1. Backward Integration
Backward integration into raw materials may occur for three main reasons: (I) to realize prospective transaction cost economies; (2) for strategic purposes; or (3) for-mistaken reasons. Transaction cost economies will warrant integration where the parties are tightly joined in a bilateral exchange relation, making problems of harmonizing the interface crucial, and where integration does not sacrifice economies of aggregation. The acquisition of the Mesabi iron ore deposits by steel companies may qualify (Parsons and Ray, 1975), though others favor a strategic explanation (Wall, 1970). Acquisitions of coal and limestone deposits by steel companies appear to lack either transaction cost or strategic purpose and were possibly mistaken. As Joskow’s recent study discloses, however, detailed knowledge of the feasible contracting options is needed (Joskow, 1985).
An illustration of what was held to be strategic backward integration, undertaken for the purpose of forestalling rivalry, is Alcoa’s acquisition of bauxite deposits and hydroelectric sites. Allegations to this effect were made in conjunction with the famous antitrust suit in which the Justice Department charged Alcoa with monopoly and monopolization.5 The definitive assessment of backward integration into bauxite deposits has since been made by John Stuckey (1983). In very gross terms, the cost difference of processing a mixed-hydrate bauxite, which is efficiently processed with a high-temperature technology, in a low-temperature refinery instead, comes to almost 100 percent (Stuckey, 1983, pp. 53-54). But the details also matter. Bauxite storage covers are needed for some ores and not for others (p. 49); residue processing costs vary greatly (p. 53); and air pollution equipment is tailored to the attributes of the bauxite (p. 60). Moreover, although smelting is less idiosyncratic, there is, nevertheless, an “art part of smelting,” which is upset if the aluminum supply is varied (p. 63).
Stuckey refers to both physical and site specificity in his summary assessment:
The bilateral-monopoly nature of bauxite-exchange relationships results from several largely immutable technical and structural factors. First, bauxite is a heterogeneous commodity, and the ore in any deposit has unique chemical and physical properties. The efficient processing of a given bauxite usually requires a tailor-made refinery with specially designed technologies for chemical processing, materials handling and waste disposal. Once a mine and its associated infrastructure is developed, and its appropriately designed refinery is constructed, ihe two plants are locked together to the economic extent of their technical complementarity. The evidence indicates that, in economic terms, the complementarity is often strong, meaning that within a wide range of bauxite transaction prices the mine and refinery are wedded together economically.
A second set of factors locking mines and refineries together includes the wide geographical spread of the world’s major bauxite deposits, the vast distances between the deposits and primary smelters, the low value of bauxite at the mine front door relative to freight rates, and the over 50-percent reduction in material volume during refining. The last three factors encourage the back-to-back location of mines and refineries for transport cost reasons. [1983, p. 290]
Stuckey also notes that information asymmetries regarding the quality and extent of bauxite deposits complicate problems of long-term contracting (pp. 291-92).
The incentive to integrate because of asset specificity of both physical plant and site specificity kinds is thus reinforced by information asymmetry considerations. The upshot is that backward integration from refining into bauxite is an altogether predictable outcome from a transaction cost economics standpoint—which presumably explains why this is the preponderant or- ganizational form.
Manufacturers appear sometimes to have operated on the mistaken premise that more integration is always preferable to less. From a transaction cost point of view, the following examples of backward integration would appear to be mistakes (and, 1 conjecture, have mainly been abandoned as renewal decisions have presented themselves): (1) backward integration by Pabst Brewing into timberland and barrel-making plants (Chandler, 1977, p. 301); (2) backward integration by Singer Sewing Machine into timber, an iron mill, and some transportation (Chandler, 1977, p. 305); (3) backward integration by the McCormick Company into timberlands, mines, twine factories, and hemp plantations (Chandler, 1977, p. 307); and (4) Ford Motor Company’s fully integrated behemoth at River Rouge, supplied by an empire that included ore lands, coal mines, 700,000 acres of timberland, sawmills, blast furnaces, a glass works, ore and coal boats, and a railroad” (Livesay 1979 p. 175).
To be sure, managers, like others, are reluctant to concede mistakes. Accordingly, mistaken integration may not quickly be undone. Moreover some, like “the River Rouge behemoth,” involve the construction of facilities in cheek- by-jowl proximity to each other. Such site specificity forces the parties into a bilateral trading relation and for this reason is apt to be continued. Distant twine factories, by contrast, can be sold off or closed down with ease.
Site preemption issues aside, backward integration that lacks a transaction cost rationale or serves no strategic purposes will presumably be recog-nized and will be undone.6 The discontinuation or sale of mistaken integration activities will occur more rapidly if the firm is confronted with active rivalry.
Source: Williamson Oliver E. (1998), The Economic Institutions of Capitalism, Free Press; Illustrated edition.