Behind the Federal Reserve Board's interest-rate
deliberations and months of sluggish stock-market performance
lies a nagging question: Why did U.S. productivity, the nation's
economic output per labor hour, suddenly accelerate in 1995
after decades of stagnation? The answer will determine whether
America is truly leading the world into a new, prosperous era,
or whether its good times reflect unique circumstances that
can, and likely will, change for the worse.
The benefits of productivity are indisputable. When more products
and services can be created from less labor, capital and raw
materials, the economy can expand without inflation. Demand
and supply roughly balance. Prices stabilize. Unprecedented
levels of employment, wealth creation and consumption become
"New economy" enthusiasts claim that information technologies
like the Internet and computers explain America's recent productivity
spurt. In the early 1990s, these technologies spread throughout
the economy. By the middle of the decade, business was wired
for success. America's digital rebirth supercharged everything
from the nation's stock markets to once-moribund manufacturing
This happy story is a staple of the business and news media.
But it's supported by surprisingly weak evidence. Some of America's
apparent productivity gain is attributable to changes, adopted
in 1995, in the way government keeps its statistics. In fast-growth
periods, moreover, capital investment usually rises faster than
work hours. This artificially boosts output relative to labor
and overstates real productivity improvements.
When adjusted for such factors, the productivity data present
a bigger paradox. It turns out that during 1995-99, all America's
real productivity gains were achieved by durable manufacturing
industries: makers of cars, furniture and computer hardware.
According to Northwestern University economist Robert J. Gordon,
productivity growth rates actually declined in the nearly 90%
of the U.S. economy outside these sectors. But nondurable industries
like trade and financial services are by far the most avid consumers
of computer and information technology. Why did their productivity
Skeptics suggest that new-economy sectors merely repackage,
rather than create, high-value information; functions like retail
sales are displaced, not better ones invented. The Internet
is counterproductive if it encourages personal e-mailing or
shopping while on the job. Real productivity gains as a result
of automating typing or databases were realized years ago when
computers were first introduced. Developments since then have
added increasingly marginal benefits. Even business-to-business
Internet links, thought to boost productivity by improving interfirm
information flows, may prove disappointing if they duplicate,
rather than enhance, fax, telephone and other existing options.
This may help explain the productivity declines suffered by
most of the U.S. economy. But can information technology account
for the productivity gains in durable manufacturing? Or do other
factors, such as historically low component and raw-material
costs and an increasingly servile labor force, better explain
Part of the productivity story in manufacturing is unquestionably
about the decline of working-class power amid the pressures
of globalization and the rise of nontraditional labor markets.
Recent Federal Reserve studies, for example, show that America's
true manufacturing work force is twice as large as officially
reported because of the skyrocketing use of "temporary" employees.
This boosts apparent productivity by severely undercounting
actual labor hours in an industry and by reducing benefit and
wage burdens. During the 1990s, in fact, real hourly manufacturing
wages rose more slowly relative to productivity growth than
at any time in the last 40 years.
The unprecedented price stability of commodities and raw materials
during the 1990s also played a big role. After the Gulf War,
energy prices annually rose by an average of just 1%, compared
with more than 13% a year in the previous two decades. Energy
and material costs are deducted from U.S. gross output to calculate
productivity. When global or political conditions keep such
key prices low, domestic productivity automatically rises.
Then there's the problem of determining which countries and
what workers actually generate the reported improvements in
productivity. The way productivity is measured in the import-dependent
high-tech sectors, which account for the vast majority of U.S.
growth, greatly complicates matters.
To measure the "true" output of U.S. computer makers, the government
uses a novel technique: It adjusts values according to product
characteristics like processor speeds or hard-drive capacities.
For example, if $ 1,000 buys a 200 MHz machine in 1997 and an
otherwise identical 400 MHz computer a year later, manufacturing
productivity is said to have doubled even though market prices
did not change.
This technique can grossly overstate the value of U.S. manufacturing.
Does doubling processor speeds, for example, really make a computer
twice as valuable? Until late last year, moreover, when the
data were partly revised, all U.S. productivity growth since
1995 was attributed to computer production alone.
Computers, like most durable sectors, are heavily dependent
on imports. As much as 80% of all computer imports are transfers
between foreign producers and U.S. companies like Apple or Compaq.
Sophisticated components like microprocessors are shipped from
country to country for manufacturing purposes before they are
assembled into a finished product.
Suppose that the reason why a U.S. company can sell a 400 MHz
machine today for the same price as last year's 200 MHz model
is that overseas suppliers were unable to raise prices on the
faster processor. The cost to the U.S. company stays the same
for a better-performing crucial component. But unless the "true"
cost of the imported processor is adjusted for quality, just
as the value of the finished product is increased to reflect
better performance, reported U.S. productivity will be too high.
The actual productivity increase would have been achieved almost
entirely by the overseas supplier's manufacturing skills or
overseas pricing constraints.
There is good reason to think that the U.S. economy may, in
large part, be benefiting from productivity subsidies of this
sort. During the last decade, worldwide overcapacity caused
manufacturing and high-tech component prices to plummet. Then
the Asian fiscal crisis forced some of the world's most sophisticated
high-tech producers to cut prices even more to earn dollars
and shore up their countries' weakened currencies. Coupled with
virtually limitless low-cost manufacturing opportunities in
the Third World, in the 1990s U.S. producers could obtain the
most advanced components and technology at unbelievably low
Statistically, there is no difference between productivity
generated by brilliant innovations like the Internet or by squeezing
lower prices from labor and suppliers of raw materials and components.
Each story implies, however, dramatically different future prospects
for the U.S.
If the happy story is correct--productivity and technology
go hand in hand--Americans can expect that current worries about
an overheating U.S. economy will be pushed aside by a tidal
wave of productivity-led growth. Tight labor markets and soaring
trade deficits may cause some short-term difficulties, but the
U.S. can literally produce its way out of these and other potential
growth bottlenecks. As other nations copy U.S. innovations,
moreover, a worldwide productivity boom will ensue.
Should it turn out that U.S. productivity gains come more at
the expense of domestic groups, like the working class, which
is disproportionately affected by globalization, or foreign
producers suffering from adverse financial conditions, the future
is far less rosy. It is true that working-class political power
is at its lowest point in decades. Also, countries like Japan,
China and most of the middle- and lower-income world appear
locked in investment and trade patterns that, for the time being,
provide the U.S. with both high-quality, low-cost products and
the capital to consume them.
The problem with building wealth from economic inequity is
that, eventually, something bad happens. Even now, Asian opinion
makers are discussing ways of eliminating their increasingly
unattractive dependence on the United States. Blue-collar labor
may now be docile, but should the economy slow and the brunt
of any hardship fall, as appears likely, on the nation's "temporary"
work force, social unrest will follow. Over time, it's difficult
to appease the less privileged and avoid conflict while preserving
the unbalanced domestic and international relationships that
feed upscale U.S. wealth. That's why the markets cheer adverse
employment data at home and stable, but unspectacular conditions
Uncertainty about the reasons for U.S. productivity growth
thus drives marketplace and policy anxiety. It's possible that
the data may yet prove a harbinger of a beneficent economic
era. Yet, the chance that they chronicle an all-too familiar,
potentially explosive social reality cannot be discounted. Everything
depends on getting the story straight.