It's time for "sell in May" to go away!!!

[nodoodahs]

This, and most other "Market Timing" strategies, just doesn't make sense. I went to Yahoo!Finance and downloaded DJI closes and adjusted closes, daily, from 1/3/1950 to 5/6/2005.

Adjusted close for 5/6/2005 = 10,345.40
Adjusted close for 1/3/1950 = 198.89

Gain is 10,345.40 / 198.89 = 52.01569 or turning $10,000 into $520,156.90. IF and ONLY IF you hold it in the market. If you take it out, you pay taxes, either short-term capital or income, on your gain every year. If you buy and hold, you pay long-term capital gains, once.

Let's download the monthly adjusted closes for the DJI and look at them. If you average the monthly returns for DJI from 1/3/1950 to 5/6/2005, you get a table like this. The "Jan" result is the gain from the first closing day of January to the first closing day of February, etc.

Jan 0.2%
Feb 0.9%
Mar 1.8%
Apr 0.1%
May -0.1%
Jun 1.0%
Jul 0.0%
Aug -1.1%
Sep 0.6%
Oct 1.7%
Nov 1.8%
Dec 1.3%

These are monthly, not annualized. Obviously the summer months have weaker returns. This is a statistic - but is there an economic benefit??? ... and is there actually a net loss over the summer???

I'm using adjusted closings (for splits and dividends) and Yahoo!Finance. Maybe that's the difference in the numbers, but I show an average 8% gain Oct 1 - May 1, and a 0.4% GAIN from May 1 through Oct 1. What gives? Is this the effect fading as it is more publicized? Or perhaps did his almanac source not used adjusted for splits and divs closings? That's possible, since my total gain is higher than what he listed.

Oh well. ASSUME a loss for May 1 through Oct 1. We could either follow the "sell in May" or "buy and hold."

So on year one, I put $10,000 in DJI on Oct 1 and pull it out on May 1, getting an average 8% return (16.6% annualized!). I then pay taxes on it. Assume 28% for my tax bracket. My gain after tax is 5.76%. I then put it into some cash holding (6-month CD's?) and pay tax on the interest. Then I put the remainder back. Hmm. This doesn't seem good. Assume an average annualized 2% on my cash, that's 1% for six months, turned into 0.72% for after tax, total return is averaging 6.52%. So the summer months would have to average a negative -1.4% to make this worthwhile. But based on his calculations the summer negative is closer to zero than that, and based on mine, the summer negative doesn't exist.

"Sell in May" is a statistical phenomenon that is not actionable, and as such, should be of no interest (other than entertainment) to a serious investor.

[HenryTo]

I think David's reasons are pretty sound - the fact that traditionally, the May to October period has underperformed relatively is definitely very interesting. For example, one can say this: if we already had four up months ending April, then the case for selling here may be very compelling, but I wouldn't trust it right now because we just had a huge down January, a down March, and also a down April.

Conversely, if the market crashes between now and October, then the case for buying equities in November becomes very compelling, but probably not as compelling if we get good performance all throughout the summer.

At this point, I am relying on my indicators for the most part but the part about mutual fund inflows/outflows and that the summer is usually a slow time for the Street is certainly interesting. I guess if a huge amount of money shifts to the Southern Hemisphere, this observation may not be as compelling.

[nodoodahs]

IRA will limit the taxes, but will also limit your opportunities - limited amounts contributed annually and not able to touch it without penalties until you are far older than I want to be when I retire (and possibly expatriate). It's an interesting phenom, and I'm curious about the reason(s) why it occurs. I do wonder why Yahoo! has different returns from your friend's reference. Either one of our (his or mine) calculations is off, or his reference material isn't adjusted for splits and divs, or Yahoo has an error. My main gripe is that it's portrayed as a "strategy" when it's more of a "curiousity." IMO, YMMV.

[HenryTo]

All this talk regarding "Sell in May" is probably overblown, but I think David wrote a very well thought-out article outlining just what the heck people are talking about out there. Whether anyone can use this to consistently make money over the long run is definitely very debatable. 100 years is also not a good enough sample size - I bet all the stat people will agree with me here.

For example, we (and David) are currently long - at least for the ST (next six to 12 weeks) anyway. January was down. So was March and April. To say that "Sell In May and Go Away" will work this year just because it has worked in other years is laughable - if this theory is working this year, then at least three of the four months from January to April would've been up, and not down.

2000 was similar. The market bottomed out in May and proceeded to move up until Labor Day. And then everything just toppled over.

By the way, one can just avoid paying the taxes from opening an IRA account. And a lot of people out there are still sitting on capital losses..... sorry, just playing with ya.

Take care,

Henry