US Private Equity Report 2017 — The Dry Powder Problem

US private equity in 2017 faced a problem that sounds like a good one: too much money.

Committed but uninvested capital — dry powder — had accumulated to record levels. Fundraising had been strong for several consecutive years, driven substantially by the institutional search for yield described in the 2016 global report: with safe assets yielding little, institutions increased allocations to private markets in pursuit of return they could not obtain elsewhere.

Capital arrived faster than assets to deploy it into. The number of companies of appropriate size, in appropriate sectors, whose owners wished to sell, did not grow at the rate that commitments did.

When capital grows faster than the asset supply, the difference appears in price. Entry multiples rose. And a higher entry multiple has direct, arithmetic consequences for return that cannot be argued away.

The industry's response was largely rational and is worth studying, because it reveals what the discipline actually consists of. Firms shifted toward add-on acquisitions, buying smaller companies at lower multiples to combine with existing platforms. They extended holding periods. They emphasised operational improvement over financial engineering. They moved into sectors and company sizes where competition was less intense.

Alongside this, the credit market financing these transactions changed in a way that mattered more than was recognised at the time. Covenant protections weakened substantially. A loan without maintenance covenants gives the borrower far more room before a lender can intervene — which reduces defaults in the short term and, plausibly, worsens recoveries when defaults eventually occur. That trade had not been tested by 2017 and, in significant respects, still has not been.

The arithmetic of a higher entry multiple

This deserves precise treatment because it is the constraint everything else in 2017 was a response to.

A private equity return can be decomposed into three sources:

• Earnings growth — the business generates more profit than when it was bought.
• Multiple expansion — the business is sold at a higher multiple of earnings than it was bought at.
• Leverage — debt amplifies the equity return, because the debt is repaid at face value while the equity captures the residual.

A higher entry multiple affects all three adversely:

• Multiple expansion becomes unlikely. Buying at a high multiple means the exit multiple must be higher still for this source to contribute. At elevated entry levels, the realistic expectation is flat to lower — a headwind rather than a contributor.
• Leverage contributes less. Lenders size debt against earnings, not against price. If the purchase price rises while earnings do not, the additional cost is funded with equity. The equity cheque grows and the leverage ratio falls.
• Earnings growth must therefore do more work. With two sources neutralised, the return depends almost entirely on improving the business.

A simplified illustration, using round numbers to expose the mechanism:

A business earning $10m acquired at 8× costs $80m, funded with $48m debt and $32m equity.

Grow earnings to $15m, exit at 8×, repay debt: proceeds $120m less $48m debt = $72m to

equity. 2.25× on the equity.

The same business acquired at 12× costs $120m. Debt is still sized to earnings — say $48m —

so equity is $72m. Same growth, same exit multiple: proceeds $180m less $48m = $132m.

1.83× on the equity.

Identical operational performance. The difference is entirely the entry price.

*Derived: illustrative arithmetic with round numbers, ignoring fees, interest and taxes.

Not a description of any actual transaction.*

That is the whole problem in one calculation, and it explains why every adaptation described below is fundamentally an attempt to avoid paying the higher multiple.

Everything a private equity firm does after the acquisition operates on the numerator. The entry multiple is the denominator, and it is set once.

Add-ons as the rational response

The dominant adaptation of 2017 was the shift toward add-on acquisitions, and it is a genuinely elegant response to the problem.

The mechanism. A firm owns a platform company. Rather than buying another platform at an expensive multiple, it buys smaller companies in the same sector and combines them.

Why it works arithmetically. Smaller companies trade at lower multiples than larger ones, for real reasons: they carry more customer and key-person concentration, have less professional management, and appeal to a smaller universe of buyers. That discount is the opportunity. Buying at 6× and integrating into a platform valued at 11× creates value immediately in the blended multiple — before any operational improvement.

Why the discount is partly real. The smaller company genuinely is riskier standalone. Integrated into a larger platform with professional management and diversified customers, much of that risk is removed. So the value creation is not purely arbitrage — some of it is real risk reduction. But some of it is arbitrage, and distinguishing the proportions is the skill.

What it changes about the firm. Add-on strategies require capability that traditional private equity did not need: a pipeline of small targets, integration expertise, and a management team capable of absorbing acquisitions repeatedly. Firms that built this operated a different business from the one they had operated a decade earlier — closer to an operating company than a financial investor.

The risks were real and became visible later:

• Integration is where the strategy fails, not acquisition. Multiple acquisitions compound integration complexity.
• Reported earnings become harder to assess, as adjustments for synergies and one-time costs accumulate across many transactions.
• Leverage builds across the platform, since each add-on typically adds debt, and the aggregate can grow faster than the earnings supporting it.

Covenants weaken

The credit market financing private equity changed materially through this period, in a way whose consequences remain partly untested.

A maintenance covenant requires a borrower to satisfy financial tests regularly — typically a maximum leverage ratio or minimum interest coverage — measured quarterly regardless of whether anything else happens. Breaching one gives lenders rights: to renegotiate, to demand repayment, or to take control.

An incurrence covenant applies only when the borrower takes a specific action, such as raising more debt or paying a dividend. A deteriorating borrower who takes no such action never triggers anything.

By 2017, a large majority of leveraged loans were covenant-lite — carrying incurrence covenants only. The shift happened because capital chasing yield had made lenders competitive on terms as well as price. When demand for loans exceeds supply, borrowers set the terms.

The consequences are asymmetric in time, which is what makes the change hard to assess:

• Defaults fall in the near term, because the technical triggers that would have caused them no longer exist. This looks like improved credit quality and is reported as such.
• Problems surface later and larger. A deteriorating borrower continues undisturbed until it cannot pay cash interest — by which point the deterioration is far advanced and the enterprise value has fallen further.
• Recoveries are plausibly worse, because lenders intervene later, when less value remains to recover.
• Reported default statistics understate risk, since the metric is measuring a different thing than it measured historically.

That last point is the one that matters for anyone reading credit data across this period. A default rate compared against a historical average is comparing two different definitions, because the trigger conditions changed. The comparison looks favourable and is not meaningful.

This connects directly to the private credit analysis in the 2024 and 2026 reports: the same question — do weaker protections reduce risk or defer it — remains open, and the asset class that grew fastest on the back of it has still not been through a full default cycle.

Holding periods extend

A quieter change in 2017 foreshadowed the structural developments of 2023.

The traditional model holds an asset three to five years. By 2017 average holding periods had extended, for several reasons operating together:

• Higher entry multiples require more operational improvement, and operational improvement takes time.
• Selling at an elevated multiple into an equally elevated market offers less advantage than selling into a rising one.
• Good assets were genuinely scarce, so firms holding one had reason to keep it rather than face the problem of redeploying the proceeds into the same expensive market.

Extended holding periods create a specific tension. Funds have finite lives. A ten-year fund holding an asset in year eight must resolve it, regardless of whether the timing is optimal.

The instruments that emerged to manage this tension — continuation vehicles, GP-led secondaries, strip sales — appeared in this period as niche solutions to a timing problem. They became standard infrastructure in 2023, when the exit drought made them necessary at scale rather than convenient at the margin.

The lesson generalises: structural innovations usually appear as workarounds before they become infrastructure. The continuation vehicle was invented to solve an ordinary problem years before the extraordinary one arrived.

What "operational improvement" has to deliver

If entry multiples neutralise leverage and multiple expansion as return sources, everything depends on earnings growth. It is worth being specific about what that requires, because the phrase "operational improvement" covers activities of very different difficulty and reliability.

Revenue growth. The most valuable source and the hardest to produce. A private equity owner can fund expansion, enter new markets, or invest in sales capacity — but underlying demand is not something ownership creates.

Margin improvement through cost reduction. Reliable and finite. Procurement, overhead rationalisation and process improvement produce real gains, and they are largely one-time. A business's cost base can be optimised once; it cannot be optimised repeatedly at the same rate.

Margin improvement through pricing. Powerful where the business has pricing power and unavailable where it does not. This is a property of the market position rather than of the owner.

Multiple arbitrage through add-ons. Buying smaller companies at lower multiples and integrating them, as described above. This is genuinely available and it depends on integration capability, which is the scarce input.

Capital efficiency. Working capital reduction, asset disposals, and better use of the balance sheet. Real, finite, and largely one-time.

The pattern across these is that the reliable sources are one-time and the repeatable source is the hardest. A firm can reduce costs once, optimise working capital once, and rationalise a portfolio once. To grow earnings over a five-year hold at a rate that compensates for a high entry multiple, the business generally has to grow revenue — which depends on the market as much as on the owner.

The implication for diligence is that a plan resting substantially on one-time improvements has a defined ceiling, and the ceiling should be computed rather than assumed. A plan resting on revenue growth is a bet on the market, and should be underwritten as one.

At a low entry multiple, a manager can be wrong about growth and still earn an acceptable return, because leverage and multiple expansion carry them. At a high entry multiple, the growth thesis has to be right. The entry price does not only reduce the return — it removes the margin for error.

What an allocator could act on

Ask for the return decomposition by source. How much came from leverage, multiple expansion and earnings growth — and within earnings growth, how much from revenue versus one-time cost improvement. A record built during a period of falling rates and rising multiples carried two tailwinds that are not attributable to the manager.

Check entry multiples against the vintage. Entry multiple is the single best predictor of vintage-level returns and it is reported by Bain annually, free. A manager deploying into a high-multiple vintage faces a headwind regardless of skill, and the appropriate response is pacing discipline rather than better selection.

Read the credit documentation, not just the leverage ratio. Covenant-lite structures defer the recognition of deterioration rather than reducing it. A default rate compared to a historical average is comparing two different trigger definitions, which makes the comparison flattering and meaningless. This is the same measurement problem the private credit reports return to in 2020, 2022 and 2024.

Assess add-on strategies on integration, not acquisition. The multiple arbitrage is real and the execution risk is entirely in integration. A firm doing many add-ons is running an operating business, and the question is whether it has the capability to absorb acquisitions repeatedly.

Watch aggregate leverage across a platform, not per transaction. Each add-on typically adds debt. The platform's total leverage can grow faster than the earnings supporting it, and the per-transaction ratios will not show this.

Treat extending holding periods as information. A manager holding assets longer than underwritten is either finding the improvement slower than expected or finding the exit market unattractive. Both are worth knowing, and the continuation vehicle — invented in this period to manage the resulting fund-life tension — became essential infrastructure by 2023.

What 2017 established for US private equity

• Capital abundance was shown to transmit directly to entry multiples, and entry multiples were shown to determine returns arithmetically.
• Add-on acquisition became the dominant deployment strategy, changing what private equity firms do as much as what they own.
• Covenant-lite became the market standard, altering what default statistics measure and deferring rather than reducing credit risk.
• Holding periods extended, creating the fund-life tension that continuation vehicles were invented to solve.
• The infrastructure that became essential in 2023 was built here, as a solution to a much smaller problem.