Buybacks Now Ride on Balance-Sheet Room, Not Cash
A widening slice of corporate repurchase programs are being sustained less by surplus recurring cash than by leverage capacity, working-capital timing, or customer float, meaning the same market that grants that room can just as easily take it away.
The comfortable story about buybacks is that they come out of leftover cash: a company earns more than it needs, and hands the rest back. Across this group of ten filers, that story is only sometimes true, and the exceptions are becoming the more interesting case. What is emerging instead is a repurchase program that survives on borrowing capacity, covenant headroom, or the accident of working-capital timing, rather than on cash a company can be sure it will have next year.
Penske Automotive Group (PAG) is the sharpest illustration. Its repurchases nearly doubled even as free cash flow fell and debt more than doubled, and the filings are explicit that capital needs, from acquisitions to inventory financing to debt refinancing, could pull a significant share of cash flow toward debt service before anything reaches shareholders. Nine-month 2025 operating cash flow of $851.9 million was already down from $962.1 million a year earlier, with management blaming working-capital timing in receivables, inventory, and floor-plan notes payable, the same financing lines that fund the vehicle inventory PAG carries. That is not a company with idle cash; it is a company whose repurchase capacity rises and falls with lender availability.
Service Corp International (SCI) sits close behind. Operating cash flow held essentially flat at $942.8 million in 2025 while repurchases jumped 42% to $464.2 million, and the arithmetic does not close cleanly: after $548.3 million of investing outflows for acquisitions, new funeral locations, and cemetery development, only about $394.5 million of cash remained before financing needs, less than the buyback itself. SCI never says it borrowed to fund the repurchase, but it did simultaneously expand its bank credit facility and describes the larger facility as adding flexibility for capital investment and general corporate purposes, all while managing to a 3.5x-to-4.0x leverage target. Repurchases there look sustainable only as long as that leverage room stays open, and management's own language, tying buybacks to market conditions and debt covenants, suggests they would flex before the dividend if credit tightened.
Booking Holdings (BKNG) complicates any tidy verdict. Its $9.0 billion of free cash flow dwarfs its $8.0 billion of buybacks, and $17.8 billion of cash, equivalents, and investments plus a fresh $20 billion repurchase authorization argue for genuine surplus. Yet Booking also raised $3.0 billion through a euro note issuance explicitly earmarked, in part, for share repurchases, and its cash generation leans on $5.3 billion to $6.3 billion of deferred merchant bookings, a float from travelers' advance payments that inflates operating cash flow independent of earnings quality. Debt-assisted and durable are not opposites here; Booking just has enough scale that the debt looks optional rather than load-bearing.
Elsewhere the pattern breaks outright. Choice Hotels International cut its buyback by 66% as free cash flow fell and debt rose, the kind of self-correction that shows repurchases can still act as a release valve rather than a fixed commitment. Burlington Stores and Onespaworld Holdings (OSW) trimmed their programs too, and Six Flags Entertainment's repurchases are too small to matter either way. D.R. Horton (DHI) and Mobileye Global carry no debt in the numbers provided, making their buybacks a different question entirely, one of inventory cycles and equity dilution rather than borrowed capital. Carvana's small repurchase, against $4.8 billion of debt and explicit reliance on capital-market access, is really a debt-management story wearing a buyback's clothes.
The upshot is not that repurchases are broadly debt-funded. It is that a subset of programs, Penske and SCI most clearly, have quietly become claims on future refinancing rather than distributions of cash already earned, and that distinction only becomes visible when credit gets more expensive.