Buy High, Sell Low - The Multibagger Series — Part 2

If you missed it, read it after this one rather than before. Part 1 gave you the machine. This one shows you the person operating it — and they are not on your side.

Two press releases. Same company.

December 2015. The board authorises a new share repurchase programme. The language is the language every board uses: confidence in long-term prospects, commitment to returning capital to shareholders. The stock is trading around $145.

28 October 2024. The same company prices 129,375,000 new shares at $143.00 — alongside $5.75 billion of mandatory convertible preferred. Total raise: $24.25 billion. The largest follow-on equity offering in the history of the United States.

Between those two dates the company spent roughly $43 billion buying its own stock — more than it earned over the period.

And the share count today is higher than it was before any of it started.

Nothing was hidden. Both releases were public. Both were covered by every wire service. Thousands of analysts with Bloomberg terminals and CFA charters read them in real time, built models around them, and published price targets.

Not one of them was graded on noticing.

Part 1 of this series ended on an identity:

Price = Earnings × Multiple.

Two numbers, multiplying. It is true, and I left something out of it — deliberately, because the missing term is not a footnote to that identity. It is the mechanism that connects the two halves of it, and it is the subject of this edition.

Earnings per share is a fraction. Every fraction has a floor.

Price = (Net Income ÷ Shares) × Multiple.

Three numbers. And unlike the first two — which are produced by customers and by the market’s collective judgement, neither of which anyone controls — the third one is set by a small group of people in a room, four times a year, using your money.

The company you own is the largest single trader in its own stock. It is the only participant that can create new inventory at will. It publishes its intentions in advance and executes them regardless of price, because the programme was authorised in dollars, not in valuation.

You are on the other side of that trade whether you know it or not.

Here is what actually happened at Boeing between 2013 and 2019, told in the order it happened rather than the order it makes sense in.

The company was performing. The 787 was ramping, the backlog was enormous, and free cash flow was extraordinary. Management did what a confident management does: it bought back stock. Aggressively, continuously, and — this is the load-bearing detail — with cash flow that peaked exactly when the share price did.

Diluted shares fell from roughly 760 million to under 570 million.

Now watch what that does to the reported numbers.

Boeing’s celebrated EPS growth in those years was not entirely Boeing’s earnings growth. A meaningful portion of it was the denominator shrinking underneath a business that was growing more slowly than the per-share figure suggested. The market saw EPS compounding at a superb rate and did what markets do: it paid a higher multiple for it.

Read that sequence again, because it is the point of this edition.

The buyback lifted EPS. The lifted EPS looked like superior growth. The superior growth justified a higher multiple. The higher multiple raised the price. And the next tranche of buyback was executed at that higher price — buying fewer shares per dollar, while making the optics better still.

The third number was not a separate lever. It was the transmission belt between the first two.

This is the part Part 1 could not tell you, because Part 1 treated the two triggers as independent. They are not. There is a mechanism that runs from one to the other, it is operated by management, it is funded by the company’s cash, and while the business is healthy it looks exactly like competence.

A flywheel that runs on capital rather than on the business is indistinguishable from a flywheel that runs on the business — right up until the capital stops.

The capital stopped.

You know the rest of the sequence: the second 737 MAX crash, the global grounding, the production halt, and then a pandemic that removed the customer base of the customer base.

But the interesting number is not the loss. It is the starting position.

Boeing entered the worst crisis in its corporate history with negative shareholders’ equity — around -$8.3 billion at the end of 2019, before COVID had happened at all. Not because the business was worthless. Because $43 billion of retained earnings had been converted into treasury stock at an average price the company would not see again for years.

A buyback is not a return of capital. A buyback is a purchase, and a purchase is not reversible. The cash left. The shares came back. The optionality that cash represented — to absorb a shock, to fund a fix, to buy time — left with it.

Five years later the company had to reacquire that optionality on the open market, from strangers, at $143 a share.

Bought high. Sold low. With the shareholder’s money, on the shareholder’s behalf, and reported in both directions as sound capital management.

The round trip, in full: roughly $43 billion spent, ~200 million shares retired, ~129 million shares reissued, plus a mandatory convertible that becomes common stock in 2027 — and a diluted count today of roughly 788 million against roughly 760 million in 2013.

Thirteen years. Forty-three billion dollars. And the denominator went up.

The business, for what it’s worth, is recovering — record backlog, improving deliveries, a return to positive earnings. Part 1’s first trigger is firing. It is firing across more shares than it would have if nobody had ever tried to help.

Now the same instrument, in different hands.

Dillard’s is a department store chain headquartered in Little Rock. Department stores were declared structurally dead somewhere around 2016 and have been declared dead approximately annually since. Nobody wrote admiring profiles of Dillard’s capital allocation.

Dillard’s has taken its share count from roughly 90 million to under 16 million.

Same tool. Same mechanics. Same line in the cash flow statement. And it produced one of the better shareholder outcomes of the last decade in American retail — out of a business that grew, at best, modestly.

The difference is not sophistication. It is price, and the discipline to act when the price is bad news rather than good news. Dillard’s bought heavily when the market had written the sector off. Boeing bought heavily when the market had decided it was a compounder.

Which gives you the only rule that matters here, and it is the same rule from Part 1 wearing different clothes:

A buyback is a bet on the multiple. An issuance is a bet against it. Management is making that bet with your capital, on a schedule set by their cash flow rather than by the price — which means they are structurally a buyer at highs and a seller at lows.

Structurally. Not because they are foolish. Because the cash to repurchase is most abundant precisely when the business is at its best, and the need to issue arrives precisely when it isn’t.

There is a second mechanism moving the third number, and it is quieter than the first.

Stock-based compensation is issued continuously — not quarterly, not on authorisation, but every pay period, to thousands of people, at whatever price the market happens to be quoting. It appears in the cash flow statement as a non-cash add-back: literally added back to earnings, on the reasoning that no cash left the building.

No cash left the building. Ownership did.

And the two mechanisms are printed within a few lines of each other on the same page, never netted, almost never read against each other. Repurchases appear gross, in financing activities. Compensation appears as an add-back, in operating. A company can spend two billion dollars retiring stock, issue two billion dollars of stock to employees, report both as positives, and leave your ownership stake exactly where it was.

You paid the compensation bill. You were told it was a dividend.

That is not a gap in disclosure. Everything above is disclosed. It is a gap in attention — and attention is the only thing in this business that is genuinely scarce.

Here is the test I actually run, given whole. It takes four minutes per holding and it requires two numbers, both free.

Most people, when they finally look at dilution, look at it the wrong way: they check whether the share count went up, decide it went up “a bit,” and move on. A bit is not a unit. Here is how to put it in dollars.

Step 1. Find the diluted share count on the date you bought. It is on the cover page of the 10-Q or 10-K covering that quarter.

Step 2. Take the company’s most recent twelve-month net income.

Step 3. Divide today’s net income by the old share count. Not the current one. The old one.

That figure is the earnings that belong to the ownership you actually paid for.

Step 4. Compare it to reported EPS. The gap is not an accounting abstraction. It is the portion of the business’s current profit that is now being collected by someone who was not on the register when you arrived.

Step 5. Multiply that gap by the current multiple. That is the dollar amount, per share, that was transferred out of your position while the business was busy doing exactly what you hoped it would.

And then the calibration, which is the part people get wrong in both directions:

• Under ~1% a year. Noise. Ignore it and go back to the first two numbers.

• 1–3% a year. The market rate for a company that pays people in equity. Not a red flag — a toll. Price it: it comes straight off your compound rate, every year, forever, and over a decade it is the difference between a 4× and a 3×.

• Above ~3% a year, sustained. You are not a shareholder in a business. You are a minority partner in a partnership that admits new partners every quarter without consulting you. This can still be a good investment. It cannot be a passive one.

Run it tonight on your largest position. If the answer surprises you, that is the edition doing its job.

The obvious conclusion from everything above is that buybacks are good and dilution is bad. That conclusion is wrong, expensive, and the mirror image of the error I’ve just spent two thousand words describing.

Royal Caribbean issued a great deal of stock to survive 2020 — its count rose by roughly a third. It was correct. A 30% smaller claim on a company that exists beats a whole claim on one that doesn’t, and every capital-intensive compounder in history has issued equity at some point. Selling stock at a rich multiple to buy assets that earn well is one of the most accretive things a management team can do. Selling it cheap to fund working capital is a slow liquidation with a press release attached.

And in the other direction: I once owned a company retiring 4% of its stock a year and read that as evidence of discipline. It was evidence of a company with nothing better to do with its cash, buying an expensive asset it happened to be standing on, with borrowed money. The denominator improved every single quarter while everything underneath it got worse. That was the most instructive position I have ever lost money in, and I’ll walk through it below.

Which means the third number cannot be read on its own. It has to be read against the second one — the multiple, the verdict from Part 1 — because a repurchase and an issuance are the same transaction with the sign reversed, and the same verdict prices both.

That is where the two editions become one system.