"there is lies, there is damn lies, and there is statistics"

i think there is a lot of truth in that quote!

Well I believe your prior argument was that you cannot trust the Beureau of Economic Analysis's data because they were lying to the public or something like that. That I disagree with. However, I do agee that there may be a large number of ways to perform statistical analysis to determine trends or forecasts. If there was one right way to do it, there wouldn't be think tanks and investment analysis groups performing their own studies. Take for example the German recession:

quote:

Meanwhile, two of Germany's best-known economic research institutes warned that the eurozone's biggest economy was shrinking. Grim Germany The Hamburg-based HWWA forecasting institute said Germany was already in a recession, "albeit a slight one." Another economic thinktank, the DIW institute, said Germany would grow by just 0.6% this year - well below a forecast of 1.3% growth it set out jointly with a consortium of six research bodies last October. The latest figures will maintain pressure on the European Central Bank (ECB) to make further cuts in interest rates in the hope of stimulating consumer and corporate spending.

Their analysis of the current situation differed from previous analyses by the German government. Does that mean the German government was lying to the public and trying to paint a better picture? Likely not. Economic figures are revised ALL the time as better data is received. But at any rate the snippet is an interesting read. I would like to read the full analysis as 4 short paragraphs doesn't do it much justice

___________________
Retro ...

Grrr I've been googling for the past 15 minutes trying to find the actual merril lynch report ... it's starting to piss me off!

___________________
Retro ...

Well, I thought I could send it to you, but First Call is changing their website and I'm having trouble finding the research reports. All of my login info is at work--I can try to address this tomorrow if your interest level is that high.

I don't think I wanted to say that the government was lying, per se, just that the data isn't always as it seems at face value. Looking at how the numbers are calculated might indicate a tendency to make the data sound more positive or negative than it really is, therefore making it not 100% trustworthy. Anyhoo, let me know if you want the report--I can probably post it (or at least a summary) if you like.

-Shakka

Fuck--couldn't seem to navigate through the new Thompson Analytics website and Bloomberg can sometimes be a bit difficult to navigate given the overwhelming amount of information available. Could take a while to track it down unless I call Merrill directly.

I also find it interesting that most sources don't list the range that result could be in, or if they do they almost never list what level of confidence interval they use.

___________________
If anyone can get me the following records please contact me:

DJ Tiesto - Battleship Grey (Miro remix)
Tilt - Invisible (Tilt's human mix)

quote:

Originally posted by dj adagnitio I also find it interesting that most sources don't list the range that result could be in, or if they do they almost never list what level of confidence interval they use.

Are you talking about economic indicators or statistics in general? Economic indicators are not statistics in the sense that they aren't sampling data and making a broader interpretation of that sample. They are more or less aggreggations of all the data collected. So for instance, when they calculate personal disposable income, they don't randomly sample some neighborhoods to determine how PDI in the entire country changed. They collect everybody's W-2s and calculate all the data nationwide to arrive at that end figure. The figures do changed and are constantly revised, but when that happens the bea usually publishes technical notes describing why the figure was revised such as this:

http://www.bea.gov/bea/newsrel/tech203f.htm

Forecasts on the other hand are simply that, educated guesses as to how the state of everything happening in the economy will affect that figure in the future. Nobody can say with absolute certainty what will happen before it has yet to happen so forecasts should be taken with the understanding that they are only that. Simply look at today's economic figures and notice the discrepancy between actual figures and the consensus predictions:

http://www.economy.com/dismal/ind_landing.asp

On that note, payrolls increased by 57,000!! Woohoo! Plus non-manuf ISM index looks to be in good shape.

Good news for Europe too in service activity.

___________________
Retro ...

I was talking specifically about statistics, which techinically only means things were a sample of a population is used to predict a larger body.

___________________
If anyone can get me the following records please contact me:

DJ Tiesto - Battleship Grey (Miro remix)
Tilt - Invisible (Tilt's human mix)