Here's the article:
Living Standards in the United States
The ninth seminar in AEI's series Understanding Economic Inequality was led by Daniel
T. Slesnick of the
University of Texas at Austin. Mr. Slesnick argued that consumption is a more relevant
measure than income for
gauging economic inequality, and he showed that consumption data suggest little change
in inequality over the past
twenty-five years. A summary of his presentation follows.
Economic inequality and its growth are almost always measured in terms of income. In
recent years, considerable
publicity has been given to studies indicating that since the early 1970s median family
income has been stagnant,
inequality has been increasing, and the poverty rate has stopped declining. But
material well-being is actually a
function of consumption, not of income. Assuming consumption could be measured
properly, what would it show
about trends in standards of living in the United States over the past quarter-century
and trends in the gap between
rich and poor?
Research into these questions yields striking results. Measures of consumption show
that--contrary to what is
generally believed--living standards for rich and poor alike have continued to rise
since 1970 and that the disparity
between rich and poor has increased very little.
Measuring trends in inequality is difficult for many reasons, and those difficulties
have an important bearing on the
controversy over inequality, as other scholars in this series have shown. Measurement
problems are particularly
significant in the case of income, but they affect analysis of consumption as well.
Consider the following issues:
How should housing be treated? A homeowner whose mortgage is paid off enjoys the use
of his house but
makes no cash outlay for that use. He consumes but does not spend. The same problem
arises on a smaller scale
for other consumer durables, such as automobiles.
What price index should one use to measure consumption in real (that is,
inflation-adjusted) terms? There are
problems with the consumer price index, including the widely acknowledged flaws in its
treatment of housing
before 1983, which exaggerated inflation.
How does one adjust for family size? Does a family of ten require exactly five times
the consumption of a family
of two? Clearly not, but what is the right adjustment for "equivalence"? What about the
households not composed
of families?
After accounting for these and other measurement problems, I have developed measures of
economic well-being
based on data on personal consumption expenditures in the national income accounts.
Real median family income, the traditional measure of economic well-being, increased by
an average of 2.7 percent
per year from 1947 to 1970 but stopped rising after 1970. Real per capita consumption,
however, kept on
growing. It rose by 2.4 percent per year from 1947 to 1970 and by only a little
less--2.1 percent--from 1970 to
1993.
What accounts for the divergence in these trends? Part of the explanation lies in
changes in family size. If a family
of four and a family of three have the same income, the standard of living--the per
capita availability of income for
consumption--of the family of three is obviously higher. This would not matter if
average family size were stable,
but in fact it has been declining over the past quarter-century. In addition, measuring
changes in real median family
income ignores the nearly one-third of the population composed of "unrelated
individuals," that is, persons not
living with family.
Another part of the answer lies in the aforementioned flaws in the consumer price
index, which exaggerated
inflation before 1983 and, as a result, decreased measures of real income. The deflator
for personal consumption
expenditures in the national income accounts, which is used to eliminate inflation from
measures of consumption,
does not have this problem.
The argument that standards of living have been rising since 1970 finds further support
in Engel's law, which with
plentiful historical evidence posits an inverse relationship between one's standard of
living and the share of total
spending one devotes to food. Reported decreases in the proportion of total expenses
households devote to food
therefore support the view that standards of living have continued to rise in the past
quarter-century.
Measures of inequality based on the survey of consumer expenditures also show a quite
different trend from those
based on income. According to the latter, inequality has increased strikingly. For
example, the ratio of income in
the top three deciles to income in the bottom three rose from 5.4 in 1961 to 7.9 in
1992. But when consumption is
substituted for income as the basis for measuring inequality, the ratio scarcely
changes--it was 3.9 in 1961 and 4.0
in 1992.
Incomes fluctuate more than consumption does from one year to the next. People do not
decrease their spending
significantly when they regard a loss of income as temporary. Material welfare is best
measured, where possible,
by lifetime or "permanent" income. Since consumption is a more accurate proxy for
permanent income than annual
income is, this is still another reason to prefer consumption to income as a measure of
material welfare.
These varied approaches to the problem of assessing economic inequality all point to
the conclusion that both the
level of economic well-being in America and its distribution have performed far better
during the past
quarter-century if consumption rather than family income measures are used. Moreover,
on several counts
consumption is a more appropriate measure of economic well-being than income.