Red Flags Ch. 2: Prosperity the Cash Never Confirmed

Investment Background

The most durable warning sign in the history of corporate collapse is one sentence long and was published in 1934.

security-analysis chapter 3 owns it, and this chapter does not restate it. Graham and Dodd's rule, in the form that chapter gives it: net income can be manipulated, cash in the bank is much harder to fake, and if reported profit is soaring while the cash generated from operations is not, run away. That chapter also owns the manipulation taxonomy, the footnote discipline, and the ten-year normalised earnings method. It is the analyst's toolkit and it has not needed revision in ninety years.

So this chapter asks the question Graham does not. The rule is one sentence. It is famous. Every institutional holder of Enron, Wirecard, Valeant and Luckin had read it, and many of them could have recited it. They held anyway, in size, for years. That is the phenomenon worth a chapter — not the rule, but its remarkable failure to be acted upon by the people who knew it best.

The Wall Street Translation

What the Divergence Looked Like From Outside

This chapter describes a shape, not a calculation. Nothing here asks the reader to compute anything, and nothing here explains how any statement is constructed.

The shape is simply this: a company reports that it is becoming steadily more profitable, and the money that a more profitable company would be accumulating does not appear. Reported prosperity goes one way. Observable cash goes another, or goes nowhere. The two lines separate, and then they stay separated, quarter after quarter, in documents anyone can obtain.

In every corpse in this book, that separation was visible for years before the end.

And in every case it was explained. That is the part that matters, and it is why the rule fails in practice despite being trivial in theory. The gap never arrives unlabelled. It arrives with a reason, and the reason is always specific, technical, forward-looking and plausible: the business is investing for growth, the accounting treatment understates economic reality, the working capital cycle is temporarily unfavourable, the industry is unusual in this respect. Each explanation is individually reasonable. What is not reasonable is the same explanation, quarter after quarter, for six years.

The Corpses

Enron reported rising profits through the late 1990s while the cash those profits implied did not arrive, and closed the gap with asset sales and financing rather than operations. The divergence was public and durable. It was also, at the time, read as sophistication rather than as a warning — the company was understood to be a new kind of business whose economics traditional measures did not capture. Being told that the ordinary measure does not apply to this company is itself the pattern.

Wirecard is the purest instance, because there the missing item was not a subtlety of measurement at all. Roughly two billion euros of cash was reported to sit in trustee accounts in Asia. In June 2020 the company's auditor declined to confirm it, and shortly afterward the company acknowledged that the cash likely did not exist. The share price went from a DAX constituent to insolvency within days. No accounting knowledge was required to understand what had happened: the money was said to be somewhere, and it was not there.

Valeant grew reported earnings through serial acquisition while presenting a heavily adjusted profit figure that chapter 3 takes up. The cash the business generated never supported the reported growth story, and the gap was financed with debt. When the acquisition machine stopped, the reported earnings stopped with it.

Luckin Coffee reported store-level sales that a Chinese research firm tested by the simplest possible method: sending people to sit in the shops and count customers, and collecting the receipts. The counted traffic did not support the reported revenue. The test required no accounting at all.

Why Knowing the Rule Did Not Help

The rule is one sentence. The people who held these companies knew it. Four structural reasons explain the gap between knowing and acting, and none of them is stupidity.

The signal is chronically early. A divergence can persist for years while the share price rises. Being right early is operationally identical to being wrong, and for a professional with quarterly performance reporting it is worse than being wrong, because it is visible.

The signal has an enormous false-positive rate. Genuinely investing companies, genuinely growing companies, and companies in genuinely unusual working-capital positions all produce the same shape. Most companies exhibiting the pattern are not frauds. A rule that fires constantly and is usually wrong gets discounted by anyone who applies it repeatedly — which is exactly what happened.

The explanation always exists and is always specific. The analyst asking the question receives a detailed, technical, confident answer from management, typically in a private meeting. Accepting it is the professional default, and refusing it requires accusing named executives of dishonesty on the basis of a pattern that usually has an innocent cause.

And the incentive runs one way. Chapter 5 is about this in full: the analyst who accepts the explanation and is wrong is one of many; the analyst who rejects it and is wrong is alone and identifiable.

Note carefully what this means. The reason these frauds ran was not that the signal was hidden. It was that the signal was public, well known, easy to state, and structurally difficult to act on. That is a much harder problem than a hidden signal, and no amount of additional analytical technique addresses it.

Division of Labor With the Rest of the Library

The boundary with security-analysis is the one to hold. Graham owns the rule and the technique. This chapter owns its failure in practice — and the failure is not analytical, which is why more analysis is not the answer to it.

Executable Trading Rules

1. Do not adopt the divergence as a screen. It fires constantly, is usually a false alarm, and acting on it mechanically will have you selling ordinary growing companies for years. This chapter is not giving you a filter to run.

2. Apply it only to what you already hold in size. The question is never "which companies show this pattern" but "does anything I own in a plan-threatening size show it, and how long has it shown it".

3. Count the years, not the quarters. One quarter of divergence is noise; four consecutive years with a fresh explanation each time is a different object entirely. Duration is the discriminating variable, and it is the one a screen cannot capture.

4. Treat "the ordinary measure does not apply to this company" as the specific phrase to be alert to. Enron was understood to be too novel for traditional metrics. That framing preceded every collapse in this chapter, and it is the moment scrutiny is being asked to stand down.

5. Prefer the physical test to the analytical one where one exists. Luckin was caught by counting customers. Wirecard was caught by asking a bank to confirm a balance. Neither required expertise, and both were more decisive than any amount of statement analysis.

Relevance to a Retirement Portfolio

The uncomfortable conclusion of this chapter is that the best-known warning sign in corporate finance did not protect the professionals who knew it. It is early, noisy, and structurally hard to act on. A retail investor applying it part-time should expect to do worse with it than they did, not better.

Which is precisely why the defence for a retirement portfolio is not analytical. You do not need to detect the divergence in a company you own a two-thousandth of. A diversified index holder was exposed to every corpse in this chapter and required no defence at all, because the position sizes made detection unnecessary.

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. The literacy in this chapter is for the concentrated position you already hold, and its practical use is narrow: if something you own in plan-threatening size has reported prosperity that the cash has not confirmed for several years running, that is a reason to reduce the position size — not a reason to become an analyst.

Chapter 3 turns to the number that replaced the reported one, and to who writes it.