Lies, Damned Lies, and Venture Performance Statistics

Don Dodge has up a post highlighting recent poor venture capital asset class performance. I responded in his comments as follows:

Don — Interesting data, and stuff I have obviously seen many times, but I’d be careful drawing many conclusions from it.

a) It is misleading to match invested capital and return proceeds in any given year. The lag between disbursements and receipts is 4-5 years, so money put out 2003 or later is pretty hard to value.

b) The skewness in returns means that most of the profits went to the top decile within the top quartile. In other words, while the asset class’s performance overall has been poor, that is not the same thing as saying that the top-performing funds — a stable and small group — have not done well.

c) Returns immediately prior to 2001 were very good indeed. 2000 and 1999 netted LPs 211% and 134%, respectively. 1999 and 2000 vintage funds, however, have been no hell, reinforcing my point in a) above.

d) While we are regressing to a performance mean after some standout years, it is true that the venture asset class has never had back-to-back negative returns years until 2001/2002.

e) The next three years will tell a great deal about the future of venture as an investment, and I’m actually fairly bullish, with the market wanting more risk, some early-stage companies performing well, and an uptick in IPO activity likely ahead of us.