Press question mark to see available shortcut keys

"'WE HAVE MET THE ENEMY... AND HE IS US': Lessons from Twenty Years of the Kauffman Foundation's Investments in Venture Capital Funds and The Triumph of Hope over Experience", May 2012 http://www.kauffman.org/uploadedFiles/vc-enemy-is-us-report.pdf

---

Excerpts:

Venture capital (VC) has delivered poor returns for more than a decade. VC returns haven't significantly outperformed the public market since the late 1990s, and, since 1997, less cash has been returned to investors than has been invested in VC. Speculation among industry insiders is that the VC model is broken, despite occasional high-profile successes like Groupon, Zynga, LinkedIn, and Facebook in recent years.

Only twenty of 100 venture funds generated returns that beat a public-market equivalent by more than 3% annually, and half of those began investing prior to 1995. The majority of funds—sixty-two out of 100—failed to exceed returns available from the public markets, after fees and carry were paid. There is not consistent evidence of a J-curve in venture investing since 1997; the typical Kauffman Foundation venture fund reported peak internal rates of return (IRRs) and investment multiples early in a fund's life (while still in the typical sixty-month investment period), followed by serial fundraising in month twenty-seven. Only four of thirty venture capital funds with committed capital of more than $400 million delivered returns better than those available from a publicly traded small cap common stock index. Of eighty-eight venture funds in our sample, sixty-six failed to deliver expected venture rates of return in the first twenty-seven months (prior to serial fundraises). The cumulative effect of fees, carry, and the uneven nature of venture investing ultimately left us with sixty-nine funds (78 percent) that did not achieve returns sufficient to reward us for patient, expensive, long-term investing.

It's become a bit of a sport among venture capital (VC) insiders and observers to assert that the venture capital model is broken. [Scott Austin, Majority of VCs in Survey Call Industry Broken, Wall Street Journal, June 20, 2009. http://blogs.wsj.com/venturecapital/2009/06/29/majority-of-vcs-in-survey-call-industry-broken/ Also see AngelBlog for links to several articles, http://www.angelblog.net/The_VC_Model_is_Broken.html ] Industry returns data show that VC returns haven't beaten the public market for most of the past decade, and the industry hasn't returned the cash invested since 1997, certainly a compelling sign that something must be wrong.

Recommendation 2: Reject the Assumption of a J-Curve: The data we present indicate that the "J-curve" is an empirically elusive outcome in venture capital investing. A surprising number of funds show early positive returns that peak before or during fundraising for their next fund. We see no evidence that the J-curve is a consistent VC phenomenon or that it predicts later performance of a fund. Committees should be wary of J-curve-based defenses of VC investing.

IRRs are influenced by the timing of investment cash flows and the length of time an investment is held, so a fund with limited capital invested and returns from early exits or early valuation write-ups can generate attractive IRRs in the short term. For example, a company that is sold and returns more than twice the invested capital in three years generates a 26% IRR, but the same multiple generated by a sale in year ten results in only a 7.2% IRR. As Josh Lerner, a professor at Harvard Business School and leading researcher on venture capital, notes, "When you look at how people report performance, there's often a lot of gaming taking place in terms of how they manipulate the IRR."5 Our portfolio analysis shows clearly that high IRR performance frequently is generated early in a fund's life, either before or during fundraising for the next fund, but that those early high IRRs do not predict a fund's eventual performance. We evaluated our funds based on the difference between maximum IRR, and final IRR and PME in order to assess whether peak IRRs early in a fund's life might foretell the outcome. What we learned is that our best-performing funds—those launched prior to 1995—did not report peak returns until the sixth or the seventh year of their lives. That pattern began to change in the late '90s, when peak returns almost always were reported during the fund's five-year investment period, usually in the first thirty-six months. We also see that peak IRRs are not consistent predictors of high final returns.

When we forget about IRR and look only at performance between the public markets and our VC portfolio, we see that our best-performing funds relative to the Russell 2000 index are pre-1995 vintage years. There is some select good performance (but overall mediocre returns) from the 1996-2000 Internet boom funds, and there are poor returns from 2000 forward...Investing in venture capital in the early to mid-1990s generated strong, above-market returns, and performance by any measure was good. What has happened since? Our colleague Paul Kedrosky asserts that the venture capital industry is too big and must shrink to effectively fund entrepreneurs and generate competitive returns.6 Longtime venture investor Bill Hambrecht notes that, "When you get an above-average return in any class of assets, money floods in until it drives returns down to a normal, and I think that's what happened.7"

During the past fifteen years of poor performance, investors have committed about $20b each year to VC, about four times the $500 million in capital committed to venture capital in total during the decade 1985-1995. The flow of capital into VC has slowed over the past few years, but still has a way to go to return to pre-1995 levels.8 If it's true that too much capital is dragging down returns, money should be flowing out of VC until returns normalize. Despite more than a decade of poor returns relative to publicly traded stocks, however, there appears to be only a modest retrenchment by LPs. We wonder: why are LPs so committed to investing in VC despite its persistent underperformance? LP hopes for VC returns are high, and those hopes fuel new money into VC funds nationwide. A Probitas Partners survey of nearly 300 institutional investors found that two-thirds of investors expect a 2x+ multiple from top quartile, early-stage VC funds.9 Contrary to those lofty expectations, Cambridge Associates data show that during the twelve-year period from 1997 to 2009, there have been only five vintage years in which median VC funds generated IRRs that returned investor capital, let alone doubled it.

VC mandates fail because generating great VC returns is entirely dependent on which funds you're in, not how many funds.15 Generating great VC returns requires access to the small group of best-performing funds. One study conducted by a fund-of-funds investment manager revealed that, from 1986-1999, a mere twenty-nine funds raised 14% of the capital in the industry, but generated an astonishing 51% of total distributions—about a 3.6x multiple. The remaining 500+ funds in the industry generated a 0.4-0.6x multiple. Put another way, the study concludes that the twenty-nine top funds invested $21 billion and returned $85 billion, while the rest of the VC fund universe invested $160 billion and returned a scant $85 billion.16 This is a surprising result given the strong venture capital returns from that time period. This performance skew is most dramatically reflected in the distribution of VC returns. In our portfolio, we find a distribution in which only sixteen of ninety-nine funds generate a VC return of 2x+. The remaining funds form a long tail of underperformance, producing an average return of 1.31x. ...Industry research conducted by Josh Lerner finds a similar relationship between IRR and fund size.21 He finds that VCs that perform well raise successively larger funds, and they often see consistent or improving returns up until the fund size grows larger than $500m, after which performance starts to degrade.

We evaluated all our venture fund investments on the same ten-year investment horizon. We centered all eighty-eight VC funds from vintage years 1995-2009 on a time zero axis and plotted both gross and net dollar-weighted IRRs. Our aggregate portfolio data reveals a trend of early positive returns that resembles the shape of an "n-curve," where net IRR peaks in month sixteen (presumably driven by increases in company valuations, which the GPs themselves determine), and retreats precipitously over the remaining term of fund life.

This structure has been the industry standard for so long that it's difficult to trace its origins or rationale. The same 2 and 20 model remains nearly universal today. One study analyzed compensation from ninety-three VC funds raised from 1993-2006 and found that 90% of the funds charged a 2% or more fee, and 95% of funds charged a 20% carry.34 In an earlier analysis, Paul Gompers and Josh Lerner reached a similar conclusion.35 It's interesting that VCs have positioned themselves as supporters, financers, and even instigators of innovation, yet there has been so little innovation within the VC industry itself. There have been changes—more funds, more money, bigger funds, and bigger deals—but very little 'creative destruction' around how funds are structured, capital is raised, or VCs are paid.

The general perception is that VCs are paid based on how well their investments do. If true, that would align the interests both of LPs, who want to maximize their returns, and VCs, who are rewarded for making high-risk, high-return investments. A closer look at compensation data shows that, while a select group of VCs remain focused on delivering great investment performance to their investors, too many are compensated like highly-paid asset managers. Public data on GP compensation amounts and structure is difficult to obtain, yet one recent study that analyzed ninety-four VC funds and estimated the amount of partner revenues from management fees and carry found that VC funds receive nearly two-thirds of their revenues from fixed fees rather than from performance-based carry.36 VC funds received a median $14.61 per $100 under management, compared to only $8.20 in carry. Another study analyzed vintage-year funds from 1986 through 1997, and also found that average VC compensation is not really performance-based at all.37 The authors found that an average VC received about half its compensation from the management fee, a surprising finding given the historically unprecedented nature of fund returns during this period of the Internet bubble in the late 1990s. Many LPs are keenly aware of the misalignment inherent in the 2% flat management fee, which pays VCs more for raising bigger funds and pays them steadily whether or not they perform. A recent Probitas Partners survey of 291 institutional LPs found that 48% of respondents identified the overall level of management fees as an area of concern.38 The same% of LPs also reported fears that fee levels were destroying the alignment of interests between GPs and investors. An Ernst and Young survey of LPs found that 89% of respondents want to see changes to the management fee.39 For smaller funds, a 2% fee might be a reasonable way to cover fund expenses. But the impact of fee income is most mis-aligning in the expanding universe of $1b+ funds, a fund size that generates $20m per year in fees from a single fund, whether there are five partners or twenty-five, one office or ten, positive returns or losses. As one GP told us: "The management fee is like heroin. No one can step away from 2 and 20."

why would GPs—top performers and emerging managers included—structure compensation systems that rely heavily on a management fee and constraints carry to a mere 20% of profits? Our interviews with General Partners blame LPs for insisting on consistent and historical fee practices. They report general rebellion by existing LPs to any change in structures that might require additional explanation or analysis for investment committees. GPs insist that they have to sell what LPs will buy and they say that LPs as a group are not at all interested in discussing any alternative structures to the 2 and 20 model. Our interviews indicate that many GPs are, not surprisingly, generally interested in increased carry and more resistant (but still open) to reduced management fees based on a budget. Our discussions with LPs confirm that it's they who consistently express strong opposition to almost any deviation from 2 and 20. One experienced GP raising his own first-time fund said he offered a budget-based management fee, and was open to a sliding carry based on performance (e.g., 25% above 2x), but felt that such a departure from industry practice sent the message to prospective LPs that the fund was desperate to attract new investors by offering unique and better-aligned economic terms. Another GP from a top-performing fund that consistently is oversubscribed (thus putting him in a position to negotiate carry) told us his firm was reluctant to "seem greedy" and risk alienating its investors by asking for higher carry on its best returns, even if the offer were made in conjunction with a reduced management fee.

The remaining question in our minds is: Do LPs have the interest, engagement, and will to actually be different and more selective investors in VC? We talked with a number of LPs who did not agree with the arguments we make in this paper, or didn't "get" why we think they're important. During our discussion about VC firm economics, one LP said that he didn't worry about management fees or firm budgets because "those guys have to make a living too," so it just wasn't a big issue for him to explore during due diligence. Another LP said that negotiating alternatives to 2 and 20 "isn't worth the energy." Several peers listened to our list of topics and responded by cautioning us that "this is a relationship business," implying a view that we are better off accepting the status quo and being in misaligned, underperforming VC relationships than pursing negotiations for better terms.

The Foundation has been an active VC investor for more than twenty years, and, in the past, we've invested in funds without strong performance or aligned terms. We have come to believe that it's a mistake to continue to do so. As recently as 2007, we had more than sixty GPs and more than 100 funds in our portfolio. Over the past few years, through attrition and select sales on the secondary market, we've reduced our portfolio to thirty GPs and sold about sixty funds that were underperforming and/or significantly misaligned with our interests. We've walked away from investments in emerging managers and re-ups in existing partnerships that were not willing to enter into better-aligned terms. We know we may need to continue to walk away, even from some of our current favorite VCs who, if they continue to be successful, may succumb to the pressure and incentives to raise bigger funds.
Shared publiclyView activity