Interesting Technical Indicator, Nine to one

Thursday March 22, 2007

A rare and bullish technical event occurred Wednesday:


By Mark Hulbert, MarketWatch
Last Update: 5:43 PM ET Mar 21, 2007
ANNANDALE, Va. (MarketWatch) -- The most bullish thing a market can do, as the saying goes on Wall Street, is to go up.
I disagree.
It is even more bullish for it to go up as it did on Wednesday.
That's because Wednesday's stock market action was so strong that it triggered a rare technical signal that, far more often than not in the past, has heralded higher stock prices over the subsequent several months.
The particular technical signal is referred to as a "Nine To One Up Day." It refers to the volume of all NYSE-listed stocks that go up on a given day, expressed as a percentage of the total volume of all stocks that rose or fell on that day. On a day when rising stocks' volume is the same as declining stocks' volume, for example, this ratio would be exactly 50%.
A "Nine To One Up Day" occurs when this ratio is 90% or higher. According to Martin Zweig, who helped to develop this indicator several decades ago, such a huge imbalance of up volume over down volume "is a significant sign of positive momentum. In other words, when daily up volume leads down volume by a ratio of 9-to-1 or more, that tends to be an important signal for stocks." The quotation comes from Zweig's 1986 book, "Winning on Wall Street."
I am familiar with Zweig's research because, until the mid 1990s when he discontinued them, he used to publish two investment newsletters. Both letters were ahead of the stock market averages at the time they were discontinued, according to the Hulbert Financial Digest.
How bullish are 9-to-1 up days?
Zweig in his book argues that, "Every bull market in history, and many good intermediate advances, have been launched with a buying stampede that included one or more 9-to-1 up days."
The relevance of all this to today's market is that there have been two 9-to-1 up days in recent weeks: one occurred on March 6, and the second one this Wednesday.
This second 9-to-1 up day adds greatly to the bullish significance of the first, according to Zweig. That's because a single 9-to-1 up day, by itself, has not always been a bullish event. Perhaps its biggest false signal came on March 16, 2000, at more or less the exact top of the market before the Internet bubble burst.
Such a spectacular failure might incline you to dismiss altogether any focus on the ratio of upside to downside volume, but that might be too hasty. In his 1986 book, Zweig acknowledged that a single 9-to-1 up day can issue false signals and that, therefore, it would be better to focus on occasions in which two such days occur relatively close to each other (Zweig used a three-month window).
Zweig called these "double 9-to-1 signals." And with Wednesday's impressive rally, we now have such a signal.
But are there any flies in the ointment?
One might be that there was a 9-to-1 down day on March 13, when down volume on the NYSE was more than 90% of combined up-and-down volume. According to Zweig, a 9-to-1 down day in the proximity of two 9-to-1 up days implies "not as much [upward] thrust" as do two 9-to-1 up days that are unaccompanied by a 9-to-1 down day.
Nevertheless, Zweig wrote that the stock market's average return is still above average following double 9-to-1 days that are accompanied by 9-to-1 down days. "The record [for such days still] provides great comfort to the bulls," as he put it in his book.
This comfort is confirmed by statistical tests conducted by David Aronson, an adjunct professor of finance at Baruch College. Professor Aronson recently wrote a book entitled Evidence-Based Technical Analysis (Wiley, 2007), in which he discusses how to use the "scientific method and statistical inference" when judging investment strategies.
In an interview, Professor Aronson told me that recently, he and the students in a class he teaches at Baruch tested the statistical basis for Zweig's confidence in double 9-to-1 signals. They did not differentiate between such signals that were accompanied by intervening 9-to-1 down days and signals that were not.
Professor Aronson told me that he and his "class used data from the beginning of 1942 through fall of 2006, and we looked at what happens in the stock market in the 60-trading-day period following a Zweig double 9-to-1 signal, versus what happens the rest of the time. In those 60-trading-day windows, S&P 500 Index
SPX1,434.54, -0.50, 0.0%) produced an average annualized return of over 22%, on the assumption that an investor entered the market on the close the day after a double 9-to-1 signal was triggered and held until the end of the 60th trading day later. In the non-signal periods, in contrast, the return averaged 4.5% annualized. The difference between these two average returns is statistically significant." (Professor Aronson told me that these calculations do not include dividends.)
The last time a double 9-to-1 signal was triggered was June 29 of last year. An investor who bought the S&P 500 at the close on June 30 and held for 60 trading days realized an annualized gain of 19%, remarkably close to the 22% average that Professor Aronson found going back to 1942.