Saturday, January 31, 2015

2015.01.30 VIX Cycle Model Chart

2015.01.30 VIX Cycle Model Chart


















The VIX cycle model seems to have predicted a multi-month turning point for volatility. The model is suggesting higher average volatility  going forward into 2015 but the caveat is that it does not capture short term fluctuations in the negative or positive directions.  A previous iteration of the model is shown below and here on the blog.



2014.12.19 VIX Cycle Model Chart

6 comments:

Jigs said...

Paulo,
Can I request you to please post similar model for s & P 500 please?

Anonymous said...

Hello Jigs,

Paolo used to post forecasts for DJIA, but they diverged significantly from the eventual observed values. Equity markets are perhaps more prone to "non-intrinsic drivers" than other more liquid markets like Forex. By that, I mean manipulation from algos driving thinly traded futures to earnings reports and macro readings.

Paolo's VIX forecast has however been outstanding and has got into the "groove" of current market action. Combined with an anticipated rise in gold and silver prices, his VIX forecast suggests a corrective period ahead.

Mr. Anon

Paolo said...

I agree with Mr. Anon regarding equity index time series. Fortunately, both commodities and VIX can be characterized by cyclical influences, though VIX also has a "hour/minute-trade" spike component to it as well which is not predictable using cyclical equations.

Anonymous said...

Hello Paolo,

May I kindly ask a question about the Z-score for VIX? I would like to know if there is any predictability of the differences between actual and predicted.

For instance, is the average Z-score over time effectively zero or perhaps there is some bias/offset? I wondering if divergences can be exploited somehow, but I guess if the Z-score is too large then maybe caution is required.

Your recent rerun of the VIX model appears little changed - does that perhaps suggest that the model has now got into the "groove" of current market action? I noticed that previous reruns over the past 12 months or so have differed significantly from each other.

Given that VIX measures the implied volatility of S&P index options, itself a complex mathematical measure, I find it fascinating that structural modelling can capture future moves in what is effectively "market fear".

I do wonder if hedge funds and market makers are doing the same?

Mr. Anon

Paolo said...

Hello Mr. Anon, please excuse my tardy reply. It's weekend catch-up time! All good questions.

The "mean Z" is by definition zero across the time series. You are right about exploiting divergences, and a good example of it is in the current XOI.X index, where the index has 'gotten ahead of itself' in reference to the model. Meaning that the model prediction and actual price has diverged close to two sigma units. In general, in a well-regulated structural model, the higher the Z=score divergence from the mean, the higher probability going forwards of a reversal to the mean.

Back to the VIX model - I have been wondering about why volatility, as an index measuring time discounting behavior, might be characterized by a cyclical phenomenon. One hypothesis is that strong emotions like fear and greed may be cyclical based on the fact that tolerance develops to strong emotions develops in humans, which introduces a lag time based on how fast the emotion develops and then recedes.

There is one cyclical influence that might also be captured by the cyclical VIX model that is related, as outrageous as the claim may be, to certain ongoing geophysical parameters. I have noticed this from a neural trading model where those parameters are used as inputs for predictive purposes. The association is strong enough that I'm certain it has been discovered by others. There is some literature out there suggesting there may be an influence of geomagnetic fluctuations on mass behavior, so maybe it is not as outrageous as it seems. Figuring out the mechanism of the putative association would be a life's work!

Anonymous said...

Hello Paolo - many thanks for the reply on VIX divergence, no problem with taking time to answer. I now understand the nature of "mean Z" and the potential for mean reversion.

Your comment on geophysical inputs to trading models is most interesting indeed. I've been following John Hampson at "SolarCycles.net" whose evidence-based work is very insighful. It might be worth having a look at as he expects a strong correction very soon indeed, which correlates with your VIX projections.

Kind regards,

Mr. Anon