Dec

20

A little chart gazing I was doing showed the following:

Not sure if it's statically significant i.e 14 years only set, and 2 month window, but I can hear a strangle coming up, and maybe even a small backpocket outright short,

March –April Turning Points in U.S. SP500

1997-2010·

5 out of 14 years, the March/April month has seen a turn that has resulted in the price closing the year at the inverse extreme.

11 out of 14 years has seen the market not break the March/April extreme in price for at least the following 3 month period.

Note: Bill Gross of Pimco has suggested that in the past 15 years, every time the fed funds rate was higher than thenominal GDP growth rate, assets such as stocks and/or housing always fell. He even suggested that the best way to price the fed funds rate would be 100 basis points below the nominal GDP growth rate.

This is not a situation at the moment.

Phil McDonnell writes: 

Remember that extrema are only known in hindsight. Also remember that extrema are governed by the ArcSine distribution which is counter intuitive once you get the normal distribution wired in your head. One of the corollaries of this is that extrema occur early or late in the given time frame but not so much in the middle time period. Also be aware that the ArcSine is a U shaped distribution. This means that ALL of the distribution is in the tails, little in the middle. So you need much larger samples than for a normal distribution.

This is not meant to throw cold water on Craig's interesting line of thought, but more an explanation of why I try to avoid extrema lines of inquiry. Ultimately you would need to test this using a randomized bootstrap simulation to see where it lies. I suspect n=14 is too small though, but might be fine for a normal dist..
 


Comments

Name

Email

Website

Speak your mind

Archives

Resources & Links

Search