A company earns the same profit as yesterday. Yet its earnings per share climbs, and we hail it as a feat. The sleight of hand is simple: it bought back its own shares to retire them. The profit has not moved, but it is now shared among fewer slices. In 2024, the largest American companies thinned out their shares at a pace never seen before. A manufactured scarcity: who benefits from it, and what did it cost?
1 The record and the grain of sand
First, the scale.
The peak
Never so much repurchased
In 2024, S&P 500 companies devoted $942 billion to buying back their own shares, an all-time record, up nearly 19% on 2023 (S&P Dow Jones Indices). The first quarter of 2025 set a new quarterly record, $293 billion in three months. Over twelve months, buybacks approach a trillion dollars, more than the dividends paid out. Since 2009, around five trillion dollars have been returned to shareholders in this form.
S&P 500 buybacks, 2024
$942bn
all-time record, +19% year on year (S&P DJI).
Quarterly record
$293bn
in the first quarter of 2025 alone (S&P DJI).
The grain of sand
And yet, in the second quarter of 2025, the engine stalled precisely where it had run hardest. The five largest capitalizations all cut back their buybacks, the total falling 20% in a single quarter. The paradox stands whole: never have so many shares been repurchased, and already the machine is running out of breath among the technology giants who kept it turning. The reason: artificial intelligence is starting to devour the cash. We will return to this.
2 Fewer shares, same profit
First, let us understand the mechanism.
The arithmetic
Dividing the same cake into fewer slices
Earnings per share is calculated simply: net profit divided by the number of shares outstanding. When the company buys back and then cancels part of its shares, the denominator falls while the numerator, the profit, does not move. Mechanically, the ratio rises. This is the so-called "accretive" effect: each remaining share represents a larger slice of an unchanged profit.
A worked example
①
Before. A company earns 100 million, spread over 100 million shares. Earnings per share is worth 1.00.
②
The buyback. It repurchases 5% of its shares; 95 million remain. The profit, for its part, stays at 100 million.
③
After. Earnings per share rises to 1.053, that is +5.3%. Without a single dollar of extra profit, without a single extra sale.
The hidden rule
Financiers have a formula for knowing when a buyback lifts earnings per share: the inverse of the price/earnings ratio need only exceed the after-tax interest rate at which the company funds itself. As the reference textbook Vernimmen sums it up, "it is purely arithmetic, and it is not proof of value creation." Let us keep that sentence: the whole debate lies in the distance between two words, accretive and value-creating.
3 The Apple case study
The most telling example fits in two figures.
The demonstration
When growth is partly an accounting mirage
Since 2012, Apple has devoted on the order of $700 billion to buying back its shares, cutting their number by about 38% in ten years. The result is striking: over the decade, Apple's net profit grew by around 145%, but its earnings per share by nearly 280%. The gap between the two is precisely the manufactured scarcity of the shares. In May 2025, Apple again authorized $100 billion in additional buybacks, after $110 billion the year before, the largest such authorization in American history.
Net profit (10 years)
+145%
the real performance of the business.
Earnings per share (10 years)
+280%
nearly double: the effect of buybacks.
What it does not say
Let us be clear: Apple is a wonderfully profitable company, and its buybacks are the use of an overflowing cash pile. But when an investor reads "+280% in earnings per share," they believe they are seeing a business that has almost tripled. Half of that gain is in fact financial, not operational, in origin. The distinction is no detail: it changes the meaning of what we call "growth."
4 Accretive is not value-creating
Hence the decisive question: does the buyback create value?
The misunderstanding
Returning cash is not creating wealth
The consensus among financiers, from McKinsey to Aswath Damodaran, is clear: a buyback does not create value in itself, it returns cash. If earnings per share rises, the price/earnings ratio adjusts downward in mirror, so that the share price itself should not move. Only improving fundamentals, more sales, better margins, create lasting value. The buyback merely redistributes the same cake among fewer guests.
The real sharing
There is, however, a case where the buyback creates value, and another where it destroys it. If it buys back undervalued shares, the company enriches the shareholders who stay at the expense of those who sell. But if it buys back overvalued shares, it operates the reverse transfer: the money of the faithful flows to those who leave. Hence the sceptics' formula: a buyback, most often, does not create wealth, it shifts it. And everything depends on the price paid.
5 The "buy high" reflex
Yet the price paid, precisely, is rarely the right one.
The wrong tempo
Buying dear, stopping when it is cheap
Common sense would have a company buy back its shares when they are cheap. The opposite happens. Buybacks are heavily pro-cyclical: they peak at market tops and collapse in the troughs. In the first quarter of 2007, just before the crisis, they topped $160 billion per quarter; in the second quarter of 2009, at the bottom, when shares were being given away, they had fallen below $25 billion. And the all-time record of $985 billion over twelve months was reached in 2022, with prices at their highest. Companies buy dear and abstain when the opportunity is there.
Buybacks at the peak, 2007
> $160bn
per quarter, just before the crash.
At the trough, mid-2009
< $25bn
when shares were at their cheapest.
The danger of leverage
More dangerous still: some companies buy back on credit. Borrowing to repurchase one's shares boosts earnings per share so long as rates are low, but weakens the balance sheet. When the turn comes, the company finds itself in debt, with shares bought back too dear and no margin left to support the price. The buyback, meant to express confidence, then becomes the symptom of an excess.
6 Who benefits from scarce slices
We must then ask who decides, and why.
The incentive
When the boss's bonus depends on earnings per share
Executive pay is often indexed to "per share" targets or to the share price. Now a buyback lifts the one and supports the other, without the business having progressed. The suspicion is old: an executive could buy back not to serve the company, but to hit the target that triggers their bonus. A study by the Pay Governance firm half confirms it: 46% of companies that buy back use a per-share metric in their incentive pay.
The nuance, in fairness
But the same study tempers the charge: among the twenty largest repurchasers, nearly three-quarters say they neutralize the buyback's effect on executive bonuses, so as not to reward them for an accounting trick. The perverse incentive exists, but it is not systematic. The honest truth is a middle ground: the conflict of interest is real, yet the most closely watched boards have learned to defuse it. To be monitored case by case, not condemned wholesale.
7 What the money does not fund
There remains the heaviest reproach: the money diverted.
The accusation
"Profits without prosperity"
In 2014, the economist William Lazonick published in the Harvard Business Review an article that became a reference, "Profits Without Prosperity." His finding: between 2003 and 2012, S&P 500 companies devoted 54% of their profits to buybacks and 37% to dividends, that is 91% returned to shareholders. What is left, he asked, for productive investment, research and wages? For him, the company had shifted from a "retain and reinvest" model to a "downsize and distribute" one.
The counterpoint, in equal measure
①
The figure is right, the causality is not. That 91% of profits are distributed is a measurement; that buybacks cause underinvestment is a thesis, and a contested one.
②
The defence. The legal scholars Jesse Fried and Charles Wang (Harvard) remind us that one must reason in net flows, and find that the research and development of listed companies is at a record level. No robust evidence of crowding-out.
③
The recycling argument. The cash returned does not vanish: the shareholder reinvests it elsewhere, in other companies, other projects. Returning money one does not know how to use is a rational allocation.
The real dispute
The debate is not settled in one camp's favour. A company with no profitable project is right to return its cash rather than squander it. But when the buyback serves first to flatter a ratio, while research or wages are pared back, it becomes what Lazonick denounced. The dividing line is not the buyback itself, but what it replaces.
8 When the law steps in
This is why the lawmaker has never been entirely neutral.
The regulatory history
From forbidden to encouraged
It is easily forgotten: before 1982, buying back large amounts of one's shares on the market risked being labelled market manipulation. It was an SEC rule, 10b-18, adopted in the Reagan era, that created a "safe harbour" effectively legalizing the practice. Buyback volumes tripled within the year. Forty years later, the Inflation Reduction Act introduced a 1% tax on net buybacks, in force since 2023; a proposal to raise it to 4% was never enacted.
The return of suspicion
The subject remains politically live. The CHIPS Act forbids using its subsidies for buybacks. And notably, in June 2026, a bipartisan provision of the defence budget, backed by Donald Trump and Senator Elizabeth Warren alike, proposes to bar Pentagon suppliers from buying back their shares or paying dividends, save by waiver. The buyback, once suspect, then commonplace, becomes an object of distrust again when public money is at stake.
9 The world buys back too
It remains to widen the frame, and to conclude.
The fashion effect
From Japan to Korea, the same grammar
Long an American speciality, the buyback is being exported. In Japan, the governance reform of the Tokyo Stock Exchange, which presses listed companies trading below their book value to reward their shareholders better, has triggered record buybacks. Korea launched a "Value-up" programme in 2024 to reduce its famous "discount," resulting in buybacks led by Samsung and SK Hynix. Everywhere, the same idea: failing to grow, thin out one's shares to raise the value of each.
The compass
①
For the investor. Tell apart the rise in earnings per share that comes from the business from the one that comes from buybacks: it is not the same quality of growth, nor the same durability.
②
The good buyback. At the right price, with cash that has no better use and without paring the future, it is healthy. At the top, on credit, or at the expense of research, it weakens.
③
The signal of the moment. That the technology giants are cutting their buybacks to fund artificial intelligence says one thing: the money has found, elsewhere, a use judged more profitable than the scarcity of shares.
What scarcity costs
In the second quarter of 2025, capital expenditure by the S&P 500 jumped 24% year on year, while buybacks fell back. Among the giants, capital is flowing to data centres rather than to shareholders. It may be the best proof that the buyback was not a fate but a choice: that of lifting earnings per share for want of anything better to fund. When a better use appears, the scarcity of shares ceases to be a priority. One question remains, for the years when it was one: how many engineering schools, laboratories, pay rises would the same sum have funded?
Reference: abbreviations & acronyms used (EPS, $bn, S&P 500, SEC, capex).