My FVE model calculates "fair value" of VIX based on several assumptions:
1) After 2008 market crash, realized volatility leads implied volatility
(estimated 70% of time)
2) Derivatives market are zero-sum games, thus you would often see
forced liquidation of losing bets (sometimes large), adding to more
frequent over/under valued situations
3) Volatility, although mean-reverting, exhibits strong trends.
3a) Many market makers do not utilize an "absolute value' model but rather
"relative value", or they piggy back off of whatever bid/ask prices are,
thus would add fuel to volatility moving from one equilibrium to
another.
4) Implied Volatility is a reflexive function of realized volatility of the underlying + statistical relationships
on supply & demand of options (as reflected in implied volatility or
VIX) based on characteristics movement of the underlying.
4a) VIX in general is inversely correlated to S&P500 index direction
4b) The more and faster investors are losing money, the more they would seek out protection
4c) I believe whether the underlying is in a trend or a range affects supply and demand for options.
FVE indicator and simple trading rules takes into account these assumptions in a
crude (yet seemingly very effective and relatively simple) model. I am not a programmer or a mathematician, so anyone with
knowledge and skills in these areas could probably
come up with a superior volatility (VIX) model. I hope my
contribution would be my insight into the market volatility.
I do believe in a
top-down (qualitative-->quantitative) approach, where one would
detail assumptions about a market (current environment), create and
apply effective algorithms, then test the algorithms in their
effectiveness. I guess the other approach would be bottom-up through
data mining.
I would love to learn about assumptions made in other models. Thanks.
No comments:
Post a Comment