Boards do not buy optimism. They buy a credible forecast, an honest
risk register and early warning when something moves.
That is not a statement of values. It is a description of what the
audit evidence rewards, and of machinery your contract may already
oblige you to operate.
The forecast
that looked worse was the better one
In January 2026 the National Audit Office reported on the New
Hospital Programme. The Department of Health and Social Care now expects
the 46 schemes to cost £60bn, of which £56bn is capital, completing in
2045-46. That is £33.8bn more capital than the plan proposed in 2023.
Torbay, Kettering and Musgrove Park hospitals will, on the NAO’s
estimate, open “nine to ten years later than planned when we reported in
2023”.
Every number moved the wrong way. The NAO’s verdict on it: the
January 2025 plan “appears to be considerably more realistic in terms of
the amount of capital funding allocated and the delivery timetable”.
Note what earned that assessment. The programme carries £12.4bn of
contingency, 21% of the total, and the report records plainly that “DHSC
expects this contingency to be needed”. It rates the risk of its own
staff vacancies delaying delivery as red. A programme that says out loud
that it expects to spend its contingency, and names its own capability
gap as a red risk, is not in trouble for saying so. It has told you
where it actually is.
The reforecast was worse and better at the same time. Those are not
in tension. They are the same fact.
Your
board pack probably reports achievement, not residual risk
The NAO’s 2020 lessons paper found, on Crossrail, that “the emphasis
on progress reports presented to the board and sponsors was on what had
been achieved, rather than on the level of risk to successful delivery
that remained in the programme”.
Read that twice, because it is not a criticism of dishonest
reporting. Everything in those reports was true. Milestones were
achieved and the achievements were real. The defect was in what the
report was about. It answered “what have we done” when the
board’s only useful question was “what could still stop us”.
If your last board report opened with milestones achieved and
put risk in an appendix, you have written the Crossrail board
pack. Most of us have. It is the natural output of asking a
delivery team to report on delivery, and it survives precisely because
nobody in the room can point to a sentence in it that is false.
The NAO is direct about the mechanism. Pressure on sponsors and
delivery bodies “can make them defensive about the programme or allow a
‘good news’ culture to develop which can undermine processes intended to
transmit accurate information”, with the result that “opportunities to
identify and mitigate serious issues may be missed, and instead emerge
suddenly and unexpectedly”.
The learning point is a question worth putting in your own governance
papers. “Decision-makers should consider whether the indicators they are
given are the right ones, and how these would alert them to emerging
issues.”
Early
warning is clause 15.1, and much of the internet will tell you
otherwise
Under NEC4, the early warning obligation sits at clause 15.1 of the
Engineering and Construction Contract, the Term Service Contract and the
Professional Service Contract, and NEC is explicit that “only the
contractor and project or service manager can notify an early warning”.
The Project Manager enters each one on the Early Warning
Register, and early warning meetings run under 15.3.
That matters because the previous generation still tops most search
results. NEC’s own May 2017 article, written before the NEC4 suite
landed, refers to “Clause 16 of most NEC contracts” and to entering
matters on the “risk register”. Both were right then. Both are wrong for
an NEC4 contract now. If a paper on your desk cites clause 16 and a risk
register against an NEC4 form, the author is working from the wrong
edition, and that is worth knowing before you rely on the rest of their
advice.
Two properties of the register are easy to miss. It is deliberately
thin: NEC’s own training material notes it “has at least 2 columns, you
may add more if you wish to”, and that “You are recording agreed actions
to be taken, it is not the action itself”. A register is a record of a
decision, not evidence that anything was done. Boards routinely read the
first as the second.
The
contract prices your silence, and not symmetrically
Fail to give an early warning an experienced contractor could have
given, and the compensation event is not barred. It is “assessed as if
an early warning had been given (61.5/63.7)”. The contractor recovers
what it would have recovered had the team been given the chance to
mitigate, and no more. Both clauses are engaged, which is worth saying
because most commentary cites only 63.7.
Now the part that concerns you more than the contractor. NEC’s
material records the Project Manager’s position in a single line: on
Project Manager failure, “Client likely suffers in terms of additional
time/money.” The obligation to warn is mutual. The sanction is not. When
your side stays quiet, there is no clause that trims the consequence
back. You simply pay for it.
And early warning is not a claim. NEC’s own position is that “there
is no direct link between early warning notices and compensation
events…”. The machine that actually extinguishes money is elsewhere,
at clause 61.3, where a contractor has eight weeks from becoming aware
that an event has happened to notify a compensation event, or seven for
a subcontractor under the subcontract. Warning early protects the
project. Notifying in time protects the entitlement. Confusing the two
loses one of them.
What this evidence does not
cover
Be careful how far you carry the audit material. NISTA’s portfolio at
31 March 2026 held 189 projects worth £924.2bn, 15% rated Green and 18%
Red. That is UK central government. The New Hospital Programme is DHSC,
and England. The NAO’s lessons are drawn from government major
programmes, not from a £20m university refurbishment or a council
leisure centre. Health, education and local government capital are
devolved, and nothing above establishes a Scottish, Welsh or Northern
Irish position.
The part that reaches your estate directly is the contract. Clause
15.1 does not care about the size of your programme.
One closing irony. Government’s own answer to late bad news is a
leading indicator: NISTA “launched a new digital AI tool this year
called the Early Warning System, which uses existing GMPP data to flag
projects at risk of transitioning to a Red rating”. That is the NAO’s
2020 recommendation, built. It is also an admission that the reports
arriving by the normal route were not doing the job.
A red rating, as NISTA puts it, “does not mean a project will fail.
In many cases it simply reflects the scale, complexity or stage of
development”. The board that can hear that is the board that gets told
things early.
