Haha. There is no common method. Every bank has its own proprietary method.
In some sense, you have to look at kurtosis- vol of vol. Actually, you need vol of vol to go from options prices to the VIX.
I'd look to maybe a Markov-switching model that has high vol and low vol regimes and start trying to monte-carlo it? Try and come up with a model that uses the same kurtosis and skew implied by the options market on the S&P. Obviously you do NOT want to assume normality in the underlying market when pricing vol of vol.
Of course, you could go the naïve route of treating the VIX as a stock. But that ignores mean reversion and fails to explain the positive skew, whereas a monte-carlo that takes kurtosis, skewness, and vol of vol into account might be able to. You could probably do this with some weird adjustments to GARCH, but you may just want to do a Markov Switching Model. If you look at a realized or implied vol chart for the SPX, you'll see that the market clearly tends to have low vol and high vol regimes.
I haven't done this before. I'm not sure I'm smart enough to do it now. I HAVE worked relatively close to the people who've needed to solve this problem. I wish I could help you more. This is a tough problem that is still worth serious money to get right, and since you're dealing with a derivative of a derivative of a derivative, it's tough to get the historical data to price this.
