Are "calendar effects" for real?

[Chart_oasis_man]

Calendar effects like Halloween indicator, January effect, Santa Claus rally, Sell in May and go away, October effect are often mentioned, but people quite rarely talk about their own experiences related to these.

I have made a little study about this (Each chart represents the monthly statistics of a particular stock.):







If there was no calendar effects at all and everything happened randomly, then I would expect that the per month averages would be near 50% for frequency of increase (like tossing a coin) and near 0% percent for the average increase, but for December on the first chart the price rose cca 10 times out of 11 (cca. 91%). It seems to be far from a coincidence...
Now I think that the impact of seasonality varies with invesments.

On the other hand there are lots of real economic factors that show seasonality, e.g.: there is an increasing demand for local currencies at popular holiday destinations when lots of people are heading there in the summer.

So what do you think? Or should we call MythBusters? :-)

[jpg7n16]

deleted

[Slug]

Time periods? Are those aggregate decades?

[dunnrobert700]

Each is a single stock? OK better than one but stocks are correlated so might as well have used an index...total market for example. Still no biggy.

So you looked over several annual periods for each stock? How many? last 5 years? 10 years? 30 years?

My biggest skepticism on this is that if it were this easy, the first guy to figure it out could make nearly unlimited returns, eliminating the effect along the way. I'm not sure consistent seasonal stock price trends are even theoretically possible.