Share buybacks generate more heated commentary than almost any other corporate action, and most of it treats the mechanism as either obviously good or obviously sinister. The mechanics are more mundane.
What actually happens
A company uses cash to purchase its own shares in the market. Those shares are either cancelled or held as treasury stock.
The company's cash falls. The number of shares outstanding falls.
Each remaining share therefore represents a larger fraction of a slightly smaller company.
That is the whole operation.
Why it is compared to a dividend
Both return cash to shareholders. The difference is in who receives it and how it is taxed.
A dividend pays every shareholder, whether they want the cash or not, and is generally taxable on receipt.
A buyback pays only shareholders who choose to sell, and holders who do not sell receive nothing immediately — their stake simply becomes proportionally larger.
Which defers the tax event to whenever they eventually sell, and gives them the choice of when.
In jurisdictions where capital gains are taxed more favourably than dividends, this is a real advantage, and it is the honest reason buybacks grew relative to dividends.
The earnings per share effect
Here is where the criticism concentrates.
Fewer shares means the same profit divided among fewer of them, so earnings per share rises.
If executive pay is linked to earnings per share, there is an obvious incentive to buy back shares regardless of whether it is the best use of the money.
That criticism is legitimate and the response has been for remuneration committees to adjust for buybacks in the metrics, which some do and some do not.
The deeper point is that a buyback does not create value by raising earnings per share. It raises the figure by shrinking the denominator, and the company is smaller afterwards.
When it does create value
The genuine case is straightforward and demanding.
If a company's shares trade below what they are worth, buying them transfers value from selling shareholders to remaining ones.
Which requires the company's own assessment of its value to be correct, and management's record at judging this is, in aggregate, poor.
Buyback volumes are consistently highest when prices are high and cash is plentiful, and lowest during downturns when prices are low, which is the opposite of what the value argument requires.
The flexibility argument
A more defensible case, and less discussed.
Dividends carry an implicit commitment. Cutting one is read as distress and is punished.
A buyback programme can be paused without the same signalling cost.
Which makes buybacks a suitable vehicle for distributing cash flow that management is not confident will persist, and dividends suitable for the portion that will.
Companies that understand this distinction manage distributions considerably better than those that do not.
The offsetting issuance
A detail that changes the picture at many companies.
Share-based compensation issues new shares to employees, which dilutes existing holders.
Many buyback programmes exist substantially to offset that dilution rather than to reduce the share count.
Which means the headline buyback figure overstates the cash returned to outside shareholders, sometimes by a great deal.
The net change in shares outstanding, which is disclosed, is the number that tells you what actually happened.
The underinvestment claim
The most serious criticism is that cash spent on buybacks is cash not spent on research, capacity or wages.
Which is true as arithmetic and does not establish that the investment would have been worthwhile.
A company with no attractive projects that returns cash is behaving correctly. A company that forgoes attractive projects to support its share price is not.
Distinguishing them from outside requires judgement about the investment opportunities available, which is exactly the hardest thing to assess.
How to read one
Look at the net share count over several years rather than the announced programme size.
Look at whether it was funded from free cash flow or from borrowing.
And look at what happened to investment spending over the same period.
Those three together tell you considerably more than any announcement does.
This describes a mechanism and is not advice regarding any company or investment.
How they are executed
The mechanics affect what an announcement actually means.
Open market purchases are conducted gradually over months, subject to daily volume limits and blackout periods around results, which exist to prevent the company from dominating trading in its own shares.
Accelerated arrangements deliver a large block immediately through a bank, which then buys back over time and bears the price risk.
Tender offers invite shareholders to sell at a specified price, generally above the market, within a window.
Each has different signalling content. An open market authorisation is permission rather than commitment, and companies frequently announce far more than they execute.
Taxation of the mechanism
Several jurisdictions have introduced or considered taxes on buybacks, generally at a modest rate on the value repurchased.
The stated aim is to reduce the tax advantage relative to dividends and to discourage the practice.
Evidence on whether modest rates change behaviour is limited, and the effect appears to be small relative to the underlying drivers.