Sell In May, and Go Away

[nodoodahs]

The "sell in May" is an oddity based on three "problems" - 1) the DJI is a 30-stock price-weighted index, 2) October 1987, and 3) indices vs individual stocks. Also, keep in mind that while the 54-year return of entering/exiting is $492,060, the 54-year return of staying put (in the DJI) year-round is $476,413. Is $15,553 worth the transaction costs and tax penalties for 108 movements in and out?

The "sell in May" doesn't apply to S&P500, Nasdaq, or S&P small cap 600. All of these have significant positive returns for May-Oct. For whatever reason, the returns for these indices are depressed May-Oct relative to Nov-Apr, but you'd have to include transaction costs and check the returns on 6-month CDs to get an economic benifit.

If you remove October 1987, the DJI average return from May 1 to October 31 is a slight positive, not negative. Basically the 10 worst months for the DJI since 1959 have included a couple of Sept, a few Aug, and some Oct. If you don't believe me, download the historical monthly closes to a spreadsheet and look for yourself. Removal of Oct 1987 helps stabilize the results of the other indices, too.

Lastly, it's easy to find individual stocks that "sell in May" doesn't apply to. The first one I checked, I downloaded adjusted (for splits and divs) closes for every month since 1980. The average return May-Oct is 14%, the average Nov-Apr is 20%. Yes, it has seasonality, but you would have lessened your return by getting away for the "poor" season.

[HenryTo]

A little bit of history on the S&P 500:

http://www.cftech.com/BrainBank/FINANCE/SandP500Hist.html

Here is something else I found interesting along the way:

http://answers.google.com/answers/threadview?id=387607

[nodoodahs]

Does the Stock Traders' Almanac provide adjusted closings for dividends and splits? The downloaded data from Yahoo!Finance showed a positive return (albeit small) for the summer months for Dow from 1950, for S&P500 and Wilshire. I did use the "adjusted closes" column for the returns. Is it possible that the Almanac referenced only prices and not total return? Or is it possible that the Yahoo!Finance data is incorrect?

What year did the S&P500 index get formed? I have found returns from 1982 on, but nothing previous. I would be skeptical of "reconstructed" data from prior to formation. I browsed the Standard and Poors website and googled the question, but did not find an answer.

Comment: I don't believe that Beta=Risk. Beta is "risk" only if you anticipate being forced to sell. A historic variability in price actually mitigates risk for a value or contrarian investor, because it increases the number of "buy" points available, when the price is appreciably below the intrinsic value.

[David Korn]

nodoodahs, thanks for the reply. Let me preface my comments by pointing out that I have no agenda with respect to the Best Six Months Strategy, so I am only interested in the facts behind this timing strategy.

A couple of points to make. I agree with you that the strategy doesn't apply to individual stocks as the vagaries of company profits wouldn't seem to be impacted by the seasons.

With respect to your comment about it only applyng to the Dow, according to the Stock Trader's Almanac, if you substitute the S&P 500 for the Dow, your $10,000 investment in the S&P 500 from May to October over that 54-year time frame would produce a gain of $349,165 versus just $7,102 for the November to April time frame. So it does seem to apply to that index as well. I haven't seen a study of the Wilshire 5000, but I would expect it to play out in a similar vein.

Your point about the difference between staying in the market versus exiting is well taken. My only response to that is that your beta or risk is much much lower if you are in cash half the year, during which time you can be earning interest. I appreciate your comments.

- David Korn
_________________
If you would like to learn about my service, just e-mail me at davidk555@earthlink.net or visit my website, http://www.BeginInvesting.com