Rank: Advanced Member
Groups: Registered, Registered Users, Subscribers, Unverified Users
Joined: 10/28/2004(UTC)
Posts: 3,111
Location: Perth, Western Australia
Was thanked: 16 time(s) in 16 post(s)
|
In the past there may have existed strong (negative) correlation between $VIX and $SP, however that relationship has been in decline and accelerating for a period from 2012-2013; although there still exists a negative correlation over "short" timeframes, generally it is much weaker than what it was. The instances when the market and the VIX move in the same direction (weak positive correlation) still happen from time to time (and allegedly this still freaks some people out?)  5 (front) to 500 (rear) bar correlation coefficients (in 5 period steps) from 2010 to the present with no lookahead (no prediction). Notice the right hand side is generally much higher than the left, showing the negative correlation weakening. Of course, if you want to see what direction the market took today, there's no point in looking at today's VIX as you can just look directly at the $SP chart; what we as traders/investors really want to know is, by looking at the VIX today, predict what will the market be doing on the next trading day, or in the next few days?  5 (front) to 500 (rear) bar correlation coefficients from 2010 to the present with 1-Period lookahead (prediction of change in $SP in the future for a current change in $VIX)
We can see the correlation is a lot less consistent, overall is weaker and many more excursions into positive territory. If we test the VIX correlation to the $SP 5 periods into the future, things get worse.  5 (front) to 500 (rear) bar correlation coefficients from 2010 to the
present with 5-Period lookahead.
The use of implied volatility is a valid trading method or tool, if, like all things, it is used correctly, with a full understanding of the capabilities and limitations. Understanding causality and inference goes a long way to preventing losses: just because the VIX moved today does not guarantee the price will change tomorrow, it just means that it has done so in the past. In my readings I've come across quite a few historical papers discussing VIX in particular; some papers say they can prove positive correlation with future index "prices" and some say they can prove negative correlation. They both cannot be correct at the same time if looking at the same periods. Whilst researching inputs to my own neural networks, I've found that measured price volatility has a more consistent correlation with future change in price than just about every other method (albeit often slightly weaker); hence it's used more widely in artificial learning systems than implied volatility. (You can test this for yourself in MetaStock, but it's painful, so I prefer to employ other applications designed with this sort of number crunching in mind, such as MatLab and Excel.) Is this weakening correlation the same for all implied volatility VIXs for all markets? I don't know, but my correlation studies for the ASX are remarkably similar to the US, so I'm going to take a guess and say, based on a sample of only two markets, probably. If you want to play with implied volatility based trading strategies, there are many web sites with information which might help you customise your system; Google is your friend to help you find information but it does not differentiate the good from the bad... like always, proceed carefully.
|
