ANNANDALE, Va. (MarketWatch) — It must be something about this time of year:
Around the turn of every year, Wall Street’s gurus succumb to a strange malady: They become extraordinarily quick to extrapolate the tiniest blip in the stock market into a forecast about where equities are headed for the year as a whole.
If the market rises strongly for a day or two, for example, the pundits confidently announce that this means that we’ll have a good year. But if stocks hit an air pocket around New Years, as they have this year, then we’re told that the year as a whole is going to be a bad one.
This tendency is what is behind several old Wall Street saws that get dusted off every New Years. One goes by the name of the “first five days of January” indicator; the other is that, “as goes January, so goes the year”–otherwise known as the “January Indicator.”
I am not aware of any reason, however, for why the market’s gyrations at this time of year should have any more significance than at any other time. When I press advisers for reasons why these patterns should prevail, I am told that they don’t need to know — the important thing is that the market has this power.
But do they really?
First five days of January indicator
Consider first the notion that the first five trading days of January are able to foretell the stock market’s direction for the rest of the year. I tested this notion by applying it to the Dow Jones Industrial Average (.DJI
) back to 1896, when this benchmark was created.
It turns out that, if the first five days of January were up, the stock market proceeded to rise 68% of the time from then until the end of the year. If the first days of January were down, in contrast, the market for the rest of the year was up 56% of the time.
To put these percentages in context, consider that the stock market on average has risen in 64% of all years since 1896. So the market’s likelihood of rising was 4 percentage points higher when the first five trading days of the year were positive, and 8 percentage points lower when those first days were negative.
Even if those differences in probability were statistically significant, I am not sure that they are big enough to make it worth investors’ while to act accordingly. But they are not statistically significant–at least at the 95% confidence level that statisticians often use to determine if a pattern is genuine.
Another telling statistic: The first five days of April and May have better forecasting records than does January, even though neither of those additional months’ forecasting records is significant at the 95% confidence level either. Yet I’ve never heard any of the talking heads refer to a “first five days of April” or a “first five days of May” indicator.
How did the “first five days of January” indicator develop a following in the first place? My hunch is that, after discovering that the indicator had an apparently good track record over certain periods, followers failed to subject it to longer-term reality checks.
For example, I’ve noticed in the blogosphere that some devotees are reporting that the indicator has had an “uncanny” record since 1942. If they had bothered to look at its record prior to 1942, however, they would have found it to be worthless.
The January Indicator?
What about the other indicator based on the stock market’s behavior during January — the notion that the stock market’s direction from February through December is foretold by its direction in January? At first blush this so-called January Indicator appears to have a better record than the “first five days of January” indicator, though there still are reasons not to make too much of it.
Since 1896, the Dow has risen 72% of the time from February through December when January was up. In contrast, when January was down, it has risen just 50% of the time over the subsequent 11 months.
Relative to the 64% frequency that the market has risen in all years, therefore, a positive January increases the odds of an up year by 8 percentage points. And a down month in January decreases those odds by 14 percentage points. Notice that these are bigger shifts in probability than the ones that emerged from testing the “first five days of January” indicator. In fact, furthermore, they are marginally statistically significant.
Bear in mind, however, that other months besides January also possess this apparent ability to forecast the market’s direction over the subsequent 11 months; November stands out in this regard. So January is not entirely unique. Why don’t Wall Street’s talking heads also talk about a November Indicator?
A second qualification: There have been significant periods in which the January Indicator was a big flop. Consider, for example, the dozen years from 1926 through 1937. The January Indicator was right in just three of those 12 years, for a dismal 25% success rate.
Take 1931, for example, when January was a positive month for the Dow and this indicator predicted that the rest of the year would be up as well. The Dow actually dropped 53.5% over the last 11 months of that year. The next year was barely better for the January Indicator: Though it called for the last 11 months of 1932 to be profitable, the Dow in fact dropped more than 21%.
The bottom line? These beginning-of-year gyrations give all of us — especially journalists — lots to talk and write about.
But basing a trading strategy on the market’s performance over the first five days of January can be dangerous to your wealth. And basing a trading strategy for the year as a whole on the market’s performance in January is only marginally a better idea.