Red Flags Ch. 3: The Number the Company Prefers

Investment Background

Every company in this book reported two profit figures, and the market used the larger one.

The larger figure was written by the company itself. It was the one in the headline of the press release, the one repeated on the earnings call, the one the sell-side models were built on, and the one the financial press quoted. The smaller, official figure was disclosed, was available, and was in practice ignored.

thirty-six-stratagems chapter 5 owns this as a single stratagem — Deck the Tree with False Blossoms, with two tactical trading rules attached. This chapter is not a longer version of that entry. The stratagem describes the tactic. This chapter is about who authors the number, what it has been trained to remove, and what a decade of the same removal looks like.

The Wall Street Translation

The Structural Oddity

Set aside every technical question about accounting standards, because this chapter does not deal in them and will not explain them.

Reduce the arrangement to its structure and something obvious appears. A company publishes an official profit figure produced under rules it does not control and which an outside firm attests to. It then publishes a second figure, produced under rules of its own devising, which it may change from period to period, and which is reliably larger. Executive compensation is frequently tied to the second one. Everyone then uses the second one.

Stated that plainly, the arrangement is strange, and it is worth holding the strangeness rather than resolving it. Nothing here claims the second figure is illegitimate — there are genuinely defensible reasons a company might present one, and most companies that do are not committing fraud. But the second figure is the company's own account of its own performance, and the entire apparatus of independent verification applies to the figure nobody is looking at.

What Gets Removed, and the Pattern That Matters

The observable pattern is not the adjustment. It is the recurrence.

A genuinely one-time item happens once. A restructuring charge because a business was reorganised, a loss on an asset sale, a legal settlement — each of these is a real event, and removing it to see the underlying business is a defensible thing to do. Removing it every single year for a decade is not.

That is the shape this chapter asks a reader to see, and it requires no accounting at all to see it. Line up the last ten years of a company's own adjustments and ask one question: which items appear every year? An item that appears every year is not an exception to normal operations. It is normal operations, relabelled.

security-analysis chapter 3 states the principle in one line — a "non-recurring" charge that recurs every two years is a regular cost of doing business. Graham owns that principle. What this chapter adds is that in the corpses, the recurrence ran for years, was fully disclosed, and was accepted anyway by the professionals whose job was to notice it.

Valeant, and Adjustment as a Business Model

Valeant Pharmaceuticals is the case where the adjusted figure was not a garnish on the results but the substance of the story.

The company grew by acquiring other pharmaceutical companies, cutting their research spending, and raising drug prices. The reported official profit was persistently weak. The adjusted figure — which removed, among other things, the costs of the acquisitions the entire strategy consisted of — was strong and grew rapidly.

The structural objection is stateable in one sentence and required no expertise. A company whose business model is serial acquisition, presenting a profit figure that excludes the costs of acquiring, is presenting the profit of a company that does not exist. The excluded item was not incidental to the business. It was the business.

It was disclosed. It was discussed. Sophisticated, highly respected value investors held the position in enormous size. When the acquisition machine stopped — pressure on drug pricing, a dispute over a specialty pharmacy relationship, a debt load that had grown with the strategy — the adjusted number stopped growing, and the equity fell more than ninety percent.

Why Disclosure Did Not Function as Protection

This is the chapter's real subject, and it generalises beyond accounting.

Everything described here was disclosed. The adjustments were itemised. The official figure was published alongside. Nothing was hidden, and any reader could have compared the two.

Disclosure protects against concealment. It does not protect against attention. The adjusted figure appeared in the headline; the official figure appeared further down. The adjusted figure was what the analyst models forecast, so it was the number that "beat" or "missed", so it was the number that moved the share price, so it was the number that mattered — regardless of which one described the company.

A market can be fully informed and still be pricing the wrong number, purely because of which number it habitually looks at. That is a failure of attention rather than of information, and no additional disclosure requirement fixes it. It is also why "but it was all in the filings" is a true statement that explains nothing.

Division of Labor With the Rest of the Library

Book	Owns
thirty-six-stratagems ch05	Stratagem 29 — adjusted figures as a tactic, with two tactical trading rules
security-analysis ch03	The principle — a recurring "non-recurring" charge is a regular cost, plus the whole analyst's toolkit
the-essays-of-warren-buffett ch04	Owner earnings — adjusting reported figures to VALUE a business you believe in
This book	The decade-long recurrence in real corpses, who authors the preferred number, and why full disclosure protected nobody

The boundary with the-essays-of-warren-buffett chapter 4 deserves a sentence, because the vocabulary is nearly identical and the purpose is opposite. Buffett adjusts reported figures to get closer to the economic truth of a business he wants to own. The companies in this chapter adjust reported figures to get further from it. The same act, in opposite directions, distinguished by who performs it and whether they benefit from the answer — which is the whole distinction, and it is not an accounting distinction.

Executable Trading Rules

1. For anything you hold in plan-threatening size, find out which number the market is quoting. If the figure in the headlines and the figure in the official statements differ substantially and persistently, you are holding a company whose price is set by a number it wrote itself.

2. Line up ten years of adjustments and look only for repetition. Any item excluded in most of those years is an ordinary cost with a label on it. This is a reading exercise, not a calculation, and it is the only exercise this chapter asks for.

3. Ask what the adjustment removes relative to the strategy. A company that grows by acquiring, excluding acquisition costs, is the Valeant structure. The test is whether the excluded item is incidental to the business model or constitutive of it.

4. Treat compensation tied to the company's own figure as an amplifier. When the number management is paid on is the number management authors, the incentive to keep the two figures apart has been made explicit rather than left implicit.

5. Do not convert this into a screen. Most companies presenting adjusted figures are ordinary companies doing an ordinary and defensible thing. The pattern here discriminates only in combination with the divergence of chapter 2 and only over many years.

Relevance to a Retirement Portfolio

A retirement investor holding index funds is not exposed to this chapter in any way that requires action. Whichever number the market quotes for any single constituent, the index holder's outcome is the aggregate, and a single name's collapse is absorbed. That is not complacency; it is the specific benefit being paid for.

The exposure is entirely in concentrated positions, and among those the most common by far is employer stock — where the reader is not only a shareholder but is being paid in the thing, has heard the adjusted figure presented internally as the real measure of the company's health, and is the last person in the world positioned to be skeptical about it. human-capital-portfolio chapter 3 owns that concentration problem, and this chapter simply notes that the adjusted number is usually how it is made comfortable.

So the standard recommendation is unchanged: a core of low-cost, globally diversified index funds, no leverage, a cash buffer covering essential spending, and withdrawal rules containing an adjustment mechanism. What this chapter adds is one habit for the reader who holds single names: know which of the two numbers you have been reasoning about, because in every collapse in this book, it was the one the company wrote.

Chapter 4 turns to the item that is hardest of all to verify from outside: whether the sale happened.