The competitive market has many employers, and its discipline is the rival always ready to bid a worker away. Remove the rival — leave one mill in the town, or bind the worker to his post with frictions costly enough that to leave is its own punishment — and the discipline vanishes, and with it the guarantee that pay tracks product. A single buyer of labor, a monopsonist, knows that to hire one more hand is to raise the wage of every hand; and so he hires too few, and pays them all less than the last is worth to him. The shortfall is not a market imperfection to be apologized for. It is the shape of his power, written into the wage — and once it is there, the oldest certainty in the book inverts: a floor set under the wage, with judgment, can lift pay and jobs at once.
Begin with the arithmetic, for the whole of the matter is in it. A mine in an isolated valley is the only work for miles; the men who would mine must mine for it or not at all. The mine wants another pair of hands. The next man will not come for the wage the others took — he has a smallholding, or a longer walk, or simply a higher price on his time — and will sign on only at something more. Here is the hinge: the mine cannot, as a rule, pay that one man more and leave the rest untouched. Wages are not so easily kept secret, nor so easily made unequal for identical work; to raise the offer for the marginal man is to raise it for every man already at the face. So the true cost of the additional miner is not his wage alone but his wage plus the raise conferred, willy-nilly, on everyone already employed. That cost — the marginal cost of labor — stands above the wage, and climbs faster than it.
A competitive employer never meets this wall, because it can hire all it likes at the going rate; the rival's standing offer flattens the supply of labor before it. The monopsonist faces the supply curve whole, and must climb it. Weighing the rising marginal cost against the value the next man would add, he stops short — sooner than a competitive employer would — and reads the wage he pays off the supply curve at that smaller number of men, which is to say below the value of what the last of them produces. Two harms fall on the worker from the one cause, and they are worth separating. Fewer are hired than a competitive market would hire. And every one who is hired is paid less than the worth of what he makes. Joan Robinson, who built the theory in 1933 and gave the buyer's power the name it still carries, called that second gap by a hard word — exploitation — meaning by it not an insult but a magnitude: the exact distance between the marginal product and the wage.
The isolated mine is the textbook's case and the world's exception. Most workers live within reach of many employers — and yet the power of the single buyer reaches them too, in diluted but pervasive form, because the argument never truly required one employer. It required only that workers cannot move for free. And they cannot. To find a better job costs time, and search, and luck; to take it costs the seniority built up where one is, the pension half-vested, the known commute, the trusted workmates, the dozen small securities the next post will not honor. Many are bound more tightly still, by clauses forbidding them to carry their skills to a competitor. Each of these frictions hands the present employer a sliver of the power the lone mine holds whole: room to let the wage fall a little below the worker's product, because the worker cannot cheaply punish him by leaving.
Seen so, monopsony is not the rare affliction of the company town but the ordinary climate of working life — a power present, in some degree, wherever quitting is costly, which is to say nearly everywhere. The applied cases economists have studied hardest are the ones where the buyer's dominance is plainest: hospitals, which in many places are the only substantial employer of nurses and have long been suspected of holding nurses' pay beneath their worth; school districts and their teachers; and, most cleanly of all, professional sport, where for the better part of a century a single clause bound each player to one club for as long as the club cared to keep him.
Now the result that makes the plate matter, and that overturns the first thing a student is taught about wage floors. The competitive model's verdict on the minimum wage is swift and grim: a floor set above the market wage prices out every worker whose product falls beneath it, and the surplus hands are turned away — a floor buys higher pay for some at the cost of work for others. That is the case argued, on the competitive market's own terms, in this volume's seventh plate. Under monopsony the same floor does something altogether different, and the difference is the whole point. The employer had held hiring back because each added worker forced up the wage of all; that restraint was the source of the harm. A floor removes it. Once the wage is fixed by law, the employer can take on another worker without raising anyone's pay — the law has already set it — and so the marginal cost of an added hand falls to the floor itself, below where it had stood. Facing a lower marginal cost, the employer hires more. A wage floor pitched between the monopsony wage and the competitive wage therefore raises pay and employment together, the very outcome the competitive model swears impossible.
But the instrument is exact, and honesty about it requires marking its limit. The happy result holds only within a band. Set the floor above the wage a competitive market would have paid, and the old logic returns: now the floor truly does price workers out, and employment falls. As the floor rises from low to high, employment does not move in one direction but traces an arch — climbing while the floor is still correcting the employer's power, cresting near the competitive wage, then falling as the floor begins to overshoot it. The minimum wage is thus neither the job-killer of the textbook nor a gift without cost; it is a tool whose effect turns on two things no diagram can supply from within itself: how much power it is set against, and how high it is pitched. To know whether a given floor will help or harm, one must leave the diagram and look.
And looking, for a long time, was the one thing the profession neglected to do, so certain was it of the competitive answer. The textbooks taught for decades that minimum wages cost jobs, and the teaching ran well ahead of the evidence. Then, in 1994, David Card and Alan Krueger looked. They compared employment at fast-food restaurants on either side of the New Jersey–Pennsylvania line after New Jersey raised its minimum wage and Pennsylvania did not, and found that employment in New Jersey had not fallen — by some measures, it had risen. The finding scandalized the discipline, and was met as scandalous findings are: attacked, reanalyzed, partly rebutted, stoutly defended, and never quite buried. The argument over the precise number runs to this day. But the weight of the evidence gathered since — across many raises, many places, many methods — is that moderate minimum wages do far less harm to employment than the competitive model foretold, and frequently none at all, which is precisely what a labor market shot through with monopsony power would lead one to expect.
The evidence has not stopped at the minimum wage. A growing body of work has set out to measure the buyer's power directly, and has found it widely: labor markets in many places concentrated in few employing hands; noncompete clauses binding something like a fifth of American workers, reaching far down into low-wage work where no trade secret could conceivably be at stake; agreements among firms, sometimes secret, not to hire one another's people. Each is a means of removing or restraining the rival whose competition was supposed to bid wages up to the worth of the work. How large the power is in aggregate, and how much of the long stagnation of wages it explains, remain genuinely unsettled questions, and the honest position is the modest one: real labor markets lie somewhere between the textbook poles of perfect competition and the single buyer, and where exactly is a matter for measurement, not assumption. What can no longer simply be assumed is the competitive case itself — that the wage is a clean readout of the worker's product, and the market that sets it a benchmark past improving.
The gap between what a worker is worth and what he is paid is not a flaw in the wage. It is the measure of who set it.
Two lessons come out of monopsony, and the first is the volume's own thesis in its hardest, most legible form. Where the buyer has power, the difference between the value of what a worker produces and the wage he receives is not noise, not friction, not the rounding error of an imperfect world. It is the power itself, made arithmetic — the precise toll the contest exacts, available to anyone willing to subtract the wage from the product. Elsewhere in the science the claim that a price encodes a hidden contest must be argued; here it can be measured. The wage is a verdict, and monopsony writes the margin of the verdict in a figure.
The second lesson is quieter, and more subversive, and it concerns the oldest tradeoff in the subject. The competitive model's deepest moral teaching is that kindness costs — that fairness and efficiency pull against each other, so that every floor under the wage and every union at the table buys a little equity at the price of a little prosperity, robbing Peter to pay Paul. Monopsony dissolves the tradeoff at its root. Where the employer's power has already bent the market, correcting that power — by a floor, by a union, by whatever countervailing strength the workers can raise — can lift wages and employment and the total product all at once, equity and efficiency moving together rather than against. The familiar tension between the decent thing and the efficient thing turns out to be, in good part, not a law of economics but a consequence of having assumed the market competitive before the argument began. Strike that assumption, and the case against helping workers loses the theoretical floor it stood on.
None of this is a warrant to set wages wherever one pleases. The band is real; past the competitive wage the floor does bite, and jobs are lost; the magnitude of the power is uncertain and varies from market to market; the competitive model still describes some corners of the economy well enough. The plate claims something narrower and harder to evade: that the reflexive defense of the market wage against every floor and every union can no longer rest on theory, because the theory it rested on assumed away the very power the evidence keeps turning up. The wage on the contract is not, after all, the worth of an hour of a life. Under the single buyer it is demonstrably less than that worth, by the exact width of the buyer's advantage — a price struck in a contest the worker entered weaker, and bearing the mark of that weakness in its size. A price is a verdict on a hidden contest, mistaken for a fact of nature; monopsony is the case where the contest leaves its measure on the verdict, in plain figures, for anyone who cares to read them. The argument remains open.