The contract commissions a thousand copies. The book sells unexpectedly well, another printing becomes worthwhile, and the agreement has said nothing about a second run. Someone must decide whether the press will turn again. Grossman and Hart begin from this sort of silence: the owner of the press holds a choice the first bargain did not allocate away. The publisher can seek another agreement, but cannot conjure access from the profit an additional printing would bring. Ownership enters where the specified promises stop. Its economic importance lies in what the parties can do while the next promise is still being negotiated.
A contract can leave a choice unsettled even when the possibility was foreseeable. Describing a contingency precisely may cost more than it is worth; the parties may observe the relevant fact but be unable to prove it to someone charged with enforcing their agreement. Incomplete-contract theory collects these difficulties into an allocation of residual control: particular uses can be assigned by agreement, while the owner retains choices over the asset that have not been assigned elsewhere. This is a way to analyze ownership, with legal and contractual limits built around it. A lien, a restriction on use or a right already granted to another party can bind an owner. A title does not make the surrounding obligations disappear.
The residual choice changes the bargain before it changes the use of the machine. If publisher and printer can negotiate without cost once the second run becomes worth doing, they can arrange it whoever owns the press. Ownership can nevertheless alter how the additional gain is divided, because it alters what each can do if they fail to agree. The distinction matters before either party has invested. A printer deciding whether to learn a difficult process, or a publisher deciding how much effort to devote to an uncertain book, considers the return that later negotiation will leave. An arrangement that uses an existing machine efficiently can still discourage the preparation that would have made the machine more valuable.
Owning the asset can protect one investment while leaving another exposed. Consider a constructed workshop with one machine and two people, Ada and Ben. Each may acquire a different skill before bargaining over production. Ada’s skill costs her ten and adds sixteen to output; Ben’s costs him six and adds ten. The increments are additive, the costs are private, and neither skill can be contracted for in advance. Whoever owns the machine can use her or his own acquired skill alone; the nonowner cannot produce without access. When they cooperate, suppose each receives the return available alone plus half of the additional gain their cooperation makes possible. These assumptions specify both what ownership changes and how bargaining distributes the result. Without them, the title on the machine would tell us too little to calculate an incentive.
Ada’s ownership makes Ada’s skill worthwhile and leaves Ben’s unacquired. She can retain the sixteen created by her own skill, which exceeds its cost of ten. Ben can obtain only half the ten his skill would add to their cooperation, and five does not cover six. The resulting net surplus is six. Give the machine to Ben and he acquires his skill, retaining ten against its cost of six, while Ada declines to spend ten for the eight her skill would return in bargaining. That arrangement yields four. Ada’s ownership is better among these two possibilities, yet neither calls forth both worthwhile skills. If both could be secured, their combined value of twenty-six would exceed their combined cost of sixteen by ten. Choosing the better owner has recovered some of the gain without reaching the whole.
The result depends on the investments the rights induce, rather than on a general rank of persons entitled to own. If Ben’s skill in this example costs four instead of six, he acquires it even without the machine: his share of five now covers the cost. Under Ada’s ownership both skills are acquired and the net gain becomes twelve, the same as the gain from securing both directly. If both investments can instead be specified and enforced in advance, the original six-and-four comparison no longer governs the incentive problem. The change that matters may be better contracting, a different technology or a less costly skill. Ownership theory asks how control alters the relevant decisions. It does not establish that every transfer must harm the nonowner’s investment or that no arrangement can ever secure both parties.
The title matters before the bargain because it changes what each party expects to keep afterward.
Buying a supplier changes the rights held by people whose motives can remain exactly as they were. A supplier’s manager does not cease to value discretion when the supplier becomes a subsidiary. Grossman and Hart press precisely this question against an effortless cure for holdup: what has integration changed that prevents the same difficulties from reappearing inside the enlarged firm? Their answer concerns the reassignment of control. The buyer gains choices previously held by the supplier; the supplier’s manager loses choices that may have sustained his willingness to invest. Preserving yesterday’s payment formula need not preserve yesterday’s position if the new owner can intervene in matters the formula never specified. An acquisition can therefore carry a cost even when its advocates promise to leave the business running as before.
Control of the means of work also affects people who hold no title. Hart and Moore study the ability to exclude particular workers from an asset: the person controlling the machine, premises or customer list can affect the terms on which another person expects to remain productive. Their distinction between dismissing one participant and abandoning an independent supplier’s whole organization clarifies why asset ownership can matter inside an employment relationship. The worker’s knowledge remains her own. An owner may need her cooperation badly enough that she retains considerable bargaining power. The theory concerns the dependence created by access to assets, and the incentives produced by that dependence; it gives no warrant for treating human judgment as another machine transferred with a deed.
A claim on the proceeds and a right to decide the asset’s use answer different questions. The previous plate followed the manager’s financial stake; this one follows what can be done when the existing agreements provide no decision. The two often travel together, but a promise to pay someone a share of proceeds need not give that person control over the machine producing them. Conversely, a person exercising control may owe much of the resulting income to others. To understand a firm’s boundary, one must follow the particular rights that move with an acquisition and the choices they influence. To understand its governance, one must then ask how the people exercising those rights are themselves constrained.
The owner decides part of what remains open, within the rights the institution actually confers. That position can call forth investment, discourage it, or change who must be persuaded before work can proceed. The firm’s boundary is consequently a distribution of future choices as well as a line around present assets. A useful arrangement places those choices where their effects justify their costs; it cannot settle the question by declaring every market dangerous or every hierarchy safe. Even after the rights have been assigned as carefully as possible, another uncertainty remains: whether the undertaking they support will prove worth doing. The next plate turns from the control of the means to the venture for which the means are assembled.