----------------- HES POSTING ----------------- 
I share Prof. Lipsey's concerns about US Keynesians using the Y=3DE diagram to derive AD. 
 
The 45-degree diagram captures a bit of the Keynesian production dynamics. First, it
emphasises that actual production takes place a period ahead, on the basis of anticipated
demand for the following period. This is a typical Keynesian feature in contrast to the
standard Walrasian model of exchange where actual production and exchange take place
simultaneously after the auctioneer announces the equilibrium prices. Second, if the sales
expectations are not realised (often possible), firms adjust quantities (from inventories)
instead of prices (as in the Walrasian markets) in the next period. Inventories give
signals for adjustments in the Keynesian model. In contrast, such signals are given by
prices in the Walrasian markets. What is important here is a discussion of the length of
these two unit time periods e.g., the Keynesian unit time is much shorter (up to a quarter
or two perhaps) than the Walrasian unit time  (say of one or two years) over which price
adjustments take place. Third, in the following period (period three), firms revise their
sales expectations and change their production plans. And this process continues until
expected demand and production equal. What is missing here is: when firms may also decide
to change prices? The Phillips curve and the mark-up principle provide some answers.
 
The modified DF model (see my previous post) is a sort of synthesis (with both quantity
and price adjustments) and gives the impression, perhaps unacceptable to the pure
Keynesians, that Keynesian results can be obtained by modifying the Walrasian GE model. In
my view, before the DF model is taught, a purely descriptive account of the differences
between the Keynesian and Walrasian systems may help students.
 
Perhaps the real world consists of both the Keynesian quantity adjustment and the
Walrasian price adjustment markets. Alternatively both quantity and price adjustments take
place  (in the unit time period) may be treated as a simplifying assumption. The question
is: What proportion of the total transactions takes place in such markets? satisfactory
answers could throw light on the relevance of the Walrasian and Keynesian approaches to
model the macro economy.
 
In a rudimentary attempt (Rao, 1993) it was found that more than 85% of the U.S. GNP
transactions take place in the Keynesian markets.  If this is further corroborated, the
standard ISLM plus the Phillips curve would be adequate for the principles texts. The DF
model is a simple diagrammatic variant of this approach. Their AS curve is the commodity
market Phillips curve in the P-Y space instead of the inflation-output space.  May I say
that it would be useful to replace or add to the current complicated diagrammatic
expositions, simple numerical dynamic simulations to explain the split between quantity
and price changes, for a given change in the so called AD. Perhaps the new generation of
textbook writers would consider my suggestion and strengthen the structure and relevance
of the macro models, instead of perpetuating confusions between the Keynesian and
Walrasian approaches.
 
 
Reference: 
 
Rao, B.Bhaskara., (1995) "The Nature of Transactions in the U.S. Aggregate Goods Market",
Economics Letters, pp.385-390.
 
 
 
------------ FOOTER TO HES POSTING ------------ 
For information, send the message "info HES" to [log in to unmask]