Company Valuations: Keep the Books

Myth 1: Accounts do not matter

It is an apparently peculiar phenomenon that the people most likely to argue
that accounts are irrelevant to valuation are accountants. Generally, the point
arises in the eternal debate over the relative merits of strict historical cost
accounting versus the practice of marking assets and liabilities to fair values.
The conservatives in this debate – those who would prefer to recognise amounts
at depreciated historical cost – produce two arguments to support their
position.

The first is that historical costs are knowable precisely whereas fair market
values are subjective. This is clearly correct. Whether it has persuasive force
is another matter, of which more below.

The second is that it is not worth trying to produce estimates of fair market
values because investors do not use accounts to value companies. Companies have
an obligation to report to investors what they have spent, and what they are
achieving by way of a return after a charge has been made for the use of the
assets they have acquired, but that is all that they can reasonably be expected
to do. Investors and other interested parties should, and do, produce their own
valuations and these should not and do not depend primarily on published
accounts.

This argument gets a significant level of support when investors are asked
how they do in fact make their investment decisions. They tend to downplay the
importance of published reports and accounts, and either cite soft issues such
as opinions of management, or refer to research reports produced by investment
banks or independent consultants, company investor relations presentations, and
other sources.

The problem with this is that it begs further questions. From where do the
investor relations departments of companies get the figures that underpin their
presentations? How does a research analyst in an investment bank model and value
a company? In both cases, the answer is predominantly that the analysis of
historical performance and of valuation is almost entirely extracted from data
produced in company reports and accounts.

There are industries in which it is feasible to value the company’s existing
assets directly. These include property companies, mineral extraction companies,
drug companies, and any company whose assets comprise a fairly small number of
separable units that can be modelled independently. A company valuation can then
be built entirely on the sum of these values, and a deduction made for financial
liabilities.

Even in this case, would the company actually trade in a free market at the
resulting valuation? Probably not, because investors would put a premium or a
discount on the theoretical value to reflect their perception of the likelihood
of the company adding to or subtracting from its value in future years. And from
where would they obtain this perception? They would have to assess its record,
as far as this could be meaningfully inferred, from its published reports and
accounts.

Myth 2: Bygones are bygones

As part of a degree in economics, one of the authors was taught that rather
than publishing the conventional profit and loss account and balance sheet, it
would be more useful if companies merely produced heavily annotated cash-flow
statements instead. The supporting argument was that the value of a company
derives from the present value of its discounted future stream of free cash
flow. Since it is cash flow that we want to discount, it is information about
cash flow that will help us best in producing our valuations.

Closely related to this argument is the notion that historical expenditures
do not matter. What we want to know about is the future, not the past. Sunk
costs do not matter. These opinions were once expressed most memorably in a
private conversation with the now retired chief financial officer of a major
international oil company. “I do not know why anyone ever bothers to read our
balance sheet,” he said. “It is merely a meaningless heap of residual items.”

What he meant by this is that if the company had spent $600m on the
development of an oilfield, and the oilfield had by now produced half of the oil
that it was estimated to be likely to produce, then it would have been depleted
in the accounts so as to have a net book value of $300m. Does this amount
represent an estimate of its fair value? Clearly not. Does it represent an
estimate of what it would cost to replace? Clearly not. Then it is indeed a
meaningless residual item.

The question is not whether strict historical cost accounting results in
meaningful numbers being recognised on balance sheets. It clearly does not. The
question is whether or not this matters. And the answer is that it matters very
much, for two reasons.

The first is that for many companies it is not practical to value assets from
the outside except with reference to figures from reports and accounts.

The second is that, as we have seen, valuations are modulated by opinions on
future prospects. And from where are investors to obtain opinions on future
prospects except from their knowledge of history? It is not so much a matter of
those who forget history being doomed to repeat it, as of those who know no
history having a very poor understanding of the present and an inability to
anticipate possible futures.

Myth 3: Goodwill write-offs don’t matter

Of all the items recognised in accounts, goodwill has probably been subject
to the fiercest debate. The abolition of “pooling” in the US generated even more
excitement among chief executives than has the more recent debate over expensing
employee stock options, or the adoption of the new accounting standards for
treatment of derivatives in Europe. The move from amortising goodwill to
capitalising it subject to annual impairment tests was greeted with relief by
companies mainly concerned to maximise their earnings per share by whatever
means.

Subsequently, those who have been required to impair a good deal of goodwill
have often argued that shareholders should not be perturbed by large impairments
of goodwill. The argument has two elements to it: the first is the
bygones-are-bygones point discussed above in myth 2. The second is more subtle.
It is that if an acquisition is financed through the purchase of the target’s
shares by issue of new shares in the bidder, then it is the relative value of
the two company’s shares that matter, and not the absolute value of the target.

This position could be (barely) caricatured, as follows. “Yes, I know that I
am buying something at about twice its fair value, but on the same analysis my
shares are trading at about three times fair value, so in relative terms I have
done a good deal.” And then, when the shares collapse: “Don’t worry about the
impairment; it is only a non-cash item. And actually our share price might well
have fallen even further in the absence of the deal. Given the values of the
day, it was a good transaction.”

To see what is wrong with this, one need only consider what would have
happened if the bidder had instead used its shares to buy something at a
reasonable price, or just raised some cash and paid off some debt. The mistake
arises through the conflation of two independent decisions: a financing decision
(using highly priced shares as a currency) and an investment decision (buying
something at twice its fair value).

Myth 4: Non-cash items do not matter

One of the more pernicious consequences of the advent of the discounted cash
flow (DCF) model has been the opinion frequently expressed in the form, “Don’t
worry about it; it is only a non-cash item.” As if to imply that its nature is
therefore entirely irrelevant to the value of a company.

The mistake arises because of the construction of most DCF models.

Unlike bonds, companies do not have finite lives, and ordinary shares are not
redeemable. This implies that all valuations are an attempt to put a figure to a
stream that extends to infinity. There is a limit beyond which it is not
practical or realistic to forecast individual line items in accounts for each
year, so models take the form of a forecast period and a terminal value. The
latter assumes that we have arrived at what might be conceived of as the
Platonic form of a normal year, from which we can assume that everything grows
at a stable state forever into the future. This makes the mathematics tractable.

The terminal value usually represents most of the value of the company. In
the late 1990s, most investment analysts working with telephone companies
produced models with up to 10 forecast years, because the business was
transforming and a long period was required before it was realistic to assume
that a stable state would be reached. Even with 10 years in the model, however,
for many companies more than 100% of the resulting values were attributable to
the terminal value part of the calculation. Among other things, this explains
why these valuations proved to be so unstable when the outlook for the industry
changed going in to the new millennium.

There is another, more subtle problem that arises in the transition from the
forecast period – let us say a five-year forecast – to the calculations that
drive the terminal value. To see why, let us review the components of the free
cash flows that we are discounting. In the absence of purchases or sales of
subsidiaries, or of disposals of assets, the flows into the company comprise
profit, depreciation and provisions. The flows out comprise capital expenditure
and changes in working capital. Free cash flow is the result – the stream that
is available to service debt and to reward shareholders.

Now imagine that profit has been reduced by a large provision for corporate
restructuring. Cash flow is not affected. Money in has simply been reclassified
from profit to provisions. And over the following couple of years, if our model
is properly constructed, the redundancy payments will be made, resulting in
substantial additional cash outflows, offset by a release of the provision. The
hit to the profit and loss account is taken in the first year, but the cash
flows out over the two subsequent years. Our cash flow model is discounting
cash, not profit, so we are indifferent to the timing of the charge to the
profit and loss account. It does not matter – it is only a non-cash item.

But what about a non-cash item, such as a charge made for stock options,
something that will be commonplace after the introduction of IFRS 2 share-based
payment? In this case, the non-cash charge cannot be ignored. It is a real cost
to the equity shareholders of the business and so must be modelled as a real
cost. The alternative is systematically to overstate the value of the existing
equity by ignoring the value that has been transferred to the recipients of the
stock options. The only sensible approach to dealing with this is to treat this
non-cash cost as if it were cash in a free cash flow calculation. Ignoring it as
it is non-cash would not reflect the economics of the business.

So, in the case of provisions that are ongoing (such as stock options), or
that will accrue after the end of our forecast period, we need to make some
adjustment to our basic model. We need to deduct from our valuation the
liability that has already accrued, even though it is not finance debt, and we
need to detach from our future stream of cash flow the future projected
provisions even though they are non-cash items. They matter very much.

Myth 5: EBITDA solves everything

The use of earnings before interest, taxes, depreciation and amortisation
(EBITDA) for the purpose of assessing business performance is now ingrained in
the financial markets. But although there are some very sensible reasons for
using EBITDA, some caution is required.

The normal reasons proffered for using EBITDA are that it is a cleaner
earnings number as it avoids being distorted by managerial policy on issues such
as amortisation periods and residual value estimation.

It is also sometimes suggested as a measure of cash flow. It should of course
be remembered that it is not cash flow, as the impact of working capital changes
or movements in provisions have not been adjusted for. It is simply nearer to
cash as it is after adding back two major non-cash items, namely amortisation
and depreciation.

The main problem with using EBITDA is that it ignores differences in the
capital investment profile of companies. It fails to distinguish between
companies that use capital efficiently and those that do not. In fact, extensive
use of this metric encourages the rather dangerous idea that depreciation is not
a real cost. If a business cannot cover its depreciation costs, it is not a
successful venture; that is clear.

Myth 6: Cash is fact, profit is opinion

This myth has grown in popularity but it is based on the very odd assumption
that cash as a real, “tangible” entity is somehow immune from manipulation.

To illustrate what is wrong with this approach, let us first offer a reminder
that earnings data is audited, and so has passed through a filter. There is
every reason to believe, especially since the collapse of Arthur Andersen, that
the quality of auditing has never been as high as it is now given the
increasingly litigious environments in which auditors operate.

Of course, cash is also subject to an audit. Therefore the auditor will
ensure that earnings numbers are generated by consistent accounting policies,
that disclosure is adequate and that the financials show a true and fair view
(or similar words, depending on what country you happen to be in). But there is
no such requirement for many cash flow figures. Therefore there is no reason to
assume that the cash flow figures are any more useful or less subject to
manipulation by the company than the earnings figures.

Suppose a company decides to stop paying suppliers for the last quarter of
its financial year. This would not alter the earnings numbers, but would
dramatically enhance cash flow. In a similar manner, offering generous discounts
to customers and slashing capital expenditure would both have a very significant
impact on the cash flow of a company, but would not be subject to any
adjustments by auditors except in the earnings numbers, where, for example,
discounts given to customers would be recorded as a cost.

It is actually quite straightforward to alter cash flow, so the idea that it
is an immutable and all-important fact is really nonsensical.

In conclusion, valuation offers far less clarity, unlike accounting, which
consists of an accepted suite of rules and standards. Part of the reason for
this is that accounts record the past, and valuation reflects an assessment of
the future. But despite the established literature on the subject, new valuation
myths develop all the time and are often repeated until dissent appears mildly
heretical. We hope we have at least encouraged the application of healthy
scepticism.

A slightly longer version of this article previously appeared in
Professional Investor, the monthly magazine for the Society of Investment
Professionals.

Nick Antill spent 16 years as an equity analyst specialising in oil and
gas companies. Email
[email protected]

Kenneth Lee is the accounting and valuation analyst at Citigroup
Investment Research. Email
[email protected]

They have recently published Company Valuation under IFRS
www.global-investor.com