E124: Is Venture Capital Entirely Based on Luck? University of Chicago and Oxford Study

Episode Summary

Executive Summary: The conversation argues that venture returns are highly persistent and best understood through primary data rather than industry myths. VenCap says small-fund superiority is overstated due to survivorship bias and missing data, while growth funds still exhibit venture-like power law dynamics, including high loss rates and fund-returning outliers. The firm’s strategy is concentrated on managers proven to back rare winners, and political shifts may help crypto and tech near term, though long-term allocation should remain disciplined.

Main Topics: Persistence of venture fund performance (Priority: 5/5): The speakers discuss evidence that top-quartile venture funds often remain top-quartile, and that poor performers can also persist at the bottom, reinforcing the idea that manager skill and repeatability matter in venture. Debunking the small-fund outperformance meme (Priority: 5/5): David argues that claims about smaller or emerging managers outperforming larger ones are not supported by the publicly available data, which is incomplete and heavily affected by self-reporting and survivorship bias. Survivorship bias in venture datasets (Priority: 5/5): PitchBook and Cambridge rely on voluntary manager reporting, which can overrepresent successful funds and understate failure rates. Carter is presented as a cleaner source for emerging managers because it records all platform-raised funds. Power law in growth funds (Priority: 5/5): VenCap’s data suggests growth-stage funds are still highly concentrated: many deals lose money, a small share produce 10x outcomes, and fund-returners appear more often than expected even at later stages. VenCap’s concentrated manager-selection strategy (Priority: 4/5): VenCap evolved from broad diversification to a concentrated portfolio focused on a small set of consistently top-performing managers who can identify and back fund-returning companies. Fund size, exit size, and long-term tech value creation (Priority: 4/5): The speakers debate whether larger funds must necessarily underperform; the response is that this depends on whether future company exit values can support larger fund sizes over the relevant 10-15 year horizon. Politics, regulation, and future fundraising conditions (Priority: 3/5): Post-election sentiment may improve crypto and tech via a more favorable regulatory environment, but LPs should avoid overreacting to short-term political cycles and stay focused on long-term venture fundamentals.

Key Arguments: Venture performance is unusually persistent: top and bottom quartile funds both tend to repeat, more so than in private equity. Public venture databases are incomplete because managers self-report, creating survivorship bias that overstates returns. Claims that small or emerging managers outperform cannot be reliably proven from PitchBook because the smallest-fund category has very low performance-data coverage. Carter may be a better future dataset for emerging managers because it captures all managers that raised via the platform, reducing survivorship bias. Growth funds are not low-risk in practice; more than 40% of deals still lose money. Late-stage investing still exhibits venture-style concentration: 10x outcomes and fund returners remain rare but meaningful. VenCap’s best-performing fund was an $800 million growth fund that generated a 13.5x multiple, showing large funds can produce exceptional returns. The key underwriting question is not whether a fund is large, but whether its manager can plausibly generate a fund-returning investment at that size. Early fund marks and TVPI are weak predictors of eventual DPI; actual fund outcomes matter more than interim paper gains. LPs should keep a steady, long-term allocation cadence and avoid trying to time politics or specific sectors too aggressively.

Data Points: Top-quartile fund persistence: Over half - University of Chicago study cited as evidence that top-quartile venture funds tend to remain top-quartile. Successive top-quartile chance: 45% - Tim Jenkinson/Oxford research referenced for consecutive top-quartile venture fund performance. PitchBook funds raised: About 14,000-15,000 - VC funds raised between 2010 and 2019 in the PitchBook dataset discussed. PitchBook funds with performance data: About 1,000 - Only a small subset of the raised funds had performance data available. Small-fund performance-data coverage: 5.1% - Funds in the $0-$99 million size band with performance data on PitchBook. $100M-$250M performance-data coverage: 13.9% - PitchBook coverage for the next fund-size tier. $250M-$500M performance-data coverage: 23% - PitchBook coverage for mid-sized funds. $500M-$1B performance-data coverage: 29.7% - PitchBook coverage for larger funds, described as much more likely to have data. Early-stage companies that do not return capital: 60% - Baseline power-law description for early-stage fund outcomes. Early-stage fund-returners: Around 1% - Share of early-stage investments that ultimately return the fund. Growth-fund losing deals: More than 40% - VenCap’s finding that a large share of growth investments still lose money. Early-stage 10x outcomes: About 5.5% - Reference point for the proportion of early-stage investments reaching 10x. Growth-fund 10x outcomes: Just under 5.5% - Growth funds were slightly below early-stage funds on 10x frequency. Growth-fund returners: 1.6% - Share of growth-fund investments that become fund-returners, higher than early stage. VenCap capital concentration: 90% - Capital invested over the last decade went to 12-13 managers. Best-performing fund size: $800 million - VenCap’s highest-performing fund in the 2010-2019 sample. Best-performing fund multiple: 13.5x - Multiple generated by the $800 million growth fund. Early-stage large-fund underwriting: 10% of a $10 billion company - Illustrative ownership needed for a $1 billion fund to return via one investment.

Pivotal Quotes: "common knowledge in venture is usually wrong" — David: Used to emphasize skepticism toward industry narratives and the importance of checking primary data. "the best founders can raise an infinite amount of capital from anybody they like" — US manager quoted by David: Explains why managers must offer more than capital to win access to elite founders. "the plural of anecdote isn't data" — David: Highlights the danger of drawing strategy conclusions from a few success stories.

Implications: LPs should prioritize manager quality, data rigor, and long-term consistency over venture clichés. Small-fund hype, reactive politics, and early paper marks can mislead; the real edge is backing repeatable winners in a power-law market.