James,

I am perfectly willing to defend the use of AS and AD without 
deriving either from ISLM.  Alan Isaac is correct that these 
ideas lie at the heart of modern time series modeling of a 
very mathematical sort used by policymakers and 
theoreticians, although often under different terminology, 
e.g. "permanent" versus "transitory" changes.

In any case, while Walrasian general equilibrium is a 
sufficient condition for general macroeconomic equilibrium of 
the AS = AD sort, it is not necessary.  Surpluses in one 
sector can be offset by shortages in another.  Likewise, one 
can have AS = AD even without IS = LM.  One simply does not 
need to go through all those horrifying contortions necessary 
to pull the AD out of ISLM.  AD is a perfectly fine empirical 
and within certain limits, theoretical construct.

Michal P.

Well, I agree that much of the appeal at the Principles level 
of AS/AD is pedagogical in the sense of its familiarity to 
students.  I would hope that people who use it do note its 
differences with what goes on at the micro level, although I 
have no doubt the majority of students do not get it.

Of course, as I have been pointing out, the micro supply and 
demand framework has some very serious theoretical and 
empirical problems, especially on the demand side.  However, 
I recognize that it is not unreasonable to assume empirical 
short run downward-sloping demand curves for pedagogical 
reasons, even though the theoretical underpinnings have all 
gone rotten, a topic you all are studiously avoiding.  
Frankly, AD is not any worse than micro D.  It is just that 
most economists have not become aware of this trenchantly.

David,

Well, as the textbook expert, I appreciate that you are clear 
that different things are more suitable at different levels.  
ISLM hangs on in the intermediate textbooks, AS/AD hangs on 
in the Principles textbooks.  What is going on in the grad 
ones is very much under reconstruction, and not as certain as 
what some on this list think it should be.  A major problem 
is indeed the math issue, so that the more dynamic Samuelson 
formulation of ISLM cannot be taught at the intermediate 
level, and the more technical version of time series AS/AD is 
not going to get taught at the Principles level.

Of course in the first edition of your Principles text you 
had a delightful figure showing a vertical AS and a vertical 
AD that did not intersect.  Somewhat before then I published 
a paper ("Indeterminacy of Macroeconomic Equilibrium in a 
Post Keynesian New Classical Model" Journal of Post Keynesian 
Economics, Fall 1991, vol. 14, no. 1, pp. 111-116) that came 
up with the same argument, but had the two lying right on top 
of each other.  The basic argument is that something like 
this is what one will get if one really rigorously carries 
through on the classical logic.  To be able to say or do 
anything about the macroeconomy one has to step back somewhat 
somewhere from a pure classical model.

Robert Leeson,

I know that you are all worked up about this business of 
positing a "government decreeing the price level to be cut in 
half while holding the money supply constant," but this is 
really a distortion of what we all do.  Please, I remind you, 
we have exactly the same problem at the micro level, and I do 
not see you or anybody else up in arms about it.  We make 
ceteris paribus assumptions and then draw these supply and 
demand curves at whatever level.  There is in both cases an 
unavoidable unreality about the whole exercise.

If one wishes to justify it, the way to do so is as I 
suggested earlier, but which you are kindly ignoring: one 
assumes that one is always in equilibrium in a reasonably 
competitive economy.  Therefore, in drawing one curve, one 
assumes the other one is shifting, with all else constant.  
Unrealistic?  Yep.  But maybe not always, as we can talk 
about a sudden backward shift of AS from an oil price shock 
and talk about moving along a relatively motionless AD curve, 
and come up with a story that fits what we observe in the 
real world, even if all kinds of things are being held 
constant and all kinds of classical assumptions are not 
holding.

Steve Kates and Pat Gunning,

OK, I may be willing to buy that Keynes invented AS and AD 
analysis.  Good for him.  

I would ask you, and Pat Gunning also, if you eschew talking 
about AD, then how do you explain the Great Depression?  Did 
the natural rate of unemployment suddenly fall?  Even if Pat 
is right that Germany just bounced back on its own after 
1932, how and why did it fall so far down in the first 
place?  Milton Friedman says it was monetary contraction, and 
I think that was a lot of it also.  But to get from that to a 
collapse of real output involves most easily assuming that 
this decline of M pushed back the AD curve along a non-
vertical AS curve.  What is the classical explanation again?  
Are we going to hear about lazy workers not willing to accept 
wage cuts?  Please.

Also, Steve, when are you going to fess up that Say did not 
always accept his own "law"/

Kevin,

Presumably you are not writing a Principles text.  Good luck, 
whatever.

Warren Young and Alan Isaac,

Thank you for your useful comments.


Barkley Rosser