Absolute and Relative Mobility

"Social mobility" is one of the few things politicians of every persuasion claim to want. It is also a term that fuses two entirely different questions, and almost every public argument about it is people answering different ones.

Question one: am I doing better than my parents did? Question two: do my parents' circumstances matter less than they used to?

These can move in opposite directions, and in the twentieth century they did — spectacularly. The first improved dramatically for an entire generation while the second barely changed at all.

Understanding how that is possible is the single most useful thing in this Topic.

Three generations in one town, and the room at the top.

Joseph was born in 1925. He left school at fourteen and went into the steelworks, as his father had. So did most of the boys in his street. In his town in 1950, roughly seven men in ten worked with their hands, and perhaps one in ten had a job that would now be called professional or managerial.

His son David was born in 1953. David also left school at sixteen, worked briefly at the same plant, then took a technical qualification at the local college and ended his career as a project manager for a construction firm — salaried, pensioned, with a company car and a house he owned.

By the time David retired, the occupational structure of his country had transformed. Manual employment had fallen from around two-thirds of the workforce to under a third. Professional, managerial and technical work had grown from about one job in eight to more than one in three.

So David rose. And so did most of his generation. Ask that cohort whether they did better than their parents and a large majority say yes, correctly.

Here is the part that gets left out.

David did not overtake anybody. He moved up into a space that had been created. The number of managerial jobs grew faster than the number of children of managers available to fill them, so the extra places had to be filled from below. Nobody had to move down for David to move up.

Now ask the second question about David's generation. Take two boys born in 1953, one the son of a steelworker and one the son of an accountant. What were their relative chances of ending up in a professional or managerial job?

That ratio barely changed across the whole period. Both boys' absolute chances improved, because there were more such jobs. The gap between them, expressed as relative odds, was close to what it had been for their fathers — and close to what it would be for their sons.

David's son Michael was born in 1981. He has a degree, which his father did not. He works in a role his grandfather would not have recognised. And the occupational structure stopped expanding at the top at roughly the rate it had.

Which changes everything about his generation's experience. When the shape of the structure stops changing, upward mobility requires downward mobility. For Michael to move up, someone from a professional family has to move down. And a cohort in that position experiences something the previous one did not: a real and rising risk of ending up below where they started — which is felt, accurately, as things getting harder, even though the underlying fairness of the competition has not changed at all.

Three generations. Enormous absolute mobility, essentially unchanged relative mobility, and a shift in experience that has nothing to do with fairness and everything to do with shape.

Two questions, and they answer to different things.

Absolute mobility asks how many people ended up in a different position from the one they started in. It depends heavily on whether the structure changed shape — if there are more good jobs than there were, more people can occupy them without anyone being displaced.

Relative mobility — also called social fluidity — asks how much origins matter. It compares the chances of people from different origins reaching a given destination, and it is deliberately constructed to be unaffected by how many places of each kind exist.

Absolute mobility is about the structure. Relative mobility is about the competition.

And the political vocabulary systematically conflates them. "Improving social mobility" nearly always means, in practice, creating more good jobs — an absolute claim. It is nearly always argued for in the language of fairness — a relative claim. Those are different projects with different instruments, and one of them has a feature nobody mentions.

The mobility table

The instrument the whole field runs on, and the two ways of reading it that get confused.

A mobility table cross-classifies origin against destination. Rows are the class position of the family a person came from; columns are the position they ended up in. Every person in the sample falls in one cell.

Two entirely different questions can be asked of it, and they are constantly mixed up.

Outflow: given where you started, where did you end up? Read across a row. Of children from working-class origins, what proportion reached professional and managerial positions? This is the question about chances, and it is the one people usually mean.

Inflow: given where you ended up, where did you start? Read down a column. Of people now in professional and managerial positions, what proportion came from working-class origins? This is the question about composition.

They can tell opposite-sounding stories from identical data. Suppose professional positions are numerous and working-class origins are common. Then a small outflow rate — few working-class children reach the top — can coexist with a large inflow figure, because there are so many working-class children that even a small proportion of them fills a lot of places.

"Most people at the top came from ordinary backgrounds" and "a child from an ordinary background has a poor chance of reaching the top" can both be true of the same society. The first is inflow, the second is outflow, and which one gets quoted usually depends on what the speaker wants to show.

Always ask which direction is being read.

Structural and exchange mobility, and why the distinction is not academic.

Total observed movement splits into two components.

Structural mobility is the movement forced by a change in the shape of the distribution. If the number of professional positions doubles and the number of manual positions halves between two generations, a great deal of movement must occur regardless of how fairly places are allocated. David's rise was substantially of this kind.

Exchange mobility — sometimes called circulation or pure mobility — is the movement that remains once the structural change is accounted for: people swapping positions. For every person who moves up in this component, someone moves down.

Why this matters enormously.

Structural mobility is not zero-sum and exchange mobility is. A society can have a great deal of the first without anyone losing anything. It cannot have more of the second without producing downward mobility in exactly the amount of the upward mobility.

Which is the fact that makes mobility politics dishonest almost by construction. Everybody is in favour of more upward mobility. In a structure that is not expanding at the top, more upward mobility means more downward mobility, and it means it for identifiable people — the children of the currently advantaged, whose parents are the best-organised and best-resourced constituency in any society and who will use every available means to prevent it (see 8.1.1 on opportunity hoarding).

No one campaigns for downward mobility. And unless someone does, an appeal for more relative mobility in a static structure is an appeal for something nobody is proposing to bring about.

The finding that organised the field

Relative mobility is remarkably stable, and remarkably similar between countries.

The great comparative studies of industrial societies produced a result that has held up better than almost anything in the discipline: relative mobility rates — social fluidity — vary far less than anyone expected, both over time within countries and between countries at similar levels of development.

Absolute mobility varied enormously, tracking each country's occupational transformation and its timing.

Relative mobility was close to constant — the phrase used for the finding was a "constant flux": a great deal of movement, with the underlying pattern of relative advantage remarkably unmoved.

Two qualifications, because the finding is often overstated.

There are real exceptions. Sweden and some other Nordic countries have been consistently more fluid than the average; state-socialist societies showed genuine increases in fluidity during particular periods, associated with deliberate exclusion of the previously advantaged from higher education. So fluidity is not a constant of nature — it is movable, and it has been moved.

And the finding is about class mobility in a particular set of countries over a particular period, measured with schemes built for those economies (see 8.2.2 on comparability). It is not a universal law.

But the core result stands, and its implication is uncomfortable for the standard political framing. The extraordinary expansion of education, the vast growth of professional employment, the transformation of the occupational structure — all of that produced enormous absolute mobility and left the relative chances of children from different origins substantially as they were.

Which tells you something about the mechanism. When the structure created more room at the top, the advantaged families kept their children's relative position by taking more of the new opportunities too — more education, more credentials, more of the new positions. The advantage was maintained by being reproduced at a higher level, which is exactly what a theory of opportunity hoarding predicts (see 8.1.1), and it is the pattern 9.2 examines in education directly.

Four things that go wrong when the distinction is dropped.

Comparing mobility rates across countries without accounting for structure. A country that industrialised recently will show high absolute mobility because its occupational structure is transforming, and this says nothing about whether its competition is fair. Cross-national comparisons must use relative measures, and press coverage almost never does.

Reading a rise in absolute mobility as increasing fairness. It usually means the economy changed.

Reading a fall in absolute mobility as increasing unfairness. It usually means the expansion stopped. This is the single most common error in current commentary about generational decline — and the underlying experience is real, which is why the error is persuasive.

And treating the mobility rate as a measure of justice on its own. A society could achieve perfect fluidity — origins predicting nothing — and still be extremely unequal, with the same steep gradients allocated by lottery instead of by inheritance. Mobility is about who gets which position; inequality is about how far apart the positions are. They are separate questions, and a society can improve on one while worsening on the other.

Which raises a point worth sitting with. Perfect mobility in a very unequal society means every family faces a high risk of severe disadvantage in every generation. Whether that is preferable to stable inequality is a genuine question, and it is not obvious that everyone would choose it — which tells you something about why mobility, rather than equality, is the goal politicians prefer to endorse.

Because the two questions have different answers and require different policies, and the confusion is load-bearing.

If the concern is absolute mobility — that people should be better off than their parents — the instrument is economic: the number and quality of jobs, productivity, and how the gains are distributed. Education policy is largely beside the point, because educating more people for positions that do not exist produces credential inflation rather than mobility (see 9.2).

If the concern is relative mobility — that origins should matter less — the instrument is whatever breaks the transmission of advantage: early childhood, the quality gap between schools, access to credentials, inheritance, informal recruitment, and the direct transfer of resources. And it necessarily involves the advantaged losing relative ground, which is why this is the harder politics and why it is so often described in the language of the other one.

Three questions for any mobility claim.

Absolute or relative? If unstated, assume absolute, because that is what is easy to measure and comfortable to report.

Outflow or inflow? "Most of our senior staff came from ordinary backgrounds" is an inflow statement and may be compatible with terrible chances.

And has the structure changed? If the answer is yes, most of what is being celebrated or lamented is a change in shape rather than in fairness.

And keep the three generations in view. Joseph could not have become David. David did not overtake anyone. Michael faces a competition that is no less fair than his father's and a great deal less forgiving — and the difference between those two facts is what this lesson exists to make visible.

Two questions get fused. Absolute mobility: did people end up in different positions from their origins? Relative mobility (social fluidity): do origins matter less? They can move in opposite directions, and in the twentieth century they did.

The post-war generation rose because the structure changed shape — manual work contracting, professional and managerial work expanding — so upward movement required nobody to move down. The relative chances of a steelworker's son against an accountant's son barely changed.

A mobility table cross-classifies origin against destination, and it answers two different questions. Outflow (read across a row): given where you started, where did you end up — the question about chances. Inflow (read down a column): given where you ended up, where did you start — the question about composition. "Most people at the top came from ordinary backgrounds" and "an ordinary background gives you poor chances" can both be true of the same data.

Total movement splits into structural mobility — forced by the changing shape of the distribution, and not zero-sum — and exchange mobility, which is. In a static structure, more upward mobility means exactly as much downward mobility, for identifiable people whose families are the best-organised constituency in any society. Nobody campaigns for downward mobility.

The central comparative finding: absolute mobility varied enormously between countries and periods; relative mobility was remarkably stable — a constant flux. With real exceptions: the Nordic countries have been more fluid, and deliberate exclusion of the advantaged from higher education raised fluidity in particular state-socialist episodes. So fluidity is movable, and it has been moved.

The mechanism the finding implies: when more room appeared at the top, advantaged families took more of the new opportunities too, reproducing their relative position at a higher level.

Four errors: comparing across countries without accounting for structure; reading rising absolute mobility as fairness; reading falling absolute mobility as unfairness — the commonest error in current commentary, and persuasive because the experience is real; and treating mobility as a measure of justice, when perfect fluidity in a very unequal society means every family faces severe risk in every generation.

Absolute mobility — the total observed movement between origin and destination positions.

Relative mobility / social fluidity — the strength of association between origins and destinations, independent of how many positions of each kind exist.

Mobility table — origins cross-classified against destinations.

Outflow / inflow — where people from a given origin ended up; where people in a given destination came from.

Structural mobility — movement forced by a change in the shape of the occupational distribution; not zero-sum.

Exchange (circulation) mobility — movement remaining once structural change is accounted for; strictly zero-sum.

Constant flux — the finding that relative mobility rates are remarkably stable over time and similar across industrial societies.

Room at the top — the expansion of higher-status positions that made mass upward mobility possible without displacement.

Credential inflation — rising qualification requirements for the same positions, produced when educational expansion outruns the growth of the positions themselves.

One — do your own three generations. Write down the occupations of your grandparents, parents and yourself. Then ask whether the movement, if any, was structural or exchange — did the positions exist in the earlier generation?

Two — spot the direction. Find a claim about mobility in your country's press and determine whether it is outflow or inflow. Then work out what the other direction would show.

Three — check the structure. Look up how the occupational composition of your country changed over your parents' working lives. Most of any generational mobility story is in that table.

Four — do the zero-sum arithmetic. In a structure of fixed shape, write down what would have to happen to the children of the professional class for the children of manual workers to reach parity. Then ask which political party has proposed it.

Five — separate mobility from inequality. Describe two societies: one with high mobility and steep inequality, one with low mobility and flat inequality. Decide which you would rather be born into, not knowing your position — and notice that the question is 8.1.3's, arriving with data behind it.