Sell In May? Empirical Evidence On A Market Adage

|
Includes: BXUB, BXUC, DMRL, EPS, IVV, PPLC, RSP, RVRS, RYARX, SDS, SFLA, SH, SPDN, SPLX, SPUU, SPXE, SPXL, SPXN, SPXS, SPXT, SPXU, SPXV, SPY, SSO, UPRO, USMC, VFINX, VOO
by: Ploutos
Summary

This article adds empirical evidence to the market adage of "Sell in May and go away".

Markets have struggled thus far in 2018, producing negative returns through April.  Does this adage suggest further weak returns ahead?

Building off previous academic research, we see that there is seasonality in market returns across geographies, asset classes, and long time intervals.

As the calendar moves to May, investors must ask themselves: "Sell in May, and go away?" Is this wise advice or folksy nonsense? In this article, I examine the striking outperformance of the S&P 500 (NYSEARCA:SPY) in the six months between November and April versus the lackluster performance between May and October.

Using more than sixty years of data, dating back to when the benchmark U.S. equity index first went to 500 constituents, I break market returns into these two semi-annual components. This study yields some surprising results. As seen in the table below, the annualized return for domestic equity investors for the six months from May to October inclusive was a disappointing 4.4%. Conversely, returns from November to April were an astounding 16.3% annualized.

The question we will seek to answer in this article is whether this calendar effect is spurious, or whether there is something to this advice that Seeking Alpha readers should heed.

As I often do when examining market phenomena, I am turning to academic research on the subject, which I will supplement with my own insights. In 2001, Dutch researchers Sven Bouman and Ben Jacobson published: The Halloween Indicator, 'Sell in May and Go Away': Another Puzzle. The authors found that this timing mechanism held in 36 of 37 developed and emerging market economies in a dataset stretching from 1970-1998. They found that the effect is particularly strong in European countries and is robust across time. In 2013, Sandro Andrade, Vidhi Chhaochharia, and Michael Fuerst published, "Sell in May and Go Away" Just Won't Go Away. Building on the Bouman-Jacobson study, the authors found that the "Sell in May" effect persisted with the same economic magnitude from 1998 to 2012.

If the market believed that there was an arbitrage to be captured after the 28 years of outperformance in the first dataset, we should not have seen continued outperformance over the next 14 years. There are always strange anomalies in a given dataset, so extending the study to out-of-sample periods adds additional credence. The Bouman-Jacobson study further showed that this seasonal effect has been in place in data in the United Kingdom stretching back to 1694! For perspective, that dataset pre-dated the birth of George Washington by 38 years.

We know that "Sell in May" has persisted across various time horizons for equity markets. Further support for this market anomaly is seen in currency and credit markets. This suggests that there is seasonality in financial markets' aggregate posture towards risk that is markedly different in these two semi-annual periods.

Why does this seasonality exist and why has it persisted? Perhaps seasonality is induced by vacations. This might suggest that investors have less demand for risky assets when they are away from the office. This possible explanation might explain why the "Sell in May effect" is even stronger in Europe where the summer holiday is longer. There is probably some merit to this calendar effect. I have shown a strong and persistent calendar effect - the so-called Santa Claus effect - in high yield bond markets in December and January that I believe is caused by a relative lack of primary debt supply in those months. While calendar effects are a violation of the Efficient Market Hypothesis, it does appear that market structure and human behavior could allow for their existence.

Another interesting observation on the seasonality of returns was captured in a 2003 paper by Mark Kamstra, Lisa Kramer, and Maurice Levi of the Federal Reserve Bank of Atlanta. In Winter Blues: A SAD Stock Market Cycle, the authors documented low returns prior to the winter solstice and strong returns following that date. They posit that seasonal affective disorder, a documented medical condition linking length of day to feelings of depression in some people, influences risk-taking and market returns. Adding support to this claim, the influence in the dataset appears stronger at higher latitudes (shorter days in winter) and is present in the Southern hemisphere at a six-month lag (opposite seasons from the Northern hemisphere).

Maybe "Sell in May" is bolstered by these calendar effects and weather's impact on mood, but these could be simply small variables adding heft to what is simply a spurious correlation in a relatively short dataset. Unfortunately, we cannot wait around for another sixty years of S&P 500 data to test this hypothesis.

What does that mean for Seeking Alpha readers? While I was surprised at the evidence supporting market seasonality, I am certainly not advocating selling your risky assets and sitting in cash for the next six months. Over this sixty-year period, investors who sat out the lower returning semi-annual period would have still sacrificed returns. I think investors, especially young investors with a long-time horizon, should capture equity risk premia over long time intervals. Timing markets consistently is a difficult proposition. Digging deeper and investing more in down markets with a long-run view is likely better advice for most investors.

For investors with a more tactical bent to the market, I hope this data was at least thought-provoking. If you want to test a market anomaly, you would want to test it across long time intervals, across geographies, and in different asset classes. As strange as it sounds, the "Sell in May" adage seems to correctly capture an abnormal period of seasonality in markets.

Disclaimer: My articles may contain statements and projections that are forward-looking in nature, and therefore inherently subject to numerous risks, uncertainties and assumptions. While my articles focus on generating long-term risk-adjusted returns, investment decisions necessarily involve the risk of loss of principal. Individual investor circumstances vary significantly, and information gleaned from my articles should be applied to your own unique investment situation, objectives, risk tolerance and investment horizon.

Exhibit: S&P 500 Returns by Month - Trailing 61 Years

Disclosure: I am/we are long SPY. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.