Question 5 of 6: Construction Delays — classification and forensic analysis

Question 5: Construction Delays — classification and forensic 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.

Excusable versus non-excusable. The first question a delay claim must answer is whether the contractor is entitled to more time. A delay is excusable when it arises from an event beyond the contractor’s control and without its fault or negligence: late issue of drawings, late site access, owner-directed changes, differing subsurface conditions, unusually severe weather beyond the contractual norm, strikes outside the contractor’s own workforce, or force majeure. An excusable delay entitles the contractor to an extension of the contract time and, critically, relieves it of liquidated or general damages for that period. A delay is non-excusable when it is the contractor’s own risk: late mobilisation, under-manning, poor sequencing, rework of defective work, failure of a subcontractor or supplier that the contractor selected, or equipment breakdown. No extension is granted, the completion date stands, and the owner may assess damages. Two practical points decide most disputes before the classification does: the delay must be shown to have affected the critical path, since delay to an activity with float consumes float rather than time, and the contract’s notice provision must have been complied with — CCDC 2 GC 6.5.4 requires written notice within a stated period, and Canadian courts have repeatedly enforced such clauses.

Compensable versus non-compensable. This second classification applies only within the excusable category and decides whether the contractor also gets money. A delay is compensable when it is caused by the owner or by someone for whom the owner is responsible — the consultant, a separate contractor, an owner-supplied equipment vendor — and the typical heads of claim are extended site general conditions and supervision, extended equipment and facility rental, labour and material escalation, loss of productivity, and head-office overhead unabsorbed during the extension. A delay is excusable but non-compensable when neither party is at fault: abnormal weather, a general strike, an epidemic, an act of a public authority. Here the risk is shared, the contractor gets time but bears its own prolongation cost, and the owner bears its own late occupancy. CCDC 2 GC 6.5 encodes exactly this split, distinguishing owner- and consultant-caused delay, which carries reasonable costs, from delay caused by labour disputes, fire, abnormally adverse weather and other neutral events, which carries an extension only. It follows that every delay must be classified twice, and that the three outcomes in practice are time and money, time only, or neither.

Concurrent versus non-concurrent. Delays are concurrent when two or more independent delay events, attributable to different parties, affect the critical path during the same period, so that removing either one alone would not have advanced completion. The usual and most consequential case is an owner-caused compensable delay running at the same time as a contractor-caused non-excusable delay. The mainstream rule in Canadian and Commonwealth practice is that true concurrency converts a compensable delay into a time-only delay: the contractor receives its extension, and therefore escapes liquidated damages, but recovers no prolongation cost, because it would have incurred that cost anyway on account of its own delay. Non-concurrent delays are sequential or affect different paths, and responsibility can be apportioned cleanly between them. Two distinctions matter in practice. Literal concurrency — two delays genuinely overlapping in time — is rarer than functional concurrency, where the delays occur in different periods within the same analysis window and both bear on the same completion date. And a pacing delay, where the contractor deliberately slows non-critical work because an owner delay has already made haste pointless, is a reasoned reaction and not a true concurrent delay, provided the contractor can show a contemporaneous decision to pace and the ability to resume.

But-for versus windows analysis. These are two families of forensic method for proving which delays actually drove the completion date. A but-for analysis, usually implemented as the collapsed-as-built method, starts from the as-built schedule, extracts the delay events attributable to one party, and recalculates what the completion date would have been but for those events; the difference is that party’s share. It is retrospective, single-pass, comparatively cheap, and intuitive to a tribunal. Its weaknesses are serious: it requires an as-built network with defensible logic, which most projects never build; it is a static calculation that cannot show the critical path shifting from one chain to another during the job; and by removing one party’s delays in isolation it tends to conceal concurrency. A windows analysis, also called contemporaneous period or time-impact analysis, divides the project into successive windows — monthly, or bounded by milestones or major events — and at each window boundary updates the schedule with actual progress before determining which activities were critical inside that window and which delays consumed the time. It captures critical-path shift and concurrency period by period, uses the contemporaneous schedule updates that a well-run project already produces, and is generally regarded by practitioners and tribunals as more reliable than a single-pass retrospective calculation, although AACE International Recommended Practice 29R-03 deliberately declines to rank any one method as best for every dispute. Its cost is the obvious drawback: it needs a validated baseline, disciplined monthly updates and considerably more expert time. The practical conclusion is that the choice of method is largely settled during construction rather than during the dispute — a project with a proper baseline and honest monthly updates can run a windows analysis, and one without them is left with a but-for argument and a weaker case.