On the Marginal Revolution blog (on August 4), Tyler
Cowen posted a link to a twitter feed,
with this header:
The link shows personal consumption expenditures as a
percent of GDP and a measure of income inequality (the Gini coefficient) from
1978. Both PCE as a percent of GDP and
the Gini coefficient have been rising—consumption has become a larger percentage
of total spending in the US and inequality has been
increasing. Cowen’s “ahem” is, I
suppose, an indication that we should be surprised that both of these have
happened at the same time.
Leaving aside the issue I raised yesterday there is also the issue of what, exactly, has been happening to consumption, if
we disaggregate households by income.
Fortunately, we can do this, using data from the Consumer Expenditure
Survey.* The CES presents measures of
after-tax income and household spending by quintiles—the 20% of households with
the lowest incomes, and so on. (Which
means each group has the same number of households in it). So we can look at what has happened both to
household (after-tax) income and to household consumption spending. And the results are interesting.
Q refers to the income quintile,
from lowest to highest. The values have
adjusted using the CPI, with 2011 used as the base year. Spending can exceed
income when households receive transfer payments or use their savings to
pay for part of their spending. Transfer
payment receipts are the obvious
source of funds in Q1 and Q2.
What this tells us is quite simple. It’s possible for consumption spending to
rise as a percentage of GDP and for income inequality to increase at the same
time. All that’s required is for the
bulk of the consumption spending to occur in high income households. And—that is exactly what happened.
Overall, income increased by about 39%...but spending
increased only by 3.8%.
While about 60%
of the total income increase went to the highest income quintile, more than 75%
of the increase in consumer spending occurred in the highest-income households.
It’s worth noting that consumption
spending actually fell
in real terms
in the lowest income quintile.