Eleven plates have examined the wage one bargain at a time — the lone exchange and its asymmetry, the marginal product, the invested skill, the buyer's power, the worker's organised answer. This last plate steps back from the single transaction to the whole economy, and puts the volume's question at the largest scale there is: of everything a nation makes in a year, what part falls to those who work for a wage, and what part to those who own? That fraction — the share of labour — was long taken for one of the fixed stars of economics, a constant of nature about which a sensible person need not worry. It is nothing of the kind. Its fall in our own time, and the strange parting of pay from productivity that has come with it, is among the largest economic facts of the age — and it is the whole argument of this volume, written at last into the national accounts.
For the better part of a century the share looked immovable. Through wars and slumps and the arrival of whole new technologies, the portion of national income paid out as wages and salaries hovered, in the advanced economies, around two-thirds, and the portion taken by capital around a third, with a steadiness that economists raised into a law — Bowley's Law, after the statistician who first marked it — and that Nicholas Kaldor enshrined among the handful of “stylised facts” any theory of growth was bound to honour. The constancy was at once convenient and consoling. Convenient, because the favourite production function of the textbooks, the Cobb-Douglas, simply assumes it. Consoling, because it seemed to settle, at the level of the whole society, the very contest this volume has tracked plate by plate: capital and labour might wrangle over the division of each firm's product, but in the aggregate they split the national one in fixed proportions, so that growth lifted both together and the quarrel had, it appeared, a standing truce. The marginal-productivity orthodoxy of the second plate was content: each factor drew its product, and the shares stood still.
And then, in the United States, around the early nineteen-seventies, something quietly came apart. For the quarter-century after the war the pay of the typical worker had risen very nearly in step with output per hour — productivity and compensation climbing the same staircase together, as both the orthodoxy and ordinary decency expected that they should. After about 1973 the two lines parted company. Productivity went on rising, decade upon decade; the compensation of the worker in the middle flattened, and then crept where once it had climbed. The widening space between them — the great divergence, in the phrase that has fastened to it — became the most argued-over chart in the whole economics of inequality, and the most quietly damning. The productivity that the second plate had called the ceiling of the wage went on rising; the wage beneath it did not.
The productivity that the second plate called the ceiling of the wage went on rising; the wage beneath it did not.
And the fixed star moved. Beginning around 1980, and quickening after the turn of the century, the share of national income going to labour fell — not in America alone, but across the advanced economies, and across much of the developing world besides. Loukas Karabarbounis and Brent Neiman, gathering the evidence from dozens of countries, showed the decline to be global and no mere statistical accident; the constant of nature was, after all, a variable, and it was sliding. As labour's share fell, capital's rose to meet it. The truce the old stylised fact had seemed to record was off — and had perhaps never been quite so firm as the textbooks taught.
Here the honest plate must slow its step, for some part of the divergence lies in the measuring rather than in the world, and a scrupulous account says as much before it reaches for villains. The headline gap narrows — though it does not close — once the comparison is drawn with care. Set the average worker's pay against productivity rather than the median's, and a good share of the gap proves to be inequality among workers, smuggled in by the back door. Count the whole of compensation, the benefits as well as the wage in the envelope. Deflate output and pay by the same index of prices, rather than by two that have drifted apart. Net out the depreciation of the capital that quietly wears away. Anna Stansbury and Lawrence Summers, pressing exactly these corrections, argued that the link between productivity and pay is more nearly intact than the starkest charts allow — that rising productivity does still lift pay, and that a large part of the worker's trouble is the plainer one that productivity itself has grown slowly. Others, at the Economic Policy Institute and beyond, hold that a wide and stubborn divergence survives every honest adjustment. The truth lies between the two camps: the link has frayed, not snapped; and a substantial gap remains, when all the corrections are made, that no measurement can explain away. It is that residue the rest of the plate must account for.
And to account for it is to watch every thread of this volume return, now woven at the scale of the nation. The thread of worth comes first: technology — the machine and the computer standing in for labour, the falling price of capital goods drawing firms to replace workers with machines wherever the two are near enough substitutes, the skill-biased change of the fourth plate enlarged into a force that reshapes the whole distribution of income. Then the thread of power: the erosion of labour's side of the bargain — the de-unionisation the ninth plate traced, the monopsony the eighth anatomised, the minimum wage left to be eaten away by inflation, the reserve army that offshoring and migration swelled into a global one — every device by which, as the first plate warned, the wage is pressed downward when the worker cannot afford to wait.
And then two threads the later groups of the table will take up in full. The first is market power: the rise of concentrated, high-markup firms — the “superstars” whose own labour shares run low and which, as they have come to rule their markets, have dragged the aggregate share down behind them. The second is capital itself: Thomas Piketty's account of a capital share swelling as the return on wealth outran the growth of the economy — though Matthew Rognlie's sharp reply, that almost the whole of the rising capital share is not factories and machines but housing and land, hands the larger part of the puzzle straight back to rent, and so to the very next volume of this table. The fall of labour's share is over-determined: not one cause but a confluence of them, and the argument over their proportions is among the most alive in all of economics.
Stand back from the quarrel, though, and the meaning of the thing is plain enough. Labour's share is not a constant of nature, and on close inspection it never quite was. It is the aggregate scoreboard of the contest this whole volume has anatomised — the bargain, the orthodoxy of worth, the skill that deepens and confounds it, the buyer's power and the worker's organised answer — the whole of it summed across an entire economy and set down as a single ratio. When labour's power holds, and the competition among employers is real, the wage tracks the productivity the second plate enthroned, and the share stands more or less still; when that power erodes and market power gathers on the other side, the wage falls behind, the share declines, and the divergence opens like a wound. The great divergence is worth and power — the volume's two answers — made visible at the scale of the nation: productivity, the measure of worth, rising on; the wage, the register of power, lagging behind; and the widening gap between them the plainest measure there could be of how, in our own age, the contest has gone.
Labour's share is not a constant of nature. It is the scoreboard of the contest, totted up across a whole society.
Return, then, a last time, to the question this volume was built to put to every wage: is it the measure of the worker's worth, or of the employer's power? Twelve plates were promised the volume, and the ones now standing trace the answer from the single bargain at the hiring-hall door to the ratio struck in the national accounts — and the answer holds at every scale between. The wage is neither pure worth nor pure power, but the settlement of a contest between them; and labour's share is that settlement totalled across a whole society — a figure that rises and falls not by any law of nature, but as the balance of power, and the pace of the machine, and the design of the institutions shift beneath it. The old fixed star of the constant share was the good fortune of an age in which labour's power happened to hold; the great divergence is the ill fortune of an age in which it has not. And here, at the close of the volume, The Wage opens at last onto the rest of the table. The rent that swallowed the larger part of capital's gain leads on to the next volume; the market power behind the superstar firm, to the failures gathered in the third group; the growth and the price level that frame the whole, to the fifth. For the wage was never an island, and was never meant to be read as one. It is a single price among the table's twenty-six — and, like every one of them, as the work has said from its first page and will say to its last, it is a verdict on a hidden contest, commonly mistaken for a fact of nature. The argument, in this volume as in the whole of the table, remains open.