Every exit deck ends on the same page: operating margin, up several hundred basis points across the hold. The room reads it as proof — value created, earned, done. It is neither. A growing share of that margin is not created; it is borrowed from the next owner's income statement, and due back the moment they own the asset. Private equity has a discipline for exactly this problem one line up, on earnings. It has none for margin. That absence is what this paper prices.
The industry's own return data explains why no one built one. For buyouts entered after 2010 and exited before the recent repricing, roughly three-fifths to two-thirds of value creation came from leverage and multiple expansion, not operating skill — McKinsey's and Bain's return-decomposition work anchors that range, with Apollo and PwC corroborating the direction independently. Margin was a supporting actor. With that engine gone, the arithmetic of a target return now demands roughly double the annual EBITDA growth it once did (Bain) — a multiplier, not an absolute; the starting growth rate varies by strategy and vintage, but the doubling holds directionally across Bain's samples. And the trajectory data is specific, not anecdotal: margins run flat, often drifting down, through most of the hold, then spike sharply in the final six quarters before a sale (MSCI) — a pattern intensifying, not fading, in the vintages now reaching market.
The old assumption made sense on those terms. A point of margin realized during a hold was treated as durable value the next owner would pay for at face value. When leverage and multiple expansion carried the return, no one needed to interrogate a margin point's timing or durability, because margin was not what the deal depended on.
That has reversed. Operating value creation is now the primary source of return, not the residual. Whatever rates do next, margin has moved from supporting actor to protagonist — and an old, tolerated imprecision about how a margin point was produced has become expensive.
Some late-hold margin is real. But margin also rises when spending stops — deferred capex, thinned support, postponed reinvestment — and on the page, earned margin and borrowed margin are indistinguishable. The difference is not cosmetic. A point produced by deferring spend is cash borrowed from the next owner, not free cash flow created. The buyer capitalizes only what they believe transfers, and where they detect the difference, conviction in the entire base weakens, compressing the multiple past the single point. The gap between mark and clearing price widens the bid-ask spread, and DPI takes the hit twice — on price, and on time. This is already visible at portfolio scale: a rising share of leveraged borrowers cannot convert reported EBITDA into free cash flow (Fitch, corroborated independently by S&P Global).
The board reviews the level of margin. Management manages to it. The buyer underwrites what survives them — and only the third is transferable value. An experienced Operating Partner doesn't ask how high margin is; the deck already answers that. The question is when each point arrived and what moved the other way to produce it. That question is what separates a business that was improved from a number that was prepared for sale, and it is rarely the question a board, a management team, or an investment committee asks on its own.
The distinction survives because no seat owns it. The operating team has the information — when each point arrived, what it cost — but that knowledge doesn't have to travel into the deck. The deal team has authority over how the number is presented, without having built it. Incentive, at the moment of sale, rewards a clean figure, not a footnote on timing. Accountability for whether the margin survives the handoff sits nowhere, until a buyer's diligence tests it, by which point it is too late to build the case. The single most economically important number in the transaction currently has no owner — which is why the only question worth asking on a near-exit asset is not how much margin improved, but would the next owner underwrite the last six quarters?
The Transferability Test turns that question into four. Timing: was it realized steadily through the hold, or in the final six quarters? Trade-off: what moved the other way when it arrived — capex, R&D, maintenance, headcount? Restoration: does the next owner have to restore anything to keep it — if yes, it was borrowed, not earned. Underwrite: would a buyer capitalize it at the full multiple, or add it back? A point that clears all four is Transferable EBITDA. A point that fails any one is Borrowed Margin — a claim carried at the risk of a discount reaching past the point itself.
The fair objection is that this is not evidence of manufactured numbers, only of timing concentration — a narrower, more defensible claim. That is correct, and it is deliberate. This paper does not claim sponsors are dishonest; most late-hold margin gains are genuine attempts to finish what the plan started. The claim is narrower and more uncomfortable: the industry currently has no mechanism to tell the honest case from the borrowed one, which means the honest operator pays the same discount as the one who sprinted. A reasonable skeptic would also ask why no industry-wide percentage is offered for how much reported margin fails the test. None exists yet. An audited measure of margin durability doesn't exist the way quality of earnings exists for the earnings line — and that absence is the finding, not a gap to paper over with an estimate.
