Sell in may and go away? On this concept, there have been literally thousands of articles written, dozens of opinions expressed. Every year, this theme is revisited. Every year, there are two camps, which battle on to argue whether there is an effect and whether it is actually useful in practice.
My opinion is that nothing is black-and-white. Even in the greatest misfortune you find good and positive effects for you. It is not about what happens to you, but how you react to it and how you take advantage of it.
Just for the record, I am not a huge fan of market-timing techniques. Generally, I hold the opinion that it is much easier to predict the exact price or price range, than to predict when exactly this price will be hit in the future. However, if some market-timing technique works for some investors, very good for them and I fully respect their accomplishment.
I hope to challenge some of your views on this phenomenon of Sell in May. I will offer an alternative point of view by showing usefulness of this effect for long-term investors. I will provide you with hard numbers that clearly prove that statistically, there is a significant Sell in May effect. I will show you truly compelling arguments for why this year, again statistically, there should be a significant sell in may effect. More importantly, I will show you specific ways, how even long-term investors, value investors and passive portfolio holders can take advantage of the Sell in may effect this year.
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. The Sell in May effect, therefore, may be better known today as the Halloween indicator, which advises to stay invested from October 31 until April 30.
An academic paper written about a study performed by Bouman and Jacobsen (2002) shows that the Sell in May effect occurred in 36 out of 37 countries examined. The first occurrence and mention of this effect dates back to the 17th century United Kingdom, where it occurred in 1694, more than three centuries ago.
Why does this effect even exist?
Academics, researchers and economists are not actually sure. Some hypotheses include the seasonality of risk in the markets, others claim the effect is irrational.
What your wife says is always right and always the most important. When I just casually mentioned to my wife that I am writing an article for Seeking Alpha about the Sell in May effect, and explained it to her that it means staying out of the market for half a year, starting in May, she didn't even take a split of a second and fired away instantly.
Well, it's obvious. People go to vacations. Plus they need the money to pay for them.
Who knows, she could just be right. She could have given me the two most significant fundamental reasons that could be behind the effect. People seasonally need more money to pay for summer adventures, and they are actually out of town for part or all of the summer. Well, it is always a magnificent idea to ask a total outsider on the opinion to get an alternative and simple, real-world opinion.
Why the Sell in May effect is highly likely to happen this year
- Psychology of individual investors - many investors got ahead of their 2007 highs (nominally). Psychological research shows that it is at least two times harder to sustain a loss than to experience a gain or a win. The investors who got ahead of the previous highs or ahead of the level at which they purchased are more likely to sell just to realize the gain and avoid the risk of falling under water again.
- Behavior of large fund managers - their funds have probably experienced spectacular runs in the last six months, one of the largest in history. They will not want to damage their stellar performance in the eyes of investors. Therefore, they will trim some of their holdings to realize some of the fat gains they are sitting on.
- Expiration of the Bernanke Put - the quantitative easing has to end or slow down sometimes. Or not. This is a controversial topic. The discussion is best left for a separate article. However, the risk that the QE will be slowed down at least temporarily, outweigh the possibility of even larger money printing. On the other hand, I realize that the world main powers have been playing a very high-stakes money-printing poker game for the past few years practically non-stop. Just yesterday, Japan raised the stakes. I am afraid the players are all prepared to go all in. Whose turn is it next? The ECB? The FED? Oh, I almost forgot that China is playing too, and it has a large stash of cash and chips. The FED better not be on a big or small blind. Or just downright plain vanilla blind. Otherwise, we can probably kiss the Sell in may goodbye this year and just keep partying.
What do the numbers say?
First, let's dig in deep to find out whether there actually is any statistically significant phenomenon of Sell in May and go away
Numerous respected studies show the following:
- The effect is centuries old (source)
- unlike many other effects, this anomaly did not cease to exist after its discovery by investors. Media bombard us every year and yet, statistically, the effect prevails, as the new study from 2012 proves.
- Analyzed on the time series of the last more than four decades (1970-2012), the average positive difference in the S&P 500 (SPY) equity returns held in Nov-Apr period compared to equities held in May-Oct period is 6.46% for the US equities. In detail, the returns in the May-Oct period are -0.39%, whereas the returns in the Nov-Apr holding period are +6.07%.
- The Sell in May phenomenon is not exclusive to the US at all. Actually, the average effect across 37 countries studies is 10%. Russia is the champion with over 20% seasonal difference.
- Unlike in case of most other market anomalies, the trading strategies based on the Sell in May phenomenon don't require too much trading. Transaction costs are extremely low as a result.
- The most salient fact, which I will emphasize again, and which sets the Sell in May seasonality effect apart from, and above, the other seasonality effects, is that it still lasts even after it has been discovered by investors. The positive US Sell in may effect has actually increased a little bit in the last 24 years as compared with the previous 18-year period. However, the rise is insignificant. 6.86% now versus 6.24% previously. Forbes late last year ran an article confirming the effect is still valid.
An important point on the above results should be mentioned. The numbers are just the average results. They don't mean to state that this year specifically, this strategy of Sell in May will generate 6% higher return than the buy-and-hold strategy. It merely says, that if you perform this strategy over many years, on average, you are likely to achieve these positive 6% above normal buy-and-hold returns of the SPY. Provided that this seasonality effect does not disappear of course.
Another point to note is that the analysis apparently does not include the dividends. It just includes the values of the market indexes. This will not change the result of the strategy in terms of the excess 6% difference in return in the two holding periods. However, including the dividends will increase the total average returns in both of the periods. For the May-Oct period this will mean moving from negative -0.39 to a positive 1% (I estimate the dividend yield of 2.39% p.a. just for the simplicity, so please take the 1% with a grain of salt. I also presume here that, on average, exactly half of the dividend amount will be distributed during each half of the year. Please remember that the 1% return is a value for just the six month period. This actually represents a return of 2% p.a.).
The implication of including the dividends, and even perhaps dividend reinvestment, into the analysis, is particularly powerful and revealing for the long-term, passive, value and dividend investors; they will still achieve positive returns if they hold during the May-Oct less favorable season.
The conclusion for the long-term investors with value and dividends in mind is that selling entire equity positions in May-Oct period and staying in cash, let alone going short equities, makes no sense if the only reason should be the Sell in May effect.
Nevertheless, smart investors should still take advantage of the apparent statistically significant Sell in May effect by different means. Let me go through them now in detail and offer some specific options for concrete actions to take in the upcoming six months.
How long-term investors can profit from the Sell in May effect?
1. Portfolio rebalancing
If you plan to rebalance long positions in the next six months like I recommend in my article about retirement portfolio, use this opportunity now in the first half of April to do that part of long-term portfolio rebalancing which involves selling any long positions in the equity markets, such the S&P 500 or DJI (DIA). Selling in the first weeks of April not only gives you the benefits of the Sell in May effect. You will also be selling into the seasonally strong weeks in the beginning of the upcoming earnings season.
Additionally, if you plan to purchase any long positions in equities in the next 6 months, either from the proceeds from the April sales of long equity positions, or from additional outside cash inflows, I suggest waiting until the end of May, to see how the Sell in May effect starts to play out this year, or until we see the first meaningful correction of this year of at least 4-5%.
2. Dynamic asset allocation
If you manage your portfolio actively, even though you definitely should not sell all your long equity positions, you can allocate smaller share of your portfolio to stocks during the May-Oct portfolio. However, make sure you can generate nominal returns over 2% p.a. elsewhere. Otherwise moving out of stocks makes little sense because on average, they still generate positive returns even in this weaker season
3. Buy protective puts
The next few weeks might be a good opportunity to buy some cheap and effective protective pus on parts of your portfolio if you have some risky high beta stocks which might fall in higher magnitude compared with the broad market. However, I have to state clearly that I am not a big fan of buying protection in the form of puts because in the long-term, the time value of the options you lose will eat into your long-term returns. If you can stand the volatility and maximum drawdown of your portfolio, I recommend just weathering the storm without buying the puts
4. Sell covered calls
Although I am not a particularly vivid proponent of permanent covered call writing on your entire portfolio because in the long-term if you write them continuously all the time, I have yet to see a compelling real world study which would convince me that there are any extra returns. If there are, aren't they eaten away by the extra trading fees, tax consequences of the short-term capital gains?
Nevertheless, even if you are not convinced that writing covered calls works if applied continuously, you might want to write from May to October and not from April to September. Such covered call writing is likely to generate positive results, because covered call writing is generally profitable if the markets seesaw, generating increased volatility, or if they are in a downright downturn.
5. Buy VIX (VXX)
There is actually no direct way how you can buy volatility itself because it is just a number based on the volatility of S&P 500 options and not an index composed of actual underlying assets. You can only buy Volatility futures. For this reason, being long VIX is a strategy recommended only for investors who understand futures and especially the effects of contango, because the VXX ETF actually represents the volatility futures value, not the spot volatility value itself.
This contango effect can have a positive or negative impact on the price of the VXX ETF. Buying the VXX or other products that are long volatility futures is recommended only for short term periods of times, such as a few weeks, mainly due to the already mentioned contango. Being long volatility is more a trading and speculative strategy than an investing tool although it can serve as a protection of your long portfolio. Moreover, volatility generally spikes as your other asset classes fall. So being long volatility can also diversify your portfolio and decrease the maximum drawdown and the volatility of your portfolio. However, beware, volatility is extremely volatile and unpredictable itself:). I truly recommend it only to highly experienced investors.
One final tip on the VIX. There is another way how to profit from volatility, among others, although it will not decrease the max drawdown of your portfolio because you will have to let your portfolio fall in value first, and only then take action below. It requires that you hold part of your portfolio in cash, or be willing to sell other assets which will not be down, such as perhaps bonds. You then actually just wait until the volatility spikes and only then sell the volatility, be short volatility with your cash. Then all you have to do is wait for the volatility to come down. Which can take just a few days, a few weeks. But it can also take a year or two. But what is one or two years for long-term, patient investors like us, right?
If you would like to explore an alternative way of looking at hedging, you can check out my previous article about natural hedging.