Chapter 9: Liquidated damages

Without the clause, an employer whose building is late must sue for damages in the ordinary way. It has to prove that the delay caused it loss, and how much. That is hard. The losses are real but spread thin: rent not received, finance carried for longer, extra professional fees, the cost of staying in old premises. Each needs evidence, and each invites argument about remoteness and mitigation. Lord Leggatt, in the Supreme Court in Triple Point (2021), called working out what delay has cost the employer an intractable task that breeds costly disputes. Fixing the sum in advance removes the argument.

The contractor gains as much. Without the clause, its liability for lateness has no ceiling. With it, the contractor knows the price of each week late and can allow for that risk in its tender. Lord Leggatt gave this as the clause's second purpose. It limits the contractor's exposure to a liability that would otherwise be unknown and open-ended, and it gives the employer certainty. Each side can then manage the risk of delay, which is why the courts are slow to disturb the clause.

The standard forms use different names for the same thing. JCT calls them liquidated damages (Design and Build 2016, cl 2.29). NEC4 calls them delay damages (Option X7), and FIDIC calls them Delay Damages (Sub-Clause 8.8). This chapter says "liquidated damages", or LDs, for all three. It calls the parties the employer and the contractor; NEC calls the employer the Client. JCT clause numbers here are from the Design and Build Contract 2016, which JCT withdrew on 31 March 2026. Check them against the 2024 edition before relying on them.

The clause has a partner: the extension of time clause. An employer that delays the work cannot fairly charge the contractor for that delay. If the contract had no way to move the date, the law would stop the employer relying on the date at all. That is the prevention principle, which chapter 3 explains. The extension clause moves the date instead. The employer keeps a fixed date and its LDs, and the contractor is not charged for delay the employer caused. The Court of Appeal in North Midland (2018) said these clauses were designed to protect employers ([12]). It also said that their primary purpose is to relieve the contractor of LDs for delays that were not its responsibility ([44]). Both are true.

Because the parties agreed the measure, the agreed sum is the only damages the employer gets for the delay it covers. The employer cannot set the clause aside and claim a larger loss. The contractor cannot insist on paying a smaller one. Either would undo the bargain in whichever direction suited the party asking. The Court of Appeal held in Temloc v Errill (1987) that such a clause is an exhaustive agreement about the damages payable for late completion (as quoted in Biffa v MEH [164]). It need not say so expressly; the question is one of interpretation (Chattan v Reigill [31]).

The rule holds even when the agreed sum is nothing. In Donne Place v McDonnell (TCC, 2011), a house-building contract said that if the contractor finished late it "will incur no penalty" ([175]). The judge held that this was an exhaustive agreement that no damages were payable ([180]). A claim for 7 months' lost rental income and wasted expenditure failed ([181]). FIDIC says the same in terms: Delay Damages are the only damages due for late completion, except where the employer terminates for the contractor's default (Sub-Clause 8.8).

The employer sets the rate before the contract is signed, and it should build the rate from the losses it expects. The SCL Protocol lists the employer's calculations of its LD rates among the tender records to keep (1.37(a)). It says such documents may be relevant to whether the LD provisions can be enforced (1.38). An employer that can show how it reached its rate is well placed if the rate is ever challenged.

A court enforces the bargain the parties made. The rule against penalties is a narrow exception. It stops one party using a clause to punish the other, which serves no interest the law protects. The Supreme Court restated the rule in Cavendish v Makdessi in 2015. A clause that bites on breach is a penalty only if it imposes a detriment out of all proportion to the innocent party's legitimate interest in performance. The innocent party can have no proper interest in simply punishing the other side ([32]).

For a long time the question was put as a choice. Was the sum a genuine pre-estimate of loss, or a penalty? Cavendish dropped that framing. The real question is whether the clause is penal. A damages clause may be neither a pre-estimate nor a penalty, or both. That matters in construction, where a rate is often set roughly, by a formula or as a percentage of the contract sum. A rough rate is not a penal one.

For an LD clause the yardstick is still loss. Cavendish said that under a straightforward damages clause the innocent party's interest will rarely extend beyond compensation for the breach ([32]). Lord Hodge gave the test for such a clause. A penalty needs an extravagant disproportion between the agreed sum and the highest loss that could possibly arise from the breach ([255]). The comparison is with what might have happened, not with what did. The rate is agreed before anyone knows what the delay will cost. So a rate that looks high against the employer's actual loss can still be perfectly valid.

The court starts on the side of the bargain. Between properly advised parties of comparable bargaining power, the strong initial presumption is that they were the best judges of what was legitimate ([35]). The party alleging a penalty must prove it. The Court of Appeal said so in 2026 in Houssein v London Credit [80], a loan case, which is a reminder that the rule is general contract law. Nor does the label decide anything. The clause in Triple Point called its sum "the penalty", and it was common ground that it was an LD clause all the same ([87]).

The presumption is hard to shift. In TTSJV v BapCo, a 2026 interim decision on a refinery contract governed by English law, the TCC applied it to carefully negotiated provisions between commercially sophisticated parties. On its own evidence at an urgent interim hearing, the contractor argued that full LDs were penal because they did not reflect a partial take-over of the plant. At that interim stage it had shown only a potentially arguable case ([37]).

Penalty challenges to negotiated construction rates seldom reach the reported cases. In Buckingham v Peel (section 4) the contractor attacked the LD clause on other grounds, and none of the declarations it sought rested on penalty ([25]). If a rate were held penal, the clause would probably fall with it, and the employer would have to prove its loss. Section 4 deals with what then happens to the cap.

LDs price one thing: the employer being without its works after the date it was promised them. So they start at the completion date, as extended, and stop when the employer gets the works. JCT charges them for the period between the Completion Date and the date of practical completion (cl 2.29.2). NEC4 runs them from the Completion Date until Completion or the Client's take-over, whichever is earlier (X7.1). FIDIC runs them to the Date of Completion stated in the Taking-Over Certificate (Sub-Clause 8.8). The words differ, but the logic is the same.

Practical completion is the end point that causes most argument. JCT does not define it, so the courts have. In Mears v Costplan (2019) the Court of Appeal said it is easier to recognise than define. It summarised it as a state in which the works are completed free from patent defects, other than trifling ones ([74]). Latent defects cannot prevent it. Whether a defect is trifling is a matter of fact and degree, measured against the purpose of letting the employer take possession and use the works as intended. That purpose is why LDs stop there. From then on the employer has what it bargained for, and later defects are a matter for the defects provisions and damages.

The other forms reach the same place by different words. NEC4 Completion needs the work the Scope requires done, and the notified Defects corrected which would have prevented the Client from using the works (cl 11.2(2)). FIDIC's Taking-Over Certificate can issue with minor outstanding work that will not substantially affect safe use of the works (Sub-Clause 10.1).

An extension of time moves the completion date, so LDs start later. The SCL Protocol puts it plainly: the main effect of an EOT is to relieve the contractor of liability for LDs for the period of the extension. The extension is measured by causation. An employer-risk event moves the date only if it delays completion, which usually means it sits on the critical path, and only by the delay it causes. A breach by the employer that holds up nothing important gives the contractor no time, because it has caused no lateness from which the contractor needs relief. Delay the contractor causes stays its own, and that is what the LDs charge for.

JCT adds a notice procedure before the employer can take the money. The contractor should know that a deduction is coming, and the Construction Act requires a payer that means to pay less than the sum due to say so. Under DB 2016 the employer must issue a Non-Completion Notice when the contractor misses the Completion Date (cl 2.28). Before the due date for the final payment, it must notify the contractor that it may require payment of LDs or withhold them (cl 2.29.1.2). Then it gives a notice requiring payment, or stating that it will withhold or deduct them (cl 2.29.2). To take them from a payment it also needs a Pay Less Notice (cl 2.29.2, footnote; Construction Act 1996, s.111). If a new Completion Date is fixed later, the Non-Completion Notice is cancelled and a further one is needed (cl 2.28).

The order of the notices matters; the gap between them does not. In S&T v Grove (2018) the employer, on an amended JCT Design and Build 2011 form with the same clause numbers, sent the last two notices one after the other. The final notice arrived 7 seconds later ([116]). Sir Rupert Jackson, for the Court of Appeal, accepted that on those facts the procedure gave no obvious benefit to anyone. But the clause required no more than notices in a set sequence, and the employer recovered its LDs ([121], [124]).

An employer can also trip over conditions it wrote for itself. In DBS v Tata (Court of Appeal, 2025) an IT contract provided for Delay Payments, akin to liquidated damages ([3]). The employer had to issue a non-conformance report first. It had not. The clause was a condition precedent, so the omission was fatal to the claim ([70]). Chapter 4 sets out the test for a condition precedent.

If the date moves after LDs have been taken, the money moves with it. Under JCT the employer shall pay or repay LDs taken for the period up to the later Completion Date (cl 2.29.3). NEC4 goes further: the Client repays the overpayment of damages with interest (X7.2). The JCT clause does not mention interest. An employer that deducts on the strength of an interim refusal takes a risk. Under DB 2016 the decision can be reviewed up to 12 weeks after practical completion (cl 2.25.5), and chapter 5 explains how that works.

One rate fits one building. Many jobs are handed over in pieces. A school opens its first block; a developer lets its first phase. Under sectional completion, each part has its own completion date and usually its own rate. With partial possession, which is less planned, the employer takes over part of the works before the whole is finished. Either way, the employer's loss from lateness falls once it has the use of part. So the rate should fall too.

The forms do this in different ways. JCT reduces the rate by the same proportion as the value of the Relevant Part bears to the Contract Sum or the Section Sum (cl 2.34). NEC4 reduces it by the Project Manager's assessment of the benefit to the Client of the part taken over, as a proportion of the benefit of the whole not yet taken over (X7.3). The FIDIC 2017 Red Book (first printing) reduces the daily rate by the value of the part, but not the maximum amount of Delay Damages (Sub-Clause 10.2). The same handover can give different numbers under different forms. Chapter 16 compares them.

In some older cases, LD clauses failed because their machinery for parts did not work. In Bramall & Ogden v Sheffield City Council (1983), a housing contract reduced LDs pro rata for each completed dwelling, and that machinery simply did not work (as described in TTSJV [31]). Judges now try harder to make a clause work. O'Farrell J in Eco World v Dobler (TCC, 2021) warned against turning the older cases into an inflexible rule. As she read them, those clauses failed because of errors in drafting (as quoted in Buckingham [73]-[74]).

A court is reluctant to hold a provision void for uncertainty and will adopt a meaning that gives effect to it if it can (Buckingham [39]). The clause is a bargain made when the contract is signed, so it is tested against what the parties then knew. Whether the machinery works is judged as at the contract date, not with hindsight (Mansion Place v Fox [97], on an amended JCT DB 2016, said obiter).

Most LD clauses also carry a cap, often a percentage of the contract sum. FIDIC puts the maximum amount of Delay Damages in the Contract Data (Sub-Clause 8.8). The cap is the contractor's half of the certainty bargain. It protects the contractor only while the clause stands. If the contractor knocks the clause out, the question is whether the cap goes with it. That is a question of construction. Did the parties agree a limit on liability for delay generally, or only a limit on the LDs?

Buckingham v Peel (TCC, 2022) shows how the question arises. The contractor argued that the LD provisions were void and unenforceable, and that its cap would then limit any general damages instead ([2]). It lost the first point: the provisions were certain and enforceable ([88]). The judge dealt with the cap anyway, while saying it was unnecessary ([90]). The cap was on "Maximum LADs", not on anything other than LADs ([98]). It would have stood or fallen with the LDs, leaving no cap on general damages for delay ([101]).

Eco World went the other way on its own words. The clause there would operate as a limitation of liability provision even if the LDs were void or a penalty (as quoted in Buckingham [91]). Both judges also said, obiter, that a void or penal LD clause is probably wholly unenforceable (Eco World [110]; Buckingham [95]). So a contractor that attacks the LD clause may be swapping a capped, certain liability for an uncapped one the employer must prove. Whether that is a good trade depends on how the cap is worded.

Lateness alone rarely justifies ending the contract. The LD clause already compensates for it, and a contractor still working towards completion is hard to accuse of abandoning the bargain. So the contract usually marks the point where delay becomes serious enough. JCT DB 2016 lets the employer give notice if the contractor fails to proceed regularly and diligently (cl 8.4.1.2). If the default continues for 14 days, the employer may terminate (cl 8.4.2). The Court of Appeal in West Faulkner v Newham (1994) described the duty as one to proceed continuously, industriously and efficiently, with proper resources (as quoted in SABIC v Punj Lloyd [18]).

The other forms set their own thresholds. Under FIDIC 2017, once the delay would take Delay Damages past the maximum, the employer may give notice of its intention to terminate (Sub-Clause 15.2.1(c)). It may terminate by a second notice if the contractor has not put matters right within 14 days (15.2.2). NEC4 has no termination reason aimed at the contractor's lateness as such. The Client usually has to bring it within R11, a substantial failure to comply with obligations, not put right within four weeks of the Project Manager's notice (cl 91.2). That is subject to any right to terminate at common law.

Outside the contract's machinery the employer needs more. In Shawton v DGP (Court of Appeal, 2005) the contractor's obligation was to finish within a reasonable time. That obligation was not a condition. The employer could terminate only after a notice making time of the essence, or if the delay deprived it of substantially the whole benefit of the contract ([32]). Where the contractor is still trying to perform, that is intrinsically difficult to show. It can be shown. In Energy Works v MW High Tech (TCC, 2022) a gasifier plant in Hull was almost 11 months late when the employer terminated ([3]). The judge upheld the termination under the contract. In the alternative, he found that the delay had gone to the root of the contract, all the more when coupled with a refusal to carry on commissioning ([301]-[302]).

What happens to LDs when the contract ends before completion? The cases had given three answers: no LDs at all, LDs up to termination, or LDs until another contractor finished the job (Triple Point [75]). The Supreme Court chose the middle one. Unless the clause clearly says otherwise, it applies to delay up to, but not beyond, the date of termination ([86]).

The reasons come from ordinary contract law. LDs accrue day by day for delay that has already happened. Termination ends the parties' future obligations; it does not take away rights already earned. Lady Arden thought it unrealistic to read the clause as making the employer's accrued right simply disappear ([36]). Lord Leggatt saw no reason, in law or in justice, for that result ([79]). The LDs stop at termination because from then on the time taken to finish is entirely outside the control of the original contractor ([85]). A daily rate for the contractor's delay no longer measures anything the contractor does. After termination the employer claims damages for breach under the general law ([35]).

The case itself was about software, not a building. Triple Point contracted to supply a software system and related services to PTT for a Phase 1 price of US$6,920,000. There was no physical construction work ([8]-[9]). It finished two stages 149 days late and none of the rest before the contract was terminated ([88]). On the Court of Appeal's reading, LDs were due on the completed stages only: US$154,662. On the Supreme Court's reading they ran on all the unfinished work up to termination: 3,220 days and US$3,459,278.40 ([88]). The rule is general. Lord Leggatt described LD clauses as a standard feature of major construction and engineering contracts ([74]).

Energy Works applied it to a construction contract. LDs were payable although Take Over was never achieved. The entitlement had accrued from time to time ([314]). They came to £23,077,331.70 ([315]). The contract capped Delay Damages at 15% of the price, and the cap had been reached 56 days before termination ([37]-[38]). That cap did not limit what came next. Losses after termination were not claims for LDs ([320]). Subject to the contract's terms, they were general damages under the ordinary rules. The employer had to prove them, and the contractor had to prove any failure to mitigate ([317]-[318]). Chapter 13 covers how such damages are measured.

The forms now say this expressly. NEC4 Option X7.1 was amended in October 2020 to stop delay damages when the Project Manager issues a termination certificate; the earlier text said nothing about termination. FIDIC 2017 charges Delay Damages up to the date of termination where the works have not been taken over (Sub-Clause 15.4(c)).

LDs need a date to run from. Suppose the employer delays the work and the contract gives no way to extend time for that delay. The employer then cannot hold the contractor to the original date. Time is at large: the fixed date is replaced by an obligation to complete within a reasonable time (Multiplex, TCC, 2007, [48]). With no date, the LD clause has nothing to run from. This is the prevention principle at work: a party cannot insist on a deadline it has itself stopped the other side meeting.

The classic case is Peak v McKinney (Court of Appeal, 1970). Defective piling had to be put right, and the remedial work took just 6 weeks, yet the job finished 58 weeks late (North Midland [13]). Part of the delay was the employer's own fault, and the extension clause did not cover it. Time was at large and no LDs could be levied. Salmon LJ said that had the clause allowed an extension for the employer's fault, the date would simply have moved and LDs would have run from the extended date (as quoted in North Midland [33]). Modern forms therefore extend time for the employer's "impediment, prevention or default" (JCT DB 2016 cl 2.26.6). Where an extension clause is ambiguous, courts read it so that the contractor gets its extension. Otherwise the date would fall away and the employer would lose its LDs altogether, so the clause protects both sides (Multiplex [49], [57]).

The contractor does not walk free when it happens. Salmon LJ said the employer is left to his ordinary remedy: the damages it can prove flow from the contractor's breach. The breach is now a failure to finish within a reasonable time, and the employer must prove it. In Shawton the employer failed. It had to establish what a reasonable time would have been, leaving out delays the contractor caused, and it had not done so ([72], upholding the trial judge at [33]). A reasonable time is judged when the question arises, in the light of all the circumstances ([69]). The original contract periods remain a relevant factor ([69]). But an employer that stops insisting on dates can lose them as a benchmark altogether ([71]). The burden sits with the employer because it alleges the breach, and the contractor answers only for delay it caused.

Contractors sometimes treat time at large as a victory. It may not be one. The contractor swaps a capped, fixed liability for an uncapped one that the employer must prove. Whether that leaves the contractor better off turns on two things. The first is whether the LD cap still limits general damages. The second is whether the employer can prove a reasonable time and its loss. Shawton shows the second can be hard. The first is open.

The same reasoning settles LDs where delay has two causes. Where an employer-risk event and the contractor's own delay are truly concurrent, English law gives the contractor its extension (De Beers v Atos [177]). The completion date moves, so no LDs run for that period. The contractor gets no money for it either. Time and money answer different questions. The extension protects the contractor from paying LDs for delay the employer helped cause. Money needs proof that the employer's event caused the loss, and that fails where the contractor's own delay would have caused it anyway. Chapter 8 has the detail.

A contract can reverse the time half. In North Midland the Court of Appeal upheld a concurrency clause that took concurrent delay out of account. Its effect was to permit the employer to levy liquidated damages for periods of concurrent delay ([45]). Two further traps sit at the edge of this chapter. Where only delay to the completion date counts, an employer delay that uses up the float can leave a later contractor delay critical. The contractor then pays LDs it would otherwise have escaped (SCL Protocol 8.3). And an employer that instructs a variation after the completion date, during the contractor's culpable delay, may lose its LDs if the contractor then accelerates (SCL Protocol 10.13; chapter 10).

The extension that moves the completion date is worked out under chapter 3. The employer's general damages, and the contractor's own money claims for delay, are in chapter 13.