Why "Sell in May and Go Away" works, and why we don't use it

You’ve probably heard of “sell in May and go away”, which states that the stock market from November-April of vastly outperforms the stock market from May-October, on average.
This phenomenon stretches back hundreds of years and applies to western stock markets. If an investor buys stocks from November-April and buys Treasury bonds from May-October, he would beat a pure buy and hold strategy of stocks in 60%+ of years.
But have you ever wondered why this seasonal pattern exists? Many investors think it’s random. It isn’t. There’s a simple logic explanation behind this.

What the BS explanation is

Some people provide the following explanation for this phenomenon:

Investors close out their books and go on holiday over the summer.

That is pure BS. Institutional fund managers and hedge fund managers do not sell everything and then chill on a beach for 6 months each year. Even when they’re on vacation, they still own stocks. Mom and pop investors are even less likely to sell their stocks before summer – they buy and hold. Quant funds like ours don’t sell everything and party in summer. Our models are always active.

What really causes this phenomenon – it’s all about the economy & earnings.

2 factors drive stocks in the medium-long term:

  1. The state of the economy.
  2. Corporate earnings

Both of these factors cause the “sell in May and go away” phenomenon to exist.
Economic data tends to dip a little in August and July because a lot of workers go on vacation and the pace of business slows down. Since the state of the economy and stock market move together in the medium-long run, this puts pressure on stocks in the summer.
In terms of corporate earnings, we do not look at 12 month trailing earnings (i.e. earnings over the past year). We look at 12 month forward earnings, which is basically earnings expectations. Earnings expectations are more important than actual earnings because this is what the media and investors focus on.
The U.S. stock market also moves in tandem with expected earnings in the medium-long run.
Every calendar year, earnings expectations:

  1. Start very high in January
  2. Plateau from January – March
  3. Slowly fall from April – May.
  4. Fall much more from June – August

This is a ridiculous game, but analysts play it ever year. They always set the bar very high for earnings expectations at the beginning of the year and then lower their expectations as the year goes on.
So as you can see, the downgrading of earnings expectations puts pressure on stocks from May – October.
Then towards November and December, analysts get excited about the upcoming year. They once again raise their earnings expectations, and the whole cycle repeats. This upgrading of earnings expectations puts upwards pressure on stocks.

Why we don’t invest based on this seasonal pattern

We completely ignore this seasonal pattern when making investment decisions and deciding our market outlook.
This pattern did not work in a lot of years. It beats buy and hold on a 50 and 100 year time frame, but often does not work during bear markets. In addition, the added benefits of using this pattern are not significant enough to be worthwhile.
You should only use this pattern if you know nothing about the stock market. Our models do not use any seasonality when determining market direction.

2 comments add yours

Leave a Comment