Thanks for posting this. It would be great if you wrote this in an op-ed piece somewhere in the MSM. The general public needs to know more about this.

thank you, this: “The problem happens when you want to predict or estimate at a higher-frequency”

is what I was getting at, the introduction of Statistics/Mathematics in the service of the impossible–here, the predicting/forecasting of human behavior.

as a, different, ex. “Retail-demand Forecasting”–Doesn’t. It is, merely, a better Information gatherer coupled with ‘Historical’-trends understanding..

Hardly the stuff of MTBF/MTTF, and other models of the physical world.
as in http://clusty.com/search?input-form=clusty-simple&v%3Asources=webplus&query=MTBF+MTTF

this is, always, an interesting Q:
Fund A returned an average of 10% PA with a standard deviation of 10%
Fund B returned an average of 11% PA with a standard deviation of 10%

they should remind us, again, that Finance is a Quantative, not Qualitative, Act.
no knowledge, whatsoever, can be derived to lead one to an understanding of “How” the results were generated, esp., without the knowledge of the Portfolio constituents..

We should remember “Investing is a ‘People’ Game”

The reality is that 2/3 of the 55+ probably have most of their wealth in real estate which is tanking. And that wealth is probably still not even enough to carry them through their 10-30 years of retirement.

I have a lot of trouble seeing government increasing taxes on those on “fixed incomes” while their assets are tanking. Not good for reelections!

So they’re going to have to find some ingeueous stealth way of transfering money without the 55+ realizing it. That means printing money and tax credits for the young.

by saying wealth is a zero sum game, i’m not strictly saying that. i’m saying that the entire planet can’t out consume it’s wealth and resources, can’t deficit spend like G20 gov’ts.

willid3, we’re back to socialism if we try to distribute the wealth. but we all know that socialism doesn’t work. and looks like capitalism doesn’t work either. not sure what does work. i think when Bruce brought up the older folks, i was thinking “Big Picture” regarding retirement. of course, companies in the past generation realized that pensions will break them, that’s why we have 401(k)s. i think even if we have Nasdaq 10,000 and Dow 30,000, it still will not be enough to sustain the masses in retirement.

MEH, regarding currency lent into existence. the Chinese are willing to do that, it’s call vendor financing. the Chinese are willing to lend to us and debase their on currency to keep their masses employed and export products. the US is willing to export our jobs to them by debasing the dollar. it’s a beautiful world.

Y-o-Y (e.g. Sept 09 vs Sept 08, or 2009 vs 2008) numbers are clean and unbiased. The problem happens when you want to predict or estimate at a higher-frequency (e.g. monthly rather than yearly.) You have three basic choices:

1. Smooth the data with a MA. This looks pretty but is not math.
2. Just look at Sept 07, 08, 09, etc. You get a number, but sigma is so high that no one cares about it.
3. Seasonally adjust, explain your reasoning, and hope you did it right.

Sampling theory (and statistics in general) is a nasty thing. Here’s a typical puzzle:

Fund A returned an average of 10% PA with a standard deviation of 10%
Fund B returned an average of 11% PA with a standard deviation of 10%

How many years of results do you need to see in order to bet that fund B is generating superior results and is not just lucky?