When should a major project stop?

The foundations are poured, contracts are signed and a delivery team has spent years turning an approved proposal into a working site. Then the next forecast arrives. Completion will take longer and cost more. The benefits case was written for a different economy, or a different government priority. Nobody wants to suggest stopping. Too much has already been spent.

That last sentence is a reason to ask harder questions, not an answer to them. Money already spent cannot be recovered by spending more. The decision facing a sponsor is whether the next dollar, and the risks attached to it, are justified by what the project can now deliver.

Stopping may be responsible. So may continuing, but with a different design, delivery strategy or objective. The difficult part is distinguishing a project that can be recovered from one whose original rationale has expired.

Completion is not the same as success

Once construction begins, finishing can become an objective in its own right. A dashboard reports physical progress, a contractor meets milestones and the next public commitment is framed around opening day. Yet a project can be delivered to its latest programme and still fail to produce the public or commercial value for which it was authorised.

Sponsors need to separate three questions. Can the asset be finished? Will it work as an integrated system? And, if it works, is it still worth the remaining cost and risk? A yes to the first question cannot settle the other two.

The business case is a live decision document. The Association for Project Management describes the sponsor as accountable for its continuing validity throughout the project life cycle. That responsibility matters most when the assumptions that justified approval begin to diverge from reality.

Look for a pattern, not a single bad month

Research by Jose Rodrigo Juarez Cornelio, Tristano Sainati and Giorgio Locatelli examined 30 infrastructure megaprojects that were terminated during delivery. They identified six categories of determinants: socio-political, environmental, financial distress, regulatory, force majeure and technological. Termination generally emerged gradually from a combination of factors. Their project health model and checklist are aimed particularly at due diligence before construction, but their account of how warning signs accumulate offers a useful prompt for reviews during delivery. The sample was weighted towards nuclear power projects and relied on publicly available secondary sources; it cannot predict the fate of every troubled scheme.

For a sponsor, a health review should bring several kinds of evidence together. Is the forecast to complete becoming less credible with each revision? Have the expected benefits changed, and who will now realise them? Do the parties still have a workable route through design, approvals and interfaces? Has the political, regulatory or market context shifted enough to change the value of the outcome? Are concerns reaching the people who can act on them?

Each warning may have a sensible explanation. A combined pattern is harder to dismiss. A review should identify the causes, the consequences if they remain unresolved and the decision date beyond which options will narrow. It should also record disagreement. A reassuring average can conceal the one unresolved dependency on which the whole scheme rests.

Account for the cost of stopping

Melbourne’s East West Link shows why an exit has to be assessed honestly. Contracts had been signed before the 2014 state election; the incoming Victorian government suspended work and agreed to terminate the project in June 2015. The Victorian Auditor-General estimated more than $1.1 billion had been paid or was expected to be paid by the state for little tangible benefit, partially offset by proceeds from selling acquired properties. The audit also found that the original business case had not provided a sound basis for committing to the investment. Termination had a substantial cost. That fact alone cannot establish that completing the project would have been better.

HM Treasury’s 2026 Green Book says sunk costs should not enter an appraisal of choices still to be made. It also asks practitioners to consider whether resources already acquired could be sold or put to another use. That does not erase the public significance of past expenditure. It means yesterday’s expenditure cannot, by itself, justify tomorrow’s.

As the earlier article on rising project costs explains, the board needs a genuine choice among continuation, redesign, suspension and exit. An exit paper should include contractual costs, safety obligations, community impacts and the value of work that can be reused. The original approval asked whether to start; the present decision concerns the best use of the resources still available.

Ask a simple question at every review: If we had the present forecast and knew what we know now, would we authorise the remaining investment? If the answer is no, identify what would need to change before more funds are committed.

Make rescue conditional

A troubled project may still be worth rescuing. But “we can turn it around” is a claim that needs a mechanism. What scope will change? Which interface will be resolved? Who can make the decision, by when, and what evidence will show that the intervention worked?

One practical approach is to set a bounded recovery period. Agree on the critical uncertainties, the actions that might resolve them, the maximum cost and time allowed for that work, and the tests for continuation. The sponsor should know in advance what happens if those tests fail. An independent review can help where the same team that prepared the optimistic forecast is being asked to judge its own recovery plan.

This is not a demand for certainty before proceeding. Major projects contain uncertainty by their nature. It is a demand to distinguish a manageable risk from an assumption that has become a substitute for evidence. A recovery plan should improve the decision, not merely move its date.

Design an orderly stop

Termination is itself a major project decision. Contracts must be administered, works made safe, staff and suppliers treated fairly, records preserved and affected communities told what happens next. Where parts of an asset have enduring value, the sponsor should decide how they will be used or maintained. Where the intended outcome remains important, a different intervention may be required.

First identify who has legal and political authority to stop the project. In three of the four termination patterns in the megaproject study, government played a substantial part in the decision. A sponsor who lacks final authority should put a clear recommendation and options before those who hold it. The decision record should explain why continuing was no longer justified, what alternatives were considered, what is being done to limit further loss, and who owns the underlying need after the construction team leaves. That record is useful even if the decision is to continue: it makes the conditions of approval and the consequences of failing them visible.

Stopping a project need not mean giving up on the problem it was meant to solve. Conversely, completing an asset does not prove that the problem was solved. The sponsor’s obligation is to keep asking which course of action, from this point forward, offers the best defensible chance of delivering the intended outcome.

Further reading

• Juarez Cornelio, Sainati and Locatelli (2023), “Digging in the megaproject’s graveyard: Why do megaprojects die, and how to check their health?”, International Journal of Project Management.
• HM Treasury (2026), The Green Book, particularly its treatment of sunk costs and options appraisal.
• Association for Project Management, “What is project sponsorship?”
• Victorian Auditor-General’s Office (2015), East West Link Project.

Related on Projects Asia-Pacific: Five questions before approving a major project.