White, still: The American upper middle class

The American upper middle class is separating along educational, economic, cultural and geographical lines. At least, that’s my contention in Dream Hoarders. (For some voices of disagreement, look here or here…) My focus in the book is on class, rather than race, but of course the two are deeply intertwined. Many of the unfair “opportunity hoarding” mechanisms used by the upper middle class have racist origins – exclusionary zoning, legacy preferences, for instance – and now serve to entrench class positions, as well as racial gaps.

In a society moving towards greater racial equity as well as greater racial diversity, we would expect to see the uppermost classes becoming more heterogeneous. So: how white is the upper middle class (defined here, as in Dream Hoarders, as the top income quintile)? Is it less white than in the past? If so, which racial group or groups are moving onto the top ladder? These are the questions we address here.

How many are upper middle class?

First we calculate the share of the people in each racial category who are in the top quintile. These have been reasonably stable over time, with essentially no change in the proportion of whites between 1980 and 2016. The proportion of black and Hispanic 40-somethings in the upper middle class has not increased in recent decades. Again the exception is for Asians, of whom 29 percent were upper middle class in 2016, up from 19 percent in 2003. In fact, Asians are now more likely than whites to be in the upper middle class (at least based on the figures from 2016):


Whites and Asians overrepresented at the top

Next we examine the racial composition of Americans in the top quintile of the household income distribution, for households headed by a person aged between 40 and 50. (We restrict our analysis to this age cohort, because there are significant differences in the age profile of different racial groups).*

The American upper middle class remains largely white: while 62 percent of Americans aged between 40 and 50 are white, 72 percent of those in the top fifth of the income distribution (for this age group) are white:

Untitled2The story is the opposite for blacks and Hispanics. Hispanics make up approximately 17 percent of the overall population aged 40 to 50, but just 9 percent of the upper middle class. For blacks, the numbers are 12 percent and 7 percent, respectively. The only minority that bucks the trend is Asians, who make up 7 percent of the overall population but 11 percent of the upper middle class.

is the upper middle class becoming more diverse? not really.

It is perhaps not surprising that the top rung of the income ladder is disproportionately white, given stubborn race gaps in upward mobility, housing segregation and its legacy, and educational inequalities. But it would be reasonable to expect the picture to have improved over time. But, with the notable exception of the ascent of Asian Americans, it has not.

The racial composition of the top quintile compared to the whole household income distribution (for our age cohort) shows considerable stability over time. The graph below shows the share of a racial group in population subtracted from its share in the top 20 percent:


Black underrepresentation in the top quintile has also remained essentially stable over time, a mirror image of the picture for whites. Asians have meanwhile increased their representation on the top income rung, while Hispanic representation has fallen (with a bigger increase in the overall Hispanic population than in the top quintile).

Of course, class is not just a question of money, but also of occupational status, culture, and education. We repeated our analysis but with a slightly narrower definition of the upper middle class: those who both have a top-quintile household income and at least a bachelor’s degree. The pattern is essentially the same, except for an increase in the over-representation of Asian Americans:


These trends likely reflect both the economic success of 2nd generation immigrants and “hyper-selectivity” (i.e. high education levels) of Asian immigrant populations. Equally, some of the decline in the Hispanic share at the top may be due to a reduced likelihood to identify as Hispanic with each passing generation. But the broad picture is clear: whites still dominate the top of the income distribution.

the myth of white eclipse

Membership of the upper middle class – the group I focus on in my book Dream Hoarders – remains overwhelmingly white; almost as white, compared to the population, as it was 25 years ago. Policies and practices that help this class, from “upside down” tax expenditures to exclusionary zoning, therefore also contribute to the race gap.

There is plenty of discussion now of the problems faced by whites, especially by comparison to other racial groups. President Trump’s most enthusiastic supporters were those who are most likely to believe that whites now face more discrimination than other races. As Ta-Nehisi Coates points out in powerful new piece in The Atlantic, “The First White President”, Trump voters had an median household income of $72,000. The idea that whites are somehow being eclipsed by other races is disproved decisively, on an almost daily basis, by almost every social and economic indicator: most recently by census data showing a median income of $65,000 for whites and $39,000 for blacks, as well as child poverty rates of 31 percent and 11 percent, respectively.

Of course our colleague Camille Busette is right that “there is practical, intellectual, and empathetic space for two types of suffering [of poor whites and poor blacks] to be acknowledged and addressed”. But it is vitally important that policymakers continue to focus on the erosion of black wealth, race gaps in multidimensional concentrated poverty, and the historical legacy of housing policy. (On which note, see Camille’s new Race, Prosperity, and Inclusion Initiative.

* This calculation takes the total adjusted (divided by the square root of household size) household income of units with heads ages 40-50 and arranges each household according to relative position. The top 20 refers to any person age 40-50 located within the 81st to 100th percentile of that ranking.