Quote from nitro:
...It should b 0 ~ ES - FV + VIX^(FractalDimension(SPX))
I thought it would be fun to take a look at this "model" again. As you may recall, the reason this relation was introduced is because we noticed that "FV" was getting really out of whack during the relentless sell off, and taking our thinking analogously to physical models, in particular conservation laws and the way particle physicists keep track of energies in a particle collision, we decided that the "FV" scalar number did not take into account all the market "energies". So, I suggested that in order to make the "model" give a complete accounting of where the energies are going, that we had to add a correction term, VIX. This gave us a correction term where our model could still keep an equilibrium of prices around "FV".
So how has this model held up? Well, we didn't just introduce a value just to make things work. It had to make sense. As we can see, "FV" and SPX have converged. Would then not the VIX scaling factor also have to go down? Has it? The answer is yes. VIX was approximately 54 when "FV" and SPX were so out of whack, and VIX is closer to 40 now. This is not surprising given our understanding of VIX, since it tends to go lower in rising markets. However, one thing occurred to me while thinking about this.
VIX is often seen as a "fear" indicator, which seems to suggest that it has a higher absolute value when the SPX is going down. But this is a bad way to think about VIX imo. What VIX measures is not so much fear, but unexpected moves away from means in some distribution [for experts, a path integral of
correlated big price swings.] So it is a single number that [indirectly to the untrained eye] measures skewness and kurtosis. It just happens that prices have greater momentum on the way down than on the way up, and hence VIX seeks a higher absolute value in this case. Further, since VIX incorporates information about the vola skew, it also a measure of correlation risk, and it is this correlation risk that markets [people] want to avoid.
One interesting thing to me is: Before the 1987 crash, there was little or no skew in SPX vola prices. Thereafter, this changed. I think VIX is not efficient in the same way that vola was not efficient pre-1987 crash in that VIX should also rise on strongly rising markets. It is an intuitive statement that I cannot prove.
Anyway, just thinking out loud.