. 43
( 126 .)


as atomic energy, advanced ceramic materials, and the Internet were originally developed
in federal laboratories. Federal government R&D spending in fiscal year 2006 amounted
to roughly $134 billion, more than half of which went through the Pentagon.
Our multipurpose Figure 3 (page 139) again illustrates the choice facing society. Now
interpret the vertical axis as measuring investments in R&D. Devoting more resources to
R&D”that is, choosing point I rather than point C”leads to less current consumption but
more growth.

Around 1973, productivity growth in the United States suddenly and mysteriously
slowed down”from the rate of about 2.8 percent per year that had characterized the
1948“1973 period to about 1.4 percent thereafter (see Figure 5). Hardly anyone anticipated
F I GU R E 5
this productivity slowdown. Then, starting around 1995, productivity growth suddenly
Average Productivity
speeded up again”from about 1.4 percent per year during the 1973“1995 period back to
Growth Rates in the
about 2.5 percent since then (see Figure 5 again). Once again, the abrupt change in the United States,
growth rate caught most people by surprise. 1948“2007
Recall from the discussion of compounding in
Chapter 6 that a change in the growth rate of
around 1 percentage point, if sustained for
decades, makes an enormous difference in living
standards. So understanding these two major
events is of critical importance. Yet even now,
SOURCE: U.S. Department of Labor at www.bls.gov/data.

some 35 years later, economists remain puzzled
Percent per Year

about the 1973 productivity slowdown, and the
reasons behind the 1995 productivity speed-up are
only partly understood. Let us see what econo-
mists know about these two episodes.

The Productivity Slowdown,
1948“1973 1973“1995 1995“2007
The productivity slowdown after 1973 was a dis-
concerting development, and economists have NOTE: Data pertain to the nonfarm business sector.

Copyright 2009 Cengage Learning, Inc. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.
Licensed to:
Part 2
144 The Macroeconomy: Aggregate Supply and Demand

been struggling to explain it ever since. Among the leading explanations that have been
offered are the following.

Lagging Investment During the 1980s and early 1990s, many people suggested that
inadequate investment was behind America™s productivity problem. Countries such as
Germany and Japan, these critics observed, saved and invested far more than Americans
did, thereby equipping their workers with more modern equipment that boosted labor
productivity. United States tax policy, they argued, should create stronger incentives for
business to invest and for households to save.
Although the argument was logical, the facts never did support it. For example, the
share of U.S. GDP accounted for by business investment did not decline during the period
of slow productivity growth. Nor did the contribution of capital formation to growth fall.
(See the box on growth accounting on the next page.)

High Energy Prices A second explanation begins with a tantalizing fact: The productiv-
ity slowdown started around 1973, just when the Organization of Petroleum Exporting
Countries (OPEC) jacked up the price of oil. As a matter of logic, higher oil prices should re-
duce business use of energy, which should make labor less productive. Furthermore, pro-
ductivity growth fell just at the time that energy prices rose, not just in the United States but
all over the world”which is quite a striking coincidence. This circumstantial evidence
points the finger at oil. The argument sounds persuasive until you remember another im-
portant fact: When energy prices dropped sharply in the mid-1980s, productivity growth
did not revive. So the energy explanation of the productivity slowdown has many skeptics.

Inadequate Workforce Skills Could it be that the skills of the U.S. labor force failed
to keep pace with the demands of new technology after 1973? Although workforce skills
are notoriously difficult to measure, there was and is a widespread feeling that the quality
of education in the United States has declined. For example, SAT scores peaked in the late
1960s and then declined for about 20 years.4 Yet standard measures such as school atten-
dance rates, graduation rates, and average levels of educational attainment all continued
to register gains in the 1970s and 1980s. Clearly, the proposition that the quality of the U.S.
workforce declined is at least debatable.

A Technological Slowdown? Could the pace of innovation have slowed in the
1973“1995 period? Most people instinctively answer “no.” After all, the microchip and the
personal computer were invented in the 1970s, opening the door to what can only be
called a revolution in computing and information technology (IT). Workplaces were trans-
formed beyond recognition. Entirely new industries (such as those related to PCs) were
spawned. Didn™t these technological marvels raise productivity by enormous amounts?
The paradox of seemingly rapid technological advance coupled with sluggish produc-
tivity performance puzzled economists for years. How could the contribution of technol-
ogy to growth have fallen? A satisfactory answer was never given. And then, all of a
sudden, the facts changed.

The Productivity Speed-up, 1995“?
Figure 5 shows that productivity growth speeded up remarkably after 1995, rising from
about 1.4 percent per annum before that year to about 2.5 percent from 1995 to 2007. This
time, the causes are better understood”and most of them relate to the IT revolution.

Surging Investment Bountiful new business opportunities in the IT sector and else-
where, coupled with a strong national economy, led to a surge in business investment

The SAT was rescaled about a decade ago to reflect this decline in average scores.

Copyright 2009 Cengage Learning, Inc. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.
Licensed to:

Chapter 7 145
Economic Growth: Theory and Policy

Growth Accounting in the United States
In this chapter, we have learned that labor productivity (output per
1948“1973 1973“1995 1995“2006
hour of work) rises because more capital is accumulated, because
Growth rate of Labor
technology improves, and because workforce quality rises. The last
Productivity 2.8% 1.4% 2.7%
of these three pillars is quantitatively unimportant in the modern
United States because average educational attainment has been Contribution of 0.9 1.0 1.0
capital formation
high for a long time and has not changed much recently. But the
Contribution of 1.9 0.4 1.7
other two pillars are very important.
The table breaks down the growth rate of labor productivity into
its two main components over three different periods of time. We
SOURCE: Bureau of Labor Statistics at www.bls.gov/data.
see that the productivity slowdown after 1973 was entirely ac-
counted for by slower technological improvement; the contribution
of capital formation did not decline at all.* Similarly, the productiv-
ity speed-up after 1995 was entirely accounted for by faster tech-
nical progress, not by higher rates of investment.

* Changes in workforce quality are included in the technology component.

spending in the 1990s. Business investment as a percentage of real GDP rose from 9.1 per-
cent in 1991 to 14.6 percent in 2000, and most of that increase was concentrated in com-
puters, software, and telecommunications equipment. We have observed several times in
this chapter that the productivity growth rate should rise when the capital stock grows
faster”and it did in the late 1990s. But then investment fell when the stock market
crashed, beginning in 2000. So, over the entire 1995“2006 period, the table in the box above
shows the same contribution of capital formation to productivity growth in 1995“2006 as
in 1973“1995. So investment cannot be the answer.

Falling Energy Prices? For part of this period, especially the years 1996“1998, energy
prices were falling. By the same logic used earlier, falling energy prices should have en-
hanced productivity growth. But, as we noted earlier, this argument did not seem to work
so well when energy prices fell in the 1980s. Why, then, should we believe it for the 1990s?
In addition, productivity continued to surge in the early years of this decade, after energy
prices had started to rise.

Advances in Information Technology We seem to be on safer ground when we look
to technological progress, especially in computers and semiconductors, to explain the
speed-up in productivity growth. First, innovation seemed to have exploded in the 1990s.
Computers became faster and much, much cheaper”as did telecommunications equipment
and services. Corporate intranets became commonplace. The Internet grew from a scientific
curiosity into a commercial reality, and so on. We truly entered the Information Age.
Second, it probably took American businesses some time to learn how to use the com-
puter and telecommunications technologies that were invented and adopted between, say,
1980 and the early 1990s. It was only in the late 1990s, some observers argue, that U.S. in-
dustry was positioned to reap the benefits of these advances in the form of higher produc-
tivity. Such long delays are not unprecedented. Research has shown, for example, that it
took a long time for the availability of electric power at the end of the nineteenth century
to contribute much to productivity growth. Like electric power, computers were a novel
input to production, and it may have taken years for prospective users to find the most
productive ways to employ them.
In summary:
The biggest pillar of productivity growth”technological change”seems to do a credible
job of explaining why productivity accelerated in the United States after 1995.

Copyright 2009 Cengage Learning, Inc. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part.
Licensed to:
Part 2
146 The Macroeconomy: Aggregate Supply and Demand

Earlier in this chapter, we observed that the relative prices of services such as
college tuition, medical care, and theater tickets seem to rise year after year.
And we suggested that one main reason for this perpetual increase is tied to
the economy™s long-run growth rate. We are now in a position to understand
precisely how that mechanism works. Rising productivity is the key. The ar-
gument is based on three simple ideas.

IDEA 1 It stands to reason, and is verified by historical experience, that real wages tend to
rise at the same rate as labor productivity. This relationship makes sense: Labor normally
gets paid more when it produces more. Thus real wages will rise most rapidly in those
economies with the fastest productivity growth.

IDEA 2 Although average labor productivity in the economy increases from year to
year, there are a number of personally provided services for which productivity (output per
hour) cannot or does not grow. We have already mentioned several of them. Your college
or university can increase the “productivity” of its faculty by increasing class size, but
most students and parents would view that as a decrease in educational quality. Simi-
larly, a modern doctor takes roughly as long to give a patient a physical as his counter-
parts did 25 or 50 years ago. It also takes exactly the same time for an orchestra to play
one of Beethoven™s symphonies today as it did in Beethoven™s time.
There is a common ingredient in each of these diverse examples: The major sources
of higher labor productivity that we have studied in this chapter”more capital and
better technology”are completely or nearly irrelevant. It still takes one lecturer to
teach a class, one doctor to examine a patient, and four musicians to play a string quar-
tet”just as it did 100 years ago. Saving on labor by using more and better equipment is
more or less out of the question.5 These so-called personal services stand in stark contrast
to, say, working on an automobile assembly line or in a semiconductor plant, or even to
working in service industries such as telecommunications”all instances in which both
capital formation and technical progress regularly raise labor productivity.

IDEA 3 Real wages in different occupations must rise at similar rates in the long run. This
point may sound wrong at first: Haven™t the wages of computer programmers risen
faster than those of schoolteachers in recent years? Yes they have, and that is the mar-
ket™s way of attracting more young people into computer programming. But in the long
run, these growth rates must (more or less) equilibrate, or else virtually no one would
want to be a schoolteacher any more.

Now let™s bring the three ideas together. College teachers are no more productive
than they used to be, but autoworkers are (Idea 2). But in the long run, the real wages
of college teachers and autoworkers must grow at roughly the same rate (Idea 3), which
is the economy-wide productivity growth rate (Idea 1). As a result, wages of college
teachers and doctors will rise faster than their productivity does, and so their services
must grow ever more expensive compared to, say, computers and phone calls.
That is, indeed, the way things seem to have worked out. Compared to the world in
According to the cost which your parents grew up, computers and telephone calls are now very cheap while
disease of the personal
college tuition and doctors™ bills are very expensive. The same logic applies to the serv-
services, service activities
ices of police officers (two per squad car), baseball players (nine per team), chefs, and
that require direct personal
many other occupations where productivity improvements are either impossible or un-
contact tend to rise in price
desirable. All of these services have grown much more expensive over the years. This
relative to other goods and
phenomenon has been called the cost disease of the personal services.

However, some people foresee a world in which some aspects of education and medical care will be delivered

long distance over the Internet. We™ll see!


. 43
( 126 .)