This is based on data analysis we did at Beacon Capital (see acknowledgements at the end of this post) and it was first presented by Chris Mairs, CBE, FREng, and Beacon Capital Venture Partner, in his keynote speech at our event “London Tech Venture Capital: An Asset Class for Global Investors”, on November 18th 2015.
There has been no shortage in recent months of inspirational (and aspirational) articles about the London venture capital and startup ecosystem (Local Globe’s post serving as a good example) pointing to the robust economic growth, world leading universities and favourable investor tax regimes to paint a picture of a thriving venture capital ecosystem and a land of opportunities for startups, VCs, and more importantly Limited Partners (LPs) – those who ultimately invest in VC funds which in turn invest in startups.
And yet, arguing that the London startup ecosystem is booming does not seem to have attracted significantly more LP investment. London venture capital (let alone the rest of the UK or Europe) as an asset class has no, or only tiny, LP allocations.
The issue is that LPs (understandably) look at historical mean performance of Venture Capital investments in different geographies and conclude that this is better in the US than elsewhere. But perhaps that is a simplistic view and it may make more sense to consider the spread of performance within a portfolio as well as the mean.
The more pertinent question therefore is whether the relative performance of the two ecosystems, the US and European/UK, are likely to “converge” and, if yes, how soon might this happen.
Relative Historic Performance of US and Europe
To try and shed some light on this question, we analysed two datasets of startups in the US and Europe that received one or more investments between 2003 and 2010.
The US dataset of nearly 6700 deals is much larger than the European dataset, which contains a little over a 1000 deals, but both sets are large enough to tell us something. The data is far from perfect and we’ve had to make various assumptions where data was missing. We only considered companies that have actually had a liquidity event – i.e. got acquired, had an IPO or went bust. For each investment we calculated the multiple return on the investment and then considered the dataset as a whole. The model is quite basic and we plan to refine this over the coming weeks but allows for some interesting observations.
NB: The X axis refers to the year the first investment was made. We did not include investments after 2010 because only very few have exited (due to the short time period) and hence this makes the analysis unreliable.
Looking at the chart on top, we can see that the US multiples (blue line) are indeed on average better than the EU (red line). However, it is very striking that the US standard deviations on the bottom chart, i.e. a measure of the spread of the returns within the dataset, are consistently higher.
What about the future?
We then applied a polynomial regression to the data to predict how the averages and standard deviations might look in the future. This is a notoriously unreliable business and of course does not take account of black swans or other unexpected global events. But so long as the relationship between macro conditions in Europe and the US follows the trends that existed over the past few years then this may tell us something about the relative performance of the two markets looking forward.
Although the US mean is still ahead, the EU is converging on the US and the EU standard deviation remains much tighter than the US. In other words, we are approaching the same level of returns but with much more predictability if we have a sensible sized portfolio.
Predicted Portfolio Performances in the US vs Europe
We then used the Central Limit Theorem to predict the performance of a portfolio containing 10 randomly selected investments per year based on these mean and spread predictions.
The red shaded area shows the performance of an EU portfolio within 99% confidence limits, and the larger blue shaded area shows the same for a US portfolio. The dark blue and dark red lines are the mean performance.
What we see is that although the EU mean performance is a bit lower, the red area falls entirely within the blue area. In other words, it is always more likely that the EU portfolio will be closer to its mean than a US portfolio, making the downside risk lower in the EU.
So one might ask is this anything more than lies, damn lies and statistics?
Well, the sanity check comes in the form of the increasing number of posts and articles arguing that the London ecosystem is becoming stronger and able to compete on a global scale but also from our very own experience having been actively involved in the local ecosystem for some time now.
A Golden Age for UK Tech Investment
Thinking about the US data for a moment, there is a very small number of mega unicorns like Uber and Whatsapp that significantly skew the data. Only the privileged few get to invest in these super outliers and it would be nonsense to build a strategy around finding the next Uber.
Coming back to the UK, a combination of the strengthening economy, presence of world class universities, the attractive investment tax regime and belief in the London ecosystem amongst investors has led to a significant increase in the private capital available to startups. This comes at a time when hands-on investors have spent many years working with entrepreneurs to build businesses efficiently in a capital constrained market. Their experience in the efficient use of capital and the fact that we have not yet suffered the early stage valuation inflation which has happened in the US is very significant. The return on an investment is the difference between the exit valuation and the entry valuation, diluted by the additional capital that the business sucks up in its journey. So buying into a capital-efficient business at an attractive early valuation is a great step towards good returns.
What does that mean for LPs?
The reality is that there is a small number of tier 1 VCs in the US who consistently get access to the best deals with consequent portfolio performance consistently towards the upper end of the blue shaded area. In a future post we will segregate the US data and we expect to see that after only excluding a very few top performing investments the mean drops significantly. If you are an LP with access to one of the top performing funds, then good luck and fill your boots! Unfortunately, these funds are almost always oversubscribed and access is tough. So, if you are not in that privileged elite then you should certainly consider the UK, where standard deviations are lower, there is more space for new LPs in well managed funds and the mean performance may well exceed all but the very best US funds.
All prudent LPs do of course take account of more than the headline numbers and should not get seduced by Unicorn valuations. Interestingly recent adjustments in publicly traded stocks including Twitter and LinkedIn  have shaken some investors’ general appetite for the Venture Capital class. However, we believe that these adjustments simply reinforce our view that markets with a smaller standard deviation, and hence less volatility, make more sense, so perhaps now more than ever before is the time for progressive LPs to refocus their investments towards the UK market.
A big thank you to the two INSEAD MBA students, Gautam Nangia and Gilles-Aime Harerimana who worked painstakingly for two weeks to get the data from various sources (Preqin, Thomson Reuters, etc) and Vasos Dramaliotis, now a graduate engineer, who performed the analysis.
Also published on Medium.