Predicting Market Directions: The Myth of the Five-Day Index

    by Mark Hulbert

    The stock market rally that began last fall must be on its last legs. Bullish advisers are scraping the bottom of the barrel in their attempts to come up with rationales as to why investors should continue betting on it.

    Consider one of the arguments that is currently widely cited among the bullish newsletters that I monitor: Because the stock market produced a sizeable gain over the first five trading days of this year, 2002 as a whole will be smartly bullish too.

    For example, John McGinley, the editor of Technical Trends, writes that when the market goes up during the first five days, “it almost always indicates that the market will go up” during the year.

    And Donald Rowe, editor of Wall Street Digest, asserts that “if the first five trading days of the year are up, the market will be up for the entire year.”

    Now, the market did indeed do well over the first five trading days of January—for instance, the Dow Jones industrial average advanced 1.3% and the Nasdaq composite index was up over 5%. But as far as I can tell, these editors’ arguments are not based in historical fact. No significant correlation exists between the stock market’s behavior over the first five days of January and its direction for the year as a whole.

    This is a cautionary tale of how, over time, certain beliefs take on a life of their own and therefore become immune to the historical scrutiny that would show them to be false.

    The Genesis of a Myth

    How could so many investment advisers have come to believe in this pattern that some refer to as the “five-day index”?

    As best as I have been able to determine, their belief is derived from a theory that first began circulating around Wall Street a few decades ago—the idea that the entire month of January is a reliable predictor of the stock market’s direction for the entire year. However, it has only been in more recent years that advisers advanced the notion that the first five days of January were a good harbinger of the entire month’s direction, and by virtue of the original theory that had by now become an established part of Wall Street folklore, that was extended to a forecast of the year as a whole.

    Unfortunately, however, neither the original theory nor its more recent amendment can withstand scrutiny. Consider first the ability of the entire month of January to foretell the remainder of the year. Over the last 105 years, as measured by the Dow Jones industrial average, there were 67 years in which January’s direction correctly foretold the stock market’s direction over the subsequent 11 months—a success rate of 64% (Figure 1).

    Figure 1. The Dow's Success Rates for Predicting the Market's Direction the Dow Jones Industrial Average: 1897 to 2001

    Because this success rate is above 50%, this may appear to be impressive.

    But it is not. Because the stock market goes up over the long run, we know that the market will rise during the majority of Januarys as well as during most years. This may create the illusion that January’s direction is predicting the full year’s course. But we also know—if you stretch back to your educational years and Philosophy 101—that when two events appear to be acting in constant conjunction with each other it does not necessarily mean that one is causing the other. Otherwise, as Scottish philosopher David Hume reminded us several centuries ago, the fact that day always follows night would mean that night is the cause of day.

    To be genuinely impressive, January would have to do even better than a simple bet every February 1 that the market would be higher in 11 months’ time. And it does not: Over the last 105 years you would have been right 68 times by making such a simple bet, one more than if you had bet according to January’s direction (see Figure 1).

    The First Five Trading Days

    That is devastating enough to the bull’s current argument, but now consider the notion that January’s first five trading days are a good predictor of the full month’s direction. Over the last 105 years, on only 49 occasions did the stock market’s direction during the first five trading days correctly foretell equities’ direction over the rest of the month. You would have done better simply by flipping a coin.

    So the five-day index is a dismal predictor of a month that itself possesses no genuine forecasting ability.

    How could these notions that are so palpably false have gained such currency?

    One factor is sloppy statistics. Several of the studies that have been cited over the years in their support were guilty of a simple error: They correlated the direction of the market during January with the direction of the market over the entire 12 months from New Year’s through December 31. That is about as fair as allowing you to continue placing bets at the racetrack even after the race has started.

    Another factor that accounts for the popularity of these patently false notions is advisers’ desire to believe in this rally. That need to believe is so strong that they have suspended their skepticism and therefore failed to subject their arguments to even a modicum of historical scrutiny.

    But perhaps the biggest reason why these false notions continue to persist in investors’ minds is the mere fact that they have been repeated so many times for so many years. Investors, therefore, have come to assume that these notions are solidly based in historical fact, and thus they never get around to demanding proof that they are.

    The Plausibility Rule

    The lesson for investors to draw is to adopt an attitude of healthy skepticism toward all the truisms that get repeated daily without question.

    While it may not be practical to subject every last thing we hear to a full and complete statistical scrutiny, there is at least one helpful rule of thumb: Always ask yourself if there is a plausible explanation for why an alleged pattern should exist in the first place. And you should give the greatest scrutiny to those notions that seem least plausible.

    The five-day index definitely fails such a plausibility test. Why should the first five trading days of the year possess any special forecasting ability? The lack of any plausible explanation for why this should be would have raised a red flag in any case, and now we know that this skepticism is completely justified.

    Mark Hulbert is editor of the Hulbert Financial Digest, a newsletter that ranks the performance of investment advisory newsletters. It is published monthly and is located at 5051B Backlick Rd., Annandale, Va. 22003; 703/750-9060;

    This column appears quarterly and is copyrighted by HFD and AAII.

→ Mark Hulbert