Beta can also refer to relative volatility. A stock with a beta of 2 has twice the fluctuation of the index.
But in terms of asset allocation, you're right: PG is actually looking for alpha. You can get all the beta you want with leverage: just buy short-term out-of-the-money calls if you want extremely high beta.
(There's actually a lot to think about, here. Startups are pure "alpha" since the beta component is an expected return of zero.)
Ah, thank you. Beta is a coefficient of index performance, not just a measure of more or less correlation. So high beta would mean that when the startup market does well you do awesome. You'd also tank when there's a bust, but because YC's downside is so limited this is irrelevant.
I meant we're looking for high risk, high reward startups (with an implicit claim that the biggest winners will seem very risky initially). I've heard beta used, possibly inaccurately, to describe that sort of volatility. Is that not right? Since you actually worked in this world I'll take your word for it. In the meantime I changed beta in the article to variability.
"Beta" has two popular definitions. One is "market beta," i.e. the return you've gotten from the market you're in versus the particular trades you make in that market. I have only heard professional investors refer to this. The other kind is the one defined by investopedia. I only hear this mentioned by retail brokers.
If PG said he wanted high alpha, it wouldn't make sense to increase the probability of failure. All else being equal, higher alpha means a lower chance of failure (at any given volatility level, the higher-alpha portfolio has a lower chance of going to zero, since e.g. if beta is -99% and the lower-alpha portfolio generates -1% alpha, that takes it to zero. Right?
Return of a security = return due to correlation to the market + return due not due to correlation to the market.
Rs = BRm + A
It doesn't speak to general volatility at all. The definition linked is inane; Beta of 2 only twice as volatile as the market if the security is highly correlated to the market.
Beta is literally cov(Rs, Rm) / var(Rm).
In this particular case, if he wanted returns highly correlated to beta he would just invest in VCs. To get a beta of 2 he'd leverage the funds somehow and invest that.
He thinks he has an edge, which is contained in alpha (return not captured by market performance.)
I don't know. Still sounds to me like he's using the linked definition. Which I agree is not a useful concept.
People do cite betas. e.g. here's a quote page for Yahoo, on a big personal finance site, that cites the "beta" as defined earlier: http://www.dailyfinance.com/quotes/yahoo-inc/yhoo/nas It's entirely possible for a typical investor to only encounter that version.
Colloquially, if someone says "I like high-beta stocks," they mean "I like really volatile stocks," not "I like stocks that have some combination of high volatility and high correlation to some index." So the statement makes sense colloquially.
A more accurate version would be "YC looks for a moderately negative beta and expects a very high alpha." Since that describes companies that fundamentally alter their target industries--e.g. by harming all the established players, but still making money.
But in terms of asset allocation, you're right: PG is actually looking for alpha. You can get all the beta you want with leverage: just buy short-term out-of-the-money calls if you want extremely high beta.
(There's actually a lot to think about, here. Startups are pure "alpha" since the beta component is an expected return of zero.)