Venture Capital Is Good for Society and Bad for Most Founders

Venture capital is good for society and bad for most founders. Both halves are documented in the empirical literature. Most writing about VC handles one half or the other; this piece handles both.

Venture capital is good for society and bad for most founders. The system funds breakthroughs that other capital does not fund at scale. The system pays for those breakthroughs by inducing many smart people to take a personally negative-expected-value financial bet. Both sentences are documented in the empirical literature. Most writing about VC says the first and not the second.

For most founders, the venture path is a negative risk-adjusted financial bet — financially bad for the typical founder. The mean return across the founder population can be positive, because rare outliers carry the distribution. But the typical founder, in expectation, does worse on a risk-adjusted basis than the salaried counterfactual. The non-financial dimensions — the autonomy, the experience, the optionality — can still be worth it. The financial part, for the typical founder, is not.

The social side, three numbers

Pooled US venture returns to LPs: approximately 14.3% net over the long horizon (Cambridge Associates US Venture Capital Index), ahead of public-equity benchmarks [STRONG].

The technologies that came through the venture pipeline include the personal computer, the smartphone, the commercial internet, modern biotech, the lithium-ion battery stack, large-scale solar, mRNA vaccines, and the foundation models the current AI wave is built on. Public R&D was foundational in several of these — the BARDA and NIH funding behind mRNA platform research, DARPA in the early internet, ARPA-E in battery and solar work, and decades of academic research the foundation models inherit from. The honest argument is not that VC invented these technologies. It is that VC was unusually effective at turning the underlying technical possibility into companies that scaled, hired, and deployed at speed. That is a different claim and harder to attack.

The alternatives have been tried. Government-led research funds early discovery well; it is weaker at company formation. State-aligned strategic capital produces industrial scale; it is slower at high-variance bets. The venture system has out-produced its capital share on the social ledger.

The individual side, three numbers

Hall and Woodward (2010), American Economic Review, 22,004 venture-backed companies: the median founder makes less over the life of their company than they could have earned in salaried employment over the same years [STRONG]. The mean is positive only because rare extreme outliers carry the distribution. The typical founder takes a financial hit.

Cooper, Woo, Dunkelberg (1988), 2,994 entrepreneurs: 33% rated their personal probability of success at 100%, against base rates closer to 50% five-year survival [STRONG]. Founders systematically over-estimate their own odds.

Freeman et al. (2019): in a 242-entrepreneur sample, founders reported 2x depression, 6x ADHD, 3x substance-use disorders, 10x bipolar relative to comparison populations [MODERATE; the directional pattern replicates across multiple adjacent surveys, including UK Sifted and EU Cerevity data]. The study is of entrepreneurs, not VC-backed founders specifically; it does not establish causation. Whether the elevated rates reflect selection (people with these traits are over-represented among founders), treatment (the founder experience produces or worsens the conditions), or both, is contested.

The composite picture: most founders absorb a financial loss relative to the salaried counterfactual, expect not to, and enter a population with materially elevated reported mental-health conditions; whether this reflects selection, treatment, or both is contested.

Why both halves are true at once

The structural argument is four steps.

Fund economics. Venture returns follow a power law. A typical 25-investment fund needs at least one extreme outlier to return capital several times over.

Selection pressure. A fund that needs outliers cannot afford the merely-good. It must reject reliably profitable companies in favour of speculative ones. This is not error; it is the math.

Founder selection. A founder pitching a fund must persuade the partner the company is upper-tail. A founder who acknowledges median outcomes will not get funded. The selection mechanism rewards over-confidence on purpose.

The recruitment narrative. For the population of pitching founders to be large enough for outlier-finding to work, the broader culture must persuade many people they could be the outlier. The pitch deck, the founder podcast, the accelerator marketing, the success-story profile — this is fund economics operating on prospective founders before any partner meets any of them.

The aggregate output the system produces depends on individual readers being persuaded that the base rates do not apply to them. Most of those readers are wrong; the system needs them to act anyway. The individual messengers are not lying. The aggregate effect of locally-rational decisions is the pattern.

The decision you are making

If you are a prospective founder, you are deciding whether to trade five to ten years of working life, with a known salaried alternative, for the base rates above. The median founder finishes worse off financially. The typical founder will rate their probability of success well above the actual rate. The cohort you would be joining shows materially elevated reported mental-health conditions, with causation contested. None of these tells you whether you specifically will be the outlier. All three should be on the table when you decide.

If you are a venture capitalist, your model requires the recruitment narrative to keep producing the population of attempts. The narrative produces real welfare costs concentrated on the founders who are not the outlier. Your incentives do not currently price those costs; the fund books returns; founders absorb the rest. The asymmetry is structural and the math does not require it to stay invisible.

If you are a policymaker, the trade-off the venture system embodies is real and not easily improved upon by the alternatives currently on offer. Acting as if it is one-sided — either by adopting the cheerleading view or by adopting the critical view — produces worse policy than acting as if both halves are real.

The honest summary is the sentence at the top. It is good for society and bad for most founders. The two halves do not cancel. A reader who comes away persuaded that VC is net-positive for humanity is reasoning correctly given the evidence. A reader who concludes that they personally should enter the venture system is making a different decision — about themselves, not about the system — and that decision is the one the recruitment messaging is designed to influence.

Where to go next

The empirical detail with primary citations: The reality of being a founder.

The recruitment-narrative argument as its own piece: For prospective founders.

The structural mechanism in detail: The power law and what it forces.

The full-length analytical version (~26,000 words; PDF and zip): VC: most fail, most suffer, some win lots.