The Commerce Department today released a set of significant revisions in
the way it calculates U.S. gross domestic product, which shows that the
economy has been growing somewhat faster than previously believed.
The revisions raise the level of GDP, both before and after adjustment
for
inflation, and the rates at which both figures have been growing. The
upward revisions in inflation-adjusted GDP should also boost the level
and
rate of increase in U.S. labor productivity -- the amount of goods and
services produced for each hour worked -- but those revisions will come
later from the Labor Department, not Commerce.
The most important of the changes coming today involves business and
government purchases of computer software, which will now be treated as
a capital investment rather than as a current expense. According to a
preliminary Commerce estimate, that change, had it been in effect, would
have raised the level of 1996 GDP by 1.5 percent, or $115 billion. Since
spending for software has continued to rise rapidly since 1996,
presumably the impact on more recent GDP figures would be even
greater.
In addition to such changes in methodology, today's releases incorporate
updated and more complete recent data on both income and spending and
include the initial estimate for GDP for the July-September period. Many
analysts expect the third-quarter GDP estimate to show that the economy
grew at an annual rate of about 4 percent, with a few predicting a higher
number.
Previously, the Bureau of Economic Analysis, the part of Commerce
responsible for GDP and national income figures, has regarded software
as what economists call an "intermediate" product, similar to the steel
going into an automobile or the paper used in a doctor's office. Neither
the
steel nor the paper shows up directly in GDP because they are consumed
in the process of making the automobile or in providing the doctor's
medical services, both of which are part of GDP.
But software isn't consumed in that fashion and it generally remains in
use
for several years, in much the same fashion as a car or truck bought and
used in a business. Instead of such a vehicle being treated as a current
expense in the year in which it is bought, its cost is written off over
the
years of its useful life -- a process called depreciation.
Treating software as an investment rather than a current expense has the
effect of reducing a business's costs without reducing its production,
which
raises the firm's profit. At the same time, the business investment portion
of GDP is increased by the amount of the software purchase.
Since federal, state and local governments also purchase software, the
new treatment means the amount of their investment also rises, giving a
small additional boost to GDP.
A second significant change concerns government employee retirement
plans. Until now, contributions to those plans by government employers
and employees have been included in the government portion of GDP.
Now the government plans are being treated like private plans, and
contributions to them are being regarded as private saving.
This change would have added about $120 billion to personal saving in
1996, according to the bureau's preliminary estimates. That would have
raised total personal saving that year to nearly $270 billion from less
than
$160 billion and reduced government saving by the same amount.
While national saving is not affected, the change has a dramatic impact
on
the level of personal saving. In 1996, for example, personal saving as
a
share of personal disposable income -- essentially after-tax income --
was
2.9 percent. The change would have lifted that to 5 percent.
Since 1996, the personal saving rate has dropped to zero -- that is,
American households have been spending more than the disposable
income they are currently earning and dipping into past saving to do so.
Many economists say consumers have been saving so little out of current
income because of their increase in wealth due to large gains in the value
of stocks they own and in the value of their homes.
Switching the treatment of government employee pension plans won't
actually increase the level of personal saving that has been going on,
but
the key savings-rate figure will be lifted well above the dramatic zero
level.
On the other hand, pension plan funds generally are not available to
potential beneficiaries until they retire and start getting benefits, and
therefore may have a limited impact on consumers' current willingness to
spend.
The bureau is revising its figures for GDP and national income back to
1929, though only revisions going back to 1959 are available today.
A third change being made also raises the GDP and lowers some of the
GDP price indexes for 1978 through 1994. For its various measures of
inflation, the bureau relies in part on the consumer price index data
provided by the Labor Department. Recently that department altered its
method of adding together the changes in prices of many individual items
covered by the CPI to make that index more closely resemble a true
cost-of-living index. In this case, the shift caused the CPI to rise more
slowly.
Since the CPI itself is never revised, the new methodology was applied
only prospectively. But the bureau regularly revises its GDP figures, and
it
had already used the new techniques for its estimates for 1995 to the
present. With prices increasing less rapidly, a larger part of each year's
rise in current-dollar spending for goods and services meant a greater
increase than previously reported for inflation-adjusted GDP.
Among other changes, the bureau also is shifting the base year for
calculating inflation-adjusted GDP to 1996 from 1992. That will raise the
reported level of inflation-adjusted GDP but not affect figures for
economic growth.
© 1999 The Washington Post Company