Sell In May And Go Away? No. Buy In May And Sell Away

Includes: SPY
by: Todd Feldman

This time last year we wrote a Seeking Alpha article titled "Sell in May and Go Away? Yes.".

Sell in May and Go Away refers to selling in May and waiting until November, thereby holding the market from November to April and remaining in cash or bonds from May to October.

Investing based on this seasonal effect has worked the past two years. Research supports that this strategy has worked for a long time (even though the research does not account for taxes, transaction costs, dividends, etc.).

Preliminary research of my own suggests that the seasonal effect is stronger in bear markets, meaning that the majority of the losses for a year in a bear market cycle occur during May to October. In a bull market cycle the majority of the gains come in November to April, however, the May to October period also experiences gains.

If one can identify we are in a bear market cycle it typically makes sense to sell in May and go away, otherwise it does not. It appears that we are still in a bear market cycle based on indicators such as unemployment, economic growth, etc.

However, for those that sell in May and buy in October we believe there is a high probability that you may be buying at a higher price than you sold in May of 2012. We are forecasting that the high for 2012 will occur in the summer. In addition, we predict a high probability of a correction by the end of 2012 to the beginning of January. Therefore, we are suggesting the opposite for this year, buy in May and sell in October.

Why do we predict such an outcome? We use our own behavioral indicators to detect inflection points and cycles in stock markets. The current evolution of our US behavioral indicator suggests that we are in a cycle very much like 1986-1987. Below is an evolution of the S&P 500 price index during three periods. The starting point of the index is at a bottom in September of 1986, August of 2010, and October of 2011. Hypothetically, let us assume we are in April of 2011. Just by looking at the overlay it would suggest that 2010-2010 was tracking 1986-1987. However, the similarity ended in April of 2011. The S&P 500 topped out in April/May of 2011 whereas in 1987 the S&P 500 kept climbing.

The important point is that our US indicator did not suggest a similarity with 1987 but another year that signaled a bottom in August. Therefore, we sold in May as written in our article last year. However, this year our indicator is signaling a similarity with 1987. (We document more specific exit and entry points in our newsletter.) Therefore, the overlay is more meaningful today.

Figure 1: S&P 500 Price Index for Three Periods

Click to enlarge

X-axis is the number of days. S&P 500 index price is benchmarked to 100.

Based on our analysis a pullback between 5-7% will occur sometime in April/May. It may be already here or it may occur in May. We suggest that one take the opportunity to buy more SPY at any 5% pullback. Then wait for the summer to end to sell or or the fall to sell before going the winter.

Disclosure: I have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours.