You might have heard that this is a great year to buy American stocks, because this is a Presidential election year. After all, it is absolutely true that in every American Presidential election year since 1928, only 4 out of those 21 years saw a negative performance by the S&P 500 Index or its equivalent. Does that sound good? These kinds of systems have many names but are most commonly known as market cycle systems. It is worth considering carefully whether they are ever worth following.
What is the System’s Logic?
The first thing to ask yourself when considering such a system is whether the logic behind the system is compelling. There are market cycle systems based upon lunar cycles. Before you laugh, you might want to consider the fact the senior police officers have gone on record saying they feel there is more violent crime on full moon nights. However, rigorous surveys have found no significant lunar influences on human behavior with the possible exception of sleep cycles.
Possibly a more intellectually compelling case can be made for buying stocks in a Presidential election year. The logic presumably goes like this: a President wants to get re-elected, or wants their preferred and associated candidate to win, and so the administration does what they can to generate an economic boom, which drives up the stock market, and which is designed to end as soon as the polls close.
This could be a plausible argument, although there are several serious flaws in it, such as the independence of the Federal Reserve. However, even if you did find the evidence compelling, you have to move to the second step, which is to take a much closer and critical look at the data supporting the market cycle system’s case.
Checking the System’s Data
We saw that 17 out of 21 years in the sample produced positive performances in the case of the Presidential election year stock-buying system. This sounds pretty good: a win rate of almost 81%.
The first question to ask is how this compares to ALL of the years since 1928. Twenty-four of these sixty years produced negative returns, which means that 60% of all the years were winners. Now while 81% seems greatly superior to 60%, we have to bear in mind that statistically sixty is a much better sample than twenty-four. In fact, samples under two hundred tend to be very unreliable statistically. So for reasons of simple statistics, you should be very wary of risking money on a system that has only twenty-four sample points, especially when the result seems not so very different from a typical year! Ideally, you want something that has been tested over many years with thousands of samples in a wide range of market conditions, backed by a very strong intellectual logic.
Another aspect of the data in our Presidential election year system that is worth taking a closer look at is the average return, not just the proportion of years giving some kind of positive return. The average return is 9.45%. As it happens, the average annual return from 1928 to 2015 was 11.41%. So even assuming a mere twenty-four data points are acceptable, just by taking a closer look at the data we can see this system is not all it is cracked up to be.
Compelling and Effective Market Cycle Systems
If you are going to seek to trade this type of system, you want to use something that has been well researched and has a lot of supporting data behind it. There are a few well-known examples.
First of all, there are several academic studies showing that a “momentum” effect exists in speculative financial markets such as stocks, commodities and Forex. This means that if the price of something has been going up or down over a period of months, then over the next few weeks it is more likely than not to continue in the same direction. This is also known as “trend following”, and the statistical truth of such market cycle systems has been established over hundreds of years in all markets using tens of thousands of samples.
The most successful Forex trend trading strategies have focused on trading USD pairs only where prices have moved in the same direction over the previous three and six month periods.
Less well-known systems exploit the “Monday effect” in stock markets, where stocks tend to rise excessively on Mondays, by buying stocks on Monday mornings in rising markets.
In Forex, my own research shows that swing and position trades in USD currency pairs are a little more profitable where trades are entered on Monday, right at the beginning of the week, or right at the end of the week on Friday, as shown in the chart below. The study was conducted over the last 14 years just entering on H4 charts after deep pullbacks began to reverse back into the direction of 3 month and 6 months trends, using the major USD pairs. The numbers are average expectancy by reward to risk ratio. We can see the more immediate rewards were superior on Mondays and Fridays. Excessive returns are highlighted in yellow:
One reason for this in the Forex market could be the triple overnight interest which is charged or paid (usually charged) by Forex brokers on Wednesday night in the middle of the week.
The “Friday effect” is not so well known, but has been observed to work in strong trends in the Forex market, exploiting the fact that many traders are frightened to hold positions over the weekend when they cannot be exited as brokerages are closed and usually will not honor stop losses triggered during weekend hours.
Conclusion
If you are going to trade market cycle systems, do make sure they are well-supported by statistical research, and that they are based upon common sense.
