The Share

We have followed the wage down to its smallest unit — the single bargain between one worker and one employer — and traced the forces that decide it: the worth that sets the ceiling, the alternatives that set the floor, the power that fixes the point between. We have watched those forces at work in the market for skill and the market for risk, in the employer's wage-setting power and the worker's answer of union and law, in the discrimination that taxes some and the machine that shadows all. One question remains, and it is the largest. Add every wage together, and every profit, and every rent, until you have the whole income of a nation in view. How is that total divided between the people who work and the people who own? The answer is a single fraction — labour's share of national income — and it is the wage seen whole.¹

For a long time the remarkable thing about labour's share was how little it seemed to move. Through booms and slumps, through war and peace, through the arrival of electricity and the motor car and the computer, the fraction of national income paid to labour hovered near two-thirds and stayed there, decade upon decade. The regularity was striking enough to be named — Bowley's Law, after the statistician who noted it — and to be enshrined among the “stylised facts” of economic growth that any theory was expected to reproduce. It was even built into the mathematics: the production function economists reach for by default, the Cobb–Douglas, has constant factor shares baked into its form, and its long popularity owed something to how well that assumption seemed to fit. The constancy of labour's share became one of the fixed stars by which the discipline navigated — a thing not to be explained so much as relied upon.²

And then the fixed star moved. Beginning around 1980, and accelerating after 2000, labour's share of income entered a sustained decline across the United States and much of the rich world — a fall of several percentage points that has not reversed. In the United States the most-watched measure, the labour share of the business sector, has slid from around two-thirds in the late 1980s to barely above half: in 2025 it reached 53.8 percent, the lowest figure since the series began in 1947. Over the same span, as labour's slice thinned, corporate profits climbed to records. The pie has grown enormously; the share of it reaching those who do the work has shrunk.³

This is the aggregate face of a divergence the previous plate named at the level of the individual worker. There we saw that productivity — what the average hour of work produces — has gone on rising while the pay of the typical worker lagged behind, the two lines parting company around 1980. Labour's falling share is the same fact viewed from above. Had pay kept up with what workers produce, labour's share would have held constant; that the share has fallen is precisely the statement that, in the aggregate, the gains of rising productivity have been flowing to capital rather than to labour. The question the previous plate left hanging — where did the productivity go? — has an answer, and the answer is: to the owners.⁴

Why has it fallen? Here the agreement ends and the argument begins, for the decline of labour's share is among the most scrutinised and least settled puzzles in economics. A roll-call of suspects has been assembled, and the striking thing — for the purposes of this volume — is that each of them is a force the earlier plates have already met.

The machine is one: automation, substituting capital for labour task by task, shifts income from wages to the returns on the equipment and software that do the work, and arms employers with the replacement threat the previous plate described.⁵ The open border is another: as trade and offshoring moved the labour-intensive parts of production to lower-wage countries, the workers who remained lost bargaining power, and income shifted toward the capital that stayed mobile — several careful studies find this the single largest driver.⁶ The giant firm is a third: a handful of dominant, high-margin “superstar” firms have come to command ever-larger shares of their industries, and because such firms pay out little of their revenue as wages, their rise drags the aggregate share down; rising markups and employer wage-setting power, the monopoly and monopsony of the eighth plate, tell the same story.⁷ And the emptying union hall is a fourth: as organised labour has shrunk and the wage floor has eroded, the institutions that once let workers claim a larger share have weakened.⁸

Honesty requires a fifth suspect, and it points the other way: measurement. A substantial part of the apparent decline may be an artefact of how the accounts are kept. Some of the fall reflects rising depreciation — capital wearing out faster — which is counted as capital's income but never reaches an owner's pocket. Some reflects the difficulty of splitting the income of the self-employed between their labour and their capital; by one careful estimate, about a third of the measured decline is an artefact of that split. And much of the rise in capital's share across the rich world has gone not to the factory-owner of caricature but to housing — to the imputed returns on the homes people live in. On the measures that strip these out, the decline shrinks, and labour's share looks closer to its historical range. The fall is real; how large it is, and therefore how much there is to explain, is genuinely in dispute.⁹

Set the measurement caveat to one side — grant that some of the fall is real, as nearly everyone does — and look again at the four causes. They are usually presented as competing hypotheses, as if the task were to discover which one is the culprit. But run them together and a single shape emerges.

What does automation do? It strengthens the employer's hand by providing an alternative to the worker. What does offshoring do? It strengthens the employer's hand by providing another alternative — a workforce abroad. What does the rise of dominant firms do? It concentrates wage-setting power in fewer hands and lifts the markups by which revenue is kept from labour. What does the decline of unions and the erosion of the minimum wage do? It removes the countervailing power and the legal floors by which labour had claimed its share. Four different mechanisms, and every one of them is a transfer of power from labour to capital. The falling labour share is not four stories. It is one story — the story of labour's weakening power to claim what it produces — told four times.¹⁰

This is why labour's share belongs at the end of this volume, as its gathering point. Every plate has argued that the individual wage is decided by power; labour's share is that argument at the scale of an entire economy. It is the sum of every markdown a monopsonist extracts, every premium a union wins, every dollar a minimum wage reclaims, every wage a machine's shadow caps — aggregated across all the workers and all the firms in a nation, and expressed as a single number. When labour's power is strong, the number is high; when it weakens, the number falls. The long stability that earlier economists mistook for a law of nature was no law. It was the visible trace of a rough and historically specific balance of power — the one struck in the mid-twentieth century by strong unions, tight labour markets, rising wage floors, and the bargaining strength of workers in a manufacturing economy. And the decline since 1980 is not the working-out of an economic law either. It is the trace of that balance coming apart.¹¹

What does this mean for the wage — the subject we began with, eleven plates ago? It means the answer to the volume's opening question has been the same at every scale. We began by asking whether a wage is what the work is worth or what power extracts, and we have put that question to the single bargain, to the market, to the institution, to the machine, and now to the whole income of a nation. At every level the verdict has held: worth sets the ceiling, but power decides how much of the space beneath it the worker actually claims. Labour's share is that verdict aggregated — the clearest statement the discipline can make of how much, across an entire society, the wage is worth and how much it is power.

And it settles, at last, the question the first plate posed and the whole volume has circled. Is the wage what the work is worth? If it were — if every worker were paid the value of what they produce — then labour's share would simply be labour's contribution to production, a technological fact, fixed by the recipe by which output is made and immune to the balance of power. It is not fixed. It moves with the strength of unions and the level of the minimum wage and the credibility of the machine's threat and the mobility of capital across borders — with power, in a word, and not with the recipe. That labour's share answers to power is the proof, written at the largest scale economics affords, that the wage was never simply worth.¹²

There is a consolation in this, and a charge. The consolation is that the falling share is not fate. Were it the verdict of technology, or the iron working-out of a law, there would be nothing to do but watch it fall. But it is the verdict of a contest, and contests can be fought again. The instruments that built the mid-century balance — the union, the wage floor, the tight labour market, and others these volumes will examine — are available still, and where they have been deployed, the share has answered. Labour's share is low because labour's power is low, and labour's power is not a constant of nature. It is something societies build, or let decay, by the choices they make.

That is the charge. A society that finds the division of its income unjust cannot console itself that the division is merely how the economics works, because it is not; the economics sets the size of the pie, but the cutting of it is a question of power, and power is arranged by law, by institution, by organisation, by choice. To say the wage is a matter of power is to say it is a matter of politics — and therefore that it is ours to settle, and not the market's alone to dictate.

And so the volume closes where it began. We were told the wage is what the work is worth, and we have found, on every scale from the single bargain to the income of nations, that worth sets only the ceiling — and that what fills the space beneath it is power. Labour's share is that truth at its largest: the running measure, for a whole society, of how much of what is made the makers are able to keep. It was never a law of nature, and it is not falling by one. It is the tally of a contest still being fought — by every union and every statute and every worker who can, or cannot, afford to walk away. The wage is not, in the end, a fact about work. It is a fact about power. And power, unlike worth, is something a society is free to arrange otherwise.¹³