Behind The Numbers:
The Treadmill Economy
Jeff Madrick
There is no denying that the United States economy has been
growing at an impressive rate over the past two years. Since the
end of 1995, gross domestic product (discounted for inflation) has
risen at an annual rate of more than 4 percent. Meanwhile,
inflation has fallen to an annual rate below 2 percent, and
interest rates are down to levels unseen in 25 years.
Productivity, we are told, is at last rising above its languishing
trend since 1973. The unemployment rate has fallen to a
quarter-century low of 4.5 percent-without generating any increase
in inflation. What enthusiasts are most heartened about is that
this resurgence of growth began late in the economic expansion
(now in its eighth year) and without the help of fiscal stimulus.
While the economy was growing before the resurgence, the rate of
expansion was unusually tepid. Even an imminent economic slowdown
may not dampen the claims that Am erica has now entered a "New
Economy" driven by the computer revolution and the unleashing of
deregulated markets.
But despite the 4 percent growth since 1995, the business cycle
that began with the last peak in 1990 remains the slowest-growing
in post-World War II history (see "Cycling Downhill," at right).
Real GDP has risen on average by 2.3 percent a year since then.
The average rate of growth since the cyclical peak in 1973, when
the slow-growth economy more or less began, has been 2.6 percent.
The Ford-Carter cycle of the 1970s and the Reagan-Bush cycle of
the 1980s, which admittedly were launched off the troughs of
particularly steep recessions, nevertheless produced an average
rate of growth of about 2.75 percent a year. Surely it is more
than a little premature to call this a New Economy-especially
relative to the good old days when a boom was really a boom. In
the 1950s and 1960s, rates of economic growth averaged 3.5 percent
to 4.5 percent a year for an entire business cycle. Even in terms
of GDP per capita, the 1990s business cycle has been growing more
slowly than any other in the post-World War II period.
The recent growth has not even been sufficient to make up for
unusually sluggish growth earlier in the business cycle. Total
hourly compensation-a measure of the salaries and wages of all
employees, which also in cludes fringe benefits-has shown the
fastest rise of any other government data. Yet it has grown at an
annual rate of only about 0.5 percent in the 1990s cycle. This is
more or less comparable to the rates of growth in the 1970s and
1980s, which were historically low. Between 1948 and 1973, by
comparison, total compensation grew by almost 3 percent a year.
Median family income is only just now catching up to its former
high in 1989 of about $43,000 a year (measured in 1996 dollars).
And this is only the case because so many spouses are working.
Median income for families with only one worker is still below the
highs of the 1970s. Even if we adjust for smaller family size,
median family income has grown at only a slow rate in the 1990s.
Median family income between 1967 (the first year the data were
compiled) and 1973, in contrast, grew by nearly 3 percent a year
even though far fewer families had two workers. About 30 percent
of the female labor force worked then; today, nearly 60 percent
does.
The average wages of production and nonsupervisory workers, who
basically comprise the lower 80 percent of earners, are still 10
or 15 percent below their 1973 highs. Poverty rates have improved
unusually slowly in this expansion and are still above the lowest
levels attained in the mid-1980s, which in turn were well above
their best levels of the 1970s. An astonishing one out of five
children still grows up in poverty today, a statistic the nation
and its policymakers shrug off like one more baseball score in a
long season.
Working Harder For Less
The heart of the problem remains labor productivity-the output of
goods and services per hour of work. It is the foundation of
growth and a rising standard of living. The New Economy boosters
say that the 1.8 percent growth rate since 1995 is the start of a
new era. But output per hour of work has risen by 1.24 percent a
year so far in the 1990s compared to 1.09 percent from the peak in
1980 to 1990 (which includes the short 1980-1982 cycle) and 1.28
percent between 1973 and 1980--all of which are well below the
long-term average of more than 2 percent since the second half of
the 1800s and nearly 3 percent since World War II (see
"Productivity Growth Rates," at right).
Even if the predicted recession does not materialize, a growth
rate of 4 percent for a couple of years is not much to call a
press conference about. Neither are a couple of years of labor
productivity growth of nearly 2 percent a year. (In fact,
according to the latest data revisions, we have really had only
one year of rapid productivity growth. In 1996, productivity
leaped by 2.4 percent. By 1997, it was growing at only 1.4
percent.) Between the low of 1982 and 1987, the economy grew by a
rate of 4 percent a year-that?s five years of 4 percent growth.
Labor productivity grew in these years by a full 2 percent a year.
But for the entire 1980s business cycle, from one peak to the
next, the rates of growth of both real GDP and productivity
remained far below historical norms just as they have in the
1990s. In the 1950s and 1960s, let?s not forget, the economy
produced annual rates of growth of 5 and 6 percent for several
years at a time; productivity often rose at 3 percent and 4
percent a year and sometimes more.
Late-cycle bursts of growth are not unprecedented either. This
economy may only now be making up for years of restraint by the
Federal Reserve and highly cautious financial markets that have
kept real interest rates unusually high and growth sluggish. The
late-cycle rise in productivity can be explained by the fact that
when the rate of GDP growth suddenly rises, productivity will also
usually rise at an above-trend rate as well. Business is
stretching thin all its resources to meet suddenly rising demand.
Workers are working harder as businesses cannot take on new hires
fast enough. In 1986, four years after the expansion began, output
per hour rose by 2.7 percent. In 1968, a full eight years after
the previous bottom, it rose by 3.5 percent. With these numbers in
mind, a sudden spurt of 2.4 percent in productivity in 1996 hardly
looks like the beginning of a New Economy. In such one- or
two-year situations, it is usually not productivity that is
pushing the economy faster, but de mand that is pulling more
productivity out of the nation?s businesses.
Some observers are also excited that the rate of growth has been
achieved without the Keynesian stimulus of higher federal
spending. But the widespread claim that the current expansion is
somehow more balanced or deep rooted because the federal budget is
in surplus requires that we ignore the precipitous climb of the
stock market. As Dean Baker and other economists have pointed out,
the (until recently) rising fortunes of Wall Street have increased
wealth, providing an economic shot in the arm the same way
Keynesian spending would-by stimulating demand. But the stock
market is harder to control. Soaring stock prices and falling
interest rates are supporting consumer spending as well as more
mortgage and credit card borrowing. The danger-as we monitor the
stomach-churning gyrations of the Dow-is what will happen if this
balloon is pricked.
The rate of productivity growth has begun to slow. In the second
quarter of 1998, labor productivity fell slightly, after having
been revised up for the first quarter. For the first half of the
year, productivity has risen by an annual rate of 1.7 percent. But
with rates of GDP growth at 4 percent, such a rate of productivity
growth is historically low. When GDP growth is above 4 percent or
so, productivity typically makes up half or more of the gain in
output, while gains in hours worked make up the other half or
less. In fact, during the rapid growth of the 1950s and 1960s,
productivity accounted for closer to 70 percent of the gain in
output, while increases in hours worked accounted for only about
30 percent. This was almost reversed in the first half of 1998.
Productivity gains accounted for only 40 percent of the growth of
business output; gains in hours worked accounted for 60 percent.
Working Longer For Less
What?s driving the noninflationary growth of the last two and a
half years is not people working more productively; it?s more
people working more hours. People are coming out of the woodwork
to take available jobs, and those who have jobs are working much
longer hours. The recent rise in productivity pales beside the
dramatic rise in total hours worked. And the sudden appearance of
this surplus labor accounts for much more of the subdued rate of
inflation than is generally realized. Enough people want to work
that firms have not had to resort to raising wages to attract
employees. This, plus falling import prices, has kept inflation
tame.
Usually by this stage in the business cycle, there are few
unemployed workers to add to the labor force. But according to the
Bureau of Labor Statistics, total hours worked have risen at an
annual rate of nearly 2.5 percent a year since the beginning of
1996-the typical pace when the economy is coming off its bottom
and unemployment is high. It also represents a greater increase in
hours worked than the fall in the unemployment rate from a level
around 5.5 percent in 1996 to 4.3 percent would suggest. With work
available, labor participation rates are again setting new
records, especially among blacks and women. Employees may feel
they ought to get all the work they can while they are still
employed (see Barry Bluestone and Stephen Rose, "Overworked and
Underemployed," TAP, March-April 1997). The high number of hours
worked may also reflect the high number of part-time workers who,
if offered full-time work or a second job, leap at the
opportunity. In my view, the willingness of people to work longer
hours is stark evidence not of a new prosperity but of the
inability of most Americans to make ends meet after years of
lagging wages. Since the beginning of 1996, total hours worked
have accounted for more than 55 percent of gains in business
output. If this were truly a New Economy, productivity-not the
sweat of labor-should be accounting for the lion?s share of gains,
as it did in the 1950s and 1960s.
The growing labor force hasn?t just depressed wages and kept
inflation under control. It has, along with the higher minimum
wage, had the beneficial effect of arresting the growth of income
inequality. Since the 1970s, earnings increases for low-income
workers have not kept pace with those of higher-income workers.
But Jared Bernstein of the Economic Policy Institute tracks a
little-watched quarterly series of weekly earnings for all
workers, issued by the Commerce Department, and he has found that
in the past two years low-income workers are at last experiencing
some serious pay hikes. In the first quarters of 1997 and 1998,
those workers in the twenty-fifth percentile of earners (that is,
those workers earning more than the bottom 25 percent of the
workforce) enjoyed a 3 percent annual hike over the earnings of a
year earlier compared to only a 0.9 percent hike for those in the
seventy-fifth percentile and a 1.3 percent increase for those in
the ninetieth percentile.
The only credible case to be made for the existence of a New
Economy is based on the possibility that government data
significantly understate productivity gains. Economists say that
the increases in the quality of some goods and especially modern
services such as health care are difficult to measure, and that
the Bureau of Labor Statistics does not adequately account for
such quality improvements when computing the Consumer Price Index
(CPI). Failing to account for these im provements would lead
inflation to be overstated, because the data would not take into
account the fact that consumers would be getting more for their
money. The Bureau of Labor Statistics may also inadequately
account for the introduction of some new products, the prices of
which often fall rapidly. If the CPI is overestimated as a result,
real output is understated and therefore productivity, which is
simply real output divided by hours worked, is also understated.
But for understatement of productivity gains to explain the
existence of a New Economy, there would have to have been
significantly more mismeasurement since 1973 than before. And
there is no evidence that that is the case. Daniel Sichel, an
economist at the Federal Reserve, calculates that even if those
alleging mismeasurement are completely right (which in my view is
highly unlikely), the errors would add at most a little less than
0.3 percent a year since the 1970s and would add still less since
1990.
What New Economy?
America is drawing the wrong lessons from recent economic
performance. Far from being proof that a productivity revolution
is underway, the evidence reveals that many more people need and
want jobs than was previously believed. With the unemployment rate
down to 4.5 percent-and serious inflation nowhere in sight-claims
that the natural rate of unemployment was 6 percent or even 5.5
percent now look naive. Ob servers who insisted that the
underground economy was providing so much opportunity that many
people didn?t want jobs at all-or that many once fully employed
Americans were hap py to retire early or to stay home with the
kids or to become independent contractors-look foolishly
incorrect. In truth, they look more than foolish-they look callous
and even cruel.
One last bit of conventional wisdom, it seems to me, is also now
under increasing suspicion. We have long been told that many
Americans were insufficiently skilled or educated to hold the jobs
of the New Economy. Skill-biased technological change has left
many workers out in the cold, especially those with only a high
school degree. This accounts for the disparity in wages between
college and high school graduates. But consider the curative
effects of a little fast economic growth. The last two years have
shown that these so-called unskilled workers can handle their new
jobs-and get raises, to boot. In fact, the skill-biased technology
argument has always rested on slim evidence. It may well be that
in slack labor markets, inequality of incomes increased when
growth was slow, largely be cause employers could choose from a
large pool of better-educated workers at relatively lower wages
than they had had to pay in the past. As demand for wor kers
rises, however, suddenly it seems that even less-well-educated
workers are more valuable than business thought only a couple of
years earlier.
There is of course a limit to how much surplus there is in the
workforce; as more people join the workforce, at some point the
danger of inflation will arise. But at the moment there is little
wage pressure overall. In my view, low unemployment rates are
finally allowing the distribution of gains in the economy to
become more equitable.
The problem remains productivity. For this economy to be truly
considered "new," it must put together three or four more years of
2 percent annual productivity growth. Perhaps this will happen. At
this writing, however, the U.S. economy looks like it is slowing
down rapidly, especially as reduced demand in Asia for imports
begins to have a greater and greater effect around the world. For
there to be long-term in creases in productivity, we must make a
serious commitment to public investment in day care, early
education, infrastructure, and basic research-all of which have
been badly neglected in the slow-growth era. A reversal in the
unequal distribution of income is also a critical component of
long-term growth. Higher incomes for low-income individuals will
enable them to educate themselves and maintain confidence in the
rewards of work. It will also expand the demand for goods and
services. All this will benefit productivity. To the extent that
rapid economic gains have been made in the past two years, they
cannot yet be attributed to the long-heralded productivity
turnaround. Rather, they have been won on the backs of Americans
who need to work more and haven?t until recently had the
opportunity to do so.
Jeff Madrick is the author of The End Of
Affluence and the editor of Challenge
magazine.
Behind the numbers
