The Rulebook

We have moved through four institutions — the firm, the contract, the money, the tax — each of which the market seemed to presuppose: the market needs firms to organize production, contracts to bind its promises, money to clear its exchanges, taxes to fund its foundation. In every case the institution turned out to be doing something the bare price could not, and in every case the market, which the price families treated as the given, was revealed to rest on something built. The fifth and last institution carries this discovery to its root, and turns it on the market itself. For we have spoken throughout of "the market" as though it were a single natural thing — a clearing where supply meets demand, arising of itself wherever people have things to trade. Alvin Roth's life work begins from the opposite premise: that there is no such natural thing, that every market is a specific human artifact, built to a specific design, and that the design is a choice.

Look closely at any market and you find not a state of nature but a thicket of rules. What may be traded, and what may not — organs, votes, and persons are excluded; most goods are not. Who may trade, and who may not — licenses, memberships, ages, accreditations. When and how the trading happens — the opening bell, the auction's format, the order in which offers are made and answered. What each party must disclose, and what each may conceal. How disputes are settled, and how the deal is enforced. Strip these away and there is no market, only a crowd of people who cannot transact, because they do not know what is permitted, what is promised, or what will be honored. The rules are not an overlay on a pre-existing market; they are constitutive of it. A market is its rules, the way a game is its rules: change them, and you have a different market, or no market at all.¹

If a market is its rules, then the rules are choices, made by someone, and different choices yield different markets. This is the premise and the provocation of market design: that the outcome of a market — who gets what, at what price, with what efficiency, with what fairness — depends not only on the competition within it but on the rules that frame it, and that those rules can be designed well or badly, deliberately, the way an engineer designs a bridge. The same buyers and sellers, the same goods, the same competition, under different rules, reach different outcomes: one set of rules clears the market smoothly, another leaves it congested; one rewards honesty, another rewards strategic concealment; one allocates efficiently, another wastes. The design is not a neutral container for the competition; it is half of what determines the result. Roth's word for the practitioner of this art is the engineer: the economist who does not merely describe a market but designs one, choosing its rules to make it work.²

That markets are designed becomes vivid where they fail. Roth studied markets that broke down not for lack of willing traders but for lack of workable rules. The market for new doctors once unraveled into chaos: hospitals, competing for the best graduates, made offers earlier and earlier — first in the final year, then years before graduation, on exploding deadlines that forced students to commit before they knew their options or their grades. The competition was fierce, but the rules were broken, and the outcome was bad for everyone. Such unraveling, Roth showed, recurs across markets — for law clerks, for college admissions, for gastroenterologists — wherever the timing of offers is left ungoverned. The cure was not more competition but better rules: a designed clearing process that gathered the offers and matched them at a single time, under a rule that no one could game by jumping early. The lesson is general: a market can fail by its design, and be repaired by redesign — which proves, as nothing else does, that the design was there all along, doing its work invisibly until it broke.³

Designing a market well, Roth found, means satisfying conditions that the rules must meet, and that bad rules violate. A good market must be thick — it must bring enough participants together at once that good matches can be found, rather than letting them peel off into private, premature deals. It must be uncongested — it must give participants the time and means to consider their options and to make offers and acceptances, rather than overwhelming them. And it must be safe to participate in straightforwardly — it must make honesty the best policy, so that participants can reveal their true preferences without being punished for it by others' strategizing. A market that is thin, or congested, or unsafe — that scatters its participants, or rushes them, or rewards concealment and gaming — will allocate badly however vigorous the competition within it. These are design criteria, like the load and span of a bridge; and meeting them is the engineer's task, not the market's own.⁴

The rulebook, then, is the foundation of this last volume and the completion of the family's argument. The market is not found but built; it runs on rules; the rules are chosen; and the choice shapes the outcome as much as the competition does. Which means that everything the earlier families took as the spontaneous working of the market — the prices that form, the quantities that clear, the gains from trade that are realized — happens inside a designed frame, and would happen differently, or not at all, under a different design. But this raises at once the questions the rest of the volume must face, and that the author sets at its center. If the rules are chosen, the design shapes the result — how much, and how exactly? The next plates show the design at work: in the auction, where the rules visibly decide the outcome, and in the match, where there are no prices and the design must do everything. And if the rules are chosen, then someone chooses them; and the deepest question is not how to design a market well but who does the designing, and in whose interest. To the power of the design, and then to the politics of it, the volume now turns. The next plate is the auction.⁵

A market is a human artifact, built to a design, running on rules that someone chose; and the rules shape who gets what as surely as the competition does. Alvin Roth made this the working premise of an engineering discipline: study how markets fail for want of good rules, and design better ones — rules that make a market thick, uncongested, and safe to enter honestly. Where the rules are well made, the market works; where they are broken, it unravels, however many willing traders crowd into it. This is the foundation of the last institution and the turning of the family's long argument: the market, which the price families treated as the given of all economic life, is itself the most made of things, an artifact of rules that are chosen. The volume must now show how much the choice matters, and then ask who makes it. The next plate asks how that building is done — the general logic by which rules can be designed so that self-interested play is made to serve the end in view, and the limits of what any such design can reach.