Friday, November 11, 2016

The Importance of Tracking Realized and Implied Correlations


When traders think of regimes, they often think in terms of bullish, bearish, or rangebound.  That is, they think directionally.  Within directional frameworks, two regime variables are particularly important:  volatility and correlation.  I have posted often about volatility, including a measure of pure volatility that assesses the amount of movement we're seeing for each unit of volume traded.  The waxing and waning of volatility often gives us clues as to where we stand in market cycles.

Like volatility, correlation can be assessed in two ways:  realized and implied.  One way I look at realized correlation is by taking the major sector ETFs for the SPX, such as XLY, XLV, etc., and take a moving average of their pairwise correlations.  Right now, that measure of realized correlation is quite low--about .56.  That is about two standard deviations below the median correlation among sectors that we've seen since 2005.  The effect of this low correlation can be seen in the bottom graphic from the excellent FinViz site, which tracks trailing one-month performance for major stock sectors.  Notice the considerable variability of returns.  Whether you've made money, stayed flat, or lost money over the past month has very much depended upon the sectors you've been in.

Implied correlation tracks the correlation among stocks implied by their options pricing relative to the pricing of options on stock indexes.  The top chart shows the CBOE Index of implied correlation ($JCJ).  Note the recent roller coaster ride and especially the recent plunge in correlation.  Options are pricing in very different paths for individual stocks relative to the index, and this has been recent--a reflection of different appraisals of the industries that might be winners and losers in the aftermath of the Presidential election.

Notice how a long/short trading strategy--one that is relatively market neutral and buys strong sectors and sells weak ones--can work well in a lower correlation regime.  High correlation regimes are more likely to be ones favoring directional trades implemented with index products.  Correlation can tell us when we're in a stock picking environment and when we're not.

The interplay of volatility and correlation tells us a good amount about where we stand in market cycles, with low volatility and low correlation at market tops and high volatility and high correlation around market lows.  When we see mixed readings in volatility and correlation, we're often transitioning from one cycle phase to another. 

Right now we're in an unusually low correlation regime, with returns driven by sector-related bets.  That sector rotation tells us that investors are not withdrawing money from stocks overall, but are reallocating their exposure.

Further Reading:  Returns from Volatility and Correlation
.