Over at The Atlantic, Annie Lowrey discusses the dark cloud that ostensibly hides the silver lining of increasing household incomes in the U.S.: “This crisis involved not just what families earned but the other half of the ledger, too—how they spent their earnings . . . . [T]he spiraling cost of living has become a central facet of American economic life.” She writes that while “Wages picked up for high-income workers [a decade ago] . . . and picked up for lower-income workers in the second half of the decade,” prices for critical purchases such as housing and health care increased even faster, resulting in many Americans being worse off despite their increasing incomes.
The problem with Lowrey’s argument is that the data showing that wages increased during the period she discusses already control for the price increases she highlights. Her argument essentially double-counts costs. This doesn’t mean that the problems Lowrey discusses don’t exist for many families in particular regions of the country—even simple averages are sometimes trickier to interpret for an entire population than they seem—but the data don’t really show that an increasing cost of living is impoverishing families.
We can see this by more closely looking at the data that Lowrey cites on wages and income. For her observation that wages “picked up for lower-income workers in the second half of the decade” Lowery links a 2017 story in The Atlantic written by Derek Thompson, “The Most Underrated Story about the U.S. Economy.” Thompson in turn draws on data from the (leftish) Economic Policy Institute and the Atlanta Federal Reserve Bank. Thompson reports that income is increasing for U.S. workers across the board, but that income growth was even faster for lower-income workers in the U.S. than for higher-income workers.
The Atlanta Fed’s data report “nominal wage growth of individuals.” “Nominal” means that the Fed’s data do not take into account the impact of changing prices on incomes. Depending on the change in prices, the Atlanta Fed’s data could be consistent with declining living standards. If prices increase faster than income, then living standards would decline despite increasing nominal incomes. Whether living standards have increased or decreased given the increasing income depends on how the cost of living has been changing relative to the reported income gains.
But the data from the Economic Policy Institute reports wages in “constant” 2016 dollars. This means that the EPI reports wages adjusted for price changes so that wages from different years can be compared relative to the same price scale.
The EPI Report includes a methodological appendix that explains the data the Report used in its calculations. It notes that “Wages are deflated by the personal consumption expenditures (PCE) index for all items, except health, which is deflated by the PCE medical index.” Further, some benefits provided by employers are deflated in the EPI report using the Consumer Price Index rather than the PCE Index. While there are differences in the way the CPI and the PCE Index are calculated I’ll discuss only the PCE Index for brevity.
Lowery discusses four particular costs in her analysis: the cost of a college education, housing costs, health-care costs, and child-care costs. As she writes:
It all adds up, and it all subtracts from families’ well-being. The price tags for tuition and fees at colleges and universities have risen twice as fast as wages, if not more, in recent years. Rental costs are outpacing wage gains by a percentage point or more a year. Health-care costs have grown twice as fast as workers’ wages. And child-care costs have exploded. These costs pressures are particularly acute on young Americans who have seen worse employment prospects and smaller raises than their older counterparts.
There is, first, the narrow methodological point that the PCE price index has already taken all of these costs into account. Secondly, the broader point is that price systems, whether prices are set by the market or by a central planning (or regulation), will always be changing. Sometimes increasing and sometimes decreasing. Selectively discussing a set of goods and services that are increasing at a rate faster than incomes are increasing does not mean that real incomes are decreasing overall. What matters is the change of the whole set of prices relative to changes in income.
The Bureau of Economic Analysis measures Personal Consumption Expenditures by Americans. As described by the BEA’s 69-page description of what goes into calculating the Personal Consumption Expenditure, the PCE “is the primary measure of consumer spending on goods and services in the U.S. economy” as well as a “comprehensive measure of types of goods and services that are purchased by households.” Basically, the PCE tracks the cost of purchasing the comprehensive “basket of goods and services” that American households purchase. That’s the point of the measure.
Because it aims be comprehensive, the PCE naturally includes household expenditures in the areas on which Lowery focuses (again: education, housing, health care, and child care).
When the PCE is used to translate Americans’ nominal income into “constant” dollars so that incomes can be compared across years, it already takes into account changes of prices in education, housing, health care, and child care, as well as changing prices in many, many other areas.
When data show an increase in income in constant dollars, as the EPI measures cited by Lowery do, this means that real incomes have in fact increased on average despite increasing prices in specific areas. When Lowery then says, “look at the increasing prices for these goods and services, this means incomes are going down,” she has her authorial thumb on one side of the scales.
That said, there is still room for a more narrowly-tailored version of Lowery’s argument. The Personal Consumption Expenditures calculate nation-wide averages. There is substantial variation by region, by class, and by other variables that belie the average.
For example, housing in specific areas like San Francisco may increase much more than the national average reflected in the PCE. It is entirely possible that housing costs have increased so significantly in San Francisco that the increase by itself results in declining real income for average residents of San Francisco relative to average gains in income.
But relative prices are always changing. Some prices increase and some prices decrease. That the data show that incomes have been increasing in constant dollars means that, overall, incomes are increasing faster than costs overall, costs that include expenditures on education, housing, health care, and childcare.
That prices are always changing is generally a good thing, because prices do not merely allocate goods and services, but also signal information to people on all sides of transactions.
There are, to be sure, supply constraints on housing in San Francisco induced by restrictive zoning that mess significantly with prices. But another part of the issue is that even without restricting zoning, housing prices in San Francisco would likely be higher than, say, prices in central Texas. The scrubby beauty of central Texas seems to be a refined taste; all things being equal, more people desire to live on the coast of the Pacific Ocean than to live on the banks of the Brazos River.
That housing prices will reflect this preference is not a problem unique to market pricing. As Deidre McCloskey reminded in her recent book, Why Liberalism Works, there was a lengthy, scholarly, and high-minded debate in the 1930s between socialist economists and Ludwig von Mises (in particular) regarding optimal pricing and production decisions in fully socialist economies. The debate ended with socialist economists recognizing Mises’ point that optimal decisions in socialist economies will require a system of socialist pricing. The outcome of the debate can be seen in the sophisticated theory of socialist pricing advanced in works such as Oskar Lange and Fred Taylor’s On the Economic Theory of Socialism. Lange opined only half tongue-in-cheek that “a statue of Professor Mises ought to occupy an honorable place in the great hall of the Ministry of Socialization or of the Central Planning Board of the Socialist state.”
How do prices work as signals regarding housing in San Francisco? On the one hand, relatively high housing prices in San Francisco signal a growing shortage, and so invite builders to build more housing (if San Francisco will allow them). Housing supply would then increase and housing costs would decrease. Simultaneously, cheaper housing in other parts of the nation can draw workers from San Francisco, thereby reducing the supply of labor and increasing wages in San Francisco to the point where ordinary workers could once again afford housing to live there.
This does not mean there’s no role for government policy. But prices provide information about government failure—such as unintended consequences of zoning and other regulations—as much as they provide information about market needs.