The January Effect
As Goes January, So Goes The Year
The first week of 2019 has pretty much repeated what it did at the start of 2018—buying stocks and selling Dollars. And yet this is a very different market. At the beginning of last year there had been little change in the market dynamic as a tax driven Goldilocks (low inflation, high growth) US economy drove most instruments positively through December into the New Year. However, this year US Fed-inspired concerns of slowing growth triggered a pre-Christmas collapse in stocks and a lagging precipitous New Year fall in risk averse USDJPY.
The early euphoric stock blowout of 2018 has been replaced by what can be seen so far as an early 2019 relief rally. The inevitable question that confronts traders going into 2019 is whether what happened into Xmas was an exaggerated, or false, panic due to illiquidity or the start of something more substantial like Bitcoin price/action proved to be after its pre Xmas 2018 collapse.
Many will typically seek the answer in the first week and indeed the first month of this year—what is known as the January Effect—where the Xmas break in sentiment leads to a new trend. 2018 proved to be a sentiment driven year characteristic at the end of a trend. Sentiment should remain an important variable, reflected by our FOREX 2019, INDEX/COMMODITY 2019 and STOCKS 2019 outlooks to be published over the next three weeks.
But, having left many fundamentals behind in the last two years, we expect markets to become more, if you like, realistic as they realign to what is actually happening in the real world rather than what people expect or hope.
This interaction between sentiment and fundamentals or perhaps traders’ more rational or evidenced view of the global economy should dominate the year. Last year started with a 17 day 7% rally in SPX and 3.7% fall in the Dollar (also fueled by NFP in the first week) only to be reversed or consolidated for much of the year. The key issue is whether the market will follow last year’s pattern or rather continue the tone set by the first week and month—The January Effect.
Today’s newsletter address the popular market aphorism is ‘as goes January, so goes the year’ and whether this will shed any light on the remaining eleven months of 2019. Our view like last year is generally No, Yes, No. A feature however of this year we believe will be a Dollar that is a mirror image of last year, a downtrend rather than an uptrend. In this respect January will be important.
The first Monday of 2018 is possibly the most statistically important single day of the year and typically very volatile. But historically it is the whole month of January that often sets the tone for much of the year and may well condition our whole approach for the ensuing 6-12 months. It is one reason why, unlike many others, we publish our 6-12 month outlooks in January rather than front run the year. After all front running caught many out for much of last year.
Context is the solid foundation of much successful analysis; that is analyzing multi time frames downwards; rather than the more common inside-out analysis, we see out there, where noise-ridden short term views drive medium term outlooks and condition many annual outlooks that are devised in November or December. It is also important because quite often these longer term views prove to be uncannily accurate, even if we do occasionally get side-tracked or thrown out by an erratic market.
Conditionality is also an important feature of successful analysis; that is each forecast (and therefore trading strategy) has certain conditions we expect to be met along the way to help ensure that it remains accurate, ready for the next step or condition—or—if the condition is not met or broken, to indicate a change/adaptation of a view and a new set of conditions.
The many Fundamental, Intermarket, Technical and Sentiment milestones we identified for the 1920s style stock bull market and indeed the 1987 style USD decline not only helped us in the early and middle stages of those trends but should help as and when they mature, particularly now we have revisited and revised those views for 2018 and have created a new far more robust set of conditions.
Statistically important time periods such as January serve as a useful guide to conditionality. In this respect, the January effect or perhaps better January Condition will guide and assess the probability of what our now confident views for 2019.
The January Effect is widely known as a stock market observation in two respects.
First, stocks tend to rise more in January than in other months. This was originally identified by Rozeff and Kinney in 1976.
Many other studies since have proven that the effect does exist. From 1950 the average January rise has been 1.79% in the S&P compared to the average month rise of only 0.65% proving this seasonality. It is particularly noticeable in small cap stocks, and in particular those whose price fell in the previous year. This links it to a practice once known as 'bed-and-breakfasting', where a stock is sold at the end of the year to crystallize a tax loss, and then immediately rebought in the New Year. Certainly the first week of the year is consistent with this practice and reflected an estimated $562bn US equity inflow, an eight year record +2 standard deviations above the historical. Studies of 1986, where US tax treatment changed, and countries whose tax year does not end on December 31st have proven a strong January effect.
Further research has shown the impact of the January Effect has fallen over time and become self-defeating. As the hypothesis has become well-known, traders try to front-run the January rise by loading up in December—to their cost this year! The net effect is that the price rises move back to December, and are just treated as part of the 'Santa Rally', a separate phenomenon with separate causes. A sign that this is happening would be a sudden sell-off just as the year closes, and quick recovery in January as we have seen so far.
Secondly, “As goes January, so goes the year”. Some have argued stock market years have a tendency to follow January’s trend. Comparison of January’s performance to the rest of the year reveals a positive correlation coefficient of 0.25, a meaningfully strong number. Although this could possibly be said of any month in a trending market, sampling every month gives a correlation coefficient of only 0.016, a significantly lower number and suggests some validity to the theory. However, before we get excited if January ends lower only 11 of the last 26 negative Januaries presaged a negative year.
However, the January Effect that foreign exchange traders use is different and its effect and implication is often the opposite of the stock January Effect. A misnomer and, from now on, we will refer to it as the January Condition. This Condition is that certain markets have a propensity to make a high or low in January that persists for at least six months and often the year. This would seem, in many cases, to imply a reversal and the opposite to 'as goes January, so goes the year'. Although it is a sufficient but not necessary condition for a reversal, sometimes the market only corrects into January before resuming the preceding trend. It has been been noted most particularly in the EURUSD (and previously US Dollar/Deutsche Mark) but is actually stronger in SPX and most notably Gold.
In the last 42 years, the Euro has set 12 lows and 14 highs for the year in January. A 62% success rate is not bad considering the demands of this condition. It would have been much better (72%) without the barren six year period from 2008-2014 that curiously coincided with Quantitative Easing.
This explanation is borne out by the data showing a higher incidence of the annual or six-monthly high or low in the first four weeks of the year. If one can reasonably prescribe probability of repetition based on the number of incidents within a sample period, then the above histogram is instructive. Bear in mind a normal distribution for a yearly high/low should be 8% and six monthly 16%. The probably that the Euro will make a six-month low in January is 34.9%, a 12-month low 23.3%, a six-month high 46.5% and a yearly high of 32.6.4% - significantly higher than normal.
Similarly, the S&P reveals an even stronger tendency for the January Condition thereby actually supporting the January Effect. There is a 37.8% historical probability SPX will make a six month low in the first four weeks; 30% for a 12 month low, and the previously noted lower 20% probability of a six month high and only 14.4% for a yearly high.
But it is Gold’s January Condition that stands out. In the last 50 years, Gold has produced annual lows 10 times in the first week. That is a staggering 20% compared to a 2% random distribution. Even though the incidence of an annual low thereafter returns to normal the probability of an annual Gold low in the first four weeks of the year still remains strong at 32% and 44% for a six monthly low. For a high, it is noticeably weaker. Only 18% of annual highs have been set in the first four weeks but 32% for the six monthly high.
Does this first week incidence apply to all commodities? No. The distribution of annual or six-monthly highs and lows for Oil is statistically normal pretty much throughout the year despite a pick up in the incidence of lows before Xmas.
Reference however to USDJPY starts to hint at an explanation for why the January Condition exists. Unlike Oil's distribution (and to a lesser extent Gold), the incidence of USDJPY highs and lows in January are similar to EURUSD and SPX. The probability of USDJPY setting a yearly high in the first four weeks is 29.8%, a six monthly high of 36.2% whereas it has set an annual low 25.5% and six monthly low 31.9% It is not surprising to find a January condition in stocks that also applies to risk aversion/appetite currency pairs such as EURUSD and USDJPY.
Where there is a strong fundamental trend with breadth, correlations will be high and such markets will move roughly together and produce similar January Effects and Conditions. However, where there is no trend or a mature trend that lacks the previous fundamental justification, then as we have seen, correlations break down. That curiously may not reduce the incidence of January highs and lows. In fact they may increase.
Seasonality exists in markets for a number of reasons. Nature is the most obvious for soft commodities as going through their harvesting cycle. The heating season and (distribution of those seeking heat!) would clearly affect heat related energy markets. Payment cycles: Amazon's recent and historical performance into and out of Black Friday is an outstanding example. Window dressing is most notable at quarter or most particularly year end, not just tax reasons for stocks but reporting reasons for funds, even currency balance sheet dressing for international corporations. The extent to which they help create a January Effect or Condition depends on their sustainability, partly on the overall market context but most importantly on sentiment.
The whole period from Thanksgiving to the first week or four weeks of January can be regarded as a season and one that is subject to strong sentiment. We have seen how front running the tradition January effect in stocks has helped create the now universally acknowledged and expected Santa rally. Its effect still applies in January but starts to fall off as the reasons for the move have less force.
The main reason is sentiment. Traders, investors and people tend to do similar but not always identical things at similar times when confronted by the same circumstances. The commercial/financial anticipation of Christmas is a boon for some. The actual holiday period presents an opportunity either for an immediate break in the possibly overdone previous trend or a reality check. The year end often produces moves that will not ultimately be sustained. And yet traders seek trend and particularly at the start of the year as a sign of things to come. They will often over-pursue year-end moves into the New Year (January Effect) and frequently produce a retracement or reversal.
Where there is fundamental justification for the trend markets will tend to retrace together and resume the trend later. However where there is little sound reason and breadth this early year over-exuberance often produces an acute correlation break down and a later reversal. Hence the January Condition and why intermarket relationships often go out the window in January only to rediscover their glue into February.
How one factors this data into actionable strategies may appear simple. Once the January range is established and broken in either February or later, then it increases the probability the direction of the break will be maintained until June 30th at least. However, even in the EURUSD or SPX case, the probability does not exceed 50% and should not be acted upon in isolation.
None of the theories and statistical evidence for either the January Effect or Condition take context into account; that is the nature of the particular market at that time. An instrument that has reached extreme sentiment leaving the fundamentals behind, displays clear maturity in its price action, and lacks breadth or support from related markets is more prone to a significant top or bottom in January than any other month of the year. And yet a trend that is sustainable or at least is complete may well provide a better opportunity to rejoin the trend in the first four weeks of the year.
As goes January, so goes the year and what a year we suspect it will be. Our outlooks for 2019 factor this in and explain why we see significant Dollar weakness emerging out of January and two-way volatility in indices that should continue to grow throughout the year leading to both significantly higher and lower prices than we have seen so far.
Here’s to making the very best.
Founder, Matrix Trade
To receive our weekly articles by email in advance of publication, please sign up for our Sunday newsletter.