The Property Rights Approach (Grossman and Hart, Hart and Moore)

Why the Property Rights Approach?

Agency theories of the firm do not distinguish between contracts within or between firms. The firm is a "nexus of contracts." Although Demsetz characterizes the "entrepreneur" as the individual with contracts with all inputs (physical and human), one could just as easily consider each individual a firm in itself, and the contracts between them and the entrepreneur a contract between firms.

"Nexus of contracts" does not delineate the boundaries of the firm. It also does not explain why firms would or would not merge, since every contract that can be written within a firm can be written between them. (Some claim that there is informational advantage to within-firm relationships, but it is unclear why this would be the case.)

Transaction cost economics assumes that agency problems (with respect to relationship specific investments) are mitigated when a transaction takes place "within the firm." But it does not explain why this is the case. Like above, why does the information structure or cost of monitoring change when one firm buys out another? After all, the same individuals, physical assets, production processes, etc., are at hand...Why does the hold-up problem diminish under under vertical integration?

To answer these questions, we need a meaningful definition of "ownership."

What is Ownership?

Ownership is defined as the person or group who has the residual rights of control over physical assets.

The firm's boundary is the set of physical assets over which this person or group has residual control rights.

One important control right is the ability to exclude others from using the asset. Another is the right to specify the way it is used. Whatever rights are not in the agreement go to the owner. (Landlords have residual control rights, as do all or most lessors.)

Firing an employee is not like firing a grocer, because when you fire an employee you deny him access to capital. You are terminating a labor contract; in the case of the grocer you have the ability to exclude use of capital.

When one firm is acquired by another, what happens is the acquired firm's manager loses the residual rights of control over their physical assets. It is transferred to the acquiring firm.

When this happens, the acquired firm's manager loses the right to deny the acquiring firm access to its physical capital -- although he still has the ability to deny it access to his human capital (he can quit).

[Since residual rights of control of human capital are non-transferrable, firms are always defined along the lines of physical capital.]

This affects the bargaining power of this manager -- his ability to extract the gains from trade -- particularly in cases where he himself can be easily replaced and where the physical capital has some degree of specificity. (If the manager is not easily replaced, he can still extract rents; if the physical capital is non-specific, he did not have bargaining power in the first place.)

Residual control rights affects bargaining power of individuals. The individual with more bargaining power is often better equipped to protect the value of his specific investments -- they become less appropriable. So reallocating ownership -- reallocating residual control rights -- affects the investment incentives of various indivduals.

Implications on Incentives

Revisiting GM and Fisher.

When GM and Fisher were separate firms, GM did not have residual control rights over Fisher's assets. This meant that Fisher's managers could withhold access to these assets (refuse to trade). Anticipating this, GM did not make efficient relationship-specific investments.

When GM bought out Fisher, it obtained residual control rights over Fisher's physical assets. While Fisher's managers could still withhold their labor, they could not prevent GM's managers from using Fisher's physical assets to produce auto bodies for GM, and could not prevent them from tailoring Fisher's physical assets to produce especially for GM.

Before integration, GM had 60% of Fisher's stock. But their residual claimancy did not provide them with residual control rights. This allowed Fisher's managers to extract rents ex post. The movement from 60% to 100% included a transfer of residual control rights. This ameliorated the underinvestment problem.

The allocation of bargaining power explains why the degree of hold up-related inefficiencies changes under different organizational structures. This is why vertical integration may improve matters. The ability to deny acquired firms' managers the rights to use the firm's physical assets limits the means through which they can appropriate quasi-rents from specific investments.


Suppose the following.

Assume investment levels non-verifiable (not contractible) and no wealth effects.

Assume that price is set in period two such that, conditional on the investments undertaken, participants split ex post gains from trade evenly.

[Let the two managers be software development and marketing people. The human capital is investment in how to create the program and how to market the product. The physical capital is the IT used in development, and the packaging materials, manuals, etc. used to market the product.

Investment in human capital is observable, but not verifiable (and not contractible). The developer sells the program to the marketer, who in turn retails the product.]

M1: [software marketer]

M2: [software developer]

Ex post total surplus:

Ex post gains from trade for each manager, conditional on i and e:


Because neither manager internalizes the full value of their investment, both will choose investments that are too low. Public good/externality problem. Therefore, neither realizes the full gains from trade.


Suppose the marginal gains from investing is greater given that trade takes place than when it does not. R'(i)>r'(i), |C'(i)|>|c'(i)|. This implies investments in your human capital are in part specific to the other manager.

[The marketing reserach is most valuable if the marketer is actually able to obtain and market the product. The software development research is most valuable if this particular manager is the one who markets the product (has special ability to market the product).]

Suppose that the marginal gains from investing are weakly greater, given that no trade takes place, the more physical assets are owned. r'(i;a1,a2)>=r'(i;a1)>=r'(i;0), c'(i;a1,a2)>=c'(i;a2)>=c'(i;0) [in absolute value].

[Suppose negotiations break down between the managers. The software developer's investments are most valuable if he can use his IT and the distribution network, packaging, manuals, etc. to patch something together to get the product on the market. The marketer's research is most valuable if he has access to the developer's IT, etc. so that he can get someone else to cobble together a similar, but inferior program. The marginal returns from investing in human capital is least if they can be denied access to both marketing and IT physical capital.]

Simple implications:

Marginal returns from investment is higher if you own both assets than if you own one, and greater if you own one than if you own neither.

You will invest more if you own both assets than one, and more with one than both, but will underinvest relative to the first best in all cases.

[Investments in human capital are better protected (have a higher marginal return if no trade takes place) if you own both assets, and are less protected if you own neither.]

Suppose M1 has both assets. Then he will invest more, and M2 will invest less, than if the assets are separately owned. Is this the best alternative?

The problem:

Which allocation of assets, ex ante, generates the investment decisions that imply the largest gains from trade?

Under what conditions is each ownership pattern the best?

When one party's investment decision is "inelastic" -- it does not change with the prospect of appropriation -- the other person should have ownership.

If he does not own the asset, he will not underinvest much if at all; the other will increase his investment considerably.

[Suppose that the software developer will (perhaps because the marginal returns do not vary much, or because he has an intellectual interest) do the same amount of research no matter what physical assets he owns. Then it makes sense for the physical assets to be owned by the marketer, who will do more research if he owns both sets of assets. This gives him bargaining power, but this will not change the developer's research agenda.]

When one party's investment decision does not affect the gains from trade, the other person should own both assets.

Underinvestment by the first manager does not affect the size of the pie; transferring the asset to the other manager will increase his investment and therefore total surplus.

[The software developer's research agenda may change a lot if bargaining power is reallocated to the marketer. But under these assumptions, additional research in product development does not make the product more valuable, but additional research in marketing may do so. So one can provide stronger incentives for the latter at little cost by allocating physical assets to the marketer.]

When one party's human capital is "essential" in that the other party's marginal returns from investment, given that no trade takes place, is the same no matter what assets they hold, then this manager should own both assets.

Essential here means that the marginal returns from investments in human capital increase only if (through trade) it is combined with the other's human capital, and not if it is combined with physical assets.

Investment decision by the manager with the non-essential human capital is independent of asset ownership, so there are only benefits from allocating all physical capital to the manager with the essential human capital.

[Suppose that if negotiations break down, the returns from the software developer's human capital are the same regardless of whether he owns (or can be excluded from) his IT, and whether he owns (or can be excluded from) marketing materials. Then allocating him these capital assets would not change his investment decision. Allocating the marketer these assets may change the marketer's investment decision, however.]

When the marginal returns from investment is the same if a manager as both assets as if he has one, but greater if he has one than if he has zero, (and this is true for both managers) one should have independent ownership

In this case, there are no gains from higher investment from the individual with common ownership, but losses from the lower investment from the one without it. This is the "competitive market for each others' physical assets result."

[Suppose that, if negotiations break down, the returns from the developer's and marketer's human capital investments are no higher if they own their marketing and development physical capital, respectively. If they are denied access to each other's physical capital, they can simply use someone else's capital. For example, the IT developer can purchase "off the shelf" marketing materials such that the return on human capital investment is not diminished, or if the marketer can buy any physical capital the software developer uses "off the shelf." Note that this doesnot mean that they can buy each others' expertise or human capital in competitive markets. But they diminish considerably if they can be denied access to their own physical capital. We are assuming that denying the software developer access to his own machines lowers the return on his human capital investment considerably, as does denying the marketer access to his own capital. Then each should own its own machines. If one has common ownership it will not increase his own investment, both it will decrease the other's.]

When the marginal returns from investment is the same if managers hold one asset as when they hold zero, but greater when they hold two than when they hold one, then one or the other should hold both assets.

Common ownership of complementary goods.

[Similarly, suppose that, if negotiations break down, the marginal returns from the developer's human capital do not depend on whether he owns the IT, but is much higher if he owns both the IT and the marketing materials (you need both to be able to effectively market the product on your own) - and likewise for the developer. In this case, buying "off the shelf" substitutes for each other's physical capital reduces the returns from one's own human capital investment considerably; given that you can be refused access to the complementary physical capital, whether you can be refused access to the capital that you commonly use does not affect your investment decision. Then one or the other should own both physical assets -- you don't diminish investment incentives of the one losing the physical asset, but you increase those of the one gaining control of it.]

What is the point?

The point is that the allocation of residual control rights can affect one or both managers' incentives to make investments which increase the gains from trading with each other relative to doing so with others. It affects their returns on investments, which in turn affects their investment decisions, which in turn affects the gains from trade.

A theory of the firm must account for how best to allocate the bargaining power which is sometimes associated with the allocations of residual control rights.

The returns from investments in human capital may depend on access to specific pieces of physical capital. When others can deny you this access, you will tend to underinvest more. If this underinvestment substantially diminishes efficiency, then this individual should generally own the physical capital.

This theory not only depicts the advantages and disadvantages of vertical integration compared to independent ownership, but it predicts who will be the owner.