Skip to content
Independent essays · the editorAbout the publication ↗
The Longer Look
Read the source check and unresolved evidence
5 May 2026
The founder experience

What the Natural Experiment Shows

Reading and evidence guide: worked examples and checks

The US, UK, and EU run variants of the same venture model under different conditions. Comparing the three lets us ask which features of the system are intrinsic to running a venture model at all, and which are local-design choices that could be different. The line between the two changes what reform is possible.

The City of London skyline, 2023
City of London, 2023. CC BY-SA 4.0, via Wikimedia Commons

Three jurisdictions run a variant of the same model. The United States runs venture capital with mostly private LPs (university endowments, family offices, sovereign wealth, with public pensions allocating modestly), high carried-interest tax preferences, deep public-market exit infrastructure, and a tolerance for high variance in founder economic outcomes. The United Kingdom runs a smaller version of the same model with more publicly-anchored LP capital, EIS/SEIS founder-side incentives, a thinner late-stage exit market, and founder share-disposal and fund-manager carried-interest rules that must be considered separately. The European Union runs a still-smaller version with substantially more state-aligned LP capital (Bpifrance, KfW, EIF), shallower exit infrastructure, and slower deployment.

The three jurisdictions are not running controlled experiments. They differ on dozens of variables that nobody designed. Comparing their outputs can suggest hypotheses, but this essay does not establish a causal natural experiment. The question it raises is: which features of the venture system are intrinsic to the model, and which are local-design choices that could be different?

What is intrinsic to the model

The essay proposes four recurring features. Recurrence across these jurisdictions does not establish that they are intrinsic or unavoidable; comparable source evidence and alternative explanations still need assessment.

The first is the power-law distribution of returns. US, UK, and EU funds all show fund-level returns dominated by a small number of extreme outliers, with the majority of investments returning less than capital invested. This is not a US peculiarity; it is what the venture model produces wherever it runs.

The second is the founder mental-health pattern. US, UK (Sifted, Cerevity), and EU surveys all document elevated reported rates of depression, anxiety, and burnout in founder populations relative to baseline. The pattern is consistent across the three jurisdictions.

The third is the broad shape of the demographic skew. Funded-founder populations in all three jurisdictions skew narrower than the underlying population on gender, ethnic, and class dimensions. The magnitude of the skew varies (see below); the direction does not.

The fourth is the basic shape of the founder economic-outcome distribution. Most founders in all three jurisdictions, on the available data, do not capture financial returns commensurate with the labour input. The mean is positive and carried by rare extreme outcomes; the median is negative or modest.

What is a local-design choice

The essay proposes six dimensions that may vary across jurisdictions. Their variation does not, by itself, identify the effect of a particular policy choice.

The first is the LP composition. The US is dominated by private LPs with high return-rate expectations. The EU is heavily state-aligned. The UK sits between. This shapes what kind of company the fund is willing to back: state-aligned capital tends to allocate more deliberately to dual-use technology, climate, and public-interest sectors than purely private capital does.

The second is the sector tilt. US venture is disproportionately weighted toward consumer attention-economy products. EU venture is disproportionately weighted toward deep-tech, climate, and industrial software. This is partly a function of LP composition and partly a function of exit-market depth.

The third is the externality profile. Because sector tilt differs, the externalities the venture model produces differ. The US-style consumer-tech externality stack (attention, mental health, social cohesion) is not equally produced by the EU-style deep-tech stack. The model is the same; what it points at is not.

The fourth is the carried-interest tax treatment. The US treats carried interest as capital gains; the UK has tightened this in 2026; some EU jurisdictions tax it as ordinary income. The wealth-concentration outcome of the venture model varies with this choice.

The fifth is the founder exit-tax treatment. UK BADR/EIS, US QSBS (post-OBBBA July 2025), and EU member-state regimes differ widely. Founder economic outcome distributions are shaped by these regimes as much as by the underlying venture math.

The sixth is the magnitude of demographic under-representation. The essay argues that the skew recurs, but has not established that it is intrinsic; comparable measures are needed. UK and EU venture have somewhat broader demographic distribution on gender than US venture; US venture has somewhat broader on certain ethnic dimensions than European venture. None of the three is well-distributed; the distribution is not equally bad.

What this means for the debate

The natural-experiment framing produces a sharper question than the single-jurisdiction framing. When someone says “venture capital produces wealth concentration,” the natural experiment lets us reply: which kind of wealth concentration? The kind produced by carried-interest tax treatment is a local choice. The contribution of skewed returns requires separate empirical analysis. This country comparison alone cannot establish whether changing that distribution requires abandoning the venture model. Conflating them produces policy debates that are about the wrong thing.

The same applies to externalities, demographic skew, and founder-welfare costs. Some of what looks like “the venture system” is the system; some is the particular national implementation of it. The comparison suggests a distinction to investigate; it does not establish that distinction causally. What the line means for any particular policy proposal depends on which side of it the proposal is operating on.

One observation worth carrying out of the natural-experiment frame: the EU model produces fewer of certain externalities and channels returns through more publicly accountable LPs. It has also produced a persistent scale and breakthrough gap relative to the US. Most companies that reached global platform scale or created entirely new categories in the past twenty years did so under US calibration. The natural experiment does not tell us this is a fair trade. It tells us the trade exists. The question of whether the absent breakthroughs would have arrived through some other channel — state R&D, philanthropy, family-controlled firms, industrial-policy mechanisms — is the question the deep treatment engages and does not fully resolve.

The work that gets done by founders inside any of the three jurisdictions, and what is held and lost across the lives of those founders, sits underneath the natural-experiment frame as a separate layer the data does not capture. That layer is the subject of the publication's sister project, orphans.ai — what is missing from the public record of how the work is actually done, and from the data the next generation of machines is being trained on.


The full natural-experiment treatment is Part V of VC: Most Fail, Most Suffer, Some Win Lots. The operational counterpart for prospective founders deciding where to build is VC across the US, UK, and EU. The carried-interest and founder exit-tax treatment specifically interacts with the UK 2026 inheritance-tax reform, which is the subject of the publication's IHT category.

Source check and limits of this comparison

Review added 8 November 2026. This AI-assisted source check corrects the distinction between founder capital gains and fund-manager carried interest. It is not independent specialist review of the essay.

Founder exit gains
These are not automatically ordinary income. HMRC’s Business Asset Disposal Relief guidance gives an 18% rate for qualifying disposals from 6 April 2026, subject to eligibility and limits.
Carried interest
HMRC’s rate guidance places carried interest received from 6 April 2026 under Income Tax and National Insurance. This is a different category from an ordinary founder’s share disposal.
What the country comparison can establish
Different outcomes across jurisdictions can suggest hypotheses. Without a defined intervention, comparable populations and a defensible causal design, they do not identify which outcomes are inevitable consequences of venture capital.
Claims still requiring evidence
Cross-country claims about typical founder returns, mental health, demographics, sector mix and externalities need comparable dated studies. This review has not validated those claims. They should be read as the essay’s arguments, not established comparative findings.

The original publication date is retained. Read the original sources and the publication’s method before relying on this material.

Reading notebook

Before you decide, check these points against the current source.

Things to check