ban at committee stage on the basis that retentions are an ineffective and problematic quality assurance mechanism, while confirming that nothing in the bill prohibits staged or interim payments, or payment arrangements through third parties such as banks, payment agents or escrow providers (Burges Salmon, 2026).

What replaces the five percent?

The instruments below are the ones the legal and surety market is now debating (RICS Construction Journal; Clyde & Co, April 2026; Weightmans, 2026). Five of them move the risk around. Only the last one removes it.

Notice what the first five have in common: every one of them is priced on the sixth. A surety deciding your retention bond premium, an insurer writing latent defects cover, a client deciding whether to accept a PCG instead of cash: all of them are asking the same underlying question. Can this business show its work is right? The provable quality regime is not one option among six. It is the thing that makes the other five affordable.

What does a provable quality regime look like?

It is not more paperwork. Paperwork is what failed at Edinburgh. It is a different definition of progress, run as a discipline. From Going for Gold, where it holds the whole delivery method honest: “Progress is measured by what is actually finished and signed off, not by what is claimed in a meeting. Digital quality sign-offs drive a line of balance, so the programme reports the truth whether or not the truth is comfortable. And the work that gets covered up gets independently checked before it disappears behind the next trade.”

Three practical components sit inside that sentence:

• 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 moment, with someone whose name goes on the record, and the latent defect problem that retentions notionally covered largely stops existing.
• An evidence trail that survives handover. This is also the direction the law is already travelling: the Building Safety Act’s golden thread requires exactly this discipline on higher-risk buildings. The retentions ban extends the logic to money: if the evidence exists, the cash hold is redundant.

A business that runs this regime walks into a surety meeting with a defects record instead of a promise. That is the difference between a retention bond priced as a formality and one priced as a risk.

What does the ban do to supply chain relationships?

More than any other measure in the bill, because retentions were never really about defects. They were about who carries whom. Sixty to eighty percent of a project’s value is delivered by people who do not work for you, and the ban rewrites the terms on which that value arrives. From Going for Gold: “Every good partner I have ever known keeps a mental league table of their clients. The contractors who pay on time, plan properly, communicate like adults and stand by them when things go wrong get the best teams, the keenest prices and the fastest response. Everyone else gets what is left over.”

The league table is about to be repriced. When nobody can hold cash, payment behaviour becomes one of the few remaining signals of what you are like to work for, and the bill makes that behaviour public and enforceable. The commercial upside is already measurable today: “The contractor down the road who pays fairly and communicates well is quietly getting five to ten percent better pricing for identical work. Not because they negotiate harder, but because their supply chain trusts them enough to sharpen the pencil.”

I have run this play from the buying side. At a previous business we were using 45 different M&E partners across 92 projects, and twenty percent of those partners went bust, with M&E running at half our total project cost. We cut the partner pool dramatically, built performance frameworks on objective data, moved procurement from lowest cost to best value, and gave partners direct sight of their own performance data, weekly, honest, specific. Within twelve months the business was profitable, and M&E insolvency has been zero since the strategy launched. None of that needed a retention. It needed the trust the retention was substituting for.

Which is the point I keep returning to: the businesses that will benefit most from this ban are the ones that do not wait for the legislation before treating their supply chain as partners rather than risks (Constructing Culture, 2026).

What should contractors do now?

1. Put a number on your exposure, in both directions. Every retention you currently hold, every retention held on you, with release dates. The ban converts that ledger into a one-off cash flow event at the last retention day. Most businesses have never seen it on one page.
2. Read your own contracts. Find where retention sits in your standard JCT and NEC amendments and in your terms with trade partners, and decide now what those clauses become when the law makes them void. Waiting for the other side’s lawyers to decide for you is a negotiating position, just not yours.
3. Open the surety conversation early. If retention bonds are part of your answer, capacity and pricing will follow your quality record, and the market will harden as the deadline approaches (Clyde & Co, April 2026). A track record of provable quality starts being cheap exactly as long before you need it as you spend building it, and the bill has just told you how long you have.
4. Build the quality evidence regime now. Digital sign-offs, progress measured by what is finished and signed off, independent checks before cover-up. It takes a full project cycle to embed, which is precisely the time the transition gives you. It is also exactly the ground the Going for Gold programme rebuilds, one business per cohort, on live projects.
5. Get your payment mechanics ready for the 60-day world. Statutory interest at 8 percent over base will no longer be something you can contract out of, and persistent poor payment will carry fines of up to 1 percent of UK turnover. Invoice verification, certification speed and payment runs stop being back-office detail and become published performance.
6. Ask your three most important partners what it is like to work for you. Then be quiet and listen. If they hesitate, that is the answer. Every instrument above is priced, in the end, on what they say.

The bottom line

A retention was a substitute for trust, held in cash. Parliament is removing the substitute. It is not removing the need.

The transition begins no earlier than 2027, and on the bill’s own arithmetic the last cash retention is released by around 2030. The firms that spend those years building the real thing, provable quality and a supply chain that wants to work for them, will find the new world cheaper than the old one: keener partner pricing, bonds priced as formalities, and jobs that finish because defects are caught before they are covered up. The firms that spend the transition lobbying for a softer landing will get to the same deadline with nothing in place. This is the end of the free credit era in construction (Constructing Culture, UK Late Payment Crackdown 2026). What replaces it is the thing the credit was standing in for.

The delivery method behind this piece, planning the finish from the first week, progress measured by what is signed off, and a supply chain treated as partners, is in Going for Gold: Constructing Project Managers, written from twenty-two years of building jobs. All royalties go to the Lighthouse Construction Industry Charity.

Sources

• Commercial Payments Bill [HL], HL Bill 4 of session 2026-27, bills.parliament.uk/bills/4128. Introduced to the House of Lords 19 May 2026; second reading 9 June 2026; committee stage completed July 2026; reprinted as amended (HL Bill 45); report stage to be announced (checked 11 August 2026).
• House of Lords Library, Commercial Payments Bill [HL], briefing LLN-2026-0028, June 2026: prohibition on deduction and withholding of retention sums; Small Business Commissioner penalty powers for breach of statutory publication requirements.
• GOV.UK, “Largest crackdown on late payments in over 25 years as landmark Bill enters Parliament”, press release, 19 May 2026: 60-day cap, 8 percent over base interest, £11 billion annual cost, 38 business closures a day.
• Government response to the late payment consultation, announced 24 March 2026 (reported by Construction News, 24 March 2026; Osborne Clarke insight, March 2026).
• Macfarlanes, “Paid in 60 days: the Government’s new Commercial Payments Bill”, 2026: commencement no earlier than 2027; 30-day terms for public authority purchasers; penalties up to 1 percent of annual UK turnover; non-excludable statutory interest.
• Beale & Co, “The Commercial Payments Bill: Key changes for UK Construction”, 2026: the three-year transition and the last retention day; the penalty of the greater of £40 or 50 percent of an unreleased retention sum.
• Osborne Clarke, “Commercial Payments Bill would ban retention and late payments in UK construction”, 2026: retention clauses void three years from the first day of the transition period.
• CMS, “The retentions ban under the microscope”, 2026: the two-year window in which retention clauses can still be agreed or varied, with release required regardless of contract terms a year later; the current all-in statutory interest rate of 11.75 percent.
• Burges Salmon, “Government shows no sign of backing down on retentions ban”, 2026: Lord Leong’s committee-stage confirmation of no exceptions and no retentions in another form; no softening amendments at committee.
• BEIS, Retentions Consultation Impact Assessment, 2017: £3.2 billion to £5.9 billion held over a year in England, central estimate £4.5 billion; £229 million a year unpaid due to upstream insolvency (2015 prices).
• Pye Tait for BEIS, Retention Payments in the Construction Industry, 2017: average retention 4.8 percent.
• Lord Aberdare, House of Lords retentions debates (via The Construction Index, “The injustice of retentions”): Carillion’s collapse and the loss of hundreds of millions of pounds of retained funds.
• Alternatives analysis: RICS Construction Journal, “What are the alternatives to retention?”; Clyde & Co, “Rethinking performance security”, April 2026; Weightmans, “Retention reform and replacement”, 2026.
• Andy Pritchard, Going for Gold: Constructing Project Managers, second edition, 2026; and Constructing Culture articles: The Construction Retentions Ban UK and Supply Chain Trust (2026); UK Late Payment Crackdown 2026: What the New 60-Day Cap Means (2026).