Protection of Dutch occupational pension plans;
an analysis of a pension guarantee systemMarien de Haan1 June 2009
Abstract
Solvency requirements protect the occupational pension plan members in the Netherlands against adverse shocks. Nevertheless, this paper wonders whether establishing a safety net as a last resort would be of added value. The research purpose is formulated as follows: to analyse a government imposed pension guarantee system for Dutch private occupational pension plans.
Both theory and practice of pension guarantee systems are considered and existing systems are reviewed. Solvency requirements are compared with a guarantee system. Moreover, relevant characteristics of the Dutch occupational pension sector are discussed and internationally compared. It turns out that the choice between solvency requirements and a guarantee system is essentially a trade-off between risk and return. Establishing a pension guarantee system in the Netherlands seems of limited added value.
Keywords: guarantee system; occupational pension plan; default; pension fund; risk JEL Classification Codes: G23; J32
1
Table of contents
Preface 4
Executive Summary 5
1 Introduction 8
2 An introduction of pension guarantee systems 10
2.1 Preliminary concepts and environment 10
2.1.1 Occupational pension plans 10
2.1.2 Pension plan funding 12
2.1.3 Funding and risk 14
2.1.4 Ways to protect against default risk 15
2.1.5 Pension guarantee system versus idiosyncratic and systematic risk 16
2.2 Rationales 18
2.2.1 Market failure and members’ diversification problems 19
2.2.2 Making government guarantees explicit 20
2.2.3 Systemic events 21
2.3 Basics of existing pension guarantee systems 22
2.4 Adverse side effects 26
2.4.1 Moral hazard 26
2.4.2 Adverse selection 27
2.5 Managing a pension guarantee system 28
2.5.1 Determining an appropriate premium 29
2.5.2 Imposing funding requirements 32
2.5.3 Enforcing asset-liability matching 34
2.5.4 Limiting covered pension benefits 35
2.6 Summary and conclusions 37
3 Solvency requirements versus a pension guarantee system 39
3.1 Assumptions 39
3.2 Protection against default risk; two possibilities 40 3.2.1 Approach 1: government imposed pension guarantee system 40
3.3 A comparison 43
3.3.1 Additional assumptions 44
3.3.2 Pension liabilities and surplus allocation 44
3.3.3 The impact of parameter γ 46
3.3.4 Trade-off between risk and return 47
3.4 Summary and conclusions 51
4 The Dutch occupational pension sector 52
4.1 Dutch private occupational pension plans 52
4.1.1 Coverage 52
4.1.2 Nature 53
4.2 Dutch pension funds 54
4.2.1 Market concentration 55
4.2.2 Legal position 56
4.2.3 Policy instruments 57
4.2.4 Financial position 58
4.3 Indexation policy 60
4.4 Dutch supervision; Financial Assessment Framework 62
4.4.1 Solvency requirements 63
4.4.2 Violation of solvency requirements 65
4.4.3 Long run analysis 66
4.5 Summary and conclusions 67
5 Reflection, discussion and conclusions 68
5.1 Risk versus return 68
5.2 Pros 69 5.3 Cons 71 5.4 Overall conclusions 72 References 75 Referred websites 80 Appendix 81
Appendix A.1 First derivative of put price with respect to strike price 81 Appendix A.2 Expected maturity value of pension assets 82
Preface
‘As our circle of knowledge expands, so does the circumference of darkness surrounding it.’
- Albert Einstein -
Although this quote was made decades ago, I feel it still very much applies to my journey through the world of pensions and pension benefit protection. Answers often led to new questions, solutions to new problems and acquired insights opened up more areas to be explored. This made writing my Master’s Thesis Economics more challenging, but especially more rewarding and instructive than I had ever held possible when I began my graduate internship at De Nederlandsche Bank.
I am grateful to the people of the Supervisory strategy department of De Nederlandsche Bank for ensuring a pleasant and open working atmosphere as well as for their help in many areas. There are some people I would like to mention in particular: my supervisor Dirk Broeders whose expertise, dedication and patience have been essential, Klaas Knot, Leo Kranenburg, Jack Bekooij, Marc Pröpper, Margreet Schuit, Robert Mosch and John Landman.
My gratitude also goes out to my friends and family for their support. A special word of thanks to Dick van Eerde, Leonie de Jongh and Janke Westra for their great hospitality, to Geert de Haan, Corrie de Haan, Tessa Pouwels, Erwin Ensing and Manuel Kampman.
Finally, I would like to thank Hilde van Eerde, her love and support really kept me going.
Groningen, June 2009,
Executive summary
The supervision of the Dutch occupational pension sector aims at protecting pension plan members against adverse shocks and is predominantly based on solvency requirements. There is no safety net if these requirements prove inadequate. The Dutch government has not established a system that rescues troubled pension funds and guarantees promised pension benefits.
This paper wonders whether such a guarantee system would be a valuable addition to the Dutch occupational pension sector. The research purpose is therefore formulated as follows: to analyse a government imposed pension guarantee system for Dutch private occupational pension plans implemented by pension funds. Specifically, the system protects against a default on the benefits promised under pension plans, called default risk.
The analysis is split up in four parts. The first part considers both theory and practice of pension guarantee systems. It is established that the funding level of a pension plan influences the positions and interactions of pension fund, sponsor, members and guarantee system. The funding level also impacts on the nature of default risk; a higher funding level diminishes the sponsor related idiosyncratic component of default risk, but attaches (relatively) more importance to the pension fund related systematic component. Idiosyncratic risk can be diversified quite easily through a guarantee system. Systematic risk is non-diversifiable but it can be hedged in the financial markets. Note that a less than perfectly hedged government imposed pension guarantee system may eventually have to ask the taxpayers to help out.
The fist part also discusses three theoretical rationales for a government imposed pension guarantee system. Such a system can counteract market failure in the labour market, it may stop systemic events originating in the occupational pension sector from growing large or may even prevent their emergence at all, and finally, a guarantee system may strengthen market discipline in the occupational pension sector. However, there are also adverse side effects associated with pension guarantee systems. These result from information asymmetries and can be grouped into cases of moral hazard and adverse selection. In order to control the side effects and to ensure a solvent guarantee system in general proper management is needed. Above all, it is essential to charge a risk based premium; each individual pension fund should compensate for the risk it imposes on the guarantee system. Other management methods are imposing funding requirements, enforcing pension asset-liability matching and limiting benefit coverage.
American system in the largest, measured both in terms of covered members as well as in amounts of assets and liabilities. The British and German systems follow at considerable distance. Except in scale, the systems differ in other characteristics too like in the way premiums are determined.
The second part of the analysis compares the current Dutch approach of protecting occupational pension plan members, solvency requirements, with a government imposed pension guarantee system on their impact on pension benefit security and the pension benefit level. A quantitative model is developed including one pension fund that implements a private occupational DB pension plan. This pension fund is at least fully funded at the start of the time frame and may hold on to its initial surplus by investing it in a diversified asset portfolio, is allowed to spend it by purchasing default insurance offered by the guarantee system, or could choose a combination of both possibilities. The development of the pension fund’s financial position over the time frame is repeatedly simulated for all possible surplus allocation decisions. The results show that exchanging solvency requirements for a pension guarantee system comes down to trading off risk and return. More spending on default insurance from the guarantee system, at the expense of the surplus invested in the diversified asset portfolio, results in higher pension benefit security (less risk) but simultaneously in lower expected pension benefits (less return).
predominantly marked-to-market. But compared to these countries the Netherlands has high required funding ratios and allows only short periods to amortise funding deficits.
The fourth and final part gives the overall conclusions and insights from the analysis. A concise summary follows.
• Choosing between the current Dutch solvency requirements (holding a surplus), and a pension guarantee system (default insurance), means trading off risk and return. Solvency requirements result in higher expected pension benefits (higher return), but also in lower pension benefit security (more risk), whereas the opposite holds for a guarantee system. • If Dutch pension funds want to keep up with their indexation ambitions, they have to
accept (investment) risks and can not afford default insurance via a guarantee system. • A solvent guarantee system can in theory be attained by both hedging it and applying four
management methods: charging a risk based premium, imposing funding requirements, enforcing pension asset-liability matching and limiting covered pension benefits.
• In practice, managing a Dutch pension guarantee system will probably prove difficult. The resulting threat of moral hazard and adverse selection may endanger its feasibility.
• Charging a risk based premium is probably hardest to do, as it requires continuous and extensive monitoring of all insured pension funds.
• A pension guarantee system unable to perfectly hedge its position can not afford to fully guarantee all pension benefits; that is, without shifting part of the insurance costs the government (taxpayers).
• Establishing a pension guarantee system in the Netherlands may be only of little added value. Individual Dutch pension funds can in principle already obtain any level of desired default protection. In fact, the system’s inherent collectiveness and uniformity could well harm general welfare, as it may lead to an inefficient allocation of pension assets.
• The benefits of a guarantee system’s abilities to solve market failure in the labour market and to improve market discipline in an occupational pension sector seem to be modest for the Netherlands. Much of this is already achieved through current Dutch solvency requirements and regulations.
1 Introduction
The Dutch three pillar pension system wins praise from all over the world. Particularly, the occupational pension sector with its well funded pension funds is admired. Even though the Netherlands is only a small actor in the global economy; in terms of accumulated pension fund assets against GDP it is a world leader.
This does not imply though that Dutch pension funds are immune for problems and crises and may rest on their laurels; the pension funds are continuously affected by their rapidly changing (economical) environment. In the years 2000 until 2002, after a period of prosperity, declining stock markets combined with low interest rates levels harmed the financial situation of pension funds, causing funding ratios to deteriorate quickly. But adjustments to occupational pension plans, measures taken by pension funds and more favourable economic conditions resulted in a period of steady recovery. But this period was ended quite abruptly by the latest financial crisis, commonly referred to as ‘credit crisis’. Once again, the financial position of Dutch pension funds is under pressure.
The supervisory regime, predominantly based on solvency requirements, protects Dutch pension plan members against adverse shocks. Pension funds are required to hold a surplus of pension assets that serves as a cushion. However, if these surpluses prove to be inadequate, there is no safety net like in certain other economic sectors; the Deposit Guarantee Scheme in the banking sector for instance. The Dutch government has not established a system that rescues troubled pension funds and guarantees the benefits promised under occupational pension plans.
This paper wonders whether such a guarantee system would be of added value to the Dutch occupational pension sector; similar systems are actually operational in a number of major industrialized countries, including the United States, the United Kingdom, Germany and Japan. The research purpose is therefore is formulated as follows:
To analyse a government imposed pension guarantee system for Dutch private occupational pension plans.
Consider two comments about this research purpose. First, the guarantee system protects against a default on pension benefits promised under the pension plans. Second, the system covers Dutch private occupational defined benefit pension plans that are implemented by pension funds.2,3
2
A minority of Dutch private occupational defined benefits pension plans is actually implemented by insurers (see also chapter 4). However, this paper focuses on a guarantee system that protects benefits promised under Dutch private occupational pension plans which are implemented by pension funds, although acquired insights may possibly also apply in a broader context.
3
Each of the next four chapters covers a part of the analysis. Chapter 2 discusses the concepts and characteristics of pension guarantee systems on a theoretical level and reviews existing systems. Chapter 3 extends the analysis by developing a quantitative model to compare the current Dutch approach of protecting occupational pension benefits, solvency requirements, with a pension guarantee system on their impact on pension benefit security and the pension benefit level. So it is asked how much security both approaches offer and at what costs. Furthermore, with a view to the research purpose, chapter 4 discusses the Dutch occupational pension sector: pension plans, pension funds, supervisory regime and associated solvency requirements. International comparisons are made on several points; most include the countries with a pension guarantee system. Finally, chapter 5 concludes this paper and gives a number of insights and conclusions about the analysis of a government imposed pension guarantee system for Dutch private occupational defined benefit pension plans that are implemented by pension funds.
2 An introduction of pension guarantee systems
This chapter considers pension guarantee systems both on a theoretical and a practical level. On the one hand, theoretical concepts and characterises of these systems are discussed based on the literature. On the other hand, existing systems are reviewed and compared on a number of points.
The chapter is divided as follows. Section 2.1 discusses occupational pension plans and introduces the government imposed pension guarantee system as one of the policy options to protect against losses of pension benefits. Section 2.2 then discusses the theoretical rationales for these systems and with that touches on the question of why it might be desirable to protect pension benefits in the first place. Section 2.3 examines the existing pension guarantee systems of seven countries. Furthermore, section 2.4 deals with potential adverse side effects of guarantee systems. Section 2.5 discusses how to mitigate these side effects through proper management. Finally, section 2.6 summarizes and concludes.
2.1 Preliminary concepts and environment
This section discusses the main characteristics of occupational pension plans. It also introduces the government imposed pension guarantee system as a policy option to protect against losses of pension benefits. Note that this section will merely give a general overview of some major concepts; it does not seek to do justice to all possible real world differences, nuances and exceptions. Later sections analyse matters in more detail.
First, subsection 2.1.1 discusses occupational pension plans as such, while subsection 2.1.2 is specifically about pension plan funding. Subsection 2.1.3 considers how the level of funding changes the nature of default risk. After that, Subsection 2.1.4 addresses policy options for default risk protection. Finally, subsection 2.1.5 discusses the extent to which pension guarantee systems are able to cope with default risk of different natures.
2.1.1 Occupational pension plans
An occupational pension plan (scheme or arrangement) is a legally binding contract between an employer (sponsor) and its (former) employees (members) aimed at providing a (supplementary) retirement income to retired employees (pensioners).4,5 Such a pension plan may be funded. That is, the sponsor and sometimes the active members accumulate assets
4
There are many possible variations and extensions on this basic representation in practice. For instance, a single pension plan may be linked to multiple employers or even a whole industry; employers may be allowed not to participate in any pension plan at all. See OECD (2005) for an extensive classification of pension plans; this classification is also consulted in behalf of this paper.
5
(contribute) intended to cover the pension benefits promised under the plan.6 The accumulated pension assets are most often administered by a pension fund. Sometimes, this task is performed by an insurance company or directly by the sponsor.
This paper focuses on the protection of a specific type of pension plan, namely the private (i.e. not implemented by the general government) occupational defined benefit (DB) pension plan.7 The literature generally distinguishes between two archetypes of pension plans: defined benefit and defined contribution (DC) (see for instance Pesando, 2000; Burtless, 2003 and Whitehouse, 2007). Benefits promised under DB plans are typically fixed. Their value is usually determined by taking variables like years of service and the wage level into account. It is not however specified how benefits are funded or how high contributions should be (Bodie and Merton, 1993). In DC pension plans it is the other way round; while contributions are specified, benefit levels typically are not. A pension fund that administers the accumulated pension assets for the purpose of a DC plan is in essence an aggregation of individual employees’ investment accounts (Bodie and Merton, 1993). Employees typically have a say in matters like investment strategy or contribution level.
DB and DC pension plans have some distinct differences; two stand out. First, the difference in investment risk allocation. Investment risk stems from uncertainty in pension assets returns. In case of a pure DB plan investment risk is carried by those ultimately responsible for paying the predetermined and fixed pension benefits (e.g. pension fund or sponsor). In a pure DC plan though, all investment risk is borne by the members; (part of) their retirement wealth simply evaporates if returns are lower than expected. Burtless (2003) extends the notion of investment risk regarding DC plans to essentially all risk associated with financial market fluctuations. Like the risk of recently retired members facing prices above expected when setting out to buy a lifetime annuity. Note that if these annuities are in nominal terms, members are exposed to inflation risk too (Burtless, 2003).
A second marked difference between DB and DC plans lies in the portability between jobs. Under a DC plan, the value of an individual’s investment account is known, making it relatively easy to remit it to another DC plan. In case of DB plans however, transfers are more complex and members could easily incur portability losses when changing jobs (see e.g. Blake, 2000; DNB, 2006b). This is caused by the fact that the benefits in these plans are typically partly conditional and inaccuracies in determining the fair value of this conditional component for a transfer may result in such losses.
6
In practice, both the sponsor and the active members may contribute. However, this paper assumes that contributions are exclusively made by the sponsor, unless stated otherwise.
7
This paper focuses exclusively on the protection of DB pension plans, as these plans offer predetermined and fixed benefits. Protecting the inherently contingent benefits of DC plans would obviously be unnecessary.8
In view of the research purpose, discussions throughout this paper presume a private occupational DB pension plan implemented by a pension fund, unless stated otherwise. Nevertheless, findings may also apply when for instance an insurer or sponsor implements the plan. Note that a pension fund is the pool of accumulated pension assets that makes up an independent legal entity (OECD, 2005). Usually, pension funds are special purpose vehicles with legal personality (e.g. foundations or corporate entities). Sometimes they are only responsible for some specific aspects of pension plan implementation. In certain countries for instance, pension funds may merely be responsible for holding the pension assets, not for investing them. Furthermore, pension funds in some cases lack legal personality and consist merely of legally separate pools of pension assets managed by financial companies (OECD, 2005).
2.1.2 Pension plan funding
Private occupational DB pension plans differ in their characteristics. A particularly important characteristic is the funding level, as it impacts on the positions of members, sponsor and pension fund.
Funding a pension plans boils down to accumulating assets intended for future payments of promised benefits (pension liabilities). A pension plan is fully funded if the value of pension assets equals that of pension liabilities, whereas the plan is unfunded if no assets are accumulated and benefits are directly paid by the sponsor and/or the active members.
Table 2.1 highlights some consequences of differences in pension plan funding levels. The first row highlights that pension plan funding actually changes the party where members are dependent on for their occupational retirement income. Members of an unfunded pension plan are directly dependent on the sponsor (their employer) for pension benefits; in a book reserve plan for instance, the sponsor merely makes a pension provision at the liability side of its balance sheet instead of making actual contributions. Members of a funded plan are fully dependent on the pension fund, whereas in case of a partially funded plan members are dependent both on the pension fund and the sponsor.
The above implies that when pension funds receive additional pension assets to administer, they automatically receive a higher proportion of the obligation to pay the promised benefits. So while employers initially accept the obligation to pay pension benefits
8
with the creation of a pension plan, they transfer this obligation to a pension fund by funding it. When a pension fund is fully funded, it has in principle received sufficient contributions, and therefore the obligation, to pay all benefits promised under the pension plan. Bodie and Merton (1993) explain the relevance of pension plan funding by drawing a parallel with an equipment trust. The authors state that assets accumulated in pension funds serve as collateral for the promised benefits, just like assets of equipment trusts serve as collateral for the related debt obligations. According to Bodie and Merton (1993), accumulated pension assets are encumbered because they serve as collateral and the sponsor is not allowed to use them for any other purposes. The same authors add that the value of the pension liability itself does not change with the funding level, but that additional funding does make the promised benefits more ‘secure’. The next subsection will go further into the relationship between funding and risk.
But first, consider two additional rows of Table 2.1; starting with the third row that is about accumulated pension assets from the sponsor’s perception. In case of a funded pension plan these assets are regarded as past contributions, since the sponsor has funded the plan through contributions of assets (cash). However, the sponsor of a (partially) unfunded plan has ‘retained’ (part of) the assets and may therefore use them as a source of funding for its own investment purposes. These assets are not ‘encumbered’, to speak with Bodie and Merton (1993). The bottom row of Table 2.1 illustrates how pension funds perceive accumulated assets. Pension funds implementing fully funded plans invest the assets based on their preferences in a certain portfolio. In case of (partially) unfunded pension plans though, the sponsors have, as it were, pre-decided to invest (part of) the assets in their own long term
Table 2.1 Consequences of pension plan funding (a stylised overview)
unfunded
(book reserve) partially funded
fully funded (or overfunded) responsible for paying
pension benefits sponsor
-sponsor (unfunded part);
-pension fund (funded part) pension fund
contributions sponsor perception
no cash outflow (reservation on
sponsor’s balance sheet)
cash outflow cash outflow
accumulated assets
sponsor perception source of funding
-source of funding (unfunded part);
-past contributions (funded part) past contributions
pension liabilities
sponsor perception long run liability
-liability (unfunded part);
-settled debt (funded part) settled debt
accumulated assets pension fund perception
long run fixed income investment in sponsora
-long run fixed income investment in sponsor (unfunded part);a -investment portfolio (funded part)
investment portfolio
Note:
It is assumed that all contributions are made by the sponsor. Source:
fixed income debt. This implies that a book reserve system can be seen as a form of funding combined with extreme asset allocation, since book reserve plans are effectively similar to funded plans investing exclusively sponsor bonds (e.g. Gerke et al., 2006).
2.1.3 Funding and risk
As aforesaid, funding impacts directly on the positions and interactions of members, sponsor and pension fund. Here, the discussion continuous by considering more specifically how funding affects the nature of risk.
The risk in question stems from the possibility that the party obliged to pay the pension benefits (pension liabilities) proves to be financially unable to do this. It is the risk of a default on the benefits promised under the pension plan; called default risk in this paper.
Default risk may actually be considered as a catch-all term. It represents the worst case scenario for pension plan members: loosing occupational pension benefits. Default risk consists of an idiosyncratic and a systematic component. Idiosyncratic (or unique) risk is defined by Brealey et al. (2001) as risk that arises from threats peculiar to a specific firm and maybe to its direct competitors. Idiosyncratic risk can in principle be diversified by pooling firms from different industries. Systematic (or market) risk is defined by Brealey et al. (2001) as risk that stems from economy wide perils that endanger all firms. Systematic risk can not be eliminated by diversification precisely because it affects all firms, but it can, in principle, be hedged in the financial markets.
Note that sponsors and pension funds, being the two parties that may default, are both exposed to a different combination of idiosyncratic and systematic risk. Pension funds deal predominantly with systematic risk and to a limited extent with idiosyncratic risk; while sponsors are exposed to more idiosyncratic risk as well as to a substantial degree of systematic risk.
An important source of systematic risk for pension funds is investment risk, like a stock market crash. Idiosyncratic risk on the other hand is much less important as pension funds are hardly exposed to (operational) factors associated with idiosyncratic risk, like product demand or input prices. One of the few real sources of idiosyncratic risk is internal pension funds problems such as fraud or mismanagement. In contrast to this, a sponsoring firm that operates a business has to deal with much more idiosyncratic risk; like fluctuations (deteriorations) consumer’s demand because of changing preferences. Moreover, sponsors also have to deal with systematic risk; interest rate fluctuations for instance affect almost all firms.
risk of diverse natures. To appreciate this, reconsider the top row of Table 2.1. As a pension plan’s funding level increases, members become less dependent on the sponsor for their pension benefits but more dependent on the pension fund instead. Or, in terms of default risk; a higher funding level diminishes the sponsor related idiosyncratic risk component but enlarges (relatively) the pension fund related systematic risk component (see Figure 2.1 for an illustration). So protecting members of differently funded pension plans requires protecting against default risk of different natures. Therefore, the suitability of a particular way of protection may depend on a pension plan’s level funding.
2.1.4 Ways to protect against default risk
This subsection considers ways of default risk protection. The question is: how to protect members from seeing (much of) their retirement wealth going up in smoke because of a default on occupational pension promises?
Most of the options have a public policy character, but some may actually be adopted directly by an individual sponsor or pension fund. For instance, Bodie and Merton (1993) put forward that sponsors could acquire a guarantee for their pension liabilities from a private-sector third-party. A pension annuity contract from an insurance company is effectively a guarantee (Bodie and Merton, 1993).
Next consider the public policy options. Pesando (2000) mentions three that should at least limit default risk to some extent. First, impose investment restrictions on pension funds. This typically comes down to requiring a match between pension assets and liabilities. Second, demanding disclosure of relevant pension fund information to members, as information helps members to put pressure on the sponsor to keep up contributions. Third, enforce funding requirements on pension funds to assure promised benefits are actually covered by pension assets. Another policy option is to give pension claims high priority in
idiosyncratic risk systematic risk
funding level→
note:
This Figure is strictly qualitative.
situations of sponsor bankruptcy (e.g. Bodie and Merton, 1993; Pesando, 1996; Stewart, 2007). The assets of the sponsor are then firstly used to settle pension liabilities.
The final policy option is that of a government imposed pension guarantee system (e.g. Bodie and Merton, 1993; Pesando, 2000). In plain words, such a system protects (or insures) the promised pension benefits. It charges a premium to sponsors or pension funds and pays out pension benefits to members whose promises are violated. Hence, the system insures against default risk. As discussed, a pension plan’s level funding changes both the nature of default risk and the relationship between sponsor, pension fund and members. Next to this, it also determines a guarantee system’s position within the occupational pension sector. If there are only book reserve plans for instance, the guarantee system effectively insures the pension liabilities of sponsors, whereas in a situation of full funding, it insures the pension fund liabilities. Note that the guarantee system will probably charge its premium to sponsors in the first case and to pension funds in the second. Figure 2.2 illustrates how a guarantee system would fit in occupational pension sectors with different levels of funding.
Consider one final note here. In view of the research purpose, discussions throughout this paper presume an occupational pension sector consisting of fully funded private DB pension plans, except otherwise specified, as this corresponds best to the Dutch situation (see chapter 4). Therefore, the guarantee system is assumed to interact directly with pension funds (see the bottom diagram of Figure 2.2).
2.1.5 Pension guarantee system versus idiosyncratic and systematic risk
As stated above, a pension guarantee system is established to protect against default risk. This section discusses to the extent to which such a system is able to cope specifically with the idiosyncratic and systematic risk components of default risk.
Idiosyncratic risk can be diversified away by pooling dissimilar firms. So pooling lowers the aggregate costs of protecting against this type of risk, as the probability of being hit by the same adverse idiosyncratic shock is uncorrelated between dissimilar firms. Note that a pension guarantee system facilitates sharing of idiosyncratic risk.
swap equity for fixed income securities to match the payouts of pension benefits (Bodie and Merton, 1993). Note that a pension guarantee system is effectively an intermediary between pension funds and the financial markets; it charges a premium to pension funds and hedges its own position in the financial markets.
Although in theory it is well possible to acquire default insurance in the financial markets; there may be problems when many pension funds actually decide to do this, especially when their liabilities are worth hundreds of billions of euros. Financial markets may then show reluctance or even inability to fully meet the demand for guarantees (see Van Ewijk and Van de Ven, 2002). So even the prices of guarantees would take all risk into account, financial markets may simply be unable to provide sufficient liquidity. Pesando
sponsors unfunded pension plans/ book reserve members pension guarantee system pension claim/pension benefits
sponsors partially funded
pension funds members
pension guarantee system
unfunded pension plans
pension claim (unfundedpart)
partially funded pension plans
sponsors fully funded pension funds members
pension guarantee system
pension claim (funded part)/pension benefits contributions pension claim/ pension benefits pension benefits in case of sponsor or pension fund default premiums
premiums
fully funded pension plans
pension benefits in case of pension fund default contributions premiums pension benefits in case of sponsor default
(1996) states that systematic risk may render private markets incapable to offer default risk insurance, even if premiums could be asked in concordance with true risk. Merton and Bodie (1992) argue in a more general sense that when risk can not be diversified or hedged in the capital markets, the government may be the appropriate party to provide guarantees.
Nevertheless, Ippolito (2004) sees an important role for the financial markets in providing default risk guarantees.9 Especially the market’s disciplining effect on sponsor and pension fund behaviour is regarded as valuable. Note that both the ability and willingness of financial markets to provide the required hedges may also depend on matters like the degree of financial market sophistication or the distribution of liabilities among pension funds. Poorly developed markets or a skewed liabilities distribution are likely to complicate things considerably.
This being said, it is important to emphasize that default risk, including its non-diversifiable systematic component, ultimately always will be borne by someone. So if there is no pension guarantee system members bear the risk. And a guarantee system bears all risk by itself if it is unable to hedge its position.
However, if such a system relies on implicit or explicit government support, it actually shifts the burden of default risk to taxpayers in general. Stewart (2007) appropriately points at the difficulty of convincing taxpayers to accept this. It will be particularly hard when many taxpayers fall outside of occupational pension plans, as they have to be persuaded to bail out those who have the privilege of being in an occupational pension plan (Stewart, 2007).
To sum up, default risk incorporates both an idiosyncratic and a systematic component. Idiosyncratic risk can be diversified away relatively easy, for instance through a pension guarantee system. Protecting against systematic risk is less clear-cut. It is ambiguous whether financial markets are able to provide the required hedges. However, it should be clear that someone ultimately bears the default risk, including its systematic component.
2.2 Rationales
This section discusses three theoretical rationales for a government imposed pension guarantee system. The previous section introduced the guarantee system as an approach to protect against default risk; but what are the rationales to actually choose such a system? In fact, why is it necessary to protect private occupational pension benefits at all? Before continuing, note that the rationales do not have to apply exclusively to guarantee systems; some may hold for other ways of pension benefit protection too. Also, the validity of a rationale may be affected by the funding level of pension plans.
9
The remainder of this section is organized as follows. Subsection 2.2.1 considers market failure and diversification problems. Subsection 2.2.2 then addresses implicit government guarantees, while subsection 2.2.3 focuses on systemic events in an occupational pension sector.
2.2.1 Market failure and members’ diversification problems10
A government imposed pension guarantee system counteracts market failure in the labour market; market failure that results primarily from information asymmetries and limited diversification possibilities for members.
According to Sharpe (1976), labour income is the sum of the present values of wages and promised pension benefits. Pension benefits can then be regarded as deferred wages: employees receive fewer cash wages now, but are promised a retirement income in return. When a violation of this promise is more likely, perfectly informed employees will demand proportionally higher wages; given the value of the promise, that is. Hence, default risk is already compensated through the current wage level (Sharpe, 1976; Pesando, 1982). This reasoning implies that default protection and pension guarantee systems are redundant. Or in other words, if rational workers discount the risk associated with promised pension benefits by demanding higher, compensating wages, the need for termination insurance is easily overstated (Pesando, 1982).
This is however based on the assumption that employees have and understand all information needed to assess default risk associated with promises of pension benefits. But in actual reality, this assumption does not seem to hold. Indeed, Mitchell (1988) studies employees’ knowledge of occupational pension plans and emphasises that this is far from complete. A survey amongst Dutch employees appears to confirm this finding (Het Financieele Dagblad, 2007).
In fact, it remains doubtful whether even employees with complete information would accept default risk, because of the diversification problems they may encounter. Employees already have a substantial non-diversified position, as their human capital value is tightly linked to the financial success of the firms they work for (Bodie and Merton, 1993). And most of them lack the assets and knowledge required to hedge such a position, let alone to hedge default risk associated with their pension promises (Bodie and Merton, 1993).
So altogether, market failure in the labour market may be a reason to establish a pension guarantee system. The system offers an extra layer of protection against defaults on pension promises by counterbalancing market failure (Stewart, 2007).
10
Note that informational problems associated with default risk assessment may be experienced by any member, regardless the funding level of his of her particular pension plan. Funding levels merely influence the nature of the problems. With a rising funding level for example, problems become increasingly sponsor-related instead of pension fund-related, as the obligation to pay pension benefits is effectively transferred from sponsor to pension fund (see Table 2.1). Hence, members of a (more than) fully funded plan will be interested in information on pension fund default and not so much sponsor default.
The funding level’s influence on diversification problems is different in the sense that only members of less than fully funded pension plans are likely to encounter such problems. Again, this has everything to do with how the responsibility of paying pension benefits is divided. Because members of a (more than) fully funded plan receive their current wage from their employer (sponsor) and their future retirement income from a pension fund, there is no great need to diversify. In contrast to this, members of a less than fully funded plan are dependent on the sponsor for their current wages as well as for (part of) their retirement income. This position does call for diversification which as aforementioned is out of reach for most members. It follows that diversification problems are especially large in case of completely unfunded, or book reserve, plans (Stewart, 2007).
Next, consider two slightly different perspectives on labour related market failure discussed so far. First, Cooper and Ross (1999) study market failure arising from market fragility: a concept closely related to informational problems. Market fragility arises when sellers (pension funds) are unable to commit themselves to a future action (pay pension benefits) and anticipating buyers (members) are therefore reluctant to do business with them. As a consequence, transactions between buyers and sellers become difficult or may even be disrupted. Cooper and Ross (1999) emphasise the positive role that guarantee funds in general can play to counteract these problems. Second, Garcia and Prast (2004) discuss market failure that results from difficulties of members in exercising market discipline on pension funds. Problems are particularly large if participation in an occupational pension plan is compulsory. Members can then typically only leave the plan by changing jobs or even industries. For most employees however, the costs of doing so are simply too high.
2.2.2 Making government guarantees explicit
An explicit guarantee in the form of a government imposed pension guarantee system may strengthen market discipline in the occupational pension sector.
managed explicit guarantees however at least give the possibility to counteract such a loss (Garcia and Prast, 2004).
Garcia and Prast (2004) focus on depositor and investor protection in the Netherlands. These authors state that in a highly concentrated banking sector individual banks may consider themselves ‘too-big-to-fail’ and could believe that their deposit liabilities are insured by the authorities.
Something similar may also be going on in an occupational pension sector. Large pension funds could take on excessive risks if they assume to be guaranteed by the government. They may for instance focus solely at high investment returns, while the (excessively) large risks are played down or ignored altogether. Risk management at such pension funds might not be optimal. The presence of a pension guarantee system could improve this; for one, because a price tag is then attached to risk taking (section 2.5 elaborates on this).
2.2.3 Systemic events
A government imposed pension guarantee system may stop systemic events originating in the occupational pension sector from growing large or may even prevent their emergence at all.
A systemic event in the narrow sense is defined by De Bandt and Hartmann (2000) as an event where a limited (idiosyncratic) adverse development (or shock) for one or more financial institutions or markets, leads, in a consecutive way, to substantial adverse effects for one or more other financial institutions or markets. Essential here is the ‘domino effect’ from one financial market or institution to another. If the definition is extended to include events where severe and widespread (systematic) adverse developments (or shocks) have simultaneous detrimental effects on many institutions or markets, it becomes a definition of systemic events in the broad sense. A systemic event is called strong if the affected financial institutions or markets in the second round or later fail because of the initial shock, and would not have done so otherwise (De Bandt and Hartmann, 2000). Note furthermore that systemic events can have a horizontal and/or a vertical dimension (De Bandt and Hartmann, 2000; Lemmen, 2003). A horizontal systemic event originates in the pension sector and subsequently leads to damage there. Take for example a situation in which persistent problems of a few pension funds lead to a loss of confidence in the occupational pension sector as a whole. Eventually, this could induce the private sector to change its economic behaviour (see also Lemmen, 2003). A vertical systemic event starts in the occupational pension sector but its subsequent adverse effects spill over to other sectors of the economy.
pension sector may easily damage confidence (horizontal event), as there is only limited collateral (pension assets) available to keep it up. The inherent entanglement of low funded pension funds, via their sponsors, with various economic sectors also causes shocks to spill over more easily (vertical event).
To sum op, default protection offered by a pension guarantee system can counteract shocks to (part of) the occupational pension sector. Therefore, it may well prevent adverse shocks from developing into systemic events.
2.3 Basics of existing pension guarantee systems
So far, the discussion has been predominantly about the theory of pension guarantee systems. But these systems do by no means exist only in theory. This section gives a concise review of the systems operating in seven major industrialised countries: the United States, the United Kingdom, Germany, Japan, Switzerland, Canada and Sweden.11 These systems cover predefined benefits promised under occupational DB pension plans.12
First of all, consider the division between occupational DB and DC pension plans in the seven countries with a pension guarantee system (see Table 2.2). Most pension assets are accumulated in DB plans. Even in the United States where DC plans are relatively important DB plans still hold nearly two thirds of the assets. Furthermore, just like the pension assets, the majority of active members are in DB plans. The United States is an exception with DC plans covering as much as 71% of the active members.
It does not show directly from Table 2.2, but in many countries, including the United States, the United Kingdom and Canada, there is (or has been) a profound shift from
11
See also Stewart (2007) for an extensive review of these seven pension guarantee systems. 12
The Swiss guarantee system actually covers occupational DC pension plans. However, these DC plans do provide predefined, minimum pension benefits.
Table 2.2 Occupational pension plans; DB versus DC
DB-DC
(pension assets)1
DB or DC
(percentages of active members)
United States 65% - 35%a 71% DCb,2
United Kingdom 78% - 22%a > 80% DBc
Germany 100% - 0%a > 90% DBc
Japan 99% - 1%a > 90% DBc
Switzerland unknown unknown
Canada 93% - 7%a > 80% DBc Sweden 95% - 5%a unknown Notes: 1 Figures: 2004. 2 Figure: 2003. Sources:
occupational DB plans to DC plans (e.g. OECD, 2007a; Lindeman and Yermo, 2007). In the United States for example, almost all employers opt for the DC type when a new pension plan is established (e.g. Bodie and Merton, 1993; CBO, 2005).13 Also, most DB plans in the United Kingdom have been closed to new members (e.g. OECD, 2007a; Lindeman and Yermo, 2007).
Next go on to the actual pension guarantee systems (see Table 2.3). Consider first how covered pension plans are funded. Five guarantee systems cover funded pension plans, while
two systems: Germany’s Pensions-Sicherungs-Verein (PSVaG) and Sweden’s
Pensionsgaranti cover also unfunded, book reserve plans. Not all occupational DB pension plans in a country are automatically covered. The American Pension Benefit Guarantee Corporation (PBGC) for example excludes plans established by charities; the PSVaG does the same with plans administered by so-called Pensionskassen (see Stewart, 2007). Furthermore, the Japanese Pension Benefit Guarantee Programme (PBG) only covers pension plans of the Employer Pension Fund (EPFs) type. Other types, like the Tax-Qualified Pension Plan (TQPP), are excluded (see Blome et al., 2007). Pensionsgaranti covers the benefits promised under ITP plans, the plan for private sector salaried employees, but only those ITP plans not administered by Alecta: a mutual life insurance corporation (see Pensionsgaranti and PRI Pensionstjänst, 2004).
The guarantee systems differ widely in numbers of covered members. With 44 million members, the PBGC covers by far the largest group. Second in line is the British Pension Protection Fund (PPF) with 12.5 million members, followed by the PSVaG that covers 10.2 million members. Pensionsgaranti brings up the rear by covering 200,000 active members.
Interference of all guarantee systems, except that of the Swiss Sicherheitsfonds BVG, is triggered by sponsor insolvency. So when a sponsor goes bankrupt, the other six systems step in and assume responsibility for the pension plan, provided that there are insufficient pension assets to cover the promised benefits or, alternatively, the compensation offered by the guarantee system.14 For the PBGC and the PBG, sponsor insolvency is not the only intervention trigger; both systems may takeover pension plans that have become a disproportionately heavy burden for the sponsor. Sicherheitsfonds BVG is the only system that is triggered by pension fund insolvency instead of sponsor insolvency.
Interfering typically means that the guarantee system either buys annuities to cover the pension liabilities, or takes over all pension assets and liabilities. Interference of the Canadian Pension Benefit Guarantee Fund (PBGF) is somewhere in between; the system makes a
13
For this reason, Ippolito (2004) believes that the American pension guarantee system is really in an endgame. 14
payment to the relevant pension fund, which it can then use to cover liabilities, for instance by buying annuities.
Table 2.3 Basics of existing pension guarantee systems
covered occupational DB pension plans; funded or unfunded number of covered members
typical trigger for intervention
intervention procedure
key financial figures
(in millions) United States Pension Benefit Guarantee Corporation (PBGC) founded: 1974 funded 44 million (2007)a -sponsor insolvency, -or plan is too high a burden on sponsor; -plan underfundedb pension assets and liabilities are taken overb assets $68,438; liabilities $82,504; net position -$14,066 (2007)a United Kingdom Pension Protection Fund (PPF) founded: 2004 funded 12.5 million (2007)c -sponsor insolvency; -accumulated assets insufficient to match PPF compensationd pension assets and liabilities are taken over d assets £4,409; liabilities £5,018; net position -£609 (2007)c Germany Pensions- Sicherungs- Verein (PSVaG) founded: 1974 -unfunded (book reserve); -funded
(support fund; direct insurance; pension fund)e
10.2 million (2007)f
-sponsor insolvency; -plan or fund can not fulfil obligationsf
annuities are boughtf
balance sheet total €2,101 (2007)f Japan Pension Benefit Guarantee Programme (PBG) founded: 1989
funded (EPF plan)e 5.3 million (2006)e
-bankruptcy of sponsor, -or worsening financial position sponsor, -or acknowledged difficulties in continuing plan; -accumulated assets insufficient to match PBG compensatione pension assets and liabilities are taken overg assets under management ¥29,792 (2005)e Switzerland Sicherheids- fonds BVG founded: 1986
funded 3.2 milliong -pension fund insolvencyg annuities are boughtg assets CHF 593; liabilities CHF 217; net position CHF 376 (2007)h Canada (province Ontario) Pension Benefit Guarantee Fund (PBGF) founded: 1980 funded > 1 million (2005)g -sponsor insolvency; -plan underfundedi payment is made to pension fund to cover liabilities g assets CAD 292; liabilities CAD 513; net position minus CAD 237 (2005)i Sweden Pensionsgaranti founded: 1989 -unfunded ( ITP plan through book reserve) -funded
( ITP plan through pension fund)j
200,000 (active members)g
-sponsor insolvencyj annuities are boughtj - SEK15,558 consolidation capital, which is 15.5% of risk adjusted insurance exposure; -target range consolidation capital SEK13,000 - 15,000 (2007) k Sources:
a PBGC (2007); b PBGC website, http://www.pbgc.gov (accessed September, 2008); c PPF (2007a); d PPF website,
The final column of Table 2.3 highlights some key financial figures. The financial situation of the PBGC is particularly notable. In 2007, the present value of the system’s liabilities exceeded that of its assets by $14 billion. Although this net position was a $5 billion improvement over 2006 and a $9 billion improvement over 2005, it is still negative and receives quite some attention in the literature.15 Bodie has done extensive work in this field. In Bodie (1996) for example, the net position of the PBGC is assessed and parallels are drawn with similarities in the problematic history of the Federal Savings and Loan Insurance Corporation (FSLIC).16,17 Bodie (1996) emphasises that equity investments are unsuitable to cover the benefits of DB pension plans, particularly in the long run.18 Moreover, Kosterlitz (2005) discusses the financial situation of the PBGC along lines of Bodie’s argumentation. Lastly, Bodie and Merton (1993) use a functional approach to learn how the PBGC could be improved. These authors focus on the functions a guarantee system should have and determine subsequently which institutional structure facilitates these functions best.
The financial situation of the PPF requires some attention too. In the few years of its existence the system accumulated a deficit of over £600 million. This is partly due to the premium income of the year 2006/07 that turned out some £300 million lower than expected. The PPF has the intention to compensate for this under-collection and to eliminate its deficit through the premium income of future years (PPF, 2007a).
The remainder of this section discusses a few noticeable characteristics of some of the other systems. To begin with, note that the PBGF operates only in the province of Ontario; one of eleven Canadian jurisdictions and the nation’s industrial centre. The PBGF was founded after a number of plants were (in danger of) shutting down (Pesando, 1982). This might foster the idea that an implicit, yet important reason behind the system’s establishment was pension costs reduction for struggling industries (Pesando, 1996).19 Bodie (1996) and Pesando (1996) state that some analysts have the same thoughts about the PBGC; it was installed to support distressed industries. Though, Bodie (1996) adds a few good reasons not to subsidise through a guarantee system; for instance because this may lead to a distorted resource allocation.
Furthermore, it is conceivable that political interference in a guarantee system is not always motivated by the desire to ensure its (long term) financial soundness. Various incentives may induce politicians to pursue for other, possibly even conflicting goals. The
15
Figure about the 2006 net position, PBGC (2007); figure about the 2005 net position, PBGC (2006). 16
The net position of the PBGC in the mid-nineties was not nearly as unfavourable as it is today. The net position was $-0.123 billion in 1995 and $0.993 billion in 1996 (PBGC, 1996).
17
The FSLIC was a governmental insurance corporation insuring deposits at savings and loan associations. When it was abolished in 1989, a large deficit was left to be covered by taxpayers (see also Bodie, 1996). 18
Subsection 2.5.3 elaborates on this subject. 19
deputy director of the PBGC, Vincent Snowbarger, believes that distance should be maintained from politicians in order to run a pension guarantee system properly.20
Finally, consider some remarks about Sweden’s Pensionsgaranti. According to Stewart (2007), this guarantee system is probably the most successful one. The president of Pensionsgaranti, Peter Lindblad, attributes the success to a number of factors.21 One factor is the effort put in assessing sponsor related risk and anticipating any changes. Another factor is that the system was founded on an initiative of a private sector and that is still supported by the large companies. Also, political interference in the system is completely absent. With the recent introduction of a new Swedish ITP plan Pensionsgaranti’s environment changed considerably. Persons born after 1978 are now covered by a pure DC ITP plan in stead of the original DB plan. On the one hand this development is unfavourable for Pensionsgaranti; the need for default insurance will eventually disappear even though short run effects are limited by the lengthy transition period (Pensionsgaranti, 2007).22 But on the other hand, the change also offers new opportunities. Pensionsgaranti could for instance capitalize on the interest of companies to create their own pension solutions adapted to the new ITP plan, like DC plans with minimum returns (Pensionsgaranti, 2008).
2.4 Adverse side effects
This section discusses the adverse side effects associated with pension guarantee systems and that form a direct threat for these systems.
The adverse side effects can be grouped into cases of moral hazard and adverse selection. These two concepts are the result of information asymmetries between market parties. Such asymmetries characterize insurance markets in general (Nicholson, 2005), and hence the insurance market effectively created through a government imposed pension guarantee system.
The remainder of this section is divided into two subsections. Subsection 2.4.1 discusses moral hazard in relation to pension guarantee systems, while subsection 2.4.2 addresses adverse selection.
2.4.1 Moral hazard
Nicholson (2005) defines moral hazard as the effect of insurance coverage on the decisions of individuals to take actions that might alter the likelihood or magnitude of losses. Typically, these actions boil down to more risk taking by the insured. After all, increases in potential
20
This statement is taken from the presentation of Vincent Snowbarger at the OECD Working Party on Private Pensions research seminar ‘reforming pension benefit protection schemes’, July 2nd 2007, Paris.
21
Peter Lindblad addressed these factors during the OECD Working Party on Private Pensions research seminar ‘reforming pension benefit protection schemes’, July 2nd, 2007.
22
losses resulting from additional risk taking do not (fully) accrue to the insured, while increases in prospective gains do. For the insurer of course, it is precisely the other way round. Moral hazard in an occupational pension sector with a guarantee system may consist of three different kinds of adverse behaviour.
First, sponsors may decide to lower contributions, especially when their own financial position is poor; in case of bankruptcy the pension plan will be taken over by the guarantee system anyway. Coronado and Liang (2005) find evidence for this behaviour. Focussing on pension plans covered by the American PBGC, it is observed that sponsors closer to bankruptcy tend to lower their contributions. The findings of Nielson and Chan (2007) underline this. These authors compare the funding levels of pension plans located in Ontario and covered by the PBGF, to the funding levels of pension plans located in three other Canadian provinces. It is observed that Ontarian pension plans are significantly lower funded on average.
Second, pension funds may have an incentive to follow an investment strategy that incorporates more or even excessive risk (e.g. Lachance and Mitchell, 2003). This is especially true for pension funds on the verge of insolvency. The reason for this is simply that it may be worth the shot; if the gamble turns out well, all profits accrue to the pension fund and in case things take a wrong turning, any losses are ultimately for the account of the guarantee system. Coronado and Liang (2005) study whether accumulated pension assets are allocated to more risky investments (equity) if the sponsor gets closer to bankruptcy. It is not observed that this is indeed the case.
Third, there could be an incentive to improve benefits promised under the pension plan. This is especially likely if guarantee system interference seems imminent, as the costs of all promises whatsoever will then almost certainly become at the expense of the guarantee system anyway.
2.4.2 Adverse selection
Adverse selection is defined by Mankiw (2002) as the tendency individuals with more information to self-select in a way that harms individuals with less information. Akerlof (1970) illustrates the concept on the basis of the market for used cars. Suppose there is an information asymmetry between sellers who know the quality of their cars, and buyers who lack all car quality information. So for buyers each car is perfectly equal to the next, meaning that both good and bad cars must sell at the same price.23 However, because this price is too low for some sellers of good cars, they withdraw from the market. This withdrawal in turn, lowers the average quality of cars for sale and therefore the prevailing market price, which
23
induces additional good car sellers to exit and so on. Hence, bad cars drive out good cars. Ultimately, the whole market could collapse, making transactions at any price level impossible.
Something similar could happen in an occupational pension sector with a guarantee system. Suppose that the guarantee system is unable (or unwilling) to distinguish between the quality of pension funds, meaning that low and high risk pension funds are charged the same premium. This inevitably causes low risk pension funds to subsidise the insurance of high risk pension funds (e.g. Bodie, 1996; Stewart, 2007). If this cross-subsidisation grows too large, low risk pension funds will try to withdraw from the system.24 This effectively means that high risk pension funds (and associated plans) drive out low risk pension funds (and associated plans); i.e. adverse selection.
A guarantee system that includes cross-subsidising can be avoided in a number of ways. First, a new occupational pension plan designed as a DC instead of a DB plan will not be covered. Of course, avoiding a guarantee system does not have to be the main reason to choose for a DC plan. Employers may simply find DB plans too expensive or may be unwilling to bear investment risk. But cross-subsidisation is likely to be a factor that promotes DC plan popularity.
Second, a DB plan can be withdrawn from a guarantee system by transforming it into a DC plan or by moving the associated pension fund system’s jurisdiction. A DB plan is effectively transformed into a DC plan by closing it to new members and simultaneously establishing a new DC plan. Physically moving the pension fund out of the guarantee system’s jurisdiction generally means leaving the country. This rather vigorous solution is actually being considered in reality. Indeed, only a few weeks after the start of the PPF British sponsors were thinking about moving the regulatory domiciles of risk low funds from the United Kingdom to particularly Ireland, in order to avoid PPF premiums (see LK, 2005).
2.5 Managing a pension guarantee system
The need to control adverse side effects underlines the importance of proper guarantee system management. This section discusses four methods of achieving this.
Merton and Bodie (1992) discuss a few interrelated management methods of guarantee systems in general. These authors emphasize that offering guarantees is a business activity which can only be viable if it is solvent and financially self-reliant in the long run.
Based on the discussion of Merton and Bodie (1992), Bodie (1996) sums up a number of management methods for guarantee systems; pension guarantee systems for example. The
24
