Angel returns work on a power law: a small number of investments generate almost all the gains, while most return little or nothing. The headline figure that matters isn't the typical deal - it's the rare outlier, because in early-stage investing the best single bet can pay back more than every other bet put together.
That one sentence quietly demolishes the way most people think about returns. We're trained to reason in averages: a "normal" outcome, a number you can plan around. Public markets more or less behave that way. Early-stage doesn't. The average angel deal is a fiction - a statistic smeared across wildly unlike results, most disappointing and a couple of them enormous.
What is the power law, in plain terms?
Picture two ways outcomes can be spread out. In a bell curve - the distribution most of us carry in our heads - results bunch near the middle. People's heights work like this: nobody is a hundred times taller than average. A power law is the opposite shape. A few results sit miles out to one side and dominate the total, while the bulk huddle near zero. Wealth behaves this way. City populations. And venture returns.
Drop ten angel cheques into ten companies and the spread won't be polite. Realistically several go to zero. A few limp back roughly what you put in, maybe a touch more. One - if you're fortunate - does something absurd: ten times, fifty times, occasionally far beyond. That single result can be larger than the other nine combined. The portfolio's fate is decided by the tail, not the middle.
You're not trying to be right often. You're trying to be in the cheque when something becomes very large.
Why does the average return mislead you?
Because an average quietly assumes the thing it's describing clusters around a centre - and a power law has no meaningful centre. Say a book of twenty investments returns 3x overall. That sounds like each company roughly tripled. It almost certainly didn't. The far likelier story is that fifteen lost money, four broke even-ish, and one returned 40x and dragged the whole average up on its own. Same headline number, completely different machine underneath.
This is why the median outcome of an angel portfolio and its mean outcome can live in different postcodes. The median deal is often a loss. The mean is rescued by the outlier. If you build expecting the median - a steady, typical company that does fine - you've designed for the part of the distribution that doesn't pay the bills.
What is a "fund-returner", and why does everyone chase one?
A fund-returner is a single investment whose exit hands back at least everything you put into the entire portfolio. Deploy £100,000 across, say, twenty companies, and a fund-returner is the one exit that on its own returns £100,000 or more. Anything beyond that is profit; everything else in the book is, in effect, the cost of buying enough lottery tickets to hold one.
Professional seed funds are explicit about this. They model each investment by asking not "is this likely to work?" - because most things aren't - but "if this works, can it be big enough to return the whole fund?" A company that can only ever be a tidy 3x is, in power-law terms, almost a distraction: it can't carry the losses the distribution guarantees. The maths rewards backing things with a credible path to enormous, even when the odds on any one of them are slim.
Why does spreading the bets matter so much?
Because if all the return lives in rare winners, your chance of catching one in a single deal is poor. Put everything into one or two companies and the most probable result, given the base rates, is that you lose - not through bad judgement, but because the outlier simply wasn't in your two picks. Spread the same money across many investments and you raise the odds that at least one tail event lands in your book: the one that makes the whole sum work.
There's no magic number of deals, and we won't hand you one - how many cheques you can sensibly write follows from how much you can genuinely afford to lose, which is a personal question for a regulated adviser. The structural point stands on its own: early-stage returns are lumpy, and a portfolio built to absorb a lot of zeros looks nothing like one built on a single conviction bet. Concentration is how you express certainty. The power law is a reminder that, this early, certainty is mostly an illusion.
How do SEIS and EIS change the maths?
This is where the UK system earns its keep. Two HMRC schemes - the Seed Enterprise Investment Scheme (SEIS) and the Enterprise Investment Scheme (EIS) - work on both ends of the distribution. They return part of your money up front, and they soften the losses the power law all but promises.
Relief on the way in
SEIS, aimed at the very youngest companies, gives 50% income tax relief on up to £200,000 invested per tax year, with a three-year minimum holding period. EIS, for the next rung up, gives 30% income tax relief on up to £1,000,000 per tax year - or £2,000,000 if at least £1,000,000 goes into knowledge-intensive companies (broadly, those doing heavy R&D) - also with a three-year hold. Relief on either can be carried back to the previous tax year. In plain terms: a meaningful slice of every qualifying cheque comes back before any company has succeeded or failed.
Relief when it goes wrong - and when it goes right
When a company fails, loss relief lets you set the capital you lost, after deducting income tax relief already given, against your income or capital gains - so the true cost of a wipe-out sits well below the headline cheque. On the winners, gains on SEIS and EIS shares are exempt from capital gains tax if held for three years or more and income tax relief was received, so the outliers that carry the book aren't taxed on the way out. SEIS adds a capital gains reinvestment relief worth 50% of a gain reinvested into SEIS shares, on up to £100,000 a year; EIS offers capital gains deferral relief. For comparison, VCTs give 20% income tax relief on up to £200,000 a year with a five-year hold and tax-free dividends, but no loss relief.
On the company side, EIS sits within defined limits on gross assets, annual and lifetime fundraising, employee count and company age, with higher ceilings for knowledge-intensive companies - and some of these figures changed from 6 April 2026, so check the current numbers in HMRC's guidance: gov.uk venture capital schemes (EIS). SEIS-qualifying companies must be trading for under three years, have fewer than 25 full-time-equivalent staff and gross assets under £350,000 at the share issue, and can raise up to £250,000 in total under the scheme. Both schemes need advance assurance from HMRC before you invest, and the company issues SEIS3 or EIS3 certificates afterwards for you to claim. You generally need to be a UK taxpayer for any of it to be worth anything - the detail is in the HMRC guidance for investors.
None of this rewrites the distribution. Most of your companies will still struggle; relief doesn't make a weak business strong. What it does is lower the cost of being wrong and leave the upside untaxed - and in an asset class defined by rare, large winners, that quietly tilts the long-run sums in the investor's favour.
Why do the returns take so long to show up?
The power law plays out on a slow clock. Until a company is sold or, far more rarely, floats, your shares are illiquid - there's no daily price, no exchange, no guaranteed buyer if you want out. A meaningful exit commonly takes seven to ten years, sometimes longer, and plenty of companies quietly fold long before any of that. The outlier that carries your book, if it arrives, tends to arrive late. Capital committed here should be money you can leave alone for a long time and, candidly, money you could lose.
So what does the power law actually mean for a portfolio?
It reframes the whole exercise. Once you accept that returns are concentrated in a few outliers, "I picked a dud" stops being a verdict on your judgement and becomes the expected texture of the asset class. The questions shift accordingly: not "will this work?" but "if it works, can it be enormous?"; not "how do I avoid losses?" but "have I written enough cheques, over enough time, for a real shot at a tail event?" That's the model the rest of the early-stage world quietly runs on - and the one most newcomers have to unlearn an average to reach.
The Carry is editorial journalism, not financial advice. This article explains how angel returns and the related UK tax schemes work; it does not recommend any investment. Early-stage investing risks total loss of capital. Seek advice from an FCA-regulated adviser before investing, and rely on current gov.uk guidance for tax figures.
Frequently asked questions
What is the power law in angel investing?
The power law describes how early-stage returns are distributed: instead of clustering around an average, a small number of investments produce almost all the gains, while most return little or nothing. In an angel portfolio the best single deal can out-earn every other deal combined, so the shape of the distribution - not the typical outcome - drives the result.
What does it mean for an angel investment to be a fund-returner?
A fund-returner is a single investment whose exit returns at least the entire amount put into the whole portfolio. If you deploy £100,000 across many companies, a fund-returner is the one exit that hands back £100,000 or more on its own. Power-law portfolios are built on the hope of catching one or two of these, because the median deal does not pay the bills.
Why do most angel investments lose money?
Early-stage companies fail at a high rate, so most individual angel cheques return less than they cost, and many return zero. That is the expected behaviour of a power-law asset class, not a sign of bad picking. The returns are concentrated in rare outliers, which is why a diversified book and a long time horizon matter more than being right on any single deal.
How do SEIS and EIS change the maths of angel returns?
SEIS offers 50% income tax relief on up to £200,000 invested per tax year, and EIS offers 30% relief on up to £1,000,000 (or £2,000,000 if at least £1,000,000 goes to knowledge-intensive companies), both with a three-year minimum hold. Loss relief cushions failures and qualifying gains are exempt from capital gains tax if shares are held three years or more with relief received. The reliefs lower the cost of being wrong and leave the outliers untaxed - they don't change the underlying distribution. See gov.uk for current rules.
How long does it take to see angel returns?
Angel returns are slow and illiquid. Shares in a private company cannot usually be sold until the company is acquired or, far more rarely, floats, which often takes seven to ten years or longer. There is no daily price and no guaranteed buyer, so capital should be money you can leave invested for a long time.