Introduction

In many developed economies networks of interlocking directorates between business enterprises are disintegrating.¹ This evolution is part of a more general trend of increasing internationalization and liberalization of economic activities and regulations, which put national production systems under pressure to move towards more market-oriented economies (Berger and Dore, 1996). Board interlocks, which occur when corporate directors sit on the board of more than one firm, increase the cohesion of the business elite, favor cooperation, facilitate coordination of economic activities and permit mutual control (Mizruchi, 1996; Windolf, 2002). Under the influence of increasing competition in more internationalized markets, such mechanisms of extra-market coordination among economic actors are increasingly called into question. In this context, cartels, corporate governance structures and corporatist arrangements have changed and are increasingly replaced by market-based forms of interaction.² The extent of these changes at the level of both national economies and companies has led different authors to designate this evolution as a 'convergence' of European countries' business systems on an Anglo-American model (see notably Hansmann and Kraakman, 2004).

Both the extent and the significance of the observed changes in inter-corporate networks have given rise to varied interpretations and disagreement among scholars. Some authors interpret the disintegration as a sign of profound changes in organized capitalism (Höpner, 2003), others consider the changes to constitute merely marginal adjustments (Heinze, 2004; Nollert, 2005), or changes in the structure of the network but less so in its function (Kogut and Walker, 2001).

Part of the disagreement arises from the fact that most studies focus purely on structural characteristics of the network, which makes it difficult to understand the reasons of the changes. In this article we attempt to move beyond the purely structural analysis of corporate networks by looking at the 'actors of change'. We investigate the disintegration of the network of interlocking directorates in Switzerland and the Netherlands over the last decade of the 20th century. The focus is on the role different categories of companies (i.e. companies which are exposed to financial market pressures and financial companies) played in the changes of the network. This allows us to better understand and interpret the significance of changes in interlocking networks. Theoretically, this approach points to the importance of completing structural analyses of the institutional setting of capitalist economies with an actor-centered approach.

Both Switzerland and the Netherlands traditionally had dense and comprehensive networks of corporate board interlocks (Helmers et al., 1975; Stokman et al., 1985; Windolf, 2002; Nollert, 2005). Dense company networks are indeed particularly developed in small, continental European countries characterized by corporatist arrangements and cooperative relations among economic actors (Katzenstein, 1985). Given the strong exposure of small states to international markets, we expect that such countries face particularly strong pressures for reform fuelled by internationalization towards a more liberal-market-oriented business system. This is all the more the case with liberal-corporatist states where a strongly developed sector of internationally oriented multi-national corporations (MNCs) exists. Owing to the small home market, the presence of a strongly developed sector of MNCs leads to high levels of internationalization (Ruigrok and Van Tulder, 1995). Katzenstein refers to Switzerland and the Netherlands as prime examples of 'liberal corporatist countries', where large firms 'have succeeded in overcoming the restrictions a small domestic market imposes on their growth through early, rapid and sustained moves towards exports first and the internationalization of production later' (Katzenstein, 1985: 107).

In this article we explore two main differences between Switzerland and the Netherlands, which – we argue – had a major impact on the changes in the interlocks network. Firstly, we consider the difference in the exposure of Swiss and Dutch companies to market pressures and the reorientation of corporate strategies towards the interests of shareholders. Secondly, we address the differences in the type and role of banks in the two countries, and the increasing reorientation of banks' strategies towards financial market-related activities such as investment banking.

Our findings show that changing bank strategies do indeed affect corporate board interlocks in the sense that more financial market-oriented strategies lead to an increasing propensity of banks to interlock. Our results are more ambiguous concerning the impact of shareholder-oriented corporate governance practices on interlocking directorates. It appears that firms which are most exposed to financial market pressures are also the most central firms, which contradicts our expectation. This suggests that 'sociological reasons' for interlocking – such as the prestige of the largest firms – may have to be taken into account in order to understand why companies are central in the interlock network.

The article is structured as follows. The next section elaborates on the internationalization of the Swiss and Dutch business systems. The following two sections discuss two central differences between the two countries: corporate governance reforms and the role of financial institutions. The fifth section discusses our data sets and methods. The penultimate section presents the empirical evidence of network change in Switzerland and the Netherlands. In the last section, we discuss our findings and conclude.

Small states under international pressure

While the Netherlands is about twice the size of Switzerland, both in terms of population (16.5 compared to 7.5 million) and in terms of its economy (with a GDP of $629.9 billion compared to $371.5 billion), both countries can be qualified as small states.³ The small size of the domestic market results in a high degree of internationalization, which is expressed by high levels of imports and exports on the one hand and by the internationalization of production on the other hand. In 2005, Dutch exports amounted to $365.1 billion, and imports to $326.6 billion. The 2004 estimates for Swiss trade report $130.7 billion in export and $121.1 billion in import revenue. Compared to GDP, levels of exports and imports in Switzerland and the Netherlands are considerably higher than in larger countries, such as Germany.

The internationalization of business in Switzerland and the Netherlands also becomes apparent in the number and the size of Swiss and Dutch MNCs. In 1990, two Swiss MNCs and two Dutch MNCs ranked among the 10 largest companies worldwide, according to total assets. In comparison, no German MNC can be found among the top ten. If we look at the ranking of the 100 largest companies worldwide, the number of Swiss and Dutch companies listed stands out in comparison to their larger neighbor, Germany (five Swiss MNCs and four Dutch MNCs, in comparison to eight German MNCs).

These MNCs constitute a considerable part of these small countries' economies. For instance, in the Netherlands, the six largest listed companies account for 64% of total market capitalization (Royal Dutch Shell alone accounted for 23% ). The importance of a handful of very large MNCs leads to a strong exposure to international market pressures as these large companies can be expected to be more responsive to international pressures than companies who can rely on a large domestic market (Ruigrok and Van Tulder, 1995). Therefore, both Switzerland and the Netherlands can be expected to face similarly strong pressures for an increasing liberalization of their economic and corporate governance structures.

A first sign of change in recent years concerns the development of financial markets. Figure 1 shows the increase in stock market capitalization as a percentage of GDP over the last decade of the 20th century in seven small European countries. The evolution of capital markets since the 1980s clearly reflects the increasing importance of capital markets for companies. It also shows that Switzerland and the Netherlands are the countries where this change was the strongest.

Figure 1.

Evolution of financial markets in small European states (1980–2000). Source: (Rajan and Zingales, 2003).

Full figure and legend (96K)

Despite these similarities, two notable differences between the two countries lead us to expect distinct trajectories of network disintegration. First, Swiss business showed a surprisingly strong response towards claims for the protection of shareholder interests. While the business system was traditionally insider-oriented and axed on the isolation from foreign influence, starting in the late 1980s corporate governance reforms took place that increased companies' exposure to market pressure. As a result, Swiss companies have been increasingly exposed to pressures stemming from financial markets. The Dutch business system, on the other hand, remained wary of legal and institutional change in corporate governance. Dutch companies remained largely isolated from direct market pressures. Owing to their stronger exposure to financial market pressures, we expect Swiss companies to create fewer interlocks than in the past, whereas Dutch companies will resist change.

Second, financial institutions play a different role in the organization of the two economies. Swiss banks resemble the German-style universal banks, while the Dutch financial system resembles more the Anglo-American model. As a result, Swiss banks were more central to the interlock network than Dutch banks. Hence a decreasing propensity to interlock through financial institutions – as it has been observed in other countries (Davis and Mizruchi, 1999) – would destabilize the Swiss network to a greater extent than in the Netherlands. The next section elaborates on the differences in corporate governance reforms and the subsequent section on the differences in the financial sectors. Following our discussion of the institutional settings of the two countries, we will analyze the corporate network dynamics.

Corporate governance reforms

Traditionally, the economic openness of small countries was accompanied by a pronounced aversion of their companies towards outsiders' influence. The existence of open economies combined with closed corporate governance structures has been called an 'associative–dissociative' relationship with international markets, which leads to the use of instruments of 'selective protectionism' (David and Mach, 2004). While both Switzerland and the Netherlands show features of this paradox between 'associative' and 'dissociative' strategies towards international markets, some differences in the degree of openness exist. Traditionally, the Netherlands was more open to foreign influence than Switzerland. While the Netherlands is actively involved in European and Trans-Atlantic cooperation, and is a founding member of both NATO and the European Union, Switzerland remained during long time aloof from most supranational organizations. Throughout the 20th century, Switzerland has been characterized by pronounced economic nationalism and a certain political isolationism due to its neutrality policy. Switzerland is not a member of the NATO, did not become a member of the UN until 2002 and it is still not a member of the EU.⁴ This isolationist attitude is also reflected at the level of individual companies which were more isolated from foreign influence than Dutch companies. While it is not unusual for bi-national companies in the Netherlands to be owned and managed jointly with foreigners (such as Dutch-British companies Unilever and Royal Dutch Shell), Swiss companies were more hostile to foreign influence. During most of the 20th century, it was almost impossible for foreign investors to take control of Swiss companies. Controlling minority structures strongly limited foreign shareholders' rights (Bebchuk et al., 2000). The most important device in this respect was the limitation of the transferability of registered shares (the Vinkulierung procedure) by which management or the board of a company could refuse any shareholder the registration in the stock ledger, thereby preventing them from exercising their voting rights. This device was effectively used as an instrument to block foreign influence, which led international investors to designate Switzerland's corporate governance system as the 'fortress of the Alps' (Monks and Minow, 1995). While Dutch companies' insiders were equally eager to keep control over 'their' firms and to limit the influence of outside shareholders through a wide variety of anti-takeover defenses, no comparable 'economic nationalism' existed in the Netherlands. Also, the instruments of insider control differ considerably in Switzerland and the Netherlands. Most importantly, corporate control in Switzerland is to a greater extent based on concentrated ownership. Ownership in the Netherlands is less concentrated, but a wide range of devices of minority control over companies exist.

Since the late 1980s, Swiss companies have undergone extensive changes in their corporate governance practices and opened up towards foreign investors. Since the first Swiss company (Nestlé) opened up its registered stock to foreigners in 1989, more and more companies followed their lead. This increased foreign investors' influence over Swiss companies. From the late 1980s onwards, this increasing openness also led to an increasing exposure of Swiss companies to hostile takeover threats. Many of the early takeovers were unsuccessful due to defensive mechanisms such as dual class shares and Vinkulierung. The business elite reacted strongly following the first hostile takeover attempts and a takeover code was adopted at the end of the 1980s. This code introduced de facto a compulsory bid rule, making takeovers more expensive (Schnyder, 2007). Yet, despite this initial rejection, hostile takeovers were soon considered a legitimate instrument of a market economy and the business elite's resistance waned. This development was favored by the abolition of complex capital structures by many companies through the introduction of a unitary share and by the abolition of restrictions to exercising voting rights. These changes reduced Swiss companies' isolation from (financial) market pressures considerably, and by the early 21st century a market for corporate control had emerged. Within two decades, Switzerland has evolved from one of the most insider-friendly systems in Europe towards a system which ranks among the top three countries in international corporate governance standards in 2006 (Aggarwal et al., 2007).

No comparable evolution can be observed in the Netherlands, where corporate governance reforms remained superficial. Practically no hostile takeovers had taken place in the Netherlands until the early 2000s (De Jong et al., 2001) and shareholders' influence over stock corporations was limited in many ways. An array of defense mechanisms protects the Dutch corporate elite from – in their eyes – excessive shareholder influence. Priority shares (prioriteitsaandeel) give certain shareholders extra votes in matters such as nominating management and directors, dividend payout and the modification of statutes. Because there is no relationship between the voting power of the share and capital investment, a share of 20 euros can be enough to control certain major decisions within a firm. Preference shares (preferente aandeel) are used by the management in order to destabilize the (voting) power distribution of current shareholders by issuing new shares to friendly investors. The most important non-statutory instrument of insider control is the use of certificates instead of shares. The company sets up an administratiekantoor, a special purpose trust office that owns most or all of the company's shares, and issues only non-voting certificates. These certificates carry all of the share's economic rights, but not the voting right. The boards of such administratiekantoren are typically populated by (former) directors of the firm whose shares are kept by the trust office, and other well-connected directors. Under the influence of shareholder activists, more and more companies who issue certificates now allow certificate holders to make use of voting rights attached to shares. However, since many shareholders typically do not participate in shareholder meetings, the administration office still maintains a decisive say.

In sum, in 2001 Dutch corporate governance was still characterized by management control. Contrary to Switzerland, growing pressures to reform corporate governance hardly led to significant changes in the Dutch business system. In 1997, a list of 40 recommendations for good corporate governance was drawn up by a committee of experts, chaired by former Aegon CEO and collector of board mandates JFM Peters (Commissie Corporate Governance, 1997). Five years later, an evaluation of these 'recommendations' made it very clear that they had not triggered any significant change in corporate practices (NCGS, 2002). Consequently, a more stringent code of corporate governance – the Tabaksblat code – was adopted (Commissie Corporate Governance, 2003). However, this code left the mechanisms of insider-control largely untouched, and also did not reconsider the Dutch stakeholder-oriented two-tiered board system, the so-called 'structured regime'. Corporate law was not revised until 2004 when a legal reform granted shareholders somewhat more influence. Only recently have some of the larger stock-listed firms become 'commodities' on the financial market. The breaking up of ABN AMRO is a prime example. But these changes are much more recent than the changes in Switzerland. As late as 2000, the Netherlands had 'more takeover defenses, fewer voting rights, and less truly independent boards than other [ ...] countries, with the exception of Japan' (Gourevitch and Shinn, 2005: 181, note 78).

Increasing exposure to financial markets can be expected to lead to a disintegration of networks of interlocking directorates as board interlocks are at odds with shareholders' interests. Three elements appear particularly important in that respect. First, the recent wave of internationalization and liberalization goes hand-in-hand with a valorization of market mechanisms over extra-market means of coordination of economic activity (Dore, 2002). Given that inter-firm networks can be seen as an alternative form of organization to markets and hierarchies (Fennema, 1982; Powell, 1990), such relations – like other extra-market forms of coordination – are incompatible with market-oriented reforms. Second, from the perspective of the outside minority shareholder, board interlocks constitute a possible source of managerial agency costs and collusion. The strong personal ties among the corporate elite can be interpreted by shareholders as a reflection of the managers' power position vis-à-vis the owners. Third, the claim for more efficient control of executives through boards has led to an increased professionalization of the boards, which in turn means that the workload associated with a board seat has increased. For this reason, directors can be expected to hold fewer board positions. Also, following the major corporate scandals of the last years and the increasing scrutiny and public criticism of the performance of board members, it has become riskier to serve on multiple boards. Therefore, we can expect that companies that are particularly exposed to market pressures will tend to interlock less then companies who are isolated from such pressures. As a result, we expect the Swiss network to show stronger signs of disintegration than the Dutch network.

Changes in the financial system

A second difference between the Swiss and the Dutch economies concerns the financial system. Throughout the history of modern corporations, financial institutions have been the linking pins in corporate networks (see notably Stokman et al., 1985 and Windolf, 2002). Owing to the importance of banks as capital providers and their interest in monitoring the companies which they financed, bankers often sat on the boards of industrial companies. In Germany, large universal banks provided long-term loans ('patient capital') to industrial companies but were also active on financial markets. Switzerland resembles the German system with a number of large and powerful universal banks that are active both in corporate lending and financial market transactions. No such system existed in the Netherlands, at least until the 1990s, and Dutch banks more closely resembled Anglo-Saxon banks, whose relationships with non-financial companies are more diffuse (see De Jong and Röell, 2005; Heemskerk, 2007).

The role of Dutch banks as creditors for companies was traditionally secondary. Rather than long-term bank loans, companies would satisfy their need for capital through retained earnings or, alternatively, short-term bank loans. This practice began to change in the 1960s when Dutch banks increasingly engaged in the lending business. Dutch banks' leeway in offering different financial services was limited by regulations. Dutch financial institutions were practically not allowed to hold significant shares in other companies. These restrictions were gradually lifted, but only as a result of the liberalization of financial services following European integration (notably as a result of the Second European Banking Directive). As a result, Dutch banks only became important owners in non-financial companies during the 1990s. This evolution goes against the trend in other countries where banks are increasingly reluctant to hold large stakes in non-financial companies (see for the German case Beyer, 2002). The liberalization of the early 1990s led to the formation of new, large financial conglomerates combining banking and insurance activities in the Netherlands, such as ABN AMRO (1991), ING (1991) and FORTIS (1990). These new corporations embraced bancassurance strategies, where banks also offer life insurance and other insurance contracts to their clients through joint ventures, takeovers of insurance companies, or simply through the development of new products.

In Switzerland, in the absence of legal obstacles, major banks developed bancassurance strategies early on. As early as 1983, UBS created its own life-insurance company SBG Leben. Crédit Suisse created a subsidiary insurance company, CS Life, in 1989. CS also began collaborating with the Winterthur Insurance company, which it eventually took over in 1997. SBC – at the time the third largest Swiss bank – and the Zurich insurance group started a joint venture in 1992 in order to develop their life-insurance business (Schaub, 1992). Therefore, the insurance business became important for Swiss banks somewhat earlier than for Dutch banks.

In parallel with bancassurance strategies, Swiss and Dutch banks increasingly reoriented their strategies towards new and more profitable activities than the traditional lending business. Investment banking activities (counseling in M&A activities, issuing of new shares, etc.) as well as asset and wealth management became increasingly important fields of activity for banks. Also, as noted above, non-financial companies increasingly use financial markets as a source of finance through the issuing of shares rather than bank credits. This 'disintermediarization' of corporate finance together with the reorientation of bank strategies reduced the dependence of banks on the traditional lending business. This is reflected in the distribution of banks' income. Figure 2 shows that in both countries, less and less of banks' income is generated by lending business and more by activities linked to financial market transactions. Yet again, despite similarities, we also find differences in the evolution of bank strategies between Switzerland and the Netherlands.

Figure 2.

Evolution of bank income (in % of total income). Source: OECD, Bank Profitability Statistics, Income Statement and Balance Sheet, Vol. 2005, release 01.

Full figure and legend (33K)

Swiss banks show a more pronounced reorientation towards new types of activities as indicated by the shift in income structure. Whereas Dutch banks realized in 1980 74.2% of their income through interest and 25.8% from commissions and fees, Swiss banks' interest income only accounted for 52.0% of total income and non-interest income for 48.0% . By 1990, the share of interest income for Dutch banks was still 71.6% of total income, while it accounted for 51.2% of total income for Swiss banks. By 2000, interest income had considerably decreased in both countries, but accounted still for more than half of Dutch banks' income (53.0% ), while the majority of Swiss banks' income was realized through non interest-related activities (62.7% ). Swiss banks appear therefore to rely more heavily on sectors of activity that are related to financial markets than Dutch banks.

Davis and Mizruchi (1999) show for the USA that strategic reorientation of banks towards financial market-related activities has had an important impact on the board interlock network. Increasing dependence on financial markets create incentives for bankers to abandon their traditional board seats in non-financial companies for two reasons. On the one hand, their presence on corporate boards is – due to the lower exposure to lending risks – no longer needed for credit monitoring. On the other hand, the presence of banks on company boards could be negatively perceived by bank clients as it may cause conflicts of interests concerning investment counseling activities. More generally, the transformation of banking entails a move away from activities where close personal ties between banks and non-financial companies were an instrument allowing banks to exert control or to coordinate their activities with non-financial institutions (relationship-based banking) towards activities where such ties are rather a handicap and where anonymous market-relations prevail (transaction-based banking). Therefore, we expect bankers to be increasingly reluctant to sit on other companies' boards, which will – given banks' traditionally very central position within the interlocks network – have an important impact on the network as a whole. Given the differences in the evolution of bank strategies in the two countries, we would expect changes in the position of banks within the network to be more pronounced in Switzerland than in the Netherlands.

Data and methods

We investigate the disintegration of the network of interlocking directorates of the largest firms in Switzerland in 1990 and 2000, and the Netherlands in 1996 and 2001. With disintegration we designate changes that lead to a sparser or less connected network. Disintegration can take place at the level of the network as a whole, and at the level of individual firms. When fewer firms are connected in the network of interlocking directorates, disintegration results in decreasing connectivity (Wasserman and Faust, 1994). The connected part of the network becomes smaller. At the level of the firm, disintegration takes place when the number of board interlocks with other firms decreases. In this case, individual firms are connected to fewer firms by shared board members. A decrease in the average number of firms' interlocks equals a decrease in network density. Both types of disintegration are not necessarily related. It might be the case that fewer firms are part of the network, but at the same time have more interlocks among each other.

Board interlock data

The Swiss sample is constituted of the 106 largest Swiss companies in 1990 and the 108 largest companies in 2000, based on the ranking published by Schweizerische Handelszeitung. Missing figures were completed based on the annual reports of the companies. We selected the 70 largest industrial and service companies according to market capitalization, the 10 largest finance companies according to total assets and the 10 largest insurance companies based on net prime income. We also included the three largest cantonal banks and three large private banks as well as three transportation companies and five electric energy supply companies. Finally, we checked if any of the 20 largest companies according to turnover or any of the 10 largest employers for both years had not been selected using the criterion of market capitalization, which led to the inclusion of some large private companies. Information on board composition was retrieved from annual reports for 1990 and for 2000 from the 'CD-Rom der Schweizer Wirtschaft' published by Orell Füssli. Also, we were in the fortunate position to make use of board composition data from Claudio Loderer, University of Bern and Urs Peyer, INSEAD to whom we are grateful.

For the Netherlands we collected a sample of the 50 largest financial firms ranked by assets, and the 200 largest non-financial companies ranked by turnover for 1996 and 2001. The sample includes both listed and non-listed firms. Subsidiaries of firms already selected have been excluded, so no full ownership relationships exist in the sample. Information on board composition and financial figures stems from annual reports as well as from REACH database published by Bureau van Dijk. The difference in sample size and in the sampling method does not constitute a major problem, since we are mainly interested in within case comparison between the two dates in each country and only in the comparison of mechanisms of changes across the two cases. The difference in time periods is due to data availability on ownership, which we discuss below.

Dutch firms typically have a two-tier governance structure with an executive board (raad van bestuur) and a separate supervisory board (raad van commissarissen). Some firms, notably large MNCs, apply a one-tier model. Swiss corporate governance is characterized by a one-tier board system where no clear distinction between executive and non-executive directors exists. In fact, the Swiss board of directors (conseil d'administration) can either delegate management to professional managers who are not board members or run the company's business themselves. Despite a trend towards an increasing separation of executive and non-executive functions, no legal obligation to delegate exists; therefore, boundaries between the two functions on Swiss boards remain fuzzy. Combined with the very limited amount of information available it was hence impossible to clearly distinguish between executive and non-executive directors (see for similar problems Nollert, 2005).

Ownership

The variable we use to measure ownership concentration is constructed following the concentration measure developed by La Porta et al. (1998). This measure takes the sum of the percentage of cash-flow rights controlled by the three largest shareholders of a given company as an indicator for the dispersion of ownership. Since data were available, we used for Swiss firms voting rather than cash-flow rights. In order to refine this admittedly rather crude measure of shareholder-orientation, we also took into account the identity of the shareholders. The rationale for this is that the presence of an institutional investor, even when holding a large stake, does not automatically isolate management from market pressures and claims for shareholder-value-oriented management practices (Useem, 1993; Fiss and Zajac, 2004). Institutional investors may use their considerable shares in order to push through shareholder-oriented reform. Therefore, if one of the three largest shareholders of a firm was a pension fund, a mutual fund, an investment company or another institutional investor known for its shareholder-oriented approach, we excluded this shareholder and took the next largest shareholder to calculate the average concentration. We did not exclude family trusts and other investors who pursue strategies oriented towards non-pecuniary interests rather than solely seeking a maximization of shareholder value (see Gilson, 2005). Although in some cases this qualitative assessment of the character of owners is difficult to make, the resulting figures on ownership concentration are a better proxy for shareholder orientation than raw ownership data.

Unfortunately, valid and precise ownership data is scarce. For Swiss firms, ownership data for 1990 stems from a survey carried out by the Bank Julius Bär (1991) and includes 112 listed companies. Seventy of these 112 companies are in the interlock sample for 1990 as well. The ownership data for the 2000 sample is from the 'CD-ROM der Schweizer Wirtschaft'. In the Netherlands, no reliable ownership data exist prior to 1996. Ownership data are retrieved from the REACH database.

Empirical evidence

Disintegration of the corporate network

First of all, we need to assess our claim that networks of board interlocks in Switzerland and the Netherlands did indeed change during the 1990s. Table 1 reports the main features of the networks in the two countries. Although the two cases differ in the extent of changes, it shows that in both cases board interlocks dropped considerably in the period under study. In Switzerland, the number of board ties decreased by almost half between 1990 and 2000. A comparable decrease – given the shorter period – can also be found in the Netherlands. In some cases, firms share multiple directors. These kind of strong interlocks between firms have all but vanished, with only 20 left in Switzerland and 33 in the Netherlands by the turn of the century. The propensity of firms to interlock reduced over the last years of the 20th century. The trend towards disintegration of the networks is confirmed by a look at the directors network, that is, when we look at ties among individual directors created by common membership on a board rather than on ties between companies created by a shared director. In both countries, directors meet fewer colleagues on the boards they are member of. The average distance between directors has increased as well, leading to a decreasing cohesion of the directors' network.

Table 1 - Main features of the interlock networks.

Full table

As expected, the drop in board interlocks results in a disintegration of the network of interlocking directorates. As explained above, disintegration can take place either at the level of the network as a whole (connectivity) or at the level of the individual node (firm). To start with the latter, we find a remarkable difference in the evolution of the average number of interlocks per firm. In Switzerland in 1990, firms shared on average directors with 8.2 other firms. Ten years later, this number had almost halved to 4.4. In the Netherlands on the contrary, interlocks per firm remained fairly stable over time at relatively high levels (8.0 in 1996 and 7.0 in 2000). By the turn of the century, Dutch firms have interlocks with nearly as many firms as Swiss corporations had 10 years earlier. This means that at the level of the individual firm, the general pattern of interlocking directorates hardly changed in the Netherlands, while Swiss firms interlocked less in 2000 than in 1990.

Because both countries show an overall drop in the number of interlocks, the stable average number of interlocks of Dutch companies hints at the fact that the Dutch network disintegrated at the level of connectivity. Table 1 shows that this is indeed the case. Connectivity of the network, that is, the number of firms that are connected with each other through board interlocks, decreased in both countries. But in Switzerland, the strong drop in interlocks per firm was not linked to a marked decrease in connectivity. In 2000, 86.1% of all corporations were part of the giant component, compared to 90.7% 10 years before. Conversely, connectivity dropped dramatically in the Netherlands. By 2001, only 54.8% of all corporations of the top 250 largest firms are connected in the network of board interlocks, compared to almost 80% in 1996. In both countries the companies that are not part of the giant component remain unconnected, that is, the network does not disintegrate into smaller components of mutually connected boards, but in one giant component and many isolated companies.

Changes in the structure of the network do not necessarily imply a fundamental change in its function. Particularly, the network's capacity to convey information through interlocks may not fundamentally change despite a decreasing number of ties. This argument has been made in connection with the 'Small World' phenomenon. Many (social) networks are 'Small Worlds', characterized by high levels of local clustering in combination with small average distances between any two members of the network (see Watts, 1999; Uzzi et al., 2007). This is true, for instance, for ownership networks in Germany (Kogut and Walker, 2001) and interlocking directorates in the USA (Davis et al., 2003). The small world statistics of a network compare the actual distance and local clustering with a random network of the same number of points and ties. A network is considered to be a small world when the observed average path length is close to the average path length that we would expect in a random graph, whereas the actual clustering coefficient is considerably higher than clustering coefficient of a random graph. Table 2 shows the figures for the Swiss and Dutch networks of interlocking directorates. The figures show that both the Dutch and Swiss networks can be characterized as Small Worlds in both 1990 and in 2000.

Table 2 - 'Small World' statistics for the network of interlocking directorates.

Full table

Despite the fact that both countries show small world characteristics at both dates, there are notable differences as well. Whereas the average distance between two boards remained stable in the Netherlands, it increased considerably in Switzerland. By 2000, the average distance in Switzerland had increased from 2.6 to 3.8, which is a relatively high figure compared to the Netherlands, where the average distance between two companies was around three steps. In the Dutch network, in terms of average distance the drop in interlocks was compensated by a parallel decrease in connectivity. In the Swiss network however, high levels of connectivity led to an increasing average distance between the nodes.

Even more striking is the difference in the clustering coefficient, which measures the extent to which firms with whom a given firm shares board members, also share board members with each other. In the Netherlands, we find a decrease in clustering, which is consistent with the fact that the number of ties in the network has decreased (see Table 1). In Switzerland however, clustering goes up despite an overall increasing distance and a neat decrease in the number of interlocks. The Swiss network thus remains well connected. As a result, the small world statistic as reported in Table 2 shows that the small world character of the Swiss network increased between 1990 and 2000, but the small world character of the Dutch network decreased between 1996 and 2001.

On the basis of our findings so far we can conclude that the Swiss corporate network of board interlocks witnessed high levels of disintegration at the level of the individual firms, that is, each firm is much less well connected. This corroborates our earlier observation that the Swiss business system was early to adapt and move towards a more market-oriented architecture. The Dutch corporate network, on the other hand, remains quite intact during the last years of the 20th century, but with less firms connected. Thus, despite the strong international orientation of Dutch firms, including full-fletched bi-national structures of some of the most prominent firms, the network of interlocking directorates remains in place, albeit on a smaller scale, that is, including fewer firms. We now turn to the role of the financial institutions in this disintegration.

The end of bank centrality?

Table 3 shows that for the Netherlands, the drop in interlocks per firm is statistically insignificant, while for the Swiss firms the decrease is statistically significant for both financial and non-financial companies (Table 3). In 1990, Swiss financial companies shared board members with 9.7 firms on average, but only with 4.7 firms in 2000. At the same time, non-financial companies shared board members with 7.7 firms in 1990, but with only 4.3 in 2000. The higher propensity of financial companies to interlock, which has been confirmed by a number of empirical studies, appears to have practically disappeared in Switzerland by 2000.

Table 3 - Average number of interlocks (edge degree).

Full table

A look at the ranking of the most central firms confirms that financial companies are no longer more central than other companies. Table 4 lists the most central firms, with financial companies underlined. The Swiss ranking in 1990 is still dominated by the three large banks UBS, SBC and CS. Ten years later, CS remains at the top of the ranking, but with considerably fewer ties than in the 1990 ranking. The new UBS – which was the product of a merger between SBC and the old UBS in 1998 – has disappeared from the ranking, although 10 years earlier the two banks had a total of 67 ties. The traditional over-representation of financial companies among the most central firms has disappeared by 2000. If we look at the same ranking for the Netherlands, we find some financial companies among the most central firms for both years. While the most central firms are still financial companies in 2001, they do not stand out in terms of the number of interlocks. The most important Dutch bank – ABN AMRO – saw its interlocks halved within 5 years. For other Dutch banks among the most central companies, however, the decrease appears to be less pronounced than in the Swiss case. ING climbed from the 11th position to second position, which indicates an increasing integration in the network. Yet, while some financial companies do remain among the best-connected firms – or accede to this place during the 1990s – they do not stand out from other well-connected firms concerning the average number of interlocks.

Table 4 - Most central firms according to number of interlocks (edge degree).

Full table

The decreasing interlocks of banks is not necessarily the result of bankers retreating from supervisory boards, but could also be due to a decreasing number of outsiders on bank boards. Table 5 shows, however, that bankers were among the most central actors in the Swiss network in 1990, but no longer in 2000. In 1990, three executives of the three large banks were among the biggest linkers holding respectively eight, seven and six seats among the companies in our sample. In 2000, only one banker with six seats was among the big linkers. Moreover, this actor – Martin Ebner – can hardly be compared to a traditional banker. His bank – the BZ Bank – was in fact not a traditional universal bank but one of the most aggressive investment companies in Switzerland during the 1990s, which specialized in hostile takeover attempts on traditional Swiss companies and was famous for its resolute pro-shareholder approach. Ebner's board seats are hence mainly explained by his presence on the board of companies in which he held large shares of ownership. Often these seats were obtained against considerable opposition of the incumbent management and served mainly to impose his shareholder-value reform agenda (Becher, 1996). Traditional bankers, on the other hand, have completely disappeared from the ranking. More generally, during the period 1980 and 2000, the number of seats held by bank representatives of the three largest banks in Switzerland decreased from 13.4 seats per person to just 3.8 seats (Schnyder et al., 2005).

Table 5 - Directors with at least five positions.

Full table

In the Netherlands, it appears that financial executives play a less central role than in Switzerland. In fact, financial executives were among the first to reduce their involvement in the network of interlocking directorates (Heemskerk, 2007). Thus, in 1996, only one financial executive can be found among the best connected board members. By 2001, none of the best-connected directors held an executive position in a financial firm. It should be noted however, that all but one of the top 10 Dutch directors in 1996 and all but two in 2001 serve as supervisory director of a financial firm. Some of them are in fact former executives or CEOs from some of the largest Dutch banks, such as Kalff, Langman and Jacobs. This indicates that the remaining interlocks of financial firms is mainly related to well-connected supervisory directors, instead of well-connected bankers, and that financial directors – but not executives – still play a certain role in the network.

Shareholder orientation and network disintegration

If shareholder influence has negative repercussions for board interlocks, we would expect that firms which are more exposed to capital markets should be less central in the network. Most authors in the field of corporate governance consider that concentrated ownership hints at the control of companies by insiders, whereas diffused and fragmented ownership implies that market pressures on the company are strong and managers are therefore responsive to shareholder interests (see La Porta et al., 1998; Cheffins, 2001). Following this reasoning, we consider that ownership concentration is a useful – even though somewhat simplistic – proxy for the influence of shareholder interests on the company's management and for the extent to which shareholder-oriented governance practices affect the company in question. Those firms that are central in the network are expected to have relatively high levels of ownership concentration. Ownership structures in both Switzerland and the Netherlands are relatively simple compared to other countries. Complex pyramid structures are rare in both countries, mainly because there are many alternative controlling minority structures and other ways to shield firms from unwanted influence. Furthermore, equity ties between firms do not show a similar structure to board interlock ties as the ownership networks in both countries are considerable sparser than the interlock network. In other words, board interlocks did not primarily serve to monitor inter-corporate ownership.⁵

A first look at the figures on ownership concentration shows pronounced differences between the two countries. Whereas in Switzerland a clear decrease in concentration can be observed (from an average of 53.4% in 1990 to an average of 40.7% in 2000), in the Netherlands concentration has increased, albeit slightly, but at a much lower level (from 22.3% in 1996 to 24% in 2000). Thus, in the Netherlands, the influence of outside shareholders on the governance of companies can be expected to have remained approximately the same during the decade under analysis. It should be noted, however, that the figures for the Netherlands do not capture possible restrictions placed on exercising control rights, which means that the low levels of ownership concentration underestimates the influence of insiders as compared to Switzerland. Be that as it may, these findings are in line with our previous discussion on market orientation of the two countries, where we argued that Switzerland had transformed its corporate governance system to a larger extent towards a more shareholder-friendly system. The concentration measure confirms this trend.

When we look at the relationship between ownership concentration and board interlocks, the tables turn. For both dates and countries, the results are opposite of what we expected. We find that the more a company's ownership is dispersed, the more central it will be in the network. For Switzerland in 1990, we find a highly significant (0.01 level) Pearson correlation of - 0.526 between average number of interlocks and our concentration measure for individual companies (not reported in the tables). In 2000, this correlation has gone down to - 0.042 and is not significant anymore. For the Netherlands, we find similar results. In 1996, a significant (0.05 level) Pearson correlation of - 0.211 exists between degree and ownership concentration. Contrary to Switzerland, this correlation actually increases in both strength and significance (to - 0.388; 0.01 level) over time. These findings suggest that there is a group of firms which is both well connected in the board interlock network, and with relatively low levels of ownership concentration. Table 6 shows the level of concentration for Switzerland and the Netherlands divided into quartiles according to degree centrality. The most central companies in our samples (first quartile) have less concentrated ownership structures than less central companies. This confirms that companies that are exposed to stronger pressures from financial markets are not less central in the network of interlocking directorates.

Table 6 - Average capital ownership of three largest owners segmented by centrality.

Full table

Change and inertia in Switzerland and the Netherlands

Our findings show that, like in many other countries, the corporate networks of board interlocks in the Netherlands and Switzerland disintegrated over the last decade of the 20th century. This contradicts Nollert's (2005) study on the Swiss and Dutch cases, which concludes that no significant change in the network structure had taken place. While his study is based mainly on data for 1993/1994, our cross-sectional data allows us to show that Switzerland and the Netherlands indeed followed the same evolution as most countries. However, we also showed that although the two countries are similar in many respects, differences in the role and function of financial institutions, and in the extent of market-oriented reforms in corporate governance have considerably influenced the dynamics of change. This led to remarkable differences in the process of disintegration in Switzerland and the Netherlands, which contradicts accounts of efficiency-driven convergence of national business systems that would lead us to expect a uniform evolution in both cases. Our study shows that – even relatively small – differences in national business systems influence the process of change in important ways. Most notably, in the Netherlands disintegration takes place at the level of the network's inclusiveness or connectivity (many companies 'drop out'), while the number of interlocks per firm is not greatly affected by the changes. The disintegration of the Swiss network, conversely, takes place at the level of the local centrality of individual firms (companies remain connected, but have fewer neighbors). Thus, in Switzerland, connectivity remains at a relatively high level, but the number of interlocks per firm decreases noticeably. Concretely, this implies that the abandoning of corporate interlock ties was a more general phenomenon in Switzerland, where most companies remained connected, but more loosely. In the Netherlands, certain companies remained as well connected as before while other lost all their ties with other companies. This can be interpreted as a resistance of certain companies to change in the Netherlands.

These findings suggest that the relative openness of companies to market pressures in Switzerland, illustrated by the extensive corporate governance reforms, lead to disintegration at the level of the firm, rather than at the level of the network. In the Netherlands, where corporate governance reforms were few and market pressures remain low, average interlocks per firm remain high due to the continuing centrality of some firms. The reasons for the decrease in connectivity of the Dutch network remain clouded, and at this point we can only speculate on the reasons. One possibility is that many of the now isolated firms are in fact subsidiaries of foreign corporations, and as such not 'interested' in corporate board interlocks with the Dutch business community. This, however, remains in the realm of conjecture.

In 2001, Dutch financial companies remain central to the network as a whole (four out of the 10 most central companies are financial companies). In Switzerland only two financial companies are left in 2000 and Switzerland's largest bank (UBS) has disappeared from the ranking all together. This could be interpreted as a greater resilience of Dutch banks to change. Since Dutch banks did not reorient their strategies as much towards financial market-related activities as Swiss banks, this resilience is in line with our initial expectations.

We should note however that Dutch financial companies increasingly interlock through supervisory board members rather than members of the executive board. Executive ties have been interpreted as a form of control and non-executive ties more as a means of communication (Carroll and Fennema, 2002). Bank interlocks with other companies therefore appear to decrease – if not in number, at least in intensity – in both countries, albeit to different degrees.

The fact that the decrease in interlocking is stronger for individual Swiss companies than for Dutch companies confirms our expectation that companies which are more shareholder-oriented will be less inclined to interlock. However, if we use ownership concentration as an indicator for market exposure and look at the relationship between centrality and concentration within each country, we find that companies that are more dependent on financial markets are not less central than companies that are more isolated from financial market pressures. On the contrary, the most central firms in both countries and for all dates are characterized by more dispersed ownership than less central ones. Therefore, 'broader reliance on capital markets' (Davis and Mizruchi, 1999: 237) does not seem to explain the decrease in interlocking as many 'exposed' companies (measured by the degree of ownership concentration) do not appear to abandon their central position. Yet, ownership dispersion might after all not be the most apt measure of shareholder influence. Alternatively, it might be that board interlocks are not as negatively assessed by shareholders as is often argued. The latter view is supported by recent research: Rondøy et al. (2006) find no significant impact of board interlocks on share price for Norway, Sweden and Denmark, and Bøhren and Strøm (2006) even find a positive effect for Scandinavian companies.

A tentative explanation for the limited importance of shareholder orientation and capital market exposure for the position within the network is that the number of interlocks is not so much linked to shareholders' influence within a company, but is mainly a function of the prestige of a company. The largest listed companies are usually the most prestigious ones and also the ones where – due to the company's market value – no large owners exist. A board seat in such a large company is sought-after and members of such a board can in turn be expected to be invited to other companies, thus increasing a prestigious company's interlocks. The predominance of such 'sociological reasons' for interlocks may be stronger in small states like Switzerland and the Netherlands, where the cohesion of the business elite can be expected to be particularly strong due to its limited size. In that sense, the characteristics of small states may account for a certain resilience of board-level interlocks despite an increasing exposure to financial market pressure. In order to confirm or reject this idea further comparative research including large and small countries would be needed.

Corporate governance reform is only one of many moving parts in a complex realignment process of national business systems. Comparing the Netherlands and Switzerland we showed how openness towards international financial markets affects strategies of interlocking directorates through changes in corporate practices. This openness however is not the result of changes in the legal framework alone. The re-orientation of the Swiss business system towards financial mechanisms is mainly the result of firm-level changes, rather than legal change (Schnyder, 2007). This means that changes in the network are affected by the interaction between specific corporate strategies, and individual actor's decisions as well as legal innovations. If we are interested in understanding the reasons that underlie the changes we see in contemporary capitalist economies, it is crucial to recognize the roles that different actors play within the realignment process.

By focusing on different actors in the network (financial companies, exposed companies), we attempted to remedy the limitations of the purely structural analysis. Our study shows indeed that changes in the network structure can be understood by looking at which actors 'drop out' and at explanations for their changing attitude towards interlocks. Thus, a changing relationship with financial markets appears to have an important impact at least on financial companies' position in the network. A refined study of non financial companies' dependence on financial markets would be needed in order to confirm that the same is the case for other categories of companies.

An increased attention to the 'actors of change' within the company network is also needed in order to better understand the 'function' of the network. In fact, it might be that the changes in the structure of the network, which we found, do not imply a change in its function. The small world statistics indicate that both Switzerland and the Netherlands are still characterized by an interlock network in which information can flow rapidly. Conversely, one could also argue that although many ties persist, the way in which actors use these ties may still have changed substantially. More qualitative investigation into how board overlaps are used by actors would allow us to shed light on these issues. In order to explore more thoroughly the reasons for the disintegration of interlock networks and to further push an actor-centered investigation into network dynamics, more detailed information on network ties and on the companies which are part of the network is needed. Especially for countries with a tradition of secrecy concerning business – of which Switzerland is certainly the prime example – data gathering proves to be the main limitation to the possibility of testing hypotheses concerning firm-level reasons for change. We therefore consider that it must be a prime objective in the study of interlocking directorates to improve the quality of comparable data sets in order to be able to deepen our understanding of how changes in national networks can be interpreted.

Notes

¹ See for the case of Germany Beyer and Höpner (2003); for the case of Norway Grønmo and Løyning (2002); for the Netherlands Heemskerk (2007); for Spain Rodriguez (2002); for Switzerland Schnyder et al. (2005).

² See for Switzerland Bonoli and Mach (2001), Mach (2006), Schnyder (2007); for the Netherlands Poutsma and Braam (2005).

³ Figures are for 2006. Source: Centraal Bureau voor de Statistiek (CBS), De Nederlandsche Bank (DNB), Eurostat, International Monetary Fund's World Economic Outlook.

⁴ It should be noted that Switzerland cannot be considered as a control case for non-EU-membership however as its domestic policies are clearly influenced by EU policy (Haverland, 2006).

⁵ In the Netherlands in 1996, only 52 interlock ties were doubled by ownership ties. Of these 52, 14 where created by an executive director and might reflect a controlling board interlock. Five years later, only 17 'double' ties remained. By that time, only five of these ties were created by executive. These few 'double' ties that do exist are mainly between financial firms, and not between financial and non financial companies. Since the ownership network is very sparse, we do not investigate network relations of shared ownership. We are primarily interested in the composition of the circle of owners of firms and in the dispersion of ownership.

References

1. Aggarwal, Reena, Iisil Erel, Rene Stulz and Williamson Rohan, 2007, Do U.S. firms have the best corporate governance? A cross-country examination of the relation between corporate governance and shareholder wealth. ECGI Finance Working Paper, European Corporate Governance Institute (ECGI), p: 41.
2. Bank Julius Bär, 1991, "Besitzverhältnisse von Schweizer Aktien". Marktübersicht Schweiz: 1–43.
3. Bebchuk, Lucian A., Reinier H., Kraakman and Georg G. Triantis, 2000, "Stock pyramids, cross-ownership and dual class equity. The mechanisms and agency costs of separating control from cash-flow rights". In R. Morck (ed.) Concentrated corporate ownership. Chicago: University of Chicago Press, pp: 295–314.
4. Becher, Jörg, 1996, Das schnelle Geld: Martin Ebners Weg zur Macht. Zürich: ABC-Verlag.
5. Berger, Suzanne and Ronald Dore (eds), 1996, National diversity and global capitalism. Ithaca, NY: Cornell University Press.
6. Beyer, Jürgen, 2002, Deutschland AG a.D: Deutsche Bank, Allianz und das Verflechtungszentrum grosser deutscher Unternehmen. MPIfG Working Paper 02.
7. Beyer, Jürgen and Martin Höpner, 2003, "The disintegration of organised capitalism: German corporate governance in the 1990s". West European Politics, 26: 179–198. | Article |
8. Bøhren, Øyvind and Reidar Øystein Strøm, 2006, Aligned, informed and decisive: Characteristics of value-creating boards. Norwegian School of Management Working Paper, Oslo.
9. Bonoli, Giuliano and André Mach, 2001, "The New Swiss employment puzzle". Swiss Political Science Review, 7: 81–94.
10. Borgatti, Steve P., Martin G. Everett and Linton C. Freeman, 2002, Ucinet for windows: Software for social network analysis. Version 6.81. Harvard, MA: Analytic Technologies.
11. Carroll, William K. and Meindert Fennema, 2002, "Is there a transnational business community?". International Sociology, 17: 393–419. | Article |
12. Cheffins, Brian R., 2001, "Does law matter? The separation of ownership and control in the United Kingdom". Journal of Legal Studies, 30: 459–484. | Article |
13. Commissie Corporate Governance, 1997, Aanbevelingen inzake Corporate Governance in Nederland. Amsterdam: Commissie Corporate Governance.
14. Commissie Corporate Governance, 2003, De Nederlandse Corporate Governance Code. Beginselen van deugdelijk ondernemingsbestuur en best practice bepalingen. The Hague: Ministerie van Financiën.
15. David, Thomas and André Mach, 2004, "The specificity of corporate governance in small states: Institutionalisation and questioning of ownership restrictions in Switzerland and Sweden". In R. Aguilera and M. Federowicz (eds) Corporate governance in a changing economic and political environment. Trajectories of institutional change on the European Continent. London: Palgrave, pp: 220–246.
16. Davis, Gerald F. and Mark S. Mizruchi, 1999, "The money center cannot hold: Commercial banks in the U.S. system of corporate governance". Administrative Science Quarterly, 44: 215–239. | Article |
17. Davis, Gerald F., Mina Yoo and Wayne E. Baker, 2003, "The small world of the American corporate elite, 1991–2001". Strategic Organization, 1: 301–326. | Article |
18. De Jong, Abe, Rezaul Kabir, Teye Marra and Ailsa Röell, 2001, "Ownership and control in the Netherlands". In F. Barca and M. Becht (eds) The control of corporate Europe. Oxford: Oxford University Press.
19. De Jong, Abe and Ailsa Röell, 2005, "Financing and control in the Netherlands: An historical perspective". In R.K. Morck (ed.) A history of corporate governance around the world: Family business groups to professional managers, National Bureau of Economic Research Conference Report. Chicago: University of Chicago Press.
20. Dore, Ronald, 2002, "Debate: Stock market capitalism vs welfare capitalism". New Political Economy, 7: 115–127. | Article |
21. Fennema, Meindert, 1982, International networks of banks and industry. The Hague: Boston Hingham, MA: M. Nijhoff.
22. Fiss, Peer C. and Edward J. Zajac, 2004, "The diffusion of ideas over contested terrain: The (Non)adoption of a shareholder value orientation among German firms". Administrative Science Quarterly, 49: 501–535.
23. Gilson, Ronald J., 2005, Controlling shareholders and corporate governance: Complicating the comparative taxonomy. ECGI – Law Working Paper: 53.
24. Gourevitch, Peter A. and James Shinn, 2005, Political power & corporate control. The new global politics of corporate governance. Princeton: Princeton University Press.
25. Grønmo, Sigmund and Trond Løyning, 2002, The strength of weak coordination. Networks of interlocking directorates in Norway 1970–1995. Paper presented at the XV World Congress of Sociology. Brisbane, Australia, 7–13 July.
26. Hansmann, Henry and Reinier Kraakman, 2004, "The end of history for corporate law". In J.N. Gordon and M.J. Roe (eds) Convergence and persistence in corporate governance. Cambridge: Cambridge University Press, pp: 33–68.
27. Haverland, Markus, 2006, "Does the EU cause domestic developments? Improving case selection in Europeanisation research". West European Politics, 29: 134–146. | Article |
28. Heemskerk, Eelke M., 2007, Decline of the corporate community. Network dynamics of the Dutch business elite. Amsterdam: Amsterdam University Press.
29. Heinze, Thomas, 2004, "Dynamics in the German system of corporate governance? Empirical findings regarding interlocking directorates". Economy and Society, 33: 218–238. | Article |
30. Helmers, Hendrika M., Robert J. Mokken, Rene C. Plijter and Frans N. Stokman, 1975, Graven naar Macht. Amsterdam: Van Gennep.
31. Höpner, Martin, 2003, "European corporate governance reform and the German Party Paradox". Max Planck Institute for the Study of Societies Working Paper, No 04/03. Cologne, pp: 1–46.
32. Katzenstein, Peter J., 1985, Small states in world markets: Industrial policy in Europe. Ithaca, NY: Cornell University Press.
33. Kogut, Bruce and Gordon Walker, 2001, "The small world of Germany and the durability of national networks". American Sociological Review, 66: 317–335. | Article | ISI |
34. La Porta, Rafael, Florencio Lopez-de-Silanes, Andrei Shleifer and Robert W. Vishny, 1998, "Law and finance". Journal of Political Economy, 106: 1113–1155. | Article | ISI |
35. Mach, André, 2006, La Suisse entre internationalization et changements politiques internes. La législation sur les cartels et relations industrielles dans les années 1990, Vol. 6. Basel: Verlag Ruegger.
36. Mizruchi, Mark S., 1996, "What do interlocks do? An analysis, critique, and assessment of research on interlocking directorates". Annual Review of Sociology, 22: 271–298. | Article | ISI |
37. Monks, Robert A.G. and Nell Minow, 1995, Corporate governance. Cambridge: Blackwell Publishers.
38. NCGS, 2002, Corporate governance in Nederland 2002. De stand van zaken. Amsterdam: Nederlandse Corporate Governance Stichting.
39. Newman, Mark E.J., Steven H. Strogatz and Duncan J. Watts, 2001, "Random graphs with arbitrary degree distributions and their applications". Physical Review E, 64: 026118. | Article | ChemPort |
40. Nollert, Michael, 2005, Unternehmensverflechtungen in Westeuropa. Nationale und transnationale Netzwerke von Unternehmen, Aufsichtsräten und Managern. Münster, Hamburg, Berlin, Wien, London: Lit Verlag.
41. Poutsma, Erik and Geert Braam, 2005, "Corporate governance and labour management in the Netherlands: Getting the best of both worlds?". In H. Gospel and A. Pendleton (eds) Corporate governance and labour management. An international comparison. Oxford: Oxford University Press, pp: 148–172.
42. Powell, Walter, 1990, "Neither market nor hierarchy: Network forms of organization". Research in Organizational Behavior, 12: 295–336. | ISI |
43. Rajan, Raghuram G. and Luigi Zingales, 2003, "The great reversals: The politics of financial development in the twentieth century". Journal of Financial Economics, 69: 5–50. | Article | ISI |
44. Rodriguez, Josep A., 2002, Revisiting the power: Changes in the Spanish structure of economic power (1991–2000). Paper prepared for SUNBELT XXII International Social Network Conference. New Orleans, 13–17 February.
45. Rondøy, Trond, Steen Thomsen and Lars Oxelheim, 2006, A Nordic perspective on corporate board diversity. Oslo: Adger University College.
46. Ruigrok, Winfried and Rob Van Tulder, 1995, The logic of international restructuring. London, New York: Routledge.
47. Schaub, Vera, 1992, Konzernpolitik im Schweizer Bankbereich. Eine Darstellung der Determinanten udn Ausgestaltungsmöglichkeiten unter Berücksichtigugn ausgewählter Schweizer Bankkonzerne. Bern: Haupt.
48. Schnyder, Gerhard, 2007, Comparing corporate governance reforms: Law, politics and the social organization of business in the case of Switzerland, 1965–2005. Unpublished doctoral dissertation, University of Lausanne, Lausanne.
49. Schnyder, Gerhard, Martin Lüpold, André Mach and Thomas David, 2005, The rise and decline of the Swiss Company Network during the 20th century. Travaux de Science Politique no. 22, University of Lausanne, Lausanne.
50. Stokman, Frans N., Rolf Ziegler and John Scott, 1985, Networks of corporate power: A comparative analysis of ten countries. Cambridge, Oxford, New York: Polity Press; B. Blackwell.
51. Useem, Michael, 1993, Executive defense: Shareholder power and corporate reorganisation. Cambridge, MA: Harvard University Press.
52. Uzzi, Brian, Luis A. Amaral and Felix Reed-Tsochas, 2007, "Small-world networks and management science: A review". European Management Review, 4: 77–91. | Article |
53. Wasserman, Stanley and Katherine Faust, 1994, Social network analysis methods and applications. Cambridge, New York: Cambridge University Press.
54. Watts, Duncan J., 1999, Small worlds. The dynamics of networks between order and randomness. Cambridge: Cambridge University Press.
55. Windolf, Paul, 2002, Corporate Networks in Europe and the United States. Oxford: Oxford University Press.

Acknowledgements

We are grateful to Bruce Kogut, Mara Yerkes, the participants in the Small Worlds of Corporate Networks group, and two anonymous reviewers for very helpful comments. Swiss data stems from a research project on the history of corporate governance in Switzerland, funded by the Swiss National Science Foundation (grant no.: 1214-068112.02/1), directed by Thomas David and André Mach, University of Lausanne.