Famous stock picking rules, honestly tested
Does "Sell in May and go away" work?
The old saying: hold stocks from November to April and stay out from May to October. Here is each year's two halves for the US and Hong Kong markets.
US 2016–2026: holding the market only from November to April made 7.1% a year; holding all year made 15.3%. Hong Kong 2018–2026: holding the market only from November to April made 6.9% a year; holding all year made 2.5%.
US, 2016–2026
November to April beat May to October in 7 of 10 years; the middle November–April made 7.3% and the middle May–October 7.4%. 10 years is a small sample: one bad summer moves the result a lot.
Each year (10), newest first
| Year | November–April | May–October |
|---|---|---|
| 2026 | 6.0% | 5.7% |
| 2025 | -1.9% | 23.7% |
| 2024 | 20.9% | 14.0% |
| 2023 | 8.6% | 1.3% |
| 2022 | -9.7% | -5.5% |
| 2021 | 28.7% | 10.8% |
| 2020 | -3.1% | 13.4% |
| 2019 | 9.7% | 4.2% |
| 2018 | 3.8% | 3.3% |
| 2017 | 13.2% | 9.0% |
Hong Kong, 2018–2026
November to April beat May to October in 5 of 8 years; the middle November–April made 7.2% and the middle May–October -7.0%. 8 years is a small sample: one bad summer moves the result a lot.
Each year (8), newest first
| Year | November–April | May–October |
|---|---|---|
| 2026 | 0.1% | -7.3% |
| 2025 | 9.9% | 19.5% |
| 2024 | 4.5% | 18.1% |
| 2023 | 35.6% | -11.2% |
| 2022 | -16.5% | -27.6% |
| 2021 | 19.3% | -9.5% |
| 2020 | -8.2% | 0.6% |
| 2019 | 19.8% | -6.7% |
The US market with dividends: the S&P 500 fund SPY. Each half runs from the last close on or before 31 October or 30 April. "Sell in May" holds only the November–April halves and earns nothing in between (no interest is counted); both are compounded over the same whole years.
Why could summer be weaker?
Explanations offered include lighter trading over the summer holidays and money flowing in at the start of each year; none is settled, and the effect varies a lot from year to year and market to market.
Where does the rule come from?
The saying comes from the London market. Bouman and Jacobsen (2002, American Economic Review, "The Halloween Indicator, 'Sell in May and Go Away': Another Puzzle") studied the effect across many countries' stock markets.