Saturday, April 25, 2015

Questioning Strong and Weak Markets

One of the best ways to generate fresh ideas is to question accepted wisdom.  What is often assumed among traders and market writers is just not the case.

Technical indicators are generally regarded as measures of market strength and weakness.  When price action has been strong over a lookback period, the technical indicator is viewed as bullish and vice versa.  At extreme values, the indicator may be regarded as an "overbought" or "oversold" measure.  

Since the middle of last year, I have been tracking the buy and sell signals for all NYSE stocks for several technical indicators.  (Raw data compiled via the Stock Charts site).  Three indicators I've looked at in particular detail are Bollinger Bands, MACD, and Parabolic SAR.  Since June, 2014, the correlations of buy signals across the indicators have been positive, ranging from +.23 (Bollinger:MACD) to +.56 (MACD:Parabolic SAR).  Similarly, the correlations of sell signals have been high, ranging from +.22 (Bollinger:MACD) to +.75 (MACD:Parabolic SAR).  Clearly, different indicators are not measuring completely different things.  Indeed, an argument can be made that indicators are simply implementations of different moving average rules.  When we look at which indicator is best, we're really fitting past market behavior to a particular moving average, which may or may not be predictive on a prospective basis.

On Friday, we had buy signals outnumber sell signals for all three of the above indicators, reflecting the recent market strength.  Going back to June, 2014, when this has occurred (N=62), the next five days in SPY have averaged a gain of only +.01%, compared with an average five-day gain of +.26% for the remainder of the sample.  Clearly, strong readings have not brought near-term strength--though neither have "overbought" readings reliably led to market declines.  The failure of strength to be followed by strength is yet one more reflection of the distinction between trend and momentum outlined in the recent post

Conversely, when we've had all three indicators yielding more sell signals than buys (N=49), the next three days in SPY have averaged a gain of +.33% versus +.04% for the rest of the sample.  Market weakness has tended to reverse in the near term, though much of that relative performance boost tends to fade over subsequent trading sessions.

It is human nature to extrapolate from the recent past to the immediate future when we are trying to anticipate events.  In the case of the stock market, failing to question "strength" and "weakness" has been hazardous to our wealth.

Further Reading:  The Breadth of Strength