AD and foreign substitution

["Peter.Dorman" <23215MGR@xxxxxxx>]

Tom W. offers an argument for the foreign substitution effect (and
therefore a downward-sloping AD curve) based on the tendency for a
one-time-only rise in P to lead to a rise in real interest rates.
This would happen, he suggests, because nominal i would rise (due
to a greater increase in demand than supply in the money market)
but the inflation premium would remain constant, due to the
instantaneous nature of the price shock.

My first response to this is that, empirically, it tends not to
work this way.  Periods in which inflation is rising are generally
periods in which real interest rates are less; presumably expecta-

tions lag behind events--just as they do when disinflation occurs
and real interest rates are very high.  Of course, this empirical
correlation is open to many doubts.  First off, it is difficult to
disentangle the various factors that are changing simultaneously
in these episodes.  In particular, it may be that lax monetary
policy often drives periods of greater inflation, and this effect
should be neutralized to approximate a ceteris paribus analysis,
as AS-AD proposes.  Second, this speaks to the apparent conse
quences of increases in the *rate* of price increases, not to the
one-time increases Tom and the textbooks talk about.  So the
numbers game is not decisive.

Here I am tempted to appeal to the expertise of Penners who know
much more than I do about monetary matters.  But talk is cheaper
than ever in cyberland, so I will speculate anyway.  I find I am
unable to accept the notion of an instantaneous increase in the
price level, unconnected with any change in the (continuing) rate
of inflation.  The price level indexes millions of individual
prices; any theory worth retaining would not *require* (although
it might employ) the assumption of simultaneous price movements.
Moreover, if one accepts some version of the pass-through story,
even an instantaneous price increase would have repercussions for
the inflation rate.  (I recall that one of my minor criticisms of
AS-AD was that it put P, not P dot, on the vertical axis.  But if
it made the switch it would be open to the criticism of treating
inflation and deflation symmetrically.)

In a world of rolling price increases which, moreover, have ripple
effects on future periods (pass-through), nominal increases in i
would not necessarily entail real increases.  To me, this seems
straightforward, and I don't see the need for a formal demonstra
tion--but perhaps I'm missing something.  Am I?

And where are our Pen-l monetary mavens?  This is not really my
turf.

Peter Dorman