"Irrational Exuberance"? Federal Reserve Bank of
San Francisco Conference, April 21, 2000
Robert E. Hall (1999), "The Stock Market and Capital
Accumulation"
Bart Hobijn and Boyan Jovanovic (2000), "The Information
Technology Revolution and the Stock Market: Preliminary Evidence"
I'm lucky here, even though I'm a discussant
of two very different papers. I'm lucky in that they do share
a common theme. Each of them tries to take the stock market very
seriously, and assumes that the stock market has a lot to tell
us about the productive economy. Bob Hall employs his patented
method of taking some piece of economic theory very seriously,
letting something usually taken as fixed--in this case dY/dK--swing
wildly, and wind up by resolving one puzzle while creating two
or three larger, deeper, puzzles. Boyan Jovanovic uses the stock
market to try to say things about creative destruction.
Let me start with Bob Hall, who assumes that
the value of corporations' stocks and bonds tells us what the
quantity of corporate capital is in the economy.
We know more or less what the replacement
cost of corporations' plant and equipment capital stock is. And
we know that what Hall calls the total quantity of capital varies
from a low of 0.7 times the replacement cost of plant and equipment
in the second half of the 1970s to a high a couple of months
ago of 1.8 times the replacement cost of plant and equipment.
Hall wants to identify the extra 0.8 times the plant and equipment
stock that is valued today as the corporate sector's aggregate
"intangible capital"--the organizational patterns and
processes and the business-created consumer attitudes that were
expensive to create and that are a key part of productive efficiency.
The story of the post-WWII evolution of the
stock market and the corporate sector as Hall tells it is as
follows: On average since 1945 the after-corporate-tax return
on corporate capital has been 10%. Of this 4% has been paid out
to shareholders and bondholders, and 6% has been reinvested.
Between 1945 and 1968 this process of compounded reinvestment--accompanied
by "noise" in asset market valuations produced by various
factors--boosted the value of corporate capital from $800 billion
to $4 trillion 1998 dollars. Then something odd happened: in
spite of substantial cash flow devoted to reinvestment, 1972
saw the real value of corporate capital unchanged from 1968:
some $1 trillion on net had been invested in expanding America's
corporate capital, and an equivalent $1 trillion of the old corporate
capital had vanished over those four years. Things got worse:
in 1978 the value of corporate capital was only $3 trillion,
instead of the $8.1 trillion that one would have obtained by
extrapolating the 1945-1968 experience in reinvestment patterns
and returns on reinvestment: some $5 trillion of corporate capital
had gone missing.
Conversely, at the start of the decade just
past the U.S. had some $5 trillion of corporate capital--an amount
that should, given the 4% average net reinvestment rate that
prevailed from 1945 to 1990, have compounded to $7.5 trillion
of our dollars by the start of 2000. Yet by the start of 2000
we had $12 trillion. An extra $4.5 trillion of corporate capital
had been created, largely in the second half of the decade just
past.
This destruction and creation of capital show
up in Bob's framework as fluctuations in the productivity of
capital: 8% per year in the 1960s, 2% per year in the 1970s,
13% per year in the 1980s, and 17% per year in the 1990s--but
if one broke the 1990s in half, one would find a productivity
of capital of some 24% per year for the last five years. That's
what you need in Hall's framework to explain the sudden creation
and destruction of large amounts of organizational capital.
So one way to look at it is that Bob has replaced
the puzzle of the extraordinary volatility of the stock market's
valuation of Lucas trees with a puzzle of the extraordinary decade-to-decade
volatility in the productivity of capital. I don't know which
puzzle we are likely to find more difficult to solve.
Boyan has an ingenious and interesting story
for where these episodes of capital destruction and capital creation
come from: the development of our new general-purpose information
technologies. And to support this story Boyan has the nicest
model of creative destruction that I have yet seen.
Boyan's story is that old vintages of capital
are always under threat because established organizations have
a hard time reconfiguring to take advantage of new technologies
and processes. In particular, in the 1970s old capital lost its
value when it was clear that the information technology revolution
was coming, and that it was about to face competition from a
vastly more efficient new vintage of capital-- organizational,
physical, technological. New capital has acquired its value as
it has become clear exactly which new firms are best at using
new this new vintage of capital based on modern general-purpose
information and communications technologies.
But I have a hard time making Boyan's story
work quantitatively as an explanation of what has happened to
U.S. equity values since 1970. The $5 trillion of corporate capital
that vanished beween 1968 and 1978 is, presumably, the result
of people between 1968 and 1978 looking forward to 1998 and suddenly
seeing the reduced value of old pre-IT capital--for there was
little IT capital (outside of airlines and insurance companies)
back in the 1970s. But given the high real required rate of return
on the market, a dollar of wealth lost in 1998 has a low present
value in 1973: seventeen cents (if you use the 7% real required
rate of return I carry in my head) or eight cents (if you use
the 10% real required rate of return that Bob Hall carries in
the back of his head). To have the shadow of competition in 1998
and thereafter destroy $1 trillion of market value in 1973 would
require that in the absence of the the new vintage of capital
the old capital would have been worth an extra amount of between
$6 and $12.5 trillion in 1998. And to have the shadow of competition
from 1998 destroy $5 trillion of capital value in 1973... I cannot
match up the capital-destruction episode characterized by Bob
Hall to a plausible effect of Boyan's model.
Moreover, even in the middle of the 1970s
it was not at all clear that old corporations would fail to profit
from new information technologies. The core of the internet today
is technology--UNIX--that was the undoubted intellectual property
of AT&T in the 1970s. The greatest profits earned from IT
in the 1980s accrued to IBM. And if you look at analysts' reports
from the 1970s for Xerox--one of the niftyest of the pre-1974
"nifty fifty", selling for 50 times earnings or more--they
stress that Xerox has a bright future in large part because of
the technologies being developed at Xerox's Palo Alto Research
Center. And, indeed, the market capitalization today of Xerox
PARC's technologies is remarkable. Ethernet walked out of PARC
inside Bob Metcalfe's head: 3COM. Postscript walkied out of PARC
inside John Warnock and Charles Geschke's heads: Adobe. Microsoft
Word walked out of PARC inside Charles Simonyi's head. When Steven
Jobs was (apocryphally) yelling at Bill Gates for "similarities"
between Windows and Macintosh, saying that it was like Gates
breaking into Jobs's house and stealing his TV set, Gates (apocryphally)
responded: "That's not true. It's like we both had a rich
neighbor named Xerox, and you just happened to steal his TV set
first."
One of the most interesting--and least well-explained--pieces
of economics is the remarkable failure of companies to keep hold
of and profit from IT research and development done at their
own facilities. Ever since the invention of the modern corporation
with its research lab companies had been pretty good at keeping
hold of and profiting from their intellectual property: think
of Alfred Nobel's dynamite company, or DuPont, or General Electric,
or Boeing.
One final thought: perhaps when we look at
the magnitude of the capabilities made possible by our general-purpose
information technology, and try to match it up to a large profit
flow that supports high stock market valuations, we are looking
in the wrong place. It is far from clear that rapid technological
change with major effects on social welfare maps clearly and
transparently into lots of profits for enterprises. New technologies
create large, permanent profits where they can be transformed
into powerful barriers to entry. And it is not clear to me that
this is the case: I tend to be an optimist about the social impact
of the technology, and a pessimist about the profits of dot-coms.
Why? Let me tell a story--the story of the
collapse of the encyclopedia market. Two decades ago I was thrilled
to get my paper Encyclopedia Britannica--a market value
of $1500. It seemed worth every penny (even though I didn't pay
for it). For a while I would go down and look at it every day,
running down a reference or flipping through random pages. It
was still the resources that I went to when I ran across something
unfamiliar until last year--until the coming of http://www.britannica.com/
to the internet.
And now my paper copy is gathering dust. It
has been replaced. Reading on a screen is hard, but so is reading
small print. On my computer I can make britannica.com's print
as large as I want, so it is actually easier to read on the screen
than in the book. And the search engines! You need a flat place
to open at least five volumes if you are going to do a comprehensive
job of searching the paper encyclopedia. By contrast, the first
search brings up all the references in the whole encyclopedia
text on the very first screen. http://www.britannica.com/ is
a superior intellectual product.
And it is free: intellectual capabilities
that were expensive--to the tune of $1500--fifteen years ago
are now effectively free to anyone with an internet connection,
and a web browser. For most people access to the Encyclopedia
Britannia is not worth the $1500 it used to cost: we know
this because most people did not buy the Encyclopedia Britannica.
Still, it is worth something. If it is worth an average of $50
per household to Americans, than we all are now $4 billion richer
as a result of the decision to put http://www.britannica.com/
online and its competitors online.
And is anyone going to profit from this creation
of $4 billion in wealth? It is hard to see how. Amazon, for example,
started out thinking that it would make money by having the best
front-end computer interface in the world to Ingram, the U.S.'s
largest book distributor. It has now decided that it will try
to make money by integrating backward--by being the best distributor
of books and other stuff. So if you look at the bulk of Amazon's
assets today, they aren't made up of new-technology goods like
servers and RAID disks and Oracle tables and Cisco routers and
Uniphase fiber. They are made up of warehouses and forklifts:
think Sears. Now maybe Amazon truly does have extraordinary organizational
capital--is much, much better at the warehouse-and-forklift business
than Walmart or Ingram, and has an advantage that others will
not be able to copy. Organizational capital is truly valuable
only when it is expensive or impossible to copy. And I don't
see how...
So perhaps there are limits to how far we
can go by starting from the premise that we should take asset
market valuations very seriously...
Professor
of Economics J. Bradford DeLong, 601 Evans Hall, #3880