Late, Then Varied: Setting a Fair Completion Date When the Contractor Is Already Behind

On construction projects everywhere, a familiar sequence repeats itself. The contractor falls behind through its own difficulties: a shortage of labour, a slow subcontractor, a sequence that did not work. The contractual completion date passes, or is about to pass, and the employer is entitled to look to liquidated damages. Then the employer issues an instruction for extra work, and that instruction, on its own, would justify more time. The question that follows is deceptively simple. What should the new completion date be?

Two instincts compete. One is to restart the clock from the instruction, so that the contractor is given whatever time the extra work needs from the day it was ordered. The other is to take the existing completion date and add only the period the extra work itself justifies. The difference between them is the difference between relieving the contractor of its own delay and allowing a fair period for the new work.

This comprehensive blog will explore why extension-of-time clauses exist and how delay is grouped, how the courts resolved the problem of an employer's instruction issued during the contractor's own culpable delay, and why a variation issued in such a period does not always add time at all. We will examine the related problems of early-completion programmes, weather, concurrency and float, and why notice, an accepted programme and event-by-event allocation must come before any new date is fixed. Whether you are a contract administrator, quantity surveyor, contractor, employer or project manager, understanding how to set a fair completion date when the contractor is already behind is key to protecting both the employer's remedy and the contractor's legitimate entitlement.

1. Three Kinds of Delay, and Why the Clause Exists

Extension-of-time clauses are often described as a concession to contractors. Historically, they are the opposite. In Holme v Guppy (1838), a contractor finished late, the employer sought the liquidated damages set in the contract, and the contractor showed that at least part of the delay was the employer's fault. Because the employer had prevented completion, time was held to be "at large": the contractor's obligation became one of finishing within a reasonable time, there was in effect no valid completion date, and the liquidated damages could not be deducted. Employers responded by introducing clauses that allowed the completion date to be extended in the event of their own default, keeping the date, and with it the right to deduct liquidated damages, alive.

The point was reinforced in Peak Construction v McKinney Foundations (1970). The contract contained no provision for extending time for the employer's default, part of the delay was the employer's, and the employer could not recover liquidated damages even for the portion of the delay caused by the contractor. Paradoxically, then, the main purpose of an extension clause is not to relieve contractors of damages but to protect the employer's right to recover them.

Delay in completion falls into three basic groups.

The first is delay caused by the contractor. The contractor's primary obligation is to complete by the due date. If it fails to do so through its own fault, it is in breach and is expected to bear the whole of the consequences, including reimbursing the employer's loss through the agreed damages.

The second is delay caused by the employer, the employer's consultant team or its directly employed contractors. Late possession or access, late drawings or information, extra work that causes delay and late supply of employer-provided materials are common examples. Without a mechanism to extend time for such events, the employer loses the right to liquidated damages, because it cannot insist on a condition that its own conduct has made impossible to fulfil. All the standard forms therefore contain comprehensive provisions for extending time in these cases.

The third is delay for reasons beyond the control of either party, such as exceptionally adverse weather, strikes, government action or force majeure. Whether and how these rank for extension is a function of the risk balance of the particular contract, and the standard forms differ considerably in which neutral events they recognise. Some extend time for adverse weather, insured perils or the actions of statutory authorities; others do not, or leave certain events to the employer's choice or the contract administrator's discretion. The list in the contract actually signed is the only reliable guide.

Force majeure illustrates the point. The term is notoriously hard to define, and any clause using it will be interpreted on its own wording; in JCT forms its meaning is probably narrower, because some events that might fall within it are listed separately. Without such a clause, an event that made performance impossible through no one's fault might frustrate the contract and bring it to an end, which on a large project is the last thing the employer wants. A force majeure clause usually keeps the contract alive instead: the party prevented from performing receives an extension of time, each party bears its own costs, and all other obligations continue.

Two procedural points apply across all three groups. First, assessment is generally made by the contract administrator after the contractor has notified the delay event, and in most cases the administrator must allocate specific extensions to specific events. Second, the administrator cannot judge what extension is fair and reasonable without knowing how the contractor proposed to carry out the work. An accepted works programme is the baseline against which delay is measured. It is also worth remembering that an extension of time does not of itself carry payment. Delay is often described in three classes for this reason: excusable and compensable, where the employer is responsible and the contractor receives both time and loss and expense; excusable but not compensable, where neither party is at fault, the contractor receives time but no money, and losses lie where they fall; and non-excusable, where the contractor receives neither and pays delay damages.

Figure 1: Three groups of delay: contractor-caused delay, for which the contractor bears the consequences; employer-caused delay, which requires extension machinery if liquidated damages are to survive; and neutral events, whose treatment depends on the form. Allocation is event by event, against a programme.

2. The Chestermount Problem: Net or Gross?

The leading illustration of the problem is Balfour Beatty Building Ltd v Chestermount Properties Ltd (1993). The contractor was already in delay through its own fault when the employer issued instructions for extra work, and those instructions themselves gave grounds for an extension of time. The completion date had already passed. The question was how the revised completion date should be calculated.

The employer argued that it should not be made responsible for the contractor's own delay. The correct revised completion date, on its case, was the original completion date plus only the period that the extra work itself justified. This became known as the "net" method.

The contractor argued that the revised completion date should be the date by which the extra work could reasonably be completed, measured from when it was instructed, so that the extension absorbed both the period of its own culpable delay up to the instruction and the period required by the extra work. This became known as the "gross" method. The contractor's point had a certain logic: it seemed illogical to say that the contractor should have finished by a date fixed by reference to an event that had not even happened when that date passed.

The judge acknowledged that fixing a completion date on the basis of an event that had not yet occurred appeared, at first sight, "distinctly peculiar". But he pointed out that the gross approach would remove from the contractor any liability for its own delay up to the time of the employer's instruction. That would be unfair to the employer, which would lose its compensation for the contractor's breach in failing to complete by the existing completion date. Noting that the contract called for extensions to be fair and reasonable, the court found in favour of the employer: the net method was the one to use.

The decision did not stand alone. Earlier authority had already favoured the net method, and the Court of Appeal had dismissed the opposite argument in vivid terms: it would be absurd for an employer to lose its claim for damages against a contractor already badly late through its own fault simply because, at the last moment, it ordered something as small as an extra coat of paint. The common thread is that an extension exists to reflect the effect of the qualifying event, not to wipe out a breach that had already occurred.

The way the question was put to the court is instructive. A gross extension was described as one that refixes the completion date at the calendar date on which the work would reasonably be expected to be completed, having regard to the date on which it was instructed. A net extension takes the date currently fixed and adds the period the architect regards as fair and reasonable. Framed that way, the answer follows naturally from the purpose of the clause.

The case also gave rise to a second proposition that is easy to overlook. It was said there that if the variation works can reasonably be carried out simultaneously with the original works, without interfering with their progress, and are unlikely to prolong practical completion, the architect might properly conclude that no extension of time was justified. A variation issued during a period of delay is therefore not automatically worth any time at all. The prior question is whether it actually affects completion.

Three cautions follow. First, the net method answers the question of where the extension is measured from; it does not tell the administrator how long the extension should be. That depends on the effect of the event, assessed against the programme. Second, the case was decided on the wording of a particular standard form, and in many of the cases in this area the judgment turned on the contract's own words, alongside principles of fairness. Third, the law has continued to develop, and anyone relying on the case for a particular result should read the report and current authority.

Figure 2: Two argued ways to reset the date: the net method adds the period justified by the extra work to the existing completion date; the gross method runs the extension from the instruction, absorbing the contractor's own earlier delay. The court favoured the net method.

3. Related Problems: Early Programmes, Weather, Concurrency and Float

The Chestermount problem sits among a cluster of recurring questions. Each has generated litigation, and each is answered, in the end, by the words of the contract and the facts of the delay rather than by a universal formula.

Programmes showing early completion. In Glenlion Construction Ltd v The Guinness Trust (1987), the contractor submitted a programme showing completion well ahead of the contractual completion date, was then prevented from achieving that earlier date because information was not provided in time, and claimed an extension. The contractor was entitled to finish "on or before" the completion date, but the architect's duty was to provide information so that the contractor could complete by the contractual completion date. There was no implied duty on the employer to produce information early, and the court refused to imply a term that would have imposed an obligation on the employer but none on the contractor. Where the contract includes an information release schedule, its dates must of course be met.

Weather. In Walter Lawrence & Sons Ltd v Commercial Union Properties Ltd (1984), the architect assessed exceptionally inclement weather against the time the work had been programmed to be done, reasoning that a contractor should not defer work for its own reasons and then recover weather delay from the employer. The contractor argued that the weather should be assessed against when the work was actually carried out. The contract required regular and diligent progress but did not oblige the contractor to programme the work or to follow a programme strictly, and the case was decided in the contractor's favour.

Concurrency. Concurrent delays, two or more competing causes affecting progress at the same time with approximately equal effect, cause the severest difficulties of causation, and judges and commentators have not been of one mind about them. In Henry Boot Construction (UK) Ltd v Malmaison Hotel (Manchester) Ltd (1999), the parties agreed that where one of two concurrent causes is a relevant event and the other is not, the contractor is entitled to an extension for the delay caused by the relevant event notwithstanding the concurrent effect of the other. The court also held that the architect, in fixing extensions, was entitled to consider other relevant factors impeding progress. For the JCT family, the approach that appears to fit best with the scheme of the forms is that a delay may be attributed to more than one cause, and that if one of them is a relevant event that makes an extension fair and reasonable, it cannot be refused simply because the delay would have happened anyway. The detail of concurrency, however, is a subject of its own and remains contested.

Float. Under the JCT 05 forms, an extension is due only if completion is likely to be delayed beyond the completion date. Employers have therefore argued that float in the programme belongs to them, while contractors have argued that it is their own contingency. The prevailing view was that float is available to whoever uses it first, and analysis of float forms part of the administrator's assessment of the event at the time it occurs. Some forms take a different position: under NEC3, delay is measured against planned completion on the accepted programme, so the contractor effectively owns the float.

All of this deserves restraint. Extension-of-time disputes have consumed inordinate amounts of consultant and court time and produced a complex and, in some cases, pseudo-scientific discipline of delay analysis. The courts have repeatedly preferred analysis that is objective, meticulous on the facts and not hidebound by theory when the facts collide with computer logic. No method is better than the data put into it.

Causation underlies every one of these problems. Before granting an extension, the administrator must be satisfied not only that a relevant event has occurred but that it is likely to delay completion of the works as a whole. In Royal Brompton Hospital v Hammond (2000), an architect granted an extension equal to the period by which access to part of the site had been late, apparently to relieve the contractor of liquidated damages for that period. The court criticised the approach: on the evidence, the late access had caused no delay to the progress of the works, and so no extension was due. Equal periods are not proof of cause and effect.

Figure 3: What must be true before time is added: a qualifying event has been notified; the programme shows completion is likely to be delayed beyond the completion date; and the new work cannot reasonably be absorbed alongside the original work without prolonging completion.

4. Notice, Programme, Event by Event

The procedural steps around an extension are not formalities. The employer's right to deduct liquidated damages has historically been regarded as dependent on the contractual procedures being properly followed, so timely and accurate compliance by the administrator matters. Equally, where a contract makes notice by the contractor a condition of entitlement, failure to give it may cost the contractor the extension. The forms differ in their timetables and in whether late action invalidates an extension, and the contract actually signed must be read for both.

Under the JCT 05 Standard Building Contract, the pattern is iterative. The contractor notifies delay when it becomes reasonably apparent that progress is being delayed, giving the causes. The administrator decides whether the events rank for extension and whether completion is likely to be delayed beyond the completion date, and issues a decision within the contractual period. The process is repeated for each delay, and after practical completion the administrator may review the extensions granted and fix the completion date as is fair and reasonable. Other forms use two-stage notices and detailed claims, or, in the case of NEC3, deal with time and money together as compensation events with their own strict notification rules.

The administrator is expected to handle extension claims in a detailed, analytical way. An impressionistic assessment of what seems reasonable has been criticised by the courts; a logical analysis, made methodically, of the impact of the relevant events on the contractor's planned programme is what is required. That does not always demand a critical path analysis. A single suspension caused by the employer may need no more than the date work stopped, the date it restarted and the reason. Competing causes over time are another matter. Throughout, the burden of proving that an event caused delay lies with the contractor.

The contractor's own conduct also matters. At common law, a delayed contractor is expected to mitigate by doing what may reasonably be required to reduce the delay and disruption, although this does not extend to spending its own money to recover delay it did not cause. The JCT forms add an express duty to use best endeavours to prevent delay, and the authorities have differed on whether that amounts to a duty to mitigate the effect of a delay or only a duty not to compound it. Either way, the contractor's response to an event is part of the evidence the administrator must weigh.

Several practical disciplines follow from these principles. Keep an accepted baseline programme and update it. Keep a log that ties every notice to a specific event. When a qualifying event arrives during culpable delay, record which approach is being used to fix the revised date and why. And record whether the new work can reasonably proceed alongside the original work without prolonging completion. A general complaint that "the job was delayed" is no substitute for allocation.

Consider a contractor on a commercial building that has already overrun its completion date because of its own labour difficulties. The employer then instructs additional work to the entrance area. The contractor writes to say that it requires a new completion date running from the date of the instruction. The contract administrator is minded instead to add to the existing completion date only the period the entrance work itself justifies, and before deciding even that, to ask whether the entrance work can be carried out alongside the finishing work already in hand without delaying completion at all. The right questions are now on the table. Has the contractor notified the event and its effect? What does the programme show about the entrance work's place in the sequence? Can it proceed in parallel? If it cannot, what period does it genuinely require? The answers will depend on the programme and the records, and the administrator's decision, whatever it is, should be written down with its reasoning.

Figure 4: From event to a revised date: notification of a specific event, review against the programme (including whether completion is likely to be delayed and whether parallel working is reasonable), then a recorded decision on the period to add and the method used to fix the date.

5. What the Extension Decision Does Not Decide

Clarity about scope prevents a sound extension decision from being stretched into something it is not.

It does not decide whether a particular instruction is a variation at all. That is a prior question, answered by the contract, the specification and the power under which the instruction was given. The analysis here assumes an event that ranks for time.

It does not decide whether the stated rate of liquidated damages is enforceable. The extension machinery keeps the completion date, and therefore the stated sum, alive; whether that sum is itself open to challenge is a separate matter.

It does not resolve concurrency. The cases offer approaches rather than a settled formula, and the outcome in a given dispute will depend on the form, the facts and current authority.

It does not turn comparisons of how the forms treat neutral events into current advice. The forms have been revised, and their lists of relevant events must be read in the edition actually used.

And it does not award money. An extension of time relieves the contractor of damages for the extended period; payment for loss and expense requires its own entitlement.

With those limits visible, five questions help the administrator reach a defensible date:

1. Which group does this delay belong to? Contractor-caused, employer-caused or neutral, and if neutral, does the contract list it as a relevant event?

2. Was the contractor already in culpable delay when the qualifying event occurred? If so, the revised date should normally be measured from the existing completion date, not from the instruction.

3. What period does the event itself require, and can the work be done without prolonging completion? If the new work can reasonably proceed in parallel, no extension may be justified.

4. Is the method stated, and has current authority been checked? The net approach is the established answer to the Chestermount problem, but the form in use and the current law must still be read.

5. Can the new date be explained event by event against the programme? If it cannot, the decision is not yet ready to be issued.

Conclusion

An extension of time exists to keep a fair completion date alive. It is not a reward for the contractor's own delay, and it is not a device for refusing time that a later qualifying event genuinely requires. When the employer orders extra work while the contractor is already late through its own fault, the established answer is the net method: take the existing completion date and add the period the new work justifies, rather than restarting the clock from the instruction. Before even that, the administrator should ask whether the new work affects completion at all, because work that can reasonably be carried out alongside the original work may justify no extension.

Around that core sit the classification of delay into contractor, employer and neutral causes; the rule that an early-completion programme does not oblige the employer to accelerate its own obligations; the assessment of weather against when work was actually done; the unsettled problem of concurrency; and the question of who may use float. None of these is resolved by a formula or by software alone. Each depends on the contract's words, on the programme and on records made at the time.

The practical discipline is therefore procedural. Require notice of specific events, keep an accepted programme, state the method used to fix any revised date, and record the reasoning event by event.