Hacker News Reader

How does early stage jump from 5.8 to 68.8% when going from 10 to 15 years? It seems there is some small sampling issues going on there.

Anyway, you can take a certain time frame, slice the start/end dates to match when certain companies sell or are the market is doing well and make your point look in either side's favor, especially for short time intervals like 1-3 years. Also curious as to how NVCA collects this data.

View on HN

Absolutely fair critique. Cambridge Associates and Preqin are primary data providers in this world, but in talking to our VC customers, we've heard that sample size, reporting bias (funds doing well may be more inclined to participate) and large time lags are a problem with judging VC performance based on their data. Some studies have also found that the same fund can have different returns in each system.

That said, it's the best of what's available today.

The best indicator or proxy that returns suck is that LPs are voting with their wallets and leaving the asset class and investing in it less.

10 year doesn't include the 1999-2000 bubble era, 15 year does. Those couple of years were among the best periods for VC returns ever, hence any period including them will have disproportionately high #s relative to other periods.

Reply on news.ycombinator.com