Does Class Still Explain Anything?

For forty years there has been a serious argument that class is finished as an explanatory concept. It is not a lazy argument — it was made by capable people with real evidence, and several of its claims are simply true.

And at the same time the classifications of 8.2.2 keep working, sorting populations into groups whose life expectancies, children's chances and job security differ steeply and consistently.

Both of those things are the case. This lesson is about how, and about what a "class effect" actually is when almost nobody uses the word about themselves.

Two sets of facts, from the same countries, over the same decades.

The first set.

Union membership has collapsed across most rich democracies — in Britain from around half the workforce at the end of the 1970s to roughly a fifth today; in the American private sector to under one worker in fifteen. The institutions through which class was organised have largely gone.

Voting has dealigned. The strength of the association between class position and party support fell sharply across Western democracies from the 1960s onwards. In several countries the relationship has not merely weakened but partly inverted: parties of the left now draw disproportionate support from the highly educated, and parties of the right increasingly from voters without degrees — a reversal of the twentieth-century pattern that shows up consistently across a large set of democracies.

And class identity is weak. Asked to place themselves, most people in most countries choose the middle, whatever their position (see 8.1.1). The industrial neighbourhoods, the working men's clubs, the shared workplace of two thousand people — the social formations in which class was lived — have substantially disappeared.

Anyone looking only at this set would conclude the concept had expired.

The second set.

In the Whitehall studies of British civil servants, mortality followed employment grade in a continuous gradient from top to bottom. Not a gap between the poor and everyone else — a step at every level. Men in the lowest grades had roughly three times the mortality of those in the highest over the follow-up period. And nobody in the study was poor. They were all employed, all salaried, all with secure jobs, pensions and health care. The gradient was not about poverty. It ran all the way up.

Children's educational transitions remain steeply patterned by parental class, at every branch point, and — the finding that matters most — the pattern persists after controlling for measured attainment. Two children performing identically at fifteen have different probabilities of continuing, depending on where they come from.

Wealth is concentrated and inherited (see 8.1.2). Security of employment differs by position. What happens to people in a downturn differs by position.

None of these gradients has closed. Several have widened.

So: class explains a great deal less than it did about how people think, vote and organise — and no less at all about what happens to them. Any account that cannot hold both of those is not describing the situation.

What is a class effect, when nobody recognises the category?

This is the real question, and it is a methodological one before it is a political one.

We are used to explanations that run through consciousness: people know their situation, form an identity, share it with others, and act. When that chain broke, it looked as though the explanation had broken.

But a category can be causally efficacious without being subjectively salient (see 7.1.3). Nobody has to identify as being in a particular position for that position to determine what resources they have, what networks they can reach, what an employer assumes about them, what they can risk, and what happens to their children.

So the honest formulation splits the question in two, and the two have different answers.

Does class explain life chances? — outcomes, resources, health, security, transmission. Yes, robustly, and it does not require anyone to think about class at all.

Does class explain collective action? — voting, organising, movements, solidarity. Only through consciousness, and where consciousness has weakened, so has the explanatory power.

Which is Marx's distinction between class-in-itself and class-for-itself (see 4.1.8), restated as an empirical finding rather than a stage of history. The first is intact. The second is not.

The case against class, taken seriously

Five claims, each with real evidence behind it.

One — structural change. Manufacturing employment collapsed in the rich economies; service employment expanded; the large single-employer workplace fragmented into contracting chains. The social formation that carried class — the industrial neighbourhood with its union, its club, its shared employer and its intergenerational occupational inheritance — has largely gone.

Two — individualisation. The argument from 5.9.1: in late modernity, risks are individualised and biographies become elective. People are required to construct their own path, and to experience failure as personal rather than structural. Beck's formulation was sharp — a capitalism with individualised inequality rather than class formations — and he later described class as a "zombie category": dead, and still walking around the discipline.

Three — consumption over production. The claim that identity is now formed through what people consume rather than what they produce, and that consumption communities cut across production positions. Stated most directly in the argument that class had died and been replaced by status-conventional groupings organised around lifestyle.

Four — political dealignment. Documented above, and not in dispute.

Five — cultural embarrassment. Overt class distinction became socially awkward; people avoid the vocabulary; "we're all middle class now" became a commonplace precisely as the distribution widened.

These are not weak claims. Four of the five are largely correct as descriptions. The question is what they establish.

What each claim actually establishes, and where it overreaches.

On structural change. It confuses the decline of one class formation with the decline of class. Industrial manual work shrank; that is a change in the composition of the class structure, not its disappearance. A warehouse worker on a zero-hours contract and a hospital consultant occupy class positions, and the distance between them is not smaller than the distance between a miner and a mill owner. What went was a particular set of institutions through which one position was organised.

On individualisation. This is the most important correction in the lesson, and Beck half made it himself. The distribution did not individualise. The biography did. People increasingly experience and narrate their situation as personal — as a matter of their own choices, effort and luck — while the distribution of outcomes remains as structured as before, and in some respects more so.

Which means individualisation is a finding about consciousness that is fully compatible with unchanged structure — and, worse, it is a mechanism by which structure is maintained, because a person who reads a structural outcome as a personal failure does not organise (see 1.1.1: this is precisely troubles and issues).

On consumption. Consumption requires resources, and consumption patterns are among the most steeply class-patterned things there are (see 5.8.1). The claim that identity has moved to consumption does not establish that identity has moved away from class — it may establish that class now operates through consumption, which is Bourdieu's argument rather than a refutation of it. And the recent evidence on the splitting of cultural capital into a highbrow form and a newer form concentrated among the young and educated (see 8.2.2) is a finding about how class-linked culture has changed shape, not dissolved.

On dealignment. Two corrections. First, party alignment is a two-sided relationship: when parties converge on economic policy, the class basis of voting weakens without anyone's position changing. Dealignment is partly a fact about what was on offer.

Second, and more decisively, the cleavage that has partly replaced it is itself class-structured. Education, housing tenure, security of employment and place are the dividing lines of contemporary politics in many democracies — and every one of them is a component or a consequence of class position. A politics divided by graduate status is not a politics that has escaped class; it is one in which a particular class asset has become the salient line.

On cultural embarrassment. A finding about discourse. The disappearance of a vocabulary is not the disappearance of a structure, and inferring the second from the first is the error this whole lesson exists to name.

Why the gradients persist: the mechanisms that need no consciousness.

If class is not an identity, what is it doing? Seven mechanisms, none of which requires anyone to think about class.

Resources and the floor. Not just income but what happens when something goes wrong. A person with savings, family who can lend, and a property behind them can take a risk, wait for a better job, retrain, or survive an illness. A person without can do none of these. The same shock produces different trajectories, and this is most of what "life chances" means.

Networks and information. Who you know determines what you hear about, who vouches for you, and what seems possible (see 2.2.7). This operates entirely without anyone's awareness of the process.

Employer practice. Recruitment, promotion and assessment respond to signals — accent, dress, school, comportment, the "fit" judgement — which are class markers being read as competence (see 7.4.2 on how this is measured).

Credential access and its costs. Who can afford the unpaid internship, the postgraduate year, the qualification, the relocation.

Place. Where you can live determines which schools, which labour market, which infrastructure, which exposures (see 7.4.2 on neighbourhood effects).

Control and autonomy at work — which the Whitehall evidence indicates is a substantial part of the health gradient. Low control over one's own work is a feature of position, and it is embodied.

And inheritance, in both senses: transferred wealth, and transferred expectations. Wealth transmits directly and compounds (see 8.1.2), which is why the wealth gradient is steeper than the income gradient and why it crosses generations more reliably.

Notice what these have in common. Every one operates through resources, relations and institutional practice rather than through what people believe about themselves. They would continue to operate in a society in which the word "class" had been forgotten entirely — which is approximately the society being described.

Three errors, and one is committed by each side.

Inferring the death of a structure from the death of a discourse. People stopped talking about class; the gradients did not move. The discourse and the structure are two objects (see 8.2.1's three meanings), and this confusion is the death-of-class argument's central weakness.

And the mirror error: insisting class explains everything. It does not explain the education cleavage on its own, or the geography of political realignment, or the specific dynamics of racial and gender inequality, which have their own mechanisms and are not reducible to position (see 8.4, 8.5, 8.7). A class analysis that treats every other axis as derivative is making an empirical claim, and it should be required to defend it with evidence rather than with priority.

And "class is back" as a substitute for measuring. The concept's return to fashion has produced a great deal of writing that asserts class matters without specifying which scheme, which outcome, or how much (see 8.2.2). The strength of the class case is that it is measurable and that it measures well. That advantage is thrown away by anyone using it as an incantation.

Because the honest position is more interesting than either side of the argument, and it is specific.

Class explains life chances without consciousness, and collective action only with it.

For mortality, morbidity, children's transitions, wealth, security and what a shock does to you, class position is among the strongest predictors social science has, it works through mechanisms that require no self-identification, and the gradients have not closed.

For voting, organising, solidarity and political identity, the explanatory power depends on whether people understand their situation in those terms — and that has weakened substantially, for reasons including the disappearance of the institutions that produced the understanding.

Three practical consequences.

When someone says class no longer matters, ask: for what outcome? The answer is usually about politics or identity, where they may well be right, and it is being generalised to outcomes where they are not.

When someone says class explains it, ask: through which mechanism? Resources, networks, employer practice, credentials, place, control, inheritance — naming one converts a claim into something checkable (see 7.9.1).

And notice which way the individualisation finding cuts. A society in which structured outcomes are experienced as personal outcomes is not a society where class has weakened. It is one where the strongest link in the chain — from position to outcome — is intact, and the weakest link — from outcome to understanding to action — has been cut.

Whether that is the death of class or its most effective form is a question this course will leave open (see 5.5.1 on hegemony, and 12.1). But it is a much better question than "is class dead", and it is the one the evidence actually poses.

Two sets of facts hold simultaneously. Union density has collapsed, class voting has dealigned and in several democracies partly inverted along educational lines, and class identity is weak. And the Whitehall studies found a continuous mortality gradient by employment grade among people none of whom were poor — roughly threefold from bottom to top — while class gradients in children's transitions, wealth and security persist and several have widened.

A category can cause outcomes without being subjectively salient. So the question splits: class explains life chances without requiring consciousness; it explains collective action only through it. That is class-in-itself and class-for-itself restated as an empirical finding.

Five claims against class, each with evidence: structural change, individualisation, consumption over production, political dealignment, and cultural embarrassment.

And what each establishes. Structural change altered the composition of the class structure, not its existence. Individualisation changed the biography, not the distribution — and is a mechanism by which structure is maintained, because structural outcomes read as personal failures do not produce organisation. Consumption is steeply class-patterned, so identity moving to consumption may mean class now operates through it. Dealignment is two-sided — partly a fact about what parties offered — and the cleavage that replaced it, built on education, tenure, security and place, is itself class-structured. And cultural embarrassment is a fact about discourse.

Seven mechanisms explain the gradients without consciousness: resources and the floor to fall on; networks and information; employer practice reading class markers as competence; the cost of credentials; place; control and autonomy at work, which the Whitehall evidence implicates directly in health; and inheritance, which compounds.

Three errors: inferring a structure's death from a discourse's death; insisting class explains everything, which is an empirical claim needing defence; and using "class is back" as a substitute for measuring — which discards the concept's main advantage.

Class dealignment — the weakening association between class position and party support.

Individualisation — the experiencing and narrating of structured outcomes as personal biography.

Zombie category — Beck's term for a concept whose object he took to be dead while the concept persists.

Social gradient in health — a continuous relationship between position and health running the whole length of the distribution, not a threshold effect at poverty.

Control at work — autonomy over one's own tasks and pace; implicated in the health gradient.

Primary and secondary effects in education — differences in attainment, and differences in choices at branch points conditional on attainment.

The floor — the resources available when something goes wrong, which convert an identical shock into different trajectories.

Class as position, identity and culture — the three objects whose confusion drives most of this argument.

Educational cleavage — the partly inverted political division by graduate status, itself class-structured.

One — split the question. Next time you meet the claim that class no longer matters, write down which outcome is meant. Then find one outcome where the claim is right and one where it is not.

Two — name the mechanism. Take a class gradient you know of — in health, education, or employment — and write down which of the seven mechanisms could produce it, and what evidence would separate them.

Three — test the floor. Think of a risk you have taken or declined — a job move, a course, leaving a situation. Write down what would have happened if it had gone wrong, and who would have caught you. Then imagine the same decision without that.

Four — check a dealignment claim. Find data on how your country voted by education and by occupation. Ask whether the two divisions are independent of each other; usually they are not.

Five — do the individualisation test. Take a structural outcome you know of — a redundancy, a rejection, a failed business — and write both accounts of it: the personal one and the positional one. Note which one the person involved gives, and which one the evidence supports.