Re: Policy in Words

[BILL MITCHELL <ECWFM@xxxxxxxxxxxxxxxxxxx>]

Barkley thanked JJG - a surrogate Paul and then dared to
predict the real Paul (are you real Paul?) - a very
dangerous thing to do in such a non-ergodic world:
[note that Paul has actually replied and i plan to
respond to him separately]

>     Actually if I were Paul, and I am guessing
>from what I have read in his PKMT

>(not sent back to the publisher-----yet :-)),

its only a matter of time. i can send you ed elgar's
address in case you've lost it !

>this is what I
>would say.  "In the short-run the economy can be
>ergodic and that is where 'my' [his] models/policy
>analysis are operative. In the long-run, well non-
>ergodicity reigns supreme and what the heck!  This
>is even very Keynesian:  In the short-run we can
>theorize and policyize, but in the long-run we are
>all dead!"  Not bad, huh? (any addenda, Paul?)
>     Actually what I have found annoying about this
>particular contretemps has been that Paul wants to
>have it both ways.  "I can be ergodic when I want
>to be, but if anybody else dares to suggest a model
>that even remotely suggests a nano-possibility of
>being ergodic, by gosh by gum, I'll blast them for
>their un-Keynesianism!!!"  I find this contradictory
>in the extreme, to say the least.

contretemps, d'ailleurs, c'est folie!

(1) policy activism is much more about catching the decay
as it appears. industry policy, incomes policy, trade policy,
fiscal position are not just reactive. And to be logical they
must assert control over the environment they are addressed at.
They must believe they can make these environments
deterministic. otherwise, there is no logic in any policy. we
are in the free market position - it just happens - so let it.

(2) uncertainty does make things difficult to predict with absolute
surety. so what? marx said a lot about that when he destroyed
say's law, and keynes said a bit later on, thinking he was destroying
say's law (but all he needed to do was quote and reference Marx).

but even if the world is intrinsically unknowable - in the future, people
and institutions of people form all sorts of rules and behavioural
"stabilities" to guide their journey into the unknown. to an extent
the future then becomes highly regulated. if people behave as though
probabilities are meaningful then their actions will reflect that belief.
even if the world doesn't know it, the people do and act accordingly.
the order that emerges from these rules of thumb and other stabilisers
is undeniable. it doesn't and cannot ever account for maverick natural
disasters, but that is hardly the point.

while the rules change as social conventions change, the nature of human
society is that this change is slow and can be traced by analysis.
the reasons for the intertia form an interesting research
programme itself - the institutionalists (dunlop, kerr etc focused
on this sort of issue w.r.t. wage determination).

there is an incredible amount of stability in the economic time series
we observe over time.

(3) this is where econometricians like me believe our skills are essential
to policy design and implementation. i exploit the fact that there is
significant inertia in human society. and if we are going to run
complex interventions to control the environment then surely we need a
framework to do sensitivities among options. there is nothin
uncertain about the dollars that we plan to spend of other people's money.

i find it incredible that all these realists carry on about non-ergodicity
and the problems of econometrics (from keynes - who after all
was a dill when it came to understanding what even tinbergen was doing*), yet
still say that the Govt. should intervene.

(* BTW - keynes knew hardly anything about econometrics despite parading himself
as an authority - he even told Tinbergen that it would be better given T has 20
observatioins to split the sample into 4 x 5 obs sub-samples and run some
regressions and test for homogeneity (stability). i just failed him in intro
stats!)

(4) non-ergodicity can arise for various reasons - it is not an homogenous
state. I have harped on this before and never got a satisfactory answer from
those who keep using the term as if it is equivalent to uncertainty. it is not!
and further even if the source of the non-ergodicity is stochastic it still
remains true that a combination of non-ergodic time series may well be
stable. that is they move together in time in a systematic manner.

if that is so then policy can exploit that as can econometric analysis.
my view of the world (maybe it is b/c i am always looking up from downunder)
is that there is a lot of these systematic relations. I go to the
northern hemisphere each year just to check it out from on top BTW.
same picture.

I am just about to finish an article on all this. i think i will send it to
the JPKE :-). it will be ready at the end of the month of may for anyone who
is interested.


kind regards
bill
*******************************************************************************
 William F. Mitchell            Telephone: +61-49-215027      .-_|\
 Department of Economics                   +61-49-705133     /     \    about
 The University of Newcastle    Fax:       +61-49-216919     \.--._/*<-- here
 Callaghan   NSW  2308                                            v
 Australia                      Email : ecwfm@xxxxxxxxxxxxxxxxxxx

 World Wide Web Home Page: http://econ-www.newcastle.edu.au/~bill/billyhp.html
*******************************************************************************