Category Archives: Investing

January Barometer

The January barometer is the hypothesis that stock market performance in January (particularly in the US) predicts its performance for the rest of the year. So if the stock market rises in January, it is likely to continue to rise by the end of December. Probably the best known is “as January goes, so goes the year”. Another says that the first 5 trading days determine the market’s returns for the whole year ahead. And another says that how January performs predicts the direction of the market for the remaining months of the year. The January barometer was first mentioned by Yale Hirsch in 1972. Historically if the S&P 500 goes up in January the trend will follow for the rest of the year. Conversely if the S&P falls in January then it will fall for the rest of the year.

If an investor believes in the ability of the January barometer to predict the equity market’s performance, he will only invest in the market in the years when the barometer predicts the market will rise, and stay out of the market when it forecasts a market pullback. It is difficult to produce excess returns based on this theory. Since the improved performance by staying out of the market during bad times can be more than offset by larger losses incurred when the barometer incorrectly predicts a bull market. January tends to be one of the market’s best months every year, rising an average of 1.3% since 1929, see our January Effect. And because the stock market rises in most years, there’s some inherent bias in the data itself. So, while the barometer works well in predicting up years (some studies even suggest that up Januaries mean up years 85% of the time), it doesn’t do so well in predicting down years. When January is down, the market continues to fall only 46% of the time.

From 1950 till 1984 both positive and negative prediction had a certainty of about 70% and 90% respectively with 75% in total. After 1985 however, the negative predictive power had been reduced to 50%, or in other words, no predictive powers at all. The following table shows the wrong prediction of January Barometer :

Year January S&P End of Year S&P Comments
1946 7% gain Loss 17.6% Start of bad misfire
1947 2.4% gain Loss 2.3% Not right, but no biggie
1948 4% loss Gain 3.5% What the f?
1956 3.6% loss Gain 6.5% Minor
1960 7.1% loss Gain 4.5%  
1966 0.5% gain Loss 13.5%  
1968 4.4% loss Gain 12.6%  
1970 7.6% loss Gain 8.4%  
1978 6.2% loss Gain 7.7%  
1982 1.8% loss Gain 16.8%  
1984 0.9% loss Gain 2.3%  
1987 13.2% loss Gain 9.9%  
1990 6.9% loss Gain 0.3%  
1992 2% loss Gain 6.6%  
1994 3.3% gain Loss 4.6%  
2001 3.5% gain Loss 13%  
2003 2.7% loss Gain 26.4% significant
2005 2.53% loss Gain 8.36% Minor disappointment
2009 8.5% loss Gain 35% Worst loss
2010 3.9% loss Gain 12.6%  
2011?      
       
       

5 wrongs for the most recent decade:

Year 2001: gain of 3.5% but loss of 13% year end.

Year 2003: The S&P has a loss of 2.7% but the S&P finished with significant  26.4% gain.

Year 2005: there was a minor misfire in 2005.

Year 2009: the January barometer is a big fake. January 2009 saw the S&P 500 fall 8.5%, only to finish with one of the best years on record as stocks soared 35% the rest of the way. Anybody who followed the barometer religiously in the year 2009 missed out on one of the most profitable market swings in a generation.

Year 2010: After a 3.9% January loss, the S&P finished with a 12.6% gain.

The year 2011 is not finished yet, what 2011 is going to be? What about 2012? Evidence also suggests that other months, such as April and November, are just as good at predicting the year as January. So why does January get all the attention? Read our other theories such as the January effect, Presidential Cycle and Halloween indicator.

Halloween indicator

The Halloween indicator is a variant of the stock market adage “Sell in May and go away,” the belief that the period from November to April has significantly stronger growth on average than the other months. In such strategies, stocks are sold at the start of May and the proceeds held in cash; stocks are bought again in the autumn, typically around Halloween. This Halloween indicator is partially related to another well known effect: The January Effect.

 

Though this seasonality is often mentioned informally, it has largely been ignored in academic circles (perhaps being assumed to be a mere superstition). Nonetheless analysis by Bouman and Jacobsen (2002) shows that the effect has indeed occurred in 36 out of 37 countries examined, and since the 17th century (1694) in the United Kingdom; it is strongest in Europe. According to the efficient-market hypothesis, this is impossible.

 

It is not clear what causes the effect. Many supporters of the Halloween indicator suggest that people taking vacations and holidays during the summer months can lead to market weakness. There are exceptions: between April 30 and October 30 2009, the FTSE 100 gained 20% (from 4,189.59 to 5,044.55).  We can not say the Halloween Indicator existed in each and every year. But that’s hardly surprising, since no indicator works all the time. The question is, over progressively longer horizons, does the success rate grew to very impressive and consistent levels (should be much higher than 50%).

 

Most interesting about the effect is that it shows that stock market returns in many countries during the period May-October are systematically negative or lower than the short-term interest rate, which also goes against the efficient-market hypothesis. Stock market returns should not be predictably lower than the short term interest rate (risk free rate).

 

Popular media often refer to this market wisdom in the month of May, claiming that in the six months to come things will be different and the pattern will not show. However, as the effect has been strongly present in most developed markets (including the United States, Canada, Japan, the United Kingdom and most European countries) in the last decade – especially May-October 2009 – these claims are often proved wrong.

 

One study which tests the Halloween indicator in US equity markets found similar results as Bouman and Jacobsen (2002) over the same time period but using futures data over the period April 1982- April 2003 and after excluding the years 1987 and 1998 no longer found a significant effect, leading these researchers to conclude that it was not an “exploitable anomaly’ during that time period in the United States.” Other regression models using the same data but controlling for extreme outliers have found the Halloween effect to still be significant. The original saying is “Sell in May and go away, stay away till St. Leger Day”, referring to the last race of the British horse racing season, however this day is unlikely to be known by non-Brits so it is replaced by Halloween (which in turn is Samhain, about one-eighth year after the equinox).

Presidential Cycle

Presidential Cycle

As we discussed in the marketing timing strategies, we will spend more time to dig deeper into the Presidential Cycle.

The first year is the weakest of all four years. Higher returns during the last two years of a Presidential term than the first years. The expectation is that as a President takes office he begins to implement his proposals and investors, hunker down waiting to see the results. During the final two years the President becomes more concerned with his re-election and will ‘prime the pump’ in order to secure re-election.

For generations, researchers have sought to make sense of the aggregated behavior. Is there any rhyme or reason behind anything that occurs in the markets or in life? The four-year U.S. presidential cycle is attributed to politics and its impact on America’s economic policies and market sentiment. Either or both of these factors could be the cause for the stock market’s statistically improved performance during most of the third and fourth years of a president’s four year term. The month-end seasonality cycle is attributed to the automatic purchases associated with retirement accounts.

Mark Hulbert of MarketWatch compiled Dow Jones data dating back to 1896 and found the following:

– the third year of a presidential cycle significantly outperforms all others, with average stock market gains of 15.5% compared to 8.8% in the first year, 0.4% in the second, and 4.1% in the fourth year.

– the market performance in the third year of a presidential cycle is not statistically correlated to the market performance of the previous year; in other words, third years have tended to outperform other years regardless of whether the second year in the cycle experienced a boom or a bust

– he also found that there is no significant correlation between stock valuations and market performances in the third year; “On average, third years perform just as well when price/earnings ratios are high as they do when those ratios are low”

Marshall Nickles of Pepperdine University found similar results in his 2004 study.

– historical stock market cycles dating back to 1942 have lasted an average of 4.02 years, which is essentially the same length as a presidential term; a cycle is defined as the time between peaks or the time it takes to go from a peak to a trough and then back to a peak

– stock markets bottomed out only once during the third year of a presidential cycle in data dating back to the FDR administration. The one time the trough occurred in the third year was in December 1987 (Reagan). The average time frame for a market trough was 1.87 years into the presidential term. Markets never bottomed out in the fourth year.

In general, incumbent politicians are more likely to be re-elected and their party remains in power if the economy has been doing well. Basically, satisfied voters have tended to re-elect incumbents. Politicians have tended to fiscal policy to buoy the economy during the campaign season.

January effect

January effect

As we discussed in the marketing timing strategies, January Effect is the oldest and more common accepted ideas. In this article, we expand the topic and look at more details.

Increase in buying securities before the end of the year for a lower price, and selling them in January to generate profit from the price differences.  The “January effect” is that American stocks rise much more in January than in any other month of the year. It is called January effect but now the date may be moved earlier because people are buying the stocks earlier to profit from it. One theory explaining this phenomenon is that income tax-sensitive investors sell stocks for tax reasons at year end (such as to claim a capital loss) and reinvest after the first of the year. The second reason is the payment of year end bonuses in January. Some of this bonus money is used to purchase stocks, driving up prices. The third reason is that may people max out the retirement contribution at the end of a year and contribute to their retirement saving at the start of the year. The January effect may not be always true; for example, small stocks underperformed large stocks in January 1982, 1987, 1989 and 1990.

 

Sidney Wachtel discovered the phenomenon in the 1940s, but it wasn’t until the 1970s that anybody took much notice. Many subsequent researchers have made many refinements and produced several ingenious explanations, usually suggesting that shares are dumped in December in response to tax or reporting requirements at year’s end. The January effect is a challenge to the efficient markets hypothesis. A reasonably bold version of that hypothesis is that you can’t beat the market without inside information. All publicly available information—including corporate accounts, price history, and what month of the year it is—is already taken into account in the market price. January Effect says “buy on Dec. 31 and sell on Jan. 31”, it just shouldn’t yield consistent returns.

 

The history may indicate a very small, inconsistent inefficiency in the stock market. Exploiting a small efficiency may be all it needs for a professional trader. For average Joe, there are many other bigger mistakes he may make. Since January Effect is well known, there may be a very large number of investors making their bets based on the January Effect. Can small number of smart investors make money from mistakes made from a larger pool of average investors? It also won’t take long for all the obvious mistakes to disappear, because they’ve been so exploited. The date of January Effect would keep shifting until you are no longer sure there is such an effect.

 

One closing thought, if I were to start a family of mutual fund or ETF, I certainly would have one for “no-January-effect Russell 2000” , another for  “January-effect Russell 2000”.

 

Easy ESPP Mathematical Formula

When you are first hired to a new company, the human resource personnel would educate you that Employee Stock Purchase Plan (ESPP) is a fantastic deal etc. She would spend hours explaining how the 15% number  would nest you a profit.

An ESPP typically works this way:

1. You contribute to the ESPP from x% of your salary. The contribution is taken out from your paycheck.

2. At the end of a “purchase period,” usually every 6 months, the employer will purchase company stock for you using your contributions during the purchase period. You get a 15% discount on the purchase price. The employer takes the price of the company stock at the beginning of the purchase period and the price at the end of the purchase period, whichever is lower, and THEN gives you a 15% discount from that price.

3. Some companies have 2 years lock period, she would have a little harder time to explain to you how that work. This is because with the x% salary and the 15%, this gets a little more complicated. Eventually you will understand the concept, but you will forget it right away because in your mind you don’t have a very concise picture of this whole business.

Today, I am going to show you a concise math formula to help those engineers to understand it, if you still can’t remember it after reading it, you should quit your engineering job :

Let date2 be the start of the ESPP cycle, date1 be the end of the cycle.  date1-date2=6 months.  We use daten denote the nth 6 months prior to date1.

your_current_price=0.85*min(previous_locked_price,price(date1),price(date2))

and then set the previous_locked_price for the next period:

previous_locked_price=min(your_current_price,

                         previous_locked_price(date2),

                         previous_locked_price(date3),

                         previous_locked_price(date4),

                         previous_locked_price(date5))