Class Reproduction in America Revisited

It’s hard to be original. Turns out, your analyst’s result on the overdose gradient had more or less already been published by health researchers in 2017. Instead of looking only at deaths due to overdose, their predictor was the common component of seven public health indicators. Like me, they found a very strong relationship between poor health indicators and the swing to Trump in the cross-section of US counties.

Donald Trump won the white working class, defined as non-Hispanic whites without a college degree, by 39 percentage points. It was the largest margin since 1980. But it did not come out of the blue. Starting in the 1990s, the white working class has increasingly abandoned the Democratic Party. Even in 2008, when Obama won with a historic majority, the GOP won the white working class vote by 20 percentage points.


This is an astonishing state of affairs. The GOP is widely regarded as the party of the business elite, and its policy platform is known to be greatly harmful to working class interests. On the other hand, Democrats think of themselves as custodians of the working class interest. But the gap that has opened up is now so vast that we must, if we do not want to delude ourselves, think of the Democratic Party as the party of the elite, and Donald Trump’s GOP as the party of the masses.

There is a story that Democrats tell themselves about how this peculiar alignment came about. We are told, again and again, that Johnson killed the New Deal coalition when he signed the Civil Rights Act into law in 1964. This story is a myth. It is often conflated with the loss of the South to the Democrats. That’s also a myth. In general, there is no evidence that the White working class turned away from the Democrats before the 1990s. So it cannot be the Civil Rights Act. But why is this myth, one so easily debunked, exercise the hold that it does on the liberal imagination? As we shall see, trying to answer this question will lead us to the real source of the present impasse of elite-mass relations in the United States.

The Civil Rights Act Story is a thinly disguised way of saying that the Dems lost the white working class because the white working class is racist. This is a longstanding perception that is reproduced over and over again; most recently, to understand the catastrophe of 2016. Hundreds of researchers have claimed to show that, to take a recent study at random, “racial prejudice, anti-immigrant sentiment, concerns about economic security, and frustration with government responsiveness may have led many white, working-class voters to support an outsider candidate who campaigned on these themes.”

This is a recycling of the same myth of working class racism. Even if it were the case that working class whites are more racist than middle class whites, although there are good reasons to doubt the claim, it would still not explain 2016. For constants cannot explain variables. And there is simply no evidence of any increase in racial resentment or anti-immigrant sentiment in working class communities in the lead up to 2016.

On the other hand, the evidence is unambiguous that the white working class has been screwed over since the 1990s. As late as 1998, the white working class’s share of family income was equal to that of the white middle class. By now, its share has dropped to around half that of the white middle class. Similarly, wealth shares of the two classes were equal in 1989. The white working class share of net worth has fallen to a third of the white middle class.


Note that the white working class is twice as large as the white middle class. Actually, that’s not quite true. We must be careful not to simplify US class structure too much. The college/no-college binary is misleading because the white working class defines itself in opposition not only to the middle class but also to what Charles Murray called the permanent underclass. I tried all sorts of fancy tricks to sort households by socioeconomic status, educational attainment, occupations, income and wealth. None of that fancy slicing and dicing adds any value to the basic decomposition we obtain from educational attainment alone.

Basically, the sorts of white people who go to college are quite different from the sorts of people who never finish high school, and both differ in moral code, aspirations, aesthetic taste, child rearing philosophy, and risk appetite from the middling bulk of the population that can be identified as the white working class. In affluent middle class families, everyone has to attend college — preferably at a prestige school. In the working class, hardly anyone ever finishes college. Certainly, some do. Just as the income of some households with no college degrees is not so modest. But on the whole, American white working class has modest means. Moreover, this is reflected in its moral economy. The white underclass, the class with truly modest means, is hedonistic and lacking in morals in the eyes of the working class. On the other hand, in the eyes of the middle class, the truly poor are in the greatest need of state help. The underclass, too, inhabits a different cultural universe. The class itself is a frontier zone; a shatter-belt of misfits of all sorts; a land of trailer parks, shady neighborhoods, red light districts, and needles. Entangled with law enforcement, this class comes into its own in prisons, where its culture coexists with the culture of its black counterpart — a fascinating story that we leave for later work. The code here is the opposite of the white working class: anything goes.

It is not a coincidence that educational attainment has the strongest class signal: in a society with an elite sorted on school prestige like the present day United States, the two diplomas provide the structural parameters of class society. We should think of the tripartite class structure as a representation of this ordering principle of American class society.

People who do not finish high school are a true underclass. Nearly half, some 49 percent, report having earned no market income in the past year, as do 23 percent of the white working class and 12 percent of college graduates. This is the most recent data from the Census Bureau.

No Earnings.png

Note that the volatility of earnings has increased since the late-1980s. Precarity cuts across class boundaries — and this is reflected in the absence from the labor market in the previous chart.

Screen Shot 2019-11-09 at 5.42.18 PM.png

We’ll work with late-1990s data in the MIDUS 1 database until further notice. They report basic occupational categories whose ranking by wages, mean years of schooling, or pay grade, is identical. The three classes predominantly do different things although there is overlap. More precisely, the working class overlaps with the domains of the professional class and the underclass. The exclusive preserve of the middle class is business management (rank 1) and the elite professions (2). College-educated whites also work as technicians (3) along with many without a college degree. In sales (4) and clerical work (5), the working class predominates. But the working class does everything it can: it shares services (6), primary production (7), and to a lesser degree, assembly and maintenance (8) and day labor (9), with the underclass.

We distinguish between households where the female breadwinner earns more than her male counterpart (“female”) than vice-versa (“male”). In all three classes, female-dominant households have higher occupational status. occupational.png

The next graph is tautological. But it does show the variation in educational attainment we have packed away with our tripartite schema. Note the higher attainment of female-dominant households.


Despite the fact that female-dominant households have higher educational attainment and occupational status, the socioeconomic status of male-dominated households is higher in the underclass, and about the same in the working class and the middle class.


The empirical income distribution of the classes has different location, scale, and skew. The median income of the middle class is $77,000, of the working class, $46,000, and of the underclass, $22,500.


There is a trend in years of schooling that we must control for.


Families reproduce class culture over time by passing on their ethos and way of life. An important reproductive mechanism is assortative mating. Controlling for the trend, the child-parent and, especially spousal correlations are very strong. Table 1 displays the correlation coefficients. We omit the P values with the understanding that all estimates in bold are significant at the 1 percent level. We can see that the trend confounds the estimates somewhat. So Table 1-B is more reliable.


Each of the coefficients is worth noting. There is no gender bias for either respondent or her spouse. The inter-generational coefficients, for both father and mother, and for both spouse and respondent, cluster around 0.4-0.45. The correlation between mother and father is slightly stronger than that between the respondent and spouse, suggesting a modest decline in the rate of assortative mating. But both are in the range 0.6-0.65. The pattern of correlations suggests a very strong signal for class reproduction if classes are defined, as we suggest, by education. What is particularly interesting is that the educational attainment of the respondent’s spouse is correlated with parental attainment even after controlling for respondent’s educational attainment. The relationship is so significant that mean parental years of schooling explains 5 percent of the residual variation in spousal years of schooling in our sample of 635,715 observations from the General Social Survey.


The trend, however, is driven by the underclass catching up. There is only a weak trend in the middle class and the working class.


Is the working class closing the educational attainment gap? Is it becoming more like the middle class over time? No. The middle class was pulling away until around 1990. Since then, the series has stabilized into a trend-less cycle. The cycle looks coupled to the macroeconomic cycle — we know that working class labor markets tighten later in the cycle. That may be what is generating the evident lagged correlation pattern.


In any case, there is no evidence to suggest convergence. In order to get to the bottom of the present impasse of elite-mass relations, we must begin with a more appropriate picture of class reproduction. There is a fourth class, that of the financial elite, that we have not looked at here, but whose reproduction we have mapped before. When social democracy capitulated upwards, what obtained was an alliance of the middle class with the financial elite around a program of socioeconomic liberalism anchored in a paradigm of neoliberal antiracism. Although that capitulation is an important part of how we got here, what prevents the restoration of elite-mass relations is not the neoliberal component of that synthesis. The blockage is rather the rise of Boasian antiracism in the thick sense, which sees the working class worldview as inherently racist. The language of class oppression that attends this attitude is deeply resented by the working class.

Remember “the authoritarians” responsible for the catastrophe of 2016? Turns out, that’s just picking up the White working class signal:

White working-class Americans are more likely than white college-educated
Americans to value deference to authority over autonomy. When asked to choose between pairs of traits that emphasize children’s autonomy versus deference to parental authority, nearly two-thirds (65%) of white working-class Americans favor authoritarian childrearing traits, compared to around 4-in-10 (41%) white college-educated Americans.

What are called “authoritarian personality types” in the sample surveys aren’t independent psycho-analytic subjects but members of the dominant everyday culture of the United States. Neither poor, nor rich, this is the middling bulk of the US population. Unlike the status-obsessed professional classes who are driven by self-actualization, the middling bulk makes meaning from family, discipline, and hard work. Their jobs involve not self-actualization but drudgery, and maybe, hopefully, some skilled work.

What I want to explore is the moral economy of the white working class. I want to understand how the white working class makes meaning. Building inter alia on Michèle Lamont’s The Dignity of Working Men, Joan Williams argued in White Working Class that, in our terms, the middle class and the working class have different cultural repertoires. What makes sense in one does not in the other. This is important because of what she politely calls the epidemic of class cluelessness.


Take the standard professional-class ice-breaker: “What do you do?” It makes sense in a class context where personal dignity stems from economic power and professional achievement. When people ask me, I reply, “I’m a law professor.” But that kind of honor is available to only a few in the working class—to firefighters, police, soldiers. For most, the dignity work affords is from what it allows you to buy and whom it allows you to support, not from the job itself. “What do you do?” is not the first question at a party. I remember attending my class-migrant husband’s high school reunion when, with a regrettable lack of code switching, he posed the “What do you do?” question to a classmate. The classmate’s face got very red as he came right up into Jim’s face and hissed, “I sell toilets.”

The working class doesn’t make meaning from self-actualization; it makes meaning from self-discipline and self-control — ‘at least he gets up and goes to work every morning’. For the working class, work is not self-actualization but, more often than not, drudgery; and valued instrumentally for provision. The core moral value of the working class is an existential strategy — you’re one job-loss away from food stamps. White working class identity is defined in opposition not only to the middle class but also to the underclass. Both are held to be immoral. The former for their duplicity, self-obsession and superiority complex; the latter for their self-indulgence and dependence; and both for neglecting family. In contrast to both, the working class is perceives itself as straight-talking, independent, disciplined, hard-working, egalitarian, family-oriented. That’s the ideology of the white working class.

The middle class conflates the underclass with the working class — this is intolerable to the latter precisely because they are not seen as status competitors but beyond the pale. The underclass is outside the moral boundary drawn around the working class. They are the ‘hard living’ whites, as opposed to ‘hard working’ whites. The working class defines itself in opposition to, and morally superior to, both the underclass and the elite. In the moral economy of the white working class, both black and white underclasses are looked down upon for the same reason — they lack the self-discipline that is so central to the white working class ethos. Class is more salient than race but both reinforce each other.

To project the meaning making of the white working class onto race relations is profoundly mistaken. It’s the common assumption of Boasian antiracism and perhaps the source of the middle class’s insulting perception of the working class as racist. The middle class has adopted the discourse of academic Boasians. But the notion that it is not racist is complete nonsense. The truth is that the middle class is racist too, including all your woke antiracist warriors. This is clear from blind tests — the Harvard study, resume responses, online dating responses, porn, you name it. The working class hasn’t learned the Boasian language, largely because it is taught in college. But Boasians are clueless about the history of Boasian antiracism. Whence the profound misunderstanding.

Unlike the middle class, the core value of the working class is not self-actualization, but dignity. Unless liberals drop the Boasian scold and treat the working class with the dignity and respect it deserves, elite-mass relations cannot be restored. And the Democratic Party will remain the party of the elite, whether or not Donald Trump is defeated in 2020.   

Postscript. The General Social Survey admits the computation of intergenerational class transition probabilities. We redefine classes by discretizing mean years of school completed of the husband and wife — less than 12, underclass;  at least 12 but less than 16, working class; at least 16, middle class. Then we compute the transition probabilities by comparing parents to their children’s households. The next chart reports the transition probabilities averaged over 2010-2018.


We can see that the children of 60 percent of working class households remain working class; 32 percent transition to the middle class; and 8 percent fall through the cracks into the underclass. In the middle class, 61 percent remain; 37 percent transition to the working class; and 3 percent fall directly to the underclass. The pattern for the underclass is very different. Only 27 percent stay within the underclass; 56 percent transition to the working class; and 16 percent transition straight to the middle class. So the flux in and out of the underclass is dramatic. The underclass is a frontier zone or counter-zone — a shatter-belt of social rejects; falling into dire straits one generation, and more often than not, pulling themselves out in the next.

The next graph shows the diachronic pattern of the transitional probabilities. What is striking is that, back in the seventies, the under class and the middle class, had less definition in the following sense. The odds were roughly even that someone born in the underclass would remain; that’s halved. The odds of dropping from the middle class to the working class were also even; they, too, have since fallen significantly.


For the working class, back in the seventies, you had a 20 percent chance of rising up or falling down. Now, working class families have a 30 percent chance of moving up in life and only a 10 percent chance of falling through the cracks. For the middle class, back in the seventies, you had 50 percent chance of falling into the working class. Now, middle class families face a 40 percent chance of doing so. The working class has maintained its high rate of reproduction. Note that these numbers are confounded by the secular trend of rising education level across classes.





16 thoughts on “Class Reproduction in America Revisited

  1. For the middle class to treat the working class with dignity they would have to humanize them, (the exact opposite of this). And that can’t happen because it would create too much cognitive dissonance with liberals who justify their lives of privilege on the categorically false assumption that this is a meritocracy; and that the only thing holding anyone back from achieving their full potential is racism, sexism, ect.

  2. Damn UserFreiendly, that was a mighty fine article that you linked to, Unfortunately it brought to my mind the question; What is the opposite of “Preaching to the Choir?”. When someone is Preaching to the Choir, they are generally pictured to be turned around “backwards” ignoring those whose lives and souls have been entrusted to them for spiritual renewal and guidance, and concentrating on those who are assumed to have achieved level of Spiritual maturity and position, (after all, is not the Choir part of the Church Establishment itself?). When someone is “Preaching to the Congregation” they are faced forward entreating these same souls and lives to move onward in their Christian walk. Rebuking when necessary but always ready to offer comfort and solace to whomever asks. But all of this is done face to face with the congregants, making an argument, entreating them, earning their trust. Unfortunately, for the author of the Vox article, the congregation has been “ex – communicated” so to speak by the very choir that sits behind him. And this choir has no intention of subjecting themselves to his sermons. That would be beneath them. The author of the Vox is article basically left alone, in his study, preaching to no one. Because no one (in this particle Church) wants to hear it.

  3. Policy Tensor, you write the following:

    “This is an astonishing state of affairs. The GOP is widely regarded as the party of the business elite, and its policy platform is known to be greatly harmful to working class interests. On the other hand, Democrats think of themselves as custodians of the working class interest. But the gap that has opened up is now so vast that we must, if we do not want to delude ourselves, think of the Democratic Party as the party of the elite, and Donald Trump’s GOP as the party of the masses.”

    I think that at this point it is rather obvious that the GOP “is” the party of the masses, and that the Democratic Party is the party of the “elite”. And the parties represent the interests of their constituents, both their economic and cultural interests. Do you disagree with this? Or do you believe that the Democratic Party is still the best party that has the “best interests” of the masses in their mind and is the party that can best carry out plans and policies ensure that the masses interests are looked after? I am not trying to be argumentative or “trollish” with you. This is a very honest question.

  4. I don’t know if this has anything to do or relates with the subject matter that PolicyTensor has been writing about for the past week or so, but I will just offer the following link as a possible example of, or data point in the rather obvious break down in Elite / mass relations.
    This just occurred in my beloved place of residence. I am not surprised or shocked that such a thing could happen, at all. Note particularly the descriptors used to identify the two protagonists in the story, “the woman” and ” right wing demonstrators”.

  5. Who are the Boasians in “But Boasians are clueless about the history of Boasian antiracism.”? Did you mean the middle class, Boasian antiracists, academic Boasians or something else?

      1. You do, above and elsewhere on this blog. I am persuaded that your analysis of elite-mass relations is correct. Why do you ask?

  6. Can you explain why you think Boasian antiracism specifically, rather than a broader set of cultural tensions surrounding education level and class, is responsible for the alienation of the working class from upper middle class liberals?

    1. Hi NPR,

      I’m saying that the ‘broader set of cultural tensions surrounding education level and class’ is due to the rise of Boasian antiracism. Boasian antiracism sensu stricto is, obviously, a honorable politico-ethical stance. But, in the thick sense, it is overlaid with snobbery and upper class racism. And in that sense it functions as class ideology.

      I think we are blinded by Donald J. Trump. He is such a caricature of the Boasian nightmare that we can’t see beyond the surface. But you’re right, of course: “We expected this of them” — we’ve known all along they were racist. This was just a matter of time, and so on. But the contempt goes deeper than the presumption that the white working class is racist. In middle class ideology, the white working class is not simply racist and bigoted, it is DUMB. This shared diagnosis of the American middle class became manifest in the reception of The Bell Curve; a highly significant and revealing event.

      Every Boasian with a pen took great offense at the single chapter that replicated known results in interracial differences in cognitive test scores — an explanandum; immediately reduced to biology by champions and critics alike. “The question” as Deparle put it in the New York Times Magazine, “is whether, in his zeal to shatter taboos, Murray has finally gone too far.” Murray would be roasted for the next 25 years for that chapter. But what was the book actually saying?

      What Murray and Herrnstein argued, if effect, is that, because we are a meritocracy, your station in life is determined to a great extent by your genomic endowment. In other words, if you went to a prestige school it’s because you were born smart. And the loser sitting on his couch, drinking beer, and watching stupid shows on the telly all day, is there, doing that, because of his poor genomic endowment. NO ONE TOOK OFFENSE AT THIS. The authors biologized class hierarchy, taking us back to the good ol’ Victorian days (‘blood tells’) — and Boasians had no problem with that. Why did no one find this offensive? No one found the reinscription of biology into class relations offensive because they agreed with it! So, already in 1994, Boasian antiracism in the thick sense, had emerged as the class ideology of the social elite.

      1. What I’m asking is why you assume that Boasian antiracism is causally central to classism among liberals? It seems more plausible to me that classism is a fairly natural byproduct of class divisions. The influx of college-educated whites into the Democratic Party since the 1960s, combined with the increasing stratification of the economy based on educational achievement, seems sufficient on its own to explain class tensions. If we were having this same conversion in 2005, we’d probably be talking about how middle class liberals stereotype working class people as religious fundamentalists, rather than as racists.

        1. These class tensions you speak of are not as natural as they look. They have their own history, which, as I’ve argued, is entangled with the intellectual history of Western elites. Stratification alone is not enough to explain the present impasse of elite-mass relations. Nor can it explain the peculiar alignment of sociopolitical forces. What explains the hostility of the white working class to the social elite (as opposed to the financial elite)? Why did it abandon the Democratic Party from the 1990s onward? Why did it throw its weight behind Donald Trump? If you think there is no puzzle to be explained then I don’t have much to offer you. In any case, you seem uninterested in revisiting your basic premises, so I won’t press the case. Revisit this issue when the breakdown of elite-mass relations seems more puzzling to you.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s