Production occurs because of the interaction of specialists. Call it a supply chain. These specialists discover and execute the terms of trade (and contracts) that link them. Alternatively they could be managed "under one roof." The latter introduces management and monitoring costs. The economic problem, then, is a matter of comparing the savings in transactions costs vs the extra monitoring costs of putting selected parts of the supply chain under one roof -- or in one shell. The entrepreneur who successfully oversees all this gets the rewards that accrue to a residual claimant.
A more recent (and surprisingly separate) literature adds location and space to the mix. Call it "urban economics" or call it spatial organization (SO). Clusters of selected activities form to economize on transactions costs -- including the possibility of internalizing externality benefits and also avoiding some externality costs.
Textbook discussions cite the original Disneyland (Anaheim, 1950s) which was developed in a rural area – and soon enriched nearby landowners. The company learned a valuable lesson and Disney World (1970s) was developed in light of the experience; the corporation bought surrounding land (or options to buy) so that the benefits of adjacency would accrue to it. Simple economics suggests that decisions on how big the newer park should be would have been affected.
The story is succinctly described in my favorite journalistic allusion, from The Economist (March 1, 1977).
“In many economics textbooks, the presence of externalities is invoked as a justification for government intervention in the marketplace. Yet the private sector often finds its own solutions to externality problems. This is the secret of the shopping mall’s success. Because a property developer owns the entire shopping complex, its profits depend on the entire mall, not on any particular shop. By choosing the right mix of tenants and charging rents that reflect each store’s contribution to the mall’s overall revenues – including the business it brings to other stores – the developer can ‘internalize’ the externality and maximize its profits.” Eric Gould et al (as well as others) eventually introduced these ideas to an academic audience.
In this example, various potential external costs between locators can be avoided by propitious spatial arrangements; various external benefits can be reaped; various transactions costs can be reduced. The mall's developer seeks the rent-maximizing (profit-maximizing) spatial arrangement by strategic choice of site rents – which can include rent subsidies to “anchor” stores -- and also charging rent premiums to nearby tenants. Foot traffic benefits as shoppers economize on search costs are noted and included. Visitors to modern shopping malls do more than shop. They may also value opportunities to mingle and hone social capital. Finally, mall land use and infrastructure plans are necessarily integrated.
Just as the firm described by Alchian and Demsetz has size and scope limits, there are size and scope limits for the mall or any similar cluster. Land assembly costs that are subject to local rules and politics come to mind.
Finally, what difference does it make? The IO and SO problems are not independent. There is the problem of determining what goes on inside vs outside the firm -- and there is the concurrent problem of determining what goes on inside vs outside the cluster. Looking at it this way, it is hard to imagine the two problems as ever separate or independent. Land consumption cannot be assumed away; spatial arrangements, not simply "density," are involved. Among the many coordination problems to be solved include the internalization of various externalities -- which can thereby be removed from the slate of problems with which we burden zoning and public policy discussions.
As a matter of public policy, make land assembly easy. The many land use rules (often adopted with an eye to controlling externalities) we have are top-heavy because the economic forces involved are misunderstood.