Halloween Stuff

I was reading about irrational decision-making and wondered if ...

>If you were irrational, right?
Me? Never! But what might affect investors' decisions, some irrational influence, like maybe ...
You got it! I thought that maybe, just maybe, Halloween Hobgoblins were at work.
Investors are frightened. They dump their stock ... and the market tanks.
And I get this, comparing the average return on the last day of October to the average of all end-of-month returns.
I look at the difference between these two averages and see that Hobgoblins actually like halloween.

With few exceptions, stocks have larger gains on Halloween than on other end-of-month days.

There's a spreadsheet to play with, here: halloween.xls
>So end-of-October ain't bad, eh?
No, but end-of-February is pretty bad:

It seems there's something called the Halloween Indicator.
It suggests you be in Cash during the Summer months (starting in May) then Buy after Halloween (Nov 1).

>Isn't that "Sell in May then Go Away"?
Yes! Have you heard of it?

Well, we decided to test the idea and got this, comparing the total (cumulative) gain fo the period (Nov-to-Apr) and that for the period (May-to-Oct).
We looked at the 10-years from Jan 1, 1999 to Jan 1, 2009 ... and got this for various country indexes:

>Uh ... I see it's getting close to the end of October. Excuse me while I Buy some stocks ...

Oh, I forgot to mention the other spreadsheet: halloween2.xls
You might like to replace country indexes with stock symbols ... just to see if'n "Sell in May then Go Away" works for stocks.