A decade ago, as the financial crisis gathered momentum, two giants of the US investment landscape,
Goldman Sachs and the insurer American International Group, were locked in an arcane but high-stakes
accounting dispute.
AIG’s fast-growing, London-based financial products arm had written billions of dollars’ worth of in
surance in the form of credit derivatives against a mountain of the investment bank’s packaged-up mo
rtgage loans. As credit conditions deteriorated in late 2007 and these mortgage-related securities s
tarted to buckle, the value of this insurance policy became central to the financial health of both
businesses.
The bank wanted to recognise the mushrooming gain it was making on its derivative position, based on
a probabilistic estimate of the underlying loans defaulting. By its calculation, AIG owed $5.1bn on
its outstanding swap positions, a large chunk of that to Goldman Sachs. Unsurprisingly, AIG took a
very different view. The insurer estimated its liability was no more than $1.5bn, a sum that helpful
ly allowed it to continue posting quarterly profits.
Both sides sought the acquiescence of their auditors for these treatments. Coincidentally, in this c
ase that meant the same firm: PricewaterhouseCoopers. And despite the iron logic that one side’s gai
n in a zero-sum trade should mirror the other’s losses, the same firm allowed these divergent — and
mutually beneficial — approaches.
Only months later, with markets still frozen, did PwC toughen its line and force its insurance clien
t to take a substantial writedown. In the event, even this proved wildly insufficient. AIG’s derivat
ive positions ultimately forced it to cough up tens of billions of dollars in 2008 — a sum it could
only pay because the US government had by then bailed it out.
AIG was forced to pay out tens of billions of dollars on its derivative positions in 2008 after it h
ad been bailed out by the US government © Bloomberg
The word audit means to survey or check. Ferreting after facts was once the auditors’ main vocation:
certifying information to assure investors that a company’s numbers were “true and fair”. But in AI
G’s row with Goldman, it is striking how little was verifiable. There were few credible market price
s, let alone transactions, to support the key valuations. Speciously precise profits and losses were
written, not on the basis of concrete observation, but mathematical calculations derived from compu
ter models.
As to the logic of such contradictory valuations, Warren Buffett’s investment partner Charlie Munger
spoke for many when he observed in a 2009 interview that they “violated the most elemental principl
es of common sense”.
“Accounts have always contained estimates; think of the provisions companies make against foreseeabl
e future losses,” says Sharon Bowles, former chair of the European Parliament’s economic and monetar
y affairs committee. “But the un-anchoring of auditing from verifiable fact has become endemic.”
Recommended
Head to HeadAccountancy
Opinion head-to-head: Should the Big Four accountants be split up
In the UK in the past three decades, standards setters have progressively dismantled the system of h
istorical cost accounting, replacing it with one based on the idea that the primary purpose of accou
nts is to present information that is “useful to users”. The process allows managers to pull forward
anticipated profits and unrealised gains, and write them up as today’s surpluses.
More recently, it is behind a string of accounting scandals involving overstatements of profit, incl
uding at the UK supermarket chain Tesco and the software company Quindell. It hangs over the insolve
ncy of the UK outsourcing group Carillion, where sudden contract restatements in 2017 erased the pre
vious six years of dividend-bearing profits. In the US, the conglomerate GE is under investigation o
ver the way it accounts for its contracts.
These events have fuelled concerns about the auditing market, and whether the Big Four accountancy f
irms— KPMG, Deloitte, EY and PwC — are too big to fail, too profit-driven and excessively compliant
to managers’ wishes. Britain has started an investigation into the effectiveness of its auditing reg
ulator, the Financial Reporting Council.
But this may be looking at the symptoms rather than the cause of the problem. That may lie in the ac
counting standards themselves.
Carillion had only impaired less than 10 per cent of the goodwill on its balance sheet when it colla
psed © Bloomberg
Modern auditing in Britain sprang from a great failure: the collapse in 1878 of the City of Glasgow
Bank. Stung by the demise of this unlimited liability institution, whose losses hit the city’s middl
e classes, more banks became limited liability ventures. With that privilege came responsibilities,
including the need to conduct independent audits.
Their purpose was to assure investors that companies’ capital was not being abused by over-optimisti
c or fraudulent managers. “At their heart, audits are about protecting capital, and thereby ensuring
responsible stewardship of capital,” says Natasha Landell-Mills, head of stewardship at the asset m
anager Sarasin & Partners.
Behind this stood a system of accounting that vaunted prudence. The principle that assets were value
d at the lower of cost or net realisable value (or the price at which it was thought they could be s
old) did not rule out estimates. But they only came into play when values had fallen. It was not pos
sible for managers to conjure up unrealised gains and profits and present these as fact.
The idea that accounts should be primarily “useful” springs from the same source as the so-called ef
ficient markets hypothesis. Indeed, it is an adjunct to that now somewhat discredited theory.
From the 1960s, academics such as William Beaver at Stanford University advanced the notion that for
markets to channel capital efficiently to the most productive outlet, accounts needed to give trade
rs of securities a clearer understanding of the current valuation of a company.
Heating up
$134bn
Combined revenues of the ‘Big Four’ auditors. They employ 945,000 people
$2.9tn
Total goodwill on balance sheets of S&P 500 companies in 2016, up from $1.8tn in 2007
60%
Remuneration based on equity for CEOs at S&P 500 groups in 2014, up from 25% in 1992
That meant abandoning inconvenient notions such as prudence and conservatism; instead, accounts had
to be “neutral” and use more up-to-date values for balance sheet items.
This was radical. Fair value accounting had been firmly shunned by the US Securities and Exchange Co
mmission for contributing to the losses of the 1929 crash. Yet contemporary events gave the hypothes
is respectability. Soaring inflation in the 1970s made historical cost balance sheets seem misleadin
gly out of whack with property values, leading to asset stripping. America’s savings and loan crisis
in the 1980s was partly blamed on these institutions having out-of-date books.
From the 1990s, fair values started to supplant historical cost numbers in the balance sheet, first
in the US and then, with the advent of IFRS accounting standards in 2005, across the EU. Banking ass
ets held for trading started to be reassessed regularly at market valuations. Contracts were increas
ingly valued as discounted streams of income, stretching seamlessly into the future.
This was also a time when managers’ pay, especially in the US, was rising through the use of market-
linked incentives. Between 1992 and 2014, equity-based pay at S&P 500 firms rose from 25 to 60 per c
ent of their total remuneration, according to database ExecuComp.
It did not take long for bosses to perceive the pecuniary possibilities of their ability to influenc
e fair values. Between 1995 and 1999, for instance, Enron’s stock underperformed the S&P 500 index.
Yet in 2000, when the US energy company’s accounting chicanery started to kick in, its shares wildly
outperformed the benchmark. In the 10 months before its collapse, the company paid out $340m to exe
cutives.
With accounting chicaery helping its shares wildly outperform the S&P500, Enron paid out $340m to ex
ecutives in the 10 months before its collapse in October 2001 © Reuters
“The problem with fair value accounting is that it’s very hard to differentiate between mark-to-mark
et, mark-to-model and mark-to-myth,” says one investor who is on the board of an audit firm.
In theory, fair value should not preclude sound audits. But it does make it harder. The greater lati
tude given to bonus-hungry management by looser evidential standards increases the pressure on audit
ors.
There is, however, little evidence of auditors rising to the challenge.Take goodwill, an accounting
item that measures the difference between the purchase price paid for an acquisition and the net val
ue of the assets actually acquired.
Until the turn of the century there was a general convention that when one company bought another, g
oodwill was an effective cost of the transaction that needed to be amortised — or written down annua
lly against group profits. Not to do so, while also counting the additional profits from the purchas
ed assets, was a form of double counting that inflated the benefits of an M&A deal. According to Kar
thik Ramanna, professor of business and public policy at Oxford university’s Blavatnik School of Gov
ernment, “it violates the basic premise of traditional accounting”.
A cheque from the City of Glasgow Bank, whose collapse in 1878 prompted tighter accounting requireme
nts © AllyD/Wikicommons
Yet egged on by Wall Street, standards setters softened the rules on goodwill in 2000. This could be
recognised in the balance sheet permanently, and only reduced if there was evidence that the discou
nted future cash flows from the underlying asset had fallen sufficiently to warrant impairment.
Puzzlingly, given the hit and miss nature of most takeovers, this evidence has proved remarkably elu
sive. Since 2007, the total goodwill on the balance sheets of S&P 500 companies had rocketed from $1
.8tn to $2.9tn by 2016, much of it used to collateralise acquisition debt.
When Carillion collapsed in January, the group had only impaired £134m of the £1.5bn of goodwill on
its balance sheet, even though at least one large acquisition had negative net assets of nearly £200
m and was only solvent because of explicit support from the parent group.
There may be little evidence of auditors getting tougher, but they have reacted in other ways to the
challenges posed by fair value accounting: principally huddling together and seeking ways to limit
their liability.
From a Big Eight in 1987, the industry consolidated to a Big Five by 1998. With the collapse of Arth
ur Andersen in 2002, their number shrank to four. These firms entirely dominate the markets for audi
ting quoted companies in the UK and the US. Many observers accept this lack of choice makes the indu
stry difficult to regulate. “It makes the Big Four too big to fail,” says Guy Jubb, an honorary prof
essor at the University of Edinburgh and a corporate governance expert.
Last year the quartet had combined revenues of about $134bn, employing almost 945,000 people, accord
ing to Statista. Scale makes the Big Four much more of a target for shareholder litigation, especial
ly in the US.
To counter this and reduce the scope for actions, auditing firms have used their lobbying power to e
rase ever more of the discretion and judgment involved in what they do. Hence the explosion of “tick
box” rules designed to achieve mechanistic “neutral” outcomes. It is a process, says Prof Ramanna,
that is tantamount to a stealthy “socialisation or collectivisation of the risks of audit”.
Into the void have stepped self-interested managers, who acquire ever more influence over the presen
tation of the numbers. Take the practice of accounting for the zero-interest credit cards that UK ba
nks issue in the hope of making money when the customer comes to the end of the interest-free period
and starts paying a high rate.
“Realistic” accounting practices allow companies to recognise revenue up front based on their estima
tes of the amount of time beyond the expiry of the free period. The more optimistic a manager’s view
s on customer retention, the higher the asset values. Given the incentives on bosses, it is hardly s
urprising that the result has been ballooning zero-interest credit card balances, and growing concer
n at the Bank of England that a bubble may be inflating.
Observers perceive the problem, but worry that the rules now preclude auditorial judgment. “How do y
ou deal with it when you get into a face-off with management and they say, ‘This is how we want to p
resent it and it is all within the rules’? Ultimately, it is very hard to argue against,” says the i
nvestor who is also an audit firm non-executive.
There is also the perception that the dominant Big Four, which are now profit-hungry professional se
rvices conglomerates, are not that worried about audit quality anyway. “They have been able to do be
tter with low quality than with high quality work,” says Erik Gordon, a professor at the University
of Michigan Ross School of Business. “It is less expensive and clients, who actually are company man
agement, not the shareholders, seem happy with audits that don’t challenge their view of how well th
ey are performing.”
Some at the largest accounting firms do recognise that trust in audit has plummeted. “There are some
legitimate reasons for that,” says a senior figure at one of the Big Four firms. “If there was noth
ing [wrong] we would not be having this debate. It would not be getting so many column inches if, at
its core, something wasn’t working.”
The Big Flaw: Auditing in Crisis
The FT examines what has gone wrong with accounting — and what might restore faith in the profession
.
Part one in the series
What has gone wrong with accounting standards?
Part two
Conflicts of interests — the impact 
Part three
Are regulators too weak?
Part four
Accountants behaving badly
Part five
How to fix the problems
There is ultimately a tension between accounts that are prudent and those that are “useful to users”
. Despite the searing experience of 2008, the accounting profession has plumped for the latter. In 2
010, in a groundbreaking conceptual paper, both America’s Financial Accounting Standards Board and t
he International Accounting Standards Board scrapped “reliability” or the need for factual verificat
ion in favour of relying on “faithful representation”, which means little more than an educated gues
s.
“It’s an indication of how standards setters, divorced from public accountability and coddled by cor
porate special interests, can manufacture their own reality,” says Prof Ramanna.
This has opened a gap between the standards and UK company law, which says the accounts of a company
should present a “true and fair” picture of its financial position and earnings.
Critics of fair value worry that standards have diverged too far from this legal requirement. Many o
f the recent changes have been stealthy, driven by international harmonisation, or a result of the “
useful for users” agenda. Ms Landell-Mills worries that investors have not absorbed the consequences
for the public interest.
“In the extreme,“ she says, “fair value accounting that treats upward revaluations as legitimate pro
fits, and ignores future foreseeable losses, can facilitate Ponzi schemes where more and more illuso
ry profits permit executives and existing shareholders to extract cash through bonuses and dividends
. It won’t be long before you come unstuck.”
This article was updated on August 1 to reflect the fact that Carillion had impaired some goodwill i
n September 2017
If you have an insight or tip about the problems facing accountancy that could inform our reporting,
please contact madison.marriage@ft.com. We want to hear from you. If your information is particular
ly sensitive, consider contacting us using one of these secure methods.
Letters in response to this article:
Bold move on audit is unlikely to come about / From Guy Dresser, Fislisbach, Switzerland
Forgotten truths about cash and liabilities / From Neil Woodcock, Edinburgh, UK