The Role of Self-Regulation in Corporate · PDF fileThe Role of Self-Regulation in Corporate...

The Role of Self-Regulation in Corporate Governance:Evidence from The Netherlands

Abe de JongDepartment of Finance

CentER for Economic ResearchTilburg [email protected]

Douglas V. DeJong*Tippie College of Business

University of [email protected]

Gerard MertensNyenrode University

[email protected]

Charles WasleySimon Graduate School of Business Administration

University of [email protected]

Current Draft: December 1, 2000Comments Welcomed.

Keywords: international economics, financial economics, law and economics, corporategovernance

JEL Classification Numbers: F36, G38, K22

* Corresponding author.

We thank Sudipta Basu, Bernard Taylor and Jerry Zimmerman; participants at the“Convergence and Diversity in Corporate Governance Regimes and Capital Markets”conference (Eindhoven, Netherlands), 1st European Conference on Corporate Governance(Brussels, Belgium), 11th Annual Conference on Financial Economics and Accounting and7th Mitsui Life Symposium on Global Financial Markets; and finance workshops at theUniversity of Iowa and Tilburg University. Part of this research was carried out while Abede Jong was visiting Florida State University and he gratefully acknowledges financialsupport from The Netherlands Organization for Scientific Research (NWO R46-406).Charles Wasley acknowledges the financial support of the Simon Graduate School ofBusiness Administration at the University of Rochester.


ABSTRACT

The purpose of this paper is to assess the effectiveness of self-regulation at promoting theinterests of investors. For two reasons the Netherlands provides an excellent opportunity togather such evidence. First, characteristics of the Dutch corporate structure have made it thefocus of attention by the European Union, the International Monetary Fund and countriessuch as Korea when deliberating issues of corporate form and governance. Second, duringthe period 1996-1998, a private sector initiative was undertaken to promote change in thebalance of power between management and investors. Not surprisingly, the Securities andExchange Commission has closely followed the Dutch “experiment” in self-regulation. Ourmaintained hypothesis is that shareholder rights are a crucial component for the success ofself-regulation. We identify corporate governance characteristics that are linked to firmvalue and find that the private sector’s recommendations had no effect on thesecharacteristics or their relationship with firm value. Event study techniques further documentthe market’s skepticism about the evolution of corporate governance practices in theNetherlands. Our results document the importance of shareholder rights when considering thedesign of a successful self-regulation process.


1. Introduction

Increased international competition and the integration of Europe through the EuropeanUnion have focused attention on industries and companies that are restructuring to meetcompetition and to promote economic growth. Accompanying this restructuring is a demandfor capital to finance such activities. With limits to conventional sources of capital (e.g.,banks and governments), attention has shifted to capital markets. Within capital markets,corporate governance plays a crucial role in determining where, in what form, and at whatcost capital is provided by outside investors (e.g., Price Waterhouse, 1997, Shleifer andVishny, 1997, La Porta, Lopez-de-Salanes, Shleifer and Vishny, 1998, and Financial Times,20 June 2000).

It is well known that agency problems are associated with the separation of ownership andcontrol in corporations (Berle and Means, 1932, Jensen and Meckling, 1976, and Fama andJensen, 1983a and 1983b). To mitigate these problems, corporate governance mechanismshave evolved to help ensure that investors earn competitive rates of return. Corporategovernance can be viewed as a mechanism design problem that is economic, legal, andpolitically based. As such, it is of interest to determine when market forces are sufficient topromote change as well as when legal/political actions are required to write and enforcecontracts between owners and managers of capital (Alchian, 1950, Stigler, 1958, and Shleiferand Vishny, 1997).

The purpose of this paper is to assess the effectiveness of market forces to promote investorinterests through self-regulation. The Netherlands provides an excellent opportunity toaddress this issue for two reasons. The first reason is its private sector initiative in corporategovernance. In 1996, a Committee on Corporate Governance was formed based on anagreement between the Association of Securities Issuing Companies and the AmsterdamExchanges. Chaired by J. Peters (retired CEO of Aegon), the committee’s members includedrepresentatives from the business community, Amsterdam Exchanges, security issuingcompanies, academics and a platform of investors (stockholder and pension representatives).The charge of the Peters Committee was to initiate debate and change in the balance ofpower between a firm’s management and investors.

In June 1997, the Peters Committee issued its recommendations designed to increase theeffectiveness of management, supervision and accountability to investors in Dutchcorporations. A key element of the report was its reliance on self-enforcement, throughmarket forces, to implement and enforce its recommendations. One year after the effectivedate of the report, the Committee initiated and then completed a project to assess the impactof the report (Monitoring Corporate Governance in Nederland, 1998).

2

The second reason for focusing on the Netherlands is the intense international interest shownby investors and policymakers in this private sector initiative and the characteristics of theDutch corporate governance structure, itself. Specifically, the Dutch economy isinternationally focused, heavily influenced by international competition, but not well knownfor strong investor rights.1 However, because of its perceived ability to balance alternativeinterests within the firm, the Dutch structure is considered by many as a prototype for theEuropean Union (e.g., Financial Times, 27 July 2000). Outside Europe, the InternationalMonetary Fund recommended, as well as offered to fund, a project for Indonesia with theDutch corporate governance model as part of its focus. The Korean government has usedthe Peters Committee report, itself, during its own deliberations on corporate form andgovernance. Finally, the Securities and Exchange Commission has closely followed theDutch “experiment” in self-regulation.2 From this interest, it is straightforward to concludethat the success or failure of self-regulation to promote and implement corporate governancechanges that enhance firm value is an important issue to regulators.

To evaluate the impact of the Peters Committee, we use Tobin’s Q to measure firmperformance. We gather data for the companies listed on the Amsterdam Exchanges over thefive years prior to as well as one-year after the Peters Committee report. The data includeinformation on a company’s organizational-form, voting rights, board characteristics, outsideblock-holders and debt characteristics. We also use event study techniques to assessinvestors’ reactions to the various events associated with changes in corporate governancepractices in the Netherlands.

Our maintained hypothesis is that shareholder rights are a crucial component for the successof self-regulation. When domestic firms in the Netherlands reach a certain size, they arelegally required to organize as a structured regime. This regime requires a supervisory boardcomprised of outsiders and this board takes numerous powers from shareholders. Forexample, the supervisory board elects the members of the management board (i.e.,management) as well as electing its own members. Due to the greater separation ofownership from control, we hypothesize that the structured regime has a negative relationwith firm value. The results support this hypothesis; average firm value measured byTobin’s Q is reduced by 0.22 (from 1.40 to 1.18) under the legally required structuredregime.

Based on a monitoring hypothesis, we hypothesize and find that major outside shareholdershave a positive effect on firm value. We also find that takeover defenses and limits onshareholder voting rights have positive effects. However, the interaction of the two has asignificant negative effect on firm value. While the benefits to major outside shareholdersand takeover defenses dominate their interaction, it nevertheless suggests that outsideowners, management and supervisory board members entrench themselves by adopting

1 When comparing stock exchange capitalization to annual GDP (gross domestic product), the Dutch economyranks sixth among developed countries in 1996 (Committee on Corporate Governance, 1997).2 Remarks by Jaap Peters in his talk at the international conference “Convergence and Diversity in CorporateGovernance Regimes and Capital Markets,” Eindhoven, Netherlands, November 4-5, 1999.

3

takeover defenses. Institutional investors fail in their monitoring role since they have anegative effect on firm value.

Our results next suggest that, to date, the recommendations of the Peters Committee have noteffected governance practices, themselves, nor have governance practices changed in a waythat has led to an increase in firm value. Our event study results also suggest that the marketis pessimistic about the substantive evolution of corporate governance practices in theNetherlands. Moreover, the market appears skeptical of the underlying premise of the PetersCommittee, that in the Netherlands market forces via self-regulation are sufficient to promotechanges in corporate governance that lead to increases in firm value. These results are astark contrast to those from the UK’s more limited initiative in self-regulation, the CadburyCommittee. Dedman, 2000, and Dahya, McConnell and Travlos, 1999, document significantchanges in board structure and management characteristics following the CadburyCommittee’s recommendations as well as an increase in the average performance of the firmsin the Dahya et. al. sample. Stiles and Taylor, 1993, further document that significantchanges took place within one year of the Cadbury Committee’s report andrecommendations. The different outcomes of the Dutch and UK initiatives are predicted byHart, 1995, who stated that market forces are required for changes to take place with self-regulation. A necessary condition for market forces is shareholder rights. Thus, thedifferential outcomes are consistent with our hypothesis since the UK has strong shareholderrights and the Netherlands does not.

From a mechanism design perspective, what is necessary for market forces to succeed in theNetherlands? One of the Peters Committee recommendations provides the starting point.Namely, an appeal to reevaluate the numerous constraints placed on the rights ofshareholders. Voting limitations should be reevaluated to increase the day-to-dayaccountability of the management and supervisory boards. Evidence suggests that this ismore efficient than relying on the market for corporate control to improve governance andperformance (Franks and Mayer, 1996, and Gugler, 1999). However, when conventionalmonitoring fails, direct/indirect takeover defenses should also be reevaluated so that themarket for corporate control is allowed to function. Since supervisory and managementboards effectively control a company’s voting rights (via the structured regime anddirect/indirect takeover defenses), it is doubtful that market forces can succeed withoutlegal/political action to restore the voting rights of shareholders.3

The next section presents a brief description of the Dutch corporate structure while section 3presents our hypotheses pertaining to the relation between corporate governancecharacteristics and firm value. Section 4 describes our data, section 5 reports our results andsection 6 concludes.

2. The Dutch corporate structure2.1. Legal structure and stakeholder rights

3 In general, La Porta, Lopez-de-Silanes, Shleifer and Vishny, 2000, are skeptical about substantivelegal/political action because of the intense opposition based upon the self-interest of the parties involved andthe lack of appreciation for the importance of investor rights.

4

Current Dutch company law was enacted in 1971, after a government committee (VerdamCommittee) issued a proposal for company law reform in 1965 and a draft law based on thereport in 1968. 4 At the time, there was a perceived monitoring problem within largecompanies and a desire to increase management’s accountability to the company’s broaderset of “stakeholders” (i.e., investors, employees and the general public). The law addressedthese two concerns by altering the legal form of large companies via the creation of the“Structured Regime" (see below). It also broadened accountability to include the possibilityof three boards, a supervisory board, a management board (technically named the board ofdirectors) and a works council, each with defined responsibilities.

As a starting point, we consider a shareholder controlled firm with a supervisory board andmanagement board. Shareholders elect members of the supervisory board and managementboard as well as approve the annual accounts. Dividend policy is set by management withthe consent of the supervisory board and formally approved by shareholders. Shareholdersalso vote on such issues as mergers and acquisitions. All votes are taken at the annuallyscheduled General Meeting of Shareholders and physical presence is required (voting byproxy is not part of the Dutch structure).

Once a company attains a certain size, it adopts the Full “Structured Regime” which islegally required for Dutch companies with more than 100 employees, a legally installed workcouncil and book value of shareholders’ equity in excess of NLG 25 million (about US$12.5M). Independent of the structured regime, the law requires a works council when acompany has more than 100 employees. The full structured regime requires a supervisoryboard that takes over the following powers from shareholders: establishing (and by defaultthe approval of) the annual accounts, the election of the management board, and the electionof the supervisory board itself (called co-optation). The supervisory board also has authorityover major decisions made by the management board.

There are exceptions to the legal requirements for the full structured regime. The prevalentexception is Dutch multinationals with more than 50% of their employees outside theNetherlands. Such companies file and obtain an exemption from the full structured regime.However, a Dutch company may then voluntarily retain the full structured regime, and Dutchmultinationals typically do, even though a company is not legally required to do so.5

4 The Verdam Committee referred to the situation as “no longer acceptable”; inadequate control ofmanagement’s activities led to their propensity to misstate the firms’ financial position and to violate theposition of shareholder, debtholders and employees (Verdam Committee, 1965, p. 119-125). This was themajor drive to restructure company law (Mertens, 1997). Slagter (1996) documents the desire for more co-determination (medezeggenschap); meaning that all stakeholders interests should be represented in a fair way.The law also dealt with financial reporting requirements, the right of inquiry, a works council and theestablishment of the enterprise chamber at the Amsterdam court (Zeff et al, 1992, p.171- 181).5 Companies required to apply the structured regime have statutes detailing the exact rights and duties of thesupervisory board. If a company no longer meets these criteria (e.g., due to its international scope) and it wantsto change to another organizational form, its statutes must be changed. The management board, supervisoryboard or the annual shareholders meeting may suggest a change in the statutes. However, the supervisory boardstill has most of the legal powers and shareholders usually have a limited say in this. This could be one of thereasons why a relatively large number of the largest publicly listed companies apply the structured regime on avoluntary basis.

5

Turning to the two boards and the works council, a Dutch company operates under a two-tiermanagement structure consisting of a Supervisory Board and Management Board.6 Thesupervisory board is “independent” of the company and comprised entirely of “outsiders.”These outsiders primarily consist of “professional managers” and can (and often do) includepast members of management. With the rare exception of a retired politician, politicians andregulators are not members of a supervisory board. Board members receive a fixed payment,depending on the firm’s size, for their services and very few hold shares in the company.Thus, reputation is important for getting and keeping such positions, which probably makessupervisory board members risk averse. The law requires that the board serve the firm’sinterest. However, under the structured regime, the supervisory board has very fewrestrictions on its ability to determine its own composition, re-appointments and otherorganizational matters including the management board. The law requires that themanagement board serve at the pleasure of the supervisory board.

The management board consists of the company’s management team and may be as small asone member, the president. The management board reports to the supervisory board and isresponsible for attaining the company’s objectives, its strategy and policy, and the ensuingresults. Labor is not required to have an “outside” representative on the supervisory boardnor is labor a member of the management board (Company Law of 1971). The legallyinstalled works council (noted above) has a right to relevant information and to advise onsuch major issues as transfers of ownership, plant closings and major investments. While thisis more than a formality, the management board decides and can overrule the advice of theworks council. The works council’s permission is only required for changes in socialarrangements (e.g., pensions, working hours, wages, safety rules). If the council disagreeswith the company’s proposals on social arrangements, the company must obtain a localjudge’s decision to proceed.

2.2. Ownership and voting rights

At the time of its organization, a company has an authorized capital structure consisting of“common” shares. Once issued, the shares are registered with the company. Such“registered” (or ordinary) shares have voting, dividend and trading rights. When thecompany’s organization and size require the full structured regime, the supervisory board isgranted the rights (detailed above) previously held by shareholders. Shareholders still voteon mergers and acquisitions and dividend policy under the structured regime. However,ordinary shareholders can have their voting power directly reduced with “Limited votingpower.” Limited voting power which is normally set at 1% implies that a shareholder canonly have a maximum of 1% of the votes irrespective of the number of shares owned.

A company can also have a second type of security called “Certificates.” In fact, under thestructured regime, the supervisory board can request the exchange of ordinary shares for

6 The creation of the supervisory board is arranged by company statute. In fact, many large companies hadalready installed a two-tier system with a supervisory board before 1971 (Slagter, 1996).

6

certificates.7 A Trust Office administers the certificates when issued or initiates acertification process where certificates are exchanged for ordinary shares. The trust office iscomprised of members from the company (supervisory board and management board) andthe “outside” (not from the company). While the chairman and majority of the trust officemembers must be outsiders, in practice, the trust office is always friendly to existingmanagement. The trust office is given responsibility for the ordinary shares associated withthe certificates. Through the process of certification, legal, but not “economic” ownership ofthe ordinary shares is transferred to the trust office (Slagter, 1996, p.210). Certificate holdershave dividend rights, can freely trade their certificates and can attend the General Meeting ofShareholders, but they cannot vote. The trust office holds all voting rights includingapproval of the dividend policy.8 Individual shareholders do not have a formal obligation toexchange their ordinary shares when a company decides to issue certificates.9 However, thecompany may decide to de-list its ordinary shares from the exchange, leaving a shareholderwith the choice of non-listed bearer shares or listed certificates (the latter being much easierto trade). The prevailing type of Certificate is the limited exchangeable certificate. Onceissued, these certificates can be exchanged for ordinary shares up to a maximum percentageof 1% of outstanding equity capital. However, once exchanged for ordinary shares, holdersloose trading privileges for the exchanged shares. Ordinary shares can be reconverted tocertificates, but then voting rights are lost.

As takeover defenses, companies may have additional types of securities in their authorizedcapital structure. The most common takeover defense is “protective preference shares.”Management can issue such shares to a friendly trust office or outside investor during ahostile takeover threat. Preference shares are sold at nominal value to the trust office orfriendly investor with an obligation to pay only 25% of the amount up front. Preferenceshares have voting rights and are restricted to a maximum of 50% or 100% of the currentoutstanding nominal capital depending on the anti-takeover amendments in place. Specialvoting privileges are also granted through “Priority shares” which give their holders specialrights in situations such as merger approval, new public offerings, charter amendments andcompany liquidation.10

If a company wants its shares or certificates traded on the Amsterdam Exchanges, there arerequirements to be met. The three most relevant for our study are minimum size, 7 The supervisory board’s authority has limits even under the structured regime. For example, if an individualowns 66% of the certificates (which translates into 66% of the votes), the individual can force the board toconvert certificates to ordinary shares.

8 Under all organizational forms, dividend policy is set by management with the consent of the supervisoryboard.

9 According to the law, it is possible for a company to have both ordinary shares and certificates outstanding.However, the Amsterdam Stock Exchange does not allow this possibility.

10 In situations outside the structured regime, priority shareholders have extensive influence. They can initiate abinding nomination for candidates to the board of directors and supervisory board at the general meeting ofshareholders. The nominations can only be defeated by a supermajority of votes present (or 50% of theoutstanding capital).

7

profitability and constraints on takeover defenses (Amsterdam Exchanges, 1997). Forlisting, the company’s book value of shareholders’ equity must be greater than NLG 11million (about US $5M) and it must have been profitable in three of the five years prior tolisting. Finally, the company can have only two of the three takeover defenses noted above(certificates, priority shares and protective preference shares). Nothing is said about “limitedvoting power.”

3. Corporate governance and firm value

Our main focus is organizational form and voting rights characteristics. The hypothesizedrelations with firm value are detailed below. To isolate the impact of these variables we mustalso recognize the monitoring role of major outside shareholders, the debt market and theeffects of cross-listing on US and UK exchanges. These hypothesized relationships are alsodetailed below. We use data from the 1992-1996 period (pre-Peters Committee Reportperiod) to estimate these relationships. We then assess the impact of the Peters’ CommitteeReport.

3.1 Organizational form and voting rights

Firm value is adversely affected by constraints placed on shareholders’ voting rights eitherpermanently or by management’s attempt to mitigate the market for corporate control (e.g.,Stulz, 1988, Malatesta and Walking, 1988). In our context, the legally required structuredregime is used to directly limit shareholder influence. (We expect no difference between afirm without a structured regime and a firm that voluntarily retains the structured regime dueto the competition these latter firms face in the international marketplace.) Explicitconstraints on shareholder influence occur through the use of certificates and limited votingpower. Takeover defenses that mitigate the market for corporate control are preference andpriority shares.

Major outside shareholders are hypothesized to constrain management’s (and under thestructured regime, the supervisory board’s) deviation from value-maximizing behavior (e.g.,Agrawal and Knoeber, 1996, Cho, 1998, Holderness and Sheehan, 1988, La Porta, Lopez-de-Silanes and Shleifer, 1999, and Morck, Shleifer and Vishny, 1988). In our tests, we considerthe influence of a major outside shareholder owning more than 5% of the shares. We alsoinvestigate the influence of major shareholdings by financial institutions (i.e., banks,insurance companies, pension funds and institutional venture capitalists) and by industrialfirms. Financial institutions can have a positive or negative impact on firm value (Pound,1988). The effect will be positive if they are more efficient monitors than atomisticshareholders. It will be negative if they collude with management. While McConnell andServaes (1990) find a positive relationship in the U.S., in the Netherlands financialinstitutions are known for their passive attitude.11 With regard to industrial firm holdings, the

11 Cantrijn and Vente (1997) sent questionnaires to Dutch institutional investors. The responses showed that theinvestors perceive liquidity to be more important than control. Exercising supervision on the firm’s investmentpolicies and the remuneration are considered to be tasks of the institutions by only 20% and 33% of therespondents, respectively.

8

effect may be positive due to improved monitoring or negative due to collusion and/orattempts to influence decisions for the benefit of their own company.

Two final factors related to monitoring and firm value are debt markets and cross-exchangelisting. Debt markets can discipline management's deviation from value-maximizingbehavior (Jensen and Meckling, 1976). Our tests use leverage and representation on thesupervisory board by financial firms (i.e., interlocking directorates) as measures of thisinfluence. We include as separate variables bank debt and interlocking directorates withbanks because we expect the role of leverage to be more pronounced for this latter type ofdebt.

When companies are listed on exchanges outside the Netherlands, it is important to recognizethe disciplining aspects this can have. For example, UK and US stock exchanges requiremore company and compensation disclosure than the Amsterdam Exchanges. Our testsinvestigate the hypothesized positive impact that cross-listing on UK and/or US stockexchanges have on the value of Dutch firms. 12

Our empirical tests study the impact of the factors outlined above on Tobin’s Q. Tobin's Q isour measure of firm value and performance (Lindenberg and Ross, 1981); it is the marketvalue of the firm divided by the replacement cost of its assets.

3.2 The Peters Committee

The Peters Committee issued its preliminary conclusions in October 1996 and its finalrecommendations (which were expected to be the same) in June 1997. The committee madea major appeal to re-evaluate the numerous constraints placed on the rights of shareholders.The committee spoke specifically to the accountability of the supervisory board (andmanagement board) under the structured regime. However, they did not address the inherentproblems of the structured regime. Rather, the committee addressed how to make thestructured regime relatively more accountable to shareholders without changing thefundamental rights of shareholders. This is likely to be a difficult task given shareholdershave very few rights under the full structured regime.

The monitoring report of December 1998 contained all the corporate governance informationthat was collected on the companies for 1997 by the committee (one-year after the release ofthe committee’s formal report). Using data from the post-Peters period (i.e., 1997), weconduct tests to assess the impact of the committee’s recommendations on the corporategovernance variables outlined above and their relationship with Tobin’s Q. We use theresults from the pre-Peters period as a benchmark.

4. Sample, variable definitions, and empirical tests4.1. Sample

Our sample contains all non-financial firms listed on the Amsterdam Exchanges from 1992-1997. We exclude financial firms because of their distinct regulatory structure. Financial 12 See Lins, Strickland and Zenner, 1999, for the arguments connected with US listing.

9

and asset structure data are obtained from a data set of Statistics Netherlands (CentraalBureau voor de Statistiek). To obtain data on board compensation and bank debt we usedthe Review and Analysis of Companies in Holland (REACH) dataset. We verified thisinformation using the firms’ annual reports. Ownership structure is obtained from the leadingDutch financial daily newspaper (Het Financieele Dagblad) that annually publishes a list ofexchange-listed firms and their stakeholders, according to the notifications for The Law onDisclosure of Shareholdings (Wet Melding Zeggenschap).13 Takeover defenses and cross-listings are from the yearly overviews of all securities listed at the Amsterdam Exchanges(Gids bij de Officiële Prijscourant van de Amsterdamse Effectenbeurs). The data for boardmembers are obtained from issues of the Yearbook of Dutch Firms (Jaarboek NederlandseOndernemingen).

Data on structured regimes as of 1992 (1997) is obtained from firms' annual reports(Monitoring Corporate Governance in Nederland, 1998). If we found a difference between1992 and 1997, we investigated all annual reports over 1993-1996. The annual reportsallowed us to investigate whether the supervisory board established (vaststellen) the annualaccounts and whether the firms met the criteria for the structured regime. In cases ofinconsistency, we contacted the firm. Our sample consists of 132 firms. Since not all firmshave data available in all years, the total number of firm-year observations is 684 (570 from1992-1996 and 114 from 1997).

4.2. Variable definitions and summary statistics

Table 1a lists the variables used in our empirical tests along with the abbreviations used torefer to them in the tables and the text. Descriptive statistics are also included. Table 1bpresents the descriptive statistics for the three sub-samples of interest, no structured regime,legally required structured regime and voluntarily retained structured regime.

Insert Table 1a and 1b Here

The dependent variable, Tobin’s Q (TQ), is measured as the book value of liabilities plus themarket value of equity divided by the replacement cost of the firm’s assets (see Perfect andWales, 1994). In the Netherlands, firms either present replacement values or historical costsin their annual reports. If replacement values are presented no adjustment is required. Ifhistorical costs are presented, we have to adjust the value to estimate replacement value. Todo this, we assume that in the base year the replacement value equals the historical cost. Foreach subsequent year, we adjust this replacement value by adding new investments andcorrections for the growth in capital good prices and subtracting depreciation. Growth incapital good prices is based upon the price index of investment goods, as provided by theStatistics Netherlands. The replacement value of the assets is the book value of assetsadjusted for these replacement value changes.

13 The Law on Disclosure of Shareholdings (Wet Melding Zeggenschap) went into effect in 1992. This law isthe Dutch implementation of the EU Transparency Directive 88/627, which allows us to collect ownershipstructure data.

10

Appearing next in Table 1a are control variables. These variables are firm size measured asthe log of the book value of total assets (BVTA), growth measured as the log of one plus(Growth) the three-year historical growth rate of the firm’s book value of assets, and leverage(LEV) measured as long-term debt divided by book value of assets. Based on prior research,we expect the coefficient on LEV and the log BVTA to be negative and that on log of oneplus GROWTH to be positive.14 We also control for the overall performance of theAmsterdam Exchanges by including dummy intercept variables for each of the years coveredby the sample.

Our first independent variable deals with cross listing. This variable, XLIST, takes on thevalue 1 (0) if the firm is (not) listed on an exchange in the UK and/or US. The organizationalform of the sample firms is addressed by the next two variables. SR takes on a value of 1 (0)if the firm is (not) a legally required structured regime while SR_V takes on a value of 1 (0)if the firm has (not) voluntarily retained the structured regime.

We capture limitations on shareholder rights by using PRIO which takes on a value of 1 (0)in the presence (absence) of priority shares, PREF which is set to 1 (0) if the company can(cannot) issue and place protective preference shares, CERT which is set to 1 (0) when thecompany has (has not) issued certificates and LVOTE which takes on a value of 1 (0) if thereis (is not) a limitation on voting, normally 1%. Two additional variables are used to capturecases where firms employ more than one of the above measures. DEF3 is the number (0, 1,2, or 3) of takeover defenses from the set PRIO, PREF and CERT that are used by the firm,and DEF4 is defined analogously after augmenting the previous set of three to includeLVOTE.

The influence of the debt market as a disciplining force is measured using BANK_ILOCK,which is the number of bank interlocking directorates on the supervisory board, andFIN_ILOCK, which is the number of interlocking directorates with financial institutions.Both are measured as the number of relationships (interlocks) with banks or financialinstitutions, with bank interlocks being a subset of financial institutional interlocks. A thirdvariable, BANK_D measures a firm’s bank debt (long-term bank debt divided by totalassets). However, because of data availability, we do not have a complete set of Bank_Dobservations.

The three final independent variables capture the concentration and identity of outsideshareholders. OSIDE_EQ is the stake of the largest outside block-holder owning 5% or moreof the shares,15 INSTI_EQ is the sum of all institutional block-holdings (banks, insurancecompanies, pension funds and institutional venture capitalists) and INDUS_EQ is the sum ofthe block-holdings by industrial firms.

14 The negative coefficient for LEV is contrary to that predicted by the monitoring hypothesis for public debtbut consistent with management entrenchment or “debt avoidance” hypothesis articulated by Zwiebel, 1996,and documented by de Jong and Veld, forthcoming.15 Because the Dutch Law on Disclosure of Shareholdings requires the notification of shareholdings whenthresholds of 5%, 10%, 25%, 50%, or 66,7% are passed, we do not have information for shareholdings below5%.

11

4.3 Regression model

The following regression model is used to test the relationships developed in section 3.

TobinQ = f (Organizational Form, Limits on Voting Rights, Debt Market, OutsideBlock-holders, and Control Variables),

where the specific variables designed to capture Organizational Form, Limits on VotingRights, Debt Market, and Outside Block-holders were discussed above. All regression t-statistics are based on White’s heteroskedasticity corrected standard errors. Estimation of theabove model is performed using OLS, which incorporates fixed-effects for each year.16

5. Results

Our results are organized as follows. We estimate a series of regressions to test therelationships hypothesized in section 3 (section 5.1). Next, we isolate the impact the PetersCommittee and its recommendations had on the relation between the corporate governancevariables and Tobin’s Q (section 5.2). Lastly, using event study techniques, we evaluate theimpact of various corporate governance-related events and announcements related to thePeters Committee, Dutch government and European Union during the 1996-1999 period(section 5.3).

5.1. Regression results for the 1992-1996 period (pre-Peters Committee)

Our initial regression results are based on the 1992-1996 period, which precedes the PetersCommittee Report. These baseline regressions serve two purposes. First, they document therelation between the corporate governance variables and Tobin's Q prior to the PetersCommittee Report. Second, they provide a means to assess the extent to which thecommittee's recommendations have had an impact on this relation, at least over the short-term (i.e., 1997). The regressions we estimate are variations of the model described insection 4.3 and the results are reported in Tables 2 and 3.

Insert Table 2 Here

Consistent with prior research (see model 1), the coefficients on the control variables havethe expected signs and are significant (negative for leverage and size and positive forgrowth). The disciplining aspects of international competition and/or disclosure and theresulting increased scrutiny is confirmed as the coefficient on the cross listing (UK and/orUS) variable is positive and significant at the 1% level.

Models (2) and (3) address the issue of organizational form and its impact. Consistent withour most important hypothesis, the legally required structured regime has a significantlynegative impact on Tobin’s Q. After controlling for other shareholder rights (Model 3), thestructure regime reduces Tobin’s Q by 0.223. On the other hand, the performance of firms 16 Because we use a data set of pooled firm-year observations over 1992-1996, our results may suffer fromautocorrelation as we have multiple observations per firm. We discuss this potential problem at the end ofSection 5.1.

12

that voluntary retain the structured regime does not differ from those firms without thestructured regime. Note that firms who voluntarily retain the structured regime areprincipally Dutch multi-national firms that are exposed to the rigors of internationalcompetition.

The impact of takeover defenses and direct limits on shareholders’ voting rights on Tobin's Qare also addressed in model (3). Taken individually, the results for the defenses and directlimits on voting rights are mixed. While the coefficients on the priority shares andcertificates variables are negative (as expected), neither is significant. Preference shares havea positive coefficient, opposite of that predicted, however the result is insignificant. Thecoefficient on the limits on voting rights variable has a significant and positive coefficient,opposite of our prediction.

To more clearly isolate the significance of the direct limitations on voting rights, weconducted two additional regressions; one with the variable DEF3 which combines the set ofthree takeover defenses (priority shares, preference shares and certificates) and another withDEF4 which simply adds the fourth (direct limits on shareholder voting). The results (nottabled) show that DEF4 is significant and positive with its significance driven by the directlimits on shareholder voting rights. Later in the paper, we consider the interaction of thesedefensive mechanisms with the organizational form and ownership structure variables, andfind that they are important and in the way predicted.

Table 3's regressions analyze ownership structure and relations with financial institutions. Inmodel (1), we focus on the monitoring role of the major block-holders. The coefficients forthe major outside shareholder and industrial block-holders are insignificant. However,financial institutions have a negative and significant effect on firm value. This is consistentwith both the collusion story of Pound (1988) and the passive attitude of Dutch financialinstitutions. Later in this section, we provide additional evidence on the collusioninterpretation.

Insert Table 3 Here

The results of model (2) in Table 3 show that bank debt and interlocking directorates withbanks are negative and significant while leverage becomes insignificant.17 Since banks are asubset of financial institutions, the two are significantly correlated with a correlationcoefficient of 0.365. To assess the implications of financial leverage and interlockingdirectorates with financial institutions, we conducted regressions that redefined bank debt aslong-term bank debt over long-term leverage (to separate the leverage effect from the bankeffect) and included (excluded) interlocks with financial institutions (banks). In theseregressions (not tabled), the coefficients are negative and significant. While inconsistentwith our hypothesized monitoring relationships, it is consistent with the managemententrenchment (debt avoidance) hypothesis articulated by Zwiebel, 1993 (see footnote 14 forfurther explanation and evidence). 17 The leverage variable is measured as long-term debt over book value of total assets. We chose to excludeshort-term debt because in many Dutch firms short-term debt is largely comprised of trade credit, and to a lesserextent bank debt and other forms of short-term credit. The disciplining role of leverage is more likely to resultfrom long-term debt obligations. The bank debt variable is long-term bank debt divided by book value of totalassets.

13

Model (3) in Table 3 addresses the relationship between ownership structure and takeoverdefenses (including direct limits on shareholder voting rights). Here ownership concentrationmay both increase entrenchment and provide monitoring. The coefficient associated with themajor outside shareholder is positive and significant, consistent with our monitoringhypothesis. The set of takeover defenses and limits on shareholder voting rights (DEF4)again has a positive and significant coefficient. However, when this set of defenses isinteracted with outside ownership, the coefficient is negative and significant. We interpretthis as an entrenched firm effect. For a given level of outside ownership, the supervisory andmanagement boards can entrench themselves by adopting a set of takeover defenses.However, the increase in firm value due to outside ownership concentration and takeoverdefenses by far dominates the negative effect that the interaction of these two forms ofentrenchment has on firm value.

Model (4) in Table 3 investigates whether institutional investors “collude” with entrenchedmanagement and supervisory board members by focusing on an important situation wherethis could occur, takeovers. Preference shares are frequently placed with friendlyinstitutional investors during takeover attempts (P. de Vries, representative of the Dutchplatform for investors, mentioned this institutional feature to us). Therefore, we expect thatownership by institutional investors is more likely to induce entrenchment in firms that canissue preference shares. Specifically, we consider preference shares and its interaction withinstitutional holdings. However, neither is significant. Thus, while equity holdings byfinancial institutions have a negative effect on firm value, there is no indirect evidence ofcollusion between the boards and institutions in potential takeover situations.

A measurement issue associated with the major outside block-holder (and to a lesser extendthe other block-holders) is that ownership is defined as cashflow (not voting) rights. While itis impossible to disentangle the two, we re-ran all of the above regressions (not tabled)including certificates and an interaction for certificates and the major outside block-holder.The coefficient on the interaction term was significantly negative which is evidence of thepotential measurement problem. However, none of the other coefficients changes signs orsignificance.

5.1.1. Summary

Our major result addresses the structured regime. For domestic Dutch firms, the legallyrequired structured regime has a negative impact on Tobin’s Q. Dutch multi-nationals thatvoluntarily retain the structured regime perform no differently than firms without thestructured regime. Increases in Tobin’s Q due to international competition and/or increaseddisclosure are also apparent when firms cross-list their securities with UK and/or US stockexchanges. Turning to the monitoring role of outside shareholders in conjunction withtakeover defenses, outside shareholders have a positive and significant effect on firm value(and so do takeover defenses). However, outside owners, management and supervisoryboard members entrench themselves by adopting defensive mechanisms that reduce the netbenefits from these two forms of limited entrenchment. Financial institutions fail in theirmonitoring role, although there is no indirect evidence of collusion.

14

Given the importance of the supervisory board and its influence over the management boardunder the structured regime and operations of the firm in general, a logical question to ask iswhether our results are affected by the omission of supervisory board and management boardcharacteristics from our regressions. For the supervisory board, we collected data on itsabsolute size, its size relative to the management board, the shareholdings and boardcompensation of its members, and the interlocking directorates the firm’s board membershave with other firms. We collected analogous data for the management board. Includingthese variables in the regressions (not tabled) does not alter the basic tenets of our results.18

We also performed some sensitivity analysis related to the regressions reported above. First,instead of pooling data across firms and years, regressions were run using five-year averagesof the dependent and independent variables for each firm. While the significance of thecoefficients was altered due to the reduction in sample size, the signs of the coefficients areunchanged. We also ran the regressions on a year-by-year basis, the results of which arecomparable to the regressions using five-year averages. None of the significant coefficientschanged signs.

When sample size is not an issue, it is conventional to run the regressions with firm fixedeffects. As an alternative, we ran the regressions with industry fixed effects using theindustry classifications from Statistics Netherlands (equivalent to SIC codes). We also ranthe regressions excluding the largest 5% and 10% of the firms (i.e., the largest internationalfirms). Our results do not change.19 Finally, the sample sizes were to small to investigatefirms that changed from no structured regime to the structured regime (and/or from thestructured regime to the voluntary structured regime) during the time period considered.

5.2. Univariate and regression results for 1997 period (post-Peters Committee)

For the firm characteristics detailed in Table 1 (and supervisory and management boardcharacteristics), we compared their values in 1996 to those in 1997 using univariate t-tests.There were no significant changes except for the increase in the proportion of firmsvoluntarily retaining the structured regime (significant at the 10% level). This increase islikely due to increased international activity by Dutch firms.

In Table 4, we perform two regressions, one for the pre-Peters period and one comparing thepre- and post-Peters periods. We only include variables that had a significant impact (at the5% level) on Tobin’s Q in the 1992-1996 period (see Tables 2 and 3). The exception is thevariable for firms that voluntarily retain the structured regime.

Insert Table 4 Here

The first regression provides an informational benchmark for the second regression. To testfor changes between 1992-1996 and 1997 in the second regression, we interact thegovernance variables with a dummy variable that has a value of 1 in 1997, and 0 otherwise.

18 The results are also robust to alternative specifications of the dependent variable. The correlation betweenTobin’s Q and the ratio of market to the book value of equity, total assets or equity plus total assets is 0.894,0.998 and 0.875 for 1992-1997 (and similarly for 1992-1996).19 The results of all our sensitivity analyses do not change when we include the data on supervisory andmanagement board characteristics.

15

The left-hand column of the regression in Table 4 contains the coefficients for the 1992-1996period while the right-hand column contains the coefficients for the variables interacted withthe 1997 dummy variable. The results in the right-hand column document that thecoefficients on institutional holdings and interlocking directorates with banks aresignificantly negative. This implies that the already negative impact of these variables onfirm value became more pronounced in 1997 when compared with the previous years. Noneof the other coefficients in the right-hand (1997) column are significant.20 Thus, the corporategovernance characteristics themselves and the impact they have on Tobin’s Q did notsignificantly change in the post-Peters period when compared to the pre-Peters period.

5.3 Stock price reactions to corporate governance events5.3.1. Background

As the above univariate analysis and regressions using Tobin's Q illustrate, it is difficult toidentify, at least in the short-term, an impact from the recommendations contained in thePeters Committee report. However, the Committee did not operate in isolation, as there wereadditional Dutch government and European Union events with the potential to influence afirm’s corporate governance structure and hence its value. In this section, we use "eventstudy" techniques to assess investors’ reactions to the various events associated with theevolution of corporate governance practices in the Netherlands. In a sense, the analysisprovides a direct market test of the premise underlying the Peters Committee, namely thatmarket forces via self-regulation are sufficient to promote changes in corporate governancethat enhance shareholder value.

Appendix A presents a list of eleven events associated with corporate governance at theCommittee, Dutch government and European Union level. Our data sources are the Dutchequivalent of the Financial Times (Het Financieele Dagblad), the preliminary and finalversion of the first report of the Committee and the monitoring report of the Committee.

5.3.2. Event study analysis

The "event study" method we use is an application of Zellner’s (1962) Seemingly-Unrelated-Regression (SUR) methodology (see Schipper and Thompson 1983 and 1985 for a detaileddiscussion of this approach). The returns-generating process for each firm is

=

+++=11

1,

kitiktikmtiiit DRR εγβα

where Rit is the return to security i on day t, Rmt is the return to the market index on day t,Dikt is a dummy variable that takes on a value of one on the day before and day of theannouncement of event k (k=1, 2, ..., 11) and zero on all other days, αi is the model interceptof firm i, βi is the slope coefficient or systematic risk of firm i, γik is the abnormal return of

20 Though significant in Table 3, model (1), Bank debt (BANK_D) was not included in the analysis because ofmissing observations. Nothing changes when we run the regression again to assess the impact of bank debt oninstitutional holdings and bank interlocks. Our conclusions also do not change when supervisory andmanagement board characteristics are included in the regression.

16

firm i associated with event k, and εit is a random disturbance. For each firm the disturbancesare assumed independent and identically distributed over time, but may be heteroscedasticand correlated in cross-section. The firm-specific parameters of the model are estimatedusing daily stock return data from January 1, 1993 to December 31, 1996. The market indexused is a value-weighted index of all firms traded on the Amsterdam Exchange (results usingalternative market indices or market adjusted returns yield similar results).

The test of interest is the significance of the mean abnormal return of the sample firms at thetime of each event. In particular,

)=(k 0, = :H ik

N

=1i0 11,...,2,1γ ,

where k denotes events and N denotes the number of firms. Since the sum is a scalarmultiple of the cross-sectional average, this test is equivalent to a test on the cross-sectionalaverage abnormal return. In addition to using this hypothesis to assess the sample-wide pricereaction to each event, we also use it to assess the abnormal returns for particular sub-samples of firms (e.g., the mean abnormal return of firms with the legally required structuredregime). The significance of the sample (and sub-sample) mean abnormal return to eachevent is assessed using the F-test outlined in Schipper and Thompson (1985).

5.3.3. Event study results

Of the eleven events listed in Appendix A, only event 8 (the release of the monitoring reportand the related corporate governance information that was collected about the companies) isassociated with a significant stock price reaction. Consequently, to save space, Table 5 onlyreports results for this event. The results reported in Table 5 document that the sample-widemean abnormal return to this event is -0.7% and the F-statistic for H0 is 6.80 (p-value =0.0001). The rejection of H0 suggests a pervasive overall negative reaction to this event andis consistent with the finding that 77% of the sample firms exhibit a negative stock pricereaction to this event. One interpretation of these results is that based on the negativecorporate governance information released with the report, the market was disappointed withthe firms' lack of progress in their governance practices. Based on personal discussions withPeters Committee staff members, this interpretation is consistent with their view that Petershimself, built up market expectations about substantive change that was not realized giventhe data that was released.21

Insert Table 5 Here

Given the overall negative impact of event 8, we next address cross-sectional variation in the

21 Event 8 also included the Minister of Finance’s announcement that legislation on proxy voting would beproposed to the cabinet of ministers. We again checked the Het Financieele Dagblad and found a discussion onthe same Communication Channel that we tested in event 6 and 7, which yielded nothing significant, and JaapPeters’ recommendation that proxy voting be considered for certificates. The Minister of Finance’s and/orPeters announcement could have been a surprise, although it was partially anticipated by the fourteen or sofirms in the private proxy initiative and most likely by certificate firms as well. Without any new information,one would expect a positive reaction by shareholders to these announcements, which works against our findingof a negative reaction.

17

reaction. Our starting point is voting rights. Shareholders must have them for self-regulationto be a viable monitoring mechanism. Without them, shareholders have no way toeffectively monitor the behavior of the supervisory board and management, nor can theyinitiate change. To evaluate this hypothesis, we compare the mean abnormal returnsassociated with event 8 for various sub-samples. The sub-samples are firms with and without:a structured regime, preference shares, certificates, priority shares, limitations on shareholdervoting, and cross-listings on a UK or US exchange.22 The sub-sample results are presented inTable 5 where we report the mean abnormal return for each sub-sample along with F-statistics for the null hypothesis that the sub-sample mean abnormal return is zero and an F-statistic comparing the mean abnormal return of the various sub-samples.

Focusing first on the structured regime, the mean abnormal return for firms with a structuredregime is -0.9%, compared to -0.6% for firms without and -0.2% for firms with a voluntarystructured regime. Only the return of -0.9% for the required structured regime issignificantly different from zero. Of note however, is that the return for the requiredstructure regime sub-sample is significantly more negative when compared to the nostructured regime and the voluntary structured regime sub-samples. This is further evidencethat the required structure regime has detrimental effects on firm value.

The results for cross-listing reveal that firms that are (are not) cross-listed exhibit a meanabnormal return of 0.5% (-1.0%). The -1.0% return for the non-cross-listed firms issignificantly different from zero and significantly less than the 0.5% return for the cross-listed sub-sample. This provides evidence that the monitoring effect of cross-listing has abeneficial effect on firm value.

Turning to priority shares, we find that firms with (without) such shares have a meanabnormal return of -0.3% (-1.0%). Firms without priority shares experienced a moresignificant negative stock price reaction than firms with priority shares. Similar results areobserved for limits on shareholder voting where the significant difference between thereturns of the sub-samples is evidence that firms that do not have limits on shareholdervoting exhibit a more negative price reaction. The results for the remaining two constraintson voting rights are not as sharp. In particular, the mean abnormal return for firms with(without) certificates is -0.9% (-0.6%), but they are not significantly different from oneanother. Similar results are observed for firms with and without preference shares.

On balance, the market’s reaction to the release of the monitoring report (event 8) is one ofdisappointment about substantive change through self-regulation. Furthermore, the marketappears to differentiate its reaction across firms in a manner dependent upon the existing setof shareholder rights currently in place within the firm. Specifically, our results here areidentical (except for priority shares) to those described in Table 2 which summarize theeffects that limitations on shareholder rights have on Tobin’s Q. The other ten events, whichgenerated no significant investor reaction, dealt mainly with the other activities of the Peters’Committee including the release of its recommendations, private and government proposals 22 Cross-listing on the UK or US stock exchanges does not directly translate into voting rights. However, thereis additional disclosure required with the cross-listing and this increases accountability. Thus, the indirectbenefits could be better pricing but not necessarily higher firm value.

18

on proxy voting, and the Minister of Finance’s reply to the monitoring report. Overall, themarket is skeptical about the substantive evolution of corporate governance practices in theNetherlands through self-regulation.

6. ConclusionsThe purpose of this paper was to gather evidence on the ability of market forces to promoteinvestor interests via self-regulation. It is of interest to determine when market forces aresufficient to monitor the managers of capital vis-à-vis when additional legal/political actionsare required to write and enforce contracts between these managers and owners of the capital.

The Netherlands provides an ideal setting to investigate the role of self-regulation. For self-regulation to have a chance to succeed, shareholders must have voting rights. Under the“pure” form of the structured regime, shareholders in the Netherlands lose their ability todirectly monitor the supervisory and management boards. However, the market for corporatecontrol still functions since shareholders vote on mergers and acquisitions. With shareholdervoting rights restricted permanently or via takeover defenses, shareholders lose their abilityto initiate change through the market for corporate control as well as through conventionalmonitoring mechanisms. These points provide the basis for our findings associated with thePeters Committee and the market’s skepticism about the evolution of corporate governance.These points also cast doubt on the Dutch corporate governance model as a prototype to beemulated by others.

It is often argued that the market provides management with the incentives to change becauseof the penalty it assesses firms with poor governance and hence performance. However,some preliminary analysis (not tabled), comparing annual (market adjusted) returns, suggeststhat the sub-sample of structured regime firms under performs the sub-sample of voluntarilyretained and non-structured regime firms by 5.5% annually. This argument is predicated onthe assumption that there are mechanisms in place that can facilitate change. With thesupervisory and management boards already controlling the voting rights, it is doubtful thatthis change will take place without legal/political action to restore voting rights toshareholders. It is equally doubtful whether the prospects for change are any different in thelong run without voting rights for shareholders (one of the major recommendations of thePeters Committee).

19

References

Agrawal, Anup and Charles R. Knoeber, 1996, Firm performance and mechanisms to controlagency problems between managers and shareholders, Journal of Financial and QuantitativeAnalysis 31, 377-398.

Alchian, Armen, 1950, Uncertainty, evolution and economic theory, Journal of PoliticalEconomy 58, 211-221.

Amsterdam Exchanges, 1997, Fondsenreglement.

Berle, Adolf and Gardiner Means, 1932, The modern corporation and private property,Macmillan, New York.

Cantrijn, Dolph and Jacco Vente, 1997, Instututionele beleggers en corporate governance inNederland, Bedrijfskunde 69, 51-60.

Cho, Myeong-Hyeon, 1998, Ownership structure, investment, and corporate value: anempirical analysis, Journal of Financial Economics 47, 103-121.

Committee on Corporate Governance, 1997, Corporate governance in the Netherlands(Secretariat Committee on Corporate Governance, Amsterdam, Netherlands).

Dahya, Jay, John McConnell and Nickolaos G. Travlos, 1999, The cadbury committee,corporate performance and management turnover, working paper.

Dedman, Elizabeth, 2000, An investigation into the determinants of UK board structurebefore and after cadbury, Corporate Governance: An International Review 8, 133-153.

de Jong, Abe and Chis Veld, forthcoming, An empirical analysis of incremental capitalstructure decisions under managerial entrenchment, Journal of Banking and Finance.

Fama, Eugene and Michael Jensen, 1983a, Separation of ownership and control, Journal ofLaw and Economics, 26, 301-325.

Fama, Eugene and Michael Jensen, 1983b, Agency problems and residual claimants, Journalof Law and Economics, 26, 327-349.

Financial Times, 20 June 2000, McKinsey report - investors pay premiums for well-governedcompanies.

Financial Times, 27 July 2000, Breaking the bank’s consensus.

Franks, Julian and Colin Mayer, 1996, Hostile takeovers and the correction of managerialfailure, Journal of Financial Economics 40, 163-181.

Gugler, Klaus, 1999, Corporate governance and economic performance, preliminary version(December), University of Vienna.

Hart, Oliver, 1995, Corporate governance: some theory and implications, The EconomicJournal, 105, 678-689.

20

Holderness, Clifford G. and Dennis P Sheehan, 1988, The role of majority shareholders inpublicly held corporations, Journal of Financial Economics 20, 317-346.

Jensen, Michael C. and William H. Meckling, 1976, Theory of the firm: managerialbehavior, agency costs and ownership structure, Journal of Financial Economics 3,305-360.

La Porta, Rafael, Florencio Lopez-de-Silanes and Andrei Shleifer, 1999, Corporateownership around the world, Journal of Finance 54, 471-517.

La Porta, Rafael, Florencio Lopez-de-Silanes, Andrei Shleifer and Robert W. Vishny, 1998,Law and finance, Journal of Political Economy 106, 1113-1155.

La Porta, Rafael, Florencio Lopez-de-Silanes, Andrei Shleifer and Robert W. Vishny, 2000,Investor protection and corporate governance, Journal of Financial Economics, 58, 3-27.

Lindenberg, Eric B. and Stephan A. Ross, 1981, Tobin’s q ratio and industrial organization,Journal of Business 54, 1-32.

Lins, Karl, Deon Strickland and Marc Zenner, 1999, Do non-US firms issue stock on U.S.equity markets to relax capital constraints?, working paper (November).

Malatesta, Paul H. and Ralph A. Walking, 1988, Poison pill securities: stockholder wealth,profitability and ownership structure, Journal of Financial Economics 20, 347-376.

McConnell, John J. and Henri Servaes, 1990, Additional evidence on equity ownership andcorporate value, Journal of Financial Economics 27, 595-612.

Mertens, G.M.H., 1997. The impact of financial reporting regulation on financial accountingmethod choice, dissertation nr. 97-38, Maastricht.

Ministry of Finance, (May 10)1999, “Kabinetsreactie op het rapport van de MonitoringCommissie Corporate Governance.”

Monitoring corporate governance in Nederland, 1998, Bericht van de Monitoring CommissieCorporate Governance en de uitkomsten van het onderzoek verricht door het EconomcischInstituut Tilburg, Kluwer.

Morck, Randall, Andrei Shleifer and Robert W. Vishny, 1988, Management ownership andmarket valuation: an empirical analysis, Journal of Financial Economics 20, 293-315.

Perfect, Steven B. and Kenneth W. Wales, 1994, Alternative constructions of Tobin’s q: anempirical comparison, Journal of Empirical Finance, 1, 313-341.

Pound J., 1988, Proxy contests and the efficiency of shareholder oversight, Journal ofFinancial Economics 20, 237-265.

Price Waterhouse, 1997, Converging cultures: trends in corporate governance (April).Schipper, K. and R. Thompson, 1983, The impact of merger-related regulations on theshareholders of acquiring firms, Journal of Accounting Research 21, 184-221.

21

Schipper, K. and R. Thompson, 1985, The impact of merger-related regulations using exactdistributions of test statistics, Journal of Accounting Research 23, 408-415.

Stulz, Rene M., 1988, Managerial control of voting rights: financing policies and the marketfor corporate control, Journal of Financial Economics 20, 25-54.Shleifer, Andrei and Robert W. Vishny, 1997, A survey of corporate governance, Journal ofFinance 52, 737-783.

Slagter, W.J., 1996, Compendium van het Ondernemingsrecht, Kluwer, Deventer.

Stigler, George, 1958, The economies of scale, Journal of Law and Economics 1, 54-71.

Stiles, Philip and Bernard Taylor, 1993, Benchmarking corporate governance: the impact ofthe cadbury code, Long Range Planning, 26, 61-71.

Verdam Committee, 1965, Herziening van het ondernemingsrecht, rapport van de commissieingesteld bij beschikking van de minister van justitie van 8 april 1960, Staatsuitgeverij, DenHaag.

Zeff, Stephen A., F. van der Wel and C. Camfferman, 1992, Company financial reporting: ahistorical and comparative study of the Dutch regulatory process, Elsevier SciencePublishers,Amsterdam.

Zellner, Arnold, 1962, An efficient method of estimating seemingly-unrelated-regressions andtests for aggregation bias, Journal of American Statistical Association 57, 348-368.

Zwiebel, Jeffrey, 1996, Dynamic capital structure under managerial entrenchment, AmericanEconomic Review, 86, 1197-1215.

22

Appendix AKey events related to Dutch corporate governance practices

The data sources are Het Financieele Dagblad, the preliminary and final report of the Peters Committee, and themonitoring report that assesses the impact of the final report of the Peters Committee (one-year after its release).

Event 1: On February 13, 1996 Van Ittersum, chairman of the Amsterdam Stock Exchanges announces acommittee for code of best practice.

Event 2: On February 28, 1996 the Ministries of Finance, Law and Economic Affairs and VvdE (shareholders)and VEUO (exchange-listed firms) agree on an arrangement over takeover defenses.

Event 3: On March 15, 1996 there is an announcement of the members of Committee Corporate Governance.Given Dutch consensus approach, all the parties are represented on committee.

Event 4: On October 28, 1996, the publication of the preliminary conclusions of Peters Committee (‘Corporategovernance in Nederland: een aanzet tot verandering en een uitnodiging tot discussie’) took place.

Event 5: On June 25, 1997, the publication of the final conclusions of Peters Committee (‘Corporategovernance in Nederland; de veertig aanbevelingen’) took place. The conclusions are similar to the preliminaryreport.

Event 6: On April 18, 1998, an announcement of a Communication channel for shareholders. A small group offirms form a private sector initiate or experiment in “voting by proxy” using a system designed and owned bythe participating firms.

Event 7: On May 19, 1998, announcement of participating firms in the Communication channel forshareholders.

Event 8: On December 3, 1998 the Peters Committee monitoring report is presented and published. This is themajor event as it contains all the corporate governance information that was collected by the monitoringcommittee on the companies. Jaap Peters presented the report to the Minister of Finance. During this meetingthe Minister of Finance announces that legislation on proxy voting will be proposed to the cabinet of ministers.The proposed legislation is independent of the private sector initiative.

Event 9 On April 29, 1999, the proposal to introduce proxy voting is approved by ‘Ministerraad’, which meansit is approved by the “cabinet” of ministers and will be sent to parliament for consideration.

Event 10: On May 10, 1999 Minister of Finance replies to Peters in a ‘nota’ to the ‘Tweede Kamer’(parliament): firms should provide more information on compensation and stock transaction by managers;proxy voting should be possible and limitations on voting power should be banned. However, no specificproposals are mentioned and the article described the reply as a ‘wensenlijstje’ (list of wishes).

Event 11: On June 23, 1999, a new European Union Directive is released which states that majorityshareholders have to make a bid on the remaining shares of the company. Certificates and preference shares areallowed in a firm’s capital structure.

Table 1a

Variable definitions and descriptive statistics for the full sample of 570 observations for 132 firms from 1992 to 1996.

Variable Description Variable Name Mean Minimum Maximum Std.Dev.Tobin’s Q Market value of total assets/replacement value of total assets TQ 1.355 0.499 7.038 0.708Total assets Book value of total assets in 1,000,000 NLG BVTA 3263 6.0 129874. 12459.Growth Three-year historical growth of total assets GROWTH 0.302 -0.647 17.072 1.122Leverage Long-term debt/book value of total assets LEV 0.160 -0.039 0.588 0.125Listing abroad Dummy variable with value of 1 for listing on a stock

exchange in the UK and/or US, 0 otherwiseXLIST 0.132 0 1 0.341

Structured regime Dummy variable with value of 1 for presence of legallyrequired structural regime, 0 otherwise

SR 0.620 0 1 0.490

Voluntary structured regime Dummy variable with value of 1 for presence of voluntarilyretained structural regime, 0 otherwise

SR_V 0.081 0 1 0.272

Priority shares Dummy variable with value of 1 for presence of priorityshares, 0 otherwise

PRIO 0.409 0 1 0.492

Preference shares Dummy variable with value of 1 for presence of preferenceshare option, 0 otherwise

PREF 0.647 0 1 0.478

Certificates Dummy variable with value of 1 for presence of certificates,0 otherwise

CERT 0.384 0 1 0.487

Limited shareholdings/voting Dummy variable with value of 1 if shareholdings or votingare limited, 0 otherwise

LVOTE 0.165 0 1 0.371

Number of takeover defenses (3) The number of takeover defenses from PRIO, PREF andCERT

DEF3 1.44 0 3 0.78

Number of takeover defenses (4) The number of takeover defenses from PRIO, PREF, CERTand LVOTE

DEF4 1.61 0 4 0.94

Interlocks with banks The number of interlocking directorates with banks BANK_ILOCK 1.239 0 10 1.965Interlocks with financials The number of interlocks with financial institutions FIN_ILOCK 1.870 1 15 2.550Largest blockholder The stake of the largest blockholder OSIDE_EQ 24.47 0.0 97.05 20.23Financial institution blockholdings The stake of block-holdings by banks, insurance companies,

pension funds and institutionalized venture capitalistsINSTI_EQ 11.36 0.0 58.28 12.31

Industrial blockholdings The stake of industrial blockholders INDUS_EQ 10.30 0.0 93.17 20.56Bank debt Long-term bank debt/book value of total assets BANK_D 0.061 0.0 0.43 0.091

Table 1b

Variables and summary statistics for selected sub-samples of no structured regime, legally required structured regime andvoluntarily retained structured regime.

No Structured RegimeN=173

Legally Required Structured RegimeN=352

Voluntarily Retained Structured RegimeN=45

Variable Mean Minimum Maximum Mean Minimum Maximum Mean Minimum MaximumTobin’s Q 1.483 0.499 7.038 1.260 0.595 5.469 1.605 0.867 4.149Total assets 7164 6 129874 1350 32 35869 3225 98 14870Growth 0.236 -0.647 2.106 0.301 -0.625 17.072 0.563 -0.062 1.752Leverage 0.157 -0.039 0.588 0.155 0 0.568 0.210 0.004 0.452Listing abroad 0.173 0 1 0.094 0 1 0.267 0 1Priority shares 0.45 0 1 0.38 0 1 0.47 0 1Preference shares 0.44 0 1 0.73 0 1 0.78 0 1Certificates 0.25 0 1 0.45 0 1 0.36 0 1Limitedshareholdings/voting

0.05 0 1 0.22 0 1 0.18 0 1

Number of takeoverdefenses (3)

1.14 0 3 1.57 0 3 1.60 0 3

Number of takeoverdefenses (4)

1.19 0 3 1.79 0 4 1.78 1 4

Interlocks with banks 1.07 0 10 1.16 0 10 2.51 0 8Interlocks with financials 1.58 0 15 1.86 0 12 3.04 0 11Largest blockholder 29.53 0 97.05 22.44 0 72.89 20.89 0 58.50Financial institutionblockholdings

7.95 0 58.28 13.32 0 51.24 9.17 0 37.80

Industrial blockholdings 11.10 0 84.54 10.31 0 93.17 7.08 0 58.50Bank debt 0.064 0 0.293 0.058 0 0.433 0.063 0 0.267

Table 2

Regression results based on shareholder rights variables. The sample consists of 570 observations for 132 firms in the 1992 to1996 period. Data definitions appear in Table 1. The dependent variable is Tobin’s Q (TQ). The regressions contain yeardummies. t-values are in parentheses and significant coefficients are indicated by * (10% level), ** (5% level), and *** (1%level) based on a one-tailed test.

Predicted Sign Model (1) Model (2) Model (3)Constant 1.400 (10.06)*** 1.454 ( 10.46)*** 1.401 (11.54)***

Y93 0.211 ( 3.19)*** 0.210 ( 3.18)*** 0.211 ( 3.29)***

Y94 0.242 ( 3.91)*** 0.234 ( 3.80)*** 0.244 ( 4.04)***

Y95 0.327 ( 4.72)*** 0.330 ( 4.83)*** 0.340 ( 5.05)***

Y96 0.464 ( 5.37)*** 0.473 ( 5.50)*** 0.500 ( 5.77)***

LOG(BVTA) - -0.056 ( -2.53)*** -0.047 ( -2.21)** -0.049 ( -2.24)**

LOG(1+GROWTH) + 0.894 ( 3.32)*** 0.804 ( 3.09)** 0.771 ( 3.12)***

LEV - -0.703 ( -2.19)** -0.746 ( -2.36)*** -0.652 ( -2.09)**

XLIST + 0.626 ( 3.92)*** 0.562 ( 3.23)*** 0.548 ( 3.25)***

SR - -0.151 ( -2.14)** -0.223 ( -2.91)***

SR_V 0 0.110 ( 1.03) 0.050 ( 0.45)

PRIO - -0.004 ( -0.07)

PREF - 0.079 ( 1.21)

CERT - -0.006 ( -0.11)

LVOTE - 0.285 ( 3.35)***

NAdj. R2

5700.133

5700.145

5700.162

Table 3

Regression results for ownership structure and financial institution variables. The sample consists of 570 observations (519 when bank debt isconsidered) for 132 firms in the 1992 to 1996 period. Data definitions appear in Table 1. The dependent variable is Tobin’s Q (TQ). The regressionscontain year dummies. t-values are in parentheses and significant coefficients are indicated by * (10% level), ** (5% level), and *** (1% level) based ona one-tailed test.

Predicted Sign Model (1) Model (2) Model (3) Model (4)Constant 1.524 (10.03)*** 1.559 ( 8.89)*** 1.201 ( 7.34)*** 1.468 (10.19)***

Y93 0.217 ( 3.27)*** 0.230 ( 3.36)*** 0.215 ( 3.32)*** 0.218 ( 3.27)***

Y94 0.237 ( 3.84)*** 0.259 ( 3.89)*** 0.250 ( 4.09)*** 0.239 ( 3.83)***

Y95 0.334 ( 4.86)*** 0.320 ( 4.71)*** 0.341 ( 5.08)*** 0.337 ( 4.88)***

Y96 0.475 ( 5.54)*** 0.507 ( 5.13)*** 0.500 ( 5.85)*** 0.483 ( 5.57)***

LOG(BVTA) - -0.053 ( -2.36)*** -0.046 (-1.65)** -0.054 (-2.33)*** -0.050 (-2.31)***

LOG(1+GROWTH) + 0.800 ( 3.03)** 0.915 ( 3.39)*** 0.778 ( 3.21)*** 0.781 ( 3.01)***

LEV - -0.766 ( -2.38)*** -0.246 (-0.60) -0.628 (-1.99)** -0.760 (-2.37)***

XLIST + 0.567 ( 3.26)*** 0.569 ( 2.49)*** 0.552 ( 3.24)*** 0.559 ( 3.23)***

SR - -0.137 ( -1.80)** -0.138 (-1.64)* -0.159 (-1.89)** -0.159 (-1.91)**

SR_V 0 0.119 ( 1.10) 0.149 ( 1.19) 0.096 ( 0.87) 0.094 ( 0.80)

OSIDE_EQ + -0.001 ( -0.71) -0.001 (-0.73) 0.005 ( 1.98)** -0.001 ( -0.44)

INDUS_EQ +\- 0.001 ( 0.91) 0.002 ( 1.28) 0.002 ( 1.02) 0.001 ( 0.94)

INSTI_EQ +\- -0.003 ( -1.86)** -0.004 (-2.27)** -0.003 (-1.47)* -0.004 (-1.41)*

BANK_ILOCK + -0.052 (-2.24)**

BANK_D + -1.642 (-4.02)***

DEF4 - 0.192 ( 3.85)***

OSIDE_EQ*DEF4 - -0.004 (-2.68)***

PREF - 0.081 ( 1.14)

PREF*INSTI_EQ - 0.0003 ( 0.11)

NAdj. R2

5700.145

5190.180

5700.167

5700.144

Table 4

Regression results for the governance variables for 1992-1996 and between 1992-1996 and 1997. The sample consists of 570observations (684 observations) for 132 firms in the 1992 to 1996 (1992 to 1997) period. Data definitions appear in Table 1.The dependent variable is Tobin’s Q (TQ). The regressions contain year dummies. t-values are in parentheses and significantcoefficients are indicated by * (10% level), ** (5% level), and *** (1% level) based on a one-tailed test.

Results for 1992-1996 Results for 1992-1996 vs. 1997:

Predicted Sign1992-1996 1997

Constant 1.172 ( 7.20)*** 1.177 ( 7.15)***

Y93 0.220 ( 3.41)*** 0.233 ( 3.57)***

Y94 0.259 ( 4.26)*** 0.279 ( 4.49)***

Y95 0.350 ( 5.27)*** 0.362 ( 5.27)***

Y96 0.495 ( 5.80)*** 0.483 ( 5.62)***

Y97 1.006 ( 2.49)***

LOG(BVTA) -0.031 (-1.24) -0.038 (-1.50)*

LOG(1+GROWTH) + 0.743 ( 3.15)*** 1.173 ( 3.81)***

LEV - -0.616 (-1.99)** -0.651 (-2.09)** -1.215 (-1.05)

XLIST + 0.571 ( 3.36)*** 0.592 ( 3.51)*** 0.548 ( 1.19)

SR - -0.171 (-2.03)** -0.164 (-1.90)** 0.080 ( 0.25)

SR_V 0 0.113 ( 1.06) 0.075 ( 0.70) -0.042 (-0.12)

DEF4 - 0.137 ( 2.72)*** 0.135 ( 2.58)*** -0.105 (-0.52)

OSIDE_EQ + 0.005 ( 1.81)** 0.005 ( 1.72)** -0.005 (-0.60)

OSIDE_EQ*DEF4 - -0.004 (-2.32)** -0.004 (-2.22)** 0.005 ( 0.81)

LVOTE - 0.154 ( 1.84)** 0.148 ( 1.73)** -0.358 (-1.17)

INSTI_EQ + -0.003 (-1.73)** -0.003 (-1.53)* -0.011 (-1.68)**

BANK_ILOCK + -0.037 (-1.85)** -0.035 (-1.72)** -0.127 (-1.73)**

NAdj. R2

570 0.174

6840.216

Table 5

Event study results for event 8 (the release of the monitoring report). Of the eleven events originally examined,only event 8 is associated with a significant mean stock price reaction. The table reports mean abnormal returnsfor the full sample and various sub-samples as well as the F-statistic for those mean abnormal returns. F-statistics comparing the mean abnormal return of the sub-samples are also provided. Due to data availabilityconstraints, organizational-form and limited voting sub-samples do not always total to the full sample of 114firms. Significance levels are indicated by *(10% level), **(5% level), and ***(1% level).

Sample Mean Abn. Return F-statistic(mean abn. return equals

zero)

F-statistic(subsample mean abn.

returns are equal)

Full Sample (N=114) -0.7% 6.80***

Sub-Samples:

Structured Regime Required (N=64) No Structured Regime (N=28)

-0.9%-0.6%

8.08***1.71 4.41**

Structured Regime Voluntary (N=16) No Structured Regime (N=28)

-0.2%-0.6%

1.421.71 1.05

Structured Regime Required (N=64) Structured Regime Voluntary (N=16)

-0.9%-0.2%

8.08***1.42 8.59***

Certificates (N=46) No Certificates (N=68)

-0.9%-0.6%

8.89***3.34* 0.01

Preference Shares (N=63) No Preference Shares (N=51)

-0.7%-0.8%

4.18** 6.15*** 0.02

Priority Shares (N=42) No Priority Shares (N=72)

-0.3%-1.0%

0.65 11.20*** 8.94***

Limited Voting (N=12) No Limited Voting (N=100)

-0.5%-0.7%

1.36 6.59** 6.20**

Cross-listed UK/US (N=18) Not Cross-listed (N=96) 0.5%

-1.0%0.90

10.11*** 12.71***

The Role of Self-Regulation in Corporate · PDF fileThe Role of Self-Regulation in Corporate...

Documents

Transcript of The Role of Self-Regulation in Corporate · PDF fileThe Role of Self-Regulation in Corporate...