Thursday, October 19, 2006

Capturing Stock Market Momentum: Relative New High Ratio

In the Trading Psychology Weblog, I keep a running tally of the number of stocks making fresh 20-day new highs and 20-day new lows. Having kept those figures for about three years now, I've seen that there is a momentum component embedded in that measure. When we get a large expansion of new highs, very often the upside momentum continues over a short-to-intermediate time frame.

In a recent post, I went back to 1990 and found that, when we measure new highs as a percentage of total issues traded, returns are superior when very few stocks are making new 52-week highs compared to when many issues are at annual peaks.

In this investigation, I create a new measure that we can call the relative new high ratio. I compute the percentage of issues traded that are making new highs and express this as a ratio. The denominator in the ratio is the average proportion of stocks making new highs for the 20 prior trading sessions. Thus, if we have a relative new highs reading of 2.0, it means that the proportion of stocks making new highs is twice as high as it has been over the prior 20 days. Conversely, a ratio of .50 means that we're seeing only half the proportion of issues making new highs as we've seen over the last 20 sessions.

Since 1990 (N = 4196 trading days), we've had 90 occasions in which the relative new high ratio has hit or exceeded 2.5. Forty trading days later, the S&P 500 Index ($SPX) has been up by an average of 3.32% (70 up, 20 down), much stronger than the average 40-day gain of 1.47% (2695 up, 1501 down). This signal only occurs, on average, a few times a year, but a win percentage of over 75% is notable. The last signal occurred on August 4th of this year, which turned out to be a great time to buy.

Interestingly, when the relative new high ratio is very low, we also get some good buy signals. Specifically, when the ratio has been .25 or under (N = 160), the next 40 days in the S&P 500 Index have been up by an average of 2.83% (120 up, 40 down)--again much stronger than the average 40-day performance, with a solid win percentage. The last time we had a signal from a low relative new high ratio was June 14th of this year--also an excellent time to buy for the interemediate-term hold.

The relative new high ratio catches bursts and contractions of new highs: times when people are pouring into stocks or bailing out of them. In that sense, it may be capturing a piece of market psychology. Can we obtain similar findings from the new low data? That study will be coming up shortly.