Re: general models

[ECWFM@xxxxxxxxxxxxxxxxxxx]

>Dear Bill Mitchell: You have gotten my position vis-a-vis Barkeley's wrong.
>Of course, nonlinear models can include linear models-- but the difference is
>that Barkeley and the multi-equilibrium people in general are implicitly
>assuming that only nonlinearities can explain less than full employment
>equilibrium. In other words, Barkeley's position must be that a linear system
>would NOT possess a less than full employment equilibrium.

Although the discussion may be at cross purposes about what is a general
characterisation of a process which generates the data we observe, I never
said, nor would I say that a linear model cannot and will not usually produce
less than full employment. Less than full employment of all resources has to be
seen as the general case with full employment of all resources (and hence
diminishing marginal products) as the special case. As an aside, our previous
discussion about DMPs arises here again. My view of a general economic model
attempting to describe the macro world of capitalism has all inputs being
varied in accordance with sales expectations (demand expectations). In this
sense, it is hard to see a place for MPT in determining employment (and
certainly not distribution). In the special case of full employment (of all
resources presumably after adjustments in the case of excess capital relative
to a fully employed labour force), it does not make much sense either, unless
perhaps we seek to expand the LF further by recruiting housepersons, retirees,
or under age school children.

I agree that Keynes' GT attempted to show that even in the case of flexible
prices (with competitive assumptions as a fixity), a persistent under full
employment state can occur if expectations of sales are such that investment
falters. In most ways, he surely did show that although I might read the same
type of message in Kalecki and aspects of it in Marx's debate with Ricardo over
JB Say.

So I am sure I agree with you on this point.

>Keynesians, multiequilibrium people, etc. can NOT explain how a linear system
>posses a stable unemployment equilibirum without some fixity. Mine  is a more
>GENERAL theory  because it can explain the existence of underemeployment
>equilibirum WHETHER THE SYSTEM IS LINEAR OR NONLINEAR. The fact that I can
>demonstrate underemployment equilibrium EVENIF THE SYSTEM IS LINEAR means
>I am providing a more general theory.

Well I appreciate the point but in the Keynes own attack on Pigou et al, a
number of implicit or explicit asssumptions are required to maintain the
persistency of the involuntary unemployment. Competition in
the product market and marginal cost pricing; investment elasticity w.r.t.
interest rates small, money demand elasticity w.r.t. interest rates large,
wealth effects in the consumption function small, propensity to consume of
debtors smaller than propensity to consume of lenders. Otherwise, price
flexibility (wages and prices) could generate demand and supply movements
commensurate with a rise in effective demand.

So it depends I think on which fixities and assumptions one wishes to think a
the more apposite.

>Once the axiom of ergodicity is overthrown, then there is no requirement for
>any long-run trend --this long un trend presumes an immutable, i.e., ergodic,
>set of structural (linear or nonlinear) relationships -- with actual observatio
>ns occuring merely to non-systematic ramdom "shocks". Ergodicity permits one to
>say: the business cycle is "transitory ...fluctuation around a deterministic
>trend". If you are dubious about this description as you suggest in your email
>posting, then you must agree with me that it is the nonergodicty that
>is the essence of the problem of less than full employment and empirical ob-
>servations cannot systematically lie along a deterministic full employment
>trend EVEN IN THE ABSENCE OF RANDOM SCHOCKS.

I offer the following here. First, ergodicity relates to the behaviour of the
mean of a process over time, which is ergodic if in the limit the sample mean
is a consistent estimator of the true mean. Ergodocity can be used to
invoke a trend-stationary interpretation of macro data with the consequence
that in the absence of imperfections (like government policy), full employment
would occur and be observed along a deterministic trend line. This is the
natural rate version of the world which denies the existence of the general
problem of involuntary unemployment (as in GT) and the possibility of general
overproduction (as in Marx and Ricardo). That is, a statistical description of
the properties of the empirical data, can be used to justify a particular
(special case) view of a unique (long-term) equilibrium, occasionally
interrupted by mean-zero (cyclical) fluctuations.

HOwever, to say that non-ergodicity is the essence of the problem of less than
full employment is another thing again. What is non-ergodic here?  First, if
the source of non-stationarity is stochastic rather than determinstic, then the
random shocks themselves are the reason for the non-ergodicity and in this
case we have a multiple equilibrium world without invoking any ad hocery.
Without the stochastic shocks there would not be any non-ergodicity in the
system.
Second, under full employment can persist in an economy which is non ergodic in
levels but ergodic in changes. If the macro economy is best described as a
series of integrated process then it can display stationary growth paths in
employment (on the special case would it be full employment) but not in levels.
Third, the individual variables in an economy can be non stationary but the
relationship between them ergodic (and stationary) if they form a cointegrated
set. The latter also can occur at less than full employment. In this case the
dynamic path of the economy is in part influenced by the equilibrium it is
heading for (any level of employment depending on conditions of effective
demand) and the shpae of the error correction component can change with policy
initiatives - that is the equilibrium is not unique.

So here I am saying that even if the equilibrium the economy is moving to, is
at, or is near, is realted (usually) with involuntary unemployment (the general
case), the equilibrium is but one of many which can be possible depending on
the conditions prevailing over the dynamic path.

It does raise one issue which I am trying to incorporate into my own work. If
one believes in a moving equilibrium world (linear or non-linear - that is not
my interest as I said yesterday - I am content to localise and rely on habit
formation among societies to forecast out a bit for FUN), and that the
hysteresis can be collared by macro policy to permanently reduce unemployment
without accelerating inflation, then what happens once we reach the irreducible
minimum unemployment rate (the special case)? My view is that the world
experiences a regime shift back into a unique equilibrium.

>Moreover, if one
>designed a policy that somehow rid the system of nonlinearities, and reduced
>it to a linear system, there is still the possibility of a stable underemployme
>nt equilibrium.

So I agree on this last point entirely.

Kind regards

bill mitchell
department of economics,
university of newcastle nsw australia
ecwfm@xxxxxxxxxxxxxxxxxxx