An interesting comment on this piece from Amon Elders, posted here with permission:
Basing myself mostly on Kauffman venture capital report of 2012: “we have met the enemy, and he is among us”. The data in your report is mostly from venture funds raised between 1980-2001. This was the most profitable period of VC investing. Afterwards returns have taken a nosedive, see kauffman report. There seems one source in your report, Robinson and Sensoy(2011), that uses a database that includes up to 2010, and median PME figures are 0.82 and 2% net IRR. Which, given that it’s an illiquid asset-class, is bad. The average seems to be around 1.0 PME, and 9% IRR. Note that this is ‘up to’, what really should be done is remove the period between 1980-2000, as the distribution of the data has shifted. Moreover, it wouldn’t surprise me that the downwards trend in VC returns has continued over the past 7 years.
Either way, the median returns are quite bad, and my question was mostly why does OpenPhil believe they can find these right opportunities. Given that it seems almost impossible to predict if a new venture capital funds is going to be successful. Which makes sense, anything beyond >5 year horizon’s seems almost impossible to predict for humans, basing myself on the book superforecasters here. And the data is highly noisy, highly limited, with high bias in venture capital investing and it might be the same for these high-impact high risk opportunities
An interesting comment on this piece from Amon Elders, posted here with permission: