The social organization of the Canadian capitalist class in comparative perspective

The social organization of the Canadian capitalist class in comparative perspective

MICHAEL ORNSTEIN York University*

J’analyse, dans cet article, la structure de la classe capitaliste canadienne. Pour ce faire, je compare le reseau de liens entre les conseils d’administration des 256 corporations canadiennes les plus grandes, a ceux existant aux Etats-Unis et dans neuf pays d’Europe. J’arrive a la conclusion que le degre de fragmentation du reseau canadien n’est pas inhabituel, et que les clivages entre capital industriel et financier, ou entre capital domestique et etranger ne sont pas particulierement prononces. Qui plus est, le capital industriel n’est ni subordonne ni peripherique au capital financier. Le reseau canadien ressemble beaucoup a ceux de plusieurs pays europeens, dont 1’Allemagne et la France, a propos desquels on ne peut pas avancer d’explication en termes de dependance ou de sous-developpement. La fragmentation des reseaux americain et britannique, par contre, est netternent plus Qlevee.

This paper assesses claims about the character of Canadian capitalist class by comparing the network of interlocking directorates among the largest 256 Canadian corporations to the corresponding networks in the U.S. and nine European nations. The analysis indicates that the Canadian inter-corporate network is not unusually fragmented, that there are no pronounced cleavages between industrial and financial capital or between foreign and domestic capital, and that industrial capital is neither subordinated nor peripheral to finance. The Canadian network is quite similar to the networks of a number of European nations, such as Germany and France, about which it is impossible to advance arguments about dependency and underdevelopment. In comparison, the networks of the u.S. and Britain are unusually fragmented.

Although they do not rule, there is now abundant evidence that capitalist classes are conscious of their class interest and organized to pursue it (Useem, 1984; Langille, 1987) . The organization of capital takes four related forms: intercorporate ownership provides direct, hierarchical links between

* I thank Beth Mintz, John Myles, Frans Stokman and especially Robert Brym for helpful comments on earlier drafts of this paper.

Canad. Rev. SOC. & Anth. / Rev. canad. SOC. & Anth. 26(1) 1989

152 MICHAEL ORNSTEIN

some corporations; interlocking directorates link the boards of the largest corporations; membership on the boards of universities, hospitals, cultural and charitable organizations, political parties, governmental advisory com- mittees and social clubg provides another meeting place; and many organizations explicitly organize and represent corporate views.' These different forms of organization serve different, although complementary needs. In- tercorporate ownership co-ordinates the activities of some groups of corporations but cannot serve to formulate and pursue a broader political strategy. The interlocking directorates that link most major corporations and business participation in cultural, charitable and social organizations provide an arena for building a political consensus, but they are not as effective at pursuing specific policy objectives as more clearly policy-oriented business organizations.

There is an urgent need for comparative studies of the social organization of capital. The known, broad and theoretically significant national variation in the structure and organization of the working class, the state and political life should correspond to important differences in the organization of capital. Because, in most contexts, capitalist classes are much less visible than the labor movement, the economy and the state, comparative studies of capital are difficult to undertake. Without minimizing the difficulties of reconciling statistics from different nations, the large and impressive body of research on the contemporary welfare state is testimony to what can be accomplished using official statistics. By comparison there is little systematic, comparative research on the social organization of capital. National ac- counts data and other economic indicators certainly provide aggregate measures of the activity of capital - but there are no grounds for assuming these are a simple reflection of the organization of capital. Comparative data on interlocking directorates, on business organizations and on the repre- sentation of business in education, cultural, governmental and other organizations cannot be looked up; they must be gathered directly.

This paper builds on one such effort to gather and analyse comparative data on the social organization of capital, the study of interlocking directorates of the 250 largest corporations in Austria, Belgium, Britain, Finland, France, Germany, Italy, the Netherlands, Switzerland and the US., which is reported by Stokman, Ziegler, and Scott (1985). Data from these 10 nations provide a context for an analysis of interlocking directorates in Canada. While there are important limitations to confining this study of Canadian capital to interlocking directorates, an analysis of interlocks is certainly critical to a broader empirical study conceptualizing national variations in the organization of capital.

Three different bodies of theory and research, which are discussed in turn below, provide insight into the task at hand. First, the recent discussion of the development of the welfare state raises some very general questions about the organization of capital. Second, the organization of the capitalist class plays an important role in the tradition of Canadian political economy. And, third, there is an important tradition of research on the control of capital.

153 SOCIAL ORGANIZATION OF CANADIAN CAPITALIST CLASS

THE CAPITALIST CLASS AND THE SOCIAL DEMOCRATIC MODEL

The richest tradition of comparative research on contemporary advanced capitalism involves the elaboration of what is variously referred to as the ‘class conflict paradigm’ (Shalev, 1983: 28) or ‘social democratic model’ (Skocpol and Amenta, 1986: 140ff.) of the welfare state. The argument is that differences in the development of contemporary welfare states reflect the relative strengths of the working classes of different nations - as indi- cated by the proportion of workers belonging to trade unions, the electoral strength of social democratic parties and their ability to form governments, and so on.2 The relative power and political organization of capital are assumed to be inversely proportional to the power of labor and the left political parties, since social reforms represent strategic victories of labor over capital. Measures of the power and organization of capital are absent from the social democratic model (see Shalev, 1983: 29) on the assumption either that they are very strongly correlated with other variables in the model, or that beyond these correlations, the organization of capital has no influence on the welfare state. As Shalev (1983: 82) observes, one justification for re- stricting the social democratic model to a focus on the power of labor involves the policy focus of these studies, ‘the policy package associated with the welfare state conventionally addresses only problems connected to em- ployment and distribution, and not the twin loci of control to which capital historically assigned priority, namely accumulation and labor process’.

The social democratic model implies that nations with weakly developed welfare states, such as Canada and especially the u.S., should have stronger and better organized capitalist classes. One obvious difficulty with this argument, however, is immediately apparent from a consideration of the balance of class power in Sweden, the exemplar of the welfare state. Al- though the development of the Swedish welfare state is unquestionably marked by strategic victories of labor over capital (see Korpi, 1978), there is also a remarkable concentration of power within the Swedish capitalist class (see Sundqvist, 1987). Furthermore, the organization of Swedish capital in the Employers Federation is both more unified and plays a much more important role than the comparable Canadian organizations. The point here, made clearly in theoretical arguments about corporatism by Panitch (19861, Schmitter (see the essays in Schmitter and Lehmbruch, 1979) and others, is that the system-wide bargaining in the welfare state requires the organization of labor and capital. Without the centralized organization of capital, extensive consultation between business and the state, let alone tripartism, is impossible. Explaining Swedish tripartism, Panitch notes that ‘the sine qua non of the whole edifice was the capacity of both the SAF and the LO [respectively the Swedish Employers Federation and the central trade union organization] to commit their affiliates to central agreements’ (1986: 57). Analysing data from the 10 nation comparative study, Ziegler provides the link between the network of interlocks and corporatism: ‘those four countries having the least centralized networks of interlocking direc-


torships (France, Great Britain, Italy and the USA) also show weak forms of corporatism’ (1985: 280).

Suggestive of the weakness of Canadian labor and corresponding disorganization of capital is the slow development of the welfare state in Canada, relative to the typical Western European experience (although, as Kudrle and Marmor (1981: 83ff.) show, after the adoption of medical insurance and maternity benefits in the early 1970s, social insurance coverage in Canada is not far from the mean for Europe). The results of Coleman’s (1986) study of business associations in Canada are consistent with this argument:

There is little doubt that the system of business associations is underdeveloped. It is characterized by congeries of isolated groups: intersectoral associations operate independently of divisional associations, divisional associations of sectoral associations ... and sectoral associations of subsectoral associations. Regional associations are paid little attention and the representativeness of associations with more general domains is suspect. (Coleman, 1986: 272)

This view of the relative disorganization of Canadian capital has not, however, gone unchallenged, especially as it relates to the Business Coun- cil on National Issues (the BCNI), whose representativeness and effectiveness is repeatedly questioned by Coleman. In a recent study Langille (1987: 70) argues that: ‘In the course of the last decade, the Business Council on National Issues has become the most powerful and effective interest group in Canada - to the point where it can now exercise hegemony over both the private sector and the state.’ Langille cites compelling evidence of the BCNI’S effectiveness in promoting free trade (also see Cameron, 19881, in designing an exit from the National Energy Programme and in shaping competition policy. In view of its supposed hegemony, however, it is surprising (as Lan- gille acknowledges) that the BCNI has been quite ineffective in achieving major reductions in social programs, government spending and the deficit.

Despite the qualifications offered by Langille, on the whole this litera- ture suggests that Canadian capital is relatively disorganized, perhaps because the weakness of labor has allowed capital to influence policy without strong organization. As the next section shows, this view is consistent with the now dominant view in Canadian political economy.

THE CANADIAN CAPITALIST CLASS: ‘UNDERDEVELOPMENT’ AND FEDERALISM

In terms of its economic strength - defined as the ability to sustain capital accumulation and defend itself from foreign and state intervention - the Canadian capitalist class is widely regarded as exceptionally weak and fragmented. The uniquely high degree of foreign ownership and correlated ‘overdevelopment’ of extractive industry and finance, relative to manufacturing, is often said to be more characteristic of underdeveloped than developed capitalist nations3 In structural terms, this proposition is self- evident. In no other advanced capitalist nation is nearly half of all manufacturing in the hands of foreign owners. The disproportionate development of extractive industries is less unusual, however, especially if Canada is com-


pared to other nations with extensive natural resources, such as Australia and the U.S. More controversial is the related argument that the structure of the Canadian capitalist class is distorted.

The strongest claim about the uniqueness of the Canadian capitalist class is advanced by Naylor (1972), who argues that Canadian capital is dominated by a ‘mercantile’ element of corporations in the financial and transportation sectors.* Clement (1975) argues that the Canadian capitalist class is divided into ‘indigenous’ and ‘comprador’ elites on the basis of the distinction between Canadian and foreign ownership. Although foreign control is concentrated in particular industries, in Canada the relationship between ownership and industry is not strong enough to make Naylor’s and Cle- ment’s claims identical. More recently, Niosi (1985: 58) has advanced a complementary, but still broader argument,

As a social group the Canadian bourgeoisie is deeply divided. With so many large foreign multinational corporations present in Canada there is a comprador counterpart to the autochthonous Canadian bourgeoisie whose job is to manage these foreign subsidiaries ... The regional character of the Canadian economy provides a second basis for cleavage within the bourgeoisie ... land1 ethnicity is a third source of cleavage ... the Canadian capitalist class is now deeply divided along linguistic and ethnic lines, with the Anglophone/Francophone split the single most important cleavage.

Taken as a description of the demographic composition of the capitalist class, Niosi’s argument is self-evident. The more important and difficult question, however, is whether these cleavages affect its behavior as a social class. While the similarities in the arguments of Naylor, Clement, and Niosi are apparent, there are important differences in their emphases. Also, they refer to different historical periods: Naylor’s theoretical argument is built around a narrative about the nineteenth and early twentieth century; Cle- ment focusses on the postwar period and most of his empirical analysis is concerned with the early 1970s; and Niosi deals with developments a t the beginning of the 1980s. Still, all three point to the importance of examining regional, ownership and industry divisions within the Canadian capitalist class. Comparative analysis of Canadian economic development, notably the work by Ehrensaft and Armstrong (1981) and Laxer (19851, is used to support this model of distorted or truncated development.

An attack on this viewpoint can be mounted from a consideration of the now extensive research on comparative development, nicely summarized by Carroll (1986: 17m, which reveals an enormous diversity. Underlying the ‘exceptionalist’ arguments, Carroll argues, is a model of idealized, or at least typical, capitalist development which is portrayed by Canadian dependency theorists as involving transitions, first, from agriculture and primary production, and then to the intense industrialization of ‘mature’ capitalism. Not only is this sequence atypical, it is clearly not the path followed by the United States, to which Canada is often implicitly compared (Singelmann, 1978).


For the empirical analysis below, the question is whether the network of interlocks reveals the unusual weakness of Canadian capital, in the form of unusual fragmentation (perhaps resulting from the high level of foreign ownership) and the absence of major industrial corporations from the core of the network. Obviously, the view that Canadian economic development is not particularly unusual suggests that the Canadian network will also not be unusual.

GENERAL MODELS OF THE CAPITALIST CLASS AND DIRECTORATE INTERLOCKS

In his introduction to the 10 nation comparative study, John Scott (1985) describes five models of the organization of capital. Two of these models por- tray the division of large corporations into distinct, although not necessarily competing, groups of corporations centred around financial institutions (usually banks, but sometimes insurance companies or financial holding companies). The Marxist variant of this approach identifies these groups of corporations with the fusion of finance and industry, known as ‘finance- capital,’ which is said to mark the highest, ‘monopoly’ stage of capitalism. In the context of the finance-capital model, interlocking directorates identify groups of corporations which are organized around, but not dominated by, the financial institutions at their centres; finance capital involves the fusion of industry and finance, not the control of industry by finance. The non-Marxist ‘co-ordination and control’ model, in contrast, leads to the interpretation of the intensive interlocking among groups of corporations as a reflection of central, financial corporations’ (or, sometimes, family-controlled holding companies’) domination of these corporate groups. Both models suggest that the interlock network will reveal cliques centred around powerful financial institutions and/or family holding companies; but, where the finance capital model suggests relatively egalitarian relations between finance and industry, the co-ordination and control model suggests that there will be evidence of financial (or family) control of industrial corporations, presumably in the form of a much stronger propensity for the executives of financial institutions to sit on the boards of industrial corporations than for the executives of industrial corporations to sit on financial boards.

According to the resource-dependence model, interlocking directorates are used by corporations to order and control their relationships with corporations that constrain them. These inter-corporate constraints are actually market relationships between industries, for which the most systematic evidence is in the form of input-output tables. The presence of interlocks is predicted to correlate with the strength of market relationships between industries and/or individual corporations. The implication is that the network of interlocks will reflect dyadic relations between corporations, that it will be relatively sparse and without prominent centres. The network is also not likely to be hierarchical, since smaller and less powerful corporations have at least as much need for control of their environments as larger corporations.

The managerial and class cohesion models give rise to the predictions that there will be a relatively featureless, although hierarchical, network of


interlocks without discrete groups of corporations. From the managerial perspective, interlocks provide general information to corporate managers (enhancing what Useem, 1984 terms their ‘environmental scan’) as well as indicating a corporation’s prestige. From the class cohesion perspective, directorships provide communications within the capitalist class, which is conceptualized more broadly as the dominant social class.

Scott (1985: 10-11) differentiates the class cohesion from the managerial model by arguing that the class cohesion model requires that

the major structural feature of the network should be a separation bctwcen an ‘inner circle’ (or ‘corporate elite’) or multiple directors and the bulk of directors. This inner circle epitomizes the relationship between social background and board membership and they are likely to be the brokers between business and the political system.

I t is not clear, however, that there need be such a sharp distinction between single and multiple directors. Similarly, while many class theorists have stressed the inheritance of wealth in capitalism, nothing in the Marxist theory of capital accumulation necessitates restricted mobility into and out of the capitalist class. The validity of Marxist class theory should rest on evidence about whether it is possible to identify a common class interest of capi- talists and whether the project of this class involves the domination and ordering of society, as well as the control of individual corporations. Thus, on the basis of information about interlocking directorates alone, it is impossible to distinguish between the class cohesion and managerial models of capital.

Although these alternative theories lay emphasis on different aspects of the organization of capital, they are not strictly incompatible. It is easy to imagine individual capitalist classes displaying combinations (if not equal combinations) of the patterns predicted by two or more of these five models. So i t is not surprising that the numerous studies of interlocking directorates provide some evidence to support each one of these theories of the organization of capital. For example, Burt (1983) shows that interlocks within u.S. manufacturing reflect market relations between industries, as predicted by the resource dependence model. Studies of interlocks between the 800 largest US. corporations by Sonquist and Koenig (1975) and between the 250 largest corporations by Bearden and Mintz (1985) reveal city- and regionally-based groups of corporations centred around major financial institutions. Depending on whether these financial institutions are believed to control these groups, the findings are consistent with the finance capital or co-ordination and control models.

Stokman and his colleagues found considerable diversity among the nations they studied. In the u S., Britain and France, groups of corporations (‘cliques’) are centred around important financial institutions (banks in the U.S. and Britain and holding companies in France), but no strong centre in- tegrates the network. In the seven other nations the networks are more centralized. Though in most cases the largest banks and/or holding companies

158 MICKAEL ORNSTEIN

act as prominent local centres, the national networks are unified by the extensive overlaps between these spheres of influence. Two previous studies of the Canadian network, by Carroll (1986) and Carroll, Fox and Ornstein (1982) suggest that the Canadian network is much closer to the second, more unified, model. Carroll describes the network of ties among the largest 100 corporations in 19765 in these terms:

capitalist interest groups are centred primarily in investment companies, built around intercorporate ownership relations and based in Montreal or Toronto. Al- though these interest groups are structurally discernible, they are themselves ex- tensively interconnected. Financial institutions are positioned particularly well in the network as articulation points between intercorporate groupings (1986: 156)

While the highly clustered, poorly integrated networks in the u.s., Bri- tain and France conform to the predictions of the finance capital model, the more commonly observed combination of local centres with overall integration appears to lend support to a combination of two theoretical models: the clusters are suggestive of the fusion of financial and industrial interests identified by the finance capital model (although some of the clusters also reflect control by families, regional concentrations, or industrial groupings), while the overall integration of the clusters into a larger network conforms to the predictions of the managerial and class cohesion models. This combination is, however, more appropriately interpreted as favoring the managerial and class cohesion models over the financial capital model, since they suggest that there should be evidence of overall integration but do not pre- clude the existence of local clusters. The finance capital model, on the other hand, provides an interpretation of the clusters in the network, but does not identify a need for overall integration.

The only previous comparative study of Canadian directorates is Cle- ment’s (1978) analysis of the density and patterns of interlocks among the ‘dominant’ corporations - which were identified on the basis of their sales, assets and market shares (see Clement, 1975: 125ff.l - in Canada and the U.S. Surprisingly, in light of his theoretical arguments about the cleavage between the indigenous and comprador elites, Clement (1978: 166) found that the Canadian network of interlocks was denser than the U.S. network and that the relationship between manufacturing and finance was stronger in Canada than in the u.s.6 Unfortunately, such a comparison between Canada and the US., the most powerful developed capitalist nation, raises as many questions about the position of the Canadian capitalist class as it answers. In light of the idiosyncrasies of the American economy, such as the relative decentralization of banking and the regionally-centred character of the US. network of interlocks (Allen, 1974; Sonquist and Koenig, 1975) it is not clear how to interpet differences between Canada and the U.S. Peculi- arities of the American economy and capital could as easily account for Canada-us. differences as ‘distortions’ of the Canadian economy. There is also some risk that Clement’s results represent a methodological artifact: his analysis is based on, what by his criteria, are 113 dominant Canadian


corporations and 194 dominant American corporations. Because the density of networks of interlocks is known to decline as the number of corporations under study increases, Clement’s study is biased in the direction of discover- ing, as he did, that the Canadian network is more densely connected.

These theoretical arguments and previous empirical research suggest a number of questions for this empirical analysis of Canadian interlocks. Most obvious is the need to test the supposition - which follows both from the comparative research on the welfare state and from the arguments about the effects of economic dependency on Canadian capital - that the Canadian capitalist class is unusually fragmented and disorganized. Second, it is necessary to locate the Canadian network in the context of the alternative theoretical models of the organization of capital. And, third, it is essential to examine the positions of the major sectors of capital within the network, focussing on the major banks and other financial institutions and on the relative positions of domestic- and foreign-controlled corporations.

DATA AND METHODOLOGY

Following the model of the 10 nation collaborative project (Scott, 19851, this Canadian study includes the 50 largest corporations in the financial sector, ranked by assets, and the 200 largest corporations outside the financial sector, ranked by sales. In the 10 nation study, allowance is made for the addition of important corporations to and under some circumstances the deletion of corporations from this sample of 250 corporations. For Canada, the addition of six major property development corporations7 increased the study population to 256 corporations.’ These listings were all taken from the 1981 rankings carried out by the Financial Post. From public records it proved impossible to locate data on the composition of the boards of nine of these 256 co orations and they are assumed to have no links with the listed corporations.

From a Canadian perspective, the 10 nations in the study by Stokman, et ai. are not the best possible selection. These nations are sufficiently varied to provide a good context for discussing the structure of Canadian capital, but one might hope for the inclusion of additional nations with populations and positions in the world economy more similar to Canada’s, such as Australia, Sweden and Denmark.

After selecting the corporations, there remains the question of which individuals should be taken as representatives of each corporation. In Canada, and also in Britain and the US., this is not problematic, since formal control of a corporation rests with a single board, usually consisting of a mixture of ‘insiders,’ who are executives of the corporation, and ‘outsiders,’ who are not. Most European nations, however, have more complex, two-tiered boards. In Austria, Germany, the Netherlands and Switzerland, the executives of a corporation appoint an ‘executive board’ and the shareholders elect a ‘supervisory board,’ while in Belgium, Finland, France and Italy an ‘administrative board’ represents the managers and an ‘auditing board,’ which is appointed by the shareholders, has the limited responsibility of oversee- ing the corporation’s financial affairs. These alternative forms of two-tiered


boards are referred to, respectively, as the ‘German’ and ‘Latin’ forms of boards. Based on an analysis of the division of labor between these boards, the comparative study included both groups on the German boards, but only the administrative members of the Latin boards.

ANALYSIS

This section is divided into three parts. The first deals with the density and other general characteristics of the interlocks between corporations, the second considers the relationships between corporations in different sectors of the economy and the third examines the distribution of directorships among individuals. The first and third parts of this section provide precise comparisons between Canada and all 10 nations in the comparative project, but the comparisons in the second part are less systematic.

Relationships Between Corporations In Canada, and all 10 other nations, most of the 250 largest corporations are joined in a single network; that is, groupings within the network are not sufficiently marked to break it into distinct pieces. In all but three of the nations, there are small numbers of corporations that are isolated from the central network but have at least one tie with another corporation that is also isolated from the central network. In Canada there are only two such corporations. There are significant numbers of corporations which are completely isolated, in the sense that none of their board members also holds a position on another board in the sample. As Table I indicates, 50 of the 256 Canadian corporations are isolated - a number which is close to the middle of the range defined by the ten nations. With, respectively 24,27,30 and 44 isolated corporations, the US., Finland, France and Switzerland ex- hibit higher levels of ‘connectedness’ than Canada, while Germany, the Netherlands, Britain, Belgium, Italy and Austria have lower levels of connectedness (ranging from 53 isolated corporations for Italy, to 90 for Austria). There is a suggestion of a curvilinear relationship between national population and the number of isolated corporations: higher levels of connectedness appear to reflect either the very large size of corporations in the biggest nations (as in the US. and France) or the social proximity of corporations in very small nations (Finland and Switzerland). A number of other nations (particularly Germany, with 62 isolates), however, do not conform to this pattern.

Partly reflecting these differences in the number of isolated corporations, there is considerable variation in the density of the national networks. In Canada, 4 per cent of the unique pairs of corporations are actually directly linked by one or more common directors. Although this density level appears to be quite low, remember that the number of unique links between pairs of 256 corporations - all of which would have to exist for the density to reach 100 per cent - is over 32,000. In addition to the directly linked corporations, 27 per cent of all the pairs of Canadian corporations are linked indirectly through one other corporation, that is they are both linked via the board of

TAB

LE I

CO

NN

ECTI

VIT

Y A

ND

DEN

SITY

OF

TH

E N

ETW

OR

K O

F C

OR

POR

ATI

ON

S.

Can

ada

Aus

tral

ia

Bel

gium

B

rita

in

Fin

land

F

ranc

e G

erm

anv

Ital

v N

ethe

rlan

ds

Switz

erla

nd

USA

Com

pone

nts

of t

he N

etw

ork

(num

ber o

f cor

pora

tion

s)

In l

arge

st c

ompo

nent

20

4 14

7 18

2 18

5 21

0 22

0 19

5 18

0 19

0 20

6 22

6

Isol

ated

50

90

80

61

27

30

62

53

56

44

24

T

otal

num

ber o

f co

rpor

atio

ns

256

241

270

250

237

250

259

247

250

250

252

In s

mal

ler c

ompo

nent

s 2

4 8

4 0

0 2

14

4 0

2

Perc

enta

ge o

f cor

pora

tion

s lin

ked

dire

ctly

A

ll co

rpor

atio

ns

4 3

3 2

5 3

4 3

3 3

3 L

arge

st c

ompo

nent

onl

y 6

8 7

3 7

4 7

6 5

5 4

dire

ctly

or

indi

rect

ly l

inke

d 31

18

20

12

38

24

28

16

23

27

22

Pe

rcen

tage

of

all c

orpo

ratio

ns

Com

pone

nts

of t

he

Net

wor

k of

Pri

mar

y Ti

es

(no.

cor

ps.)

In

larg

est

com

pone

nt

179

126

137

84

187

168

174

155

141

151

190

In s

mal

ler c

ompo

nent

s 8

7 24

10

4

13

6 4

9 12

13

Isol

ated

69

10

8 10

9 15

6 46

69

79

88

10

0 87

49

Typ

es o

f int

erlo

cks

(per

cent

ages

) Pr

imar

y 30

33

36

48

32

25

28

39

24

21

29

In

duce

d by

pri

mar

y ti

es

27

43

39

25

29

26

34

29

22

16

16

No

prim

ary

tie

invo

lved

43

23

25

28

38

50

38

32

53

63

55

T

otal

10

0 10

0 10

0 10

0 10

0 10

0 10

0 10

0 10

0 10

0 10

0

Mul

tiplic

ity o

f O

ne

82

78

67

94

75

85

81

70

87

81

84

Lin

es

Tw

o

13

15

18

5 15

10

14

17

11

12

13

(p

erce

ntag

es)

Thr

ee

3 4

7 2

5 3

4 7

2 5

2 F

ouro

r m

ore

2 3

8 0

5 2

2 7

1 2

1 T

otal

10

0 10

0 10

0 10

0 10

0 10

0 10

0 10

0 10

0 10

0 10

0

Mea

n no

. int

erlo

cks

wr

line

1.27

1.

36

1.68

1.

09

1.50

1.

25

1.27

1.

55

1.20

1.

31

1.20

'Exc

ept

for

Can

ada,

dat

a are

from

Sto

kman

, Zie

gler

and

Sco

tt. 1

985:

25-

27


some third corporation. Counting both direct and indirect ties Canada ranks second after Finland among the 11 nations.

Naturally, nations with more isolated corporations tend to have lower interlock densities, although the relationship is not perfect: for example, while Canada has more than twice as many isolated corporations as the u.s., the overall density of the Canadian network is considerably higher. Britain, with only 2 per cent of all possible pairs of corporations directly linked and 12 per cent indirectly linked, stands out as having the lowest interlock density.

Since isolated corporations lower the density of the network, it makes sense to provide an alternative measure of network density of ties which excludes the isolated corporations (and also excludes the small number of corporations in small components which are unconnected to the largest component in the network). After adjusting for the number of isolated corporations, the 6 per cent of the possible ties in the largest component of the Canadian network actually exist. By this standard, Canada is near the middle of the range of densities among the ten nations.

In analyses of interlocking directorates it is also common to distinguish ‘primary’ ties, which involve an executive of one of the linked comporations from ‘secondary’ ties, which do not. Secondary ties may be further subdivided according to whether or not they involve an executive of another corporation in the network.” According to Stokman and Wasseur (1985: 39), the balance among the three types of ties is indicative of the relative importance of relationships between individual corporations, which are primarily reflected in primary ties, and the network’s role in the more general economic and political co-ordination of capital, which is best served by directors with many positions but no primary allegiance to one corporation. Secondary ties may be subdivided into two categories according to whether they are ‘induced’ by primary ties, that is whether they link two corporations which have primary links with a third corporation. Clearly, secondary ties that are not induced by primary ties are the ones most likely to serve the general interest of the corporations in the network.

Table I shows that 179 out of 204 Canadian corporations which are included in the largest component of the network, are also included in the largest component of the network defined on the basis of the primary ties alone. Thus the cohesion of the Canadian network does not depend on pro- fessional directors, who are not executives of any of its member corporations. In this respect, Canada is typical. Only in Britain, where just 84 of the 250 British corporations have any primary tie, is the network of primary ties much smaller than the overall network. Of course, there can be dense networks of both primary and secondary ties - and in Canada, this proves to be the case. Thirty per cent of the interlocks between Canadian corporations involve primary ties, another 27 per cent are induced by those primary ties and 43 per cent of the ties do not involve a primary tie. Again Canada is near the middle of the range of the wide range distributions of types of ties: in France, the Netherlands, Switzerland and the U S . , (uninduced) indirect ties account for half or more of all the links between corporations,

11


while in Austria, Belgium and Britain the corresponding proportion is about 25 per cent.12

Finally, Table I provides some further information on the relative importance of ties linking particular pairs of corporations, in the form of the frequency distribution of the ‘multiplicity’ of ties, that is the distribution of the numbers of ties between unique pairs of corporations. Links involving two or more interlocks, like primary ties, are not likely to have occurred by chance. Eight-two per cent of the links between Canadian corporations involve a single common director, 13 per cent involve two directors, 3 per cent involve three directors, and 2 per cent involve four or more directors. Canada is again near the middle of the range where, at one extreme, 94 per cent of British ties involve only one director and only 2 per cent involve three or more directors, and in Belgium and Italy, by contrast, about 15 of all the links between corporations involve three or more directors.

The findings of this general comparison between the Canadian and the 10 other networks are strikingly consistent. The high degree of foreign ownership and other peculiarities of Canadian capital clearly have not resulted in an unusually fragmented network. Depending on the criterion em- ployed, Canada falls in the middle of the range of variation among the 10 nations. The American and British networks are very weakly integrated, while the Canadian pattern is more similar to those exhibited by a number of European nations, including France and Germany, which are certainly not ‘undeveloped’ and ‘dependent’. Of course, these findings do not prove that weaknesses of Canadian capital, such as the high level of foreign control, do not decrease the integration of the network, only that there are there must be compensating factors for any such effects. In the next section, an examination of the relations within Canadian capital addresses this issue.

Intersectoral Differences For this analysis the corporations were initially divided into the financial, non-financial, and property development sectors. The Canadian-controlled non-financial corporation are divided into five groups: management controlled public companies;’’ corporations controlled by identifiable individuals or family groups; federal Crown corporation^;'^ provincial Crown corporations; and a remainder consisting of corporations with identifiable, but non-family, owners of a controlling interest (termed ‘non-family control’ in Table 11). The foreign-controlled non-financial corporations were divided according to whether control rested in the US., Britain, or elsewhere. The financial corporations were divided only into Canadian- and foreign- controlled. Beca se there are only six development corporations these were not subdivided.

From Table 11 it is immediately apparent that there are striking differences between the 11 categories of corporations. First, the foreign controlled corporations are much more likely to be isolated from the network: among the non-financial corporations, 31 of the 71 US. corporations are isolated, as are 3 of the 14 British corporations and 1 of the 8 from other nations; 3 of the 14 foreign financial corporations are also isolated. These

I!!

TABLE

I1

NU

MBE

R O

F Tl

ES A

ND D

ENSI

TY O

F TI

ES B

ETW

EEN

ELE

VEN

IND

UST

RY B

Y O

WN

ERSH

IP C

ATE

GO

RIE

S

Non

-Fin

anci

al C

anad

ian

Cor

pora

tions

N

on-F

inan

cial

For

eign

F

inan

cial

Sec

tor

Prop

erty

W

idel

y-

Fam

ily-

Fed

eral

Pr

ou.

Non

-fan

. C

orpo

ratio

ns

Deu

elop

- he

ld

Con

trol

C

row

n C

row

n co

ntro

l A

mer

ican

B

ritis

h O

ther

C

anad

ian

For

eign

m

ent

Den

sity

'of

Tie

s N

on-F

inan

cial

Can

adia

n C

orpo

ratio

ns

Wid

ely-

held

Fa

mily

con

trol

Fe

dera

l C

row

n Pr

ovin

cial

Cro

wn

Non

-fam

ily c

ontr

ol

Non

-Fin

anci

al F

orei

gn

Cor

pora

tions

A

mer

ican

B

ritis

h O

ther

Fina

ncia

l Sec

tor

Can

adia

n Fo

reig

n

Prop

erty

Dev

elop

men

t

Num

ber

of c

orpo

ratio

ns

Num

ber

of c

orpo

ratio

ns

292 95

31

52

84

57

37

400

16

69

30

53

61

35

57

23

12

6

48

11

0 0

70

31

63

31

314

135

106

53

31

20

12

12

42

83

28

42

16

36

6 12

77

28

10

20

20

13

53

26

10

71

150

59

49

115

186

8

5 5

27

42

2

48

12

0 21

10

6 24

0

38

71

13

n 36

14

6

outs

ide

the

mah

com

pone

nt

1

2 0

4 3

31

3

1

2 3

1

Mean n

umbe

r of

ties

w

ith o

ther

cor

pora

tions

29

.1

13.8

12

.5

6.5

15.3

6.

4 5.

9 12

.1

30.3

4.

3 11.3

*The

entr

ies are

equa

l to

the

num

ber

of in

terl

ocks

bet

wee

n co

rpor

atio

ns in

the

two

cate

gori

es d

efin

ed b

y th

e ro

w a

nd c

olum

n of

the

tabl

e di

vide

d by

the

num

ber

of u

niqu

e pa

irs

of c

orpo

ratio

ns, m

ultip

lied

by 1

000


findings are not unusual: for Austria, Ziegler, Reissner and Bender (1985: 82) find that 30 of the 60 foreign-controlled corporations are isolated and for Belgium, Cuyvers and Meeusen (1985: 155) report that half of the 108 foreign-controlled corporations are isolated. Although US. corporations make up just over one-quarter of all the corporations in the sample, 62 per cent of the isolated corporations are American. Besides the foreign-controlled corporations, only among the provincial Crown corporations is a significant proportion (4 out of 12) entirely outside the network.

Of course the proportion of isolated corporations is a crude measure of their attachment to the network, so Table 11 also gives the mean number of ties between the corporations in each category and all other corporations in the network. With means of 29.1 and 30.3 ties, respectively, the widely-held Canadian non-financial corporations and the Canadian financial corporations have about twice as many ties as any other group. This finding, which is consistent with a number of previous studies (Ornstein, 1976; Carroll, Fox and Ornstein, 1982; Carroll, 19861, demonstrates that it is not the Canadian financial corporations alone which dominate the network, but a mixture of Canadian non-financial and financial corporations. The similarity of the mean numbers of ties conceals a large difference in their variances. The five ‘big banks’ had an average of 85 ties, so that the mean number of ties of the remaining 31 Canadian financial corporations was oniy 21.4; the five non- financial corporations with the most interlocks had an average of 54.6 interlocks.

It is interesting to compare the most central corporations in Canada and the U.S., which for 1962 are gwen by Mintz and Schwartz (1985: 150). Of the 20 American corporations with the most interlocks, 12 are banks, five are insurance companies, two are railways and one is a manufacturer. Mizru- chi (1982: 67) and Bearden and Mintz (1985: 234ff.l show that this pattern of bank domination of the network extends from the turn of the century to the present. In Canada, the top 20 included seven banks, three insurance companies and 10 non-financial corporations. Not only is the prominence of Canadian banks not unusual in comparison with the US., but Ziegler (1985) demonstrates that Canada is quite similar in this respect to a number of other national networks, notably those of Switzerland, the Netherlands and Germany - all of which have only three or four dominant banks. In Eng- land, Scott and Griff (1985: 219) show that nine of the 10 corporations with the highest adjacency (i.e., links to the greatest number of other corporations) are banks.

In an intermediate group with between 11.3 and 15.3 ties per corporation are the Canadian family and non-family controlled corporations, federal Crown corporations, non-family controlled corporations, ‘other’ foreign non-financial corporations and development corporations. Least integrated into the network are the provincial Crown corporations, American and British non-financial corporations, and foreign controlled financial corporations, with means between 4.3 and 6.5 interlocks. Of course, the presence of a large number of isolated corporations in a category decreases the mean numbers of interlocks in that category. Adjusting the means to eliminate


the effect of isolated corporations produces only one major change in the mean number of interlocks: the mean of American controlled non-financial corporations increases from 6.4 to 11.4 interlocks. Thus, American corporations are more likely to be isolated; but those which are not isolated are only somewhat less likely to be connected to the network than their Canadian counterparts.

In Table 11 the density of ties between the various categories of corporations is given by the number of interlocks between each pair of categories divided by the number of unique pairs of corporations (multiplied by 1000, for convenience). For example, the density of ties between widely-held and Canadian financial corporations is equal to 314, which is equal to the number of ties between the two categories of corporations (181) divided by the number of possible pairs which can be formed out of the 16 widely held corporations and 36 Canadian financial corporations (576, or 16 times 361, multiplied by 1O0O.l6 This value is actually one of the largest in the Table. There are two other large, important values: the density of ties between pairs of widely held Canadian non-financial corporations is 292; and the density of ties between pairs of Canadian financial corporations is 186 - these are mostly ties between banks and trust and insurance companies, since the banks are forbidden to interlock with each other. Thus there are very dense ties within the widely-held non-financial corporations and between them and the Canadian financial sector. Except for the provincial crown corporations, the Canadian financial corporations have strong ties with all the types of Canadian controlled corporations and with development corporations.

Using a larger number of corporations (256 versus 100) and more recent data, this analysis provides strong confirmation of Carroll, Fox and Orn- stein’s (1982) and Carroll’s (1986) findings that, at the centre of the network are the largest, Canadian controlled financial and industrial corporations. While the five ‘big banks’ have the most interlocks, a number of large industrial corporations immediately follow them in ranking. This finding is clearly at odds with the predictions of the co-ordination and control model of interlocks, in which large financial institutions are seen as being at the centres of competing spheres. In the context of Canadian political economy, the findings raise doubts about the domination of ‘mercantile’ capital.

Both family-controlled and non-family-controlled non-financial Cana- dian corporations - the two categories of corporations for which it is possible to locate an individual or corporate controlling interest - have much fewer ties than the mana ement-controlled corporations whose ownership is ‘widely -distributed’! Combined with our findings about the relative isola- tion of foreign-controlled corporations, this suggests that corporations which are controlled by their managements (as are all the Canadian banks and a number of the major insurance companies) make the greatest use of the network. Confirming this argument is the finding that corporations with family or corporate controlling interests are more likely to have outgoing than incoming primary ties - in other words, their directors are more likely to serve on the boards of other corporations than are outsider executives to sit on their boards.


Foreign-controlled corporations lower the integration of the network and, in their absence, the Canadian network would be more highly integrated than the networks of all 10 other nations. Foreign corporations do not, however, form an alternative centre. The largest group of foreign-controlled non-financial corporations, those owned in the US., exhibits exceed- ingly weak internal ties (the density of ties between pairs of these firms is only 10); it has much stronger ties with Canadian corporations. Instead of forming an alternative, competing centre, as Clement and Niosi suggest, the foreign-controlled corporations are peripheral to a network dominated by the largest Canadian corporations.

Corporate Directors No analysis of the network of interlocks between corporations can be complete without consideration of the individuals whose membership on two or more boards constitute the links between the corporations. While there is necessarily a relationship between the characteristics of the network, just examined, and the distribution of positions among the directors, it is far from perfect. For example, because the number of links formed by an individual varies as the square of the number of positions he or she holds, small differences in the distribution of positions among individuals can result in very different network structures. Of course it is these directors who provide the link between capital - understood as the collectivity of capitals identified with corporations - and the capitalist class, the collectivity of individuals which controls capital and whose class interests involve not only control of the accumulation process itself, but the existence and reproduc- tion of the larger complex of institutions which constitute capitalist society.

By definition an analysis of the individuals who link pairs of corporations is restricted to multiple directors, whose interests cannot be entirely identified with a single corporation. At issue in the international comparison are differences in what might be termed the autonomy of the corporate network, that is the extent to which the network can be said to represent more than the interests of individual corporations. Compared to individuals with only two positions andlor inside directors, individual directors with positions on three or more boards and who are not executives of one of the corporations in the network are better able to serve the general interest of the corporations.

According to the measures of network autonomy in Table 111, the Canadian network is among the most autonomous. The ratio of positions to directors, 1.39, is slightly higher than the ratio observed in all the other 10 nations. In terms of the proportion of directors with two or more positions (relative to directors on the board of only one corporation), Canada, along with France and Switzerland, ranks highest. With 464 multiple directors Canada ranks second in the number of multiple directors, behind Finland and the U.S., both with 564. Twenty-two per cent of the Canadian multiple directors occupy 3 positions, 13 per cent occupy 4 positions, 7 per cent occupy 5 positions and 5 per cent occupy 6 or more positions. In these terms, Britain and the U.S. are exceptional: only 1 per cent of the British and Amer-

TABL

E 11

1 D

ISTR

IBU

TIO

NS

OF

DIR

ECTO

RS

AN

D P

OSI

TIO

NS‘

Can

ada

Aus

tria

B

elgi

um

Bri

tain

F

inla

nd

Fra

nce

Ger

man

y It

aly

Net

herl

ands

Sw

itzer

land

U

SA

Num

ber

of p

osit

ions

3205

Num

ber

of d

irec

tors

2298

Rat

io o

f pos

itio

ns

to d

irec

tors

1.39

Num

ber

of m

ulti

ple

dire

ctor

s 464

Prop

ortio

n of

mul

tipl

e di

rect

ors

0 0.20

Num

ber

of ‘b

ig li

nker

s’

118

Num

ber

of ‘n

etw

ork

spec

ialis

ts’

40

Num

ber

of p

osit

ions

Tw

o 53

held

by

mul

tipl

e T

hree

22

dire

ctor

s Fo

ur

13

(per

cent

ages

) Fi

ve

7 Si

x or

Mor

e 5

Total

100

Num

ber of

inte

rloc

ks

Tw

o

14

‘car

ried

’ by

dir

ecto

rs T

hree

17

wit

h di

ffer

ent n

os. o

f in

terl

ocks

(p

erce

ntag

es)

Four

69

Tot

al

100

2939

2430 1.21

271 0.11

50

11

65

17 9 4 4

100 14

11

75

100

3000

2203 1.36

373 0.17

92

28

57

19 9 6 8

100 10

10 81

100

3091

2682 1.15

282 0.11

27 4 69

21 6 3 1

100

33

30

37

100

4178

3110 1.34

564 0.18

110 41

61

20 6 6 7

100 15

15

70

100

2625

1931 1.36

378 0.20

80

42

60

19 9 6 6

100 17

16

67

100

4727

3943 1.20

420 0.11

82

32

60

20 9 5 6

100 16

16

69

100

2358

1737 1.36

322 0.19

65

16

63

17 7 5 8

100 14

13

73

100

2950

2321 1.27

357 0.15

68

34

64

17 8 6 5

100 19

16

65

100

3681

2999 1.23

405 0.20

56

31

67

19 6 2 6

100 21 17

62

100

3976

3108 1.28

564 0.18

69

39

64

24 8 3 1

100 28

31

41

100

*Exc

ept f

or C

anad

a, d

ata are

from

Sto

kman

, Zie

gler

and

Sco

tt, 1985: 22-24


ican multiple directors have six or more positions, compared to at least 4 per cent in the other nations. Stokman and Wasseur (1985: 23) suggest that the structures of the boards might account for this difference, since, among the 10 nations in the comparative study only Britain and the U.S. had single board systems. The addition of the Canadian case, however, shows that there is no necessary relationship between board structure and the distribution of directorships.

The distribution of the numbers of positions held determines the relative contribution to the network of individuals with different numbers of positions. Again because the number of interlocks created by an individual varies as the square of the number of positions he or she holds, individuals with large numbers of positions contribute disproportionately to the network. Al- though in every nation more than half of all multiple directors occupy just two positions, these directors are responsible for an average of only 20 per cent of all ties. In Canada, the 25 per cent of all multiple directors who occupy four or more positions - labelled ‘big linkers’ by Stokman et al. - account for 69 per cent of all interlocks. In this respect Canada is quite similar to eight of the 10 other nations; only in Britain and the US. do the occupants of four of more positions account for less than 60 per cent of all ties. Trans- lating these percentages into numbers, we find that 118 individuals account for more than two-thirds of all the interlocks between the largest Canadian corporations. Among the 10 nations, the number of big linkers varies from only 27 in Britain to 110 in Finland, with the mean about 80.

Stokman and Wasseur also distinguish ‘network specialists’ who have positions on the boards of the (four or more) corporations, but are not executives of any these corporations. Because the ‘network specialists’ have no primary affiliation with a corporation in the network, they may play a special role in unifying a system which is subject to cleavages reflecting the interests of economic sectors and individual corporations. Again there is wide international variation. In Britain, there are only four such ‘network specialists,’ while the Canadian, Finnish, French, and American networks include about 40.

Turning from the interlocks between corporations to the distribution of positions among individuals it is apparent again that the Canadian network more closely resembles the networks of the eight continental European nations than the networks of the U.S. or Britain, which is quite the opposite of what would be predicted from arguments about Canadian exceptionalism. In Canada, there is a very cohesive network, which is tied together by a relatively small number of individuals with many board positions.

CONCLUSIONS

The analysis indicates that the Canadian inter-corporate network is not unusually fragmented, that there are no pronounced cleavages between industrial and financial capital or between foreign and domestic capital, and that industrial capital is neither subordinated nor peripheral to finance. The Canadian network looks like the networks of a number of other countries, such as Germany and France, about which it is impossible to advance argu-


ments about dependency and underdevelopment. Paradoxically, it is the networks of the U.S. and Britain which are atypically fragmented. While, as a means of studying the organization of capital, there are limitations to focussing exclusively on interlocks, these findings are very damaging to the claims about the fragmentation and distortion of the Canadian capitalist class.

In terms of the more general questions about the organization of Canadian network, the analysis lends support to the models of capitalist class relations which emphasize unity and not the formation of discrete, potentially competing centres. As argued above, our data cannot be used to decide between the managerial and class cohesion models of capital. The strong relationships between corporate size and numbers of interlocks suggest that influence in the network is a reflection of economic power. The finding that the corporations most closely identified with individual capital- ists - the family-controlled non-financial corporations - are less integrated than the management controlled corporations is clearly at odds with an interpretation of the class cohesion model which emphasizes the role of inheritance and family relations. Rather than ruling out a class model, however, these findings suggest that the capitalist class can be characterized by considerable mobility. That the capitalist class is capable of gaining and integrating new recruits does not, however, make it less of a social class, (although it is further evidence against Porter’s and Clement’s characteri- zation of Canada as a low mobility society).

Whatever the strengths of the resource dependency model of the network - and a much more extensive analysis would be required to assess these effects - it is apparent that this model cannot explain the main features of the Canadian network. I t is not clear how to account for the effect of ownership on interlocking within the resource dependency framework, since in that framework interlocks are conceived as reflecting efforts to deal with exter- nal threats from other corporations. A further obstacle is posed by the very large number of interlocks involving financial corporations - because they supply capital, financial corporations are, in a way, dependent on all other corporations. Finally, it is difficult to account for the strong, positive relationship between corporate size and interlocking in the context of resource dependency theory. Within an industry, the vulnerability of a corporation to competition should not be a positive function of size; smaller corporations should be more vulnerable than their larger counterparts and should therefore require more interlocks.

If the structure of the Canadian network is not particularly unusual, Canada is an exception to the association between network centralization and corporatism observed by Ziegler. Although in terms of the relatively slow development of the welfare state Canada resembles the U.S. and Bri- tain, in terms of the network of corporate interlocks Canada resembles the more corporatist European nations. Of course, Canada conforms to the pattern identified in the social democratic model of the welfare state: the weakness of corporatism and slow development of the Canadian welfare state is correlated with the relative weakness of the Canadian labor movement and


of the national and most provincial social democratic parties. As well as lend- ing support to this model, the Canadian findings suggest an interpretation - or rather limitation - of the association between the development of the welfare state and the social organization of capital. As Panitch and others have noted, the development of the welfare state requires, and perhaps brings about, the organization of capital. The relative disorganization of capital in the U.S. and Britain suggests that without corporatism a highly concentrated network of interlocks is unlikely to develop. But this analysis of the Canadian network demonstrates that a strong interlock network can develop in the absence of corporatist pressures.

On the methodological side, this study demonstrates the risks of gener- alizing on the basis of a small number of national cases. For example, Stokman, et al. (1985: 28) find that:

Four countries had only a small number of isolated corporations and a large component of 200 or more corporations: the u.s.A., France, Finland and Switzerland. The small numbers of isolated corporation was mainly due to the relatively small numbers of subsidiaries of foreign enterprises among the selected corporations in these countries. In all countries these subsidiaries tended to bc isolated or only very loosely connected with their corporations.

But Canada is clearly an exception. Although foreign controlled-corporations are peripheral to the Canadian network, by a variety of different quan- titative standards, the Canadian network is among the most densely integrated.

The Canadian case also proves disturbing for another generalization sug- gested by the 10 nation study. One apparent source of variation among nations is the organizational form of the board itself. The two nations whose corporate networks are most distinctive (and least integrated) are Britain and the US. , which alone among the 10 nations have single corporate boards to perform the functions which in the eight other nations are divided between two boards. Scott (1985: 17) describes these single boards as ‘meeting frequently and regularly to take responsibility for determining and moni- toring corporate strategy.’ The suggestion is that the Anglo-American board system encourages a focus on the needs of individual corporations and dis- courages interrelations between them. In Canada the single board system is accompanied by a very high degree of corporate integration.

Summarizing the results of the 10-nation study, Ziegler (1985: 267) concludes,

I t has sometimes been said about studies of interlocking directorships that they have resulted again and again in the same finding: the vast majority of corporations are tied into a large component, with the biggest companies and especially the financial institutions most densely connected. Despite this expecta- tion, the great variations found among the ten nations studied are striking.

172 MICKAEL ORNSTEIN

In the context of these ‘great variations,’ the Canadian network is among the most centralized and dense. This characteristic of the Canadian network cannot be the result of some simple difference between Canada and the 10 other nations. For example, the Canadian network is denser than the networks of some larger nations (Britain, Italy and the US.) and some smaller nations (Belgium, the Netherlands and Switzerland). That this high degree of centralization has brought high levels of foreign ownership, resource dependency, low levels of research and development, etc., is unarguable - but these weaknesses do not develop out of a distorted, finance-dominated or fragmented capitalist class.

NOTES

The Canadian research includes Porter (1965: 298ff.); Barkans and Puppo (1974); Cle- ment (1975); Coleman (1986); Langille (1987); and Ornstein (1988). The outstanding American work is by William Domhoff (1979; 1980; 1983). Some key works in this tradition include Korpi (1978; 1983); Castles (1978); Bjorn (1979); Korpi and Shalev (1979a; 1979b); Stephens (1979); and Hollingsworth and Han- neman (1982). For an interesting effort to reconcile inconsistencies among these studies see O’Connor and Brym (1988). Two good summaries of these arguments are presented by Brym (1985: 3ff.l and Carroll (1986: lff.). This point is taken up first by MacDonald (1975) and Moore and Wells (1975) and then by a number of other scholars, who argue that Naylor incorrectly classifies the transportation sector as ‘mercantile’. Of course, in Volume 3 of Capital. Marx clearly classifies transportation as a part of industrial capital. The more difficult question, however, is not whether Naylor (who does not claim to be an orthodox Marxist) misuses Marxist theory, but whether there are unusually strong links between finance and transportation in Canada and whether this has any relationship to the extent and nature of Canadian economic development. Developments after 1981, which is the year to which the data discussed in this paper per- tain, suggest the emergence of family centred groups in a network which earlier studies describe as relatively undifferentiated (Francis, 1986). Surprisingly, in view of the stronger links between finance and manufacturing in Canada, Clement (1978: 167) concludes: ‘In Canada elite members in the sphere of circulation are thoroughly integrated with each other but have tenuous connections with the sphere of production, which is foreign controlled’. In both Canada and the U.S. the ‘trade’ sector, which is also part of the sphere of circulation, is relatively weakly linked to other sectors of the economy. The six are Cadillac Fairview Corp., Nu-West Development Corp., Daon Development Corp., Oxford Development Group Ltd., Trizec Corp. Ltd. and Campeau Corp. A more detailed consideration of the sampling issues should also consider the problem of ranking of corporations. The decision to select the largest 50 financial corporations, ranked by assets, and the largest 200 non-financial corporations, is arbitrary in a number of respects. First, the appropriateness of the 50:200 ratio of financial to non-finan- cia1 corporations in the sample depends on whether the 50th largest financial corporation and the 200th largest non-financial corporation are roughly comparable in economic importance. Of course, this depends on both the relative importance of these two sectors of the economy and on the degree of concentration in each - both of which are known to vary across nations. Second, within the non-financial sector, the decision to rank the firms by sales, rather than by assets, tends to exaggerate the importance of


corporations and industries whose sales are high relative to their assets. Reasonably, both a corporation’s sales and its assets are indicators of its importance in the economy. This is why corporations in the commercial sector, including department and food store chains and large-scale wholesalers are usually quite peripheral to the network of interlocks - both their assets and their profits are small relative to their sales. At the opposite extreme, extractive industries (especially when commodity prices are low) and utilities tend to be underrepresented in rankings based on sales. Because the number of interlocks of a corporation is known to vary strongly with corporate size, there is a potentially expensive, but not arbitrary solution to the problem of determining the optimal size of a sample of corporations for network analysis. As the number of corporations selected becomes larger and, therefore, each new selection becomes smaller, the number of interlocks added declines. Sooner or later, the new corporations add little or nothing to the network. If data on a sufficiently large number of corporations are available it is possible to determine where the network ends - or a t least reaches some lower limit of density. Thus the nature of the sample is determined empirically, rather than being arbitrarily specified as some round number.

9 All but two of the nine were foreign-owned corporations with highly centralized management, they include CIL Inc., Kraft Inc., PCL Construction Ltd., Pratt & Whitney Aircraft of Canada Ltd., Chevron Canada Ltd., Rockwell International of Canada Ltd., Coca-Cola Ltd., Eaton Yale Ltd., and Oxford Development Group Ltd. Since it is extremely unlikely that these corporations would have large numbers of ties with other firms in the network, the addition of their directors would do very little to influence the present findings.

10 For this analysis and subsequent discussion of corporations executives, the chair of the board is included among the executives, on the ground that the position of the chair is usually a full-time employee of the corporation and is intimately involved in its opera- tion. A very small number of primary ties are non-directional in that they involve individuals who hold executive positions (or one executive position and the chair of another board) on two different corporations.

11 There is considerable controversy over the relative strength of primary and secondary ties (for opposing views see Ziegler, 1985: 281 and Ornstein, 1984), there is general agree- ment that primary ties are indicative of a stronger relationship between corporations, since executives are in a much stronger position than outside directors to influence the behavior of a corporation.

12 Given that in four of the ten nations in their study secondary interlocks unassociated with primary ties are responsible for at least half of all ties, it is surprising that Stok- man and Wasseur (1985: 43) should conclude that: ‘the contours of an overall national network were determined by the pattern of its primary interlocks ... The primary interlocks can, therefore, be interpreted as the skeleton around which the body of a network is formed and which thus determines its shape’ (1985: 43). This interpretation suggests a prejudice in the direction of an inter-organizational interpretation of the network.

13 These corporations are identified as management-controlled because their shares were ‘widely distributed,’ according to the Financial Post.

14 For convenience, if not consistency, following the Financial Posf classification, the Federal Business Development Bank (a federal crown corporation) is classified with the Canadian financial corporations.

15 Although these 11 categories may appear to be a bit disorderly, the more log.lcal, complete classification of industries by ownership would have too many empty cells to be usable.

16 Note that because two or more interlocks may occur between a given pair of corporations, this ratio may exceed l and it tends to overstate the proportion of all pairs of cor-


porations which have at least one link. Correcting for the multiplicity of ties would do little to change the pattern in Table 111.

17 Of course it is possible that this effect of ownership might reflect size differences between the groups of corporations in the various ownership categories. A regression analysis (not presented here) showed that controlling for size - measured by both assets and sales - lowered the difference by only about one third.

REFERENCES

Allen, Michael 1974 ‘Interorganizational elite cooptation.’ American Sociological Review 39: 393-

406 Barkans, John, and Norene Puppo 1974 ‘The Board of governors and the power elite: A case study of eight Canadian

Bearden, James, and Beth Mintz 1985 ‘Regionality and integration in the American interlock network.’ Pp. 234-49

Bjorn, L. 1979 ‘Labor parties, economic growth, and the redistribution of income in five

Brym, Robert J. 1985 ‘The Canadian capitalist class, 1965-1985.’ Pp. 1-20 in Robert J. Brym (ed.),

The Structure of the Canadian Capitalist Class. Toronto: Garamond Burt, Ronald S. 1983 Corporate Profits and Cooptation. New York: Academic Press Cameron, Duncan 1988 ‘The dealers.’ This Magazine 21(8): 18-23 Carroll, William K. 1986 Corporate Power and Canadian Capitalism. Vancouver: University of

Carroll, William K., John Fox, and Michael Ornstein 1982 ‘The network of directorate interlocks among the largest Canadian corpora-

tions.’ The Canadian Review of Sociology and Anthropology 19: 44-69 Castles, Francis G. 1978 The Social Democratic Image of Society. London: Routledge and Kegan Paul Clement, Wallace 1975 The Canadian Corporate Elite: An Analysis of Economic Power. Toronto:

1978 Continental Corporate Power: Economic Linkages Between Canada and the

Coleman, William D. 1986 ‘Canadian Business and the State.’ Pp. 243-90 in Keith Banting (ed.), The

State and Economic Interests. Toronto: University of Toronto Press Coleman, William D., and Henry J. Jacek 1983 ‘The roles and activities of business interest associations in Canada.’

Cuyvers, Ludo, and Wim Meeusen

universities.’ Sociological Focus 7(3): 81-98

in Stokman, Ziegler and Scott (1985)

capitalist democracies.’ Comparative Social Research 2: 93-128

British Columbia Press

McClelland and Stewart

United States. Toronto: McClelland and Stewart

Canadian Journal of Political Science 16: 237-80


1985 ‘Financial groups in the Belgian network of interlocking directorships. ’ Pp.

Domhoff, G. William 1979 The Powers that Be. New York: Vintage 1983 Who Rules America Now? A View for the Eighties. Englewood Cliffs, N.J.:

Prentice-Hall Domhoff, G. William (ed.) 1980 Power Structure Research. Beverly Hills: Sage Ehrensaft, Philip, and Warwick Armstrong 1981 ‘The formation of dominion capitalism: economic truncation and class struc-

148-65 in Stokman, Ziegler and Scott (1985)

ture.’ Pp. 99-155 in Allan Moscovitch and Glenn Drover (eds.), Inequality: Essays on the Political Economy of Social Welfare. Toronto: University of Toronto Press

Fox, John, and Michael Ornstein 1986 ‘The Canadian state and corporate elites in the post-war period.’ Francis, Diane 1986 Controlling Interest: Who Owns Canada? Toronto: Macmillan Hicks, Alexander M., Roger Friedland, and Edwin D. Johnson 1978 ‘Class power and state policy.’ American Sociological Review 43: 302-15 Hollingsworth, J. Rogers, and Robert A. Hanneman 1982 ‘Working-class power and the political economy of western capitalist

Korpi, Walter 1978 The Working Class in Welfare Capitalism: Work, Union and Politics in

1983 The Democratic Class Struggle. London: Routledge and Kegan Paul Korpi, Walter, and Michael Shalev 1979a ‘Strikes, industrial relations and class conflict in capitalist society.’ British

1979b ‘Strikes, power and politics in the Western Nations, 1900-1976.’ Political

Kudrle, Robert T., and Theodore R. Marmor 1981 ‘The Development of welfare states in North America.’ Pp. 81-124 in Peter

societies.’ Comparative Social Research 5: 61-80

Sweden. London: Routledge and Kegan Paul

Journal of Sociology 30: 164-87

Power and Social Theory 1: 299-332

Flora and Arnold J. Heidenheimer (eds.), The Development of Welfare States in Europe and America. New Brunswick, N.J.: Transaction Books

Langille, David 1987 ‘The Business Council on National Issues and the Canadian State.’ Studies

Laxer, Gordon 1985 ‘The Political Economy of Aborted Development: The Canadian Case.’ Pp.

67-102 in Robert J. Brym (ed.), The Structure of the Canadian Capitalist Class. Toronto: Garamond

in Political Economy 27: 41-85

MacDonald, Larry 1975 ‘Merchants against industry: An idea and its origins. ’ Canadian Historical

Mintz, Beth, and Michael Schwartz 1985 The Power Structure of American Business. Chicago: University of Chicago

Review 56: 261-81

Press


Mizruchi, Mark S. 1982 The American Corporate Network, 1904-1974. Beverly Hills, Cal.: Sage Moore, Steve, and Debi Wells 1975 Imperialism and the National Question in Canada. Toronto: privately

Naylor, Tom 1972 ‘The rise and fall of the third empire of the St. Lawrence.’ Pp. 1-41 in Gary

published

Teeple (ed.), Capitalism and the National Question in Canada. Toronto: University of Toronto Press

Niosi, Jorge 1985 ‘Continental nationalism: The strategy of the Canadian bourgeoisie.’ Pp. 53-

66 in Robert J. Brym (ed.), The Structure of the Canadian capitalist class. Toronto: Garamond

O’Connor, Julia S., and Robert J . Brym 1988 ‘Public welfare expenditure in the OECD countries: towards a reconciliation

Ornstein, Michael 1976 ‘The boards and executives of the largest Canadian corporations: Size, com-

position and interlocks.’ Canadian Journal of Sociology 1: 411-37 1984 ‘Interlocking directorates in Canada: Intercorporate or class alliance?’ Ad-

ministrative Science Quarterly 29: 210-31 1988 ‘Corporate involvement in Canadian hospital and university boards, 1946-

1977.’ The Canadian Review of Sociology and Anthropology 25: 365-88 Panitch, Leo 1986 ‘The tripartite experience.’ Pp. 37-119 in Keith Banting (ed.), The State and

Economic Interests. Toronto: University of Toronto Press Porter, John 1965 The Vertical Mosaic. Toronto: University of Toronto Press Schmitter, Phillippe, and G. Lehmbruch (eds.) 1979 Trends Towards Corporatist Intermediation. Beverly Hills: Sage Scott, John 1985 ‘Theoretical framework and research design.’ Pp. 1-19 in Stokman, Ziegler

Scott, John, and Catherine Griff 1985 ‘Bank spheres of influence in the British corporate network.’ Pp. 215-33 in

Shalev, Michael 1983 ‘Class politics and the western welfare state.’ Pp. 27-50 in Shimon E. Spiro

of inconsistent findings.’ British Journal of Sociology 39: 47-68

and Scott (1985)

Stokman, Ziegler and Scott (1985)

and Ephraim Yuchtman-Yaar, Evaluating the Welfare State: Social and Political Perspectives. New York: Academic Press

Singelmann, Joachim 1978 From Agriculture to Services. Beverly Hills: Sage Skocpol, Theda, and Edwin Amenta 1986 ‘States and Social Policies.’ Annual Review of Sociology 12: 131-57 Sonquist, John, and Thomas Koenig 1975 ‘Interlocking directorates in the top U.S. corporations: A graph theory ap-

Stephens, J. D. proach.’ Insurgent Sociologist 5: 196-230


1979 The Transition from Capitalism to Socialism. London: Macmitlan Stokman, Frans N., and Frans W. Wasseur 1985 ‘National networks in 1976: A structural comparison.’ Pp. 20-44 in Stok-

Stokman, Frans N., Rolf Ziegler, and John Scott 1985 Networks of Corporate Power. Cambridge: Polity Press Sundqvist, Sven-Ivan 1987 Agarna och Makten i Sveriges Borsforetag 1987 (Bilingual, in Swedish and

English, titled Owners and Power in Sweden’s Listed Companies). Stock- holm: Dagens Nyheter

man, Ziegler and Scott (1985)

Useem, Michael 1984 The Inner Circle. New York: Oxford University Press Ziegler, Rolf 1985 ‘Conclusion.’ Pp. 20-44 in Stokman, Ziegler and Scott (1985) Ziegler, Rolf, Gerhard Reissner, and Donald Bender 1985 ‘Austria incorporated.’ Pp. 73-90 in Stokman, Ziegler and Scott (1985)

mguisti-language Behavior &a& Fast, economical )o/vrnent delivery available.

Documents

The social organization of the Canadian capitalist class in comparative perspective