Re: Davidson seminar -A final response (fwd)

[bill mitchell <ECWFM@xxxxxxxxxxxxxxxxxxx>]

paul raised this issue again:

>Moreover, Bill  still does not explain why when he tried the
>random number generator experiment I suggested, several times (I
>think 4 out of 20 if I remember() bill found a second difference
>that appeared to have a "structure" (What Yaglom calls an nth
>difference structure function) despite the fact that the original
>series of numbers were obtained by a random number generator.
>Bill if you can get 4 out of 20 with a 2nd difference function,
>how many out of 20 tries would you have found a structural
>equation in n tries where n---> infinity. And if the number of
>"structural" difference equations discovered increases with n,
>then how do you decide whether this an "empirical" fact or not?
>

you seem to have forgotten what i said at the time this little exercise was
performed. it is along the lines of storks and babies. at present i can give
you another example. it is pissing down rain right here and i have to ride for
3 hours this morning - obsessive. the black cockatoos have been about at the
weekend. they only come when it rains. the aboriginal legends depict them as
"bringing" the rain. of-course they do not being merely correlated.

so there is is a lot of correlation in the system. and lots of spurious
relationships could be found from otherwise random events. the skill is to see
that the econometrics we do is intrinsically embedded in the economic or other
theory that we develop or invent. one goes with the other. so while i will
always get a high r2 and other diagnostics on the rain-black cockatoo
regression and i will probably find them CI and whatever fancy tests i might
know about, why would i bother. it is not a mindless obsession with tools. i
save that for my bicycles.

but otherwise i liked the seminar and paul did put in. although i have to say
the use of the word "vulgar" to describe a word in the english language is a
bit over the top. i prefer to distinguish intent and substance from style.

kind regards
bill
         ####    ##        William F. Mitchell
       #######   ####      Head of Economics Department
     #################     University of Newcastle
   ####################    New South Wales, Australia
   ###################*    E-mail: ecwfm@xxxxxxxxxxxxxxxxxxx
   ###################     Phone: +61 49 215065
    #####      ## ###             +61 49 215027
                           Fax:   +61 49 216919
                  ##
WWW Home Page: http://econ-www.newcastle.edu.au/~bill/billyhp.html