Re: Leftists knocking Friedman ?

On Feb 19, 12:02=A0pm, stuff_st...@[EMAIL PROTECTED]
 wrote:
> On Tue, 19 Feb 2008 09:11:57 -0800 (PST), Lysander
>
>
>
>
>
> <lysan...@[EMAIL PROTECTED]
> wrote:
> >On Feb 16, 11:27 am, stuff_st...@[EMAIL PROTECTED]
 wrote:
> >> On Sat, 16 Feb 2008 10:50:02 -0600, "John Galt"
>
> >> <whoisjohng...@[EMAIL PROTECTED]
> wrote:
> >> >> (This stability is one of the reason markets and the economy
perform=

> >> >> better under Democratic administrations that push regulatory
policie=
s
> >> >> that increase stability. )
>
> >> >Perhaps a bit of a reach here. You'd have to wait until the end of
the=
 next
> >> >business cycle to determine if such intervention was in fact
working, =
and
> >> >there's a pretty good chance that event will occur in somebody
else's =
term.
>
> >> We've had this discussion, but the numbers bear it
out.http://www.coba.=
unr.edu/econ/wp/papers/UNRECONWP06008.pdf
>
> >> (He also does it with various lag times.)
>
> >Correlation does not prove causality.
>
> Chuckle. That's all your left with. The numbers are clear the markets
> and economy do much better under Democrats in the HW. It's true before
> WWII, it's true after WWII.
>

Only clear to someone who does not understand stastics or how easy
spurious correlation is to find. You have to prove the links not just
two things occur at the same time. Even sophisticated regressions
often suffer sample selection bias when examined more closely. I would
strongly argue the shortness of the time series is a problem here. See
the last issue of the Southern Economic Journal to see how a short
time series can lead to false conclusion. Furthermore there is serious
sample selection bias. I doubt you know how to control from either
problem.

What I would like to see as evidence, at the very least, is a probit
model. All this shows is if you eyeball the data there appears a
pattern. There appears to be a pattern between who wins the superbowl
and how the stock market does. You seriously do not believe simply
because the stock market does better when the NFC wins that the Giants
winning the superbowl will cause the stock market to be better. Do
you?

A probit or logit regression would be a start to the question. So far
you have a pattern that has no theory or reasoning as to why it may
hold. We are supposed to believe the democrats have a magic touch and
just electing any democrat makes the economy better. Sorry this is no
better than the spurious correlation of the superbowl and the S&P
index. If you care to debate it is their policy then look at the
effect of similar policies. For instance Kennedy and Reagan both had
similar tax cuts. Then control for other business cycle factors that
congress has little control over like oil prices, food prices,
expected inflation, and investment. Then you might have something.

So far your numbers are garbage in and garbage out. They amount to the
Hoodoo belief that the superbowl mystically affects the stock market.


>
> The markets and economy perform better and stronger under Democrats
> than republicans.
>
> > It is easy to hypothesize =A0 . . .
>
> Yeah, that's why we're posting facts and figures and your spewing
> platitudes and hypotheses.- Hide quoted text -
>
No the hypothesis is based solely on the data. You make the elementary
mistake that correllation proves causation something very common in
those who do not understand statistics. You need to prove the links
not that two things happened simultaneously. Until you do that you are
open to spurious correllation. This is nothing more than spurious
correlation until you can find a regression that controls for other
business cycle factors and ideally approaches the question from a
probit or logit model.

Your cherrypicked numbers do not mean squat.They are the very most a
reason to study the possible relation****p in more detail. Not any
"proof" in of themselves. Are they even any statistics to see if the
GDP growth is stastically significant over administrations? I highly
doubt it. Surpisingly if you have few obversations you may find a
negative statistic is not significantly different from a positive
stastic. As I said before this is garbage in garbage out. A starting
point at best.