Multiple arbitrage is buying small companies at a low multiple of earnings and owning that profit inside a larger company that buyers value at a higher multiple. GF Data's 2024 medians for US private deals were 4.5x EBITDA for $1–10M companies, 6.5x for $10–100M and 9.0x for $100–500M, so the same dollar of profit can be worth roughly twice as much at scale. As of October 2026, GF Data's latest published quarter, Q2 2026, put the average private equity purchase multiple at 7.0x adjusted EBITDA. The gap is only real if the combined company is genuinely less risky and better run; overpaying, debt, overhead and failed integration can erase it.
A multiple is a price for risk and growth. A buyer pays more per dollar of profit when the profit is more likely to continue and grow.
Larger companies usually score better on the things buyers worry about:
The GF Data spread reflects this. In 2024, median EBITDA multiples rose from 4.5x for $1–10M enterprise value to 6.5x for $10–100M and 9.0x for $100–500M. More recent GF Data figures are averages, not medians, and they cover private-equity-sponsored deals. Its Q2 2026 report, released August 27, 2026, put the average purchase multiple at 7.0x trailing adjusted EBITDA, down from 7.3x in Q1 2026. Mercer Capital's Fall 2026 summary of GF Data gives about 6.8x for $10–50M deals and about 7.6x for $50–250M. Averages run above medians, so compare each set only with itself. GF Data's Q3 2026 figures were not yet published at this review; its Q3 2025 report came out in November 2025. Main Street businesses trade lower still. BizBuySell data from 2021 to Q1 2026 shows an all-sector average of 2.57x seller's discretionary earnings (SDE), which includes the owner's pay.
A roll-up buys several small companies in one industry or adjacent ones. It combines them under a platform with shared management, systems and reporting. If the platform earns a higher multiple than the pieces cost, the difference is the arbitrage.
The site's own illustration: buy a business earning $1M a year at 4x, or $4M, paid with debt, seller notes and equity. Inside a platform valued at 10x, that profit supports $10M of value before any growth. That is the core reason I think holding companies create value. But the 10x is illustrative. In the GF Data medians, a multiple near that level shows up only at $100–500M of enterprise value.
Assume a buyer acquires five companies, each earning $1M of EBITDA, at the $1–10M median of 4.5x. The overhead and debt figures are assumptions for illustration, not benchmarks.
Now add debt. Assume 60% of the $22.95M is borrowed ($13.77M) and $9.18M is equity.
The same debt that makes the good case look better makes the bad case worse. This ignores debt paydown and growth, which help, and integration costs, which hurt.
Roll-up returns come from three places. Separate them before you believe a projection.
A useful test: if the plan only works when the exit multiple is higher than the entry multiple, you are betting on the arbitrage. If it still works at the entry multiple because earnings grew and debt was repaid, the arbitrage is upside.
At scale, a public holding company can pay sellers partly in stock and raise equity in the market. If the market values the platform above what it pays for each business, issuing stock to buy profit can be accretive. The academic data is a caution, though. A study of 6,000+ IPO registrations from 1990 to 2024 found that, among newly public firms, stock-paid acquisitions underperformed cash deals by 20–30 percentage points over 3–5 years and were about 50% more likely to end in delisting. Public stock is a currency, not a free one.
Where these numbers come from: GF Data 2024 median EBITDA multiples by deal size, as summarized by Carry and Conquer; GF Data Q2 2026 average multiples from the ACG release of the GF Data Q2 2026 reports; size-band averages from the Mercer Capital Fall 2026 Middle Market Transaction Update; the Q3 2025 release date from GF Data news; BizBuySell 2021–Q1 2026 SDE multiples via Beancount; deal-cost survey from Big Deal Small Business; stock-paid deal performance from the CLS Blue Sky Blog (July 14, 2026). Overhead and debt figures in the example are assumptions.
Not if the platform really is a better business. It becomes financial engineering when the only change is the label and the debt load. Buyers at exit will look for management depth, clean reporting and organic growth.
The GF Data medians step up at $10M and again at $100M, but they are medians, not rules. Quality, growth and industry move individual prices more than size thresholds do. Size helps mostly because it tends to bring better management and reporting.
It can, but compare like with like. An SDE multiple includes the owner's pay, so once you hire a manager to replace the owner, the effective EBITDA multiple you paid is higher than it looks. Run the math on EBITDA after a market-rate manager.
Yes. A private platform can be sold to a larger buyer or recapitalized at the higher multiple. Going public is one route, and it fits only holding companies with real earnings and books an auditor can work with.
Once the arithmetic makes sense, the practical question is sequence: which thesis, which first platform, what capital and what team. That is covered in how do you run a roll-up as a small business owner?
Meraki Dealmakers is a service of Meraki Partners, LLC, an advisory firm. It is not a broker-dealer, investment adviser, business broker, law firm or accounting firm. This article is general information, not legal, tax or investment advice.
Start with the playbook. The first conversation is free.
Meraki Dealmakers is a service of Meraki Partners, LLC.
Meraki Partners, LLC is an advisory firm. It is not a broker-dealer, investment adviser, business broker, law firm or accounting firm, and does not solicit investors or receive compensation based on capital raised.
© 2026 Meraki Partners, LLC