> These discussions with Chris always leave me wondering if he really
> thinks his debating tactics are fair. I'll point out below some
> curious editing and misreadings.


In Rob's opinion, is the unstated caveat.  I am always happy to let
readers decide for themselves if my editing is "fair."  I have, for
instance, edited liberally here.

> On the other hand, I find his reactions to mine curious. If I say
> Chris is mistaken about X or seems not to understand Y, I don't
> think I am making a "snide little dig." Nor am I challenging him
> to list those areas of economics that he does understand.
> Furthermore, if I suggest an improvement of phrasing for some
> of his statements, I think I am disagreeing in a nuanced and
> courteous fashion. His reactions seem not to conform to such
> standards of politeness.


I wonder if Rob really thinks relentless accusations of 
misunderstanding, ignorance, and technical error really constitute 
"nuanced and courteous" discussion.

>>> Chris' insistence on not addressing the text to which he is responding
>>> is his problem. Notice that Chris doesn't even attempt to provide any
>>> evidence that I claimed to demonstrate the possible existence of
>>> upward-sloping labor demand curves. 

>>> The title of this thread and my choice of newsgroups was designed to
>>> achieve a certain audience.


>> The second quoted paragraph adequately refutes the first.  Rob, if you
>> don't want to claim you're showing the existence of upward-sloping labour
>> demand curves, don't use that terminology.  Quite simple, really.

> I wanted to point out some literature that uses that terminology. I wanted
> to do this in a way that brought this literature to the attention of
> readers that had seen some of my previous discussions on this
> topic. That seems an adequate reason for the title.


I went back and reread one of Rob's oft-posted essays.  Rob, please do
explain 1) the title of this thread, 2) comments like

> What, then, is the rational basis for assuming downward-sloping
> labor demand curves?


and 3) the continual accusation that economists teach "exploded dogma"
when they convey the idea that input demand scedules slope down, if
you weren't trying to argue that labour demand curves can slope up.
I see no reason to be polite here: Rob knows damn well he's abusing
terminology (I explained it to him before), and now he's trying to
weasel out of his disingenuousness.

> Saying that L( w, r ) is a labor demand function is not an outline
> of how to construct the curve. Chris did say the derivation of this
> curve comes from a profit-maximizing firm. But in my first post,
> I mentioned "equilibrium of the firm" and stated my opinion that
> this means that, in a long-run context, the firm should be on the
> "factor price frontier".


Rob has a lot of nerve accusing others of technical incompetence.
Here, he's confusing a firm's problem with market equilibria.  A
firm's labour demand schedule is derived without any assumptions
regarding how prices are determined -- if the firm is on its
demand function, then it is in "equlibrium" even if the market
is not.  In Rob's context, one can derive a "long run" input
demand schedule for the firm without any knowledge of the "factor
price frontier," and requiring that the economy be on that frontier
in the firm's problem implies that whatever solution falls out is
no longer an input demand function.

> We have not seen "baseless accusations of technical incompetence." I
> merely pointed out where Chris was mistaken in places where he was
> mistaken.


I suppose Rob forgot about:

>>> (2) Chris' comments tend to contain technical errors. In this post

>>> What is the firm's profit maximization problem in my example? 


>> It isn't there, which is one of the reasons Rob's story doesn't provide
>> enough structure to characterize general equilibrium.

> But I gave a labor-demand relation in the post to which Chris was
> responding.


No, Rob, you gave a *conditional* labour demand schedule.  Are you ever
going to grasp the distinction? ...

> Chris is always free to specify what additional structure he thinks
> he needs. My example can generate "this object," and I have
> generated it.


I guess not.  Rob, please look it up in the discussion of the theory
of the firm in any micro text.  In your problem, you need at least
the demand curve the firm faces -- I can't recall, but you have perfect
competition in this model, it's inconsistent with your constant returns
assumption.

> Since he provided this clarification, he should
> have noted I was asking him to outline how to construct conditional
> labor demand curves, if he did not think a labor demand curve could
> be constructed for my example.


Look it up in any micro text, Rob.

>>> In the opinion of some economists (Schefold 1990, Woods 1990),
>>> describing the effect I have been describing as an upward-sloping
>>> labor demand curve is legimate.


>> None of the text Rob has quoted contains such an assertion.


[ quotes from: ]

>        -- Schefold, 1990
>
>      --Woods, 1990


I don't believe Rob has quoted that text before.  If he did, I stand
corrected.  In any case, both these authors are abusing terminology,
and I would hope that somewhere in the text they acknowledge this.
If they do, then Rob is deliberately misrepresenting these pieces.
If they don't, these authors deserve to be chided, not unlike Rob,
for misusing basic terminology.

[ Much bickering deleted.  I am not, in particular, going to discuss
  dynamic economic modelling with Rob when he himself admits he isn't
  familiar with the literature, ie, the vast majority of economic
  theory developed in the past three decades.  ]


>> One can only see this as an attempt to divert the discussion.  Rob,
>> do you have any evidence whatsoever that endogenous changes in the
>> interest rate more than offset own-price induced changes in labour
>> demand when the minimum wage is increased?  Did you grasp my (deleted)
>> point that it is easy to control for this effect statistically? 

> I have always said that the effect illustrated in my example was meant
> to make a logical point.


Much like Rob's refusal to actually either defend or dismiss the idea
that labour demand curves can slope up, Rob doesn't seem to want to
either defend or dismiss the idea that the mechanism he presents is
responsible for counter-intuitive effects of a minimum wage.  Contrary
to his sentence above, his lengthy oft-posted essay uses the effects
of a minimum wage to motivate the discussion.  Well, Rob, are you going
to take a position or are you not?

> I objected to Chris' use of statistics to
> minimize the effects of a change in the level of the minimum wage. That's
> the context of this "attempt to divert the discussion."


Fine, then are you willing to defend the notion that endogenous changes
in the interest rate more than offset own-price induced changes in labour
demand when the minimum wage is increased?  If not, will you stop reposting
this essay every time the subject of minimum wages comes up in a thread?


>> Both "process alpha" and "process beta" describe Leontief production
>> functions.  The firm can choose process alpha or process beta.  I would
>> like Rob to explain why my statement is not entirely accurate.

> Sure. A Leontief production function is of the form:
> 
>    Q = min ( L/a0, X1/a1, ..., Xn/an )
> 
> Alternatively, one can define a Leontief production function as the
> solution of the following Linear Program:
> 
>     Max Q
>     such that
>          a0 Q <= L
>          a1 Q <= X1
>               .
>               .
>               .
>          an Q <= Xn
> 
>     where Q >= 0
> 
> Now consider my example. Let L, X1, and X2 be the labor, steel, and
> corn used by the steel sector as inputs. The production function in
> the steel sector in my example is the solution of the following Linear Program:
> 
>     Max Q = Q1 + Q2
>    such that
>          0.19321 Q1 - L1 <= 0
>          0.033594 Q2 - L2 <= 0
>             L1 + L2 <= L
>          0.35 Q1 - X11 <= 0
>          0.13329 Q2 - X12 <= 0
>            X11 + X12 <= X1
>          0.0095553 Q1 - X21 <= 0
>          0.15590 Q2 - X22 <= 0
>             X21 + X22 <= X2
> 
> 
>    where Q1 >= 0, Q2 >= 0, L1 >= 0, L2 >= 0, X11 >= 0
>          X12 >= 0, X21 >= 0, X22 >= 0
> 
> Q1 is the tons steel produced by the alpha process, and Q2 is the
> tons steel produced by the beta process. Notice that L, X1, and X2
> are parameters, not decision variables found by solving this LP.
> Thus, one can express the solution value of the objective function
> as a function of these parameters:
> 
>                 Q = f( L, X1, X2 )
> 
> Since this second LP is not of the same form as the first, the
> production function f( L, X1, X2 ) is not Leontief.


Well, first, 

 Q = f( L, X1, X2) 

is, in fact, a special case of

        Q = min ( L/a0, X1/a1, ..., Xn/an )

so Rob's pointlessly elaborate "proof" doesn't cut it.  But it is 
also intending to prove wrong an assertion I never made -- I said 
the technology here can be described as "a choice between two 
Leontief production functions,"  *not* "a Leontief production 
function."


>> Rob's context -- he hasn't provided the information necessary.  Perhaps
>> if he reflects on the difference between "labour demand schedule" and
>> "conditional labour demand schedule" which I've attempted to hammer 
>> home he'll realize that.

> But I have constructed that schedule in my context. Besides, if it's
> so difficult to construct, how did he know "this object is an ungainly
> step function."


1) There is no labour demand schedule in Rob's example, there is a 
conditional labour demand schedule.  2) I know it's an step function
because the isoquants, given the technology, are kinked -- there will
be switch points as the price ratio passes the relevant thresholds.
I refuse, however, to pull out a calculator and work through Rob's
unwieldy numerical example to present those thresholds, nor do I see
any point in his insistence that I do so.

> I like that Chris seems to find it awkward to refer to it without
> using the phrase "labor demand function."


And yet I don't use that term because doing so would be incorrect and
misleading.  Many models generate relationships which superficially
look like defined theoretical objects but we don't just misapply 
jargon right and left to make the text a little shorter -- that would
hinder rather than aid communication.

> I don't understand Chris'
> statement, but I wonder how he manages to distinguish long-run and
> short-run effects in his econometrics.


Neither do I, in this context.  That's one reason the explicitly dynamic
models the profession has long since adapted are superior to the ad hoc
"dynamics" in the old models Rob continually mischaracterizes as mainstream
tools.

The econometric point is really quite simple.  Ignoring endogeneity and
other potential econometric problems, if we estimate a relationship such
as 

 labour =  constant + b(wage) + (other stuff) + residuals

then the paramter b is the change in labour induced by a change in the
wage rate holding "other stuff" equal.  If "other stuff" does not include
prices of other inputs which vary endogenously with the wage rate, then
b picks up all these endogenous second-order effects as well.  It is then
the slope that can be positive in Rob's model.  If however, "other stuff"
includes the prices of other inputs (such as the interest rate/rental
rate of capital), then b is the impact of a change in the wage rate on
labour demanded holding all those other prices fixed.  It is then 
purged of changes in the interest rate or other prices induced by changes
in the wage, and theory unambiguously predicts it should be negative.


>> Quite true.  As a matter of historical record, Rob is quite free to
>> repost his essay another couple of hundred times.  No one will stop
>> him, although others may occasionally note its errors and lack of 
>> relevance to modern economic thought.

> I haven't seen anybody note any errors in that essay. It doesn't
> surprise me that within a couple of posts after agreeing that my
> numerical example is correct, Chris is now insinuating that it
> is mistaken.


I am?  In what way?  Have I not pointed out other assertions I 
believe are erroneous?

> Anybody reading this thread might guess other reasons for Chris'
> "frustation" than those he mentioned. He might as well declare
> victory and quit.


Grand idea that.  If Rob has nothing interesting to say in response
to this note , I don't see much point in continuing the discussion.  I
think it has been very well established that Rob does not show that
labour demand curves can slope up.  If he starts reposting his essay
again under the current title, or including the jabs about "exploded
dogma" and the like, I'll make a habit of reposting my initial 
rejoinder in this thread.