Quiz Question V1
Robert Vienneau
Markku,
I'm back from vacation.
My Leontief example does indeed illustrate my contention that a
higher wage may be associated with an equilibrium in which firms
choose to adopt a more labor-intensive technique.
Hahn and Solow's chapter 2 model assumes
o A well-behaved "one-good" aggregate production function
o Overlapping generations
o An ability of households to lend savings to the industrial sector or
hold savings as money.
o A Clower constraint in which a household planning to spend a certain
amount on consumption in a given period must start that period with
a certain percentage of that planned consumption in the form of
money.
You may say that a recession in output happens in this model because
the assumptiosn of a Clower constraint or of nonnegative interest rates
are market imperfections. But these are very weak forms of market
imperfections - if they can be considered so at all, certainly when
compared with other possibilities.
Problems with instabilities seems to be a fairly general feature of
overlapping generations models, anyways.
--
Robert Vienneau Try my Mac econ simulation game,
rv...@future.dreamscape.com Bukharin, at
ftp://csf.colorado.edu/econ/authors/Vienneau.Robert/Bukharin.sea
Whether strength of body or of mind, or wisdom, or virtue, are always
found...in proportion to the power or wealth of a man [is] a question
fit perhaps to be discussed by slaves in the hearing of their
masters, but highly unbecoming to reasonable and free men in search
of the truth. -- Rousseau
Markku Stenborg
O 'ear ye mortals, the infinite wisdom of rv...@dreamscape.com ...
>
>Markku,
>
>I'm back from vacation.
>
>My Leontief example does indeed illustrate my contention that a
>higher wage may be associated with an equilibrium in which firms
>choose to adopt a more labor-intensive technique.
The use of Leontieff prod fn sounds like cheating in a model where the aim is
to discuss input prices and labor-intensitivity of chosen technology. It, at
min, creates an impression that the results heavily depend on the
particularities of Leontieff.
Why don't you use Cobb-Douglas instead? Or is L the driving force behind your
conclusion? Even a simple numerical version of C-D would have more credibility
than L.
How about explaining the intuition of the conclusion, esp. what drives the
result (of higher wage, more labor-intensive tek). Yes, you've e-mailed me a
"home-work" version of your model, but this is not what I'm asking for. The
logic of your reasing isn't all that transparent in it (I can't view or print
post-script files, something to do w/ our LAN here, I s'pose).
>Hahn and Solow's chapter 2 model assumes
>
> o A well-behaved "one-good" aggregate production function
>
> o Overlapping generations
>
> o An ability of households to lend savings to the industrial sector or
> hold savings as money.
>
> o A Clower constraint in which a household planning to spend a certain
> amount on consumption in a given period must start that period with
> a certain percentage of that planned consumption in the form of
> money.
>
>You may say that a recession in output happens in this model because
>the assumptiosn of a Clower constraint or of nonnegative interest rates
>are market imperfections. But these are very weak forms of market
Mkt imperfections aren't really causes of recessions. They may be the causes
that prolong recession or increase the depth of recessions, I guess.
Recessions occur due to simultaneous contractions on several mkts, say, or due
to coordination problems when agent's play an inferior equilibrium.
>imperfections - if they can be considered so at all, certainly when
>compared with other possibilities.
Not having read the book, can't really evaluate; but see below.
>Problems with instabilities seems to be a fairly general feature of
>overlapping generations models, anyways.
Yep, fairly nice looking OLG models can have quite ugly equilibria.
________
Markku Stenborg
OFC & Turku Business School
<mailto: markku....@finofc.fi>
<mailto: mar...@utu.fi>
Robert Vienneau
In article <5d9hl7$b...@idefix.eunet.fi>, real.a...@bottom.of.msg
(Markku Stenborg) wrote:
> O 'ear ye mortals, the infinite wisdom of rv...@dreamscape.com ...
>
> >My Leontief example does indeed illustrate my contention that a
> >higher wage may be associated with an equilibrium in which firms
> >choose to adopt a more labor-intensive technique.
>
> The use of Leontieff prod fn sounds like cheating in a model where the aim is
> to discuss input prices and labor-intensitivity of chosen technology. It, at
> min, creates an impression that the results heavily depend on the
> particularities of Leontieff.
>
> Why don't you use Cobb-Douglas instead? Or is L the driving force behind your
> conclusion? Even a simple numerical version of C-D would have more
credibility
> than L.
Keep in mind I had more than one Leontief process available in the steel
sector from which firms can choose. Also notice my data did not include
endowments of produced goods (steel and corn).
The phenomenon in the labor market in which I am interested can arise
in models in which there are many available Leontief processes in each
sector. The book of blueprints from which firms choose can be quite thick.
I think it can also arise in "Austrian" capital models with continuous
production functions, but no intermediate goods which are used directly
or indirectly in their own production. I am not sure about this claim,
though. I think the introduction of a nonlinearity with the rate of
interest may be essential for my conclusion.
Notice that the availability of many Leontief processes in each sector
may be able to approximate a continuous production function in some
sense. So I don't assume "fixed coefficients" since they can vary. But
it may not be the case that one can trade off inputs of corn and labor
without also changing the input of steel.
On the other hand, I can construct a model with traditional "substitution"
behavior by embedding the same assumptions about production possibilities
in a model of intertemporal equilibrium. Endowments become data, prices
are forward prices in some sort of futures markets, and it is not
clearly meaningful to talk about *the* rate of interest (although the
numeraire has own rates of interest for various periods).
The question is the structure of various equilibrium models, their
logic, and their appropriateness for looking at capitalist economies.
It is now clear to many (although Hollander tries to lead a
counter-revolution against this insight) that neoclassical economics
is improperly named. It is not a more rigorous development of
classical economics. Classical economics has its own structure and
has been reborn in recent years.
Furthermore, the neoclassicals between 1870 and, say, the 1930s
(when theories of intertemporal and temporary equilibrium were
becoming more well known) were logically inconsistent. It is
not clear that neoclassicals can offer a theory of an economy in
which investors are free to move their financial capital from one
sector to another in search of the highest rate of return.
There's an argument along these lines in
G. Dumenil and D. Levy, "The Classicals and the Neoclassicals: A
Rejoinder to Frank Hahn," _Cambridge Journal of Economics_, 1985,
V9, pp. 327-345.
They consider a (deterministic) model of temporary equilibrium. Suppose
the prices appropriate for intertemporal equilibrium happen to prevail
at some point of time for some unknown reason. Demand and supply are
equated in spot markets, and, assuming a lack of perfect forsight, prices
would not be changed. But since intertemporal equilibrium requires prices
to change, the economy would depart from equilibrium immediately. On the
other hand, if the firms calculate the rate of return using one set of
prices, they would want to change their outputs, but not as in
intertemporal equilibrium.
Dumenil and Levy also show that the predictions of spot prices embedded
in an infinite period intertemporal equilibrium approach classical prices
of production, if they converge at all.
Their concluding paragraph is:
"The way equilibrium is described in the neoclassical paradigm remains
fascinating only because the issue of dynamics is overlooked. The
trouble with the Walrasian perspective is not only that the
*tatonnement* is not realistic, but also and more importantly that the
project of making it more realistic has no real meaning. The Walrasian
equilibrium *by definition* cannot be dissociated from the *tatonnement*,
or the assumption of perfect foresight. Any attempt to incorporate the
search for equilibrium in the neoclassical model ruins the significance
of the preasymptotical equilibrium. Only the limit to infinity can be
preserved; this limit embodies the basic features of the classical
target of the convergence process. Everything which does not correspond
to prices of production in the neoclassical appartus does not have the
least economic significance. Formulating the problem in the most
favorable way for the neoclassicals, it would be possible to contend
their model describes a very particular line of convergence of the
classical dynamics which has no reason to be followed by any real
decentralized economy."
So the models in which neoclassical intutitions about substitution are
locigally consistent are inapplicable to our economy. Models that may be
applicable do not conform to those intutitions.
> How about explaining the intuition of the conclusion, esp. what drives the
> result (of higher wage, more labor-intensive tek). Yes, you've e-mailed me a
> "home-work" version of your model, but this is not what I'm asking for. The
> logic of your reasing isn't all that transparent in it [...]
If the above is not helpful, try
Luigi L. Pasinetti, _Lectures on the Theory of Production_, Columbia
University Press, 1977.
Heinz D. Kurz and Neri Salvadori, _Theory of Production: A Long-Period
Analysis_, Cambridge University Press, 1995.
In my opinion, your education should have exposed you to techniques for
solving my problem and noted the possibility of my phenomenon. After all,
I'm outlining the data and asking you to find the optimal solution. Isn't
this something neoclassical economists are trained to do? And these are
long lasting disputes in the literature. I don't think your education is
atypical, though.
[deleted stuff about Hahn and Solow]
> Mkt imperfections aren't really causes of recessions. They may be the causes
> that prolong recession or increase the depth of recessions, I guess.
>
> Recessions occur due to simultaneous contractions on several mkts, say,
or due
> to coordination problems when agent's play an inferior equilibrium.
OK, so they cannot be removed by removing all market imperfections. That is,
the way to clear recessionary unemployment is not to try to ensure
wages and prices are perfectly flexible.
I think the Hahn and Solow book is essential reading for all macroeconomists.
Since they consider in later chapters how to embed modern ideas about
labor markets (e.g. efficiency wages) in macroeconomic models, a researcher
in models without perfect information and with signalling properties might
find it of interest as well.
Markku Stenborg
O 'ear ye mortals, the infinite wisdom of rv...@dreamscape.com ...
>
>In article <5d9hl7$b...@idefix.eunet.fi>, real.a...@bottom.of.msg
>(Markku Stenborg) wrote:
>
>> O 'ear ye mortals, the infinite wisdom of rv...@dreamscape.com ...
>>
>> >My Leontief example does indeed illustrate my contention that a
>> >higher wage may be associated with an equilibrium in which firms
>> >choose to adopt a more labor-intensive technique.
>>
>> The use of Leontieff prod fn sounds like cheating in a model where the aim
i
>s
>> to discuss input prices and labor-intensitivity of chosen technology. It,
at
>
>> min, creates an impression that the results heavily depend on the
>> particularities of Leontieff.
>>
>> Why don't you use Cobb-Douglas instead? Or is L the driving force behind
you
>r
>> conclusion? Even a simple numerical version of C-D would have more
>credibility
>> than L.
>
>Keep in mind I had more than one Leontief process available in the steel
>sector from which firms can choose. Also notice my data did not include
>endowments of produced goods (steel and corn).
Yes. That's why I did say "sounds like cheating" and "creates an impression"
rather than, umm, "is cheating". But still, how vital is the L fn for the
issue? My impression, w/o going through all the motions, is that it is.
>The phenomenon in the labor market in which I am interested can arise
>in models in which there are many available Leontief processes in each
>sector. The book of blueprints from which firms choose can be quite thick.
[snip]
>Notice that the availability of many Leontief processes in each sector
>may be able to approximate a continuous production function in some
>sense. So I don't assume "fixed coefficients" since they can vary. But
Yup, did notice.
>it may not be the case that one can trade off inputs of corn and labor
>without also changing the input of steel.
See, this is the prob of presenting a model w/o explaining the intuition and
making a distinction b/w the crucial and the boringly unimportant but
simplifying assumptions. The details muddy the main reasoning. Perhaps you've
not explained the crucial driving force to yourself, either?
>On the other hand, I can construct a model with traditional "substitution"
>behavior by embedding the same assumptions about production possibilities
>in a model of intertemporal equilibrium. Endowments become data, prices
>are forward prices in some sort of futures markets, and it is not
>clearly meaningful to talk about *the* rate of interest (although the
>numeraire has own rates of interest for various periods).
So why isn't this the interest?
>The question is the structure of various equilibrium models, their
>logic, and their appropriateness for looking at capitalist economies.
Yes, this is one of the more important questions.
>It is now clear to many (although Hollander tries to lead a
>counter-revolution against this insight) that neoclassical economics
>is improperly named. It is not a more rigorous development of
The name of neoclass trad is less important than the quality of work.
>classical economics. Classical economics has its own structure and
>has been reborn in recent years.
I sure welcome multitude and diverse approaches to looking at economic issues.
Its the usefulness of the insights provided and the quality of empirical
support that should sort the good out of the lousy.
>Furthermore, the neoclassicals between 1870 and, say, the 1930s
>(when theories of intertemporal and temporary equilibrium were
>becoming more well known) were logically inconsistent. It is
This I don't buy. There might have been some logically inconsistent models
published, but I doubt, at least until demonstrated otherwise. Besides, even
if there were, who cares, as long as we got rid off the inconsistencies?
>not clear that neoclassicals can offer a theory of an economy in
>which investors are free to move their financial capital from one
>sector to another in search of the highest rate of return.
Why not?
I recall you've claimed something similar before, but I don't recall the
explanation.
>There's an argument along these lines in
>
> G. Dumenil and D. Levy, "The Classicals and the Neoclassicals: A
> Rejoinder to Frank Hahn," _Cambridge Journal of Economics_, 1985,
> V9, pp. 327-345.
>
>They consider a (deterministic) model of temporary equilibrium. Suppose
>the prices appropriate for intertemporal equilibrium happen to prevail
>at some point of time for some unknown reason. Demand and supply are
>equated in spot markets, and, assuming a lack of perfect forsight, prices
>would not be changed. But since intertemporal equilibrium requires prices
>to change, the economy would depart from equilibrium immediately. On the
>other hand, if the firms calculate the rate of return using one set of
>prices, they would want to change their outputs, but not as in
>intertemporal equilibrium.
This makes no sence. There's deterministic world, but lack of perfect
foresight. Why would intertemporal equilibrium require prices to be changed?
Further, which prices? Spot or future? Is the particular concept of
intertemporal equil reasonable here?
>Dumenil and Levy also show that the predictions of spot prices embedded
>in an infinite period intertemporal equilibrium approach classical prices
>of production, if they converge at all.
>
>Their concluding paragraph is:
>
> "The way equilibrium is described in the neoclassical paradigm remains
> fascinating only because the issue of dynamics is overlooked. The
> trouble with the Walrasian perspective is not only that the
> *tatonnement* is not realistic, but also and more importantly that the
Nobody has ever claimed that tatonnement is realistic. Besides, tatonnement
has next to nothing to do w/ dynamics, it is a metaphore for the stability of
a static equilibrium concept.
> project of making it more realistic has no real meaning. The Walrasian
The project of making it more realistic, if there ever was one, is definitely
not in the heart of neoclass Econ. Tatonnement is all but ignored.
> equilibrium *by definition* cannot be dissociated from the *tatonnement*,
> or the assumption of perfect foresight. Any attempt to incorporate the
Sorta true.
> search for equilibrium in the neoclassical model ruins the significance
> of the preasymptotical equilibrium. Only the limit to infinity can be
> preserved; this limit embodies the basic features of the classical
> target of the convergence process. Everything which does not correspond
> to prices of production in the neoclassical appartus does not have the
> least economic significance. Formulating the problem in the most
> favorable way for the neoclassicals, it would be possible to contend
> their model describes a very particular line of convergence of the
> classical dynamics which has no reason to be followed by any real
> decentralized economy."
>
>So the models in which neoclassical intutitions about substitution are
>locigally consistent are inapplicable to our economy. Models that may be
>applicable do not conform to those intutitions.
I take you mean "intuitions"? I fail to see how you can draw such conclusions.
At best, the guys presented *one* model they claimed to be in line w/ neoclass
models where, in equil, I presume, some atypical behavior occurs. I can come
up w/ models that sorta looks like using neoclass assumptions but has weird
equil. This does not make neoclass Econ useless, equally as if I'd come up w/
a set of assumptions that look a like Shraffian but has unShraffian
conclusions establishes logical inconsistency of Shraffian Econ.
>> How about explaining the intuition of the conclusion, esp. what drives the
>> result (of higher wage, more labor-intensive tek). Yes, you've e-mailed me
a
>
>> "home-work" version of your model, but this is not what I'm asking for. The
>> logic of your reasing isn't all that transparent in it [...]
>
>If the above is not helpful, try
>
> Luigi L. Pasinetti, _Lectures on the Theory of Production_, Columbia
> University Press, 1977.
>
> Heinz D. Kurz and Neri Salvadori, _Theory of Production: A Long-Period
> Analysis_, Cambridge University Press, 1995.
>
>In my opinion, your education should have exposed you to techniques for
>solving my problem and noted the possibility of my phenomenon. After all,
>I'm outlining the data and asking you to find the optimal solution. Isn't
>this something neoclassical economists are trained to do? And these are
>long lasting disputes in the literature. I don't think your education is
>atypical, though.
Yep, I can solve w/o problems. I asked you to provide an intuition for your
story, not to help me solve the equil, but to see what all's going on. You're
much more involved w/ the nitty gritty of the idea than I am, so you've a
comparative advantage in explaining the big picture. A paragraph-lenght
explation would suffice.
If I were to review a paper for a journal, I'd expect the authors to spell out
the intuition of the model and the driving forces behind the results to
evaluate its worth. For instance, a model where the results crucially depend
on the shapes of prod. fns is lousy for analyzing input mkts. A model where
money demand is assumed, is lousy in explaing the existence of money. But
in the opposite cases, the assumptions could be perfectly justifiable.
[snip]
Robert Vienneau
In article <5da566$p...@idefix.eunet.fi>, real.a...@bottom.of.msg
(Markku Stenborg) wrote:
> [...]
> Yes. That's why I did say "sounds like cheating" and "creates an impression"
> rather than, umm, "is cheating". But still, how vital is the L fn for the
> issue? My impression, w/o going through all the motions, is that it is.
It can be the the case that the optimal technique varies continuously
with the wage and that at some values of the wage a higher wage is
associated with a choice of a more labor-intensive technique.
I suspect that this association between a more labor-intensive
technique and a higher wage is also possible in models with continuously
differentiable production functions, but no basic goods.
What sort of assumptions are needed to get well-behaved neoclassical
stories of substitution is a matter of research, and you can examine
the literature yourself.
[...]
>
> See, this is the prob of presenting a model w/o explaining the intuition and
> making a distinction b/w the crucial and the boringly unimportant but
> simplifying assumptions. The details muddy the main reasoning. Perhaps you've
> not explained the crucial driving force to yourself, either?
So what has this discussion been about?
Perhaps you might want to state your assumptions.
Sometimes a comparison of long run positions will find that a more labor
intensive technique will be preferred at a higher wage. Sometimes a
comparison will find that a less labor intensive technique is preferred.
The question for anybody familiar with the literature is less why this
comes about, but rather why so many economists don't know about this
possibility and why they refuse to accept these implications of optimizing
behavior. The answer will revolve around history and the sociology of
knowledge. And you won't like my hypotheses since in one way of telling
the story the events of 1848 are prominent.
> >On the other hand, I can construct a model with traditional "substitution"
> >behavior by embedding the same assumptions about production possibilities
> >in a model of intertemporal equilibrium. Endowments become data, prices
> >are forward prices in some sort of futures markets, and it is not
> >clearly meaningful to talk about *the* rate of interest (although the
> >numeraire has own rates of interest for various periods).
>
> So why isn't this the interest [rate]?
Here you present your lack of knowledge of a literature as an argument.
Which good we select as numeraire is just for the convenience of the
analyst. There is no money in the Arrow-Debreu model and no good is
picked out by the structure of the model for numeraire. The own-rate of
interest of the numeraire will vary with the choice of numeraire.
Since all produced goods have the same equilibrium own rate of interest
in my homework example, the rate does not vary with the choice of
numeraire. Nor does it vary in different time periods.
The classical economists had a story about competitive markets. A sort
of leveling process is assumed to be at work to smooth out differences
in wages for the same type of labor and differences in prices for the
same good. Also, those with investments in sectors making a lower rate
of profit will tend to remove their capital and invest it in sectors
making a higher rate of profit (making due allowances for risk, etc.)
Prices of production or "natural" prices show a system of stationary
relative (spot) prices in which these tendencies do not exist
because these leveling processes have been completed. The claim is
that at any point in time, natural prices are exerting some sort of
gravitational attraction on market prices.
Arrow-Debreu equilibria prices are not prices of production. The
theories simply have a different structure. A more interesting question
is whether intertemporal equilibria can represent that process of
gravitational attraction of market prices to prices of production.
The consensus seems to be against this idea in any meaningful way.
> >The question is the structure of various equilibrium models, their
> >logic, and their appropriateness for looking at capitalist economies.
>
> Yes, this is one of the more important questions.
>
> >It is now clear to many (although Hollander tries to lead a
> >counter-revolution against this insight) that neoclassical economics
> >is improperly named. It is not a more rigorous development of
>
> The name of neoclass trad is less important than the quality of work.
Not the point. The point is the name carries the mistaken implication
that classical economics was just a not-fully-worked-out forerunner,
not a genuine alternative.
> >classical economics. Classical economics has its own structure and
> >has been reborn in recent years.
>
> I sure welcome multitude and diverse approaches to looking at economic
issues.
> Its the usefulness of the insights provided and the quality of empirical
> support that should sort the good out of the lousy.
Are you asking me to demonstrate the empirical applicability of Leontief
input/output matrices?
> >Furthermore, the neoclassicals between 1870 and, say, the 1930s
> >(when theories of intertemporal and temporary equilibrium were
> >becoming more well known) were logically inconsistent. It is
>
> This I don't buy. There might have been some logically inconsistent models
> published, but I doubt, at least until demonstrated otherwise. Besides, even
> if there were, who cares, as long as we got rid off the inconsistencies?
It has been demonstrated, and I gave references last post to textbook
presentations of a consensus of a body of scholarship. If you want to
assert the consensus is wrong, shouldn't you be obligated to present
an argument?
And where's your consistent model of neoclassical long run equilibrium?
Try to include more than one produced good and at least one capital good
that is not consumed.
> >not clear that neoclassicals can offer a theory of an economy in
> >which investors are free to move their financial capital from one
> >sector to another in search of the highest rate of return.
>
> Why not?
>
> I recall you've claimed something similar before, but I don't recall the
> explanation.
See above and below.
> >There's an argument along these lines in
> >
> > G. Dumenil and D. Levy, "The Classicals and the Neoclassicals: A
> > Rejoinder to Frank Hahn," _Cambridge Journal of Economics_, 1985,
> > V9, pp. 327-345.
> >
> >They consider a (deterministic) model of temporary equilibrium. Suppose
> >the prices appropriate for intertemporal equilibrium happen to prevail
> >at some point of time for some unknown reason. Demand and supply are
> >equated in spot markets, and, assuming a lack of perfect forsight, prices
> >would not be changed. But since intertemporal equilibrium requires prices
> >to change, the economy would depart from equilibrium immediately. On the
> >other hand, if the firms calculate the rate of return using one set of
> >prices, they would want to change their outputs, but not as in
> >intertemporal equilibrium.
>
> This makes no sence. There's deterministic world, but lack of perfect
> foresight. Why would intertemporal equilibrium require prices to be changed?
I said "deterministic" just to avoid arguing about stochastic versions
of rational expectations.
> Further, which prices? Spot or future? Is the particular concept of
> intertemporal equil reasonable here?
I assume you can translate from the complete forward prices and period-0
spot prices in intertemporal equilibrium to the series of spot prices
that would prevail in a sequence of temporary equilibria, assuming that
sequence corresponds to an intertemporal equilibrium. You should
see that, in general, relative spot prices will change from week to
week.
If supply and demand are equal at the start of a week, how do
prices change so as to be at the appropriate values at the start
of the next week? If they don't change, the sequence of temporary
equilibria will depart very quickly from the corresponding intertemporal
equilibrium.
"We probably misunderstand Frank Hahn's argument: maybe the issue...
is the construction of Socialism."
-- Dumenil and Levy
> >Dumenil and Levy also show that the predictions of spot prices embedded
> >in an infinite period intertemporal equilibrium approach classical prices
> >of production, if they converge at all.
> >
> >Their concluding paragraph is:
> >
> > "The way equilibrium is described in the neoclassical paradigm remains
> > fascinating only because the issue of dynamics is overlooked. The
> > trouble with the Walrasian perspective is not only that the
> > *tatonnement* is not realistic, but also and more importantly that the
>
> Nobody has ever claimed that tatonnement is realistic. Besides, tatonnement
> has next to nothing to do w/ dynamics, it is a metaphore for the stability of
> a static equilibrium concept.
Huh? You think it coherent to analyze the stability of equilibria without
examining dynamics? Maybe you better discuss this with a physicist.
> > project of making it more realistic has no real meaning. The Walrasian
>
> The project of making it more realistic, if there ever was one, is definitely
> not in the heart of neoclass Econ. Tatonnement is all but ignored.
Whatever. But see the "Hahn process," the "Edgewoth process", and
a certain book by Franklin Fisher.
> > equilibrium *by definition* cannot be dissociated from the *tatonnement*,
> > or the assumption of perfect foresight. Any attempt to incorporate the
>
> Sorta true.
>
> > search for equilibrium in the neoclassical model ruins the significance
> > of the preasymptotical equilibrium. Only the limit to infinity can be
> > preserved; this limit embodies the basic features of the classical
> > target of the convergence process. Everything which does not correspond
> > to prices of production in the neoclassical appartus does not have the
> > least economic significance. Formulating the problem in the most
> > favorable way for the neoclassicals, it would be possible to contend
> > their model describes a very particular line of convergence of the
> > classical dynamics which has no reason to be followed by any real
> > decentralized economy."
> >
> >So the models in which neoclassical intutitions about substitution are
> >locigally consistent are inapplicable to our economy. Models that may be
> >applicable do not conform to those intutitions.
>
> I take you mean "intuitions"? I fail to see how you can draw such
conclusions.
> At best, the guys presented *one* model they claimed to be in line w/
neoclass
> models where, in equil, I presume, some atypical behavior occurs.
No. They examined the structure of various equilibrium models, their logic,
their interrelationships and their appropriateness for looking at
capitalist economies.
> I can come
> up w/ models that sorta looks like using neoclass assumptions but has weird
> equil. This does not make neoclass Econ useless, equally as if I'd come up w/
> a set of assumptions that look a like Shraffian but has unShraffian
> conclusions establishes logical inconsistency of Shraffian Econ.
Speaking of spelling...
No. If you make Sraffian assumptions, you will obtain Sraffian
conclusions (except perhaps in some interesting cases of the choice
of technique in the fully general case of joint production where
Sraffa may not be correct). If you add some special-case assumptions to
Sraffian assumptions you will obtain special cases of Sraffian conclusions,
e.g. the labor theory of value as a theory of relative price. If
your assumptions are merely "like" Sraffian assumptions, I would
expect somebody competent in economics to be able to tell you where
your assumptions vary from Sraffian assumptions and how this variation
drives whatever un-Sraffian conclusions you may have obtained.
Are the assumptions in my homework example neoclassical? If not, where
do they vary?
>[...]
>
> >In my opinion, your education should have exposed you to techniques for
> >solving my problem and noted the possibility of my phenomenon. After all,
> >I'm outlining the data and asking you to find the optimal solution. Isn't
> >this something neoclassical economists are trained to do? And these are
> >long lasting disputes in the literature. I don't think your education is
> >atypical, though.
>
> Yep, I can solve w/o problems. I asked you to provide an intuition for your
> story, not to help me solve the equil, but to see what all's going on. You're
> much more involved w/ the nitty gritty of the idea than I am, so you've a
> comparative advantage in explaining the big picture. A paragraph-lenght
> explation would suffice.
I think I've provided an explanation. So do you now agree that a comparison
of long run equilibria with given production possibilities can find
a more labor-intensive technique preferred at a higher wage? And that
the low wage equilibrium can be transformed to the high wage equilibrium
by varying the parameters, say, in utility functions?
> If I were to review a paper for a journal, I'd expect the authors to
spell out
> the intuition of the model and the driving forces behind the results to
> evaluate its worth. For instance, a model where the results crucially depend
> on the shapes of prod. fns is lousy for analyzing input mkts. [...] But
> in the opposite cases, the assumptions could be perfectly justifiable.
I agree neoclassical economics is lousy for analyzing the distribution of
income.
I think the techniques I have for analyzing the optimal choice of technique
are perfectly general and capable of being applied under any reasonable
set of assumptions.
Markku Stenborg
O 'ear ye mortals, the infinite wisdom by Robert Vienneau of
rv...@dreamscape.com ...
>
>In article <5da566$p...@idefix.eunet.fi>, real.a...@bottom.of.msg
>(Markku Stenborg) wrote:
>
>> [...]
>> Yes. That's why I did say "sounds like cheating" and "creates an impressio
>n"
>> rather than, umm, "is cheating". But still, how vital is the L fn for the
>> issue? My impression, w/o going through all the motions, is that it is.
>
>It can be the the case that the optimal technique varies continuously
>with the wage and that at some values of the wage a higher wage is
>associated with a choice of a more labor-intensive technique.
>
>I suspect that this association between a more labor-intensive
>technique and a higher wage is also possible in models with continuously
>differentiable production functions, but no basic goods.
Only w/ some mechanisms that provide strong-enough effects that more than
balance the substituition effects. In neo- and neo-neoclass Econ, income
effects are the reason the usual intution on prices and quantities do not
necessarily apply in GE.
Leontieff prod fn is the easiest to use and find these other effects that
dominate substitution. There's no substitution at all, so any effect that
works to the other direction than substitution will do. In your example, there
were 2 possible labor-kapital combinations, and some substituition is
available, so you need to have stronger effects. But, still, the shadow of a
doubt remains.
>What sort of assumptions are needed to get well-behaved neoclassical
>stories of substitution is a matter of research, and you can examine
>the literature yourself.
Totally besides the point. And I've examined the litarature.
>> See, this is the prob of presenting a model w/o explaining the intuition a
>nd
>> making a distinction b/w the crucial and the boringly unimportant but
>> simplifying assumptions. The details muddy the main reasoning. Perhaps you
>'ve
>> not explained the crucial driving force to yourself, either?
>
>So what has this discussion been about?
Recently: What is the driving force that leads to higher wage, more labor
intensive production? What is the intuition for this? Which assumptions are
crucial for this, and which ones are just plain simplifications irrelevant for
results?
>Perhaps you might want to state your assumptions.
Assumptions regarding what? I haven't presented a model that has unintuitive
behavior in equil w/o providing explanation for that behavior.
>Sometimes a comparison of long run positions will find that a more labor
>intensive technique will be preferred at a higher wage. Sometimes a
>comparison will find that a less labor intensive technique is preferred.
Now you're describing the equil again. I can read, so I knew this beforehand,
I'm not quite that stupid, but the intuition why this happens is not at all
clear. What is the mechanism that over-weighs substitution effects? This is
something you've never explained.
>The question for anybody familiar with the literature is less why this
>comes about, but rather why so many economists don't know about this
>possibility and why they refuse to accept these implications of optimizing
>behavior. The answer will revolve around history and the sociology of
The question is not that. It is, Why don't you answer the questions presented
to you?
Economists are loathe to accept any type of silliness as long as it comes from
optimization and the intuition and the forces driving the results can be
explained so that no obvious cheap tricks are involved.
>knowledge. And you won't like my hypotheses since in one way of telling
>the story the events of 1848 are prominent.
Which hypotesis? Hypothesai are for statistical testing, so there's not much
liking involved, 'xept re: uses of statistics.
Which story, which events on 1848? If they are beside the point, don't care to
elaborate.
>> >On the other hand, I can construct a model with traditional "substitution
>"
>> >behavior by embedding the same assumptions about production possibilities
>> >in a model of intertemporal equilibrium. Endowments become data, prices
>> >are forward prices in some sort of futures markets, and it is not
>> >clearly meaningful to talk about *the* rate of interest (although the
>> >numeraire has own rates of interest for various periods).
>>
>> So why isn't this the interest [rate]?
>
>Here you present your lack of knowledge of a literature as an argument.
>
>Which good we select as numeraire is just for the convenience of the
>analyst. There is no money in the Arrow-Debreu model and no good is
>picked out by the structure of the model for numeraire. The own-rate of
>interest of the numeraire will vary with the choice of numeraire.
So, models that lack financial mkts are not the ones to be used for discussing
financial things, such as interest rates?
>Since all produced goods have the same equilibrium own rate of interest
>in my homework example, the rate does not vary with the choice of
>numeraire. Nor does it vary in different time periods.
So there's a proxy for interest that is presumably equal to discount rate?
[snip]
>> The name of neoclass trad is less important than the quality of work.
>
>Not the point. The point is the name carries the mistaken implication
>that classical economics was just a not-fully-worked-out forerunner,
>not a genuine alternative.
Not to me, nor to economists in general, I guess. Quality of theoretical work,
intuition, and results, the empirical evidence, and the like, are the
important factors. If classical econ fulfills these criteria not much worse
than neo-neoclass, I have no problems.
>> I sure welcome multitude and diverse approaches to looking at economic
>issues.
>> Its the usefulness of the insights provided and the quality of empirical
>> support that should sort the good out of the lousy.
>
>Are you asking me to demonstrate the empirical applicability of Leontief
>input/output matrices?
No. Howabout evidence for higher wage, more labor intense prod tech. This is
the interesting conlusion different from std. neoclass intuition. And I'd be
surprised if this isn't different from classical intuittion, too.
[snip]
>> This makes no sence. There's deterministic world, but lack of perfect
>> foresight. Why would intertemporal equilibrium require prices to be change
>d?
>
>I said "deterministic" just to avoid arguing about stochastic versions
>of rational expectations.
Isn't this esides the point? Perhaps I didn't make myself clear. If there's no
stochastics whatsoever, why cannot the agents understand the links b/w past,
present and future in the model?
[snip]
>> Nobody has ever claimed that tatonnement is realistic. Besides, tatonnemen
>t
>> has next to nothing to do w/ dynamics, it is a metaphore for the stability
> of
>> a static equilibrium concept.
>
>Huh? You think it coherent to analyze the stability of equilibria without
>examining dynamics? Maybe you better discuss this with a physicist.
Not if the model is totally static, as Arrow-Debreu is. There are no dynamics
involved, nor dis-equil, there are only equilibria. These can be treated as if
they were static: know the endowments of the economy, you know the equil up to
infinte time, though you have no idea how to select one equil path over the
other. The model are totally quiet on things outside equil. There's no dynamic
moving toward equil, only jumps from one to another. Tatonnement is used to
decide which equil are stable. It is a lousy description of dynamics of real
economies, on this we agree.
>> > project of making it more realistic has no real meaning. The Walrasian
>>
>> The project of making it more realistic, if there ever was one, is definit
>ely
>> not in the heart of neoclass Econ. Tatonnement is all but ignored.
>
>Whatever. But see the "Hahn process," the "Edgewoth process", and
>a certain book by Franklin Fisher.
Umm, the same applies to all ad hoc dynamics added on top of what is a
fundamentally static model.
[snip]
>> You're
>> much more involved w/ the nitty gritty of the idea than I am, so you've a
>> comparative advantage in explaining the big picture. A paragraph-lenght
>> explation would suffice.
>
>I think I've provided an explanation. So do you now agree that a comparison
No. Lemme try again. What is the force that more than balances the
substitution effects? If this was a property of a model I designed, I should
be able to tell it in few sentences.
>of long run equilibria with given production possibilities can find
>a more labor-intensive technique preferred at a higher wage? And that
I don't know what you mean by "preferred". If you mean that "it could happen
that, in GE, there's an equil where wage is higher and more labor is hired
than in some other equil", I'd say that this is a conclusion not in conflict
w/ neoclass econ.
>the low wage equilibrium can be transformed to the high wage equilibrium
>by varying the parameters, say, in utility functions?
Again, depending, a qualified "Yup".
Robert Vienneau
Markku Stenborg writes:
> O 'ear ye mortals, the infinite wisdom by Robert Vienneau of
> rv...@dreamscape.com ...
>
> [...]
>
> >I suspect that this association between a more labor-intensive
> >technique and a higher wage is also possible in models with continuously
> >differentiable production functions, but no basic goods.
>
> Only w/ some mechanisms that provide strong-enough effects that more than
> balance the substituition effects. In neo- and neo-neoclass Econ, income
> effects are the reason the usual intution on prices and quantities do not
> necessarily apply in GE.
Your suggestion leaves out the case of inputs that are complements, not
substitutes. For example, suppose labor inputs in non-contiguous time
periods were complements.
[...]
> >What sort of assumptions are needed to get well-behaved neoclassical
> >stories of substitution is a matter of research, and you can examine
> >the literature yourself.
>
> Totally besides the point. And I've examined the litarature.
Your comments indicate, at best, limited support for this claim. Certainly
we understand differently whatever overlap in literature we may have
read.
How would you explain the "nonsubstitution" theorem?
>[...]
>
> >Perhaps you might want to state your assumptions.
>
> Assumptions regarding what? I haven't presented a model that has unintuitive
> behavior in equil w/o providing explanation for that behavior.
Assumptions for a long-run equilibrium model that is guaranteed to act in
accordance with your intuitions. Or don't you care whether you can show
your prejudices are logically consistent?
> >Sometimes a comparison of long run positions will find that a more labor
> >intensive technique will be preferred at a higher wage. Sometimes a
> >comparison will find that a less labor intensive technique is preferred.
>
> Now you're describing the equil again. I can read, so I knew this beforehand,
> I'm not quite that stupid, but the intuition why this happens is not at all
> clear. What is the mechanism that over-weighs substitution effects? This is
> something you've never explained.
There is a general model, which I did not invent, of the optimal choice of
technique in long run positions. I presented a specific numerical example,
which I *did* invent, that has behavior that you consider conter-intuitive.
Apparently, neoclassical theory, as you understand it, is only
consistent with a special case of this general model. It is up to
neoclassical economists who share this understanding to specify what
assumptions they want to add to the general model, to specify how they
can distinguish between non-neoclassical models and the conjunction of
this special case and its completion with other neoclassical assumptions
(e.g. intertemporal utility maximization), and to produce some
empirical support for their models. Strangely enough, this is a
challenge that they refuse.
On the other hand, I have provided references in the past that show
slight empirical superiority for non-neoclassical theories of distribution
that are consistent with the general case of production theory.
> >The question for anybody familiar with the literature is less why this
> >comes about, but rather why so many economists don't know about this
> >possibility and why they refuse to accept these implications of optimizing
> >behavior. The answer will revolve around history and the sociology of
>
> The question is not that. It is, Why don't you answer the questions presented
> to you?
Why hand out problem sets? One reason is to give the student some examples
in which to apply the techniques he is being taught. But other reasons
might exist for the advanced student. Examples in which one clause of
a theorem and its conclusions fail may provide the student with an
understanding of why that clause is there. (But what theorem does
Markku want to invoke?) Another reason is to train the student's
intuition to what is possible behavior, given the assumptions of the
theory. A student trained in this way will be more convinced than one
who just receives assertions.
Nevertheless...
In a comparison of long-run equilibrium positions in a pure circulating
capital model, a lower wage will be associated with a higher rate of profit.
The techniques in such models can be reduced to a dated series of labor inputs.
The effects of compounding interest charges on these dated labor inputs
are nonlinear. Thus the possibility for reswitching, capital reversing,
and "perverse" behavior in the labor market arises.
There's another important point in my example. A lower price of a
produced input (steel) can be associated with an equilibrium in which
a technique is chosen that uses that input *less* intensely. Although
my example does not demonstrate this, this phenomenon can arise at a switch
point that does not behave "perversely" in either the capital or the
labor markets. This phenomenon does not seem to have anything to do with
continuously differentiable microeconomic production functions; that is, it
may be consistent with their presence or their absence.
Substitution just does not characterize the logic of comparing
long run positions.
> Economists are loathe to accept any type of silliness as long as it
comes from
> optimization and the intuition and the forces driving the results can be
> explained so that no obvious cheap tricks are involved.
So now my example has obvious cheap tricks? Next you will be telling
me that neoclassical economists reject Linear Programming.
> >knowledge. And you won't like my hypotheses since in one way of telling
> >the story the events of 1848 are prominent.
>
> Which hypotesis? Hypothesai are for statistical testing, so there's not much
> liking involved, 'xept re: uses of statistics.
No. "Hypothesis" is an English word that can be used in other contexts than
formal statistical tests.
> Which story, which events on 1848? If they are beside the point, don't
care to
> elaborate.
Your intuition seems to be based on
(1) Unformalized person-on-the street views
(2) The inappropriate extension of models where they work to other
frameworks
(3) The mathematical errors of your predecessors
I claim that perhaps these errors became widespread and are still held on
to today because of political influences.
> >> >On the other hand, I can construct a model with traditional "substitution
> >"
> >> >behavior by embedding the same assumptions about production possibilities
> >> >in a model of intertemporal equilibrium. Endowments become data, prices
> >> >are forward prices in some sort of futures markets...
> [...]
>
> >> The name of neoclass trad is less important than the quality of work.
> >
> >Not the point. The point is the name carries the mistaken implication
> >that classical economics was just a not-fully-worked-out forerunner,
> >not a genuine alternative.
>
> Not to me, nor to economists in general, I guess. Quality of theoretical
work,
> intuition, and results, the empirical evidence, and the like, are the
> important factors. If classical econ fulfills these criteria not much worse
> than neo-neoclass, I have no problems.
>
> >> I sure welcome multitude and diverse approaches to looking at economic
> >issues.
> >> Its the usefulness of the insights provided and the quality of empirical
> >> support that should sort the good out of the lousy.
> >
> >Are you asking me to demonstrate the empirical applicability of Leontief
> >input/output matrices?
>
> No. Howabout evidence for higher wage, more labor intense prod tech. This is
> the interesting conlusion different from std. neoclass intuition. And I'd be
> surprised if this isn't different from classical intuittion, too.
"...the issue is whether [the classical economists] envisaged the
relationships of a systematic and general nature between those variables
which are the necessary logical foundation of the decreasing demand
schedule for labour in neoclassical theory. In other words the question
is whether we can say that the classical economists thought that a rise
in the wage rate would always set in motion a change in production
methods and/or in consumption patterns such as to lead (other things
being equal) to a reduction in the employment level (and vice versa).
Concerning direct substitution between labour and other inputs, it is
fairly widely recognized that such a mechanism was not envisaged by the
classical economists, for techniques tended to be regarded as given
in any normal position of the economy, and as gradually changing with
the accumulation process because of innovations and returns to scale.
...However, despite the recent revivals of interpretations following
the original example of Marshall, which tend to emphasize the similarity
of the classical and neoclassical theories..., many scholars have denied
the existence of these notions in the body of classical economic
theory. Indeed, even those who emphasize similarity often admit that
substitution mechanisms were not really formulated by classical
economists, but maintain that their theory would benefit from being
integrated and 'rationalized' by introducing such mechanisms."
-- Antonella Stirati, _The Theory of Wages in Classical Economics_,
1994
>
> [snip]
>
> >> This makes no sence. There's deterministic world, but lack of perfect
> >> foresight. Why would intertemporal equilibrium require prices to be change
> >d?
> >
> >I said "deterministic" just to avoid arguing about stochastic versions
> >of rational expectations.
>
> Isn't this esides the point? Perhaps I didn't make myself clear. If
there's no
> stochastics whatsoever, why cannot the agents understand the links b/w past,
> present and future in the model?
I notice you don't tell a story of how spot prices change appropriately
from week to week in a sequence of temporary equilibrium.
Markku Stenborg
>Markku Stenborg writes:
>
>> O 'ear ye mortals, the infinite wisdom by Robert Vienneau of
>> rv...@dreamscape.com ...
>>
>> [...]
>>
>> >I suspect that this association between a more labor-intensive
>> >technique and a higher wage is also possible in models with continuously
>> >differentiable production functions, but no basic goods.
>>
>> Only w/ some mechanisms that provide strong-enough effects that more than
>> balance the substituition effects. In neo- and neo-neoclass Econ, income
>> effects are the reason the usual intution on prices and quantities do not
>> necessarily apply in GE.
>
>Your suggestion leaves out the case of inputs that are complements, not
>substitutes. For example, suppose labor inputs in non-contiguous time
>periods were complements.
Yes, complements tend to improve each other's effects. The point is? Anyways,
this has nothing to do w/ your model, this isn't the driving force for your
result, right?
[snip]
>> Assumptions regarding what? I haven't presented a model that has
unintuitive
>> behavior in equil w/o providing explanation for that behavior.
>
>Assumptions for a long-run equilibrium model that is guaranteed to act in
>accordance with your intuitions. Or don't you care whether you can show
>your prejudices are logically consistent?
I have not presented a model of some long-run equilibrium.
One of my intuitions regarding GE is that income effects may more than balance
the substitution effects, eg, so that higher wage can lead to more people
employed. These are the assumptions you're interested? I don't think so.
Which of my prejudices are you trying to move this discussion to?
You came up w/ claim "higher wage, more labor intensive production". I pointed
out that this is against the usual econ intuition of substitution. I asked you
to explain the intuition that provides effects that more than balance the
substitution effects in that model. I've provided you plenty of opportunities
to explain the workings of your model. I won't be holding my breath, but I
thought you'd at least have an idea how your your model operates.
>> >Sometimes a comparison of long run positions will find that a more labor
>> >intensive technique will be preferred at a higher wage. Sometimes a
>> >comparison will find that a less labor intensive technique is preferred.
>>
>> Now you're describing the equil again. I can read, so I knew this
beforehand,
>> I'm not quite that stupid, but the intuition why this happens is not at all
>> clear. What is the mechanism that over-weighs substitution effects? This is
>> something you've never explained.
>
>There is a general model, which I did not invent, of the optimal choice of
>technique in long run positions. I presented a specific numerical example,
>which I *did* invent, that has behavior that you consider conter-intuitive.
>Apparently, neoclassical theory, as you understand it, is only
>consistent with a special case of this general model. It is up to
No. And we're not even discussing this.
Actually, I prefer counter-intuitive results to the repetition of old textbook
stories. I have no problems w/ models that have counter-intuitive results. I'm
merely asking what are the forces in your model, or in that "general long-run
positions" -- what ever those are -- model. Is that too much? Apparently.
>neoclassical economists who share this understanding to specify what
The "this understanding" being? That substitution follows from optimization?
>assumptions they want to add to the general model, to specify how they
>can distinguish between non-neoclassical models and the conjunction of
>this special case and its completion with other neoclassical assumptions
>(e.g. intertemporal utility maximization), and to produce some
>empirical support for their models. Strangely enough, this is a
>challenge that they refuse.
No. And completely besides any point.
>On the other hand, I have provided references in the past that show
>slight empirical superiority for non-neoclassical theories of distribution
>that are consistent with the general case of production theory.
As I recall, the empirical references you've offered turned out to be
worthless. Or have I missed something?
>> >The question for anybody familiar with the literature is less why this
>> >comes about, but rather why so many economists don't know about this
>> >possibility and why they refuse to accept these implications of optimizing
>> >behavior. The answer will revolve around history and the sociology of
>>
>> The question is not that. It is, Why don't you answer the questions
presented
>> to you?
[snipped some silly excuses for not answering a simple question]
>Nevertheless...
>
>In a comparison of long-run equilibrium positions in a pure circulating
>capital model, a lower wage will be associated with a higher rate of profit.
>The techniques in such models can be reduced to a dated series of labor
inputs.
>The effects of compounding interest charges on these dated labor inputs
>are nonlinear. Thus the possibility for reswitching, capital reversing,
>and "perverse" behavior in the labor market arises.
Now this is bit more like what I've been asking for, although you're still
basically describing the result and how to get to it. I still fail to see the
logic of the model or the intuition behind it, perhaps you've not just used to
communicating w/ economists.
This is like describing the ituition and driving forces of neoclass GE models
as: "Mkts clear, prices equal marginal cost, and profits tend to zero. There
are many equil. We can solve equil by fixed point theorems. We have no idea
which of these will be selected." If the listener of this stoty is not an
economist, the above is next to meaningless.
[snip]
>> Economists are loathe to accept any type of silliness as long as it
>comes from
>> optimization and the intuition and the forces driving the results can be
>> explained so that no obvious cheap tricks are involved.
>
>So now my example has obvious cheap tricks? Next you will be telling
No. "Cheap triks" refer to assumptions not spelled out that actually drive the
results.
>me that neoclassical economists reject Linear Programming.
>> >knowledge. And you won't like my hypotheses since in one way of telling
>> >the story the events of 1848 are prominent.
>>
>> Which hypotesis? Hypothesai are for statistical testing, so there's not
much
>> liking involved, 'xept re: uses of statistics.
>
>No. "Hypothesis" is an English word that can be used in other contexts than
>formal statistical tests.
Then, which hypothesis of yours I don't like? Why is my alledged non-liking
explained by events of 1848? Which events?
>> Which story, which events on 1848? If they are beside the point, don't
>care to
>> elaborate.
>
>Your intuition seems to be based on
>
> (1) Unformalized person-on-the street views
My intuition is based on articles on, in no particular order, and not
restricted to, JET, AER, JPE, Econometrica, JEMS, GEB, Rand, QJE, ...
> (2) The inappropriate extension of models where they work to other
> frameworks
Eh?
> (3) The mathematical errors of your predecessors
Eh?
>I claim that perhaps these errors became widespread and are still held on
>to today because of political influences.
Yeah, right.
You have no idea re: my political views?
[snip]
>> >> This makes no sence. There's deterministic world, but lack of perfect
>> >> foresight. Why would intertemporal equilibrium require prices to be
change
>> >d?
>> >
>> >I said "deterministic" just to avoid arguing about stochastic versions
>> >of rational expectations.
>>
>> Isn't this esides the point? Perhaps I didn't make myself clear. If
>there's no
>> stochastics whatsoever, why cannot the agents understand the links b/w
past,
>> present and future in the model?
Lemme restate:
Agents live in non-stochastic and dynamic world, and interact in some manner,
perhaps on mkts. Is is sensible to restrict the agents from understanding any
of the links b/w past, present and future?
>I notice you don't tell a story of how spot prices change appropriately
>from week to week in a sequence of temporary equilibrium.
Why would I?
Robert Vienneau
> O 'ear ye mortals, the infinite wisdom by Robert Vienneau of
> rv...@dreamscape.com ...
>
> >Markku Stenborg writes:
> >
> >> O 'ear ye mortals, the infinite wisdom by Robert Vienneau of
> >> rv...@dreamscape.com ...
> >>
> >> [...]
> >>
> >> >I suspect that this association between a more labor-intensive
> >> >technique and a higher wage is also possible in models with continuously
> >> >differentiable production functions, but no basic goods.
> >>
> >> Only w/ some mechanisms that provide strong-enough effects that more than
> >> balance the substituition effects. In neo- and neo-neoclass Econ, income
> >> effects are the reason the usual intution on prices and quantities do not
> >> necessarily apply in GE.
> >
> >Your suggestion leaves out the case of inputs that are complements, not
> >substitutes. For example, suppose labor inputs in non-contiguous time
> >periods were complements.
>
> this has nothing to do w/ your model, this isn't the driving force for your
> result, right?
Do you first read to the end of a post before responding?
Since we started, I've looked up a reference by Tatsuo Hatta that shows how
to get switches with "perverse" capital-deepening with continuously
differentiable neoclassical production functions in an Austrian context.
These switches would also show "perverse" behavior in the labor market. So
that claim is established in the literature in some articles that
frankly I've never fully understood.
They do suggest, though, that I may have been too hasty in dismissing
your claim that my results depend on not assuming "gross substitution,"
but I'm not sure. I suspect you were relying on some theorems about the
uniqueness and stability of equilibria in the Arrow-Debreu model. This
does not apply to my results since
o My example is not an example of the Arrow-Debreu model
o A "perverse" switch can be an unique equilibrium
o As I think Edwin Burmeister would agree, the stability of these
perverse switches is not clear. That is, it's not clear they are
unstable.
A switch with "non-perverse" behavior in the "capital" and labor markets
can still behave "perversely" with regard to the price of a produced
good. That is, the "perverse" behavior of the price of steel in my
example can be independent of the "perversity" of the capital and
labor markets. I don't know that this behavior is ruled out by gross
substitution.
Most seriously, models with fixed capital cannot be reduced so neatly
to dated inputs of labor, but can still exhibit all these "perversities"
and more. Does anybody even know how to apply Hatta's criteria of
complementarity in these cases?
Don't you have an unanswered question about your understanding of
the "non-substitution" theorem?
> [snip]
>
> >> Assumptions regarding what? I haven't presented a model that has
> unintuitive
> >> behavior in equil w/o providing explanation for that behavior.
> >
> >Assumptions for a long-run equilibrium model that is guaranteed to act in
> >accordance with your intuitions. Or don't you care whether you can show
> >your prejudices are logically consistent?
>
> I have not presented a model of some long-run equilibrium.
Correct. You have not given a hint of how to formalize your intuitions
in a model of long run equilibrium.
> One of my intuitions regarding GE is that income effects may more than
balance
> the substitution effects, eg, so that higher wage can lead to more people
> employed. These are the assumptions you're interested? I don't think so.
You're right that I'm not interested in income effects and that my example
is not an example of Arrow-Debreu GE or of momentary equilibrium.
> Which of my prejudices are you trying to move this discussion to?
I think "intuitions" that you give no argument for and no empirical
evidence for in the context under discussion can be fairly characterized
as "prejudices."
> You came up w/ claim "higher wage, more labor intensive production". I
pointed
> out that this is against the usual econ intuition of substitution. I
asked you
> to explain the intuition that provides effects that more than balance the
> substitution effects in that model. I've provided you plenty of opportunities
> to explain the workings of your model. I won't be holding my breath, but I
> thought you'd at least have an idea how your your model operates.
Assume I'm an ignoramus about how this example I magically conjured out of the
air works. If you are really curious about it, why don't you look up
one of my textbook references, say, Kurz and Salvadori? Then you could make
more intelligent comments than "Eh?"
> >> >Sometimes a comparison of long run positions will find that a more labor
> >> >intensive technique will be preferred at a higher wage. Sometimes a
> >> >comparison will find that a less labor intensive technique is preferred.
> >>
> >> Now you're describing the equil again. I can read, so I knew this
> beforehand,
> >> I'm not quite that stupid, but the intuition why this happens is not
at all
> >> clear. What is the mechanism that over-weighs substitution effects?
This is
> >> something you've never explained.
> >
> >There is a general model, which I did not invent, of the optimal choice of
> >technique in long run positions. I presented a specific numerical example,
> >which I *did* invent, that has behavior that you consider conter-intuitive.
> >Apparently, neoclassical theory, as you understand it, is only
> >consistent with a special case of this general model. It is up to
>
> No. And we're not even discussing this.
You don't seem to be discussing much of anything.
On the other hand, I continue to wonder what are neccessary special-case
assumptions that will yield long-run models that are guaranteed to behave
in accordance with neoclassical beliefs on substitution. Or if that's too
much to ask, what sufficient special-case assumptions are neoclassical
economists inclined to make? Furthermore, I think that the one arguing
for the special case should present some empirical evidence. And I am quite
aware that some neoclassical economists are quite happy to develop
models in the general case. I wonder if they can be said to have a
neoclassical long run theory of value, though.
> Actually, I prefer counter-intuitive results to the repetition of old
textbook
> stories. I have no problems w/ models that have counter-intuitive
results. I'm
> merely asking what are the forces in your model, or in that "general long-run
> positions" -- what ever those are -- model. Is that too much? Apparently.
It would help if you showed more willingness to read.
> >neoclassical economists who share this understanding to specify what
>
> The "this understanding" being? That substitution follows from optimization?
>
> >assumptions they want to add to the general model, to specify how they
> >can distinguish between non-neoclassical models and the conjunction of
> >this special case and its completion with other neoclassical assumptions
> >(e.g. intertemporal utility maximization), and to produce some
> >empirical support for their models. Strangely enough, this is a
> >challenge that they refuse.
>
> No. And completely besides any point.
So where is your formal model of long run equilibrium in which
your intuitive substitution behavior follows from your assumptions?
And where's the empirical evidence for it? Oh, only one of us has to
provide an argument - I see.
> >On the other hand, I have provided references in the past that show
> >slight empirical superiority for non-neoclassical theories of distribution
> >that are consistent with the general case of production theory.
>
> As I recall, the empirical references you've offered turned out to be
> worthless. Or have I missed something?
Not my recollection at all. I don't recall that you or anybody else
has offered any comments at all on empirical results in Marglin (1984) or
Ochoa & Glick (1992). I am cautious above in claiming anything definite about
empirical results on these difficult questions and have been so in the past.
> >> >The question for anybody familiar with the literature is less why this
> >> >comes about, but rather why so many economists don't know about this
> >> >possibility and why they refuse to accept these implications of optimizing
> >> >behavior. The answer will revolve around history and the sociology of
> >>
> >> The question is not that. It is, Why don't you answer the questions
> presented
> >> to you?
I do, and repetitively. You don't seem to answer questions put your way
though.
> [snipped some silly excuses for not answering a simple question]
>
> >Nevertheless...
> >
> >In a comparison of long-run equilibrium positions in a pure circulating
> >capital model, a lower wage will be associated with a higher rate of profit.
> >The techniques in such models can be reduced to a dated series of labor
> inputs.
> >The effects of compounding interest charges on these dated labor inputs
> >are nonlinear. Thus the possibility for reswitching, capital reversing,
> >and "perverse" behavior in the labor market arises.
>
> Now this is bit more like what I've been asking for, although you're still
> basically describing the result and how to get to it. I still fail to see the
> logic of the model or the intuition behind it, perhaps you've not just
used to
> communicating w/ economists.
Which part do you have a problem with?
Some economists (many are not Sraffians) would find this old hat. Isn't it
kind of arrogant to assume your gaps in understanding and your intuitions
are representative of the entire profession?
> This is like describing the ituition and driving forces of neoclass GE models
> as: "Mkts clear, prices equal marginal cost, and profits tend to zero. There
> are many equil. We can solve equil by fixed point theorems. We have no idea
> which of these will be selected." If the listener of this stoty is not an
> economist, the above is next to meaningless.
No. Profits *tending* to zero is a dynamic story that you previously agreed
does not apply in the Arrow-Debreu model. Anyway, my example has markets
that clear, prices equalling marginal cost in the sense that applies to
discrete models, and no pure economic profits. But having studied
economics somewhat, I understand your summary.
> [snip]
> >Your intuition seems to be based on
> >
> > (1) Unformalized person-on-the street views
>
> My intuition is based on articles on, in no particular order, and not
> restricted to, JET, AER, JPE, Econometrica, JEMS, GEB, Rand, QJE, ...
And my intuition is based on articles in the Review of Economic
Studies, the Quarterly Journal of Economics, the Journal of Political
Economy, the Journal of Economic Literature, etc. You have a point?
[...]
> >I claim that perhaps these errors became widespread and are still held on
> >to today because of political influences.
>
> Yeah, right.
Nice to see that your attitude towards hypotheses about political
influences on the development of economics cannot be characterized
by emotional terms such as "liking" or "disliking."
>
> You have no idea re: my political views?
Perhaps. But I don't care for the purpose of this thread.
> [snip]
>
> >> >> This makes no sence. There's deterministic world, but lack of perfect
> >> >> foresight. Why would intertemporal equilibrium require prices to be
> change
> >> >d?
> >> >
> >> >I said "deterministic" just to avoid arguing about stochastic versions
> >> >of rational expectations.
> >>
> >> Isn't this esides the point? Perhaps I didn't make myself clear. If
> >there's no
> >> stochastics whatsoever, why cannot the agents understand the links b/w
> past,
> >> present and future in the model?
>
> Lemme restate:
>
> Agents live in non-stochastic and dynamic world, and interact in some manner,
> perhaps on mkts. Is is sensible to restrict the agents from understanding any
> of the links b/w past, present and future?
Who said anything about restricting the agents to not understanding *any*
of the links between past, present, and future? How can the agents tell
from observing that present spot markets clear that their plans are
sufficiently coordinated that spot markets will clear next week if
they follow through with their plans? Are you sure that you don't have
a centrally planned economy in mind?
> >I notice you don't tell a story of how spot prices change appropriately
> >from week to week in a sequence of temporary equilibrium.
>
> Why would I?
You did not accept my restatement of an argument about why a sequence of
temporary equilibrium would not follow a path analogous to an intertemporal
equilibrium. I'm glad to see you've dropped your objections.
Less snidely, I'm really not sure we have a substantive disagreement
over intertemporal or temporary GE models.
Markku Stenborg
>> O 'ear ye mortals, the infinite wisdom by Robert Vienneau of
>> rv...@dreamscape.com ...
>>
>> >Markku Stenborg writes:
>> >
>> >> O 'ear ye mortals, the infinite wisdom by Robert Vienneau of
>> >> rv...@dreamscape.com ...
[snip]
>> >Your suggestion leaves out the case of inputs that are complements, not
>> >substitutes. For example, suppose labor inputs in non-contiguous time
>> >periods were complements.
>>
>> Yes, complements tend to improve each other's effects. The point is?
Anyways,
>> this has nothing to do w/ your model, this isn't the driving force for your
>> result, right?
>
>Do you first read to the end of a post before responding?
Not always, but I did read your previous post first and only then responded.
The intuition of and/or the driving forces behind the results is the fact
that, in your version of a temporal model, labor inputs in non-contiguous time
periods are complements?
Now, if this were a general property, ie, applying to all inputs at all
levels of uses, we've found the "free lunch", we could indefinitely increase
the wealth by increasing input prices. This cannot be the case in you model.
>Since we started, I've looked up a reference by Tatsuo Hatta that shows how
>to get switches with "perverse" capital-deepening with continuously
>differentiable neoclassical production functions in an Austrian context.
>These switches would also show "perverse" behavior in the labor market. So
>that claim is established in the literature in some articles that
>frankly I've never fully understood.
Yep, that Austrian stuff I never understood. It seems to make sense for some
guys.
>They do suggest, though, that I may have been too hasty in dismissing
>your claim that my results depend on not assuming "gross substitution,"
>but I'm not sure. I suspect you were relying on some theorems about the
>uniqueness and stability of equilibria in the Arrow-Debreu model. This
Not really.
>does not apply to my results since
>
> o My example is not an example of the Arrow-Debreu model
This has been clear all along. You might have wanted to point out where the
differences lie, though.
[snip]
>> I have not presented a model of some long-run equilibrium.
>
>Correct. You have not given a hint of how to formalize your intuitions
>in a model of long run equilibrium.
Yes. If I had presentd some results regarding long run behavior, say, I sure
would explain what drives the results and what is the intuition behind. I
would be carefull to spell out all assumptions that are or could be seen to be
crucial for the results, and clearly mark them from the ones that are mere
simplifications w/o effects to qualitative results. Especially, if the results
would go agaist the grain, I'd be carefull to convice the reader that this is
not due to some unrealistic behavioral assumptions or to strange restrictions
posed on the agents or their interactions. I would discuss the info and the
decisions rules of the agents, the transitions mechanisms from period to
period, etc.
>> One of my intuitions regarding GE is that income effects may more than
>balance
>> the substitution effects, eg, so that higher wage can lead to more people
>> employed. These are the assumptions you're interested? I don't think so.
>
>You're right that I'm not interested in income effects and that my example
>is not an example of Arrow-Debreu GE or of momentary equilibrium.
Arrow-Debreu could also be interpreted as a intertemporal model. Each good on
different time period is a different commodity, etc. But this is besides the
point.
>> Which of my prejudices are you trying to move this discussion to?
>
>I think "intuitions" that you give no argument for and no empirical
>evidence for in the context under discussion can be fairly characterized
>as "prejudices."
No. Substitution is a result that seems to pop out regradless of the details
of the model, and it makes intuitively a lot of sense. If something becomes
more expensive, you want to reduce its use and substitute it w/ something
else. Usually, this isn't the only effect, and as I've stated zillion times,
these other effects could more than balance substitution. Perhaps you think in
manner where the above sounds unreasonable, and do not undestand the question
of expalantions for these other effects.
>> You came up w/ claim "higher wage, more labor intensive production". I
>pointed
>> out that this is against the usual econ intuition of substitution. I
>asked you
>> to explain the intuition that provides effects that more than balance the
>> substitution effects in that model. I've provided you plenty of
opportunities
>> to explain the workings of your model. I won't be holding my breath, but I
>> thought you'd at least have an idea how your your model operates.
>
>Assume I'm an ignoramus about how this example I magically conjured out of
the
>air works. If you are really curious about it, why don't you look up
>one of my textbook references, say, Kurz and Salvadori? Then you could make
>more intelligent comments than "Eh?"
You have a comparative advantage in explaing the model you've spent some time
in thinking about.
[snip]
>> >There is a general model, which I did not invent, of the optimal choice of
>> >technique in long run positions. I presented a specific numerical example,
>> >which I *did* invent, that has behavior that you consider
conter-intuitive.
>> >Apparently, neoclassical theory, as you understand it, is only
>> >consistent with a special case of this general model. It is up to
>>
>> No. And we're not even discussing this.
>
>You don't seem to be discussing much of anything.
I've been asking you to explain the model you've set forth, esp. the driving
forces behind "higher wage, more labor intensive production".
[snip]
>> Actually, I prefer counter-intuitive results to the repetition of old
>textbook
>> stories. I have no problems w/ models that have counter-intuitive
>results. I'm
>> merely asking what are the forces in your model, or in that "general
long-run
>> positions" -- what ever those are -- model. Is that too much? Apparently.
>
>It would help if you showed more willingness to read.
Another "Eh?" is called for.
Should we have a vote on this? Which one of youse guys understand the
intuition behind and/or the forces that drive "higher wage, more labor
intensive production" result? Perhaps the "Yeay" voters could enlighten the
rest of us.
>> >neoclassical economists who share this understanding to specify what
>>
>> The "this understanding" being? That substitution follows from
optimization?
>>
>> >assumptions they want to add to the general model, to specify how they
>> >can distinguish between non-neoclassical models and the conjunction of
>> >this special case and its completion with other neoclassical assumptions
>> >(e.g. intertemporal utility maximization), and to produce some
>> >empirical support for their models. Strangely enough, this is a
>> >challenge that they refuse.
>>
>> No. And completely besides any point.
>
>So where is your formal model of long run equilibrium in which
>your intuitive substitution behavior follows from your assumptions?
>And where's the empirical evidence for it? Oh, only one of us has to
>provide an argument - I see.
Should I present a model, I'll promise I'll ecplain exactly what I've been
asking from you.
Besides, forces that drive substitution shouldn't call that much explanation?
[snip]
>> >In a comparison of long-run equilibrium positions in a pure circulating
>> >capital model, a lower wage will be associated with a higher rate of
profit.
>> >The techniques in such models can be reduced to a dated series of labor
>> inputs.
>> >The effects of compounding interest charges on these dated labor inputs
>> >are nonlinear. Thus the possibility for reswitching, capital reversing,
>> >and "perverse" behavior in the labor market arises.
>>
>> Now this is bit more like what I've been asking for, although you're still
>> basically describing the result and how to get to it. I still fail to see
the
>> logic of the model or the intuition behind it, perhaps you've not just
>used to
>> communicating w/ economists.
>
>Which part do you have a problem with?
Perhaps you've used to discussing w/ Sraffian economists more than w/
neoclass.
Lets start w/ (1) "pure circulating capital", (2) "lower wage will be
associated with a higher rate of profit", (3) "effects of compounding interest
charges on these dated labor inputs are nonlinear".
The (1) seems like something from ye olde labor value theory. Other than that,
I don't understand what it means, esp. how it differs from neoclass prod fn
idea.
Point (2) seems to state that, ceteris paribus, lower wage means higher
profits. I doubt this interpretation is the complete content of it?
What you say in (3) is quite vague. I can imagine different plenty non-linear
functions. Are they exponential as compounding might suggets? Monotically
increasing? Or do they go wildly all over the phere? What does "interest
charges on these dated labor inputs" mean in the first place?
Further, (3) sounds bit like income effects. You compound interest and become
richer. Right?
>Some economists (many are not Sraffians) would find this old hat. Isn't it
>kind of arrogant to assume your gaps in understanding and your intuitions
>are representative of the entire profession?
Not really. I challenge you to find, say, a reputable PhD program (other than
the New School), where Sraffian stuff is even mentioned.
>> This is like describing the ituition and driving forces of neoclass GE
models
>> as: "Mkts clear, prices equal marginal cost, and profits tend to zero.
There
>> are many equil. We can solve equil by fixed point theorems. We have no idea
>> which of these will be selected." If the listener of this stoty is not an
>> economist, the above is next to meaningless.
>
>No. Profits *tending* to zero is a dynamic story that you previously agreed
>does not apply in the Arrow-Debreu model. Anyway, my example has markets
But it could. It is a sorry excuse for a dynamic story but it sure could be
interpreted as a one.
>that clear, prices equalling marginal cost in the sense that applies to
>discrete models, and no pure economic profits. But having studied
>economics somewhat, I understand your summary.
>
>> [snip]
>
>> >Your intuition seems to be based on
>> >
>> > (1) Unformalized person-on-the street views
>>
>> My intuition is based on articles on, in no particular order, and not
>> restricted to, JET, AER, JPE, Econometrica, JEMS, GEB, Rand, QJE, ...
>
>And my intuition is based on articles in the Review of Economic
>Studies, the Quarterly Journal of Economics, the Journal of Political
>Economy, the Journal of Economic Literature, etc. You have a point?
My point: How many person-on-the street regularly reads these pieces
blockbuster entertainment to base their intuitions?
>> >I claim that perhaps these errors became widespread and are still held on
>> >to today because of political influences.
>>
>> Yeah, right.
>
>Nice to see that your attitude towards hypotheses about political
>influences on the development of economics cannot be characterized
>by emotional terms such as "liking" or "disliking."
I take expectations over integrals b/c and use sequential equilibrium concept
b/c I didn't vote for communists?
Economists have a variety of political views. Your political views influence
what and how you do. Satisfied?
[snip]
>> Lemme restate:
>>
>> Agents live in non-stochastic and dynamic world, and interact in some
manner,
>> perhaps on mkts. Is is sensible to restrict the agents from understanding
any
>> of the links b/w past, present and future?
>
>Who said anything about restricting the agents to not understanding *any*
>of the links between past, present, and future? How can the agents tell
You said they don not have perfect foresight. If everything is deterministic,
and if the links b/w time periods are as straightforward as I guess they are,
perfect foresight = understanding links between past, present, and future.
Right? Or if there are some issues that muddle this, as you indicate below,
change that perfect foresight to rational expectations.
>from observing that present spot markets clear that their plans are
>sufficiently coordinated that spot markets will clear next week if
>they follow through with their plans? Are you sure that you don't have
>a centrally planned economy in mind?
No. Besides, why would we even need mkt clearance for rational expectations?
Here, if there are doubt on mkt clearance, RE would have things like
probalility of being able to purchase quantity Q_i,t+k at price P_i,t+k at
time t+k.
Robert Vienneau
Markku Stenborg writes:
> O 'ear ye mortals, the infinite wisdom by Robert Vienneau of
> rv...@dreamscape.com ...
[snip]
> The intuition of and/or the driving forces behind the results is the fact
> that, in your version of a temporal model, labor inputs in
non-contiguous time
> periods are complements?
Some neoclassical economists think that that is the driving force behind
the "perverse" behavior in the labor and capital markets.
Suppose we want to produce one widget on January 1, 1997. We have two
techniques available with the following inputs:
94 95 96
Technique 1 0 7 Workers 0
Technique 2 2 Workers 0 6 Workers
Assume workers are paid at the end of the year. We have to pay interest
charges on the wages we pay to the workers until we sell our product.
Thus, the cost of the first technique is (in obvious notation):
7 w (1 + r)
The cost of the second technique is:
2 w (1+r)^2 + 6 w
The difference between the cost of the second and the cost of the
first is:
2 w (1+r)^2 - 7 w (1+r) + 6 w
When this quantity is positive, the 1st technique is cheaper. When
it is negative, the 2nd technique is cheaper. When zero, either
technique may be adopted (a switch point).
Since the first technique employs less total workers in the 3 years, it
is cheapest at r = 0. One can see that the above quantity is zero at
r = 50% and r = 100%. Thus we have the following map from interest rates
to cost minimizing technique:
1st technique 2nd technique 1st technique
|<--------------->|<----------------->|<-------
0 50% 100%
Suppose we compare stationary states producing 1 widget each year. For the
first technique, the output-labor ratio is 1/7 Widgets per worker. It
is 1/8 Widget per worker for the second technique. Note that around an
interest rate of 100%, a lower interest rate is associated with a lower
output-labor ratio. (Isn't your intuition that a lower interest rate
implies more capital per head implies greater output per head?)
I'll let you figure out to construct a wage-rate of profits curves for
these techniques and to demonstrate that a higher wage may be associated
with an adoption of a more labor-intensive technique.
I'll also let you figure out how to convert the data on techniques in
my previous example to inputs consisting solely of dated quantities of
labor. I'll give you a hint: the resulting sequence will not contain a
finite number of terms.
The above example was first created by Paul Samuelson about three decades
ago and was even in his intro text in the 70s. So why don't economists
have their intuition trained to understand it?
[snip]
> >Since we started, I've looked up a reference by Tatsuo Hatta that shows how
> >to get switches with "perverse" capital-deepening with continuously
> >differentiable neoclassical production functions in an Austrian context.
> >These switches would also show "perverse" behavior in the labor market. So
> >that claim is established in the literature in some articles that
> >frankly I've never fully understood.
>
> Yep, that Austrian stuff I never understood. It seems to make sense for some
> guys.
Hatta's claim is that "perverse" switches are possible with well-behaved
production functions of the form f( L1, L2, L3 ). This is Austrian like
Bohm-Bawerk was Austrian; nothing mystical here.
[snip]
> > o My example is not an example of the Arrow-Debreu model
>
> This has been clear all along. You might have wanted to point out where the
> differences lie, though.
I have. Maybe another example will help:
---------------------------------------------------------------------------
This example is from J. E. Woods, _The Production of Commodities: An
Introduction to Sraffa_, 1990.
Consider a simple two-good model economy. Assume only one good, corn, can act
as a consumption good. Both goods, corn and steel, are used as inputs into
production processes. All production processes require a year to complete,
and totally use up their inputs of corn and steel. (So this is a circulating
capital model.)
Assume only one process is known for producing corn, as given by the
Leontief production function:
Q2 = min[ L2, 2*X12 ]
where Q2 is the bushels corn produced for sale at the end of the year, L2 is
the person-years of labor services hired in the corn industry at the beginning
of the year, and X12 is tons steel bought at the beginning of the year.
Two processes are available for producing steel. The alpha process is
specified by the following Leontief production function:
Q1( alpha ) = min[ 2*L1( alpha ), 2*X21( alpha ) ]
where Q1( alpha ) is the tons steel produced by the alpha process for sale at
the end of the year, L1( alpha ) is the person-years of labor services hired
in the alpha process at the beginning of the year, and X21( alpha ) is bushels
corn also bought at the beginning of the year.
The beta process for producing steel is specified in obvious notation by
the following function:
Q1( beta ) = min[ (1/3)*L1( beta ), 8*X21( beta ) ]
Although laborers start working at the beginning of the year, they are
paid at the end of the year.
Consider (long run) competitive equilibria with these properties:
o Spot prices are constant with time.
o No firm earns pure economic profits (and no firm can earn pure
economic profits by using some other process than the one they have
chosen)
o No firm operates a process in which costs exceed the price of the output.
o Quantity flows are appropriate for a stationary state.
QUESTIONS AND HINTS:
1. Let corn be numeraire. Find equilibria prices and quantities for any
given value of the rate of interest r. What range must r be in for an
equilibrium to exist?
2. How does the equilibrium wage vary with the rate of interest? How does the
price of steel vary with the rate of interest?
3. Define a technique as a combination of the corn-producing process with one
of the steel-producing processes. Which technique will be adopted at each
rate of interest? Define a switch point as a set of prices (the rate of
interest, wage, and price of steel) at which both techniques, or linear
combinations of techniques, are eligible for adoption by cost-minimizing
firms. Where is the switch point?
4. In Questions 2 and 3, you determined the choice of technique and the
relations between the rate of interest, the wage, and the price of steel.
Determine quantity flows for each technique such that the net output of
the economy consists of one bushel corn. Plot the wage versus the
labor-intensity of this vertically-integrated corn industry. Note that
around the switch point a slightly lower wage is associated with the
adoption of a more labor-intensive technique than the technique adopted
at a slightly higher wage.
5. Repeat Question 4, but plot the price of steel versus the steel-intensity
of the vertically-integrated corn industry. What prices of steel are
associated with more than one value for steel-intensity?
6. Define "substitution" as the adoption of a techinque that uses an input
more intensively when its price falls. Note that the price of steel at
the switch point can be associated with the adoption of the less steel-
intensive technique and that a slightly lower price can be associated
with an equilibrium that uses the more steel-intensive technique. This
relationship can go the other way since the preferred technique is not
a function of the price of steel. If substitution is taken as central
to neoclassical economics, discuss its logical coherence in light of
your answers to Question 5.
7. Conclude the failure of substitution to characterize the relation of
the price of a produced input (e.g. steel) to the factor-intensity of
the technique is independent of the existence of "perverse" relations
in the market for a non-produced good (labor) or in the "capital" market.
---------------------------------------------------------------------------
For ease of reference, Markku, we might as well call my original example
the "Vienneau example," the first example above the "Samuelson example"
and the latter the "Woods example." If you look at the two steel-producing
processes in the Woods example, you will notice that they trade off
two inputs, labor and corn, in a manner consistent with your intuition.
But the adoption of one or the other technique will require an
economy-wide change in the amount of steel used as an input since the
scale of the corn-producing industry will need to be changed. It looks
like we might be able to get the highlighted behavior with smooth
production functions - I don't know.
Notice that endowments of corn and steel are not data, but are
determined in the model (if we know the endowment of the labor force).
> [snip]
>
> >> I have not presented a model of some long-run equilibrium.
> >
> >Correct. You have not given a hint of how to formalize your intuitions
> >in a model of long run equilibrium.
>
> Yes. If I had presentd some results regarding long run behavior, say, I sure
> would explain what drives the results and what is the intuition behind. I
> would be carefull to spell out all assumptions that are or could be seen
to be
> crucial for the results, and clearly mark them from the ones that are mere
> simplifications w/o effects to qualitative results. Especially, if the
results
> would go agaist the grain, I'd be carefull to convice the reader that this is
> not due to some unrealistic behavioral assumptions or to strange restrictions
> posed on the agents or their interactions. I would discuss the info and the
> decisions rules of the agents, the transitions mechanisms from period to
> period, etc.
What would you do if you were quite aware that the literature was full
of economists of different schools who accept the possibility of your
qualitative results, but who disagree about the driving mechanisms and
their correct characterization?
> [snip]
>
> >I think "intuitions" that you give no argument for and no empirical
> >evidence for in the context under discussion can be fairly characterized
> >as "prejudices."
>
> No. Substitution is a result that seems to pop out regradless of the details
> of the model, and it makes intuitively a lot of sense. If something becomes
> more expensive, you want to reduce its use and substitute it w/ something
> else. Usually, this isn't the only effect, and as I've stated zillion times,
> these other effects could more than balance substitution. Perhaps you
think in
> manner where the above sounds unreasonable, and do not undestand the question
> of expalantions for these other effects.
Yes, I do not think substitution is a reasonable expectation or
characterization of comparisons of steady states. So I don't think I
need to find other overriding effects in this case. My long quote from
a book examining Classical theory suggests that there are other
economists who share my expectations and that some of them think the
Classical economists (particularly Ricardo, IMO) shared these expectations
too. Finally, if you were well-taught, your intuition should match
mine at least as far as the possible workings of these models. (And I
don't think you're atypical.)
> [...]
>
> >Assume I'm an ignoramus about how this example I magically conjured out of
> the
> >air works. If you are really curious about it, why don't you look up
> >one of my textbook references, say, Kurz and Salvadori? Then you could make
> >more intelligent comments than "Eh?"
>
> You have a comparative advantage in explaing the model you've spent some time
> in thinking about.
You'd rather accept the authorithy of a Usenet poster to a book with a
cover blurb by Samuelson and a generally positive review by Burmeister
(last summer's (?) JEL)?
[snip]
> Should we have a vote on this? Which one of youse guys understand the
> intuition behind and/or the forces that drive "higher wage, more labor
> intensive production" result? Perhaps the "Yeay" voters could enlighten the
> rest of us.
Uh, Markku, you know that I think most of the academic posters here share
your lack of logic and lack of knowledge of the literature. But I'd
gladly hear interjections by others who understand this stuff. I
know where in cyberspace to find academics who do understand it.
[snip]
> >So where is your formal model of long run equilibrium in which
> >your intuitive substitution behavior follows from your assumptions?
> >And where's the empirical evidence for it? Oh, only one of us has to
> >provide an argument - I see.
>
> Should I present a model, I'll promise I'll ecplain exactly what I've been
> asking from you.
I see you refuse the challenge. Aren't models of long-run equilibrium
common in the literature, e.g. from Walras to the New Growth Theory?
> Besides, forces that drive substitution shouldn't call that much explanation?
They should in this context if your intuition was properly trained. (My
explanation of how to train your intuition you dismissed as "silly.")
> [snip]
>
> >> >In a comparison of long-run equilibrium positions in a pure circulating
> >> >capital model, a lower wage will be associated with a higher rate of
> profit.
> >> >The techniques in such models can be reduced to a dated series of labor
> >> inputs.
> >> >The effects of compounding interest charges on these dated labor inputs
> >> >are nonlinear. Thus the possibility for reswitching, capital reversing,
> >> >and "perverse" behavior in the labor market arises.
> >>
> >
> >Which part do you have a problem with?
>
> Perhaps you've used to discussing w/ Sraffian economists more than w/
> neoclass.
No. I'm used to discussions elsewhere with economists of a variety of
schools who are aware of the contents of the literature treating issues of
contention for decades.
> Lets start w/ (1) "pure circulating capital", (2) "lower wage will be
> associated with a higher rate of profit", (3) "effects of compounding
interest
> charges on these dated labor inputs are nonlinear".
>
> The (1) seems like something from ye olde labor value theory. Other than
that,
> I don't understand what it means, esp. how it differs from neoclass prod fn
> idea.
Are you aware of wool-mutton examples of joint production? Now consider
a case of a process that uses inputs of labor, a new machine, and perhaps
fuel and raw material to produce (after some period of time) something-
or-other and an older-and-used machine. The fact that the machine can
be used again in a process (perhaps with a different efficiency)
following this one makes this a case of "fixed capital." If the machine
were all used up in the first process, it would be an example of
"circulating capital." I hope my description makes it clear why fixed
capital can be seen as a special case of joint production. (So can
land.)
Where's the logical connection with the labor theory of value? I don't
see why neoclassicals cannot use this framework (although they should
then recognize that their usual formula for depreciation is a special
case). Johnny von Neumann used it too, and he understood depreciation.
> Point (2) seems to state that, ceteris paribus, lower wage means higher
> profits. I doubt this interpretation is the complete content of it?
Your restatement is inaccurate. A lower wage would be associated with
a change in equilibrium prices. So the value of capital would be
changed. How can you be sure that
(Profits)/(value of capital)
is higher?
Are you aware of the use of Perron-Frobenius theorems in economics?
Since Samuelson invented one formulation of this relation in defending
the Solow growth model (badly), how can it be non-neoclassical? On the
other hand, I think here's where my model can be characterized as a
surplus-based theory of value.
> What you say in (3) is quite vague. I can imagine different plenty non-linear
> functions. Are they exponential as compounding might suggets?
Yes.
> Monotically
> increasing? Or do they go wildly all over the phere? What does "interest
> charges on these dated labor inputs" mean in the first place?
The Samuelson example at the start should give you some obvious hints.
> Further, (3) sounds bit like income effects. You compound interest and become
> richer. Right?
I don't think so. It's just a matter of present value calculations.
> >Some economists (many are not Sraffians) would find this old hat. Isn't it
> >kind of arrogant to assume your gaps in understanding and your intuitions
> >are representative of the entire profession?
>
> Not really. I challenge you to find, say, a reputable PhD program (other than
> the New School), where Sraffian stuff is even mentioned.
Although you do rule out my strongest U.S. example, I could easily
answer your question if I thought it relevant. I think the following
economists, at least, have a good grasp of my arithmetic: Samuelson,
Solow, Hahn, Burmeister, and Tobin (after tutoring by Steedman). Don't
you think a good PhD program should train a student to be able to
follow their arguments? (I don't claim that those economists share
what some sci.econ readers might think are my judgements.)
[snip]
> >> >Your intuition seems to be based on
> >> >
> >> > (1) Unformalized person-on-the street views
> >>
> >> My intuition is based on articles on, in no particular order, and not
> >> restricted to, JET, AER, JPE, Econometrica, JEMS, GEB, Rand, QJE, ...
> >
> >And my intuition is based on articles in the Review of Economic
> >Studies, the Quarterly Journal of Economics, the Journal of Political
> >Economy, the Journal of Economic Literature, etc. You have a point?
>
> My point: How many person-on-the street regularly reads these pieces
> blockbuster entertainment to base their intuitions?
Clearly I find them entertaining, but then I received a No Fear t-shirt
for X-mas that read "For those who see the world differently."
Anyways, my intuition was trained by my reading and thinking.
> >> >I claim that perhaps these errors became widespread and are still held on
> >> >to today because of political influences.
>
>[...]
> I take expectations over integrals b/c and use sequential equilibrium concept
> b/c I didn't vote for communists?
BTW, I put 1989 in the same sequence that includes 1848.
> Economists have a variety of political views. Your political views influence
> what and how you do. Satisfied?
No, I'm not. What views get accepted, promoted, and taught? Are there
political influences there? Think of how many forerunners we can
find for Jevons. Was it an accident that neoclassical economics caught
on then, and not before? Was there a new element in Jevon's theory? Was
it something to do with the outside world - perhaps the most rigorous
restatement ('til then) of Classical economics by Marx? Did this have
something to do with the class situation, say, working class concerns
seemingly opposed to capitalist interests? (1848 could have been a
turning point here - leading economists to have been reluctant to have
developed the full Ricardian logic). I don't even need Jevons to
have been aware of Marx for the suggested story (although it's
interesting how many early neoclassicals wrote on Marxism.)
Or let me take a less debatable case. Let's not explain Keynesianism
by Keynes' world-view. Think rather why Keynesianism became temporarily
accepted. The Great Depression is an obvious external influence on
the development of economics. It's not just because it was a
disinterested scientific puzzle to be explained.
But I turn to another thread for a contemporary example. Don Dale answers
William Hummel's thoughts on slow growth:
> >Second, even if [capital]
> >did what you say, there is no assurance that it would flow its most
> >productive potential uses (a vague term, at best).
>
> Not vague at all. The marginal product of capital will in equilibrium be
> equal to the real interest rate adjusted for market perceptions of risk. If
> one investment offers a rate of return higher than others, capital will flow
> to that investment and away from others, lowering the marginal productivity
> of additional investment in the former and raising it in the latter.
I claim that Don's comments are an example of a logical error, and an
error that was exposed decades ago. Note the movement between finance
and "real" capital in that thread. Note Don's context - he is trying
to show something about how "capital" is "productive." Are you sure I
cannot read this persistence in exposed error as having something to do
with politics, and even more so, politics related to my suggested
Jevons vs. Marx/Ricardo story?
I know history of economics is not to your taste, so feel free to snip the
above paragraphs.
[snip]
Markku Stenborg
O 'ear ye mortals, the infinite wisdom by Robert Vienneau of
rv...@dreamscape.com ...
>
>Markku Stenborg writes:
>> O 'ear ye mortals, the infinite wisdom by Robert Vienneau of
>> rv...@dreamscape.com ...
>
>[snip]
>
>> The intuition of and/or the driving forces behind the results is the fact
>> that, in your version of a temporal model, labor inputs in
>non-contiguous time
>> periods are complements?
>
>Some neoclassical economists think that that is the driving force behind
>the "perverse" behavior in the labor and capital markets.
I do have some idea what some economists think about driving forces behind
some results. I was expecting a simple "Yes" or "No", with perhaps some
qualifiers regarding your model. This is what mean when I complain about
communication w/ you.
So, does this mean "Yes, if labor inputs in non-contiguous time periods are
complements, the model is temporal and competitive, and given the assumptions
that ...., higher wage level will be associated w/ a more labor intesive
production in equilibrium"?
[snip]
> 1st technique 2nd technique 1st technique
>
> |<--------------->|<----------------->|<-------
> 0 50% 100%
>
>Suppose we compare stationary states producing 1 widget each year. For the
>first technique, the output-labor ratio is 1/7 Widgets per worker. It
>is 1/8 Widget per worker for the second technique. Note that around an
>interest rate of 100%, a lower interest rate is associated with a lower
>output-labor ratio. (Isn't your intuition that a lower interest rate
>implies more capital per head implies greater output per head?)
There isn't any capital in the model, or if there is, its level cannot be
changed? Anyways, this is again the Leontieff model, and even w/ 2 kinks, I
find L to be not the appropriate choice for models where the crucial issues
are inputs and their prices and especially when the model is supposed to
discuss long run positions.
Anyways, to inject some realism, first, I find interest rates in excess of few
dozen percentage points unappealing (there's no risks here). We could fool w/
the numbers to have cut-off interest rates less than 50 %?
Second, here, the interest rate is important b/c the hiring and manufacturing
occurs before the item is payed for. So, the model applies only (i) to items
that take a long to produce and (ii) in mkts where the buyer cannot or will
not pay before delivery. If production takes days or weeks instead of years,
the interest charges are next to irrelevant. If the buyer pays ex ante or as
the costs of the project incur, the interest charges are next to irrelevant.
In these cases, the first tech would always be preferred, and no switching
would occur.
[snip]
>The above example was first created by Paul Samuelson about three decades
>ago and was even in his intro text in the 70s. So why don't economists
>have their intuition trained to understand it?
I, at least, understand it. My main complaint is still: Leontieff is not a
good choice if you want to make arguments on inputs and their prices.
[snip]
>> >> I have not presented a model of some long-run equilibrium.
>> >
>> >Correct. You have not given a hint of how to formalize your intuitions
>> >in a model of long run equilibrium.
>>
>> Yes. If I had presentd some results regarding long run behavior, say, I
sure
>> would explain what drives the results and what is the intuition behind. I
>> would be carefull to spell out all assumptions that are or could be seen
>to be
>> crucial for the results, and clearly mark them from the ones that are mere
>> simplifications w/o effects to qualitative results. Especially, if the
>results
>> would go agaist the grain, I'd be carefull to convice the reader that this
is
>> not due to some unrealistic behavioral assumptions or to strange
restrictions
>> posed on the agents or their interactions. I would discuss the info and the
>> decisions rules of the agents, the transitions mechanisms from period to
>> period, etc.
>
>What would you do if you were quite aware that the literature was full
>of economists of different schools who accept the possibility of your
>qualitative results, but who disagree about the driving mechanisms and
>their correct characterization?
Not much, I guess. I can live w/ diversity of views and modelling choices. At
least I wouldn't claim that unless they agree w/ me, their school of thought
is dead or necessarily wrong.
[snip]
>> >Assume I'm an ignoramus about how this example I magically conjured out of
>> the
>> >air works. If you are really curious about it, why don't you look up
>> >one of my textbook references, say, Kurz and Salvadori? Then you could
make
>> >more intelligent comments than "Eh?"
>>
>> You have a comparative advantage in explaing the model you've spent some
time
>> in thinking about.
>
>You'd rather accept the authorithy of a Usenet poster to a book with a
>cover blurb by Samuelson and a generally positive review by Burmeister
>(last summer's (?) JEL)?
No. The authority I accept is clear and logical reasoning and emprical
evidence.
You've spent some time thinking about these examples, so I expected you can
explain the workings of the models, and separate assumptions that are crucial
from the one's that aren't.
Polite cover blurbs are, mmm, polite cover blurbs. Burmeister's review I
haven't seen, so I have no comment on it.
>[snip]
>
>> Should we have a vote on this? Which one of youse guys understand the
>> intuition behind and/or the forces that drive "higher wage, more labor
>> intensive production" result? Perhaps the "Yeay" voters could enlighten the
>> rest of us.
>
>Uh, Markku, you know that I think most of the academic posters here share
>your lack of logic and lack of knowledge of the literature. But I'd
Now which logic do I lack? Which rules of which logic have I violated?
I admit my knowledge of Sraffian literature is non-existent, and if you're a
reprentative example, I'll be quit happy to remain ignorant. I don't see much
interesting in those examples.
>gladly hear interjections by others who understand this stuff. I
>know where in cyberspace to find academics who do understand it.
>
>[snip]
>
>> >So where is your formal model of long run equilibrium in which
>> >your intuitive substitution behavior follows from your assumptions?
>> >And where's the empirical evidence for it? Oh, only one of us has to
>> >provide an argument - I see.
>>
>> Should I present a model, I'll promise I'll ecplain exactly what I've been
>> asking from you.
>
>I see you refuse the challenge. Aren't models of long-run equilibrium
No. My formal models are those of neo-neoclass econ. I've told you zillion
times, even in neoclass econ, my intuition is that there are forces that more
than balance substitution effects.
>common in the literature, e.g. from Walras to the New Growth Theory?
Yes. Should I present a model of endogenous growth, say, I promise I will
explain the intuition, the forces that drive the results, and make a
distinction b/w crucial and merely simplifying assumptions.
How about closing your model by including labor and capital mkts to determine
w and r, given the data?
>> Besides, forces that drive substitution shouldn't call that much
explanation?
>
>They should in this context if your intuition was properly trained. (My
>explanation of how to train your intuition you dismissed as "silly.")
Eh? I don't recall dismissing anything as silly, and esp. don't see this on
your post. Care provide the evidence?
>> [snip]
>>
>> >> >In a comparison of long-run equilibrium positions in a pure circulating
>> >> >capital model, a lower wage will be associated with a higher rate of
>> profit.
>> >> >The techniques in such models can be reduced to a dated series of labor
>> >> inputs.
>> >> >The effects of compounding interest charges on these dated labor inputs
>> >> >are nonlinear. Thus the possibility for reswitching, capital reversing,
>> >> >and "perverse" behavior in the labor market arises.
>> >>
>> >
>> >Which part do you have a problem with?
>>
>> Perhaps you've used to discussing w/ Sraffian economists more than w/
>> neoclass.
>
>No. I'm used to discussions elsewhere with economists of a variety of
>schools who are aware of the contents of the literature treating issues of
>contention for decades.
Your explantions of these examples are far from what I've learned to expect
from economists who present models, and esp. if they want to claim that their
results are not what economists would expect.
Usually, it is required that the crucial assumptions are marked as such and
not quietly left out of discussion, and that froces driving the results are
explained. In your case, this would've meant spelling out the restrcitions on
production and selling, explaing the lacks of labor and capital mkts, etc.
>> Lets start w/ (1) "pure circulating capital", (2) "lower wage will be
>> associated with a higher rate of profit", (3) "effects of compounding
>interest
>> charges on these dated labor inputs are nonlinear".
>>
>> The (1) seems like something from ye olde labor value theory. Other than
>that,
>> I don't understand what it means, esp. how it differs from neoclass prod fn
>> idea.
>
>Are you aware of wool-mutton examples of joint production? Now consider
Yes.
>a case of a process that uses inputs of labor, a new machine, and perhaps
>fuel and raw material to produce (after some period of time) something-
>or-other and an older-and-used machine. The fact that the machine can
>be used again in a process (perhaps with a different efficiency)
>following this one makes this a case of "fixed capital." If the machine
>were all used up in the first process, it would be an example of
>"circulating capital." I hope my description makes it clear why fixed
>capital can be seen as a special case of joint production. (So can
>land.)
So your result requires that the capital is completely eaten up in the
production process?
>Where's the logical connection with the labor theory of value? I don't
>see why neoclassicals cannot use this framework (although they should
>then recognize that their usual formula for depreciation is a special
>case). Johnny von Neumann used it too, and he understood depreciation.
I thought they used this framework, among others?
>> Point (2) seems to state that, ceteris paribus, lower wage means higher
>> profits. I doubt this interpretation is the complete content of it?
>
>Your restatement is inaccurate. A lower wage would be associated with
>a change in equilibrium prices. So the value of capital would be
>changed. How can you be sure that
>
> (Profits)/(value of capital)
>
>is higher?
>
>Are you aware of the use of Perron-Frobenius theorems in economics?
Perron doesn't ring a bell.
There's at least a Frobenius Theorem on diff eqs. I recall it states something
along the following lines: The symmetry of 1st order partial derivative matrix
is necessary and sufficient condition for the existence of its solution. This
is what you're referring to?
[snip]
>> Not really. I challenge you to find, say, a reputable PhD program (other
than
>> the New School), where Sraffian stuff is even mentioned.
>
>Although you do rule out my strongest U.S. example, I could easily
>answer your question if I thought it relevant. I think the following
You claimed my ignorance on Sraffa is not representative.
>economists, at least, have a good grasp of my arithmetic: Samuelson,
>Solow, Hahn, Burmeister, and Tobin (after tutoring by Steedman). Don't
I have never been complaing about your arithemetic.
>you think a good PhD program should train a student to be able to
>follow their arguments? (I don't claim that those economists share
Yes. They should also train them to think critically and not to take opinions
of past authorities as anything else than opinions, even their names were
Sraffa or Marx.
[snip]
>> Economists have a variety of political views. Your political views
influence
>> what and how you do. Satisfied?
>
>No, I'm not. What views get accepted, promoted, and taught? Are there
Ideally, the ones that have best arguments.
>political influences there? Think of how many forerunners we can
Some, but not much. Not all Econ depts are caricatures of Chicago.
>find for Jevons. Was it an accident that neoclassical economics caught
>on then, and not before? Was there a new element in Jevon's theory? Was
No, but I don't that was crucial. A whole lot was needed to get neoclass econ
started.
>it something to do with the outside world - perhaps the most rigorous
>restatement ('til then) of Classical economics by Marx? Did this have
>something to do with the class situation, say, working class concerns
>seemingly opposed to capitalist interests? (1848 could have been a
>turning point here - leading economists to have been reluctant to have
>developed the full Ricardian logic). I don't even need Jevons to
>have been aware of Marx for the suggested story (although it's
>interesting how many early neoclassicals wrote on Marxism.)
>
>Or let me take a less debatable case. Let's not explain Keynesianism
>by Keynes' world-view. Think rather why Keynesianism became temporarily
>accepted. The Great Depression is an obvious external influence on
>the development of economics. It's not just because it was a
>disinterested scientific puzzle to be explained.
But this has nothing to do w/ polital views? Depression is depression
regardless of your political thoughts or lack thereof.
>But I turn to another thread for a contemporary example. Don Dale answers
>William Hummel's thoughts on slow growth:
>
>> >Second, even if [capital]
>> >did what you say, there is no assurance that it would flow its most
>> >productive potential uses (a vague term, at best).
>>
>> Not vague at all. The marginal product of capital will in equilibrium be
>> equal to the real interest rate adjusted for market perceptions of risk.
If
>> one investment offers a rate of return higher than others, capital will
flow
>> to that investment and away from others, lowering the marginal productivity
>> of additional investment in the former and raising it in the latter.
>
>I claim that Don's comments are an example of a logical error, and an
>error that was exposed decades ago. Note the movement between finance
>and "real" capital in that thread. Note Don's context - he is trying
>to show something about how "capital" is "productive." Are you sure I
>cannot read this persistence in exposed error as having something to do
The error being? His mind is fixed w/ idea that amount of capital can be
adjusted as the costs change?
>with politics, and even more so, politics related to my suggested
>Jevons vs. Marx/Ricardo story?
>
>I know history of economics is not to your taste, so feel free to snip the
>above paragraphs.
________
Robert Vienneau
In article <5durlj$h...@idefix.eunet.fi>, real.a...@bottom.of.msg (Markku
Stenborg) wrote:
In article <5durlj$h...@idefix.eunet.fi>, real.a...@bottom.of.msg (Markku
Stenborg) wrote:
> O 'ear ye mortals, the infinite wisdom by Robert Vienneau of
> rv...@dreamscape.com ...
> >
> >Markku Stenborg writes:
> >> O 'ear ye mortals, the infinite wisdom by Robert Vienneau of
> >> rv...@dreamscape.com ...
> >
> >[snip]
> >
> >> The intuition of and/or the driving forces behind the results is the fact
> >> that, in your version of a temporal model, labor inputs in
> >non-contiguous time
> >> periods are complements?
> >
> >Some neoclassical economists think that that is the driving force behind
> >the "perverse" behavior in the labor and capital markets.
>
> I do have some idea what some economists think about driving forces behind
> some results. I was expecting a simple "Yes" or "No", with perhaps some
> qualifiers regarding your model. This is what mean when I complain about
> communication w/ you.
You seem to need repetition.
> So, does this mean "Yes, if labor inputs in non-contiguous time periods are
> complements, the model is temporal and competitive, and given the assumptions
> that ...., higher wage level will be associated w/ a more labor intesive
> production in equilibrium"?
Yes, with "will" changed to "may." If
1) Technology can be represented as a choice of techniques with a point
output and inputs consisting solely of dated labor inputs
2) If labor inputs in non-contiguous time periods are complements
3) If firms take prices as given (are competitive) and know the
technology and prices (there are no information problems here)
4) Firms minimize cost
Then
In comparing steady states, one may find that a higher wage level will
be associated with a more labor-intensive cost minimizing technique.
My qualifications were in previous posts. To repeat -
(1) is not necessary. There are examples where it does not hold and
the "perverse" conclusion exists. (2) does not seem meaningful in these
examples where (1) does not hold. Furthermore, there are "perversities"
in the relationship of prices and the intensity of produced commodities
that can exist even when the capital and labor markets are not "perverse"
(See the Woods example).
Again, where have I made any assumption that is not neoclasscial?
> [snip]
>
> > 1st technique 2nd technique 1st technique
> >
> > |<--------------->|<----------------->|<-------
> > 0 50% 100%
> >
> >Suppose we compare stationary states producing 1 widget each year. For the
> >first technique, the output-labor ratio is 1/7 Widgets per worker. It
> >is 1/8 Widget per worker for the second technique. Note that around an
> >interest rate of 100%, a lower interest rate is associated with a lower
> >output-labor ratio. (Isn't your intuition that a lower interest rate
> >implies more capital per head implies greater output per head?)
>
> There isn't any capital in the model, or if there is, its level cannot be
> changed?
Models with inputs consisting solely of dated labor flows have been
a standard method of analysis in capital theory for over a century. The
value of capital (at a given rate of interest) varies between the
two techniques.
> Anyways, this is again the Leontieff model, and even w/ 2 kinks, I
> find L to be not the appropriate choice for models where the crucial issues
> are inputs and their prices and especially when the model is supposed to
> discuss long run positions.
The representation of techniques in the Samuelson example is not in
Leontief production functions, although I could map it into a model
with 5 Leontief production functions and 4 industries.
Anyways, from the post that you are supposedly responding to:
:: Hatta's claim is that "perverse" switches are possible with well-behaved
:: production functions of the form f( L1, L2, L3 ).
So the perversity is claimed in the refereed literature not to depend
on Leontief functions. I admitted that I did not fully understand the
argument. But don't you think that if you want to contradict the
literature, you might look it up and present an argument?
> Anyways, to inject some realism, first, I find interest rates in excess
of few
> dozen percentage points unappealing (there's no risks here). We could fool w/
> the numbers to have cut-off interest rates less than 50 %?
Yes, although we might need more periods. Think of the "years" in the
Samuelson example as decades, if you like.
> Second, here, the interest rate is important b/c the hiring and manufacturing
> occurs before the item is payed for. So, the model applies only (i) to items
> that take a long to produce and (ii) in mkts where the buyer cannot or will
> not pay before delivery. If production takes days or weeks instead of years,
> the interest charges are next to irrelevant. If the buyer pays ex ante or as
> the costs of the project incur, the interest charges are next to irrelevant.
> In these cases, the first tech would always be preferred, and no switching
> would occur.
If production takes a day, but the inputs consist of labor and a produced
good, then the reduction to dated labor inputs will consist of labor for
more than one day. That is, you will need labor from before the first
day to manufacture the produced input. This process can be repeated to
continue to find more labor inputs. From the post that you are
attempting to respond to:
>> I'll also let you figure out how to convert the data on techniques in
>> my previous example to inputs consisting solely of dated quantities of
>> labor. I'll give you a hint: the resulting sequence will not contain a
>> finite number of terms.
You find production with unassisted labor for only a few days or weeks
realistic? Curious, indeed, for a first-worlder.
It doesn't matter when the buyer pays. The fact that the physical world
imposes some time before at least one input and the output means that
somebody will have to do some time-discounting.
> [snip]
>
> >The above example was first created by Paul Samuelson about three decades
> >ago and was even in his intro text in the 70s. So why don't economists
> >have their intuition trained to understand it?
>
> I, at least, understand it. My main complaint is still: Leontieff is not a
> good choice if you want to make arguments on inputs and their prices.
Why don't you present an argument that what goes on at the perverse switch
depends on the assumption of Leontief production functions?
[snip]
> >> not due to some unrealistic behavioral assumptions or to strange
> restrictions
> >> posed on the agents or their interactions. I would discuss the info
and the
> >> decisions rules of the agents, the transitions mechanisms from period to
> >> period, etc.
> >
> >What would you do if you were quite aware that the literature was full
> >of economists of different schools who accept the possibility of your
> >qualitative results, but who disagree about the driving mechanisms and
> >their correct characterization?
>
> Not much, I guess. I can live w/ diversity of views and modelling choices. At
> least I wouldn't claim that unless they agree w/ me, their school of thought
> is dead or necessarily wrong.
OK. Long run equilibrium prices are not scarcity indices and comparisons of
steady states need not be consistent with substitution. If you disagree,
present an argument. I'll even take a reference to an argument that
addresses Pasinetti's mid-70s argument.
[snip]
> >> Should we have a vote on this? Which one of youse guys understand the
> >> intuition behind and/or the forces that drive "higher wage, more labor
> >> intensive production" result? Perhaps the "Yeay" voters could
enlighten the
> >> rest of us.
> >
> >Uh, Markku, you know that I think most of the academic posters here share
> >your lack of logic and lack of knowledge of the literature. But I'd
>
> Now which logic do I lack? Which rules of which logic have I violated?
I find it difficult to find correct arguments in your posts. In fact, most
of your assertions are unargued. So you cannot be said to follow any
rules of logic at all in presenting assertions without argument.
> I admit my knowledge of Sraffian literature is non-existent, and if you're a
> reprentative example, I'll be quit happy to remain ignorant. I don't see much
> interesting in those examples.
You seem to think that if you label something "Sraffian," you can quite
happily continue to deny that "1 + 1 = 2."
A little while ago you were asserting that you found counter-intuitive
results of interest. Glad to see you don't worry about a foolish consistency.
> >gladly hear interjections by others who understand this stuff. I
> >know where in cyberspace to find academics who do understand it.
> >
> >[snip]
> >
> >> >So where is your formal model of long run equilibrium in which
> >> >your intuitive substitution behavior follows from your assumptions?
> >> >And where's the empirical evidence for it? Oh, only one of us has to
> >> >provide an argument - I see.
> >>
> >> Should I present a model, I'll promise I'll ecplain exactly what I've been
> >> asking from you.
> >
> >I see you refuse the challenge. Aren't models of long-run equilibrium
>
> No. My formal models are those of neo-neoclass econ. I've told you zillion
> times, even in neoclass econ, my intuition is that there are forces that more
> than balance substitution effects.
So what? I'm not depending on income effects, information assymmetries,
principal agent problems, signalling, prestige goods, etc. Where do
you see the necessity of substitution in comparisons of long-run
equilibrium before adding your counter-balancing forces?
>
> >common in the literature, e.g. from Walras to the New Growth Theory?
>
> Yes. Should I present a model of endogenous growth, say, I promise I will
> explain the intuition, the forces that drive the results, and make a
> distinction b/w crucial and merely simplifying assumptions.
Walras assumed steady-state prices, equal rates of profits, and given
endowments. His model was logically inconsistent, and he realized he
had problems.
Endogeneous growth models, from what I understand, are 1-good models.
Where are logically consistent long run models to be found that behave in
accordance with your intuition about the behavior of many goods?
> How about closing your model by including labor and capital mkts to determine
> w and r, given the data?
My first example had a specified labor force and an overlapping generations
approach to close the model with labor and capital markets. Obviously
I have dropped that out in later examples because I did not think
it essential. It is needed to close the model in some way if you want
to formally analyze stability. I think a good way of closing it should
have a story behind it that does not require a special case of production
theory. At least those who want to tell such stories should present an
argument for their special case and clearly specify their assumptions.
> >> Besides, forces that drive substitution shouldn't call that much
> explanation?
> >
> >They should in this context if your intuition was properly trained. (My
> >explanation of how to train your intuition you dismissed as "silly.")
>
> Eh? I don't recall dismissing anything as silly, and esp. don't see this on
> your post. Care provide the evidence?
Several posts ago you wrote:
:: [snipped some silly excuses for not answering a simple question]
> >> [snip]
> >>
> >> >> >In a comparison of long-run equilibrium positions in a pure circulating
> >> >> >capital model, a lower wage will be associated with a higher rate of
> >> profit.
> >> >> >The techniques in such models can be reduced to a dated series of
labor
> >> >> inputs.
> >> >> >The effects of compounding interest charges on these dated labor inputs
> >> >> >are nonlinear. Thus the possibility for reswitching, capital reversing,
> >> >> >and "perverse" behavior in the labor market arises.
> >> >>
> >> >
> >> >Which part do you have a problem with?
[snip]
> Usually, it is required that the crucial assumptions are marked as such and
> not quietly left out of discussion, and that froces driving the results are
> explained. In your case, this would've meant spelling out the restrcitions on
> production and selling, explaing the lacks of labor and capital mkts, etc.
I start out with a (possibly infinite) list of techniques that the
firms can employ. I assume competition. I assume cost-minimization. I
could assume utility-maximization (as I did in my first example). I
assume everybody knows all the data. There aren't any special restrictions.
The only way there would be crucial assumptions here is if you introduced
them so as to ensure comparisons of steady states behaved in accordance
with your intuitions. I suspect that your bad temper isn't to be
explained solely by my tone, but by your inability to believe my
claim, which I think fairly well demonstrated by my examples and
your continuing inability to point out the special assumptions.
> >> Lets start w/ (1) "pure circulating capital", (2) "lower wage will be
> >> associated with a higher rate of profit", (3) "effects of compounding
> >interest
> >> charges on these dated labor inputs are nonlinear".
> >>
> >> The (1) seems like something from ye olde labor value theory. Other than
> >that,
> >> I don't understand what it means, esp. how it differs from neoclass
prod fn
> >> idea.
> >
> >Are you aware of wool-mutton examples of joint production? Now consider
>
> Yes.
>
> >a case of a process that uses inputs of labor, a new machine, and perhaps
> >fuel and raw material to produce (after some period of time) something-
> >or-other and an older-and-used machine. The fact that the machine can
> >be used again in a process (perhaps with a different efficiency)
> >following this one makes this a case of "fixed capital." If the machine
> >were all used up in the first process, it would be an example of
> >"circulating capital." I hope my description makes it clear why fixed
> >capital can be seen as a special case of joint production. (So can
> >land.)
>
> So your result requires that the capital is completely eaten up in the
> production process?
No, it is a simplifying assumption. A couple of posts ago, I wrote:
:: Most seriously, models with fixed capital cannot be reduced so neatly
:: to dated inputs of labor, but can still exhibit all these "perversities"
:: and more. Does anybody even know how to apply Hatta's criteria of
:: complementarity in these cases?
I might as well point out that these "perverse" behavior is possible
in models with land as well.
> >Where's the logical connection with the labor theory of value? I don't
> >see why neoclassicals cannot use this framework (although they should
> >then recognize that their usual formula for depreciation is a special
> >case). Johnny von Neumann used it too, and he understood depreciation.
>
> I thought they used this framework, among others?
How is that responsive to the question about a *logical* connection?
Ricardo used the distinction. Marx criticized it, and argued that the
distinction between what he called "variable capital" and "constant
capital" was more fundamental.
[snip]
> A lower wage would be associated with
> >a change in equilibrium prices. So the value of capital would be
> >changed. How can you be sure that
> >
> > (Profits)/(value of capital)
> >
> >is higher?
> >
> >Are you aware of the use of Perron-Frobenius theorems in economics?
>
> Perron doesn't ring a bell.
>
> There's at least a Frobenius Theorem on diff eqs. I recall it states
something
> along the following lines: The symmetry of 1st order partial derivative
matrix
> is necessary and sufficient condition for the existence of its solution. This
> is what you're referring to?
There are several Frobenius theorems - sometimes his name being coupled
with Perron. The ones I know are about matrics, but not differential
equations. They have something to do with the positivity of eigenvalues
and eigenvectors. They are set forth in Pasinetti (1977) and their
connection with my previous question is explained. I found them useful
for understanding some results I saw set forth in a technical report
using Markov chains.
> [snip]
>
> >> Not really. I challenge you to find, say, a reputable PhD program (other
> than
> >> the New School), where Sraffian stuff is even mentioned.
> >
> >Although you do rule out my strongest U.S. example, I could easily
> >answer your question if I thought it relevant. I think the following
>
> You claimed my ignorance on Sraffa is not representative.
I never did. Several posts ago, I wrote:
:: Some economists (many are not Sraffians) would find this old hat. Isn't it
:: kind of arrogant to assume your gaps in understanding and your intuitions
:: are representative of the entire profession?
Where do you find a complaint about your ignorance of *Sraffianism.* See
below for non-Sraffians that understand capital theory. (And I avoided the
word "ignorance" on purpose.)
> >economists, at least, have a good grasp of my arithmetic: Samuelson,
> >Solow, Hahn, Burmeister, and Tobin (after tutoring by Steedman). Don't
>
> I have never been complaing about your arithemetic.
Your characterizations of my arithmetic I find bizarre and unargued.
> >you think a good PhD program should train a student to be able to
> >follow their arguments? (I don't claim that those economists share
>
> Yes. They should also train them to think critically and not to take opinions
> of past authorities as anything else than opinions, even their names were
> Sraffa or Marx.
This comment does not seem to apply to me. You're the one who brought up
the name of Sraffa in this thread. I wanted to discuss possible behavior
under neoclassical assumptions. And where have I taken Marx as offering
anything but suggestions?
[snip]
> >No, I'm not. What views get accepted, promoted, and taught? Are there
>
> Ideally, the ones that have best arguments.
Obviously I don't think that happens.
> >political influences there? Think of how many forerunners we can
>
> Some, but not much. Not all Econ depts are caricatures of Chicago.
Who was talking about Chicago exclusively?
[snip]
> >But I turn to another thread for a contemporary example. Don Dale answers
> >William Hummel's thoughts on slow growth:
> >
> >> >Second, even if [capital]
> >> >did what you say, there is no assurance that it would flow its most
> >> >productive potential uses (a vague term, at best).
> >>
> >> Not vague at all. The marginal product of capital will in equilibrium be
> >> equal to the real interest rate adjusted for market perceptions of risk.
> If
> >> one investment offers a rate of return higher than others, capital will
> flow
> >> to that investment and away from others, lowering the marginal productivity
> >> of additional investment in the former and raising it in the latter.
> >
> >I claim that Don's comments are an example of a logical error, and an
> >error that was exposed decades ago. Note the movement between finance
> >and "real" capital in that thread. Note Don's context - he is trying
> >to show something about how "capital" is "productive." Are you sure I
> >cannot read this persistence in exposed error as having something to do
>
> The error being? His mind is fixed w/ idea that amount of capital can be
> adjusted as the costs change?
No, that is not the error. I continue to produce models in which the
value of capital goods and the produced inputs vary with prices,
including the prices of inputs. Yet somehow the interest rate is not
equal in equilibrium to the marginal product of capital (adjusted
for risk). My longest example was posted in January when you were
away, perhaps someplace with more sun in January.
So what's your resistance to reading textbooks that explain all this?
Markku Stenborg
>Stenborg) wrote:
[snip]
>> So, does this mean "Yes, if labor inputs in non-contiguous time periods are
>> complements, the model is temporal and competitive, and given the
assumptions
>> that ...., higher wage level will be associated w/ a more labor intesive
>> production in equilibrium"?
>
>Yes, with "will" changed to "may." If
>
> 1) Technology can be represented as a choice of techniques with a point
> output and inputs consisting solely of dated labor inputs
>
> 2) If labor inputs in non-contiguous time periods are complements
>
> 3) If firms take prices as given (are competitive) and know the
> technology and prices (there are no information problems here)
>
> 4) Firms minimize cost
>
>Then
>
> In comparing steady states, one may find that a higher wage level will
> be associated with a more labor-intensive cost minimizing technique.
Result that states "... may ..." is rather weak. I can recall one instance
where this "... may ..." stuff was publised paper. Usually, propositions take
the form "Given ..., the ... follow". I though you had something along these
lines, esp. since this was but one example based on Leontieff prod fn,
time-to-build techniques but no contracts b/w byuers and sellers, etc, not a
theory.
How about:
"Consider a production economy where production technologies are Leontieff for
each n > 1 industries. On industry i, there are i_k technologies available,
where k >= 1. Production on industry i lasts for i_t periods, t >= 1. All
labor and capital equal. Labor receive wages at the end of each period, but
the final product is sold and payed for only after i_t periods.
All mkts behave competitively, firms max profits, and workers a well-behaving
max preference map. There is a common discount factor.
Then there's a measurable(*) set of equilibria such that higher wage rate *is*
associated with a more labor-intensive cost minimizing technique, provided the
following auxiliary assumptions are satisfied."
(*) Measurable wrt exogenous data: prod fns, amount of labor and capital
endowments, tastes, etc.
This would move your perhaps zero-measureable example to the level of theory,
and youd have a definite conlusion instead of this "...may..." thing.
>My qualifications were in previous posts. To repeat -
>(1) is not necessary. There are examples where it does not hold and
>the "perverse" conclusion exists. (2) does not seem meaningful in these
>examples where (1) does not hold. Furthermore, there are "perversities"
>in the relationship of prices and the intensity of produced commodities
>that can exist even when the capital and labor markets are not "perverse"
>(See the Woods example).
>
>Again, where have I made any assumption that is not neoclasscial?
I guess not.
Have you made the most sensible assumptions is a better question. Neoclass
Econ, to many of us at least, is not a subset of medieval-style syllogism
repeating the accepted assumptions, but an empirical science.
Second, from purely formal point-of-view,
[snip]
>Models with inputs consisting solely of dated labor flows have been
>a standard method of analysis in capital theory for over a century. The
>value of capital (at a given rate of interest) varies between the
>two techniques.
The amount of capital can not be altered or moved from industry to another was
my concern.
[snip]
>Anyways, from the post that you are supposedly responding to:
>
>:: Hatta's claim is that "perverse" switches are possible with well-behaved
>:: production functions of the form f( L1, L2, L3 ).
>
>So the perversity is claimed in the refereed literature not to depend
>on Leontief functions. I admitted that I did not fully understand the
>argument. But don't you think that if you want to contradict the
>literature, you might look it up and present an argument?
I don't think I have contradicted any literature w/o an argument?
Lemme repat: IMHO, Leontieff, or stuff that can be transformed into L, is not
the appropriate choice for the model or example that concentrates on input
mkts. If L is not a necessary assumption, why do you insist on it?
>> Anyways, to inject some realism, first, I find interest rates in excess
>of few
>> dozen percentage points unappealing (there's no risks here). We could fool
w/
>> the numbers to have cut-off interest rates less than 50 %?
>
>Yes, although we might need more periods. Think of the "years" in the
>Samuelson example as decades, if you like.
I'd be willing to accept the idea that production takes some time so that
you'd need to take interests ínto account, but decades? Are we discussing
forestry? Transforming dinoraurs into oil? If such slow-paced production is
the case, usually, there're contracts that require the buyer to pay up as the
project evolves. Hence, the interest charges on dated labor input phenomen all
but disappear.
>> Second, here, the interest rate is important b/c the hiring and
manufacturing
>> occurs before the item is payed for. So, the model applies only (i) to
items
>> that take a long to produce and (ii) in mkts where the buyer cannot or will
>> not pay before delivery. If production takes days or weeks instead of
years,
>> the interest charges are next to irrelevant. If the buyer pays ex ante or
as
>> the costs of the project incur, the interest charges are next to
irrelevant.
>> In these cases, the first tech would always be preferred, and no switching
>> would occur.
The following in no manner answer my concern above.
>If production takes a day, but the inputs consist of labor and a produced
>good, then the reduction to dated labor inputs will consist of labor for
>more than one day. That is, you will need labor from before the first
>day to manufacture the produced input. This process can be repeated to
>continue to find more labor inputs. From the post that you are
>attempting to respond to:
If I understand this line-of-thought correctly, the dated labor has received
their wages and the producers of the intermediate stuff have received their
pay awhile back. Your result in that example seems to hinge partially on the
misalignement of labor inputs and the payments for the final good. If this
misalignement is removed, the "perversity" is removed, barring income or
wealth effects?
>>> I'll also let you figure out how to convert the data on techniques in
>>> my previous example to inputs consisting solely of dated quantities of
>>> labor. I'll give you a hint: the resulting sequence will not contain a
>>> finite number of terms.
>
>You find production with unassisted labor for only a few days or weeks
>realistic? Curious, indeed, for a first-worlder.
Umm, so it is an infinite seuquence of labor inputs who have already received
their wages and need not to discounted anymore?
>It doesn't matter when the buyer pays. The fact that the physical world
>imposes some time before at least one input and the output means that
>somebody will have to do some time-discounting.
Yes. If you need huge interest rates to get the result, you need huge time
lapses. The interest charges for a month's salaries, say, is zilch.
>> [snip]
>>
>> >The above example was first created by Paul Samuelson about three decades
>> >ago and was even in his intro text in the 70s. So why don't economists
>> >have their intuition trained to understand it?
>>
>> I, at least, understand it. My main complaint is still: Leontieff is not a
>> good choice if you want to make arguments on inputs and their prices.
>
>Why don't you present an argument that what goes on at the perverse switch
>depends on the assumption of Leontief production functions?
Leontieff is non-responsive to changes in relative prices, so that there is
no subsititution. Any effect, then, that works against substitution will
dominate.
[snip]
>> Not much, I guess. I can live w/ diversity of views and modelling choices.
At
>> least I wouldn't claim that unless they agree w/ me, their school of
thought
>> is dead or necessarily wrong.
>
>OK. Long run equilibrium prices are not scarcity indices and comparisons of
>steady states need not be consistent with substitution. If you disagree,
>present an argument. I'll even take a reference to an argument that
Oh, come on. Present a general model along the lines I suggest on the top
that allows the possibility of substitution, and not an example based on 50%
interest rate and Leontieff.
>addresses Pasinetti's mid-70s argument.
[snip]
>> >Uh, Markku, you know that I think most of the academic posters here share
>> >your lack of logic and lack of knowledge of the literature. But I'd
>>
>> Now which logic do I lack? Which rules of which logic have I violated?
>
>I find it difficult to find correct arguments in your posts. In fact, most
>of your assertions are unargued. So you cannot be said to follow any
I have asserted that your example is not satisfactory, and my arguments for
this should be transparent to all.
>rules of logic at all in presenting assertions without argument.
I've been pointing out the obvious problems w/ your example, not presenting
arguments to support some views. The burden of proof usually rests on the one
who is arguing, not on the one who points out the problems w/ the argument.
>> I admit my knowledge of Sraffian literature is non-existent, and if you're
a
>> reprentative example, I'll be quit happy to remain ignorant. I don't see
much
>> interesting in those examples.
>
>You seem to think that if you label something "Sraffian," you can quite
>happily continue to deny that "1 + 1 = 2."
No. Try some reading comprehension.
[snip]
>> >They should in this context if your intuition was properly trained. (My
>> >explanation of how to train your intuition you dismissed as "silly.")
>>
>> Eh? I don't recall dismissing anything as silly, and esp. don't see this on
>> your post. Care provide the evidence?
>
>Several posts ago you wrote:
>
>:: [snipped some silly excuses for not answering a simple question]
'Cause I asked a simple question which you did not answer, if I recall
correctly.
[snip]
>The only way there would be crucial assumptions here is if you introduced
>them so as to ensure comparisons of steady states behaved in accordance
So you don't know what the expression "crucial assumption" means? I mean the
assumptions that guarantee the result, w/o which the qualittative result would
not hold. OTOH, since you only have "...may...", nothing is much crucial b/c
the result is vague enough.
Anyways, it seems that the misalignement of wage payments and paymments from
the sales of final goods is one crucial assumption. I thought that Leontieff
prod fn is also one, but you seem to think otherwise.
>with your intuitions. I suspect that your bad temper isn't to be
Bad temper?
>explained solely by my tone, but by your inability to believe my
>claim, which I think fairly well demonstrated by my examples and
>your continuing inability to point out the special assumptions.
See above.
[snip]
>> I have never been complaing about your arithemetic.
>
>Your characterizations of my arithmetic I find bizarre and unargued.
Doesn't still ring a bell.
[snip]
Robert Vienneau
In article <5e1cvi$5...@idefix.eunet.fi>, real.a...@bottom.of.msg
(Markku Stenborg) wrote:
> Result that states "... may ..." is rather weak. I can recall one instance
> where this "... may ..." stuff was publised paper. Usually, propositions take
> the form "Given ..., the ... follow". I though you had something along these
> lines, esp. since this was but one example based on Leontieff prod fn,
> time-to-build techniques but no contracts b/w byuers and sellers, etc, not a
> theory.
I think I'll try several shorter posts making several points. This is
POST I in this series.
In 1965 David Levhari, a MIT economist, published, under encouragement
from Paul Samuelson, a theorem of the form Markku approves of: p -> q.
The q was "reswitching cannot happen." The theorem was mistaken and
Levhari's proof was in error, something like writing "less than" when
he should have written "less than or equal to" or vice versa.
A number of economists quickly created counterexamples to the
theorem and pointed out the mistake. The Samuelson example was part
of Paul Samuelson's concession that the critics were correct.
Now what would a counter example look like? Obviously it would have
p and not q. This is not equivalent to showing that p implies not
q. Rather it only shows that if p happens, q may happen.
Now when one has a result that some may find surprising, one might
spend some time constructing examples to better understand it. How
many more conditions can one add to p and still have the surprising
result? What other conclusions can be drawn? Can examples be constructed
that have some of these conclusions and not others? This sort of
"exploratory mathematics" is a typical heuristic approach for suggesting
theorems. Markku is correct that a lot of this is rarely published
(see Lakatos 1976), though I wonder if he can come up with some examples
from General Equilibrium that led economists towards believing that
the assumption of gross substitutability could not be easily relaxed without
loss of some desirable properties.
Finally, consider two models A and B. B contains all of A's assumption.
B also has some more assumptions. Then A is the general model and
B is the special case. One might be able to exhibit examples of A
that are not B. That is, one might be able to exhibit
behavior that may occur in A, but not B. In this case, one cannot say
A contains any "crucial assumptions" implying this behavior. Rather
B has the "crucial assumptions" (the additional special case assumptions)
that imply the interesting behavior is impossible.
I don't know of any adequate special case neoclassical model of long run
equilibrium in which substitution is guaranteed to characterize comparisons
of long run equilibrium. I know of some work which I consider unsatisfactory.
Robert Vienneau
This is POST 2 in my series of hopefully shorter points. I don't think
Markku has the mathematical apparatus to answer the questions about
generalizations here, although I assume he could understand the
references I previously gave.
Consider a simple example where corn is grown on rent-free land with
inputs of only seed corn and labor. The seed corn is bought on 1 Jan
and must be paid for then. The labor is also hired on 1 Jan, but the
wage w (in units of bushels corn) is paid on 31 Dec. The corn is
harvested on 31 Dec and immediately sold for a price of unity. (Corn
is the numeraire.) The yearly interest rate is r.
There is some production function that relates inputs of seed corn
and labor to the amount of corn harvested. I assume constant returns
to scale. I assume that the firm takes the wage and rate of interest
as given, and that these happen to be compatible with a stationary
state. The firm chooses some relative proportion of inputs of labor
and corn. I am not modeling that decision today. But however that
decision was made, I assume that I can observe the results. I
express these results by saying that a0 hours of labor and a1 bushels
of corn are hired per bushel of corn produced. Notice I am not assuming
fixed coefficients or a Leontief production function, just exploring
some necessary relationships in a stationary state.
(In a multicommodity pure circulating capital model, we would replace
a0 and a1 by a(i, j); i = 0, 1, ..., n; j = 1, ..., n; where a(i, j)
is the amount of input i used in the production of 1 unit of good j.
Let a0 represent the first row of this matrix and A represent the
remaining square matrix. One might refer to A as a Leontief matrix,
but I still haven't made any assumptions about the production
function other than CRS.)
Now let's discuss quantity relationships. Assume that the net output
consists of y bushels corn. (In the multicommodity case, y would be
a column vector.) y is net of the replacement of seed corn. That is
it consists of wages and the remaining output that could be consumed
while leaving an ability for the economy to reproduce itself at the
same level. This remainder is called profits by many, although some
Neoclassicals will prefer the term "interest." So net output consists
of the sum of wages and profits. By the assumption of a stationary
state, all y bushels are consumed.
We make an assumption than the chosen technique is "viable," that is,
that there is some net output after replacing the seed corn. In
symbols, 0 <= a1 < 1. Viability in the multicommodity case would mean
that there exists some level of operation of each of the n processes
in A such that there is some positive net output of at least one
commodity and all inputs are replaced. How can one state a mathematical
condition on A for viability?
Let q be the gross output of corn. We have the following relationship:
y = q - a1 q = (1 - a1) q (1)
Or
q = y/(1 - a1) (2)
So given the net output, we can figure out what gross output must
be. What would Equation 2 look like in the multicommodity case? Why
does one know that the matrix inverse exists and that all elements of
q are non-negative? What condition on A ensures they are all strictly
positive?
The amount of labor employed in the given year is (a0 q), but these
workers are not all employed directly in producing corn for immediate
consumption. Some are employed in producing seed corn for use next year in
producing corn to be consumed next year or the in the following years.
How many workers are involved in employing corn for immediate consumption?
The answer is (a0 y) workers, and (a1 y) bushels of seed corn were bought
as inputs for use by these workers. If this technique were used in
the immediately proceeding year, (a0 a1 y) workers would have been
employed in working up (a1^2 y) bushels of corn into this year's seed
corn. In general, the number of workers employed k years ago
(k = 0, 1, ...) in producing this year's net output would have been
a0 a1^k y (3)
This is a notional calculation. I do not care about the actual history
of this economy in performing it. I like to think of the results of
this calculation as displaying the number of workers employed this
year in producing corn available for consumption in subsequent years,
as shown in the following table:
1994 1995 1996 1997 1998 1999
a0 a1^3 y a0 a1^2 y a0 a1 y a0 y
... a0 a1^3 y a0 a1^2 y a0 a1 y a0 y
... a0 a1^3 y a0 a1^2 y a0 a1 y a0 y
... a0 a1^3 y a0 a1^2 y a0 a1 y
... a0 a1^4 y a0 a1^3 y a0 a1^2 y
.
.
.
Each row shows the labor inputs needed in a stationary state to
produce the net output available at the end of the year in which the
entries of the row end. It's a way of thinking about the column under
1997 when the economy is coordinated so as to be able to reproduce itself
forever. There are interesting questions here about whether this conception
can handle technical change, but those who have experimented with my
game will have seen some ways of handling this.
Consider the sum of the labor inputs in the 1997 column:
a0 (1 + a1 + a1^2 + ... ) y = a0 y/(1 - a1) (4)
Equation 4 agrees with our previous result for labor hired in the
given year.
So we see that a sequence of dated labor inputs is just another way of
looking at the input/output relationships in a single year. The
sequence is merely a method of expressing that some current labor is
spent on producing commodities not for immediate consumption, but for use
in producing goods in later years.
We can also imagine many firms working in this example. They buy inputs
at the beginning of the year and sell their outputs at the end of the
year. Since production takes a year here, they will earn some
interest/profit on the costs of the seed corn. Assuming a uniform
rate of profit, the price equations in this very simple example will
be:
a1 (1 + r) + ao w = 1 (5)
(In the n good case, we have
p A (1 + r) + ao w = p (6)
where p is a row vector of prices. Setting the numeraire leaves one
degree of freedom.)
Expressing the inputs as a sequence of dated labor inputs suppresses
transactions between firms where they immediately sell their corn once
is it produced. Including these transactions, however, will not make
interest charges negligible; in fact, the size of interest charges
will be unaffected.
I hope this shows Markku more about the relationship between my first
example and the Samuelson example. I still claim his characterizations
of my arithmetic are bizarre. The above will make more apparent why.
Robert Vienneau
POST 3 of my response and then I quit. Here I consider the choice of
technique for a 2 good model. Let a process for producing the first
good be represented by ( a0( alpha ), a( alpha ) ) where a0( alpha )
is the labor input per unit output and a( alpha ) is a 2 element
column vector. Let a process for producing the second good be in
obvious notation ( b0( beta ), b( beta ) ).
A technique is represented by the row vector
a0 = ( a0( alpha ) b0( beta ) )
and the 2x2 Leontief matrix
A = ( a( alpha ) b( beta ) )
I still have not assumed Leontief production functions. If alpha and
beta are index variables from an uncountably infinite sets, this
representation could be well-behaved continuously differentiable
production functions. They could also be index variables from finite
sets. In that case, we still don't have "fixed coefficients" since
one could vary the coefficients by choosing another process.
Assume spot prices are stationary. Let p be the 2 element row vector
of prices. Let w and r be as usual. A process for producing the jth good
pays extra profits if
p x( gamma ) (1 + r) + x0( gamma ) w < p( j )
where x = a, if j = 1
b, if j = 2.
A process will incur extra costs if
p x( gamma ) (1 + r) + x0( gamma ) w > p( j )
A long run equilibrium is defined by
o At least one process in each industry is operated.
o No process can pay extra profits.
o No process in operation incurs extra costs.
One can use this definition to find the technique(s) that will be
operated at any rate of profit r. Assuming something like viability
conditions this technique will exist. For each chosen technique, the
following price equations will hold:
p A (1 + r) + a0 w = p
Given r, p and w will be unique (even though the technique may not be).
Since no process can pay extra profits, the technique(s) is cost-minimizing.
It's also the case that given r, the cost minimizing technique will
maximize the wage.
In review, I have outlined how given net ouput, the technology (the set
of processes), and either w or r; then prices, the other distributive
variable, and gross outputs are determined. Notice that prices are
indpendent of y; they depend only on one or the other of w or r. (This
is the "non-substitution" theorem.) Changes in demand for final outputs
cannot affect prices except as they affect w or r. But as r varies,
the price of a produced output need not vary as in well-behaved
substitution relationships (see the Woods example of a couple of
posts ago). Nor need labor-intensity or capital-intensity vary in
the proper manner with w or r.
What happens if we relax the assumptions of the non-substitution
theorem, say by considering fixed capital, land, and non-constant
returns to scale? Then variations in the composition of net output
will influence prices. But we still may find a higher price of
a produced input may be associated with a choice of a technique
that uses input more intensively. And the "perversities" in the
labor and capital markets may still exist.
So there's a general theory in which long run equilibrium prices
are not scarcity indices and in which substitution relationships
need not characterize comparisions of steady states.
If we want to call "Neoclassical" a theory in which these "perversities"
cannot exist, then the long run Neoclassical theory of production is
a special case of the above. It would need additional assumptions, and
some economist should state what the assumptions are. The neoclassical
model would close this theory of production by assumptions about utility
maximization, although other non-neoclassical models exist. Perhaps its
sufficient for neoclassical theory to close this general model of
production with assumptions about utility maximization. But then
neoclassical economists should be taught that their intuition about
substitution does not apply to long run equilibria.
This is all textbook stuff.
Markku Stenborg
[snip yet another boring example]
>In review, I have outlined how given net ouput, the technology (the set
>of processes), and either w or r; then prices, the other distributive
>variable, and gross outputs are determined. Notice that prices are
>indpendent of y; they depend only on one or the other of w or r. (This
>is the "non-substitution" theorem.) Changes in demand for final outputs
>cannot affect prices except as they affect w or r. But as r varies,
And this is a property you do not find curious?
>the price of a produced output need not vary as in well-behaved
>substitution relationships (see the Woods example of a couple of
>posts ago). Nor need labor-intensity or capital-intensity vary in
>the proper manner with w or r.
There's no y in the example above?
So, the model is still (a) based at least on disparity in the timing of
payments which then brings these interest charges and (b) under determined in
that from exogenous data and behavior rules you cannot determine all
non-exogenous variables, such as all (relative) prices, profits and other
incomes?
>What happens if we relax the assumptions of the non-substitution
>theorem, say by considering fixed capital, land, and non-constant
>returns to scale? Then variations in the composition of net output
>will influence prices. But we still may find a higher price of
>a produced input may be associated with a choice of a technique
>that uses input more intensively. And the "perversities" in the
>labor and capital markets may still exist.
They are hardly perversities, since here w is not really the (marginal) price
of labor as in more std neoclass models.
>So there's a general theory in which long run equilibrium prices
>are not scarcity indices and in which substitution relationships
>need not characterize comparisions of steady states.
Your w and r are not the prices of marginal labor and capital as they are in
textbook econ models, so this is not a critique of conclusions presented in
those models. The marginal price of labor also needs to have these interest
charges included. Now do the comp stat w/ this wage variable, nor w/ w alone.
>If we want to call "Neoclassical" a theory in which these "perversities"
>cannot exist, then the long run Neoclassical theory of production is
Why would we want to have this characterization for "neoclass"?
>a special case of the above. It would need additional assumptions, and
No. It seems all I need to do is to *drop* the assumption that wages are payed
before payments for final goods are received. This would remove these interest
charges, and leave only the neoclass results -- where some other effects still
might more than balance substitution.
>some economist should state what the assumptions are. The neoclassical
>model would close this theory of production by assumptions about utility
>maximization, although other non-neoclassical models exist. Perhaps its
I would close this by leaving no variable undetermined.
>sufficient for neoclassical theory to close this general model of
>production with assumptions about utility maximization. But then
>neoclassical economists should be taught that their intuition about
>substitution does not apply to long run equilibria.
>
>This is all textbook stuff.
This is all textbook stuff.
Markku Stenborg
>This is POST 2 in my series of hopefully shorter points. I don't think
>Markku has the mathematical apparatus to answer the questions about
>generalizations here, although I assume he could understand the
>references I previously gave.
I didn't think the issue of this thread was my ability to read math?
[snipped boring example]
>I hope this shows Markku more about the relationship between my first
>example and the Samuelson example. I still claim his characterizations
>of my arithmetic are bizarre. The above will make more apparent why.
My characterizations of your arithmetic? I don't think I have uttered a
syllable on arithmetic?
I've characterized your example as "ruling out substitution by assumption",
"under determined", "based on disparity b/w payments". None of these refer to
any issue re: arithmetic.
Markku Stenborg
>(Markku Stenborg) wrote:
>
>> Result that states "... may ..." is rather weak. I can recall one instance
>> where this "... may ..." stuff was publised paper. Usually, propositions
take
>> the form "Given ..., the ... follow". I though you had something along
these
>> lines, esp. since this was but one example based on Leontieff prod fn,
>> time-to-build techniques but no contracts b/w byuers and sellers, etc, not
a
>> theory.
>
>POST I in this series.
>
>In 1965 David Levhari, a MIT economist, published, under encouragement
>from Paul Samuelson, a theorem of the form Markku approves of: p -> q.
This the usual form for theorems, irrespective of my humble approval or lack
there of. The other usual form is that p is necessary and sufficient for q.
>The q was "reswitching cannot happen." The theorem was mistaken and
>Levhari's proof was in error, something like writing "less than" when
>he should have written "less than or equal to" or vice versa.
>
>A number of economists quickly created counterexamples to the
>theorem and pointed out the mistake. The Samuelson example was part
>of Paul Samuelson's concession that the critics were correct.
>
>Now what would a counter example look like? Obviously it would have
>p and not q. This is not equivalent to showing that p implies not
>q. Rather it only shows that if p happens, q may happen.
In this simple logic, the counterexample for (p => q) is [p and (not q)]? You
must mean: if p happens, q need not happen?
>Now when one has a result that some may find surprising, one might
>spend some time constructing examples to better understand it. How
>many more conditions can one add to p and still have the surprising
>result? What other conclusions can be drawn? Can examples be constructed
>that have some of these conclusions and not others? This sort of
>"exploratory mathematics" is a typical heuristic approach for suggesting
>theorems. Markku is correct that a lot of this is rarely published
All very true.
But from an example or even from a collection of examples I would not draw
sweeping conclusions like "prices are not scarcity indexes, substitution does
not occur in long term", etc. Nor would I draw the opposite conclusion from
examples of similar quality.
The example(s) you've been constructing are rigged for the conclusion you want
to establish, so they are neither surprising nor do they lend much support for
the conclusions.
If you want to criticize neo class econ for over emphasizing substitution, it
should be done in a model that at least allows substitution. Your Leontieff
prod fn models allow a minimal substitution b/w labor and kapital b/c you have
two kinks instead of one, the model seems to be under determined, etc.
Somebody w/o your convictions is unlikely to be impressed by the fact that not
much substitution occurs in model where there's not much room for
substitution.
>(see Lakatos 1976), though I wonder if he can come up with some examples
>from General Equilibrium that led economists towards believing that
>the assumption of gross substitutability could not be easily relaxed without
>loss of some desirable properties.
>
>Finally, consider two models A and B. B contains all of A's assumption.
>B also has some more assumptions. Then A is the general model and
>B is the special case. One might be able to exhibit examples of A
>that are not B. That is, one might be able to exhibit
>behavior that may occur in A, but not B. In this case, one cannot say
>A contains any "crucial assumptions" implying this behavior. Rather
>B has the "crucial assumptions" (the additional special case assumptions)
>that imply the interesting behavior is impossible.
True. Which assumptions then rule out substitution?
>I don't know of any adequate special case neoclassical model of long run
>equilibrium in which substitution is guaranteed to characterize comparisons
The special cases are too easy to construct: make sure all other effects are
small enough, so that nothing dominates substitution. But this is really
boring.
>of long run equilibrium. I know of some work which I consider unsatisfactory.
Again you resort to this silly innuendo "I know of some work which I consider
unsatisfactory." I also know plenty work I consider unsatisfactory. The
interesting questions are Which work? The reason for their unsatisfactoryness
is what?
Robert Vienneau
In article <5ebtqa$c...@idefix.eunet.fi>, real.a...@bottom.of.msg
(Markku Stenborg) wrote:
> O 'ear ye mortals, the infinite wisdom by Robert Vienneau of
> rv...@dreamscape.com ...
>
> [snip yet another boring example]
>
> >In review, I have outlined how given net ouput, the technology (the set
> >of processes), and either w or r; then prices, the other distributive
> >variable, and gross outputs are determined. Notice that prices are
> >indpendent of y; they depend only on one or the other of w or r. (This
> >is the "non-substitution" theorem.) Changes in demand for final outputs
> >cannot affect prices except as they affect w or r. But as r varies,
>
> And this is a property you do not find curious?
No, because I've thought about it. But I thought you would. That's why
I've asked you to summarize your understanding of this theorem several
times. Several economists were surprised by it when this theorem was
first formulated in the 50s.
To help you out, here are a couple of mathematical facts. The
nonsubstitution theorem still holds when the production functions in
each industry are well-behaved and continuously differentiable. And
it does not depend on the choice between the price equations discussed
below; it holds in both cases.
So how do you account for this theorem?
> >the price of a produced output need not vary as in well-behaved
> >substitution relationships (see the Woods example of a couple of
> >posts ago). Nor need labor-intensity or capital-intensity vary in
> >the proper manner with w or r.
>
> There's no y in the example above?
The point of the theorem is that price and quantity relationships
are independent. So I can consider variations in prices, the
choice of technique, and the labor-intensity of a vertically
integrated industry with variations in, say, w, without caring
about the composition of final demand or changes in this composition.
> So, the model is still (a) based at least on disparity in the timing of
> payments which then brings these interest charges and...
As far as (a) goes, I think you are mistaken. My price equations were
p A (1 + r) + a0 w = p (*)
This system shows wages being paid at the end of the year/day/whatever
at the same time that the final product is sold. So wages are paid
simultaneously with the receipt of profits. Physical inputs other
than labor, though, are paid for at the beginning of the period.
Now the equation you seem to want is:
(p A + a0 w) (1 + r) = p (**)
I just use (*) as a matter of tradition. It's not a matter of principle,
and my results go through. My first example and the Samuelson example
continue to be "perverse" in the labor and capital markets if you
assume (**).
By the way, my first example, I remind you does show the same set of
goods being used as inputs, but in different proportions, in both
processes available in the steel industry. The number of available
processes can be expanded to any finite number with all processes
using the same set of inputs (in varying proportions), and reswitching
is still possible. The analysis shows which technique the cost-minimizing
firm will choose. This can be cast in the form of a Linear Program,
which is a traditional method to analyze the choice of activity
levels. How have I ruled out substitution by assumption? And what
are the effects that override substitution?
Furthermore, the claim that "reverse capital deepening" and
"perverse" behavior in the labor market are compatible with
continuously differentiable production functions is independent
of the specification of (*) or (**).
And the association of a lower price of a produced input with
a cost-minimizing technique that uses that input less-intensely
can arise under (**), as well as (*). I don't know how the particular
Woods example works out though, having not bothered to sit down
and do the calculations.
Finally, all valid marginal productivity productivity relationships
are satisfied in whichever equation you assume. Hahn showed this
for (**) in the paper to which D. and Levy were responding, and Hahn
seemed to agree no deep matter of principle was involved.
So there go your claims that the perverse behavior depends on "disparity
in the timing of payments" and that if I used w as in the traditional
marginal productivity theory, these perversities would necessarily
vanish.
> ...(b) under determined in
> that from exogenous data and behavior rules you cannot determine all
> non-exogenous variables, such as all (relative) prices, profits and other
> incomes?
As for (b), the model is complete in that all endogeneous variables
are determined (except one has more than one cost-minimizing technique
at a switch point). It's just that some economists take as exogeneous
either w or r, a decision I'm sure you don't think highly of. Are
there any variables that are treated as endogeneous in my 3-part
presentation that you think should be exogeneous data?
If I wanted to take both distributive variables as endogeneous...
There are several non-neoclassical methods of closing the model.
And, once again, I have given you references to a couple of
empirical comparisons of these other models with a neoclassical
model.
[snip]
> >If we want to call "Neoclassical" a theory in which these "perversities"
> >cannot exist, then the long run Neoclassical theory of production is
>
> Why would we want to have this characterization for "neoclass"?
So as to have a theory that behaves in accordance with your intuition
on substitution behavior. Your intuition has a historical basis in that
tradition which is normally labeled "neoclassical." If you cannot find
a long run theory that behaves appropriately, perhaps you should modify
your intuition accordingly.
I did deliberately, though, state that as "if" so as to allow
you to argue as you pleased.
> >a special case of the above. It would need additional assumptions, and
>
> No. It seems all I need to do is to *drop* the assumption that wages are
payed
> before payments for final goods are received. This would remove these
interest
> charges, and leave only the neoclass results -- where some other effects
still
> might more than balance substitution.
No. See above.
> >some economist should state what the assumptions are. The neoclassical
> >model would close this theory of production by assumptions about utility
> >maximization, although other non-neoclassical models exist. Perhaps its
>
> I would close this by leaving no variable undetermined.
But not by introducing intertemporal utility maximization? Care to
expand?
> >sufficient for neoclassical theory to close this general model of
> >production with assumptions about utility maximization. But then
> >neoclassical economists should be taught that their intuition about
> >substitution does not apply to long run equilibria.
> >
> >This is all textbook stuff.
>
> This is all textbook stuff.
My comments really are textbook stuff in that they are summaries of
the textbooks that I have already referenced.
Bored with the thread yet? Is this a fair summary:
I think that the general theory of production in long run equilibrium
does not behave in accordance with your intuitions on substitution.
It's logic is simply not a matter of substitution behavior. Accordingly,
these "perversities" I have been pointing out do not need to be
explained. What does need to be explained is what special case
assumptions need to be introduced to get a long run equilibrium
model to behave in accordance with your intuition on substitution.
You seem to think that substitution is so natural as to not need
explaining. What, according to you, does need explaining is what
overrides substitution in my examples. You are not sure that you
have found the cause, but think that I should do the explaining
since I created the examples. You probably think I'm being
obstinate in not explaining what goes on in non-mathematical terms.
Both of us seem to think that we have the stronger mathematical
case. I think your position would be enhanced if you didn't respond
as if you were seeing some of this for the first time. I've given
you textbook references. But apparently you don't find the economic
issues of sufficient interest to pursue them. And my questions about
generalizations to matrix formulations haven't convinced you
that there's any cool math here either.
Markku Stenborg
O 'ear ye mortals, the infinite wisdom by Robert Vienneau of
rv...@dreamscape.com ...
>
>In article <5ebtqa$c...@idefix.eunet.fi>, real.a...@bottom.of.msg
>(Markku Stenborg) wrote:
>
>> O 'ear ye mortals, the infinite wisdom by Robert Vienneau of
>> rv...@dreamscape.com ...
>>
>> [snip yet another boring example]
>>
>> >In review, I have outlined how given net ouput, the technology (the set
>> >of processes), and either w or r; then prices, the other distributive
>> >variable, and gross outputs are determined. Notice that prices are
>> >indpendent of y; they depend only on one or the other of w or r. (This
>> >is the "non-substitution" theorem.) Changes in demand for final outputs
>> >cannot affect prices except as they affect w or r. But as r varies,
>>
>> And this is a property you do not find curious?
>
>No, because I've thought about it. But I thought you would. That's why
>I've asked you to summarize your understanding of this theorem several
>times. Several economists were surprised by it when this theorem was
>first formulated in the 50s.
Again, I don't find the results "perverse" or surprising b/c you've rigged the
model to not to generate substitution at all.
Further, you've defined "substitution" as a reaction to spot prices, while the
marginal costs or prices, or the relative prices, relevant for decisions in
this economy are not the spot prices but rather the ones w/ the interest
charges added. So, substitution or the lack thereof should be defined relative
to these true marginal prices, not relative to spot prices.
And most if not all economists since 50s have found something more relevant to
do w/ their time than to fool around w/ these math possibilities, which, I
presume, accounts for the lack of these types of stories or even references to
these.
Besides, neither I nor neoclass econ have never ever indicated that
substitution is the only game in town.
>To help you out, here are a couple of mathematical facts. The
>nonsubstitution theorem still holds when the production functions in
>each industry are well-behaved and continuously differentiable. And
>it does not depend on the choice between the price equations discussed
>below; it holds in both cases.
>
>So how do you account for this theorem?
As I account Giffen goods, say. [You don't say you see Giffen goods daily?] It
a mathematical possibility, but its econ interest and relevance is at the
level of "Yawn, somebody wake me up if the opera ends".
And see above.
>> >the price of a produced output need not vary as in well-behaved
>> >substitution relationships (see the Woods example of a couple of
>> >posts ago). Nor need labor-intensity or capital-intensity vary in
>> >the proper manner with w or r.
>>
>> There's no y in the example above?
Umm, you said "prices are indpendent of y" w/o defining or explaining what
that y is. Income? Output?
>The point of the theorem is that price and quantity relationships
>are independent. So I can consider variations in prices, the
Pretty much my point, if you mean that price and quantity relationships are
independent of each other.
>choice of technique, and the labor-intensity of a vertically
>integrated industry with variations in, say, w, without caring
>about the composition of final demand or changes in this composition.
>
>> So, the model is still (a) based at least on disparity in the timing of
>> payments which then brings these interest charges and...
>
>As far as (a) goes, I think you are mistaken. My price equations were
>
> p A (1 + r) + a0 w = p (*)
>
>This system shows wages being paid at the end of the year/day/whatever
>at the same time that the final product is sold. So wages are paid
>simultaneously with the receipt of profits. Physical inputs other
>than labor, though, are paid for at the beginning of the period.
I thought you've said that wages are payed before final products are sold? At
least one of your examples was based on labor working for 3 yrs before the
output was sold, and the interest rate was 50%.
>Now the equation you seem to want is:
>
> (p A + a0 w) (1 + r) = p (**)
[snip]
So why do keep using Leontieff then? If it has no bearing on results while
simultaneously creating the impression that input substitution is all but
ruled out by assumption in a model that tries to argue that substitution does
not characterize long-term behavior.
>And the association of a lower price of a produced input with
>a cost-minimizing technique that uses that input less-intensely
>can arise under (**), as well as (*). I don't know how the particular
>Woods example works out though, having not bothered to sit down
>and do the calculations.
Out of few examples you reach sweeping generalizations w/o even understanding
how your examples work?
>Finally, all valid marginal productivity productivity relationships
>are satisfied in whichever equation you assume. Hahn showed this
>for (**) in the paper to which D. and Levy were responding, and Hahn
>seemed to agree no deep matter of principle was involved.
>
>So there go your claims that the perverse behavior depends on "disparity
>in the timing of payments" and that if I used w as in the traditional
>marginal productivity theory, these perversities would necessarily
>vanish.
Umm, you're the one who claimed that the result is due to interest charges on
dated labor inputs. I only pointed out that, in this economy, these are
equivalent to disparities in timing of payments.
>> ...(b) under determined in
>> that from exogenous data and behavior rules you cannot determine all
>> non-exogenous variables, such as all (relative) prices, profits and other
>> incomes?
>
>As for (b), the model is complete in that all endogeneous variables
>are determined (except one has more than one cost-minimizing technique
>at a switch point). It's just that some economists take as exogeneous
>either w or r, a decision I'm sure you don't think highly of. Are
Yep. If the point of the model is to look at input choices in a GE model, I
find it curious to keep some spot prices exogenous.
>there any variables that are treated as endogeneous in my 3-part
>presentation that you think should be exogeneous data?
No, the conventional wisdom is that the less stuff is kept exogenous the
better the model.
>If I wanted to take both distributive variables as endogeneous...
>There are several non-neoclassical methods of closing the model.
>And, once again, I have given you references to a couple of
>empirical comparisons of these other models with a neoclassical
>model.
Care to elaborate? Which is the "neoclassical model"?
>[snip]
>
>> >If we want to call "Neoclassical" a theory in which these "perversities"
>> >cannot exist, then the long run Neoclassical theory of production is
>>
>> Why would we want to have this characterization for "neoclass"?
>
>So as to have a theory that behaves in accordance with your intuition
>on substitution behavior. Your intuition has a historical basis in that
Which intuition? That substitution can occur? That is is not crazy to
substitute more expensive good by less expensive ones? That there are other
things that work to opposite direction to substitution?
>tradition which is normally labeled "neoclassical." If you cannot find
>a long run theory that behaves appropriately, perhaps you should modify
>your intuition accordingly.
But I do change my intuition w/ the results of the models. I have never ever
claimed that substitution is the only possibility.
Besides, the only "long-run" in the examples you've presented is the presence
of interest. There are no financial mkts, no saving, no intertemporal stuff.
The rate of interest seems to pop down from the sky, ie, it is neither
determined in the model nor based on intertemporal discounting. You find a
model along these lines to be an appropriate representation of issues involved
in long run econ?
[snip]
>> >some economist should state what the assumptions are. The neoclassical
>> >model would close this theory of production by assumptions about utility
>> >maximization, although other non-neoclassical models exist. Perhaps its
>>
>> I would close this by leaving no variable undetermined.
>
>But not by introducing intertemporal utility maximization? Care to
>expand?
If the model is intertemporal, why would you rule out intertemporal utility
maximization?
[snip]
>Bored with the thread yet? Is this a fair summary:
Not only w/ the thread, but the whole issue and models seem so boring I find
it hard to believe these are the things that make you tick.
>I think that the general theory of production in long run equilibrium
>does not behave in accordance with your intuitions on substitution.
Your simple examples are a far cry from general theory.
>It's logic is simply not a matter of substitution behavior. Accordingly,
>these "perversities" I have been pointing out do not need to be
>explained. What does need to be explained is what special case
>assumptions need to be introduced to get a long run equilibrium
>model to behave in accordance with your intuition on substitution.
>
>You seem to think that substitution is so natural as to not need
>explaining. What, according to you, does need explaining is what
>overrides substitution in my examples. You are not sure that you
No. Substitution seems to pop out of models regardless of the modelling
details, unless ruled out by assumption. The idea, though, is so natural to
even a complete ignorasmus of econ that it shouldn't need explanations: you
want to use less of something that becomes relatively dearer, and more of what
becomes relatively cheaper. Many a non-economist needs explanations for
effects that work against substituition. Many an economist needs explanations
for, or at least mentioned, the forces that work against substitution.
>have found the cause, but think that I should do the explaining
>since I created the examples. You probably think I'm being
>obstinate in not explaining what goes on in non-mathematical terms.
You should have the relative advantage in explaining the model, both w/ and
w/o math terms since you're the one who's been thinking about it. This is the
std requirement in econ seminars for instance. The other participants are not
supposed to explain your example to you.
>Both of us seem to think that we have the stronger mathematical
>case. I think your position would be enhanced if you didn't respond
I believe I have the stronger *econ* case.
>as if you were seeing some of this for the first time. I've given
No, the question is your understanding of the issues involved.
>you textbook references. But apparently you don't find the economic
>issues of sufficient interest to pursue them. And my questions about
Yep. This seems really boring stuff which is probaly the explanation why other
economists have substituted away from this to more interesting threads.
>generalizations to matrix formulations haven't convinced you
>that there's any cool math here either.
________
Robert Vienneau
I don't care to continue this thread since I find it difficult to
believe that Markku is even reading all of my responses.
Markku Stenborg
>believe that Markku is even reading all of my responses.
Whatever. As good of an exuse as any other.