S&P 500/VIX Correlation And The Perils Of Visual Charting

Another indicator of market tops surfaced yesterday: When the 10-day correlation between the S&P 500 and VIX is positive, then this indicates a price top. The problem is that this does not work in practice unless one is just a market spectator.

Updated on October 14, 2017. Click here for the update section.

Below is a chart of S&P 500 since 2015. A chart of VIX is shown in the first indicator pane and the 0 lag, 10-day correlation is shown in the bottom pane.

A visual chartist may get the impression from the above that when the 10-day correlation of S&P 500 and VIX rises above 0 is an indicator of short-term market top.

But this is 1980s technical analysis.

This is 2017, the year of the algos. 

Nowadays traders are interested in whether an indicator can be used to make profits. Surely, there are market spectators who use all forms of indicators but in most cases the signals are random.

In the context of an evidence-based analysis framework, we have to determine whether this indicator is useful to traders. This means that we have to use it in a strategy. Then, we have to analyze the strategy and determine whether it worked in the past in acceptable ways. Otherwise, this is only talk and no substance.

The simplest strategy is buying the index (hypothetically, in practice via an ETF) when the S&P 500/VIX 10-day correlation is positive and exiting after a number of days.

Below are backtests results for number of days to exit equal to 5, 10, 20 and 60. Click on images to enlarge.

It may be seen that this strategy has been unprofitable. This is a common problem with visual chartists because they tend to pick signals of indicators that confirm their bias. But in most cases the indicators fail in the context of an evidence-based approach.

Note that if instead of price level correlation we use returns correlation, then for a 5-day exit there is a small gain but performance is indistinguishable from random. In the case of a 10-day exit, performance is negative. Click on images to enlarge.

The conclusion is that the 10-day correlation between S&P 500 and VIX is not a good indicator of price tops, as it was claimed. Evidence-based analysis says that this is not so. Visual chartists may think otherwise and this is fine. But this is not the 1980s.

Update on October 14, 2017:

Below is a chart showing both price level and daily returns correlations of the two indexes:

It may be seen that although price level correlation is +0.246, returns correlation is negative at -0.753, as expected. The price level correlation is an ambiguous measure of anything.

Below is a chart that shows the 2010 and 2011 corrections:

The indicator based on price level correlation failed to warn about the 2010 correction and had very poor timing of the 2011 correction.


If you found this article interesting, I invite you follow this blog via any of these methods: RSS or Email, or follow us on Twitter

If you have any questions or comments, happy to connect on Twitter: @mikeharrisNY

Charting and backtesting program: Amibroker


Technical and quantitative analysis of Dow-30 stocks and 30 popular ETFs is included in our Weekly Premium Report. Market signals for longer-term traders are offered by our premium Market Signals service. Mean-reversion signals for short-term SPY traders are provided in our Mean Reversion report.

Copyright Notice

This entry was posted in Technical Analysis and tagged , , , , . Bookmark the permalink.

Leave a Reply