150 The Multinational Enterprise as an Economic Organization
locational pressures—the bauxite mine where the bauxite is, bauxite converted
to alumina at the mine because the process is strongly weight-losing, and the
smelter that converts alumina into aluminum near a source of low-cost electric
power. The question is, why do they come under common administrative control?
The proprietary-assets model is not necessary, because neither upstream nor
downstream production unit need bring any distinctive qualification to the parties’
vertical consolidation. Some proprietary advantage of course could explain which
producer operating at one stage undertakes an international forward or backward
vertical integration.
Models of Vertical Integration
Until the rise of transaction-cost economics the economic theory of vertical
integration contained a large but unsatisfying inventory of special-case models.
Some dealt with the physical integration of production processes: If you make
structural shapes out of the metal ingot before it cools, you need not incur the
cost of reheating it. Such gains from physical integration explain why sequential
processes are grouped in a single plant, but they neither preclude two firms sharing
that plant nor explain the common ownership of far-flung plants. Another group
of traditional models regard vertical integration as preferable to a stalemate between
a monopolistic seller and a monopolistic buyer, or to an arm’s-length relation
between a monopolistic seller and competitive buyers whose activities are distorted
due to paying the monopolist’s marked-up price for their input. Some models
explain vertical integration as a way around monopolistic distortions, while others
explain it as a way to profit by fostering such distortions.
The theory of vertical integration has been much enriched by the same
transaction-cost approach that serves to explain horizontal MNEs. Vertical
integration occurs, the argument goes, because the parties prefer it to the ex
ante contracting costs and ex post monitoring and haggling costs that would
mar the alternative state of arm’s-length transactions. The vertically integrated
firm internalizes a market for an intermediate product, just as the horizontal
MNE internalizes markets for proprietary assets. Suppose that there were pure
competition in each intermediate-product market, with large numbers of buyers
and sellers, the product homogeneous (or its qualities costlessly evaluated by
the parties), information about prices and availability in easy access to all parties
in the market. Neither seller nor buyer would then have reason to transact
repeatedly with any particular party on the other side of the market. When these
assumptions do not hold, however, both buyers and sellers acquire motives to
make long-term alliances. The two can benefit mutually from investments that
each makes suited to special attributes of the other party. Each then incurs a
substantial fixed cost upon shifting from one transaction partner to another. Each
seller’s product could be somewhat different, and the buyer incurs significant
costs of testing or adapting to new varieties, or merely learning the requirements
and organizational routines of new partners. The buyer and seller gain an incentive
to enter into some kind of long-term arrangement.