FRBSF Economic Letter
97-30; October 17, 1997
Assessing the Benefits of Economic Growth
On September 29, 1997, the U.S. Census Bureau released its annual report
on family income and poverty in the United States. The report showed that,
by official calculations, economic growth is finally benefitting a majority
of Americans. While this news prompted much fanfare, there is reason to
suspect that most Americans have been benefitting for much longer than
these official numbers would suggest. This Economic Letter reports
on recent research by Burkhauser, Crews, and Daly (1997) (hereafter BCD),
which uses the decade of the 1980s to highlight the sensitivity of family
income analyses to measurement issues. BCD find that studies of family
income in the United States frequently exaggerate the losses of middle
and low income persons by mixing cyclical with secular differences, using
a narrow definition of income sharing units, and applying an upwardly
biased cost-of-living index. Employing alternative measures BCD find that,
even in today's highly competitive global economy, growth improves the
economic well-being of most Americans.
Counting economic winners and losers
The simplest way to examine how the benefits of economic growth during
the 1980s were distributed throughout the population is to compare the
distributions of real income by percentile for selected years. Figure 1 shows two
related methods of making this comparison, using years commonly discussed
in the family income literature, 1979 and 1992. First, the two black lines
show that the 1979 and 1992 distributions cross at about the 56th percentile
of income. This means that one must go to the 56th percentile before real
income in 1992 exceeds the real income for an equivalent person in 1979.
According to this view, the lower 56 percent, or more than one-half of
the population, were economic losers in 1992.
The second way of making the comparison is to take the difference between
the first two lines--this difference is shown by the dotted line. The
benefit of this measure is that it shows the size of the gains and losses
experienced by those on both sides of the cross-over point. Although these
diagrams typically report the percentage change in real income on the
vertical axis, the absolute difference in real income is reported here.
This "cross-over point" methodology provides a simple knife-edge
categorization of winners and losers from which the effects of changes
in measurement suggested by BCD are easily seen.
Characterizing the 1980s expansion
If real growth during the 1980s occurred for all portions of the income
distribution, the cross-over point in Figure 1 would not
be at the 56th percentile, but at zero. Using this fact, numerous authors
have argued that the benefits of economic growth during the 1980s did
not trickle down to the poor and middle class (for an example, see Danziger
and Gottschalk 1995).
However, BCD argue that this rather bleak picture of economic expansion
during the 1980s is extremely sensitive to the years compared, the income
sharing unit chosen, and the inflation adjustment employed. Their analysis
shows that issues of definition and measurement, although frequently relegated
to an appendix or left undiscussed entirely, are important components
of accurately describing income changes over time. Parallel research by
Jencks and Mayer (1996), on changes in child poverty rates over the past
three decades, confirms this view.
To demonstrate these sensitivities, BCD use data from the March Current
Population Survey (CPS) and the cross-over point methodology employed
in Figure 1.
The March CPS is a nationally representative sample of more than 50,000
households and is used frequently in studies of income distribution. Income
in this analysis refers to the combined pre-tax, post-transfer real resources
of all members of an income sharing unit. To account for the fact that
$20,000 a year provides a higher standard of living for a single person
than it does for a family or household with multiple members, all incomes,
including those in Figure 1, are adjusted
by the size of the income sharing unit. See Daly (1997) for a more detailed
description of this adjustment.
displays the results of the analysis using different years, different
income-sharing units, and an alternative inflation index.
Although most economists take for granted that any examination of changes
in the income distribution over time will be sensitive to the years being
considered, BCD find that research in this area frequently has failed
to distinguish between changes associated with movements in the business
cycle and changes that occur between two similar points in the business
cycle. While there are no formal rules for choosing comparison years,
a brief review of the economic fluctuations over the past two decades
illustrates the potential problem with selecting analysis years randomly.
The early 1980s were marked by the worst recession since the Great Depression.
Real median family income fell from a 1979 business cycle peak of $39,227
to a 1982 business cycle trough of $36,326. In addition, unemployment
rose from 5.8 percent in 1979 to 9.7 percent in 1982. However, seven years
of uninterrupted economic growth followed, so that by the next peak in
the business cycle in 1989, real median family income had increased to
$40,890, and unemployment had fallen to 5.3 percent. Although the business
cycle of the 1990s has not fully played out, in terms of unemployment
it appears 1992 is the trough year with unemployment rising to 7.4 percent
and real median income falling to $38,635. In 1993, real median income
fell to $37,905, but unemployment decreased to 6.8 percent.
The effect of the business cycle on the income distribution and on the
assessment of who won and who lost during the 1980s can be seen by comparing
and the curve labeled "Families 79-89" in Figure 2. One reason
that the cross-over point in Figure 1 is so high
in the income distribution is that 1979 and 1992 are peak and trough years,
respectively, of different business cycles. In contrast, the "Families
79-89" curve in Figure 2 shows the
cross-over point in the income distribution when business cycle peak years
1979 and 1989 are compared. Using these years, the cross-over point becomes
the 36th percentile. Hence, for almost two-thirds of the population, family
size-adjusted real income increased between the two 1980s business cycle
peaks. This is a more favorable economic outcome than is implied by the
peak-trough comparisons of Figure
Resource sharing. Another factor affecting characterizations
of economic progress is the treatment of resource sharing among individuals
living in a common residence. In the CPS, one can use the family or the
household as the unit of income sharing. This choice determines how the
income of unrelated individuals is combined. Under the family definition,
only those related by blood or marriage are assumed to share resources.
Anyone else who lives in the dwelling is treated as a single-person "family."
Applying this definition, two people cohabitating and sharing resources,
but not officially married, would be counted as two separate single-person
families. Since there are economies of scale from living together--two
can live together for less than the sum of each one living separately--treating
cohabitators as separate, single-person families may overstate the number
of people in the lower tail of the income distribution.
The curve labeled "Households 79-89" in Figure 2 illustrates
the importance of the choice of income sharing units. Using the household
sharing unit, BCD find that the cross-over point moves from the 36th to
the 26th percentile. Under a household sharing unit definition, three-quarters
of all persons gained from economic growth between the two peak years
of the 1980s.
Cost-of-living adjustments. Finally, it is
also important to recognize the effect that the cost-of-living adjustment
has on the analysis. Increasingly it is argued that the CPI cost-of-living
index overstates inflation. Boskin (1995) offers the most systematic criticism
of the CPI and proposes alternative indices for the 1980s. Following Boskin,
BCD apply an index based on yearly measured price changes that are on
average 1 percentage point lower than those reflected in the current CPI
for 1979 and 1989. The curve labeled "Boskin 79-89" in Figure 2 shows that
under a Boskin-CPI index the cross-over point goes to zero. All individuals
gained from economic growth between the two peak years of the 1980s.
The distribution of economic
Unlike previous periods of economic growth, the decade of the 1980s
and to some extent the current expansion of the 1990s frequently are characterized
as periods when the wealthiest have gained while those in the middle and
at the bottom have stagnated or lost. However, BCD have shown that this
view of recent periods is sensitive to the years compared, the income
sharing unit chosen, and the inflation index used. Their analysis shows
that previous studies using peak-to-trough years of comparison, a narrow
definition of the income sharing unit, and an upwardly biased cost-of-living
index have exaggerated the number of people who lost ground during the
1980s. When other reasonable measures are used, the 1980s is more similar
to previous decades in which economic growth lifted most, if not all,
What does this tell us about the 1990s? Simply this: while official
statistics are just now revealing the benefits of the current expansion,
these benefits may have been present for some time.
Mary C. Daly
Boskin, M. 1995. "Toward a More Accurate Measure of the Cost of
Living." Interim Report to the Senate Finance Committee from the
Advisory Commission to Study the Consumer Price Index, September 15.
Burkhauser, R.V., A.D. Crews, and M.C. Daly. 1997. "Recounting
Winners and Losers in the 1980s: A Critique of Income Distribution Measurement
Methodology." Economics Letters 54: 35-40.
Daly, Mary C. 1997. "The 'Shrinking' Middle Class?"
FRBSF Economic Letter 97-07.
Danziger, Sheldon, and Peter Gottschalk. 1995. American Unequal.
Cambridge, MA: Harvard University Press.
Jencks, Christopher, and Susan E. Mayer. 1996. "Do Official Poverty
Rates Provide Useful Information about Trends in Children's Economic Welfare?"
Opinions expressed in this newsletter do not necessarily reflect
the views of the management of the Federal Reserve Bank of San Francisco,
or of the Board of Governors of the Federal Reserve System. Editorial
comments may be addressed to the editor or to the author. Mail comments
Federal Reserve Bank of San Francisco
P.O. Box 7702
San Francisco, CA 94120