What’s Broken – Venture Capital or Venture Perceptions?

There is an oft-repeated mantra than venture capital is broken.  The evidence cited is the dismal absolute median returns of the venture capital industry.   It is true that the venture asset class has become more competitive in the economic sense, i.e., excess returns are more difficult to achieve.  That is to be expected.   Excess profits attract competitors and returns decline.   But that doesn’t mean all competitors converge to the equilibrium return. It just means that the winners will either have the same strategy as every one else with better execution, or simply better strategies.

Cambridge Associates is a leading consulting firm that benchmarks the venture capital and private equity industry.  Nearly everyone uses their data to assess whether investment performance of a VC firm is “top quartile” or not.  The reason is simple.  Top quartile performers garner the lions’ share of the profits.   Between 1994 and 2006 the returns of the top quartile funds compounded in value at rate nearly 7X that of the median venture capital firm. Cambridge recently compiled some statistics about the venture business which document this. The data are proprietary to Cambridge, but I will share with you a high level view of the results, with their permission.  Of course, the following discussion reflects only my own conclusions.

Once Upon a Time, Everyone Was Top Quartile
The 1994-1996 period was an exceptional time to invest. If you were the first investor in a company in that time period your average internal rate of return (IRR) was over 100%.  Repeat -- your AVERAGE IRR was over 100%!  Not surprisingly, when these excess returns became apparent in 1997-2000, the number of venture firms exploded, as did the amount of capital.  We all know this, and have observed the consequences for the past nine years. The period of 1994-1996 was the Wonder years.  Post-2000 has been more like the Wonder What Happened? years.

The prevailing wisdom from the pre-2000 period was that early stage venture capital is the place to be.  That same wisdom today is be anyplace but venture capital because the old model doesn’t work.  The facts from Cambridge shine a bright light on the alleged wisdom and illuminate a more nuanced, insightful understanding of what drives performance. 

Timing is Everything - Who Knew?
The data in the table below are a paraphrase of Cambridge’s findings.  (If you want the original data, you'll need to become a Cambridge client).  They categorized over 22,000 venture capital initial rounds of investment in a thirteen-year period.  Each investment round was classified from Seed stage to Public, and they calculated the IRR of the initial VC investments made in each year. For example, if Crosslink led a B round investment in NewCo in 1998 as our initial investment, the outcome of our initial and subsequent investments in NewCo would be attributed to the original 1998 investment decision.

I have simplified their presentation to indicate which stages in each year led to IRRs higher or lower than the average IRR of all investments in that year.  Basically, it is scorecard of what kinds of investments you should have been making each year.

The pattern clearly shows that the miracle years of 1996-1999 were all about early stage.  Here’s where the perception that early stage is everything became axiomatic.  Note that as the business cycle shifted in 2000, the best investments were no longer early stage. They were restarts, PIPEs and some late stage deals.  As the economy began to re-heat in the 2005-2008 period, the pendulum again swung back to earlier stage deals.

These data are consistent with our own experiences at Crosslink.  We raised Crosslink Ventures IV in 2000 and Crosslink Ventures V in 2006.  (Elsewhere I have discussed our sector-focused, stage-independent strategy of growth equity investing.)  Our results are consistent with the Cambridge data on venture performance over the same period.    We have found that an ability to move with the market, both by stage and by sector, is a critical factor in achieving top quartile performance or better.

History is Bunk
What are we to conclude from the data?  While market and entrepreneur selection remain fundamental to venture manager success, risk management, time-to-money, and inflection point investing have emerged as perhaps equally determinative of  top quartile performance.  We all know the ‘hits’ nature of the business has been dampened, perhaps forever, but that doesn't mean venture capitalists cannot achieve premium rates of return. The Cambridge data show the business has transformed into a mature asset class where relative performance varies across the business cycle. What’s broken is not the industry, but the outdated perceptions of what strategies drive top quartile performance.

The venture industry has disappointed for a long time, or so it seems.  Certainly that is true when measured against the exceptional period where the average IRR was 100%.  Popularity and profits are incompatible concepts over the long run.  Nevertheless, top quartile venture performance is still exceptional, relative to other asset classes. Below is a comparison of the SP500 performance over the past nine years to the venture funds of vintage year 2000.

Some might counter that the SP500 investment is 100% liquid and venture numbers are merely private equity accounting entries.   However, the $1 invested in the SP500 would be a fully "realized" $0.65 at 2008 year end, while even the median venture fund would have returned $0.48 of cash, plus an illiquid remainder.  The top quartile fund would likely have returned $0.90 or more, trumping the SP500, and still have a substantial upside.  

Are We Blinded By A Remembrance of Things Past?
So is the venture asset class broken?   It doesn't appear to be.  It seems our assumptions of what constitutes "venture performance" are anchored in an obsolete memory of absolute performance, when the world of investing is all relative.  The mythology around venture investing reinforces this belief. Like every business, the successful venture investment strategies require continuous re-invention, not just a repetition of what worked before. Venture capital certainly isn't a license to print money anymore, but it is still the best game in town.  It's just that the rules of the game have changed.

Comments

[Tony Bain]

Changing the rules of the VC game of course has the impact of changing the rules for the whole start up sector. What do you see as being the most significant knock on effects?

[Cem Sertoglu]

Great insight, Peter. Thanks. I wonder what this would look like for European VC.

[Don Geddis]

Your By-Stage IRR chart almost suggests that there were just a handful of successful companies. Roughly speaking, a seed stage investment in 1998 becomes an early stage in 1999, becomes expansion in 2000, becomes public in 2001. That almost seems to imply that there were zero good new ideas, from 2000-2005.

OK, the analogy isn't perfect. But it seems awfully suspicious that every year, the best category to invest in is a company about a year older. That makes it seem like the diagonal line from 1998 to 2002 is really just the exact same tiny set of companies, every single year.

[Peter Rip]

Don:

There are >1000 investments tracked in each year of the data. So the suggestion that it's a few companies each year that persist as best deals is highly unlikely. Also, I think the assumption of stage of maturity equating to one year is also too short. Since 2000, the average time between rounds has stretched, as the prevailing view became that capital too expensive in 2001-2005 and funding takes longer to complete. Also, the mean time to liquidity has stretched from four to eight years since 2000. So I am guessing that stages are close to two years than one.

Actually, I think the data understate the real stage-based variation. The IRRs probably have a log-normal distribution. So the means are probably skewed high. If we did the same analysis on median IRRs, you might actually see a slightly more pronounced effect because the marginally positive stages each year would go red from green.

[Healy Jones]

Is the first chart vs the mean of other VC-type investments or is it vs the mean of all available investments (such as bond funds, S&P, etc?) What is the right investment class against which to compare venture funds, given their illiquidity, long hold periods and fees? I certainly like the second chart, vs the S&P, because this seems like a real asset against which to compare the venture class. However, how does this analysis change after the first quarter of 2000 is cut out? There were some big exits early in 2000 that could be potentially be skewing the results for venture? I don't know the answer, just asking the question.