‘Sell in May’ was a costly mistake. Should you buy stocks now at record highs?

When should you get back into stocks if you followed the market-timing adage to “Sell in May and go away?” I’m referring to the six-months on, six-months-off seasonal pattern that is also referred to as the “Halloween Indicator.”

Investors who went to cash last May 1 are undoubtedly sorry they did, since they have missed out on powerful stock-market gains. The Dow Jones Industrial Average DJIA is ahead 12.5% since then on a total-return basis (through Sept. 26).

Should followers of this seasonal pattern stick to their guns and avoid stocks until Oct. 31? Or do they throw in the towel now? And if so, should they jump back in immediately or make a bet on which day is just right to do so?

The first step in answering these questions is reviewing the Halloween Indicator’s historical record. Since the Dow was created in the late 1890s, it has produced an average price-only gain of 5.3% between Halloween and the subsequent May 1 (the “winter” months) and 1.9% in the other six months of the year (the “summer” months). The Halloween Indicator’s statistical foundation rests on the significant difference between these two averages.

This summer’s gain of 12.5%, while above the long-term average, is hardly unprecedented. Since the 1890s, there have been 21 summers in which the Dow did better than it has so far this summer. So there’s no statistical reason to give up on the strategy just because the stock market this summer has been as strong as it has.

Here’s another important feature of the Halloween Indicator’s historical record: Even though the U.S. stock market typically doesn’t do as well during the summer months, it nevertheless does not lose money on average. This is especially the case when you take dividends into account. So a follower of the Halloween Indicator should expect to underperform a simple buy-and-hold strategy. The follower’s realistic hope is outperforming the market on a risk-adjusted basis.

The Halloween Indicator strategy jumps over this hurdle, as you can see from the table above. Its Sharpe Ratio, which is a measure of risk-adjusted performance, is 0.66 — 11% better than the 0.59 for buying and holding. (The table focuses on performance since May 2002, which is how far back data extend for all three of the strategies listed in the table.)

Getting a jump start on Halloween

Some investment professionals have tweaked the Halloween Indicator in hopes of getting into the stock market in October at better prices than what prevails on Halloween, along with getting out of stocks in April at prices better than those in May.

One such investment adviser is Jeffrey Hirsch, editor of the Stock Trader’s Almanac newsletter. He uses a trend-following indicator known as Moving Average Convergence Divergence (MACD) to determine the optimal time for making stock-portfolio moves. The table above shows that Hirsch’s strategy has reduced risk by even more than the traditional version of the Halloween Indicator, though with slightly less return. On a risk-adjusted basis, the strategy beats a buy-and-hold approach and essentially equals that of the Halloween Indicator itself.

This year presents a special challenge to Hirsch’s modified Halloween Indicator strategy, since the market is already in an uptrend and therefore already on an MACD buy signal. Per the rules he has specified for his strategy, he issues a buy signal only when a new MACD buy signal is triggered on or after Oct. 1.

The market’s strength notwithstanding, Hirsch writes in the October issue of his newsletter that such a buy signal could still happen relatively soon. 

A loss of upside moment and “a mild consolidation of recent gains” could create the preconditions for a buy signal “within a matter of days.”

By Mark Hulbert, Sept. 30, 2024