I want to tell you about a job you did not know existed, being done badly by companies you cannot name, in a way that quietly affects your life.
Stay with me. There is a chatbot scandal at the end and I promise it is worth it.
First, a job you have never thought about
Every big company in the world tells you how much money it made last year.
Here is the obvious problem with that. The company is the one telling you.
If I run a business and I tell you it earned twenty million dollars, why on earth would you believe me? I have every reason to lie. My investors want good news. My bank wants good news. My bonus depends on good news.
So somebody invented a solution, and it is genuinely clever. You hire an outsider. A completely separate firm comes in, opens your books, checks that the numbers are real, and then publicly puts their name to it. They say: we looked, and this company is telling the truth.
That is called an audit. That is the whole job.
It only works because of one thing. The outsider must be independent. If the referee is on the payroll of one team, the referee is not a referee. He is a fan with a whistle.
Remember that. Everything that follows is that sentence being slowly, expensively violated.
How four firms ended up doing all of it
In the 1970s, companies went global. IBM, Ford, Coca-Cola, suddenly operating in forty countries at once, every country with different tax laws and different accounting rules.
That created a real headache. Who can possibly check books spread across four continents?
The answer: a few enormous firms with offices everywhere. They became known as the Big Eight.
By the late 1980s, those eight firms audited about 98% of America’s public companies. Almost every major business on earth was being checked by one of eight organisations.
Then eight became five. Then five became four.
Today they are called the Big Four: Deloitte, PwC, EY, and KPMG. You have walked past their offices. You have never thought about them once. They employ well over a million people between them and they are woven into the financial plumbing of the entire planet.
Of the 500 biggest companies in America, 498 are audited by these four firms.
Four hundred and ninety-eight. Somewhere out there are two beautiful rebels doing their own thing, and I would like to shake their hands.
Why nobody ever fires them
Here is the trap, and it is not a conspiracy. It is just human cowardice, industrialised.
Say you run a big company. Audit time. You have two options.
Option one: a new firm nobody has heard of, possibly excellent, possibly cheaper.
Option two: the giant firm that has checked your books for thirty years, audits all your competitors, and works for half the governments of Europe.
You pick option two. You will always pick option two. Because if option one messes up, it is your fault for hiring them. If option two messes up, well, everybody uses them.
Nobody was ever fired for hiring Deloitte.
That sentence is the entire business model. You are not really buying an audit. You are buying an excuse. It is much easier to say “we hired McKinsey” in a board meeting than “we hired a firm nobody has ever heard of,” and that difference is worth billions of dollars a year.
The result is a market that has quietly sealed itself shut. When a company picks one of these firms, it takes an average of twenty-three years to switch. Canada’s largest telecoms company has used the same auditor for 144 years.
A hundred and forty-four years. Empires have risen and fallen. Zimbabwe has been through the entire colonial project, a liberation war, independence, and several currencies, and in all that time BCE Inc. has not once thought, you know what, let us get a second opinion.
What happens to a company that cannot lose
Now we arrive at the actual lesson, which is not about accounting at all.
When any organisation becomes genuinely impossible to remove, the quality does not collapse in a dramatic way. It leaks out slowly, over years, and nobody notices until something explodes.
In 2001, something exploded.
Enron was one of America’s most admired companies, an energy giant supposedly worth tens of billions. Its auditor was Arthur Andersen, then one of the Big Five. In the year 2000 alone, Andersen earned 25 million dollars auditing Enron, plus another 27 million advising it.
Meanwhile Enron was hiding roughly 20 billion dollars of debt using a web of shell companies designed to make it look far richer than it was.
Catching precisely that was Arthur Andersen’s one job. The single thing they were paid for. They missed all of it.
When it came out, Arthur Andersen was destroyed within nine months. Thousands lost jobs. Enron employees lost their pensions.
Now, you would think the lesson was obvious: this industry is far too concentrated, and we should break it up.
Instead, the Big Five became the Big Four and everyone carried on with their lives.
The response to a market being dangerously small was to make it smaller. I have watched governments respond to a fuel crisis with more competence than this.
The numbers, which are worse than you think
You might assume that after Enron, standards tightened. Let us look.
Regulators grade audits. A “deficiency” means the auditor did not gather enough evidence to justify the opinion they signed. Not that they were necessarily wrong, just that they could not show their working, and signed anyway.
In 2022, 43% of EY’s audits were found deficient.
Across all four firms, the average deficiency rate more than doubled in two years, from 12% in 2020 to 26% by 2022.
One in four audits. From the four companies who check almost every major business on earth.
Or take a simpler example. In 2014, in a routine job, PwC managed to overstate the profits of Tesco, a British supermarket chain, by 250 million pounds. Roughly 30% above reality.
If I misjudged a client group by 30%, I would arrive at a lodge with fourteen people and nine beds and I would be fired by lunchtime. These firms send an invoice.
The bit where they start marking their own homework
Then, in the 1980s, someone had an idea that broke the whole thing.
These firms noticed their reputations had become so strong that they did not need to sell a service at all. They could sell advice.
That is what consulting is. You pay an enormous firm to tell you how to run your business. Deloitte Consulting, PwC Strategy, EY Parthenon, KPMG Advisory, all of them grew out of the audit business like a second head.
By 2023, the Big Four made 95 billion dollars from advisory work, more than the 66 billion from auditing.
Now look carefully at what just happened.
The job of an auditor is to independently check whether a company’s numbers are honest.
The job of a consultant is to improve how that same company runs.
So the same firm now gives the advice, and then grades how well the advice worked.
That is marking your own homework. That is refereeing a match your brother is playing in, wearing his shirt, having bet on the result.
There is a permanent quiet incentive to sign off the books a little too kindly, because a generous audit makes the consulting arm look brilliant.
And by the way, what does all that expensive advice actually look like? Here is real language from a real consultancy slide: “develop value creating partnerships,” “build a clear mission,” “develop strategies to create sustainable related opportunities.”
That is it. That is what millions of dollars buys. Put it in a nice font, add a graph, choose a calming colour scheme, and it becomes wisdom. I have heard more actionable strategy from a taxi driver in Bulawayo, and he did not charge me a consulting fee, though he did have strong views on the national team.
The one where they got caught doing something genuinely outrageous
In 2013, the Australian government did a sensible thing. It hired experts to help write new laws to stop multinational companies dodging tax.
One of those experts was a PwC partner named Peter Collins. He signed confidentiality agreements. Several of them.
He then shared that confidential government information with colleagues at his firm.
They used it to help clients get ahead of the very law PwC was helping to write. PwC later identified 76 current and former partners connected to the affair. Reporting indicated that Uber and Facebook restructured their operations just weeks before the law took effect, and the tax office was reportedly stunned at how fast companies managed to rearrange themselves.
Read that sequence slowly, because it is beautiful in the way a car crash is beautiful.
The government hired a firm to help close the loopholes. The firm used its inside knowledge of the loopholes to help corporations sprint through them before they closed.
The eventual consequences were real. PwC Australia sold its entire government consulting business, worth over 100 million dollars, for one Australian dollar. Its chief executive resigned. Federal Police opened an investigation.
One dollar. That is what a century of reputation is worth on the day it runs out.
And now, the chatbot
Which brings me, finally, to the thing that made me want to write this.
In 2024, the Australian government wanted a review of its welfare system, which had been automatically issuing penalties to poor people. Grim stuff, and exactly the sort of thing that deserves serious independent scrutiny.
They hired Deloitte. They paid 440,000 Australian dollars, about 290,000 US dollars.
Deloitte delivered a 237-page report.
It contained references to academic papers that do not exist. It quoted a Federal Court judgment that was never said. It cited respected legal scholars on work they had never written.
Deloitte had used a generative AI tool, and the AI had simply invented sources, which is a known behaviour charmingly called “hallucination.”
Nobody at the firm checked.
They were caught not by a regulator, not by a Senate committee, but by one academic at Sydney University named Chris Rudge, who read the footnotes and noticed the books were imaginary.
Deloitte quietly published a corrected version, added a disclosure admitting the AI use, and refunded part of the fee.
I coach teenagers. I have received homework that was clearly not written by the teenager holding it. The difference is that when a seventeen-year-old hands me an invented essay, he looks at his shoes, and he does not invoice me a quarter of a million dollars first.
Why any of this should matter to you, in Zimbabwe
Fair question. Why should a Zimbabwean care about Australian accountants?
Two reasons, and the second one should annoy you.
The first is that this is our story in a nicer accent. We know exactly what happens when a small group of well-connected players capture something essential and cannot be dislodged. We know what a tender going to the same firm for the twentieth year looks like. We know what happens to quality when nobody can be replaced.
When it happens here, we call it corruption. When it happens there, they call it market consolidation. The machinery is identical.
The second reason is more direct. These are not abstract foreign companies. They are the same firms that audit the mining companies operating on our continent, that advise African governments on tax and privatisation, that produce the reports used to justify decisions affecting millions of us. Investigations like LuxLeaks and the Panama Papers repeatedly put these firms near the centre of the global tax avoidance machinery, the system that helps multinationals dig things out of poor countries and book the profits somewhere sunny with a low tax rate.
And here is the part that should keep you up. When one of these firms gets sloppy in Sydney, an Australian academic reads the footnotes and the story goes global.
When one gets sloppy in Harare, or Lusaka, or Kinshasa, who exactly is reading the footnotes?
The scandals we hear about are the ones that happened in countries equipped to catch them. That is not comforting. That is the opposite of comforting.
What happens now
Two things are finally moving.
Governments are stirring. In the UK, since 2024, all four firms have had to operationally separate their audit arms from the rest of the business. Australia has held inquiries. Even the US Senate has discussed a breakup. This frightens them, because the two halves depend on each other: consulting is where the money is, but it only commands those prices because a century of auditing built the credibility.
And then there is the technology that produced our chatbot disaster.
For decades, part of what made these firms untouchable was cheap, brilliant labour. The best graduates funnelled into accounting and consulting almost automatically. At Harvard, roughly 57% of the class of 2022 went straight into finance, consulting, or tech. A small firm could never compete for that talent.
But the actual grunt work, the document review and the data sorting, is exactly what AI does fastest. And that technology is not theirs. A firm of thirty can now produce what once needed three hundred.
So the moat is draining. Clients are bringing work in-house or moving to smaller firms. By 2024, only 13% of businesses felt consultants were doing more good than harm.
Thirteen percent. That is not a reputation problem. That is a customer base that has quietly stopped believing in you.
The lesson, which is much older than accounting
Here is what I keep turning over.
These firms were not built by villains. They were built to solve a real problem, and for a long time they solved it well.
What ruined them was not greed exactly. It was safety.
Once they could not lose, they stopped needing to be good. Once they stopped needing to be good, they stopped being good. And because nobody could replace them, it took twenty years for anyone to find out.
That is not an accounting principle. It is a political one, and it is close to the only one I fully trust. Power that cannot be removed will rot. Always. It does not matter whether it is a government, a liberation movement, a monopoly, or a firm that has audited the same company since 1882. The mechanism never changes, and neither does the cure: competition, transparency, and a real chance of losing your position.
I take a strange kind of hope from all of this. Not from the regulators, who slept through two decades. From the fact that a firm eventually embarrassed itself so publicly that people finally looked.
And when they finally looked, it was not a watchdog that caught them. It was one man in Sydney who bothered to read the footnotes.
That is usually who gets them. Not the system. Just somebody paying attention.
Which leaves the only question that really matters, and I ask it of my own country far more than anyone else’s.
Who is reading our footnotes?
