Note · Opinion, not rated
The price makes the buyback.
Buybacks are praised as returning cash and condemned as financial engineering, usually by people looking at the same repurchase. Both crowds are skipping the only input that decides which one it was.
The claim
A share repurchase is not a return of capital to shareholders. It is a purchase of ownership from the shareholders who leave, made on behalf of the shareholders who stay, using money that belonged to both. Like any purchase, it is judged by one thing: the price paid against the value received. A buyback below intrinsic value creates value for the people who remain. A buyback above it destroys value for them, dollar for dollar. Nothing about the announcement, the size, or the stated intention changes that arithmetic.
One company, one value, three prices
Take a company worth $150 a share by your own honest estimate, with 100 million shares outstanding, so $15,000 million of intrinsic value in total. Management spends $1,000 million on repurchases. The only variable is the price the market happens to be charging.
At $100, below value. The company retires 10.0 million shares. Intrinsic value per remaining share rises to $155.56, a gain of +3.7%. The holders who sold at $100 handed over something worth $150 for $100, and the $500 million they left behind now belongs to everyone who stayed.
At $150, exactly fair value. Per-share value afterward is $150.00. Nothing was created and nothing was destroyed. The transaction was a dividend wearing different clothes: cash left the company, ownership concentrated, and every party got precisely what they gave.
At $200, above value. Per-share value falls to $147.37, a loss of 1.8% for everyone who stayed. The sellers captured $250 million more than they surrendered, paid for by the holders the buyback was supposedly serving.
The bookkeeping closes exactly. At every price, what the remaining holders gain or lose equals what the selling holders lose or gain, to the dollar. A buyback moves value between two groups of the company's own shareholders, and the direction of the move is set entirely by the price.
Whether a buyback returned capital or burned it is not a matter of opinion or intention. It is the sign of one subtraction: intrinsic value minus price paid.
The accretion trap
The standard defense of a repurchase is that it is accretive to earnings per share, and the defense is usually true and almost always irrelevant. In the example above, give the company $600 million of earnings, so $6.00 per share, with the cash earning a small after-tax return while it sits. Now run the buyback at $200, the price we just showed destroys value.
Earnings per share rise to $6.21, an increase of +3.5%, at the very same moment that intrinsic value per share falls 1.8%. The repurchase is accretive and value-destroying simultaneously, because accretion only compares the earnings yield of the stock to the yield on idle cash, and almost every stock clears that bar at almost any price. If accretion is the test, there is no such thing as an overpriced buyback, which should tell you what the test is worth.
The offset that is not a buyback at all
One more repurchase deserves its own category: buying shares to mop up the ones issued as stock compensation. We wrote a full note on why that cost is real, so here it is in one sentence. A repurchase that holds the share count flat against ongoing grants is not capital return at any price. It is compensation expense being settled in cash a quarter late, and the honest measure of what a company returns is the net figure: repurchases minus stock issued, read off a share count that is actually falling.
What we watch in practice
In our own coverage this shows up as two habits. First, the net test: in the Uber initiation we flagged that quarterly repurchases had fallen sharply while a much larger sum went into total return swaps, because capital allocation that is getting more complicated is usually capital allocation that is getting harder to grade. Second, the price test: a management team that repurchases aggressively near highs and goes quiet after drawdowns is running the arithmetic of this note in reverse, and the share count over a full cycle will say so.
What this note is not saying
None of this is an argument against buybacks. Below intrinsic value, with no better use for the cash inside the business, a repurchase is among the best decisions a management team can make: it compounds the ownership of every continuing holder without a tax event, which a dividend cannot do. The argument is narrower and, we think, harder to dispute. A buyback is an investment the company makes on your behalf, and investments do not become good because of what they are called. They become good because of what was paid.