Basically, risk spreading suffers from unintended consequences. There are however other alternatives to the portfolio approach that do make more sense. My favorite is the concept of a keiretsu ( http://en.wikipedia.org/wiki/Keiretsu ). This approach makes sense, especially when you have potential co-dependencies between startups in a portfolio. The YC portfolio is generally large enough and the group activities create enough comraderie between startups that it functions like a keiretsu because I often hear about one startup using the services of another startup.
I think it could make sense at the level of investor portfolios. If I were accepted into YC, I would be open to the idea of giving up a small percentage into a YC "insurance" fund. The same would apply to a few investors (Sequoia, Kleiner, Benchmark, A16Z, Greylock, etc.), but for anyone other than the top funds, I think such a fund would be a losing proposition.
That cartoon is lampooning the MBS pooling that let agencies rate bundles of loans as AAA even though all the underlying securities were more risky. Bundling actually does reduce risk, but in the case of the mortgage meltdown of 2008 it was missing the forest (systemic market-wide mispricing of loans) for the trees (slightly reduced risk).
In this case, the founders already are invested in the dead cow, and so it can make sense to diversify to reduce risk. It's doesn't increase the value of their shares, it just makes the overall portfolio less risky.
The analogue to the MBS fiasco would be if the U.S. hit another depression, the whole YC class would be likely to flop, so the diversification wouldn't help, but in "normal" market conditions, the whole group would benefit from the few winners and get a payout in more scenarios.
http://dilbert.com/strips/comic/2008-12-13/
Basically, risk spreading suffers from unintended consequences. There are however other alternatives to the portfolio approach that do make more sense. My favorite is the concept of a keiretsu ( http://en.wikipedia.org/wiki/Keiretsu ). This approach makes sense, especially when you have potential co-dependencies between startups in a portfolio. The YC portfolio is generally large enough and the group activities create enough comraderie between startups that it functions like a keiretsu because I often hear about one startup using the services of another startup.
I think it could make sense at the level of investor portfolios. If I were accepted into YC, I would be open to the idea of giving up a small percentage into a YC "insurance" fund. The same would apply to a few investors (Sequoia, Kleiner, Benchmark, A16Z, Greylock, etc.), but for anyone other than the top funds, I think such a fund would be a losing proposition.