The best available UK data, NESTA's Siding with the Angels study from 2009, found that 56% of exited angel investments returned less than the capital invested. Put plainly: on most individual deals that reached an exit, the angel got back less than they put in, and a large share of those were a total or near-total loss.
That is the number behind every honest conversation about angel risk, and it is the one this page is about. The live explainer on the real risks of angel investing tells the story; here we stay with the figures, where they come from, and how old they are. Because the awkward truth is that the strongest UK loss data is now well over a decade old, and nothing newer has replaced it.
Most individual angel deals lose money. The portfolio survives on the few that don't.
The headline number, and where it comes from
56% of exited UK angel investments returned less than the money put in. That figure comes from NESTA's Siding with the Angels (2009), a study of 158 UK angels and 406 exits, with data running to late 2008. It remains the most-cited realised loss rate for British angels.
Read it the right way and it isn't a verdict of doom. The same study put the mean return at 2.2x over 3.6 years, so the average angel in the sample more than doubled their money across a portfolio. The high failure rate and the positive mean sit together precisely because of how early-stage returns are shaped: a lot of small losses, a few large wins. One does not cancel the other. The loss rate describes the typical deal; the mean describes the book.
And the data is old. Late 2008 is a different funding world from 2026, with different cheque sizes, a far larger angel population and tax reliefs that have moved since. The honest position is that this is still the best UK realised number we have, not that it is current.
The full distribution: losers, winners and the tail
The 56% has a mirror image: 44% of exits made a gain (NESTA 2009). So the picture is not that almost everything fails, but that more than half of exited deals come back light, and a minority pay. What turns that minority into a positive book is the shape of the winners.
About 9% of exits returned 10x or more (NESTA 2009), and that thin tail produced most of the study's total return. This is the power law in raw form, not as a slogan: a single 20x or 30x outcome can outweigh a dozen write-offs, which is exactly why the median deal can lose money while the average deal makes it. If you strip the top few exits out of the sample, the returns collapse.
An earlier UK study points the same way. Mason & Harrison's Is it Worth It? (2002) found 34% of investments were a total loss, with roughly 47% of exits negative on one cut. Older still, and from a smaller informal-investment market, but a useful corroborating read: across two independent UK datasets, a decade apart, the loss rate sits in the same broad territory.
Why the loss rate is structural, not bad luck
A first-time angel who loses on more than half their exits has not necessarily picked badly. They have met the base rate. Early-stage companies fail often: they run out of money, the market doesn't show up, the founding team breaks, a bigger player arrives. None of that is unusual, and a model built on backing them has to assume most won't make it.
This is the part the headline number can obscure. The 56% isn't a sign the asset class is broken; it is the asset class working as designed. The return doesn't come from being right most of the time. It comes from being very right occasionally, on a deal that pays for all the others. That is why a single angel investment is closer to a lottery ticket than a savings account, and why the qualitative risk narrative matters as much as the arithmetic.
The practical reading sits one level up, at the portfolio. Angel portfolio maths walks through what the realised loss distribution does to a notional book of names, and why a power-law return needs enough shots to have a chance of catching the tail. The loss rate and the spread of a book travel together; you can't think about one without the other.
What softens a loss without changing the rate
Here is a distinction that trips people up. SEIS and EIS loss relief makes a failed investment cheaper. It does not make failure less likely. The two are easy to blur, and blurring them flatters the maths.
The mechanism is factual: when a qualifying company fails, loss relief lets a UK taxpayer set the net loss (after any income tax relief already claimed) against income or capital gains, which recovers part of the capital at risk through the tax system. On an SEIS deal the combined effect can leave a higher-rate taxpayer carrying only a fraction of the original cheque on a write-off. That is a real cushion, and it is one reason angels can stay in the game across a string of failures. The mechanics are set out in the live piece on how loss relief cushions a failed investment.
But notice what it does and doesn't do. It changes your downside on a deal that fails; it does nothing to the 56%. A portfolio with generous loss relief still sees the same share of companies go under. Relief reduces the cost of being wrong. It is not a return, and it is not protection against the loss rate itself.
Reading the number honestly, and what this isn't
Two cautions before you lean on any of this. First, the data is dated, and openly so: the 56% and the 34% come from 2009 and 2002. Newer UK sources, including the British Business Bank's first angel section in its 2025 equity research, report activity and sentiment, not realised loss rates, so they can't give you a current failure figure. Where you see rosier numbers (the ERC's 2015 survey, for instance), check whether they describe expected returns rather than realised ones; expectations have consistently run ahead of outcomes.
Second, this piece is not a steer. It does not tell you angel investing is too risky, or that loss relief makes it safe, or what share of a deal you should risk. It reports what the named studies found, and flags how old they are. Reliefs and rules shift with each Budget, and the figures here need confirming at source. Read HMRC's guidance on the venture capital schemes, and take FCA-regulated advice before you commit capital.
Frequently asked questions
What percentage of angel investments fail?
The best UK data, NESTA's Siding with the Angels study from 2009, found that 56% of exited angel investments returned less than the money put in, with most of those a total or near-total loss. An earlier UK study, Mason and Harrison in 2002, put total losses at 34% of investments. Both figures are dated, so confirm the current picture at the source, but they remain the strongest UK realised numbers available.
Do most angels lose money on most deals?
On most individual investments, yes. The NESTA 2009 data shows 56% of exited deals returned less than capital. But the portfolio can still come out ahead, because a small tail of big winners carries it: about 9% of exits returned 10x or more in that study, and that handful produced the bulk of the total return. The same study put the mean return at 2.2x over 3.6 years.
Does EIS loss relief change the failure rate?
No. SEIS and EIS loss relief reduces the after-tax cost of a failed investment by letting a UK taxpayer offset the loss, but it does nothing to whether the company fails. The failure rate is unchanged; only your downside on a failure is softened. Treating loss relief as protection against the loss rate is a common mistake.
Is there more recent UK loss data than 2009?
Not for realised losses. The strongest UK realised loss figures are from 2009 (NESTA) and 2002 (Mason and Harrison). Newer sources, such as the British Business Bank's 2025 equity research and the ERC's 2015 survey, report activity, sentiment or expected returns rather than realised loss rates, so they cannot give a current failure figure. The 2009 number still stands as the best read, dated as it is.
Is this article financial advice?
No. This is general information about published research on angel investment loss rates, not financial advice. It does not recommend whether to invest, how much to risk, or which schemes to use, and the figures are dated and need confirming at source. Tax reliefs depend on your circumstances and change with each Budget. Check the current rules at GOV.UK and take advice from an FCA-regulated adviser before committing capital.