Sell in May?

While I’m lying on the beach (and I go in the Atlantic every year this weekend), here’s a very interesting look at the old axiom, "Sell in May and then go away." 

Does it have any scientific basis? My South African friend Prieur du Plessis of Plexus Asset Management notes that the long-term statistics support the notion that the best time to be invested in equities is the six months from early November through to the end of April, while the “bad” periods normally occur over the six months from May to October. Surprisingly, this is true globally also:

"A study of the MSCI World Index, a commonly used benchmark for global equity markets, reveals that since 1969 "good" periods returned 8.4% per annum while investors were actually in the red during the "bad" periods by -0.4% per annum. Interestingly, this phenomenon – of "good" period returns outperforming those of "bad" periods – applied to all 18 markets where MSCI computed index returns.

“Sell in May and go away” also holds true for the US stock markets. A study by Plexus Asset Management of the S&P500 Index shows that the returns of the “good” six-month periods from January 1950 to December 2006 were 8,5% per annum whereas those of the “bad” periods were 3,2% per annum.

A study of the pattern in monthly returns reveals that the “bad” periods of the S&P500 Index are quite distinct with every single one of the six months from May to October having lower average monthly returns than the six months of the good periods."

A review of the basic monthly returns since 1950 show the weaker periods:

Monthly_spx
chart courtesy of Plexus Asset Management

>
The summer months are particularly slow; also impacting returns: the crashworthy tendencies of September and October.

Print Friendly, PDF & Email

What's been said:

Discussions found on the web:
  1. Christopher Laudani commented on May 23

    Hey Barry,

    What about the new market axiom: “Buy everyday and never go away!”

  2. Michael C. commented on May 23

    Tony Crescenzi on RM has an interesting commentary on recent jobs data:

    “A key reason for the overcounting is because each month the government has been adding in jobs it thinks are being added through new and expanding business, failing to account for the fact that previously expanding industries — the construction sector in particular — are now contracting.

    A case in point is the April payroll report. The BLS added 317k jobs to the total, believing that net business formation for the month followed the pattern of previous Aprils. Of that 317k, the government assumed that the construction sector added 49k jobs. In other words, the government added 49k jobs to the total to adjust for job growth it believes it missed in its payroll survey because of increases in net business formation in the construction trade. This is based on the behavior of construction jobs in previous years. With the industry now contracting, this is a false assumption. The government is overstating the job count, and it is showing up in the spending figures. “

  3. James Bednar commented on May 25

    While I’m lying on the beach (and I go in the Atlantic every year this weekend)

    Barry,

    Please, spare us any pictures unless whoever you might meet is both famous and beautiful.

    The last thing I need to ruin a perfect weekend is a picture of you and Kissinger in matching Speedos.

    jb

  4. Leisa commented on May 25

    While this “Sell in May…” aphorism is interesting, it would be more beneficial to look at the sectors that benefit from this time frame–so while one is laying on the beach there portfolio does not have to take a full vacation.

  5. Aubrey commented on May 25

    Sell in May but don’t go away. Invest elsewhere; like emerging markets.

  6. ~ Nona commented on May 26

    I dunno.

    Kissinger and Barry in matching Speedos? Could be interesting.

    I vote for seeing the picture.

  7. Tarik commented on May 31

    May to October looks like a good time to put index options, if this axiom holds up over the long term.

Posted Under