Evidence from the record · What it means for leaders
Carillion
The warning that never left the room.
The one failure we can actually read
Most business failure stories are reconstructed afterwards, from interviews and memory. This one isn't.
Because Carillion went into compulsory liquidation rather than administration, its records passed to the Official Receiver. Two parliamentary committees obtained the board minutes and board papers, and published them.
So we are not guessing at what the board knew. We can read what it was told, and when. That makes Carillion the best-documented board failure in modern British business.
It became better documented still in 2026, when the financial regulator concluded its own action against the company and three of its former directors.
What happened
The sequence matters more than the numbers.
The signal is created
Inside the UK construction division, managers track what they call “hard risks” — money sitting inside their own forecasts that they privately believe will never be collected. It is written down. It goes up to the finance director and the chief executive.
The accounts are signed
The 2016 results are published and audited without qualification. A record dividend of £79 million is declared — the sixteenth consecutive annual increase.
Someone says it out loud
A new finance director for UK construction, six weeks into the job, raises concerns about the accounts. The board minutes describe her as a whistle-blower who did not feel she had been listened to. A contract review is commissioned. The board’s initial intention to give that review an independent element is later reconsidered.
£55 million goes out of the door
The bulk of the dividend is paid to shareholders, four weeks before the write-down is announced.
The truth arrives at full size
The board agrees an £845 million contract provision. The chief executive steps down. By September the figure reaches £1,045 million.
Liquidation
The largest corporate failure in modern British history.
The regulator concludes
The FCA publicly censures Carillion and fines three former executive directors, having found that the 2016 risk reports were never passed to the board or the audit committee.
The information existed, in writing, inside the company, nearly a year before shareholders were told.
Nothing new was discovered in July 2017. What changed was that it could no longer be held.
How much was hiding in how little
The scale of the July write-down suggested a company-wide collapse in trading. It wasn't.
July 2017
A large share of the UK provision came down to a handful of individual contracts — the kind of concentrated exposure that is entirely knowable, if anyone is looking.
“A story of recklessness, hubris and greed”
Joint report of the Business, Energy and Industrial Strategy and Work and Pensions Committees, House of Commons, May 2018.
A verdict few boards recognise themselves in
It is a fair verdict. It is also close to useless as a lesson, because no board reading it recognises itself.
“Recklessness, hubris and greed” allows other boards to place Carillion into a different category: a story about unusually bad people behaving unusually badly.
The more uncomfortable reading is that the failure was produced through decisions, incentives and deferrals that could each appear rational at the time.
External assurance
The problem was never detection
The people closest to the work spotted the problem, quantified it, named it and sent it up the line roughly a year before it surfaced.
The organisation's eyes worked perfectly.
The signal died in the middle. It reached the executive layer and stopped there.
Detection worked. Transmission didn't.
Everything the organisation lost was lost in a gap of one layer.
Holding it cost nothing that month
It stopped for a reason that was entirely rational in the moment.
The 2016 numbers were carrying a great deal of weight: the dividend, the executive bonuses, the banking position and the government contract pipeline all rested on them.
Passing the bad news upward would have brought all four down at once. Holding it cost nothing that month.
This was not a single act of deception. It was a series of individually survivable deferrals, each one making the next admission more expensive than the last.
By July 2017, the truth had become unaffordable — which is precisely why it arrived all at once, and at full size.
And the board is not let off by this
It would be comfortable to conclude that the directors were deceived and the board was blameless. The evidence doesn't allow it.
The board approved a record payout on numbers it had not interrogated.
It watered down the one review that might have given it an independent answer.
It had spent sixteen years being told good news, and had built no habit of asking for anything else.
A board that never demands uncomfortable information will be supplied with comfortable information.
And it will not notice the exchange being made.
Being uninformed was not something that happened to this board. It was a state it had made comfortable.
The tension is the finding
The conventional account says the board ignored the warnings.
The regulator found that the board never received them.
Both are supported by the evidence.
That contradiction is not a flaw in the record. It is the most useful thing in it.
An organisation can fail to inform its leadership and fail to have leadership that insists on being informed — and the second failure is what makes the first one possible.
Dependability performed, not held
Sixteen consecutive years of rising dividends functioned as proof of stability.
Underneath, cash was being generated by stretching suppliers to 120-day payment terms and then charging them for the privilege of being paid sooner.
There was also no point at which anything stopped by rule rather than by argument.
Loss-making contracts ran on because ending one required a conversation nobody wanted to have.
What was missing
The Shortboard model describes twelve attributes an organisation builds in three cumulative waves.
Wave One is Lean Dependability — the ability to do what you say you will do, honestly and without drama.
Wave Two enables new growth. Wave Three creates perpetual relevance.
The waves are cumulative. Wave Two capabilities built on a hollow Wave One do not hold, because there is nothing underneath them to take the weight.
Each wave contains four attributes which work together to create the capability required at that stage of the model.
Read Carillion against the twelve attributes and something striking emerges.
Every capability it lacked sits in Wave One.
This was a business with Wave Two ambitions — acquisitions, new territories and new markets — built on a foundation it had never actually secured.
Wave One
Lean Dependability
Absent at Carillion
conflict Surfacing and working through disagreement so that difficult truths can influence decisions.
Wave Two
New Growth
Not reached
Wave Three
Perpetual Relevance
Not reached
Carillion mistook mass for stability.
Five billion in revenue. Tens of thousands of staff. Sixteen straight years of rising dividends. Every visible measure said solid.
None of them measured whether the thing could still respond.
The four Wave One gaps
Each of these gaps is visible in the published record. None is inferred simply from the outcome.
Carillion did not have one isolated weakness. The evidence reveals a failure across the whole foundation of Lean Dependability.
When a senior finance leader raised a serious concern, the organisation treated her as a problem to be handled rather than a signal to be examined.
The one review that might have produced an independent answer was softened. Nobody was rewarded for saying the difficult thing, and everybody could see that.
The people who understood the true position had no route to act on it and no ownership of the outcome.
They reported the risk and it left their hands. Reporting a problem is not the same thing as being able to do something about it.
Carillion kept bidding into thin markets, kept acquiring, and never exited anything.
Loss-making contracts ran on for years because ending one required an argument nobody wanted to have. There were no pre-agreed points at which something stops by rule rather than by debate — and so nothing ever did.
Its cash came from stretching suppliers and charging them to be paid sooner.
That is not resourcefulness. It is borrowing dressed as operating strength, and it bought time at the direct cost of the supply base the business ran on.
Not one gap. The whole foundation.
This is the failure mode the cumulative logic predicts.
An organisation can grow for years on Wave Two ambition while its Wave One quietly rots — and the rot is invisible precisely because the growth is real.
Where the capacity to see reality is actually set
Most leaders worry about whether their business is picking up the right signals.
Carillion suggests that is the wrong worry. The signals are usually there, and somebody usually has them.
What determines whether they arrive is whether the people in the middle can afford to pass them on.
That is a question about incentives and conflict tolerance. It is not a question about information systems, and it will not be solved by buying one.
An organisation's ability to see reality is set at the point where information is passed on, not the point where it is gathered.
Three questions for your own team
Do not answer these in principle. Identify the most recent, specific example in your own organisation.
What is the last piece of bad news that reached us late, and who knew it first?
Which of our current numbers would be expensive to correct — and who else knows that?
When we last commissioned a review of something uncomfortable, did we make it independent? If not, why not?
Ask what it costs to tell you.
Every organisation has a price attached to delivering bad news upward — in credibility, in relationships, in someone's number.
Most leadership teams have never calculated theirs.
The teams that survive contact with reality are the ones that have deliberately made it cheap.
How we know this, and what we don't
These studies are only worth reading if the evidence behind them is stated plainly, including where it is weak.
Carillion gives us an unusually detailed record of how the failure developed. It does not give us a perfect or neutral account.
Contemporary internal evidence
Carillion's own board minutes and board papers were obtained and published by Parliament.
These are contemporaneous documents, written before anyone knew the final outcome.
The evidence also includes a joint parliamentary inquiry with powers to compel evidence, and regulatory findings concluded in 2026 after examining the internal reporting directly.
The record is not neutral
Everything here was gathered by bodies looking for fault after a known outcome. Hindsight shapes every document, including this one.
The regulatory findings were settled rather than tested in court. All three directors withdrew their appeals rather than contest them.
Minutes record what was written down, not everything that was said.
There is no counterfactual.
We cannot know what a board that asked harder questions in October 2016 would have found in time to matter.
Treat this as evidence of how the failure worked — not as proof of what would have prevented it.
Business, Energy and Industrial Strategy and Work and Pensions Committees, Carillion, House of Commons, May 2018, together with the board minutes and papers published alongside it.
Financial Conduct Authority Final Notices concerning Carillion plc and its former directors, January and February 2026.
The Pensions Regulator, intervention report.