Seven Years In, and the Exit Hasn't Come. Now Re-Underwrite the Operator.

A sponsor buys a mid-market industrial services business in 2019. Forty million in revenue, six million of EBITDA, a five-year plan, a CEO hired to execute it. It is now 2026 and the company is still in the fund. The plan expired two years ago. The CEO did not.

That company is not unusual. It is one of roughly 32,000.

Bain & Company released its 17th annual Global Private Equity Report on 23 February 2026, and the headline was genuinely good news: buyout deal value up 44% year over year to $904 billion, exit value up 47% to $717 billion, both the second highest on record. The recovery is real.

Underneath it sit four numbers that describe a different situation entirely.

And one more, which is the one that should change a hiring decision. Bain coined a rule of thumb for it: "12 is the new 5."

During private equity's golden decade, a typical deal needed roughly 5% annual EBITDA growth to deliver a benchmark 2.5x multiple on invested capital over a five-year hold, and about a 20% IRR alongside it. Cheap debt and rising multiples did the rest. In today's environment — borrowing costs of 8% to 9%, leverage ratios of 30% to 40%, purchase multiples in record territory, no multiple expansion to lean on — Bain finds the same deal requires 10% to 12% average annual EBITDA growth for the same return.

Read those two paragraphs as a job description and the problem becomes obvious. The person hired to deliver the first number is being asked to deliver the second one, and nobody has said so out loud.

How the hold actually unfolded

• 2019 · The underwrite Five-year hold. Roughly 5% EBITDA CAGR in the model. Meaningful multiple expansion assumed. The CEO is specified against that: a steady operator who can run a playbook, protect margin, manage a board, and not break what already works. That is the correct spec for the deal as written.
• 2021 · The plan works Valuations peak. The thesis is on track and the leadership team is performing against the number they were given. Nobody has any reason to revisit anything.
• 2022–2024 · The window closes Rates move. Exit routes narrow. Bain records fund-raising declining for a fourth consecutive year — buyout fund-raising down 16% to $395 billion in 2025, with the number of funds closed down 23%. Distributions to LPs stay under 15% of NAV year after year. The asset stays in the fund, not because it is broken but because there is nowhere for it to go.
• 2025 · A narrow recovery Deal value jumps 44% — but the buyout deal count falls 6% to 3,018 transactions, and just 13 megadeals above $10 billion account for 30% of global deal value. Exits total $717 billion across 1,570 transactions, a count 2% lower than the year before. If your asset is not a megadeal, the recovery happened somewhere else.
• 2026 · Year seven Three live options: sell into a thinner sub-$10 billion market, roll the asset into a continuation vehicle, or hold and compound. Two of the three extend the hold. All three require the company to grow earnings at a rate the original model never asked for.

Note what never happened anywhere on that timeline: a moment where anyone formally re-specified the leadership requirement. The hold changed length. The return math changed shape. The job description stayed in a 2019 folder.

Two different jobs wearing the same title

It is tempting to treat 5% and 11% as the same task at different intensity. They are not.

Five percent a year is multiplication. The business model works. The job is to run it well, hold price, keep the good people, add a bolt-on when one appears, and avoid unforced errors. The operator who does this best is often someone who has held a similar seat before and knows exactly where the landmines are.

Ten to twelve percent a year with no multiple expansion is transformation. It means rebuilding pricing architecture, redesigning the commercial motion, integrating acquisitions rather than merely closing them, and taking cost out of a structure that was already run lean. It requires an operator who has done a rebuild before and is comfortable being disliked for six months while it happens.

A CEO can have been exactly the right hire in 2019 and be the wrong fit in 2026 without ever having failed at anything.

That sentence is the reason this is a fit problem and not a performance problem, and the distinction matters enormously to how a sponsor should respond. Firing someone for failing to do a job they were never hired for is expensive, slow, and — in our experience — the single most common way a good asset loses eighteen months. We have written about where that leads in the year-two CEO exit: the pattern is a fit failure, dressed as a talent failure, priced as a search fee.

What the gap is actually worth

Take the company from the opening paragraph and run it out. Entry: $40M revenue, 15% EBITDA margin, so $6M of EBITDA. Bought at 9x, so roughly $54M of enterprise value.

Seven years at the original 5% CAGR: $6M × 1.05⁷ = $8.44M of EBITDA.
Seven years at 11% — the midpoint of Bain's new requirement: $6M × 1.11⁷ = $12.46M.

The difference is $4.0M of annual EBITDA. At a flat 9x exit multiple, that is roughly $36M of enterprise value — about two-thirds of the original purchase price of the entire business.

This is a directional calculation and should be treated as one. It holds entry and exit multiples constant, models no debt paydown, ignores add-on acquisitions, and applies Bain's economy-wide rule of thumb to a single hypothetical asset. Change the multiple assumption and the number moves substantially. What survives every version of the assumptions is the shape of the answer: the spread between the operating profile you underwrote and the one the extended hold now requires is worth more than any search fee, retention package, or advisory engagement you will ever consider. It is frequently worth more than the equity cheque.

The continuation vehicle makes it urgent

Continuation vehicles are no longer a workaround. Bain reports GP-led continuation vehicles growing 62% year over year, on 37% annual growth since 2022. A quarter of respondents to a recent StepStone/Bain survey had initiated or completed one in the past two years, and around 40% planned to explore one in the next 12 to 24 months. More than half said generating liquidity was the primary motive.

A continuation vehicle is a decision point that looks like a formality, and it is where the leadership question gets settled by default.

Here is what it looks like from the operator's chair. They have run the business for seven years. Their management incentive plan was struck against a 2019 model and a 2024 exit. In a long hold, accrued shareholder debt and additional funding rounds routinely push that equity out of the money — Latham & Watkins documents the resulting MIP resets as a standard sponsor problem, precisely because unrewarded management is a retention risk. So the CEO is being asked to commit to years eight through eleven, on a repriced package, against an earnings requirement roughly double the one they signed up for.

Most sponsors handle this as a compensation exercise. Reset the plan, refresh the vesting, move on. That solves the retention question and leaves the fit question completely untouched — which is how a fund ends up three years into a continuation vehicle discovering that the person who is now well-incentivised is well-incentivised to do a job they have never done.

Three tests, run before the vehicle closes

The re-underwrite is not complicated. It is just rarely done, because it feels like disloyalty to a team that has performed.

1. Which job is this now — multiplication or transformation? Answer it explicitly, in writing, with the required EBITDA CAGR in the sentence. If the answer is "some of both," name which one dominates the first eighteen months. Most sponsors have never written this down, which is why the leadership conversation stays vague and personal instead of specific and operational.

2. Has this operator done the dominant job before — not a version of it, the actual thing? A CEO who has taken 400 basis points of margin out of a services business has done it. A CEO who has managed a business that already had good margins has not, and there is no shame in that. This is the same specification discipline that decides whether an external hire works at all: the résumé shows what they have done, but the fit question is whether they have done this, under this kind of pressure, reporting to this kind of board.

3. What does the seat need at month 36 of the extended hold? Not month one. If the vehicle runs four more years, the company at the end of it is materially larger and structurally different from the one being handed over today. Hire — or retain — against that company.

The default answer to a failed test is to replace the CEO, and it is usually the wrong one. Bain's own conclusion is that value creation is now the differentiator, and the reflex to buy leadership rather than build it is already well documented: roughly three-quarters of portfolio leadership roles are filled externally, at a cost almost no entry model prices in. Often the correct move is narrower — add the capability the sitting CEO was never hired to carry, alongside them, for the specific stretch where it is needed. That is what the fractional bridge is for, and it is considerably cheaper than discovering in year nine that you replaced a good operator instead of completing one.

Bain's framing of the new era is that most firms "will have to substantially raise their value creation game," and that winners will build differentiation into "a system, not a slogan." Every sponsor we speak to accepts that at the portfolio level. Very few have translated it into the only lever that actually delivers EBITDA growth, which is the twelve people who run the company on a Tuesday.

Thirty-two thousand companies. Roughly seven-year holds. Almost every one of them is being run by a leadership team hired against an underwrite that no longer exists. The exit will eventually come. The question is whether the person who has to produce the number between now and then was ever asked to.