What the Kauffman Foundation Learned from Twenty Years of Venture Capital - Index Fund Advisors, Inc.

In May 2012, three members of the Ewing Marion Kauffman Foundation's investment staff, Diane Mulcahy, Bill Weeks, and Harold Bradley, published a review of roughly twenty years of the Foundation's venture capital investing. The Foundation began investing in VC funds in 1985, had been a limited partner in more than 100 funds run by more than sixty general partners, and as of February 29, 2012 held an endowment of $1.83 billion, with about $249 million invested in or committed to venture capital and growth equity. The paper's data run through 2011. The authors placed responsibility largely on institutional investors, including the Foundation, arguing that investment committees and trustees should "shoulder blame for the broken LP investment model."

The Foundation's data covered 99 venture funds. Of these, 50 had a net multiple below 1x, 33 fell between 1x and 2x, and 16 reached 2x or better, six of them at 3x or more. The average multiple was 1.31x, and the paper does not separate recent funds that had not matured. The authors found that none of the Foundation's funds raising more than $500 million returned more than twice invested capital after fees. They also cited a Silicon Valley Bank study of 850 funds from 1981–2003 vintages, which found 51% of funds larger than $250 million failed to return capital after fees.

The authors preferred public market equivalent (PME) analysis against the Russell 2000 over top-quartile and vintage-year rankings, arguing that venture investing is riskier and less liquid than public equities and must beat public markets to compensate. Their hurdle was an excess return of about 3 percentage points a year, roughly a PME of 1.34 for a ten-year fund. Of 96 funds from 1989–2010 vintages, 25 cleared it: 10 of 14 funds raised before 1995, 8 of 48 from 1996–2000, and 2 of 21 from 2001–2005. Harris, Jenkinson, and Kaplan (2014) found similar results across nearly 1,400 funds, with venture outperforming public equities in the 1990s but underperforming in the 2000s. The paper also notes the PME method does not adjust for differences in risk or liquidity.

On fees, the paper describes the standard "2 and 20" structure: a management fee of about 2% of committed capital during a five-year investment period, plus about 20% carried interest. A study of funds raised 1993–2006 found about 90% charged fees of 2% or more and 95% charged 20% carry. Metrick and Yasuda (2010) estimated that about two-thirds of expected manager revenue across 238 funds came from components not dependent on performance. The authors also noted that some managers charged additional fees on older portfolios and objected to general partner commitments of about 1%, while the Foundation said it expects 5% to 10%.

On repeatability, the paper found that reported net IRRs in aggregate tended to peak around month 16, and managers typically returned to market about 27 months later, while only 25 of 86 funds (29%) followed the traditional J-curve. Of the 22 funds with VC PME at 27 months, only six retained it at the end, while 53 of 66 funds lacking it at that point never achieved it. Academic findings are more mixed: Kaplan and Schoar (2005) documented strong persistence, while Harris, Jenkinson, Kaplan and Stucke found venture persistence continued but weakened after 2000, and was reduced when limited to information available at fundraising. The paper estimates only about 10 to 50 top-performing venture funds exist out of several hundred.