Numerous studies have shown that since about 1900 the largest industrial, financial, and transportation companies have interlocked extensively through a sharing of directors and, much less often, officers (e.g., Bunting and Barbour, 1971; Bunting, 1976a, 1976b; Dooley, 1969). In fact, this interlocking has been so extensive that a virtual network exists whereby nearly any large corporation in principle is able to participate either directly or indirectly, once or twice removed, in the top-level policy deliberations of any other large concern (Mizruchi, 1982; Pennings, 1980).

Little is known about the origins of this network. Most research implicitly assumes that the network has resulted from some relatively recent decline in competition and subsequent movement toward economic concentration. This conclusion follows from the commonly accepted proposition that competition precedes monopoly in industrial development. Scherer (1979:47), an authority on modern American industrial organization, cites Marx for the essence of this notion: “‘One capitalist always kills many’ (creating) a ‘constantly diminishing number of the magnates of capital, who usurp and monopolize all the advantages of this process.’” In less colorful but more factual terms, Burns (1936:1-42) and many others have described the “decline of competition” and the factors leading to the domination of many industries by relatively few large corporations (literature reviewed in Scherer, 1979:67-70; Blumberg, 1975:16-83).

Many social scientists have documented that the American corporate capitalist class is highly organized and cohesive, both in terms of formal structural characteristics (Domhoff, 1967; Useem, 1979; Scott, 1979) and informal social patterns (Baltzell, 1964; Domhoff, 1974; Mills, 1956). One of the major structural features integrating corporate capitalists is interlocking directorates, which have been shown to form a single, continuous, cohesive network (Sonquist and Koenig, 1975; Allen, 1974; Mariolis, 1975; Mizruchi, 1982; Pennings, 1980; Scott, 1979; Fennema and Schijf, 1978).

The period around the turn of the twentieth century has been called the “era of finance capital” in American history. It was during this period that the first of the great trusts were formed, combining in some cases hundreds of corporations into one. Investment bankers, in particular J. P. Morgan, were the major orchestrators of these schemes, but other leading financiers and industrialists were also active in controlling several corporations apiece. George F. Baker and James Stillman in banking; Edward H. Harriman, James J. Hill, and the Moore brothers in railroads; and John D. and William Rockefeller in oil and other industries were prominent among the leading capitalists of the age, and they dominated much of the business scene at that time.

One of the remarkable aspects of this period is the rare unanimity among scholars and journalists about its main characteristics. There is general agreement that financial corporations were the most powerful of the period, that a relatively small number of individuals dominated the corporate world, and that if it was ever possible to speak of a ruling class in American society, this was the period (e.g., Riesman, 1953; Bell, 1960; Parsons and Smelser, 1957).

Many recent studies of interlocking directorates have paid special attention to the possible existence of interest groups or cliques within the corporate world (Allen, 1978; Dooley, 1969; Mariolis, 1977; Mintz and Schwartz, 1981a; Mizruchi, 1982b). The search for economic groupings of this type dates from the early part of the twentieth century when financial domination of industrial firms was thought to characterize American business. Bank control theory, in particular, posited the coalescence of individual companies into a number of competing groups, each organized around banking interests. Thus, Baran and Sweezy (1966: 17) define interest groups as “a number of corporations under common control, the locus of power being normally an investment or commercial bank or a great family fortune.”

Early work in this area, therefore, attempted to identify separate groups of companies organized around either family or financial interests. Rochester (1936), for example, used information on stock ownership, banking dependence, service relationships, capital investments, and joint ventures to identify many autonomous cliques headed by founding families. Sweezy (1939) used a similar set of criteria to identify eight dominant financial interest groups. Perlo (1957) replicated these results 20 years later.

Numerous network studies demonstrate that interlocking directors unite corporations in the United States into a tight-knit corporate community (e.g., Sonquist and Koenig, 1975; Mariolis, 1975; Mizruchi, 1982). In addition, nonnetwork studies of foundations, universities, and other nonprofit organizations suggest that corporate officers and directors play a large role in the governance of such institutions (e.g., Hartnett, 1969; Dye, 1976; Useem et al., 1976; Salzman and Domhoff, 1980; Useem, 1981). However, there are no studies that cast light on the structural position of nonprofit institutions in relation to the well-established corporate network.

Two different theoretical frameworks, the interorganizational and the intraclass, guide empirical research on the origin and function of corporate interlocks. Conclusions from this research often have been used as the basis for more general claims about the structure of power in the American economy. To a great extent, these claims are based on the implicit assumption that a majority of interlocks facilitate one type of interfirm relationship—formal coordination.

This article presents evidence about the continuity of interlocks which have been disrupted accidentally by such events as death and retirement. What happens after an interlock is broken reveals something of the extent and manner in which different types of interlock ties facilitate formal coordination. This in turn provides the basis for assessing the applicability of the interorganizational and intraclass approaches to the phenomenon of interlocking.