The Extraordinary “January Barometer”
By: Jeffrey A. Hirsch & Christopher Mistal
|
January 03, 2017
|
|
Devised by Yale Hirsch in 1972, the January Barometer (JB) has registered nine major errors since 1950 for an 86.4% accuracy ratio. This indicator adheres to propensity that as the S&P 500 goes in January, so goes the year. Of the nine major errors Vietnam affected 1966 and 1968. 1982 saw the start of a major bull market in August. Two January rate cuts and 9/11 affected 2001.The market in January 2003 was held down by the anticipation of military action in Iraq. The second worst bear market since 1900 ended in March of 2009 and Federal Reserve intervention influenced 2010 and 2014. Last year, DJIA slipped into an official Ned Davis bear market in January. Including the eight flat years yields a .742 batting average.
 
As the opening of the New Year, January is host to many important events, indicators and recurring market patterns. U.S. Presidents are inaugurated and present State of the Union Addresses. New Congresses convene. Financial analysts release annual forecasts. Residents of earth return to work and school en mass after holiday celebrations. On January’s second trading day, the results of the official Santa Claus Rally are known and on the fifth trading day the First Five Days early warning system sounds off, but it is the whole-month gain or loss of the S&P 500 that triggers our January Barometer.
 
And yet for some reason, every February or sooner if January starts off poorly our January Barometer gets raked over the coals and every attempt at disparaging this faithful indicator comes up lame. It never ceases to amaze us how our intelligent and insightful colleagues, that we have the utmost professional respect for and many of whom we consider friends, completely and utterly miss the point and fallaciously argue the shortcomings of the January Barometer. However, this year we are not waiting until this happens again. Instead, here is why the January Barometer is relevant and important.
 
1933 “Lame Duck” Amendment—Why JB Works
 
Many detractors refuse to accept the fact the January Barometer exists for one reason and for one reason only: the Twentieth “Lame Duck” Amendment to the Constitution. Passage of the Twentieth Amendment in 1933 created the January Barometer. Since then it has essentially been “As January goes, so goes the year.” January’s direction has correctly forecasted the major trend for the market in many of the subsequent years.
 
Prior to 1934, newly elected Senators and Representatives did not take office until December of the following year, 13 months later (except when new Presidents were inaugurated). Defeated Congressmen stayed in Congress for all of the following session. They were known as “lame ducks.”
 
Since 1934, Congress convenes in the first week of January and includes those members newly elected the previous November. Inauguration Day was also moved up from March 4 to January 20. 
 
January’s prognostic power is attributed to the host of important events transpiring during the month: new Congresses convene; the President gives the State of the Union message, presents the annual budget and sets national goals and priorities. 
 
These events clearly affect our economy and Wall Street and much of the world. Add to that January’s increased cash inflows, portfolio adjustments and market strategizing and it becomes apparent how prophetic January can be. Switch these events to any other month and chances are the January Barometer would become a memory.
 
JB vs. All
 
Over the years there has been much debate regarding the efficacy of our January Barometer. Skeptics never relent and we don’t rest on our laurels. Disbelievers in the January Barometer continue to point to the fact that we include January’s S&P 500 change in the full-year results and that detracts from the January Barometer’s predicative power for the rest of the year. Others attempt to discredit the January Barometer by going further back in time: to 1925 or 1897 or some other arbitrary year. 
 
After the Lame Duck Amendment was ratified in 1934 it took a few years for the Democrat’s heavy congressional margins to even out and for the impact of this tectonic governing shift to take effect. In 1935, 1936 and 1937, the Democrats already had the most lopsided Congressional margins in history, so when these Congresses convened it was anticlimactic. Hence our January Barometer starts in 1938. 
 
In light of all this debate and skepticism we have compared the January Barometer results along with the full year results, the following eleven months results, and the subsequent twelve months results to all other “Monthly Barometers” using the Dow Jones Industrials, the S&P 500 and the NASDAQ Composite.
 
Here’s what we found going back to 1938. There were only 10 major errors. In addition to the nine major errors detailed above: in 1946 the market dropped sharply after the Employment Act was passed by Congress, overriding Truman’s veto, and Congress authorized $12 billion for the Marshall Plan.
 
Including these 10 major errors, the accuracy ratio is 87.3% for the 79-year period. Including the 9 flat year errors (less than +/– 5%) the ratio is 75.9% — still effective. For the benefit of the skeptics, the accuracy ratio calculated on the performance of the following 11 months is still solid. Including all errors — major and flat years — the ratio is still a respectable 68.4%.
 
Now for the even better news: In the 48 up Januarys there were only 3 major errors for a 93.8% accuracy ratio. These years went on to post 16.1% average full-year gains and 11.7% February-to-December gains.
 
[January Barometer vs. All]
 
Let’s compare the January Barometer to all other “Monthly Barometers.” For the accompanying table we went back to 1938 for the S&P 500 and DJIA — the year in which the January Barometer came to life — and back to 1971 for NASDAQ when that index took its current form.
 
The accuracy ratios listed are based on whether or not the given month’s move — up or down — was followed by a move in the same direction for the whole period. For example, in the 79 years of data for the S&P 500 for the January Barometer, 60 years moved in the same direction for 75.9% accuracy.
 
The Calendar Year ratio is based on the month’s percent change and the whole year’s percent change; i.e., we compare December 2015’s percent change to the change for 2015 as a whole. By contrast the 11-month ratio compares the month’s move to the move of the following eleven months. February’s change is compared to the change from March to January. The 12-month change compares the month’s change to the following twelve months. February’s change is compared to the change from March to the next February.
 
Though the January Barometer is based on the S&P 500 we thought it would clear the air to look at the other two major averages as well. You can see for yourself in the table that no other month comes close to January in forecasting prowess over the longer term.
 
There are a few interesting anomalies to point out though. On a calendar year basis DJIA in January is slightly better than the S&P. 2011 is a perfect example of how the DJIA just edges out for the year while the S&P does not. For NASDAQ April, September and November stick out as well on a calendar year basis, but these months are well into the year, and the point is to know how the year might pan out following January, not April, September or November. And no other month has any basis for being a barometer. January is loaded with reasons. 
 
Being the first month of the year it is the time when people readjust their portfolios, rethink their outlook for the coming year and try to make a fresh start. There is also an increase in cash that flows into the market in January, making market direction even more important. Then there is all the information Wall Street has to digest: The State of the Union Address, FOMC meetings, 4th quarter GDP, earnings and the plethora of other economic and market data. 
 
Myths Dispelled
 
In recent years new myths and/or areas of confusion have come to light. One of the biggest errors is the notion that the January Barometer is a stand alone indicator that can be used to base all of your investment decisions for the coming year on. This is simply not true and we have never claimed that the January Barometer should or could be used in this manner. The January Barometer is intended to be used in conjunction with all available data deemed relevant to either confirm or call into question your assessment of the market. No single indicator is 100% accurate so no single indicator should ever be considered in a vacuum. The January Barometer is not an exception to this.
 
Another myth is that the January Barometer is completely useless. Those that believe this like to point out that simply expecting the market to be higher by the end of the year is just as accurate as the January Barometer. Statistically, they are just about right. In the 79-year history examined in this article, there were only 23 full-year declines. So yes, the S&P 500 has posted annual gains 70.9% of the time since 1938. What is missing from this argument is the fact that when January was positive, the full year was also positive 87.5% of the time and when January was down the year was down 58.1% of the time. These are not the near perfect outcomes that true statisticians prefer, but once again, the January Barometer was not intended to be used in a vacuum.