The Company Bought Back Billions. Did You Actually Benefit?

Imagine owning a laundry shop with nine friends.

Ten equal owners. One business. Plenty of socks that somehow return without their partners.

One owner wants out. The shop uses its cash to buy back his stake.

“Good news,” the manager announces. “There are only nine of you now. Your ownership has increased.”

Everyone celebrates.

Then someone asks two annoying questions.

“How much did we pay him?”

“And didn't we just promise the new manager an ownership stake?”

Suddenly, the celebration needs a spreadsheet.

Stock buybacks deserve the same questions. A company can spend billions repurchasing shares while barely reducing its share count. It can also reduce shares, increase earnings per share and still make continuing owners poorer.

The useful question is what remains for each share after the money has left.

Follow both doors, not just the exit

A buyback uses company resources to acquire its own shares. If outstanding shares fall and you keep yours, your percentage ownership rises.

But shares can enter through another door: employee compensation, option exercises, acquisitions or other issuance.

Consider an illustrative company starting with 100 million shares. It repurchases five million and issues four million during the same period. It ends with 99 million.

The buyback removed 5% of the starting shares. The net reduction was only 1%.

Both statements are true. They tell different stories.

If you owned one million shares throughout, your stake rose from 1% to roughly 1.01%, not to the 1.053% you would own if five million disappeared with no new issuance.

This does not mean the buyback was useless. Without it, the same issuance would have taken the count to 104 million and diluted your percentage.

Preventing dilution can matter. It just isn't the same achievement as making each remaining owner substantially larger.

Back at the laundry, buying out one partner while admitting another may preserve everyone's percentage. The outgoing partner still took real cash home.

A $12 billion example hiding in plain sight

Microsoft's fiscal 2024 annual report provides a useful historical example, not a claim about its current valuation.

For the year ended June 30, 2024, its repurchase-program table shows 32 million shares bought for $11.96 billion. Yet its share-count reconciliation shows 7,432 million shares at the start, 34 million issued and 32 million repurchased, leaving 7,434 million at year-end.

A large buyback programme. A slightly larger ending share count.

Scope matters: that programme table excludes $5.3 billion of withholding-related repurchases. Mixing its $12 billion with a broader cash-flow number would misstate the comparison. Microsoft annual report, Note 16.

This is not evidence that Microsoft wasted the money or that its shares were unattractive. It demonstrates why the headline expenditure cannot answer the ownership question by itself.

You need the reconciliation.

“Non-cash” does not mean nobody pays

Giving employees shares can be sensible. Skilled people create products, win customers and build businesses. Ownership can help retain them.

The mistake is treating that compensation as free because no salary payment left the bank account on the grant date.

Someone receives a claim on the business. Existing owners share future outcomes with them. If the company later buys shares to offset that dilution, it spends cash to keep the ownership base from expanding.

Think of the laundry's new manager.

If she improves operations enough to create more value than her compensation costs, everyone can benefit. If she receives a stake and contributes very little, calling her pay “non-cash” does not improve the bargain.

There are three separate measurements here: the accounting expense for stock compensation, the shares actually issued and the cash required to offset dilution. Different dates, prices and award terms mean those numbers need not match.

Avoid a shortcut in either direction. Don't ignore compensation economics. Don't subtract every related number again from earnings and accidentally count the same cost twice.

A useful owner-level test is whether business value per share is growing after the full cost of employing people and maintaining the operation. A flat share count alone cannot settle that question.

EPS can improve while the deal gets worse

Here is a deliberately simplified example. All numbers are illustrative, not a valuation of a real company.

A business has 100 shares and estimated total equity value of $1,000, including its cash. It earns $100 annually. Assume the cash earns no interest and the buyback changes neither operating profit nor taxes.

That gives estimated value of $10 per share and earnings per share of $1.

Management repurchases ten shares at $15 each.

It spends $150. Estimated equity value falls to $850. Ninety shares remain.

For a full year after the repurchase, EPS would be about $1.11 because the same $100 profit is divided among fewer shares.

But estimated value per remaining share falls to about $9.44: $850 divided by 90.

The earnings-per-share slide looks better. The continuing owners' bargain got worse because management overpaid.

Now change the purchase price to $8. Buying ten shares costs $80, leaving $920 of estimated value across 90 shares, or about $10.22 each.

Same number of shares removed. Same simplified EPS improvement. Opposite economic result.

Warren Buffett made this price-dependent argument in his 2016 shareholder letter. The difficult part in practice is estimating business value. It isn't printed on the share certificate.

Real-world calculations also need lost interest income, financing costs, taxes and timing. Our example isolates one mechanism; it does not predict a stock's next move.

The denominator has its own calendar

There is another trap: treating every “share count” as the same number.

Period-end outstanding shares describe a particular date. The weighted-average count used for EPS reflects how long shares were outstanding during the reporting period. Diluted EPS also incorporates relevant potential shares under the accounting rules.

Microsoft illustrates the distinction: its fiscal 2024 diluted weighted-average count fell from 7,472 million to 7,469 million even though its period-end outstanding count edged up. Annual report, Note 2.

Neither number is necessarily wrong. They answer different questions.

For ownership change, compare consistent period-end figures and investigate issuance. For EPS growth, inspect the weighted-average denominator alongside the change in total earnings.

Otherwise, a December buyback can look unimpressive in annual EPS even though it changes the year-end share base. Timing becomes mistaken for effectiveness.

Also adjust comparisons for splits. Five shares becoming twenty through a split is a change of units, not evidence of dilution or failed buybacks.

A permission slip is not a receipt

“Board approves a $10 billion buyback” sounds like $10 billion has already entered the market.

An authorization establishes capacity. Actual purchases establish execution.

Salesforce's fiscal 2024 report, for example, explicitly states that its authorization does not oblige it to purchase a specific number of shares. Annual report, Note 11.

Read the wording. Is the figure a new authorization, an increase to an existing one, the amount remaining or completed purchases?

Then ask about pace and price. A programme spread over years has different near-term implications from purchases already completed last quarter.

Announcements can plausibly communicate confidence or flexibility. They can also attract attention without requiring immediate expenditure. You cannot establish management's motive from the headline alone.

Likewise, a share-price rise after an announcement does not isolate the buyback's effect. Earnings, guidance and market conditions may have changed at the same time.

A laundry manager saying “we may buy out partners” hasn't handed anyone a cheque yet.

The overlooked cost: what the money cannot do next

Suppose the laundry buys out a partner just before three washing machines need replacing.

The ownership percentage improved. The business now struggles to serve customers.

A repurchase competes with other uses of cash: maintenance, expansion, debt repayment, acquisitions, dividends and reserves for difficult periods.

Buffett's 2016 discussion also notes that internal funding needs or a superior investment opportunity can take priority over repurchasing undervalued shares.

This is why “more buybacks” cannot be a universal quality score.

A mature business with limited attractive expansion opportunities may sensibly return surplus capital. A business with valuable projects may sensibly retain it. Either can make poor decisions.

Funding matters too. Borrowing to repurchase shares replaces some equity funding with debt obligations. Whether that is sensible depends on the price paid, financing cost, business resilience and remaining flexibility.

The incentive deserves a look. If executives are rewarded against EPS, a smaller denominator can help that metric. That is a reason to read the compensation framework, not proof of manipulation.

Prefer evidence over a villain story. Ask whether the company explains its price discipline and competing priorities, rather than assuming every buyback is either generous or dishonest.

Read a buyback in five questions

Before celebrating the next large announcement, work through this short sequence.

1. What happened, rather than what was permitted? Separate authorization, remaining capacity and actual repurchases. Use the same reporting period throughout.

2. How much did the ownership base change? Reconcile beginning shares, issuance, repurchases and ending shares. Check splits and acquisition-related issuance before drawing conclusions.

3. What drove EPS growth? Compare total earnings growth with the weighted-average diluted share count. Improving operations and a shrinking denominator are different contributors.

4. Was the price sensible? Examine average repurchase prices against a reasonable valuation range, not merely today's price or the previous all-time high. State the assumptions that would make the purchase good or bad.

5. What was sacrificed? Inspect cash generation, investment needs, debt and liquidity. A larger percentage of a weaker business may be a poor exchange.

For a short-term trader, this analysis provides context, not an entry trigger. A buyback does not guarantee a price floor. Keep the actual chart, invalidation and risk budget separate from the corporate story.

For a longer-term owner, the question is whether management is improving what each remaining share can reasonably claim.

Back at the laundry, count the owners. Open the cash box. Check the machines.

Then decide whether to celebrate.

A smaller ownership register can be good news. A better business bought at a sensible price is better news.