What a Stock Buyback Actually Does, and Who It Helps

There are two lazy takes on stock buybacks, and they're both wrong. The first is that buybacks are financial engineering, a trick executives use to juice their own pay while the real business goes nowhere. The second is that buybacks are always great, a company “returning cash to shareholders” and therefore a green flag you should buy. Neither holds up, because a buyback isn't good or bad on its own. It's a decision about what to do with a pile of cash, and like every capital-allocation decision, it's only as good as the price paid and the alternative passed up.

The frustrating part is that almost every headline treats the announcement as the event. “Apple authorizes $100 billion buyback.” That number is a ceiling, not a purchase. It tells you what the board allowed, not what got bought, at what price, or whether it beat just holding the cash. The execution is the story. The announcement is the press release.

The mechanic that makes it work, and makes it suspect

Here's the whole thing in one idea. A company's earnings per share is net income divided by shares outstanding. A buyback doesn't touch the top of that fraction. The business earns what it earns. It shrinks the bottom. Fewer shares, same profit, so the number you get when you divide goes up. EPS rises even if the company sold one more widget than last year, or fewer.^[3]

Run the numbers and it's obvious:

Net income (profit):        $1,000,000,000   (unchanged)
Shares outstanding before:     500,000,000
EPS before:  $1,000,000,000 / 500,000,000  = $2.00

Company spends $2B buying back 40,000,000 shares
(at $50/share) and retires them.

Net income (profit):        $1,000,000,000   (still unchanged)
Shares outstanding after:      460,000,000
EPS after:   $1,000,000,000 / 460,000,000  = $2.17

EPS up ~8.7%. The business did not grow. The denominator shrank.

Now you can see why people are suspicious. If a chunk of an executive's bonus is tied to hitting an EPS target, and EPS can be lifted without growing the business, the incentive practically writes itself. Spend cash, shrink the count, clear the hurdle, collect. That happens. It's a real conflict and worth watching for.

But notice what the math also says. The buyback only created value for the remaining shareholders if those shares were worth more than $50. Buy a dollar of value for fifty cents and everyone left holding stock is richer. Pay a dollar for a dollar-fifty and you just lit money on fire to make a per-share number look better. The mechanic is neutral. Price is what flips it from good to bad.

The legit case: it's a flexible, tax-friendly dividend

Strip away the suspicion and a buyback is just one of two ways a company hands cash back to owners. The other is a dividend. Both move money from the corporate balance sheet to shareholders. They just do it differently, and the differences actually matter.

Tax is the big one. A dividend is a taxable event the moment it lands. You owe tax on it this year whether you wanted the cash or not. A buyback creates no taxable event for the shareholders who keep their shares. Your slice of the company quietly grew, and you owe nothing until you choose to sell. You control the timing. That's a genuine edge, especially for long-term holders who'd rather compound than get handed cash and a tax bill every quarter.^[6]

Flexibility is the other one. A dividend is close to a promise. Once a company starts paying one and raising it, cutting it is read as a distress signal, so boards defend the dividend even when cash gets tight. A buyback carries no such expectation. A board can authorize a big program, then buy aggressively in a good quarter and barely at all in a bad one, with nobody screaming. That optionality is valuable. It lets a company return cash without locking itself into a payment it might regret.

And when the stock is genuinely cheap, a buyback is one of the best uses of cash a company has. Warren Buffett has made this point for years. Repurchases “make sense for the continuing shareholders only if the shares are bought at a price below intrinsic value,” he wrote. Do that, and “the continuing shareholders are reaping a benefit as value flows to them from the departing shareholders.”^[5] A company that knows its own worth and buys below it is doing exactly what a good investor would do. That is not financial engineering. That is buying low.

The critique that actually lands: terrible timing

Here's where buybacks earn their bad reputation, and it's not the conspiracy people reach for. It's something more mundane and more damning. Companies are awful at timing. They buy the most when they have the most cash and the stock is high, which tends to be near the top of a cycle. Then a downturn hits, the stock gets cheap, cash gets scarce, and they stop, exactly when buying would actually pay off. They buy high and sit out the lows. The mirror image of what you'd want.

The clearest proof is the cliff in 2020. In the run-up, with markets near records, S&P 500 companies were repurchasing more than $700 billion of their own stock a year.^[4]When the pandemic crash arrived and shares were on sale, buybacks collapsed. Companies hoarded cash to survive, some after years of buying their stock far higher. Banks suspended programs. The pattern was “buy when we're flush and confident, freeze when we're scared,” which is precisely backwards for anyone trying to buy value.

Two more failure modes deserve a name, because they get dressed up as shareholder-friendly when they're nothing of the sort.

Debt-funded buybacks.Borrowing money to buy back stock isn't returning excess cash. It's swapping equity for leverage and calling it a gift. It can be defensible when rates are low and the stock is cheap, but it raises the company's risk, and a more fragile balance sheet is the bill that comes due in the next downturn. A buyback funded by profits and a buyback funded by a bond issue are not the same act, even though the EPS line moves the same way.

Buybacks that just offset dilution.Companies that pay employees in stock are constantly printing new shares. Stock-based compensation quietly grows the share count every quarter. A lot of repurchasing exists only to soak up those new shares and keep the count flat. That's a real cost of doing business, but it isn't shrinking your slice of the company. It's running up a down escalator. If a company buys back $5 billion and issues $5 billion in new stock to employees, the headline buyback did nothing for you, even though it got announced like a win.

Why buybacks won, and the 1% tax that followed

Buybacks weren't always the default. For a long stretch of the twentieth century companies leaned on dividends, partly because heavy open-market repurchasing looked uncomfortably close to a company manipulating its own stock price. That changed in 1982, when the SEC adopted Rule 10b-18. The rule created a “safe harbor.” Stay inside its conditions on timing, price, volume, and which broker you use, and a company is shielded from manipulation liability for its repurchases.^[1] It took the legal cloud off buybacks, and they took off.

The scale today is enormous. In recent years S&P 500 buybacks have run in the neighborhood of $800 billion to $1 trillion a year, often outpacing the dividends those same companies pay.^[4] For a big chunk of corporate America, the buyback is the primary way cash gets returned, with the dividend as the steady supporting act.

That scale is exactly why buybacks became a political target. Starting in 2023, the Inflation Reduction Act imposed a 1% excise tax on the net value of shares a public company buys back, net meaning repurchases offset by new shares issued.^[2] The framing from supporters was that buybacks enrich shareholders and executives while doing nothing for workers, so a small tax both raises revenue and nudges companies toward investment instead. Critics countered that 1% is too small to change behavior and just skims a tax off returning capital that was already taxed as profit.

My honest read: 1% is a rounding error next to the real driver, which is price. No CFO worth the title is going to skip a buyback that's clearly accretive because of a 1% toll. What the tax did accomplish is rhetorical. It cemented “buyback” as a loaded word in political language, the shorthand for companies favoring Wall Street over Main Street. That framing is too blunt. It treats a financing decision as a moral category. A buyback below intrinsic value helps every long-term owner, including the pension fund holding the stock on behalf of those exact workers.

How to actually read a buyback

Forget the announcement. A board authorizing $50 billion is a permission slip, not a deadline. Programs routinely run for years or never get fully used. Here's what I look at instead.

1. Shares outstanding over time.This is the whole game. Pull up the diluted share count over five or ten years. If it's falling steadily, the buybacks are real and your slice is genuinely growing. If it's flat or rising despite years of “record buybacks,” the repurchases are treading water against stock-based comp.
2. The price they paid versus what it's worth.You want a company that buys more when its stock is cheap and eases off when it's expensive. The opposite, piling in at the highs and freezing at the lows, is the tell of a board that treats buybacks as a cash-disposal habit rather than an investment decision.
3. Where the money came from.Funded by free cash flow is a company returning surplus. Funded by new debt is a company changing its risk profile. Both can be fine, but they're different decisions and you should know which one you're applauding.
4. What got passed up.Every dollar spent buying stock is a dollar not spent on R&D, a factory, an acquisition, or paying down debt. Sometimes the buyback really is the best available option, when a company is mature and out of high-return projects. Sometimes it's a company out of ideas papering over stalled growth.

Put it together and the verdict on any given buyback is never automatic. A mature, cash-rich company with a cheap stock and a shrinking share count is doing right by you. A leveraged company buying at the highs to clear an EPS bonus target and offset its own dilution is doing right by its executives. Same press release, opposite reality. The only way to tell them apart is to ignore the announcement and read the execution, which is the part nobody puts in a headline.

So no, buybacks aren't financial engineering, and they aren't an automatic green flag. They're a capital-allocation decision, the same kind a good investor makes every day. Buy low, and you win. Buy high to flatter a number, and you don't. The mechanic never changes. The price is the whole story.