Can Private Equity Winners Be Picked in Advance?

Kaplan and Schoar's 2003 analysis of private equity fund returns has been cited as support for choosing managers with strong prior records. The "winners" in that research are funds that land in the top quartile or top third of their peers, not necessarily funds that beat the stock market. A closer reading, and the research that followed, is more nuanced.

Why this study still gets cited

Private equity is private by design, which makes its results hard to study. In 2003, Steven Kaplan of the University of Chicago and Antoinette Schoar of the MIT Sloan School of Management circulated a working paper that took an early look at fund-level results. It was later published in the Journal of Finance in 2005, and it continues to be cited in debates over whether private equity earns its place in institutional portfolios, including in the 2020 follow-up discussed below.

The setting matters. According to the paper, which cites Venture Economics data, annual commitments to private equity partnerships grew from under $10 billion in 1991 to more than $180 billion at the 2000 peak. Institutions were moving quickly into an asset class with limited public disclosure.

The authors used a database from Venture Economics, which relied on voluntary reporting by fund managers (general partners) and their investors (limited partners). Their core sample was 746 funds that were largely liquidated, mostly started before 1995. About 78% were venture capital funds and 22% were buyout funds. All returns were measured net of fees.

The main yardstick was the public market equivalent, or PME. It asks what an investor would have ended up with had the same cash flows gone into the S&P 500 Index instead. A PME above 1.00 means the fund beat the index. A PME below 1.00 means it trailed. (The S&P 500 is an unmanaged, market-capitalization-weighted index of large U.S. companies and cannot be invested in directly.)

Finding one: the average fund roughly matched the index, and the typical fund trailed it

Across the full sample, the average fund's net return was approximately equal to the S&P 500. Weighting each fund equally, the average PME was 0.96. Weighting by committed capital, it was 1.05. Venture capital funds averaged 1.21 on a capital-weighted basis, while buyout funds averaged 0.93. The authors attributed the gap to timing and size: the larger venture funds of the 1990s outperformed the smaller venture funds of the 1980s, while buyout funds showed the opposite pattern.

Averages can hide the experience of the typical investor. The median fund, the one in the middle of the pack, had an equal-weighted PME of just 0.74 and a capital-weighted PME of 0.82. Because the average sits above the median, results were skewed toward stronger funds.

The spread between funds was wide. On an equal-weighted basis, the fund at the 25th percentile had a PME of 0.45, while the fund at the 75th percentile had a PME of 1.14. Results also swung with the year a fund began investing (its vintage). Measured on a capital-weighted basis, the PME was below 1.00 in 10 of the 16 vintage years from 1980 through 1995 (two of them at 0.99) and above 1.00 in the other six, reaching 1.45 for 1994.

The authors were careful about limits. Reporting to Venture Economics was voluntary, and they said any bias would likely run toward better-performing funds. Their sample leaned toward larger funds, which tended to do better, and roughly half of all funds in the database did not report performance at all. PME also does not adjust for risk, and the authors noted it would overstate results if private equity carries more market risk than the index. For these reasons they described their conclusions on average returns as uncertain.

Finding two: performance persisted from one fund to the next

The paper's most influential result concerned repeatability. When a manager's earlier fund did well, its next fund tended to do well too. In the authors' regressions, a 1% better result in the prior fund was associated with a 54 basis point (0.54 percentage point) better result in the next one. When they sorted funds into thirds by performance, a manager whose previous fund landed in the top third had a 55% chance of landing there again, and one in the bottom third had a 44% chance of staying there. By chance alone, each outcome would be about one in three.

"Returns persist strongly across funds raised by individual private equity partnerships."

Steven Kaplan and Antoinette Schoar, abstract, MIT Sloan Working Paper 4446-03. © 2003 Steve Kaplan and Antoinette Schoar.

The authors contrasted this with mutual funds, where persistence has been hard to find. They tested whether differences in risk, industry focus or selective reporting could explain the pattern and concluded that such explanations were unlikely. They also noted something important for any investor hoping to benefit: better-performing funds did not grow proportionally as much as average ones, and most limited partners describe the best funds as oversubscribed. Repeatable skill is of little use if investors cannot get access to it.

Why repeatability is a harder question than the headline suggests

Two developments complicate the picture. The first concerns the data. In a 2020 follow-up, Robert Harris, Tim Jenkinson, Steven Kaplan and Ruediger Stucke noted that the Venture Economics data used in the earlier paper had since been shown to be flawed. Kaplan is a co-author of both studies, and the 2020 authors describe their data as higher quality than what was available earlier.

The second concerns what investors can actually observe. Using cash-flow records from a large sample of institutional investors, covering 2,222 U.S. funds (893 buyout and 1,329 venture capital) with vintage years from 1984 through 2014 and data through June 2019, the 2020 study confirmed persistence when it measured each fund by its final results. But a committee deciding whether to commit to a new fund does not have final results for the manager's last fund. It has interim figures, and those proved an unreliable guide: only 57% of buyout funds that reported top-quartile results at the time of fundraising ended up in the top quartile (63% for post-2000 funds). Using only information available at the time of fundraising, they found little or no evidence of persistence for buyouts, including after 2000. For post-2000 buyouts, the rule of thumb of backing managers whose prior funds were in the top quartile did not hold up.

"we find little or no evidence of persistence for buyouts"

Harris, Jenkinson, Kaplan and Stucke, NBER Working Paper 28109 (2020). Measured using fund performance known to investors at the time of fundraising.

The evidence is not one-sided. The same study found that venture capital persistence remained even with information available at fundraising, though it declined after 2000. Repeatability appears to differ by strategy and by era. For buyouts, persistence that is visible in hindsight largely disappears when measured with fundraising-date information. For venture capital, it remains, though weaker after 2000.

The fee drag

Kaplan and Schoar did not have fee data for individual funds. They relied on typical terms of the period reported in earlier research: an annual management fee of 1.5% to 2.5% plus 20% of profits. They also cited earlier work estimating that, for venture funds, the discounted value of management fees alone equaled 16% to 19% of committed capital. Because they found average gross returns above the index but net returns roughly equal to it, their own arithmetic suggests that fees absorbed the gross excess return.

Later studies put numbers on that gap. Ludovic Phalippou and Oliver Gottschalg, writing in the Review of Financial Studies, argued that earlier industry figures were overstated because of inflated valuations of unsold investments and a bias toward better-performing funds. Using a base sample of 852 funds raised between 1980 and 1993, and after corrections, they estimated that funds outperformed the S&P 500 by about 3% a year before fees but underperformed it by about 3% a year after fees, with fees estimated by the authors at about 6% a year.

"We estimate fees to be 6% per year."

Phalippou and Gottschalg, abstract, Review of Financial Studies (2009).

A separate line of research looks at how fee contracts are built. Andrew Metrick and Ayako Yasuda analyzed 238 funds raised between 1993 and 2006 and estimated that about two-thirds of managers' expected revenue comes from fixed components that do not depend on performance. In a 2020 paper, Phalippou estimated that private equity funds of 2006–2015 vintages had collected roughly $230 billion in performance fees (carried interest, the managers' share of profits) as of the end of 2019, and argued that investors' net returns since 2006 had matched those of public equity indexes.

The other side of the debate

A balanced reading has to include the research that reaches different conclusions. In a 2014 study, Harris, Jenkinson and Kaplan, using institutional investors' cash-flow data, found that the average buyout fund in their sample of nearly 1,400 U.S. buyout and venture capital funds (Burgiss data from more than 200 institutional investors, cash flows through March 2011) outperformed the S&P 500 net of fees by more than 3% a year, while venture capital funds outperformed public equities in the 1990s but underperformed in the 2000s.

"Private equity returns after fees matched public market returns"

Ludovic Phalippou, University of Oxford, as quoted in Institutional Investor's 2020 press coverage of his paper "An Inconvenient Fact: Private Equity Returns & The Billionaire Factory." Quotation taken from the press coverage, not from the paper. Other researchers, including those above, reach different conclusions.

The 2020 persistence study points the same way on average returns. Across its 1984–2014 vintages, measured through June 2019, buyout funds had an average net PME of 1.20 and venture capital funds 1.22, with lower averages for more recent buyout vintages, and buyout funds in every previous-fund quartile, as well as first-time funds, returned more than the S&P 500. The difficulty it identifies is not whether private equity has beaten the index on average, but whether an investor can pick the managers who go on to do so.

Phalippou has separately argued that much of the apparent outperformance depends on the benchmark chosen. Researchers also differ on how to treat unsold investments and how to adjust for risk and illiquidity.

What the studies have in common is narrower. Results vary widely across funds. Averages are sensitive to the data and benchmark used. In these studies, fees were large and largely fixed. And the later evidence finds weaker support than the 2003 paper's headline for picking tomorrow's winners from yesterday's record, particularly for buyouts. Venture capital persistence remains, though weaker after 2000.

Summary: Kaplan and Schoar (2003) and later research

Sources: the studies listed under Sources below. The studies use different samples, periods and methods and are not directly comparable. Actual historical results of third-party funds as reported by the authors; not hypothetical or back-tested, and not IFA performance. Past performance does not guarantee future results.

What this means for an investment committee

For a foundation, endowment or nonprofit, the research suggests several questions worth asking of any private equity proposal:

• Is the average the right expectation? The typical fund in the Kaplan and Schoar sample did worse than the average.
• Can we identify the winners in advance? Persistence measured with hindsight is not the same as persistence an investor can use.
• Can we get in, and what does it cost? The funds with the best records may be closed to new money. In the Phalippou and Gottschalg estimates, fees exceeded the gross excess return.

IFA OCIO takes a different approach. It builds diversified portfolios from index funds, which hold the securities of a market index rather than seeking managers expected to outperform, and it pays close attention to costs. The comparison here is meant to contrast two ways of meeting the same institutional need, and they differ in important ways, including liquidity (private equity funds are typically closed-end partnerships with lives of ten to twelve years), fees, access to managers, and the type of risk taken. IFA's advisory fees are described in Item 5 of its Form ADV Part 2A brochure and on the Fees page of IFA.com.

Index funds do not remove market risk: when markets fall, an index portfolio falls with them, and an index-based portfolio will sometimes trail other approaches. Diversification does not guarantee a profit or protect against loss in declining markets. Lower costs do not guarantee better results.

Sources

• Kaplan, Steven and Antoinette Schoar. "Private Equity Performance: Returns, Persistence and Capital Flows." MIT Sloan School of Management Working Paper 4446-03, November 2003 (published in the Journal of Finance, 2005). © 2003 Steve Kaplan and Antoinette Schoar.
• Harris, Robert S., Tim Jenkinson, Steven N. Kaplan and Ruediger Stucke. "Has Persistence Persisted in Private Equity? Evidence from Buyout and Venture Capital Funds." NBER Working Paper 28109, November 2020.
• Phalippou, Ludovic and Oliver Gottschalg. "The Performance of Private Equity Funds." Review of Financial Studies 22(4), 2009, pp. 1747–1776.
• Metrick, Andrew and Ayako Yasuda. "The Economics of Private Equity Funds." Review of Financial Studies 23(6), 2010, pp. 2303–2341.
• Phalippou, Ludovic. "An Inconvenient Fact: Private Equity Returns & The Billionaire Factory." SSRN 3623820, June 2020; Institutional Investor coverage, 2020.
• Harris, Robert S., Tim Jenkinson and Steven N. Kaplan. "Private Equity Performance: What Do We Know?" Journal of Finance 69(5), October 2014, pp. 1851–1882; NBER Digest summary, July 2012.