Question 5 of 6: Construction Delays — Classification and Analysis

Question 5: Construction Delays — Classification and Analysis (20 marks)

Question text not reproduced: the examination questions are © Engineers and Geoscientists BC. Open the official past paper (linked at the top of this page) to read the question, then follow the worked solution below.

Delay analysis rests on two separate questions that are easily conflated. The first is legal and contractual: who bore the risk of the event that caused the delay? The second is factual and forensic: did that event actually push out the completion date, and by how much? The four pairs in this question map onto those two enquiries — the first three classify the entitlement, the fourth chooses the method for measuring it.

Excusable versus non-excusable delays. An excusable delay is one caused by an event outside the contractor’s control and not attributable to its fault or negligence, and it entitles the contractor to an extension of the contract time. The usual examples are owner-directed changes, late or defective owner-supplied information or materials, differing site conditions, abnormally adverse weather beyond the norms recorded for the locality, strikes, and acts of a public authority. A non-excusable delay is one the contractor caused or assumed — under-manning, late procurement, defective work requiring rework, poor coordination of its own subcontractors, or an unrealistic as-planned schedule. The practical consequence is stark: an excusable delay stops the clock and relieves the contractor of liquidated damages for that period, while a non-excusable one leaves the contractor exposed to those damages and obliged to accelerate at its own cost. CCDC 2 GC 6.5 frames this in the familiar Canadian form — an extension of time “as the Consultant may decide in consultation” — and conditions it on written notice within a defined period, which is where many otherwise good claims are lost.

Compensable versus non-compensable delays. This is a further division of the excusable category, and it asks whether the contractor gets money as well as time. A compensable delay is one for which the owner is contractually responsible — a change order, an owner’s failure to give timely access, suspension of the work for the owner’s convenience, or errors in the owner-furnished design. Because the risk lay with the owner, the contractor recovers its extended field overhead, unabsorbed head office overhead, escalation and idle-equipment costs in addition to the time extension. A non-compensable but still excusable delay is one for which neither party is responsible — the classic case is abnormal weather, along with epidemics, general strikes and force-majeure events. Here the risk is shared by design: the contractor gets time, so it escapes liquidated damages, but bears its own extended costs, and the owner bears the cost of receiving the facility late. A no-damage-for-delay clause attempts to push even owner-caused delay into this category; Canadian courts read such clauses narrowly and will not generally allow them to excuse an owner’s own breach or bad faith.

Concurrent versus non-concurrent delays. Concurrency arises when two or more independent delaying events, for which different parties are responsible, affect the critical path over the same period. It is the single most contested issue in delay litigation because it decides whether the contractor loses its compensation, the owner loses its liquidated damages, or both. The prevailing approach — reflected in the Society of Construction Law Delay and Disruption Protocol and followed in Canadian practice — is that true concurrency of an owner-caused and a contractor-caused critical delay yields an extension of time but no compensation: the contractor is relieved of liquidated damages, because the owner cannot recover for a period it would have lost anyway, but recovers no prolongation cost, because it would have incurred that cost in any event. Non-concurrent delay is the simpler case in which one cause alone drives the critical path in the period under review, and the entitlement follows the classification of that single cause. Two refinements matter in practice: concurrency must be assessed on the critical path, since a delay to a float-rich activity delays nothing, and it must be assessed over the same window, since sequential delays by different parties are not concurrent merely because both occurred in the same month.

But-for versus windows analysis. These are the two dominant families of forensic method. A but-for analysis — also called collapsed as-built or impacted as-planned depending on its direction — takes a single schedule and asks a counterfactual question: what would the completion date have been but for the events in issue? In the collapsed as-built form the analyst begins with the as-built programme and extracts the owner-caused delay events one by one, and the amount by which the schedule contracts is the claimed entitlement. It is inexpensive, intuitive and well suited to a small number of discrete events, but it is static: it assumes the baseline logic held throughout, ignores changes in the critical path, and can be manipulated by the choice of which events to extract and in what order.

A windows analysis, by contrast, is dynamic and contemporaneous. The project is divided into consecutive periods — typically the monthly update cycle, or windows bounded by significant events — and within each window the analyst identifies the then-current critical path, measures the slippage actually recorded, and attributes it to causes occurring in that window. The updated programme at the end of each window becomes the baseline for the next. Because it tracks the critical path as it migrates and uses the schedules the parties actually relied on at the time, a windows analysis is the method most likely to be accepted by a court, an arbitrator or a referee, and it is the only one that handles concurrency and pacing convincingly. Its costs are real: it requires complete, reliable and contemporaneously updated schedules, and it is time-consuming and expensive to prepare and to rebut. The practical rule is to match method to evidence — a windows analysis where good monthly updates exist and the delay pattern is complex, a but-for analysis where the events are few, discrete and the schedule record is thin, and in either case a narrative explaining why the chosen method fits the facts.