Does Technology Create Jobs?
Two leading economists,
MIT's Paul Krugman and the Hoover Institution's
David R. Henderson, debate whether jobs lost
to technology are met by a net increase in jobs elsewhere in a
more productive economy. Krugman, a noted liberal, says maybe
in the long run, but for now ordinary workers see their wages
falling. Henderson, a conservative, says that the problem is not
the elimination of jobs through technology but a workforce with
for ordinary folk
By Paul Krugman
Even the early
stages of the Industrial Revolution quickly made England the wealthiest
society that had ever existed, but it took a long time for the
wealth to be reflected in the earnings of ordinary workers. Economic
historians still argue about whether real wages rose or fell between
1790 and 1830, but the very fact that there is an argument shows
that the laboring classes did not really share in the nation's
again. As with early-19th-century England, late-20th-century America
is a society being transformed by radical new technologies that
have failed to produce a dramatic improvement in the lives of
ordinary working families--indeed, these are technologies whose
introductions have been associated with stagnant or declining
wages for many. The Industrial Revolution was based on iron and
steam, while we are living through a revolution based on silicon
and information; but in a deep sense the story is probably much
As far back
as 1817, the great economist David Ricardo explained how technological
progress can raise productivity yet hurt workers; his analysis,
suitably reinterpreted, remains valid today.
Here is a
modernized version of Ricardo's story: imagine that initially
our economy uses a technology requiring that each worker be supplied
with $50,000 in capital equipment. And suppose that the current
level of savings and investment is just enough both to replace
old capital as it wears out and to equip new workers with the
same level of capital as those already employed. In such an economy,
there will be more or less full employment and a stable distribution
of income between capital and labor.
a new technology comes along--one that raises the productivity
of the average worker dramatically, say by 75 percent. The only
drawback is that to use the new technology, a worker must be equipped
with much more capital--say $100,000's worth. If wages are a great
enough share of costs, companies will find the new technology
well worth introducing in spite of the extra cost, but what will
it do to the workers?
is that, at least at first, workers will be hurt, because the
economy will no longer have enough savings to maintain full employment
at the going wage. An investment that would have added two jobs
will now add only one, so there will no longer be enough jobs
created. The new technology will begin destroying jobs instead
of creating them.
Now it's true
that the law of supply and demand can still work its magic. In
a free-market economy, the prospect of unemployment will drive
down wages, and at sufficiently lower wages, employers will find
it profitable to offer more jobs after all. But the point is that
these will be worse, lower-paying jobs even though the economy
as a whole is richer.
true that higher profits generated by the new technology will
lead to more investment, and this may eventually mean higher wages.
But the operative word is eventually. If history is any guide,
it may be decades before the fruits of a better technology are
fully reflected in higher wages. There are, admittedly, some important
differences between the early 19th century and the late 20th,
but they are less fundamental than they may seem.
the Industrial Revolution bad for wages was that it was not only
labor saving but also, to use technical jargon, “capital
using,” because the new technology meant replacing small-scale
artisan production with capital-intensive factories, creating
a shortage of capital and a scarcity of jobs. Information technology,
however, is not especially capital using. Indeed, it often seems
to economize on capital as much as it economizes on labor.
of modern technology, rather, is that it is human-capital using;
it greatly increases the demand for highly educated and exceptionally
gifted people. Never in human history have so many people become
so rich so quickly, and the rewards to skill and talent have never
been larger. But for every Bill Gates or Marc Andreessen, there
are thousands who find that technology has made it harder, not
easier, to earn a living. Just as the physical-capital-using technology
of the Industrial Revolution initially favored capital at the
expense of labor, the human-capital-using technology of the information
revolution favors the exceptional (and lucky) few at the expense
of the merely intelligent and hardworking many.
We could not
stop the information revolution even if we wanted to. And in the
long run, new technology will undoubtedly raise everyone's standard
of living. But that is then and this is now, and as John Maynard
Keynes famously pointed out, in the long run we are all dead.
(firstname.lastname@example.org) is a professor of economics at MIT and winner
of the 1992 John Bates Medal. He has served as senior international
economist on the staff of the Council of Economic Advisers and
is the author of The Age of Diminished Expectations: U.S. Economic
Policy in the 1990s (1990) and Pop Internationalism (1996), reviewed
in October's Herring (see “Everything You Know Is Wrong”).
for everyone but the unskilled
By David R. Henderson
and I agree that as long as wages are flexible--and we agree that
in the United States they are--technological change cannot destroy
jobs on net. The reason: even if the demand for labor falls, wage
rates can and will fall, keeping workers employed. The one exception
would be very unskilled workers, some of whom would be priced
out of work by the minimum wage. Krugman and I also agree that
“capital using” technological change can reduce real
wages for workers.
But a theoretical
possibility is not the same as a fact. The important question
is not whether the information revolution can reduce real wages
for workers, but whether it does. This Krugman has failed to establish.
that real hourly wage rates for employees have fallen gradually
over the last 23 years. Based on data from the president's Council
of Economic Advisers, I compute that the average real wage for
production and nonsupervisory workers in the private sector peaked
in 1973 at $14 (in 1996 dollars) and is now about $12.13. But
these data have two big shortcomings; the effect of both is to
understate current real wages.
the last 23 years, an increasing portion of workers' pay has taken
the form of benefits--pensions, health insurance, etc.--none of
which are counted in hourly wages. Although the Bureau of Labor
Statistics reports overall compensation for all employees, not
just for production and nonsupervisory workers, the data are illuminating.
Since 1980, real benefits, valued at the employer's cost, have
risen by 20 percent. Average real employee compensation, including
benefits valued at cost, has risen by about 4 percent.
problem with the standard data on real wages is that the consumer
price index (CPI), used to adjust for inflation, overstates inflation.
According to the 1995 Report by the Advisory Commission to Study
the Consumer Price Index, between 1987 and 1995 the CPI overstated
the inflation rate by between 1 and 2.7 percentage points annually.
The CPI does not adjust for the fact that people buy more of those
goods whose price has fallen and less of those whose price has
It also fails
to adjust for quality improvements and to capture the “Wal-Mart
phenomenon”--that consumers can now purchase goods at large
chains for lower prices than they used to pay at local mom-and-pop
stores. These three factors alone, according to a recent study
by Northwestern University economist Robert J. Gordon, bias the
CPI upward by about 1.2 percent a year. Assuming this same 1.2
percent bias for every year since 1973, real hourly wages have
actually increased from $14 to about $16.50, and real employee
compensation has increased by about 40 percent. One of the main
reasons for quality improvement, incidentally, is the revolution
in technology that has improved cars, made movies available on
demand at a fraction of the previous cost, and slashed transportation
and communication costs.
fringe benefits should not be valued at employer cost because
they are typically worth less. The employer's portion of social
security taxes, for example, is mandated by the federal government
and is less valuable to employees than the cash that they could
have invested in stocks and bonds. Benefits that are not mandated,
such as health insurance, are probably worth less than their cost
but are provided because they are a form of tax-free income. Therefore,
the picture I painted of rising real compensation is rosier than
the reality. But let's put the blame where it lies: not on the
information revolution, but on actions like the federal and state
governments' increase of social security taxes.
may well be true that very unskilled workers earn lower real wages
than they did 20 years ago. But the reason is that they have fewer
skills than their counterparts did two decades ago. A recent study
in Review of Economics and Statistics, by two economists from
Harvard and one from MIT, concludes that “a high school
senior's mastery of skills taught in American schools no later
than the eighth grade is an increasingly important determinant
of subsequent wages” (italics theirs). It finds that those
who graduated from high school in 1980 are noticeably less skilled
than their class-of-1972 counterparts. What are these skills?
Not rocket science, but simple computation with decimals, fractions,
and percents and recognition of geometric figures.
spending on schools is not the solution. The government's approach
to schools is the problem. What are we to think of a president
of the United States proudly stating his ambition for every student
to know how to read by the end of the third grade? Only about
half of the nation's high school seniors have mastered eighth-grade
skills, the study's authors note. When a firm has only a 50 percent
success rate on the basics, most of us think the customer should
R. Henderson (email@example.com) is a research fellow at the Hoover
Institution and an economics professor at the Naval Postgraduate
School in Monterey, California. He was a senior economist with
President Reagan's Council of Economic Advisers. He writes regularly
for Fortune and the Wall Street Journal and edited The Fortune
Encyclopedia of Economics.