Ask an employer what actually protects them if a contractor walks off site three months before completion, or a key subcontractor collapses mid-package, and the honest answer is usually a stack of paperwork almost nobody has read closely: bonds and guarantees. They’re two of the most common security instruments in construction contracts, and two of the most commonly misunderstood – teams use the words interchangeably, assume every bond pays out the same way, and often only discover what their security actually says on the day they try to call on it.
This guide sets out what bonds and guarantees actually are, the main types you’ll meet on a project, how on-demand and conditional bonds differ, how NEC and JCT each deal with them, what they cost, and the mistakes that quietly leave an employer under-protected.
A bond and a guarantee aren’t automatically the same thing
Both are promises that a third party will make good a loss if the contractor doesn’t perform, but the label on the document tells you surprisingly little about how it actually behaves. A bond is typically a three-party arrangement: a bank or surety issues an undertaking to the employer on the contractor’s behalf, in return for a fee. A guarantee, in the form most QSs meet day to day – the parent company guarantee – is a promise from a related party, usually the contractor’s parent, to answer for the contractor’s obligations, rather than an independent institution charging a premium for the risk. The more important distinction cuts across both: whether the instrument pays out on demand or only once default has been proven. As practitioner guidance on this puts it, the name of the document isn’t reliable – it’s the wording inside it that decides how, and how easily, the security actually pays out.
The main types of bonds you’ll meet on a project
Different bonds cover different stages of a project’s life, and more than one is often running at once:
- Bid (tender) bond – taken out at tender stage to guarantee a successful bidder will actually enter into the contract on the terms they tendered; if they withdraw, the surety compensates the employer, usually for the cost of re-tendering.
- Performance bond – the one most QSs deal with most often, guaranteeing the contractor’s performance for the duration of the contract. Conventionally set at around 10% of the contract value, though that figure is negotiable and should reflect the project’s actual risk rather than being lifted unchanged from the last job.
- Advance payment bond – required wherever the employer makes an upfront payment (commonly for mobilisation or long-lead materials) before value has been earned on site; it protects that payment against the contractor failing to perform or becoming insolvent before it’s worked off.
- Retention bond – issued in place of cash retention, guaranteeing an agreed sum in the event of non-performance or insolvency instead of the employer physically holding money back from certified payments. Popular with contractors because it improves project cash flow.
- Payment and off-site materials bonds – less universal, but worth knowing: a payment bond covers sums due further down the supply chain, and an off-site materials bond secures payment for materials manufactured or stored away from site before delivery.
Parent company guarantees: cheaper, but only as strong as the parent
A parent company guarantee is often presented as a lower-cost alternative to a third-party bond, and in pure premium terms it usually is – there’s no bank or surety charging a fee for taking on the risk. But that lower cost is the trade-off: a PCG is only ever as good as the parent’s own balance sheet, and if the parent becomes insolvent alongside (or instead of) the subsidiary, the guarantee is worth the paper it’s written on and nothing more. Under JCT 2024, both performance bonds and PCGs are optional provisions, required only in the form specified in the Contract Particulars – the employer has to actively decide to ask for one and specify what it should say. Before accepting a PCG as sufficient security, do real due diligence on the guarantor’s financial standing rather than treating “a guarantee from head office” as automatically equivalent to a bank-backed bond.
On-demand vs conditional bonds
This is the distinction that actually matters when a bond needs to be called. An on-demand bond is a primary obligation: the surety pays out on written demand, without the employer having to prove the contractor is in breach or quantify a loss first. It’s fast and simple for the employer, which is exactly why contractors resist it – it can be called speculatively, with the contractor left to argue afterwards, often through litigation, that the demand was unjustified. A conditional (or default) bond is a secondary obligation instead: it only pays once the employer has demonstrated an actual breach and, usually, the resulting loss. UK domestic construction leans heavily towards conditional bonds – the ABI Model Form is a common example of the wording used – while on-demand bonds are far more common internationally, often driven by funder or lender requirements on projects using forms like FIDIC. It’s one more way the major contract suites genuinely diverge in practice, covered in more detail in our comparison of NEC, JCT and FIDIC.
How NEC and JCT actually deal with bonds
Under NEC4, a performance bond is incorporated through secondary option X13 in most of the main contracts (the Alliance Contract and Facilities Management Contract use option X4 instead). The bond amount is stated in the contract data – conventionally 10% of the contract value – and the contractor must provide it within four weeks of the contract start date, with the Project Manager required to accept the guarantor’s creditworthiness before it counts. NEC doesn’t publish its own standard bond wording, as the NEC’s own guidance confirms, so the employer has to specify the form in the scope. Fail to provide it, and the client can terminate after notice and the standard four-week cure period.
JCT takes a similar “optional but specified” approach: bonds and PCGs are only required where the Contract Particulars call for them, in whatever form is stated there. JCT does publish its own standard Advance Payment Bond, but there’s no single mandated form for performance bonds generally, and many contracts are amended to let the employer withhold a percentage of the contract sum until the required bond or PCG is actually in place – because the employer is exposed for every day it hasn’t been provided.
What it actually costs
Bond premiums are typically priced between 1% and 8% of the bond value, plus a flat administration fee commonly in the region of £700 to £1,500, with the exact rate driven by the contractor’s financial standing, the contract’s risk profile, sector and track record. That cost isn’t abstract to the QS: under priced contracts it’s a billable item that should show in the tender price, and under cost-reimbursable arrangements it belongs within the contractor’s fee percentage rather than quietly absorbed.
Where QSs get caught out
- Wording that doesn’t match the underlying contract. A bond referencing the wrong contract value, an out-of-date completion date, or defined terms that don’t line up with the building contract is a genuine gap in security – check it against the executed contract, not the tender draft.
- Bond expiry that doesn’t track the programme. If practical completion slips, or the rectification period runs on, a bond written to expire on the original completion date can lapse while real exposure is still live. Extending it is a renewal to diarise, not something automatic.
- Varying the contract without telling the surety. Materially changing the underlying contract without the bondsman’s consent can discharge the bond entirely under English law, unless the wording includes an indulgence clause specifically allowing variations while keeping the surety on the hook. Flag any material amendment to whoever issued the bond, not just to the contractor.
- Treating a PCG as equivalent to a bank bond. A guarantee from a parent with a thin balance sheet, or one based overseas and hard to enforce against, is materially weaker security than it looks – and that only becomes visible at the worst possible moment, when it’s actually needed.
- Letting work start before the security is in place. A bond protects the employer’s exposure from day one; a security document still “being finalised” three months into the works isn’t protecting anything, much like an insurance policy that hasn’t incepted yet doesn’t cover a loss that happens the week before it does.
What this means for you
Bonds and guarantees are one of the few areas of a construction contract where the paperwork genuinely is the protection – there’s no fallback if the wording doesn’t say what you thought it said. That makes three habits worth building into every project, whichever form of contract you’re on: read the actual bond or guarantee document rather than assuming it matches the standard clause it’s supposed to support, check its value, expiry and coverage against the current programme rather than the one at contract signing, and flag any variation to the underlying contract to the party who issued the security. None of that is glamorous, but it’s the difference between a bond that’s genuinely there when you need it and one that turns out, at the worst possible moment, to be protecting a project that no longer exists.