Structural Redundancy and Multiplicity within Networks of US Corporate Directors

Abstract

The argument of this paper is that social ties beyond the boardroom matter. At the structural level, social ties from non-profit foundations, cultural organizations, university boards, policy planning organizations, and private social clubs add unique ties to those established by interlocking directorates. In addition, the multiple configurations of these social affiliations increase the centrality and reduce the social distance among the corporate directors. Although visualizations of these social networks underscore the changing roles these extra-corporate ties play in uniting corporate directors, they cannot precisely capture the unique effects of ties created by social affiliations outside the boardroom. Utilizing the concepts of structural redundancy and multiplicity, the paper describes the unique contribution each set of ties makes in constructing the social mosaic of corporate America during the second half of the 20th century. The results indicate a dramatic reduction in the role private clubs play in supplementing corporate ties, and an equally significant rise in the importance of social ties generated by policy planning organizations.

Keywords

power structure research interlocking directorates network analysis policy planning organizations social clubs

Introduction

This paper advances G. William Domhoff’s argument that there exists a small but cohesive upper class by examining the social ties among US corporate directors during the last half of the 20th century. In his seminal work, Who Rules America, Domhoff (1967) shows how this relatively small group of individuals who share common values and world views exert an enormous amount of power on society. Such influence is made possible by the group’s cohesion, which is maintained through shared memberships in social institutions, corporate boards, and an array of non-profit organizations. The results within this paper document the changes in this upper class community by systematically identifying the unique contribution made by a given set of ties across multiple venues between 1962 and 1995. Politically, this 33-year period begins with efforts by business interests during the Johnson administration to lower corporate tax burdens, increase payroll taxes, and design and operate social programs in ways that benefited corporate America (Ferguson and Rogers, 1986). However these efforts pale in comparison to the intense political mobilization by the right under the Reagan administration when business interests enjoyed further tax cuts, increased military spending, decreased social spending, attacks on labor, and deregulation. The empirical argument here is that during this historical period, the ties existing beyond the boardroom mattered. Structurally, these social ties increase the number of unique connections among the group of corporate directors and shrink the social space among the directors of some of the most powerful US corporations.

Interlocking Directorates as Foundational

Early documentation and mapping of the inter-organizational ties among the directors of the country’s major corporations (Domhoff, 1967, 1970, 1983, 1998; Mills, 1956) has been foundational to the contemporary analyses of the exercise and structure of economic and political power in the United States.¹ Dye’s work (2001) documenting ‘top down policymaking’ in the United States begins with the national, i.e. corporate, elite. Similarly, Useem’s concept of the ‘inner circle’ emphasizes how interlocking corporate directors obtain information on ‘contemporary business practices and the general business environment’ (1984: 85). Moving the analyses from people to positions, Mizruchi’s (1982) longitudinal study of corporate interlocks demonstrates that, by 1974, financial corporations had begun to occupy a more general role in the network due to their central positions, rather than the exercise of power over specific nonfinancial corporations. Similarly, Scott and Griff (1984) found that intercorporate relations were less a matter of personal power, but instead a function of the ‘institutional power’ or the structural positions of the enterprises. Therefore, within networks of interlocking directorates, certain directors become important by virtue of the number of ties they have with other directors and their close proximity to their colleagues. However, as Burris (2006) has emphasized, corporate interlocks should be ‘only one thread in a much denser fabric of social ties among corporations and corporate elites’. Burris further cautions that one must not ‘reify or isolate director interlocks from other social networks within which they are embedded’. This paper represents my response to this charge to generate a deeper ‘sociology of corporate directors’.

Beyond the Boardroom: The Social Mosaic of Corporate Directors

Complementing the research on interlocking corporate directors, work has also examined the inter-organizational ties between corporate directors and non-profit foundations, cultural institutions, research universities, policy planning organizations, and elite social clubs. Beginning with the non-profit foundations, these organizations emerged at the turn of the 20th century as a ‘private sector alternative to the welfare state’ (Hall, 1987: 12). However, as Hall also notes, the establishment of much more open-ended foundations generated fears that ‘foundations, through their ability to channel huge amounts of money toward charitable objects at will, might become major instruments through which “the interests” could influence public policy’. Given the historical record, these fears were certainly not without merit. On the conservative side of the political spectrum, The Rockefeller Foundation has long supported the Council on Foreign Relations (Dye, 2001), an organization with a moderately conservative ideology (Burris, 1992). On the other hand, not all foundations operate to promote a conservative political agenda. The Ford Foundation has contributed funds to such environmental organizations as the Audubon Society, the National Resources Defense Council, and the liberal ‘think-tank’ the Institute for Policy Studies (Dye, 2001). Nonetheless, Arnove and Pinede (2007) argue that foundations such as the Ford, Rockefeller and Carnegie are essentially elitist and technocratic institutions that merely ameliorate the inequalities and injustices generated by the global capitalist economy.

The nation’s fine art organizations are not only repositories of artifacts and venues for classical music; sociologically, these organizations represent the institutional means by which culture is reproduced (Wolff, 1983). During the past 40 years, research has demonstrated a fairly consistent level of overlap between corporations and cultural institutions. Useem (1979) reports that 11.9 percent of the directors who held only one corporate board membership also served as a trustee or director of a cultural organization. This figure increases to approximately 21 percent for directors who sit on two or three boards and 27 percent for directors serving on four or more corporate boards of directors. Salzman and Domhoff (1983) arrived at comparable conclusions regarding the ties between corporate directors and seven cultural organizations – including the Metropolitan Museum, Museum of Modern Art, the Kennedy Center, and the Metropolitan Opera Guild. Similarly, Galaskiewicz (1985: 60) found a high level of mutual affiliations between the ‘corporate philanthropic elite’ and major CEOs within the Twin Cities metropolitan areas. More recently, Wu (2002) reports that the presence of the business sector on the Tate Museum’s board of trustees grew from 40 percent in the 1970s to over 80 percent in the 1980s.

As one of the main institutions dedicated to the production and distribution of knowledge, social scientists have long been interested in the ties between corporations and universities (McGrath, 1936; Veblen, 1918). According to Goldschmidt and Finkelstein (2001), corporations select particular university presidents to belong to their board for their knowledge, to exert influence over a particular university, or to foster corporate-community relationships. This recent observation is similar to Miliband’s earlier point (1969) that university professors and academic officers are increasingly utilized as business consultants. Alternatively, scholarship has also documented the growing number of individuals from the business community being recruited as university trustees or directors. Useem (1984) notes that as early as the turn of the 20th century there was higher representation on university boards by manufacturers and financiers, while Hartnett (1974) found that there were increasing incidences of trustees possessing a business background. Moore (1979) reports that there is a ‘central circle’ of key leaders of which 15.4 percent were from business and 7.5 percent were from academia. Using a more direct measure, Useem (1980) reports that approximately half of the directors who sit on corporate boards of directors also serve as a trustee for some college or university and that ties with colleges and universities are more prevalent among directors who serve on the boards of large industrial corporations and commercial banks.

Policy planning organizations are important to this mix of social ties because it is through such organizations as the Business Roundtable or the Council on Foreign Relations that members of the business community articulate their interests and exert influence. Useem (1984: 71) notes that although no single organization unanimously speaks for American business interests, ‘the fragmentation is not a competitive one … for the key associations perform largely complementary tasks’. Similarly, in reference to the Council on Foreign Relations, Domhoff (1970: 112) concludes that such an organization ‘is the key middle term, so to speak, between the large corporations on the one hand and the federal government on the other’. Particularly noteworthy is the work on 12 major policy planning organizations (Burris, 1992; Peshick, 1987). In addition to nicely summarizing the origins of these policy planning organizations, Burris categorizes these organizations according to whether they are primarily lobbying groups or research oriented, and whether the policy planning groups are moderate or conservative in their political ideology. More importantly, his work shows how these groups have influenced public policy: the Heritage Foundation’s Mandate for Leadership served as a catalyst for such initiatives as ‘Star Wars, enterprise zones, and supply-side economics’ (Burris, 1992: 119). More recently, Domhoff (2014) has reviewed documents from the Council on Foreign Relations to demonstrate how ‘the Council’ was deeply involved in both the escalation of the Vietnam war in 1964 and the withdrawal four years later. Therefore, analyzing the participation of the corporate elite in policy planning organizations is a cornerstone for understanding how elite corporate directors exercise political influence.

Finally, exclusive social clubs have historically been a private venue where members of the upper class meet, including corporate directors. Despite their steadfast protection of members’ privacy, many people have studied the affiliations within and between elite private social clubs. In analyzing the club membership for the directors of the 20 largest corporations in the 1960s, Domhoff (1983) was able to identify the clubs with the most directors, as well as which corporations had the highest number of social club memberships. Domhoff’s (1983) research has also examined the social club memberships for the presidents of the top 100 industrial firms in 1958 and found that 70 percent of the presidents belonged to one or more upper-class club. Earlier, Baltzell (1958) reported that 64 percent of the men in his Philadelphia study claimed to be in at least one club; but among businessmen, 76 percent reported to be in at least one club. Soref (1976) studied the connection between corporate directors and the upper class by looking into directors’ club memberships as an indicator of being a part of the upper class. Although social clubs are an elusive feature of American business culture, they serve important roles in contemporary capitalism – as locations to execute business deals (Sheehan, 1958), occasions to formulate positions on public policy (Domhoff, 1974), or simply providing seclusion where members of the corporate elite reinforce social ties.

Given this array of social affiliations, the following analyses provide a systematic accounting of the effects of the different sources of social ties among corporate directors between 1962 and 1995. The analyses not only examine the unique contribution of each type of tie over time, they will also aggregate these ties to reveal their cumulative effect on the social network formed among these corporate directors. The results indicate that non-corporate ties remain an important feature of the social network that exists among the directors of the nation’s top corporations (although the set of ties have become increasingly corporate in character). Secondly, the set of non-corporate ties have changed from being characterized by exclusive private clubs to ties dominated by shared affiliations with policy planning organizations. In short, the social mosaic continues to exist, albeit in a different configuration.

Data and Method

The data for this paper encompasses six different sources of social ties: the board of directors for a wide array of corporations, the governing bodies of the leading non-profit foundations, the boards of directors of the leading cultural organizations, the trustees or governors of the largest research universities, the boards of directors for the dozen most influential policy planning organizations in the United States, and finally, self-reported membership in a widely recognized set of exclusive social clubs across the country. Given the challenges of compiling the data on non-corporate affiliations (especially for the earlier time periods of this study), it was not possible to complete both the archival research required to locate and record all of the affiliation data and simultaneously extend the data set to include more recent cross-sections. Replicating this methodology to include more recent time periods is certainly in order and a direction for future research.

Corporations

While many important studies of interlocking directorates in the United States observe the convention of using all of the corporations in the Fortune 500, e.g. Davis et al. (2003), I wanted to be certain to capture the array of corporations across types of industries, regardless of revenues. Therefore, I include individuals if they were board members of the top four corporations within the various industry surveys published by Standard and Poor’s. The data also include men and women who were directors of any corporation that was in the top 100 of the Fortune 500 in terms of revenues and the top 20 commercial banks, and the top 15 insurance companies. However, given that the network analyses require complete networks, I omit directors belonging to ‘isolated’ corporations and members of the boards of pairs of corporations tied only to each other. This strategy yields the following ‘samples’: 1962 – 239 corporations and 2850 directors; 1973 – 2754 directors and 227 corporations; 1983 – 2732 directors and 235 corporations; and 1995 – 226 corporations and 2280 directors.

Social Organizations

Using the directors in each of the four time periods, I next identified the non-corporate organizations of interest using a ‘reputational’ method which balances the need to keep the number of organizations manageable with ensuring that the most significant non-corporate organizations are included. For non-profit foundations, I identified the top 25 foundations by assets in each of the four time periods using the Foundation Directory (1960, 1964, 1971, 1975, 1983, 1995). Among the foundations in this study are the Ford and Rockefeller foundations noted above in terms of their alignment with the Council on Foreign Relations and the National Resources Defense Council respectively. Because organizational affiliations are spatially located (Kono et al., 1998) I selected the top ‘cultural institutions’ by first identifying the 20 largest cities in 1998 and then used the largest art and nongovernmental museums by attendance in those cities according to the American Art Directory 2001–2002, and the major symphony orchestras and opera companies by operational budgets as reported in Musical America: International Directory of the Performing Arts, 2003 edition. This procedure generated a list of nearly 50 organizations in each time period. For research universities, I focused attention on the top 50 research institutions according to the National Science Foundation’s reports on federally funded grants received between 1972 and 1995. This procedure captures 46 unique boards of trustees or directors. With the assistance of Val Burris, I assembled the names of those individuals who occupy leadership positions in the 12 major policy planning organizations as recognized by previous research (Burris, 1992).² Finally, I selected the social clubs for this study based on an inductive and reputational basis. For all the directors with listings in the various Marquis Who’s Who publications, I recorded all self-disclosed memberships and limited the ‘sample’ to the 53 private clubs in the major metropolitan areas in the United States – the vast majority of these clubs were included in the national editions of the Social Register of exclusive clubs in 1980 and 1992. Of the nine clubs not listed in the Social Register, seven were listed by Bill Domhoff (2010: 238) as indicators of upper-class standing. Furthermore, by adopting this selection method these data capture social club affiliations in such cities as Atlanta, Detroit, Milwaukee, Minneapolis, Portland, and Seattle. A list of all US corporations, non-profit foundations, cultural organizations, research universities, policy planning organizations and, finally, exclusive social clubs is available on my Social Mosaic website (Barnes, 2015; see Table 1).

Table 1.

Number of organizations by type, 1962–1995.

	Year
	1962	1973	1983	1995
Corporations	239	227	235	226
Non-Profit Foundations	25	25	25	25
Cultural Organizations	47	51	53	51
Research Universities	46	46	46	46
Policy Planning Organizations	8	12	12	12
Private Social Clubs	53	53	53	53

Visualizing the Social Mosaic of Corporate Directors

Once the corporations and social organizations had been identified and listings of their directors obtained, I created a series of two-mode affiliation matrices for each of the time periods. A two-mode affiliation network is essentially a matrix of the individual directors in rows by their organizational affiliations in columns. For each of these two-mode networks, I then created director by director networks based on each set of affiliations. Using these ties, it is straightforward to visualize both sets of ties by treating corporate ties as arcs in one direction and the ties from, say, cultural organizations as arcs in the opposite direction. For example, consider the series of graphs for 1962 that begin with corporate ties alone and then add each of the other non-corporate ties to these arcs (Figures 1a –f). In addition one can illustrate the changing configuration of ties from corporate ties and another source of social affiliations over time. Figures 2a –d illustrate the changing set of ties formed through corporate and university boards between 1962 and 1995.

Figure 1a.

Total network of directors by level of interlocking, 1962.

Figure 1b.

Network of directors tied through corporate and foundation boards by level of interlocking, 1962.

Figure 1c.

Network of directors tied through corporate and cultural organization boards by level of interlocking, 1962.

Figure 1d.

Network of directors tied through corporate and university boards by level of interlocking, 1962.

Figure 1e.

Network of directors tied through corporate and policy planning boards by level of interlocking, 1962.

Figure 1f.

Network of directors tied through corporate and social club membership by level of interlocking, 1962.

Figure 2a.

Network of directors tied through corporate and university boards by level of interlocking, 1962.

Figure 2b.

Network of directors tied through corporate and university boards by level of interlocking, 1973.

Figure 2c.

Network of directors tied through corporate and university boards by level of interlocking, 1983.

Figure 2d.

Network of directors tied through corporate and university boards by level of interlocking, 1995.

Network Density and Structural Redundancy

While these graphs provide a quick overview of the contribution each set of non-corporate ties has on the set of ties formed through interlocking directorates, they do not provide precise measures of the added connectivity generated by the various pairs of ties across institution type, or the different configurations of ties longitudinally. To analyze the effects of the various sources of social affiliations among these corporate directors, I utilize a straightforward set of techniques which revolve around the concepts of structural redundancy and multiplicity of ties (Barnes and Burkett, 2010). Structural redundancy is best thought of as the opposite of a unique effect which occurs when new ties emerge in a network of interlocking corporate directors once the affiliations from other non-corporate organizations are included. To quantify these effects, I use network density – or the number of ties present divided by the total number of possible ties in the network. Because additional ties from shared membership in the same cultural organization may reinforce an already existing corporate tie, the first set of analyses reports the changes in network density when ties from each of the organizations are added to the corporate network. The larger the increase in the density, the greater the number of unique ties added by the organization.³

Table 2 presents the number of possible, unique, non-corporate and redundant ties, as well as network density, and the percent change when compared to the network of corporate ties alone. First, the most striking feature of the results in Table 2 is the overwhelmingly large effect of the ties generated by social club membership in 1962 and 1973. The percent change in network density when social club memberships are added to the corporate network is +166.7 and +121.3 percent in 1962 and 1973 respectively. This effect dissipates in 1983 where the addition of social club ties only increases network density by +51.9 percent. By 1995, the increase of +13.5 percent is no longer the largest effect, but is second to the role of policy planning ties.

Table 2.

Changes in network density with the addition of non-corporate ties, 1962–1995.

	Year and Source	Ties					Percent Change in Density
		Possible	Unique	Org Alone	Redun-dant	Network Density	Percent Change in Density
1962	Corporation Ties Alone	4,059,825	30,318			0.0075
	Corporation and Non-Profit Foundation		30,521	247	44	0.0075	0.00
	Corporation and Cultural Organization		31,235	1,126	209	0.0077	2.67
	Corporation and University		31,978	1,829	169	0.0079	5.33
	Corporation and Policy Planning Organization		31,269	1,002	51	0.0077	2.67
	Corporation and Social Club		81,023	53,188	2,483	0.0200	166.67
1973	Corporation Ties Alone	3,790,881	28,391			0.0075
	Corporation and Non-Profit Foundation		28,565	188	14	0.0075	0.00
	Corporation and Cultural Organization		29,509	1,319	201	0.0078	4.00
	Corporation and University		30,088	1,857	160	0.0079	5.33
	Corporation and Policy Planning Organization		30,079	1,763	75	0.0079	5.33
	Corporation and Social Club		63,013	36,316	1694	0.0166	121.33
1983	Corporation Ties Alone	3,730,546	29,376			0.0079
	Corporation and Non-Profit Foundation		29,528	174	22	0.0079	0.00
	Corporation and Cultural Organization		30,681	1,449	144	0.0082	3.80
	Corporation and University		31,087	1,867	156	0.0083	5.06
	Corporation and Policy Planning Organization		32,073	2,874	177	0.0086	8.86
	Corporation and Social Club		44,740	16,078	714	0.0120	51.90
1995	Corporation Ties Alone	2,598,060	19,188			0.0074
	Corporation and Non-Profit Foundation		19,259	80	9	0.0074	0.00
	Corporation and Cultural Organization		19,764	628	52	0.0076	2.70
	Corporation and University		20,283	1,180	85	0.0078	5.41
	Corporation and Policy Planning Organization		22,399	3,329	118	0.0086	16.22
	Corporation and Social Club		21,939	2,862	111	0.0084	13.51

Now, it is important to bear in mind that the time period for this study encompasses both the civil rights movement and the rise of feminism. Therefore, it is quite possible that the decline in social club ties observed is an artifact of self-reporting and not a substantive change in the affiliation patterns among these directors. To address this concern, I consulted membership lists for the Bohemian Club in San Francisco for the years 1964, 1972, 1986 and 1990, which Bill Domhoff graciously shared. Using these 4180 unique names across these four time periods, I used Who’s Who on the Web to determine rates of self-reporting membership in this all-male social club. There are two important results from these analyses. First, the rate of appearing in Who’s Who on the Web actually increased from 11 percent to 31 percent between 1964 and 1990. Second, among those members that appeared in Who’s Who on the Web, the level of disclosing membership in the Bohemian Club declines only slightly from 63 percent in 1964 to 58 percent in 1990. Therefore, while there is a modest reduction in the level of reporting one’s membership in the Bohemian Club, the magnitude of this decrease cannot completely account for the changes presented in Table 2.

There are two additional noteworthy results embedded in Table 2. First, it is clear that although there were unique ties added with the inclusion of affiliations from non-profit organizations, these non-redundant ties essentially added no new connections within the network in any of the four time periods. In contrast, the inclusion of cultural organizations and ties from university board memberships represent modest increases in network density in all four time periods. In 1962 and 1995, the addition of ties from cultural organizations increased overall network density by approximately 2.7 percent. In 1973 and 1983, the impact was larger with percentage increases in network density of 4.0 percent and 3.8 percent respectively. In terms of the effects of adding unique ties from shared membership on research university boards, the results in Table 2 show a consistent increase of around 5 percent in all time periods. The second feature of Table 2 is the steady increase in the importance of the ties added by policy planning organizations between 1962 and 1995. In 1962, the addition of ties from membership in policy planning boards increased overall network density by only 2.7 percent. The following time period, the contribution from policy planning affiliations increased network density by 5.3. This upward trend continues in 1983 with an 8.9 percent increase in overall network density. Finally, by 1995, overall network density increased by 16.2 percent – and for this last time period, this change represents the largest increase from any other source of social ties.

While the results in Table 2 show how the addition of the non-redundant ties from a given non-corporate organization decreases overall network density, the comparisons are based on the addition of one organization at a time. However, when one considers the ties from all of the organizations simultaneously, new redundancies may emerge. Therefore, to gauge the effect of a given organization on the total mosaic of social ties, one needs to use a different base-line – the density of the network from all ties. To determine the unique effects of one type of tie while controlling for the effects of all other non-corporate ties, one compares this network density to the density of the network of all ties except those from the organization under investigation. If the organization of interest adds unique ties to the configuration of ties from the five other sources, there should be a lower density value. The lower the density for a second network, the greater the unique contribution from the organization in question. The results from this set of comparisons are in Table 3.

Table 3.

Changes in network density with the exclusion of non-corporate ties, 1962–1995.

	Source	Ties		Unique Ties From Excluded Org	Network Density	Percent Change in Density
		Possible	Unique	Unique Ties From Excluded Org	Network Density	Percent Change in Density
1962	Corporation and All Other Organizations	4,059,825	84,014		0.0207
	All Ties Excluding Non-Profit Foundations		83,857	157	0.0207	0.00
	All Ties Excluding Cultural Organizations		83,431	583	0.0206	−0.48
	All Ties Excluding Universities		82,630	1,384	0.0204	−1.45
	All Ties Excluding Policy Planning Organizations		83,166	848	0.0205	−0.97
	All Ties Excluding Social Clubs		34,005	50,009	0.0084	−59.42
1973	Corporation and All Other Organizations	3,790,881	66,888		0.0176
	All Ties Excluding Non-Profit Foundations		66,741	147	0.0176	0.00
	All Ties Excluding Cultural Organizations		66,094	794	0.0174	−1.14
	All Ties Excluding Universities		65,434	1,454	0.0173	−1.70
	All Ties Excluding Policy Planning Organizations		65,465	1,423	0.0173	−1.70
	All Ties Excluding Social Clubs		32,998	33,890	0.0087	−50.57
1983	Corporation and All Other Organizations	3,730,546	50,146		0.0134
	All Ties Excluding Non-Profit Foundations		50,002	144	0.0134	0.00
	All Ties Excluding Cultural Organizations		49,013	1,133	0.0131	−2.24
	All Ties Excluding Universities		48,593	1,553	0.013	−2.99
	All Ties Excluding Policy Planning Organizations		47,628	2,518	0.0128	−4.48
	All Ties Excluding Social Clubs		35,170	14,976	0.0094	−29.85
1995	Corporation and All Other Organizations	2,598,060	26,644		0.0103
	All Ties Excluding Non-Profit Foundations		26,581	63	0.0102	0.98
	All Ties Excluding Cultural Organizations		26,174	470	0.0101	−1.94
	All Ties Excluding Universities		25,681	963	0.0099	−3.88
	All Ties Excluding Policy Planning Organizations		23,496	3,148	0.0090	−12.62
	All Ties Excluding Social Clubs		24,059	2,585	0.0093	−9.71

In 1962, while the exclusion of foundation ties has no effect on the overall network density, the exclusion of ties formed by shared social club memberships decreases overall network density by nearly 60 percent. Less extreme are the reductions observed when the ties from cultural organizations, university boards and policy planning organizations are removed. These results indicate that there exists a fair degree of structural redundancy and that the ties from each of these organizations appear to reinforce each other. The story for 1973 is essentially the same as that for 1962; however, by 1983 the narrative changes. First, although the exclusion of social club ties decreases network density by nearly 30 percent, this effect is much smaller than those of the previous two time periods. Secondly, the exclusion of ties from cultural organizations and university boards has larger relative effects than in 1962 and 1973. Finally, the exclusion of ties from policy planning organizations, even after we include the ties from corporate board memberships and all other non-corporate organizations, decreases by nearly 4.5 percent – that is four times the effect in 1962 and more than double its effect in 1973.

Multiplicity of Network Ties

The above discussion has focused on structural redundancy or the overlap between two different sources of social ties. As noted, there appears to be a fair amount of structural redundancy, especially in the first two time periods for the ties involving the shared affiliations from elite club memberships in 1962 and 1973. Given the occurrence of these multiple ties, the next set of analyses describes the different configurations, or multiplicity, of ties when the six sources of affiliations are examined simultaneously. Unlike the analysis of structural redundancy, where one intentionally ignores multiple ties, the multiplicity of ties recognizes that the number of ties between a given pair of directors matters and that the configuration of those ties is meaningful. By extending the multiplicity of ties beyond the number of ties, this paper identifies those directors possessing only corporate ties, those connected by both corporate and one other non-corporate tie, and those directors that are linked through corporate and two or more other organizational ties. By capturing the configuration of ties, one is able to distinguish if a given pair of directors met only in boardrooms, or whether there were opportunities to meet at a monthly regents’ meeting or on the golf course during the weekend. The clear pay-off of this method is that it provides a researcher with the capacity to understand the social context in which ties are formed and to assess which configurations of ties are most important.

Table 4 displays the descriptive statistics on the multiplicity of ties among the corporate directors from all sources of social ties. The first panel reports the strength of all of the ties among corporate directors where the distinct trend is an increase in the percent of ties with a strength of one, and a corresponding decrease in the percent of ties with a strength of two, and three or more. In 1962, nearly 90 percent of the 84,014 ties had a strength of one whereas, in 1995, 96 percent of the 26,644 ties were constituted by a single tie. So, not only does the density of ties decrease between 1983 and 1995, it is also the case that the ties that remain are, relatively speaking, weaker.

Table 4.

Multiplicity of ties among corporate directors, 1962–1995.

	Year
	1962		1973		1983		1995
	N	%	N	%	N	%	N	%
Total Ties	84,014		66,888		50,159		26,644
Strength of Ties
Single	75,537	89.91	59,820	89.43	46,884	93.47	25,604	96.10
Two	7,205	8.58	5,813	8.69	2,809	5.60	885	3.32
Three or more	1,272	1.51	1,255	1.88	466	0.93	155	0.58
Multiplicity of Types of Ties
Single	80,479	95.79	64,090	95.82	48,582	96.86	26,064	97.82
2 Types	3,376	4.02	2,657	3.97	1,485	2.96	538	2.02
3 or More Types	159	0.19	141	0.21	92	0.18	42	0.16
Configuration of Ties
Corporate Only	27,498	32.73	26,382	39.44	28,258	56.34	18,835	70.69
Corporate and Other Types of Ties	2,820	3.36	2,009	3.00	1,131	2.25	353	1.32
Non-Corporate Ties Only	53,696	63.91	38,497	57.55	20,770	41.41	7,456	27.98

The second panel reports the distribution of the number of different types of ties. Here, the analyses are not interested in the number of ties between directors, but rather on the type of tie. For example, imagine three different pairs of directors. The first pair had three different corporate ties; the second possessed two different corporate ties and one tie from a cultural organization; and the third involved three distinct types of ties – corporate, university and social club. All three pairs would have tie strengths of three, but the configurations of these multiple ties are quite different. In 1962, 95.8 percent of the ties are constituted by only one type of tie. In 1973, the percentage remains the same with 95.8 percent of the 66,888 ties being of only one type of affiliation. In the next time period, 96.8 percent of the 50,159 ties stemmed from one source of social affiliation. And by 1995, this percentage increases to 97.8 percent. On the other hand, the percentages of ties that are made up of multiple types of ties decrease accordingly. Therefore, not only are the ties ‘thinning’, but it also appears to be the case that they are becoming slightly more homogeneous.

The last panel of Table 4 describes three basic configurations of ties: ties generated by corporate ties alone, non-corporate ties alone, or from both corporate and non-corporate affiliations. This is perhaps the most interesting set of results in Table 4 for it clearly shows that the social mosaic of ties is becoming increasingly corporate in character. Conversely, the percentages of ties that involve either corporate and one additional type of non-corporate tie, or the percent of ties that are strictly non-corporate, both decease. In 1962, nearly 33 percent of the ties were formed by corporate affiliations only. In 1973, this percentage increases slightly to 39.4 percent. However, by 1983, the percent increases to 56.3 percent and by 1995, over 70 percent of the ties are from corporate affiliations only. The percent of corporate ties with an additional type of tie is quite small, ranging from 3.3 percent in 1962 to 1.3 percent in 1995. In contrast to the trend in corporate ties alone, the percent of ties from non-corporate affiliations alone decreases over the four time periods of the study. In 1962 and 1973, over half of the ties were from serving on non-corporate boards or (primarily) belonging to the same social clubs (63.9 percent and 57.6 percent respectively). In 1983, however, the percentage of non-corporate ties only falls to 41.4 percent and by 1995, 28 percent of the ties are from social affiliations other than shared corporate boards. While this decrease is substantial, it is important to note that even in the last time period, nearly one-third of the ties in the social network are from non-corporate sources. This means that such organizations as non-profit foundations, cultural organizations, universities, policy planning organizations and shared social club membership remain an important (albeit diminishing) feature of the overall social mosaic.

Table 4 shows that other types of ties play an important role in ‘filling-in’ the network of corporate ties. This overview, however, does not indicate which types of non-corporate ties are most important. Table 5 elaborates upon the last line in Table 4 by analyzing the distribution of the different types of non-corporate ties that comprise the social mosaic of all ties. In 1962, the incidence of non-corporate ties that do not involve any corporate affiliation is quite high, over 50,000. However, just like the configuration of corporate and non-corporate ties, the vast majority of these social affiliations were comprised of only one type of non-corporate tie. In fact, between 97.0 percent and 98.7 percent of the ties were of a single type in all four time periods. Nonetheless, the results in Table 5 show that in the earlier two time periods, the vast majority of non-corporate ties were formed by shared social club memberships. In 1962, over 94 percent of these non-corporate ties were constituted through club affiliations while in 1973 the percent decreased only marginally to approximately 90 percent. By 1983, nearly three-quarters of the non-corporate ties were from shared memberships in private clubs (73.7 percent). However, by 1995, only a little over a third (35.8 percent) of the non-corporate ties were accounted for by social club memberships.

Table 5.

Distribution of non-corporate ties, 1962–1995.

	Year
	1962		1973		1983		1995
	N	%	N	%	N	%	N	%
Total Non-Corporate Ties	53,696		38,497		20,770		7,456
Two or More Non-Corporate Ties	715	1.33	789	2.05	446	2.15	227	3.04
One Type of Non-Corporate Tie	52,981	98.67	37,708	97.95	20,324	97.85	7,229	96.96
Foundation Only	157	0.30	147	0.39	144	0.71	63	0.87
Cultural Only	583	1.10	794	2.11	1,133	5.57	470	6.50
University Only	1,384	2.61	1,454	3.86	1,553	7.64	963	13.32
Policy Only	848	1.60	1,423	3.77	2,518	12.39	3,148	43.55
Club Only	50,009	94.39	33,890	89.87	14,976	73.69	2,585	35.76

In contrast to the relative decline in the occurrence of social club ties, the other sources of non-corporate links – with the exception of non-profit foundations – increased between 1962 and 1995. Ties from cultural organizations went from 1.1 percent in 1962 to 6.5 percent in 1995, while the proportion of ties from serving on the same research university boards increased from 2.6 percent in 1962 to 13.3 percent in 1995. The most dramatic increase, in both relative terms but also in numbers, is the percent of ties from policy planning organizations. In 1962, only 1.6 percent of the 52,981 ties were formed through policy planning organizations. In 1973, the percent increased to 3.8 percent of the 37,708 non-corporate ties. By 1983, there were over 2500 ties formed through policy planning boards out of a total of 20,324 for a percentage of 12.4. Finally, by the last time period, social ties from policy planning boards constituted 43.6 percent of the 7229 non-corporate ties. In short, the type of non-corporate tie changed from being one cultivated in private clubs to one formed in policy planning organizations: a shift, so to speak, from relations forged in ‘smoke filled back rooms’ to groups of politically engaged agents promoting specific public policy agendas.

The Effects of Multiplicity on Director Centrality and Social Distance

Up to now, the discussion of the different trends in the types, strengths, and configuration of ties has operated at the network level – be it the effects of these different types of ties on overall density, or the distribution of the specific types of ties within the network. This last section considers how these changing configurations of ties affect the characteristics of the directors themselves. Given the existence of unique ties from non-profit foundations, cultural organizations, university boards, policy planning organizations, and private social clubs, do these ties translate into significantly higher levels of centrality or decreased social distance (geodesic distance) for the directors in the network? Conversely, does the removal of a specific type of tie from the basket of all ties significantly reduce the degree centrality or increase the social distance between the directors?

Director degree is the number of different corporate directors tied to a given corporate board member. This value is different from the number or configuration of the ties within the networkexplored thoroughly above. While there are 84,014 ties in the 1962 network of directors, the analysis of the multiplicity of ties does not tell us how these unique ties – however configured – translate into this common measure of a director’s importance in the network. Table 6 reports the average degree centralities for the directors in each of the four time periods – first with only corporate ties and then with the addition of each non-corporate tie, and secondly with all ties and then the exclusion of the non-corporate tie of interest. The t-values for the difference between means is reported so that a negative t-value means that Mean₂ > Mean₁. Consequently in the top half of Table 6 we would anticipate negative t-values with the addition of ties, whereas in the bottom half of Table 6 we would expect positive t-values since we are subtracting ties from the network. In 1962, the 2850 directors were tied to an average of 21.28 other directors. With the inclusion of the ties from non-profit foundations, the average degree centrality increased to 21.42. Although this relatively small increase yields a statistically significant t for the difference between means, the magnitude of the t-value (-7.31) is the smallest among the five comparisons. Therefore, while the addition of foundation ties increases average director degree, this is the weakest effect. In 1962, the addition of policy planning ties also makes a relatively small increase in average director degree. In contrast, the addition of social club ties greatly increases average director degree from 21.28 to 56.86!

Table 6.

Average degree centrality of directors, 1962–1995.

	Year
	1962, N=2850		1973, N=2754		1983, N=2732		1995, N=2280
	D	t	D	t	D	t	D	t
Corporation Ties Alone	21.28		20.62		21.51		16.83
Corporation Ties
& Non-Profit Foundations	21.42	−7.31	20.74	−6.39	21.62	−5.86	16.89	−5.45
& Cultural Organizations	21.92	−11.99	21.43	−14.23	22.46	−13.31	17.34	−10.49
& Universities	22.44	−13.14	21.85	−13.96	22.76	−13.91	17.79	−13.08
& Policy Planning Organizations	21.94	−9.26	21.84	−11.03	23.48	−11.75	19.65	−10.50
& Social Clubs	56.86	−26.99	45.76	−22.49	32.75	−18.25	19.24	−12.44
Corporation and All Organizations	58.96		48.58		36.71		23.37
All Ties Excluding
Non-Profit Foundations	58.85	6.86	48.47	6.01	36.60	5.75	23.32	5.28
Cultural Organizations	58.55	11.00	48.00	13.42	35.88	13.06	22.96	10.94
Universities	57.99	12.66	47.52	13.57	35.57	13.77	22.53	13.27
Policy Planning Organizations	58.36	9.17	47.54	11.03	34.87	11.73	20.61	10.52
Social Clubs	23.86	26.98	23.96	22.47	25.75	18.27	21.10	12.45

Note: All paired samples t-tests for the difference between means among directors are significant at the 0.001 level.

Table 6 reports similar results for 1973 and 1983 with foundations marginally adding to average director degree, and social club memberships greatly increasing this measure of centrality. The other three sources of non-corporate ties demonstrate greater increases in degree compared to 1962 (given the magnitude of their t-values). This suggests that these social ties are contributing relatively more to the average degree of directors than in the initial time period. By 1995, the ties from social clubs no longer have the largest effect on average degree centrality. While shared social club membership increased average degree to 19.24 from 16.83, the ties from policy planning organizations increased the average degree to 19.65. As with the earlier time periods, ties from non-profit foundations had little effect on average degree and the ties from cultural organizations and research universities had a moderate effect in increasing average degree.

The bottom half of Table 6 begins by reporting the average degree of the corporate directors based upon all affiliations. In 1962, the average degree from all sources of social ties was 58.96.Given the thinning of the networks noted above, it is not surprising that the average degree decreases between 1962 and 1995. In 1973, the average degree falls to 48.58, and by 1995 the average degree for corporate directors from all sources of social ties was less than half of the level of 1962. The bottom portion of Table 6 also reports the average degree of the corporate directors only when the social tie of interest is excluded. In 1962, 1973 and 1983, it appears that the exclusion of non-profit foundations, cultural organizations or research universities does not substantially reduce the average degree when each of these sources of social ties is excluded. In all of these three earlier time periods, the removal of the ties from social clubs is the source that makes the greatest difference. However, by 1995 this pattern no longer holds as the exclusion of the ties from policy planning organizations has the greatest reduction of average degree centrality – from 23.37 to 20.61. Consistent with earlier findings in this paper, by 1995 corporate directors are tied to each other more through policy planning organizations than by private social clubs.

The last analysis considers the effects of the various sources of social ties on a standard measure of cohesion – geodesic distance. This measure captures how far away a given director is to all other directors in the network. This measure of social distance is closely aligned with the ‘small world’ phenomena popularized by Milgram (1967), who determined that it only took, on average, six steps to get a letter to an unknown addressee through existing acquaintances. Applied to a network of ties among corporate directors, a director with an average geodesic distance of three would be that many handshakes away from every other director – or everyone would be no further away than a friend of a friend of a friend. In the analysis of geodesic distance, we would expect positive differences since the inclusion of a given set of ties should decrease the social distance between directors Mean₂ < Mean₁. Conversely, in the bottom half of Table 7 where we begin with the average geodesic distances from all sources, we would expect that the exclusion of any given set of ties would increase the social distance producing negative differences between means. The first row of Table 7 shows that the average geodesic distance for corporate directors based on corporate ties alone in 1962 was 3.68, and this level of social distance remains stable for the subsequent three time periods. The inclusions of ties from non-profit foundations in 1962 had a very small effect on average geodesic distance, reducing the average number of steps from 3.68 to 3.65. Relatively larger reductions are seen for ties from universities and policy planning organizations. But the greatest difference is seen with the inclusion of ties from social clubs. When the ties formed though shared membership in social clubs are added, the average geodesic distance for the corporate directors declines to 2.85 – a reduction of almost a full step! Similar results are seen for 1973 and 1983 with social club ties having the greatest effect on increasing social cohesion among the corporate directors. However, by 1995 the differences between the various sources of social ties become less pronounced. Furthermore, it is very important to note that social club ties remain the most important set of non-corporate ties for reducing social distance. While other social ties decrease average geodesic distance by 0.02 to 0.23 steps, social clubs reduce average social distance by over half a step, from 3.69 to 3.13.

Table 7.

Average geodesic distance of directors, 1962–1995.

	Year
	1962, N=2850		1973, N=2754		1983, N=2732		1995, N=2280
	G	t	G	t	G	t	G	t
Corporation Ties Alone	3.68		3.70		3.69		3.69
Corporation Ties
& Non-Profit Foundations	3.65	20.92	3.68	22.60	3.67	17.63	3.67	19.20
& Cultural Organizations	3.61	41.08	3.60	44.86	3.58	40.68	3.58	37.40
& Universities	3.51	51.31	3.54	51.01	3.50	42.49	3.50	60.42
& Policy Planning Organizations	3.55	48.10	3.54	54.06	3.46	49.69	3.46	61.51
& Social Clubs	2.85	123.56	2.93	109.71	3.13	79.46	3.13	59.72
Corporation and All Organizations	2.80		2.87		2.99		2.99
All Ties Excluding
Non-Profit Foundations	2.81	−7.69	2.87	−7.96	3.00	−13.26	3.00	−10.64
Cultural Organizations	2.81	−15.22	2.88	−16.80	3.01	−23.79	3.01	−25.66
Universities	2.83	−20.06	2.90	−22.99	3.03	−30.43	3.03	−34.28
Policy Planning Organizations	2.82	−20.53	2.89	−21.85	3.04	−34.02	3.04	−39.70
Social Clubs	3.38	−98.14	3.37	−89.35	3.29	−63.64	3.29	−44.72

Note: All paired samples t-tests for the difference between means among directors are significant at the 0.001 level.

The bottom half of Table 7 examines the effects on average geodesic distance when a specific set of ties is removed from the basket of all social ties. For all time periods, the average geodesic distance among the corporate directors is between 2.8 and 3.0 when you consider all corporate and non-corporate ties. In the first time period, the removal of social club ties is the only type of tie that substantially increases average geodesic distance. This general pattern holds for all subsequent time periods – even in 1995 where the removal of social club ties has the most sizeable increase in average geodesic distance, even greater than policy planning organizations. Therefore, unlike the narrative that emerged from Table 6, social clubs remain the most important facet of the social mosaic in terms of reducing the social distance among corporate directors. This is in contrast to the finding regarding degree centrality in which, by 1995, policy planning ties emerged as the largest contributor to director centrality and social club affiliations had a secondary effect.

Conclusion

The first contribution of this paper is that the analyses above represent an extensive application of the concepts of structural redundancy and multiplicity to document the changing social mosaic of social relations among US corporate directors during the last half of the 20th century. This method preserved important information about the context and configuration of ties while still allowing for ease of analysis with standard algorithms such as measures of density, degree centrality and geodesic distance. As presented above, these analyses permit greater insight into the context in which corporate and non-corporate ties coexist and provide researchers with the ability to evaluate the relative importance of the various social ties. Consequently, this method and analyses represent one step towards a richer sociology of corporate directors. Given the time frame of this study, a clear direction for future research would be to update these affiliation data from the six different sources of ties to include more recent time periods. Also, future research should consider the emergence of a transnational/transatlantic capitalist class and the role of global policy planning organizations such as the International Chamber of Commerce and the European Round Table (Carroll, 2010; Carroll et al., 2010).

There are two substantive findings that emerge from this analysis of structural redundancy and multiplicity of social ties among US corporate directors. First, there is a significant shift in the relative number of social ties and their configurations between 1962 and 1995. Given the relative decline in the frequency of social club ties and the emergence of policy planning ties by 1995, it is nonetheless true that social clubs persist in being the most important set of ties in reducing the social distance among the corporate directors – though social club ties did not remain the most important source of social ties contributing to the directors’ average degree centrality in 1995. This highlights the second finding that that there may be a division of labor among ties, with the more public affiliations formed through policy planning organizations contributing to director centrality, while the ties formed through exclusive social clubs continue play a more prominent role maintaining cohesion. Theoretically, one can fruitfully conceptualize these empirical differences in degree centrality and average geodesic distance as the classic distinction between a class in itself versus class for itself. As Domhoff notes, social cohesion among members of the upper class is promoted by small group interaction through their membership in ‘social clubs, retreats, resorts, and social gatherings’ (2010: 51). Such venues reduce social distance and promote a class in itself among these directors. In contrast, the corporate directors in this study simultaneously emerge as a class for itself through their increasing participation in policy planning organizations. The increase in average degree centrality is one indication of the directors acting as a class for itself as it pursued the right wing political agenda noted above.

While this study clearly demonstrates that ties from non-corporate affiliations matter, it is also the case that their roles change over time. Initially social club memberships were important for the network as a whole – density and distribution of different configurations of ties – but this changes in 1995 with policy planning organizations becoming most important. This shift from interacting within the secluded and private setting afforded by elite social clubs to the participation in public policy planning organizations may reflect a shift in the modus operandi of how the corporate directors exerted political influence in the last four decades of the 20th century. Burris (2008: 10) has characterized the middle two time frames of this study as a period of intense political mobilization by the right, while the last cross-section of the study, 1995, is noted for a more ‘routinized’ and ‘less strident’ force in policy making, i.e. the consistent support of free trade policies by the National Association of Manufacturers and the Chamber of Commerce.

The changes in the level and style of political mobilization among the corporate directors are also partially reflected in the changes documented in the analyses of the average degree and social distance reported for the individual directors in this study. Consistent with the arch of conservative political mobilization documented by Burris (2008), policy planning ties have the greatest effect on degree by 1995. However, it is also the case that social clubs continue to be the most important source of social ties for reducing social distance. In short, while the contours of the social mosaic of corporate directors may change over time, these analyses have shown that all of the facets within the social mosaic have contributed to increasing centrality and social cohesion among US corporate directors during the last half of the 20th century.

Footnotes

Acknowledgements

I wish to sincerely thank members of the Politics and Interlocking Directorates research community for their numerous comments and support over the course of this research project – in particular, Eelke Hemskerk, Bill Carroll, Val Burris, and Bill Domhoff. I would also like to thank the anonymous reviewers for their very constructive criticisms.

Funding

Support for this research has been provided by several Research Initiative Partnership grants through the Office of Research, University of Michigan-Flint.

Notes

References

Arnove

Pinede

(2007) Revisiting the ‘Big Three’ foundations. Critical Sociology 33(3): 389–425.

Baltzell

(1958) Philadelphia Gentlemen: The Making of a National Upper Class. Glencoe, IL: Free Press.

Barnes

(2015) Social Mosaic Project. Available (consulted 2 March 2015) at: www.umflint.edu/cas/social-mosaic-project.

Barnes

Burkett

(2010) Structural redundancy and multiplicity in corporate networks. Connections 30(2): 4–20.

Burris

(1992) Elite policy-planning networks in the United States. Research in Politics and Society 4: 111–134.

Burris

(2006) Interlocking Directorates and Political Mobilization among Corporations and Corporate Elites: Evidence from the U.S. Paper presented at the Politics and Interlocking Directorates Conference, University of Barcelona, 28–29 September.

Burris

(2008) The interlock structure of the policy-planning network and the right turn in US state policy. Politics and Public Policy 17: 3–42.

Cárdenas

(2014) Are Latin America’s corporate elites transnationally interconnected? A network analysis of interlocking directorates. Global Networks. DOI: 10.1111/glob.1270.

Carroll

(2010) The Making of a Transnational Capitalist Class. London: Zed.

10.

Carroll

Fennema

Heemskerk

(2010) Constituting corporate Europe: A study of elite social organization. Antipode 42(4): 811–843.

11.

Davis

Yoo

Baker

(2003) The small world of the American corporate elite, 1982–2001. Strategic Organization 1(3): 301–26.

12.

Domhoff

(1967) Who Rules America? Englewood Cliffs, NJ: Prentice-Hall.

13.

Domhoff

(1970) Higher Circles. New York, NY: Random House.

14.

Domhoff

(1974) The Bohemian Grove and Other Retreats: A Study in Ruling Class Cohesiveness. New York, NY: Harper and Row.

15.

Domhoff

(1983) Who Rules America Now? Englewood Cliffs, NJ: Prentice-Hall.

16.

Domhoff

(1998) Who Rules America?: Power and Politics in the Year 2000. Mountain View, CA: Mayfield Publishing Company.

17.

Domhoff

(2010) Who Rules America?: Challenges to Corporate and Class Dominance. Englewood Cliffs, NJ: Prentice-Hall.

18.

Domhoff

(2014) The Council on Foreign Relations and the Grand Area: Case Studies on the Origins of the IMF and the Vietnam War. Class, Race and Corporate Power 2. Available (consulted 6 February 2015) at: http://digitalcommons.fiu.edu/classracecorporatepower/vol2/iss1/1.

19.

Dye

(2001) Top Down Policy Making. New York, NY: Chatham House Publishers.

20.

Ferguson

Rogers

(1986) The Right Turn: The Decline of the Democrats and the Future of American Politics. New York, NY: Wang and Hill.

21.

Galaskiewicz

(1985) Social Organization of an Urban Grants Economy: A Study of Business Philanthropy and Nonprofit Organizations. Orlando, FL: Academic Press.

22.

Goldschmidt

Finkelstein

(2001) Academics on board: University presidents as corporate directors. Academe 87(5): 33–7.

23.

Hall

(1987) A historical overview of the private nonprofit sector. In: Powell

(ed.) The Nonprofit Sector: A Research Handbook. New Haven, CT: Yale University Press, 3–26.

24.

Hartnett

(1974) College and university trustees: Their backgrounds, roles, and educational attitudes. In: Useem

Useem

(eds) The Educational Establishment. Englewood Cliffs, NJ: Prentice Hall, 19–40.

25.

Heemskerk

(2013) The rise of the European corporate elite: Evidence from the network of interlocking directorates in 2005 and 2010. Economy and Society 42(1): 74–101.

26.

IJCS (2012) Special Issue: Interlocking corporate directors in comparative perspective, ed. Barnes R. International Journal of Comparative Sociology 53(4): 247–342.

27.

Kono

Palmer

Friedland

(1998) Lost in space: The geography of corporate interlocking. American Journal of Sociology 103(4): 863–911.

28.

McGrath

(1936) The control of higher education in America. Educational Record 17(2): 259–272.

29.

Milgram

(1967) The small world problem. Psychology Today 2(1): 60–7.

30.

Miliband

(1969) Higher Education and Capitalist Society. New York, NY: Basic Books.

31.

Mills

(1956) The Power Elite. New York, NY: Oxford University Press.

32.

Mizruchi

(1982) The American Corporate Network 1904–1974. Beverly Hills, CA: SAGE.

33.

Moore

(1979) The structure of a national elite network. American Sociological Review 44(5): 673–692.

34.

Peshick

(1987) Policy-Planning Organizations. Philadelphia, PA: Temple University Press.

35.

Salzman

Domhoff

(1983) Nonprofit organizations and the corporate community. Social Science History 7(2): 205–216.

36.

Scott

Griff

(1984) Directors of Industry: The British Corporate Network 1904–1976. Cambridge: Polity Press.

37.

Sheehan

(1958) The executive lunch. Fortune 57(1): 120.

38.

Soref

(1976) Social class and a division of labor within the corporate elite: A note on class, interlocking, and executive committee membership of directors of U.S. industrial firms. The Sociological Quarterly 17(3): 360–368.

39.

Useem

(1979) The social organization of the American business elite. American Sociological Review 44(4): 553–572.

40.

Useem

(1980) Which business leaders help govern? The Insurgent Sociologist 9(2–3): 107–120.

41.

Useem

(1984) The Inner Circle. New York, NY: Oxford University Press.

42.

Veblen

(1918) Higher Learning in America: A Memorandum on the Conduct of Universities by Business Men. New York, NY: B.W. Huebsch.

43.

Wolff

(1993) The Social Production of Art, 2nd edn. New York, NY: New York University Press.

44.

(2002) Privatising Culture: Corporate Art Intervention since the 1980s. London: Verso.