Behind the Numbers: Spin Cycle

The many economists who criticized Bob Dole's campaign
pledge to cut taxes by 15 percent this fall cited the record of
the 1980s as a convincing example that such a deep tax cut would
not generate enough economic growth to pay for itself. As federal
deficits grew through the 1980s, what we got instead, of course,
was a quintupling of the national debt, high real interest rates,
an overvalued dollar, and, to the ongoing consternation of Reagan
revolution boosters, a low rate of savings. What's more, when
measured over the business cycle, gross domestic product (GDP)
continued to grow as slowly as it had in the 1970s, and there
was no improvement in labor productivity growth.

Defenders of President Reagan's record would have none of this.
They mounted a predictably swift response. But, as hard as they
tried, they could come up with no new weapons to fight the battle.
Rather, they fell back on what has become an old standby, which
is to ignore distortions in the record caused by the ups and downs
of the business cycle. Books like The Seven Fat Years,
by the Wall Street Journal editorial chief Robert L. Bartley,
were widely criticized for painting the Reagan record in a glorious
light by only counting the good years and excluding the severe
1982 recession. A book called The Eight Fat Years would
have had significantly different results. But chastisement has
not deterred the supporters of the Reagan record—including some
eminent scholars from the nation's great universities—from doing
it all over again.

Alan Reynolds, a charter supply-sider and now the director of
economic research for the Hudson Institute, presented probably
the classic example of this rebuttal technique in the October
14, 1996, issue of the National Review. The irony is that
Reynolds, perhaps humbled a bit, went out of his way in this piece
to make the claim that he would adjust for the business cycle.
In fact, he did nothing of the sort. "Graph: A Tall TaleTo avoid being misled
by the business cycle, growth should be measured from one business-cycle
peak to the next," Reynolds quoted one of the vociferous
critics of supply-side economics, the Concord Coalition. So far,
so good. He then went on to complain that the Concord Coalition
is distorting the record when it compares the annual rate of real
GDP growth between 1969 and 1980, which was about 3.4 percent,
to the slow growth rate of the 1980s. He has a point. The rate
of economic growth during the business cycle be tween 1973 and
1980 was markedly slower than between 1969 and 1973. He delightedly
reported that real GDP grew from the cycle peak in the fourth
quarter of 1973 to the next peak in the first quarter of 1980
by only 16.2 percent. By contrast, real GDP grew by 32.8 percent
between 1980 and 1990. On this he rested his case.

But something is fundamentally amiss here: Reynolds was comparing
growth over a six-year period to growth over a ten-year period.
In the 11 years between 1969 and 1980, GDP grew by well more than
40 percent, for example. In order to adjust for the business cycle,
all one needs to do is compute the annual rate of growth over
each period. Did the Reagan revolution improve the annual rate
of growth compared to the record of the latter two-thirds of
the 1970s? Reynolds didn't bother to run the numbers, even though
he gave himself a decided ad vantage by citing the wrong real
GDP data for the first quarter of 1980—he reported it as $4,574
trillion when in fact it was $4,674 trillion. Whether this was
a simple mistake or wishful thinking, it certainly helped his
cause. Real GDP rose by 19 percent between 1973 and 1980 rather
than by the 16 percent he reports.

And the annual rates of growth? Real GDP grew by 2.9 percent a
year between the quarterly peak in 1973 and the peak in 1980.
It grew by only 2.8 percent a year between the quarterly peak
in 1980 and the peak in 1990. So much for the corrective power
of the Reagan tax cuts. No wonder Reynolds failed to do the computation
for us.


One might expect such slick and sloppy analysis from editorialists
and Washington policy lobbyists. But well-respected scholars from
the nation's best universities have also felt free to neglect
the effects of the business cycle when it comes to justifying
their political points of view. Robert Barro, a professor of economics
at Harvard, presented one of the more memorable examples of this
kind of oversimplification in September of 1992. In an op-ed for
the Wall Street Journal, Barro graded presidential administrations
for economic performance based on a version of the so-called misery
index, which combines the unemployment rate and the inflation
rate. Because reductions in inflation and unemployment are given
equal weight in this computation, the sharply falling inflation
during Reagan's years helped rank his terms first and second,
according to Barro, out of the eleven presidential terms of the
post-World War II period. No matter that the Federal Reserve deserved
most of the credit.

Barro went on to supplement the traditional index with a bonus
for lowering interest rates as well—even though interest rates
and inflation are closely correlated, so when one falls so usually
does the other. Try to get that one by a Ph.D. dissertation committee.
But this bit of double counting nudged the Bush record a little
higher, and Barro awarded him a "gentleman's B," even
though Barro admitted that under Bush the economy grew more slowly
than under any other administration. The more important point
was that Barro didn't take the business cycle into account at
all. Instead, he simply measured performance over the four years
of each administration compared to the years that preceded it.
President Carter ranked worst on Barro's list. But his performance
would have been significantly better had it been adjusted for
the business cycle, because the Carter years did not include the
first year of the economic recovery, in which growth is usually
the fastest and the unemployment rate falls the most.

Subscribe to The American Prospect

Barro may have had his tongue slightly in cheek. But John
Taylor, the Stanford University economist and advisor to presidential
candidate Bob Dole, recently committed to paper one of the more
egregious examples of this sort of analysis, and one that significantly
damages our understanding of recent economic history. In the final
weeks of the presidential election campaign last October, Taylor
wrote an unusually disturbing piece for the editorial page of
the Wall Street Journal (Octo ber 18, 1996). His objective
was to make a case for lowering taxes and reducing regulations.
The evidence he chose to cite was the record of labor productivity
growth in the Reagan-Bush years, which was indeed higher than
it was in the Carter and Clinton years.

Let's remember that Taylor was on Bush's Council of Economic Advisers,
so a little bias is understandable. But to ignore completely the
influence of the business cycle over productivity growth rates,
as Taylor did, goes well beyond acceptable bounds. If we were
to choose a single measure for rating the nation's economic performance,
I would agree that productivity should be the one. Labor productivity
is simply the output of goods and services per hour of work, and
it has been the source of the growing standard of living ever
since the industrial revolution began.

But labor productivity is tied to the business cycle. In the first
year of a business recovery, for example, labor productivity almost
invariably rises the fastest as GDP bounces off the bottom of
the cycle and as companies remain cautious about adding workers
(keeping output per worker high) until they are certain business
has truly recovered. As the expansion lengthens, productivity
gains begin to peter out and sometimes productivity even shrinks
before the next recession begins. Finally, during recessionary
years, productivity does indeed usually fall (negative growth)
as GDP falls. Thus, to find the trend rate of productivity growth,
you should measure it over the course of an entire cycle. If a
president's four-year term does not include a full cycle, you
can't measure the true course of productivity without making adjustments.
And few four-year terms have included such a complete cycle.

But Taylor computed the annual labor productivity growth rates
for each administration by calendar years, without business cycle
adjustments. My own computations of these non-business-cycle-adjusted
productivity rates under the last four administrations differ
only slightly from Taylor's (see "The Real Deal"). Over
Reagan's two administrations, labor productivity grew by a rate
of 1.3 percent a year. In President Bush's administration it
grew by 1.2 percent a year. Graph: The Real DealIn the Carter and Clinton years (up
to mid-1996), labor productivity grew at only 0.4 percent a year.
Taylor compared the Clinton performance to the bad old days before
the Industrial Revolution. This is "close to the zero growth
of pre-Industrial Revolution days," he wrote. "At that
rate it would take 10 generations to double a person's income."
And he, of course, attributed the better performance of the Reagan-Bush
years to lower taxes and less regulation. (Incientally, under
Barro's report card system, which ignores productivity, I am pretty
certain that Clinton would have earned an A.)

When we properly adjust for the business cycle, however, the results
are nothing like what Taylor purported they are. First, let's
consider the two Reagan administrations. Reagan left office in
1988, two years before the cycle ended-years during which productivity
growth flagged. Between when Reagan took office in 1981 (a cycle
peak) and 1990, labor productivity grew at a rate of only 1.1
percent a year.

Oddly enough, the four Bush years do indeed represent a fairly
complete business cycle, though an inverted one of sorts. Bush's
term began with the modest growth of the tail end of the long
Reagan expansion, then dropped into a moderate recession, followed
by recovery in 1992. In that last year, labor productivity grew
rapidly, as it typically does in the first year of expansion.
Over the four Bush years, labor productivity grew by 1.2 percent
a year, as noted.

The Carter years between 1977 and 1980 include a recession but
exclude 1976, the first full year of recovery, in which labor
productivity grew by 3.6 percent. If we include that year in the
Carter record, which would complete the cycle, we would find that
labor productivity grew by 1.1 percent a year. Similarly, the
first Clinton term does not include 1992, the first full year
of recovery, in which productivity rose by 3.2 percent. If 1992
were included, productivity in the Clinton years would have risen
by 1.1 percent a year. (Although the Clinton record, it should
be noted, is as yet recession-free.)

Non-economists may well ask why Carter and Clinton should
be credited for recoveries that did not begin on their watch.
(There is, of course, a serious question about whether any president
should be credited for igniting a business recovery even when
it occurs in his administration.) But the goal is not to give
Carter or Clinton credit for economic expansion that had nothing
to do with their policies. The point is that if we are to compare
records on productivity growth, it must be done over the same
part of the business cycle, and preferably it should be over an
entire business cycle if we are to determine the long-term trend.
Others may also insist that we should give Reagan credit for the
length of his business expansion. But let's keep in mind that
the durability of the 1980s expansion, unlike the one in the 1960s,
came at the expense of building a mountain of debt. And, again,
the question is whether the long-term course of productivity growth
was altered under Reagan's administration. The answer is that
it was not. The 1.6 percent productivity growth from the 1982
bottom of the cycle, so frequently cited by the Reagan boosters,
takes into account neither the drop in productivity during the
Reagan recession nor the flagging rate of growth in the final
two years of the Reagan expansion presided over by George Bush.
In fact, the final GDP data for 1990, which showed that productivity
had grown at only an annual rate of 0.4 percent over the four
years between 1986 and 1990, ought to have caused John Taylor
many sleepless nights. But somehow I doubt that he rushed to phone
Bush's Council of Economic Advisers to warn his colleagues that
their policies were responsible for a rate of productivity growth
no higher than the pre-Industrial Revolution era.


What is so damaging about Taylor's analysis is that it ignores
what, in my view, is the most important fact about the economy
over the past quarter century. Under Democratic and Republican
administrations alike, labor productivity has continued to grow
at a rate of about 1 percent a year. This is less than half the
average rate of productivity growth since 1870, which economic
historians place at 2.25 percent a year. It is well below the
rapid rate of productivity growth of nearly 3 percent a year reached
in the 25 years after World War II. The Reagan revolution could
not raise the peak-to-peak rate of productivity growth above 1
percent a year between 1979 and 1990. Since the cyclical peak
in 1990, it is still growing at about 1 percent a year.

In sum, this historically slow rate of productivity growth
has held through periods of inflation and disinflation, high and
low real interest rates, soaring and falling oil prices, an overvalued
and an undervalued dollar, tax cuts and tax hikes, union strength
and union weakness, deregulation and even a bit of reregulation.
This is the central fact of our current economic circumstances.
Such a history over a quarter of a century suggests that we may
well have to adjust to slow labor productivity growth indefinitely.
But it also suggests that we should search imaginatively for ideas
that have not yet been tested.

Economists sometimes seem to be infinitely flexible; they can
find support for their theories whatever the historical facts.
Some will claim that productivity did not rise more rapidly because
we did not cut taxes enough in the 1980s. Others will say the
Federal Reserve has yet to create an environment of low real interest
rates without high inflation. My own preference is to think about
what we have not had. We have badly neglected long-term, high-risk
private investment as well as basic research. We have also underinvested
in key public goods, from daycare to infrastructure, urban centers,
education for the less well off, and federal research and development.
What we should not pretend is that there is evidence from the
past 25 years that demonstrates which policies have worked to
sustain and enhance productivity. Certainly, there is no evidence
to support Taylor's contention that a combination of income tax
cuts, capital gains tax cuts, and fewer regulations has improved
America's economic prospects.

The expansion of the 1990s may yet hold more bad news. The next
recession could push the trend rate of productivity growth lower
than 1 percent a year. Some observers still plead that the productivity
of services is being underestimated and that inflation is overstated.
But even if this were true, there is no reason to think that these
mismeasurements are significantly greater in the 1990s than they
were in the 1980s or throughout the post war period. Pro duct
and services quality rose rapidly then too—is MRI technology a
more significant advance than the polio vaccine? Had John Taylor
devoted himself to demonstrating how little progress we have made
in the 1990s toward improving labor productivity, despite the
Clinton expansion, he would have done a service. Instead, he has
distorted the record. The respected economist adds his name to
a lengthening list of those who are willing to ignore the most
basic conventions of economic analysis to make their case.

You need to be logged in to comment.
(If there's one thing we know about comment trolls, it's that they're lazy)