A company can increase earnings per share while making its remaining shareholders worse off.

That is the problem with treating every buyback announcement as good news. The number of shares goes down, earnings are divided among fewer owners, and the result looks better. But the company spent real money to achieve that result. What it paid matters. Buffett made this distinction explicitly in Berkshire’s 2022 letter: repurchases below value benefit continuing shareholders, while overpaying does the opposite.

I like businesses that can grow, maintain a strong balance sheet, and gradually reduce their share count. Owning a larger percentage of a better business over time is an attractive prospect.

But I would rather receive a dividend than watch management repurchase expensive shares just to meet an earnings target.

When I look at capital allocation, I want to understand what the company could have done with the money instead.

The choice does not begin with buybacks versus dividends. It begins with the opportunities inside the business.

In Berkshire’s 2012 letter, Buffett put reinvestment in existing operations first: improving products, expanding the business, increasing efficiency, and strengthening competitive advantages. Returning cash is one option among several, not an obligation to pursue regardless of the alternatives.

Suppose a company has a credible opportunity to invest $100 million and eventually earn an additional $25 million a year after tax. Depending on the risks, timing, and durability of those earnings, I may prefer that investment to either a dividend or a repurchase.

The word credible matters. A management presentation can make almost any project look attractive.

I want evidence that customers need the product, that the company can execute, and that the projected returns survive less favorable assumptions.

I also separate returns on existing capital from returns on new capital. A business may earn exceptional returns from assets and relationships built years ago without having many opportunities to repeat that success.

High historical returns do not give management permission to keep every dollar.

My own cleaning businesses are a small-scale example. I take distributions from them and invest in public markets instead of automatically putting every dollar back into expansion. I want those businesses to generate cash without taking more and more of my time. Keeping the operations healthy comes first, but owning a business is not enough reason to keep giving it more capital.

For the businesses I want to own, the attractive combination is strong returns on additional investment and relatively modest cash requirements. The company can fund worthwhile growth and still have money left over.

That is when the payout decision becomes especially interesting.

Consider a hypothetical company earning $1 billion a year with 100 million shares outstanding. Each share represents $10 of annual earnings.

Now assume total earnings grow 8% annually and the share count falls 3% annually, after accounting for any new shares issued. Assume the company can afford both the investment needed for growth and the repurchases without weakening its finances.

Here is what happens:

Hypothetical calculations using year-end shares to illustrate earning power. Reported earnings per share normally use a weighted-average share count.

Total earnings compound at 8%, but per-share earnings compound at approximately 11.3%. The calculation is 1.08 divided by 0.97, minus one.

Without the share reduction, the same operating earnings would produce about $21.59 per share in year ten, rather than $29.28.

An investor who never sells also owns a larger percentage of the company. A starting 1% interest becomes roughly 1.36%.

That is the appeal of a business steadily buying back its own shares. A patient owner can participate in both operating growth and increasing ownership.

But this is not a complete comparison of investment returns. The company that did not repurchase shares would still have the cash, would have invested it elsewhere, or would have paid dividends. Those dividends could themselves be reinvested.

The chart illustrates the share-count effect. It does not prove that repurchases created value or beat the alternatives.

A second hypothetical example makes the distinction clearer.

Assume a company has an estimated equity value of $10 billion, including $2 billion of genuinely surplus cash. There are 100 million shares, making estimated value $100 per share.

The operating business earns $500 million annually. For simplicity, the surplus cash earns nothing, and we ignore taxes, transaction costs, and changes in the business.

Management spends the entire $2 billion on repurchases.

Original hypothetical illustration. Before the repurchase, annual earnings were $5 per share and estimated value was $100 per share. After buying at $80, $100, or $125, annual earnings per share would be $6.67, $6.25, or $5.95, respectively.

In every case, earnings per share increase.

Yet the outcomes for continuing shareholders are very different.

At $80, the company retires shares for less than their estimated worth. Each remaining share becomes more valuable.

At $100, the transaction is neutral to estimated per-share value under these assumptions. Shareholders own a larger percentage of a company with less cash.

At $125, earnings per share rise by approximately 19%, while estimated value per remaining share falls by almost 5%.

A management team could celebrate the earnings increase while having made an expensive capital-allocation mistake.

Of course, the estimated value is not something I can observe with certainty. The $100 figure in this example is an assumption. In an actual investment, that uncertainty is a reason to demand room for error.

There is an important distinction between investing in operations and repurchasing ownership in those operations.

A company earning 30% on its operating capital does not automatically earn 30% when it buys its shares.

Buying stock at 30 times annual earnings means paying $30 for each $1 of current annual earnings. That is a starting earnings yield of about 3.3%, not 30%.

Future growth can make that purchase attractive. So can durable earnings, modest reinvestment requirements, and a long period of competitive strength. But those expectations must justify the purchase price.

The company’s operating returns and the return available on its shares are related, but they are not interchangeable.

The practical issue is that expensive shares also make each repurchase dollar less effective. At a $100 share price, a $1 billion budget buys 10 million shares. At $200, it buys 5 million.

Buffett explained this advantage for continuing owners in Berkshire’s 2011 letter: when a business keeps repurchasing stock, lower purchase prices allow the same money to retire more shares. He also emphasized that operational funding and financial strength must come first.

I would welcome that opportunity only if the business remained healthy. A falling share price does not help much when the future earnings supporting the valuation are disappearing too.

There is nothing inferior about receiving cash.

A dividend lets shareholders decide where to put the money. They can reinvest in the same company, buy something else, meet spending needs, or hold cash.

That flexibility becomes more valuable when management has few attractive reinvestment opportunities or the shares are expensive.

Return to the hypothetical company worth $100 per share, including $20 of surplus cash.

Instead of repurchasing shares, it could distribute the cash as a $20 dividend. Under the same simplified assumptions, shareholders would then own a business worth $80 per share and hold $20 in cash.

Their combined value would still be $100 before taxes and costs.

The fair-value repurchase also left continuing shareholders with $100 of estimated value per share. It produced higher earnings per share, but that alone did not make it the superior outcome.

The dividend is not free money, either. Cash has moved from the company to its owners.

Taxes can affect the comparison. In a U.S. taxable account, dividends generally create taxable income when received. A continuing shareholder who does not sell into an ordinary open-market repurchase generally avoids a current personal tax event from that repurchase. The eventual tax outcome depends on the investor’s circumstances and subsequent transactions.

My preference would be a sustainable regular dividend alongside flexible repurchases, provided both fit the company’s opportunities and finances. I would not want management preserving either policy at the expense of the business.

Visa’s fiscal 2025 cash-flow statement provides a useful example of a company using both methods:

Source: Visa’s fiscal 2025 annual report. Figures rounded. Cash after investment is operating cash flow less purchases of property, equipment, and technology. Total payouts and the funding gap are calculated from those figures.

The combination is attractive to me: substantial operating cash generation and relatively modest spending on property, equipment, and technology.

But repurchases and dividends totaled approximately $22.95 billion, exceeding operating cash flow after those purchases by about $1.37 billion. Visa also reported borrowing proceeds and other cash movements during the year. The distributions therefore cannot all be described as that year’s leftover operating cash.

This is not, by itself, a criticism. It is a reminder to reconcile where the money came from before praising where it went.

I still want to understand how the payouts fit alongside debt, acquisitions, and other obligations.

Visa repurchased roughly 54 million shares in fiscal 2025 at an average cost of $335.44 per share.

At that price, Visa paid about 29.2 times adjusted earnings of $11.47 per share.

For a simple test, I assume 12% annual earnings-per-share growth for five years and a 25x ending multiple. That would put year-five earnings near $20.21 and the share price around $505. Including dividends, the implied annual return is about 9.4%.

That does not look compelling to me given the uncertainty in those assumptions.

A weaker case, 8% earnings growth and a 20x ending multiple, produces roughly 1% annual returns. A stronger case, 15% growth and a 30x multiple, produces roughly 16%.

My judgment: the repurchases were defensible, but not an obvious bargain. I would prefer a smaller, price-sensitive program over spending a fixed amount regardless of valuation.

I care more about the net change in ownership than the headline repurchase budget.

Suppose a company starts with 100 million shares, repurchases 5 million, and issues 4 million through employee awards or acquisitions. The net reduction is only 1 million shares.

That may still be sensible. Employees need compensation, and acquisitions can create value. But I would not describe the outcome as a 5% increase in continuing shareholders’ ownership. The net share count fell by 1%.

The difference deserves attention over several years, not just one quarter.

I also want to understand management’s incentives. A compensation plan tied heavily to earnings per share may reward reducing the denominator without adequately testing the purchase price. The potential conflict between repurchases and executive incentives has been examined publicly, including in a 2018 speech by securities regulator Robert Jackson.

For me, the useful questions are practical: did the business improve, did my ownership increase, and what did the company spend to achieve both?

A large authorization does not answer any of them.

I am not automatically opposed to debt-funded repurchases. I am opposed to treating them as equivalent to purchases funded with money the company truly does not need.

Debt brings interest expense and repayment obligations. Using existing cash also has a cost: the company gives up whatever income or flexibility that cash could have provided.

Berkshire’s 2011 repurchase discussion explicitly put adequate operating funds and liquidity ahead of buying back stock. That is a useful discipline even for businesses with very predictable revenue.

I would ask how the plan holds up through a period of weaker earnings. Can the company still invest, service its debt, and meet obligations without depending on a favorable refinancing market?

The bad outcome is easy to imagine: a company borrows to retire expensive shares, encounters trouble, and later issues new shares cheaply to repair its finances.

The initial reduction in share count would have concealed a much worse long-term result.

For the companies I own, I would rather tolerate a smaller repurchase program than sacrifice the ability to withstand a difficult period.

The capital-allocation discussion I find useful would answer four questions.

What opportunities inside the business are going unfunded? Before celebrating cash returned, I want to know whether product development, customer service, systems, or necessary capacity have been neglected.

What makes the shares attractive at the purchase price? I do not need management to publish a precise valuation target. I do want evidence that price affects the decision, rather than a commitment to spend the same amount regardless.

What happens to ownership after new shares are issued? The repurchase budget should be considered alongside employee compensation, acquisitions, and other sources of dilution.

How much flexibility remains afterward? A business should be able to stop repurchasing shares when conditions change. I would consider that a sign of discipline when the price or financial circumstances no longer justify buying.

These questions matter more to me than whether management calls a policy shareholder-friendly.

They also apply to dividends. A payout is not attractive simply because it has increased for years. The business still has to support it.

I like the idea of holding a strong business while my ownership gradually increases. That is one reason I pay attention to companies that can reduce their share count consistently.

But the most attractive version is not a company shrinking itself as quickly as possible.

It is a business that continues serving customers well, invests where returns justify it, and buys back shares when the price makes sense. Over time, I can own more of something that is also becoming more valuable.

When the shares are expensive and the company has no better use for the cash, I would rather receive a dividend. When the business has exceptional reinvestment opportunities, I would rather it pursue them. When financial strength is uncertain, I would rather it keep the money or reduce debt.

There is no permanent winner between buybacks and dividends.

What I want is management that recognizes the choice can change.

A falling share count is useful evidence. What matters is whether the value of my remaining shares, together with the cash I receive, is increasing.

Disclosure: I own shares of Visa. This article reflects my personal views and is not investment or tax advice. The numerical examples are hypothetical illustrations, not forecasts. They exclude taxes and transaction costs unless stated otherwise.