[PEN-L:3799] RE: stock market & investment

["Gary Dymski" <dymski@xxxxxxxxxxxx>]

> The other day, I asked loyal Keynesians to comment on Morck, Shleifer, and
> Vishny's 1990 paper showing the stock market to have little effect on
> investment - a very damaging empirical refutation of JMK's line on the
> malign influence of sentiment on the real world.

Let me add a couple of points to Chris Niggle's nice response.  The
disjuncture between stock-market prices and investment has also been found
in the years-long effort by many to find empirical verification of Tobin's
'q'.  Tobin's q, which comes from his JMCB 1969 paper on banking
equilibrium, is in turn drawn from -- forgive me Paul, but time is short
today -- somewhere in Chapter 17 in Keynes.  It is by way of a suggestion
that the gap between financial-market assessments of firm value, and the
replacement cost of firm assets, should determine ceteris paribus whether it
is cheaper to get a factory by buying the parts and building it, or by
buying the company that has it.

This suggestion comes in Keynes' book which is, by all accounts, dripping
with suggestions about the relationships of the financial and real.  Doug
always comments sensibly on the economy, and I know he doesn't want to make
the same blunder in theory as do some of our neoclassical colleagues -- that
is, to cook Keynes down to a single formula.  The most egregious example of
this approach is in the "Keynes effect" that has been discussed in "modern
macro" (see Blanchard and Fischer).

Tobin's use of this (q) ratio was quite another thing, because Tobin yoked
the idea to his porfolio equilibrium model of financial and real markets.
So his q was one of the few levers available for "financial structure to
matter" (given that he had along the way thrown out fundamental uncertainty,
myopia, missing information, etc.).  In effect, tests of 'q' became more or
less tests of the efficient-markets hypothesis.  Their weakness in empirical
estimation didn't suggest that "financial factors don't matter", but rather
than "nobody knows what to make of this 'q' or how to measure it."  There
was a long and futile debate on "marginal" vs. "average" q, etc.

Finally some good Keynesians (such as Steve Fazzari) found other variables
that show "financial structure matters" -- in SF's case, debt structures of
firms "matter" empirically.

I too don't specifically remember the Vishny et al paper you cite, but I
know the work; that is, asymmetric information/game theoretic modelling of
principal-agent problems in ownership/credit markets.  Here some Keynesian
features are introduced, in a manner of speaking, but only just.  Paul
Davidson, Roy Rotheim, and others on this net have discussed the limitations
of bringing in Keynes this way.

Now you propose the following: if Keynes is right and stock-market
sentiment matters, independent of fundamentals, then a measure of
sentiment should be a determinant of investment.  But this doesn't follow,
for two reasons.  First, the whole point of Keynes' Chapter 12 discussion of
stock-market insanity is that there IS a disconnect between what goes on in
the Bourse and the conditions of real production.  It is a game independent
of other games.  Second, distinguishing "sentiment" from "fundamentals"
would be a very tricky business indeed.  Shiller's work on market volatility
shows that you cannot independently test a theory of market prices (are they
right or wrong) without testing a theory of expectations (how are they
formed).  Shiller wanted to "disprove" rational expectations, but couldn't
really do it -- couldn't show that prices moved around "too much" -- because
of this Heisenberg uncertainty problem.  It seems to me the same trap is
open here to lure us in.

So in sum, I don't see where the failure of stock-market prices to matter
in determining investment invalidates Keynes or modern-day Keynesian
economics in any sense.  Other measures of "financial
structure" consistently _DO_ matter in such equations -- the sorts of
cash-flow and debt-load variables that Hy Minsky and others have emphasized
in their interpretations of Keynes.  And it isn't clear how one could
separate out "sentiment" from "fundamentals" in a way that avoids the sort
of trap that has frustrated Shiller for all these years.

So Doug, as for whether Keynesians today are intellectually dead, I would
simply reply (with a small smile) in the words of the Great Communicator:
"There you go again ..." (!)

Trying hard to avoid intellectual death on the Left Coast,
Gary D.
Gary A. Dymski                                      dymski@xxxxxxxxxxxx
Dept. of Economics                                   fax (909) 787-5685
Univ. of California, Riverside CA 92521   voice (909) 787-5037 ext 1570