Well, it is probably a waste of everybody's time to get too far
off into the actual discussion of whether or not either the multiplier
or AD analysis is worthwhile or anything more than a myth.  But,
these rather adamantine statements by both Kates and Gunning
do call for some kind of response.
      So, regarding AD, well, Fred Foldvary notes that even if one
is very classical, one can get a downward-sloping AD curve simply
by positing a fixed MV, which then gives a hyperbolic AD curve.
One can then shift it due to either a change in money supply, or
if one is willing to accept that fiscal policy can increase V (not a
traditional classical view, but V certainly has been very unstable
since the mid-1980s for the US economy) then via fiscal policy.
      Of course if one argues that the AS is vertical at the level of
real output associated with the natural rate of unemployment, then
all these AD changes are merely affecting the price level.  But
the experience of the 1990s suggests that the natural rate is at
best endogenous, and certainly does not equal the NAIRU, as
originally argued by Friedman, if there is even such a thing as
the NAIRU.  I realize that it is now widely entrenched in many
textbooks, especially above the Principles level, to assert that
the AS is vertical at the natural rate of unemployment, which then
is asserted to equal the NAIRU.  But the evidence in reality for
this is pretty pathetic.  (I would also note that Keynes himself in
Chap. 21 of the GT, especially pp. 300-01, argued for a short
term AS that looks like the sort shown in some Principles texts,
with a horizontal "Keynesian" zone, an upward-sloping "intermediate"
zone, and a vertical "classical" zone at full employment, although
he did not use these terms, of course).)
      Regarding Kates's repeated remarks about Say's Law, I
shall repeat myself again.  Several years ago when all this came
up I provided to this list some lengthy quotations from Say's
original work showing that he himself recognized a rather long
list of possible exceptions to his own law, even if he stated it as
a general fact repeatedly elsewhere in the same work.  I do not
wish to repeat this exercise, but I would appreciate it if Kates and
others who are fixated on the idea of the absolute truth of Say's
Law would recognize this truth about Say's own remarks about
the matter, please, unless they wish to argue that Say himself was
wrong about the matter in this case, and that his Law is absolutely
true everywhere and always despite his own clearly stated caveats
to it.
     As real world examples of changes in AD leading to substantial
changes in both real output and employment, I would simply note
Germany in the 1930s after Hitler came to power, and the US in
WW II.  In 1933, Germany had an unemployment rate of around
30%, highest in the world.  By 1939, prior to the actual outbreak
of WW II, it was near zero.  Was this due to some technological
change that shifted the AS curve?  I doubt that any serious reader
of this list would believe such an idea.
     Finally, regarding the multiplier, presumably the argument against
it is premised on the notion that one is on a vertical AS curve, and
therefore it gets ruled out in real terms a priori, even if might be
operating on a successive round of price effects.  However, I would
note that most macro-forecasting models of the US GDP do assume
a multiplier, generally estimated to be a bit below 2 in value.  Also,
it is clearly a serious reality for local economies, where the first round
of the multiplier effects of any change in local output/employment in
export base industries generally show up locally, although I recognize
that a macroecnomists might say that these changes are offset by
changes in the opposite direction somewhere else in the economy.
      Finally, for those who wish to argue about the trend of economic
thought on all this, I would point out that Nobel Prize winner George
Akerlof just gave a presidential address at the AEA meetings that is
getting widely quoted and cited all over the blogosphere in which he
argued that the implication of norm-following behavior on macroeconomics
is to bring about an outcome that looks a lot like the sort of 
"neo-Keynesian"
analysis one would have found in a Samuelson Principles text back in the
1960s, with some exceptions, such as his argument that the long-run
Phillips Curve is downward-sloping for low inflation rates, but upward-
sloping for higher inflation rates.  In any case, it would appear, that 
Krugman
aside, Keynes is definitely back and relevant, along with AD and multiplier
effects.
     And one simply cannot repeat "Say's Law is true, Say's Law is
true," over and over, without recognizing that Say himself said it was not
always true.


Barkley Rosser