In June, the Institute for Private Capital published the first large-sample empirical study of independent sponsor investments. Gregory Brown and William M. Volckmann II collected data on 846 transactions from the members of the Small Business Investor Alliance and its Independent Sponsor Forum, benchmarked them against matched buyout investments, and reported the results. For a segment that has operated for two decades largely on assertion and anecdote, it is the first time anyone has had a large-sample figure to cite.
The figure that circulated came from the trade association’s announcement rather than from the paper: a median gross internal rate of return of 23.8% against 18.5% for matched buyouts. The paper prints the first of those figures but not the second. It reports a median excess of 5.3 points, and 18.5% is the result of subtracting one from the other. Because a median of differences is not the difference of medians, that subtraction approximates the benchmark rather than reporting it. The comparison is directionally sound and the paper supports it, but it is the least informative result in the study, and the way it was derived is a reason to read the rest carefully.
Three findings beneath it did not appear in the announcement. The outperformance shows up in the rate of return but not in the multiple. The paper’s attempt to explain which investments did well found nothing. And the sample, read carefully, describes a specific part of the capital stack rather than the market as a whole. Together they show that the category outperformed its benchmark and that the data cannot identify which members of it will.
What the study is, and what it is a study of
The paper begins with a definition. An independent sponsor sources, negotiates, and diligences an acquisition without a pre-raised, committed pool of capital, then raises the money for that specific transaction. The work is the same as a fund’s; the difference is that the capital decision happens once per investment, with the investor underwriting the sponsor and the asset at the same time.
The data came from two surveys fielded in the summer of 2025, one to sponsors and one to the capital providers who fund them. The authors received 846 investment records, 697 from investors and 149 from sponsors, and note separately that the underlying contributions were 677 and 138 across 35 investors and 38 sponsors, for 815 after duplicates and irregularities. Of the 846, 542 were still active at the response date and 304 were fully exited.
Two restrictions then narrow the sample that produces the performance numbers, and both are sound choices. The authors use only investor-reported performance and discard the sponsor-reported figures, on the grounds that a sponsor reporting on its own investment is the more biased source. They exclude investments entered in 2022 or later unless already exited, because an unexited investment that young is a valuation opinion rather than a result. What remains is 328 observations for the multiple and 283 for the rate of return.
The benchmark is carefully constructed. Each independent sponsor investment is matched against United States buyout investments of no more than $100 million in investment size, drawn from the MSCI Private Capital Universe, entered in the same year, with at least 20 comparable investments per match. The comparison is therefore vintage-controlled, which matters for the window question discussed below.
The authors are also candid about what they cannot fix. Participation was voluntary, which raises the possibility that survivors are overrepresented. They note that two established and highly successful investors declined to participate over confidentiality concerns, which cuts the other way. They state that they cannot isolate or correct the bias, and that caveat should accompany any use of the figures.
The outperformance shows up in the rate before the multiple
The first finding is in a single table.
On the multiple of money, the median independent sponsor investment returned 2.10 times invested capital, and the median excess over its matched benchmark is 0.08, which the authors report as not statistically distinguishable from zero. On the rate of return, the median independent sponsor investment produced 23.8%, the median excess is 5.3 points, and that difference is significant at the 1% level.
Read those two rows together. On the multiple of money, the typical independent sponsor investment in this sample cannot be distinguished from a comparable buyout. On the annualized rate, it can. The same capital coming back in the same proportion at a higher annualized rate points to shorter holding periods or earlier distributions, which is the natural explanation, and it should be labeled as an inference rather than a finding. The paper does not decompose the result, reports no holding period for the benchmark against which the independent sponsor holds could be compared, and carries a majority of the sample at interim marks rather than realized values.
The averages differ from the medians, and the difference describes the shape of the distribution. The mean multiple is 2.86, with a mean excess of 0.73, significant at the 1% level. The mean return is 29.3%, with a mean excess of 11.4 points, also significant. When the mean multiple outperforms decisively and the median multiple does not, the outperformance comes from the right tail.
The category outperformed its benchmark, and the data cannot identify in advance which investments will.
The paper’s own dispersion figures make this concrete. Complete write-offs occurred in 4.7% of investments. Nearly a quarter returned more than four times invested capital. The standard deviation of the reported rate of return is 41.3 points against a mean of 29.3%. This is a distribution with a long right tail and a floor, and describing it by its median understates the upside while describing it by its mean overstates the typical outcome.
Loss incidence is the one place where a widely repeated claim holds up under inspection, with one correction to how it is usually stated. Independent sponsor investments show a loss rate of 23.5% against 21.6% for the matched set, which is the higher of the two, and the difference is not statistically significant at any loss threshold, including losses of 50% or more. The accurate reading is that no difference in downside is detectable at this sample size. That is a weaker claim than the one in circulation, that independent sponsors achieve the returns without taking on more risk of loss, and it is the one the data support.
The composition of that loss rate is more informative than the rate itself, and the announcement did not report it. Total losses, meaning capital fully written off, occurred in 4.7% of independent sponsor investments against 7.3% for the benchmark. Losses of 75% up to 100% run the other way, 8.4% against 5.2%. The samples are small enough that this could be noise, and the authors do not claim otherwise, but in this sample independent sponsor investments went to zero less often and lost most of their value more often. For an investor holding minority equity and junior debt, those two outcomes are not equivalent.
The most useful finding in the paper is a null result
The authors then tested what any reader would want tested next. They regressed performance and excess performance against the characteristics expected to matter: enterprise value at entry, entry valuation measured as enterprise value to earnings, industry, and vintage. They tried several robust methods, given how skewed the data are.
They found nothing. In their words, they could not identify any statistically significant factors, and could find no reliable evidence that larger or smaller, more or less profitable, or industry-specific investments performed differently. They attribute this to limited statistical power and expect a larger dataset would help.
This null result has more practical content than anything else in the study. The paper establishes that the category outperformed its benchmark on average, and that the outperformance came from the right tail. It then finds nothing observable at entry that predicts which investments land in that tail. The characteristics a screening memo is built from did not explain the outcomes.
The authors draw the conclusion themselves in the body of the paper: superior performance for investors is likely to depend on the ability to identify sponsors who can reliably deliver it. A study of transactions concludes that investors should underwrite the sponsor, because the transaction characteristics did not explain the results.
Who is in this sample
Three details define the population, and each one narrows what the study can be said to describe.
The first is the capital source. Small Business Investment Companies are the primary investor in 58.1% of the investments. The authors flag this directly as likely an artifact of the survey being facilitated by the Small Business Investor Alliance, which is the trade association for that program. Non-SBIC private equity accounts for 13.2%, funds of funds 9.4%, non-SBIC private credit 7.4%, family offices 6.5%, institutional limited partners 2.0%, and other sources the remaining 3.3%. This is a picture of the SBIC-funded portion of the independent sponsor market, and a different sampling frame would produce a different picture.
The second is the instrument. Weighted by investment rather than by investor, which is how the paper’s table is built, 93.1% of the capital provided came from investors who make minority equity investments, 68.1% from investors who make mezzanine debt investments, and 53.4% from investors who make unitranche investments. Only 22.4% came from investors who make control equity investments, and the categories overlap, because a single investor can do several. The performance figures in this paper are equity-component returns earned mostly from a minority position alongside junior capital, which is a different risk from a control position, and that context should accompany the numbers whenever they are quoted.
The third is the size of the tables that describe the sponsors themselves. The frequently quoted finding that 88% of independent sponsors have prior private equity experience comes from a table with 24 respondents; the figure is 21 people. The companion table on partner counts has 24 respondents, and the table on closed investment counts has 23. Those tables also contain a detail the summaries did not carry: in the closed-investment table, 26.1% of responding sponsors had completed exactly one investment and one respondent had completed none. The paper’s narrative states separately that two-thirds of transactions were done by sponsors with three or more investments to date, which is a transaction-weighted claim and a different measurement from the sponsor-weighted table beneath it. Both can be true at once, and they describe different things.
The study therefore rests on two samples: 846 transactions and about two dozen sponsors. Its findings about transactions rest on the first, its findings about sponsors rest on the second, and the second is too small to generalize from.
The vintage question, and why the benchmark answers it
The obvious objection to any private markets return study is that the time window flatters it. This sample’s mean entry year is 2017 and its mean exit year is 2020, with a median entry in 2018 and the middle half of investments entered between 2014 and 2020. Reported entry multiples sit at a median of 5.5 times earnings and exit multiples at a median of 8.0. The median investment expanded its multiple by 2.6 turns, which is a median of the per-investment changes and not the difference between the two medians above, and more than 75% of investments expanded at all. Buying at 5.5 times and selling at 8.0 times accounts for a substantial part of a 2.1 times outcome, and the period in question was broadly favorable for that kind of expansion.
The benchmark addresses most of this objection, because it is matched by entry year. Comparable buyouts entered in 2018 benefited from the same repricing. Whatever the window contributed, it contributed to both sides of the comparison, and the excess return is measured after that. The finding that survives is relative, and relative performance is what the matched benchmark is designed to measure.
The absolute figures do not survive as an expectation. A median 23.8% and a median 2.10 times describe what a particular set of investments, mostly entered in the second half of the 2010s and mostly funded by a particular kind of investor, has reported so far, with a majority of them not yet exited.
One number in the study has nothing to do with returns and received little attention. The median investment added 10 employees over the holding period, a median increase of 17%. The study does not attribute this to ownership: it has no counterfactual for employment and does not benchmark it, and the years in question were broadly expansionary for small industrial and service firms. It is still worth knowing, because the sample is concentrated in business and consumer products and services, which together account for 69.0% of investments, with technology, healthcare, and financial services adding up to only 20.9%. Manufacturing accounts for 23.5% on the paper’s own memo line, which cuts across several of those industry categories rather than sitting beside them.
What follows for your own decisions
If you provide capital into this market, the study’s null result is the finding to carry into your next underwriting. Entry size, entry valuation, industry, and vintage did not explain outcomes in the largest sample published to date. Whatever explains outcomes lies outside that list, which leaves the assessment of the sponsor and the specific situation as the main underwriting work. In practice, ask for a sponsor’s entire history rather than its exits, because a distribution with a 4.7% write-off rate and a long right tail cannot be evaluated from its best outcomes.
If you are considering selling a company in this range, the data describe your likely buyer more precisely than anything published before. The typical investment in the study has an enterprise value between $10 million and $50 million and earnings between $2 million and $10 million; 14.5% of investments were under $10 million of enterprise value and 12.1% were in companies with earnings above $10 million. The buyer is probably not drawing on a committed fund. They are assembling capital for your transaction specifically, most often alongside a Small Business Investment Company, and more often taking a minority equity position than a control one. That changes which questions are worth asking during a process, starting with who is funding the transaction and whether they have funded this sponsor before.
If you advise these companies, one comparison deserves care. A separate survey published by Citrin Cooperman in 2025, drawing on 172 sponsor and capital provider responses, reports that 44% of independent sponsors now target companies with earnings above $10 million, against 4% in 2017. The Institute for Private Capital sample shows 12.1% of investments at that size. Those two numbers are not directly comparable, and the reasons matter: one measures current intentions and the other measures completed investments, and the completed investments are concentrated in years when the whole market was smaller. Some of the difference is sponsors moving upmarket over a decade, and some is the ordinary gap between what a buyer is looking for and what a buyer closes. Neither the intentions nor the outcomes tell you which of those applies to the sponsor in front of you, so ask.
The study’s contribution is not the 23.8%. It gives a segment that previously argued about itself a shared set of facts, including the uncomfortable ones. The next version of this research will have more exits and more statistical power, and the result to watch is whether the null finding holds. If a larger sample still cannot explain which investments worked, investors should treat sponsor selection, rather than deal characteristics, as the main source of returns in this segment.
Works Cited
- Independent Sponsors: Investment Characteristics and Performance | Gregory Brown and William M. Volckmann II, Institute for Private Capital and UNC Kenan-Flagler Business School, June 22, 2026 (full paper, PDF)
- Independent sponsor deals deliver higher returns than traditional lower middle market buyouts, new Institute for Private Capital study finds | Small Business Investor Alliance, June 22, 2026
- Citrin Cooperman’s 2025 Independent Sponsor Report | Citrin Cooperman Transaction Advisory Services, 2025