1
Home Bias Levels of Dutch Pension Funds
Political and Societal Pressure Affecting Pension Fund’s Investment Decisions
David Looijschelder
11945982
Supervisor:
dr. D.J. (Joost) Berkhout
Second Reader:
dr. M.C. (Marcel) Hanegraaff
June 2019
2
Abstract
Dutch society is aging rapidly. Over the last decade, the national pension pot has doubled and the number of retirees has grown quickly. This has increased the societal and political relevance of how pension funds invest their assets. In the aftermath of the financial crisis, multiple political actors publicly demanded from pension funds to increase the amount of domestic investments. However, the amount of investments in the Netherlands was already disproportionately high relative to the size of the national economy; a so-called home bias was already visible. By combining political and
economic views on the issue of home bias, this research studies the external forces affecting pension fund’s investment decisions. Hereby, I attempt to fill the research gap that has emerged in the existing literature on agenda-setting, regulatory capture, and the revolving door mechanism. These scientific subfields study business influence on policy output extensively, but largely ignore influence in the opposite direction: how politicians affect corporate decision making. My findings show that political connectedness is highest among the two largest pension funds (ABP and PFZW), but this does not translate into higher home bias levels compared to their non-connected peers. I argue this is caused by the negative relation between home bias and the size of assets under management of a pension fund. Second, the findings show that industry-wide pension funds have higher political connectedness and home bias levels than corporate pension funds, indicating that industry-wide pension have a larger societal role. Third, the falling home bias trend prior to the financial crisis has not continued in the post-crisis period. Although home bias levels are not much higher than pre-crisis figures, this finding is remarkable as existing literature presents a wide variety of reasons why home bias levels should be diminishing. I argue that the discontinuation of falling home bias levels is partly caused by external forces steering pension funds into a more societal role, but this is certainly also stimulated by economic conditions and profitability of certain Dutch asset classes. In some cases, politicians see pension funds as a tool and instrument to boost the national economy, to gain political momentum, and to satisfy their electorate. Political pressure, and interaction between political and business actors in general, is not normatively wrong. However, it is paramount that pension fund boards keep being aware of their primary duty: serving the interests of the participants of their pension scheme. If pension funds start serving the interests of the government instead, a reversed regulatory capture emerges.
3
Contents
I. Introduction 4
II. Literature Review 10
III. Theory 13
Home Bias Explanations: Beyond Economic Models 13 Omissions in Existing Literature on Political-Corporate Connections 14
Politically Connected Boards 16
IV. Research Design 23
V. Results & Discussion 30
VI. Conclusion 46
VII. Literature 50
4
I. Introduction
Similar to most other Western European countries, Dutch society is aging and is do so more rapidly: currently, nearly half of society - more than 7 million citizens - is over 50 years old (CBS, 2014). There are nearly 3.5 million retirees in the Netherlands (CBS, 2018), while there were only 2.5 million pensioners at the turn of the century (CBS, 2012). This trend explains why pension funds are
becoming increasingly important; not just as social institutions taking care of the pension payments for Dutch citizens, but also as gigantic financial and even powerful political actors. Pension funds’
responsibilities include the management of both liabilities and assets. Liabilities of a pension fund are current and future pension payments to retirees, active participants (currently accruing a pension), and deferred participants (early leavers e.g. due to switching jobs) of the pension fund (Blake & Inkmann, 2007). Pension funds invest their assets in order to finance future obligations. In this context they are referred to as institutional investors, which are generally resourceful, sophisticated, and above all powerful players in the financial sector (Menkoff, Schmeling & Schmidt, 2009). To put the Dutch national pension pot – one of the largest in the world and covering 91% of the employed workforce - in perspective: the total amount of assets in the national pension pot is approximately $1517 billion (Willis Towers Watson, 2019), which equals circa 170% of the Gross Domestic Product (GDP) of the Netherlands. Currently, the amount of pension assets as a percentage of GDP is almost twice as much as it was ten years ago (89% of GDP in 2008). The size of the pension pot shows the enormous interests involved with pension funds’ activities, not only for its retirees and its active participants, but also for policymakers and the social partners such as employee and employer organisations.
The board members of each pension fund bear the final responsibility for the management of assets and liabilities. On a daily basis, liability management includes e.g. determining the amount of pension payments – discussions on indexation or cutting of pension payments - and the amount that active participants of the pension scheme contribute every month. Within a pension fund various
stakeholders might have opposing interests: e.g. retirees generally demand indexation of pension payments, while this might force the pension fund into an unsustainable situation, which would jeopardize future payments to active participants; a discussion on intergenerational solidarity. The management of assets on the other hand, which is the primary focus of this research, includes
decisions of how the assets are invested. Considerations on how to allocate the assets can both be short term (tactical asset allocation) and long term (strategic asset allocation). The tactical asset allocation (TAA) focuses on taking advantage of market conditions; resulting in short-term tactical adjustments. The strategic asset allocation (SAA) is primarily based on ideology, investment beliefs, and long term investment objectives of a pension funds. The SAA prescribes the boundaries for the allocation of the investment portfolio across various asset classes such as: cash, fixed income (government and
5
aspect of the SAA is the regional diversification: e.g. the amount of investments allocated to Asia, Latin America, Europe, or even more precisely: the Netherlands.
Most recent data (2013) published by the Dutch central bank, De Nederlandsche Bank (DNB), show that Dutch pension funds invest 14.7% of their assets in the Netherlands, while the Dutch economy is only 1.08% of the global economy (PP 33746-4, 2013). In 2013, the investments in Dutch equity equalled 1.12% of the total portfolio; 2.6% was allocated to Dutch real estate and 4.73% of the portfolio consisted of Dutch government debt (PP 33746-4, 2013). So Dutch pension funds tend to allocate substantial segments of their assets to the Netherlands, also called orange investments. Proportionately, the amount of orange investments by pension funds has even been much higher in the past. Until the mid-1990s, international diversification possibilities were largely ignored by Dutch institutional investors, with a typical pension fund portfolio consisting of 25% to 30% exposure to foreign markets (Wilcox & Cavaglia, 1997). This has changed due to increased global integration, the increasingly large size of the national pension pot, higher levels of capital flow mobility, and fewer political restrictions (Rubbaniy, Van Lelyveld & Verschoor, 2014). Also, pension funds needed to diversify their investment portfolio to find enough investment opportunities and to seek higher return on investments. Rubbaniy, Van Lelyveld, and Verschoor show that the percentage of orange
investments has declined between 1992 and 2006: from 37% to 13% of the total portfolio (Rubbaniy, Van Lelyveld, Verschoor, 2014). Still, there exists a bias towards orange investments among Dutch pension funds: a so-called home bias. Home bias is the presence of a higher proportion of domestic holdings in an international portfolio compared to what should be expected considering standard portfolio theory and the related risk and return expectations (Rubbaniy, Van Lelyveld & Verschoor, 2014; Darvas & Schoenmaker, 2017).
First of all, why is it relevant to study home bias in pension fund’s portfolios? What the scientific and societal relevance of this issue is, becomes clear if one knows the potential consequences of a large home bias. Home bias implies incomplete usage of diversification possibilities and is therefore not in line with the optimal risk and return measures for an investment portfolio (Li, Gao, Huang, 2007). Generally, international diversification offers higher returns relative to a non-diversified portfolio (Grubel, 1968). Existing literature suggests this could result in underperformance of the portfolio, which could ultimately harm the already low coverage ratios of Dutch pension funds. This can potentially for ce pension funds to cut back on future pension payments. DNB warns for a ‘triple whammy’ due to a large home bias in case of a national economic downturn (PensioenPro, 2018). In that case, the Dutch real estate and job market are negatively influenced, but this is now accompanied by pension investments getting hit hard due to a proportionally large allocation to the Dutch market (PensioenPro, 2018).
6
Partly, home bias can be explained by the fact that Dutch government bonds are AAA-rated (highest credit worthiness) Eurozone bonds (Lemay & Dyck, 2019). Pension funds are generally seen as defensive investors seeking for a safe long-term investment opportunities. However, only investing in the Netherlands for the safe nature of government bonds would not explain the 13-14% of orange investments in the portfolio.
Pension funds are highly sophisticated and professional investors, capable of largely overcoming traditional barriers to invest abroad (Chan, Covring, Ng, 2009; Lewis 1999; Menkoff, Schmeling & Schmidt, 2009). Therefore, the presence of home bias is particularly surprising among pension funds. This makes it an intriguing phenomenon, which has to be analysed from a broader perspective than the scope of common home bias research. Dealing with pension funds as if they are purely financial institutions without a societal and political relevance would be naive and erroneous. Existing literature finds various reasons for the presence of a home bias: different scientific disciplines have suggested multiple financial, cultural, and behavioural explanations. However, a convincing explanation for a home bias with a focus on pension funds and governance structures is still missing. This thesis intends to add to the discussion on the presence of a home bias among pension funds by investigating decision making processes within pension funds from a governance, strategic decision making, and political angle; it aims to generate an insight into the connection between pension funds and political actors. Why is this angle the best starting point to examine home bias in pension fund’ portfolios? First, to understand possible influence exercised by external forces on the pension fund’s decision making, one should know the composition of pension fund boards. According to the Pensioenwet (PW), the Dutch pension law implemented in 2007, pension fund boards can be formed in three manners: members being appointed on the basis of parity, independence or a hybrid version (PW Article 99). Paritary boards, in which all stakeholders - employees, employers and retirees - are represented as balanced as possible, are most common in the Netherlands (PW Article 100). Board members can be former politicians, such as Corien Wortmann-Kool, the chairman of the largest pension fund of the Netherlands: ABP. Naturally, this does not necessarily imply any political influence, but having politically-connected boards makes it academically relevant to study the effect of their presence on pension fund’s decisions. Collectively, all of the board members bear the shared responsibility for the execution of the pension scheme (PW Article 102). Each member must take the interests of all stakeholders into account. Although the professionality of board members has improved recently, partiality of each representative still plays an important role. In most cases, the representatives are member of one of the Dutch social partners (employee and employer organisations). These organisations, like VNO-NCW, FNV and CNV participate in discussions with the government on working conditions and wages (CAO negotiations) as well. Therefore, feuds related to CAO negotiations can be transferred to pension fund boards, thereby politicizing the pension fund’s discussions. Concluding, the way in which pension fund’ boards are composited makes pension fund
7
decisions of political relevance and shows why the governance-political angle of this research is most valid.
Second, Dutch politicians - including (prime) ministers - tend to see pension funds’ investments as a tool to boost the economy in times of a lack of financing available for the national economy. Possibly, this is amplified by the ‘special and social’ role that the European Court of Justice ascribed to
industry-wide pension funds in 1999 (Case C-67/96), thereby implying that pension funds have a societal role with a broader scope than just their participants or retirees. In the aftermath of the 2008 financial crisis, pension funds were encouraged to complement or even substitute banks, common creditors of the economy, that were subject to new strict financial regulation curtailing their lending possibilities. Henk Kamp, former Minister of Economic Affairs and Climate regularly criticized pension funds for not putting the Dutch national and societal interests first (Wolzak, 2017 I & II; Daalder, 2017). In 2013, Kamp initiated Catshuis talks with key pension funds and their asset managers, which ultimately resulted in the foundation of the Nederlandse Investerings Instelling (NLII) (Kamp, 2013). This organisation was designed to be complementary to the banking sector by providing credit to infrastructural projects, small and medium-sized enterprises (SMEs), private housing market, school buildings, and health care real estate (PP 33746-1, 2013). Although, the institution failed and was dismantled in 2018, its intention was clear: to serve as a connection between supply and demand of capital and to select, bundle, and standardise appropriate investable projects (PP 33746-1, 2013). NLII forms an example of political support and demand for more orange investments. Some pension funds and politicians however argue that this restricts pension funds’ autonomy and that pension funds should not be seen as an ‘ATM’ (Wolzak I, 2017).
The academic relevance for political science of the issue of this thesis becomes clear by linking the topic to the research gap this thesis attempts to fill. In common agenda-setting literature and political economy research, the revolving door mechanism, regulatory capture, and biased policy output are frequently studied. Regulatory capture ‘occurs when bureaucrats, regulators and politicians cease to serve some notion of a wider collective public interest and begin to systematically favour specific vested interests, usually the very interests they were supposed to regulate and restrain for the wider public interest’ (Baker, 2010, 648). The revolving door mechanism is the flow of people between public and private sectors in both directions, leading to ‘colonization’ of regulatory agencies and dysfunctional incentive structures for regulators (Baker, 2010, 652). The revolving door, together with skewed distribution of resources and intellectual capture, result in regulatory capture, according to Baker. Existing studies point to an interconnectedness between political and business actors, ultimately causing policy output to be biased in favour of narrow business interests. There is ample research showing the influence and agenda-setting capabilities of business actors within the political arena. However, the direction of influence the other way around, politicians affecting corporate decision making, is insufficiently studied. This is remarkable, because the revolving door mechanism
8
is a two-way process, as described by Baker. Furthermore, there is no clear reasoning why regulatory capture would only work in one direction either. The pension sector, being on the border of the public-private domain and consisting of a few semi-public organisations, is particularly suitable to study the influence of politicians on corporate decision making. Thereby this study attempts to fill an important research gap in the existing political economy literature. The omission in the existing literature on political-corporate interconnectedness is more extensively described in Section III.
The image that emerges from this section is that asset allocation decisions can for pension funds be based on more than just financial risk and return expectations. This thesis studies what precedes regional allocation decisions by a pension fund. It investigates the governance structures of pension fund boards and researches if external forces (such as politicians, members of the social partners, or society as a whole) can influence these processes. It does so by comparing the home bias of the fifteen largest Dutch pension funds and relating this to the politically connectedness of their boards. It is examined why some pension funds have home bias levels over thirty percent while others have a home bias of less than ten percent. Also, this study attempts to explain the downward trend in home bias prior to the crisis and studies whether this trend has continued after the crisis. The dependent variable home bias is studied over a timeframe of five years (2013-2017). I argue that home bias levels of pension funds is relevant to study as it is a highly sensitive topic for pension funds, politicians, and media: it shows the societal role of pension funds beyond providing solid income for its retirees. The issue is highly politicized and seems to permanently be an issue of public debate. In this thesis I therefore study reasons for home bias levels beyond pure economic models. As a fundament of this study, I formulate the following research question:
Why do Dutch pension funds have different levels of home bias?
Section II and III together form the literature review part of this thesis. Section II briefly outlines the institutional settings of Dutch pension funds. It shows trends of interconnectedness, herding behaviour and elaborates on the legal context in which pension funds operate. It deals with the different types of pension funds and clarifies what the relation is between pension funds and their asset managers. This section provides the reader with the necessary knowledge to fully grasp the reasoning in the remainder of this study.
Section III describes what explanations the existing literature has for the presence of a home bias beyond pure economic models. Furthermore, this section deals with omissions in the existing literature regarding the agenda-setting capabilities of political actors in corporate boards. Most agenda-setting theories and literature on influence, regulatory capture, and interaction between business and political actors focus on the effects that lobbying activities have on (biased) policy output. However, there is little or no research showing the effect of the revolving door mechanism working in the opposite direction: how government officials affect corporate decision making. Also, this section deals with
9
what the rationale is for firms to appoint politically connected members in their board and connects this to the composition of pension funds and its governance structures. Finally, in Section III I formulate three hypotheses, which forms the fundament for the remainder of this thesis.
Section IV explains the used mixed-method research design. In this section I show how this thesis builds upon three pillars of various types of research methods. First, I analyse annual reports of pension funds to find a potential trend in orange investments over time. This first pillar makes a quantitative comparison between pension funds. Second, I study the role of political pressures
influencing orange investments. Pillar II aims to paint a comprehensive image of political pressure on pension fund decision making. I study the background of pension fund board members to find out whether they have a political background. Pillar III contains interviews with key players in the
political-pension fund arena to retrace, verify, and build a context around the findings from Pillar I and II. This includes (former) politicians, high-level employees of interest associations or pension funds, and their asset managers.
Section V presents the findings. Again, the findings section consists of three pillars of data, but this is structured along the times of the three hypotheses. This section also includes the discussion of the findings: it shows the impediments in terms of threats to external/internal validity, reliability and measurement validity. Subsequently, it is discussed how this effects the possibility to draw
conclusions from my findings. Section VI, the final part of this thesis, contains the conclusions of this research. This includes the scientific relevance, major findings and most relevant shortcomings of this thesis.
10
II. Literature review
Institutional Settings of Dutch Pension Funds
To be able to understand governance structures within pension funds and decision making processes ultimately resulting in investment outcomes, it is paramount to grasp the institutional and regulatory settings in which Dutch pension funds operate. The Dutch pension sector has experienced a
consolidation over the last decades. The number of pension funds declined from 833 in 2007 to 297 in 2016 (Bauer, Bonetti & Broeders, 2018, 2). This has a couple consequences for the nature of
governance structures within pension funds, which are extensively described in Section III and V. There can be identified three types of different pension funds:
- Industry-wide pension funds: these funds serve a whole sector e.g. the civil service,
construction industry, the hotel and catering industry. Dutch examples - 52 in total - are: ABP, PFZW, PMT, PME, BpfBOUW, and Pensioenfonds Vervoer.
- Corporate pension funds are company specific pension funds (166 in the Netherlands): Stichting Shell Pensioenfonds, Pensioenfonds ING, ABN AMRO Pensioenfonds, and Philips Pensioenfonds.
- Pension funds for independent professionals such as medical specialists and dentists. This subgroup of pension funds is not taken into account in the research part of this thesis due to the modest size and rareness of this type of pension funds.
Table 1 gives an overview of the fifteen largest pension funds in the Netherlands (PensioenPro, 2018). These fifteen pension funds form the basis of the data section of this thesis and include both corporate as well as industry-wide pension funds.
Table 1. List of 15 largest Pension fund of the Netherlands and their Assets under Management
Pension funds AuM (x 1 mil)
ABP 477.289
PFZW 207.003
PMT 71.791
BpfBOUW 63.383
PME 46.962
Stichting Shell Pensioenfonds 28.051 Pensioenfonds ING 27.261 ABN Amro pensioenfonds 26.698 Pensioenfonds PGB 25.633 Rabobank Pensioenfonds 25.449 Pensioenfonds Vervoer 24.795 Pensioenfonds Detailhandel 20.249 Philips Pensioenfonds 17.531 BPL Pensioen 17.134 Spoorwegpensioenfonds 16.436
11
If pension funds have a standard asset manager, this asset manager is referred to as a fiduciary manager or fiduciary. Fiduciary management, a system pioneered in the Netherlands, involves outsourcing day-to-day management of a pension scheme to a lead manager. This process includes a high degree of transparency so that fiduciary decisions can be scrutinized easily. For example: APG is the asset manager/fiduciary for ABP and BpfBOUW, PGGM is the fiduciary for PFZW, MN for PMT and PME, and Blue Sky Group for KLM Pensioenfonds. This construction is relevant as multiple interviewees mention these asset managers. Also, one of the interviewees is a managing director of APG, the largest fiduciary of the Netherlands. They are not left out of the analysis of this study since fiduciaries can be even larger than pension funds in terms of assets under management. Fiduciaries are however not the main unit of analysis in this thesis. Although fiduciaries make the investment
decisions on a daily basis, final formal responsibility for the investments still lies in the hands of the pension funds.
Before analysing informal or indirect influence by political actors, it is important to understand the formal regulatory environment in which pension funds operate. Roughly, regulating institutional investors can be done in two ways: quantitative investment restrictions or regulation on the basis of the prudent person principle (PP 33746-4, 2013). Quantitative investment restrictions are specific rules, set up by the government or the supervisor, to quantitatively restrict investments in certain investment categories (PP 33746-4, 2013). The prudent person principle on the other hand, is based on the premise that pension funds should be prudent investors (PP 33746-4, 2013). Generally, this does not include quantitative restrictions in specific investment categories, but rather builds on pension funds’ prudence, internal control mechanisms and governance structures during the investment process (PP 33746-4). In practice, these two ways of regulation are somewhat indistinct and there exists a continuum of quantitative restrictions and the prudent person principle (PP 33746-4, 2013). The Dutch Financial Toetsingskader (FTK), which is part of the PW and the European IORP II directive outlines the regulatory environment for pension funds to operate in. For both laws, the prudent person principle is the prevalent regulatory philosophy (PP 33746-4, 2013). Article 19 of the IORP II directive (and FTK Article 13) on the investment rules for pension funds stipulates that investments should be in line with the prudent person principle and – relevant in the context of this thesis - that ‘the assets shall be invested in the best long-term interests of members and beneficiaries as a whole’ (IORP II Article 19.1a) and ‘in the case of a potential conflict of interest, an IORP, or the entity which manages its portfolio, shall ensure that the investment is made in the sole interest of members and beneficiaries’ (IORP II Article 19.1a). Evenly important here is that ‘the assets shall be properly diversified in such a way as to avoid excessive reliance on any particular asset, issuer or group of undertakings and accumulations of risk in the portfolio as a whole. Investments in assets issued by the same issuer or by issuers belonging to the same group shall not expose an IORP to
12
excessive risk concentration;’ (IORP II, Article 19.1f; FTK 13.5). This is interpreted by Henk Kamp in his role as Minister of Economic Affairs and Climate as: ‘if it is evident that the risk-return proportion is better for investments outside of the Netherlands, pensions funds ought to invest abroad. This follows directly from the prudent person principle.’ (PP 33746-4, 2013).
13
III. Theory
Home Bias Explanations: Beyond Economic Models
In Section I the concept of home bias in pension fund’s portfolios is introduced. Section II describes the institutional and regulatory context in which pension funds operate. Section II intends to build the context around Section III, which discusses the existing literature on home bias. It deals with different arguments provided by various scientific subfields and explains what the omissions in these existing theories are. This is complemented by a newly developed determinant of home bias: external and political influence on pension fund’s investment decision making processes.
To start this section with, I briefly outline the wide variety of explanations for the presence of a home bias, as provided by existing literature. This includes reasoning from economic studies, but also more intermediary disciplines such as behavioural economics and studies taking into account the cultural aspects of investing. I explain and refute the three commonly mentioned reasonings, but briefly address multiple other explanations for home bias as well. First, home bias can be caused by high cross border investing costs (Lewis, 1999) and barriers to international diversification (Errunza & Losq, 1985). These costs or barriers can include: transaction costs, international taxes, or restrictions. If costs of investments in foreign countries outweigh expected returns, an investor chooses to invest domestically (Lewis, 1999). I however argue this argument is less valid for pension funds and their fiduciaries for three reasons. First, pension funds commonly have strong international ties, use foreign asset managers, and sometimes even have offices abroad (such as APG’s offices in New York and Hong Kong). Second, pension funds are well placed to take advantage of diversification possibilities (Rubbaniy, Van Lelyveld & Verschoor, 2014) and have sufficient means to overcome international barriers due to scale advantages and high levels of sophistication. Third, the relevance of barriers to foreign financial markets has declined over the last decades. Van Nieuwenburgh and Veldkamp state: ‘while restrictions on international capital flows may have been a viable explanation for the home bias thirty years ago, they no longer are today’ (Van Nieuwenburgh & Veldkamp, 2009, 1187).
A second reason for the existence of home bias is the informational advantage that domestic investors have over foreign investors (Suh, 2005; Fedenia et al, 2013; Rubbaniy, et al, 2014). This information-based theory can be refuted by taking ‘information choice’ into account. Information immobility is low, access to information is almost infinitive, so investors have an information choice: they are free to choose which information they want to collect (Van Nieuwenburgh & Veldkamp, 2009). Cross-border information flows could offset initial informational advantages that domestic investors generally have. Again, since pension funds are resourceful and highly sophisticated investors, it is highly unlikely this forms the main reasons behind the existence of a home bias.
Strong and Xu (2003) determine a potential third reason for home bias: the tendency of managers to be more optimistic about the domestic market than they are about foreign markets (Strong & Xu, 2003).
14
Interesting here is not how this theory can be refuted, but rather how this optimism emerges. I argue optimism could be the result of social networks between pension fund board member and the boards of companies, in which they are likely to invest in.
Furthermore, scholars from various scientific subfields provide additional explanations causing home bias, such as: familiarity (Huberman, 2001), cultural variables like risk averseness (Anderson et al, 2011), loyalty (Cohen, 2007), patriotism (Morse & Shive, 2008), overconfidence (Graham, Harvey and Huang, 2006), inflation hedging, and hedging of human capital (Wheatley, 2001). The problem of the lion’s share of these arguments is however that it starts from the perspective of US investors. The scarcely available research with a focus broader than the US commonly insufficiently explains the dynamics within mature markets such as the Dutch financial market (Rubbaniy, Van Lelyveld & Verschoor, 2014).
Omissions in existing literature on political-corporate connection
Concluding, there is elaborate and extensive research done to analyse domestic investments by asset managers. Although each explanation is viable within its own subfield of research, the reasons are characterised by fragmentation. Each scientific discipline finds its own explanation for home bias without connecting this to other subfields of research. I argue that a broader approach is needed, which takes into account the complex governance structures within pension funds and external forces
affecting investment decisions. To do so I deal with agenda-setting theories, as well as theories on the benefits of having politically-connected corporate board members. First, I start with a rather surprising omission in the literature on political-corporate influence and interest group activities. Subsequently, I deal with the rationale for companies to appoint politically-connected board members; Second, the institutional differences between industry-wide and corporate pension funds and the related effects on their societal role; Third, I provide examples of post-crisis political influence on pension fund decision making regarding orange investments.
In the broader context of interest group activity and interconnectedness between corporate and
political actors, the prevalent paradigm argues that business interest groups create biased policy output in favour of ‘narrow’ business interests (Schlozman et al, 2015; Schattschneider, 1960). This bias emerges from the process of regulatory capture, which ‘occurs when bureaucrats, regulators and politicians cease to serve some notion of a wider collective public interest and begin to systematically favour specific vested interests, usually the very interests they were supposed to regulate and restrain for the wider public interest’ (Baker, 2010, 648) According to Baker, regulatory capture is based on the revolving door mechanism, intellectual/cognitive capture, and a skewed distribution of resources and lobbying activities (Baker, 2010, 648). The revolving door mechanism is by Baker (2010) defined as the flow of people between public and private sectors in both directions, leading to ‘colonization’ of regulatory agencies and dysfunctional incentive structures for regulators (Baker, 2010, 652). The US
15
example of this process is the Washington-Wall Street corridor: here the usual route includes that individuals move from the industry to a public policy role and subsequently return to the industry. This creates a system of revolving doors and ‘professional ecologies’ (Baker, 2010, 653).
Consequently, the distinction between public and private sectors has become blurred, creating a ‘like-minded’ policy community (Seabrooke & Tsingou, 2009). Intellectual and cognitive capture on the other hand is based on repeated interaction, personal connections and networks between public and private sector individuals (Baker, 2010, 653). This gives personnel of leading banks the possibility to influence the thinking and the mindset of regulators (Baker, 2010, 653).
More specifically, in the context of financial regulatory policymaking Pagliari and Young (2016) find four reasons for low levels of interest group plurality, which causes biased policy outcomes in favour of the ‘narrow few’ (Mattli & Woods, 2009). First, the financial issue area is highly complex and technical, which leads to a heavily asymmetric distribution of information among interest groups (Pagliari & Young, 2016). Second, the wider public has little interest in financial issues during non-crisis times. Low levels of salience and ‘quiet politics’ leave the business interests relatively uncontested (Pagliari & Young, 2016). Third, the increasingly transnational institutional context in which financial regulatory bodies operate is beneficial for the ‘narrow few’ (Pagliari & Young, 2016). Fourth, financial interests are commonly represented by ‘insiders’, who are able to access key
policymakers in an early stage of the policymaking process. Financial business interests therefore benefit from the so-called ‘first-mover advantage’(Pagliari & Young, 2016).
This direction of influence includes corporate political strategies: the proactive actions by firms to create a public policy environment that is favourable to them. But surprisingly little research is done on the opposite direction of influence: the way in which policymakers affect corporate decision making beyond influence via formal legislation. Politicians might experience financial or electoral gains by pressing private companies in a certain direction. In the existing literature, the question remains unanswered why ample research in done on revolving doors, intellectual capture and close networks creating biased policy output, but why on the other hand influence in the opposite direction (from public to private) is barely studied. I argue, the opposite way of influence is just as plausible given the two-way nature of the connection between politics and business. Baker clearly describes regulatory capture as the result of close networks, interconnectedness, the revolving door mechanism, and intellectual capture. I argue it is naive to ignore the other direction of influence: politicians affecting corporate decision making. This direction of influence is equally plausible since politicians can use pension fund’s assets as a tool to generate an economic boost: stimulating domestic
investments, employment and the housing market. This can help the economy flourish and can ultimately translate in electoral gains for a politician.
16
Although authors like Shleifer and Vishny touch upon the personal gains for politicians when influencing business (Shleifer & Vishny, 1994), the works are insufficiently explanatory about the effects of policymakers influencing corporate decision making. Existing literature shows the
interconnectedness between public and private networks, based on e.g. educational background, elite clubs, and the revolving door mechanism, which ultimately results in a biased policy output in favour of business relative to societal interests. This thesis examines similar networks, but intends to provide a clear context in which influencing happens in the opposite direction too.
Politically Connected Boards
In this section I deal with the rationale for the presence of politically-connected members in corporate boards and link this to legislation on the composition of pension fund boards. Subsequently, I deal with three institutional characteristics of pension fund boards potentially forming channels via which external influence could be exercised by politicians or other external actors: political connectedness of individual board members; the presence of representatives from employee and employer organisations in the boards; and the interconnectedness of pension funds, possibly causing a snowball-effect of influence by external actors. Together both political science theories and institutional settings of pension funds translate into the first hypothesis of this thesis. This reasoning has both inductive and deductive origins: based on practices and laws concerning the pension sector (inductive) and agenda-setting aspects of political science theory (deductive).
First, this section determines what the benefits of appointing politically-connected board members are. For pension funds, but also for other types of businesses, recruitment of (former) politicians –
parliamentarians, ministers, and civil servants – offers a large number of benefits, which can be sources for future economic gains (Gonzalez-Bailon et al, 2013, 4). Appointment of politicians within corporate boards results in specific knowledge of the policy process and decision making processes within politics, which grants valuable knowledge on the inner workings of politics (Gonzalez-Bailon et al, 2013, 4). Furthermore, politicians in corporate boards deliver political connections, prestige, and reputational advantages (Gonzalez-Bailon et al, 2013, 7). Prestige and reputational advantages are translated into investors assigning a higher value to a company. This is studied by Goldman (2009), who finds that the announcement of a politically connected person for a board position is followed by an abnormal positive stock return, even before the individual takes office (Goldman, 2009). Goldman argues that political connections therefore not only create future economic gains, but add immediate value as well (Goldman, 2009).
If a firm is politically connected, the benefits can however be even broader, ranging from preferential treatment by State Owned Enterprises (SOE) to lighter taxation, relaxed oversight, or preferential policy adjustments (Faccio, 2006). This may also be the case in the context of pension funds, where
17
close ties between pension funds and their board members on one side and government and DNB on the other side might lead to relaxed oversight and a loosened regulatory leash.
As argued by Hillmann (2005), who studied the revolving door mechanism between Washington and corporations, resource dependence theory states that firms can improve survival rates and performance by creating linkages with key policymakers since these linkages can buffer the firm from
environmental fluctuations (i.e. policy changes) (Hillmann, 2005). Firms seek to ‘co-opt’ the government by appointing board members with political connections (Hillmann, 2005, 465). Above mentioned political economy reasoning explains the rationale for pension funds (and other corporations) to appoint politically connected board members. How does this incentive for pension funds translate into tangible influence for policymakers in the daily investment decisions of pension funds? I argue that politicians can influence pension fund’s policy through three channels based on the institutional design of pension fund boards. First, the appointment of former political actors within pension fund board. Second, the presence of representatives from social partners within pension fund boards. Third, the interconnectedness of pension fund boards.
First, according to the PW it is not formally ruled out that members of a pension board can be former politicians or have strong political ties. Although the PW stipulates that boards must be composited on the basis of parity between all stakeholders and that each individual board member has to take the interests of all stakeholders into account when taking decision, partiality still is not ruled out (PW Article 100). Due to the above mentioned reasons it can be in the interest of pension funds to appoint politically-connected board members. To give two examples from the two largest pension funds in the Netherlands: The chair of the board of ABP is: Corien Wortmann-Kool, who has been member of European Parliament for Dutch political party CDA from 2004 until 2014. Hans Alders, former Minister of Housing, Spatial Planning and the Environment, was chairman of the board of PFZW from 2001 to 2017.
Naturally, being politically connected does not per definition imply partiality, it does however clarify how close ties between pension fund boards and politics can be. Classic interest group literature would focus on the effect these connections have on public policy output. I rather focus on the other direction of influence: how the revolving door mechanisms and interconnectedness affects pension fund’s investment decisions.
Second, based on the PW the representatives in a pension fund board collectively bear the shared responsibility for the execution of the pension scheme (PW Article 102). Each member must take the interests of all stakeholders into account. Although the professionality of board members has improved recently, partiality of each representative still plays an important role in the decision making process of a pension fund. The representatives are commonly member of the Dutch social partners (employee or employer organisations) (PW Article 100), who’s members also take place in discussions with the
18
government, to negotiate on working conditions and wages (CAO negotiations). Therefore, these regular CAO negotiations and related feuds can be taken into the pension fund boards; thereby politicizing the pension fund’s discussions even more. The presence of representatives of employee and employers organisations like FNV, CNV and VNO-NCW (but there all multiple branch specific organisations too) gives an extra dimension to the agenda-setting power of politics within pension fund boards. Hans de Boer, chairman of VNO-NCW for example regularly openly criticises pension funds (Visser, 2017; Tamminga, 2019), stating that institutional investors should have a more active role as shareholder of Dutch listed companies. He calls this ‘fertilization of your own soil’ (Telegraaf, 2019).
Third, connections can however be broader than just between political and business actors. Intersections between organisations - pension funds in this case - and individuals are commonly referred to as interlocks and can potentially result in exchange of information and diffusion of norms and practices across the firms (Breiger, 1974). Interlocks are studied in the context of the Dutch pension system by Bauer, Bonetti, and Broeders (2018). Bauer, Bonetti and Broeders (2018) have investigated herding behaviour of pension funds. Herding behaviour in the investment context means as much as trading in the same direction (Sias, 2004). They conclude that if one observes the
investments done by Dutch pension funds, clear trends emerge and figures show that they tend to move together in terms of where they invest in. This is due to the interconnectedness of their board members (Bauer, Bonetti. Broeders, 2018). The authors describe the pension sector as an
interconnected sector, with a small number of managers forming a close network of interconnected individuals and institutions, which is amplified by the earlier mentioned trend of consolidation in the pension sector.
Board members can even take place in multiple pension fund boards, thereby transferring their
individual investment beliefs across multiple pension funds (Bauer, Bonetti, Broeders, 2018). The only restriction here is, according to Dutch pension law (Besluit uitvoering Pensioenwet en Wet Verplichte Beroepspensioenregeling, Article 35a), that an individual cannot be in more pension fund boards than the equivalent of a full time working week (VTE ruling). This means for example that an individual can take place in three small pension fund boards (0.2 VTE), and in one large pension fund board (0.4 VTE) at the same time. Also, an individual can take place in two large pension fund boards and one small pension fund board. The law defines a small pension fund as a fund with less than €10 billion of assets under management (Besluit Uitvoering Pensioenwet en Wet Verplichte
Beroepspensioenregeling, Article 35a). However, the consequence is that an individual can take place in up to five pension fund boards at the same time. Social network is defined as nodes of people that are interconnected through various social relationships that range from casual to close bonds (Bauer, Bonetti, Broeders, 2018, 10). Due to the interconnectedness of increasingly consolidated Dutch pension sector, close networks of board members emerge. These networks affect the investment
19
beliefs of pension funds and ultimately result in herding behaviour by pension funds. The consolidated Dutch pension sector, with a small number of influential board members, makes influencing the investment beliefs of the whole sector easier for political actors. One could expect that a political actor only has to convince board members of the large and leading pension funds, who are ahead of the curve, on the importance of orange investments. Also, board members are apparently susceptible to persuasion by other individuals and fear to negatively deviate from their peers. Via the
interconnectedness of the pension sector this investment belief can automatically cascade to other pension funds: a snowball effect. Cohen (2008) adds evidence of the relevance of social network theory in the context of investing. He finds that investors place larger weights to investments in companies to which they are interconnected through an education network (Cohen, 2008). These three characteristics of the composition of pension fund boards open channels to influence pension fund decision making in terms of their regional allocation via three channels. These institutional settings should however be seen in the context of the explained opposite direction of influence (from policymaker to business) following from the revolving door mechanism and interconnectedness between political and business actors. Thereby, this section combines both inductive and deductive reasoning by merging practices and laws regarding the pension sector with agenda-setting and political-corporate influence theories. In combination with the described incentives for pension funds to appoint former politicians, this results in the following hypothesis:
H1: Politically connected pension funds have higher home bias levels relative their non-connected peers.
As discussed in Section II, the Dutch pension system consists of three different types of pension funds: industry-wide pension funds, serving a whole sector such as the civil service or the health care
industry; corporate pension funds, managing pensions of a single company like Shell, ING, or Philips; and pension funds for independent professionals, e.g. medical specialists. I leave this latter type of pension fund disregarded in the remainder of this thesis due to its rareness and modest size.
There are two reasons why one could expect industry-wide pension to take a more social role within its own industry, but also in society as a whole. First, I argue this is due to the broader mandate and broader range of clients of an wide pension funds. The nature of the corporate and industry-wide pension funds therefore differs. Corporate pension funds could have a more autonomous nature relative to large industry-wide pension funds due a more international nature of the underlying firm. A company (predominantly) consisting of expats might be less connected to long term sectoral or societal interests compared to a pension fund serving the teachers and civil servants.
The divergent nature of both types of pension funds is also formalised via rulings of the European Court of Justice. The rulings of the European Court of Justice in 1999 in Case C-67/96 between Albany International BV and Stichting Bedrijfspensioenfonds Textielindustrie have made the social
20
and special role of industry-wide pension funds binding to justify the low levels of competition for these kind of pension funds (Case C-67/96): “Compulsory affiliation of all workers in a given sector to a supplementary pension scheme pursues an essential social function within the pension system applicable in the Netherlands because of the extremely limited amount of the statutory pension
calculated on the basis of the minimum statutory wage. Provided that a supplementary pension scheme has been established by a collective agreement within a framework laid down by law and affiliation to that scheme has been made compulsory by the public authorities, it constitutes an element of the Netherlands system of social protection and the sectoral pension fund responsible for management of it must be regarded as contributing to the management of the public social security service” (Case C-67/96, 73). I.e. industry-wide pension funds’ monopolistic position within a specific sector, translated into compulsory participation, is based on the social task that these pension funds have. This social task has a broad scope, according to the ruling: “It is therefore necessary to consider whether, as contended by the Fund, the Netherlands Government and the Commission, the exclusive right of the sectoral pension fund to manage supplementary pensions in a given sector and the resultant restriction of competition may be justified under Article 90(2) of the Treaty as a measure necessary for the performance of a particular social task of general interest with which that fund has been charged” (Case C-67/96, 98).
This argument is also made in the explanatory memorandum of the Dutch Act that makes participation in industry-wide pension funds obligatory (Wet BPF). “The cabinet believes that industry-wide
solidarity, strengthened by compulsory participation, justifies lack of perfect competition in the second pillar” (PP 27073-3) This forces industry-wide pension funds to have high levels of solidarity and a social role.
Third, politicians might strategically attempt to convince the big fish first as these can be considered the leaders and forerunners of the pension sector (industry-wide pension funds), hoping the big fish will cascade this investment belief on to other pension funds via the social network of board members and the tendency to herd. This potential snowball effect and interconnectedness of the pension sector could incentivize politicians and other external actors to target the large pension funds first. I argue that the earlier described herding behaviour among pension funds amplifies the effect of convincing a few large pension funds instead of the whole sector: the small pension funds will automatically follow the larger forerunners, as the small pension funds trust the highly sophisticated large pension funds. H2: industry-wide pension funds have higher home bias levels than corporate pension funds.
Pre-crisis data show a clear downward trend of home bias in pension fund’s portfolios over the last decades. Rubbaniy, Van Lelyveld, and Verschoor (2014) show that the fraction of orange investments within the national pension pot has decreased from 37% in 1992 to 13% in 2006 (Rubbaniy, Van Lelyveld & Verschoor, 2014). The authors argue that the declining fraction of domestic investments in
21
the portfolios was driven by growing integration of financial markets, increased financial knowledge and more risk awareness (Rubbaniy, Van Lelyveld &Verschoor, 2014, 983). Also two external processes helped facilitating the diversification process during the 1990s. The introduction of a common currency significantly reduced exchange rate risk for foreign investments within Europe (Rubbaniy, Van Lelyveld & Verschoor, 2014, 983). Furthermore, the DNB loosened restrictions on capital mobility for pension funds, which made it possible for ABP to adopt a new foreign investment policy (Rubbaniy, Van Lelyveld & Verschoor, 2014, 983). The implementation of a new foreign policy was imitated by other pension funds. This again, shows the interconnectedness of the pension fund sector, but moreover exemplifies the leading role of large sophisticated pension funds like ABP (Rubbaniy, Van Lelyveld & Verschoor, 2014).
In the aftermath of the financial crisis, demand for large amounts of credit was high due to insufficient credit provision by banks (PP 33746-1, 2013). Former Dutch Minister of Economic Affairs and Climate Henk Kamp repeatedly called upon pension funds to fulfil their societal and national duty: providing credit to the demanding parts of the Dutch economy (PP 33746-1, 2013). Pension funds were seen as the solution to solve borrowing issues during a phase of ‘economic downturn and balance sheet recovery for banks’ (PP 33746-1, 2013). Lenders had become more prudent and cautious; partially forced by stricter regulation. Consequently, profitable projects did not receive funding. This formed a serious impediment for economic recovery (PP 33746-1, 2013).
In 2013, conversations between a delegation of the Cabinet, pension funds and their fiduciaries, and insurers started (Letter of Intent, 2013). These conversations were initiated Kamp and were intended to promote domestic investments by institutional investors (Letter of Intent, 2013). During these
‘Catshuis’ meetings they discussed what role institutional investors could play in the process of strengthening the financing capacity of the Dutch economy (PP 33746-1, 2013). Kamp stated that pension fund capital was especially needed to cope with the following societal challenges: supporting SMB, renewable energy, private housing market, investments in health care real estate, school buildings, energy infrastructure, and the mortgage market (PP 33746-1, 2013). Kamp was expecting a structurally larger role for institutional investors as long-term lenders. In a letter of intent, signed by pension funds ABP and PFZW and fiduciaries APG, PGGM and MN, the pension funds
acknowledged their societal role and stated that they were aware of the fact that their participants would benefit from economic recovery and a powerful and future-proof economy (Letter of Intent, 2013).
Although institutional investors showed their willingness to invest more in the Netherlands, they were witnessing serious impediments at the time (PP 33746-1, 2013). In some cases, the size of investments projects was too small (which makes transaction costs relatively higher); in other cases, the problem was a lack of standardisation and specific knowledge and expertise among pension funds to perform
22
operational due diligence for SME loans (PP 33746-1, 2013). Therefore, an intermediary institution was developed, which attempted to mobilise institutional investors to play a supplementary and complementary role to the banking sector for the Dutch economy: the Nederlandse
Investeringsinstelling (NLII). Twelve large Dutch pension funds, fiduciaries, and insurers participated in the initiative (Letter of Intent, 2013).
Initially, NLII was expected to raise €3 billion of orange investments, but ultimately raised only €360 million (Wolzak, 2018). Less than four years after NLII was established, its activities were already shut down. According to CEO Loek Sibbing, this was due to a lack of suitable and profitable projects (Wolzak, 2018). NLII’s reasons for failure were fivefold. First, profitable projects were finding less difficulties to seek funding since the economic conditions had changed and supply of credit had grown (Wiebes, 2018). Second, institutional investors generally tend to invest directly into projects or funds, which makes an intermediary like NLII unnecessary (Wiebes, 2018). Third, even after the bundling of projects, the size of these investments was too small for the largest institutional investors (Wiebes, 2018). Fourth, NLII did not have any investment capital itself, resulting in the fact that the institution was not taken seriously (Straver, 2019). Fifth, the development of projects proved to require a lot of effort (Wiebes, 2018). Although, the NLII initiative can be seen as a failure, the intermediary institution exemplifies how political forces try to promote investments in the Dutch economy. Invest-NL, a new institution managed by former Minister of Finance Wouter Bos, is currently being developed. Although NLII and Invest-NL are alike, they differ on some crucial aspects according Minister of Finance Eric Wiebes: Invest-NL is not like NLII just an intermediary, but is rather a development- and investing institution, which can participate in projects of societal and economic importance with its own capital (Wiebes, 2018). Invest-NL contains €2.5 billion of government funding, but it is not Invest-NL’s intention to finance sustainable projects on its own. The institution attempts to persuade other investors to participate as well. To persuade banks and pension funds, it can be decisive that Invest-NL can bear relatively a lot of risk itself, according to Bos (Straver, 2019). Kamp and Wiebes predict that €2.5 bn will be supplemented by €10 billion from other investors (Konig, 2017). In contrast, VNO-NCW argues this figure could go up to €50 billion from other investors (VNO-NCW, 2017). Implicitly, this shows the role for pension funds again.
The two institutions exemplify how political forces attempt to steer the pension fund’s investments decisions in a certain direction: more domestic and social investments. This is not done in a formal legislative way, but rather via the indirect and informal channels related to H1. As the economy was in need for credit provision after the financial crisis, due to a higher degree of prudence, I predict that the following hypothesis holds:
23
IV. Research Design
Sections II and III summarise relevant parts of different subfields of research related to home bias, agenda-setting, and government-corporate influence; the sections are furthermore intended to show the omissions in the existing literature in terms of their explanation of the same topics. The data Section V of this thesis consists of three pillars of data. These three pillars use different methodological
approaches based on combining both quantitative and qualitative research techniques: therefore, this study is a mixed method research. By combining quantitative and qualitative research methods, this thesis intends to present a comprehensive picture of external forces influencing investment decisions made by Dutch pension funds. As shown in Section III, there is abundant research attempting to explain the presence of home bias, providing a wide variety of explanations, such as: transaction costs, an information advantage that domestic investors have over foreign investors, and overly optimistic views on domestic stocks by investors. However, none of these studies investigate the political importance of high domestic investments and the potential incentive that politicians might experience to affect investment decisions by pension fund board members.
I have chosen to use a mixed-method research design as the most appropriate manner to investigate the issue. Alternatively, this issue could be studied in multiple ways. First, a purely qualitative research method, solely being based on interviews trying to paint the picture of political-corporate interaction and influence, would omit the foundation around which the interviews build the context. The data section is needed to form a fundament to be able to determine the angle from which home bias is studied. Interviews alone could paint a good and comprehensive picture of government-corporate influence, but would not be able to connect this to empirical data.
Oppositely, a purely quantitative research method, solely relating home bias figures to political connectedness would be insufficiently explanatory in terms of the way both variables affect each other. Especially since political pressure and influence are generally difficult to distinguish and to measure due to the indirectness of interaction between pension fund board members and politicians. Similarly, studying this issue by conducting a survey would overly simplify the topic as context and qualitative reasoning is necessary to sketch ways of interaction, connections, channels for influence, and decision making processes within a pension fund. The interviews improve the internal validity of the study as the context around the findings helps to be sure that the relationship between the
independent (political connectedness) and the dependent variable (home bias) is a genuine causal relationship and is not produced by a third cofounding variable.
Alternatively, one could alter the number of units of analysis. In this thesis I study the fifteen largest pension funds. Preferably, one would include all Dutch pension funds. This would improve the quality of my findings, but is practically unfeasible since sector-wide data are not publicly accessible and resources are limited. Another possibility would have been to focus on one unit of analysis: a case
24
study, in which I would study a single pension fund’s home bias and interaction with political actors. This would enable the researcher to make an in-depth analysis of channels of influence and decision making processes. However, this would also make it practically impossible to draw conclusions for the whole pension sector since the external validity of such a study is low. Pension funds are very
dissimilar due to the unique nature of each pension fund in terms of size, and the culture of the underlying sector or company.
This Section IV extensively explains how three pillars of data collection are combined as mixed method research to answer the research question of this thesis.
Pillar I: Quantitative Analysis of Orange Investments
The first pillar of data contains a quantitative analysis of the amount of orange investments. This section presents home bias levels as a portion of each pension funds’ total investment portfolio. The fifteen largest Dutch (Table 1) pension funds are the unit of analysis of this thesis and are studied over a five year period (2013-2017). The list of the fifteen largest pension funds entails both industry-wide pension funds and corporate pension funds. Previously cited research by Rubbaniy, Van Lelyveld, and Verschoor (2014) used sector-wide portfolio and balance sheet information obtained from DNB to analyse home bias among Dutch pension funds from 1992-2006. Unfortunately, DNB is unwilling to collaborate with this study by sharing similar information of more recent years. They generally do not provide data to master theses if the research questions are not formulated in collaboration with DNB. Furthermore, they argued that the issue was not on their agenda at the moment. Therefore, data on home bias and orange investments are obtained from each pension fund’s specific annual report. Naturally, this increases the chance of measurement errors since I depend on the definition used by each individual pension fund to calculate and present the amount of investments allocated to the Netherlands. In the below section I exemplify how definitions on geographic allocation might differ across pension funds or even over years by the same pension fund. This forms a threat to the measurement validity of my findings as in some cases it is unclear whether the measures really represent the concept they are supposed to be taping (orange investments). A priori, six remarks should be made regarding the data collection of Pillar I. These remarks are relevant for Section VI of this thesis, in which I draw conclusions from my findings:
First, the external validity of this section is fairly limited. The generalizability to other sectors is low as the pension sector is a unique sector, consisting of a unique set of (semi-public) corporations with a clear societal, but also political and financial importance. This makes the sector different from most business sectors, but particularly interesting to study. The financial sector would however be most similar to the pension sector. Especially large too-big-to-fail banks have a societal and political importance similar of the largest pension funds. This could incentivize politicians to protect societal
25
interests in the board rooms of banks, which would create a government-corporate interaction and influence similar to the pension sector.
Second, although both types of pension funds (industry-wide and corporate) are represented in the top fifteen largest pension funds, they are not distributed equally: the list of pension funds entails ten industry-wide pension funds, and five corporate pension funds. The five largest and most powerful pension funds are all industry-wide pension fund. This could create a selection bias, but I argue this does not form an additional threat to the validity nor the reliability as both pension funds are represented and the numbers of pension funds studied is limited anyway.
Third, in this thesis I present data on orange investments of ten of the fifteen largest pension funds. Only six (ABP, PFZW, PMT, BpfBOUW, PME and Philips) out of fifteen pension funds publish country specific allocation to the Netherlands in their annual report. However, as I contacted all remaining nine pension funds, four more pension funds (BPL, Detailhandel, Spoorwegpensioenfonds and Rabobank) were willing to share their data.
Fourth, I expect the modest number of pension funds (small N) to cause issues regarding the scientific significance of my findings. Especially due to the large differences in terms of size between the pension funds. For example, ABP and PFZW, the two largest pension funds, together manage more than €684 billion of pension money. This is not only almost half of the total national pension pot, it is also 63% of the assets investigated in this thesis. These substantial differences in size of the studied pension funds, imply unalike: decision making processes, internal governance structures, and political/societal importance. A priori, large variances in outcomes are therefore expected; in combination with the small N of this research, this will cause home bias outcomes to not be scientifically significant. However, I argue this does not necessarily harm this study a lot since the quantitative findings are not intended to serve as self-contained or autonomous data, but rather form the foundation on which Pillar II (political connectedness of pension fund boards) and III (interviews) are built. The findings in Pillar I primarily form the starting point of discussing decision making processes within pension fund boards as this study is not simply an analysis of home bias trends over time, but rather aims to present a comprehensive picture of external forces affecting decision making by pension funds.
Fifth, the timeframe studied in this thesis is five years: from 2013 to 2017. Since not all 2018 annual reports have been published yet, this year is not taken into account. Due to practical reasons, the preferred fifteen year timeframe, which would include pre- and post-crisis data, was unfeasible: pension funds only started publishing their country specific geographic allocation in their annual reports from 2013 onwards. This fact alone is already an interesting finding and should not be seen as a coincidence. As mentioned in Section III, the issue of orange investments became salient after the financial crisis. Due to a lack of financing capabilities for large segments of the national economy,
26
salience of the issue rose. A high degree of salience and Catshuis talks ultimately resulted in the formation of NLII in 2013. 2013 was not only the starting point of new initiatives to facilitate domestic investments by pension funds, it is also highly likely that publishing data on orange
investments was part of a wider movement by pension funds to attempt to make publicly visible what they were already doing in the Dutch societal interests. This is also stated by Gert Dijkstra, one of the interviewees: “We were already doing quite a lot in the Netherlands, but we made it insufficiently visible for the wider public … Making visible what we were already doing for our clients has been step one a couple of years ago.”
Sixth, since the data are not collected from a central database, differences in definitions to measure orange investments might occur across pension funds. The importance of the definition of orange investments becomes clear if one observes the orange investments of ABP in 2017: ABP invested 15.0% of their portfolio in the Netherlands, while only 4% of their portfolio was actually invested in the real Dutch economy. Contrary to the former figure, the latter figure “only include direct
investments (such as equity and fixed income investments), indirect investments (in infrastructure and real estate, among other things), and investments in participations in external investment funds ” (ABP Annual Report, 2017) and does not include derivatives, swaps, money market products, short
positions, and cash positions (ABP Annual Report, 2017). Definitions are however not specified in the annual reports of most other pension funds. Since ABP generally uses the former figure to report orange investments, I do so too. Furthermore, I must assume that the other pension funds use the same definition. Although 2018 data are not included in this thesis, the following example clarifies even further how important definitions of orange investments are, even for the same pension fund. If one observes ABP’s figures of 2017 and 2018, a strong decline in orange investments (from 15% to 7% in just one year) becomes clear. According to ABP’s explanation, this halving is caused by the
dismantling of the Treasury Center, responsible for ABP’s derivative contracts. Since ABP does these contracts directly from now one, they are now counted as foreign investments.
Pillar II: Analysis of the Background of Pension Fund Board Members
Data Pillar II consists of an analysis of the political connectedness of pension fund boards. This includes a study of the background of each individual pension fund board member over the same timeframe studied as in Pillar I (2013-2017), based on what is publicly available. During this timeframe, the fifteen largest pension funds have had 269 board members. The analysis intends to show whether the presence of former politicians in pension fund boards forms a channel via which political forces can influence pension fund’s decision making. Furthermore, the section aims to study a potential relationship between political connectedness of pension fund boards and the amount of orange investments of these pension funds. I define political connectedness in this context as: an
