My idle thoughts on tech startups
Why Average VC Returns Don’t Really Matter
The VC industry (both the GP part and the LP part) pays attention to the sector’s returns, but the broader tech ecosystem only occasionally tunes in. Typically it’s when Cambridge Associates releases their benchmark data on the VC asset class (here’s a 2010 example article from TechCrunch) or an organization like the Kauffman Foundation publishes a white paper (a 2012 example article from Business Insider here). And to be fair, the sector overall has done poorly in terms of returns in the last 10-12 years despite the fact that some funds (both established and new) have done very well.
But the reality is that average VC industry returns don’t really matter to many people. This is true for both LPs and GPs for several reasons:
1) No Tradable Index – VC funds are private investment vehicles, making investments in private companies. LPs investing in venture hold a subset of all the funds in the VC universe by design (see #4 & 5). And even if they wanted the index return, there is essentially no way to buy (or sell) a broad-based basket of VC funds in the way you can trade the S&P 500 or Russell 2000 or other public equity index.
What about fund of funds (FoF), you ask? Aren’t they a form of bundled investment into the VC asset class? It’s true that FoFs provide LPs a way to purchase VC funds in a basket, but by design these are comparatively narrow actively-managed investment funds rather than broad-based passive vehicles. FoFs have a range of strategies of course, but broadly speaking LPs that invest in FoFs pay them a management fee and carried interest (on top of the fee & carry of underlying VC funds they invest in) for access, diversification, active management or a combination of all three. Access in the sense that some FoFs have relationships with top-performing funds that aren’t open to new would-be investors. Diversification in that a small investor (by large institutional standards), say a foundation with $25M in assets that wants to commit $1M (4%) of their assets to VC, can get exposure to 10+ individual VC funds through a FoF but might only be able to inveset in 1-2 if they tried to do it directly. And active management in the sense that many institutional investors feel that they are better served with having a specialist manager overseeing their allocation to VC rather than trying to do it with their in house staff. But no matter how you slice it, FoFs don’t really represent a broad-based way to invest in VC.
2) No Synthetic Alternative – If an LP can’t “buy” VC as an index, could they replicate the returns of an index some other way? For example, if there were no S&P 500 ETFs or index mutual funds an investor could easily buy a “synthetic” equivalent by buying shares of all the S&P 500 stocks in proportion to each company’s weighting in the index. The same is not true for venture capital of course, since the underlying startups VCs invest in aren’t publicly selling their equity. Yes accredited investors can theoretically buy shares in a small number of very late stage startups through secondary exchanges, but these are a tiny slice of the VC-backed startup ecosystem and of course the largest returns accrue to the earliest investors in these companies not the latest ones.
3) GP’s Care About Absolute, Not Relative, Returns – Of course most GPs pay attention to the broader returns of the market, but at the end of the day absolute returns are what matter. They matter in the sense that GPs only earn carried interest if they generate absolute gains on their portfolio, regardless of how their fund does relative to other VC funds or even other asset classes. Even if you’re top decile for your vintage year (e.g. in comparison to all other VC funds raised in a given year) or you blow away the S&P index over the same time period, a 1.0x fund or marginally profitable fund doesn’t generate any carry for the GPs that are investing it.
4) Venture Is Driven By Exceptional, Not Average Outcomes – The entire VC business is based upon outlier outcomes, not the average. This is acutely true at the underlying startup investment level… across a broad range of vintage years and fund sizes, the bulk of VC fund returns are driven by a small minority of the portfolio companies that produce 10x+ returns for their investors. By definition, the “average” investment in a VC fund that produces a “good” return (2.5-3.0x+ cash on cash) is profitable. But the median investment is almost certainly a middling return if not a modest loss. Startup outcomes are a power law distribution rather than a standard distribution.
This impact of exceptional returns is diluted some at the fund level due to portfolio effect, so the distribution between the best and worst funds is far narrower than the distribution of startup outcomes. The best startup investments are routinely 50-100x+ returns whereas a grand slam fund return is something more like 5-10x. So LPs aren’t necessarily trying to pick the one or two VC firms that will have the standout return of a given vintage year. But most typically are trying to assemble a roster of VC managers and select those groups that they believe have a disproportionate advantage for sourcing, selecting, and winning the underlying startup investments that are exceptional “head end” outcomes.
5) Long Term Alpha-Seeking LPs – A good majority of LPs actively investing in venture take a long-term view of the asset class. Yes, some investors come and go (tending to sell low and buy high) and over the decades some new categories of LPs may enter the investment category. Within a large institution, the VC asset class may be “competing” for allocation with a very broad range of investment types. And even long-term LPs committed to investing in venture will periodically rebalance their VC portfolio, swapping out underperforming managers for better ones or shifting allocations between different segments of the VC sector (different industries, stage focus, geography, etc). But these long-term LPs are by definition using active managers in search of alpha (i.e. better than the market average), rather than trying to achieve an index return within the VC asset class.
So the average returns of the venture industry don’t have a meaningful impact on either GPs or LPs, and in reality they probably have little direct impact on entrepreneurs. In the very long term (5-15+ years), sustained increases in average VC industry returns may attract more capital in aggregate into the system just as bad returns may decrease the total capital committed to the asset class. But even though it’s episodic, the ecosystem seems to pay more attention to average VC industry returns than is probably warranted.