The Quiet Squeeze

Why Adjusted EBITDA keeps growing, who has to deliver the difference, and what happens when they can't.

Imagine you are the CEO of a mid-sized company that was just bought by a private equity firm. You did not do the buying. The fund did the buying. You inherited a chair, a board, a budget, and a thing called the value-creation plan. The value-creation plan is a slide deck. Page one says the company earned $70 million last year. Page two says it will earn $110 million next year. You will be measured against page two.

Page two is partly real. Page two is also, in part, a number someone made up.

The made-up part has a name. It is called the addback schedule. It is the list of items the fund and the lender agreed to treat as if they had not happened, will not recur, or will be more than offset by savings the company has not yet found. Five years ago, more than half of the addback schedule was made up of things that had already happened — transaction fees that had actually been paid, restructuring charges that had actually been incurred. Today, more than half of it is made up of things that have not yet happened. The schedule is the same size. What is inside it has changed.

This post is about why that change occurred, and about whose job it is to make page two come true.

So it goes.

The trend

S&P Global has tracked EBITDA addbacks since 2018. Their dataset covers about 700 leveraged buyouts and large acquisitions. The headline finding people usually quote is that addbacks make up roughly 29% of management-adjusted EBITDA. That number has been remarkably stable. It was 27% in 2018. It peaked at 32% in 2021. It is 26% now. The story most coverage tells is that the total addback stack grew and is now easing.

The interesting story is inside the stack.

Addbacks come in six broad categories. Three of them are defensible. Transaction fees actually happened — they show up on real invoices. Pro-forma acquisition earnings come from the income statements of real acquired businesses. Restructuring charges relate to real money the company spent on something it has now stopped doing. None of these will repeat next year, so removing them from EBITDA gives you a cleaner picture of steady-state earnings. This is what addbacks were originally for.

The other three categories are speculative in increasing degrees. “Non-recurring items” covers everything from pandemic disruption to one-off litigation — usually backward-looking, but the “non-recurring” label is a forecast in disguise. “Expected synergies and cost savings” are forward-looking by construction. These are dollars the company has not yet saved. The biggest category in S&P’s bucket list is the one called “other adjustments.” Other adjustments are what you put in when you cannot come up with a better category. It is a residual. It has no owner.

In 2018, defensible items made up about 58% of the typical addback stack. Speculative items made up the remaining 42%. By 2022, those percentages had reversed. Speculative addbacks were 51% of the stack. Defensible ones were 49%. The categories driving the reversal were “expected synergies” — sponsors promising more future cost-outs than ever — and “other adjustments,” which nearly tripled as a share of the stack between 2018 and 2022, briefly becoming the single largest addback category.

The total stack was about the same size. The portion of it that an auditor could sign off on shrank. The portion that someone, somewhere, still had to deliver grew.

It is worth asking why this happened.

The cause

The simplest answer is that the math of a leveraged buyout changed. Bain & Company laid this out in their 2026 Global Private Equity Report.

In 2015, a typical deal entered at roughly 10× EBITDA, financed with about 50% debt at a 6% interest rate. Exit assumptions allowed for meaningful multiple expansion — say, 12.5× — because asset prices were still climbing. A sponsor targeting a 2.5× return on the fund’s money over five years needed the underlying company to grow EBITDA at about 5% annually. Five percent is a number a competent operator can deliver from organic growth and modest operational improvement. The addback schedule could afford to be honest and small, because it did not have to do much work.

Now it is 2025. You buy the company at 14× earnings. You borrow 36% of the price at 8.5%. Your exit assumption is 15× — barely more than you paid, because the market is not what it was. Same investors. Same five-year hold. Same return target.

The math now requires the company to grow its earnings by about 12% per year.

This is a different kind of company. A company that grows earnings 12% annually for five consecutive years is rare. A mature private business in year three of a hold period, in a portfolio of companies that have already had most of the easy cost takeout done by the previous owner, is not that company. It cannot grow 12% organically. So 12% has to come from somewhere else.

It comes from the addback schedule. And because the defensible categories of addback — transaction fees, real restructuring, pro-forma acquisitions — are bounded by reality, the only way to make the schedule bigger is to lean harder on the speculative categories. Run-rate synergies that have not yet been realized. Cost programs that are in flight but not yet delivering. The “other” bucket that does not require a name.

This is not malice. This is arithmetic.

Time moves on

The math above assumed a five-year hold. It is now closer to seven. The average private equity portfolio company is being held for 6.7 years, against a 20-year historical average of 5.7. There are about 28,000 PE-backed companies waiting to be sold globally. They represent roughly $3.6 trillion in unrealized value. The companies sit in the funds. The funds sit on the companies. The LPs sit on the funds. Nobody moves.

A longer hold makes the composition problem worse. Every additional year is another chance for a run-rate synergy promised at deal inception to not have materialized yet. When it does not materialize, it gets rolled forward into the next year’s plan. Or it gets quietly migrated into the “other adjustments” bucket. The defensible portion of the stack shrinks naturally as time passes — transaction fees age out, pro-forma acquisition earnings get absorbed into the base. The speculative portion grows by default.

And so it does.

The receipts are starting to land

S&P stops at the 2023 cohort because the firm needs two years of post-deal data to evaluate whether the marketing case held up. The 2024 vintage hasn’t been studied yet. But Lincoln International — which values about 5,000 sponsor-owned U.S. private companies every quarter and now has the most comprehensive operating dataset in private markets — has been tracking what happens to the addback promises mid-hold.

The picture is unflattering. Across every deal vintage going back to 2019, average leverage has increased by roughly half a turn since deal close. The 2019 cohort drifted the most — they’ve had the longest to fail. The 2024 cohort drifted the least — they’ve barely had time to start. Every cohort in between went in the wrong direction. This is the exact opposite of the underwriting case for every leveraged buyout ever done: borrow at X turns, pay down debt with cash flow, exit at X minus something. Instead, sponsors are entering at X turns and arriving at X plus 0.5 turns several years later.

Lincoln’s own attribution: the drift reflects “EBITDA failing to translate into cash flow as a result of the unrealized pro forma adjustments and increased interest expense.” In their most recent update they were more specific: “lower realization of synergies and limited free cash flow generation due to years of high rates.”

Translation: the synergies didn’t materialize. The cash flow didn’t show up. The leverage went the wrong way.

This is the operating-side confirmation of the composition-side trend. The speculative addbacks that grew between 2018 and 2022 — the run-rate synergies, the unallocated “other adjustments” — are now being measured against reality. Reality is winning.

So it goes.

Whose job is this, actually

This is the part the marketing decks never address.

The sponsor underwrites the deal. The sponsor negotiates the addbacks with the lender. The sponsor sells the marketing EBITDA to LPs. The sponsor takes the management fees. None of this is the same as actually making the company earn the number on page two of the value-creation plan.

That part is your job. You are the CEO. You inherited the chair.

The plan says your company will earn $110 million next year. About $70 million of that comes from things the income statement already produced last year. About $40 million of it comes from things the income statement has not yet produced. Of that $40 million, half or more is now made up of items that nobody has actually paid for, restructured, or acquired. These items live in the “expected synergies” line and the “other adjustments” line. They are the line items with the least definition and the highest growth.

Your bonus is struck against the budget. The budget is built off page two of the value-creation plan. Page two includes the addbacks. So your bonus is contingent on delivering items that the underwriting case did not specify, against headwinds the underwriting case did not anticipate, over a hold period that keeps quietly getting longer, with the share of speculative content in your target growing every year.

When you miss, one of four things happens. The addback rolls forward. The budget gets quietly reset. You get replaced. Or all three.

The sponsor’s clock keeps ticking.

This is the quiet squeeze. It does not show up in the marketing materials, because the marketing materials are the source of the problem. It does not show up in the credit agreement, because the lender’s covenant tests against a more disciplined version of the same number. It shows up in operator turnover, mid-hold leadership changes, and the occasional restructuring that surprises everyone except the people who were already running the company.

So it goes.

What this means

The story of the last five years is not that Adjusted EBITDA grew. It is that Adjusted EBITDA stayed roughly the same size while becoming structurally less defensible. The portion of the headline number that the income statement has already produced got smaller. The portion that someone still has to deliver got bigger.

This does not reverse without one of three things. A meaningful decline in the cost of debt. A recovery in exit multiples that lets the market do some of the work again. A return to higher leverage that reduces the equity check sponsors have to underwrite.

None of those is imminent. Rates are coming down slowly. Exits remain narrow. Lenders have not rediscovered their old enthusiasm.

In the meantime, the speculative share of the addback schedule keeps doing the work the underwriting case requires. And the Portco CEO keeps being asked to deliver it.