Three years is a very short timespan from which to make claims about asset classes. In most cases, seven years is considered the minimum for a true sampling of baseline performance; ten years is better, and more than ten is better still. Obviously a16z hasn't been around for ten years, so metrics like three-year IRR are the best we have. That said, it's silly to take a three-year IRR and benchmark that confidently against something like the S&P 500.

On the other hand, I would strongly suspect that the top VC firms massively outperform the VC industry as a whole, due to any number of factors, including deal access, ability to secure favorable terms, ability to make new rounds or exits happen, etc. In time, a16z's longitudinal performance may well beat the market. But it's way too early to call the ball.

I agree that you can't really look at the aggregate performance of the entire VC industry. It's probably a highly skewed distribution, with almost all the big returns going to a handful of funds.

I don't know about that. Over that time period, post 2008 crisis, many funds (PE, fixed income focused HF's and equity funds) did very high returns as most asset classes bounced back from the crisis depths. Helped along, of course, by unprecedented money printing and credit expansion by the Fed.