Governing pension funds

(1)

by

A

RNOUD

R

INGELBERG

Masterthesis for obtaining the degree of Msc of Business Administration

Faculty Management and Organization

Supervisor: prof. dr. ir. R. Goodijk Co-reader: dr. D. Veltrop

(2)

Page 2 Version 2.2 “Yet good pension fund governance is not rocket science.”

Keith P. Ambachtsheer

“A paradigm shift takes place when [our] world is qualitatively transformed and quantitatively enriched by fundamental novelties of either fact or theory...”

(3)

Page 3 Version 2.2

Preface

In this master thesis, I have combined my expertise in the field of pension funds with knowledge I gathered in the past three years during the Business Administration master programme at Rijksuniversiteit Groningen. I was eager to gain new knowledge; for me, the knowledge on new theories meant not only a shift in my professional development, but also a change in my views as an expert on the pensions industry. I am convinced pension organizations and their governance could be improved by the application of business administration theories. In a British article on innovative models for pension fund governance, I read that “...pension funds have tended to be

much slower evolving institutions than the financial services industry. Pension funds often seem unable to deal with the institutional costs of change, are slow to adapt, adopt or innovate, and tend to rely on past practices, notwithstanding the uncertainties of global financial markets. Inertia tends to dominate – individually and collectively.” (Clark & Urwin, 2010: 66). Although

pension funds and their executive boards are doing their utmost to comply with legal requirements and all relevant stakes, there is often not enough attention for change management and alignment with external contingencies. For this reason, I agree with Clark & Urwin. Nevertheless, I am pleased to be able to make a contribution towards the unavoidable changes within the sector by means of this study of pension fund governance models.

During my studies, I was always motivated, and determined to pass my exams the first time, trying for a score of 96%. Working on new theories and numerous case studies was, above all, fun, although sometimes quite complicated in combination with a full time job as selfemployed interim manager in the pensions industry. I am very grateful to my beloved wife Petra and my children Martijn, Daniël, Isa and Julian. Without their flexibility, support and patience I would never have been able to complete this master’s programme. Special thanks also to prof. dr. ir. Rienk Goodijk, whose enthusiasm and profound knowledge of corporate governance made me wonder to what extent theories of the firm could be applied to pension funds. I thank dr. Dennis Veltrop for co-reading this master thesis. Finally, I am pleased to have been working on a topic that is so much in the spotlights in the pensions industry.

Although completing this programme while working full-time was a big task, it was worth every minute. Not only did I reach several useful conclusions, I also encountered a number of promising possibilities for future research on governance and macro organization in the pensions industry. Food for thought and new discoveries to reveal...!

Arnoud Ringelberg December 2012

(4)

Executive summary

“Pension funds are financial institutions”. This hypothesis is placed between quotation marks, because we are convinced that the regulator, the legislator and the pensions industry are not yet agreed on this issue. The regulator has organized its supervision as if pension funds were financial institutions, but at the same time, a bill on pension fund governance has been presented in which the representation and authority of social partners, involved in the establishment of job conditions, is the key feature. Pension funds are considered to be social financial institutions, introduced in the joint letter sent by Pensioenfederatie and Labour Foundation from 19th November 20121. Despite this recent letter, there does not yet seem to be a match between the two, which lead us to wonder how governance models for pension funds are designed, and which factors are relevant. From a business administration point of view (a field featuring numerous studies on governance and corporate governance) we would expect pension funds and their organizations to have learned from useful insights gained from, and common theories regarding, firms.

This report briefly discusses four main theories of the firm: the agency theory, the theory of transaction cost economics, the shareholder theory and the stewardship theory. The agency theory focuses principally on the diverging interests of shareholders and managers of publicly held corporations. The theory of transaction cost economics which, roughly, addresses the issue of the cost effectiveness of management decisions (for example: to outsource or not to outsource?), is closely related to the agency theory. At the opposite side of the spectrum is the stakeholder theory. This theory assumes that a company takes into account the stakes of all individuals or entities who may be affected by a business, and who may, in turn, bring influence to bear upon that business (Wheeler & Sillanpää, 1997). Finally, the stewardship theory perceives an organization as a coalition of vested interests; conflicts of interest are minimized because key players such as management, directors, and shareholders form a strategic coalition aimed at maximizing shareholder wealth.

We have indicated in brief the potential relevance of each theory for pension fund governance. We made deductions regarding relevant factors, thereby answering part of the main research question: what factors are relevant to the design of a pension fund governance model? To this end, we designed a research model based on existing literature on corporate governance, and explored it with the help of several interviews with a selection of leading operatives in the pensions industry. Their answers helped us to understand the main governance issues and to validate the explorative research model. Upon reviewing the interview reports we were obliged to make some changes to our research model, but basically the main design was maintained.

1

(6)

With respect to the theories of the firm, we learned that despite their usefulness, the pensions industry does not seem to benefit from the multitude of existing insights. Although the use of corporate governance is not as suitable for pension funds as it would appear (as stated by a number of interview participants), the theories of the firm could indeed shed light on the way in which a governance model for pension funds is chosen.

Apart from making some changes to the explorative research model, our findings include that:

• One-tier boards are considered suitable for pension funds. The assumption that one-tier boards are smaller than two-one-tier boards is a misunderstanding from a corporate governance point of view – one tier boards usually encompass the larger number of directors.

• A one-tier board construction introduces, among other things, a higher degree of involvement of executive directors. Specifically in the case of pension funds, the one-tier board can potentially combine several of the responsibilities governance bodies such as the members’ council, accountability body and internal supervisory body nowadays have.

• However, research shows that one-tier boards usually have more board committees than two-tier constructions, possibly because more resources are needed to meet the minimum standards of the sponsor company.

• In a corporate one-tier board construction, shareholders have more influence on day-to-day operations because they are directly represented in the board. Given the variety in opinions of the participants in our interviews, the question remains who is actually considered to be a ‘shareholder’ of the pension scheme and therefore deserves an influential position on the board.

• Two-tier boards must continue to exist, for a number of pension schemes have indicated that the one-tier construction does not match the sponsor company’s culture.

• A remarkable finding is that a minority of the individuals interviewed stated that they consider the long term and the short term to be incompatible.

• Both literature and several of the interviews show that pension schemes could learn from institutions in other sectors, for example health care or education, which are organized in a similar way.

(7)

We conclude our report by arguing that the stewardship theory is best suited for organizing pension fund governance. The board, the representatives of the pension fund members, and internal supervisors are considered to be collectively responsible for the pension fund’s performance in both the short and the long term, and the governance model has been designed along this line. Regarding the great and widespread interests that are at issue we recommend that pension fund managers periodically make a systematic analysis of stakeholders and their stakes in the pension fund.

Key words

Corporate governance, shareholder model, stakeholder model, stakeholder approach, governance models, company pension funds, pension fund governance.

List of abbreviations

For the reader’s convenience, we list several relevant abbreviations used in this report in the following table.

AFM Autoriteit Financiële Markten, regulatory body in The Netherlands CSR Corporate Social Responsibility

DB Defined Benefit DC Defined Contribution

DNB De Nederlandsche Bank, regulatory body in The Netherlands OECD Organization for Economic Cooperation and Development PFG Pension Fund Governance

(8)

1 Introduction

“Pension funds are financial institutions”.

A decade ago, this hypothesis was not necessarily true. Pension funds were meant to secure our pension savings. Pension funds were led by experts, who were imbued with the necessity of results and were keen on obtaining the best possible results. They were supported by the favourable economic conditions during the past decades. However, early in the millennium, blemishes started to appear on the glossy exterior as a result of the collapse of the Internet bubble. Numerous pension funds experienced a serious decline in their funding ratio, and needed years to recover. In addition, the recent worldwide financial crisis did not make things easier for pension funds, and society began to wonder about the way in which the administration of pension funds is organized. Are board members sufficiently expert? Are they competent in the field of investment management, one of the main responsibilities of pension funds? Do board members acknowledge that a pension fund is a nexus of contracts? Do pension fund boards acknowledge the rights and duties of all stakeholders? The appearance of a specific Pension Fund Governance Code may be a good development for answering these questions. Since this Code is not established yet, and from a very recent date2, it not subject in our research.

One of the two Dutch regulatory bodies in the field of pensions, De Nederlandsche Bank, stated that pension funds must be considered to be financial institutions and must be managed accordingly. Demand for specific expertise in the boards changed correspondingly and led to a rising need for specialized consultancy. Consolidation in the industry is an easily identified development: in 2004 there were around 750 pension funds; 8 years later, “only” 450 were left. To manage pension funds in a proper and professional way, further consolidation is needed, if only because of the fact that there are not enough experts with specific competences and capabilities for managing our pension funds to go round (Schuit & Maatman, 2012).

Pension Fund Governance is in the spotlights, and has been discussed by numerous scholars and authorities in the pensions industry in recent years (i.e. Ambachtsheer et al., 2006; Ambachtsheer, 2007; Allen & Overy/Boer & Croon, 2004; Clark & Urwin, 2010; Maatman & Schuit, 2012). The connection between business administration, theories of the firm, and PFG was always indirect, although several decades of thorough studies and numerous papers shed light on PFG. These studies introduced various main theories of the firm such as the agency theory (Berle & Means, 1932; Fama & Jensen, 1983a), the stakeholder theory

2

(9)

(Freeman, 1984), the transaction cost economics theory (Williamson, 1975, 1984), and the stewardship theory (Donaldson & Davis, 1991). Several variations have since seen the light. In this respect, the arrival of a bill aimed at strengthening pension fund boards3 is interesting. In fact, the bill led to fierce discussions between pension fund experts and stakeholder organizations in the field, and is expected to result in adjustments to the Pensions Act. As a result of this new legislation, most pension funds will have to revise their governance structure, which in general meets the often inefficient legal requirements (board, accountability body, members’ council, internal supervisory body) mentioned above. The literature surveyed in this research makes a contribution in that it tests an explorative theoretical framework pension fund boards and their advisors could use to find the most suitable governance model. A number of factors relevant to the selection of a suitable governance model are introduced. In practice, the governance models of smaller and larger pension funds will differ, but it is nevertheless necessary to select and build governance models systematically (Stewart & Yermo, 2008: 236). In what way could literature on corporate governance and, more importantly, on theories of the firm be of any help in this respect?

The main question in this report concerns the factors that are relevant in choosing and designing a governance model. We use existing theories of the firm to answer this question. At the same time, we look into the specific pension fund issues that could identify relevant factors.

In practice, several pension funds debated on the boundaries between the responsibilities of the board, the accountability body, and the members’ council (see also section 1.1). The division of tasks between the three was not always clear for all involved. Could the administration of pension funds be organized more efficiently?

Scientifically, our research tried to fill an existing gap by connecting pension fund governance and business administration. At the same time, an important connection was made with the earlier-mentioned bill that introduces three standard governance models for pension funds. Is that number sufficient? Do the proposed governance models suit all pension funds?

1.1 Terminology

For a clear understanding of some of the technical details, a short elaboration of Dutch pension fund administration is necessary. Legislation on pension fund administration is laid down in the 2007 Pensions Act. Topics covered include stakeholder representation, accountability issues and internal supervision. Generally, pension funds are administered by a

3

(10)

board (in Anglo-Saxon areas this would be the board of trustees but under Dutch law such a body is unknown; board members are called directors). Accountability to stakeholder groups is delegated to the accountability body (“verantwoordingsorgaan”), which controls the decision-making process within the board. Several pension funds have a separate member’s council4 that controls certain decisions, such as changes regarding legal documents or executive arrangements (actuarial tables, communications plan, etc.). Finally, pension funds must appoint an internal supervisory body that may consist of permanent members (usually in the case of structural supervision) or non-permanent members. In the latter case, supervision normally occurs once every three years, complying with the minimum legal requirements for internal supervision. The bill on PFG mentioned above was issued under pressure from both society (transparency of governance and accountability by board members) and the sector (governance model complexity),.

This research will make use of several definitions of organizational concepts and specific pension concepts that require clarification. An organization is a tool people use to coordinate their actions to obtain something they desire or value – that is, to achieve their individual or collective goals (Jones, 2010: 24). In reaching these goals, people or organizations are involved or influenced: the so called stakeholders or stakeholder groups. Stakeholders or stakeholder groups are defined as any group or individual who can affect or is affected by the achievement of an organization’s objectives (Freeman, 1984: 46). In addition, we will refer to certain types of pension schemes in relation to governance models: defined benefit and defined contribution arrangements. In a defined benefit scheme, a participant in the scheme is guaranteed the accrued pension rights. We will refer to these schemes as DB schemes. By contrast, in a defined contribution scheme, a participant in the scheme is guaranteed a certain premium; on the expiration date, the participant must convert the accrued capital to a periodic payment, usually with a life insurance company. We will refer to these schemes as DC schemes. Finally, pension fund governance (PFG) is a key topic in this paper. The PFG concept is defined as follows:

• Meticulous administration by a board that is in control and that uses a thorough decision-making process.

• Internal supervision of the board, the running of the pension scheme, and of investment management.

• Accountability to all stakeholders based on a dialogue on policies and execution.

• Expertise of board members, executive bodies and those assigned supervisory tasks.

• Transparency in decision making, reporting and results.

• Well considered and clear communications on all the above concepts.

(Allen & Overy/Boer & Croon, 2004)

4

(11)

Before beginning with a short review of literature on corporate governance models, we will define the boundaries of this research. In general, the research boundaries are determined by the type of pension fund involved: the funds included in our research involve company pension funds linked to one particular company (i.e. Shell Nederland BV, Unilever NV and Akzo Nobel NV, but also smaller companies such as Technische Unie BV or Grontmij NV). These pension funds are distinct from sector-wide pension funds involving companies in a specific sector (i.e. the metal or bakery industries).

Bookmark

(12)

2 Literature review

2.1 Introduction

This chapter explores relevant literature on theories of the firm. We will link each theory to pension fund governance (PFG). To begin with, the model in figure 2.1 is used to indicate four key concepts. The original model by Mallin (2010) will be expanded to include several connections between the stewardship theory, the agency theory and the transaction cost economics theory (TCE). A number of connections between the three theories could be deduced from a combination of Mallin’s model and the work of Maassen (2002) on corporate governance. In fact, the stewardship theory draws on the assumptions underlying the agency theory and TCE (Donaldson & Davis, 1991). The agency theory, the stewardship theory and TCE can therefore be viewed as theories that consider governance primarily within organizational boundaries.

FIGURE 2.1: major concepts in corporate governance

Source: Mallin, 2010

In table 2.1, the main characteristics of the four concepts are presented.

(13)

TABLE 2.1: major concepts in corporate governance Theory name Characteristics

Agency theory Agency theory identifies the agency relationship where one party, the principal, delegates work to another party, the agent. In the context of a corporation, the owners are the principal and the directors are the agent.

Transaction cost economics

Transaction cost economics views the firm itself as a

governance structure. The choice of an appropriate governance structure can help align the interests of directors and

shareholders.

Stakeholder Stakeholder theory takes account of a wider group of

constituents rather than focusing on shareholders. Where there is an emphasis on stakeholders, then the governance structure of the company may provide for some direct representation of the stakeholder groups.

Stewardship Directors are regarded as the stewards of the company’s assets and will be predisposed to act in the best interest of the shareholders.

Source: Mallin, 2010

These first drawings of theories of the firm are far from complete and yet uncriticised. Before we can apply these theories to pension funds, they must be elaborated. The next sections therefore explore the earlier-mentioned literature on the theories of the firm and also give some comment on them.

At this point, we refer to the first set of research sub-questions supporting our key question, which will be discussed in chapter 3. The questions concern the connection that may exist between PFG and commonly used theories of the firm:

How could major theories of the firm be applied to pension funds?

Which specific corporate governance model could apply to pension funds? How could corporate governance models be combined with PFG?

2.2 Agency theory

(14)

Means, who observed in the early 1930s that the ownership of large corporations is so “...widely distributed that no individual or small group has even a minority interest large enough to dominate the affairs of the company”. In fact, the theory acknowledges the importance of the separation of ownership, decision-making and decision control, as stated by Adam Smith as early as 1776. Potential agency conflicts were discussed thoroughly by Fama & Jensen (1983a, 1983b) in their breakthrough papers on decision management and ownership. Inadequate representation of stakeholders allows managers to maximize their own interests, which do not necessarily coincide with those of the shareholders. If the organization is considered as a nexus of contracts (Fama & Jensen, 1983a: 302), then the agency theory would comprise the contracts of principals (shareholders) with agents (managers). These contracts would be set up to secure the rights and stakes of the principals. Adoption of strategies or policies which have a negative impact on the wealth of residual claimants (such as the principals) must be avoided. A critique in this field says that an organization must always reflect on how far the contracts should reach. There are (inter)dependencies between the type of contract and several secondary factors, such as:

• firm ownership of the agent;

• the principals level of information;

• the origin of contract (behaviour-based or outcome-based);

• risk aversion from both parties;

• goal conflicts,

• the length of the agency relationship, and

• the type of business (high-tech, low-tech, etc.). (Eisenhardt, 1989a)

These factors must all be considered in the set-up of principal/agent contracts. One specific type of contract relevant for monitoring executive behaviour is that with the members of the board of directors. From an agency perspective, boards may be used as monitoring devices for shareholder interests, because they ratify and monitor the organization’s most important decisions and hire, fire, and compensate top-level decision managers (Fama & Jensen, 1983b: 332).

(15)

Jensen (1983b) with those of the critics of pension fund boards. Indeed, if boards are to function as monitoring devices, the criticisms regarding pension fund boards are not very helpful. The potential conflicts of interest of boards members are another reason to be concerned about the quality of PFG from an agency perspective. Although a UK survey demonstrated that, since 2006, confidence in managing conflicts of interest has increased significantly, when finance gets tough, conflicts of interest may arise (Stewart & Yermo, 2008: 230). In such cases, impartial board members and a sound, permanent internal supervisory body are indispensible.

2.3 Transaction cost economics

Transaction cost economics (TCE) is often viewed as closely related to agency theory. The theory, expounded in particular in the works of Williamson (1975, 1984), views the firm as a governance structure, whereas the agency theory views the firm as a nexus of contracts (Mallin, 2010). The rationale behind TCE is, that a firm may choose to keep certain procedures in-house rather than outsource them, because insourcing is cheaper. As a result, the internal costs for managing a new procedure will increase, but not as much as outsourcing costs would. At a certain point, the additional costs of insourcing another procedure will be higher than outsourcing because of the additional bureaucracy, caused by all the insourced activities. It is argued that firms may become less efficient the larger they become; equally, all changes which improve managerial technique will tend to increase the size of the firm (Coase, 1937). Williamson (1984) builds on the earlier work of Coase, and provides a justification for the growth of large firms and conglomerates, which essentially provide their own internal capital market. The cost of any misaligned actions may be reduced by a judicious choice of governance structure rather than merely realigning incentives and pricing them out.

(16)

relationship with the firm is broken. Moreover, such transactions are likely to extend over such a long period of time, and are sufficiently complex and unpredictable, that important aspects of future transactions cannot be reduced to contract in advance but rather must be dealt with over time according to experience (compare also Mallin, 2010, at the start of this section). The type of contract, or its duration, can be influenced by the mutual suspicion of opportunistic behaviour. Those circumstances will be an incentive to close a long-term contract. On the other hand, such contracts can become a plain gamble between the parties involved, inefficiently creating risks for both where there is actually little or no underlying social risk. In all cases, the issue is who will pay the costs of ownership contracts. Here, too, there is a connection with the type of market. Generally, the heavier the competition in a market, the more difficult it is to make consumers bear those costs. However, regardless of who bears the cost, there is an incentive to reduce those costs wherever possible by reorganizing the firm with a more efficient form of ownership (Hansmann, 1996) – that is, governance.

What does TCE entail for PFG? In general, when TCE is considered to involve the sum total of the costs of internal contracting, as the agency theory elaborates, it could be argued that PFG is not a very efficient system. As a result of legal requirements, a large number of contracts must be drawn up to put all mandatory or voluntary committees and governance bodies in place. As history shows, there is a high risk of ambiguity in responsibilities. The governance models resulting from the bill5 will probably mitigate this risk. However, the future implementation of the changes to the Dutch Pensions Act will still have to prove its worth. TCE in terms of interrelated questions on insourcing and outsourcing, and the existence of an internal capital market, are less important for pension funds. The establishment and management of a pension fund is, quite simply, not as capital-intensive as a multinational would be. The costs of transactions within a pension fund are primarily caused by the nexus of contracts that forms the pension fund.

2.4 Stakeholder theory

The recognition of stakeholders goes back as far as 1776, when Adam Smith identified external interests to the firm, consumers being external members who were affected by and had an interest in the firm. Barnard (1938) suggested that employees were an important factor in a firm’s success and that, as such, their interests should be carefully assessed, another early example of the existence of stakeholders to the company (Key, 1999: 320). The development which recognizes employees as a stakeholder group cannot be separated from

5

(17)

another important development, the rise of Corporate Social Responsibility6 (CSR1), Corporate Social Responsiveness (CSR2) and Corporate Social Rectitude (CSR3) (conceptualization by Frederick, 1994). Although developments concerning CSR1 are important to the developments in stakeholder approach, they must be considered separately, for CSR conceptualizations have often been prescriptive while the stakeholder theory is generally descriptive (Key, 1999: 319; Donaldson & Preston, 1995: 66). This does not correspond with the conclusions of Kristofferson et al., who report that “...stakeholder theory was quickly adopted as a theoretical foundation for a contemporary conceptualization of CSR” (Kristofferson et al., 2005: 13). In our view, CSR is to be considered as a lever that has resulted in the successful development and quick adoption of stakeholder theory. Since 1984, numerous scholars have explored the theory and made useful contributions to the original theory, which must often be considered as variations on the theory (i.e. Fassin, 2009; Rowley, 2008; Mitchell et al., 1997; Post et al., 2002)

R. Edward Freeman is one of the main founders of the stakeholder theory. His breakthrough work on stakeholder management (originally intended as yet another book on strategic management) was the incentive for the development of a totally new field of study. The stakeholder theory is fundamentally a theory on how business works at its best, and how it could work, taking into account the integration of ethics and business, and responsibility (Freeman et al., 2010: 9). The stakeholder theory was debated by a number of scholars, who expanded the theory or introduced new variants (e.g. Wheeler & Sillanpää, 1997; Rowley, 1997; Podnar & Jançiç, 2006; Fassin, 2009). Freeman himself developed the stakeholder theory over the past decades, both alone and in joint effort with other scholars, among other things extending it to corporate social responsibility and sustainability (Wheeler et al., 2006).

The stakeholder view has implications for the governance models of organizations pursuing a stakeholder view of the firm and wishing to operationalize the influence of various stakeholders. Research shows that a loyal and close relationship with stakeholders is one of the most important foundations of commercial viability or organizational success. A sound and structured way to increase the influence of stakeholders is by dividing the stakeholders into four categories: social, non-social, primary and secondary7. In addition to this, cycles of inclusion and cycles of continuous improvement are used to analyse the salient stakeholders and their engagement. Cycles of inclusion refer to processes of diagnosis, dialogue and audit

6

Although several different definitions of CSR are used, we choose for the definition of the World Business Council on Sustainable Development: “the continuing commitment by business to behave ethically and contribute to economic development while improving the quality of life of the workforce and their families as well as of the local community and society at large” (Kristoffersen, 2005: 2). 7

(18)

aimed at securing the effective participation and active inclusion of stakeholders in the affairs of the company. Cycles of continuous improvement refer to more technical processes where diagnoses tends to be factually based, and include processes for the optimal management of occupational safety and health, quality, environmental protection and animal welfare (Wheeler & Sillanpää, 1997). Figure 2.2 illustrates the composition of a generalized cycle of inclusion.

FIGUUR 2.2: a generalized “cycle of inclusion”

Source: Wheeler & Sillanpää (1997: 169)

As far as pension funds are concerned, the stakeholder theory is not commonly followed, although several elements of the theory could be applied. For instance, the legislative governance models in the earlier-mentioned bill on better PFG prescribe the (indirect) representation of the main pension fund stakeholders. In each model, it is mandatory to incorporate representatives of the sponsor, employees and pensioners in the board of directors. Although several of the stakeholder theory’s components can be identified, the theory does not exclusively characterize PFG. Pension funds do not have just long-term stakes; for certain stakeholder groups (e.g. pensioners), short-term stakes are more important. But it does indeed appear possible to design a governance model that uses elements of several theories of the firm.

Leadership & commitment

Review or adoption of policy

Agreements of indicators Stakeholder consultation Stakeholder surveys Internal audit Preparation of accounts

and internal reports Agreement of

objectives

External verification and certification

(19)

2.5 Stewardship theory

As stated in the introduction to this literature review, both agency theory and TCE are sources of stewardship theory. The theory views the corporation as a coalition of vested interests. Proponents of stewardship theory oppose the notion that conflicts of interest exist between management, directors and shareholders. The three key players are considered as a coalition, willing to co-operate and to bargain to achieve long-term growth, stability and profitability. If one assumes that there are no conflicts of interest between these key players, there is no need to establish corporate boards to monitor management and secure the maximization of shareholder wealth. Instead, the stewardship theory sees corporate boards as valuable strategic devices to maximize shareholder wealth in situations where authority structures are unified and boards are composed of experienced executive directors are not impeded by information asymmetries and unnecessary bureaucratic structures that may paralyze the strategic decision-making processes of corporations (Maassen, 2002: 19). Management and boards of directors are seen as stewards that act in the best interest of the organization.

From a decision-making point of view, the proposed governance structure offers more opportunities to make and monitor decisions, because directors and management make decisions together. However, attention must be paid to the formal independence of corporate boards, which may hinder the involvement of non-executive directors in the initiation and implementation of decisions. This is called the paradox of board involvement (Maassen, 2002: 39).

Stewardship theory, therefore, does not view board organization from a conflict perspective. In other theories, such as the agency theory, principal/agent conflicts of interest are the basis for designing a governance model. Although these factors are certainly relevant, it is believed that a ‘stewardship design’ of the governance model harmonizes better with the present business era than the agency theory does. Decision-making is often a process in which several members of an organization participate. In general, long-term shareholder objectives will not differ from those of the management and directors of an organization. However, it is always important to pay attention to the role of non-executive directors, who have an important task in supervising the decision-making process. Without good supervision, management can make serious mistakes with great impact on shareholder value (see Goodijk, 2012).

(20)

scheme. Although PFG may borrow elements from the stewardship theory and consider the PFG system as a co-operating coalition, it is incorrect to speak about shareholders or owners of a pension fund.

Finally, before presenting our research model, we will look into the differences and similarities of the Anglo-Saxon and Rhineland governance models, often implemented as one-tier or two-one-tier board models. Their differences are important for analysing PFG systems and for reproducing and analysing the outcome of the research interviews (see chapters 4 and 5). One-tier and two-tier model dichotomy plays an important role in the conclusions and recommendations of this research.

2.6 Anglo-Saxon vs. Rhineland models

Governance models in the Anglo-Saxon world (mainly North America, Great Britain and part of Australia) are driven by individual and financial short-term profit, and pursue shareholder value optimization. These basic principles are the opposite of the Continental European or Rhineland model, which seeks collective success, consensus and long-term stakeholder value. Both models have historic origins that go back to the beginning of the 19th century. As in the case of management models, a large part of the conceptual framework was established during the World Wars that afflicted Europe in the 20th century. Although the two models are found in different parts of the world (Goodijk, 2009: 12), they are seen as opposites for the sake of this research.

The origin of the Anglo-Saxon way of governance lies with great scholars such as Thomas Macaulay and Adam Smith, who believed in the manufacturability of both society and economy. In this field of governance, certain preconditions must be met: fellow-feeling, fair share, fair trade and absolute transparency. In practice, these preconditions are based on the assumption that the world is a perfect and harmonic environment; this is considered to be the SOLL-situation. It is clear that the real world and the real economy are an IST-situation, and much will have to be done to make it similar to the (unrealistic?) SOLL-situation. The (simple) belief in the SOLL-situation is, however, still very much alive (Brouwer & Moerman, 2005: 26)!

(21)

accomplished. This is what the Germans call Auftragstaktik. They had confidence in the competences of their soldiers (employees) and the management skills of the commanders (management). In turn, the commanders ignored the chaos of the battle field and taught their soldiers to cope with it. In other words: the commanders took certain uncertainties for granted. This mission-orientation is an important element in the Rhineland way of thinking and has an impact on several aspects of organizations, such as:

• Decentralisation of responsibilities and powers;

• Emphasis on cooperation between departments;

• A situational style of leadership, aimed at the growing competences of employees;

• A clear and simple management and communications structure, where much is left to the initiative of subordinates;

• Introduction of organizational development and quality within the organizations’ management.

At the opposite end of this interesting and holistic spectrum, there is the Anglo-Saxon ‘way of life’, which was far more dominant in the decades after WWII. The Anglo-Saxon concept of management was derived from the scientific management concept mentioned earlier and was the winning concept in WWII. It dovetailed perfectly with the circumstances of the American army at that time: a fast-growing organization with semi-educated soldiers and mediocre management. The ingredients of an, at first glance, average army characterised the American management concept as orientated towards simple assignments, or Befehlstaktik. This concept encompasses among other things:

• Working with detailed plans that are to be executed meticulously;

• Hierarchic structures;

• Centralisation of responsibilities and powers;

• A meticulous planning & control cycle;

• Emphasis on control and rationalisation of procedures;

• Emphasis on standardization and uniformity.

Americans acted on the belief that developments in organizational surroundings can be analysed and controlled. On that basis, it is possible to make plans that can, in turn, be controlled by means of a planning & control cycle. They act upon uncertain certainties and

manufacturability of contingencies (Brouwer & Moerman, 2005: 48). The differences

(22)

prototypes from the point of view of decision making, since the separation of control power from managing power is closely related to the decision-making process.

Decision-making process

To finalize the discussion on the dichotomy of governance models, we will look into the background and differences of one-tier and two-tier boards. Governance models are closely related to either the Anglo-Saxon world (one-tier board) or the Rhineland (two-tier board). This specific distinction, expressed by the degree in which board members are involved in the decision-making process, will prove to be important in the discussion on the results and conclusions of the research interviews. To formalize the boards’ involvement in responsiveness to various stakeholders, theoretical models on corporate decision-making processes generally identify a sequence of decision-making steps. In general, these steps concentrate on the formulation, implementation and evaluation/monitoring of decisions. The widely applied model of Fama & Jensen (1983a: 303) recognizes the following four steps in decision making:

FIGURE 2.3: The decision-making process

Source: Fama & Jensen (1983a)

The term ‘decision management’ is used to refer to the combination of the initiation and implementation steps in decision-making (steps one and three). The authors use the term ‘decision control’ to describe the combination of the ratification and monitoring steps in the process of decision-making (steps two and four). The model is used to describe the formal independence of corporate boards of directors. Basically, design strategies that are applied to enhance the formal independence of corporate boards seek to separate decision management from decision control in decision-making (Maassen, 2002: 31). In practice, this separation is implemented through the introduction of a two-tier board, in which the decision-making steps are formally separated. Decision management is delegated to the managing directors in the executive management board. Decision control lies in the hands of the non-executive supervisory directors in the supervisory board. One-tier boards are formally based on a structure that integrates the four steps in decision-making (see figure 2.3). It is common for corporations to have one or more committees to support both the

1. Initiation – the generation of proposals for resource utilization and structuring of contracts;

2. Ratification – choice of the decision initiatives to be implemented

3. Implementation – execution of ratified decisions; and

4. Monitoring – measurement of the performance of decision agents ans implementation of rewards.

Decision control Decision

(23)

operational and supervisory bodies in the organization. Research shows that one-tier board structures generally have more committees with delegated tasks than two-tier board structures8 (Maassen et al., 2005: 650).

FIGURE 2.4:

The Formal Separation and Integration of Decision-making Steps in Board Model Prototypes

Source: Maassen, 2002: 31

Fama & Jensen (1983a) link decision management and decision control, agency problems (“residual claimants bearing”) and complex or non-complex organizations. The degree of complexity of organizations involves the degree of concentration of relevant decision-making information to one or several agents. In pension funds, such information often is concentrated in a small number of people who help boards make decisions that affect all stakeholders, i.e. anywhere between a few hundred to thousands of pension scheme participants. Although a pension fund as a whole is large, the decision-making unit is, relatively, quite small. In such cases, it is best to allocate both decision control and decision management to the small group of agents that has access to the relevant information (Fama & Jensen, 1983a: 305), i.e. a one-tier board. But is a one-tier board model the right model for all pension funds? What agency problems or possible conflicts of interest can be expected? What implications does a one-tier board have? We will leave these questions for now, but will return to them in chapter five in a discussion on the theoretical and practical issues of PFG. The discussion on one-tier boards and two-tier boards is very relevant to PFG: the bill on PFG reform referred to above allows pension funds to choose from three models - one based on the one-tier board and two based on the two-tier board. In chapter three we will introduce our research model, which may help in designing a sound governance model that takes all internal and external contingencies into account.

8

On average, one-tier boards have 3.4 committees and two-tier boards have 2.7 committees (Maassen, 2010: 27).

The SupervisoryBoard

In charge of decision control

The Management Board

In charge of decision management

The Board of Directors

In charge of decision management and

decision control

(24)

Convergence

Over the past decades, Anglo-Saxon and Rhineland corporate governance models have converged. This development appears to be important for the discussion on PFG models as the effectiveness of boards is at stake. In the sixties, research conducted by Mayo and McGregor proved that human motivation probably was a more important factor for a high rate of production than the focus on scientific management (Brouwer & Moerman, 2005). Famous were the Hawthorne studies9 in which the productivity rate of employees that were not directly involved in the research equalled that of the group of employees that was involved. Increasing interest in employees in general was enough to make all employees work harder. This increased interest in the ‘human factor’ is very much in line with the Rhineland way of governing organizations.

On the other hand, attention for control and planning cycles increased in Rhineland corporations. Developments such as quality control (ISO, TQM), Six Sigma and LEAN provided a far more scientific view on management, a characteristic of Anglo-Saxon models. Other developments that dovetail better with Anglo-Saxon models but get more attention in the Rhineland areas are:

• A (mostly short-term) shareholder value approach;

• A trend towards the one-tier board as the one and only model;

• The hierarchic leadership structures and disappearance of more horizontal structures;

• A jurisdictional conflict approach that gains attention;

• An instrumental human capital approach. (Goodijk, 2009: 18)

These developments paved the way for the one-tier board model and for ousting the Rhineland approach of two-tier structures, without abandoning the useful elements of the latter.

Specific two-tier rules for PFG in Germany (it is mandatory to combine a managing and supervisory board) tend to converge to form the widespread system of management by a single management body because of the intensive interaction of the two boards in a two-tier model (Stewart & Yermo, 2008: 238).

9

(25)

2.7 Conclusions

Looking at common corporate governance models, a number of important factors can be identified for finding the right model. An important first step in this process is to assess the degree of bureaucracy and stakeholder involvement that directors desire within their organization.

Our literature research into theories of the firm allows us to deduce a limited number of relevant factors. The stakeholder involvement mentioned earlier and the cost of governance (TCE) are important, but we wonder what other factors could be applied to choose the correct governance model for a pension fund. In that search, we have identified more relevant factors, which we will elaborate in the next chapter.

At the same time, the emerging convergence of Anglo-Saxon and Rhineland governance models, leading to new and innovative designs, must be taken into account. Is it possible to design a governance model that incorporates the best of both worlds? In answering this question, it is important to learn from existing research that shows that one-tier boards make more use of board committees. This may mean that more alignment is necessary because more committee members will be active. PFG specifically implies that several specialized committees will be in place for the preparation of various policies for pension fund decision making. A board delegation may participate in some or all of the committees. The degree of board participation in administering the pension fund is yet another factor to be assessed for the choosing of a suitable governance model. For example, what is the influence on supervision? What is the link with other important aspects of pension fund governance?

Much is to be learnt from the stakeholder perspective of governance, which provides a structured method for the identification of prominent stakeholders. Various scholars have developed useful ways for operationalizing this search (compare Mitchell et al., 1997; Donaldson & Preston, 1995 or Rowley, 1997).

(26)

3 Research model

3.1 Introduction

In recent years, interest in PFG has grown rapidly, and includes numerous publications on PFG all over the world. Because of the maturity and quality of the pensions industry in the various areas, we have concentrated on literature from Canada and the European Union, although the Australian pensions industry is known to be well developed as well10. Ambachtsheer (2007) notes that Australia and several European nations have “radically reshaped the pension paradigm”. The Dutch pension system is named as example for other countries around the world (Ambachtsheer, 2007: 42).

Scholars have argued that the governance of non-corporates deserves more attention (see Maassen, 2002: 200 and Mallin, 2010: 345). A remarkable detail is, that the number of publications on PFG has increased sharply in recent years. Most of the specific papers and books are no older than five to ten years (e.g. Stewart & Yermo, 2008; Ambachtsheer, 2007; Clark & Urwin, 2010), which indicates that even scholars and other authorities are still analysing and testing the possibilities of applying optimal governance models to pension funds. The time is ripe for examining specific aspects of PFG more closely and to carry out general and specific research on relevant factors for choosing the correct pension fund governance model.

3.1.1 Research questions

The central question in this research is:

Which factors are relevant to the design of a pension fund governance model?

We gave the first set of research sub-questions in section 2.1, before elaborating on theories of the firm. The following sub-questions concern our research model as presented in this chapter:

Are the right sub-factors identified?

What kind of issues lead to the identification of general factors?

In what way are (internal) supervision and the decision-making process interrelated?

10

(27)

3.2 Research model

Our research model is presented at the end of this section. The model presumes that there are four relevant main factors for choosing or designing a PFG model, based on the conclusion from our literature research in chapter two and the analysis in this section. Three factors are endogenous in nature and depend on the internal contingencies of pension funds, one factor is exogenous.

It is important to specify that our research model is explorative. As mentioned earlier, the research field is too young to be able to use proven theories that have been studied in-depth by scholars. Instead of a hypothetical model, we inferred a preliminary model that enables further research in the field of pension fund governance models. This explorative research method made it possible to reassess the model in the concluding chapter, after having validated the main factors and sub-factors used in setting up the research model (see section 5.1.2). This precautionary observation means that although our research model resembles common practice it may, at the same time, be possible that we are forced to conclude that more or fewer factors are eligible for inclusion in or exclusion from the model. In addition, factors may have to be combined to achieve a more compact model.

The next sections will set out the general factors and identify several factors. Most sub-factors are inferred from literature and have been elaborated in the previous chapter.

3.2.1 Endogenous factors

Endogenous aspects generally consist of two types of factors: those that can be explained with the aid of literature and those that must be explained using the characteristics of a specific pension fund.

Internal contingencies

The contingencies that usually influence pension funds’ governance models concern the following aspects:

• Sponsor company size;

• Sponsor company governance model;

• Type of pension scheme;

• Legal entity.

(28)

normally adopt a specific governance model for their worldwide operations. Corporate governance models may differ locally, as do governance models in Europe and North America, for example. Apart from the sponsor company governance model, the size of the company may influence the governance model of the related pension fund. In corporate governance, one-tier boards tend to be larger than two-tier boards11. In combination with the fact that one-tier board systems tend to have more committees (Maassen, 2005), this leads to the plausible assumption that small companies will choose for a two-tier system. Apart from the board system, directors (executive and non-executive) have to possess the correct and necessary skills and competences. For that reason, some of the governance tasks of directors may be assigned to external board members, who work alongside internal board members. In practice, most members of internal supervisory bodies of small and large pension funds are contractors appointed by the pension fund’s board.

From a risk-sharing point of view, the type of pension scheme affects PFG. Pension funds with DB schemes normally absorb 100% of the risk of poor investment performance, mostly in combination with an obligation for the sponsor company to supply additional funding. Usually, accrued pension rights are not cut back as a consequence of poor asset performance. DC scheme risks, however, borne only by the participants in the pension scheme; they usually have capital at their disposal that is funded by the premiums employers pay into the pension fund, and that is administrated to the individual “account” of a participant. Potential negative performance of investments is not refunded by the employer but is the sole risk of scheme participants. These contingencies may have consequences for the involvement of scheme participants in the administration of DC scheme pension funds. Employer representation is less important because the employer’s sole responsibility is to fulfil its duty to pay premiums. In DB schemes, the related responsibilities of employer and scheme participants (employees) are different. We assumed that the representation of pensioners in DC or DB pension funds will be similar.

Finally, it is clear that the legal form chosen has implications for the governance model. The composition of the board or board committees of a foundation is different from, for example, those of a cooperative.

Accountability & Transparency

Accountability and transparency are two key characteristics of corporate governance (Maassen, 2002: 87). It is surprising that, where PFG is concerned, transparency as such is not

11

(29)

mentioned in official government publications concerning PFG (OECD, 2009; Allen & Overy/ Boer & Croon, 2004). Nevertheless, all publications emphasize the need for an independent supervisory body. Exceptionally transparent decision-making board procedures may be necessary for the execution of this task, introducing transparency into the board structure. In our research model, we assumed that there is a strong connection between transparency and the supervisory function within the pension fund (see also ‘good governance’).

A number of publications cite the need for specific expertise as a requirement for governing pension funds (i.e. Goodijk, 2006: 5; Ambachtsheer, 2007: 88). This requirement is expected to lead to a higher degree of accountability and a higher decision-making quality. Indeed, as expert knowledge on pension fund management improves, so will decisions made by board members. In fact, several countries have introduced requirements regarding professional skills and qualifications, the majority of which concern economics or law (Stewart & Yerno, 2008: 240). The Anglo-Saxon areas seem to lag behind the other developed countries, although Australia has introduced drastic legal reforms over the past few years.

The decision-making process can boost both the accountability of pension fund boards and administrative transparency. Although pension fund decisions are always prepared by committees whose members include representatives of the main stakeholders, the efficiency of decision making is still open to improvement. Research and case-studies of institutional decision making suggest that bounded rationality is a common feature of financial decision making. There is even evidence that financial decision making is strongly influenced by past commitments and current relationships (Clark & Urwin, 2010: 66). It has therefore been argued that pension fund boards should work on their decision-making processes and even further optimize them by innovating them. The decision-making process gives a strong boost to transparency and, therefore, accountability: if a decision-making process is transparent, pension fund members will know who can be held responsible for important decisions concerning their accrued and future pension rights. Several of the recommendations of the OECD paper on PFG (Stewart & Yerno, 2008) concern decision-making processes for pension fund boards. Codes of conduct and specific delegation of tasks are important features.

(30)

Good governance

Good governance is another key factor for choosing a suitable governance model. It is operationalized by the running of the pension scheme, (internal) supervision, and potential conflicts of interest. It is virtually impossible to administrate a pension fund unless these three factors are sustainably incorporated in its governance. Above all, good internal supervision is essential to ensure transparency of board processes. Good governance has since been defined for all kinds of organizations in official publications, for example in the Cadbury Report of 1992, the 1995 Greenbury Report and the Myners Reports of 2001 and 2008 (Mallin, 2010: 29). Obviously, history has shown that requirements for corporate governance are perishable and are determined by the era in which they were introduced. In The Netherlands, the Code Tabaksblat, drawn up by a governmental commission chaired by Morris Tabaksblat, a former Chief Executive Officer of Unilever and Reed Elsevier, is a fine example of corporate governance regulations. The Code consists of more than 100 rules on the tasks, task delivery, and payment of executive committee members.

There are also good governance rules that apply exclusively to non-profit organizations and charity organizations, for example in the UK, Good Governance Standard for Public Services:

i. Good governance means focusing on the organizations’ purpose and on outcomes for citizens and service users;

ii. Good governance means performing effectively in clearly defined functions and roles; iii. Good governance means promoting values for the whole organization and

demonstrating the values of good governance through behaviour;

iv. Good governance means taking informed, transparent decisions and managing risk; v. Good governance means developing the capacity and capability of the governing

body to be effective;

vi. Good governance means engaging stakeholders and making accountability real. (Mallin, 2010: 50)

Is it a coincidence that all of these guidance rules could be applied to pension funds?

(31)

Community Sector12. In the field of good governance, pension funds may learn from best

practices in the sectors mentioned above. Moreover, research in the United Kingdom concluded that pension fund trustees are comfortable with the idea of adopting corporate governance practice, especially now that company law has been reformed to diminish apparent conflicts of interest on boards (Clark & Urwin, 2010: 68).

Although the (international) debate on corporate governance focuses on the executive board, supervision by non-executives, interaction between the two, and stakeholder groups, the discussion on good governance is undecided. On the one hand, in the USA, the focus is on shareholders and the boards’ responsibility for an increase in shareholder value. On the other hand, the OECD takes into account the responsibilities of boards towards different stakeholders. And finally, EU member states combine the two and take into account the responsibilities of boards towards both shareholders and stakeholders (Goodijk, 2009: 23). Nevertheless, all good governance rules focus on processes and their quality both within boards and between board members, as well as on the board’s responsibility for all stakes, whether inside or outside the organization.

For a specific yet anonymous pension fund, Goodijk (2006) provides a list of key aspects to consider when choosing a governance model. The list is based on the factors mentioned above, as well as on countervailing power, checks and balances, and avoiding conflicts of interest. The design of internal supervision is another point of consideration, since a pension fund has several options to implement this task. For this reason, supervision is named as one of the sub-factors of good governance. In practice, there appears to be a strong tie between internal supervision and the decision-making process. One of the main tasks of internal supervisors (usually independent, externally recruited experts) is to evaluate the board’s decision-making processes critically (Nationaal Register, 2011: 76). Therefore, effective internal supervision must be linked to the board’s decision-making process, one of the sub-factors of Accountability & Transparency.

Before discussing the final sub-factor, conflicts of interest, we will briefly name the types of conflicts that may arise. The first may arise, as we have seen earlier on, in the relationship between principals and agents. An agent may, in his organizational role, take decisions that are not in the principal’s best interest. Two further types of conflicts of interest are described as confusions of interest, which occur when board members confuse their personal and business interests or when two or more board members confuse their joint and personal interests. Although all sorts of conflicts of interest will most certainly be avoided by good internal supervision of the board’s performance and processes, this factor does play a role in

12

(32)

designing a governance model. As stated in the earlier discussion on transaction cost economics, the avoidance of conflicts of interest (and even possible agency problems) has its costs: it does not come cheaply. In establishing a governance model, certain measures must be taken, for example the introduction of a code of conduct. Conflicts of interest can then be resolved effectively (Stewart & Yermo, 2008: 225).

Summarizing, there are a number of ways to operationalize good governance. The focus should be on the following three aspects:

• Board processes (including selection and nomination) and processes between board members with a focus on possible conflicts of interest;

• Clear task delegation and specialization;

• Internal supervision.

Several studies (Ambachtsheer et al., 2006, 2007; Clark & Urwin, 2007; Mercer) indicate that good governance is linked to good performance. Good governance also reduces regulatory costs.

At this point, we must point out the potential overlap between the key factors Accountability & Transparency and Good Governance. These two factors appear to be closely related. Why not align them and make the research model easier to understand and to apply? The answer lies in the specific responsibility fields of PFG. In our view, it is important to separate board (decision making) processes from running a pension scheme. The latter is one of the primary tasks of the board and may easily be done without taking serious account of pension fund members stakes. However, these stakes have recently gained importance and have resulted in specific requirements for board members involving expertise on the technical details of pension schemes that is also required for well-informed decision making. Nowadays, a pension fund be run without such skills; the requirement of expert pension knowhow is also supported by the only exogenous factor: pressure by external parties.

3.2.2 Exogenous factors

The last key factor to be considered for establishment of the PFG model involves the external environment of pension funds.

Governing pension funds

A

R

Preface

Contents

Executive summary

1

Introduction

2

Literature review

3

Research model