Saturday, March 29, 2014

A Look at Some Interesting Market Relationships

Perhaps a bit too much to think on an early Saturday morning!  Anyway, here are a few of the more cogent thoughts:

One of the measures I track is something I call market "stability".  It basically looks at the uniformity vs. dispersion of correlation and volatility within the stock market.  I notice that stability has been breaking down in recent sessions.  The last time this happened was at the very end of February/very beginning of March, which preceded a market dip.  Indeed, since the start of 2012, the top quartile of stable markets have averaged a five-day gain of +.72%.  The bottom quartile of stable markets have averaged a five-day loss of -.02%.  In the middle quartiles, we've averaged five-day gains of +.52% and +.23%.  We're not at that bottom quartile presently, but we've moved out of the top half of the distribution. 

Interesting post from Paststat recently in which he builds on this blog's earlier observation and investigates the implications of underperformance of small caps.  Again, we can look by quartiles since the start of 2012 and it turns out that when smaller caps (IWM) have most underperformed the S&P 500 index (SPY) over a ten day period, the next five days in SPY have averaged a gain of +.85%.  The next three quartiles average, respectively, +.05%, +.09%, and +.53%.  Interestingly, we've seen momentum when small caps have meaningfully outperformed large caps and reversal when they've meaningfully underperformed.  

*   I don't see a similar relationship between SPY outcomes and the degree to which NASDAQ shares (QQQ) underperform.  Indeed, following the instances in the top quartile of QQQ underperformance over a 20-day period since 2012, the next 20 days in SPY average a gain of only +.60, compared with an average 20-day gain of 1.61% over the remainder of occasions.

*   Sentiment matters.  Since 2012, the top quartile of days based on single day stock index put-call ratios have averaged a next 20-day gain of 2.22%.  The bottom quartile of days (lowest index put/call ratios) have averaged a next 20-day gain of .78%.  The top quartile of days based on single day single equity put-call ratios have averaged a next 20-day gain of 2.11%.  The bottom quartile of days have averaged a gain of .81%. 

Volatility matters as well.  Since 2012, the lowest quartile of VIX days have averaged a next five-day gain of only .03%.  The highest quartile of VIX days have averaged a next five-day gain of .87%.

Divergences?  The good news is that SPY has been only about 2% off its all-time highs during the past three trading sessions.  The bad news is that stocks making new monthly highs fell short of those making fresh monthly lows by 376 to 993 three days ago; by 212 to 1461 two days ago; and by 328 to 646 yesterday.  There has been much more weakness than the large cap indexes suggest.

Further Reading:  A Different Way of Looking at New High and Low Data