Smith noticed, and did not trouble much to hide, that the masters were always combined — quietly, tacitly, by the mere fact of being few and like-minded — while it was the workers' combinations the law saw fit to forbid. The union is the correction of that asymmetry: the many, who are weak singly, acting as one, so that the worker who cannot afford to outwait a single employer can, with his fellows, afford to. John Kenneth Galbraith called the principle countervailing power — strength on one side of a market calling up strength on the other to meet it. Whether this restores a fair balance to the bargain or merely stages the contest on new ground is among the most argued questions in the subject; what is not in doubt is that the wage emerging from a bargain between organized equals is a different figure from the one a lone worker carries away.
The asymmetry Smith named was not a turn of phrase but the plain arrangement of the law. For most of the era in which modern wages took shape, a worker who combined with his fellows to raise his price committed a crime — Britain's Combination Acts of 1799 and 1800 made the very act of joining together a criminal conspiracy — while the masters' standing, tacit understanding not to bid each other's labor up drew no notice and no penalty at all. The union is what the workers built against that arrangement: an attempt to do openly, together, and at last lawfully what the employers had always done quietly and alone — to turn a crowd of individually powerless sellers of labor into a single voice that could bargain as the buyer bargained.
What drives a worker to combine is not a wish to argue but the discovery that argument is beside the point. The previous plate showed how a wage comes to sit below the worth of the work: where the employer faces no rival ready to bid the worker away, he simply declines to raise the wage to the value of what the worker produces, and keeps the difference as the measure of his power. Against that, a better case avails nothing, for the worker's product was never in dispute; what is in dispute is who shall keep the gap between that product and the wage. The only reply to power of that kind is power of the answering kind — and the union is the worker's instrument for manufacturing it, out of the one resource the lone worker lacks and the many possess: the capacity to withhold their labor all at once.
John Kenneth Galbraith made this the center of an argument about the whole modern economy. In a world of large firms, he wrote in 1952, the classical discipline of competition is often feeble or absent, and what actually restrains market power is not more competitors but opposed power — concentrated strength on one side of a market summoning concentrated strength on the other to face it. He called it countervailing power, and the union was his leading case. The reframing matters, and it follows straight from the previous plate. Where the employer is a monopsonist, holding the wage below the worker's product, a union that pushes the wage back toward the value of that product does not distort a healthy market; it corrects a bent one. And — the same surprise the minimum wage sprang — it can do so without the loss of jobs the competitive model would predict, even raising employment, because it is undoing a restriction the employer had every reason to impose. Against employer power, the union is not the thumb on the scale. It is the second hand that brings the scale level.
Carry the logic to its limit and the wage loses, openly, the last of its pretensions to be a fact. When a union that speaks for all the workers faces an employer who is the only buyer of their labor, economics offers no determinate wage at all — only a range, with a floor at what the employer might get away with and a ceiling at what still leaves the firm able to trade, and within that range nothing but the relative strength and nerve of the two parties to fix where the wage falls. This is the bargain with its mask off. There is no marginal product to read the answer from, no market to defer to, no figure written into the nature of things; there are two powers in a room, and a wage that will be whatever their contest makes it. Everywhere else in the science the claim that the wage is a negotiated verdict must be argued against the appearance of a fact. Here the appearance is gone, and both sides know precisely what they are doing.
But to leave the matter there — the union as the worker's just counterweight — would be to show one of its two faces and hide the other, and the honest treatment, the one the modern study of unions is built upon, insists on both. The first face is the one the textbook has always drawn: the union as a labor monopoly. Like any monopolist it raises the price of what it sells — its members' labor — above what an open market would pay, and like any monopoly price that raise has costs: fewer jobs in the union sector than there would otherwise be, workers pushed into non-union work where they crowd the wage down, labor held in the wrong places by the premium, the friction of restrictive rules and the deadweight of strikes. This face is real, and a serious account does not wave it away. A union that wins its members a fifth above the market has, on this face, done so partly at someone's expense — the consumer who pays more, the outsider who does not get the job.
The second face is the one the monopoly story cannot see, and it begins from a simple fact about how a discontented worker can act. He has, in the main, two options: he can leave — quit, and look for something better — or he can stay and try to change what is wrong. An ordinary labor market runs almost entirely on the first; its remedy for a bad job is the door. But exit is costly and blunt, and it throws away all that the worker knows about the place and the place knows about him. The union supplies the second option, the one the lone worker effectively lacks: a collective voice, a way to mend grievances from within rather than by walking out. And voice, it turns out, does things the monopoly face never imagined. It lowers turnover, and with it the great hidden cost of forever hiring and training anew; it improves the flow of information between floor and office; it can lift morale, and with it output. A workplace with a voice can be a more productive workplace, not a less one.
Between the two faces lies the union's effect on the shape of wages, and here the evidence is least ambiguous and most overlooked. Unions raise their members' pay — the union wage premium, long measured at something near a sixth in the United States — but more consequentially they compress it, lifting the bottom more than the top, narrowing the gaps within a workplace and between workplaces, and, through the mere threat that an industry might organize, dragging non-union wages upward too. Whatever one concludes about jobs and productivity, this much is firm: where unions are strong, wages are more equal, and not only for those who carry a card. It is a fact that will matter greatly when the volume turns, in its final plate, to what has become of labor's share of the whole.
The costs must be kept honestly in view, and they reach beyond the monopoly face's misallocation. A union strong enough to raise wages is strong enough to defend the jobs of those who hold them at the expense of those who do not: the employed insiders bargain over their own wages and have small reason to weigh the unemployed outsider, and a labor market organized around insiders can lock joblessness in — a leading explanation of Europe's long unemployment in the closing decades of the last century. Strong unions bargaining firm by firm can also chase wages and prices upward together, the wage-push inflation that haunted the 1970s. And public-sector unions raise a difficulty all their own, for they bargain not against a profit-disciplined firm that competition will punish for overpaying, but against a government spending other people's taxes, with no rival and no bankruptcy to mark the limit.
Which is why the largest truth about unions is that there is no such thing as the union, only the particular institution a country has made. The word covers arrangements that scarcely resemble one another and behave nothing alike. The American union is adversarial, organized firm by firm, and in long decline; the European is often sectoral, its bargains extended by law across a whole industry so that coverage stays high even where membership is modest; the German seats workers' representatives in the boardroom itself; the Scandinavian hands the unions the running of unemployment insurance, and membership stays high in consequence. Each produces a different economics. Whether organized labor raises wages by distorting or by correcting, whether it lowers employment or leaves it untouched, whether it drags on productivity or lifts it — these are not properties of unions as such, but of the shape a society gives them, and of the question Galbraith asked first: how much employer power there was to be countered in the first place.
And the American union, the one whose adversarial firm-by-firm form makes it weakest, has been vanishing. Where a third of private-sector workers carried a union card at the middle of the last century, fewer than one in ten workers do now, and in the private sector fewer than one in fifteen. The disappearance has not been costless, and its timing is hard to ignore: the long decline of union density tracks, decade for decade, the great widening of wage inequality and the slow fall in labor's share of the national income — the very divergence the volume's final plate takes up. Whether the fading of the union caused that divergence or merely kept it company is argued, as everything here is argued; but the case that the loss of countervailing power is part of why the wage came unmoored from the worth of the work has grown steadily harder to dismiss.
The union invents nothing about the wage. It only refuses to pretend the wage was ever a fact.
This is the union's deepest lesson, and it is the volume's thesis stated by an institution rather than an argument. When workers and employers sit down to bargain collectively, neither side pretends the wage is a number read from the worker's marginal product; both know it for what it is — a settlement, struck between unequal powers, of how the value the work creates shall be divided. The union does not turn the wage into a contest. The wage was always a contest; the marginal-product story was the means of not seeing it; and the union is simply the worker's refusal of that story, and the apparatus for entering the contest on better terms. As to whether it restores a fair balance or merely restages the fight on new ground — the question the plate began with — there is no answer that holds in every case, and pretending otherwise is the surest sign of an axe being ground. Against the great employer power the previous plate described, the union corrects, and equity and efficiency move together. In a labor market already competitive, it distorts, and they part. Which case one stands in is a matter for looking, not for doctrine.
What the union changes, then, is not the nature of the wage but our willingness to see it. It drags the hidden contest into the open, gives it a table and a name and a body of law, and seats the weaker party where he can dispute the verdict instead of merely receiving it. The wage that emerges is no more a fact of nature than the one a lone worker is handed — but it no longer pretends to be, and that honesty is the union's quiet gift to the science: it shows, by making one, that every wage is a settlement. A price is a verdict on a hidden contest, mistaken for a fact of nature; the union is the institution that strips away the mistake, and lets the contest be fought in the light. The argument remains open.