Capturing Value and Momentum in the Stock Market

In mid-2014 I hit upon an idea for analyzing the strength and weakness of the overall stock market. Suppose we took every stock in the New York Stock Exchange and assessed whether it gave a buy signal, a neutral signal, or a sell signal for a standard technical indicator, such as Bollinger Bands. Such a measure would capture the breadth of strength and weakness for stocks as a whole, not just for the index itself.  Would this be a useful measure?  It turns out that the measure was indeed useful and I began collecting the data daily from the Stock Charts website.

Then I hit upon another idea.  The signals from cumulated stock performance on one indicator (such as Bollinger Bands) were different from the signals from other indicators (such as RSI and Parabolic SAR).  Might it be useful to create an indicator of indicators? This would show occasions when we have strength and weakness across all stocks *and* all indicators.  

The resulting cumulative indicator measure is charted above from 2016 forward (indicator in red; SPY in blue).  Even within the considerable uptrend we've had over that period in SPY, we've seen relative periods of overbought and oversold in the measure.  Note that we currently stand at a significantly oversold level.

Going back to June of 2014, when I first began accumulating these data, next ten day returns in SPY have averaged +.01% when we have been in the top half of the distribution for the cumulative measure.  When we have been in the bottom half of the measure, next ten day returns in SPY have averaged +.63%.  This is a significant value effect.  Returns have been significantly better over a swing period when we've been oversold than when we've been overbought.  If we break down returns by quartiles, the upside returns are even more striking in the weakest (most oversold) quartile, which is where we stand now. Interestingly, when the indicators have been simultaneously strong, we've seen superior upside returns over the same ten day horizon.  

In other words, the cumulative measure is capturing both a value effect (buy when things have gotten weak) and a momentum effect (buy when there is a broad thrust higher). Returns have been subnormal if we are not broadly weak or broadly strong.

This is a nice illustration of the value of "big data" and especially the value of well-conceived unique data sets.  As a discretionary trader, I find it crucial to be quantitatively informed.  I observe that integration of discretionary and quantitative among the great majority of the successful traders and portfolio managers I work with.  Even for longer time frame active investors, timing market entries and exits with shorter-term measures that capture value and momentum can meaningfully enhance returns.

Further Reading:  The Foundation of Trading Success

.