Retention is one of the oldest security mechanisms in construction contracting, and one of the most resented. A slice of every certified payment – typically 5% – is deducted before it reaches the contractor’s account and held back, sometimes for years, as insurance against defects and non-performance. For subcontractors especially, that money is genuinely theirs: working capital sitting idle in someone else’s bank account until a release event finally arrives, if it arrives at all. It’s no surprise retention disputes are a recurring feature of adjudication referrals, or that the UK government has now committed to banning the practice outright.
This guide covers what retention is actually meant to protect, how much is held and under what mechanics, how release is supposed to work (and where it routinely goes wrong), the UK’s proposed retention ban announced in March 2026, and the quite different statutory model already in force in parts of Australia.
What retention is actually for
Retention exists to give the paying party – employer or main contractor – security that the works will actually be finished and any defects put right, without having to chase the money separately through the courts or an insolvent balance sheet. It’s a safeguard against non-performance, not a float the paying party is entitled to trade with or use to shore up its own cash flow, however tempting that becomes on a tight project. Treating retention as free working capital rather than someone else’s security is exactly the habit that has driven the UK’s current reform push, and it’s worth QSs on the paying side remembering that distinction even before legislation forces the point.
How much, and how it’s structured
Most UK standard forms default to 5% of the value certified, though the actual rate is whatever the parties agree and state in the contract particulars – subcontractors should always check the main contract rate and make sure their own subcontract doesn’t apply a higher one. The mechanics differ by form:
- JCT treats retention as optional but conventional: the percentage and a Retention Release Date are stated in the Contract or Sub-Contract Particulars, and release is paid on the next interim payment date after that date arrives, provided there are no outstanding defects.
- NEC4 doesn’t carry retention in its core clauses at all – it’s added through secondary option X16, with the percentage set in the Contract Data. X16 is unusual in letting the client specify a “retention-free” amount, so most progress payments can run retention-free with a higher rate applied only to the last few, and NEC4 also lets the contractor provide a retention bond instead of cash retention being deducted at all – worth weighing against the more familiar performance bonds and parent company guarantees used elsewhere in the contract.
- FIDIC forms typically retain a percentage of each payment up to a stated cap, with half released on taking-over and the remainder at the end of the defects notification period – the same half-now, half-later logic most UK forms use, just under different clause numbers.
- Release simply doesn’t happen on time. Practical completion or the end of the defects period passes and the retention isn’t paid – often because nobody on the paying side diarised it, or because it’s being used, consciously or not, as leverage on an unrelated dispute.
- Insolvency wipes it out. Cash retention held by an insolvent main contractor or employer is, in the absence of a trust arrangement, an unsecured debt – the subcontractor queues with every other creditor for money that was always meant to be theirs.
- Nobody can see where the money actually is. Without a ledger or separate account, retention is just a number in a ledger entry on the paying party’s own books, with no visibility for the party it belongs to.
- Disagreement over whether defects justify withholding more than the retention covers. This is where retention disputes most often end up in formal process – and statutory adjudication is usually the fastest route to a binding answer, given retention claims are typically well short of justifying litigation.
Release mechanics – and where the rules actually come from
The standard pattern splits retention in two. The first half is released on practical completion – the point the works (or that portion of them) are complete enough to be used for their intended purpose, a test with no judicial definition, which is itself a frequent source of disagreement. The second half is due at the end of the defects liability period, once any defects notified have been made good.
One technical point trips up a lot of subcontracts: under the Housing Grants, Construction and Regeneration Act 1996 as amended from October 2011, a subcontract’s payment and release provisions have to be triggered by an event under that subcontract itself, not by a certificate issued under the main contract above it. A subcontract that still ties the second half of retention to, say, the main contractor’s certificate of making good defects is non-compliant, and the Scheme for Construction Contracts’ default terms step in instead. It’s a clause worth checking on any older template still in circulation. Most forms also set a minimum retention balance – commonly £250 if nothing else is agreed – below which the whole remaining sum must be released once the first half is paid, rather than being carried as a trivial balance indefinitely.
Where retention disputes actually happen
A handful of patterns account for most of the friction:
The reform that’s coming, and why it changes the calculus now
On 24 March 2026, the Department for Business and Trade announced it will ban the withholding of retention payments under construction contracts outright, describing it as the most significant legislation to tackle late payment in over 25 years. It followed a 2025 consultation that weighed an outright ban against simply requiring retentions to sit in third-party trust accounts; the government chose the ban. It arrives alongside a wider late-payment package: a mandatory 60-day maximum payment term for large firms paying smaller suppliers, statutory interest of 8% above the Bank of England base rate on commercial late payments, and expanded powers for the Small Business Commissioner to investigate and fine persistent late payers.
None of this is law yet. It needs primary legislation, a further consultation on implementation is still to come, and legislation is expected later in 2026 with an anticipated transition period of twelve to twenty-four months once it lands. But the direction is settled enough that it’s worth planning around now rather than waiting: employers and main contractors will need alternative security – performance bonds, defects bonds, parent company guarantees or latent defects insurance – standard forms including NEC4, JCT and FIDIC will all need amending, and contract administrators will be relying on inspection and defect-management discipline rather than a cash lever they can simply withhold. Industry bodies have also flagged a real risk that paying parties restructure payment schedules to replicate the effect of retention by other means, so anti-avoidance provisions are likely to follow.
Australia already runs a different model
While the UK debates whether to ban retention, several Australian states already ring-fence it by law. In New South Wales, head contractors on construction projects worth $20 million or more must hold retention money – typically up to 5% of the subcontract value – in a dedicated trust account at an authorised deposit-taking institution, notify the regulator of the account details within 14 days of opening it, and provide subcontractors with a ledger of deposits and withdrawals at least every six months. The point is exactly the one the UK reform is chasing: if the head contractor becomes insolvent, the retention sits in a protected account rather than the general pool of assets available to creditors. Other states run their own project trust account and security-of-payment regimes aimed at similar protection, with different thresholds and mechanics, so it’s worth checking the specific state legislation on any project rather than assuming NSW’s rules travel.
What this means for you
Whichever side of the retention line you sit on, three habits are worth building in now. Diarise release against the contract’s own trigger event – practical completion, the Retention Release Date, the end of the defects liability period – rather than relying on someone remembering an upstream certificate that may not even be a valid trigger any more. Track retention explicitly as its own line in your cost value reconciliation rather than letting it blur into general debtors, so everyone can see exactly what’s held and why. And start reviewing which of your contract templates will need amending once the UK’s retention ban actually lands – because on current signals, it’s a question of when, not if.