Netherlands, Abe de Jong, industrial firms, pension funds, Ailsa R�ell, dual-class shares, insurance companies, stock exchange, corporate control, The Netherlands, interlock, supervisory board, equity stakes, control structure, the twentieth century, nineteenth century, Dutch banks, managerial control, Dutch industry, management board, board members, listed companies, shareholders, shareholder, corporate law, total assets, ordinary shares, interlocks, Sweden, Dutch corporation, VOC, corporate laws, Ailsa Roell, corporate governance, industrial development, stock market, French occupation, corporate financing, corporate ownership, banking system, preference shares, RvC, family firms, industrial firm, preferred shares, certificates, bearer certificates, investment banking, insurance company, Nederlandsche Bank, the stock exchange, defensive measures, universal banking
This PDF is a selection from a published volume from the National Bureau of Economic Research Volume Title: A History of Corporate Governance around the World: Family Business Groups to professional managers
Volume Author/Editor: Randall K. Morck, editor Volume Publisher: University of Chicago Press Volume ISBN: 0-226-53680-7 Volume URL: http://www.nber.org/books/morc05-1 Conference Date: June 21-22, 2003 Publication Date: November 2005 Title: Financing and Control in The Netherlands: A Historical Perspective Author: Abe de Jong, Ailsa Roell URL: http://www.nber.org/chapters/c10275
8 Financing and Control in The Netherlands A Historical Perspective Abe de Jong and Ailsa Rцell L'ambition de la rйpublique est de s'enrichir et non de s'agrandir. --Denis Diderot, Voyage en Hollande, 1780 Introduction The goal of this paper is to place the current structure of Dutch owner- ship and control in a historical perspective. The historical development
of Dutch financial markets and institutions is somewhat idiosyncratic. It mixes elements such as a stock exchange culture dating back to the Dutch golden age of seaborne trading dominance, a legal system handed down from a brief period of French occupation, and strong influences from neighboring Germany as well as England and the United States. The paper first sets out, in section 8.1, to describe in brief the historical development of Dutch industrial finance. The remainder of the paper then turns to a comparative analysis of Dutch listed firms over the course of the twentieth century by focusing on three years spaced at thirty-five-year intervals: 1923, 1958, and 1993. A general description of the data and their sources is given in section 8.2, focusing on a wide array of financial characteristics of the firms. This is followed in section 8.3 by a closer analysis of corporate control mechanisms and, in particular, shareholder rights and defenses against hostile takeovers. Networks of influence are the focus of section 8.4, and the main themes discussed in that section are the nature and composition of the supervisory and management boards: the degree to which there are interlocking directorships with banks and other industrial firms, Abe de Jong is an associate professor in the Department of Financial Management at Erasmus University, Rotterdam. Ailsa Rцell is professor of finance at Tilburg University and a senior research scholar at Princeton University. We would like to thank Randall Morck, Jan Luiten van Zanden, and an anonymous referee for helpful comments. We thank Henry van Beusichem and Esther Koomen for excellent research assistance. 467
468 Abe de Jong and Ailsa Rцell and the presence of identifiable founding-family members
on the board. Section 8.5 concludes. 8.1 Historical Overview 8.1.1 General Introduction The Dutch have some claim to a pioneering role in stock exchange capitalism. The first shareholdings in a Dutch corporation came into being in 1602, when the Vereenigde Oostindische Compagnie (VOC), the first great limited-liability joint stock company in the world, was founded. The initial investors were, in 1602, unaware of their destiny: ostensibly, they were contributing money to a limited-term partnership that would send out a series of merchant ships to the East Indies, with a liquidating dividend promised at the end of twenty years. To the investors' dismay (and despite their vociferous protests), in 1622 the company's directors (who reported to the government rather than to the shareholders) decided to prolong the company's charter, thus shelving the liquidation and keeping this astonishingly lucrative1 enterprise going for many years. By the middle of the seventeenth century
the Netherlands had developed an active shareholding culture, with speculation in VOC shares and even derivatives trading a widespread popular pursuit. In the eighteenth century, the fortunes of the Dutch East India trade declined, and the VOC finally went under in 1799. Even so, the wealth amassed by the Dutch during the Golden Age was still largely undissipated and primarily invested in a wide range of international government securities. A spate of defaults, notably by the French government, reduced this wealth and seriously undermined confidence in securities investment, but even in the nineteenth century there were still many wealthy rentier families whose riches were primarily held in the form of securities.2 In the early nineteenth century the Dutch nation emerged from the French occupation of 17951813; it assumed its present geographical contours with the separation of Belgium from the Netherlands in 1830. The first half of the nineteenth century was a period of continued economic stagnation: Dutch investment in infrastructure and the new steam-driven manufacturing technologies was minimal, and the country's industrial development
lagged far, far behind that of Belgium, Germany, France, and, 1. By the time of its last dividend in 1782, an initial investment of f 100 in the VOC would have yielded f 360,033.33 in payouts (Steensgaard 1982). Steensgaard gives an insightful discussion of how the novel corporate form of the VOC made this enduring profitability possible--for example, by facilitating long-term investments in the military protection of trading routes and monopolies. 2. The rentier class were popularly referred to as coupon-cutters: "ces rentiers hollandais que le peuple appelle ironiquement coupon-knippers, parce qu'ils n'ont rien а faire, sauf а detacher les coupons semestriels de leurs fonds publics" (de Laveleye 1864, p. 329).
Financing and Control in The Netherlands 469 of course, England. This period of retarded growth has been studied intensively by economic historians, and the consensus now seems to be that it cannot be attributed to a shortage of capital or to Dutch investors' supposed preference for foreign investments above domestic industry. Other factors seem more likely culprits. One was the disarray of government finances: the new Kingdom of the Netherlands inherited from the French a crushing debt burden of 420 percent of net national income, with concomitantly high interest rates on government paper; the situation was not brought under control until around 1850 (see Jonker 1996). Another was the need to redefine the traditional division of labor
within the low countries: the southern provinces, now Belgium, had traditionally specialized in manufacturing while the North focused on commerce. Thus, there was no strong manufacturing base to build on. Then there were the steep transport costs related to the extra cost of providing a proper infrastructure, with adequate drainage and flood defenses, in such low-lying and waterlogged territory; and various other factors such as the high cost of raw materials, and the high wage levels and the poor education of the citizenry. Industrial development started coming to life in the second half of the nineteenth century, with new shareholder capital raised for a number of enterprises such as railway construction, albeit rather laboriously, buffeted by the vicissitudes of international political development
s and the business cycle. The main source of capital for industry during that period seems to have been retained earnings, supplemented with contributions by members of the founding families and closely connected wealthy individuals. Interestingly, the rather meager contribution of publicly raised equity was not offset by long-term bank loan finance: such financing was also very scarce throughout the country's industrialization. The long period of stagnation of the eighteenth and early nineteenth centuries, and the short period of French hegemony, create a natural break in capital market traditions and institutions. Only in the late nineteenth century did substantive modern industrial development get off the ground. Although several institutions were already present in the Dutch Golden Age, we take this revival as a starting point for our analysis. We turn now to a few specific themes that are of central importance for the genesis of today's landscape of corporate finance and control: first, the evolution of the Dutch framework of company law, and second, the role of the stock exchange, banks, and private financing in providing capital for industry. 8.1.2 Evolution of the Public Limited Company Public shareholder finance requires an appropriate legal basis, and at the start of the seventeenth century, there was little in the way of precedent to draw upon. The earliest Dutch joint-stock enterprises of the seventeenth century (in addition to the VOC, several other trading companies and a number of insurance companies emerged) were explicitly created to further
470 Abe de Jong and Ailsa Rцell the public interest, with trading monopolies granted by the government and control exercised by public appointees. Almost from the start, Dutch shareholding culture was embroiled in a series of corporate governance skirmishes, as conflicts of interest became apparent and their resolution was hammered out.3 The legal form of the Dutch corporation evolved over time from the early days of the VOC.4 Around 1720, the legal status of the limited company or naamloze vennootschap (NV for short) was largely remodeled along the precedent set by English company law; and the setting up of companies whose primary purpose was private profit, rather than the service of the public interest, became the norm. By and large, the companies set up at the time in Holland were reputable, unlike some of their English counterparts spawned by the prevailing stock market
bubble. One Dutch innovation of the time was the Amsterdam broker Abraham van Ketwich's creation, in 1774, of the world's first investment trust:5 the Negotiatie onder de Zinspreuk "Eendraagt Maakt Magt" (Investment under the Motto "Unity Is Strength"). The subsequent collapse of company profits and share prices led to a slowdown in the creation of new limited companies (and of new investment trusts: after 1779, there was a ninety-year hiatus). Following the French occupation of the turn of the nineteenth century, Dutch civil law was codified along lines closely following the French civil code of 1804. The Wetboek van Koophandel (commercial code) of 1838 set the legal parameters for public limited companies. From the start, it was felt to be inadequate to its purpose. At first, there was particular resistance to the "foreign" notion that the founding of a public limited company 3. These disputes bear an amusing resemblance to the issues that are still being debated today. Those initial VOC shareholders who were not actively involved in the running of the company--known as the long-suffering or dolerende shareholders--had many reasons for complaint. Their objections are vividly preserved in the company's archives. The initial complaints centered around payout policy (when the interim dividend payouts were passed or fell short of the amount stipulated in the company's charter, and when the promised liquidating dividend of 1622 was shelved) and the murkiness of the company's accounts: letters and pamphlets calling for financial disclosure and speaking of abuses and damaging disorders were circulated, but they were ignored by the Heeren XVII--the "seventeen gentlemen" directors--until the strength of shareholder outrage prompted the government to require full and open accounts for 1622. Even so, a groundswell of protest about inadequate financial disclosure continued for decades after. Later documents regulate the conflation of management's personal interests with those of the company proper (there were numerous company directives reminding its employees that they were under no circumstances allowed to transport or trade goods on own account, and the Heeren XVII brought out a report in 1741 on abuses by company management at home and abroad). There is even the seventeenth-century equivalent of the corporate jet (directors' travel on company business by inland yacht, and the declaration of travel expenses, was carefully regulated--for example, in a document dating from 1698). See Frentrop (2003) for an English-language history of Dutch corporate governance. 4. Our description is based on the introductory chapter of van der Heijden's (1992) handbook of Dutch company law. 5. Albeit one containing a somewhat curious lottery element, intended to stimulate speculative interest.
Financing and Control in The Netherlands 471 Fig. 8.1 Caricature of a Dutch supervisory board, by J. Braakensiek, 1898 (currently in the Gemeente Archief Amsterdam) Commissioner A (to his neighbor): "Is everything in order over there, with that safe. . . . ?" Commissioner B: "Now listen here, that is up to the management. We have our hands full supervising the company; if we have to start looking after the safe as well . . ." would require royal approval, even if the conditions that would ensure such approval were set down in the law. Camfferman (2000) mentions that, in particular, the relevant government ministry's practice of asking that financial accounts be sent in on an annual basis was very unpopular. The law also failed to address a number of issues such as the personal liability of founders, issuers, management, and directors; the shareholders' obligations with regard to paying in their capital,6 and nonmonetary contributions to the company. The last quarter of the nineteenth century saw a spate of company bankruptcies, some of which involved the outright looting of company funds. The weaknesses of corporate governance safeguards in protecting investors, and in particular the inadequacies of monitoring by boards of directors, was already an open matter of public concern, as evidenced by figure 8.1, an 1898 cartoon depicting a supervisory board in action. After a very long period of public debate, with legislative proposals submitted, withdrawn, and resubmitted regularly from 1871 onward, a new, more comprehensive and flexible company law was finally enacted in 1928. Preventive government scrutiny was retained: the minister of justice would vet the proposed charter of an NV before it could be registered with the 6. For example, in the case of the NCS railway initial public offering (IPO) in 1860, described in the appendix 1.
472 Abe de Jong and Ailsa Rцell chamber of commerce and thereby officially founded. The new regime was based on four principles (Van der Heijden 1992, para. 28, p. 19): 1. Preventive government monitoring, including the possibility of judicial suppression 2. Transparency of the internal organization and division of powers (including financial reporting) 3. Protection of the capital against excessive payouts to shareholders 4. Strengthened liability of founders, management, and directors One of the most controversial issues was the openness requirement, in particular the obligation to publish full annual accounts (a balance sheet and a profit and loss statement) open to the general public. Traditionally many companies had kept this information private within a small inner circle--for example, by allowing only a small number of shareholder delegates to look at the accounts. Almost immediately, a commission was set up to examine if the obligation to publish accounts could be weakened. The law was criticized for not distinguishing between large, open companies that placed securities with the general public and closed or family companies that did not. Others countered that limited liability requires, in principle, openness of the financial situation of both kinds of NV. Other objections concerned the law's restrictions on oligarchic clauses, the rights of redress awarded to minority shareholders, and the strengthened liability of management and directors. Company law was again fully revised in 197071. The main impetus was twofold. Firstly, there was the need to adjust to the European Economic Community's First Directive on Company Law of 1968. The biggest change in this regard was to create a new, separate type of limited company, following the law of surrounding countries (Germany, France, and Belgium): the besloten vennootschap (BV) or closed company, in addition to the traditional NV. The impact of this change was immediate. The great majority of smaller companies converted from NV to BV, primarily as a result of the lower level of financial disclosure required of the latter (NVs were now required to make their annual accounts readily available to the public at large by depositing them at the offices of the handelsregister). In addition, new arrangements were set in motion for the protection of minority shareholders (through enquкterecht: the right to ask for a judicial enquiry under certain conditions). The second force driving change was the wish to increase the influence of employees. Dutch attitudes to the role of corporations had evolved over the course of the twentieth century. In the beginning of the century, corporations were seen as vehicles for shareholder wealth creation. Over the course of the century, firms became seen as more independent entities oriented toward continuity, stability, and the interests of multiple stakehold-
Financing and Control in The Netherlands 473 ers, as expressed in a salient Hoge Raad (Supreme Court
) decision of 1949. It is perhaps the relative homogeneity of the Dutch population that has fostered a sense of solidarity, expressed in a preference for consensus decisionmaking and a generous welfare system. The corporatist model of centralized, consensual economic decision-making, known as the poldermodel, was very successful in the reconstruction of the Dutch economy after World War II. In particular, centralized collective bargaining
made possible a lengthy period of wage restraint that contributed substantially to economic growth
. In return, employee representation in decisions regarding job security
and employment is considered appropriate. And indeed, any corporate restructuring that involves the loss of jobs imposes a significant cost on the public purse in the form of unemployment and/or disability pay. This means that corporate decision making has a direct public interest dimension. Not surprisingly, the stakeholder view of corporate governance, which sees shareholders as just one of many interested parties entitled to a say in decision making, dominates Dutch Public Opinion
. The structuurregime or "structured regime," introduced in 1971, was designed to increase worker participation by imposing a carefully defined control structure on all larger firms (roughly speaking, those with at least 100 employees). Such firms must set up an ondernemingsraad (OR) or company council, a body created to represent and consult the views of employees.7 These and other large firms (those with capital and reserves of at least f 25 million) are also obliged to set up a supervisory board (raad van commissarissen, RvC) with some powers that might otherwise be held by the shareholders' meeting. Such a board appoints new members itself by co-optation (unless the shareholders' meeting or council objects), and the statutes may determine that one or more are to be government appointees. The board supervises important managerial decisions, appoints and dismisses the management board (raad van bestuur, RvB), and establishes and approves the yearly accounts (De Jong et al. 2004). A perhaps unintended side effect of the structuurregime is that, because it gives shareholders almost no say in the appointment or removal of supervisory board members and management, it protects entrenched management to an excessive degree. The co-optation system is currently the topic of intense public debate and is unlikely to survive in its current form. The most recent developments in the Netherlands are two best practices codes for publicly listed firms. The first code is a product of the Peters Committee, named after former Aegon chief executive officer (CEO) Jaap Pe- 7. It has a right to relevant information, a right to advise on major decisions (e.g., transfers of ownership, relocation, and important investments); it can delay decisions it disagrees with for one month and appeal to the ondernemingskamer (company chamber) of the Amsterdam Court. Its permission is required for changes to social arrangements (pensions, working hours, wages, safety rules), and if it disagrees the employer must obtain a local judge's decision to go ahead.
474 Abe de Jong and Ailsa Rцell ters. This code contains forty recommendations, about the role of management, supervisory boards, and, most important, a reconsideration of the role of capital in governance. As thirty-nine (out of the forty) recommendations did not involve legal changes, the code's implementation draws on self-regulation. De Jong et al. (2004) show that this effort failed, as no observable changes were present and stock market reactions, if present, were negative. After the irregularities with Ahold an initiative was taken to restore investors' confidence in the Dutch market. In March 2003 a committee chaired by Morris Tabaksblat, former CEO of Unilever, started a new code and had already released the final draft in December 2003. Following the successful U.K. codes, the comply-or-complain principle is introduced, forcing firms to explain to shareholders any deviations from the best practice. Although the contents of the code largely overlap with Peters's ideas, the enforcement is more promising. 8.1.3 Equity Financing and the Role of the Stock Market in Industrial Finance The Amsterdam stock exchange was a sophisticated and active market throughout the nineteenth century. The prolongatie system funneled large amounts of savings to the market. The market was overcrowded, open and competitive: the principle of unrestricted public access was carefully upheld by the city authorities, and premises were shared with commodities trading. However, the stock exchange did not initially play much of a direct role in the financing of industry. The bulk of the official list seems to have been made up of foreign state loans, American railway stocks, American industrial shares, and colonial securities. The first date at which domestic industrial stock was officially listed on the Amsterdam stock exchange is generally reported to be a brewery listing in 1889, though Jonker (1996) suggests this date is misleading; four industrial issues (from a sugar refinery, a shipyard, and an engineering firm) were already quoted in the early 1880s. In any case a listing meant little before 1903, when listing requirements and a vetting process by the Vereeniging voor Effectenhandel (set up in 1876 to oversee the market and instill investor confidence) were formalized. Meanwhile there was a large and active unlisted securities market on which domestic securities were both auctioned and directly placed; an example of a prime unlisted stock traded there during the last decades of the nineteenth century is Heineken. Shares were often initially privately placed, and Jonker (1996) cautions that a lack of domestic industrial stock exchange listings should not be interpreted as a definitive indicator of investor disinterest. A number of NVs set up in the 1840s and 1850s found ready backers; they did not seek a listing until the end of the century. By 193739, private placements still encompassed 16.6 percent of bond issues and 4.8 percent of equity issues, and private "underhand" loans remained
Financing and Control in The Netherlands 475 important right up until the eve of World War II: in 1938, institutional investors' portfolios still contained equal amounts of underhand loans and listed securities (Renooij 1951, pp. 186 and 190). Clearly, then, the Amsterdam stock exchange was not the sole venue for primary issues or for secondary trading. The dearth of domestic industrial listings cannot be interpreted as a sign of structural impediments to equity financing. Van Zanden (1987, 1998) points out that external finance, albeit not obtained from the general public, played a major part in the industrialization of Amsterdam. Initially, money for capital-intensive new ventures would be supplied by the city's traditional trading elite. For example, merchants set up two companies for steamship transport and shipbuilding in hopes of stimulating trade. Similarly, rich and successful entrepreneurial dynasties would move into related industries: for example, the profits from sugar refining were plowed back into beer brewing and flour milling concerns. Meanwhile, the government and King William I at times provided crucial credit lines. And in 1883 Amsterdam's financial elite contributed a capital of f 0.5 million for a banking venture, the Finantieele Maatschappij voor Nijverheidsondernemingen, whose explicit purpose was to provide finance for industry in the form of credit, in anticipation of repayment when a public share issue was completed. Still, it is fair to say that infusions from a network of family, friends, and business associates, complemented by retained earnings, were, in the Netherlands as in most other countries, the dominant source of risk capital for much of industry in the late nineteenth century. For example, the textile industry developing in the East and South of the country was almost exclusively financed in this way. The exception, rather than the rule, were large, capital-intensive infrastructure projects like railways, which typically relied on an initial primary issue of shares to the general public, sometimes combined with some form of limited government support, to get off the ground. The appendix describes the initial share ownership structure following four nineteenth-century railway flotations. 8.1.4 The Role of Banks A surprising feature of Dutch financial history (particularly when contrasted with the emergence of powerful universal banks in Germany in the late nineteenth century) is the limited role played by banks in the financing of industrial growth, not just in the early period of industrialization of the late nineteenth century but well into the twentieth century. Dutch economic historians attribute the patchy record of late nineteenth-century banking initiatives--banks were set up, but many failed, and the industry remained exceedingly fragmented well into the twentieth century--to a number of causes. One major cause was the dominance of the prolongatie system of financing, which flourished throughout the late nineteenth and early twentieth
476 Abe de Jong and Ailsa Rцell centuries. Prolongatie refers to short-term callable margin loans, on the face of it a rather unlikely source of industrial finance. As a legacy from the successes of the Golden Age, the nineteenth-century Netherlands still had a strong stock market culture and a well-developed network of local agents (notaries, lawyers, and brokers) who would collect savings from wealthy individuals and channel them to the stock exchange. Much of the money was not invested in securities directly but made available to firms or other investors in the form of short-term margin loans. These, though of course callable at short notice, were typically rolled over or "prolonged," whence their name. They were backed by securities, commodities, or other exchange-traded collateral. Thus industry and trade in effect obtained direct short-term capital in a very fragmented way, via margin loans provided by investors without the intermediation of a banking system. The prolongatie loans were considered safe; the interest rate was attractive and roughly tracked the London discount rate (hovering between 3 and 5 percent between 1820 and 1860; see Jonker 1996, figure 12.4, p. 96). The system worked so smoothly that intermediation and liquidity transformation by a nascent banking system was effectively crowded out. This remained the case well into the twentieth century, as argued by Jonker (1995). On the eve of World War I, the amount outstanding on prolongatie at any point in time was around 400 million guilders, more than double the known deposits of all the banks taken together. Jonker (1996; see figure 9.2, p. 191) points out that the short-term interest rate on the Amsterdam exchange remained at or above the yield on government bonds until nearly 1920, effectively precluding substantive profitable deposit taking by banks. The prolongatie market did not disappear until short rates fell dramatically toward the end of the 1920s. Another brake on banking development was Dutch savers' distrust of financial institutions. The sovereign bond defaults of the late eighteenth century and the parlous state of government finances in the early nineteenth century (with government debt
hovering around a staggering 400 percent of national income) meant that even the paper money circulated by the Nederlandsche Bank (set up in 1814 at the behest of King Willem I, an energetic supporter of initiatives to revive the Dutch economy) was long considered an unsafe substitute for specie. Private banking institutions were considered even more dubious, a view confirmed when the first wave of new banking ventures of the 1860s was followed by several banking failures in the long recessionary period starting in 1870. The industrial boom that started in 1895 precipitated a period of intense interest in industrial finance in the early twentieth century, right up until 1920. During this period many new companies were listed and public share offerings were readily absorbed. Banks, for this short period only, were prepared to offer long-term financing to industry. Meanwhile, a wave of banking consolidation from 1911 onward, together with a major shakeout of minor and regional banks in the crisis that started in 1920 (in 192022 a to-
Financing and Control in The Netherlands 477 tal of bad debts amounting to nearly 10 percent of the assets of the biggest five banks was written off), left the general banking industry dominated by the "Big Five" banks. Financing for industry completely dried up in the deflationary 1920s and did not revive until after World War II. Banks' reluctance to provide longterm financing for industry was the subject of intense debate; while large companies could fill the gap by issuing stocks and bonds, small and medium-sized enterprises were seriously constrained. The government went so far as to attempt to set up a bank for industrial finance in 1935 (it succumbed to the bad economic climate). The banks limited their role to collecting deposits (though, as Jonker 1995 shows, in the interbellum years Dutch banking deposits, and in particular time deposits, were still extraordinarily low relative to the total money supply compared with neighboring countries), making short-term loans (maturities of over three months were avoided as much as possible), and underwriting new issues. While they dominated the new issue market from the 1930s onward, they acted only as a conduit, never retaining equity stakes in industry or making long-term loan commitments. In short, the Dutch banks most resembled the British banks, not the universal banks of neighboring Belgium and Germany, as stressed in Van Goor and Koelewijn's (1995) overview of Dutch banking in the twentieth century. Dutch bankers focused on mercantile finance and consistently veered away from long-term commitments. As "general" banks they did do a lot of underwriting and investment banking business (also carried out by some private specialized firms); there was no counterpart of the GlassSteagall Act formally mandating the separation of commercial and investment banking. In 1945, the Herstelbank (bank for reconstruction), a joint venture between the government and the financial sector, was set up to fill the perceived gap in finance for long-term investment by providing long-term loans (a subsidiary, the Nationale Participatie Maatschappij, was created to take equity stakes). It played an important role in the recovery of Dutch industry over the decade following World War II. Perhaps its example (and that of its various successors), together with other government policies, stimulated the commercial bank
s' slow evolution toward medium- and long-term lending in the 195060 period. Meanwhile, banks did adhere to the fundamental principle of nonengagement in industry; indeed, industry spokesmen at the time explicitly expressed reservations about bank influence on commercial and strategic decision making. The boom years of 1955 to 1970 saw a period of increased diversification, as specialized institutions such as the mortgage banks lost ground. A spate of large-scale bank mergers led to a fear that banks had too much market power and were exposing themselves to an unacceptably wide range of risks. Thus, starting in 1971 the Nederlandsche Bank, as industry regulator, put
478 Abe de Jong and Ailsa Rцell out a number of unofficial directives (some of which were later codified in the Wet Toezicht Kredietwezen 1978) that prohibited mergers of general banks with insurance companies or mortgage banks, restricted bank participation in the equity of other companies (financial or nonfinancial) to 5 percent without explicit permission from the NB, and limited the value of share stakes held by banks to 60 percent of their capital. The 1980s were a difficult period of retrenchment for the banks, and again the accusations that banks were excessively cautious led to the adoption of various government measures (such as loan guarantees) to encourage the provision of risk-taking capital. Meanwhile the international expansion of Dutch industry brought with it a continuing trend toward the formation of large banking conglomerates offering a wider range of financial services. In 1990 banking and Insurance Regulation
was radically loosened. Participation in the European Union (EU) has meant that Dutch banks' market power is no longer considered a threat. As an immediate consequence, more mergers in 1991 created the three current giant banks (ABN-Amro, ING Bank, and Rabobank). And restrictions on banking-insurance alliances were lifted in accordance with EU practice. This has led to the formation of conglomerate groups holding substantial share stakes in large numbers of companies. Thus, a gradual trend away from the Anglo-Saxon model and toward a more Continental style of banking is in evidence. 8.1.5 Nonbank Institutional Investors: Insurance Companies and Pension Fund
s Insurance companies and pension funds have played a role in taking equity stakes, absorbing bond issues, and providing long-term loans at least since the beginning of the twentieth century. Our data for 1993 show that both the ING Bank and the Aegon insurance group had substantial longterm stakes in other companies (note that ING was formed in 1991 by a merger involving, among others, the large insurance company Nationale Nederlanden). Institutional investors rose to a prominent place in the Dutch capital markets during the early decades of the twentieth century.8 Traditionally, nineteenth-century life insurers had invested primarily in securities that 8. Renooij (1951, p. 63) reports figures illustrating the rising importance of such investors: between 1900 and 1939, deposits with private savings banks rose from f 80 to f 515 million and those with the state Rijkspostspaarbank from f 85 to f 670 million, while the capital of the life insurance companies rose from f 130 to f 1,359 million. Meanwhile, various social insurance funds were founded in the first quarter of the century in response to social legislation, and by 1939 the Algemeen Burgerlijk Pensioenfonds (the government employees' pension fund) held f 794 million in assets, the railway workers' and miners' pension funds held a combined f 203 million, and private industry's Ongevallenfonds and Invaliditeits- en Ouderdomsfonds together some f 491 million, while the self-employed workers' voluntary Ouderdomsfonds B held f 68 million.
Financing and Control in The Netherlands 479 were judged to be particularly safe and liquid; many of them invested exclusively in Dutch government bonds, and indeed many were restricted to do so by their statutes. The twentieth century saw a gradual lifting of these restrictions, but investment in private issuers' securities remained only a small fraction of their investments. In the preWorld War I burst of enthusiasm for industrialization, a typical life insurer, Eerste Nederlandsche, invested as much as 4 percent of its assets in banking and 7 percent in manufacturing securities; these were predominantly bonds rather than equity. Interest in privately issued securities then dwindled down to almost zero, until it revived in the late 1930s; by 1939, the precursor companies of Aegon held about 5 percent of their assets in manufacturing company securities, while over time the balance had shifted from bonds to equity (Gales 1986). Still, around 1950 life insurers' investment in industrial securities remained modest, indeed, the proportion was lower than at the turn of the century. Insurers did also make some contributions to industrial finance in the form of direct long-term loans (onderhandse leningen)9 and mortgages. But the trend toward equity and nongovernment bonds did not gather force until the second half of the twentieth century. Regarding pension funds, to illustrate their contribution to equity financing, consider the combined Philips pension funds, founded in 1913, described in appendix 2d of Van Nederveen Meerkerk and Peet (2002). Equity comprised a mere 2 percent of the fund's total investment in 1925; most of the fund was invested in (government) bonds. By 1950, equity took a 7 percent share, rising to 28 percent in 1975 and 46 percent in 2000. By then, the Philips pension fund was holding f 16,771 million in equity, together with f 106 million in venture participations. Here again we see very modest interest in risk-bearing capital in the first half of the century, with a marked shift toward investment in corporate equity in the second half of the century. In any case, it does not seem to be the case that institutional equity ownership has been matched by an active role in corporate decision making. The discussion surrounding the recent management crises at Ahold and other major Dutch companies gives some insight into why the independent, public-sector employee pension fund Algemeen Burgerlijk Pensioenfonds (ABP) is one of the few Dutch institutional investors to attempt an activist stance. As pointed out by an insurance company spokesman, banks and insurance companies are not only shareholders; for them, the firm in which they invest is at the same time a (potential) client: "You are in a difficult position if you want to present a new contract to the management board whilst you have voted against one of their proposals the day be- 9. These were exempt from stamp duty until 1939 and hence a popular substitute for bonds in the interwar period. The major place taken by direct long-term private loans in institutional investors' portfolios is distinctive to the Netherlands and Germany.
480 Abe de Jong and Ailsa Rцell fore."10 Meanwhile, activism by private companies' pension funds is likely to be reined in by the parent company's management, in return for reciprocal restraint by their counterparts' pension funds. Institutional shareholder activism thus remains somewhat limited in scope and potential. 8.2 Empirical Analysis: The Data 8.2.1 The Sample of Firms Our study focuses on all domestic firms that have equity officially listed on the Amsterdam Stock Exchange in the years examined. It should be pointed out that this concept is somewhat different from the usual definition of "listed firms" for the Netherlands, which also includes firms whose bonds only are listed. Traditionally, many of the security issues listed and traded on the Amsterdam exchange have been bonds; though the proportion of listed firms that list only their bonds and not their stock as well is relatively small (for example, 17 percent in 1910). One reason to exclude these firms from our sample is that they are somewhat less likely to comply with the obligation to publish annual accounts. The universe of firms for which we present data also excludes the financial sector. In 1923 this sector comprised mainly banks and mortgage banks. In the second half of the century insurance firms and collective investment vehicles such as mutual funds are important additional constituents of this group. Our main data sources for information about the nonfinancial firms are Van Oss's Effectenboek for 1923 and 1958 (Van Oss 1924, 1959) and the electronic database REACH (Review and Analysis of Companies in Holland) for 1993. It should be noted that many of the largest Dutch firms are not listed, so that our sample cannot be said to represent all the most important Dutch companies. Sluyterman and Winkelman (1993) identify the 100 largest Dutch firms in terms of their assets. Even though they point out that their methodology probably underrepresents privately held firms because their balance sheet data are harder to obtain and their accounting practices are generally more conservative, they still find that only about three-fourths of these firms are listed. Agricultural firms (and their food-processing outgrowths) in particular are often organized as cooperatives, as are the banks that specialize in agricultural loans (the Rabobank and its precursors). 8.2.2 The Three Sample Years Our data were gathered for three years spaced at thirty-five-year intervals: 1923, 1958, and 1993. In choosing these particular years we were influenced by three considerations. 10. The speaker is D. Brilleslijper, Delta Lloyd spokesman, in FEM Business, 20 September 2003.
Financing and Control in The Netherlands 481 First, we would like to have years that were in some sense typical of an epoch. The year 1923 comes toward the end of the first great boom in industrial development; it is still a year of relative prosperity, predating the subsequent collapse in share prices, the Depression, and the Second World War. In 1958 the economic dislocation wrought by the war has receded: postwar reconstruction is virtually complete, and a new era of prosperity and growth has set in. Meanwhile, 1993 is a year in which the impact of EU membership has already shaped many developments. A second consideration is the availability of data. For example, large ownership stakes were only available following the 1991 disclosure law (Wet Melding Zeggenschapsrecht), which came into effect in February 1992 (De Jong et al. 2001). This makes 1993 an interesting year to study. Finally, our aim was to try to pick years that as much as possible enabled us to complement rather than duplicate the available body of work on Dutch economic history. 8.2.3 Data Availability For most limited-liability companies, the publication of annual accounts was not legally required in the Netherlands until 1928. However, from 1909 the stock exchange's Fondensreglement required all companies that wished to list their stocks or bonds to make available to shareholders annual published accounts comprising the balance sheet and profit- and loss statements. By 1910, about 80 percent of listed firms complied in whole or in part, though the level of compliance was considerably lower (around 50 percent) among manufacturing firms. By 1923, our first sample year, compliance (as measured by the availability of accounts in Van Oss' Effectenboek) had risen considerably. Information on share ownership prior to the share ownership disclosure law of 1991, the Wet Melding Zeggenschapsrecht (WMZ), is very hard to obtain because as a rule Dutch public listed companies issue bearer, not registered, shares, and we have no easy access to public registries to trace share ownership. In principle, some information about share ownership can be retrieved from company archives. In particular the records of shareholder meetings would give insight into, at least, the identities of shareholders actively involved in decisions about the company. Such archival research, however, is far beyond the scope of this paper. Thus, for 1923 and 1958 the only way we investigate family influence and control is by tracking the identities of the management and the board of directors, both available for much of the late nineteenth and the twentieth centuries from published sources.11 11. This information is in principle available for all public limited companies (naamloze vennootschappen) from both the yearbooks of NVs compiled by Van Nierop and Baak over the period 18801948 and from the yearbooks relating to listed securities, Van Oss's Effectenboek 190378 (later continued as Effectenboek).
482 Abe de Jong and Ailsa Rцell
8.2.4 Summary Statistics As shown in table 8.1, the number of firms on the Amsterdam Stock Exchange's official list has actually declined over the last few decades studied. The decline in numbers is offset by a substantial increase in size; the average book value of assets increased more than a hundredfold over the sev-
Book value total assets ( f 1,000) Past three-year growth book value total assets No. of observations Tobin's q No. of observations Return on assets No. of observations Four-year standard deviation ROA No. of observations Payout ratio No. of observations Debt to total assets Fixed assets to total assets Cash and liquid assets to total assets Age No. of observations Managerial board size Supervisory board size Family firm: (former) firm name equals board member's surname (%) Family firm: at least two board members with same surname (%) Family firm based on both criteria (%) Family firm based on at least one criterion (%) No. of firms
13,673 (3,158) 0.077 (0.074) 303 0.372 (0.338) 214 0.073 (0.047) 317 n.a. 0.375 (0.311) 300 0.300 (0.280) 0.552 (0.578) 0.114 (0.050) 21.80 (18) 317 2.158 (2) 4.874 (5) 28.1 31.5 16.4 43.2 317
79,700 (9,314) 0.161 (0.102) 318 0.421 (0.411) 245 0.159 (0.113) 321 0.037 (0.024) 298 0.716 (0.440) 323 0.339 (0.325) 0.404 (0.352) 0.124 (0.084) 47.12 (46) 333 2.318 (2) 4.540 (4) 27.6 29.1 16.2 40.5 333
2,286,000 (360,000) 0.170 (0.080) 141 1.270 (1.132) 143 0.073 (0.074) 143 0.032 (0.023) 141 0.369 (0.363) 143 0.535 (0.536) 0.381 (0.355) 0.107 (0.041) 48.75 (36) 84 2.776 (2) 5.167 (5) 6.3 5.6 1.3 10.4 143
Notes: Medians are reported in parentheses below the means. n.a. not available. ROA return on assets.
Financing and Control in The Netherlands 483 enty years from 1923 to 1993, a period during which prices (as measured by the gross domestic product [GDP] deflator) rose by a factor of 12. To some extent these trends are attributable to mergers and consolidation, but a tendency to limit the exchange's official list to very large and liquid companies may also play a role. The data regarding the three-year growth in assets show that 1923 followed upon a difficult period; indeed, there had been a serious economic downturn, and overall stock market equity prices had fallen by about onehalf during the immediately preceding decade. That Tobin's q was extremely low in 1923 is not surprising; however, the low average value of 0.421 for 1958 is less easily explained. Tobin's q is measured as the ratio of market value of total assets to book value of total assets. The market value of total assets is measured as book value of total assets minus book value of equity plus market value of equity. A problem arises here, because the book and market values of equity need to be "comparable." Especially in 1923, many firms have multiple types of equity. The market value is not available for each type of equity. In 1923 and 1958, we leave the types of equity for which we have no market prices in book value terms and attributed reserves to equity types on a pro rata and book value basis. The median return on assets (ROA) fluctuated between a low of 4.7 percent in 1923 and a high of 11.3 percent in 1958, down to 7.4 percent in 1993. The ROA in 1923 is the ratio of net profits to equity; in other years it is operating income to book value total assets. The median payout ratio of 0.31 that we obtain for 1923 is somewhat on the low side in both historical and international perspective. In the nineteenth century, the norm was to pay out most or all of earnings, perhaps with some retentions from extraordinary profits to create a reserve for use in smoothing dividends in bad years. In the first two decades of the twentieth century, it gradually became accepted practice to retain earnings for the purpose of expansion. However, payout ratios were generally still very high, and Post (1972, table 5) cites a payout ratio of 0.78 in 1923 for all Dutch NVs (not just listed ones). One point to note is that 1923 was not a good year for the economy, coming at the end of the depression of 1921 23. Many of the firms in our data set made losses, and nearly half of the firms passed their dividend; the median payout ratio for the firms that did pay out a nonzero dividend was 0.75, which is very close to Post's figure. Our data source, Van Oss' Effectenboek, sometimes gives fairly detailed information about the disposition of profits, both as stipulated in the company charters and as carried out ex post. A striking feature of the 1923 data is the substantial proportion that is statutorily destined for the executives and directors in the form of tantiиmes or profit-sharing agreements. The norm for statutory payouts of this nature is in the region of 15 percent of profits, which suggests that such payments should perhaps be interpreted
484 Abe de Jong and Ailsa Rцell in part as a reflection of the ownership rights of the individuals concerned rather than just as remuneration for executive effort. But in practice the actual payments made often fall far short of the profit-sharing payouts stipulated ex ante in the company statutes. By 1958 the mean (median) payout ratio was 0.72 (0.44), declining to 0.37 (0.36) in 1993. Payout ratios declined secularly until the 1980s, as firms chose to retain earnings to finance expansion. A probable contributing factor was the introduction of a corporation tax, phased in around 1941. A classical system is in force: corporate earnings are taxed at 35 percent, whether distributed or not, and dividends are subsequently taxed as personal income at a heavy 60 percent marginal rate, while there is no capital gains tax. Indeed, of the thirty-three countries studied by La Porta et al. (2000), the Netherlands tax regime has the rock-bottom ratio of net-of-tax payout from dividends relative to capital gains. Accordingly, one would expect Dutch personal investors to have little enthusiasm for dividend payouts and a preference for retained earnings. Such a preference would be less likely on the part of those institutional investors that are exempted from income tax. While the Dutch tax system does not attempt to mitigate the double taxation of dividends at the corporate and personal income tax levels, it has traditionally been exceptionally careful in ensuring that intercorporate dividends are not double-taxed at the corporate level. This feature of the tax regime is one reason why the Netherlands (and in particular the Netherlands Antilles) is popular as a base for international holding companies. Leverage as measured by the ratio of debt to total assets exhibits a marked increase from a median of 0.32 in 1958 to 0.54 in 1993. Again, the corporate tax shield from debt may explain this increase in leverage in the postwar half-century. The sizes of the managerial and supervisory boards remained fairly constant over the seventy-year period studied. Meanwhile, founding family influence seems to have declined dramatically. Our proxies for family influence are two: one is the presence of board members with the founding family surname; the other is multiple board members with a common surname. These indicators of family presence declined only slightly from 1923 to 1958, but there was a large reduction from 1958 to 1993. Both criteria for family influence dropped by a factor of about 5, from roughly 30 percent to 6 percent, leaving a total of only 10.4 percent of firms in 1993 still exhibiting one or both indicators of family influence. 8.3 Oligarchic Clauses and Takeover Defenses 8.3.1 Description of Takeover Defenses and Shareholder Rights Dutch corporations are insulated against the threat of hostile takeovers by an array of unusually strong and somewhat idiosyncratic defense mech-
Financing and Control in The Netherlands 485 anisms. In this section we will describe the main devices currently in use12 and attempt to trace their historical origins. It should be pointed out at the outset that ever since 1881, Dutch corporate law does not permit the use of nonvoting or lower-voting shares, thus ruling out one obvious means of detaching control from ownership. Moreover, early Dutch corporate law from 1838 onward mandated voting caps in order to protect minority shareholders from oppression by a dominant shareholder: one person should not have more than six (three) votes in a company with more (less) than a hundred shares. This means that before the new law of 1928, pyramids or large majority stakes were not a secure means of entrenching control, necessitating the development of alternative safeguards. As a small country surrounded by powerful and at times warring neighbors, it should not be surprising that vulnerability to foreign influence has always been a source of serious concern among Dutch industrialists, particularly in the early part of the twentieth century. A number of defensive measures have been rationalized on this basis. As mentioned earlier, the Dutch stakeholder model (poldermodel ) also induced a movement that shifted shareholder power to independent supervisory board members. Statutory Defenses By statutory defenses we mean those that are enshrined in the company's statutes. Among the statutory defenses, those that restrict the powers of the algemene vergadering van aandeelhouders (AVA) or shareholders' meeting are known as oligarchische regelingen (oligarchic measures/arrangements/ devices). Such clauses give all or part of the control of the company to others than to the shareholders representing the majority of the capital at the shareholder meetings. The most prominent oligarchic device is the use of prioriteitsaandelen or priority shares13 with statutorily defined extra powers of decision within the corporation. Such shares were first introduced in 1898, when the main Dutch oil company operating in the Netherlands Indies (the progenitor of Royal Dutch/Shell) changed its statutes to ward off the threat of foreign influence. Such shares are often associated with a right of bindende voordracht (binding proposal) in the nomination of management and directors. Other oligarchic devices include arrangements to allocate decisionmaking powers to another organ of the company (such as the board or the priority shareholders) for explicitly specified important classes of decisions that would normally require shareholder approval: such matters can include the composition of management and board, their remuneration, dividend payout policy, modification of company statutes, or dissolution 12. Our description is based on Voogd's (1989) detailed investigation of Dutch companies' statutory defenses. 13. Confusingly, such shares were initially known as "preference shares," but this usage is now no longer allowed. They are also sometimes called "founders' shares" or, say, "A-shares."
486 Abe de Jong and Ailsa Rцell of the company. Finally, there are devices such as voting restrictions and strengthened supermajority and quorum requirements for shareholder meeting decision making. Since World War II the issue of shares into friendly hands, and in particular of preference shares ( preferente aandelen), has developed into a major defensive strategy. This is a nonoligarchic statutory device in that it attempts to influence the composition of the shareholder meeting rather than restricting its powers. From 1949 to 1981 there were some twenty-six instances where companies defensively issued ordinary shares to friendly individuals, banks, institutional investors, potential merger partners, or allied foundations. The motive for such defensive issues was to dilute the power of large shareholders, preserve independence in the face of a hostile takeover attempt, or ensure takeover by a white knight. The use of ordinary shares for defensive purposes waned after the mid-1970s because the issue of ordinary shares is costly in terms of cash requirements (the issue price must be fair to existing shareholders, and for registered shares, a down payment of at least 25 percent of the nominal value plus 100 percent of the agio, the difference between the issue price and the nominal value, is required), frowned upon by legal commentators, and much circumscribed by the adjustments to Dutch company law made in 1981 to implement the Second European Directive on Company Law. In particular, the new law gave preemptive rights to participate in ordinary share issues to existing shareholders, unless the shareholders' meeting explicitly waived the right; and a five-year expiration limit was placed on any allocation of the power to make issue decisions to organs other than the shareholders' meeting.14 In the early 1970s preference shares quickly replaced ordinary shares as the instrument of choice for defensive issues. Provisions for issuing preference shares for defensive purposes first appeared in the statutes of a Dutch company, Rijn-Schelde, in 1969. There were two reasons for the switch to preference shares. First, under the new law, ordinary shareholders do not automatically have preemption rights to new issues of preference shares (though the stock exchange did attempt to impose on listed companies a shareholder approval requirement for issues of preference shares of more than 50 percent of the existing capital). Second, preference shares can be designed to provide a much larger ratio of voting power to paid-in capital than ordinary shares; indeed, the net outlay can be made essentially negligible. Preference shares can be issued more or less at par, if liquidation 14. Even so, Voogd (1989) finds that on January 1, 1988, 59 percent of the companies on the stock exchange's official list had statutes empowering an organ other than the shareholders' meeting to issue ordinary shares (in 76 percent of these cases, the management; in 15 percent, the priority shareholders; in 8 percent, the board of directors; and in 1 percent, the board and management jointly), while 51 percent of companies had made similar arrangements for the power to deny preemptive rights to shareholders (distributed 74 percent, 17 percent, 8 percent, and 1 percent respectively among the various alternative organs).
Financing and Control in The Netherlands 487 rights are limited to the paid-in capital and the preferred dividend is suitably tied to the market interest rate. If the legal minimum of 25 percent of par value is paid in, the number of votes obtained for any paid-in sum of money is maximized. But that is not all. The preference shares are generally placed with financial institutions, institutional investors, or a foundation specially set up for the purpose. For such a foundation to be selffinancing it would need to borrow the amount required for paid-in capital; therefore the dividend on the preference shares must be carefully tied to the required interest on the loan, and cumulative preference rights are necessary to ensure that the foundation can reasonably be expected to meet its obligations. Voogd (1989) found that 48 percent of the listed companies he examined had defensive preference shares, defined as preference shares that were op naam (registered) and not aan toonder (bearer) shares, not fully paid in, with limited dividend and liquidation rights, and with dividend rights tied to the market interest rate. Of the companies issuing defensive preference shares, 66 percent had issued preference shares equal to 100 percent of the authorized ordinary shares, thus carrying 50 percent of voting power (20 percent of companies had preference shares ranging between 50 and 100 percent of the ordinary shares, and only 14 percent of companies issued 50 percent or less). A further device for influencing the composition of the shareholder base is the issue of registered (op naam) shares15 together with limitations on the transfer or ownership of such shares. Such blocking devices (blokkeringsregeling) can include a requirement for permission from a company organ for the transfer of shares, a requirement to offer shares to fellow shareholders before selling them to third parties, or statutory limitations on who can own the shares (Dutch nationals, residents, etc.). Finally, an important statutory defensive device is the X percent rule (X percent-regeling), which limits the ownership of shares (usually the ordinary shares, which are normally the ones that are listed and that thus change hands often) by a single shareholder. Voogd (1989) finds that 25 percent of listed companies (excluding mutual funds) have such a rule in their statutes. Usually the company's shares are registered (op naam) and placed with a specially created foundation or administratiekantoor, which issues nonvoting bearer certificates that are listed on the stock exchange. These are freely exchangeable into voting shares, but only up to the specified X percent boundary. Other less common statutory defensive measures include voting limits, though as these can be circumvented by the use of straw men, they are now out of favor. All twelve officially listed companies that included voting caps in their statutes in early 1988 were ones already in existence before 1929; 15. Such shares cannot be listed; typically, these companies issue bearer certificates that are traded on the stock exchange.
488 Abe de Jong and Ailsa Rцell taken together, these companies represented around 40 percent of the market value of Amsterdam listed companies (Voogd 1989). Some corporate statutes include a varied brew of other measures limiting voting rights to long-term shareholders, Dutch nationals, and so on. Nonstatutory Defenses A classic and quite common nonstatutory defense mechanism is the use of an administratiekantoor (AK), typically a special-purpose foundation that owns all or most of the company's shares and issues nonvoting certificates to the general public. The certificates carry all the underlying shares' economic rights (dividends, liquidation value, etc.) but no control rights. Especially in cases where these certificates are niet royeerbaar--that is, not exchangeable for ordinary vote-carrying shares--the effect is to give all voting power to the trustees of the AK, who are typically closely intertwined with the company's management, although the stock exchange imposes some independence requirements on the AK.16 From the mid-fifties onward the increasing use of certification of this kind has been roundly criticized from many quarters, including the legal profession and the Vereniging voor de Effectenhandel, the securities dealers' association running the stock exchange. Since 1992, listings of niet-royeerbare certificates are not allowed anymore. In a recent adaption of Dutch company law, all certificate holders are allowed to vote by proxy with their certificates. Only under special circumstances (in case voting by certificate holders interferes with the general interests of the firm) can the proxy voting be refused or limited. The use of pyramidal holding companies to concentrate control is relatively rare in the Netherlands, given that certification is a readily available means of securing control without any appreciable outlay of capital. However, a small number of such holding company constructions do exist,17 and with certification likely to be phased out, pyramids may become more prevalent. Similarly, cross-shareholdings along the French model are unusual but not unknown in the Netherlands. The Structured Regime In 1971, the "structured regime" was imposed on all large companies with a large number of employees in the Netherlands. The primary reason for its introduction was to give workers some power of consultation and in- 16. For example, at a chaotic shareholder meeting for Ahold in September 2003, 97 percent of votes supported a remuneration package for the incoming CEO that was widely denounced as excessively generous. No representatives of the AK were present; before even knowing the broad scope of the remuneration proposal, they had already authorized the secretary of the management board to exercise their votes, representing 50 percent of the total. 17. Most notably, Heineken, where the Heineken family has 50.01 percent control of the unlisted Heineken Holding NV, which in turn controls the listed firm Heineken NV with 50.01 percent.
Financing and Control in The Netherlands 489 fluence through the ondernemingsraad (workers' council). In addition, some of the powers normally given to the shareholders' meeting (such as management appointments and the approval of the annual accounts), as well as the power to approve a set of other important management decisions, were vested in the raad van commissarissen (supervisory board), which appointed its own members by a system of co-optation that basically bypassed any shareholder influence. An exemption for the structured regime is allowed for multinational companies with a majority of employees working abroad. Also, companies that do not meet the criteria for compulsory subjection to the structured regime can still voluntarily apply it. Many have chosen to adopt the regime voluntarily or not to abolish the regime when as a result of international expansion the percentage of foreign workers passed the 50 percent threshold. The structured regime gives corporate insiders much more freedom at the expense of shareholder rights. Under very specific conditions firms have to adopt the mitigated structured regime, where the powers to appoint management and approve annual accounts would normally remain with the shareholders' meeting, although the co-optation system for supervisory board appointments remains in place. Recently, the structured regime has been a topic of public debate. The influence given to employees via the ondernemingsraad is quite weak; a recently adopted proposal for the revision of the regime includes reserving positions on the supervisory board for employee appointees, a move that will clearly enhance worker power. At the same time the structured regime's allocation of shareholders' normal powers to an unaccountable, self-perpetuating supervisory board is the target of heavy criticism. A prominent Dutch legal scholar, Jaap Winter, has gone so far as to describe the structured regime as a "cynical compromise"18 that transfers shareholder rights to corporate insiders without giving employees or shareholders any real decision-making powers. 8.3.2 Data and Analysis Our data enable us to give an overview in table 8.2 of the takeover defenses employed by the companies in our sample; in the Netherlands, takeover protection has traditionally been very strong. Our sources are Van Oss's Effectenboek for 1923 and 1958 (S. F. van Oss 1924, 1959) and for 1993 the Gids bij de Officiлle Prijscourant (J. H. de Bussy 1993a). One of the most prominent mechanisms, priority shares, has increased dramatically in importance; by 1993 43 percent of firms had such shares.19 18. "The starting point was the idea that labor and capital were equally valuable, and both should have equal power. In reality a cynical compromise was reached: the heart of their powers has been taken away from the shareholders, while little more was received by employees" (FEM Business, 13 September 2004). 19. The low figure for 1923 should be treated with some caution, as the nomenclature for priority or founders' shares was somewhat less clearly established.
490 Abe de Jong and Ailsa Rцell
Takeover defenses and ownership structure (%)
Priority shares Voting limits Certificates Limited or fully exchangeable certificates Not exchangeable X arrangement Certificates and traded ordinary shares Joint ownership construction Preference shares (antitakeover) Structured regime Compulsory Voluntary Mitigated Ownership concentration Largest outside blockholder All outside blockholders Ownership identity of blocks Banks
2.52 (By law) 11.67 8.52 10.76
28.23 0.30 24.92 18.02 5.71 21.62 2.10
Managerial board members
Supervisory board members
No. of firms
1993 42.66 6.29 38.46 3.50 10.49 2.10 3.50 60.14 53.15 41.96 9.79 1.40 24.49 43.10 7.16 (77 nonzero) 2.75 (50 nonzero) 0.73 (11 nonzero) 0.61 (3 nonzero) 12.58 (51 nonzero) 5.31 (20 nonzero) 2.47 (12 nonzero) 143
Meanwhile, voting limits were, in 1923, still a feature of all firms by law. Their prevalence in the statutes of listed firms had fallen to 6.3 percent by 1993, and in most cases these were firms surviving from the pre-1928 period when statutory voting limits were mandatory.20 The use of certificates or depository receipts has increased substantially over time; 38 percent of firms had some measure of certification present in 1993, rising steadily from 12 percent in 1923. A joint ownership construction was present in 3.5 percent of firms by 1993. 20. In 1958 it is possible that Van Oss does not contain complete information about voting limits. Therefore, the percentage reported is likely to underestimate the actual presence.
Financing and Control in The Netherlands 491 The issue of preference shares is a crucial defensive strategy in takeover situations. The use of preference shares for defensive purposes was initiated in 1969; by 1993, 60 percent of listed industrial firms had this defensive mechanism in place. The structured regime, which gives some influence to the workers' council and devolves much of the authority of the shareholders' meeting to a self-constituted supervisory board, was introduced in 1971. By 1993, 53 percent of listed industrial firms were subject to the structured regime, and 10 percent of these had voluntarily chosen to have the structured regime apply. Table 8.7 in section 8.4.3 compares and contrasts the prevalence of takeover defenses in family and nonfamily firms. The main distinction is that in family firms there are more likely to be priority shares, conferring upon the holders of these shares a varying set of decision-making powers that would otherwise fall upon the ordinary shareholders' meeting. For 1993, ownership data are available, and it is possible to investigate the interactions between takeover defenses and ownership structure. Table 8.8 in section 8.4.3 shows that, on the whole, takeover defenses and concentrated ownership are substitute control mechanisms and thus negatively correlated. Large outside block holders are negatively correlated with all defense mechanisms considered, and significantly so with the use of defensive preference shares and priority shares. Similarly, when management board members hold large stakes, certificates are less likely to be used. The results regarding the structured regime need to be interpreted with caution as it is generally compulsory for the largest firms. Such firms are less likely to be heavily management owned and more likely to be partially owned by a bank. The finding that takeover defenses and concentrated ownership are substitutes rather than complements agrees with earlier work by De Jong and Moerland (1999). Table 8.9 in section 8.4.3 explores the impact of takeover defenses on corporate performance by regressing Tobin's q cross-sectionally on dummies for the presence of the various common defense mechanisms (the third column of results for each of the three sample years in table 8.9). Earlier research by De Jong, Moerland, and Nijman (2000) on a cross section of fifty listed Dutch firms suggests that defense mechanisms such as certificates, defensive preference shares, and, most significantly, the structured regime do reduce other performance measures
such as the stock market return and the return on equity; they find that only the size of the supervisory board has a significant, negative impact on Tobin's q. De Jong et al. (2004) confirm these results for a sample of all Dutch listed firms over 199399. In our larger sample, again, there is not much evidence of an impact of defense mechanisms on q, though in 1958 the presence of priority shares seems somewhat detrimental.
492 Abe de Jong and Ailsa Rцell 8.4 Networks of Influence: Interlocking Board Memberships 8.4.1 Boards and Networks In this section we will focus on the phenomenon of interlocking directorates--that is, of having the same individual occupy board seats in multiple firms. Two aspects of this practice will be looked at. First, the number of appointments per board member is studied. Members with multiple appointments may have reputational capital; that is, they may be excellent managers or monitors. On the other hand, multiple appointments may reduce the time available for individual firms, reducing the effectiveness. Ferris, Jagannathan, and Pritchard (2003) provide recent evidence in U.S. firms and find no negative effects of multiple appointments. For the Netherlands, there is no evidence relating network relationships to firm performance but a wealth of descriptive evidence regarding interlocking directorates. To name but two prominent studies, Schijf (1993) describes networks in 1886 and 1902 and Stokman, Wasseur, and Elsas (1985) focus on networks in 1976 in the context of an international comparative project. A second aspect of interlocking directorates that is of particular interest is the relation between banks and nonfinancial firms. Bank relations may bring expertise to the board of nonfinancial firms. Besides, bank relations may offer monitoring, which reduces contracting costs. On the other hand, banks may abuse their power and information to expropriate wealth from other lenders and shareholders; recent studies on U.S. firms are Booth and Deli (1999) and Kroszner and Strahan (2001). The relations between banks and nonfinancials have been studied in the Netherlands by, among others, Van den Broeke (1988) and Jonker (1989). Van den Broeke selects four industrial firms and one bank and describes the interlocking directorships. The bank, Rotterdamsche Bankvereeniging, has joint board members with three out of four firms through eight interlocks in the period 191839, even though throughout this period it did not make a single long-term loan to any of the firms concerned, in line with Dutch banking practice at the time. Jonker (1989) selects eight banks and measures interlocks with nonfinancial exchange-listed firms in 1910, 1923, 1931, and 1940. For example, in 1923, the eight banks had forty-three board members and these persons held 431 board positions outside the banks. Interlocking directorships can involve both executive and supervisory board members. Dutch firms have dual board systems on the German model. The first tier comprises the executive board (Directeuren or Raad van Bestuur), the management team that is responsible for the firm's strategy and daily operations. These executives are supervised by the second tier, the supervisory board (Raad van Commissarissen or Raad van Toezicht).
Financing and Control in The Netherlands 493 In 1923, supervisory boards were not a legal requirement (Bos 1923, p. 34). Nonetheless, all exchange-listed firms in 1923 do have a supervisory board. The members are normally appointed by the shareholders' meeting. In special cases, the owners of preferred shares, priority shares, or bonds have the right to appoint all or a limited number of supervisors. Intermediate arrangements existed where other parties than the shareholders propose members, while the shareholders can reject the proposal. In 1993, a supervisory board is a legal obligation for firms that adopt the so-called structuurregeling or structured regime, introduced in 1971. This regime is compulsory for firms that meet size criteria (in particular, those that have more than a cutoff number of domestic employees). In 1993, again, all the listed firms have supervisory boards. 8.4.2 Data Sources Our aim is to describe the relevance of interlocks for nonfinancial firms. First, we describe the interlocks with other nonfinancial firms. Second, we focus on interlocks between banks and nonfinancials. The focus for nonfinancial firms is simply on all exchange-listed firms. For 1923 and 1958 we use Van Oss's Effectenboek (S. F. van Oss 1924, 1959). For 1993 we mainly use REACH and Jaarboek van Nederlandse Ondernemingen (J. H. de Bussy 1993b). For the identification of board members of banks we do not want to restrict ourselves to listed banks because, especially in 1923, several important banks were unlisted partnerships. Therefore we select the largest banks. For 1923 we use the Financieel Adresboek voor Nederland issued by J. H. de Bussy (1923). This book contains the section Financiлele instellingen in Nederland, which includes for each financial institution its name, its placed equity and reserves, and the names of its board members. The book includes listed and nonlisted institutions. For 1958 we use the same book (J. H. de Bussy 1958) and collect bank information from the section Banken credietwezen. For 1993 we use Omzetcijfers 1993, issued in 1994 by Het Financieele Dagblad. This guide contains the banks and other financial institutions in the Netherlands, including total assets. The board members of most banks are in the Jaarboek van Nederlandse Ondernemingen and, if not, are obtained from Annual Report
s. For 1923 we identify 504 banks, of which 423 banks have available a book value of equity (placed equity plus reserves). Total equity value is 1,319 million guilders. The first 60 firms have 1,213 million guilders of equity value, or 92 percent of the total. The smallest firm in the selection of 60 has equity worth 200,000 guilders. Of these 60, 32 are listed on the Amsterdam stock exchange. The 5 largest banks have 49 percent of the total equity value, and the 10 largest have 67 percent. In 1958, we traced 148 banks, with total equity value of 1,099 million guilders. The largest 50 banks have 96 percent (1,061 million guilders). The
494 Abe de Jong and Ailsa Rцell
5 largest banks have 48 percent of the total equity value, and the 10 largest have 69 percent. In 1993, we have seventy-one banks (general and savings banks), and for fifty-six we have a book value of total assets. Total value is 1,309,788 million guilders. We select the ten largest banks but exclude two banks for governmental financing. We also include three smaller banks that are known for long-standing relations with nonfinancials. The eleven banks have a total asset value of 1,084,151 guilders, or 91 percent (excluding governmental banks). The difference with 1923 is striking and in particular caused by the dominance of three large banks: ABN-AMRO, Rabobank, and ING. 8.4.3 Results and Analysis Table 8.3 describes the interlocks of board members of nonfinancial listed firms for 1923, 1958, and 1993. The first six rows in table 8.3 describe our sample of (nonfinancial) firms and banks. The average board size has fluctuated somewhat: the average total number of board members per firm decreased slightly from 7.03 in 1923 to 6.86 in 1958, increasing by 1 to 7.94 by 1993. It is important to notice that the number of banks in our sample declines from fifty-seven to
Boards and interlocks
Number of firms Number of managerial board members Number of supervisory board members Number of banks Number of managerial board members Number of supervisory board members Firms: number of managerial board members With one interlock With two interlocks With three interlocks With four interlocks With five interlocks With more than five interlocks Total interlocks Average number of interlocks Firms: number of supervisory board members With one interlock With two interlocks With three interlocks With four interlocks With five interlocks With more than five interlocks Total interlocks Average number of interlocks
1923 317 684 1,545 57 238 432 137 56 32 39 11 61 1,248 1.82 371 205 136 141 49 170 3,440 2.23
1958 333 772 1,512 50 159 361 127 38 21 6 19 25 599 0.78 328 175 131 69 77 220 3,606 2.39
1993 143 397 739 12 60 122 38 13 6 3 0 0 94 0.24 170 89 90 32 6 0 776 1.05
Financing and Control in The Netherlands 495
twelve over the seventy-year period studied; as mentioned in our discussion of the data selection procedure, ongoing concentration in the banking system means that the proportion of total banking equity value represented by our sample remains roughly constant at over 90 percent. Not surprisingly, as the banks in the 1993 sample are so much larger, they have more board members: 15.2 on average, as opposed to 11.7 (10.4) in 1923 (1958). Meanwhile, for both banks and industrial firms, the ratio of supervisory board members to management board members remained fairly steady, ranging between 1.82 and 2.27. Table 8.3 shows us whether board members have more or less additional board seats. For managerial board members (including the chairman), our findings indicate that members in 1923 held many more positions than in 1958 or 1993: the average number of interlocks dropped from 1.82 in 1923 down to 0.24 in 1993. For supervisory board members the average number of interlocks decreased less dramatically: in the postwar period the average number fell by roughly one-half. In 1923 we also find quite a few board members with more than five interlocks; by 1993, no board member had more than five additional seats. In the remainder of this section we focus on industrial-firm board members who have affiliations with banks. Table 8.4 contains the frequency distributions
of bank interlocks in firms. Banking interlocks were more widespread in the earlier periods of our investigation: the proportion of firms with no bank interlocks was 40 percent (39 percent) in 1923 (1958), rising to 55 percent in 1993. Thus, in 1923 and 1958, the presence of bankers was more widespread than in 1993. In 1923, twelve firms even had ten or more bankers on the board. The average number of board members with a bank affiliation decreases from 0.60 (0.61) in 1923 (1958) to 0.45 in 1993. However, it should be noted that the significant concentration in the banking industry over the 195893 pe-
Frequency distribution bank interlocks
% of firms with:
No bank interlocks One bank interlock Two bank interlocks Three bank interlocks Four bank interlocks Five bank interlocks Six bank interlocks Seven bank interlocks Eight bank interlocks Nine bank interlocks Ten or more bank interlocks Average number of bank interlocks
40.38 22.08 12.30 7.89 5.05 4.42 1.26 0 0.63 2.21 3.79 0.596
1958 39.34 26.43 13.81 6.61 5.11 3.00 0.90 1.20 0.60 0.90 2.10 0.607
1993 55.24 25.87 9.09 5.59 3.50 0.70 0 0 0 0 0 0.447
496 Abe de Jong and Ailsa Rцell
Banks and their interlocks
1923 (over 15) Rotterdamsche Bankvereeniging Nationale Bankvereeniging Bank voor Indiл De Twentsche Bank Nederlandsche Handel-Maatschappij Hollandsche Bank voor Zuid-Amerika Koloniale Bank De Nederlandsche Bank Kas-Vereeniging Amsterdamsche Bank Bank-Associatie Wertheim & Gompertz 1834 en Credietvereeniging 1853 Nederlandsch Indische Handelsbank 1958 (over 10) Rotterdamsche Bank N.V. De Nederlandsche Bank N.V. De Twentsche Bank N.V. Nederlandsche Handel-Maatschappij N.V. Amsterdamsche Bank N.V. Nationale Handelsbank N.V. Bank voor Handel en Scheepvaart N.V. N.V. Export-Financiering-Maatschappij N.V. Nederlandsche Bankinstelling voor Waarden belast met Vruchtgebruik en Periodieke Uitkeringen Van Mierlo en Zoon N.V. Nederlandse Overzee Bank N.V. N.V. Hollandsche Disconteeringsmaatschappij van 1939 N.V. Hollandsche Koopmansbank Kas-Associatie N.V. Maatschappij voor Middellang Crediet N.V. Hollandsche Bank Unie N.V. 1993 (over 10) Abn-Amro Internationale Nederlanden Bank (ING) Nationale Investeringsbank MeesPierson
Interlocks 119 56 55 50 43 31 31 26 26 19 18 16 149 73 63 51 46 20 20 19 19 19 14 12 12 11 11 11 34 18 18 14
riod would have led to a decline in the number of bank board members available for positions on industrial firm boards. The use of interlocks by banks is illustrated in table 8.5, which lists all banks with at least ten (fifteen) interlocks in 1958 or 1993 (1923). It is clear that there has been a substantial decline in the latter period of the century in the number of major-bank board members who sit directly on industrial-firm boards.21 21. Our data do not allow us to determine whether all or part of this decline may be offset by the placement of bank officials from below the board level on industrial firm boards.
Financing and Control in The Netherlands 497
Interlocks at firm level
Own firm--other firm--type of other firm
Supervisory board--supervisory board--industrial Supervisory board--management board--industrial Management board--supervisory board--industrial Management board--management board--industrial Supervisory board--supervisory board--bank Supervisory board--management board--bank Management board--supervisory board--bank Management board--management board--bank
5.997 (83.9) 1.079 (46.7) 1.530 (31.5) 1.000 (25.6) 1.202 (46.7) 0.461 (29.0) 0.293 (11.4) 0.287 (3.8)
6.327 (79.9) 0.901 (42.3) 0.921 (24.6) 0.366 (12.3) 1.285 (53.7) 0.198 (17.4) 0.201 (8.4) 0.012 (0.9)
3.441 (74.1) 0.454 (37.8) 0.454 (20.3) 0 (0) 0.622 (40.6) 0.077 (7.7) 0.084 (7.7) 0 (0)
Note: Average number of interlock and in parentheses percentage of firms with at least one interlock.
Table 8.6 further documents the decline in interlocks, contrasting banks' and other industrial firms' board members' roles on industrial-firm boards. Industrial-firm interlocks have declined steeply over the seventy years of our investigation; the overall decline in the average number of interlocks is by roughly a half. While multiple supervisory board memberships are still very common, interlocks involving management board members in particular have fallen steeply. Indeed, by 1993 there was no industrial firm in our sample sharing a common management board member with a bank or other industrial firm. Meanwhile, the role of banks in industrial firm board interlocks was falling even more rapidly than that of industrial peers. Again, bankindustry interlocks involving a management board member fell very steeply, far more so than those involving two supervisory boards. A further decline in bank interlocks over the period 197696 is documented by Heemskerk, Mokken, and Fennema (2003), who find that finance-industry interlocks declined by almost 40 percent over that period, outpacing the 25 percent decline in overall interlocks. Table 8.7 compares and contrasts the prevalence of interlocks in family and nonfamily firms; the criterion used to define family firms in this table is a board member with a surname that matches the firm's original name. In 1928, the only significant difference was that members of the management board of nonfamily firms were much more likely to be on the board of other industrial firms. By 1958, this difference had largely disappeared, as nonfamily firms' board members became more like those of family firms. In 1993, the situation had reversed, as the management board
Characteristics of family firms
Family No family Family No family Family No family
Book value total assets Past three-year growth book value assets Tobin's q Return on assets Four-year standard deviation ROA Payout ratio Debt to total assets Fixed assets to total assets Cash and liquid assets to total assets Age RvB size RvC size Dummy certificates Dummy priority shares Dummy preferred shares Dummy structured regime Dummy interlock RvC-- RvC/industrial Dummy interlock RvC-- RvB/industrial Dummy interlock RvB-- RvC/industrial Dummy interlock RvB-- RvB/industrial Dummy interlock RvC-- RvC/bank Dummy interlock RvC-- RvB/bank Dummy interlock RvB-- RvC/bank Dummy interlock RvB-- RvB/bank Ownership largest outside blockholder Ownership all outside blockholders Ownership banks Ownership RvB members Ownership RvC members No. of observations
12,043 14,309 67,937 84,190 1,885,229 2,312,999
0.151 0.400 0.030
0.048 0.362 0.090
0.289 0.464 0.152
0.112 0.405 0.162
0.372 1.284 0.076
0.331 0.306 0.436
0.392 0.298 0.596
0.040 0.511 0.398 0.331
0.035 0.795 0.316 0.432
0.093 13.57 2.52 4.52 0.090 0.045
0.122 25.01 2.02 5.01 0.130 0.018
0.080 37.28 2.91 4.36 0.20 0.450
0.140 50.88 2.09 4.61 0.27 0.220
0.036 0.257 0.537 0.417 0.068 65.80 2.67 4.22 0.22 0.56 0.44 0.33
0.18 0.37 0.20 0.27
0.08 0.33 0.09 0.14
34.67 7.79 20.00 10.32
0.159 1.269 0.073 0.032 0.376 0.534 0.379 0.109 47.67 2.78 5.23 0.40 0.42 0.61 0.54 0.75 0.37 0.19 0.00 0.41 0.07 0.07 0.00 24.64 43.67 7.12 4.32 1.94 134
Significant at the 1 percent level. Significant at the 5 percent level. Significant at the 10 percent level.
Financing and Control in The Netherlands 499
Relations between takeover defenses and interlocks and ownership
Priority Preferred Structured
Certificates Certificates shares Certificates shares
Dummy certificates Dummy priority shares Dummy preferred shares Dummy structured regime Dummy interlock RvC-- RvC/industrial Dummy interlock RvC-- RvB/industrial Dummy interlock RvB-- RvC/industrial Dummy interlock RvB-- RvB/industrial Dummy interlock RvC-- RvC/bank Dummy interlock RvC-- RvB/bank Dummy interlock RvB-- RvC/bank Dummy interlock RvB-- RvB/bank Ownership largest outside blockholder Ownership all outside blockholders Ownership banks Ownership RvB members Ownership RvC members
1.000 0.079 0.093 0.218 0.144 0.113 0.071 0.087 0.134
1.000 0.188 0.120 0.080
0.188 1.000 0.009 0.012
0.120 0.009 1.000 0.323
0.080 0.012 0.323 1.000
0.079 0.012 0.066
0.050 0.016 0.069
0.207 0.074 0.074
0.205 0.061 0.166
0.084 0.005 0.165 0.038
0.195 0.241 0.006 0.057
0.173 0.153 0.022 0.083
0.024 0.198 0.273 0.085
Significant at the 1 percent level. Significant at the 5 percent level.
members of nonfamily firms became even less likely to hold supervisory board positions elsewhere. To complete our description of the prevalence of interlocks, table 8.8 illustrates their relation to takeover defense mechanisms. In both 1923 and 1958 interlocking directorships, especially those of the management board, show a strong positive association with certification of shares. By 1993 this was no longer the case. Instead, supervisory board crossdirectorships were associated with the use of defensive preferred shares, and most types of interlocks were associated with subjection to the structured regime, which may simply indicate that these are the larger and less multinationally oriented firms. As an exploratory enquiry into the impact of interlocks on industrial performance, in table 8.9 the second of each year's set of regressions considers the impact of interlocks on Tobin's q. The impact is insignificant in 1993, but in the two earlier years the association between interlocks and q
0.184 0.236 0.208 1.126 1.015 1.191
Log (book value)
0.061 0.062 0.028
Debt to total assets
(1.18) (1.61) (1.70)
0.846 0.841 0.858 0.951 0.948 0.958 0.028
(19.14) (20.10) (18.29) (18.49) (17.61) (0.11)
Past three-year growth book value total assets 0.015 0.016 0.019
Fixed assets to total assets
0.078 0.068 0.073 0.642 0.602 0.569
(0.30) (0.27) (0.12)
(2.41) (2.27) (2.09) (2.21)
0.009 0.009 0.008 0.018 0.020 0.017
Cash and liquid assets to total assets
(.87) (0.87) (0.87)
0.0001 0.002 0.002 0.003 0.002
(0.02) (0.25) (0.88) (1.05) (0.83)
Dummy priority shares
Dummy preferred shares
Dummy structured regime
Dummy interlock RvC--RvC/industrial Dummy interlock RvC--RvB/industrial Dummy interlock RvB--RvC/industrial Dummy interlock RvB--RvB/industrial Dummy interlock RvC--RvC/bank Dummy interlock RvC--RvB/bank Dummy interlock RvB--RvC/bank Dummy interlock RvB--RvB/bank Dummy family firm Adj. R2 No. of observations
0.005 (0.22) 0.033 (2.06) 0.016 (0.86) 0.003 (0.19) 0.011 (0.60) 0.027 (1.75) 0.015 (0.64) 0.008 (0.30) 0.002 (0.12) 0.78 193
Notes: White heteroskedasticity-consistent t-values. Observations with payout 0 removed. Significant at the 1 percent level. Significant at the 5 percent level. Significant at the 10 percent level.
0.009 (0.48) 0.015 (0.93) 0.014 (0.60) 0.042 (2.33) 0.005 (0.36) 0.012 (0.89) 0.033 (1.22) 0.041 (1.43) 0.002 (0.17) 0.74 224
0.072 (0.97) 0.06 139
0.012 (0.11) 0.002 (0.02) 0.130 (0.83) 0.117 (1.22) 0.041 (0.20) 0.514 (1.08) 0.057 (0.37) 0.08 139
502 Abe de Jong and Ailsa Rцell is negative whenever it is significant. This is weak evidence that interlocks, and especially those that involve management board members of other industrial firms or banks, were not beneficial in 1923 and 1958. 8.5 Conclusions Our paper gives a bird's-eye overview of financing and control of Dutch listed firms over the past century. Regarding the influence of families in firms, our data suggest a clear trend toward professional management taking hold in the second half of the twentieth century. The role of banks in the control and financing of Dutch industry seems to have been rather secondary, and more British than German in nature. While employees have been given some voice in corporate decision making in the last few decades of the century, again, their power is not as strong as in Germany. Real decision-making power currently seems to rest very strongly with a set of selfperpetuating management insiders, entrenched behind a quite formidable array of takeover defenses. But the ongoing process of convergence toward a common European model is slowly but surely eliminating some of the idiosyncrasies of Dutch corporate governance. Appendix Railway Finance in the Nineteenth Century The flotation of a number of railway issues in the middle of the nineteenth century seems to have been fairly easy, with the exception of the Nederlandse Centraalspoorweg Maatschappij in 1860, which many industry insiders realized in advance would be unprofitable because it did not connect major industries or population centers. These flotations are of additional interest because the disposition of the shares has been investigated, giving some insight into their initial ownership structure. Van den Broeke (1983, 1985) documents in detail how initial finance was raised. As a case in point, take the 1863 flotation of the Maatschappij tot Exploitatie van Staatsspoorwegen. The initial shareholders were 244 in number, holding a total of 24,000 shares of f 250 each (f 6 million in total). The largest stake reported by Van den Broeke is 3,000 shares held by a Paris bank, Hottinguer & Cie; the second largest, 2,765 shares, by Wurfbain en Zoon, an Amsterdam securities brokerage house. Four other stakes of 1,000 shares and above are mentioned, all held by banks in Amsterdam, London, and Brussels. There is no sign that any of these shareholders were motivated by a de-
Financing and Control in The Netherlands 503 sire to take a controlling stake in the venture: the largest stake was no more than 12.5 percent. The largest stakes were all held by banking houses or securities firms. Many of these were based abroad and therefore in no position to exercise meaningful control. In total, 74.5 percent of the capital was taken up by the banking/financial sector, and the Dutch banks never developed an active role in the management of industry as in the German model. Nor is there any sign that the government saw shareholding as an attractive means of ensuring control. The king and his entourage, and various politicians, government officials and members of the judiciary contributed for less than 600 shares in total. When efforts to raise a further f 6 million in the subsequent five years seemed to founder, the government repeatedly declined to step in and only came up with a loan of f 2.5 million, to be paid off as soon as new equity was raised.22 There seem to have been a couple of shareholders with direct commercial ties to the railway business: a shipping line connecting England to Vlissingen (Flushing), for example. 22. Even this very modest form of government
support was considered too much in some quarters, to judge by a pamphlet published in Breda in 1866, entitled May the money, that is contributed by the Dutch citizen as taxes, be lent to a private company for its own profit? A word to the Dutch people, by Someone (Mag het Geld, dat door den Nederlandschen Burger als Belasting Wordt Opgebragt, Worden Geleend aan eene Maatschappij van Partikulieren, Tot Haar Eigen Winstbejag? Een Woord aan het Nederlandsche Volk, van lemand).
Dutch railway finance
Hollandsche Ijzeren Spoorweg Maatschappij
Nederlandsche Rhijnspoorweg Maatschappij
Nederlandse Centraalspoorweg (NCS) Maatschappija
Maatschappij tot Exploitatie van Staatsspoorwegen
Year foundedb Initial equity capital (placed) Number of shares Nominal value No. of initial shareholders No. of stakes 2% Largest stake Largest shareholders and no. of shares held Foreign ownership Leverage (total capital)d 1850 1860 1870 1880 1890
1837 f 1.24 millionc 1,240 f 1,000 140 9 11.3% 140 stockbroker Amsterdam 110 stockbroker Amsterdam 79 stockbroker Amsterdam 66 stockbroker Amsterdam 65 stockbroker Amsterdam 49 stockbroker Amsterdam 35 manufacture Amsterdam 25 trade Amsterdam 25 unknown Amsterdam 2.4% 0.66 (f 10.8 million) 0.71 (f 11.1 million) 0.89 (f 17.2 million) 1.50 (f 37.5 million) 1.68 (f 60.4 million)
1845 f 24 million 100,000 f 240 35 9 14% 14,000 merchants London 14,000 merchants London 14,000 merchants/bankers London 14,000 manufacturers The Hague 14,000 manufacturers Haarlem 12,500 directors SWRail London 6,250 bankers London 6,250 bankers Liverpool 3,545 merchants Amsterdam 67.0% 0.49 (f 14.0 million) 0.38 (f 26.5 million) 0.41 (f 41.6 million) 0.64 (f 49.9 million) 0.84 (f 56.2 million)
1860 f 5 million (authorized), f 4.1 million (subscribed) 20,000 f 250 105 at least 2 at least 16% 3,200 railway construction Paris 1,600 businessmen Paris 90% -- -- 1.90 (f 10.0 million) 5.26 (f 9.9 million) 2.35 (f 11.4 million)
1863 f 6 million 24,000 f 250 244 9(?) 12.5% 3,000 bank Paris 2,765 stockbrokers Amsterdam 2,000 bank Amsterdam 2,000 bank London 1,730 bank Amsterdam 1,400 railway equipment company Utrecht 1,000 bank Brussels 500 bank Frankfurt 500 bank Basel 35.2% -- -- 0.75 (f 10.5 million) 0.43 (f 18.0 million) 1.31 (f 40.2 million)
Source: Van den Broeke (1985). aThe data for NCS are less meaningful because the initial offering was underscribed, and many initial shareholders subsequently reneged on their obligation to fully pay up. Stakes are those attending/represented at shareholder meetings. The French stakes were somewhat involuntary, as they were payments for construction services and materials. bThe year given is the date of incorporation and of raising capital (or attempting to) from the public. Concessions were generally granted a few years in advance to a small group of entrepreneurial individuals (concessionarissen) who then set up the company and raised capital. cRaised to f 6.2 million within the same year by a 4-for-1 rights issue that was heavily oversubscribed and entirely taken up by the initial shareholders. dDebt is measured as long-term debt; equity excludes reserves. Leverage is debt-equity ratio; total capital is the sum of debt and equity.
Financing and Control in The Netherlands 505 References Booth, J. R., and D. N. Deli. 1999. On executives of financial institutions as outside directors. Journal of Corporate Finance 5:22750. Bos, J. W. 1923. Over commissarissen van naamlooze vennootschappen. PhD diss., Rijksuniversiteit Groningen. Broeke, W. van den. 1983. De financiering en de financiers van de spoorwegen in Nederland: De Maatschappij tot Exploitatie van Staatsspoorwegen 18631870. Economisch- en Sociaal-Historisch Jaarboek 46:2644. ------. 1985. Financiлn en financiers van de Nederlandse spoorwegen 18371890. Zwolle, The Netherlands: Uitgeverij Waanders. ------. 1988. Vermogensstructuren en netwerkrelaties in het Nederlandse bedrijfsleven 18901940. Jaarboek voor de geschiedenis van bedrijf en techniek 5:15471. Camfferman, Kees. 2000. Jaarrekeningpublicatie door beursgenoteerde naamloze vennootschappen in Nederland tot 1910. Nederlandsch Economisch-Historish Archief (NEHA Jaarboek, 71103. De Jong, A., and P. W. Moerland. 1999. Beheersingsmechanismen in Nederland: Substituut of complement? Maandblad Accountancy en Bedrijfskunde 73:499 512. De Jong, A., P. W. Moerland, and T. Nijman. 2000. Zeggenschapsverhoudingen en financiлle prestaties. Economisch-Statistische Berichten 85:36871. De Jong, Abe, Doug DeJong, Gerard Mertens, and Charles Wasley. 2004. The role of self-regulation in corporate governance: Evidence and implications from the Netherlands. Journal of Corporate Finance, forthcoming. De Jong, Abe, Rezaul Kabir, Teye Marra, and Ailsa Rцell. 2001. Ownership and control in the Netherlands. In The control of corporate Europe, ed. Fabrizio Barca and Marco Becht, 188206. Oxford, UK: Oxford University Press. Ferris, S. P., M. Jagannathan, and A. C. Pritchard. 2003. Too busy to mind the business? Monitoring by directors with multiple board appointments. Journal of Finance 58:10871111. Frentrop, Paul. 2003. A history of corporate governance 16022002. Haarlem: Deminor. Gales, B. P. A. 1986. Werken aan zekerheid: Een terugblik over de schouder van AEGON op twee eeuwen verzekeringsgeschiedenis. The Hague: Aegon Verzekeringen. Goor, Linda van, and Jaap Koelewijn. 1995. Le systиme bancaire nйerlandais: Йtude rйtrospective. In Les banques en Europe de l'Ouest de 1920 а nos jours, ed. Maurice Lйvy-Leboyer, 15375. Paris: Ministиre de l'Йconomie et des Finances, Comitй pour l'histoire йconomique et financiиre. Heemskerk, E., R. Mokken, and M. Fennema. 2003. From stakeholders to shareholders? Corporate governance networks in the Netherlands 19761996. University of Amsterdam, Amsterdam School for Social Science Research. Unpublished manuscript. http://users.fmg.uva.nl/heemskerk/publications/stakeholder Preprint_2003.pdf. Heijden, E. J. J. van der. 1992. Handboek voor de naamloze en de besloten vennootschap. 12th ed. Adapted by W. C. L. van der Grinten. Zwolle: W. E. J. Tjeenk Willink. J. H. de Bussy. 1923. Financieel adresboek voor Nederland [Financial address book for the Netherlands]. Amsterdam: J. H. de Bussy. ------. 1958. Financieel adresboek voor Nederland [Financial address book for the Netherlands]. Amsterdam: J. H. de Bussy. ------. 1993a. Gids bij de officiлle prijscourant van de Amsterdamse Effectenbeurs
506 Abe de Jong and Ailsa Rцell 1993 [Guide with the official price list of the Amsterdam Securities Exchange 1993]. Amsterdam: J. H. de Bussy. ------. 1993b. Jaarboek van Nederlandse ondernemingen 199293 [Yearbook of Dutch companies 199293]. Amsterdam: J. H. de Bussy. Jonker, Joost. 1989. Waterdragers van het kapitalisme: Nevenfuncties van Nederlandse bankiers en de verhouding tussen bankwezen en bedrijfsleven, 1910 1940. Jaarboek voor de geschiedenis van bedrijf en techniek 6:15890. ------. 1995. Spoilt for choice? Statistical speculations on banking concentration and the structure of the Dutch money market, 19001940. In The evolution of financial institutions and markets in twentieth-century Europe, ed. Y. Cassis, G. D. Feldman, and U. Olsson, 187208. Aldershot, UK: Scolar Press. ------. 1996. Merchants, bankers, middlemen: The Amsterdam money market during the first half of the nineteenth century. Amsterdam: NEHA. Kroszner, R. S., and P. E. Strahan. 2001. Bankers on the boards: Monitoring, conflicts of interest, and lender liability. Journal of Financial Economics 62:41552. La Porta, R., F. Lуpez-de-Silanes, A. Shleifer, and R. W. Vishny. 2000. Agency problems and dividend policies around the world. Journal of Finance 55:133. Laveleye, Emile de. 1864. L'йconomie rurale en Nйerlande. Revue des Deux Mondes 49. Nederveen Meerkerk, Elise van, and Jan Peet. 2002. Een peertje voor de dorst: Geschiedenis van het Philips pensioenfonds. Amsterdam: Aksant. Post, J. G. 1972. Besparingen in Nederland 19231970: Omvang en verdeling. PhD diss., University of Amsterdam. Renooij, Dirk Cornelis. 1951. De Nederlandse emissiemarkt van 19041939. Amsterdam: J. H. de Bussy. S. F. van Oss. 1924. Effectenboek voor 1924: Binnenland [Securities guide for 1924: Domestic]. The Hague: S. F. van Oss. ------. 1959. Effectenboek voor 1959: Binnenland [Securities guide for 1959: Domestic]. The Hague: S. F. van Oss. Schijf, Huibert. 1993. Netwerken van een financieel-economische elite: Personele verbindingen in het Nederlandse bedrijfsleven aan het eind van de 19e eeuw. Amsterdam: Het Spinhuis. Sluyterman, Keetie E., and Hйlиne J. M. Winkelman. 1993. The Dutch family firm confronted with Chandler's dynamics of industrial capitalism, 18901940. Business History 35 (4): 15283. Steensgaard, Niels. 1982. The Dutch East India Company as an institutional innovation. In Dutch capitalism and world capitalism
, ed. Maurice Aymard, 23557. Cambridge: Cambridge University Press. Stokman, Frans N., Frans W. Wasseur, and Donald Elsas. 1985. The Dutch network: Types of interlocks and network structure. In Networks of corporate power: A comparative analysis of ten countries, ed. Frans N. Stokman, Rolf Ziegler, and John Scott, 11230. Cambridge, UK: Polity Press. Voogd, R. P. 1989. Statutaire beschermingsmiddelen bij beursvennootschappen. Kluwer: Deventer. Zanden, Jan L. van. 1987. De industrialisatie in Amsterdam 18251914. Bergen: Octavo. ------. 1998. The economic history of the Netherlands 19141995: A small open economy in the "long" twentieth century. London: Routledge.
Financing and Control in The Netherlands 507 Comment Peter Hцgfeldt It is natural to compare the Netherlands and Sweden since they are the two smallest countries surveyed in this volume, but also because their financial systems and corporate control structures have developed along different historical paths from initial points that could hardly have been further apart. The free city of Amsterdam provided the fertile ground for the first modern hub of international financial markets and advanced intermediation that, for example, helped underdeveloped Sweden to finance imperial wars against neighboring countries. The Swedish students, sailors, and businessmen who visited Amsterdam in the seventeenth and eighteenth centuries were attracted by the city's openness and dynamics. The philosopher Renй Descartes, who went the other way to enlighten the court of Queen Christina, was taken by the poverty, isolation, and coldness of Stockholm compared to the opulence and modernity of Amsterdam, even if the initial shock was not the immediate cause of his death in the winter of 1650 shortly after his arrival. When the political map of Europe was significantly redrawn during the following centuries, the relative decline of Amsterdam and the fast industrialization of peaceful but very poor Sweden after 1870 evened out the economic differences between the two countries. Today the two countries are small, open, and export-oriented economies dominated by very large transnational companies. But the conspicuous institutional differences between the two countries' financial systems and corporate control structures reflect the strong, historical path dependence of their developments. Abe de Jong and Ailsa Rцell's chapter illustrates this very nicely by painting a broad picture of the Dutch financial developments that is suitable for a comparative analysis with Sweden. They also present interesting analyses of how characteristics (for example, leverage, payout ratios, and Tobin's q) of firms listed on the Amsterdam Stock Exchange have changed over time and of interlocking board memberships. Given the thick veil of secrecy that by tradition protects Dutch firms, the authors have done an excellent job
when collecting their data. Even if God's hand may be in the details, I will focus my comparative analysis on three major characteristics of the Dutch financial system and discuss possible causal links in the historical development using Sweden as an alternative institutional setting. Peter Hцgfeldt is a research associate of the European Corporate Governance Institute (ECGI) and an associate professor in the Department of Finance, Stockholm School of Economics.
508 Abe de Jong and Ailsa Rцell The Limited Role of Banks in Corporate Financing For an outside observer perhaps the most surprising feature of the Dutch financial system is "the limited role played by the banks in the financing of industrial growth" throughout the industrialization as well as later. Although the banks were not prohibited from operating as universal banks, they specialized in the traditional short-term mercantile financing like British banks and stayed out of long-term, industrial financing, and did not hold equity stakes in their clients. This started a long tradition of bank noninterventionism in Dutch corporate governance. Instead, the industrial firms' demand for capital was satisfied via the prolongatie system of short-term callable margin loans collected from wealthy private persons through an old, local network of agents outside the banking system. At first sight this seems the most unlikely source of long-term industrial financing, but it was evidently competitive enough to preempt the banks from entering the market for a long time. The semi-market character of the decentralized system seemed to have circumvented the general public's distrust of financial institutions due to very bad experiences in the past. The wealth accumulated in the Golden Ages and controlled by wealthy rentier families thus financed investments several centuries later. This is a nice example of path dependence in the development of a financial system, despite the turmoil created by the Napoleonic Wars. The Dutch case thus shows that development of a universal banking system is not necessary for financing of industrial growth when the private, pecuniary wealth of a country is in the hands of a small but sufficiently large number of wealthy families that also have access to nonintermediated networks for investments. The story becomes more intriguing when we compare it to what happened in Belgium, the industrialized and relatively poorer southern part, after the separation from the Netherlands. Despite initially having the same legal origin and identical banking laws as well as the same corporate law, universal banking became the Belgian solution, and, unlike in the Netherlands, pyramiding became extensive. Although the stronger French influence in Belgium points in the direction of universal banking, I conjecture that the lack of a larger class of very wealthy families combined with a substantial demand for long-term industrial investments explains why collection of savings from the broader base of all people in the society via a universal banking system became the solution. A well-developed deposit banking system may collect the necessary savings at a relatively low cost but may not be able to intermediate efficiently without developing special competencies in industrial financing. Direct equity ownership may be one way to make the monitoring of clients more efficient, but recurrent industrial crises may also explain why banks became owners of large equity positions. When dual-class shares are prohibited, pyramiding may be an at-
Financing and Control in The Netherlands 509 tractive solution to control firms for both banks and other interests, as the Belgian example shows. The development of Swedish banking from a pure deposit system to a universal banking system when the demand for industrial financing increased substantially at the beginning of the twentieth century illustrates this very nicely. Lacking a sufficiently broad base of wealthy people, the national political debate was about how to collect the citizens' savings via a national banking system in order to finance the necessary industrial investments; the infrastructural investments were primarily financed via international public bonds. Down the road the banks later became the controlling owners of the largest industrial firms in Sweden via pyramiding, most often combined with the use of dual-class shares. If the financial system in general and the stock markets in particular are underdeveloped due to lack of a sufficiently large group of wealthy individuals at the early stage of industrialization, corporate financing via universal banking may thus become the dominating interface instead of stock marketbased corporate financing. While universal banking seems to point in the direction of bank and shareholder control via separation of ownership from control through use of either dual-class shares or pyramiding or both, the Dutch case without universal banking seems to lead to dispersed ownership but with very entrenched managerial control via use of other legal devices to separate ownership from control when dual-class shares are prohibited. It thus seems to be the combination of a wealthy class of investors with investment options outside the banking system, and the prohibition of dual-class shares that in the Dutch case implies managerial control with dispersed ownership. The passive rentier attitude of Dutch equity investors has perhaps reinforced this effect. By being counterfactual to the convention in Continental Europe, the Dutch case thus in effect supports the causal link between universal banking and shareholder corporate control via separation of ownership and control. I thus conjecture that when the initial level of wealth in a country is low and nonintermediated forms of financing are rare or nonexistent, intermediation of industrial financing via a universal banking system is more likely to occur, which seems to later imply shareholder control via strong separation of ownership and control. The Dutch case seems to be the exception that supports the generality of this conjecture. The Bulwark of Takeover Defenses While it seems relatively straightforward to identify when the Dutch tradition to use very elaborate entrenchment devices started, it is much more difficult to understand why it happened. It would have been very interesting to have more information and analysis of the political background to the corporate law of 1881, and to know why it did not permit "the use of nonvoting or lower-voting shares, thus ruling out one obvious mean of de-
510 Abe de Jong and Ailsa Rцell taching control from ownership," in particular since the future corporate laws have consistently used the same design principle. Since the earlier corporate law of 1838 that mandated voting caps in order to protect minority shareholders and also became the standard feature of later corporate laws, there seems to have been a common underlying principle to limit the power of large block shareholders. But why did this early and strong aversion against shareholder control occur? Was it a delayed reaction to the manifest and persistent expropriation of noncontrolling shareholders in the Vereenigde Oostindische Compagnie (VOC)--the Dutch East India Company--or was it an attempt to protect the Dutch firms from hostile takeovers by German and French firms, as the authors suggest? The latter alternative is less plausible as it does not explain why the use of dual-class shares, the single most efficient antitakeover defense, was prohibited or what would stop the new controlling owners from getting around the law by taking the firm private or forming a merger. Or was it because the ordinary shareholders viewed themselves as passive rentiers primarily interested in dividends that are shared in proportion to their capital contribution, and not in corporate control? It seems as if the Dutch investors behaved more like long-term bondholders rather than as typical shareholders because of their large and old private wealth. I have no specific answer to these questions, but it seems pivotal to understand why if we want to understand the current Dutch control structure since the first corporate laws that carefully limited the larger shareholders' opportunities to maintain and exert corporate control set the standard for the future laws and, thus, shaped the future control structures through path dependence. For example, were the legislators and the larger shareholders, who probably exerted political influence, (fully) aware at the time that the corporate law down the road opened up for co-optive managerial control protected by a plethora of antitakeover devices, which became a Dutch specialty long before U.S. lawyers perfected it in recent decades with the help of the state of Delaware? If they were aware of the consequences, why the manifest intention to limit the power not only of the controlling shareholders but also de facto of the noncontrolling ones? The relatively strong protection of minority shareholders against expropriation by controlling shareholders in the Dutch corporate laws thus implies both (a) (relatively) dispersed ownership and (b) very strong managerial entrenchment via direct control often without ownership. The first implication appears consistent with Burkart, Panunzi, and Shleifer's (2003) legal minority protection theory of corporate ownership, while the second is inconsistent since the shareholders are poorly protected against managerial expropriation as well as against inefficient decisions by the management (agency costs); hostile takeovers are hardly efficient threats! The missing nonlegal element in the Dutch case seems to be that banks were
Financing and Control in The Netherlands 511
passive and did not get involved in long-term industrial financing and corporate ownership, which opened up for managerial control instead. Hence, strong legal protection of minority shareholders does not rule out exceptionally strong managerial entrenchment through co-option. There is a fundamental difference between (a) mechanisms that primarily protect the large shareholders' interests by separating ownership from control, like dual-class shares, pyramiding, and cross-shareholding, and (b) other devices like the administratiekantoor, preference shares, the X rule, voting limits, and so on that entrench management control by diluting and limiting the value of shareholder control. The first type of defenses implies reinforced shareholder control via separation of ownership and control, while the latter type opens up for and supports managerial control by weakening shareholder control. The two sets of protective mechanisms are, however, substitutes in the following sense: if a Swedish or a U.S. IPO firm uses dual-class shares it uses none or very few of the other mechanisms, while firms without such shares often have a long list of complementary antitakeover defenses, but not as extensive a list as the Dutch arsenal; see Field and Karpoff (2002) and Holmйn and Hцgfeldt (2004). The empirical results thus imply that because of the long-standing prohibition of dualclass shares, the single most powerful mechanism against takeovers, the Dutch firms use a diversified portfolio of other, weaker antitakeover defenses; pyramiding is of limited use since it supports shareholder control. As a comparison, table 8C.1 shows the extreme simplicity and trans-
Use of control mechanisms by controlling owner (%; from Agnblad et al. 2001)
Dual-class Right of
Voting Mandatory Shareholder
Sample (% of total)
shares preemption restriction bid rule
Whole sample (100)
Buyout investor (1)
Mutual fund (6)
Source: Aktiemarknadsbevakning (AMB), Sundin and Sundqvist (1998), company charters, and Patent- och Registreringsverket (PRV). Notes: Table shows frequency of different control mechanisms for 304 firms listed on the Stockholm Stock Exchange and the Stockholm Bцrsinformation list in October 1998. The sample is split into subsamples based on the characteristics of the controlling shareholder and type of mechanism.
512 Abe de Jong and Ailsa Rцell parency of the control structure of Swedish firms listed on the Stockholm Stock Exchange; see Agnblad et al. (2001). The dual-class share design is the most commonly used mechanism to control firms, in particular for family firms, since 63 percent of the listed firms use it. Only 13 percent of the firms have the right to preemptively redeem nonlisted A-shares that have been passed on to a new owner. This is the second most common control mechanism; the others are very infrequently used. The use of dualclass shares to separate ownership from control has very strong political support in Sweden since it is a very efficient protection against foreign takeovers; see Holmйn and Hцgfeldt (2004). Another important element of the Dutch control structure is the strategic use of co-optive (family) foundations that by legal design are very opaque, almost impenetrable for an outsider like the tax authorities, and not subject to taxation on corporate dividends. This is an important advantage since dividends to regular shareholders are disadvantaged by a relatively high tax, which implies a low payout ratio of profits and support for use of retained earnings as a primary source of financing. Since shareholders discount the levered and opaque control structure when the firms need to raise external capital from the capital markets, the relatively high tax on dividends seems like a logical element of the Dutch control structure to lock in capital into the existing firms. The strong legal support and protection of the very secretive foundations is another example of the extreme nature of the Dutch control structure. The historical preference for secretive, private decision making among a small number of business partners and for co-optive control already in place in the golden days of Amsterdam seems to have been propagated through time and taken to its extreme--another interesting example of path dependence. The strong aversion for centuries to making annual reports and accounting information available to general shareholders and to the public is another example. It is thus logical that the founding families of the two most successful firms founded in Sweden during the last fifty to sixty years, IKEA (the Kamprads) and Tetra Pak (the Rausings), have moved their fortunes out of reach of the Swedish tax authorities and have kept their firms fully private by using the very favorable Dutch legislation for private foundations as holding entities before paying out rents to personal foundations for the family members in Liechtenstein. The heavy entrenchment, very comfortable secrecy (no questions asked), and low or nonexistent taxes on corporate dividends are thus very convenient features of the Dutch foundations, also for foreign families interested in locking in control for generations. The historical irony in the case of the Netherlands is that the strong public aversion against corporate power in the hands of large shareholders as well as in the hands of financial institutions like banks has generated
Financing and Control in The Netherlands 513 perhaps the most extreme concentration of corporate power in the hands of very heavily entrenched and co-optive management teams and foundations that operate behind a legal veil of secrecy. But where did all the shareholders go, and why did they give up their power so easily without a fight? The answer seems to lie in the original character of the Dutch financial system and the strong historical path dependency in its development. The Politics of Corporate Control Since political ideologies and decisions shape and affect the development of a country's corporate control system, a deeper understanding requires an analysis of how politics and corporate financing interact. The public acceptance of a corporate governance structure in a society ultimately depends on its politically viability; without manifest political support, an extreme control structure will not survive. An analysis of the politics of Dutch corporate governance would thus have been even more interesting. The authors have, however, decided to leave this out of their already very rich chapter. But they stress that the small Dutch welfare society, like the even smaller Swedish society, is strongly consensus oriented. Despite this political affinity, the two countries' control structures have developed along different paths over the past thirty to forty years. There is, however, a common theme. The vigorous political ambitions in both countries since the late 1960s to reform the traditional control structures and make them more "democratic" by giving firms' stakeholders more voice has had the opposite result: the entrenchment of the controlling interests has increased in recent decades. But the two countries followed very different roads. In the Netherlands the trade unions did not unexpectedly join forces with the management and short shrift the shareholders by transferring pivotal decision-making powers from the shareholders' annual meeting to a co-optive (corporative) supervisory board dominated by management and labor appointees. The already entrenched managerial power was thus reinforced by a political measure that was supposed to achieve the opposite. Were there any strong political protests voiced against the structured regime, or was it done in consensus behind a veil of secrecy? It would have been interesting to know how this actually happened since it seems to have been a relatively recent, pivotal event. In Sweden employees were granted formal representation in the boards but with very limited decision-making power; the primary motivation was to have access to pertinent corporate information and an opportunity to give voice. The traditional skepticism toward managerial capitalism because of its perceived short-sightedness, combined with the political consensus between the leading capitalism and the Social Democrats, instead resulted in stronger political support for the incumbent owners in control. The new corporate law as well as the political rhetoric stressed the pivotal importance of firms having well-defined and strong owners in control; in-
514 Abe de Jong and Ailsa Rцell creased use of dual-class shares was the primary means to obtain the objective. A more realistic and sinister objective was to ascertain that the leading listed firms remained under Swedish control when capital markets became deregulated and capital demands increased as the international competition became more vigorous. The two consensus-oriented welfare societies thus handled the new historical situation very differently by (not unexpectedly) reinforcing the incumbent management in the Netherlands and the controlling shareholders in Sweden. The path dependence in the development of the two countries' control structures was therefore reinforced rather than weakened. The longer historical perspective that pinpoints the path dependence, however, also accentuates rather than moderates the impression of how much has happened in recent decades! In both countries, however, the historical compass points in the direction of more entrenched control structures rather than toward more flexible ones. Final Thoughts The case of the Netherlands is very interesting by itself because Amsterdam gave birth to the first modern, advanced financial system. I think, however, that the Dutch case is even more interesting since it nicely illustrates how its historical roots via path dependence have shaped future developments without making the outcomes predictable: a mixture of random and nonrandom factors representing Anglo-Saxon, German, and French influences of a political, legal, and economic nature has affected the actual path followed by a small country at the geographical crossroads. For example, there was managerial control of the largest listed firms like in the United States, although with a distinctive Dutch control twist, and a banking system that focused on short-term mercantile financing as in the United Kingdom rather than universal banking with long-term industrial financing as in continental Europe. The Dutch financial system is thus not a clean example that easily fits into the civil law country camp. The standard dichotomy is simply too coarse when we really want to understand the development and characteristics of the Dutch financial system; the differences versus other civil law countries like Sweden are simply more interesting than the similarities. It is thus not surprising that the EU, dominated by civil law countries, has failed conspicuously to harmonize takeover codes and eliminate antitakeover defenses despite ambitious attempts. The comparison between the Netherlands and Sweden, however, shows that the developments of the national control structures over time have a common element--the strong historical path dependence since established control structures reproduce and even reinforce themselves over time despite changing conditions--but the Dutch case takes it to the extreme. Civil law countries seem to be conducive to such dependencies and causalities since their political organization and decision-making pro-
Financing and Control in The Netherlands 515 cesses are often very centralized. The early political support in the Netherlands for prohibition of dual-class shares in the corporate law as well as of pyramiding seems down the road to (logically) imply managerial control protected by a plethora of antitakeover defenses. There is significant political support in Sweden to instead allow and even encourage the use of control mechanisms that rigidly separate votes from capital, which points in the future direction of maintained shareholder control via increased separation of ownership from control over time but without using any other special antitakeover devices. It is thus not surprising that large listed firms in the Netherlands are controlled by co-optive management teams while in Sweden the controlling owners ultimately make the pivotal decisions. Since the Dutch case is an extreme exception to the typical continental European corporate control structure, I am still puzzled by three enigmas: First, why have shareholders passively accepted that their control powers have been transferred to co-optive and heavily entrenched management teams often without direct ownership? Second, how efficient is such a rigid and opaque corporate control structure over time, in particular when subject to structural changes in a competitive international environment? And third, why were dual-class shares prohibited in the first Dutch corporate laws, and how has this affected the development of the Dutch corporate control structure? References Agnblad, Jonas, Erik Berglцf, Peter Hцgfeldt, and Helena Svancar. 2001. Ownership and control in Sweden: Strong owners, weak minorities, and social control
. In The control of corporate Europe, ed. Fabrizio Barca and Marco Becht, 22858. Oxford: Oxford University Press. Burkart, Mike, Fausto Panunzi, and Andrei Shleifer. 2003. Family firms. Journal of Finance 58:21672202. Field, Laura C., and Jonathan M. Karpoff. 2002. Takeover defenses at IPO firms. Journal of Finance 57:185789. Holmйn, Martin, and Peter Hцgfeldt. 2004. A law and finance analysis of initial public offerings. Journal of Financial Intermediation
13:32458. Sundin, Anneli, and Sven-Ivan Sundqvist. 1998 Owners and power in Sweden's listed companies. Stockholm: Dagens Nyheter.
A De Jong, A Roell