Costs and Benefits of Collective Pension Systems
O. W. Steenbeek S. G. van der Lecq (Editors)
Costs and Benefits of Collective Pension Systems
With 24 Figures and 28 Tables
123
[email protected]
[email protected]
Dr. Onno Steenbeek is financial economist Dr. Fieke van der Lecq is economist Both editors are affiliated with the Erasmus University Rotterdam, The Netherlands
Library of Congress Control Number: 2007933404
ISBN 978-3-540-74373-6 Springer Berlin Heidelberg New York This work is subject to copyright. All rights are reserved, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilm or in any other way, and storage in data banks. Duplication of this publication or parts thereof is permitted only under the provisions of the German Copyright Law of September 9, 1965, in its current version, and permission for use must always be obtained from Springer. Violations are liable to prosecution under the German Copyright Law. Springer is a part of Springer Science+Business Media springer.com © Springer-Verlag Berlin Heidelberg 2007 The use of general descriptive names, registered names, trademarks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. Production: LE-TEX Jelonek, Schmidt & Vöckler GbR, Leipzig Cover-design: WMX Design GmbH, Heidelberg SPIN 12111558
134/3180YL - 5 4 3 2 1 0
Printed on acid-free paper
Preface
In January 2006, the Dutch Association of Industry-wide Pension Funds (VB) told us about their plan to prepare a book on solidarity in collective pension systems. We were intrigued by this topic, both because of our interest in the pension sector and because of the connection with solidarity in a cost-benefit approach. After some discussions with VB director Peter Borgdorff, we decided to start a project with leading scholars and practitioners, which was to result in a book. We hoped that the researchers could investigate the extent of value transfers within collective pension funds, so that quantitative indications of this institutionalized solidarity would become publicly available. While the book was in progress, the discussion on solidarity and mandatory pension systems became very topical, and so were the results of the analyses. When the book was released in Dutch1, the chapter on costs differentials between pension funds and insurance companies also drew much attention. In the early months of 2007, the political and professional debates continued, with increasing attention from pension experts from abroad. This made us decide to try and arrange a translated version of the book. We were happy to find Springer Verlag, and collaborate with their enthusiastic publisher Dr. Niels Peter Thomas. We are also grateful to the Pension Science Trust (Stichting Pensioenwetenschap) for subsidizing the translation by Language Lab, and to the earlier mentioned VB for their generous collaboration in getting the international edition released. Now, the international pension sector colleagues can also profit from this assessment of the Dutch second pillar pension system. The Dutch pension sector has much to offer in terms of expertise and institutional experience, as the various distinguished authors in this volume show. The debate on collective pension funds remains fierce, both in the Netherlands and elsewhere. Nearly everyone has a stake in it, and a tremendous amount of money is involved. Those with established interests defend the 1
S.G. van der Lecq and O.W. Steenbeek, 2006, ‘Kosten en baten van collectieve pensioensystemen’, Deventer: Kluwer publishers.
VI
Preface
system, while relatively new stakeholders sometimes criticize it. Our aim is not to settle the debate, nor to present our own opinion about the matter. Instead, we set up the book as a collection of facts and arguments, so that the reader can form his/her own opinion, for instance on the dilemma’s in the concluding chapter. May evidence, solid arguments and wisdom prevail. Rotterdam, July 2007
Onno Steenbeek and Fieke van der Lecq
Executive summary
This volume takes stock of the costs and benefits of collective pension systems. The concept of solidarity between participants is key. Which groups show solidarity with which other groups, and how much money is involved? Part 1 concentrates on the question as to which aspects of solidarity are relevant with respect to collective pension funds. Several groups show solidarity with each other, such as the old and the young, men and women, healthy and ill participants, et cetera. Solidarity amongst these groups is also present in health care insurance, and this solidarity is compared to that of the second pillar collective pension systems. In part 2, several aspects of collective pension arrangements are quantitatively assessed. It turns out that the costs of execution of pension agreements differ substantially between different types of pension systems. The cost differences between the average collective pension arrangement on the one hand and executed by insurance companies on the other hand are very large, but differences among collective systems can be significant as well. Solidarity is about distribution: who receives the benefits of collective systems and who pays the costs? Costs and benefits are not shared equally and tentative calculations present an impression on these redistributive effects. They indicate that particular forms of solidarity bring about substantial ex ante value transfers between groups of participants. The value transfers between generations are largest. Part 3 discusses the mandatory participation by companies and individuals, and explores what would happen if participation would no longer be mandatory. As for individuals, experience from abroad indicates that people who are not obliged to participate in a pension scheme, tend to save too little, invest their savings poorly and cash out whenever they get the chance. The mandatory participation by companies in industry wide pension funds – whenever possible – has resulted in a reduction of ‘blank spots’. It has also prevented firms from competing on the labour market via their pension contributions. Collective pension agreements provide substantial advantages above individual pension plans. The execution costs are much lower, and risk
VIII
Executive summary
sharing within and across generations improves the average result to the participants. However, the benefits are not distributed evenly among individual participants. This brings about distortions on the labour market, which also imply costs. Still, on the macro level the benefits of collective systems exceed their costs.
Table of contents
1
Introduction ...................................................................................... 1 S.G. van der Lecq and O.W. Steenbeek
Part 1. The concept of solidarity 2
Solidarities in collective pension schemes ............................... 13 J.B. Kuné
3
Solidarity: who cares? ................................................................... 33 P.P.T. Jeurissen and F.B.M. Sanders
Part 2. Quantifying solidarity 4
Operating costs of pension schemes .......................................... 51 J.A. Bikker and J. de Dreu
5
Optimal risk-sharing in private and collective pension contracts........................................................................................... 75 C.G.E. Boender, A.L. Bovenberg, S. van Hoogdalem, and Th.E. Nijman
6
Intergenerational value transfers within an industrywide pension fund – a value-based ALM analysis ................. 95 R.P.M.M. Hoevenaars and E.H.M. Ponds
7
Intergenerational solidarity in the uniform contribution and accrual system....................................................................... 119 T.A.H. Boeijen, C. Jansen, C.E. Kortleve, and J.H. Tamerus
X
8
Table of contents
Everyone gains, but some more than others........................... 137 K. Aarssen and B.J. Kuipers
Part 3. Mandatory participation 9
Why mandatory retirement saving? .......................................... 159 P.J.A. van Els, M.C.J. van Rooij, and M.E.J. Schuit
10 Mandatory participation for companies.................................. 187 P.H. Omtzigt
Part 4. Conclusion 11 Macroeconomic aspects of intergenerational solidarity....... 205 J.P.M. Bonenkamp, M.E.A.J. van de Ven, and E.W.M.T. Westerhout
12 Summary and conclusions ......................................................... 227 S.G. van der Lecq and O.W. Steenbeek
About the authors................................................................................ 237 Subject Index........................................................................................ 243
1
Introduction
S.G. van der Lecq and O.W. Steenbeek Employee solidarity is central to the second pillar of the Dutch pension system. This solidarity is given shape in collective schemes implemented by industrywide, company and professional group pension funds. The present book sets out to explain how solidarity works within collective pension schemes and to answer the questions: what groups participate in the solidarity mechanism and what are the financial stakes for each group? After reading this book the reader will be in a better position to form his own opinion as to whether the current collective pension system is desirable or not. The concept of solidarity looms large in the public debate on pensions, particularly now that the funding of old-age provisions is under pressure. Solidarity is a complex issue, however, and gives rise to many questions. First of all: what groups are expected or required to show solidarity to what other groups? Secondly: how must that solidarity be organised? Answering these two questions is the purpose of this book. In many developed economies, solidarity in pensions is institutionalised through a state pension system in the first pillar and supplementary collective pension schemes in the second pillar. First pillar systems are often a general statutory scheme funded by contributions from the overwhelming majority of the population, so in this case mutual solidarity is virtually universal. This system of universal solidarity is not the subject of the present book, however. Our focus here is on the supplementary collective pension schemes that form part of remuneration packages. Each collective pension scheme makes its own choices, where some participants are perceived to benefit more than others. This inequality in the distribution of costs and benefits is one reason why opinions are divided as to the desirability of maintaining and/or modifying the collective pension systems. This book seeks to provide a firm foundation for that debate by putting the costs and benefits of collective pension systems into clear focus. These costs and benefits are quantified where possible.
2
1.1
S.G. van der Lecq and O.W. Steenbeek
Backgrounds to this book
This is not the first time that the costs and benefits of collective pension systems have come under scrutiny. An earlier study by the WRR (Scientific Council for Government Policy) also compared the costs of solidarity within a pension fund (i.e. the value transfers between participants and disruptive effects in the labour market) with the benefits of collectivity (risk-sharing and economies of scale). In addition, a comparison was drawn between the cost-benefit ratio of collective schemes and that of individual pension schemes (Boender et al., 2000). Both the quoted study and another WRR report calculated that collective pension funds lead to a substantial increase in wealth, benefiting the total economy (Jansweijer et al., 2001). The present book can in a certain sense be seen as an update of the WRR studies in the light of the current debate about solidarity in the Dutch pension system. It also digresses on the several subgroups within the population of active participants and pensioners, to provide an insight into the value transfers between these subgroups. Put simply: who pays and who gains? Another relevant aspect in the framework of this book is the debate surrounding the transition from defined benefit system to defined contribution schemes. This is a particularly interesting development given that the latter system is much more individualized and therefore has a much smaller solidarity component. The answer to the question as to whether the benefits of a different system outweigh its costs will thus decide the future direction of the pension sector. Leaving aside the actual design of the system, the very principle of solidarity in the field of pensions is under pressure. Two causes for this can be identified. First of all, ageing is making heavy demands on the younger generations, also in relation to supplementary pension provisions. One example of this in the Netherlands concerns the transition arrangement for the abolition of early retirement and the phasing out of the pre-retirement pension, where the younger generations of employees are expected to foot part of the bill. They are up in arms over this issue, partly because they themselves will never enjoy the benefits of such schemes (Roscam Abbing et al., 2005). This critical attitude of young people is reinforced by the advent of the transitional labour market. Lifelong salaried employment is now a thing of the past, which means that pension fund participation is also becoming intermittent – with potentially dire consequences for pension accrual rates (Muffels et al., 2004). Some initiatives are already being taken in the Netherlands to counter this threat. One such step concerns the creation of a special ‘open’ pension fund for self-employed persons by the
1 Introduction
3
AVV, a dedicated union for the self-employed and freelancers. The AVV happens to have a relatively high proportion of young members – which is precisely the group of people currently contending with the flexible labour market where there is the constant threat of losing pension rights due to changing jobs. To sum up: those who are being hit in their pockets by the changing labour landscape are simultaneously expected to show solidarity with the ageing population. A second reason as to why solidarity is under fire can be found in the growing availability of information on this issue. Pensions are now headline news. The drive for a future-proof pension system has become the subject of public debate and the widespread dissemination and sharing of information is giving people a better understanding of the pension system, including its built-in solidarity mechanisms. There is an inherent danger in this process. If, for instance, such information feeds fears of eroding solidarity and the subsequent collapse of the entire system, a self-reinforcing process may be set in motion (De Beer, 2005). Nevertheless the authors of this collection of essays believe that information is far preferable to ignorance – for only an informed debate can put the choices of the past and the choices for the future in clear perspective. The following chapters will provide examples of this through the quantitative elaboration of some central solidarity issues. Facts are the building blocks of an informed opinion. So to help the reader form, substantiate or modify an opinion on the current debate, our first step will be to provide clear definitions of the central concepts of solidarity and collectivity. Next, we will outline the design of the book and indicate the close interrelationship between the issues raised in the various chapters.
Solidarity The term “solidarity” is used widely in this book, for different aspects of collective pension arrangements. The most important distinction is between types of solidarity that are mutually beneficial and those that are not. The first type is identical to an insurance contract: when a participant in the pool suffers a damage, the other policy holders will compensate the loss. This is usually “risk sharing”, which is an important element of all collective arrangements. Less obvious is “one-sided solidarity”, which refers to elements in collective pension arrangements where it is clear ex ante who will subsidize whom. For example, in current collective pension systems, value is transferred from young participants to old participants. The term “value transfers” is usually applied to these elements of collective pension arrangements.
4
S.G. van der Lecq and O.W. Steenbeek
1.2
How are solidarity and collectivity related?
SOLIDARITY De Beer defines solidarity as ‘a positive sense of shared fate: the fate of one person is positively bound up with the behaviour of the other. A person shows solidarity if his behaviour benefits the other.’ (De Beer, 2005, page 56) This, however, is a very broad definition which could also cover all sorts of gifts, for instance. The same applies to his description of solidaritybased behaviour as ‘every form of behaviour that deliberately benefits the other, without immediately requiring a benefit in return.’ (De Beer, 2005, page 55). The addition of the terms ‘deliberately’ and ‘immediately’ point in the direction of a further narrowing down. The term ‘deliberately’ suggests that solidarity can be an aim in itself: the action is consciously designed to benefit the other. No immediate benefit in return is expected, which is the big difference between solidarity and a barter transaction. This is borne out by another definition of solidarity as ‘a positive sense of shared fate between individuals or groups. That is, a situation where social relationships centre on the stronger helping the weaker or on promoting the communal interest’ where the term ‘help’ is used (Van Oorschot, 1997, quoted in Beltzer and Biezeveld, 2004, page 41).
DIMENSIONS IN SOLIDARITY To classify different forms of solidarity, two dimensions are distinguished (De Beer, 2005, pp. 56-58): • obligatory and voluntary solidarity; • one-sided and two-sided solidarity. Voluntary solidarity can be informal and unorganised. This form is also known as warm solidarity. Examples are caring for family and donating to charity. Sometimes this solidarity is one-sided, e.g. towards children, and sometimes two-sided, e.g. between friends. Voluntary solidarity can also be institutionalised, and is then known as cold solidarity. Another, probably better known term for cold or voluntary solidarity is “risk sharing”, where it is not clear in advance who is the recipient and who is the payer. This then is two-sided. An example concerns private insurance. One-sided institutionalised solidarity can be found in tax-funded social security services where the payers and the beneficiaries are known in advance. The various forms of solidarity are worked out in greater detail in the first chapters of this book and return in the final conclusions.
1 Introduction
5
Empirical research shows that one-sided solidarity is retreating while two-sided solidarity is advancing (De Beer, 2005, pp. 61-70). Particularly institutionalised (cold) one-sided solidarity is decreasing in terms of the share of gross domestic product allocated to it. Informal (warm) one-sided and two-sided solidarity are more or less stable1. Remarkably, institutionalised (cold) two-sided solidarity is on the increase. This suggests that people are becoming more calculating in their behaviour and want to guarantee that their solidarity is returned in equal measure by means of formal arrangements. Solidarity thus takes on the nature of an insurance policy. With pensions, as with insurance, people want to know what the costs and benefits are.
RISK SOLIDARITY AND SUBSIDISING SOLIDARITY With pensions a distinction can be made between risk solidarity and subsidising solidarity. The latter is ex ante, i.e. it is clear in advance who shows solidarity to whom. The former is ex post, i.e. this only becomes clear retrospectively. Risk solidarity or risk sharing involves a pooling of risks, such as with fire insurance. Risks are thus effectively shared, as is the case with collectivities. As a result, the communal premiums paid are sufficient to cover the communal risk whereas the individual premium would nowhere near cover the costs of the individual risk should that risk actually materialise. In the pension system the investment risk is borne collectively, so that on balance more investment risk can be taken than if each participant were to try to optimise their investments individually. Subsidising solidarity involves value transfers, where no attempt is made to exactly match the individual premium with the individual risk. As a result, good risks offset bad risks, such as between employed and retired people, men and women and the sick and the healthy. In these cases it is clear in advance that value will flow from one group to the other. Subsidising solidarity goes beyond the sharing of risks that is inherent in collectivity-based risk solidarity. For this reason, subsidising solidarity is often seen as ‘real’ solidarity. Pension funds combine both types of solidarity, while insurers specialise in risk solidarity. 1
Note that there is no question of ‘warm’ solidarity being substituted by ‘cold’ solidarity. So there is no question of an analogy with the substitution of intrinsic motivation by extrinsic motivation (Frey, 1997). Warm solidarity remains as much in evidence as ever.
6
S.G. van der Lecq and O.W. Steenbeek
Different authors define and classify solidarity in different ways. So the above definition and classification need not correspond exactly with those used in the following chapters.2
DEVELOPMENTS The advantages of collectivity or risk solidarity are, by their very nature, not open to question. However, the desirable scale of such collectivity arrangements within society is currently the subject of intense study and debate, also in this book. Regarding the risk solidarity encapsulated in institutional arrangements, we noted earlier that the one-sided variant is retreating in favour of the two-sided variant. People are prepared to show solidarity to those in need, provided the favour is returned when they themselves require assistance. This can only be guaranteed through agreements that lay down clear rights and obligations. Basically, people’s deep-felt need for formal solidarity arrangements reflects uncertainty over the future rather than a fundamental lack of solidarity as such. This is consistent with the finding that warm solidarity, which manifests itself informally both in one- and two-sided relations, is not decreasing. In fact, security provided by institutional arrangements may actually serve to facilitate and encourage informal acts of solidarity: “Voluntary, spontaneous solidarity and enforced organised solidarity are therefore not substitutes but largely complementary”(De Beer, 2005).
SOLIDARITY IN PENSIONS Within the pension sector mandatory solidarity takes shape via the collectivity. Within this collectivity, value is transferred from certain groups of participants to other groups of participants, e.g.: • from employees to retirees and inactives; • from young to old;
2
In the insurance industry, the term ‘risk’ is used differently from the pension sector, for instance with respect to subsidies from ‘bad’ to ‘good’ risks. Here, the term risk refers to people who run a particular risk to a greater or lesser extent. These people with different risk profiles show solidarity towards one another, thus giving rise to the term risk solidarity. In the context of pension funds, by contrast, risk solidarity is equated with risk sharing because economists operationalise the term risk on the basis of distribution of probabilities.
1 Introduction
7
• from employees to early retirees; • from men to women; • from those who die young to those who live long • from singles to married persons (in the case of mandatory surviving dependant’s pension); • from healthy to unhealthy (in the case of non-contributory pension accrual during disability); • from employed to unemployed (if pension accrual continues during unemployment); • from businesses to businesses (different numbers of pension members). Not all forms of solidarity occur at each pension fund to the same degree. Each pension scheme has its own solidarity profile. This book deals with solidarity between some of the groups mentioned above. Intergenerational solidarity is currently one of the central issues and is therefore particularly highlighted, drawing on recent data and new computational models.
1.3
Structure of the book
The first part of this book explores the concepts of solidarity and collectivity. Chapter 2 elaborates these in the context of the pension sector, while chapter 3 compares solidarity and collectivity in the pension sector with the health sector. The second part of the book looks at the costs and benefits of solidarity in collective pension systems on the basis of qualitative and quantitative analysis. Chapter 4 dissects the costs of pension funds and pension insurers, thus highlighting the efficiency differences between collective systems mutually as well as between collective and individual pension schemes. Chapter 5 answers the same question for the investment policy of the pension funds, revealing that collective schemes also yield better investment returns than individually tailored schemes. Chapter 6 views the generations separately, making it clear to what extent solidarity exists between young and old and indicating the effects of specific policy adjustments within existing schemes. Chapter 7 centres on the use of the uniform contribution to promote solidarity and indicates the costs and benefits of this instrument. Chapter 8 homes in on solidarity from the perspective of the
8
S.G. van der Lecq and O.W. Steenbeek
individual. On the basis of a number of representative participants it shows in what situations an individual is a net contributor or net beneficiary of the solidarity arrangements in the current pension system. Collectivity in the pension system is assured through the mandatory participation of the individual (Chapter 9). In addition, many sectors operate mandatory industry pension schemes for affiliated employers (Chapter 10). Both types of mandatory participation are discussed in terms of costs and benefits in the third part of this collection of essays. The various analyses of the way in which solidarity is given shape in the collective pension system are evaluated in the final part of the book. This takes place in Chapter 11, where the analyses are discussed from a macro-economic perspective. Chapter 12 contains a summary as well as conclusions. The bulk of the material in this book is new. More important than that, however, is the thematic coherence between the chapters, which should allow the reader to form a well-founded opinion on the costs and benefits of solidarity in collective pension schemes.
Literature Beer, P.T. de, 'Hoe solidair is de Nederlander nog?, in: E. de Jong en M. Buijsen (ed.), Solidariteit onder druk?, Nijmegen: Valkhof Pers, 2005, pp. 54-79. Beltzer, R. and R. Biezeveld, De pensioenvoorziening als bindmiddel, Amsterdam: Aksant, 2004. Boender, C.G.E., S. van Hoogdalem, E. van Lochem and R.M.A. Jansweijer, ‘Intergenerationele solidariteit en individualiteit in de tweede pensioenpijler: Een scenario-analyse’, WRR (Scientific Council for Government Policy) report, 114, The Hague: WRR, 2000. Frey, B., Not just for the money; An economic theory of personal motivation, Cheltenham: Edward Elgar, 1997. Jansweijer, R.M.A and A.J.C.M. Winde (ed.), Intergenerationele solidariteit en individualiteit in de tweede pensioenpijler: een conferentieverslag, work document 116, The Hague: WRR (Scientific Council for Government Policy), 2001. Muffels, R., P. Eser, J. van Ours, J. Schippers and T. Wilthagen, De transitionele arbeidsmarkt. Naar een nieuwe sociale en economische dynamiek, Tilburg: OSA, December 2004.
1 Introduction
9
Roscam Abbing, M., and many others, FNV is niet solidair met jonge ambtenaren, De Volkskrant, 18 July 2005, in: G.M.M. Gelauff et al. (ed.), KVS Jaarboek 2005/2006, The Hague: Sdu uitgevers, 2006, pp. 21-22. WRR, Generatiebewust beleid, particularly chapter 6: Intergenerationele risicobeheersing in de pensioensfeer, WRR (Scientific Council for Government Policy) Report 55, The Hague, 1999.
Part 1. The concept of solidarity
2
Solidarities in collective pension schemes
J.B. Kuné1 This opening contribution to the book Costs and Benefits of Collective Pension Systems, an initiative of the Association of Industry-Wide Pension Funds and the Erasmus University of Rotterdam, first discusses a number of general solidarity aspects. Next it deals with more specific issues that have an important bearing on the pension sector. The collective pension system draws its strength and justification from collectiveity and solidarity. The conditions required to ensure broad-based consensus and acceptance must therefore be fulfilled at all times. Hence, the question arises what forms and degree of solidarity are desirable. Where does undesirable solidarity start? The outcomes will differ from one fund to the other and will also vary over time. In the past few years pension funds have made important progress in introducing pension contracts. These contracts specify the relationship between the fund’s financial position on the one hand and its contribution policy and indexation of accrued pension rights (of retirees and workers) on the other. In the coming years further arrangements will be made regarding diverse types of income and value transfers, also known as solidarities. Numerous questions arise waiting for an answer. Some tentative recommendations are made for the way forward.
2.1
Solidarity and social cohesion
The notion of solidarity is closely intertwined with the creation and evolution of the national welfare state. Solidarity appears in many guises. Where the principle of reciprocity plays an essential role, we speak of horizontal solidarity (Teulings, 1985: 48-73). This involves (homogeneous) groups of people making mutual rather than individual arrangements in order to gain cost and other efficiencies. A certain equality among the 1
This article was written in a personal capacity.
14
J.B. Kuné
participants is an important condition for realising such a contract. Solidarity with the group thus implies exclusion of others. On the opposite side of the spectrum we find vertical solidarity, which is not predominantly based on reciprocity and self-interest. The Dutch welfare state is a good example. Here, the better-off surrendered part of their wealth to assist more vulnerable groups, usually with a view to protecting society by avoiding social unrest and crime, and promoting public order, health and housing. The nation state plays a crucial role in this vertical solidarity by extracting donations from the population. Given this coercive element, one could say that the concept of enforced solidarity contains an internal contradiction. In this case too, outsiders must be excluded so that the nation can provide welfare to its vulnerable citizens. A sense of shared community, values and identity is necessary to legitimise this solidarity (Van Oorschot, 1997: 162-174). Vertical solidarity thus functions within the borders of the nation state as the moral cement of society. The AOW state pension is a prime example of this in the Netherlands. The survival of this system depends largely on sustained and stable consensus within society. Without this consensual acceptance, the state pension would lose its ‘equal rights for all’ character and, like supplementary pension schemes, become dependent on the individual’s employment and salary history (Kuné, 2004). The more extreme form of solidarity, based largely on the compassion of one human being for another, is less important in value transfer terms and is also more tenuous in nature. For instance, an immigration society characterised by increasingly heterogeneous population groups will find it also increasingly more difficult to continue allocating all conceivable rights to everyone residing within the territory of the nation (Entzinger and Engbersen, 2004: 41-55). As noted, solidarity draws its lifeblood from a shared sense of identity. It presumes affinity with an in-group whose members have a common history and heritage and are therefore worth a sacrifice. It follows that there is an out-group for whom the community is less willing, or even totally unwilling, to make a sacrifice. Solidarity thus always has its limits. It can be extended, but not infinitely. Beyond a certain level, solidarity becomes subject to various conditions. In the ethnically diverse United States, for instance, there is clearly less solidarity than in European countries with more homogeneous populations.
SOCIAL COHESION Various studies reveal a picture of fairly stable solidarity (in attitude and behaviour) in the Netherlands (Beltzer and Biezeveld, 2004; De Beer, 2006). However, western societies are set to become more heterogeneous in the
2 Solidarities in collective pension schemes
15
coming century. We therefore foresee more segregation along ethnocultural lines, with a growing number of smaller homogeneous subpopulations whose main allegiance is to their own circle. Fault lines are thus more likely to arise between socio-economic classes and ethnic groups than between generations. In a society where there is less cohesion and hence less support for all sorts of solidarity, increasingly more but smaller homogeneous groups will emerge. These, in turn, will build up their own internal solidarity structures based on common values and interests. Pensions is one area where this development will come to the fore. An additional factor is that the advent of genetic identification will provide yet another means of demarcating a larger number of small, more homogeneous groups. This may further reduce the appetite for society-wide solidarity, with individuals who do not match the group DNA simply being excluded. Thus, genetic similarity – together with more traditional criteria such as membership of a certain industry, company, professional group or social class – can also serve as a basis for defining specific groups of people who are willing to insure certain risks. These circumstances will induce a shift towards smaller collectivities (Aarts and De Jong, 1999). Income is already widely used as a differentiation criterion, where identifiable groups of people whose earnings exceed a defined threshold start their own supplementary pension schemes based on cost-price funding. Below this income threshold, everyone continues to participate in the same collective scheme. After these introductory observations, let us return to the real subject of this chapter: solidarities within supplementary pension schemes. It’s worth noting, incidentally, that the terms ‘risk-sharing’ and ‘risk-sharing arrangements’ are increasingly being used to denote situations which actually involve a transfer of funds or solidarity from one generation to the other or from one group of participants to another group. The next section describes several aspects of solidarity in the supplementary pension system. Section 2.3 then tentatively formulates some principles providing a basis for making a distinction between desirable and undesirable forms of solidarity. As will be clear, these principles are to some extent personally coloured. The appendix finally provides a summary of the different types of solidarity present in supplementary pension schemes.
2.2
Income and value transfers
This section casts light on some aspects of the types of income and value transfers occurring within the supplementary pension system. These transfers are also known as solidarities.
16
J.B. Kuné
THREEFOLD DISTRIBUTION MECHANISM A pension system realises a threefold distribution of funds: two are simultaneous and one is sequential. The pension system thus acts as a threefold distribution mechanism. At macro-level the pension system ensures a (simultaneous) distribution of funds between older and younger generations in each time period. The distribution mechanism depends on the applied method of funding: pay-asyou-go, funded or any combination of the two. With pay-as-you-go financing, the pensions are paid out of the earned income of the working population, while with a funded system the pension income is obtained from the accumulated pension assets. At a micro-level the pension system realises a (sequential) distribution of funds over the active and post-active period of a person’s life. The distribution mechanism exclusively concerns the financing via the funded system. The ownership of pension capital, accumulated during the active period, provides the individual during his post-active period with an economic claim on national product or a claim on production of foreign countries if the pension assets have been (wholly or partly) invested there. In the words of the WRR (Scientific Council for Government Policy) ‘… To slightly overstate the case, this (dissaving in the post-active phase) effectively entails that the future generations of young people dedicate themselves to labour-intensive caring for the elderly and are rewarded for this with cars, computers, video recorders and so on that are imported from Asia (and paid for with the aforementioned dissavings).’ (WRR, 2000: 109). At an intermediate or meso-level, the Dutch pension system encompasses numerous redistribution mechanisms which lead in each time period to (simultaneous) income transfers (in different directions) between all those belonging to the same age group or generation. Most of these value transfers are found in the first pillar state pension (AOW), but they are also present in the supplementary pension system. Little is known about the total amount involved in these solidarity-based value transfers; there is not much transparency. Note that the state pension was primarily developed as an income policy instrument. At an individual level there is little or no relationship between the amount and distribution of the contribution payments on the one hand and the size of the state pension payments on the other. The solidarities in the supplementary pension schemes are significantly more diffuse. It is generally observed that the Dutch pension system cannot survive without (subsidising) solidarity. But this claim rests mainly on emotional notions such as the idealistic belief that members of society should be pre-
2 Solidarities in collective pension schemes
17
pared to pay for one another. (Aalberse et al., 2004). So far there has been no rationalisation of the required solidarities in terms of e.g. circumstances and conditions, beneficiaries and contributors, and realistic limits.
NUMEROUS SOLIDARITIES We can identify numerous solidarities which lead to a frequently significant redistribution of funds between various groups of participants. These include the solidarity between actives and post-actives, between young and old (in a stable situation the young make a sacrifice now and benefit in their old age), between individual companies within an industry, between workers and the disabled (who continue accruing a non-contributory pension), between those who die young and those who become old, between employees with divergent career paths, between males and females, between participants with different marital status, between different types of pension, solidarity arising from the working of the financing system which determines how the financing of new obligations is distributed over time across the existing population of actives (particularly the complex redistribution arising from the level premium that is applicable at all times to all participants), between spouses in respect of the pension exchange option, etc. See also the appendix. Jointly all these solidarities ensure a simultaneous and sequential settlement of pension outcomes between different groups of participants. The calls for transparency throughout society as well as in the world of pensions will grow louder as society continues to become more heterogeneous. This certainly applies to such a sensitive issue as the distribution of pension costs. It is therefore inevitable that all sorts of open and (largely) hidden subsidy flows, i.e. solidarities, will eventually be made explicit and transparent. Greater transparency – more knowledge and a better understanding of how ‘it works’ – can either strengthen or erode the consensual support within society. In the latter case, the solidarity framework must be redesigned in order to put the system back on a viable basis. We can assume that a rational person who knows that collective schemes mostly yield a better pension result than individual schemes will – if he has a choice in the first place – voluntarily opt for the former, provided that undesirable ex-ante solidarities are sufficiently removed from the scheme. This raises the question: where does desirable solidarity stop and undesirable solidarity start? Research into the size of subsidising financial flows has yet to begin. The participants will want to form an opinion about the desirable and undesirable solidarities: what type of solidarity, by whom
18
J.B. Kuné
and for whom, how much, when, for how long and what are the limits? The answers to these questions will vary between pension institutions and populations and will also be subject to change over time.
RISK SHARING OR SUBSIDIES Within supplementary pension schemes, risk sharing mainly refers to solidarity regarding the duration of life. Substantial differences in life expectancy exist between men and women, between socio-economic classes and between the healthy and less healthy. Moreover, each group is characterised by individual differences in life expectancy. For risk-averse participants this type of solidarity or risk sharing usually has a significant wealth-enhancing effect as they are more likely to benefit from an indexed pension until well into old age. This risk solidarity within homogeneous groups is probably greater – because the differences in life expectancy are smaller – than between various socio-economic groups and between men and women. Gender solidarity is incorporated in European legislation, in this case the antidiscrimination legislation. Within collective pension schemes, life duration solidarity is realised by levying the same level premium on all participants. New entrants cannot be excluded, nor can they decide to withdraw from the fund. Participation, in other words, is mandatory.
VALUE TRANSFERS The justification of the aforementioned risk sharing is not widely questioned, though even here selective behaviour and moral hazard may arise. With subsidising solidarity, settlement of the pension outcome already exists ex ante in a certain manner. By far the most important category of subsidising solidarity concerns the solidarity in relation to the investment returns on assets. This form of solidarity takes the shape of income and value transfers that are triggered when actual returns undershoot expected returns, while the benefits and pension accrual rates are adjusted to a lesser degree or not at all.2 This is
2
The solidarity resulting from equity and fixed-income price movements can be mentioned in one breath with the aforementioned solidarity in respect of investment performance. The latter entails that when interest rates fall the obligations – apart from perfectly matching investments – usually increase more than the fixed-income investments. Finally we should note that no mutual redistribution exists between pension schemes in the Netherlands. Evidently there is such a close – though unspoken – cohesion and/or special bond between the partici-
2 Solidarities in collective pension schemes
19
also referred to as intergenerational solidarity. Lower-than-expected returns affect all participants. The question then is: how must the pain be distributed over the participants. What is reasonable and fair?
Protect retirees? It is widely agreed that the group of retirees must be protected against the impact of a negative investment shock as much as possible. They, after all, have no way of taking corrective action. Does the same apply to active participants who are approaching retirement? This would mean that younger participants would have to bear the entire burden of absorbing the threatening fall in retirement income. During the recovery period they will continue paying contributions, whilst accruing less or no pension rights for themselves. Asking for this sacrifice from young people is often considered justified given that they may be able to benefit from higher-than-average returns in the future. What’s more, their incomes will rise over time, giving them an opportunity to buy extra pension rights if necessary. Nevertheless, this sacrifice is greater than it seems. Contributions paid at a young age can be placed in higher-risk investments for a longer period of time. Sacrificing this time advantage means that young people will need to pay higher contributions to accrue the same amount of pension. In addition, demanding an extra contribution from the young to repair the incomes of retirees may provoke negative labour market effects and send the national economy into a downward spiral. The principle of the level premium entails only a weak relationship between the contributions paid during the active years and the pensions received upon retirement. Instead of an insurance it is really a provision for the collective group, just as the state pension (AOW) is a provision for all residents. The fact that this scheme is not a personal insurance but a collective provision is an argument in favour of charging contributions to retirees, too.3
pants in a particular scheme that a certain degree of redistribution within the scheme is considered justified. However, there is no basis for an income redistribution between different occupational groups such as between the less well-off cigar makers and rich professional groups such as notaries, (consulting) actuaries, medical specialists, pharmacists, and so forth. 3
Added to that, it can be argued that the elderly should show more solidarity to one another (WRR, 1992).
20
J.B. Kuné
Some generations may contribute substantially more than they receive during their lifetime, while the opposite applies to other generations. But is this always justified? After a fall in the value of the assets new entrants will not readily volunteer to join a pension scheme and saddle themselves with the obligation of repairing the pension incomes of the elderly. So some persuasion will be required to coax the younger participants into shouldering collective responsibility. Fortunately, strong arguments do exist for asking young people to make a sacrifice in the form of a ‘solidarity contribution’ (see section 2.3.)
RECIPROCITY Where advantages and sacrifices are distributed more or less equally within successive subperiods over the lifetime of successive generations, intergenerational solidarity is a universal good that makes everyone more prosperous. Here, there is a strong degree of reciprocity. However, where certain generations make systematic sacrifices over their entire lifetime while others benefit equally systematically – and there is consequently no reciprocity – the fairness of the system is open to question. This holds all the more so if the conditions for the respective generations are otherwise equal. One example of this situation is the pre-retirement pension. In the transition period towards a fully funded pre-retirement pension, many will make substantial contributions towards early-retirement and similar (transitional) schemes without ever being able to enjoy the fruits of these schemes themselves. In the light of the ageing population, the Dutch government had good reason to intervene in this mechanism. Unfortunately, society has shown little understanding and sympathy for this policy. The government has evidently failed to persuade the public of the fairness and necessity of these measures.
PERVERSE SOLIDARITY Some forms of solidarity are desirable, others are undesirable or unintended (inverse solidarity) and still others are unnatural (perverse solidarity).4 The (fair) – simultaneous and sequential – distribution of benefits and costs over generations is also known as ‘generational accounting’. 4
Our opinion about (a certain form of) solidarity depends on how we see society. A strict viewpoint is that solidarity cannot be taken for granted and must be earned. Only those who have done everything to avoid becoming dependent are worthy of solidarity.
2 Solidarities in collective pension schemes
21
One example of an accumulation of (partly inverse, partly perverse) solidarities concerns the transfer of income between a fictitious man and woman. The man started working at a young age, is untrained and unmarried (with a flat or even declining career path). On top of all this, he dies at a relatively young age shortly after being declared unfit for work. The woman started working at a later age, is highly educated (with an upward career path), is married to a younger man, and lives to a ripe old age. In the current system, significant ‘solidarity’ value transfers take place from the first man to the woman. Clearly the chosen rate of time preference or discount rate is crucial when comparing advantages and sacrifices over time. The rate of time preference is time-dependent and probably also group and persondependent (think e.g. of socio-economic class and age). The current tendency, for both scientific and practical purposes, is to select a discount rate of about 1% (Davidson, 2004: 290-293). A high discount rate (about 5% or higher) attaches substantially less weight to the future and to the interests of future generations.5
SOLVENCY MARGINS A pension fund obviously must determine the value its future obligations (i.e. pension payments) as accurately as possible. On top of this estimated amount, it will want to maintain a (solvency) margin or buffer to absorb setbacks. The size of the buffer depends on the severity and duration of the shock against which the fund wants to protect itself. The desirable degree of protection must be determined taking into account that the accumulation of such a buffer demands sacrifices from the current generations. An alert pension fund responds rapidly and adequately to economic shocks by changing pension benefits and accrued rights of actives. For such fund, a comparatively small solvency margin suffices. Conversely, a pension fund with a generous indexation policy will need a bigger buffer. The upside of a higher buffer is that it yields the pension fund correspondingly higher investment revenues. The resulting surplus return – i.e. the return in excess of the return required to maintain the desired buffer – can be used to apply catch-up indexation and/or reduce the contributions. Buffers give pension funds more room for manoeuvre. In the event of a shock, for instance, they can opt to raise contributions gradually rather 5
A payment of 100 by future generations 30 years hence that is transformed to today obtains a weight of 0.74 at a discount rate of 1% and a weight of 0.23 at a discount rate of 5%.
22
J.B. Kuné
than abruptly. Building, maintaining, and releasing buffers involves numerous redistributional aspects and, depending on the speed at which this takes place, has implications for several generations.
PENSION CONTRACT Scheme participants are increasingly demanding adequate information about the operation and performance of their pension fund. Openness and transparency are important to avoid alienating the critical and inquisitive participant from the fund. Participants also frequently want advice. The Pension Act deals extensively with the provision of information to participants. One new phenomenon is the ‘pension contract’, which lays down arrangements concerning the relationship between the recovery of the fund’s financial position (i.e. the funding ratio: the value of assets divided by the value of liabilities) on the one hand, and the contribution policy and indexation of accrued rights (of retirees and the active participants) on the other. This integrated contribution and indexation policy must be implemented according to a defined graduated scale (CPB, 2000). The pension contract must make clear in advance when pensions will and will not be indexed, and to what extent. In the past, contribution and indexation measures were usually taken on an ad hoc basis (incomplete contract). Within the context of the FTK (Financial Assessment Framework) the nominal funding ratio for a standard pension fund has been set at about 130% and the recovery period at a maximum of 15 years. When the real obligations are fully funded, gradual contribution reduction is permitted above a funding ratio of 130%. If the funding ratio rises even further, contribution refunds can even be considered. If the funding ratio falls below 130%, catch-up contributions are charged and the rights of actives and retirees are reduced if necessary. It should be noted, however, that even in a moderate inflation environment a funding ratio of 130% leaves little room for indexation. The exact content of the pension contract thus determines the degree of – simultaneous and sequential – solidarity between the various groups of beneficiaries.
SOCIAL SOLIDARITY This form of solidarity concerns the participants’ involvement in and responsibility for the proper working of the national economy. This commitment to society as a whole is the most abstract form of solidarity: it is
2 Solidarities in collective pension schemes
23
the furthest removed from people’s personal lives and therefore rests on the most fragile consensus. Increasing contributions has much greater macro-economic consequences than reducing the indexation of pension payments and accrued rights of the working population (CPB, 2004). However, pension funds give little or no attention to the macro-economic consequences when establishing their pension policy. Therefore, it is up to the supervisor to ensure that the funds operate in the best interests of the national economy. To this end, its likeliest course of action is to impose restrictions on the use of the contribution instrument, whilst leaving room for reducing or even stopping indexation. When serious shortfalls occur, accrued rights may even be reduced. In solidarity terms, this policy line spares younger employees, new entrants in particular, whilst demanding a greater sacrifice from retirees and older employees. Alternatively, the supervisor may opt for a policy where young people pay the level premium whilst accruing little or no pension rights. In this case, the contributions largely benefit the elderly.
2.3
Setting the standard for solidarity
This section formulates a standard for the desirable and undesirable degree of solidarity in the supplementary pension system.6 This standard leads to certain conclusions, two of which are discussed here. Some degree of personal preference is of course inevitable here.
STANDARD The basic premise is that every generation of new entrants should finance its own pension provision (including an agreed solidarity contribution) on an ex-ante basis. In other words, every generation must be self-supporting by
6
Views on solidarity also differ in the pension world. Here are two quotations of two (ex-) chairmen of the Association of Industry-Wide Pension Funds and two ex-directors of the PGGM pension fund and the Philips pension fund. The first quotation is of G. Beuker and J. Wennekus: ‘…solidarity between generations has a wealth-enhancing effect. A pension provision financed on a level premium basis that is accessible to everyone – which is what we want – is only possible in a system with solidarity between the various risk groups.’ (Aalberse et al., 2004). The second quotation is from D.J. de Beus and D. Snijders: ‘… pension funds should not bury their heads in the sand by continuing to brandish a term like solidarity. The day will come when participants simply won’t buy that any more.’ (C. Petersen, 2002).
24
J.B. Kuné
always paying the cost-price contribution in advance. But how strictly or loosely should this principle be put into practice? Or, in other words, how much solidarity between successive generations is reasonable, responsible and desirable as well as acceptable to all those involved? This is one of the key questions that must be carefully considered. The answer will provide valuable input for making clear and unambiguous ex-ante arrangements.7 Note, however, that ex-post discrepancies are inevitable: reality is inherently unpredictable and surpluses and shortfalls will always occur.
CONCLUSIONS The first conclusion is that there should be no hidden re-distributing solidarities within the supplementary pension schemes. Everyone must know what solidarities are involved and how much they cost. The participants can then form a standpoint as to which solidarities are necessary or unnecessary, and to what extent. They thus, knowingly and willingly, enter into a commitment to pay the price of the solidarities that they collectively wish to preserve. Given that pensions are basically deferred income, it is reasonable to argue that they should be redistributed in the same way as the current income. If this is combined with the premise that income policy is a matter for politics and government, then it follows that neither pension institutions nor other social partners (e.g. employers or trade unions) should have any influence in this case. Income policy is conducted through taxation, subsidies and so on. Consequently, any solidarity elements in pension schemes should not be open-ended (and therefore uncontrollable) but closed-ended with pre-determined quantitative floors and ceilings. A second conclusion is that any surpluses at a pension institution basically belong to, or are attributable to, those participants who paid for these surpluses. Therefore surpluses only benefit these participants.8 New entrants have no legal or moral claim to an existing surplus. Conversely, shortfalls are the responsibility of the participants who caused or allowed these deficits to arise during their participation and should not burden new entrants. In other words, new entrants should neither benefit from
7
You could for instance agree that a generation that has made a solidarity sacrifice over a certain period of time will be the first in line for the next solidarity bonus in a subsequent period.
8
Part of the buffers was formed by ex-participants who are no longer alive. That part can go to the benefit of the entire collectivity, including the new entrants.
2 Solidarities in collective pension schemes
25
existing surpluses nor be burdened with existing shortfalls at the time of entering the scheme. However, as they start accruing rights in the scheme, they must share in the costs and benefits of shortfalls and surpluses on a pro rata basis. Having said this, there are circumstances in which intergenerational settlement of increased or reduced contribution rates as well as surpluses and shortfalls may make sense. But this must always be based on sound arguments and consensual agreement between both parties. This agreement is codified in the pension contract. Future generations clearly have no say in the arrangements made by earlier generations. Consequently, generous promises may be made to existing participants at the expense of future participants. An independent party can protect the rights of future generations. This may be a role for the DNB (Dutch central bank) in its capacity as supervisor of the pension sector. The arrangements must obviously be reasonable and fair. But these criteria are notoriously hard to quantify. Good information, sound analysis and equitable judgement are therefore crucial to arrive at a fair deal for both current and future generations. Nothing must be left to incomplete information or chance.
AGE-RELATED PENSION ACCRUAL AND CONTRIBUTION PAYMENTS In the current system, young and old pay the same percentage of their income (say 14%) to the pension fund and they accrue pension rights that are the same percentage of their income (say 2%), in spite of the fact that the contribution of the younger participant will generate returns over a much longer period. This so-called “uniform contribution and accrual system” is in place to limit competition in the labour market between the young and old. Would differentiation in contribution levels be a good idea? In other words, is the solidarity between young and old implicit in equal contribution levels always justified? The following case is instructive. A 45-year-old participant A leaves his employer’s pension scheme (as active employee) to pursue a self-employed activity. During the previous 20 years this participant paid a relatively high pension premium. As a selfemployed person he will in the future continue saving for an individual pension by paying a cost-price contribution (which is relatively high compared to the level premium). All in all, A pays a relatively high amount of money for his pension. Compare this with participant B who, at the age of 45, decides to bid farewell to his 20-year freelance career to take a steady job and joins his employer’s pension scheme. In the previous 20 years this
26
J.B. Kuné
participant paid a cost-price pension contribution (which was relatively low compared to the level premium) and in the coming 20 years will be able to benefit from the relatively low level premium that he is required to pay. All in all, B pays relatively little for his pension. This raises the following question: is it reasonable and fair that participant A receives a certain refund on leaving the scheme while participant B starts paying more than the uniform premium on entering the scheme? The above gives rise to two issues for debate: (1) the age or intergenerational solidarity inherent in the level premium and (2) the pension accrual rate and contribution level in relation to age. The central question concerns the desirability of an age-related level premium or an age-related pension accrual rate. It looks like the current system is on a collision course with the Equal Treatment Act.9
SOLIDARITY CONTRIBUTION In a situation of severe or prolonged underfunding, younger participants may be prepared to make a sacrifice to help out older employees and retirees, but only up to a certain degree. They, after all, also have their own future to think about and want to be sure their contributions are sufficient to finance their own future pension income. Tentatively and intuitively we suggest that the younger generation might be prepared to pay an extra solidarity contribution of 3 to 5% to lend their elders a helping hand. Anything above that level would probably be stretching the system beyond its limits.
Solidarity sacrifice Assuming a negative shock on capital markets, the CPB Netherlands Bureau for Economic Policy Analysis has calculated that in an indexed average-pay scheme some 16% of this shock would impact on new participants while 84% would be absorbed by the present participants (CPB, 2004). Of the latter, the active population would pay the lion’s share of 78% and the retirees only 6%. The solidarity sacrifice of the new participants is also limited. In a DC scheme, future entrants are obviously not charged for existing shortfalls. The shortfall arising from a shock on capital markets is borne proportionally by the generations involved: 75% by the actives and 25% by the retirees.
9
The aspiration for equal treatment (gender, age, etc.) frustrates the process of eliminating undesirable forms of solidarity from pension schemes.
2 Solidarities in collective pension schemes
27
There are good arguments for persuading the younger generation to accept paying a relatively high solidarity contribution. In the first place, they too will be old one day and will then benefit from the solidarity mechanism. Secondly, they will benefit from a huge production potential and infrastructure that was created by many earlier generations. The future is being presented to them on a platter of gold. Thirdly, at least a part of the higher pension income of the elderly today will ultimately come to them in the form of gifts and inheritances. So perhaps there is no conflict of interest after all! These arguments provide a foundation for maintaining the solidarity mechanisms. Opinions may differ as to the acceptable level of the solidarity contribution and the rules to be observed in times of relative hardship or prosperity. But these issues can no doubt be resolved provided the parties are willing to engage in an open and constructive discussion.
VALUE TRANSFERS An employee changing jobs can decide to have his accrued pension rights transferred from the old to the new pension provider. The way in which the value transfer is determined has significant solidarity implications.
Dilemmas To illustrate the solidarity dilemma, let’s assume a new entrant whose pension rights (based on 20 years of service) are transferred. He joins a fund with a very large surplus, say a funding ratio of 200%. Should this high funding ratio also apply to the new entrant who will probably also benefit from a low contribution rate? And conversely, suppose he joins a fund with a funding ratio of only 50%. Will the value of his transferred pension rights be halved at one stroke? And will he also be required to pay more than the cost-price contribution into the bargain? Current practice seems to be developing in the following direction. If one of the two funds has a funding ratio lower than 105%, then no value transfer takes place. In all other cases, outgoing value transfers are based on the lower of the two funding ratios. As a result, neither incoming nor outgoing transfers affect the financial position of the ‘poorest’ fund. The financial position of the ‘richest’ fund, hence, improves with an outgoing transfer and deteriorates with an incoming transfer.
28
J.B. Kuné
2.4
Conclusion
The issue of redistributing or subsidising solidarity in supplementary pension schemes needs to be treated with great care. Valid arguments can be put forward for risk sharing. The creation of consensual support for diverse forms of subsidising solidarity within supplementary pension schemes is all the more crucial if the abolition of mandatory participation makes pension exit a real possibility. Put differently, limits must be set on the extent to which new entrants can be charged for the shortfalls that have arisen before their time. To maintain the required level of acceptance among all participants, it is advisable to check for each type of solidarity where desirable solidarity ends and undesirable solidarity starts.
Literature Aalberse, B. et al., ‘De waarde van solidariteit; spanning in de tweede pijler van het pensioengebouw’, SISWO Cahiers Sociale Wetenschappen en Beleid, no. 6, Amsterdam, 2004. Aarts, L.M.J. and P.R. de Jong, Op zoek naar nieuwe collectiviteiten; sociale zekerheid tussen prikkels en solidariteit, The Hague: Elsevier Bedrijfsinformatie, 1999. Beer, P. de, De ontwikkeling van de solidariteit in Nederland, unpublished paper UvA/AIAS, 2006. Beltzer, R. and R. Biezeveld, De pensioenvoorziening als bindmiddel, socialecohesie en de organisatie van pensioen in Nederland, Amsterdam: Aksant, 2004. Beus, D.J. de and D. Snijders, ‘Toekomstvisie pensioenfondsen’, in: Bestuur en management van pensioenen, regelingen, beleggingen en uitvoering (C. Petersen editor), The Hague: SDU, 2002. Boer, J. de et al., Impact van flexibilisering op solidariteit, Woerden: Actuarieel Genootschap/ASIP, 2002. CPB, Solidariteit, keuzevrijheid en transparantie, The Hague, 2000. CPB, Naar een schokbestendig pensioenstelsel, verkenning van enkele beleidsopties op pensioengebied, Document no. 67, The Hague, October 2004. Davidson, M.D., ‘Discontovoet voor klimaatschade behoeft politieke keuze’, Economisch-Statistische Berichten, 89, 25 June 2004.
2 Solidarities in collective pension schemes
29
Entzinger H. and G. Engbersen, ‘Immigratie en solidariteit’, in: W. Arts et al. (ed.), Verzorgingsstaat vaar wel, Van Gorcum, Assen, 2004. Kuné, J.B., Op weg naar één nationale pensioeninstelling, oration text 4 June 2004, Amsterdam: Vossiuspers UvA. Oorschot, W. van et al., Solidair of selectief, een evaluatie van toepassing van het selectieve-marktmodel in de sociale zekerheid, Deventer/Zeist: Kluwer/ Sovac, 1996. Oorschot, W. van, ‘Solidair en collectief of marktgericht en selectief?, Nederlanders over sociale zekerheid’, Beleid en Maatschappij, 24, 1997. Petersen, C., Bestuur en management van pensioenen, regelingen, beleggingen en uitvoering, The Hague: SDU, 2002. Teulings, C., ‘Solidariteit en uitsluiting, de keerzijden van een zelfde medaille’, in: Engbersen, G. and R. Gabriëls (eds.), Sferen van integratie; naar een gedifferentieerd allochtonenbeleid, Amsterdam: Boom, 1985. Vorselen, L. van, Solidariteit en pensioen, denkbeelden over een solidair ouderdomspensioen, Deventer/Zeist: Kluwer/Sovac, 1993. WRR (Scientific Council for Government Policy), Ouderen voor ouderen, The Hague: SDU, 1992. WRR (Scientific Council for Government Policy), Generatiebewust beleid, The Hague: SDU, 2000.
Appendix: Different types of solidarity within supplementary pension schemes (‘solidarities’) Many forms of real solidarity and quasi-solidarity can be identified. Some are open and visible; others hidden and rather difficult to detect. The various forms are summarised below. 1.
Risk solidarity (risk sharing): this principle underlies all insurance contracts, whether individual or collective. The participants bear one another’s burden insofar as these arise from bad risks. As opposed to this, there are numerous subsidising solidarities that typically result from level-premium financing.
2.
Solidarity between workers and retirees (in any given time period): this mainly concerns the exposure to real investment risk. The pension provision is usually determined on the basis of an actuarial interest rate of about 3%. This means that if the real return on pension capital
30
J.B. Kuné
– nominal return less the indexation of pension rights and payments – is greater than the actuarial discount rate, then the population of workers will benefit in the form of a lower contribution. The greater this differential, the more the retirees constitute a gold mine for the actives. Conversely, if the real return on assets is below the actuarial interest rate, the actives must make a sacrifice in the form of a higher contribution rate to maintain the pension payments at the desired level. 3.
Younger versus older employees: the price of a euro pension income is considerably higher for an older employee than for a younger worker, even though they all pay the same premium. The actuarially fair premium (with an actuarial interest rate of 4%) for a euro pension is on average about 1¼% of the pensionable salary for a 24-year-old. With a level premium of 9% this therefore entails an implicit additional charge of 7¾%. For a 64-year-old the actuarially fair premium is about 17½% and he therefore receives an implicit subsidy of 8½%. As with the AOW state pension scheme, this solidarity is based on a hidden or implicit social contract where the younger generation trusts that the next generation of young people will do the same for them when they are old. In the case of stable relations in terms of age distribution and income position, this social contract is not at risk. But the situation changes when the number and age of active participants is no longer stable. In a rapidly shrinking sector, the young people of yesterday have paid high premiums only to find out upon retirement age that there are no longer sufficient young people to pay the major part of the cost-price contribution for them. In this respect large collectivities are at an advantage compared to smaller collectivities.
4.
Individual companies within a certain sector: insofar as the profile of the participants differs from the (average) profile of the total collectivity, mandatory participation in the fund will lead to subsidising cash flows. This effect occurs in all industry-wide pension schemes.
5.
Actives versus disabled employees: disabled employees usually continue to accrue pension rights in the same manner as actives (on the basis of the indexed income at the time of becoming disabled or otherwise unfit for work) but no longer pay contributions. These pension accrual costs are included in the level premium paid by the actives.
6.
Divergent career paths: these result in what is frequently called perverse solidarity, i.e. those who make a relatively steep career, particularly towards the end of their working life (known as ‘pension promotions’), benefit considerably from participants with a flatter career. This income redistribution effect occurs irrespective of the level of in-
2 Solidarities in collective pension schemes
31
come. In recent years many schemes have been transformed from final pay to average pay schemes. 7.
Socio-economic class: the difference in life expectancy between persons of diverse socio-economic classes is substantial and can be as high as five years. This causes a redistribution of funds from lower to higher socio-economic classes. This too can be considered as a form of perverse solidarity.
8.
Gender solidarity: women aged 65 are expected to live 5 to 6 years longer than men, while all pay the same level premium. Women also run a higher risk of occupational disability. One special form of solidarity concerns the option of exchanging (usually towards the age of 65) the surviving dependant’s pension for a higher or earlier old-age pension. The principle of equal rights demands gender-neutral exchange factors, leading to a weighted average of male and female factors that generally work out favourably for women and less favourably for men. This means again solidarity between men and women, which is further reinforced by the selection effect.
9.
Solidarity related to marital status: participants with marital or equivalent status benefit from the participation of participants without a partner. The level premium paid by the latter also includes the risk premium of a surviving dependant’s pension in the event of death prior to retirement. The aforementioned exchange option mitigates this form of solidarity to a certain extent. Analogously there is solidarity between participants with dependant children and those without.
10.
Solidarity between the various types of pension occurring in a pension scheme: this occurs when the actual costs of the aforementioned types of pension differ from the expected costs and the difference is credited or charged to the (rest of the) collectivity. Where the pension package is indivisible (diffuse), there is by definition no advantage or disadvantage for any of the participants. However if we look at a pension package on the basis of its different components, the subsidising effect can be seen. Subsidies occur to the extent that distinct groups of participants (at sector and subsector level) make different use of the different components of the scheme without these being separately priced and charged.
11.
Solidarity between collective pension products and supplementary optional products: this occurs where supplementary optional products are not priced strictly according to the self-financing or self-supporting principle. No scheme or scheme component is ever entirely self-supporting.
32
J.B. Kuné
12.
Buffer solidarity: the presence of buffer assets of, say, 30% or more of the provision for accrued rights is a major source of financing. This makes it possible to keep the contribution relatively low and gives participants a disincentive to exit the scheme – assuming they are free to do so – and makes it very attractive for prospective participants to enter the scheme rather than make individual arrangements. Clearly, the participants (and former participants) who accumulated the buffer have made a sacrifice, even when the creation of the buffer took place almost imperceptibly (as was the case with many funds in the nineties).
13.
Solidarity resulting from the actual operation of the financing system: the financing system brings about a certain redistribution over time of the costs – and thus the costs for the participants present in any given time period – of the expected flows of future payments. Depending on the manner in which the distribution of costs over time is realised, some generations will contribute more or less than they receive during their lifetime. They benefit from or pay for others. On an ex-post basis the contributions paid over the lifetime period of a generation will usually not be exactly equal to the total amount of the payments that the generation receives. One crucial factor (with a strong influence on the outcome) is the rate of time preference, the discount rate used to calculate the present value of future cash flows. The method of financing a pension scheme strongly influences the premium differential and thus the extent of the value transfers and solidarity between the generations. It follows that contribution stabilisation over a certain period is not neutral for the extent of redistribution over generations, viewed from both a simultaneous and sequential perspective.
14.
Solidarity generated by and from a pension package with options: this basically involves a pension menu system. Financial neutrality is usually impossible when participants are able to choose ‘a bit more of the one and a bit less of the other’. Things become even more complicated with flexible reward packages that permit all sorts of exchange options without it being clear which components belong to the pensionable salary. The process of selection and moral hazard will inevitably lead to cost increases. The solidarity here will therefore often be of a perverse nature. But even an (ostensibly innocent) extension of the working week (as arranged in the Collective Labour Agreement) from e.g. 36 hours to 38 or 40 hours at the end of the employee’s working life will – assuming a corresponding increase in income – results in a considerable cost increase in a final pay system.
3
Solidarity: who cares?
P.P.T. Jeurissen and F.B.M. Sanders This contribution describes the principal forms of solidarity in the healthcare sector. We discuss the concept of solidarity and its diverse roles as well as the design of the existing solidarity framework and trends for the future. Solidarity in healthcare is under pressure: the costs are rising and the distribution of solidarity transfers is becoming increasingly uneven, socio-cultural trends are sending out mixed signals and many think it is fair to ask people with unhealthy lifestyles to pay more. However, a fully funded system is less suited to healthcare than to supplementary pensions. In healthcare, more so than in the pension sector, solidarity is nurtured by feelings of community and justice. At the same time, egalitarian outcomes are increasingly difficult to achieve due to the evermore uneven distribution of the health cost burden, the enormous supply of healthcare products and the large mutual differences in the production process.
3.1
Introduction
This chapter contains a discussion of the position of solidarity in the health sector. We show how solidarity is given shape in healthcare and why solidarity is pivotal in achieving equitable and efficient healthcare provision. The outcomes of this analysis are compared with the solidarity arrangements in the pension sector. The health sector and the pension sectors each have their own dedicated institutions with little mutual overlap. One common factor, however, is the aim to reduce ‘risk’ and ‘uncertainty’. Collective pensions mitigate our fears of insufficient income in old age, while healthcare decreases our concerns over treatment and care when sick – a situation that often coincides with old age. Healthcare and pension are thus the mainstays of existential security for the elderly: healthcare in the form of transfers in kind and pensions in the form of transfers in cash. This contribution starts with a reflection on the meaning of the term solidarity. What do we understand by the term solidarity and what is its
34
P.P.T. Jeurissen and F.B.M. Sanders
relationship with collectivity (section 3.2)? Next we consider how solidarity is given shape in the financing and implementation of healthcare (section 3.3). Section 3.4 shows which aspects of solidarity are most under fire and explains why. Most reform proposals concentrate on a combination of individual responsibility and system incentives; both of which are aimed at making business processes more efficient. In this respect healthcare (the no-claim rebate, regulated competition) is no different from the pension sector (growing trend of defined contribution schemes). We describe what issues play a role in this connection in the healthcare sector and discuss whether individual schemes also have a role to play (section 3.5). This contribution ends with a brief comparison of the most important characteristics distinguishing the insurance-based healthcare market from the collective pension sector (section 3.6).
3.2
Solidarity
Solidarity is a concept which we almost all endorse, but which is also subject to diverse interpretations. Solidarity suggests a sense of community and the willingness to bear the consequences of community membership – it implies a certain bond. The classic sociologists Durkheim and Weber saw solidarity as the social cohesion arising from a sense of shared fate between individuals and groups. In doing so, they made a distinction between culture-based solidarity which sprang from (shared) identity and utility-based solidarity (Widdershoven, 2005). Schuyt posits that: ‘Solidarity, as a social phenomenon, means sharing of feelings, interests, risks and responsibilities’ (Schuyt, 1998). He thus adopts a slightly more specific approach, with feelings and responsibility referring to the cultural or identity-based solidarity and interests and risks encompassing the utility aspect. Both definitions are descriptive rather than normative: they place solidarity in a sociocultural and economic context, but give no indication as to whether, and how much, solidarity is desirable. A tension exists between solidarity described in empirical terms and solidarity interpreted in normative terms of right and wrong (Verstraeten, 2005). This differentiation is of fairly recent date. Originally the normative approach prevailed: solidarity, quite simply, was the right thing to do (Verburg and Ter Meulen, 2005). More recently, however, this moral stance has been challenged by empirical arguments. In the current debate, the proponents of welfare retrenchment tend to come up with empirical arguments (rising costs, eroding support), while the opponents adhere to normative principles (‘you either have 100% solidarity or no solidarity at
3 Solidarity: who cares?
35
all’). After briefly discussing the most important normative views on solidarity, we will turn to the empirical views that may fundamentally alter the solidarity landscape.
HOW MUCH SOLIDARITY? The political philosopher John Rawls tried to define how much solidarity a society needs. In his book entitled ‘A theory of justice’, he describes a method for arriving at a fair distribution of goods (Rawls, 1972). In a hypothetical original position, citizens agree on a social contract through a process of negotiation. These negotiations take place behind a ‘veil of ignorance’. According to Rawls, the participants will follow rules that permit a fair distribution of resources precisely because of their lack of relevant knowledge of their position in society. In this context Rawls presumes a situation of moderate scarcity: the living conditions of the least well-off can be improved without causing any severe disadvantage to the betteroff. Conflict between different groups is thus avoided. In reality, this situation existed during the establishment of the welfare state when strong economic growth diminished scarcity and greatly increased everyone’s personal purchasing power, despite rising taxation. Rawls distils two fundamental rules of distribution from this original position. First, everyone has equal rights to the most extensive basic freedoms, provided that others enjoy equal freedom. Second, social and economic inequalities are only acceptable if these demonstrably (also) serve to benefit the least privileged and if there is an open opportunity structure: people of equal ability have an equal chance of achieving a certain position (Verburg and Ter Meulen, 2005). Though such reasoning underpins the welfare state, it does not lead to entirely egalitarian distribution as the freedom principle has the greatest priority. Inequality could, for instance, promote the prosperity of the least well-off by stimulating productivity improvements that are in everybody’s interest. The Rawlsian principles also entail that the current generation should not squander the wealth of future generations ‘and require one generation to save for the welfare of future generations’ (Kukathas and Pettit, 1990). Rawls’ theory assumes rational behaviour: uncertainty over their personal position combined with risk aversion ‘compels’ citizens to show a substantial degree of solidarity. Rawls’ approach to solidarity is thus essentially rational. By contrast, the culture- and identity-based community spirit of Durkheim and Weber and the feelings and responsibilities of Schuyt attach a more open character to solidarity. Solidarity, in their view, is not the utilitarian outcome of a
36
P.P.T. Jeurissen and F.B.M. Sanders
social contract based on a ‘veil of ignorance’ but a measure of the moral quality of interhuman relationships in society. This tradition is deeply rooted in history and resonates with the Christian concept of charity. Notions of personal or voluntary charity are combined with a religious or ideological instruction to do good, such as the Biblical duty to care for the poor, orphans and travellers (Buijsen, 2005). This form of solidarity is also referred to as ‘vertical’, ‘one-sided’ or ‘warm’ solidarity. Over the years these views on solidarity as an individual moral obligation acquired a more rational dimension as arguments such as enlightened self-interest started to gain currency. In the literature this process is frequently illustrated by the insurance world, where ‘cold’ utilitarian interests progressively displaced the ‘warm’ sense of social duty in the solidarity framework (Widdershoven, 2005). In the healthcare sector this development is evident from the fact that many now perceive health as an enshrined ‘right’ (Starr, 1982). This evolution from subjective duty to objective right has strongly influenced the way solidarity is organised in our society of today.
SOLIDARITY VERSUS COLLECTIVITY Weighing up the costs and benefits of collective pension schemes is the core theme of this book. What is the relationship between solidarity and collectivity? Insurance economics mainly perceives collectivity as a means of achieving economies of scale by spreading individual risks. In sociology, however, a collectivity is usually defined as a cohesive group with a shared culture, values and experiences. Society comprises a wide array of different collectivities: civilians, the insured, patients, professionals and pension fund members, to mention but a few. The interests of these collectivities can conflict in all sorts of institutions. Health insurers, for instance, have dealings with the insured, who want the lowest premium, and patients who want the best (read: most expensive) healthcare. In the course of their lives individuals may switch between the roles of insured and patient any number of times. In the pension sector the switch between collectivities only occurs once, namely on retirement. Consequently, collective interests here are more clear-cut and less muddled. The emotional sense of belonging to a collectivity promotes support for solidarity. But so does the rational awareness that individual risks can be shared in the collectivity. Insurance schemes where people with an equal risk pay the same premium are essentially rational. More emotional types of solidarity include helping sick relatives (informal care) or voluntary com-
3 Solidarity: who cares?
37
munity work. The state, acting in its capacity as over-arching body, orchestrates these diverse collectivity mechanisms and – through its legislative monopoly – is able to enforce solidarity where appropriate: the young are obliged to contribute towards the state pension (AOW) and pre-pension, high earners pay welfare benefits for low earners and the healthy help to pay the costs of the sick. In practice, there is a considerable degree of overlap between the twin concepts of solidarity and collectivity: ‘no collectivity, no solidarity’ and ‘no solidarity, no collectivity’. In insurance economics the concepts of solidarity and collectivity are separated from each other via the equivalence principle (equal risks pay equal premiums). However, even in the commercial market, few products are entirely without value transfers (“good risks” pay for “bad risks”) and operate fully according to the equivalence principle. This is partly due to economic market imperfections, such as moral hazard, adverse selection and actuarial inaccuracies. But solidarity considerations also enter into the equation. Formerly, for instance, sickness funds rarely excluded nonpayers from their health schemes (Widdershoven, 2005) and private insurers hardly ever charged totally risk-based premiums (Schut, 1995). When these ‘spontaneous’ mechanisms enhancing solidarity came under pressure, the government took regulatory action to ensure that the system continued to be supported by a combination of economic (collectivity-based) and social (solidarity-based) motives.
3.3
Health solidarity
The most important solidarity concepts in healthcare are income solidarity and subsidising solidarity.1 When high earners make a more than proportionate contribution to the financing of health costs, we speak of income solidarity. Subsidising solidarity assumes that everyone pays the same premium, even though some predictably represent a higher risk than others (community rating). Those with a low health risk thus help to cover the costs of those with a higher health risk. This is achieved through solidarity transfers between the diverse risk groups – which obviously do not automatically come about in a free market. The old have a higher health bill than the young, so subsidising solidarity implicitly contains an intergen-
1
As mentioned in Chapter 1, footnote 2, the concept of risk is used slightly differently in the (health) insurance industry. As this book focuses on the pension sector, we prefer using the term as used in the pension industry.
38
P.P.T. Jeurissen and F.B.M. Sanders
erational solidarity component – as will become increasingly evident in our ageing society. Our main focus so far has been on the financial side of solidarity. In the healthcare sector, however, solidarity is also reflected in the actual provision of the services; a healthcare system that treats all patients according to medical urgency and need is considered to show more solidarity than a system which sets different priorities. We refer to this as solidarity in healthcare outcomes (RVZ, 2005). According to this solidarity criterion, the Dutch health system, due to need related referrals by general practitioners, compares favourably with that of other countries (Van Doorslaer and Masseria, 2004).
THREE COMPARTMENTS The health insurance system consists of three different compartments, each with its own solidarity levels and institutions. These are listed and tentatively compared with pensions in Table 1. Pension solidarity was extensively discussed in the previous chapter in this book. The Exceptional Medical Expenses Act (AWBZ) provides a national social insurance for uninsurable risks as well as residential care. This scheme, like the state pension (AOW), is funded by employees who pay income-dependent premiums up to a certain threshold. The AWBZ is a pay-as-you-go scheme in which subsidising solidarity plays a significant role, particularly in relation to the physically and mentally handicapped. In long term care the distribution of risk is more age-related. As with the state pension (AOW), ageing will put pressure on the financial robustness of pay-as-you-go schemes; intergenerational solidarity is thus also an important issue in the healthcare sector. Similarly, subsidising solidarity occurs in co-payment schemes. In the ABWZ these are income-dependent, so that the degree of subsidising solidarity is smaller towards high-earning patients than to low-income patients. In the new private obligatory health insurance scheme with publicly regulated conditions, subsidising solidarity plays a central role. The insured pay a nominal risk-independent premium; employers pay a wage related contribution that covers fifty percent of the total costs. Everyone pays virtually the same price for the basic benefit package, regardless of their health status and insurers are not allowed to reject applicants. But the principle of subsidising solidarity does not apply in full. There is a noclaim rebate of € 255 per insured adult; those who incur no health costs are refunded this amount. In addition, the insured can opt for a deductible of € 500 at maximum.
3 Solidarity: who cares?
39
Table 1. Solidarity in the financing of healthcare and pensions
Collective
Risk solidarity
Subsidising solidarity
Income solidarity
Intergenerational solidarity
AWBZ (Exceptional Medical Expenses Act)
o
++
++
+++
State pension (AOW)
o
+
++
++
Mandatory Insurance
o
+++
+
++
Collective pensions
+
+
o
o
Supplementary Health Insurance
+
++
o
+
Private pensions
++
+
o
o
Regulated
Free market
Insurers are compensated for bad risks from a central fund. This takes place both ex-ante (risk adjustment) and ex-post (risk sharing), so that the premium discounts given to people with such a deductible as well as general premium differences are limited. The fund is filled with wagedependent employer premiums; lower-income earners also receive a health allowance. A certain degree of income solidarity has thus been maintained. Insurers may award collectivities a discount up to a maximum of 10% of the premium, but this is really a high-volume discount and has little bearing on the expected cost of claims. As older people are more frequently ill and generate more costs, intergenerational solidarity is also amply present in the new private health insurance scheme. In the ‘free market’ people can take out supplementary insurance. There is no income solidarity here. So insurers are allowed to apply experience based premiums and refuse applicants. Such rejections are quite common with supplementary dental insurance. Sometimes eligibility for supplementary insurance is subject to age limits. But even this segment displays a fair degree of subsidising solidarity. This is related to the large demand for supplementary policies (95%), which limits adverse selection effects and establishes an implicit ‘linkage’ between this insurance and the basic obligatory policy. Consequently, competition between insurers on these policies is limited. This is also reflected in the fact that the old supplementary health insurance schemes of the sickness funds fetched a high gross margin of 20% (RVZ, 2005).
40
P.P.T. Jeurissen and F.B.M. Sanders
3.4
Solidarity under pressure?
Though health solidarity can still rely on broad social support, it is no longer unchallenged: the no-claim rebate included in the new private obligatory health insurance scheme and plans for a radical overhaul of the Exceptional Medical Expenses Act underline this point. Pressure to make choices will no doubt continue to grow. Both financial-economic and socio-cultural aspects play a role in this respect. Within the pension sector, inflation, increased longevity and the return on investments are the main uncertainties. The first two also play a role in health insurance. Investment performance is of less significance here, though the need to meet solvency requirements remains a challenge to some of the private health insurers. One specific characteristic of the healthcare sector is the substantial uncertainty over the financial impact of advancing medical technology. In the past, technological innovation (and socio-cultural trends) accounted for almost half of health expenditure growth (Spaendonck and Douven, 2001). Medical advances drive up costs by creating treatments for new (older) target groups. In contrast with the pension sector, healthcare is entirely financed on a pay-as-you-go basis. If premiums continue to be undifferentiated by risk, ageing will lead to larger transfers between generations.
INCREASINGLY SKEWED DISTRIBUTION OF GROWING HEALTH EXPENDITURES With the number of old people continuing to grow, health expenditure is set to soar. The economic scenarios of the CPB Netherlands Bureau for Economic Policy Analysis estimate that about 20% of GDP growth will be spent on healthcare in the coming decades (RVZ, 2005). If this growth is to be entirely collectively funded, with health expenditure accounting for a constant share of the economy, this percentage will double. In this case healthcare will swallow up a large portion of the extra budget, unless society is prepared to accept higher taxation, lower purchasing power or substantial spending cuts in other policy spheres. This has never been necessary in the past. The competition for a slice of the budget will intensify; the Rawlsian assumption of moderate scarcity, which helps to ensure a fair distribution of resources, no longer holds water. Healthcare financing is now overwhelmingly based on the principle of subsidising solidarity (EIM, 2002). A combination of rising health spending and an increasingly skewed distribution of health costs has greatly increased the transfers between risk groups in the past decades; the costs of the 10% most expensive insured persons has jumped from 43% of total
3 Solidarity: who cares?
41
curative expenditure in 1953 to 70% in 2002 (Cutler and Meara, 2002). In the sphere of exceptional medical expenses this distribution is even more lop-sided: the costs of residents of nursing and care homes, handicapped institutions and psychiatric care provisions (1.6% of the population) amount to about 74% of the total exceptional medical care expenses. This evermore skewed cost distribution is largely due to advancing technology (which mainly benefits limited groups), ageing and several sociocultural trends, such as unhealthy lifestyles, patient empowerment and medicalisation of inconvenience and distress. These cost-increasing factors can be easily assimilated within the general healthcare framework because the basic health package is formulated in general terms (that which is regarded as normal in professional circles) rather than restrictive terms. This trend towards subsidising solidarity is unlikely to weaken. Far from it, in fact. Due to a combination of factors, these transfers must continue to grow at an accelerating rate in the coming decades to maintain subsidising solidarity in its current unlimited form. We have already mentioned advancing technology as one reason for this process. In this context the National Institute for Public Health and the Environment (RIVM) also mentions the impact of epidemiological trends: the occurrence of expensive chronic illnesses such as depression and asthma will continue to rise rapidly until 2020. The higher incidence of age-related illnesses (e.g. dementia and stroke) compared to disorders that are more prevalent among younger age groups (e.g. as mental handicaps and pregnancy complications) means that subsidising solidarity will show a stronger correlation with intergenerational solidarity (RIVM, 2005). The ageing affluent post-war baby boom generation will become an evermore voracious net recipient of intergenerational transfers. Moreover, this generation will be more demanding in terms of quality than the current generation of retirees and that will further accelerate health expenditures. Young people, by contrast, will be confronted with wealth stagnation in the coming decades (McKinsey, 2005). So, up to what quality level is this generation willing and able to continue financing solidarity? One possible solution to this growing problem would be to maximise the collective contribution towards health treatments at a certain amount per quality adjusted life year (QUALY) (RVZ, 2006).
LESS COMMUNITY, MORE RECIPROCITY? The scope of this contribution does not permit a detailed account of the impact of socio-cultural change on health solidarity. Verburg and Ter Meulen note that the sustainability of solidarity hinges on more factors
42
P.P.T. Jeurissen and F.B.M. Sanders
than the increasingly skewed distribution between the contributions to and use of healthcare that we highlighted above. They point to an ongoing process of differentiation and individualisation that may erode our collective consciousness and sense of society. As a consequence, the limits of solidarity will be more sharply drawn and claims will be increasingly tested against criteria of equality, reciprocity and personal responsibility (Verburg and Ter Meulen, 2005). Not everyone agrees that individualisation is undermining the existing base of support for solidarity. Some challenge the view that solidarity is necessarily at odds with individualisation (De Beer, 2005). But it cannot be disputed that the financing of healthcare has increasingly become the domain of insurers. The launch in the Netherlands of private health insurance to replace the former social health insurance scheme as the dominant system corroborates this point – and suggests that reciprocity and collectivity may be in the ascendancy at the expense of solidarity. Civic acceptance Subsidising solidarity in the healthcare sector used to draw support from the notion that illness is a question of bad luck and that nobody goes to hospital or a nursing home for fun – the underlying assumption being that opportunistic (i.e. increased-risk) behaviour is rare. This also explains why, from a historical perspective, under the old mutual sickness funds monetary benefits such as sick pay (where opportunistic behaviour was assumed to exist) were paid out much more reluctantly than medical claims (Widdershoven, 2005). Scientific research has shown that a high level of health solidarity exists as long as these assumptions hold sway, particularly in highly cohesive societies where strong affinity is felt with the patient (e.g. through close social relationships). But this sympathy evaporates when unhealthy behaviour is perceived to be the cause of ill health. When asked about the right to collectively insured healthcare and the justification of co-payments, respondents tend to weigh lifestyle factors more heavily. People with an unhealthy lifestyle and high earners are deemed less eligible for collective financing of treatment for certain diseases (Hansen, Arts and Muffels, 2005). Society attaches conditions to solidarity: high earners and those with risky lifestyles should pay more. So unlimited solidarity is certainly not to be taken for granted, as is also evident from research of Bongers et al.; 50% of the insured are prepared to pay a higher contribution to maintain solidarity, but 39% are no longer prepared to do this (RVZ, 2006).
3 Solidarity: who cares?
43
The end of the ‘veil of ignorance’? This trend is compounded by an entirely different phenomenon: new scientific insights into the causes of disease, including genetic factors and behaviour, and the use of healthcare increase the capacity to predict who will be net payers and who will be net recipients. Broad-based social support for a high degree of subsidising solidarity seems to derive partly from the uncertainty among net payers whether they too may be net recipients one day. The more this ‘veil of ignorance’ is lifted, the more unwilling net payers will become to endorse unlimited subsidising solidarity (Rosanvallon, 2000).
3.5
Is there a role for individual schemes?
Health policy-makers have so far opted for a policy strategy of cost control (efficiency cuts, budgeting) in combination with patient co-payments and certain reductions in statutory cover. Besides this, the reforms are concentrated on promoting greater efficiency through market forces. So far, the solidarity mechanisms have been left largely untouched. This is a sensitive issue and the media are always waiting in the wings to cry foul over any moves towards welfare retrenchment. As a result, the sparse cautious attempts undertaken to reduce subsidising solidarity either failed or were reversed (e.g. co-payment of medicines, specialist treatment, national health and IVF treatment). The recently introduced no-claim rebate has survived so far, but for how long? All things considered, the upshot is that economic, demographic and socio-cultural factors as well as growing scientific insights into the causes of disease entail that the current solidarity arrangements will not be left intact without discussion. All aspects of subsidising solidarity, income solidarity and intergenerational solidarity will be taken on board in the debate. Here, however, we will concentrate mainly on intergenerational solidarity, which provides the most interesting basis for comparison with pensions.
INTERGENERATIONAL SOLIDARITY: COLLECTIVE VERSUS INDIVIDUAL SCHEMES? When comparing the health and pension sectors, intergenerational solidarity is particularly important. This exists to a maximum degree in healthcare, but plays virtually no role in formal solidarity institutions as it effectively coincides with subsidising solidarity. The big question is: can this remain the case in the future? Many different aspects come into play here, but we will
44
P.P.T. Jeurissen and F.B.M. Sanders
focus on the accrued savings and personal assets, the way premiums are determined and, finally, the option of individual savings schemes. Today, personal wealth not only depends on income but increasingly on savings and other assets. Even so, these assets play virtually no role in the financing of collective healthcare arrangements. The current baby boom generation is affluent (RVZ, 2005). Whether younger generations will ever be in a similar position is uncertain (McKinsey, 2005). This is due to the growing costs of ageing, which diminish purchasing power growth. Future generations may enjoy increased longevity but, according to the most recent CPB analyses, will not be much richer than the current generations (Jacobs and Bovenberg, 2006). This raises the question whether the affluent elderly should be expected to make an extra (wealth-dependent) contribution to the rising cost of healthcare. This could be achieved by (partly) moving health premiums out of the social security sphere and into the tax sphere. Payable premiums could then not only be based on the individual’s income, but also on his or her assets. So far, however, reciprocity and insurance-based arguments have prevailed to create the opposite effect. One notable example is that of affluent people who pay a much lower exceptional expenses premium because they have retired to Mediterranean countries that provide less elderly care than the Netherlands. A second way of renewing intergenerational solidarity within the payas-you-go system runs via the premium levy system. If insurers are permitted to factor age into the payable premiums, the elderly will be charged a higher premium. This, incidentally, was already common practice in the old private health insurance system. Hitherto, higher premiums for the elderly were always justified on the grounds that they entailed an increased health risk. But now that the elderly are richer than ever, personal wealth could be put forward as a further reason for age-related premium differentiation. The less affluent elderly could then receive a health allowance to compensate for this higher premium. Yet another redistribution of the costs can be achieved within a pay-asyou-go system. This, in fact, is exactly what Germany did a few years ago with its Exceptional Medical Expenses Scheme (Pflegeversicherung). Intergenerational solidarity was invoked in justification of the measure, the argument being that people without children should pay a higher premium than people with children. The German Supreme Court ruled that child-rearing made a contribution to the sustainability of social insurance for elderly care based on pay-as-you-go financing (Di Fabio, 2005). The debate about making premiums more dependent on claims risk also touches on intergenerational solidarity, as a higher age entails a higher risk
3 Solidarity: who cares?
45
of disease. However, with this policy option, other risk factors such as obesity can also be taken into account. People who generate extra health profits and/or cost reductions by leading a healthy lifestyle could then qualify for a premium discount or lower co-payment. Substantial wealth gains can in principle be achieved in this way (Bhattacharya and Sood, 2005). The National Institute for Public Health and the Environment (RIVM) recently concluded that higher excise on smoking is the most effective measure from a wealth perspective (Feenstra et al., 2006). The most fundamental recalibration of intergenerational solidarity in healthcare could take the form of a fully funded system. This system would involve every individual having their own health savings account (De Kam, 2001) to pay for specific health expenditures in the future. This could be done either individually or in a group context. Such a system would be best suited for smaller health expenditures and more or less fixed costs to be incurred in the more remote future, such as GP, homecare and elderly care costs. But savings schemes also have disadvantages. Amongst other things, it is unclear whether sufficient capital can be saved up before the costs arise. Also, due to the large inter-individual variation in costs (e.g. for hospital care), people may run into problems if their savings prove to be insufficient. Therefore, it will be clear that individual savings schemes can never provide a full alternative to health insurance (RVZ, 2005). NYFER has mooted pension-cum-health policies as an interesting option. By including mandatory old-age health insurance in the pension scheme, participants will already start saving for intensive old-age care from a young age. Apart from being cheaper, this also avoids the selection and acceptance problems that arise when people put off taking out old-age cover until a later age (NYFER, 2005). These are interesting options from the insurer’s perspective as an unhealthy lifestyle increases the risk of higher annual healthcare costs but also leads to decreased life expectancy and a lower pension risk.
3.6
Conclusion
This contribution set out to discuss the most relevant solidarity concepts for the healthcare sector. We showed how solidarity is categorised and explained the relationship between solidarity and collectivity. Collectivity is both a precondition for solidarity and a means of achieving economies of scale. A tentative comparison with the pensions sector shows that subsidising solidarity is the most important type of solidarity in both sectors; in the healthcare sector there is also a strong correlation between risk and intergenerational solidarity.
46
P.P.T. Jeurissen and F.B.M. Sanders
Solidarity is crucial in healthcare, much more so than in the pensions sector. Without a certain minimum solidarity, vulnerable groups can easily lose access to healthcare. For this reason, substantial redistributions of wealth are necessary between the healthy and ill, young and old, and higher and lower incomes. The argument of accessibility (no solidarity, no healthcare for some) is less relevant in the pensions sphere. After all, people save for their own supplementary pension. Consequently, solidarity is not so much positioned in terms of redistribution, but is above all relevant for achieving economies of scale (e.g. because the participants in collective pension schemes are relatively insensitive to the time horizon). The direction of redistribution in pension systems is therefore mainly the (unintended) consequence of such factors as the achieved return on investments and wage inflation (WRR, 1999). The financial sustainability of pensions is relatively strong thanks to the full funding system. The current generations are only dependent to a limited extent on the will of future generations to help cover their pension costs. Healthcare is a different story: the reduction of uncertainty about the future financing of healthcare rests primarily on the perceived solidarity of future generations. If this solidarity remains strong, a pay-as-you-go system is probably the most efficient financing method. However, this contribution has also shown that we cannot simply assume that the strong healthcare solidarity we know today will remain intact in the future. With health expenditures set to continue rising, the health sector will make a structural claim on a large portion of the available economic growth. What’s more, the rising health expenditures will be concentrated on a relatively small group. The empirical ‘veil of ignorance’, which long concealed the relationship between health and genetics, behaviour and lifestyle, is rapidly wearing thin. This is likely to influence the degree of support for solidarity arrangements in society where differentiation and individualisation are still gaining in significance. Views differ as to the exact impact of these trends, but few expect them to strengthen support within society for solidarity. From this we can distil a policy agenda that is more focused on the individualisation of (existing) arrangements, including in the healthcare sector. This is already visible in the increasing material importance of supplementary health insurance and certain recent measures, including the noclaim rebate, the optional health insurance deductible and the possibility of joining a reduced-risk collectivity. Going forward, the future could bring more fundamental reforms such as a partial abolition of the ban on risk related premiums, supplementary savings schemes, and intergenerational redistributions to make the (affluent) elderly pay more for healthcare.
3 Solidarity: who cares?
47
Literature Beer, P. de, ‘Hoe solidair is de Nederlander nog?’, in E. de Jong and M. Buijsen, Solidariteit onder druk? Over de grens tussen individuele en collectieve verantwoordelijkheid, Nijmegen: Valkhof Pers, 2005. Berg, B. van den and F.T. Schut, ‘Het einde van gratis mantelzorg?’, Economisch Statistische Berichten (88) 4413: pp. 420-422, 2003. Battacharya, J. and N. Sood, Health insurance and the obesity externality, Working Paper 11 529, NBER, 2005. Buijsen, M.A.J.M., ‘Solidariteit, rechtvaardigheid en de zorg om gezondheid’, Filosofie & Praktijk (26)5, pp. 5-18, 2005. Cutler, D.M. and E. Meara, The medical costs of the young and the old: a forty year perspective, Working Paper 6114, NBER, 1997. Di Fabio, Die Kultur der Freiheit, München: Verlag C.H. Beck, 2005. Doorslaer, E. and C. Masseria, Income related inequality in the use of medical care in 21 OECD countries, Health working paper 14, OECD, 2004. EIM-onderzoek voor bedrijf en beleid, Solidariteit in het ziektekostenstelsel, inkomens- en risicosolidariteit in het tweede compartiment, Zoetermeer: EIM, 2002. Feenstra, T. L., P.H.M. van Baal, R.T. Hoogenveen, S.M.C. Vijgen, E. Stolk and W.J.E. Bemelmans, Cost-effectiveness of interventions to reduce tobacco smoking in the Netherlands, An application of the RIVM chronic disease model, Bilthoven: RIVM, 2006. Hansen, J., W. Arts and R. Muffels, ‘Wie komt eerst? Een vignetonderzoek naar de solidariteitsbeleving van Nederlanders met patiënten en cliënten in de gezondheidszorg’, Tijdschrift voor Sociale Wetenschappen (48) no. 1/2, pp. 31-59, 2005. Jacobs, B. and A.L. Bovenberg, ‘Voortschrijdend inzicht in de vergrijzing’, Tijdschrift voor Openbare Financiën (38)2, pp. 62-79, 2006. Kam, C.A. de, ‘Samen voor ons eigen’, Economisch Statistische Berichten (86) 4336, pp. D15-D16, 2001. Kukathas, C. and P. Pettit, Rawls: a theory of justice and its critics, Cambridge: Polity Press, 1990. McKinsey, The Coming Demographic Deficit: How Aging Populations Will Reduce Global Savings, McKinsey, 2005.
48
P.P.T. Jeurissen and F.B.M. Sanders
NYFER, Blijvende Zorg: economische aspecten van ouderenzorg, Zoetermeer: RVZ, 2005. RIVM (National Institute for Public Health and the Environment), Risicosolidariteit en zorgkosten, Zoetermeer: RVZ, 2005. RIVM, Kosten van ziekten in Nederland 2003, RIVM 270 751 010, Bilthoven, 2006. Rosanvallon, P., The New Social Question, Princeton: University Press, 2000. RVZ (Council for Public Health and Health Care), Houdbare Solidariteit in de Gezondheidszorg, Zoetermeer: RVZ, 2005. RVZ, Briefadvies, Houdbare solidariteit in de gezondheidszorg, Zoetermeer: RVZ, 2006. RVZ, Zinnige en Duurzame Zorg, Zoetermeer: RVZ, 2006. Schut, F.T., Competition in the Dutch Health Care Sector, thesis Erasmus University, Ridderkerk, 1995. Schuyt, K., ‘The sharing of risks and the risks of sharing. Solidarity and social justice in the welfare state’, Ethical Theory and Moral Practice (1):297-311, 1998. Spaendonck, T. and R. Douven, Uitgavenontwikkelingen in de Gezondheidszorg, Memorandum no. 16, The Hague: CPB, 2001. Starr, P. The social transformation of American medicine, The rise of a sovereign profession and the making of a vast industry, New York: Basic Books, 1982. Ter Meulen, R., R. Verburg, M. Offermans and H. Maarse, ‘Hoe verdelen we de schaarse zorg? Een vergelijkende analyse van opvattingen over criteria voor toegang tot zorgvoorzieningen in medische ethiek, beleid en surveyonderzoek’, Tijdschrift voor Sociale Wetenschappen (48) nr. 1/2, pp. 85-107, 2005. Verburg, R. and R. ter Meulen, ‘Solidariteit of rechtvaardigheid in de zorg? Een spanningsveld’, Tijdschrift voor Sociale Wetenschappen (48) nr. 1/2, pp. 11-31, 2005. Verstraeten, J., ‘Solidariteit in de katholieke traditie’, in: E. de Jong and M. Buijsen, Solidariteit onder druk? Over de grens tussen individuele en collectieve verantwoordelijkheid, Nijmegen: Valkhof Pers, 2005. WRR (Scientific Council for Government Policy), Generatiebewust beleid, The Hague: SDU-uitgevers, 1999. Widdershoven, B.E.M., Het dilemma van solidariteit. De Nederlandse onderlinge ziekenfondsen, 1890-1941, Amsterdam: Aksant, 2005.
Part 2. Quantifying solidarity
4
Operating costs of pension schemes
J.A. Bikker and J. de Dreu1 This chapter examines what type of pension scheme has the lowest operating costs. We first analyse the operating costs of Dutch pension funds, broken down by administrative and investment costs. Various cost-influencing factors are identified, including scale, pension fund type, plan type, outsourcing and reinsurance. Economies of scale are shown to be dominant in explaining differences in costs across pension schemes, leading to the conclusion that the consolidation of small pension funds could improve cost efficiency. In addition, the costs per participant of mandatory industry-wide pension funds turn out to be significantly lower than those of company pension funds. Next, the costs of pension schemes offered by pension funds and life insurers in the Netherlands are compared in an effort to distinguish between collective and private schemes. We find that the operating costs per participant of collective pension funds are many times lower than those of private schemes.
4.1
Introduction
The fall of equity prices in 2000-2002 combined with persistently low longterm interest rates and an ageing population led to a worldwide crisis in the pension industry. Since then, higher contributions and lower pension accrual rates, as well as a rebound of equity prices and interest rates con1
The Dutch central bank (DNB), Supervision Policy Division, Strategy Department,
[email protected]. Jan de Dreu wrote this article when working for DNB; currently he is employed by ABN AMRO,
[email protected]. Views expressed are those of the individual authors and do not necessarily reflect official positions of DNB. The authors are grateful to Dirk Broeders, Aerdt Houben and Wil Dullemond for valuable comments and suggestions. The first part of this article is based on Bikker and De Dreu (2007).
52
J.A. Bikker and J. de Dreu
tributed to the recovery of the financial position of pension funds.2 Despite the financial problems sketched above, the operating costs of pension funds as a potential source of savings received little attention. However, the cumulative effect of these costs can have a strong impact on the size of pension benefits. Figure 1 shows the impact of annual costs on pension benefits for a fictitious pension scheme (for a single person or a group of persons). Here, annual costs of 1% of total assets lead to a reduction in the pension payments of 27% in a defined contribution (DC) system or an increase in the costs or contribution of over 37% in a defined benefit (DB) system.3 Operating costs per participant vary strongly between pension funds, mainly due to scale effects and inefficiencies. In addition, operating costs differ significantly between pension funds and life insurers. In this context, one should bear in mind that the different types of pension schemes are not fully comparable. The costs of collective schemes of pension funds and collective contracts of life insurers, on the one hand, and the costs of private pension schemes at life insurers, on the other, differ in nature. As a result, the costs of private schemes as a percentage of the contribution are typically five times higher than those of collective arrangements.4 In private schemes, unutilised economies of scale is the dominant factor explaining relatively high costs. These cost differences illustrate the importance of selecting the right organisational form for pension provisions. This chapter first examines the operating costs of pension funds and the main factors that determine these costs. A distinction is made between administrative and investment costs. Key cost determinants include the size of pension funds, their organisational form, the type of pension plan and the degree of outsourcing of the administration, asset management and risks. This analysis makes it possible to determine characteristics of an ideal pension fund in terms of efficiency and to identify which policy or form of market organisation can help to improve the efficiency of existing funds. For this analysis we use data of all (one thousand) Dutch pensions funds during 1992-2004. Next we turn to the role of life insurers as pension providers, both as a service provider to pension funds and as an independent provider of collec-
2
Risks that were shifted from employers to participants also play a role here insofar as employees were not compensated for this.
3
See Bateman and Mitchell, 2004, and Bateman, Kingston and Piggot, 2001.
4
One should be careful in interpreting these figures, as the comparison is complicated.
4 Operating costs of pension schemes
53
Reduction of pension benefits
0% -10% -20% -30% -40% -50% 0%
0,5%
1%
1,5%
2,0%
Administrative and investment costs
Fig. 1. Erosion of pension benefits due to annual operating costs Note: To simulate the impact of operating costs on annual pension payments, we assume annual wage growth of 3%, annual inflation of 2%, a nominal investment return of 7%, uninterrupted contribution payments over 40 years and a pension payout period of 20 years. Source: Bikker and De Dreu, 2007.
tive and private pensions. Attention is also devoted to the cost differences between pension funds and life insurers as well as between private and collective schemes. We thus establish the characteristics of pension funds and pension schemes that are best suited to provide efficient pensions.
4.2
Operating costs of pension funds
The operating costs of pension funds consist of administrative costs and investment costs. Administrative costs relate to all operational tasks excluding asset management, such as record keeping, communication with participants, policy development and compliance with regulatory and supervisory requirements. These costs include salaries, rents and fees charged by third parties such as actuaries, accountants and lawyers. Investment costs are discussed in section 4.4. We use data of Dutch pension funds over the past thirteen years as reported to the Dutch central bank (DNB) for prudential purposes. The number of pension funds gradually decreased from 1131 in 1992 to 742 in 2004 (see also Table B.1 in the appendix). Tables 1 and 2 present key statistics on administrative costs in 2004 for, successively, different size categories, types of pension funds and types of pension plans. Size is measured
54
J.A. Bikker and J. de Dreu
by the number of participants or total assets. Participants consist of contributing employees, inactive participants and pensioners. Data from earlier years (1992-2003) lead to comparable figures as shown in the tables below and are not presented separately. Though all pension funds are independent legal entities, many small and some mid-sized company pension funds utilise staff and facilities of the sponsor company. The associated costs are often not fully charged to the pension fund and consequently also not reported. About 12% of the pension funds report no administrative costs.5 These funds are therefore excluded from the statistics presented in this chapter. In addition, many (mainly small) company pension funds underreport their administrative costs. For example: 65% of these funds report no wage costs. Evidently these costs are either borne by the sponsor company or are included in ‘other costs’ (and remain part of the administrative costs). Such underreporting does not occur among industry-wide pension funds, as these are unable to shift costs to their sponsors. Later we will see that these imperfections in the data are systematic (occurring mainly at smaller company pension funds) and therefore do not significantly impair our analyses. Without this distorting effect, the observed dominant influence of economies of scale and differences in costs between the different categories of pension funds would only be greater. The upper part of Table 1 shows the average administrative costs of pension funds for different size categories in terms of participant numbers. The table indicates that the (weighted) average of administrative costs as a percentage of total assets declines sharply as the number of participants increases: from 0.59% for the smallest funds to 0.07% for the largest funds. The average administrative costs per participant fall even more sharply as the number of participants increases, namely from an average of € 927 per year for the smallest funds to about € 30 for the two largest size categories. As noted earlier, the cost differences between the size categories are actually even greater than shown in these figures. This is due to the underreporting of costs, mainly by the smallest company pension funds (see also fourth column). Almost half of the category of smallest funds consists of personal pension vehicles for director-owners and director funds for a limited number of (former) board members and members of the supervisory board. This explains why, on average, this category has much higher assets per participant than the other categories. 5
The data was collected for prudential supervision purposes, where costs only play a minor role.
4 Operating costs of pension schemes
55
Table 1. Annual administrative costs of pension funds by size category (2004) Size categories of pension funds based on:
Administra- Administra- Total assets Funds that Total numtive costs/ tive costs per partici- do not report ber of total assets per participant wage costs participants (%) pant (€ 1000) (%)a (1000) (€)
Number of funds
1. number of participants < 100
0.59
927
157
88
2
56
100-1000
0.46
302
66
82
104
225
1000-10 000
0.23
156
68
55
809
264
10 000-100 000
0.17
86
50
18
2 774
87
100 000-1 million
0.24
28
12
30
7 146
20
> 1 million
0.07
33
46
0
5 611
3
Average / total
0.15
48
33
61
16 446
655
2. total assets (€ million) 0-10
1.23
159
13
85
37
105
10-100
0.55
129
23
71
508
289
100-1000
0.27
51
18
45
3 532
209
1000-10 000
0.17
45
27
23
4 929
44
> 10 000
0.10
43
45
25
7 439
8
a
Note that mainly (small) company pension funds sometimes underreport wage costs.
Source: Bikker and De Dreu, 2007.
Economies of scale result from high fixed costs and other operating costs that increase less than proportionally with pension fund size. Examples include the costs arising from policy development, data management systems, reporting requirements and the hiring of experts such as actuaries, accountants, lawyers and consultants. The lower part of Table 1 presents the (weighted) average administrative costs for different size categories in terms of total assets. The table shows that administrative costs expressed as a percentage of total assets are negatively related to the size of pension funds. While the smallest funds have operating costs of 1.23% of total assets, this percentage is only 0.10% for the largest funds. Figure 1 shows the impact of a 1% difference in annual operating costs on pension benefits. In summary, Table 1 shows that the operating costs of pension funds are characterised by strong economies of scale, irrespective of whether the size of the institution is expressed in terms of participant numbers or total assets.
56
J.A. Bikker and J. de Dreu
The upper part of Table 2 presents administrative costs for different types of pension funds. We distinguish three main types: company pension funds, industry-wide pension funds and professional group pension funds. Company pension funds provide pension schemes to employees of the sponsor company. They are legally independent of the sponsor company and are managed by the employer and employee representatives. Industry-wide pension funds provide pension schemes to employees in a sector based on a Collective Labour Agreement (CLA) between the employers and labour unions in this sector. There are two types of industrywide pension funds: mandatory and non-mandatory. Mandatory funds are based on a binding CLA, making participation mandatory for all employers and employees working in the respective sector. Non-mandatory funds are based on a CLA that allows employers to choose whether to participate in the collective fund or not. Professional group pension funds provide pension schemes to professional groups such as general practitioners and notaries. Apart from these three main groups, there are also other types of funds such as savings funds. The administrative costs of company pension funds average € 138 per year, which is high compared to industry-wide pension funds whose annual Table 2. Annual administrative costs by type of pension fund and type of pension plan (2004) Type:
Administrative Administrative Total assets Total number costs/total costs per per participant of participants assets participant (€ 1000) (1000)a (%) (€)
Number of fundsa
Average number of participants (1000)
Pension fund Industry-wide (all)
0.13
33
26
14 072
95
148
– mandatory
0.12
31
26
13 557
76
178
– non-mandatory
0.16
66
40
515
19
27
Company
0.19
138
71
2 167
524
4
Professional group
0.10
221
221
71
11
6
Average / total
0.15
48
33
1 446
655
25
Mainly DB
0.14
49
34
15 546
590
26
Mainly DC
0.37
25
7
672
51
13
Other
0.36
33
9
221
12
18
Pension type
a
The pension type of 21 pension funds is not known; four funds are savings funds.
Source: Bikker and De Dreu, 2007.
4 Operating costs of pension schemes
57
costs average only € 33. As noted before, the actual differences are even greater due to the aforementioned underreporting of costs by company pension funds. Professional group pension funds have the highest costs per participant, namely € 221. Company pension funds and professional group pension funds usually manage more assets per participant, which leads to higher costs. This may be due to e.g. more generous pension schemes or a relatively large number of older participants (whose accrued pension assets are obviously larger than those of younger participants). Consequently, the administrative cost difference between the various types of pension funds is smaller per invested euro (or as a percentage of total assets) than per participant. Pension schemes provided by company pension funds are generally much less standardised and much more customised to the preferences of the employer and employees than schemes provided by industry-wide pension funds. In addition, the services to the participants can be of a higher quality. However, this explicit choice for customisation and extra service results in higher operating costs. Table 2 shows that most pension funds are company pension funds, but that these serve only a small number of the participants. One major advantage of industry-wide pension funds is that when employees change employers within the same sector, the accrued pension assets can often remain within the fund. As a result, lower costs are incurred than when the assets need to be transferred between company pension funds. The (mandatory) industry-wide pension funds have by far the largest number of participants. The lower part of Table 2 shows the administrative costs for different types of pension plans. We see significantly higher average costs for DB pension plans of € 49 per participant per year, as compared to € 25 for DC pension plans. However, the total assets per participant are much higher in DB pension funds than those of DC funds, presumably because the participants in the latter funds are a lot younger and have therefore accrued much less pension assets.6 In addition, the number of DC participants has been fairly limited so far. These cost differences are probably also partly determined by scale effects. Overall, scale effects appear to be the dominant explanation for cost differences between pension fund categories, whilst the organisational form possibly has some, albeit smaller, effect. In order to identify the impact of different factors that capture the various organisational forms, we need to perform a multivariate regression analysis so that all factors can be taken into account simultaneously.
6
Note that more assets also involve higher costs.
58
J.A. Bikker and J. de Dreu
4.3
An empirical model for administrative costs
To establish the impact of scale, organisational structure and pension plan types on the operating costs of pension funds, we use a multivariate regression model. The left-hand column of Table 3 provides the estimates for the impact of variables that explain administrative costs in our model.7 The scale of pension funds is represented by the number of participants (in logarithms). This term is also included quadratically to account for the possible nonlinearity of scale effects. The coefficient of 0.63 indicates that a 1% increase in the number of participants leads to a cost increase of only 0.63%. This implies that there are substantial unutilised economies of scale averaging 37% per unit of extra production. This observation was also made in relation to DB and collective DC pension funds in the United States (Caswell, 1976 and Mitchell and Andrews, 1981), Australia (Bateman and Mitchell, 2004), and in relation to DC pensions in sixteen countries around the world (Whitehouse, 2000; Dobronogov and Murthi, 2005, and James, Smalhout and Vittas, 2001). The quadratic term indicates that these economies of scale are greater for small funds and smaller for large funds. In 2004, all existing funds were below the theoretical optimum size, where the economies of scale turn to diseconomies of scale. While controlling for other factors, mandatory industry-wide pension funds are found to operate at the lowest costs.8 As noted before, this can partly be explained by their generally standardised and less generous pension schemes, which are simpler to administer. An additional advantage is that when employees change employers the accrued pension assets can often remain within the fund, so that less transfer costs are incurred. Nonmandatory industry-wide pension funds and company pension funds occupy the middle ground in terms of efficiency, while professional group pension funds are the least efficient. Their costs may be higher mainly because these funds cater to lots of individual participants instead of a single employer, which, for instance, makes the collection of contributions more cumbersome. We find that pension funds with a DC plan have lower administrative costs than funds with a DB plan. This applies to Dutch DC pension funds in which participants (1) are unable to select and switch between pension funds so that no marketing costs need to be incurred and (2) have only a limited choice in terms of the investment mix so that information costs are 7
Administrative costs are expressed here in logarithms.
8
Significantly lower than company and professional group pension funds.
4 Operating costs of pension schemes
59
Table 3. Estimates of the administrative and investment cost models (1992-2004) Administrative costs Coefficients Number of participants (in logarithms) Total assets (in logarithms)
Investment costs
t-values
Coefficients
t-values
0.63
105.1
–
–
–
–
0.83
76.4
0.05
38.1
0.03
10.0
Mandatory industry-wide pension funds
-0.56
10.5
-0.24
3.5
Non-mandatory industry-wide pension funds
0.49
6.8
-0.25
2.6
Company pension funds
0.56
17.1
0.14
3.0
Professional group pension funds
1.24
18.1
0.05
0.5
Defined Contribution pensions (DC)
-0.20
4.7
0.05
0.8
Outsourcing of the administration
1.08
33.2
Complete reinsurance of liabilities
-0.77
19.2
-0.30
4.9
Partial reinsurance of liabilities
-0.12
2.9
-0.09
1.8
Total assets (in € 1000) per participant
0.07
3.0
Percentage of pensioners
0.62
11.5
-0.09
1.1
Reported investment costs
-0.45
17.8
Constant
-0.45
8.9
-5.17
50.3
10 119
7.4
4 986
Ditto, squared
a
Number of observations R
2
0.71
0.75
Respectively, the number of participants (in logarithms) and total assets (in logarithms). Note: Almost all coefficients are significant at the 99% level; italics indicate ‘no significance’. All variables are expressed in 2004 prices.
a
limited.9 Compared with DB funds, DC funds require no or less actuarial advice, which should imply lower costs. Outsourcing of the administration seems more expensive, but that is probably a distortion due to the aforementioned underreporting of administrative costs. With outsourcing the invoice puts the full costs on the table, whereas without outsourcing part of the costs can remain concealed, at
9
Marketing costs constitute a major part of the operating costs in countries where participants can switch between funds (Dobronogov and Murthi, 2005). In addition, since many participants have no idea how to invest their pension assets properly (e.g. see Van Rooij et al., 2007), they should be provided with information and advice if they can choose to select their own investment mix for their pension.
60
J.A. Bikker and J. de Dreu
least for company pension funds. It is found that both full and partial reinsurance of insurance and investment risks, which is often accompanied by the outsourcing of administration and asset management, lead to lower operating costs. However, it is probable that part of the operating costs is included in the contributions that are paid to the insurer. We are therefore unable to conclude that reinsurance increases efficiency. Next, we look at three control variables. As expected, the costs are slightly higher if more pension assets are managed per participant. Costs also increase with a growing number of pension recipients. Finally, costs are lower if the fund also reports investment costs. Evidently, investment costs are sometimes partly stated as administrative costs. This does not distort the total operating costs, but does influence the distribution over the two cost categories. Insufficiently accurate reporting by a (small) portion of the funds has evidently not prevented clear regression results. All coefficients are significant at a very high reliability level. Also, if the regression model is estimated for different subsets (e.g. all industry-wide pension funds, all company pension funds or only the data of 2004), the results show the same signs for the coefficients and, for most variables, the same high level of significance of 99%. The first important conclusion is that substantial unutilised economies of scale occur in the management of small and medium-sized pension funds. The same applies after controlling for the option to achieve economies of scale through outsourcing at life insurers and pension providers. Size, therefore, is a crucial determinant of the efficiency of pension funds. The second important conclusion is that, on average, mandatory industry-wide pension funds have significantly lower administrative costs than company pension funds. The organisational form of pension funds is evidently also essential for efficiency purposes. Due to the systematic underreporting of administrative costs by mainly small company pension funds, both effects are actually expected to be somewhat higher than observed in this analysis.
4.4
Investment costs of pension funds
Investment costs arise from investment analysis, risk management and trading, and include salaries of analysts and portfolio managers, brokerage fees and charges for the use of electronic trading facilities.10 The reported 10
The literature on mutual funds (which have comparable investment activities) shows that higher investment costs are not (sufficiently) compensated by higher returns (e.g. Jensen, 1968, Malkiel, 1995, and Malhotra and McLeod, 1997). Large
4 Operating costs of pension schemes
61
investment costs amount to approximately 40% of total operating costs. Actual investment costs are probably somewhat higher because part of the investment costs is immediately deducted from the returns. About 24% of the pension funds report no investment costs. These funds do not occur in the tables and estimates used below. Sometimes these costs are included in the reinsurance premiums. In addition, the investment costs may have been deducted directly from the investment returns or included in administrative costs. The upper part of Table 4 presents the average investment costs of pension funds for different size categories, expressed in numbers of participants. Investment costs as a percentage of total assets decrease as the number of participants increases from about 0.14% for the three smallest fund classes to 0.08% for the biggest funds. The average investment costs per participant decrease even more sharply with the number of participants, namely from € 270 for the smallest funds to € 13 and € 39 for the two largest fund categories. Note that the investment costs per participant are the lowest for pension funds in the second-largest category, which contains most participants. Once again, the real cost differences between size categories are even greater than shown in Table 4, as non-reporting of investment costs is much more common among small funds than the large funds. The lower half of Table 4 shows a comparable picture for the various size categories on the basis of total assets. The analysis of the investment costs of pension funds reinforces our earlier finding that scale has a strong impact on operating costs and that industry-wide pension funds operate at significantly lower costs. This conclusion is confirmed if the earlier regression analysis is repeated with a model for investment costs, where the size of total assets is used as the scale variable (see the right-hand column of Table 3). The results show similar coefficients and comparable conclusions. On the investment side, large unutilised economies of scale are found to exist, though these are smaller than for administrative costs (17% versus 37%). For investment costs it is found that – after controlling for other factors – industry-wide pension funds again operate at significantly lower costs than company and professional group pension funds. The coefficients of the other explanatory variables are less significant.11 funds with an extensive investment apparatus generate no or insufficient excess returns to compensate for higher costs. Therefore it makes sense to reduce the costs to an optimal level. Note that bid-ask spreads are not part of the investment costs. See for this e.g. Bikker et al. (2007). 11
In an alternative specification (not shown here) with participant numbers as the scale variable, the other variables do all prove to be highly significant, with the same signs as in the administrative cost model.
62
J.A. Bikker and J. de Dreu
Table 4. Annual investment costs of pension funds by size category (2004) Size category of pension funds based on:
Administrative Administrative Total assets Total num- Funds that Number of costs/total costs per per particiber of report no funds assets participant pant participants wage costs (%) (€) (€ 1000) (1000) (%)a
1. number of participants < 100
0.13
270
208
1
52
27
100-1000
0.14
101
72
75
33
151
1000-10 000
0.14
97
71
672
21
209
10 000-100 000
0.11
45
41
2 469
13
76
100 000-1 million
0.13
13
10
6 847
10
18
> 1 million
0.08
39
46
5 611
0
3
Average / total
0.10
31
31
15 676
26
484
0-10
0.15
25
17
16
53
49
10-100
0.14
31
22
418
28
209
100-1000
0.14
25
18
3 163
14
179
1000-10 000
0.10
24
24
4 809
7
41
> 10 000
0.10
39
41
7 270
25
6
2. total assets (€ million)
Source: Bikker and De Dreu, 2007.
4.5
Life insurers as providers of pension schemes
The second part of this chapter examines the cost differences between private and collective pension schemes. The present section discusses the role of life insurers as a provider of both private and collective pension schemes. Section 4.6 then takes a closer look at the cost differences between life insurers and pension funds.
COLLECTIVE PENSION SCHEMES Companies that do not belong to an industry with a mandatory industrywide pension fund can choose to arrange their employee pension scheme through a life insurer. Forty insurers provide such ‘direct schemes’ to some 1.8 million participants working for around forty thousand companies.
4 Operating costs of pension schemes
63
In addition, pension funds can reinsure their insurance and investment risks at life insurance companies and outsource their administration and the management of their investments to more specialised institutions, including life insurers. A pension fund can even outsource all its activities to a life insurer, in which case it exclusively acts as a middleman. Where insurers or other institutions are better equipped to bear certain risks of pension funds or are able to perform certain activities more cost-effectively, pension funds can increase their efficiency by reinsuring risks and outsourcing activities. This actually happens on a considerable scale. In 2004 pension funds outsourced on average 36% of their activities in cost terms. Over a third of the pension funds (principally smaller institutions) outsourced more than 50% of their activities. In the same year, 20% of the (mainly smaller) funds were fully reinsured.12 In this way, pension funds seek to maximise the efficiency of their pension activities, with (mainly small) pension funds benefiting from economies of scale at life insurers and pension providers in cases where their own scale is too limited. Direct schemes and outsourcing thus mean that (at least some) market efficiencies are still achieved in providing pension schemes.13
PRIVATE PENSION SCHEMES Besides collective schemes of pension funds and collective contracts of life insurers, there are also private pension schemes. These are important for a large number of self-employed people who are not in salaried employment and are not members of an professional group pension fund. In addition, many people choose to supplement their employee pension with additional savings in the third pillar of the Dutch three-pillar system, e.g. to repair a loss of pension rights due to a change in employment or to enjoy a higher pension. Apart from privately managed assets (e.g. savings or investment accounts), this mainly concerns life insurance policies. These are either (deferred or immediate) annuities or endowment insurance policies with annuity clauses. Premiums for both types of insurance are eligible for 12
On average these funds have a balance sheet total that is only one tenth of that of the other funds.
13
Outsourcing of pension activities to insurers is accompanied by additional agency costs: the pension fund or the responsible employer needs to check whether the insurer or pension provider fulfils all its obligations.
64
J.A. Bikker and J. de Dreu
income tax deduction, subject to certain conditions.14 In the case of endowment insurance, savings are built up in order to purchase an annuity: for instance, an annuity payable on death before retirement for the benefit of surviving dependants or a single lifetime annuity. Table 5 provides an overview of the administrative costs of life insurers, consisting of operational costs and acquisition costs (marketing and selling costs, including commissions for intermediaries).15 Note that these figures indicate the average costs of life insurers for their entire portfolio of products, which include both collective and private policies, insurance policies where the investment risk is borne or not borne by the policyholders, life annuities, endowment insurance, and so forth. In addition, it is important to remember that cost comparisons between pension funds and insurers are difficult to make (see section 4.6). The first point worth noting is that large unutilised scale effects also occur for life insurers. The total costs as a percentage of gross premiums (a measure we use to permit comparison with pension funds) vary from Table 5. Administrative costs of life insurers and pension funds by size category (2004) Life insurers Size category based on total assets (€ million)
Pension funds
Administrative Gross Administrative costs/gross profits/gross costs plus gross premiums premiums (%) profits/gross (%) premiums (%)
Number of insurers
Administrative Number of costs/gross pension funds premiums (%)
0-10
37.9
0.6
38.5
12
11.9
80
10-100
36.1
11.8
47.9
11
7.8
277
100-1000
17.2
13.4
30.6
26
5.0
206
1000-10 000
13.2
11.4
24.6
24
3.9
43
> 10 000
12.4
13.0
25.4
8
2.6
8
Average/ total
13.1
12.6
25.7
81
3.5
614
14
A proposal (annuity saving bill of Depla-De Vries) has been made to the Dutch Lower Chamber to extend the application of such fiscal facilities to old age savings via blocked bank accounts.
15
The data are described in Bikker and Van Leuvensteijn (2007). Investment costs are discussed later in this chapter.
4 Operating costs of pension schemes
65
12.4% for the largest life insurers to 37.9% for the smallest. In proportional terms therefore the costs of small insurers are three times higher than those of large insurers. On average almost half of the costs consist of acquisition costs; the percentage is somewhat higher for small insurers. For the period from 1995 to 2003, Bikker and Van Leuvensteijn (2007) calculated unutilised scale effects of on average 21%, varying from 10% for the 25% largest insurers to 42% for the 25% smallest insurance firms. These unutilised scale effects are therefore somewhat lower than those of pension funds. In addition, a portion of the contributed premiums goes to gross profits. This compensates shareholders for bearing certain risks such as the longevity and investment risk. The profit margin in 2004 seems to have been more or less equal across the size categories. It should be noted that this profit margin relates to the entire portfolio. There are indications that the margin for the new production is smaller than for older policies.16 On average, administrative costs and gross profits jointly account for a quarter of the gross premium at the larger insurers and almost half of the gross premium at the smaller insurers. Insurers are partly unable to avoid these costs while pension funds, being non-profit institutions, do not charge profit margins. It is again noted that the above analysis provides information on the average costs of all life insurance products. We lack the data required for a more refined analysis. It is plausible, however, that large cost discrepancies will occur for different types of products, such as collective versus private contracts. With collective contracts the costs will decrease relatively strongly as the size of the contract increases, e.g. in terms of number of participants. Alongside the aforementioned administrative costs, insurers also incur investment costs. In the financial figures that life insurers are required to report to the Dutch central bank (DNB), the investment costs are aggregated with interest charges. Averaging 0.31% of total assets, these costs are higher for life insurers than pension funds (0.10% of the total assets). This is probably (partly) due to interest charges. However, as the basis of comparison, i.e. total assets, is not identical at life insurers and pension funds, no further conclusions can be drawn from this.
16
This is evident from e.g. embedded value calculations where, for instance, the profit on new policies is determined over the entire term.
66
J.A. Bikker and J. de Dreu
4.6
Administrative costs of life insurers and pension funds compared
Comparing the administrative costs of pension schemes provided by life insurers with those provided by pension funds gives rise to numerous complications.
DIFFERENT PRODUCTS The first question that arises is whether insurers deliver, or are able to deliver, the same products as pension funds. This is not the case. Most pension funds provide DB pension plans where the size of the pension benefits is fixed (long) in advance on the basis of the final or average salary, in a few cases with guaranteed price or wage indexation, at least until the time of retirement (see Bikker and Vlaar, 2007). Insurers do not provide such pensions (and, in fact are not allowed to, at least not at fixed contributions). They cannot distribute the investment, inflation and longevity risks over different generations by varying contributions. In general, insurers provide nominal pensions where surplus profit sharing creates the possibility – but not the certainty – of applying indexation.17 Incidentally, the aforementioned DB pensions can be offered by an insurer if the employer undertakes to pay the additional contribution required for indexation and supplements up to a certain percentage of the final salary (socalled “back service” in final pay schemes). Besides direct schemes and collective contracts that are comparable with pension fund schemes, life insurers also provide reinsurance contracts and private policies for individuals, including both pension schemes as well as other types of insurance. Such products cannot be provided by pension funds.
MANDATORY PARTICIPATION Moreover, from a cost perspective, the position of pension funds and that of life insurers is not always comparable. First of all, the mandatory participation at pension funds leads to a strong reduction in costs. Almost half of the administrative costs of life insurers consist of acquisition costs made up of marketing and selling costs, including commissions for intermediaries. Insurers need to incur these costs to acquire customers, while pension
17
Sometimes partial indexation is guaranteed.
4 Operating costs of pension schemes
67
funds can avoid these as a result of the mandatory participation. It is worth noting that these costs are not entirely without benefit for clients, as they are partly incurred to advise on the need for, and best method of, saving for a pension.18 Mandatory participation in pension schemes yields large social savings in terms of reduced educational and search costs. Note, incidentally, that collective contracts of life insurers also benefit from mandatory participation, as acquisition costs can then be avoided.
ADVERSE SELECTION The absence of mandatory participation with private policies of life insurers also leads to costs due to adverse selection. People with poor health and therefore a greater risk of death are, on average, more likely to take out life insurance payable on death. Similarly, people in good health are more likely – on average – to take out a lifetime annuity. In order to limit the effects of adverse selection, applications for life insurance involve a costly medical examination and selection process. Due to the mandatory participation, costs related to adverse selection play no role for pension funds. In addition, buyers of annuities tend to be more highly educated and remunerated people who, on average, are healthier and have a longer life expectancy. This will be taken into account in the pricing process.19
DIFFERENT ORGANISATIONAL FORM The difference in organisational form also leads to unequal costs. Insurers tend to be profit-oriented companies while pension funds are non-profit institutions. Gross profits averaged 12.6% of the contributions in 2004. For comparison purposes, the corporation tax on the profit and surplus profit must be included in the calculation as costs for the policyholder. Whether the net return on equity should also be included in the calculation is open to question. For pension funds a portion of the paid contributions is used to fund buffers. In a sense the participants in a pension fund must themselves contribute a kind of share capital. In the long term, however, they will eventually benefit from this capital as the returns earned on the buffer will be used for e.g. indexation (see Bikker and Vlaar, 2007). The buffer itself, however, will be shifted to the following generation.
18
The quality and reliability of such advice is sometimes disputed (CPB, 2005).
19
This does not influence the administrative costs of insurers.
68
J.A. Bikker and J. de Dreu
DIFFERENT REGULATORY REGIME Finally insurers must cover the risks on insurance contracts by maintaining capital, so that costs of capital (or profits before taxation) become part of the cost price.20 Pension funds are required to cover their nominal liabilities for 105% and also to maintain a solvency buffer for investment and longevity risks. The required buffer for an average fund is approximately 30%. Until the Dutch Financial Assessment Framework took effect on 1 January 2007, pension funds were permitted to base their calculations on an actuarial interest rate of 4% at maximum,21 whereas insurers were required to use 3% for new contracts since 1997. This difference in interest rate does not lead to widely divergent pension contributions in the long term, but may do so in the short term, for instance in a recovery period when buffers need to be repaired. Different regulatory regimes can disturb the optimal allocation of pension provisions over pension funds and life insurers. Though some regulatory convergence is likely in the near future, differences between the regulation of the two sectors will continue to exist on account of the profit objective of most life insurers and the corporation tax on their profits as well as the intrinsic differences between pension funds and life insurers.
COST DIFFERENCES Table 5 shows the differences in the administrative costs of Dutch pension funds and life insurers by expressing these costs as a percentage of the gross premiums.22 The size categories are not relevant for the comparison; these only give information about the distribution of the costs. On average, administrative costs account for 3.5% of the gross contributions at pension funds in 2004,23 while the percentage at insurers is over 13% excluding profit margins and almost 26% including profit margins. These data can differ from year to year due to e.g. fluctuations in profits or changes in
20
The supervisor sets a minimum solvency requirement for life insurers, which, incidentally, is much lower than the capital that insurers maintain in connection with their own operational targets.
21
Many pension funds did not use 4% but 3.7% at year-end 2005.
22
Insurers do not report investment costs separately. These costs are included in the investment charges item.
23
This figure may be a fraction higher due to partial underreporting of costs by mainly small pension funds.
4 Operating costs of pension schemes
69
Table 6. Administrative costs of life insurers and pension funds (2000-2004) Life insurers
Pension funds
Year Administrative costs
Gross profits
Administrative costs and gross profits
Administrative costs
As % of gross premium 2000
12.4
14.7
27.1
5.8
2001
12.8
12.8
25.6
5.6
2002
13.1
1.7
14.8
4.2
2003
13.0
13.0
26.0
3.9
2004
13.1
12.6
25.7
3.5
Average
12.9
11.0
23.9
4.4
As % of total assets 2000
1.20
1.42
2.63
0.13
2001
1.31
1.32
2.63
0.15
2002
1.31
0.17
1.48
0.18
2003
1.26
1.26
2.52
0.17
2004
1.23
1.19
2.42
0.14
Average
1.27
1.08
2.35
0.15
the contributions. For this reason, Table 6 presents the same data for the past five years while the administrative costs are also reported as a percentage of total assets.24 The conclusions remain the same. The comparison indicates that due to (i) the frequently individual scale, (ii) the need for acquisition (promotion, distribution and advice), (iii) the costs that are caused by adverse selection and (iv) the profit objective, insurers generally incur higher costs for the provision of pension schemes than pension funds.25 Life insurers play a vital social role in offering insurance
24
Because pensions are build up over a very long time, pension funds maintain comparatively more assets, which further reduces their cost margins expressed as a percentage of total assets.
25
The annual continuing costs per policy amount to about € 50-100. The one-off costs per life insurance, including medical examination, equal about € 300-500 as opposed to € 1500-2000 per policy for endowment policies (e.g. for mortgages) and annuities (immediate annuities and endowment policies with an annuity clause). The latter include advice.
70
J.A. Bikker and J. de Dreu
products and their policies can yield significant benefits for individuals, partly due to the possibility of providing customised products. However, collective pension schemes based on mandatory participation can be offered at significantly lower (administrative) costs. One area where a cost comparison could conceivably be made between life insurers and pension funds is that of collective contracts. Unfortunately, the absence of separate data on the administrative costs of these collective contracts implies that we are unable to make such a comparison. However, we do know more about one specific cost item, namely profits. Insurers report profits on both collective and private insurance products, and it turns out that both yield comparable profit margins in both market segments (see tables B.2 and B.3 in the appendix). These, however, are the profit margins on the existing portfolio, i.e. the production from the past. Another source of information consists of embedded value calculations, where the profitability of the portfolio of both existing and new collective and other contacts in the remaining term until maturity is calculated. These often turn out to be loss making in the sense that the targeted return on equity is not entirely achieved. Apparently, this part of the market, where actuarial knowledge is present on both sides of the table, has become a fiercely competitive market. Smaller and medium-sized pension funds take out reinsurance contracts on a reasonably large scale, while smaller and medium-sized companies take out collective contracts. Evidently this is more cost-effective in these cases, where economies of scale will often be a decisive factor. Finally we can make a statement about the costs of private policies. Though we do not have separate administrative cost data, we can see in Table B.2 of the appendix that more than half of the contributions and provisions relate to private policies. The number of private policies greatly exceeds the number of collective policies (by 36 million).26 Given that the costs are strongly determined by scale, we conclude that most of the insurers’ costs are allocated to private policies. Nevertheless, we prefer to use conservative estimates by assuming average costs as a percentage of the gross contributions for private policies. Finally, we assume that there is no significant difference between the administrative costs for endowment insurance and pension and annuity insurance products.
26
The number of collective arrangements is limited and comprise less than 5 million insured persons.
4 Operating costs of pension schemes
4.7
71
Conclusions
This chapter shows that the administrative costs of collective pension schemes offered by pension funds constitute only a fraction of the operating costs of private pension schemes offered by insurers (over the last five years an estimated 4.4% versus 12.9% of the gross contributions). The difference becomes almost twice as large when the gross profit margin of insurers is also taken into consideration: 4.4% versus 23.9%. These differences are explained by, among other things, scale effects, adverse selection, acquisition costs and institutional structure. With some provisions for cost comparison problems (averages need not apply to sub-categories), we conclude that collective schemes are much cheaper than private schemes. From a cost-efficiency perspective, collective schemes are superior to private schemes. Furthermore this study shows that the operating costs of pension funds are strongly influenced by scale. The operating costs of small funds are more than ten times higher per participant than those of very large funds. Some employers and employees may deliberately opt for a small pension fund to obtain extra service and customisation (where pension schemes are designed to accommodate non-standard choices), but whether they are sufficiently aware of the resulting higher operating costs is open to question. More transparency to stakeholders about operating costs could help in this respect. The conclusion here is that the consolidation of small pension funds would lead to efficiency gains. Lastly, the above analysis has shown that some types of pension funds are more efficient than others (after controlling for economies of scale), even though the cost differences in view of the above comparison are modest. Industry-wide pension funds, particularly the mandatory ones, have significantly lower operating costs per participant than company and professional group pension funds. Standard pension schemes that are less generous and simpler to administer yield extra efficiency gains. In this context efficiency also refers to factors that cannot be influenced, such as the lower costs of value transfers for industry-wide pension funds. In addition, more extensive services can be provided to participants. The aforementioned efficiency gains cast a different light on the recent discussion about the desirability or undesirability of mandatory industry-wide pension funds and the possible expansion of the number of participants at this type of funds.
72
J.A. Bikker and J. de Dreu
Literature Bateman, H., G. Kingston and J. Piggot, Forced saving: Mandating private retirement incomes, Cambridge: Cambridge University Press, 2001. Bateman, H., O.S. Mitchell, ‘New evidence on pension plan design and administrative expenses: the Australian experience’, Journal of Pension Economics and Finance 3, 2004, pp. 63-76. Bikker, J.A., M. van Leuvensteijn, ‘Competition and efficiency in the Dutch life insurance industry’, Applied Economics, forthcoming, 2007. Bikker, J.A., P.J.G. Vlaar, ‘Conditional indexation in defined benefit pension plans’, Geneva Papers on Risk and Insurance, forthcoming, 2007. Bikker, J.A., J. de Dreu, ‘Pension Operating costs of pension funds: the impact of scale, governance and plan design’, Journal of Pension Economics and Finance, forthcoming, 2007. Bikker, J.A., L. Spierdijk, P.J. van der Sluis, ‘Market impact costs of institutional equity trades’, Journal of International Money and Finance, forthcoming, 2007. Caswell, J.W., ‘Economic efficiency in pension plan administration: A study of the construction Industry’, Journal of Risk and Insurance 4, 1976, pp. 257-273. CPB, ‘Competition in markets for life insurance’, CPB Document no. 96, The Hague: Netherlands Bureau for Economic Policy Analysis, 2005. Dobronogov, A. and M. Murthi, ‘Administrative fees and costs of mandatory private pensions in transition economies’, Journal of Pension Economics and Finance 4, 2005, pp. 31-55. James, E., J. Smalhout, D.Vittas, ‘Administrative costs and the organization of individual retirement account systems: A comparative perspective’, in: Holzmann, R. and J.E. Stiglitz (eds.) New ideas about old age security: Toward sustainable pension systems in the twenty-first century, Washington, DC: World Bank, 254-307/ Policy Research Working Paper Series no. 2554, World Bank, Washington DC, 2001. Jensen, M.C., ‘The performance of mutual funds in the period 1945-1964’, The Journal of Finance 23, 1968, pp. 389-416. Malkiel, B.G., ‘Returns from investing in equity mutual funds 1971 to 1991’, The Journal of Finance 50, 1995, pp. 549-572. Rooij, M.C.J. van, C.J.M. Kool, H.M. Prast, ‘Risk-return preferences in the pension domain: are people able to choose?’, Journal of Public Economics, forthcoming, 2007.
4 Operating costs of pension schemes
73
Appendix: Key data of pension funds and life insurers Table B.1. Key data of pension funds (1992-2004) Year
Number of funds, total
Number Number of Number of Total Number of Total costs/ Total costs of industry- company funds in assets, participants, total assets per particiwide penpension sample average average (%)b pant (€)a,b sion funds funds (€ million)a (1000)
1992
1131
82
1029
781
197
11
0.19
34
1993
1123
82
1021
820
209
11
0.18
34
1994
1111
82
1009
819
223
11
0.19
37
1995
1098
81
997
823
237
12
0.18
38
1996
1090
83
987
823
409
15
0.15
42
1997
1059
82
957
805
468
15
0.14
41
1998
1040
85
938
816
545
16
0.15
52
1999
1014
93
904
784
651
17
0.13
49
2000
986
92
877
791
658
17
0.14
51
2001
961
100
843
773
644
19
0.15
51
2002
924
102
804
727
613
22
0.18
51
2003
873
103
753
702
696
23
0.17
52
2004
841
104
718
655
826
25
0.15
48
a
In 2004 prices;
b
Weighted averages.
Source: DNB.
74
J.A. Bikker and J. de Dreu
Table B.2. Technical provisions and gross premiums of life insurers and pension funds (2004; € billion) Private
Collective
Total
Endowment insurance - insurers
9.0
0.9
9.9
Pension and annuity insurance
0.1
3.0
3.1
Total life insurersa
9.1
3.9
13.0
– of which for pensionsb
1.2
3.9
5.1
–
22.8
22.8
101.2
7.9
109.0
21.4
84.8
106.2
122.6
92.6
215.2
–
446.9
446.9
Gross premium
Insurers
Pension funds Technical provision Endowment insurance – insurers Pension and annuity insurance Insurers Total life insurers Pension funds a
Excluding annual deposits in savings banks;
Data from Statistics Netherlands (CBS), subject to differences in definition. Distributed over private and collective, according to the authors’ own judgment. Note that a proportion of the private endowment insurance policies includes an annuity clause and is intended for pension purposes. b
Source: DNB, Financial data life insurance companies; and Statistics Netherlands (CBS), National Accounts. Table B.3. Gross profits of life insurers (2004) Gross profits (€ million) Private insurance Collective insurance Totala a
Gross profits /gross premiums (%)
Gross profits / technical provisions (%)
1 192
12.3
1.0
624
15.9
0.7
1 816
13.9
0.9
Excluding the item ‘not to be categorised’.
Source: DNB, Financial data of life insurance companies.
5
Optimal risk-sharing in private and collective pension contracts
C.G.E. Boender, A.L. Bovenberg, S. van Hoogdalem, and Th.E. Nijman Pension solidarity can no longer be taken for granted. Due to demographic changes – and hence a growing retiree/employee ratio – additional contributions offer steadily fewer opportunities for clearing pension shortfalls. Together with the growing costs of contribution volatility and the trend towards short-termism, this means that the added value of solidarity is increasingly being called into question. A carefully argued and well-substantiated answer is therefore in order. What is the added value of solidarity and what is an ‘optimal’ pension contract? This contribution seeks to provide a survey of what we can learn about these issues from the current academic literature and to identify those areas where further in-depth research is warranted. The starting point consists of the private and collective pension contracts that are perceived to be optimal in the academic literature. However, the practical questions regarding pension funds and the economic environment in which pension funds operate are considerably more complex than assumed in the literature. Additional research is necessary to answer the central questions concerning the added value of pension solidarity and the optimal form of pension contracts. This contribution analyses how the assumptions and findings of the WRR study (Boender et al., 2000) relate to the customary assumptions in the academic literature. It specifies what we can learn from this about the added value of pension solidarity as calculated in that study. The insights in this contribution do not result in a single uniform answer regarding the exact added value of pension solidarity and the precise form of optimal pension contracts. Our aim here is rather to arrive at a number of concrete research questions in order to gain a deeper understanding of the underlying considerations and to be able to build a bridge in the near future between the academic literature and complex reality.
76
C.G.E. Boender, A.L. Bovenberg, S. van Hoogdalem, and Th.E. Nijman
5.1
Introduction
Solidarity in pension systems exists in many different guises.1 Several contributions in this book discuss examples of solidarity and seek to define what form of solidarity leads to the most desirable cost-benefit distribution in pension contracts. An important aspect is that the pension contract is designed to meet the preferences and circumstances of all individuals in an optimal way. This is often referred to as risk solidarity, where the pension contract is designed to protect the vulnerable elderly against a sudden loss of pension capital due to economic shocks. Young people are better able to absorb these shocks, because of their available human capital and longer investment horizon. In return, young people must receive an adequate reward for their role as shock absorber. This reward is all the more vital now that solidarity in the pension system is under pressure. One important cause of this pressure is the growing retiree/employee ratio. In the year 2006 pension assets in the Netherlands had already grown considerably above the value of the gross domestic product (GDP) and, when ageing reaches its peak, pension assets will be more than twice as large as the Dutch GDP. Assuming there is no structural decrease in annuity rates and returns in an ageing society and no further increase in estimated longevity, the pension system will not become more expensive due to ageing and a growing retiree/employee ratio, but it will become less risk-resistant. Back in the seventies and eighties, the national wage bill and GDP were so large compared to the accrued pension capital, that a pension capital loss of e.g. 10% could be easily made up for by charging the working population limited extra contributions. In the year 2006, however, this passing of responsibility to the employed would soon cost more than 10% of GDP, rising to over 20% of GDP in the year 2030. This is an important and objective cause of the growing doubts within society regarding the sustainability of the pension system. New pension regulations have highlighted this reduced risk-resistance, which is positive in itself. But this increased awareness has also divided opinions as to who should bear responsibility for the pension system’s greater vulnerability. The debate took on an ever sharper edge in the wake of the equity and bond slump at the start of the new millennium. The upshot, in short, is that solidarity is under pressure, particularly among the young who fear their current contributions are predicated on solidarity
1
An extensive description of the many different types can be found in the appendix of Chapter 2 in this volume, by Jan Kuné.
5 Optimal risk-sharing in private and collective pension contracts
77
principles that may no longer hold sway when they reach old age. For this reason, pension solidarity can no longer be taken for granted and must be shown to offer an economic win-win proposition for young and old alike.
PREVIOUS RESEARCH The value of pension solidarity in collective pension systems is demonstrated in a study carried out on behalf of the Scientific Council for Government Policy (WRR) (Boender et al., 2000). This report provides quantitative evidence that an individual within a collective scheme achieves a significantly better pension than an identical individual who is entirely responsible for making his own pension arrangements. Within the pension solidarity debate this report is often cited as an argument against switching to defined-contribution (DC) systems where all pension risks are offloaded onto the individual participants. However, the risk-sharing assumptions made in this report do not correspond with the pension policy that is demonstrated to be optimal in simplified theoretical models that are explicitly based on individual utility functions (see e.g. Teulings and De Vries, 2005). The outcomes of the WRR report have therefore attracted strong criticism, raising doubts as to whether pension solidarity yields genuine economic benefits. Clearly therefore, there is a great practical need for a more explicit understanding of what type of pension solidarity delivers what economic benefit under what assumptions. This contribution is a first step in that direction.
STRUCTURE OF THE ARGUMENT This chapter is built up as follows. In section 5.2 we will first establish what the optimal pension contract looks like according to the recent academic literature. The central question is: what investment and contribution decisions are optimal in the event of shocks on the financial markets and given the age of an individual? This section describes what pension policy is optimal if a number of simplifying assumptions are met. The assumptions of the theoretical model are not consistent with practical reality. The extent to which this is the case and the resulting consequences for, respectively, the optimal pension policy for an individual and a collective are discussed in sections 5.3 and 5.4. The optimal pension policy given the assumptions made in the academic literature diverges significantly from the outcomes of the aforementioned WRR study. With the assumptions applied in the WRR study, it is
78
C.G.E. Boender, A.L. Bovenberg, S. van Hoogdalem, and Th.E. Nijman
found that, notably, risk solidarity between generations delivers substantial economic benefits and that buffers have great added value. This solidarity comes about because people in the workforce – if necessary – pay extra contributions and temporarily accept underfunding in order to protect the pensions of the elderly against inflationary erosion. In return, the elderly put a buffer at the disposal of the young. However, with the assumptions made in the academic literature, the economic benefits of solidarity and of buffers is much smaller because the volatility of contributions is also included in the costs. This creates confusion over the actual importance of these two pillars within optimal pension contracts. For this reason, a number of assumptions underlying the academic literature and the WRR study are put under the microscope in section 5.5 to make a qualitative analysis of the consequences of these differences for the resulting valuation of risk solidarity. Quantification of the consequences of the differences in the applied assumptions is high on the research agenda. In section 5.6 this leads to a summary of the most important research questions. Section 5.7 outlines the principal conclusions.
5.2
Optimal risk-sharing in theory
A BENCHMARK This section describes how a pension fund, given certain simplifying assumptions, would provide an optimal pension service to its participants. An optimal pension policy for the individual consists of a combination of contribution, indexation and investment policy. To define the explicit characteristics of the optimal pension policy for the individual, it is assumed that the individual exclusively saves for retirement via the pension fund.2 The optimal pension fund policy in relation to the contribution, indexation and investments obviously depends strongly on the objective function of the individual. One common basic assumption in the literature is that an 2
For instance, the contributions to this book by Hoevenaars and Ponds (Chapter 6) as well as that by Boeijen et al. (Chapter 7) make precisely the opposite extreme assumption, namely that the individual has optimal access to the capital market and can and will trade all undesirable risks at no cost. Only the market value of the pension commitment is relevant in this case. The two approaches are complementary.
5 Optimal risk-sharing in private and collective pension contracts
79
individual maximises the utility of the expected consumption in each future year. Apart from taxation, consumption in the working period is equal to the salary less the pension savings and after retirement the level of consumption is determined by the pension. The utility increases in each period with the consumption, but the extra utility of an extra unit of consumption decreases with the level of consumption. Another basic assumption is that the more stable the development of this consumption, the greater the positive utility awarded by the individual to the future development of consumption. Moreover, an individual also discounts future consumption. This implies that the further in the future the consumption, the less utility it carries for the individual. For this reason in this model young people attribute relatively little utility to their consumption in retirement. The absolute value of the sensitivity of the marginal utility in relation to the consumption level is known as the relative risk aversion of the individual. Other frequently used basic assumptions are: • this relative risk aversion is constant. It does not depend on the level of consumption; • the contributions can be constantly optimally adjusted to new information; • interest rates and inflation are constant and equity returns have no memory (no ‘mean reversion’).
OUTCOME With these basic assumptions the optimal investment policy of an individual is surprisingly simple. At each moment the same portion of the total assets must be invested in equities, where the total assets consist of financial capital and human capital (the discounted value of future earnings3). The share of financial capital in the total assets increases with the individual’s age. In other words, with the passage of time the individual steadily converts his human capital into financial capital. According to this theory, the portion of the financial capital that is invested in equities decreases with age. A very simple example will clarify this. Assume that, according to this theory, an individual should invest 50% of his total assets in equities; and also that this young person’s total capital consists of 10% financial capital, whereas that of an old individual consists of 100% financial capital.
3
Uncertainty about future earnings is ignored in this simplest model.
80
C.G.E. Boender, A.L. Bovenberg, S. van Hoogdalem, and Th.E. Nijman
The young individual must then invest 500% in equities and the old individual 50%. One result of the optimal investment policy is that all cohorts (given equal risk preferences) lose the same percentage of consumption over the rest of their lives as a result of a negative shock on the financial markets. Assuming e.g. a 10% underperformance in any given year, this implies that the consumption and the pension during the active period and after retirement are reduced to such an extent that a fixed percentage is expected to be relinquished in each future year. In this pension model this reduction is the same for each age group. Young people invest more in equities, but spread lower- and higherthan-expected returns over a longer period. Elderly people invest less in equities and spread the results over a shorter period, in such a way that the pension consequences in relation to consumption are the same for everyone. The optimal investment behaviour in complete capital markets thus creates solidarity between the various age groups. Though young people suffer a larger loss in euro terms, they can also spread that loss over a longer horizon. They basically have longer to recover than the elderly. The economy’s loss of capital is spread as equally as possible over all age groups and also over each person’s remaining life. Younger generations entering the labour market do not yet have any financial capital. In this model, therefore, it is optimal for younger generations to borrow from the older generations in order to invest in risk-bearing capital. The optimal situation for the elderly is that the young do this by issuing indexed bonds to the elderly. In this way the young give the older generations the greatest possible certainty that they can enjoy an indexed pension, while the young profit at an early age from the risk premium on equities. In this theoretically optimal pension model, the young basically own an insurance company for the elderly. Put differently: they invest the pension capital of the elderly at their own risk and provide an indexed pension in return. The basis for risk-sharing can be even further expanded in this theoretical pension model by also including future generations in the risk-sharing mechanism. In this context, the term ‘future generations’ refers to generations who do not yet participate in the labour market, including generations who are not yet born when a shock on the financial markets occurs. Basically these generations are then already investing in the financial markets before they start paying contributions. This increases the opportunities for wealth-creating trading between the elderly and the young. The elderly are entirely dependent on their financial capital and therefore are more vulnerable to financial risks. The young can still use their human
5 Optimal risk-sharing in private and collective pension contracts
81
capital to absorb risks and thus benefit from the reward for risk. In this case risk-sharing takes place between non-overlapping generations: the shocks are borne not only by the cohorts who are alive when the risks occur, but also by cohorts who must still enter the fund in the future.
EXTENSIONS OF THE BASIC MODEL We will discuss a number of well-known extensions of the basic model before turning in the subsequent sections to explore how practical, implementable private and collective pension contracts relate to the recommendations made by the simplest theoretical model. In the simplifying assumptions underlying this optimal pension policy, equity returns have no ‘memory’, i.e.: there is no question of mean reversion. Mean reversion entails that equity returns become predictable up to a certain point, because the chance of higher returns increases as the period with lower returns lengthens. In these circumstances, equities are less risky over a longer horizon, because the returns average out to a certain extent over time (see, for instance, Siegel, 2002; Steehouwer, 2006). The differences in optimal investment policy between different age groups become stronger if mean reversion is taken into account. Older people often want to maintain the standard of living they enjoyed when younger, which makes them even less inclined to take investment risk. Above all, they want to minimise the risk of sinking below their accustomed – relatively high – level of consumption in times when returns are low. Young people will take even more risks than posited in the basic assumptions if they are not only able to adjust their contribution levels, but also the number of hours they work. Basically they have a larger stock of human capital, which they can use as a buffer in the financial markets. The assumption that older people place a smaller portion of their financial capital in equities is reasonably robust to changes in other assumptions in this theoretical pension model. As for young people, however, there are certain circumstances in which they should invest relatively little in equities: e.g. where, contrary to the basic assumptions, their salaried earnings are uncertain and strongly correlated to equity returns. In such cases, the human capital of young people already has much in common with a high-risk asset such as equities. In this theoretical pension model, they will therefore invest a smaller proportion of their financial capital in equities. Liquidity restrictions can also make young people risk-averse investors. These also play no role in the basic assumptions. If young people are unable to borrow against their human capital to adjust their con-
82
C.G.E. Boender, A.L. Bovenberg, S. van Hoogdalem, and Th.E. Nijman
sumption level to their expected future earnings, they will be inclined to take less risk - for this would have a direct adverse impact on their current consumption level and they would be unable to spread negative shocks over a longer period. The optimal portfolio for the elderly will contain a relatively large equity component if the elderly can count on a relatively high state pension that is not correlated with financial risks. Basically, the elderly then have a relatively certain pension claim via the state, so that they can afford to place a large part of their financial capital in high-risk investments. The optimal investment behaviour will vary not only across the life cycle but also between individuals. For the share of total assets placed in equities depends in part on individual risk preferences as well as on the nature of the human capital within a household. The pension income of or through a partner will also play a role, as will any assets held in addition to the pension capital.4
5.3
Private pension savings in practice
In practice private pension savings differ in several ways from the policy that is optimal given the assumptions in section 5.2. In the first place, the real economic environment confronting pension funds is much riskier and more dynamic than assumed in section 5.2. Consequently, pension contracts that are optimal in practice may diverge from the optimal pension contracts in section 5.2. In addition, more and more information is becoming available (Van Rooij et al., 2004) about how people value possible future developments concerning their pension savings and pension payments. Discrepancies may therefore be found between the utility functions assumed in section 5.2 and actual practice. This has obvious consequences for the pension contracts that are optimal in real-life conditions. Moreover, in contrast to the assumptions in section 5.2, ‘adverse selection’ and ‘moral hazard’ generally make it difficult, if not impossible, for young people to borrow against the value of their human capital. This limits their ability to benefit from the risk premium on equities.5 In addition, unborn generations are unable to trade with current generations. This basically entails the absence of a public market for trading risk, so that 4
A more extensive discussion of the basic model and the many ways in which it can be refined is provided in Bovenberg et al. (2007). This paper also presents additional empirical results on the value of risk sharing.
5
See e.g. Constantinides et al. (2002).
5 Optimal risk-sharing in private and collective pension contracts
83
young people are unable to take over the risks of elderly people in exchange for a reward. The risk-trading capabilities of capital markets are also limited in other ways. The market for index-linked loans, for instance, is at best embryonic, particularly in relation to the indexation of Dutch inflation. The same applies to the trading of longevity risk through longevity bonds. With this type of bonds the interest paid by the issuer increases with the percentage of people of a pre-determined age group who live longer than expected. There is also fundamental uncertainty regarding not only the set of possible outcomes, but also the objective probability distribution. Certain types of macro-economic shocks (such as political uncertainties) are inconceivable to us, let alone that negotiable products exist to insure such risks. In short: by no means can all risk factors be traded. Adverse selection also results in financial markets that are inadequate or even non-existent. The market for annuities, for instance, suffers from adverse selection because providers will try to bar relatively healthy elderly people from this market. Examples of non-existent markets are: • Borrowing by younger and even unborn generations using their human capital as collateral; • Sufficient availability of index-linked loans, notably for Dutch price inflation and industry-specific wage inflation; • Sufficient availability of longevity bonds; • Insurance products to protect against macro-economic shocks, such as political risk; • Financial guarantees, such as put options on stock exchange indexes, with a very long term of several decades; • Availability of complex derivatives strategies which are available to institutional investors, but not to individuals. In practice, an individual is also unable to approximate optimal investment behaviour because constant trading leads to excessive transaction costs or because certain markets are entirely closed to individuals and only accessible to large institutional investors (e.g. complex derivatives strategies that pension funds use to optimise performance). In addition, individuals often lack the expertise to save and invest rationally. People have difficulty making complex decisions in uncertain circumstances. The re-
84
C.G.E. Boender, A.L. Bovenberg, S. van Hoogdalem, and Th.E. Nijman
cent literature describes various ways in which individual investors make systematic errors and often do not know what is best for them.6 Buying professional knowledge is also problematic, because this market is often opaque and involves high transaction and marketing costs. Every percentage point of the invested capital that an individual loses annually as a result of these impediments and irrational behaviour translates into a decline in the pension payments of about 25% (see the contribution of Bikker and De Dreu in this volume). Suboptimal private pension management consequently implies a substantial deterioration of the pension result. In this light it is obviously of crucial importance to eliminate these impediments insofar as possible.
5.4
Collective pension funds in practice
CREATION OF HITHERTO NON-EXISTENT MARKETS Collective pension funds seek to overcome the imperfections of individual behaviour and the incomplete capital markets mentioned in the previous section. The more successful they are in achieving this objective, the closer actual practice can approximate the theoretically optimal pension contract. Pension funds can absorb these market imperfections in various ways. Through collective DB systems they organise risk-sharing between overlapping and non-overlapping generations that is not (yet) possible in capital markets. This takes the form of young people paying catch-up contributions as and when necessary. This opportunity for absorbing negative shocks enables the fund to take more risks in the investment portfolio and thus generate a higher return than would otherwise have been possible. Shocks in financial markets thus do not directly undermine the inflationproof nature of the pensions paid to pensioners. In collective DB systems, where pensions are linked to prices (wages), the young basically issue (wage-)indexed longevity bonds to older participants that are not yet for sale on the financial markets. Depending on the investment behaviour of the fund, the young invest the obtained funds at their own risk in the capital market. So here, the practical working of pension funds corresponds with the theory described in section 5.2, where the young are basically the shareholders of an insurance company for the elderly. Pension funds thus construct financial instruments (such as wage-indexed annuities with a 6
See e.g. Munnell and Sundèn (2004) and also Van Els et al. (Chapter 9) in this book.
5 Optimal risk-sharing in private and collective pension contracts
85
long term which cover the longevity risk) that are not yet (readily) available in the financial markets. By filling these gaps in the financial markets, risk-sharing between the generations can be made more efficient: the young share in the financial risks of the elderly and the elderly share in the wage risk of the young. Mandatory participation can further reinforce the funding base for intergenerational risk-sharing, so that future generations can also be involved in intergenerational risk trading. In addition pension funds can create value which is, in itself, separate from the selected degree of risk sharing. Economies of scale dampen the management and marketing costs, thus closing the gap between their contracts and the optimal pension contracts given the assumptions in section 5.2. Moreover, the funds provide employees with access to complex investment strategies that few individuals could use if left to their own devices. In addition, they also protect employees against unwise savings and investment decisions by offering professional asset management (see Van Els et al. in this volume). Their non-profit character boosts the confidence of participants in the fund’s policy: the participants are also the owners, so there are fewer conflicts of interest between pension fund management and participants. Finally, the mandatory participation of employees prevents adverse selection in the market for life insurance and annuities. All this reduces the implementation costs. Due to their close ties with the social partners (employer and employee representatives) as managers of the human capital in a sector, pension funds are also able to make optimal use of the buffer function of human capital in undertaking financial risks – for instance by enabling employees to take out loans against their human capital. Even if the financial markets start offering more risk-sharing instruments (especially wage-indexed bonds and longevity bonds), pension funds will continue to play a vital role in offering cheap, professional management of human - and, above all, financial capital, forcing people to save for retirement and preventing adverse selection in longevity risk insurance.
COSTS OF INSTITUTIONS Collective pension funds are not able to create all non-existent markets. This is mainly due to the fact that even if participation is made mandatory, young people can still evade this obligation by choosing to work in a different sector or company or as a self-employed person. They can also decide to work less. The greater the labour mobility and wage elasticity of the labour supply, the more catch-up contributions will distort labour market behaviour and be translated into compensating wage differences. In simple terms: in a labour market where people can change jobs quickly, it
86
C.G.E. Boender, A.L. Bovenberg, S. van Hoogdalem, and Th.E. Nijman
is more difficult to charge employees catch-up contributions as they will then simply defect to companies where no catch-up contributions are levied. This danger is smaller with larger collectives where it is less easy for participants to switch between collectives. Moreover, apart from voting with their feet, participants can also exert influence on the management of the fund. Delegating decision-making powers to a pension fund results in collective decision-making which always involves certain political risks. Older participants, for instance, are vulnerable to the risk that younger participants will refuse to pay large catch-up contributions. The greyer the fund, the greater this risk. The limited freedom of choice over the contributions and investments within collective systems protects participants against unwise decisions, but prevents these same participants from adapting individual behaviour to personal circumstances. The availability of pension benefits accrued by a partner, the nature of the human capital (is working longer an option?) or the risk attitude of the participant usually play no role in the pension contract. By offering more freedom of choice or by basing the pension on more information about the participant’s individual circumstances, collective pension funds could provide more customised pensions. This, however, has a price tag in the form of higher transaction and information costs. Most current collective systems (still) offer little in the way of customisation because they impose homogeneous contributions, investment portfolios and indexation on heterogeneous participants. Freedom of choice raises the danger of individuals making unwise choices. But, there too, delegating complex decisions to collective funds also inevitably creates extra transaction costs. These are incurred because the participants must be sure that the professional managers and investors who are acting on their behalf are genuinely working in their interest. Clear arrangements about governance, risk monitoring, and investment performance evaluation by a mandated supervisor are therefore of crucial importance.
5.5
Reconciling the WRR study and the theory
The WRR study (Boender et al., 2000) mentioned in section 5.1 quantifies the added value of solidarity that is realised by means of catch-up contributions and buffer formation. However, the added value of this solidarity and the optimal pension contracts based on this value do not correspond with the optimal pension policy, given the assumptions made in section 5.2. As a result, pension policy-makers are in the dark about one crucial point.
5 Optimal risk-sharing in private and collective pension contracts
87
This section therefore looks in greater detail at the differences between the starting points underlying the WRR study and the assumptions of the theoretical pension model from section 5.2. In addition, we will try to establish what consequences these differences have for the resulting optimal pension policy. A future study will seek to quantify the consequences of the differences in the applied assumptions. To permit a proper analysis of the differences between the assumptions of the WRR study and those of the academic literature, we will briefly describe the design and results of the WRR study: Design of WRR study • The study simulates the life cycle of an individual who starts saving an entirely self-managed pension at the age of 25 and compares this with an identical individual in a solidarity-based collective whose participants differ in age only. In the basic policy, both the individual and the collective pay a single premium which, given a fixed actuarial interest rate, is sufficient for a nominal 70% average-pay oldage pension and 49% survivor’s pension. • If financial market volatility (inflation, interest and returns) leads to a lower-than-expected pension accrual rate, both the individual and the collective adjust the contribution. The central control variable here is the funding ratio given a 4% discount rate. This implies that an individual is stronger than the collective at a young age and weaker at a later age. Within the context of the collective, this premium mechanism means that the elderly can continue benefiting from the equity risk premium via catch-up contributions. If necessary, these are paid for by the young. • Next, a follow-up policy is analysed for the individual, who takes less investment risk as he grows older. Specific policy variants are analysed for the collective, where buffers are built up and then passed on to subsequent generations. • The results are evaluated on the basis of a large number of stochastic scenarios whose characteristics (uncertainties, correlations, memory) are based on historical figures (1966-1998) for the applied factors (inflation, interest and return). The central return expectations in the long term are determined according to the insights applicable at the time.
88
C.G.E. Boender, A.L. Bovenberg, S. van Hoogdalem, and Th.E. Nijman
• The applied evaluation criteria are important. These concern, on the one hand, the average pension during the retirement period and, on the other hand, the downside deviation from the pension enjoyed in retirement compared to the 70% real average-pay pension with survivor’s pension. In plain language, this means that a twofold increase in the negative difference between the received and envisaged pension will be felt 2 x 2 = 4 times more severely by the individual.
RESULTS OF WRR STUDY With these starting points and evaluation criteria, it is found that an individual within a collective realises an approximately 30% better pension result than an identical individual who carries full responsibility for his own pension saving scheme. In other words: given the same average contributed pension capital and the same expected pension result, the individual within the collective fund runs 30% less pension risk during the retirement period than the identical individual outside the collective fund. It may happen (and actually does happen in the stochastic scenarios) that the collective fund takes an advance on the future by granting its participants indexation even when insufficient funds are available to continue doing this in the future. In that case the solidarity within the collective fund could start to crumble. This is not discounted in the calculation.
DIFFERENCES The first difference in the starting points applied in the WRR approach and the most common academic literature concerns the weighting of the contribution volatility. In the WRR approach this plays no role in the evaluation, while contribution fluctuations are relatively strongly penalised in the literature. As a result, the WRR approach assigns a higher expected added value to solidarity than the models in the literature. A second difference concerns the measurement of the difference between the envisaged and the realised pension. In the WRR study, the fact that the actual pension undershoots the target pension is penalised quadratically. This entails that a retiree receiving a pension that is lower than the 70% indexed average pay will feel this four times more severely if the shortfall were to double. This measure for the downside pension risks implies a specific form of aversion to loss of consumption levels during the pension (see Tversky and Kahneman, 1992). Only downside volatility is penalised where, due to the squaring operation, larger deviations carry
5 Optimal risk-sharing in private and collective pension contracts
89
relatively more weight. If people also valued upside deviations in the pension result, as in most utility functions applied in the literature, then the added value of solidarity will obviously decrease. The WRR study shows that it is more optimal for younger employees to invest in equities than for older employees or retirees. This advantage is mainly due to the assumed ‘mean reversion’ of equity returns and the longer time horizon. By implication, the cause does not lie in the greater propensity of young people to carry risk, because even in the collective system, shocks must be absorbed within one year. The WRR study assumes a uniform investment mix for the collective fund, irrespective of the participant’s age. The added value of solidarity as calculated in the WRR study is sensitive to divergent assumptions regarding the degree of mean reversion in equity returns, the incorporation of the risk characteristics of human capital and a longer recovery term for pension shortfalls. Finally, the WRR study applies a richer description of the financial market risks by including not only equity risks but also inflation and interest rate risks. These risk factors are not taken on board in the simple theoretical models. Thanks to the solidarity within collectives, risks such as inflation risk and longevity risk, which are difficult if not impossible to trade on financial markets, can be implicitly traded within the fund by the participants. This turns out to be an important determinant of the added value of collective contracts, which is probably why the simple academic models arrive at a lower added value for solidarity than the WRR study. By contrast, the fact that longevity risks were ignored in the WRR study means that theoretical models that do incorporate longevity risk can indicate in a higher added value of solidarity than estimated in the WRR study.
ADDED VALUE OF SOLIDARITY On the basis of the above comparison there is no way of telling in advance whether the added value of pension solidarity as calculated in the WRR study will be higher or lower if the assumptions are adjusted to correspond more closely to the most common assumptions in the academic literature. This is an important reason for carrying out a follow-up study. In addition, the WRR study needs to be deepened further even though it contained a richer description of economic uncertainty. Among other things, a negative weight must be assigned to the contribution volatility during the working life as also happens in the models from the literature. In addition, products that were not yet applied in 2000 in the pension world, such as derivatives, should be added to the analysis.
90
C.G.E. Boender, A.L. Bovenberg, S. van Hoogdalem, and Th.E. Nijman
5.6
Research agenda
The previous section showed that the WRR study must be enriched in certain areas. The relevant research questions focus on the way the added value of collective pension systems develops if 1. financial markets become more complete through the introduction of option contracts and indexed bonds or because participants can act dynamically and create certain options. In that case younger and older individuals can in principle also trade in risk via the financial markets without the intervention of a pension fund. The question remains whether individuals have the required expertise or can hire this at low costs. These new financial instruments also enable collective pension funds to control certain risks via negotiable securities in the financial markets instead of via implicit trading in non-negotiable claims within the pension funds between participants; 2. the preferences of the participants correspond more closely (or less closely) with the assumptions used in the academic literature. Recent experiences suggest that contribution volatility can be costly for the sponsor of a pension scheme. If this is the case, the optimal pension contract could shift from DB towards DC. By contrast, if participants are less concerned about contribution volatility and contribution pressure and attach more importance to keeping pensions inflation-proof, then the principle of inflation-proof pensions must be maintained insofar as possible (DNB, 2004). Further research into the preferences of participants in an ageing society is therefore of great importance; 3. the funding ratio on the basis of a 4% discount rate is replaced by a funding ratio at market value; 4. the labour market distortions caused by catch-up contributions are included in the analysis. Even in collective pension systems loans taken out by young people against the future value of their human capital are not entirely cost-free -- for young people are able to evade catch-up contributions by choosing to work less or elsewhere; 5. the possibility of a more flexible retirement age is included in the analysis. By utilising the retirement age as a risk buffer, people in the workforce can afford to take more investment risk and the optimal pension contract will thus change.
5 Optimal risk-sharing in private and collective pension contracts
91
In this case the participants take a risk in relation to their personal health (at age 40 one doesn’t know how healthy one will be at age 65), but disability insurance can be taken out to cover this risk; 6. an estimate of the difference in costs between private and collective pension contracts is taken on board. It is clear that the costs of collective contracts are considerably lower and will therefore become relatively more attractive; 7. realistic assumptions are made concerning the actual behaviour of individual decision-makers. It is well-known that individuals often save insufficiently for the future and tend to maintain undiversified investment portfolios, which impairs the quality of the pension result; 8. heterogeneity of the participants is assumed. Collective systems are generally characterised by an identical contribution and indexation rule for all participants, irrespective of their age, risk aversion, realised pension accrual, etc. This can easily lead to wealth loss as compared to the optimal scheme for the individual participant whose characteristics deviate strongly from the average participant; 9. pension contracts (policy ladders) are determined optimally. Above, we have already looked at the optimal risk-sharing contracts given the evaluation criteria used by Boender et al. (2000); 10. longer recovery terms become possible if the pension accrual rate is lower than expected due to unfavourable developments in the financial markets. This would mean that surpluses and shortfalls can be spread more evenly over the remaining life. Collective systems can thus add even more value and become less procyclical than when shortfalls are eliminated as quickly as possible (as is also the case in the WRR study); 11. risk solidarity in relation to longevity risk is included in the analysis. In this case, collective systems can add value as there are virtually no opportunities for trading longevity risk in financial markets.
5.7
Conclusions
Collective and private pension solutions differ in many dimensions. To permit a choice between the two and, above all, to further optimise the existing pension solutions, it is important to answer the central question: under what assumptions are specific private or collective pension contracts the most suitable and efficient way of achieving risk solidarity?
92
C.G.E. Boender, A.L. Bovenberg, S. van Hoogdalem, and Th.E. Nijman
Above all, this contribution sought to sum up the current state of affairs regarding this central research question in the literature. The adopted approach was to test the reality value of the assumptions underlying pension contracts that are assumed to be optimal in the literature. The conclusion is that reality is so much more complex and complete than the assumptions made in the literature and the WRR study, that extra research is necessary to answer the central question regarding the added value of pension solidarity. In addition we also analysed how the assumptions and findings of Boender et al. (2000) relate to the assumptions that are more common in the academic literature, and what we can learn from this regarding the added value of pension solidarity as calculated in this study. The table below compares the strengths and weaknesses of private and collective pension schemes as discussed in sections 5.3 and 5.4. Table 1. Strengths and weaknesses of private and collective pension schemes Strengths Private pension saving
Weaknesses
• Individual customisation of investment and contribution policy
• Suboptimal choices due to low pension awareness
• Competition between providers
• Suboptimal choices due to behavioural effects • Adverse selection • Not all financial products can be accessed
Collective pension savings
• Creates hitherto non-existent markets
• Continuity of solidarity not guaranteed
• Enables young people to take much more risk (elimination of restrictive conditions)
• Not geared to heterogeneous participants
• Low costs
• Ownership rights are not transparent
• Professional investors
This study does not produce full answers to the central question regarding the value of pension solidarity. This is because private and collective pension schemes both have strengths and weaknesses (see table) and because the assumptions underlying the assertions about the optimal pension contract as made in the literature may still be too far from reality. Instead, this study provides a list of concrete questions to be addressed in follow-up research. This will seek to gain a deeper understanding of the underlying considerations and thus provide a more well-founded basis for the further optimisation of pension contracts.
5 Optimal risk-sharing in private and collective pension contracts
93
Literature Boender, C.G.E., S. van Hoogdalem, E. van Lochem and R.M.A. Jansweijer, Intergenerationele solidariteit en individualiteit in de tweede pensioenpijler: Een scenario-analyse, WRR (Scientific Council for Government Policy) report 114, The Hague: WRR, 2000. Bovenberg, A.L., R.S.J. Koijen, Th.E. Nijman and C. Teulings, ‘Saving and investing over the life cycle and the role of collective pension funds’, Netspar panel paper, Tilburg University, 2007. Constantinides, G., J. Donaldson, and R. Mehra, ‘Junior cannot borrow. A new perspective on the equity premium puzzle’, Quarterly Journal of Economics, pp. 269-296, 2002. Cui, J, F. de Jong and E. Ponds, The value of intergenerational transfers within funded pension system, Working paper presented at the 4th RTN Workshop on ‘Financing Retirement in Europe: Public Sector Reform and Financial Market Development’, Louvain, België, 2005. Munnel, A. and M. Sundèn, Coming Up Short: The Challenge of 401(k) Plans, Washington: The Brookings Institution, 2004. Rooij, M.C.J. van, C.J.M. Kool and H. Prast, Risk-return preferences in the pension domain: are people able to choose?, DNB working paper 25, Amsterdam: De Nederlandsche Bank, 2004. Siegel, J,J., Stocks for the Long Run, New York: McGraw-Hill, 2004. Steehouwer, H., Macroeconomic Scenarios and Reality, thesis, VU University Amsterdam, 2005. Teulings, C.N. and C.G. de Vries, ‘Generational Accounting, Solidarity and Pension Losses’, De Economist, 154 (1), pp. 63-83, 2006. Tversky, A. and D. Kahneman, ‘Advances in Prospect Theory: Cumulative Representation of Uncertainty’, Journal of Risk and Uncertainty, 1992, pp. 297-323.
6
Intergenerational value transfers within an industry-wide pension fund – a value-based ALM analysis
R.P.M.M. Hoevenaars and E.H.M. Ponds1 Intergenerational solidarity is an important topic in the increasing interest in collective pension schemes. How great is this solidarity? Is there a balanced sharing of costs and benefits across age cohorts? The long-term sustainability of any pension scheme stands or falls by the willingness of members to continue to participate; the attitude of younger persons is crucial in this regard. In this chapter we set out a method by which we can illustrate the way in which the value transfer between generations within an industry-wide pension fund occurs. This method – which we term value-based generational accounting – is ideally suited to investigating how far current policy itself, and changes to that policy, result in a balanced sharing of costs, benefits and risks across the generations participating in the pension fund. The method thereby also forms a good basis for justifying (in advance and in retrospect) the policy that is pursued. We begin the chapter by explaining the method of value-based generational accounting. We deduce from this that a pension fund can be characterised as a ‘zero-sum game’. A change in policy does not create extra value, but does result in a redistribution of value between the parties involved in the pension fund. We then examine the generational effects for a standard industry-wide pension fund the pension fund policy regarding investments, contribution rate setting and indexation policy. We pay no attention on transfers between members as a consequence of the operation of the uniform contribution rate. We regard this practice as a given. The contribution by Boeijen et al. in this book deals specifically with the pay-as-you-go 1
We owe a debt of gratitude to Niels Kortleve for our many discussions on this topic, to Roderick Molenaar for the construction of the deflator set and to Alexander Paulis and Jo Speck for the allocation of the actuarial cash flows to the various generations.
96
R.P.M.M. Hoevenaars and E.H.M. Ponds
element from younger to older employees, making use of the technique explained in that chapter. In addition, value transfers can also occur within a cohort. This topic is the focus of the contribution of Aarssen and Kuipers in this book. It is our view that the proposed method is a valuable addition in the evaluation of current policy and policy variations. The approach of value-based generational accounting should therefore form a part of the decision process regarding the financing policy of the fund. This can prevent undesirable and/or unintended value transfers between generations. The proposed method can assist in searching for a set of policy parameters whereby transfers do not take place, or if they do, they are of acceptable size.
6.1
Introduction
In this chapter we set out the method of value-based generational accounting. This method gives us an insight into the size and direction of value transfers between generations within the pension fund in the event of a policy change. The aim of this method is to evaluate the current financing structure or changes in this structure from the angle of a balanced sharing of costs and benefits across generations. Intergenerational value transfers are the result of risk sharing between generations within an industry-wide pension fund. Recent studies reveal that intergenerational risk sharing within a pension fund is welfare-improving for the plan members vis-à-vis an optimal indiviual pension plan, even under ideal market conditions (Cui et al. 2006, Gollier 2006, Teulings and De Vries 2006). The analysis of transfers of value between generations is based on generational accounts. A generational account can be formulated for each age cohort within a pension fund. A pension fund faces an uncertain future. Each projection of benefits and contributions is therefore shrouded in uncertainty. In formulating generational accounts we need to bear such uncertainty in mind. ABP and PGGM have worked together to develop the method of ‘value-based generational accounting’ (Kortleve, 2003, Ponds, 2003, Kortleve and Ponds, 2006). The generational accounts method is supplemental to the current ALM analysis. A classic ALM study helps to answer the question how realistic and/or desirable a policy variant is in terms of, for example, level of contribution rate, indexation allowed and the development of the financial position and the uncertainties involved. Value-based generational accounting makes clear how changes in the financing structure and the risk allocation specified in the pension contract can lead to value transfers between generations.
6 Intergenerational value transfers
97
The structure of this chapter is as follows. We first explain the method of value-based generational accounting. We then discuss various policy changes for a stylised industry-wide pension fund and examine the possible value transfers between members thereby involved.
6.2
Explanation of the ‘value-based generational accounting’ method
6.2.1 PENSION FUND AS ZERO-SUM GAME A pension fund is a structure of stakeholders with diverging interests. The content of the pension contract determines the value of the claim of each of the stakeholders on the assets of the fund. This claim is a combination of accrued rights, indexation terms, premium contributions and claims on the fund’s buffer. The sum of all claims is always equal to the value of the assets in the fund, since the total value of all claims of members at any one time can never be greater or smaller than the total assets in the fund at that time. A change in the financing structure (adjustment of the asset mix or contribution methodology) or a change in the rules of risk allocation between members (for example, a change to the rules regarding conditional indexation or a change in the tempo of catch-up indexation) will result in the claims of interested parties changing in value. As the total value of claims at the moment that a policy is altered does not change, the pension fund can therefore in this context be characterised as a zero-sum game. The assets do not increase or decrease as a result of a change in policy, but a change in policy usually will lead to a change in value of the individual claims of interested parties. Since the pension fund is a zero-sum, then what one party gains through an increase in value will be at the expense of one or more other parties suffering a loss in value. Policy changes lead to value transfers between the members. In a company pension scheme value transfers are mostly transfers between the company’s shareholders on the one hand and the plan members of the pension fund on the other hand (Chapman et al. 2006, Steenkamp 1998). In the case of an industry-wide pension fund we are dealing principally with value transfers between generations (age cohorts). These value transfers can now be analysed using the value-based generational accounting method. This method is, in fact, a combination of generational accounting and value-based ALM. We explain this in the following sections.
98
R.P.M.M. Hoevenaars and E.H.M. Ponds
6.2.2 GENERATIONAL ACCOUNTS Generational accounting is a method developed by public finance economists as a tool for investigating the intergenerational distributional effects of fiscal policy (Auerbach et al., 1999, Kotlikoff 2002). Generational accounting is based on the government’s intertemporal budget constraint, which requires that either current or future generations pay for government spending via taxation. The government’s net wealth (including debt) plus the present value of the government’s net receipts from all current and future generations, must be sufficient to pay for the present value of the government’s current and future consumption. The generational accounting method can be employed for calculating the present value changes in net life-time income of both living and future generations that result from changes in fiscal policy. Generational accounting reveals the zero-sum feature of the intertemporal budget constraint of government finance. In other words, what some generations receive as an increase in net lifetime income must be paid for by other generations who will experience a decrease in net lifetime income. Planned increase or decrease in government debt can be used for tax smoothing over time in order to realize a sustainable fiscal policy (Auerbach et al. 1999, Van Ewijk et al., 2006). Similarly, the method of generational accounting may be of use in evaluating the policy of pension funds to cover both current and future participants. Two similarities with public finance can be discerned. Firstly, pension funds also face an intertemporal budget constraint, as the promised benefits must be backed by current and future contributions and returns on paid contributions. Secondly, as the government uses the tax instrument to close the budget over time, adjustments in contribution and indexation rates are the instruments used by pension funds to square the balance over time.
6.2.3 UNCERTAINTY Generational accounts focus on the long term. Future projections are thereby shrouded in a great deal of uncertainty. We must take this uncertainty into consideration when valuing future benefits and costs. In applying generational accounting to government finances, economists face the problem of how to deal with these uncertainties. Kotlikoff, the one of the godfathers of generational accounting, explains this problem as follows: ‘In the realistic case in which countries’ tax revenues and expenditures are uncertain, discerning the correct discount rate is even more difficult. In this case, discounting based on the term structure of risk-free rate is no longer theo-
6 Intergenerational value transfers
99
retically justified. Instead, the appropriate discount rates would be those that adjust for the riskiness of the stream in question. Since the riskiness of taxes, spending and transfer payments presumably differ, the theoretically appropriate risk-adjusted rates at which to discount taxes, spending and transfer payments would also differ. (…) Unfortunately, the size of these risk adjustments remains a topic for future research. In the meantime, generational accountants have simply chosen to estimate generational accounts for a range of discount rates.’ (Kotlikoff, 2002). Virtually all studies in the field of government finances therefore in the first instance assume a world of certainty, and then analyse uncertainty based on a sensitivity analysis for alternative core parameters, including the discount rate applied to the present value calculation. The recent study by the CPB Netherlands Bureau for Economic Policy Analysis on the longer-term development of Dutch government finances is a good example of this (Van Ewijk et al. 2006). An analysis of uncertainty that is better than this sensitivity analysis of the discount rate is an analysis that uses stochastic discount rates (also termed deflators). This has recently received a lot of attention in the literature (Cochrane, 2001; De Jong, 2004; Ang and Bekaert, 2004; and Nijman and Koijen, 2006). A feature of stochastic discounting is that favourable economic variants in calculating current value are weighted less than unfavourable economic variants. Individuals and policy makers avoid risk and in a poor economic climate attach greater value to payments and revenues than they do when the economy is flourishing. Through the use of stochastic discount rates, the risk aversion is amply reflected in the valuation of uncertain cash flows, so that in valuing future cash flows there is a correction made for risk. This approach is much used in the valuation of insurance contracts and derivatives traded on the financial markets, such as options and futures. Cash flows out of and into the government, as well as pension funds, can also be valued as derivatives, since in the case of government finance, the payment out of tax revenues and government expenditure are endogenous to a number of underlying fundamental factors in the economy, such as growth and inflation. With regard to a pension fund, indexation payments in pension contracts depend on underlying variables such as investment returns and inflation. Such agreements in financial contracts can therefore be seen as derivatives that are dependant upon the underlying variables. Stochastic discounting with the help of deflators2 has found 2
Although derivatives are in general valued as risk-neutral, the deflator approach is more useful in the valuing non-tradables such as taxes and indexation. The
100
R.P.M.M. Hoevenaars and E.H.M. Ponds
acceptance in the context of pension funds with the development of the socalled value-based ALM, which we discuss below.
6.2.4 FROM CLASSIC ALM TO VALUE-BASED ALM The usual Asset Liability Management (ALM) studies of pension funds, which we label classic ALM analysis, are intended mostly to identify the risk distribution of significant pension fund variables. The uncertainty surrounding future variables are analysed, and choices are then made with regard to the fund policy to be adopted on the basis of an evaluation of the expected values and risk for the current policy and alternative policies. An evaluation of pension fund policy on the basis of risk distribution alone provides insight into the possible value transfers between generations. An understanding of these transfers with the aid of generational accounts is important in order to prevent unintended and/or undesirable transfers. Such transfers can be the result of an imbalanced distribution of burdens (risks, contributions) and benefits (undertakings, indexation) amongst the interested parties. Partly as a result of discussions between British actuaries, we have seen the development of value-based ALM studies alongside the classic ALM studies (Chapman et al. 2001). This development means that the policy of a pension fund is valued not only on the basis of risk distributions, but also in terms of economic value. A value-based ALM applies the stochastic discount rates discussed above. The approach here is as follows: by determining all contributions, benefits and investment returns for each future variant and by calculating current value using the stochastic discount rate applicable to each variant, we arrive at the current economic value of these cash flows. ‘Economic value’ means here the cash value in euro’s at today’s date of uncertain future cash flows. The extending of the classic ALM analysis to a value-based ALM and the technique of generational accounts therefore potentially offers a solution to the problem of how to deal with uncertainty surrounding the results of generational accounts. Classic ALM and value-based ALM are therefore both extremely useful. Classic ALM allows us to calculate how a policy variant performs under significant variables such as expected values and risks concerning the level of contributions, indexation and the stochastic discount rate varies with the underlying stochastic risk factors such as investment returns and inflation. This means that the deflator is dependent on the risk in the payment out, and therefore more suitable for the valuation of risky payments than a fixed discount rate.
6 Intergenerational value transfers
101
development of a financial position for the fund as a whole. Value-based ALM allows us to see the economic value in euro’s of participation in the fund by various members, and how large the possible value transfers between members of a pension fund are.
6.2.5 EXPLICT PENSION CONTRACT AS A BASIS FOR VALUE-BASED ALM In valuations applying value-based ALM techniques it is important that benefits and costs for the members are defined explicitly. Until recently, pension contracts for Dutch pension funds were characterised in practice by poorly-formulated ownership rights, especially with regard to the question of who shares in the gains and losses in the funding process, when and in what proportion, and who owns the funding surplus or shortfall (overfunding or underfunding) of the pension fund. In the nineties, this led to a lot of discussion between stakeholders regarding the allocation of what at the time was regarded as a high level of overfunding. The discussion centred on whether this surplus should be used for contribution holidays, extra indexation or the financing of early retirement. During the pensions crisis of a few years ago, it was not clear which party was responsible for making up for the underfunding. Partly as a result of this problem, pension contracts have since then been drawn up in more explicit detail. Over recent years, a large number of pension funds have implemented policy ladders with explicit rules governing indexation allowances and conditions relating to contributions. This makes it clear how the relevant parties share gains and losses. Based on such policy ladders it is possible to arrive at explicit economic values for contributions to be paid and indexation to be received. On the other hand, many funds have not yet properly regulated the allocation of the surplus. An important element of the value-based generational accounting method is that there must here be a explicit contract. It is therefore important to formulate a closure rule regarding the question to whom, when and in what proportion a surplus can be allocated. This study assumes as closure rule that at any point in time the surplus is to be allocated amongst the members in proportion to their nominal claims. Alternative closure rules are of course possible, such as allocation in proportion to the actual value of the pension commitments or in proportion to the amount of contribution. These alternatives would not lead to significantly different results. Thanks to the move to draw up explicit rules governing the allocation of risks and the funding surplus or shortfall, the pension contracts of Dutch
102
R.P.M.M. Hoevenaars and E.H.M. Ponds
pension funds have, in fact, been transformed into financial contracts and the claims by members of these pension contracts can be valued as if they are being traded on financial markets (Kocken 2006, Frijns, 2006). As Dutch pension funds have moved to explicit pension contracts, valuebased ALM can be applied.
6.2.6 VALUE-BASED GENERATIONAL ACCOUNTING Basic variant The set up below shows how a generational account is drawn up (refer to Hoevenaars & Ponds (2007) for more details). This calculation can be applied to pension funds with intergenerational solidarity. The set up covers the period 2006-2025 for the generation born in the year ‘19xx’. Here, all terms are expressed in euro as at 2006. Using stochastic discount rates, the value of all cash flows over the period 2006-2025 is calculated back to 2006, including an adjustment for the risk that is implied in these cash flows. In general, this means: the more volatile the cash flow is expected to be, the higher the discount rate and the lower the current cash value. In fact, these terms are the current prices (option premiums) that the market would be willing to pay for risky future cash flows. Risk adjustment means that the risk aversion of the market is reflected in the valuation of the cash flows. Cash flows in poor economic climates are valued more heavily than cash flows in buoyant economic conditions.3 Generational account for age cohort 19xx (in euro as at 2006) = +
Value of accrued benefits in 2006
+
Increase in benefits due to new accrual and indexation 2006 - 2025
+
paid-out benefits 2006-2025-/- written-off accrued benefits 2006-2025
-/-
Contributions to be paid 2006-2025
+
Claim on surplus in 2025 -/- Claim on surplus in 2006
3
In this chapter we do not deal with the technique of valuing in line with market conditions with a correction for risk. We merely indicate that the ALM projections are based on a vector autoregressive model (see Hoevenaars et al. 2005) and that an economic valuation of these projects makes use of a pricing kernel (see Nijman and Koijen, 2006 and De Jong, 2005). See also Hoevenaars and Ponds (2006) for a description of the techniques underlying value-based generational accounting.
6 Intergenerational value transfers
103
The generational account is the sum of a number of components. The first component is the value of accrued benefits of the relevant age cohort. Over the next 20 years, these accrued benefits increase as a result of further accrual and indexation. This increase in accrued benefits enlarges the value of the generational account. The contributions payable by the generation are a minus item in the generational account. In addition, the cohort receives benefits over the period in question. As a rule, the actual benefits will tie in with the written-off benefit obligations that can be forecast actuarially. In line with the closure rule we have assumed for the calculations that the generations have a claim upon the funding residu in proportion to the value of their nominal claims. This interpretation of the ownership right to the surplus means that the funding balance at the end of 2025 will be allocated to the cohorts in proportion to the value of their nominal claims at that time. This increases the value of the generational account. On the other hand, the funding residu at the beginning of the period is at the expense of the generational account. The right of ownership of the funding balance as at 2006 is, as it were, surrendered, in exchange for a right of ownership of the funding balance as at 2025. Policy variant We can also draw up such generational accounts for a policy variant, i.e. for an alternative pension contract. The generational accounts can then be compared with each other. This is shown in Table 1. The first column indicates the age of the cohorts in 2006. The oldest generation is 105 years. The youngest generation will be born in two years’ time and will reach the age of 18 in 20 years’ time. Members of this generation will then become pension fund members for the first time. The second column indicates the generational accounts for the basic variant, i.e. existing policy. The third column indicates the results of the generational accounts for an alternative policy variant. The fourth column shows the difference between the results of the generational accounts for the alternative, and those for the basic variant. A very important characteristic is that these differences add up to zero. This reflects the notion that in terms of economic values the pension fund is a ‘zero-sum game’: a change to policy may lead to a positive value effect for one age cohort, but has a negative value effect of the same order on at least or more age cohorts.
104
R.P.M.M. Hoevenaars and E.H.M. Ponds
Table 1. Outline of generational accounts
Age – 2yrs.
Generational account standard variant X
Generational account alternative variant Y
Increase /decrease
X-2
Y-2
X-2 – Y-2
X30
Y30
X30 – Y30
X60
Y60
X30 – Y60
X105
Y105
X105 – Y105
Sum = 0
Sum = 0
Sum = 0
… Age 30yrs. … Age 60yrs. … Age 105yrs
6.3
Analysis using value-based generational accounting
6.3.1 BASIC VARIANT FINANCING STRUCTURE We have carried out a classic ALM study for a standard industry-wide pension fund with a conditionally index-linked average-wage scheme. We calculated over a 20-year period from 2006 to 2025. We did this for 5,000 variants. Then, based on the results, we drew up the generational account for each age cohort using the value-based generational accounting method. The pension fund’s investment portfolio consists of 40% equities, 40% bonds and 20% alternative investments (including commodities, hedge funds and private equity). Figure 1 shows the policy ladder operated by the pension fund. The contribution rate is stable and is fixed on the basis of the real return (prudently estimated) on the investment portfolio. The indexation policy is conditional. There is no indexation if the funding ratio is less than 100%. If the funding ratio is greater than 140% then there is full indexation. For a funding ratio between 100% and 140% there will be partial indexation according to the formula: (FR-100%)/140%-100%) x wage increase, where FR stands for funding ratio. Figure 2 shows the generational accounts. The horizontal axis represents the age of the relevant cohorts in 2006. The vertical axis represents the economic value of the generational accounts, expressed as a percentage of the total fund liabilities in 2006. The sum of all generational accounts is zero. The calculations all assume that the funding ratio for the pension fund at
6 Intergenerational value transfers
105
Costprice contribution rate
Full indexation
indexation
0% 100%
Funding ratio
140%
Fig. 1. Policy ladder – basic variant 1.5%
% total liabilities
1.0%
0.5%
0.0%
-0.5%
-1.0%
-1.5% -2
3
8
13
18
23
28
33
38
43
48
53
58
63
68
73
78
83
88
93
98 103
age cohort
Fig. 2. Generational accounts in basic variant as % of total liabilities in 2006
the beginning of 2006 is equal to 120%. According to the ladder, therefore, there is a cut back of the indexation at the start. The results of the generational accounts are heavily determined by the use of the uniform contribution rate. All employees pay the same uniform contribution as a percentage of their pensionable salary, but the value of the rights that are acquired for any year of service depends on age. For a 48-year-old the value of the new sum accrued is virtually the same as the contribution that this member pays in. The value of rights of employees under the age of 48 is well below the amount of contribution they pay, whereas the older members accrue many more rights than the contribution
106
R.P.M.M. Hoevenaars and E.H.M. Ponds
they pay. The contribution from Boeijen et al. in this book deals with this subject in detail. A 48-year-old has the highest positive generational account. For almost 20 years this member has profited from the fact that the value of the new sum accrued over this period is larger than the contributions paid over the same period. A 25-year-old has the largest negative generational account. For this member, the difference between the value of the contributions and the value of the sum accrued is the biggest. For a member aged 38, the generational account is approximately zero. In the first 10 years the value of the sum accrued remains less than the amount of the contributions, whilst in the second half of the period this is exactly compensated for by the accrued sum that exceeds the value of the contributions made. Those in retirement have a slightly negative generational account. This reflects the uncertainty surrounding future indexation. The question that arises is whether we can show on the basis of figure 2 whether the current pension contract is fair or not to the relevant parties. The answer is no: we are unable to make any judgment on the degree of fairness. Firstly, we are looking 20 years into the future and the history of the members’ contribution payments and indexation rates is not included in the analysis. We need this additional information to obtain a broad picture of the total contributions paid by the generations and the total benefits being received by the generations. Secondly, the approach does not include welfare aspects regarding the pension scheme in the analysis. Membership of the pension fund’s pension scheme gives the prospect of pension income being related to wage developments. This thereby provides an insurance against risks concerning future inflation and actual growth in income, whereby we have the prospect that on reaching pensionable age our standard of living is related to the standard of living we enjoyed before reaching such age. This form of insurance cannot be purchased on the financial markets. It may be that a generation has to accept a loss in value terms, but that nevertheless – in welfare terms – there is a positive-sum game (cf. Cui et al. 2006). For welfare analysis we have to apply a utilitybased ALM, but for the time being this is not useful yet for practical applications.4
4
The academic literature contains a number of initiatives for formulating an objective function for a pension fund and for utilising this in formulating an optimum policy. This approach, which we call ‘utility-based ALM’ can, however, as yet only be applied to a very stylised characterisation of an actual pension fund.
6 Intergenerational value transfers
107
Value-based generational accounting is not therefore suitable for issues concerning fairness and wealth analysis. The method is primarily suited to identifying value transfers between generations resulting from changes in policy. This is the subject of the following sections.
6.3.2 POLICY ALTERNATIVES In this section we study a number of policy variants, in particular the effects in terms of generational accounts when there is a switch from the existing financing structure to these policy variants. We begin by noting the results of the classic ALM analysis for a number of core criteria. Then, we analyse the consequences of a change in the policy for the generational accounts. The effect on the generational account of a policy change can be split into the following components: Change in generational account for x through policy change = + Change in value of accrued benefits of cohort x for 2006-2025 – Change in value of contributions by cohort x for 2006-2025 +
Change in value of the claim on the surplus of cohort x in 2025
Here, the total across all the cohorts must always be zero, as we showed in section 6.2 above, since the pension fund is a zero-sum game in terms of economic value. The total value at any moment is equal to the value of assets at that moment. A change in policy does not result in more or less wealth in the pension fund. As a rule, it will, however, lead to a redistribution of risk between the members and therefore a redistribution of the value. We will examine three policy changes: 1. a change in investment policy; 2. a transfer to a DB final-salary scheme with unconditional indexation and variable contribution; 3. a switch to a collective DC scheme. Changes to investment policy We will look at two variants of investment policy: a switch to 100% bonds and a switch to 100% equities. We will first discuss the results of the classic ALM analysis, and then the effects for generations in value terms.
108
R.P.M.M. Hoevenaars and E.H.M. Ponds
Classic ALM The following table shows the results for a number of core variables of a classic ALM analysis of the existing policy and two policy variants. Table 2. Results of classic ALM analysis for investment variants FR* 2024
FR < 105% 2005 – 2024
Median
Contribution 2005 – 2024
Indexation as % of wage 2005 – 2024
Median
Median
<80%
Standard variant
161%
0.4%
18%
98%
22%
100% bonds
150%
0.0%
34%
98%
11%
100% equities
158%
8.3%
16%
98%
33%
*FR = Funding Ratio
The median of the funding ratio (FR) in the standard variant is 161% in 2025, well above the upper limit for the policy table of 140%. There is, however, a spread around this result. For example, the risk of underfunding (funding ratio less than 105%) over the period 2006-2025 is on average 0.4%. The contribution is 18%, based on a fixed discount rate of 3% (prudently estimated real return on the investment portfolio). Finally we indicate the results of the standard variant with regard to indexation. The median of the allocated indexation is 98% of what was promised. There is a wide spread around this result, since the indexation is related via the indexation table to the funding ratio. The spread in the funding ratio means that in 22% of cases the indexation is less than 80% of the full indexation. The switch to 100% bonds firstly results in the need to adjust the contribution rate significantly upwards. Investment in bonds is anticipated to lead to a lower nominal return. What is lost in terms of investment returns must be recouped via higher contributions. In this investment variant the contribution increases from 18% to 34%. The median of the funding ratio is 151%, well above the upper limit of the indexation table. The risk of underfunding is thereby reduced to 0%. Whilst a mix of 100% bonds does lead to a low return, it also implies little investment risk and therefore little funding ratio risk. A smaller funding ratio risk will also lead to a smaller spread in indexation. the risk of given indexation being less than 80% of full indexation decreases by half in the case of the standard variant to 11%.
6 Intergenerational value transfers
109
The switch to 100% equities reveals a completely opposite picture to the switch to 100% bonds. In the standard variant the contribution is here reduced to 16% due to the return on 100% equities being somewhat higher than the return on the strategic mix. The median for the funding ratio is now 158%, somewhat lower than in the standard variant. The 100% share mix is very risky. Accordingly, the spread around the median of the funding ratio is considerable. The risk of underfunding thereby increases dramatically, from 0.4% in the standard variant to over 8% for the 100% equity mix. Consequently, the spread of the indexation is also significantly greater. The risk of the indexation being less than 80% of the full indexation increases in the case of the standard variant from 22% to 33%. Value-based generational accounting What are the consequences for the generations of these changes to investment policy? Figure 3 shows the effects of the switch to 100% bonds. The horizontal axis represents the age of the cohorts 2006. The oldest member is aged 105, but it is certain that one year later he will be no longer in the pool. The youngest member is aged –2 and is still included in the comparison, because this cohort will be aged 18 in 2025 and then be in the pool of members. The vertical axis represents the change in the value of the generational accounts as a percentage of the value of the liabilities in 2006. 2.5% 2.0%
% total liabilities
1.5% 1.0% 0.5% 0.0% -0.5% -1.0% -1.5% -2.0% -2
3
8
13 18 23 28 33 38 43 48 53 58 63 68 73 78 83 88 93 98 103 age cohort
Change in generational accounts
Change in value (accrued benefits - contributions)
Change in value of claim to surplus 2025
Fig. 3. Effects for generations on switch to 100% bonds as % of total liabilities 2006
110
R.P.M.M. Hoevenaars and E.H.M. Ponds
100% bonds We begin by looking at the effect for the change in the generational accounts, represented by the black line in figure 3. The picture shows that whilst young employees contribute value, the older employees and those receiving pension are receiving value. Figure 3 also shows how this comes about. The grey line represents the change in value of the accrued benefits less the change in value of the contribution paid in. Employees face a significance increase in the contribution. This leads to a loss in value compared to the standard variant. Those receiving a pension benefit slightly more. This can be explained by the fact that this investment strategy gives less spread in the funding ratio and thereby leads to greater certainty with regard to the indexation allocated. This effect also affects employees, but this gain in value far from balances out the loss value due to the higher contribution payments. The interrupted line represents the change in the claim to the surplus at the end of 2025. This variant does not in itself lead to a higher average overfunding compared to the standard variant. Furthermore, the size of the claim of the cohorts to this surplus does not substantially change. However, this surplus is less risky and therefore has more value than in the case of the standard variant. Consequently, the value of the claim of the cohorts to this end surplus also increases. 100% equities The picture in value terms for the generations in the case of a switch to 100% equities is the opposite of that for the switch to 100% bonds. Here it is young employees who win value, whilst older employees lose out. Compare here the black line in figure 4. The explanation is as follows: we first interpret the direction of the grey line, which is a combination of the change in value accrued benefits and change in value contributions paid in. The lower contribution paid in leads to value gains for employees compared to the standard variant. The indexation has less value as a result of the higher investment risk. This reduces the value of the indexation for all ages. Those receiving a pension thereby lose value. The indexation loss for older employees is greater than the value gain due to the lower contribution payments, so that on balance the difference for them between the change in value accrued benefits and change in value contributions paid in is negative. For employees under the age of 50, this difference is positive. The interrupted line represents the change in the claim to the surplus at the end of 2025. Compared to the standard variant,
6 Intergenerational value transfers
111
0.20% 0.15%
% total liabilities
0.10% 0.05% 0.00% -0.05% -0.10% -0.15% -0.20% -0.25% -2
3
8
13
18
23
28
33
38
43
48
53
58
63
68
73
78
83
88
93
98 103
age cohort
Change in generational accounts Change in value of claim to surplus 2025
Change in value (accrued benefits - contributions)
Fig. 4. Effects for generations in the case of a switch to 100% equities as a % of total liabilities 2006
the sum of higher investment returns and lower contribution revenues does not result in a greater overfunding. However, this overfunding is more risky, and therefore less valuable. This also translates into a less valuable and therefore lower claim to the surplus for all cohorts. Switch to a traditional DB scheme This variant is more or less a return to the traditional pension scheme existing in the Netherlands from the post-war period up to the turn of the century. The indexation in this variant is unconditional and the risk is primarily covered by the contribution, meaning that the contribution fluctuates quite significantly. Figure 5 shows the characteristics of the policy ladder in this variant. There is always full indexation. The contribution is calculated according to a contribution table. If the funding ratio is 140%, the costprice contribution rate is required. If the funding ratio is higher, then a reduction can be made that is proportionate to the overfunding. The contribution cut can even be so large that a negative contribution (i.e. refund of contribution) is possible. If the funding ratio is below the upper limit, then a supplement on top of the costprice contribution rate will be payable, which increases according to how much lower the finding ratio is. There is no maximum supplement.
112
R.P.M.M. Hoevenaars and E.H.M. Ponds
Contribution rate Costprice contribution rate
indexation
Funding ratio
140%
Fig. 5. Policy ladder for traditional DB variant. Table 3. Results of classic ALM analysis for traditional DB variant FR* 2024
FR < 105% Contribution 2005-2024 2005-2024
median
Indexation as % of wage 2005-2024
Volatility of contribution rates
median
median
< 80%
Standard variant
161%
0.4%
18%
98%
22%
median ± 0%
Contribution variant
155%
1.3%
17%
100%
0%
3.3%
*FR = Funding Ratio
Classic ALM analysis Table 3 compares the results of the classic ALM analysis of these policy variants with those of the standard variant. The results reveal that the average contribution is somewhat lower than for the standard variant. Naturally, in this variant the indexation allocated is always equal to the increase in wages. The variant leads to a considerable volatility of contribution rates. The last column shows the average annual change in contribution rate from year to year. This is 3.3%, which is very large. Value-based generational accounting Figure 6 indicates the changes in the generational accounts for a switch to a traditional DB final-salary scheme with unconditional indexation and variable contribution. The generational accounts show that younger employees gain value; employees between the ages of 38 and 58 lag behind and those receiving
6 Intergenerational value transfers
113
0.20% 0.15%
% total liabilities
0.10% 0.05% 0.00% -0.05% -0.10% -0.15% -0.20% -0.25% -2
3
8
13
18
23
28
33
38
43
48
53
58
63
68
73
78
83
88
93
98 103
age cohort Change in generational accounts Change in value of claim to surplus 2025
Change in value (claims - contributions)
Fig. 6. Effects for generations in the case of a switch to a traditional DB scheme with unconditional indexation and variable contributions as a % of total liabilities 2006
pensions do well, as shown by the black line. The explanation is as follows. We first examine the direction of the grey line, which combines the change of value accrued benefits and the change in value of the contribution payments. This is positive for all generations. Why? All members see the value of their accrued benefits increase due to the switch from a conditional to an unconditional indexation policy. Furthermore, employees pay, on average, a smaller contribution rate. This variant therefore results in the pension fund providing greater value to its members than the standard variant. This is reflected by the direction of the interrupted line, which represents the change in the claim to the surplus at the end of 2025. The surplus is smaller compared to the standard variant, so that all cohorts lose value. The combination of all these effects reveals a picture of younger employees and members over the age of 58 gaining value on balance. Older employees between the ages of 38 and 58 make value gains through the higher claims and lower contributions paid, but these do not appear to balance out the value loss through the lower claim to the surplus, so that on balance they must accept a value loss. Collective DC This variant exchanges the existing indexation ladder for a ladder that is appropriate for a collective DC. Figure 7 shows the ladder for the collective DC variant. Here the indexation is always related to the financial
114
R.P.M.M. Hoevenaars and E.H.M. Ponds
Costprice contribution rate
Full indexation
indexation
0% 100%
140%
Funding ratio
Fig. 7. Policy ladder for collective DC variant
position of the pension fund. The indexation is proportionally linked to the available indexation reserve. There will be full indexation where the funding ratio is 140%. If the funding ratio is greater than 140%, then additional indexation will be given in proportion to the extra funding over 140%. In the case of a funding ratio of 100%, the assets are equal to the nominal claims. There is then no indexation reserve and the indexation is therefore zero. The indexation will be negative if the assets fall below the nominal liabilities. Classic ALM analysis The results of the classic ALM analysis in Table 4 reveal that this variant results in a lower funding ratio than the standard variant. This is because this variant, on average, produces a high indexation, even greater than that for an unconditional indexation policy. This indexation structure therefore gives more to the members in the form of extra benefits than the current variant. This does imply that the overfunding will be less high than in the standard variant. Furthermore, we can establish that in this variant the indexation is subject to greater volatility than in the standard variant. Table 4. Results of classic ALM analysis for a collective DC FR* 2024
FR < 105% Contribution 2005-2024 2005-2024
median
Indexation as % of wage 2005-2024
Indexation volatility
median
median
< 80%
Variation per year
Standard variant
161%
0.4%
18%
98%
22%
1.6%
Indexation
151%
0.4%
18%
113%
24%
3.6%
*FR=Funding Ratio
6 Intergenerational value transfers
115
Value-based generational accounting The results for the classic ALM analysis are reflected in the value analysis. All members gain value through acquiring extra claims via higher indexation (grey line in figure 8). There is a value loss because the overfunding is less (interrupted line). On balance, (black line) the switch to collective DC is to the advantage of older members and to the detriment of younger members.5 0.4% 0.3%
% total liabilities
0.2% 0.1% 0.0% -0.1% -0.2% -0.3% -2
3
8
13
18
23
28
33
38
43
48
53
58
63
68
73
78
83
88
93
98 103
age cohort
Change in generational accounts Change in value of claim to surplus 2025
Change in value (claims - contributions)
Fig. 8. Effects for generations in the case of a switch to a collective DC scheme as % of total liabilities 2006
6.4
Conclusion
Nowadays, many industry-wide pension funds in the Netherlands have a financing structure characterised by a conditional indexation policy based on an indexation ladder, a contribution rate that is stable and not used – or only restrictively used – as a control mechanism, and an investment mix that is roughly made up of 40% equities, 40% fixed income securities and 20% alternative investments.
5
This conclusion is important for the discussion concerning the substance of the policy for indexation correction.
116
R.P.M.M. Hoevenaars and E.H.M. Ponds
In this chapter we have analysed the generational effects of various policy variants for a standard industry-wide pension fund compared to the current practice. We have established that in terms of economic value a pension fund has to be characterised as a zero-sum game. Each policy adjustment will therefore result in there being generations that benefit from such changes and generations that lose out. For example, more risky investment (greater investment in equities) leads to value gains for younger employees and value losses for older employees and people receiving pensions. A less risky investment policy (more fixed income investments) will provide value gains for older members and value losses for younger members. Older generations benefit from a switch to a policy ladder in which there is no upper or lower limit to the indexation that can be allocated (collective DC). Future developments may give reason for implementing changes to the financing structure. Such changes could lead to unintended and undesirable value transfers between generations. With the proposed method of value-based generational accounting, we can identify such incidental generational effects where policy changes are made. The challenge is then to search for such combination of policy parameters that the resulting generational effects can be regarded as acceptable and justifiable to the members.
Literature Auerbach, A.J., L.J. Kotlikoff, and W. Leibfritz (eds), ‘Generational Accounting around the World’, NBER, Chicago: Chicago University Press, 1999. Ang A. and G. Bekaert, The Term Structure of Real Rates and Expected Inflation, Working Paper, Columbia Business School, 2005. Chapman, R. J., T. J.Gordon, and C.A.Speed, ‘Pensions, funding and risk’, British Actuarial Journal, 7(4): 605–663, 2001. Cochrane, J.H., Asset Pricing, Princeton: Princeton University Press, 2001. Cui J., F. de Jong and E.H.M. Ponds, Intergenerational risk sharing within funded pension schemes, Working paper, Netspar, Universiteit van Tilburg, 2007. De Jong F., ‘Deflators: An Introduction’, VBA Journaal, pp. 22-26, 2004. De Jong F., ‘Pension Fund Investments and the Valuation of Liabilities under Conditional Indexation’, Netspar Discussion Papers 2005 - 024, December 2005, Netspar, Universiteit van Tilburg, 2005.
6 Intergenerational value transfers
117
Ewijk C. van, N. Draper, H. ter Rele and E. Westerhout in collaboration with J. Donders, Ageing and the Sustainability of Dutch Public Finance, CPB report no. 61, The Hague, can be accessed on-line: www.cpb.nl, 2006. Frijns J., ‘Collectief DC in huidige vorm niet duurzaam’, Tijdschrift voor Pensioenvraagstukken, April 2006, pp. 38-44. Gollier C., Intergenerational risk sharing and risk taking of a pension fund, Working paper Toulouse, 2006. Hoevenaars R.P.M.M., R.D.J. Molenaar, P.C. Schotman, and T.B.M. Steenkamp, Strategic Asset Allocation with Liabilities: Beyond Stocks and Bonds, LIFE working paper, Maastricht University, 2005. Hoevenaars, R.P.M.M and E.H.M Ponds, ‘Valuation of intergenerational transfers in funded collective pension schemes’, Netspar working paper, 2006, forthcoming in: Insurance: Mathematics and Economics Kocken T., ‘Curious Contracts - Pension Fund Redesign for the Future’, PhD thesis Free University Amsterdam, Tutein Nolthenius, 2006. Kortleve, N., ‘De meerwaarde van beleidsopties’, Economisch Statistische Berichten, 12 December 2003, pp. 588-590. Kortleve N., ‘De marktwaarde van beleggingsopties’, VBA Journaal, no. 2, Summer 2004, pp. 32-36. Kortleve, N. and E.H.M. Ponds, Pension Deals and Value-based ALM, in: Kortleve e.a. (2006), Chapter 10, 2006, pp. 181-209. Kortleve, N., Th. Nijman and E.H.M. Ponds (eds), Fair Value and Pension Fund Management, April 2006, Amsterdam: Elsevier, 2006. Kotlikoff, L., ‘Generational Policy’, in: Handbook Public Economics, vol. IV, Amsterdam: Elsevier, 2002. Nijman, Th. E. and R.S.J. Koijen, ‘Valuation and risk management of inflation-sensitive pension rights’, in: Kortleve e.a. (2006), Chapter 6, 2006, pp. 84-117. Ponds, E.H.M., ‘Pension Funds & Value-Based Generational Accounting’, Journal of Pension Economics and Finance, vol. 2, nr. 3, November 2003, pp. 295-325. Ponds, E.H.M., ‘Waardeoverdrachten tussen generaties’, Economisch Statistische Berichten, 23 September 2005, pp. 415-417, 2005. Steenkamp, T.B.M, Het ondernemingspensioenfonds in een corporate finance perspectief, thesis VU, Amsterdam, 1998. Teulings, C.N. and C.G de Vries, ‘General accounting, solidarity and pension losses’, De Economist 154, no. 1, March 2006, pp. 63-83.
7
Intergenerational solidarity in the uniform contribution and accrual system
T.A.H. Boeijen, C. Jansen, C.E. Kortleve, and J.H. Tamerus1 In the current uniform contribution and accrual system, all members - irrespective of their age – receive the same pension accrual for the same contribution rate. This leads to large transfers between various groups of members, which makes the pension system vulnerable. The decreasing accrual system that we have analysed, in which each member pays the same contribution rate and in return receives a decreasing accrual depending on their age, does not suffer from these transfers. However, a switch to this system may entail undesirable social effects. Compensation for insufficient pension accrual for active members – upto a maximum of 20% of the liabilities – must be taken into account.
7.1
Reasons
Collectivity and solidarity are the vital characteristics of industry-wide and other pension funds. They can even be used as touchstones for the issue of whether pension funds should be allowed to offer supplementary and other products. Nevertheless, criticism against certain aspects of solidarity is increasing. This criticism originates both inside and outside the sector. Within industry-wide pension funds, members as well as employers are becoming more and more critical of aspects in which solidarity is far too one-sided in their view. The external criticism comes mainly from economists (Bovenberg and Jansweijer, 2005). In this context, they make an explicit distinction between risk solidarity and subsidising solidarity. Economists support risk solidar1
Views expressed are those of the individual authors and do not necessarily reflect official positions of PGGM..
120
T.A.H. Boeijen, C. Jansen, C.E. Kortleve, and J.H. Tamerus
ity (also called chance solidarity), because it pays to share unforeseeable, reciprocal risks with each other. It is therefore in everybody’s interest to insure against these risks in a group scheme. Things are different with subsidising solidarity. This solidarity goes further than sharing reciprocal risks and leads to future-oriented structural value transfers from the one group to the other. In the current climate of transparency and accountability, subsidising solidarity is increasingly a subject of debate, especially when the transfers become too large or tend to flow in one direction. That is why economists in particular ask for these transfers to be identified in order to make a new conscious consideration about whether or not to retain them. The current uniform contribution and accrual system is a symbol of collectivity and solidarity. This system, in which each member receives the same rights (uniform accrual) for the same price (uniform contribution), is considered the ultimate embodiment of the pension fund phenomenon. At the same time, the intergenerational solidarity in the uniform contribution and accrual system is a prominent target for criticism. Due to their very nature, the combination of uniform contribution and uniform accrual leads to systematic transfers during the careers of active members. After all, young people pay more contribution than the actuarial cost price, whereas old people actually pay less than the cost price. The intergenerational solidarity in the uniform contribution and accrual system reinforces the intergenerational debate. In our opinion, the fire of this debate is mainly fanned by the perception that young people pay too much in the form of catch-up contributions to work off shortfalls and to fund legal transition rights. We believe, however, that the criticism from economists on the systematic transfers is sufficient reason to investigate the intergenerational solidarity in the uniform contribution and accrual system in more detail.
7.2
Design of study
FOCUS ON INTERGENERATIONAL SOLIDARITY In the current uniform contribution and accrual system, each member – irrespective of gender, age, health or civil status – receives the same pension rights for the same price (both accrual and contribution are a calculated as a percentage of gross wage). The uniform contribution system therefore contains many forms of solidarity. We restrict ourselves to the
7 Intergenerational solidarity in the uniform contribution
121
solidarity between young and old, the so-called intergenerational solidarity. Using numerical examples, we identify the impact of this solidarity quantitatively.
FICTITIOUS INDUSTRY-WIDE PENSION FUND The examples in this chapter use a fictitious industry-wide pension fund with an average pay scheme and conditional indexation. The fund is in a balanced situation. There is therefore no question of catch-up contributions or contribution discounts. Because of the indexation ambition, the uniform contribution rate is calculated on the basis of the real interest rates.
RESULTS DEPEND ON ASSUMPTIONS For the calculations, we have made assumptions with respect to real interest rate, mortality rate, pay rises, and member population, which are based as much as possible on current pension practice. We have also examined the sensitivity to alternative assumptions. The scope of the solidarity differs; its direction and order of magnitude do not.
OLD-AGE PENSION For clarity and legibility, we restrict ourselves to the old-age pension. The retirement age is 65.
EMPLOYEE BENEFITS Pension costs play an important role in several examples in this chapter. We assume that the employer and the employee share these pension costs. The exact allocation ratio does not matter. What does matter is that the contribution is paid from the available margin for wage increases (‘loonruimte’). One final observation. For ease of reading, we speak of ‘he’ instead of ‘he or she’ if we are talking about stakeholders (members, people).
7.3
Transfers in the uniform contribution and accrual system
Figure 1 shows that the actuarially required contribution rate is age related. For old members, the required contribution rate is four times as high
122
T.A.H. Boeijen, C. Jansen, C.E. Kortleve, and J.H. Tamerus
contribution as a percentage of salary
30% 25% 20% 15% 10% 5% 0% 25
35
45
55
65
age actuarially required contribution contribution actually paid (uniform contribution)
Fig. 1. Young people pay more than necessary; older people pay less than necessary.
as for young members2. The reason for this is the time value of money: the contribution paid by a 25-year-old can yield for 40 years, whereas the contribution paid by a 64-year-old person can only yield for one year.3 Young people pay more in the uniform contribution and accrual system than is actuarially required. Older people pay less than is needed. The break-even point is at an age of 46 years old. So each year, until the age of 46, a member supplies a subsidy to the mutual pension fund. From the age of 46 he annually receives a subsidy in return. The younger member thus transfers a part of his contribution to the current generation of old people. He assumes that later, when he himself is old, there will be new young people who subsidise him. In this way, the uniform contribution and accrual system provides a pay-as-you-go element within the funded pension system. Figure 2 shows the size of this pay-as-you-go element. We can see that a member at age 46 has ‘prepaid’ for an amount of some 70% of his annual salary. 2
In section 7.2, we already noted that all the results depend on the assumptions made; the higher the real interest rate, the steeper the curve becomes.
3
This is the greatest effect. Also, it should be taken into account that the chance of a 25-year-old reaching the age of 65 is smaller than the chance of a 64-yearold reaching the age of 65.
7 Intergenerational solidarity in the uniform contribution
123
accumulated prepaid contribution as a percentage of annual salary
80% 70% 60% 50% 40% 30% 20% 10% 0% 25
35
45
55
65
age
Fig. 2. Pay-as-you-go element in the uniform contribution and accrual system accrues up to 70% of an annual salary
Explanation for the level of prepaid contributions In figure 2 we examine a 25-year-old, who is a member of a pension fund based on uniform contribution and accrual. Figure 1 shows that this member pays too much contribution until he reaches the age of 46. Suppose that we put the excess of the contributions paid aside each year in a bank account. This will create an annually accruing capital sum. After he is 46, the member pays too little contribution. Suppose that this shortfall is withdrawn annually from the bank account. The accrued capital then decreases until finally there is nothing left on the bank account. The accrued capital on the bank account is equal to the accumulated prepaid contribution. We have expressed this on the vertical axis of figure 2 as the percentage of the annual salary for the corresponding age on the horizontal axis. We can see that, at the age of 30, the member has prepaid around a quarter of his annual salary as a result of the uniform contribution and accrual system. When he is 46, he has prepaid an amount equal to 70% of the annual salary that he is earning then. Figure 2 shows the situation for a member who exactly earns back all the prepaid contributions. So in this case there is reciprocal solidarity. However, the following prerequisites must be fulfilled for this reciprocity 4: 4
In the contribution of Bonenkamp et al. in this book, it is shown that one does not earn back the prepaid contribution entirely, because the current young people
124
T.A.H. Boeijen, C. Jansen, C.E. Kortleve, and J.H. Tamerus
The member has a uniform career • He works from the age of 25 until he is 65, without interruptions and always with the same number of working hours; • He works his entire career in same industrial sector or in a sector with the same pension accrual and contribution; • He has average growth of salary. No discontinuities within the pension fund • The pension scheme remains unchanged; • The composition of the membership of the pension fund does not change; • The actuarial and economic assumptions do not change. The cases 1 to 5 illustrate what can happen if these conditions are not met. Then, the uniform contribution and accrual system can lead to large subsidies from one group of participants to the other. These solidarity subsidies are of a non-reciprocal nature.5 Uniform contribution and accrual system leads to mutual subsidies Cases 1 to 3 concern the prerequisite of a uniform career. What happens if a member does not have this, deliberately or unintentionally? Case 1: late entrant receives subsidy Member A works abroad until the age of 46. Then he takes a job with a Dutch company that has a pension scheme based on uniform contribution and accrual. He continues work for this company until the age of 65. Suppose that at age 46, A has a salary of € 50,000 and average pay rises after this. At age 65, with accrued interest, he will then have paid € 290,000 in contributions. If he had paid the actuarially required contribution, this amount would have been € 350,000. Member A has therefore received a subsidy of € 60,000 in the uniform contribution and accrual system.
still contribute to the one-off subsidy that was granted to the older generations at the commencement of the uniformly priced pension scheme. We ignore this in our calculations. 5
Kuné explains several types of subsidizing solidarity in Chapter 2 of this volume.
7 Intergenerational solidarity in the uniform contribution
125
Case 2: premature leaver pays subsidy Member B works at an average salary of € 27,500 between the ages of 25 and 35. Then he leaves to go and work somewhere else, in a company without a uniform contribution and accrual system. B has paid € 36,500 in contributions up to the age of 35. If he had paid the actuarially required contribution, this amount would have been € 22,000. Member B has therefore paid a subsidy of € 14,500 in the uniform contribution and accrual system. Case 3: late career maker receives subsidy Member C has worked for the same company from the age of 25. He started with an annual salary of € 20,000 and since then always had pay rises that were the average for the sector in which he works. When he is 50, C gets a big promotion: his salary increases by € 20,000 that year. At the age of 65, he has paid € 580,000 in contributions. If he had paid the actuarially required contribution, this amount would have been € 610,000. Member C has therefore received a subsidy of € 30,000 in the uniform contribution and accrual system. Cases 4 and 5 concern possible discontinuities in the pension fund. The members can not exercise any control over this, but still run this risk, however. Case 4: active members pay subsidy with economies in pension scheme From the age of 25, member D (starting salary: € 20,000) has been accruing pension in a uniform contribution and accrual system. The annual accrual amounts to 1.5% of the annual salary. When this member is 40, the pension fund’s board decides to economise on the scheme. The new accrual percentage becomes 1.0%. D works until the age of 65. At that age, he has paid contributions of € 496,000. Had he paid the actuarially required contribution, this amount would have been € 474,000. Member D has therefore paid a subsidy of € 22,000 in the uniform contribution and accrual system. Case 5: increase in the average age of the contribution payer leads to subsidies The 25-year-old member E has a salary of € 25,000 and accrues a pension in a sector with an aging membership. In the coming 40 years, the average age of the contribution payers will become 5 years higher. This results in the uniform contribution increasing during the course of the years. As a result of this, up to the age of 65, E has paid € 740,000 in contributions. Had he paid the actuarially required contribution, this amount would have been € 685,000. Member E has therefore paid a subsidy of € 55,000 in the uniform contribution and accrual system.
126
T.A.H. Boeijen, C. Jansen, C.E. Kortleve, and J.H. Tamerus
The subsidies in the uniform contribution and accrual system take place not only between generations, but also within them. Consider, for example, two members who are both 65 years old. Member X has worked in the same sector from the age of 25 until 46, member Y from the age of 46 until the age of 65. Then member X has indirectly paid a generous subsidy to member Y. Both members belong to the same generation. However, a mutual subsidy occurs precisely because of the different careers of X and Y. In other words: the intergenerational solidarity in the uniform contribution and accrual system leads to mutual subsidies between populations of employees with different careers. Examples of populations that receive subsidies in the uniform contribution and accrual system: • members who enter late; • members who receive big salary increases at an older age; • members who continue working after the age of 65. Examples of populations that pay subsidies: • members who leave prematurely, or work less when they are older; • members who have a flat salary curve; • members who start work at a younger age.
7.4
Creation of the uniform contribution and accrual system
In 1949, the Industry-Wide Pension Fund Act came into force. At that time, the first pension schemes also emerged. They combined uniform contribution and uniform accrual, for the following reasons (Lutjens, 1999):
YOUNG AND OLD HAVING EQUAL OPPORTUNITIES ON THE LABOUR MARKET Thanks to the uniform contribution, an old employee and a younger person pay the same amount for a euro of pension. The uniform contribution and accrual system therefore ensured that older employees on the labour market did not experience a competitive disadvantage compared to young people due to pension costs.
7 Intergenerational solidarity in the uniform contribution
127
NO COMPETITIVE DISADVANTAGE WITHIN INDUSTRIAL SECTORS Thanks to the uniform contribution, a company with many older employees has the same contribution burden as a company with a number of young employees. The uniform contribution and accrual system ensures, therefore, that companies with a relatively many old employees do not experience a competitive disadvantage in the area of pension costs relative to companies with a younger population.
PROPER PENSION ACCRUAL FOR OLD PEOPLE Thanks to uniform accrual, in terms of percentage gross pay, old people annually accrue as much pension as young people. Usually, the pension rights are also defined in terms of percentage gross pay. So in the years after World War II, the uniform contribution and accrual system ensured that the old people, carried on the shoulders of the young people, could still accrue a reasonable pension income in a short time frame. The post-war years were characterised by a great sense of togetherness. It was no coincidence that during this time of reconstruction that all social security schemes got off the ground. The uniform contribution and accrual system is the embodiment of this togetherness. Mutual non-reciprocal subsidies were not perceived as a problem, and, in view of the pattern of employment, probably occurred less at that time. The breadwinner model and the lifelong solidarity between employer and employee ensured that the majority of employees did have a uniform career. Moreover, the young working population and the baby boom implied there were no subsidies as result of ageing.
7.5
Is the uniform contribution and accrual system vulnerable?
In section 7.3, we saw that within the uniform contribution and accrual system large mutual subsidies are a possibility. Section 7.4 presented the rationale that led to the introduction of the uniform contribution and accrual system in 1949, in spite of this possibility. In this section, we wonder whether the mutual subsidies can lead to pressure on the uniform contribution and accrual system, now or in the future.
128
T.A.H. Boeijen, C. Jansen, C.E. Kortleve, and J.H. Tamerus
ARE THE MUTUAL SUBSIDIES BECOMING LARGER AND MORE NUMEROUS? Because of the increased international and domestic mobility of labour and the emergence of defined contribution pension schemes, it is conceivable that there are less and less members who accrue their pension in a uniform contribution and accrual system from the age of 25 until the age of 65. Changing patterns of work can cause the uniform career to disappear. The cases in section 7.3 show that this can lead to large mutual subsidies. The current ageing of the population has an amplifying effect on these subsidies.
ARE THE MUTUAL SUBSIDIES PERCEIVED AS A PROBLEM? Solidarity of a non-reciprocal nature is vulnerable (Kuné, 2005). Such solidarity is always under tension, especially if there are no good reasons to justify it. When society becomes individualistic and increasingly more commercial, it is always more difficult to find this justification. Increasing transparency can cause to people perceive mutual subsidies as a problem. In this chapter, we do not examine the question of whether the uniform career is disappearing. Nor do we examine the issue of whether society is becoming more individualistic and increasingly commercial. But if this is the case, we must prepare ourselves that the logic and the willingness to pay a uniform contribution rate may continue to decrease. When the durability of the uniform contribution and accrual system comes under pressure, the pressure on solidarity as a whole will also increase. A possible result of this is a flight to more individual schemes, in which no one shares risks with other people. This is undesirable, from both an economic and social point of view. In the next section, we investigate whether there are alternatives for the current uniform contribution and accrual system. Is it possible to make adjustments to this system to ensure that it becomes less vulnerable to possible changes in society? Or, in other words, can we make adjustments to the system technique to make the pension more solid and protect the desired solidarity better?
7.6
Can we make the uniform contribution and accrual system less vulnerable?
In this section, we examine whether there are adjustments possible in the uniform contribution and accrual system, which make the system less vulnerable without harming the advantages of collectiveness and solidarity. We examine two possible alternatives:
7 Intergenerational solidarity in the uniform contribution
129
1. The progressive contribution system In the progressive contribution system, each member annually accrues the same percentage of pension regardless of age. Depending on his age, he pays the actuarially required cost-based contribution. The previous figure 1 shows the level of this contribution.
accrual as a percentage of annual salary
2. The decreasing accrual system In the decreasing accrual system, each member pays an equal contribution rate, irrespective of his age. In return, he receives an agerelated accrual. The accrual is calculated such that the contribution covers the costs of it exactly. Young people therefore accrue more than old people. Figure 3 shows the levels of the accrual percentages for the various ages. 3.0% 2.5% 2.0% 1.5% 1.0% 0.5% 0.0% 25
35
45
55
65
age
Fig. 3. Decreasing accrual: young people accrue more than old people
We compare the two alternatives with the current uniform contribution and accrual system on the basis of the following questions: a. Do members share the risks in the system with each other on a collective basis? b. Is there leeway within the system for subsidising solidarity on the basis of gender, health or civil status? c. Does the system have mutual non-reciprocal subsidies as in section 7.3?
130
T.A.H. Boeijen, C. Jansen, C.E. Kortleve, and J.H. Tamerus
Table 1. Comparison of three pension systems Uniform contribution Progressive and accrual system contribution system
Decreasing accrual system
Collective risk sharing?
YES
YES
YES
Leeway for subsidising solidarity?
YES
YES
YES
Mutual non-reciprocal subsidies (as in section 7.3)?
YES
NO
NO
Does contribution burden depend on average age?
YES
YES
NO
Competitive disadvantage within sector?
NO
YES
NO
Does each individual receive the same accrual?
YES
YES
NO
d. Does the contribution burden depend on the average age of the contribution payers? e. Does the system lead to competitive disadvantage within the business sector for older employees and companies with a population of older employees? f. Does the system provide each individual member with the same accrual percentage? Sub a, b and c: there is collective risk sharing in each of the three systems. Each of the three systems also has leeway for subsidising solidarity on the basis of gender, health and civil status. The uniform contribution and accrual system additionally has subsidising solidarity on the basis of age. The progressive contribution and decreasing accrual systems do not have this solidarity. That means that none of the members need to pay contributions in advance, so there is no mutual non-reciprocal subsidies as seen in section 7.3. Sub d: in the decreasing accrual system the contribution burden does not depend on the average age of the contribution payers. In the progressive contribution system, however, this is the case, because each member pays an age-related contribution. Moreover, in the uniform contribution and accrual system, the level of the uniform contribution is set at the average of the actuarially required contributions of all contribution payers. Figure 1 shows that the actuarially required contribution is age related. For this reason, therefore, the level of the uniform contribution depends on the average age of the contribution payers.
7 Intergenerational solidarity in the uniform contribution
131
Sub e: because in the uniform contribution and accrual system and in the decreasing accrual system each member pays an equal contribution, there is no question of cost of labour competition within business sectors. Young and old also have equal opportunities on the labour market. In the progressive contribution system, each member pays an age-related contribution. This leads to competitive disadvantage within the business sector for older employees and companies with a population of older employees. Sub f: in each of the three systems, pension accrual is independent of gender, health and civil status. However, in the decreasing accrual system, the accrual is indeed age related. The progressive contribution system does not appear to be an attractive alternative for the uniform contribution and accrual system. It is true that in this system there is no question of mutual non-reciprocal subsidies, but the competitive disadvantage within the business sector, for companies with a population of older employees as well as the older employees, is a considerable disadvantage. The decreasing accrual system appears to be an attractive alternative for the uniform contribution and accrual system. Retaining both collectiveness and solidarity, it makes retirement less vulnerable to social and demographic developments. Furthermore, it prevents competitive disadvantage within the business sector. In section 7.7, we examine the possible transitional issues with a switch to the decreasing accrual system. In section 7.8, we examine the structural effects that could accompany such a switch. In that section, we also ask ourselves to what extent it is inconvenient that each individual does not receive the same accrual.
7.7
Transitional issues
Suppose that all pension funds in the Netherlands immediately want to abandon the uniform contribution and accrual system to switch to the decreasing accrual system. What problems would they face?
CURRENT LEGISLATION IS INADEQUATE The Equal Treatment Act does not allow an age-related accrual, on the grounds of age at work. Moreover, within the current law tax, there are ceilings for the accrual percentage of wage-related pension plans. These ceilings are exceeded in the decreasing accrual system. Therefore, various legislative amendments are required to make a system with decreasing accrual possible.
132
T.A.H. Boeijen, C. Jansen, C.E. Kortleve, and J.H. Tamerus
COMPENSATION FOR PREPAID CONTRIBUTIONS A large number 6 of the current members who are not yet retired have paid contributions in advance. For the very youngest cumulative contributions are not large, because they have not been paying contributions for very long. For the oldest too, because they have already earned back a major proportion of the prepaid contributions. See figure 2. If, from one day to the next, Dutch pension funds replaced the uniform contribution and accrual system by the system with decreasing accrual, they would disadvantage a large number of members who have not yet retired. It seems obvious to compensate them for this. The question is: how? This simple question consists of a lot of subquestions. We cite a few: 1. Which members receive compensation and which not?
Do all members receive a compensation? Or will conditions be attached to this? Example: compensation only for members between the ages of 30 and 60. 2. How would the compensation amount be calculated for each member?
In retrospect, for each member, it can be calculated how much excess contribution he has paid. Looking ahead, it can be calculated how much accrual each member will miss out on in the future. At a collective level it probably makes very little difference, but at individual level this can be a big variation. Compare, for example, a 50-year-old who has been a member since the age of 25 with a 50-year-old who has only been a member since the age 46. If we look forwards, both members receive the same compensation. But retrospectively, the first member does receive compensation and the second member does not. 3. When does the member receive his compensation?
Is the compensation awarded at one time? Or do current members receive a part of their total compensation annually only and as long as they remain a member? 4. How is the total compensation amount funded?
A switch to the decreasing accrual system will possibly involve a large amount of money. This is illustrated in case 6. When this is funded via additional contribution, the future and current active members pay the costs. Funding via the buffer affects the pensioners as well as the future and current active members. 6
Not all members have paid contributions in advance. Take, for example, a member who is now 60 and has only been a member from the age of 50 until the age of 60. As figure 1 shows, this member has paid too little contribution all that time, and therefore certainly has not paid contributions in advance.
7 Intergenerational solidarity in the uniform contribution
133
Case 6: What amount is the total compensation for the prepaid contribution? We calculate the total compensation amount if the PGGM pension fund currently were to replace the uniform contribution and accrual system by the decreasing accrual system. It is purest to look retrospectively, so that every member receives a compensation for the contribution excess paid by him. But this is very difficult in practice, because we then have to identify the precise history of every individual member. For the purpose of illustration, we calculate the level of the compensation by looking forward. For each member therefore we calculate how much accrual he will lose in the future and its current value. The assumption in this is that each member works until he is 65. For the question ‘Which members receive compensation and which do not?’ we distinguish three age intervals. The total compensation amount is equal to the value of the lost accrual for all members combined. Table 2 shows the compensation amounts in euros and in terms of funding ratio. Table 2. Switch costs a maximum of 19.5% of funding ratio Ages that receive the compensation
Level of the total compensation amount in €
in % points of funding ratio
25 to 65
11.9 billion
19.5%
30 to 60
11.3 billion
18.5%
35 to 55
8.9 billion
14.6%
The amounts in table 2 must be seen as a maximum. We always assume that all the members continue working in the business sector until the age of 65. That will probably not be the case in reality. However, it is noted that the compensation amount largely depends on the composition of the membership. For a young membership, compensation is inexpensive, and for an old membership as well. For a pension fund such as PGGM, a switch to the decreasing accrual system is relatively expensive. The average age of the PGGM membership is actually 45, and many of the active members (30%) are aged between 40 and 50.
134
7.8
T.A.H. Boeijen, C. Jansen, C.E. Kortleve, and J.H. Tamerus
Structural effects
Suppose that all pension funds in the Netherlands were to abandon the uniform contribution and accrual system to switch to the decreasing accrual system. What possible structural effects could this bring about?
LESS REDISTRIBUTION IN THE PENSION SYSTEM At the end of section 7.3, we have seen that there are various employees who experience disadvantage from the uniform contribution and accrual system. They pay a subsidy to the pension fund because, deliberately or unintentionally, they have followed a certain career. However, there are also various employees who benefit from the uniform contribution and accrual system. They receive a subsidy from the pension collective because, voluntarily or involuntarily, they have followed a certain career. In the decreasing accrual system, these subsidies disappear. Cases 7 and 8 show the possible consequences of this.
Decreasing accrual system has less redistribution CASE 7: YOUNG PEOPLE WITH INSUFFICIENT PENSION ACCRUAL ARE WORSE OFF In the decreasing accrual system, a younger person accrues more pension than in the uniform contribution and accrual system. This means that a member in the decreasing accrual system who has not accrued pension as a young person, loses more than in the uniform contribution and accrual system. As an illustration we consider two 65-year-old members, A and B. Member A has accrued pension in the uniform contribution and accrual system, member B in the decreasing accrual system. Both members did not start accruing their pension when they were 25, but only when they were 35 years of age. Now that they are 65 years of age, they have a ‘pension gap’ of 10 years. Member A must work 21 months longer to make up his pension gap, member B 39 months: a difference of 18 months. CASE 8: OLD PEOPLE WHO CONTINUE WORKING LONGER ARE WORSE OFF In the decreasing accrual system, an older person accrues less pension than in the uniform contribution and accrual system. This means that an older person who decides to continue working longer, has most benefit in the uniform contribution and accrual system in this context. As an
7 Intergenerational solidarity in the uniform contribution
135
illustration we consider two 65-year-old members, C and D. Member C accrues his pension in the uniform contribution and accrual system, member B in the decreasing accrual system. Both decide to continue working for one additional year. That provides member C with a pension increase of 9.8%; member D receives a pension increase of 8.5%.
SHOULD WE BE AFRAID OF THE FRAYED-EDGE THEORY? The frayed-edge theory states that removing a part of the subsidising solidarity will also lead to other forms of solidarity being put under pressure. As a result, these forms are also removed, until the system ultimately contains absolutely no form of solidarity. The analogy of a loose thread in a piece of knitting is telling. If someone pulls the thread, he pulls the whole piece of knitting apart. Following a different line of reasoning, the analogy would be: by cutting off the loose thread with proper care, we can make the knitting less vulnerable. The loose thread is the pay-as-you-go element in the uniform contribution and accrual system. If the mutual subsidies resulting from that become too numerous and too large, someone can and possibly will pull this thread. But if we step out of the uniform contribution and accrual system and with careful policy switch to an alternative, we can make our pension system less vulnerable.
7.9
Further research is required
We have seen that the current uniform contribution and accrual system leads to large transfers between the various groups of members. This makes the pension system vulnerable, because these transfers cause increased pressure on solidarity as a whole. A possible result of this is a flight to more individual schemes, in which no one shares risks with other people. This is undesirable, from both an economic and social point of view. In that respect, the criticism of economists in particular is understandable. The decreasing accrual system appears to be an attractive alternative for the uniform contribution and accrual system. Retaining both collectiveness and solidarity, it makes retirement less vulnerable to social and demographic developments and, just like the uniform contribution and accrual system, prevents competition within the business sector.
136
T.A.H. Boeijen, C. Jansen, C.E. Kortleve, and J.H. Tamerus
However, a switch to the decreasing accrual system may entail undesirable social effects. Working longer, to a later retirement age, becomes less attractive, for instance. Moreover, it costs up to 20% of the liabilities to compensate current active members for a lower future pension accrual. We therefore end with the question in return to the economists: what do they think of this?
Literature Bovenberg, A.L. and R. Jansweijer, ‘Doorsneepremie bedreigt pensioenstelsel’, Het Financieele Dagblad, 1 July 2005. Kuné, J.B., ‘Billijkheid en doelmatigheid in het systeem van de (aanvullende) pensioenvoorziening’, Financiële en monetaire studies, volume 23, no. 3, 2005. Lutjens, E., Een halve eeuw solidariteit, Rijswijk/The Hague: VB (Association of Industry-Wide Pension Funds), 1999.
8
Everyone gains, but some more than others
K. Aarssen and B.J. Kuipers Within collective pension plans, the uniform contribution rates cause considerable redistribution. Women benefit more from the retirement pension than men, while benefiting less from the partner’s pension. Single people benefit less from schemes than partnered members. The option of exchanging the accrued partner’s pension for supplementary retirement pension has made the differences smaller, however. Additionally, the inequality between employees with longer and short careers has decreased due to the transition from final pay to average earnings schemes. When calculating their premiums, life insurers do make distinctions between characteristics of the members. However, the costs of private insurance products are so high, that nevertheless nearly everyone is cheaper off with a collective pension.
8.1
Introduction
Collective pension schemes use a uniform contribution and pension accrual. Every year, each employee in an enterprise or business sector accrues the same percentage of pension for one and the same uniform contribution. The uniform contribution rates in pension schemes lead to substantial redistribution. Some members pay more for the pension scheme than what they get back, and vice versa. Young people, for example, pay too much pension contribution and old people too little. Women benefit more from the retirement pension than men, because they have a longer life expectancy. On the other hand, however, because women live longer than men, partnered men benefit more from the partner’s pension. Single people do actually take part in paying for the partner’s pension, but until recently they did not benefit from it. For a few years now, however, pension funds have been required to offer to exchange the accrued partner’s
138
K. Aarssen and B.J. Kuipers
pension for supplementary retirement pension. These are only a few examples of the many forms of redistribution within a collective pension scheme. Elsewhere in this volume, Kuné (Chapter 2) provides an extensive summary of the different solidarities in collective pension schemes. Furthermore, Hoevenaars and Ponds (Chapter 6), Boeijen et al. (Chapter 7), as well as Bonenkamp et al. (Chapter 11), examine the transfer of accrued benefits between generations in more detail. Little is known about the size of the redistribution within generations. Hári, Koijen and Nijman (2006) calculate the value of a simple average earnings scheme for various members, in which they analyse the effect of age, gender and level of education. In this chapter, we calculate the benefit that specific groups of employees have from a more detailed average earnings scheme. In this typically Dutch pension scheme members at retirement age have the option to exchange the accrued partner’s pension for supplementary retirement pension. The scheme offers employees, during their career, a survivors’ pension insurance in case of death, and contribution-free pension accrual in case of disability. We quantify the redistribution that is caused by the uniform contribution rate by comparing the actual price or the actuarially fair contribution for a number of representative members. We do this for men and women, single and partnered employees, young people and old people, and employees with a long or a short career. We do not distinguish between levels of education, as in the aforementioned study. When calculating their premiums, life insurers distinguish between age and gender groups. Naturally, employees who want to insure a partner’s pension must pay an actuarially fair contribution as well. Furthermore, for private products, insurers can be selective with acceptance and may require a medical examination. However, it cannot be concluded from this that members who pay the price for the redistributions in a collective pension scheme would be better off with an individual pension product. Mandatory participation enables the pension funds to benefit from economies of scale. For example, considerable cost advantages can be achieved with the administration of the pension scheme and the asset management (see the contribution by Bikker and De Dreu in this volume). The rest of this chapter illustrates the economies of scale that collective pension funds offer. We show how much contributions would rise if employees placed their pension individually with an insurer. We commence with a description of the stylised pension scheme and the representative members.
8 Everyone gains, but some more than others
8.2
139
A stylised pension scheme
There are almost eight hundred company and industry-wide pension funds in the Netherlands. The basic principle of the pension schemes is the same. Employees and employers use the pension fund to save for retirement pension and survivor’s pension. The exact elaboration of the schemes, however, can diverge depending on the preferences of employers and employees involved. A pension fund often administrates several schemes for various groups of members, because changes in the pension scheme often lead to transitional arrangements. A good example of this is the recent transfer of early retirement schemes to additional retirement pension, which has led to transitional regimes at many pension funds. The pension scheme that we use in our calculations is representative for schemes with Dutch pension funds, although some elements may differ significantly.
RETIREMENT PENSION Our stylised pension scheme is an average earnings scheme (see table 1). The retirement pension is a supplement to the General Old-Age Pension Act benefit (state pension, AOW). For this reason, employees only accrue pension for their income above a certain amount, the so-called franchise. The pensionable income above the franchise is called the pension basis. In our stylised pension scheme, we assume a franchise of € 10,000. Many pension schemes, have a low franchise and a high accrual percentage, so that employees can also retire before they reach 65 with a reasonable pension benefit. In our calculations, we assume an annual accrual percentage of 2%. The accrued rights and the pension benefits are annually indexed with the development of contractual wages. The level of indexation tends to depend on the financial position or the funding ratio of the pension fund. It is therefore worth bearing in mind that, in our (deterministic) analysis, the final pension result is surrounded by some uncertainty. The average earnings scheme is currently the dominant pension scheme in the Netherlands. Five years ago, the majority of members in a pension fund were still in a final pay scheme. At present, no less than threequarters of the members of a pension fund fall under an average earnings scheme, while only 10% still fall under a final wage scheme. The stock market crash and declining interest rates at the start of this century led to a serious deterioration of the funding ratios, as a result of which many pension funds saw themselves forced to switch from a final wage scheme to an economised average earnings scheme.
140
K. Aarssen and B.J. Kuipers
Table 1. Characteristics of the stylised average earnings scheme Retirement age
65 years
Accrual percentage
2% of salary less franchise
Franchise
€ 10 000
Indexation
based on development of contractual wages
Partner’s pension
70% in case of death during participation (risk-based) 70% in case of death after retirement (accrual)
Exchange of partner’s pension
15% higher retirement pension
Orphan’s pension
14% per child until the age of 21
Disability
100% contribution-free pension accrual
With a final wage scheme, the members each year accrue a certain percentage of their last earned wage. With a usual accrual percentage of 1.75%, this results after 40 years of service, in a benefit of 70% of final earnings. With an average earnings scheme, on the other hand, each year a percentage of wage earned in that year is accrued. Often, a higher accrual percentage is used than in a final pay plan. The accrued pensions are usually indexed with the development of contractual wages. It is easier for a pension fund with an average earnings scheme to absorb a sudden deterioration in the financial position. This way, the omission of indexation not only affects the deferred members and pensioners, but also the active members and consequently leads to a more balanced sharing of the costs. Another advantage is that it reduces the inequality between members with longer and shorter careers (see section 8.4).
SURVIVOR’S PENSION In our stylised scheme, partnered employees have a survivor’s pension insurance. If the member dies before the age of 65, the partner is entitled to 70% of the pension. This is a risk-based insurance. This means that the employee is only insured for partner’s pension during participation. When the member leaves the company or retires, this insurance is cancelled. The calculation of the partner’s pension is based on the retirement pension that the deceased employee would still have accrued up to the retirement age of 65. The children of a deceased member are entitled to an orphan’s pension of 14% of the retirement pension up to their 21st birthday. When a pensioner dies, the partner is entitled to 70% of the accrued old age pension. Since 1 January 2002 pension funds are legally required to
8 Everyone gains, but some more than others
141
offer all members the option between partner’s pension and supplementary retirement pension. However, this only applies to accrual-based and not to risk-based partner’s pension. In our stylised pension scheme, at retirement, single and partnered members can opt to exchange the accrued partner’s pension for an added 15% of retirement pension.
DISABILITY In case of disability members are entitled to continued, contribution-free accrual. So if the member is no longer able to work due to disability, his normal pension accrual for his last-earned salary continues without him or his employer having to pay contributions. In case of partial disability, of course, partial waiver of contribution payments is granted. The stylised pension scheme contains no supplementary disability benefit, although such cover is actually provided by many pension funds.
RETURN ON INVESTMENT An important source of funding for a pension scheme is of course the return on the invested contributions. In our calculations, we assume a nominal return on the invested capital of 6% (see table 2). One way of looking at this is that the pension fund invests half in government bonds and the other half in equity. At a return of 4½% on government bonds and 7½% on equity – an equity premium of 3% – the expected portfolio return results in 6%. If we adjust for the annual indexation – with a contractual wage increase of 3% – an actual return of 3% results. Table 2. Assumed return on investment and costs of pension scheme Return on Investment
6%
Contractual wage increase
3%
Real return on investment
3%
Investment costs
¼%
Net real return on investment
2¾%
Collection costs
¼%
of the pension basis
Pay- out costs
1½%
of the pension benefit
142
K. Aarssen and B.J. Kuipers
A pension fund incurs costs for asset management and administrative expenses. We assume that the investment costs reduce the annual return by 25 basis points. The administrative expenses are generally subdivided into collection and pay-out costs. The pay-out costs (1½% of the pension benefit) are related to the administration of the pension plan for non-actives (including pension payments), whereas the collection costs (¼% of the pension basis) are related to the other administration expenses. The costs for our pension scheme are in line with the costs that Bikker and De Dreu find in their contribution to this volume.
PENSION CONTRIBUTION In addition to the return on assets, the scheme is funded by a uniform pension contribution. The uniform contribution is first and foremost dependent on the composition of the membership. As we shall see below, women are more expensive than men for the retirement pension and cheaper for survivor’s pension, partnered members are more expensive than single people, and old people are more expensive than young people. With a membership composition representative for the Netherlands, the breakeven contribution for the scheme will amount to approximately 21% of the pension basis. It should be taken into account that, in practice, pension funds do not always charge a break-even uniform contribution. At the end of the 1990s, the contributions were often below a break-even level due to the favourable equity returns. During recent years, however, solvency surcharges were actually necessary to help recover the financial position of pension funds. The pension contribution depends heavily on the assumed return on assets. Had we chosen an actual return of 2% instead of 3%, the breakeven contribution would have been approximately 28% of the pension basis. On the other hand, with a 1 percentage point higher return, the break-even contribution of 16% would have ended up considerably lower. The actual return may vary because of macroeconomic conditions. In times of a low real interest rate, the pension scheme will be relatively expensive. However, the pension fund itself also has an effect on the expected return by investing with a higher or lower degree of risk. The higher the risk, the higher the expected return and the lower the breakeven contribution. At the same time, periods with solvency surcharges and indexation shortfalls are more likely, or even periods with contribution cuts and catch-up indexations (see the contribution by Hoevenaars and Ponds).
8 Everyone gains, but some more than others
8.3
143
Actuarially fair contribution and representative members
All the members pay the same uniform contribution for the pension scheme as a percentage of the pension basis. However, the benefit from the pension scheme will differ from member to member. We determine the value of the scheme for various members on the basis of the actuarially fair contribution rate. The actuarially fair contribution is the contribution that, during his active life, the individual member would have had to pay at each moment in that career if the pension fund had charged the actual costs. The actuarially fair contribution depends in the first place on the amount that must be put aside for the annual pension accrual, so that at retirement age sufficient capital has been accrued to meet the pension benefit. Thus the contribution depends on the salary development of the employee, whether or not there is a continuous contract of employment, the presence of a partner, etc. The amount of the single premium further depends on the return on investment and the administrative expenses. A lower return or a higher level of costs means that more money must be put aside each year. In addition, the amount of the single premium depends on the probability of death before retirement, the remaining lifespan of the member after retirement, and the life expectancy of a possible partner. A second component of the actuarially fair contribution concerns the premium for the survivor’s and disability insurance. This premium depends on the probability that the insurance will have to pay out and the value of the benefit.
REPRESENTATIVE MEMBERS We calculate the actuarially fair contribution for a number of representative members. We distinguish between men and women, single and partnered members, and age. The value of the pension scheme depends to a high degree on the life expectancy of men and women. A higher life expectancy means that the pensioner can benefit longer from the retirement pension. Our calculations are based on the most recent life-expectancy tables (1995-2000) from the Dutch Actuarial Association. The survival probabilities have then been corrected for the trend in the increase of life expectancy according to the forecast from Statistics Netherlands (CBS) until 2050. For example, a 25-year-old male member who now joins the pension scheme is ex-
144
K. Aarssen and B.J. Kuipers
pected to live another 17.5 years after the retirement age of 65, against 20.2 years for a female member. For the value of the partner’s pension, the average number of years the partner still lives after the death of the member is important. Male members have on average a three-year younger partner. This means that at retirement age the partner will survive the male member by an average of 7.8 years. Female members, on the other hand, have on average a three-year older partner. This means that, after retirement age, the partner will survive the female member by only 2.4 years. The risk of disability for women is higher than for men and increases with age. Furthermore, the career development is important for the actuarially fair contribution. In our calculations, male members build a stronger career path than female members. We base this on CBS data from 2004 (see figure 1). Men and women commence at a young age with virtually the same starting salary, but men tend to earn approximately 25% more than women at the end of their careers. We assume that the representative members have full-time jobs. However, women in particular often work part-time. Part-timers accrue less pension in proportion to their part-time factor, but the actuarially fair contribution does not change because of this, as it is a percentage of the pension basis on a part-time basis. For this reason, any part-time employment is abstracted from. euro 50.000
40.000 30.000
20.000
men 10.000
women
0 25
35
45
55
65 age
Fig. 1. Career development of men and women (based on the gross wage of fulltime employees in five-year age brackets in 2004 from the CBS)
8 Everyone gains, but some more than others
8.4
145
Redistribution within the pension scheme
As indicated above, the uniform contribution (and uniform accrual) leads to redistribution within the collective pension scheme. In this section, we examine successively the redistribution between men and women, single and partnered members, young people and old people, and members with long and short careers. To this end, we compare the value of the pension scheme or the actuarially fair contribution rate for the different employees during the time of their membership (see table 3). The actuarially fair contribution is expressed as a percentage of the pension basis, i.e. that part of the income for which the member also accrues pension. In addition, we present the level of pension benefit that the members receive from the age of 65. Table 3. Actuarially fair contribution rate and level of pension benefit for representative members a Man Benefit from 65 b
Woman Actuarially fair contribution c
Benefit from 65 b
Actuarially fair contribution c
Single, entry age 25 – no career development
14 700
16.7%
14 700
19.8%
– average career development
29 200
17.9%
24 000
20.7%
– no career development
12 800
20.7% (19.4%)
12 800
20.2% (21.5%)
– average career development
25 400
22.0% (20.6%)
20 900
21.0% (22.4%)
– no career development
15 200
23.0%
15 200
27.9%
– average career development
16 100
23.1%
15 700
27.9%
– no career development
13 200
27.8% (25.8%)
13 200
27.7% (29.6%)
– average career development
14 000
27.9% (25.9%)
13 600
27.8% (29.7%)
Partnered, entry age 25
Single, entry age 45
Partnered, entry age 45
a For 25-year-olds, the starting salary is € 26, 000, for 45-year-olds € 43, 000. b In euros and corrected for general wage inflation. c In percentages of the pension basis. It is assumed that a partnered member does not opt to exchange the partner’s pension for supplementary retirement pension. The actuarially fair contribution is shown in bracket, if the partnered member does choose supplementary retirement pension.
146
K. Aarssen and B.J. Kuipers
MEN AND WOMEN Single women benefit more from the pension scheme than single men. The actuarially fair contribution for a man, who accrues pension from the age of 25 to the age of 65, amounts to 17.9% of the pensionable salary. The value for a woman, at 20.7%, is 2.8%-points higher. The most important cause is the almost three years longer life expectancy of the woman, which means she can enjoy the retirement pension longer. Moreover, women benefit relatively well from the insurance for contribution-free pension accrual with disability, because they have a higher risk of becoming incapacitated for work. In contrast to that is the disadvantage that women on average develop careers less strongly than men. Single men and women paying contributions to the pension fund starting from the age of 25, will have accrued a pension of 80% of the average pension basis after 40 years of service. With an equal initial salary of € 26,000 and no career development, this means a supplementary pension benefit of € 12,800 [= 80% x (€ 26,000 – € 10,000)]. None of these has a partner. They will therefore exchange the accrued partner’s pension for a 15% higher retirement pension, so that the total benefit amounts to € 14,700 each year (see table 3). With a normal career development, of course, the pension benefit is a bit higher. The woman then accrues less pension, because she makes less career progress than the male member. However, the lower pension benefit cannot be considered as a disadvantage, because she also pays less contribution with a fixed contribution rate. Partnered women actually have less benefit from the pension scheme than partnered men. The value of the pension scheme for a woman with a partner, who is a member from the age of 25, amounts to 21.0% of the pension basis. This is 1%-point lower than for a partnered man. The partnered man has a lot of benefit from the partner’s pension. After his death, after all, his partner receives 70% of the retirement pension until her death, surviving on average for almost 8 years. A partnered woman has much less benefit from the survivor’s pension, because her partner only survives her by an average of 2.4 years. It is more beneficial for women to choose supplementary retirement pension. Even if she has a partner, the value of her accrued benefits in that case is higher (22.4%) than when she does not select the exchange (21.0%). It therefore seems obvious to select supplementary retirement pension, if her partner does not consider this supplement as really necessary, for example, if he already has a satisfactory retirement pension. If this is not the case, the woman can also choose to personally insure an individual partner’s pension separately. The behavioural finance literature shows, however, that people often take the default option if they are faced with a choice (see the
8 Everyone gains, but some more than others
147
contribution by Van Els et al. in this volume). Furthermore, it is cumbersome for many members to manage the options for additional insurance and this can concern substantial supplementary premiums. In practice, it has been found that few partnered members wish to exchange the partner’s pension. However, the choice offered still involves relatively modest amounts, because with many pension funds only the accrual with effect from 1 January 2002 can be exchanged. Therefore it is possible that this situation will change in the future. In our calculations we assume that the members remain healthy from entry age to retirement age. If the member becomes disabled or dies before retirement age, then there is a very heavy redistribution (see box).
Redistribution with disability and death Within our stylised pension scheme, redistribution occurs by means of the insurance cover for waiver of contribution payments with disability and the survivor’s insurance for employees. Disabled members continue accruing a pension on the basis of the last-earned salary without paying a contribution. On death of an employee, the partner receives a partner’s pension of 70% of the pension that the employee would eventually have accrued. The redistribution via both insurances usually only takes place in retrospect. Everyone runs a risk of becoming disabled or of dying prematurely. Only after such an event has taken place the redistribution occurs. This redistribution with insurance is also referred to as risk solidarity or cold solidarity. Nevertheless, there is also warm solidarity, because no selection is applied and all members, young and old, sick or healthy, pay the same contribution for this cover. The transfers on death and disability are very large. This is shown for our representative person who joins the scheme at the age of 25 (see figure 2). At the age of 30, he has only accrued pension for five years. The future pension benefit payments then only have a value of € 10,000. If the man becomes permanently disabled for work at that age, then that is € 110,000. The difference of € 100,000 represents the value from the contribution-free pension accrual and risk cover up to his age of 65. The value of the insurance increases until an age of just over forty years. The number of contribution-free years does indeed decrease, but in contrast to this is the fact that the income on which pension is accrued at a young age strongly increases. After the age of forty the effect of a shorter contribution-free period becomes dominant.
148
K. Aarssen and B.J. Kuipers
The value of the partner’s and orphan’s pension amounts to no less than € 375,000 if the man dies when he is 30 years of age. In this context, the value of the insurance reaches a peak around an age at death of just under 40. If the man dies just before he is 65, the insurance is then worth less than the accrued retirement and partner’s pension. The long benefit payment duration of the partner’s pension then no longer weighs against the fact that the rights to retirement pension lapse. 600
1000 euro
500 400 300 200 100 0
age 25
30
35
40
45
dismissal or resignation
50
55
disability
60
65
death
Fig. 2. Discounted value of pension rights upon dismissal or resignation, disability and death (Partnered man with average career development and entry age of 25)
SINGLE AND PARTNERED EMPLOYEES Single people have less benefit from the pension scheme than partnered employees, even though they can transfer the accrued partner’s pension into a 15% higher retirement pension. We saw previously that the pension scheme for a single man, who has accrued a pension since 25 years of age, represents a value of 17.9% of the pension basis (sees table 3). This is 4.1%points lower than for a partnered man, for whom the value is 22%. One cause of this difference is that, when calculating the exchange percentage of the partner’s pension, pension funds are not allowed to distinguish between men and women. Therefore they assume an average value of the partner’s pension in calculating the exchange percentage. Because of this,
8 Everyone gains, but some more than others
149
the 15% higher retirement pension does not compensate the partner’s pension, which has a relatively high value for men. Moreover, single men do not benefit from the survivors’ insurance on death during their working life. Single women also benefit less from the pension scheme than partnered women. The difference is minor, however: the annual value of the scheme amounts to 20.7% of the pension basis for a single woman, compared to 21.0% for a partnered woman. For the partnered woman it would also be more advantageous to choose 15% extra retirement pension at retirement age because of the fixed exchange percentage for the partner’s pension.
Less disadvantage for single people Since 1 January 2002, pension funds have been legally required to offer of the right to transfer the accrued partner’s pension for supplementary retirement pension. This has considerably reduced the redistribution from single members to partnered employees. A single man in former days – when the partner’s pension could not yet be exchanged for supplementary retirement pension – had a benefit of only 15.6% of the pension basis (see figure 3). For the same pension contribution, the single man would annually have profited as much as 6.4%-points less than a partnered man. Now single men in our scheme still benefit 4%-points less than partnered men. The difference between single and partnered women has become marginal. However, we assume that partnered women choose the default option of keeping the partner’s pension.
% pension basis
25 20 15 10 5
Man previously
Man - now Single
Woman previously
Woman now
Partnered
Fig. 3. Value of the scheme with and without exchange of partner pension (Joining at the age of 25 with average career development)
150
K. Aarssen and B.J. Kuipers
The value of her pension rights then increases by 1.4%-points. That the single woman nevertheless still experiences a small disadvantage, is caused by the fact that she has no benefit from the survivors’ insurance on death before retirement age. So, single people benefit less from the pension scheme than partnered employees. The difference, however, is not as big as it used to be (see box).
YOUNG AND OLD The largest redistribution within the pension scheme takes place between young people and old people. Young and old members pay the same pension contribution and annually have the same pension accrual. However, the actuarially fair contribution for a year’s pension accrual is much lower for a young member than for an older member. The further the employee is away from of his retirement age, the more return on investment the pension fund can still make on the assets. The value of one year’s pension accrual in our stylised scheme for a member aged 25 amounts to approximately 11% of the pension basis (see figure 4). For a member aged 60, that is 35% to 40%, thus approximately 3.5 times as much. Men and women who do not join the pension fund until they are 45 years of age, therefore have a bigger advantage than employees who start to pay contributions at 25 years of age (see table 3). Naturally, it does not happen often that people only start to accrue a pension at age 45. This happens, for instance, to self-employed people who enter employment at a later age or for people who have worked abroad up to the age of 45. 45 40
% pension basis
35 30 25 20 15 10 5 25
30
35
40 man
age
45
50
55
60
woman
Fig. 4. Actuarially fair contribution for one year of membership for partnered employees (Entry at the age of 25 with average career development)
8 Everyone gains, but some more than others
151
For employees who have been members of a pension fund for their entire working life, with a balanced and stable membership, this redistribution is not a major problem. Employees then pay too much at a young age and too little at an older age in relation to the actuarially fair contribution. It is also considered as a benefit that this mechanism encourages old people to continue working longer. Indeed, an extra year continuing to work produces a higher pension benefit for a very low contribution. However, in their contribution to this volume, Boeijen et al. discuss a number of cases in which the redistribution between young and old lead to considerable advantages or disadvantages for the employee concerned.
LOW AND HIGH INCOMES Within our stylised pension scheme, no direct redistribution takes place between low and high incomes. Someone with a higher income, of course, accrues a higher top-up pension measured in euros. In percentage terms of the pension basis, the benefit that low and high incomes enjoy is the same. Because the pension contribution is also a fixed percentage of the pension basis, the different income brackets receive the same percentage in exchange for the deposits. It is, however, true that employees with a steeper career development benefit more, generally the upper-income groups. It means that a relatively large part of the pension accrual takes place close to the retirement age. And the closer to retirement age, the more expensive the pension accrual becomes. The value of the pension scheme for a single man with an average career development is 1.2%-points higher than for the same man without any career development: 17.9% of the pension basis versus 16.7% (see table 3). For a single woman, the difference is a little smaller (0.9%-point), because on average she has a less strong career development. The benefit for career makers is considerably less in average earnings schemes than in final salary schemes (see box). In our calculations, we make no distinction on levels of education. There tends to be a relationship between life expectancy and the level of education enjoyed. Employees with a higher education, and usually a higher income, on average have a more healthy lifestyle, have access to better medical care and, as a result, have a higher life expectancy than low and semi-skilled people. Hári, Koijen and Nijman (2006) estimate that a highly educated man aged 65 has two to three more years to live than a low or semi-skilled person, as a result of which they benefit more from the retirement pension.
152
K. Aarssen and B.J. Kuipers
Less advantage for career makers During recent years, many pension funds have changed from a final pay scheme to an average earnings scheme. Currently only 10% of the members of pension funds are still under a final salary scheme. Employees with a steep career development are favoured much more strongly in a final pay scheme than in an average earnings scheme. With an individual pay rise in a final wage scheme, the benefits accrued in the past also increase, the so-called backservice. This is relatively expensive for a pension fund. The increase of the rights occurs closer to the retirement age, so that there is less time for returns to be accrued. To illustrate this, we have calculated the advantage a single man with an average career development has over a single man without any career development in a final wage scheme. For this we assume an annual accrual percentage of 1.75%, but in other respects the final wage scheme is the same as the average earnings scheme. In an average earnings scheme, a single man with average career development benefits 1.2%point more than the same man without any career development (see table 3). In a final pay scheme, with 3.6%-points, this difference is three times as large. Of course, we could have chosen a more extreme example, such as an employee who receives another considerable pay rise just before his retirement, a so-called pension promotion.
8.5
Collective and individual pensions
It is clear that some members benefit more from the pension scheme than others. Thus, the value of the scheme is high for partnered men and low for single men. However, it would be a misconception to immediately conclude that single men, for example, would be better off without the obligation to save with a pension fund. Due to mandatory participation, pension funds can provide a pension product at much lower costs than individual insurers. Pension funds can benefit from economies of scale and do not need to incur marketing and sales expenses. The operating costs, including profit, of life insurers are more than seven times as large as the costs of pension funds because of this (see the contribution by Bikker and De Dreu). Also, summaries regularly published, by the Dutch consumers’ association for instance, show that the costs for individual pension products are high. Around the middle of 2006, the top 10 providers of a guaranteed
8 Everyone gains, but some more than others
153
annuity offered a man aged 60 an average implicit return of 2.6% with a comparable long-term interest rate of 4.5% (Consumentenbond, 2006a). This is the implicit return for a man with average probabilities of death. Since insurers will take into account expected improvements in the life expectancy, and with the fact that the very healthier men buy such an annuity (‘adverse selection’), the effective costs are actually lower than 1.9%, however. The top 25 most sold investment funds charged – apart from buying and selling costs – annual management fees of an average 1.2% of the capital (Consumentenbond, 2006b). Small differences in annual costs lead to very large differences in contribution rates. Contributions finance roughly half of the pension rights; the earned return on the invested contributions the other half. A 1% increase in the annual costs means that the net return comes out more than 35% lower. This will have to be compensated with a considerable increase in the contribution. To illustrate this, we have calculated what the costs would be if a partnered man took out our stylised pension scheme individually with a life insurer. We assume for this that the annual investment costs amount to 1¼% with a life insurer instead of the ¼% with the pension fund. Partnered male members, who accrue pension for 40 years, would have to pay an insurer an annual contribution of 29.7% of the pension basis (see table 4). With a pension fund, the same scheme costs only 22.0%. If the pension fund would ask for a break-even contribution, the employee with the insurer would therefore be in a more expensive position of no less than 7.7%-points of the pension basis or approximately 35% worse off. In practice, the pension contribution will never be exactly break-even, because of the many forms of redistribution within a pension fund or, for example, because of solvency surcharges to improve the financial position of the fund. In contrast to this, after a positive return shock, it is also possible that pension contributions temporarily arrive below a break-even level, such as at the end of the 1990s. Table 4. Actuarially fair contribution for a partnered man at a pension fund and an insurera Pension fund
Insurer
25 years
22.0
29.7
45 years
27.9
34.5
Entry age
a
in percentages of the pension basis.
154
K. Aarssen and B.J. Kuipers
The solvency surcharges reveal another advantage that collective pension schemes have compared to individual schemes. In our calculations, we have not taken into account uncertainty concerning equity returns, interest and wage developments. With disappointing results on for example equity in individual schemes the individual members will bear the full burden. Lower than expected equity returns then lead directly to lower benefit payments after retirement. Collective pension funds, on the other hand, have the possibility of spreading risks over several generations. A disappointing return on investments is not entirely carried by the current generations, for example, by omitting indexation or imposing higher contributions. It is partially shifted to future generations and vice versa. The possibility of shifting surpluses and shortfalls implies, for the members, a lower risk for the same contribution or lower contribution for the same risk. A recent study by the CPB Netherlands Bureau for Economic Policy Analysis estimated the advantage from intergenerational risk sharing in an indexed average earnings scheme to be 0% to 8% of the contribution or a maximum of 1.5% of the salary (Westerhout et al., 2006). Cui et al. (2006) calculate that – depending on the individual scheme with which the average earnings scheme is compared – the welfare benefit can amount to 4%. Teulings and De Vries (2006), finally, arrive at a benefit of 6% of the salary with the best possible risk sharing between generations.
8.6
Conclusion
In this chapter, we have quantified the redistribution that takes place within pension funds between various members. We have assumed a stylised pension fund and representative employees, so that the results must be interpreted with the necessary prudence. The Netherlands has countless pension schemes and the personal situation of employees differs widely. Single people benefit less from the pension scheme than partnered employees. The option that pension funds offer nowadays for accrued partner’s pension to be converted into supplementary retirement pension has, however, considerably reduced the differences. Partnered members do still enjoy the benefit of risk cover before the pension commences. Single women benefit more from the pension scheme than single men, because they have a longer life expectancy. The partner’s pension is very valuable for men, so that partnered men generally enjoy a larger benefit than partnered women. The change from final pay to average earnings
8 Everyone gains, but some more than others
155
schemes has led to more equal results between employees with much and less career development. The largest redistribution within a pension fund is undoubtedly between young and old people. However, with a balanced and stable membership during the course of the years, young people naturally receive the excessively paid contribution back again at a later age. The redistribution within a pension fund can certainly not be ignored. With a uniform contribution rate of 21%, for example, single men annually overpay 3% of the pension basis, and partnered men underpay 1%. These differences, however, pale into insignificance when we compare the contribution with the pension fund against the contribution with a life insurer. Due to mandatory participation, pension funds can benefit from economies of scale in providing a pension scheme at very low costs. Employees would, on average for an individual scheme with an insurer, annually pay an average of 7% more of the pension basis. With an average salary for a full-time employee of almost € 40,000 this comes out to approximately € 2,000 per annum. To put it briefly: everyone gains with collective pensions, but some win more than others.
single man partnered man single woman partnered woman
pension fund insurer 5
10
15
20
25
30
% pension basis
Fig. 5. Summary of redistribution and advantage of collective versus individual pensions (Shows the actuarially fair contribution for members who enter the pension fund at age 25. For the pension fund and the insurer, the weighted average of these members is shown.)
156
K. Aarssen and B.J. Kuipers
Literature Consumentenbond, ‘Lijfrente naar uw partner’, Geldgids, July edition, pp. 40-42, 2006. Consumentenbond, ‘Parels uit de eerste divisie’, Geldgids, July edition, pp. 46-49, 2006. Cui, J., F. de Jong, E. Ponds, Intergenerational risk sharing within funded pension schemes, unpublished paper, 2006. Hári, N., R. Koijen and Th. E. Nijman, The determinants of the money’s worth of participation in collective pension schemes, unpublished paper, 2006. Teulings, C.N., and C.G. de Vries, ‘Generational accounting, solidarity and pension losses’, De Economist, vol. 154, no. 1, 2006, pp. 63-83. Westerhout, E., M. van de Ven, C. van Ewijk and N. Draper, Naar een schokbestendig pensioenstelsel – verkenning van enkele beleidopties op pensioengebied, The Hague: CPB Document no. 67, 2004.
Part 3. Mandatory participation
9
Why mandatory retirement saving?
P.J.A. van Els, M.C.J. van Rooij, and M.E.J. Schuit1 More than 90% of employees in the Netherlands compulsorily accrue pensions via their employer. Experiences abroad, supplemented by empirical research among Dutch households, suggest that, without this automatism, large groups of employees would build up much less pension. Procrastination, self-control problems, and limited financial knowledge and skills frequently lead to low pension savings and low returns on the accrued pension capital. Mandatory retirement saving prevents these problems. The Dutch mandatory participation works well and there is no reason for drastic modifications. What can be studied, however, is how mandatory retirement saving for the self-employed can result in a better pension build-up.
9.1
Introduction
In the majority of countries, participation in a public pay-as-you-go pension system is mandatory, whereas participation in privately organised, supplementary funded pension schemes is voluntary. In a number of countries, on the other hand, mandatory participation has also been introduced for private pension schemes. In these cases, various mandatory participation systems can be identified. This chapter closely examines the Dutch mandatory participation and views it from an international perspective. Section 9.2 describes the Dutch system of mandatory participation, discussing both its aim and results in section 9.2.1. Subsequently, section 9.2.2 examines which alternative sys-
1
Views expressed are those of the authors and do not necessarily reflect official positions of De Nederlandsche Bank (DNB). Comments by Aerdt Houben, Fieke van der Lecq, Onno Steenbeek and Job Swank on earlier versions of this contribution are gratefully acknowledged. We thank Gita Gajapersad and Rob Vet for statistical assistance.
160
P.J.A. van Els, M.C.J. van Rooij, and M.E.J. Schuit
tems of mandatory participation are used in other OECD countries while section 9.2.3 addresses the level of participation this leads to. In the second part of this chapter, mandatory participation is analysed in more detail. In section 9.3 we discuss the literature about and experiences with voluntary pension schemes, focusing on the United States in particular. The central issue here is whether mandatory participation is actually necessary. Research on individual financial planning and decision making, in the field of behavioural finance, shows that people find it difficult to make prudent choices in the domain of saving and investing. In line with this, we proceed by reviewing the results of recent empirical research among the Dutch public in section 9.4. Section 9.4.1 deals with their preferences concerning retirement saving. The problems that would emerge if Dutch people were given more autonomy in the pension domain are identified in section 9.4.2. Subsequently, in section 9.5, we discuss experiences in the Netherlands with restricted freedom of choice related to the use of the life-course savings scheme and individual savings products in the third pillar of the Dutch pension system. The chapter closes with several conclusions in section 9.6.
9.2
Mandatory participation in the Netherlands and elsewhere: the facts explained
9.2.1 DUTCH MANDATORY PARTICIPATION IN A PENSION PLAN The Dutch system does not feature a direct statutory obligation to participate in a pension plan. An employer is in principle not forced by law to grant a pension to employees, and employees are in principle not obliged to save for their pension. The Dutch system does have an indirect mechanism of mandatory participation by means of the mandatory participation for companies (see the contribution by Omtzigt in Chapter 10 of this book) and mandatory participation for individuals.2 Mandatory participation for companies and for individuals The mandatory participation for companies relates to the legal possibility to oblige companies within a certain industry or business sector to affiliate 2
In addition, the compulsory membership effective under the Compulsory Professional Pension Scheme Act (Wet verplichte beroepspensioenregeling – WVP) can also be identified. This law can oblige professional people to be members of a professional pension scheme (see Chapter 10).
9 Why mandatory retirement saving?
161
with the industry-wide pension fund. Social partners, employers and employees or their trade unions, can submit an application for this to the Dutch Minister of Social Affairs and Employment. Neither employers nor employees can in principle withdraw from this mandatory participation in the industry-wide pension fund.3 Moreover, under this system of mandatory participation, employers are obliged to withhold pension contributions from their employees’ wages and pay these to the pension fund that administers the pension scheme. Self-employed can also be included under the mandatory participation if the social partners include this in their application.4 Besides the mandatory participation for companies, there is the mandatory participation for individuals: employees who are covered by a collective labour agreement (CAO) are compelled to participate in the pension scheme that is associated with the relevant CAO. The Dutch mandatory participation system, as a result of which both employers and employees can be compelled to participate in pension schemes, therefore comes about by means of collective contracts and can be typified as ‘quasi-mandatory participation’ (OECD, 2005), or mandatory participation by means of collective contracts. Legal basis The possibility for mandatory participation in industry-wide pension funds, the mandatory participation for companies, has existed since 1949 and is based on the Act on mandatory participation in a company pension fund (Wet betreffende verplichte deelneming in een bedrijfspensioenfonds - hereafter: BPF Act). The initial backdrop of this mandatory participation was to restrict the process of free wage determination after the Second World War. By making pension schemes mandatory within an industry, the wage costs for every enterprise within the industry would be similar, thus containing undesired competition on the basis of wage costs. Besides this economic motive, nowadays a second issue plays an important role: the social aspect. This became clear during the preparation of the new Act on man-
3
Exceptions are a granted exemption or dispensation, and the possibility due to a poor investment performance (the so-called ‘z score’) to select a different pension administrator. In these situations, however, the pension scheme remains compulsory.
4
For example, the industry-wide pension funds for the building industry (BPF Bouwnijverheid) and the industry-wide pension fund for painters, finishers, and glass installers (BPF Schilders) also include certain categories of self-employed under their compulsory membership.
162
P.J.A. van Els, M.C.J. van Rooij, and M.E.J. Schuit
datory participation in an industry-wide pension fund (Wet verplichte deelneming in een bedrijfstakpensioenfonds), the successor to the 1949 Act, which came into force on 1 January 2001. The aim of the new act is to reduce the number of ‘blank spots’. This term is used for employees who are unable to take part in a pension scheme because the employer does not offer one, or because they are excluded from the employer’s pension scheme. The latter occurs on the basis of gender (women), working part-time, a temporary contract, age (too young), waiting time provisions, or salary (too low). Preventing blank spots developed into a social objective. Blank spots Research into the size of this blank spot in 1985 showed that 18% of all employees between the ages of 25 and 65 were not accruing supplementary pensions. Ten years later, in 1996, this share had decreased to 9% (Ministry of Social Affairs and Employment, 2002). Of these 9%, 7 percentage points were caused by employees being excluded from participation, as described above. Research into the level of exclusion of employees from pension schemes clearly shows that the number of pension schemes with grounds for exclusion has further decreased since the turn of the century (Social and Economic Council of the Netherlands (SER), 2002). This is undoubtedly related to the fact that a number of grounds for exclusion have meanwhile been prohibited by law (exclusion of women, part-time workers and employees with a temporary contract). Self-employed The government’s aim is that every employee can accrue a supplementary pension. This supplementary pension must be sufficient in combination with the General Old Age Pension (state pension, AOW) to reasonably maintain the person’s standard of living level after retirement. Enforcing mandatory participation is an important instrument to reduce the number of employees without a supplementary pension (Ministry of Social Affairs and Employment, 2005). The bulk of the self-employed, however, do not compulsorily save for a pension. It is estimated that 13% of the self-employed participate in a professional or industry-wide pension fund5 while 5
These are own calculations based on the total number of active members in professional group pension funds and an estimate of the number of selfemployed people participating in industry-wide pension funds for the construction industry (BPF Bouwnijverheid) and for painters, finishers, and glass installers (BPF Schilders).
9 Why mandatory retirement saving?
163
the remainder are not compelled to save for a pension. They are expected to supplement their state pension (AOW) through additional voluntary savings themselves. The question is whether they are capable of doing this, or whether a form of mandatory participation must be introduced for them.
9.2.2 MANDATORY PARTICIPATION: AN INTERNATIONAL PERSPECTIVE The Dutch indirect form of mandatory participation has been described in the previous section as an indirect mandatory participation by means of collective contracts. Social partners determine the contents of the pension scheme themselves. In a number of other countries, however, direct statutory mandatory pension schemes exist for which the law also sets the content or the level of the contribution. Two different versions of direct mandatory participation can be identified. In the first alternative, the direct mandatory participation is arranged via employers: employers are legally obliged to arrange a collective pension scheme for their employees, in which legal minimum requirements (as regards the level of the pension contribution or pension result) are prescribed. Employees are then compelled to participate in the employer’s pension scheme. In the second alternative, the direct mandatory participation is arranged via the employees: employees are legally required to save for a pension. They are free to choose a pension administrator themselves. Employers are usually required to make a minimum contribution to their employees’ pension savings. Table 1 shows a summary of the OECD countries with mandatory private pension schemes. This includes the three different mandatory participation alternatives mentioned above: direct statutory mandatory participation via the employer, direct statutory mandatory participation via the employee, and quasi-mandatory participation by means of collective contracts. Of the total of 30 OECD countries, nine countries have mandatory private pension schemes of the types mentioned above. The other countries have no mandatory private pension schemes. Three OECD countries have statutory mandatory participation in which the employer is required to provide a pension scheme for its employees: Australia, Iceland and Switzerland. In Switzerland, the employer has a number of options, including setting up its own pension fund or affiliating with an existing pension fund. Moreover, a minimum employer’s contribution and a minimum rate of interest are legally regulated. The latter was 4% for a long time, but was reduced to 3.25% in 2003. In addition, the government also prescribes the level of the annuity ratio, if the total
164
P.J.A. van Els, M.C.J. van Rooij, and M.E.J. Schuit
Table 1. Mandatory funded pension schemes
Countries
Statutory mandatory participation via employer
Statutory mandatory participation via employee
Mandatory participation via collective contracts (Quasimandatory participation)
Australia
Denmark
Denmark
Iceland
Mexico
Netherlands
Switzerland
Hungary
Sweden
Poland Sweden
Source: OECD (2005), page 23.
amount is converted into a pension at retirement age. This system therefore cannot be described as a defined benefit (DB) or defined contribution (DC) scheme, but is a hybrid scheme. In Iceland, employers are required to provide a pension scheme with a high DB content, whereas, on the other hand, the Australian pension schemes are DC in nature. Incidentally, employees in Australia have had the option of choosing a fund themselves since 1 July 2005. In five of the 30 OECD countries, a system of statutory mandatory participation has been introduced that is typified here as statutory mandatory participation via the employee. Chile was the first country, not a member of the OECD, to introduce a statutory mandatory funded scheme in which employees are legally required to open a private retirement savings account. Employees can select themselves with which private pension administrator they open this account. As a result, the pension scheme is mainly DC in nature, but the government has built in a guarantee system through which a minimum pension is guaranteed. The latter coincides with the fact that this system has been introduced as a replacement for the pay-as-yougo state pensions. Also, in other Latin American countries, including OECD member Mexico, and in a number of Eastern European countries, including Poland and Hungary, pension reforms have been carried out because of the expected increase of the pension costs as result of the ageing of the population. This has led to a shift from mandatory participation in pay-as-you-go state pensions to privately implemented funded pension schemes with mandatory participation. The Danish and Swedish pension models have a statutory mandatory private pension account as well as supplementary funded pensions that are provided via the employer. These supplementary funded schemes are achieved by means of collective contracts between employers
9 Why mandatory retirement saving?
165
(or employers’ organisations) and employees (or unions). The Swedish model is most similar to the Dutch model. In Sweden, the collective agreements made by the social partners can be made mandatory for a complete sector and the pension schemes, as in the Netherlands, are mainly DB in nature.
9.2.3 MANDATORY VERSUS VOLUNTARY PENSION SCHEMES The Dutch system of quasi-mandatory participation by means of collective contracts has resulted in almost all employees being covered by supplementary funded pension schemes. To what extent does mandatory participation play a role in this high level of participation? To be able to assess this, a comparison can be made between participation levels of private pension schemes in a number of countries. This will provide an indication of the extent to which the absence of a direct or indirect mandatory system leads to a lower level of participation. Figure 1 shows the levels of participation in tax-facilitated pension schemes in OECD countries. Clearly, among the countries with a high participation level, the majority has a system of statutory mandatory participation or quasi-mandatory participation. The Icelandic statutory mandatory 120
100
% of employees
80
60
40
20
ak ia Po rtu ga l
Ja pa n
Sl ov
in
d
ex ic o M
Sp a
es
la n Ire
om
St at
Ki ng d
U
ni
te d
a Po la nd U
ni
te d
ay w
an ad
C
or N
nd s Au st ra lia Sw ed en Sw itz er la nd
ar k m
he rla
N et
D en
Ic el a
nd
0
Fig. 1. Participation levels in tax-facilitated pension schemes (Percentage of the number of employees) Note: Hungary is not included in this table, because in Hungary the mandatory private pension schemes do not enjoy any tax credits. Source: Pearson and Martin (2005).
166
P.J.A. van Els, M.C.J. van Rooij, and M.E.J. Schuit
participation ensures a participation level of 100%. It is striking that this is not the case in Australia or Switzerland. This is because employers are not required to provide a pension scheme for employees with a low income. In Switzerland this is related to the fact that these employees accrue a sufficiently high pension by means of the basic pension from the first pillar, and then no longer qualify for a supplementary pension via the employer. In countries with voluntary schemes, such as the United States (US), the United Kingdom (UK), Canada and Norway, the participation in pension schemes is above 50%. In these countries, relatively many employers offer their employees a voluntary supplementary pension scheme. In Canada and Norway, employees are compelled to participate if a collective contract has been agreed between employers’ organisations and employees’ organisations. In the US, employees are not compelled to participate in the pension schemes offered by the employer. Most of the employees can choose whether they become members of the so-called 401(k) plans6 that many employers offer. These tax-facilitated, private, work-related, voluntary pension savings schemes (DC schemes) have largely replaced traditional DB schemes in the past twenty years. Of the employees who were offered 401(k) plans, in 2004 almost 80% actually became members, of whom only 11% paid in the legally permitted maximum (Munnell and Sundén, 2006). In the UK, since 1987, employees were no longer compelled to participate in the pension scheme of the employer (opting out) or in the supplementary state pension (state second pension). Instead of this they could choose a ‘Personal Pension’. This is a private defined-contribution scheme that is contracted with an insurer. Approximately 25% of the employees left the state pension or the employer pension scheme in the period 1988-1992, and chose such an individual DC plan (Disney et al., 2003). A large proportion of these employees changed on the wrong grounds, because they were misinformed, which was known as the ‘misselling’ scandal in the UK. In countries where the statutory mandatory participation works via the employee, Mexico and Poland, a large share of employees is still not participating in a funded pension scheme. This is caused by the transitional situation in which these countries find themselves. The funded pension schemes were introduced in Poland in 1998. The mandatory participation was introduced there for employees who were born after 1968, whereas participation is voluntary for those born between 1948 and 1968. In Mexico, the reform of the pension system was implemented in 1997, in which 6
These plans are named after section 401(k) of the US Internal Revenue Code of 1978, which sets rules for the offering of savings plans.
9 Why mandatory retirement saving?
167
statutory mandatory participation is only imposed on new members. It is expected that because of mandatory participation, the future participation levels in these countries will increase to 100% (Antolin, 2004). No figures are available in the majority of countries concerning the participation of the self-employed in pension schemes. Also, in countries where mandatory participation has been implemented via the employees, participation of the self-employed is optional in most of the cases. Reasons put forward for this is that the income of the self-employed is more difficult to calculate and that mandatory participation is harder to enforce (World Bank, 1996). The comparison shows that high participation levels are only achieved in statutory mandatory or quasi-mandatory schemes. In countries with voluntary schemes, the participation level of employees is lower, as a result of which fewer employees accrue a supplementary pension. This is the reason that in several countries discussions are currently going on concerning the possibility of introducing mandatory participation (Ireland, UK), or concerning ways to provide more guidance to participants in pension schemes (US). This type of guidance (libertarian paternalism) can be achieved by making participation in the employer’s pension plan automatic, unless the employee explicitly requests otherwise. More generally, this raises the issue of whether autonomy leads to lower or even inadequate savings for retirement? This is the main theme of the second part of this chapter.
9.3
Is mandatory participation necessary?
The current Dutch pension system is predominantly DB in nature. The pension funds play a central role in the system. Employees have virtually no influence on the level of the pension contribution, nor on the investment policy. In an individual DC system, employees themselves have much more control of the contribution and the asset mix. An advantage of this freedom of choice is that, in principle, employees can more easily bring their pension savings into line with their personal preferences and circumstances, provided they are capable of doing so. At the same time, employees themselves are responsible for the planning and adequacy of their pension savings and they bear the risk of disappointing investment results. In this section, we examine the problems with which the public see themselves confronted in choosing between consuming and saving, and investing their personal assets. We do this on the basis of the literature on household financial decision making, which is mainly focused on the ex-
168
P.J.A. van Els, M.C.J. van Rooij, and M.E.J. Schuit
periences in the US. The recent behavioural finance literature on households’ financial planning and decision making shows that individuals have difficulty with making consistent and prudent choices regarding saving and investing. Underlying forces are inertia, procrastination, inconsistent saving and investment behaviour, and insufficient financial knowledge on the part of the public. Procrastination and self-control Inertia and procrastination with respect to financial decisions can be an expression of a self-control problem. The theoretical basis of this has been elaborated in more detail by Thaler and Shefrin (1981). Inertia and procrastination frequently lead to insufficient saving. It is illustrative that taking part in saving for retirement strongly depends on the system used: is participation the standard option or default (participation is automatic unless the employer opts out) or does it require an active action by the employee to apply or confirm? In the first case, participation in pension savings schemes turns out to be considerably higher. This follows from a survey by Madrian and Shea (2001) based on data from a large US listed company in the healthcare and insurance sector. The size and nature of savings also heavily depend on the default options concerning the level of the contribution and the allocation of the investments. Bütler and Teppa (2005) show that pensioners who have the option of taking out the accrued pension capital as a lump-sum payment, as an annuity, or as a mixture of both, almost always select the alternative they are offered as default. Individuals who are aware of their lack of self-control can go in search of mechanisms to cope with it. They can decide, for example, other than might be expected on the basis of portfolio theory, not to consider their financial reserves simultaneously. In their financial planning, they then keep separate ‘mental accounts’ in the hope that they do not dig into their retirement savings for consumption in the short term (Prast, 2004), or they seize opportunities to commit themselves to save for their retirement (see Thaler and Benartzi, 2004). Procrastination may not only be the result of a lack of self-control, but can also be related to the phenomenon that people discount the benefit of future consumption in a hyperbolic manner (Laibson, 1996). In the economic literature, this is associated with myopia. Experimental research among test subjects indeed confirms that the ratio between the benefits of consumption now and in the future varies depending on the horizon to which the choice problem is related (Thaler and Benartzi, 2004).
9 Why mandatory retirement saving?
169
Inconsistencies in investment behaviour Inconsistencies in saving and investment behaviour can be the result of poor insight into how investment decisions translate into accrual of assets and pension benefits. Benartzi and Thaler (1999) show, for example, that individuals, even though pension decisions are about investing for the long term, place too much emphasis on the short-term risk of bad investment results. Benartzi and Thaler relate this behaviour to myopic loss aversion, loss aversion combined with short-sightedness. For this reason, decisions made are not always consistent with people’s actual individual preferences. Individuals are furthermore susceptible to the way in which alternative investment portfolios are presented – a phenomenon known as framing (see Kahneman and Tversky, 1984) – and are inclined to choose for the median portfolio, even if this clearly deviates from the actual portfolio they prefer (Benartzi and Thaler, 2002). The median investment portfolio is composed in such a way that half of the individuals choose more shares and the other half chooses less shares in the portfolio. The same applies for the other investment categories, such as bonds. Experiences with DC schemes in the US The situation in the US shows that participants in individual DC schemes are usually underinsured. They deposit too little, take too risky investment decisions (e.g. by investing partly in shares of the company they work for) and, with a change of job, are tempted to spend accrued pension savings that are paid out, instead of transferring these to another pension savings scheme (Diamond, 2004). Of the employees who changed jobs in 2004, 45% took their money, in spite of having to pay a penalty of 10% over it (Munnell and Sundén, 2006). Additionally, Mitchell, Utkus and Yang (2005), and Engelhardt and Kumar (2005), show that employees in the US make no use or not the best possible use of the opportunities offered to them. Not only does participation in DC pension schemes, sponsored by employers, of the so-called 401(k) type turn out to be far from complete, but also the deposits of employees are relatively insensitive for whether or not the employer contributes to the savings (employer matching contributions). This means that employees leave money lying on the table, as it were. These and other studies indicate that complete freedom of choice and high flexibility in the US are accompanied by insufficient pension income. In the US, where previously employees were more or less left to their own devices, employers are now encouraged in the new Pension Protection Act to choose a more guiding approach whereby, through offering standard
170
P.J.A. van Els, M.C.J. van Rooij, and M.E.J. Schuit
options, more employees accrue pension savings. It is still too early to judge whether this approach leads to good results without damaging the individual autonomy that is traditionally considered of paramount importance in the US.7 Financial illiteracy and retirement planning in the US Poor financial knowledge and expertise, also known as financial illiteracy, play an important role in individuals failing in the domain of saving for retirement. For the US, Hilgert, Hogarth and Beverly (2003) identify a positive relationship between financial education on the one hand and financial behaviour on the other. Americans who deliberately plan their pension savings, generally have more capital available (see Lusardi, 2003; Ameriks, Caplin and Leahy, 2003). Financial illiteracy is very common. Even in the US, where traditionally many people invest, financial expertise seems to be limited. In a recent study based on a survey among Americans aged 50 or over, Lusardi and Mitchell (2005) observed that a large part of this group has no understanding of economic concepts such as equity risk, inflation and compound interest.8 Moreover, it is found that people with more understanding of these basic financial concepts not only more often have a pension plan, but also are more successful in sticking to this plan. It could be concluded from this that a higher level of financial literacy leads to better financial planning of household finances and pension savings. Thus, financial literacy contributes to a more balanced pattern of spending over the life cycle. In an ageing society, this is likely to foster economic stability.
9.4
Empirical research for the Netherlands
What would happen if Dutch citizens were completely autonomous in their saving behaviour for retirement? Do employees want autonomy and, if so, would they be able to cope with it? On the basis of the existing litera7
A cause for concern in this is that many of these ‘default’ options assume a low pension contribution. The fear that a higher contribution will be at the expense of the total number of participants does play a role in this. Besides this, companies offer a standard investment alternative with few shares included. The danger of this is that ultimately many employees may indeed accrue a pension, but that this pension is relatively modest and expensive (Munnell and Sundén, 2006).
8
In particular, this applied to women, ethnic minorities and less educated people.
9 Why mandatory retirement saving?
171
ture and experiences in the US particularly, question marks about the capacity of individuals to make judicious choices themselves with respect to the size and composition of pension savings are in order. Roughly speaking, the relevant research for the Netherlands to be discussed here can be split into two questions of interest: 1. Do individuals prefer (more) autonomy in the pension domain (section 9.4.1), and 2. Can they cope with this autonomy, in other words, are they capable of saving for a pension on their own (section 9.4.2)?
9.4.1 DO DUTCH CITIZENS WANT TO BE ABLE TO CHOOSE AND DECIDE FOR THEMSELVES? DNB Household Survey Below we discuss recent research into the financial behaviour of individuals in the Netherlands. In doing so, we base ourselves on various studies that have been performed using the DNB Household Survey (DHS). In the DHS, formerly known as CentER Savings Survey, each year approximately 1,500 households are questioned about their financial characteristics and behaviour.9 In addition, there is room for questionnaires and experiments that have been specifically tailored to a particular topic. Due to a balanced selection of the panel members, the panel is representative for the Dutch population.10 In the studies discussed below, the response is generally around 1,500 individuals or 1,000 employees, depending on the target group. Opinion on mandatory participation The preference among employees for autonomy in the pension domain has been more closely analysed for the Netherlands by Van Rooij, Kool and Prast (2007) on the basis of the DHS.11 They asked employees whether they considered the mandatory nature of pension savings as an advantage or a disadvantage. As it turns out, three-quarters (77%) of the respondents 9
See www.uvt.nl/centerdata/dhs for more information about the DHS and the CentERpanel on which it is based.
10
The results presented here have, moreover, been weighted for age, gender, education, and income.
11
See also Prast, Van Rooij and Kool (2005).
172
P.J.A. van Els, M.C.J. van Rooij, and M.E.J. Schuit
Fig. 2. Opinion concerning mandatory participation (Percentage of the number of employees) Source: Van Rooij, Kool and Prast (2007) and own calculations.
consider this as an advantage, 12% as a disadvantage, whereas 11% pronounced no preference or said they did not know (figure 2). Four out of seven supporters of the mandatory nature of pension savings welcome the fact that they did not have to invest time and effort considering their own pension contribution. In addition, supporters of mandatory pension savings fear procrastination and a lack of discipline on their own behalf. This emerged from the fact that a third of them stated that left to themselves they would not save sufficiently for retirement. Pension preferences Since 2003, the DHS has contained several annually recurring questions and propositions that provide more detailed insight into the attitudes and preferences of the Dutch concerning their pension. See Van Els, Van den End and Van Rooij (2004) for an extensive analysis of the first results from 2003. Figure 3 provides a graphical impression of the results for both 2003 and 2005, the last year for which full information is available.12 The interest in the private pension provision has not changed much since the baseline measurement of 2003. In 2005, a small minority of 7% of the respondents keep themselves well informed on developments regarding their pension, 43% say they don’t worry about it now (‘we’ll see to that when we come to that’), whereas 40% indeed attach importance to the
12
The first issue of the provisional data for 2006 seems to indicate that the picture at the aggregate level is more or less the same as that for 2005.
9 Why mandatory retirement saving?
173
Fig. 3. Propositions concerning private pension provision (Percentage of the number of respondents) Source: DNB Household Survey, 2003 and 2005.
174
P.J.A. van Els, M.C.J. van Rooij, and M.E.J. Schuit
pension being well organised, but don’t want to know the details. The changes compared to 2003 are minimal. There has also been very little change in the public’s preferences concerning the certainty of pension benefits. Of the respondents, 53% would rather pay more contribution if this were for a guaranteed pension. This is a smaller group than in 2003.13 The group that would rather pay less contribution for a pension without guarantees has increased somewhat to 17% (was 15%). Concerning the influence on their own pension arrangements, 53% of the respondents still say that they gladly leave the pension build-up to the pension fund of the employer. Moreover, the proportion of respondents that want a say in how the pension deposits are invested has dropped from 21% to 17%,14 whereas an unchanged small minority of less than 10% would like to be able to select a pension fund themselves for the management of their pension savings. It can be concluded from this, that the need for autonomy has tended to decrease rather than increase during recent years. The proportion of respondents that provide no answer has increased by 4 percentage points. Van Rooij, Kool and Prast (2007) have also polled the preference for individual autonomy, but in a hypothetical situation of a DC pension scheme. It seems plausible that someone who bears the investment risk would also want to take responsibility for investing the pension contributions. Nevertheless, around half of the respondents would rather leave the investment to a pension fund. More than a quarter opted for more influence on the investments, the rest were indifferent (approximately 10%) or didn’t know (15%). Employees who attribute good financial skills to themselves and demonstrate a lower risk aversion, want more often to decide on the investment portfolio themselves. Apparently those who have a lot of trust in their own financial expertise assume that they can accrue a pension that matches their own preferences better. Pension preferences of the self-employed The respondents underlying figure 3 also include approximately 75 selfemployed. Although the analysis is not aimed primarily at this group, it can be explored, however, whether their preferences differ substantially from those of the average respondent. It was found that in 2005 the self13
A comment here is that the question does not contain a quantification of how much more contribution would be necessary for this (see also DNB, 2004).
14
In this context, the level of the eventual pension benefit payments depends on their own decisions.
9 Why mandatory retirement saving?
175
employed were even less occupied with their pension than the average Dutch person: half of them said they did not worry about their pension arrangements (compared with 43% for the average respondent). Selfemployed people have a stronger preference for a pension without guarantees against a lower contribution, although in absolute terms the number of self-employed with a stated preference for a certain pension against a higher contribution is still larger. In addition, the self-employed comparatively want more influence on their pension scheme then the average Dutch person. Incidentally, almost half of the self-employed stated that the issue of ‘autonomy versus leaving pension build-up to a pension fund’ was not considered applicable. Possibly, the fact that a large share of the selfemployed has no pension scheme arrangement at all plays a role in this. All in all, these findings of recent survey research indicate that a majority of Dutch employees can happily live with the mandatory nature of pension savings, appreciate the security of a DB system, and have no preference for more autonomy in the pension domain.
9.4.2 ARE DUTCH PEOPLE CAPABLE OF SAVING FOR THEIR PENSIONS THEMSELVES? Suppose that people would increasingly be given more autonomy in the pension domain, would they then be capable of making sensible choices? More generally, the question is whether the risks of autonomy, such as procrastination and self-control problems, inconsistent choices with saving and investment decisions, limited financial skills and the like, which emerge in the international literature, would apply to the same extent among Dutch employees. Substitution of retirement savings Indications of procrastination also playing a material role in the Dutch pension context follow from the answers that DHS respondents give to the question of whether they expect to adjust their saving behaviour if existing pension schemes were retrenched (see table 2). Approximately 35% of the respondents do not adapt their saving behaviour, but postpone a decision about this under the motto ‘I’ll see to that when I come to that’. Moreover, more than 20% do not know what to do in such a situation. This group has not become smaller since 2003, despite the major attention paid by the media during recent years to the possible insufficiency of Dutch retirement savings.
176
P.J.A. van Els, M.C.J. van Rooij, and M.E.J. Schuit
Table 2. Saving behaviour if pension schemes were to be retrenched, 2003 and 2005 results (Percentage of respondents) Would you adjust your saving behaviour if pension schemes were to be retrenched? Yes, I would put aside extra money myself for my pension
No, I’ll see to that when I come to that
No, I can fairly easily Don’t know/no opinion manage with my pension
2003
25
36
18
21
2005
29
33
15
23
Source: DNB Household Survey, 2003 and 2005.
Inconsistencies in investment decisions A similar picture emerges regarding investment decisions. Van Rooij, Kool and Prast (2007) asked Dutch individuals how, within a DC scheme, they would themselves divide their investment portfolio between bonds and stocks, conditional on the basic assumption of an unchanged contribution. The average stock percentage respondents desired in the portfolio was 30% and that was also the median. In a follow-up survey, the respondents were shown two different retirement income schemes in which the future pension income for each respondent depends on the returns on the investment portfolio during their working life. Each scheme showed a monthly pension income with successively a very favourable, an average and a very unfavourable return on the investment portfolio. The calculation of the pension benefits was based on the current income from employment, forty years contributions paid in, and realistic assumptions on the level and volatility of real stock and bond returns. Without the respondents being informed of it, the calculation of one of the pension income schemes was based on their own choice for the desired stock percentage in the portfolio. The other scheme had been based on the outcomes of the median portfolio (30% stocks and 70% bonds). The respondents could state their appreciation for both income schemes on a scale of 1 (very unattractive) to 5 (very attractive). Those people who initially had chosen a conservative investment portfolio (20% or less in stocks) assessed the outcomes of the more risky median portfolio with 30% stocks on average higher than that of their own portfolio (3.6 against 2.8). People who had originally chosen a relatively risky portfolio (40% or more in stocks) gave this a higher rating than the median portfolio (average 3.5 versus 3.0). After seeing the two pension income schemes, roughly 60% of the respondents preferred the results of the most risky investment portfolio; the other 40% were indifferent or chose the less risky portfolio. Thus the choice for the percent-
9 Why mandatory retirement saving?
177
age of stocks in the investment portfolio in this experiment does not correspond with the preferences with respect to the expected pension benefit and the risks therein. Knowledge of their own pension arrangements Another important topic of applied research in the field of pension autonomy that emerges prominently in the international literature is financial illiteracy. Research into the Dutch situation shows that people’s knowledge of their own pension arrangements leaves much to be desired and, in a more general sense, that their financial knowledge and skills are limited. Moreover, this problem is recognised by Dutch employees themselves. When asked, two-thirds of them consider themselves to be more or less financially ignorant (Van Rooij, Kool and Prast, 2007). Van Els, Van den End and Van Rooij (2004, 2005) show that Dutch citizens are not very consciously occupied with their pension. This follows from the answers to a 2003 questionnaire asking what Dutch citizens know about their own pension arrangements (see table 3), which have been repeated in subsequent years. In 2003, 40% of the respondents did not know whether his or her pension scheme was based on final or average pay, or depended on the returns on the contributions paid. Almost half of them (44%) did not know whether the pension was indexed. Despite the fact that they had received a pension overview, 60% said they did not know the amount of pension rights accrued so far, and 63% had no idea of the expected level of the pension benefit that would be paid from 65 years of age. The figures for 2005 show that the knowledge concerning the type of pension scheme and the level of the pension benefit has increased slightly compared to 2003.15 The reverse applies to the knowledge concerning indexation and accrued pension rights. That said, the results clearly indicate that the knowledge concerning their own pension arrangements increases with age. Other personal characteristics that positively affect pension knowledge are education, income and stock ownership. Furthermore, men appear to be better informed than women.
15
Although the knowledge of employees about their own pension arrangements was limited in 2003, it seemed the public were properly aware of the discussion surrounding the sustainability of pension schemes; see Van Els, Van den End and Van Rooij (2004). A large proportion of the respondents said then that they took into account that within more or less ten years time the General Old Age Pensions (AOW) would be retrenched.
178
P.J.A. van Els, M.C.J. van Rooij, and M.E.J. Schuit
Table 3. Knowledge of own pension arrangements, 2003 and 2005 results (Percentage of respondents) Age (in years)
Proportion of respondents who know about: Type of pension scheme
Indexation
Accrued pension rights
2003
2005
2003
2005
2003
16-24
30
22
22
11
0
25-34
38
48
35
30
35-44
52
53
53
45-54
63
70
59
55-64
73
78
65 and older
81
Total
60
2005
Level of pension 2003
2005
0
19
12
22
21
27
39
42
31
29
35
41
53
47
38
41
46
64
67
66
64
48
54
84
81
81
n/a
n/a
n/a
n/a
64
56
53
40
36
37
43
Source: DNB Household Survey, 2003 and 2005.
Financial knowledge Besides knowledge about their own pension scheme, a certain level of general financial literacy helps the public taking decisions in the pension domain. This applies even more if individual autonomy and responsibility concerning the size and investment mix of the retirement savings will become more important in the future. Van Rooij, Lusardi and Alessie (2007) conclude, however, that the general financial knowledge of the Dutch public is limited. They derive this from the answers to five simple questions concerning interest rates and inflation and eight questions concerning investing for testing the most basic financial knowledge and skills. No more than 40% of the respondents gave the correct answer to all the questions concerning interest rates and inflation. Underlying this, there were especially poor scores on the questions concerning compound interest, inflation and money illusion. In addition, the basic knowledge of Dutch people in the area of investing is certainly no better. Only 6% of respondents answered all 8 questions correctly, and less than 20% gave the correct answer on at least 7 questions. In particular, the knowledge about bonds is very poor. This is most probably related to the fact that only 5% of Dutch citizens invest in bonds. It is remarkable, furthermore, that half of the respondents did not know that normally speaking, over a longer period of, say, ten or twenty years, stocks provide a higher return than bonds or savings accounts. They also did not know that investing in the stock of a single company provides a less certain return compared to a mutual fund. A more
9 Why mandatory retirement saving?
179
detailed analysis showed that men, higher educated people and those who actively invest in stocks have better financial knowledge. People who score well on financial knowledge questions, moreover, have a more accurate picture of macroeconomic developments as measured by economic growth and inflation (DNB, 2006). It can be assumed that a more accurate perception of economic growth and inflation contributes to the quality of the decisions that households make concerning their financial future.
9.5
Experiences with freedom of choice in the Netherlands
Life-course savings scheme The introduction of the life-course savings scheme (‘Levensloopregeling’) from 1 January 2006, and the response of employees to this, can be considered as a sort of test case for the exploring to what extent the behavioural inadequacies discussed above apply in practice to the pension planning of Dutch citizens. After all, this introduction brings with it a large degree of individual autonomy concerning early retirement. As a result of the new Act on amendment of tax treatment of early retirement and the introduction of the life-course savings scheme (Wet Aanpassing fiscale behandeling VUT/prepensioen en introductie levensloopregeling – VPL Act), many employees are confronted with a retrenchment of the scheme for early retirement applicable up until then (there is a transitional scheme for employees born before 1950). At the same time, the life-course savings scheme has been introduced as an attractive possibility for saving individually in a ‘taxfriendly’ manner, in order to fund the intended early retirement. Nevertheless, only 8% take part in the life-course savings scheme, of whom 58% say they are saving to be able to retire early.16 This followed from a representative survey in July 2006 among the members of the DHS panel. The tax treatment forms an important reason for participating in the life-course savings scheme (table 4). Moreover, the contribution of the employer plays a large role, certainly when it is considered that a third of the life-course participants receive no employer’s contribution. It should be noted that the employer’s contribution should in fact not be an issue. After all, the employer is required to also pay a contribution to employees who do not participate in the life-course savings scheme. For one in ten partici16
Saving for a sabbatical (15%) or parental leave (10%) are the other most important reasons.
180
P.J.A. van Els, M.C.J. van Rooij, and M.E.J. Schuit
pants, the fear that they would otherwise use the contribution for other purposes plays a role. Reasons for not participating are many and various. One important reason is the forced choice between either participation in the another taxfriendly contractual savings scheme (‘Spaarloonregeling’) or in the lifecourse savings scheme. More than 30% of the employees having a birth year of 1950 or later state the most important reason not to participate in the life-course savings scheme is that the contractual savings scheme is more attractive for them. A quarter has problems with the restrictions that the scheme imposes. The majority of them prefer to save themselves in order to keep their options open, but others find themselves too old to still take part, find the scheme too complicated or too expensive, expect that the life-course savings scheme is not fated to have a long life, or object to the dependence on permission from the employer for taking the compensated leave. On top of these, almost 10% postpone the decision, because they want to see how the life-course savings scheme will develop in practice. One in ten employees has anyway hardly or not at all taken the effort to examine the life-course savings scheme in depth. More than 10% of employees state the most important reason for not taking part is that the contributions cannot easily be missed. The broad range of reasons of whether or not to participate in the life-course savings scheme illustrates that Dutch citizens too are sensitive to the factors that complicate choice in the pension domain. It thus confirms the earlier discussed findings from the international literature and the experimental research on the basis of hypothetical survey questions for the Netherlands. Self-employed Finally, the benefit of the mandatory participation can be examined on the basis of the differences between the responses of employees and the selfemployed to the question of whether they consider the level of their pension accrual (state pension (AOW) combined with a company pension) adequate, or have made additional provisions. The majority of the selfemployed, after all, are not subject to a mandatory participation. It should be noted that the results of this recent survey in June 2006 among the members of the DHS panel is only indicative, because the number of selfemployed in this analysis, including freelancers and liberal profession groups, is limited with only 50 respondents. As to employees, half of them consider the level of the accrual to be good as it is, and a quarter find the pension accrual too low or far too low. As could be expected, there is a large difference with the opinion of the self-employed: 60% of the self-employed
9 Why mandatory retirement saving?
181
Table 4. Reasons for whether or not participating in the life-course savings scheme (Percentage of the number of employees) Reasons for participating in life-course savings scheme instead of themselves saving for compensated leave (n=62)1 – Tax benefits
65
– Contribution from my employer
32
– Otherwise I wouldn’t save
11
– I expect I will more easily receive permission for leave from my employer
6
– Other reasons
6
– Don’t know
10
Most important reasons for not taking part in the life-course savings scheme (n=626) – Other contractual savings scheme is more attractive
31
– Saving themselves and keeping options open
16
– Cannot easily miss the money
13
– Complicated / expensive / too old / dependence on employer / no faith in scheme’s continued existence
10
– Not considered it / forgot / too much trouble / didn’t know it existed / didn’t know how
10
– Isn’t necessary
8
– I’ll wait and see
7
– Other reasons
2
– Don’t know
3
Source: DNB 1) Because respondents could indicate several reasons, the sum of the percentages exceeds 100%.
consider the pension accrual too low (and usually much too low). In both cases, virtually 20% of the respondents say they don’t know. It is striking that there are no differences between employees and the self-employed concerning schemes that have been joined voluntarily for supplementing the mandatory provisions. Approximately 40% have arranged supplementary measures, mostly by means of annuities and single-premium policies. More than half of the self-employed who have not arranged supplementary measures, however, state that they actually should save more for their pension. These results indicate that people who are not under the mandatory participation are generally less satisfied with their pension accrual; at the same time, in far from all cases do they succeed in setting sufficient money aside themselves for their pension provision.
182
P.J.A. van Els, M.C.J. van Rooij, and M.E.J. Schuit
Summarising the above, on the basis of empirical research, it can be observed that procrastination and self-control problems not only affect the American, but also the Dutch pension savers. There are also indications that, with individual autonomy, the Dutch savers will be sensitive to the way in which the alternative investment mixes are presented to them. Finally, research has shown that the financial knowledge of the Dutch leaves a lot to be desired. This concerns not only the knowledge of pension provisions, but also financial knowledge in general terms concerning topics such as interest, inflation and investing.
9.6
Evaluation
The Netherlands has no direct statutory mandatory participation for retirement savings, but a system of quasi-mandatory participation via collective contracts. Because of this, in practice, more than 90% of Dutch employees automatically save for their pension. Experiences in other countries show that some form of mandatory participation is necessary to get the participation percentage up to a high level. Moreover, households and individuals encounter difficulties with their pension planning that go further than the decision on whether to save or not. At least equally important are decisions concerning the level of the pension contributions and the management of the pension reserves. In countries with a large freedom of choice, many people accrue an inadequate pension. The empirical research performed for the Netherlands shows that there is no reason to assume that Dutch employees would actually behave differently. The knowledge of basic financial concepts is limited and investment expertise and skills are poor. Moreover, adequate knowledge of own retirement and pension arrangements is sadly lacking. In addition, Dutch people also demonstrate the behavioural traits that emerge prominently in behavioural finance literature. Procrastination, self-control problems and myopia frequently stand in the way of accrual of a proper pension. This means that, for a system with a large degree of individual autonomy in pension accrual to be successful in practice, a number of important basic conditions are currently missing. Moreover, the Dutch public is well aware of this. They rather prefer to take no risks concerning their pension accrual, and a majority supports mandatory participation and a system with little autonomy and as much security as possible.
9 Why mandatory retirement saving?
183
Can we conclude from all this that the current Dutch mandatory participation for employees by means of collective contracts and the elaboration of this by social partners and pension funds are optimal? In a narrow sense, the Dutch system scores well against the experiences abroad; the participation level is high, the pension deposits are considerable and the pension capital is managed by experts. In a broader sense, however, that conclusion cannot be drawn on the basis of the analyses in this chapter. A proportion of employees as well as many of the self-employed still fall outside the mandatory participation. In this context, therefore, discussion is actually still possible concerning which form of compulsion is better: via the employer, via the employee or a possibly adjusted version (quasimandatory participation) via collective contracts. Moreover, the question whether the system of collectiveness works to increase prosperity for the society as a whole goes beyond the scope of this chapter. In this context, the impact of collective pension agreements on international competitiveness, competition on national markets and labour participation is also important. More fundamental is the question of how far government and social partners must or can go in taking important pension decisions out of people’s hands. In contrast to those who benefit from this, however, there is also a small group of people who want freedom of choice to have their pension provisions more in line with their own preferences and who do not need to be protected against themselves. It is interesting in this respect that in many other countries there are experiments with pension systems that rely to varying degrees on the consumers’ individual responsibility. It is clear that there is a broad spectrum of possibilities between completely mandatory participation and full individual autonomy, in which it is not immediately clear, depending on the formulated objectives, which alternative will lead to the best results. This chapter has shown that the current Dutch system of mandatory participation leads to beneficial results for the average consumer. The research results discussed here and experiences abroad do not provide any reasons to introduce drastic changes in pension autonomy. Nevertheless, the coverage rate of the mandatory participation is still open for improvement (certainly if the aim is also to include the self-employed). A small step towards more freedom of choice would possibly open the opportunity for tailoring pension provisions a little more to the personal situation of pension scheme members, while they are nevertheless safeguarded from the major ‘slips’ in the pension domain.
184
P.J.A. van Els, M.C.J. van Rooij, and M.E.J. Schuit
Literature Ameriks, J., A. Caplin and J. Leahy, ‘Wealth Accumulation and the Propensity to Plan’, Quarterly Journal of Economics 68, pp. 1007-1047, 2003. Antolin, P, A. De Serres, A. and C. De La Maisonneuve, ‘Long-term budget implications of tax favoured retirement plans’, OECD Economics Department Working Paper no. 393, 2004. Benartzi, S., and R. Thaler, ‘Risk Aversion or Myopia? Choices in Repeated Gambles and Retirement Investments’, Management Science 45(3), pp. 364-381, 1991. Benartzi, S., and R. Thaler, ‘How Much Is Investor Autonomy Worth?’, Journal of Finance 57(4), pp. 1593-1616, 2002. Bütler, M., and F. Teppa, ‘Should you take a lump-sum or annuitize? Results from Swiss pension funds’, CEPR Discussion Paper 5136, 2005. De Vos, K., R.J.M. Alessie and P.F. Fontein, ‘Pensioenpreferenties’, Economisch Statistische Berichten, 15 May 1998, pp. 398-399. DNB, ‘Financial behaviour of Dutch households’, DNB Quarterly Bulletin, De Nederlandsche Bank, September 2004, pp. 71-84, can be downloaded from www.dnb.nl. DNB, ‘Financial behaviour of Dutch households’, DNB Quarterly Bulletin, De Nederlandsche Bank, June 2006, pp. 47-54, can be downloaded from www.dnb.nl. Diamond, P., ‘Social Security’, American Economic Review 94 (1), pp. 1-24, 2004. Disney, R., C. Emmerson and S. Smith, ‘Pension Reform and Economic Performance in Britain in the 1980s and 1990s’, NBER Working Paper 9556, 2003. Engelhardt, G., and A. Kumar, ‘Employer Matching and 401(k) Saving: Evidence from the Health and Retirement Study’, DNB Working Paper 79, 2005. Hilgert, M.,J. Hogarth and S. Beverly, ‘Household Financial Management: The Connection between Knowledge and Behavior’, Federal Reserve Bulletin, pp. 309-322, 2003. Kahneman, D., and A. Tversky, ‘Choices, Values and Frames’, American Psychologist 39, pp. 341-350, 1984. Laibson, D., ‘Hyperbolic discount functions, undersaving, and savings policy’, NBER Working Paper 5635, 1996.
9 Why mandatory retirement saving?
185
Lusardi, A., ‘Planning and Saving for Retirement’, Working paper, Dartmouth College, 2003. Lusardi, A., and O.S. Mitchell, ‘Financial literacy and planning: implications for retirement wellbeing’, DNB Working Paper 78, 2005. Madrian, B.C., and D.F. Shea, ‘The power of suggestion: inertia in 401(k) participation and savings behavior’, Quarterly Journal of Economics 116, 2001, pp. 1149-1187. Ministry of Social Affairs and Employment, Nationaal Strategierapport Pensioenen 2005, The Hague, July 2005. Ministry of Social Affairs and Employment, Nationaal Actieplan Pensioenen Nederland 2002, The Hague, August 2002. Mitchell, O.S., P. Utkus and T. Yang, ‘Better plans for the better-paid: determinants and effects of the 401(k) plan design’, Pension Research Council Working Paper 2005-5, Pension Research Council, Wharton School, University of Pennsylvania, 2005. Munnell, A.H., and A. Sundén, ‘401(k) plans are still coming up short’, Issue in Brief, March 2006, Number 43, Center for Retirement Research at Boston College, 2006. OECD, Pensions at a Glance, Public Policies across OECD Countries, OECD Publishing, 2005. Pearson, M., and J.P. Martin, ‘Should we extend the role of private social expenditure’, OECD, Employment and Migration Working Papers 23, 2005. Prast, H.M., ‘Psychology in Financial Markets: an Introduction to Behavioural Finance’, Financial Monetary Studies, Amsterdam: NIBE-SVV publishers, 2004. Prast, H.M., M.C.J. van Rooij and C.J.M. Kool, ‘Werknemer kan én wil niet zelf beleggen voor pensioen’, Economisch Statistische Berichten, 22 April 2005, pp. 172-175. SER, ‘Witte vlekken op pensioengebied, quick scan 2001’, SER, 12 February 2002. Thaler, R. and H. Shefrin, ‘An economic theory of self-control’, Journal of Political Economy 89, pp. 392-406, 1981. Thaler, R.H., and S. Benartzi, ‘Save More Tomorrow: Using Behavioural Economics to Increase Employee Saving’, Journal of Political Economy 112 (1), pp. 164-187, 2004.
186
P.J.A. van Els, M.C.J. van Rooij, and M.E.J. Schuit
Van Els, P.J.A., W.A. van den End and M.C.J. van Rooij, ‘Pensions and public opinion: a survey among Dutch households’, De Economist 152(1), pp.101-116, 2004. Van Els, P.J.A., W.A. van den End and M.C.J. van Rooij, ‘Financial behaviour of Dutch households: analysis of the DNB Household Survey. In: Investigating the relationship between the financial and real economy, BIS Papers 22, pp. 21-40, 2005. Van Rooij, M.C.J., C.J.M. Kool and H.M. Prast, ‘Risk-return preferences in the pension domain: are people able to choose?’, Journal of Public Economics 91, pp. 701-722, 2007. Van Rooij, M.C.J., A. Lusardi and R.J.M. Alessie, ‘Financial Literacy and Stock Market Participation’, DNB Working Paper, forthcoming, 2007. World Bank, Viewpoint, Designing Mandatory Pension Schemes, World Bank, February 1996.
10 Mandatory participation for companies
P.H. Omtzigt1 In the Netherlands, when representative organisations apply for it, the Minister of Social Affairs can decree a pension scheme as mandatory for a complete industrial sector. The majority of the employees accrue pension in a scheme that is imposed by this mandatory participation for companies. This chapter describes the creation and content of this mandatory participation, as well as the advantages and disadvantages. A comparison with other countries shows that the mandatory participation for companies succeeds well in its objectives, as it encourages saving behaviour, prevents ‘blank spots’ and avoids competition on labour costs.
10.1 Introduction For a hundred years already, one question has remained central to the social security in the Netherlands: should the state organise an aspect of this, or can social partners do it together? And if a task is left to the social partners, should the state make insurance mandatory, and should the insurance be publicly or privately implemented? In the Netherlands, employers and employees and industries have had a large role in formulating and executing social security schemes, much more so than in neighbouring countries. The industry-wide pension provision is one of the clearest examples of this.
1
The author thanks Renske Biezeveld and Joos Nijtmans for their major support in writing of this article. Furthermore, he thanks Erik Lutjens, Fieke van der Lecq and Onno Steenbeek for comments on an earlier version. This article is written in a private capacity.
188
P.H. Omtzigt
Mandatory participation for companies has existed in the Netherlands since 1949. This stipulates that, at the request of a sufficiently representative proportion of an organised industry, participation in the pension fund can be made mandatory for all employers in that industry or business sector. Non-organised employers are also obliged to contribute to the industrywide pension fund for their employees in this manner. At the heart of the debate is, on the one hand, the emphasis on a large scope of application for supplementary pensions and industry-wide solidarity and, on the other, the lack of freedom of choice for an individual and lack of competition and a level playing field between pension funds and insurers.
HISTORY The company pension is one of the oldest social securities in the Netherlands. The first pension scheme was introduced as long ago as 1879: that was for the Nederlandse Gist- en Spiritusfabriek Van Marken (Van Marken Dutch yeast and methylated spirits factory), based on collective participation and a jointly redistribution of the contributions. Other industrial enterprises also set up their own pension funds (Stork in 1881 and Philips in 1913). Some time later funds were also set up in several industries. Lutjens (1999) mentions, among others, the Coöperatief Verzekeringsfonds (cooperative insurance fund - 1917), the Algemeen Mijnwerkerspensioenfonds (general mineworkers’ pension fund - 1918), two regional pension funds in the cigar industry (1933) and a pension fund for the bookbinding industry (1934). All this time there was not really a legal framework. The Dutch insurance supervisory board (De Verzekeringskamer), set up in 1923, did not yet have supervisory powers over pension funds (Bakker et al., 1998). The immediate reason and political pressure to introduce an act governing company pension funds was the large number of funds that were set up soon after World War II. Furthermore, social partners were about to set up a fund for the farming industry. The number of members in that fund could well exceed half a million. It was therefore important to rapidly pass legislation which protected the rights of the members. A fund in the farming sector, with lots of small companies and relatively low incomes, is always difficult to regulate on a voluntary basis. The solidarity between employers was an explicit justification for mandatory participation. In addition, mandatory participation prevents competition on payroll costs or conditions of employment within an industry. The mandatory participation also limits the blank spots (the number of employees without pension facilities), because all employers in a business sector are compelled to provide a pension scheme. The pension law also stipulates that one cannot
10 Mandatory participation for companies
189
exclude part-timers or females. New employees must start accruing rights within two months. In 1954, the Pension and Savings Funds Act became effective, regulating the definition of pension funds and their supervision. But even prior to that, in 1949, the Act on Company Pension Funds, which regulated mandatory participation in a pension scheme for the employers in an industry, was passed. The only exception was for employers who participated in a company pension fund or had already set up an insurance scheme at least six months before the submission of the application for industry-wide mandatory participation. That exemption also only applied if the pension was at least equivalent. The bill received a lot of attention in parliament. Ultimately, only the communist party voted against it. The number of pension funds and the number of members rapidly expanded during the 1950s and 1960s. In 1960, 64 industry-wide pension funds were operating, of which 48 funds had mandatory participation. These developments have led to the Netherlands currently having a unique position in the world: more than 90% of employees accrue supplementary pensions, while there is no mandatory participation in a pension plan. In the Netherlands there are a number of tax advantages that makes the accrual of a pension attractive: the tax reversal rule means that contributions are deductible and the pension benefit is generally taxed at a lower rate. Pension capital is not subject to a 1.2% yield tax per annum, which, accumulated over a period of 40 years, makes a significant difference. Furthermore, accrued rights with a pension fund are not taken into account when applying for rental allowance or social security benefit. Besides these tax advantages, which occur elsewhere in comparable forms, the government in the Netherlands also stimulates the accrual for pension in another way. The mandatory participation for individuals, which stipulates that employees are required to become members of a pension scheme offered by their employer, has been examined in the previous chapter. Equally important is the mandatory participation for companies: a pension scheme and a pension fund will be made mandatory for all employees in this sector, whenever employers and employees in an economic sector apply for it. Only under strict conditions, discussed later in this chapter, can an employer obtain dispensation. In addition, the Act stipulates mandatory participation of professionals like GP’s, notaries and physiotherapists in their respective funds. The table below shows the relationships between the various pension funds.
190
P.H. Omtzigt
Table 1. Pension funds in the Netherlands, 2004
Mandatory participation for individuals
Company pension fund
Mandatory participation for companies and individuals
Industry-wide pension fund
Act on professional pension Professional pension fund funds
Number of funds
Active members
Assets (billions)
627
900 000
€ 169.9
98
5 320 000
€ 365.2
15
48 000
€ 19.0
The mandatory participation for individuals can also be placed with an insurer instead of a pension fund. Source: CBS, 2006.
Since the 1980s, the real returns of pension funds have increased considerably and the funds have also become major players on the capital markets. The privatisation of the civil servants’ fund ABP in 1996 means that this fund also counts as an industry-wide pension fund. As a result, total assets of all Dutch pension funds together have exceeded national debt since 1998.. The assets even remained at a good level during the extremely miserable investment years at the start of the new millennium, as table 2 shows. Table 2. Development of industry-wide pension funds No. of funds
Capital (billions)
in % GDP
Active Members (thousands)
1950
4
€ 0.04
0.5%
286
1960
64
€ 1.24
6.5%
978
1970
87
€ 3.93
6.8%
1 231
1980
82
€ 17.20
10.7%
1 528
1990
79
€ 50.34
20.7%
1 882
2000
91
€ 316.20
75.7%
4 425
2004
98
€ 365.20
74.6%
5 320
Since 2000: including the ABP, privatised in 1996. Source: Lutjens, 1999 and CBS, 2006.
10 Mandatory participation for companies
191
10.2 Why mandatory participation? The most important reason for mandatory participation is that it dramatically increases the number of people accruing a pension. In an ideal world, every individual saves sufficiently for the situation in which he or she is too old or too sick to work. In a properly operating insurance market, everyone would then, at retirement, buy lifelong benefit, an annuity. Without an obligation to save, it is found that short-sightedness and poor financial knowledge frequently lead to insufficient saving for retirement (see also Van Els et al. in this volume). That is why many OECD countries and a majority of the other countries set up pension systems, in which participation is mandatory. Individuals are hardly capable of assessing their pension scheme and making rational choices. The pension misselling scandal in the United Kingdom, in which from 1988 people could exchange a collective scheme for a private product, illustrates this problem. The costs of the private product were much too high, as were the promised returns. Aggressive selling techniques led millions transferring and they generally suffered a substantial loss. Furthermore, there is also the selection impact in the annuities market. In the world of rational economic models, people who themselves know that their life expectancy is relatively low, should not participate in pension funds. People with higher incomes tend to live longer and should insure the most. Thus, those with the ‘worst risks’ should insure themselves. Pension funds and insurers can and should not make any distinction between groups of people on the basis of their income and their health situation. The latter is legally prohibited. That is why there are uniform rates. For someone with a short life expectancy, this means that it is extraordinarily unattractive to take out a pension. Finkelstein and Poterba (2002) show that this obligation for equal rates has a heavy impact on the benefits. Without mandatory participation, this impact doubles due to the selection effects that then occur. Furthermore, insurers more often use waiting and threshold times, or other mechanisms, to prevent people insuring themselves after a risk has already arisen. But this, of course, also disadvantages people who become sick or die shortly after commencing employment. The mandatory participation helps to mitigate this impact. After all, people must always be insured and cannot wait until they think that they will soon need it. But this, of course, also disadvantages people who become sick or die shortly after commencing employment. It can also be rational not to save, because: ‘the government will provide. And if you save, you ‘lose’ your
192
P.H. Omtzigt
rights to benefits. But this, of course, also disadvantages people who become sick or die shortly after commencing employment. Mandatory participation also means that very little needs to be spent on acquiring customers (see also the contribution of Bikker and De Dreu, elsewhere in this volume). The objective of having as many people as possible accrue a pension has indeed been achieved in practice. Since 2001, the percentage of over-65 households with a low income has been lower than the same percentage for the category of under-65 households. This makes the Netherlands the only country in the world where relative poverty among old people occurs less than poverty in the entire population (SCP, 2005).
10.3 Solidarity Pension funds have various forms of risk sharing and solidarity. For example, there is equal treatment of men and women, a medical examination prohibition, and contribution-free accrual with disability. In addition, there is intergenerational solidarity and investment risks are shared between the generations. Mandatory participation industry-wide pension funds, however, have yet another form of solidarity applicable, specifically between employers. For individual members, the mandatory participation guarantees an affordable and accessible pension scheme. In the paper on greater flexibility and mandatory participation, the Cabinet proposed that the mandatory participation also serves two important Cabinet objectives: reintegration of occupationally disabled and chronically ill patients, and increasing the labour participation of older employees by levying a uniform contribution (House of Representatives of the Dutch Parliament, 1996-1997a). The principle of uniform contribution, as described in other chapters in this book, can in fact be carried through to companies. Companies with a relative young workforce always have solidarity with companies with a relatively old workforce. An industry-wide pension fund also ensures that the members in the pension fund do not carry the individual operating risk of bankruptcy, as is the case with company pension funds. Even after bankruptcy, the members with rights already accrued in the fund will benefit from indexation for pay rises or price inflation in the same manner as the members from companies that continue to exist.
10 Mandatory participation for companies
193
However, there are also risks that are not absorbed by an industrywide pension fund. There are industries that are ageing heavily. With disappointing investment results, a group of members that is becoming smaller must maintain the indexation and benefits for pensioners by paying higher contributions. Adjustment in such a fund is more difficult. In theory, a fund with few new members joining could reduce many risks by merging with a pension fund in a thriving sector. If the returns are better than expected, the assets of the closed pension fund would be surrendered to the younger fund. If the returns fall short of expectations, an increase of contribution with the younger fund ensures an extra contribution to the older fund. The larger the collective, the more risks can be shared. A state pension fund should be in an even better position for this. But then again, this would create other problems: it is not possible to customise pension plans for sectors. And almost all state pension funds that started as funded, during the course of time have become pay-as-you-go schemes due to political intervention. One example is the ‘social security’ in the United States. Even then country-specific risks, such as risks of hyperinflation, stagnation of the national economy or even war, still exist. In theory, Pan- European pension funds could be an answer. In practice, there are many legal and other issues making this inconceivable in the short term. However, multinationals can set up funds that operate in several countries.
10.4 Resistance to mandatory participation Beside the benefits of mandatory participation, there are also potential disadvantages. Since mandatory participation pension funds are not subject to the free market, they have theoretically less incentives to efficient operation. Yet in practice it is found that these funds do better than insurers concerning cost efficiency, among other things, due to economies of scale and the absence of marketing costs. Moreover, the absence of a profit motive means that returns fully benefit the members. Although it is not possible for participants to leave their pension fund, pension funds increasingly aim towards improving service and transparency concerning their performance, to legitimise the lack of freedom of choice. Also the test on the z-score, which is discussed later in this chapter, encourages funds to obtain sufficient investment performance. The discipline of the free market only works to a certain extent for insurers too; because the lack of complete transparency makes it more difficult for customers to make a choice
194
P.H. Omtzigt
(House of Representatives of the Dutch Parliament, 2000-2001a). The compulsory element is also used as an argument against the mandatory participation, particularly by insurers. They would gladly provide people or individual companies with pension products. Whether or not you consider this as a problem depends on your perspective. The further the pension scheme is removed from your ideal scheme, and the worse the performance of the pension fund, the greater the desire for either an individual member or an employer to find a pension scheme outside the pension fund. There have been many appeals with reference to European regulations. In particular it has been argued by the insurers that freedom of choice is too severely restricted and that pension funds are de facto unfair competitors of insurers because of tax and cost advantages (Thijssen, 2000). From a number of rulings from the European Court of Justice (Albany, Brentjens and Drijvende Bokken), it has meanwhile become clear that the mandatory participation in an industry-wise pension fund is not contrary to the European competition rules. Agreements that social partners have entered into in the context of collective negotiations and that are meant for improvement of the terms and conditions of employment, because of their nature and objective fall outside the ban on cartels (Lutjens, 2000). The other objections to the mandatory participation also apply to mandatory participation for individuals. Mandatory solidarity in a single scheme leaves little room for individual freedom of choice. From a theoretic perspective, the pension accrual can also be too high. If someone is single, moreover, the survivor’s pension can be exchanged for a higher old-age pension. Because people aged 65 and over pay a lower tax rate and social security contribution, one can say they have a higher net income, when there is talk of a generous pension scheme. That is sub-optimal, because such a single person perhaps would have wanted more disposable income during his or her working life. A number of pension funds try to compensate for the lack of freedom of choice by offering more options. Some options, such as the legal right to exchange survivor’s pension for a higher old-age pension, must be offered by a pension fund. Funds also offer to separate the cover for survivor’s and disability pension, and extra can be saved on a voluntary basis. This reduces the pressure from the mandatory participation. These possibilities are, however, limited. On the one hand, this is due to the relatively high ambition level of the pension schemes, but, on the other, because of the legally embedded demarcation lines between pension funds and insurers.
10 Mandatory participation for companies
195
10.5 Expanding the opportunities for exemption In addition to voluntary exemption, the original Act on professional pension funds from 1949 provides one possible exemption from the industrywide pension fund (BPF): there must be an existing pension scheme, which was at least equivalent. In 2000, the successor to the BPF Act was handled in the House of Representatives of the Dutch Parliament. In 51 years, the pension system had developed into one of the most advanced and richest in the world. The joint pension facilities in the Netherlands meanwhile amounted to more than the gross domestic product. Via the pension funds, Dutch workers in fact now owned the capital: a situation that Marx and Engels had not considered achieving via this path. The opposition against mandatory participation did not now come from the left, but from the centre-right VVD party; the state secretary Hoogervorst (VVD) posed the question ‘are no alternatives conceivable, with which a touch of the mandatory participation is settled, but that perhaps in another manner nevertheless meet the objective that we both want, specifically that people can accrue as good a pension as possible?’ (House of Representatives of the Dutch Parliament, 2000-2001b). Eventually the law was adopted, in spite of being voted against by the VVD party. One of the most important amendments extended the number of grounds for exemption (2000 Exemption decree, BPF Act). The new act provides four grounds for exemption from participation, namely the old basis of exemption related to existing pension provision, exemption in connection with formation of a group, exemption in connection with an own collective labour agreement, and exemption related to inadequate investment performance. In all these situations, exemption must be granted, provided further set conditions are met. Thus a company’s own pension scheme, except when this has been agreed with collective labour agreement (CAO), should be at least equivalent to the scheme of the industry-wide pension fund. An industry-wide pension fund can itself also grant an exemption on other grounds.
MUTUAL SETTLEMENT With all new exemption grounds, the industry-wide pension fund (BPF) can ask for a reimbursement for the technical insurance disadvantage that the fund suffers. This disadvantage is related to the fact that for employers
196
P.H. Omtzigt
with a membership that is younger than the average within the industry, withdrawal from the BPF can produce a contribution advantage. For a younger membership, after all, a lower uniform contribution must be paid. In this case, the BPF can ask a reimbursement for this disadvantage. The reimbursement of the technical insurance disadvantage prevents enterprises with good risks taking their pension provisions elsewhere and placing them at a lower contribution for this reason. Otherwise, this would lead to an increase of the uniform contribution for the BPF, making it more attractive for the remaining enterprises with better risks also to withdraw from the BPF, and so on. Over the course of time this would lead to the end of the industry-wide pension fund (House of Representatives of the Dutch Parliament, 1996-1997a).
THE PERFORMANCE TEST The exception with bad investment result is possibly the most striking ground for exemption. The performance test, also known as the z-score, measures the extent to which the formulated investment objective of a fund has been achieved, compared to a benchmark chosen by the pension fund itself (the standard portfolio) over a rolling period of five years. The standard portfolio, in which the investment mix is fixed, is determined annually by the fund’s governing board given the risk profile that it is set on the basis of studies of the expected development of the obligations in relation to those of the assets. The requirement for the test on the z-score has the positive effect of providing pension funds an incentive to compare their achieved results with those of similar investors. However, the test also has disadvantages, which will emerge even more clearly when the new Financial Assessment Framework (FTK) takes effect on 1 January 2007. In the existing test, there are only very limited possibilities for adjusting the standard portfolio. With heavy fluctuations in market returns (e.g. the interest) the pension fund can be forced to revise its policy under the FTK. However, this means deviating from the benchmark, creating the possibility that either a pension fund undeservedly fails the z-score test, or it takes investment decisions that are not optimal given the circumstances. One solution to this would be to grant pension funds more opportunities to adjust their standard portfolio. A total of three funds have failed the test since the first five-years test. This has nevertheless only resulted in one case of an employer withdrawing.
10 Mandatory participation for companies
197
10.6 Alternatives for the mandatory participation for companies If you take the desirability of the widest possible participation in a pension scheme as a basic principle, are there alternatives for mandatory participation for companies? A paper concerning alternatives for the mandatory participation and job demarcation of the Cabinet (House of Representatives of the Dutch Parliament, 2000-2001a), listed four alternatives, in which the number of participating employees would not decrease. This could be: 1. a statutory participation pension scheme obligation, with which all employers are compelled to form a collective pension scheme with a legally prescribed minimum level, as in Switzerland; 2. a statutory participation obligation with a uniform pension scheme, with which employees have a free choice for a pension administrator, as in Chile; 3. enforcement of the current mandatory participation to a maximum salary, above which employees have the freedom to arrange a supplement; 4. mandatory participation at industrial sector level, with which it is not the participation in the pension fund, but the participation to a pension scheme at industrial sector level that is mandatory. Employers could then have the freedom to choose an administrator for the scheme. All these alternatives have their own advantages and disadvantages. In the first two examples it is clear that the options for the employee do not increase. In all examples, solidarity will be reduced or disappear unless supplementary measures are taken for this, although the freedom of choice for the individual does indeed increase. The then coalition government therefore stuck to its point of view that ‘the mandatory participation is still a good instrument for achieving and maintaining a high funding ratio in the field of supplementary pensions.’ (House of Representatives of the Dutch Parliament, 2000-2001a). This actually returned to earlier suggestions concerning restriction of the mandatory participation (House of Representatives of the Dutch Parliament, 1996-1997b), which incidentally had also already been rejected by the Social and Economic Council of the Netherlands (SER). These suggestions consisted mainly of restriction of the accrual per annum and restriction of the mandatory participation to the maximum daily pay for social security purposes.
198
P.H. Omtzigt
ABROAD The US and UK have chosen to promote participation mainly with tax arrangements. This leads to a participation rate of 50% to 60%, compared to more than 90% in the Netherlands. It is noticeable that it is actually the relative weaker groups in American society who remain deprived of pensions, as is clearly shown in table 3. Only half of employees take part in a pension scheme. Especially among small firms and in sectors such as farming, participation rates are low, which is a major difference with the Netherlands. Table 3. Membership in pension plans of employees (21 to 64 years old) in the United States All employees, 1987
46.1%
All employees, 2004
48.3%
Man, 2004
49.4%
Woman
47.2%
Fulltime employees, (working the full year)
56.6%
Part-time employees, (working the full year)
25.7%
Companies, more than 1000 employees
59.2%
Companies, less than 10 employees
16.0%
Government
75.8%
Farming, mining and construction
28.6%
Source: Copeland, 2005.
Besides encouraging participation in a pension fund, the continuity of schemes is a cause for concern for governments. Various governments have attempted to form a fund. Social Security in the US was once intended to be a funded scheme, but has become a pay-as-you-go system. In addition, the government in the US has set up a safety net for pension funds that get into problems: the pension benefit guarantee corporation. With an inadequate funding ratio, this organisation takes over the pension commitments, though partially reduced. This fund is indeed financed by the pension funds, but in extremis is covered by the government. But it is that individual operating risk that is actually shared in the Netherlands in an industry-wise pension fund. Crucial for the success in Chile is the fact is that the pension funds are private bodies at a distance from the government. Especially in countries that have regularly had to deal with nationalisations and changes of regime, such as in Chile, a clear ownership structure is important. Norway is one of the few countries which have succeeded in retaining a funded scheme within the public sector. The country is assisted in this by
10 Mandatory participation for companies
199
the enormous oil revenues. Nevertheless, even there the pressure is great to grant more rights entitlements and loosen the reins. Finally, Sweden also has an interesting hybrid system. In addition a to a large pay-as-you-go system, employees contribute 2.5% of their salary and invest it in one of around 600 approved investment funds. This does create a funded scheme, but without any solidarity. Solidarity could be brought into the system, however, for example by an equalisation system, such as the one that exists in the Dutch healthcare system, in which the employers’ contributions are shared. For a chronically ill patient, an insurer gets a lot of extra money from the ‘pot’. Cross-subsidising should take place between the individual accounts: then there would be explicit crosssubsidising from men to women (longevity risk) and from young to old (time-weighted proportional accrual of rights). Such an equalisation could, however, still be technically complicated and would very explicitly show that young men with low salaries transfer money to, for example, older women with high incomes. When survivor’s pension is also insured, this is mitigated slightly because the life expectancy of dependants is reversed. Yet designing the system would be difficult and loaden with practical and political problems. It seems that ultimately only the Netherlands and Switzerland have succeeded in getting a pension system off the ground with almost universal participation and a funded system, of which the Dutch scheme is a lot more flexible. This shows that the mandatory participation for companies has provided a major contribution to a flexible and future-proof system.
10.7 Final considerations: challenges and opportunities Industry-wide pension funds performed well compared to company pension funds. The increased requirements for professionalisation of pension funds paved the way for taking advantage of economies of scale. This has led to the position where industry-wide pension funds are now the largest owners of capital in the Netherlands. This finds a major responsibility in their role as shareholders. In Norway, this pressure is already so great, that the state’s oil fund can no longer invest in certain enterprises, such as the Franco-Dutch Thales company. Public pressure is building up on Dutch pension funds to cease investing in companies that produce for instance landmines or cluster bombs or exploit child labour.
200
P.H. Omtzigt
The accounting guidelines can provide further support. Companies must include investment gains and losses of company pension funds in their operating result. According to the Dutch RJ 271 guideline, this is not necessary with industry-wide pension funds. The question to what extent this obligation applies to listed companies is still being discussed. One challenge is the larger variety in members’ careers: where the majority of schemes have been set up with a view to breadwinners who work fulltime, the current population is often more varied. Part-time work, divorces, remarriages and greater mobility on the labour market make a uniform scheme increasingly constrictive. For this reason the legislator increased the number of exchange rights. Section 2b in the Pension and Savings Funds Act has provided the right to exchange survivor’s pension for old-age pension since 2001. Section 55, an implementation of the OmtzigtDepla motion, provided the right to the reverse of this, exchanging old-age pension for survivor’s pension. The first is very important for single people, the second can be very valuable in case of remarrying or when the pension scheme does not provide for survivor’s pensions. Pension funds and social partners, however, do not need to wait for legislation to obtain greater flexibility. They can provide tailored solutions to industrial sectors for the risks of old age, disability and death. Shortly after the Second World War, mandatory participation for companies became effective and this contributed greatly to a supplementary pension provision being arranged in the Netherlands for almost all employees. Also, without there necessarily being an empathetic solidarity with an industrial sector, the economies of scale of a large collective, coupled to mandatory risk sharing within an industrial sector, provide benefits which would be difficult to arrange in any other manner. Precisely in an increasingly complex society, government intervention to help people with pension accrual continues to be necessary. Mandatory participation is then a smart form, which does not directly intervene in the employment relationships, but nevertheless ensures that sufficient money is saved. Because these savings are outside the immediate sphere of government intervention, creation of a funded scheme in the Netherlands has been a success.
10 Mandatory participation for companies
201
Literature Bakker, R., E. Haaksma, T. Kool, K. Verhagen and C. de Wijs, Toezien of toekijken?, verzekeringskamer 75 jaar, Apeldoorn: Stichting de Verzekeringskamer, 1998. Copeland, C., ‘Employment-Based Retirement Plan Participation: Geographic Differences and Trends, 2004’. EBRI Issue Brief 286 (Employee Benefit Research Institute, October 2005), 2005. Dutch Social and Cultural Planning Board, Armoede monitor 2005, The Hague, 2005. Finkelstein, A., and J. Poterba, Selection Effects in the United Kingdom, 2002. House of Representatives of the Dutch Parliament, Flexibilisering en verplichtstelling, 25 014 no. 1, 1996-1997a. House of Representatives of the Dutch Parliament, Nota Werken aan zekerheid, 25 010 no. 2, 1996-1997b. House of Representatives of the Dutch Parliament, Taakafbakening tussen pensioenfondsen en verzekeraars, 26 537 no. 4, 2000-2001a. House of Representatives of the Dutch Parliament, Proceedings, 5, 235-251, 2000-2001b. Individual Annuities Market, The Economic Journal, 112, pp. 28-50. Lutjens, E., Een halve eeuw solidariteit, 50 jaar wet betreffende verplichte deelneming in een bedrijfstakpensioenfonds, Amsterdam: VU, 1999. Lutjens, E., ‘Taakafbakening pensioenfondsen-verzekeraars’, Tijdschrift voor Pensioenvraagstukken, no. 2, 2000, pp. 31-35. Statistics Netherlands (CBS), CBS Statline, www.cbs.nl/statline, 2006. Thijssen, W.P.M., ‘Van verplicht gesteld naar individueel pensioen’, Het Verzekeringsarchief, no. 1/2, 2000, pp. 36-44. Vrijstellingsbesluit Wet BPF 2000, Staatsblad, no. 633, 2000.
Part 4. Conclusion
11 Macroeconomic aspects of intergenerational solidarity
J.P.M. Bonenkamp, M.E.A.J. van de Ven, and E.W.M.T. Westerhout1 This chapter addresses the macroeconomic gains and losses of intergenerational solidarity. The benefits are mainly in better risk sharing, the losses are particularly formed by a distortion of the labour market. Besides risk sharing, intergenerational solidarity also leads to intergenerational redistribution, of which the benefits are not always clear. On the basis of the current insights, the benefits seem to exceed the losses.
11.1 Introduction Dutch supplementary pensions are characterised by a large degree of solidarity, both between and within generations. Solidarity is a great benefit. It unites various individuals and groups in society (see Jeurissen and Sanders in this volume). Solidarity, however, also has negative macroeconomic side effects. It cannot be excluded in advance that the side effects dominate and, on balance, solidarity does more harm than good. Moreover, there is still also the ageing of the population that is advancing: there are fewer and fewer young people to have solidarity with a growing group of old people. This drives benefits and losses of solidarity apart, and can result in something that initially worked well, reversing and becoming a disadvantage. There is, in short, sufficient reason to closely re-examine the costs and benefits of solidarity in the supplementary pension schemes. This chapter chooses a macroeconomic perspective for this. It starts with a classification of solidarity in general, and afterwards focuses on intergenerational soli1
The authors would like to thank Peter Kooiman, Casper van Ewijk and both editors for their comments on earlier versions.
206
J.P.M. Bonenkamp, M.E.A.J. van de Ven, and E.W.M.T. Westerhout
darity. This is followed by a qualitative exploration of the positive and negative effects of intergenerational solidarity. Subsequently we attempt to identify the scope (in cash flows) of intergenerational solidarity. It will be clear that unambiguous answers are difficult to provide, but an indication of the order of magnitude of various effects is actually possible. After this we attempt to quantify two important aspects of solidarity: the value of intergenerational risk sharing and the costs of distortions of the labour market. All identified elements come together in the final considerations.
11.2 Aspects of intergenerational solidarity CLASSIFICATION As Jeurissen and Sanders make clear, there is broad support for the idea of solidarity in society. However, it is not always clear in advance what this term means. In this chapter, we interpret solidarity as financial transfers between several (groups of) people. The adjective financial is added here to emphasise that, with pensions, it exclusively concerns financial transfers. Given this definition, we can distinguish between two forms of solidarity.2 First of all there is solidarity which comes into operation after a person for whatever reason suffers a loss and other people entirely or partially compensate him or her for that. This form of solidarity is the basic principle of insurance (e.g. fire insurance). Since this solidarity is conditional to the occurrence of an event, this also called ex post solidarity. In addition, there is a type of solidarity that is independent of a certain event occurring, but in advance leads to a certain redistribution between participants. This type of solidarity is not related to the insurance idea and is called ex ante solidarity.3 There are many types of solidarity in supplementary pension schemes (see Kuné in this volume). These are not all equally visible; also little is known about their meaning for the identified groups. In this chapter we will examine intergenerational solidarity in more detail. This intergenerational solidarity is related to the financial transfers be2
This distinction is based on De Laat et al. (2000).
3
De Laat et al. (2000) also distinguish a third type of solidarity: income solidarity. This occurs, among other things, with the General Old Age Pensions Act (state pension, AOW). The AOW benefit is independent of income, but the contributions (up to a cap) are income-related. With the supplementary pensions, this type of solidarity plays a smaller role since not only contributions, but also pension benefits are income-related.
11 Macroeconomic aspects of intergenerational solidarity
207
tween current and future generations, between working and retired generations, and between various cohorts of working generations. As with solidarity in general, it is also possible to distinguish between ex ante and ex post intergenerational solidarity. Ex post intergenerational solidarity occurs for example after the funding ratio of the defined benefit (DB) pension fund has endured a negative shock and the younger and future generations mainly absorb this shock, whereas older and retired generations remain mainly protected. In contrast, windfalls mainly benefit the younger generations. In a defined contribution (DC) system, everyone saves for themselves, so there is no question of ex post solidarity. Ex ante intergenerational solidarity operates independently of an uncertain event occurring. As we will see below, the uniform contribution rate leads to ex ante intergenerational solidarity. The distinction between ex ante and ex post solidarity is not always clear and, in a certain sense, is artificial. A simple example may make this clearer. Suppose that the investments of the pension fund decrease in value. Without a pension fund, mainly the cohorts with relatively large amounts of risk-bearing pension assets would be affected. By charging an extra contribution and cutting indexation, however, the pension fund can redistribute between cohorts. After a decline in share prices, the cohorts with relatively little risk-bearing pension assets will have solidarity with the groups of members with relatively large amounts of risk-bearing pension assets. This occurs because those mentioned first transfer to the latter groups. Suppose then, that the same shock occurred in the past and that the pension fund has not yet fully repaired its funding ratio. Generations that enter the labour market, and join this pension fund, can currently calculate that according to expectations they should pay in more than they will ever receive from the fund. These generations are, as it were, in solidarity with older generations by taking over a proportion of the funding deficit of the pension fund from them. The first case is undoubtedly a case of ex post solidarity. There are two possible lines of reasoning for the second case. As we will see later in this chapter, in existing literature the second case is also considered as ex post solidarity. After all, a pension contract contains an implicit agreement with future generations to always communicate all unexpected surpluses and deficits. You can counter this by saying that a generation only enters into a pension contract at the moment that it effectively enters the labour market. Reasoning along these lines, existing deficits and surpluses are more a type of ex ante solidarity than ex post solidarity for future generations.
208
J.P.M. Bonenkamp, M.E.A.J. van de Ven, and E.W.M.T. Westerhout
Note that in this example ex post solidarity includes an implicit sharing of the investment portfolio among the various generations. If, for example, the pensioners would be spared entirely from a negative assets shock, then that implicitly means that their pension rights are fully funded by risk-free investments. The working and future generations then implicitly ‘own’ all risk-bearing investments from the pension fund portfolio (see also Boender et al. in this volume). In this view, the intergenerational risk sharing between older, younger and future generations replaces a pension fund’s explicit transactions on the capital market. In other words, the pension contract specifies implicitly how the proprietary rights to the pension assets are shared among several generations.
11.3 Intergenerational solidarity in the second pillar Intergenerational solidarity is interwoven into various places in our system of supplementary pensions. The articles of association and regulations of pension funds can or do lead to intergenerational transfers. Moreover, the government’s legislation and regulations also play an important role. In the second pillar, we can distinguish three sources of intergenerational solidarity, in particular solidarity as a result of: • funding deficits; • the uniform contribution and accrual system; • deferred taxation system. The most striking is the intergenerational solidarity that operates when pension funds have underfunding. Through a combination of increases in contributions and cuts in indexation, pension fund deficits are shared among currently employed people, pensioners and future generations. In the previous section we saw that the distinction between ex ante and ex post solidarity is not always clear for this form of solidarity. This is different for the uniform contribution and accrual system.4 Dutch pension funds are up to a certain level legally obliged to charge contribution rates on the basis of the uniform contribution and accrual system. Be4
In imitation of Boeijen et al. (Chapter 7), we define the uniform contribution and accrual system to be a combination of a uniform contribution with a timeweighted proportional pension accrual.
11 Macroeconomic aspects of intergenerational solidarity
209
cause of the time-weighted proportional basis of the pension rights accrual, this leads to young people paying more than the value of the right they obtain. Old people actually pay less than the value of the obtained right. This type of solidarity is independent of the financial position of a pension fund and is therefore a pure form of ex ante solidarity. This also applies to the tax grant that is locked into the second pillar. By means of the deferred taxation system, the government grants a considerable tax allowance to retirement saving. Responsible for this are the lower rate for people over 65, the progression in the tax system, and the exemption of the assets of pension funds for the capital yield tax. Whether these subsidies imply redistribution, and in what form, is not clear. This depends entirely on how the government finances these subsidies. If the financing is a burden on the same generations as those who receive the subsidy, these government subsidies are neutral. A tax that is only paid by people in employment, is a good example: after all, it is the working people who enjoy the subsidy. With other sources of financing, however, there is an aspect of redistribution, namely in favour of the working generations. The mandatory participation5 does not in itself lead to intergenerational solidarity, but it is an important pre-requisite to make intergenerational solidarity possible. It is impossible for pension funds to bind future generations without the mandatory participation. This would also mean that the possibility of having deficits partly or wholly paid for by future generations would be lost. Future generations could then always decide not to participate in the pension fund.
11.4 The benefits of intergenerational solidarity IMPROVED RISK SHARING Intergenerational risk sharing can increase welfare for the same reason that an insurance can increase the welfare of insured people: by spreading shocks over a larger group, the individual contribution can be lower than when everyone would have to form their own provision. The scope of the pension insurance is broad. In a pure DB pension system, the members’ pensions are protected from bad returns, due to share prices and interest rates. Pension insurance also offers protection against 5
For more information about the mandatory participation in supplementary pension schemes, see the contributions from Van Els et al., and of Omtzigt elsewhere in this volume.
210
J.P.M. Bonenkamp, M.E.A.J. van de Ven, and E.W.M.T. Westerhout
the risk of living longer. Moreover, if the pension has been set up as welfare-linked, it offers protection against shocks in the growth rate of wages. If the pension has been set up as inflation-proof, it protects against shocks in the general price level. However, this does not directly justify the raison d’être of a pension fund. With properly operating capital markets, there is very little to argue in favour of a role for pension funds. After all, as has already been observed, intergenerational risk sharing is nothing more than an implicit sharing of the pension investments among the various members. This sharing could also explicitly take place via the capital market. However, capital markets do not work well, at least not as well as in the textbooks.6 Private life insurance markets have to contend with adverse selection, as a result of which people with a low life expectancy will less rapidly take out an annuity insurance.7 Furthermore, the market for index-linked bonds is still poorly developed. Only a limited number of countries issue price-indexed bonds, wageindexed bonds are not available at all. It is also frequently impossible for young people to invest in shares with borrowed money, because the only collateral they have is their future earned income. For banks, however, this collateral is much too risky to lend money on. It is completely impossible to deal with unborn future generations on the private market.
HETEROGENEITY OF BUSINESS SECTORS AND ECONOMIES OF SCALE The organisation of intergenerational solidarity by means of large industry-wide pension funds gives economies of scale and also guarantees the heterogeneity between business sectors. The economies of scale and the non-profit nature of Dutch pension funds seem to even out the disadvantages of the limited competition. The picture is not completely clear, but research seems to indicate lower costs and better performance of pension funds than of insurers (see De Laat et al., 2000, for a summary; see the contribution of Bikker and De Dreu in this volume for a comparison of the implementation costs of pension providers). Note, however, that this benefit coincides with the collectivity and can also manifest itself in collective DC schemes. 6
See the contribution of Boender et al. in this volume for a more comprehensive summary of the market imperfections, sometimes called ‘missing markets’, which are partly or fully resolved by pension funds.
7
See Westerhout et al. (2004) for a discussion of empirical research into adverse selection. Brown (2004) discusses explanations for the limited demand for annuity insurance policies on the market in the US.
11 Macroeconomic aspects of intergenerational solidarity
211
COMPETITION Although there is no explicit competition between pension funds, there is, however, a question of implicit competition. Thus, for several years the investment performance of mandatory industry-wide pension funds has been mutually compared by means of the so-called z-scores. If a fund performs worse than a standard portfolio over a period of five years, an employer can apply for exemption from the mandatory participation. This also applies to collective DC schemes.
11.5 The costs of intergenerational solidarity LABOUR MARKET DISTORTIONS When a pension fund finances its funding deficit with income-related contributions, intergenerational risk sharing has consequences for the labour market. If a generation is asked to pay a higher contribution than it would expect to receive over its lifetime in benefits, then the excessive contribution acts as a tax. Pension savings are currently coupled to work: more contracted hours of work means that higher pension rights are accrued and that more contribution is paid. As a result, the tax on pension savings works de facto as a tax on labour. If a funding deficit obliges a pension fund to charge catch-up contributions, then that also reduces the labour supply, with lower employment and production as a result.8 If the funding is realised by charging income-independent contributions, the labour market effects mentioned above do not occur. Excessive contributions are in that case a loss that cannot be avoided by reducing the number of working hours offered. Another labour market effect can occur, however: people can resign. In general, such a response is less obvious than reducing the number of hours that people work. Nevertheless, stopping working actually is a realistic option for people with a relatively low income or a relatively high reservation wage. For old people with a relatively large preference for leisure, a limited increase in
8
Developments that lead to a funding surplus have an opposite effect on the labour force participation. However, we must attach more weight to funding deficits than to funding surpluses. This can be argued both on historical grounds and on theoretical grounds. On balance, it must be concluded that there are negative labour market effects.
212
J.P.M. Bonenkamp, M.E.A.J. van de Ven, and E.W.M.T. Westerhout
contribution can be a reason to retire early. Instead of no longer participating on the labour market, a member can also decide to seek employment elsewhere, in another company for instance (in case of a company pension fund) or in another business sector (in case of an industry-wide pension fund). Also, the possibility of giving up being an employee in exchange for becoming self-employed or emigrating to a foreign country should be kept in mind. If pension funds concentrate more on cuts in indexation instead of increasing contributions, labour market distortions are also reduced. Indexation cuts always affect the rights already accrued. The loss because of this cannot be reduced by working fewer hours. If it is expected that indexation cuts would be applied in future years, however, this can have a labour market effect: the currently accrued rights then always become less valuable due to the future indexation cuts, which distorts the current labour supply. Also, the principle of the uniform contribution rate can distort the labour market. The funding of pensions with a uniform contribution rate implies that younger cohorts pay more than they accrue in rights. For this reason, they could be encouraged to limit their labour supply. The reverse applies for older cohorts of employees: the cheap accrual of pension rights could be a reason for them to postpone retirement another couple of years and in this way contribute to enlarging the labour supply. In the exceptional case that the labour market elasticity of several cohorts is equally large, on balance, this results in no impact on the labour supply at macro level. In the event that labour market elasticity increases or decreases with age, the uniform contribution rate has a promoting or obstructing effect on the labour supply. However, we do not know of hard empirical evidence for these case-study positions. Elsewhere in this volume, Van Els et al. demonstrate that the knowledge of pension accrual becomes greater as the age progresses. This suggests a positive relationship between labour market elasticity and age. Lastly, the favourable fiscal treatment of supplementary pensions encourages labour force participation. Although the subsidy is mainly intended to encourage retirement saving, because of the mandatory participation, it is a subsidy on labour. This subsidy does not need to exist, effectively. If increasing the tax rate on income from labour finances the costs of it, de facto there is no question of subsidy. With other forms of funding, there is actually an implicit subsidy for labour.
11 Macroeconomic aspects of intergenerational solidarity
213
DISCONTINUITY RISKS When contribution payers or sponsoring companies withdraw, the continuity of the pension scheme is threatened. This discontinuity risk mainly occurs, as we will see below, when a fund is underfunded. The discontinuity risk is higher the more intergenerational risk sharing is organised by a larger number of small pension funds. For the larger the number of pension funds, the simpler it is to avoid the required payments by offering labour in another sector or company, or by exchanging employee status to become self-employed. If intergenerational solidarity were to be organised at national level, the discontinuity risk would be mitigated but not removed. After all, members of pension schemes may raise the mandatory participation for debate at political level or, in the worst case, decide to leave the Netherlands by emigrating. There are also discontinuity risks associated with the uniform contribution rate, because of the transfers from young to old that are embedded in it.
POLITICAL RISKS Capital that has been accrued within large collective pension schemes to fund future obligations, has an almost irresistible attraction on policymakers who are in search of financial resources. Arguments for creaming off pension assets are usually rapidly found: the future really is a very long way away; generations that live then will be richer than currently living generations; the need now really is very high, etc. Certainly, if there is to be any discussion on how obligations must be valued, the political risk lurks that policy-makers will want to appropriate a portion of the pension capital for other purposes.9 This is an argument for not allowing intergenerational risk sharing to be managed by the government, but by locally organised pension funds that have been placed at a distance from the government. But discussions can also arise within pension funds, concerning the sharing of the plusses and minuses among the various members. Consider, for example, the discussions around the sharing of the alleged surplus reserves of pension funds at the end of the 1990s.
9
But also the sector itself, from a short-term perspective, can plead for a high discount rate. For example, see the discussion concerning the discount rate for supplementary pensions in the US (The Economist, 2004).
214
J.P.M. Bonenkamp, M.E.A.J. van de Ven, and E.W.M.T. Westerhout
11.6 The scope of intergenerational solidarity In this section, we illustrate the scope of intergenerational solidarity. In doing so, we distinguish ex ante and ex post solidarity. We illustrate the scope of ex ante solidarity on the basis of the uniform contribution rate; the scope of ex post solidarity is considered on the basis of a shock on the stock market.
INTERGENERATIONAL REDISTRIBUTION: UNIFORM CONTRIBUTION RATE Size of intergenerational transfers The discrepancy between contributions paid in and rights accrued with uniform pricing causes intergenerational transfers. In order to provide an idea of the size of these intergenerational transfers, we compare the uniform contribution rate with an actuarially fair contribution rate. In the case of an actuarially fair contribution rate, each member pays an annual contribution that is equal to the cash value of the pension right. This actuarially fair contribution is break-even at individual level and thus also at fund level, but the uniform contribution is only break-even at fund level. The pension accrual in our calculations takes place according to a conditionally index-linked average-salary scheme. For this we use the following basic principles:10 • The pension fund guarantees a wage-indexed pension commitment. This means that the cost-covering pension contribution also funds future indexations. • The pension fund is in balance. This means that the fund has no funding deficit or surplus, so that no indexation cut occurs and no catch-up contribution is charged. Pension accrual and benefits are fully indexed to nominal wage growth of 3.7% per annum, divided into 2% inflation and 1.7% productivity growth.
10
In the contribution by Boeijen et al. to this volume (Chapter 7), comparable calculations are made. In a qualitative respect, the results are similar, but quantitatively they are not similar because of differences including probabilities of death, accrual percentage and membership. Furthermore, Boeijen et al. omit the survivor’s pension and constantly take the average contribution instead of the marginal contribution as used by us.
11 Macroeconomic aspects of intergenerational solidarity
215
• We use a stable population, both in size and in age composition. Age-specific probabilities of death are assumed constant. There is no migration. • The pension fund can invest in a single asset with a guaranteed real return of 3%. • We assume fair value of pension rights. Because we assume that there are no investment or other risks, the real discount rate of the pension rights is equal to the real return. • Each year, members accrue 2.25% of their contribution base as a pension right. This also includes an allowance for the survivor’s pension. The contribution base is defined as the gross wage less a General Old Age Pensions Act (AOW) franchise.11 The franchise increases annually with the nominal wage growth. The gross wage, on the other hand, increases not only because of the nominal wage growth, but also because of age-related career moves. • The marginal uniform contribution rate is determined by the total increase of rights divided by the aggregate contribution base. • Within a cohort, all members are homogeneous. Between working cohorts, members differ on three points: age, labour force participation and career. Between retired cohorts, of course, members only differ in age. • The contribution payments are fully attributed to the employee. This assumption implies that the employer’s part of the pension contribution is passed on to the employee.12 Figure 1 shows the uniform contribution rate and the actuarially fair contribution rate, both expressed as the percentage of the contribution base (marginal) and of the gross wage (average). Because of the constant population, Figure 1 can be interpreted in two ways, specifically as the division of contributions over the different age cohorts in a certain year, but also as 11
Because an employee also receives a statutory old-age pension benefit (AOW) from the age of 65, the supplementary pension is not accrued over the full gross wage.
12
For a small open economy such as the Dutch one, this assumption is justifiable for the longer term. After all, the prices of labour and capital are determined on the global markets. Because of the competitive position, the Netherlands cannot remain out of step in the long term. In the short term, however, there can be cases of incomplete transfers.
216
J.P.M. Bonenkamp, M.E.A.J. van de Ven, and E.W.M.T. Westerhout
% 40 35 30 25 20 15 marginal actuarially fair 10
marginal uniform average actuarially fair
5
average uniform 0 20
24
28
32
36
40
44
48
52
56
60
64
age
Fig. 1. Profile of the actuarially fair contribution rate and the uniform contribution rate
the contribution profile of a group of members over their career. Figure 1 shows that the marginal uniform contribution rate is constant for all ages and equal to 26% of the contribution base. The marginal fair contribution rate increases with age, starting at 19% at 20 years of age, up to 37% for a 64-year-old member. The increase of the marginal contribution rate reflects the increasing costs of the pension accrual during the career. The difference between the uniform contribution rate and the actuarially fair contribution rate is large for the young and old members. For a 20-year-old entrant, more than a quarter of his or her contribution payment is transferred to older generations. This means less than three-quarters of the pension contribution is used for pension accrual. A 65-year-old member receives, on the other hand, a subsidy equal to 41% of his or her contribution payment. The age at which the transfer turns into a subsidy in this analysis is around 44 years of age.13 For illustration, figure 1 also contains the average contributions as a percentage of the gross wage. The average contributions are calculated by
13
We have also examined the sensitivity of the results for alternative assumptions with respect to interest, productivity and career path. The size of the subsidies varies, the direction does not. In addition, the age of reversal remains virtually unchanged.
11 Macroeconomic aspects of intergenerational solidarity
217
multiplying the marginal contributions by the age-related ratio between the contribution base and the gross wage. Because we have assumed that the franchise grows with nominal wage growth and the gross wage with wage growth together with age-related career moves, the ratio between the contribution base and the gross wage increases with the member’s age. This explains the increase of the average uniform contribution rate. From the age of 54, the average uniform contribution rate remains constant at a level of 18%, because it has been assumed that from that age on a member makes no more career moves. The ratio between the uniform contribution rate and the actuarially fair contribution rate is the same for the average contributions and for the marginal contributions. Life-cycle redistribution versus intergenerational redistribution It is interesting to examine what the uniform contribution rate implies for the total contribution and benefit flows over the career. Does a member break even, or does the implicit burden in the young years weigh heavier or actually less heavy than the subsidy at an older age? The first case would mean that the uniform contribution is ultimately nothing else than an individual life-cycle redistribution. Figure 2 compares the uniform contribution rate with, on the one hand, the actuarially fair uniform contribution rate and, on the other, the actu% 40 35 30 25 20 15 actuarially fair contribution
10
uniform contribution 5
actuarially fair uniform contribution
0 20
24
28
32
36
40
44
48
52
56
age
Fig. 2. Life-cycle redistribution and intergenerational redistribution
60
64
218
J.P.M. Bonenkamp, M.E.A.J. van de Ven, and E.W.M.T. Westerhout
arially fair contribution rate as seen so far. The difference between the actuarially fair contribution rate and the actuarially fair uniform contribution rate is that the former is actuarially fair on an annual basis, whereas the latter is actuarially fair over the career. Using figure 2, we can more accurately attribute the intergenerational redistribution of the uniform contribution rate. A deviation between the uniform contribution rate and the actuarially fair uniform contribution rate implies that the uniform contribution rate is not only redistributive over the individual career, but also between generations. More specifically, the difference between the actuarially fair uniform contribution rate and the actuarially fair contribution rate is the transfer over the member’s own career; the difference between the uniform contribution rate and the actuarially fair uniform contribution rate is intergenerational transfer. Let us concentrate on a 20-year-old entrant. In the previous section, we concluded that the contribution payments of an entrant have 27% transferred to the older generations. From figure 2, it can be derived that this 27% can be split into 25% individual life-cycle redistribution and 2% pure intergenerational redistribution. The importance of life-cycle redistribution is therefore by far the largest in the uniform contribution rate. The intergenerational transfer amounts to 2% for all ages. In this regard, our analysis does not take account of transfers within generations, which also play a role in the uniform contribution and accrual system (see contribution by Aarssen and Kuipers in Chapter 8). Negative net benefit The actuarially fair uniform contribution rate is less than the uniform contribution rate. This means that a member on a contract with a uniform contribution rate over his or her career pays more contributions than he or she receives in pension benefits. In other words, the net benefit for this member, the balance of all benefits received and contributions paid, is negative. These extra contributions correspond to approximately 2% of all contributions that a member pays during his or her career. This amount is a pure redistribution from current generations to earlier generations. The negative net benefit of an entrant reflects back on the introduction of the uniform contribution rate, once in a distant past. At the moment the uniform contribution rate was introduced, a one-off subsidy was granted to the older generations. After all, they have fully benefited from the uniform contribution rate favourable for them, and did not need to make additional payments during their young years. The account for this subsidy was, probably unconsciously, left for the future generations.
11 Macroeconomic aspects of intergenerational solidarity
219
x 1000 € 9
7
5
3
1
-1
-3 100 90
80
70
60
50
40
30
age of current generations
20
10
0
10
20
30
40
50
60
70
80
90 100 110 120
future generations
Fig. 3. Change in net benefit with introduction of uniform contribution rate
Figure 3 shows the distribution of net benefits over the generations after introduction of the uniform contribution rate. On the horizontal axis to the left-hand side are the ages of the current generations, to the right-hand side the birth year of a future cohort, expressed in the number of years after introduction of the uniform contribution rate. On the vertical axis is the absolute difference in net benefit under a uniform contribution rate compared to a fair contribution rate. A 65-year-old member and all older members experience no effect from a transition to a contract with a uniform contribution rate, since these cohorts no longer have to pay pension contributions. The participants with an age between 27 and 65 benefit from a changeover to a uniform contribution rate. These cohorts have paid a low fair contribution in their young years, and in their later career are faced with a uniform contribution rate that is relatively low for them. The current generations that are younger than 27, and all future members, pay the price for the subsidy that was supplied to the current generations. Their net benefit is negative. The loss appears to flatten off over the course of time, but this flattening is attributable to the discounting. The sum of all benefits (in cash) is zero. That means that the introduction of the uniform contribution in itself was a zero-sum game. This conclusion is closely in line with the observation of Sinn (2000) that each pension system whether it is a funded system, a pay-as-you-go system, or a combination of both, is a zero-sum game for all participating generations.
220
J.P.M. Bonenkamp, M.E.A.J. van de Ven, and E.W.M.T. Westerhout
INTERGENERATIONAL RISK SHARING: ASSET SHOCK An important characteristic of our collective pension system is the risk sharing between different generations. The value of risk sharing is taken into consideration in the next section. Here we restrict ourselves to illustrating its scope. We take a negative assets shock as our example. Suppose that, as result of a price fall on the stock exchange, the pension assets decrease by 15%. Up to that moment, the fund is in equilibrium. This means the funding ratio is 100% of the real rights, no catch-up contribution has been charged, and the pension rights are fully indexed to the nominal wage growth. We assume a conditionally index-linked averagesalary scheme. That means that the pension fund has two instruments to work off funding ratio deficits: charging an increase on the break-even contribution (a catch-up contribution) and reducing indexation of pension rights and benefits according to an indexation table. As a basic principle for the catch-up period of the funding ratio, a standard called the Financial Assessment Framework (FTK) applies that stipulates that a fund must have a nominal funding ratio of 130% within a period of fifteen years after a shock. To identify the intergenerational transfers after the assets shock, we compare the average-salary scheme against a pure DC scheme. In this DC scheme, each generation bears the impact of the asset shock itself in its accrued assets and thus no intergenerational transfers occur at the time a x 1000 € 12 10 8 6 4 2 0 -2 -4 -6 -8 -10
average-salary scheme DC scheme difference
-12 -14 -16 100 90
80
70
60
50
40
30
20
10
0
10
20
30
40
age current generations
Fig. 4. Change in net benefit after an asset shock of 15%
50
60
70
80
90 100 110 120
future generations
11 Macroeconomic aspects of intergenerational solidarity
221
shock takes place. As in the previous section, we assume that the pension fund can only invest in one asset. This means that in the DC scheme there is no possibility to reduce the financial vulnerability of old people by adapting the investment mix for each cohort. Figure 4 presents the change in net benefit of pension funds after an asset shock. The solid line represents the average-salary scheme, the dotted line the DC scheme. The thin line is the difference between these two lines and shows the intergenerational transfers. In both pension systems, the total decrease of the net benefit is obviously equal to the asset loss of 15%. The distribution over the generations diverges. In the DC scheme, our benchmark, the blow is fully absorbed by the current generations. Members who are 65 years of age at the moment of the shock have the most pension capital and therefore experience the heaviest blow. The entering generation has not yet accrued any pension capital, and therefore is not affected by the shock. In the DC scheme, future generations therefore are not affected by the shock. On the basis of the distribution of the net benefit among the cohorts, it can be quantified that 37% of the capital loss is carried by the pensioners and 63% by the active members. In a conditionally index-linked average-salary scheme, the cost of the shock is partly shifted onto future generations in the form of catch-up contributions and indexation cuts. The older generations profit from this; the loss in net benefit for these generations is significant less than in the DC scheme. In the average-salary scheme, approximately 74% of the shock is carried by the active members (compared to 63% in DC), only 10% is borne by the pensioners (37% in DC), while 16% is shifted onto the future generations (0% in DC). The intergenerational transfers are shown by the thin line. A 65-year-old member profits most from risk sharing. This member only pays 27% of the actual loss on his or her pension capital. Here, the insurance aspect of a pension contract with risk sharing is therefore strongly reflected. This risk is shifted onto the younger and future generations. The worst off is the cohort of 27 year olds for whom the loss of net benefit is more than seven times higher than the capital loss in a DC scheme.
11.7 The value of intergenerational risk sharing The organizing of intergenerational risk sharing, however, is considered as the raison d’être of pension funds. Ultimately, the question is whether everyone is better off because of it. In other words: what welfare gain does it achieve? Calculations of the benefits of intergenerational risk sharing are
222
J.P.M. Bonenkamp, M.E.A.J. van de Ven, and E.W.M.T. Westerhout
very scarce in the literature, however. Moreover, the methods of approach are so different that they make mutual comparisons difficult. For this reason we confine ourselves here to a summary of this literature. In addition, we present our own indicative calculation of both the benefits and the costs of intergenerational risk sharing.
THE BENEFITS OF INTERGENERATIONAL RISK SHARING One of the first calculations of the benefits of intergenerational risk sharing is from the Dutch Scientific Council for Government Policy (WRR, 1999). This study ended up with a value of 25% of the pension capital. This is the extra amount that an individual saver must put aside to have the same probability of underfunding as he or she has in a collective pension fund.14 An indicative calculation in the CPB (Netherlands Bureau for Economic Policy Analysis) study into the shock resistance of the Dutch pension system (Westerhout et al., 2004) ends up with a value for the risk sharing between current and future generations of between 0% and 8% of the capital. The mutual risk sharing between current generations is disregarded in this. Teulings and De Vries (2006) approach the problem of intergenerational risk sharing from optimal investment theory. Intergenerational risk sharing is nothing more than the optimal insurance for risks on financial markets over generations. A generation that wants to take more risks, can include more risk-bearing capital in its portfolio; a generation that wants less risk, does the opposite. Market prices ultimately determine how much risk is handled in this manner. Pension funds can improve the risk sharing by acting on behalf of future generations. In a numerical example in Teulings and De Vries, the pension fund acts on behalf of the generations that enter the labour market in the coming 15 years. However, a condition is that these generations subject to mandatory membership of the pension fund at the moment they enter the labour market. After all, an investment in shares is uncertain and there is the possibility that the value of the portfolio will have decreased, as a result of which it is not attractive to join this pension fund. Teulings and De Vries calculate a welfare gain of pension funds of approximately 6% of the life-cycle income of a generation. As with the calculation in the CPB study, this calculation also takes no account of the risk sharing between current generations.
14
The contribution of Boender et al. goes deeper into this WRR calculation and indicates which points of this calculation can be improved.
11 Macroeconomic aspects of intergenerational solidarity
223
Gollier (2007) compares the situation of an individual DC scheme with a collective DC scheme in which all current generations transfer their pension savings to a pension fund that shares risks among the generations. Potentially, this could lead to a welfare gain of more than 30% of the capital. This can only be realised by assuming mandatory participation (as Teulings and De Vries also do) or by limiting the investment policy of the fund so that a minimum return can be promised to future generations. This introduces a DB element, because a guarantee of a minimum return automatically means that a minimum benefit is promised. Gollier takes account of a minimum return of 0%, i.e. that a benefit equal to the contributions paid in is guaranteed. Under this supplementary condition, a welfare gain of almost 10% of the capital is the result. Cui et al. (2006) make a comparison between an optimised individual DC scheme and a number of collective pension schemes that differ in the way in which shocks to the pension capital are absorbed. There is a collective DB scheme in which the benefits are defined and shocks are absorbed by means of variable contributions. In a collective DC scheme the contributions are constant and shocks are actually absorbed via the benefits. Lastly, there are two hybrids schemes, in which both contributions and benefits are used to absorb shocks. The welfare gain of collective pension schemes can run up to 4% in terms of the certainty equivalent consumption, which means the consumption level that produces the same benefit in a world without uncertainty as in an uncertain world. The CPB has a stochastic model with two overlapping generations (see Bonenkamp and Van de Ven, 2006). Individuals work during the first period and are pensioners in the second period. Their income in the second period depends on the return on their own free savings and a pension benefit. This pension benefit can be both a DB and a DC benefit. Participation in the pension fund is mandatory. When choosing their free savings, members take into account the uncertainties to which they are exposed in the second period. This concerns the return on their free savings. In addition, they are uncertain concerning the aggregated life expectancy. Furthermore, they are uncertain about the growth in productivity with which the DB pension is indexed. In a calculation with this model, a comparison is made between membership of an individual DC scheme and of a collective DB scheme. The result is an added value of DB schemes of approximately 10% of the capital.15
15
A sensitivity analysis concerning the risk aversion and the time preference leads to an added value of DB schemes between 8% and 12% of the capital.
224
J.P.M. Bonenkamp, M.E.A.J. van de Ven, and E.W.M.T. Westerhout
THE COSTS OF INTERGENERATIONAL RISK SHARING As discussed previously, the repayment of underfundings, uniform contribution rates and favourable fiscal treatment distorts the labour market. Welfare losses that correspond with the labour market distortions identified above are hard to quantify exactly. The order of magnitude can be assessed, however. We confine ourselves here to the case in which the labour market elasticity is independent of the age of the workers and in which the implicit government subsidies to supplementary pensions translate into higher tax rates on labour, so that there is no question of subsidising. We are then left with the situation in which shocks translate into catch-up contributions that put the labour supply under pressure. Shocks can go in two directions. For this reason we consider the case that there is a certain risk of a shock that requires a catch-up contribution and an equally high risk of a shock that implies an equally large discount on the break-even contribution. To be precise: suppose that there is a 50% risk of a shock that implies a catch-up contribution amounting to 30% of the total amount of wages and an equally high risk of an equally large catch-up contribution with the opposite sign. How large then is the certainty equivalent, i.e. the amount that households would be prepared to give up to avoid that they arrive at a situation with two possible outcomes for the contribution rates, and the possibility of welfare losses as a result of labour market distortions? An initial calculation produces a certainty equivalent of 0.8% of the life-cycle income. Although clearly positive, this nevertheless is an impact of minor scale. The value of this welfare loss also proves to be extremely susceptible to the size of the labour market elasticity used and the size of the contribution risk.16 One could object that, with the application by pension funds of indexation cuts, this argument has become less important; our figure of 0.8% of the life-cycle income is then an overestimate of the welfare loss cause by labour market distortions. However, this is only correct insofar as indexation cuts are unexpected. However, with the arrival of indexation tables, such cuts can actually be anticipated, at least in part. In our opinion, the overestimate in our result is thus of minor significance.
16
Sensitivity analysis of labour supply elasticity provides results for the welfare loss of 0.24%, 0.8% and 6.87% of the life-cycle income when assuming values for this elasticity of respectively 0.0625, 0.20 and 1. If the contribution shocks are set at 10% or 50% of the wage total, the results are then 0.08% or 2.8% instead of the 0.8% that results with shocks of 30% of the wage total.
11 Macroeconomic aspects of intergenerational solidarity
225
11.8 Concluding observations The Dutch supplementary pensions have been impregnated with intergenerational solidarity: because of the uniform contribution rate, the government subsidies to pensions and the way in which pension funds finance funding deficits. The resulting intergenerational financial transfers are far from small. The uniform contribution rate obstructs proper operation of the labour market, discourages labour force participation at a young age, but promotes a late retirement. The net effect is not entirely clear. The government subsidies to pensions, on the other hand, seem favourable for labour force participation in general. There is a lot of uncertainty concerning the benefits of intergenerational risk sharing. Nevertheless, current research indicates that these benefits can be significant. The costs of labour market distortions are equally laborious to quantify. Our indicative calculations, however, result in costs that are less than the benefits of intergenerational risk sharing. Although the sum of all these plus and minus items is not in the least clear, the benefits of solidarity seem to exceed the costs of it. The question of whether this solidarity could be better organised by the government than by a large number of pension funds still remains unanswered.
Literature Bonenkamp, J.P.M. and M.E.A.J. van de Ven, A small stochastic model of a pension fund with endogenous saving, CPB Memorandum 168, The Hague: CPB (Netherlands Bureau for Economic Policy Analysis), 2006. Brown, J.R., ‘Life annuities and uncertain lifetimes’, NBER Reporter: Research Summary, Spring NBER Website, www.nber.org/reporter/spring04/ brown. html, 2004. Cui, J., F. de Jong, and E. Ponds, Intergenerational risk sharing within funded pension schemes (unpublished), 2006. De Laat, E.A.A., M.E.A.J. van de Ven and M.F.M. Canoy, Solidariteit, keuzevrijheid en transparantie – de toekomst van de Nederlandse markt voor oudedagsvoorzieningen, The Hague: CPB (Netherlands Bureau for Economic Policy Analysis), 2000. The Economist, ‘Pension Pork’, The Economist, 15 April 2004.
226
J.P.M. Bonenkamp, M.E.A.J. van de Ven, and E.W.M.T. Westerhout
Gollier, C., Intergenerational risk-sharing and risk-taking of a pension fund, CESifo Working Paper, No. 1969, 2007. Sinn, H.W., ‘Pension Reform and Demographic Crisis. Why a Funded System is Needed and Why It is Not Needed’, International Tax and Public Finance, 7, 2000, pp. 389-410. Teulings, C. N. and C.G. de Vries, ‘Generational accounting, solidarity and pension losses’, De Economist, 154, 2006, pp. 63-83 Westerhout, E.W.M.T., M.E.A.J. van de Ven, C. van Ewijk and D.A.G. Draper, Naar een schokbestendig pensioenstelsel – verkenning van enkele beleidopties op pensioengebied, CPB Document 67, The Hague: CPB (Netherlands Bureau for Economic Policy Analysis), 2004. WRR (Scientific Council for Government Policy), Generatiebewust beleid, The Hague: Sdu Publishers, 1999.
12 Summary and conclusions
S.G. van der Lecq and O.W. Steenbeek In this volume a large number of renowned researchers and policy-makers discussed a broad range of topics. Together they offer as complete a picture as possible of the costs and benefits of collective pension systems. The concept of solidarity was extensively examined and the extent to which the existing types of solidarity play a role within collective pension schemes was analysed. Now it is clear where the inherent costs of these systems lie and which elements are responsible for the benefits. Where possible, these aspects are quantitatively substantiated. A debate that is based on facts is preferable to ignorance. In this book, we aimed to present sufficient clarity and factual material, so that the discussion concerning the costs and benefits of collective pension schemes can be continued at a higher level.
12.1 Summary This book contains contributions by researchers from the policy practice and academia. In collective pension schemes, risk sharing and value transfers are the two manifestations of solidarity. The authors discuss several concepts of solidarity, the costs and benefits of solidarity and the macroeconomic consequences of solidarity within collective pension schemes. The results of choices made for groups of participants on this subject are analysed. The book is divided into four parts. Part 1 (chapters 2 and 3) provides a general summary of the various concepts of solidarity. This concerns both risk sharing and value transfers between different groups of participants. Chapter 2 discusses the forms of solidarity within collective pension schemes and chapter 3 draws the parallel with the healthcare sector. Both in healthcare and in pensions, the advantages of collective schemes become unequally shared among groups of participants. Part 2 (chapters 4 to 8) pro-
228
S.G. van der Lecq and O.W. Steenbeek
vides quantitative information on the degree of risk sharing and value transfers within collective pension schemes. Part 3 (chapters 9 and 10) goes deeper into the mandatory participation in a collective scheme. Part 4 (chapters 11 and 12) presents the conclusions, in which chapter 11 mainly focuses on macroeconomic consequences of the structure of collective pension funds. In this chapter, we will get back to the theme of solidarity within collective pension schemes.
PART 1. THE CONCEPT OF SOLIDARITY In chapter 2, Jan Kuné provides an extensive summary of solidarity within collective pension schemes. The most important distinction is that between, ex post solidarity (also: risk solidarity or risk sharing) and, on the other, ex ante solidarity (also: subsidising solidarity or value transfers). Ex post solidarity refers to risk sharing as in a fire insurance: you don’t know in advance whether or not your house may burn down, but should it happen, the collective will bear the costs. Only after a certain event has taken place do we see who is paying and who is the receiving party. Within a collective pension, this mainly concerns shared investment risks. With ex ante solidarity, it is already clear in advance that the one participant benefits and the other one pays. Thus, by application of the uniform contribution rate or with final-salary schemes, value is transferred e.g. from participants with a flat career pattern to the career makers. Participants in collective pension funds may become less willing to accept these types of ex ante value transfers in the future. In chapter 3, Patrick Jeurissen and Floris Sanders discuss the Dutch healthcare system, in which comparable aspects of solidarity play a role. The healthcare sector and the pension sector each have their entirely individual institutions with little mutual overlap, but in both systems ‘risk’ and ‘uncertainty’ play an important role. Collective pensions reduce people’s uncertainty about having sufficient income if they are retired, while the healthcare system reduces their uncertainty concerning the treatment or care that they need if they are sick. Older people run a larger risk of getting ill, and both healthcare and pensions are therefore important provisions for old people. Solidarity is under pressure in pensions as well as healthcare, because the costs increase and the value transfers become increasingly skewed. Especially young people pay for this. In healthcare this leads to discussions, for example, on the question of whether people who live an unhealthy lifestyle should pay higher contributions. New reform proposals
12 Summary and conclusions
229
therefore concentrate on a combination of individual responsibility and system incentives. The first part of this book offers a picture of the existing concepts of solidarity within the collective pension schemes. These partly coincide with those within the healthcare sector. The importance of various types of solidarity differs between the sectors.
PART 2. QUANTIFYING SOLIDARITY In chapter 4, Jacob Bikker and Jan de Dreu examine cost differences between industry-wide pension funds, company pension funds, professional group pension funds and schemes run by insurers. A further distinction is made between large and small pension funds. Large cost differences are found, for which economies of scale prove to be a dominant explanatory factor. It is especially due to these economies of scale that the implementation costs of mandatory industry-wide pension funds are significantly lower than those of company pension funds. Lower implementation costs can be translated directly into a substantial higher or cheaper pension product. One conclusion is that the consolidation of small pension funds can substantially increase efficiency. When the comparison is made between collective and individual schemes, the cost differences become even more pronounced, especially if insurance companies run these individual schemes. On average, approximately 25% of the contributions paid to insurance companies go to operating costs and profit margin, whereas with collective pension schemes this average is 3.5%. In the next chapter, Guus Boender, Lans Bovenberg, Sacha van Hoogdalem and Theo Nijman discuss the question of what an ‘optimum pension contract’ theoretically implies from the participant’s perspective. Such a pension contract determines how much contribution must be paid and how this must be invested to achieve a specific pension benefit. Then they analyse the extent to which pension funds are able to approach this optimum and what obstacles individuals encounter if they would want to implement the optimal investment policy themselves. After the WRR (Dutch Scientific Council for Government Policy) study of 2000, academic literature has been published that is based on other assumptions. However, definite conclusions concerning the added value of pension solidarity cannot yet be drawn on the basis of this. The authors do offer a complete summary of the aspects that a collective pension fund is better capable of organising than the individual For example, a pension
230
S.G. van der Lecq and O.W. Steenbeek
fund organises risk sharing between generations, as a result of which it is able to take more investment risk. In addition, the fund arranges de facto contracts between young and old, in which younger generations borrow from older generations and invest these loans in equity. Roy Hoevenaars and Eduard Ponds, in chapter 6, present a method of assessing policy changes at collective pension funds. Using this method, they demonstrate that even limited policy changes can have substantially differing results for various generations within a fund. If, for example, a fund decides to invest a larger proportion of its assets in equity, this is beneficial for the younger participants, because they then have to pay a lower contribution. For older participants, however, it is unfavourable, because shares are riskier than bonds and their own pension benefit would become less certain. The results of a transfer to another type of pension scheme can also be examined using this method. The authors use examples to show how large the transfers are from one generation another. They recommend applying this method with all possible amendments to the financial policy of a fund, and consider the consequences for several age brackets in the final decision-making. In chapter 7, Dick Boeijen, Corné Jansen, Niels Kortleve and Jan Tamerus go deeper into the intergenerational debate and investigate the uniform contribution and accrual system as part of this. This system is legally vested and provides that all participants pay the same percentage of their gross income as contribution pays and in exchange for that receive the same percentage of their income as pension accrual. Because a pension accrual of a euro over 40 years that is paid out today already costs less than an accrual of a euro that will be paid out in 5 years time, de facto, younger participants pay too much for that euro and the older participants pay too little. For this reason, there is an element of pay-as-you-go included within the funded pension schemes. This appears to involve substantial amounts. When participants remain for their entire career with a single or a similar fund, this not an issue: the surplus that you pay at a younger age is recovered later. On the other hand, however, it becomes an issue for a participant who decides to resign halfway through his or her career and become self-employed, for example. In fact, substantial value is transferred from this employee to other employees who join the fund at a later age, such as formerly self-employed people, and people re-entering the labour market. Furthermore, this system slows down the labour mobility of young people. Scrapping the uniform contribution and accrual system would
12 Summary and conclusions
231
lead to younger employees becoming cheaper and older employees relatively expensive, as a result of which old people will encounter more difficulty participating on the labour market. The authors indicate that adjustment of the uniform contribution and accrual system is possible, but also that this is not simple. Not only would laws have to be amended, but a transfer to a different type of pension accrual also has consequences for the labour market. The latter point is examined more closely in chapter 11. In chapter 8, Karin Aarssen and Barthold Kuipers analyse the redistribution within collective pension funds on the basis of representative participants. Women live on average three years longer than men and, as a result of this, benefit more from the old-age pension, but correspondingly less from the partner’s pension. Single people benefit less from the scheme than partnered participants. The inequality between employees with a steep career and their colleagues with a flat career is reduced by the transition from final pay to average wage schemes. Life insurers, however, when setting their premiums, can make distinctions between characteristics of the participants. The costs of private insurance products are so high, however, that everyone proves to be provided for cheaper in a collective pension scheme. The contribution for the same pension with private pension accrual would be approximately 40% higher on average. Part 2 offered a specific picture of the level of the solidarity transfers that take place within collective pension schemes. Transfers were measured in euro as possible. The different forms of ex ante value transfers prove to be extremely varied in size. Through a careful analysis, the correct focus emerges in the debate concerning the future of collective pension schemes.
PART 3. MANDATORY PARTICIPATION In chapter 9, Peter van Els, Maarten van Rooij and Margreet Schuit study the mandatory participation of employees in the pension scheme of the employer. Experiences from abroad show that, without this mandatory participation, employees would save less for their pension. Moreover, they would not invest and manage their pension money optimally. Therefore, a certain measure of paternalism seems to be in its place. The results from a recent survey among Dutch households mainly confirm this picture. For example, it appears that the average Dutch person’s knowledge of basic financial concepts is limited and the investment skills are moderate. Moreover, their knowledge of their own retirement benefits is sadly lacking. In addition, Dutch people generally prefer not to take
232
S.G. van der Lecq and O.W. Steenbeek
risks with respect to their pension accrual, and a majority support a system where participation is mandatory. Citizens have a preference for a system with little autonomy and as much security as possible. Pieter Omtzigt analyses the mandatory participation for companies in chapter 10. This refers to the legal possibility for individual enterprises in specific industries to have mandatory participation in the pension scheme of the sector imposed, if this is available. The most important conclusion is that this regulation has contributed substantially to encouraging saving behaviour, the prevention of ‘blank spots’ and limiting competition between employers on labour costs. Omtzigt therefore argues for retaining the mandatory participation for companies, where applicable. In part 3, the spotlight was on mandatory participation: a striking and crucial element in the design of the Dutch pension system. The backgrounds are still legitimate and the benefits appear to outweigh the costs.
PART 4. CONCLUSION In chapter 11, Jan Bonenkamp, Martijn van de Ven and Ed Westerhout examine the macroeconomic aspects of solidarity, with a lot of attention for the generation debate again. From this perspective they look back to the various contributions in the previous chapters. The authors consider the organisation of risk sharing as the foremost benefit of collective pension schemes. This allows a pension fund to take more investment risk and this leads again to lower costs and more stable pension benefits. The costs are mainly formed by potential disruptions on the labour market. What is more, the uniform contribution and accrual system discourages labour participation at a young age, but it promotes a late retirement. Although it is not in the least clear what addition of all these plus and minus items produces, the benefits of solidarity seem to exceed the costs of it.
12.2 Answer to the central question In this chapter, we return to the central question of this book: what are the most important benefits and what are the most significant costs that come with collective pension schemes? Who shares risks with whom and who transfers how many euro to whom? The most important benefits of the collective come first of all from the way risk sharing is organised between generations. Because unexpected
12 Summary and conclusions
233
gains and losses on financial markets are distributed among all generations, the collective is prepared for and able to accept more risks. The uniformity of the collective pension product, moreover, enables funds to provide their services at low costs to their participants. The most important costs of collective pension schemes are first of all found in the restriction of freedom to match the pension savings to the personal situation. The financial policy of a collective scheme is tuned to the average participant and does not need to be optimal for everyone. Besides this, a collective scheme contains forms of subsidising solidarity, as a result of which not all participants profit to the same extent from the benefits mentioned above. Also, the current system has a disruptive effect on labour participation and mobility on the labour market.
12.3 Dilemmas The logical follow-up question is: how could the costs of collective pensions be reduced, without putting the benefits in danger? To this end we discuss a number of dilemmas that sooner or later confront collective pension funds.
1. FREEDOM VERSUS SOLIDARITY By scrapping the mandatory participation, pension savings can be better matched to the personal situation. Participants themselves probably have a better picture of their most likely career paths than a pension fund, the value of their own human capital and their preferred time to retire. This pleads for less paternalism and more freedom to choose. However, a large majority of Dutch consumers say that they are not eager for more freedom. Instead, they consider more options as another problem that must be resolved, rather than an opportunity for tailoring their financial affairs to their own wishes. Furthermore, foreign surveys reveal the mistakes people make when they have more freedom: they not only save too little for a proper pension, but also invest the savings in the wrong way. Also, in an individual system they cannot profit from the substantial benefits that come with sharing risks among generations. This risk sharing enables participants to take more investment risk; probably even more than the youngest participant would take in his theoretical optimum portfolio. Finally, it appears that the investment costs are many times higher for individuals investing for themselves than for people investing for their
234
S.G. van der Lecq and O.W. Steenbeek
retirement via a collective scheme. Although the investment mix may not be optimal, in view of the higher costs that accompany investment freedom, a collective investment fund seems nevertheless the best alternative on balance from this perspective, too.
2. CUSTOMISED VERSUS ECONOMIES OF SCALE An alternative for scrapping mandatory participation is to provide more customisation within collective schemes themselves. Many pension funds already meet some of these wishes, for example through flexibilisation of the retirement age and the introduction of additional products. However, the better accommodation of the personal situations must always be critically considered against the higher implementation costs, to avoid the risk of throwing out the baby with the bath water. This concerns both the administrative processing of individual arrangements and the adjustment of the investment policy for them. The economies of scale that can be realised due to the uniformity of the product within a collective are simply huge.
3. ABOLISH SUBSIDISING SOLIDARITY? In this book we have identified several forms of subsidising solidarity. In many cases, the cash flows concerned have diminished during recent years. For instance, there less value is transferred from participants with flat career paths as a result of the transition from final-salary to averagesalary systems. There is also less subsidising from single persons to partnered participants now that many schemes allow the partner’s pensions to be converted to higher pension benefits for the participants themselves. What is legally vested is that, following a Court of Justice ruling, women and men must be treated equally by funds, in spite of the fact that women have a longer life expectancy and are therefore more expensive for a pension fund. The subsidy from men to women, however, is relatively limited. Furthermore, there is the solidarity with disabled people who accrue pension without paying contributions.
4. SOLIDARITY WITHIN OR BETWEEN GENERATIONS? An important aspect of collective schemes is the uniform contribution and accrual system, in which both the contribution and the pension accrual are tuned to the average participant. This system consists of the uniform contribution and the uniform pension accrual. This means in practice that
12 Summary and conclusions
235
there is a pay-as-you-go element in the collective scheme, implying value transfers from younger to older participants within the fund. As long as all participants stay within the same collective scheme for their entire career, there is no reason to change this uniform contribution and accrual system. However, when people switch from one pension scheme to another, it can happen that they accrue more or actually less pension. This slows down the labour market mobility. Revision of this system requires amendment of the law and costs money. Alternative systems have different effects on the labour market. If, for example, an age-related contribution were to be selected, the contribution for a 64-year-old person would become four times as high as for a 25year-old. In that case there would be less solidarity between generations. If, on the other hand, an equal contribution is selected, but with agerelated pension accrual, an extra year’s work at the end of the career provides little in extra pension accrual, and old people could decide to retire early.
5. TRANSPARENCY VERSUS SOLIDARITY It is difficult to be an opponent of more transparency. Moreover, since new legislation in the Netherlands and many other advanced economies was introduced, there is no way back: transparency is a central theme in the regulatory supervision. Besides the Dutch central bank (DNB), the Netherlands Authority for the Financial Markets (AFM) has been appointed as the second supervisory body, specifically charged with assessing the communication with participants. Nevertheless, greater transparency can have consequences for the life expectancy of the current system, because a lot of solidarity exists thanks to ignorance. The more clarity there is about the degree of solidarity, the more specific the debate about the desirability or need for it can be. This book makes a contribution to this debate. The information provided can assist the reader to form an opinion about the costs and benefits of solidarity in collective pension schemes.
About the authors
K. (Karin) Aarssen (1968) works at the Financial & Risk Policy Department of ABP Pension Fund, focusing on pension policy and risk management. She studied Operations Research at the Erasmus University and Actuarial Science at the University of Amsterdam. After graduating she worked at KPMG Brans&Co (now: Watson Wyatt) and ING Group. J.A. (Jacob) Bikker (1952) is senior scientific researcher at the Supervisory Policy Division of DNB (Dutch Central Bank). He obtained a PhD in Econometrics and performs research in the fields of banks, insurance, pensions, financial market competition, financial institution efficiency, regulatory procyclicality and risk management. T.A.H. (Dick) Boeijen (1977) studied Technical Mathematics and Actuarial Science and works as an actuarial officer at PGGM. C.G.E. (Guus) Boender (1955) studied Business Econometrics at the Erasmus University Rotterdam and obtained a PhD at the same university. He is currently Professor in Asset Liability Management at the Free University Amsterdam and Director at ORTEC. At ORTEC he shares responsibility for ALM, strategic asset allocation, risk management and performance attribution which ORTEC performs at pension funds, insurers and housing corporations. J.P.M. (Jan) Bonenkamp (1979) works as a research officer at the Labour Market and Welfare State Sector of the CPB Netherlands Bureau for Economic Policy Analysis. In this capacity he focuses mainly on pension and ageing issues. Jan studied General Economics at the University of Groningen. A.L. (Lans) Bovenberg (1958) studied Econometrics at the Erasmus University Rotterdam and obtained his PhD at the University of California, Berkeley. Bovenberg is currently Professor in Economics at the University of Tilburg. In 2003 he won the Spinoza Prize, the most important science award in the Netherlands. He used the prize money to set up Netspar, a private-public knowledge network on pensions, ageing and retirement.
238
About the authors
Bovenberg started his career at the International Monetary Fund, worked as policy officer at the Ministry of Economic Affairs and as Deputy Director at the CPB Netherlands Bureau for Economic Policy Analysis. He also held the position of Scientific Director of the Center for Economic Research (CentER) at the University of Tilburg. Bovenberg mainly publishes on public finance and taxation, but also carries out research in other fields, including environmental, macro, institutional and financial economics. J. (Jan) de Dreu (1981) wrote this chapter while working at the Supervisory Policy Division of DNB (Dutch Central Bank). His research interests include pension fund investment policy, lending technologies of microcredit institutions and the interaction between deposit insurance and market discipline. Since September 2006 Jan de Dreu has been working at ABN AMRO. P.J.A. (Peter) van Els (1959) is departmental head of Economic Research at the Economics & Research Division of DNB (Dutch Central Bank). He studied General Econometrics and Mathematical Economics at the University of Tilburg and has worked since 1986 in various positions within the research environment of DNB. In 1995 he obtained his PhD at the University of Amsterdam. His research specialisation covers macroeconomic modelling, monetary transmission and household financial behaviour. R.P.M.M. (Roy) Hoevenaars (1978) is senior portfolio manager at the global tactical asset allocation fund of ABP Investments. Previously, he was head ALM modelling and senior researcher ALM at the Financial and Risk Policy Department of ABP Pension Fund. He was researcher at the Research Department of ABP Investments in the field of strategic asset allocation, ALM and quantitative equity models. Roy is also affiliated to the University of Maastricht where he is finalizing a PhD dissertation on strategic asset allocation and ALM. He studied Econometrics and Operations Research at the University of Maastricht. S. (Sacha) van Hoogdalem (1970) studied Business Econometrics at the Erasmus University Rotterdam and is currently Director at ORTEC. She focuses mainly on ground-breaking ALM studies in the pension sector. She also carries ultimate responsibility for ensuring that the ALM models are adequate for obtaining more insight into ALM issues in theory and practice. C. (Corné) Jansen (1978) studied Econometrics and works as actuarial officer at PGGM.
About the authors
239
P.P.T. (Patrick) Jeurissen (1969) is senior health care analyst at the Council for Public Health and Care. His work centres on financial-strategic issues in the healthcare sector. He is particularly interested in financing and distribution issues as well as in the development of the difference between for-profit and non-profit institutions. He is also preparing a thesis on this subject. He studied Public Administration at the Erasmus University Rotterdam, specialising in public finance. C.E. (Niels) Kortleve (1965) studied Econometrics and is currently Actuarial Projects & Special Accounts Manager at PGGM. B.J. (Barthold) Kuipers (1972) works at the Financial & Risk Policy Department of ABP Pension Fund. He focuses on the policy, strategy and risk management of the pension fund. Previously he worked at CPB Netherlands Bureau for Economic Policy Analysis. He graduated in Econometrics at the University of Amsterdam. J.B. (Jan) Kuné (1946) is senior researcher at Stichting Pensioenfonds ABP and part-time Professor in Pension Actuarial Science at the Economics Faculty of the University of Amsterdam. Kuné started his career at the Europa Institute of the University of Amsterdam in 1968. In 1970 he joined DNB (Dutch Central Bank). In 1975 he accepted a post at the Royal Tropical Institute and worked as an epidemiologist in Indonesia and Kenya. In 1978 he returned to the University of Amsterdam where he carried out his PhD research and also lectured in Pension Actuarial Science. Since 1988 he has worked for ABP where (the introduction of) an adequate funding methodology was long his main task. Since 2003 he has again been attached to the University of Amsterdam as part-time Professor in Pension Actuarial Science. S.G. (Fieke) van der Lecq (1966) took degrees in Economics and Business Economics at the University of Groningen where she obtained her PhD on a thesis in monetary economics. After appointments at the Ministry of Finance and the CPB Netherlands Bureau for Economic Policy Analysis, she was editor-in-chief and publisher of ESB, a journal for economists, as well as director of the ESB publishing house. Alongside her other current affiliation, she works 1 day a week as associate faculty at the Economics group of the Erasmus University Rotterdam. Th.E. (Theo) Nijman (1957) is Professor in Econometrics of the Financial Markets as well as Professor in Investment Theory on the F. Van Lanschot Chair. Both Chairs are at the University of Tilburg. Theo is chairman of the
240
About the authors
Scientific Board of Netspar (www.netspar.nl). He is also Scientific Director of the Tilburg Center of Finance and Academic Coordinator of Inquire Europe, the European meeting forum for science and institutional asset managers. Theo Nijman obtained his PhD in 1985 at VU University Amsterdam. Since then he has been in the employ of the University of Tilburg, including five years as researcher of the Royal Netherlands Academy of Arts and Sciences. Theo has been involved in more than ten theses as PhD supervisor. During the period 2000-2004 Theo was Scientific Director of CentER, the research institute of the Tilburg Economics Faculty. P.H. (Pieter) Omtzigt (1974) is member of the House Of Representatives Of The Dutch Parliament for the Christian Democratic Party and is attached to the Faculty of Economic Science and Econometrics of the University of Amsterdam. In the House Of Representatives he is spokesman for, among other things, public and private pensions and the new healthcare system. E.H.M. (Eduard) Ponds (1958) has worked for ABP since 1995 at, successively, the Actuarial Department, Asset Research, and currently the Financial and Risk Policy Department. In addition, he is affiliated with Netspar. At ABP he is engaged in projects on pension plan design, ALM and risk management. Before joining ABP, Eduard worked at the University of Tilburg and the Open University. He graduated in Economics. In 1995 he obtained his PhD entitled: “Supplementary pensions, intergenerational risk-sharing and welfare”. Since September 2007 he is part-time Professor in Economics of Collective Pension Plans at the University of Tilburg. Email:
[email protected] M.C.J. (Maarten) van Rooij (1970) works as researcher at the Economics & Research Division of DNB (Dutch Central Bank) and is affiliated to Netspar. He graduated in Econometrics at the University of Tilburg, specialising in Monetary Economics, General Econometrics and Operational Research. Currently, he is writing a PhD-thesis on the financial behaviour of households. F.B.M. (Floris) Sanders (1958) is radiologist at the Diakonessenhuis Utrecht – Zeist – Doorn Hospital. He has held many administrative posts in the healthcare sector. From 1997 tot 2000 he was chairman of the Order of Medical Specialists, and from 2002 to 2005 he was chairman of the Council for Public Health and Care, an important advisory body of the Minister of Health, Welfare and Sport. His particular interests are issues relating to the (financial) viability of collective insurance schemes in the healthcare sector.
About the authors
241
He is currently a member of the Supervisory Boards of the Erasmus MC and of Altrecht, a large mental healthcare institution. He fulfils positions in the Executive Committee of the NIVEL research institute and in the programme committee on demand-based delivery of ZonMw. In addition, he fulfils various advisory positions, including membership of the Advisory Council of Health Insurers in the Netherlands. M.E.J. (Margreet) Schuit (1967) works as policy officer at the Supervisory Strategy Department of DNB (Dutch Central Bank). She studied Economics at the University of Amsterdam, specialising in Monetary Economics, General Economics and Econometrics. After graduating in Economics she focused on various subjects including compensation & benefits policy (in the employ of the FNV trade union) and in recent years mainly on pension policy. O.W. (Onno) Steenbeek (1967) is head of Corporate ALM and Risk Policy at ABP Pension Fund since 2007. Onno studied Financial Economics and Modern Japanese Studies at the Erasmus University in Rotterdam, where he also obtained his PhD in finance. In 2001 he joined ABP Investments as senior strategist, until he switched over to the Financial and Risk Policy Department of ABP Pension Fund in 2005. He is affiliated with the Department of Finance of the Economics Faculty of Erasmus University for one day a week. J.H. (Jan) Tamerus AAG (1953) studied Actuarial Science, is internal actuary at PGGM and Director of the Corporate Actuarial Services & ALM Department. M.E.A.J. (Martijn) van de Ven (1966) works as a research officer at the Labour Market and Welfare State sector of the CPB Netherlands Bureau for Economic Policy Analysis. His work centres on pension and ageing issues. Martijn studied Econometrics at the University of Tilburg, where he obtained his PhD in 1996 with a thesis about political decision-making on intergenerational redistribution. E.W.M.T. (Ed) Westerhout (1965) is programme leader in Ageing and Pensions at the Labour Market and Welfare State sector of the CPB Netherlands Bureau for Economic Policy Analysis. Ed studied Economics at the University of Tilburg and obtained his PhD in 1997 at the University of Amsterdam on a thesis about imperfect substitution on capital markets, capital gains tax and the EMU. In earlier positions at the CPB Netherlands Bureau for Economic Policy Analysis he focused on e.g. healthcare, disability and the labour market. In addition, he lectured for many years in General Economics at the University of Amsterdam.
Subject index
A access to healthcare 46 adverse selection 37, 39, 67, 69, 71, 82, 85, 153, 210 annuity 63, 64, 67, 69, 70, 74, 76, 83 – 85, 153, 163, 168, 181, 191, 210 Asset Liability Management 96, 100, 102, 104, 106 – 108, 112, 114, 115, 239, 240, 242, 243 average-wage system 26, 31, 121, 215, 220 – 222, 233, 236
B behavioural finance 146, 160, 168, 182
Collective Defined Contribution (CDC) system 58, 107, 113 – 116, 211, 223, 224 company pension fund 51, 54 – 58, 60, 73, 161, 188, 189, 192, 199, 200, 212, 231 contribution rate 25, 27, 30, 95, 96, 104, 108, 111 – 113, 115, 119, 121, 129, 137, 145, 146, 153, 209, 212, 214, 217 – 220, 225 actuarially fair 138, 143 – 146, 150, 151, 155, 214, 216 – 218 contribution volatility 75, 88 – 90 cost differences 52 – 54, 57, 61, 62, 71, 231 costs administration 53 – 61, 64 – 71 operation 51 – 53, 55, 57 – 61, 71, 152, 231
blank spots 162, 187, 188, 234
C
coverage rate 183
catch-up contributions 22, 84 – 87, 90, 120, 121, 212, 215, 220, 222, 225
D
catch-up indexation 21, 97, 142 cold solidarity 4, 147 collective contracts 52, 63, 65 – 67, 70, 89, 91, 161, 163 – 165, 182, 183
decreasing accrual system 119, 129 – 136 Defined Benefit (DB) system 2, 52, 56 – 59, 66, 84, 90, 107, 111 – 113, 164 – 167, 175, 207, 210, 223, 224
244
Subject index
Defined Contribution (DC) system 2, 26, 34, 52, 56 – 59, 77, 90, 113, 128, 164, 166, 167, 169, 174, 176, 207, 221 – 224 deflators 95, 99 direct schemes 62, 66 disability 7, 31, 91, 138, 141, 143, 144, 146, 147, 192, 194, 200, 244 discontinuity risk 213 discount rate 21, 30, 32, 87, 90, 98 – 100, 102, 108, 214, 215 stochastic (deflators) 99, 100, 102 distribution mechanism 16
E economies of scale 2, 36, 45, 46, 52, 54, 55, 58, 60, 61, 63, 70, 71, 138, 152, 155, 193, 199, 200, 211, 231, 236 employer matching contributions 169 endowment insurance policies 63, 64, 69, 70, 74 equivalence principle 37
financial illiteracy 170, 177 financial planning 160, 168, 170 first pillar pension 1, 14, 16, 19, 30, 37, 38, 82, 139, 162, 164, 166, 180, 193, 206 flexible retirement age 90 framing 169 frayed-edge theory 135 funded pension system 16, 33, 45, 122, 159, 164 – 166, 199, 220, 232
G generational accounting 20, 96 – 100, 102 – 107, 109, 110, 112, 115
H heterogeneity 91, 211 horizontal solidarity 13 Household Survey 171, 173, 176, 178 human capital 76, 79, 81 – 83, 85, 86, 89, 90, 235
ex-ante solidarity 17, 206 – 209, 214, 230
I
explicit pension contract 102
income solidarity 37, 39, 43, 206
ex-post solidarity 206 – 208, 214, 230
incomplete capital markets 80, 84, 210 indexation 13, 21 – 23, 30, 66, 67, 78, 83, 86, 88, 91, 95 – 106, 108 – 116, 121, 139 – 142, 154, 177, 192, 193, 207, 208, 212, 215, 220, 222, 225
F fair value 78, 90, 215 final-wage system 32, 66, 107, 112, 139, 140, 152, 230 Financial Assessment Framework (FTK) 22, 68, 196, 221
conditional 97, 115, 121 unconditional 107, 112 – 114
Subject index
indexation policy 21, 22, 95, 104 individual arrangements 13, 32, 236 individual pension scheme 52, 63, 71, 159, 163, 165 individualisation 42, 46 industry-wide pension fund 30, 51, 54, 56 – 62, 71, 73, 95 – 97, 104, 115, 116, 119, 121, 139, 161, 162, 188 – 190, 192, 193, 195, 199, 200, 211, 212, 231 inertia 168 informal care 36 insurance company 45, 60, 62, 63, 66, 74, 80, 84, 138, 153, 155, 166, 190, 199, 231 intergenerational risk sharing 96, 154, 206, 208, 210, 211, 213, 214, 222 – 224, 226 intergenerational risk trading 85 intergenerational solidarity 19, 20, 26, 38, 39, 41, 43 – 45, 102, 120, 121, 126, 192, 205 – 211, 213, 214, 225 investment policy 7, 78 – 81, 107, 109, 116, 167, 223, 231, 236, 240
J
245
life expectancy 18, 31, 45, 67, 137, 143, 146, 151, 153, 154, 191, 199, 210, 224, 236, 237 life insurance policy 63 life-course savings scheme 160, 179 – 181 longevity bond 83 – 85 longevity risk 66, 68, 83, 85, 89, 91, 199
M mandatory participation 8, 28, 30, 66, 67, 70, 85, 152, 155, 159 – 167, 171, 172, 180 – 183, 187 – 195, 197, 199, 200, 209, 211, 213, 223, 230, 233 – 236 companies 160, 161, 187, 189, 197, 199, 200, 234 indirect 163 individuals 160, 161, 189, 190, 194 mental accounting 168 misselling scandal 166, 191 moral hazard 18, 32, 37, 82 mutual subsidies 124, 126 – 128, 135 myopia 168, 169, 182
John Rawls 35
N
L
net benefit 219 – 222 no-claim rebate 34, 38, 40, 43, 46
labour market distortions 59, 90, 205, 206, 211, 212, 224 – 226 labour mobility 85, 128, 200, 232, 235, 237 libertarian paternalism 167
O old age pension 140 outsourcing 51, 52, 59, 60, 63
246
Subject index
P pay-as-you-go 16, 38, 40, 44, 46, 95, 122, 135, 159, 164, 193, 198, 199, 220, 232, 237 pension basis 139, 141 – 146, 148 – 151, 153, 155 defined 145 pension benefit guarantee corporation (PBGC) 198 pension contract 13, 22, 25, 75 – 78, 81, 82, 84 – 86, 90 – 92, 96, 97, 99, 101, 103, 106, 207, 208, 222, 231 pension-cum-health policies 45 perverse solidarity 20, 30, 31 policy ladder 91, 101, 104, 111, 116 procrastination 168, 172, 175, 182 professional group pension fund 1, 56 – 58, 61, 63, 71, 162, 231 profit margin 65, 68, 70, 71, 231 progressive contribution system 129 – 131
Q quasi-mandatory participation 161, 163, 165, 182, 183
R reciprocity 13, 14, 20, 41, 42, 44, 123 recovery period 19, 22, 68 redistribution 16 – 18, 30 – 32, 44, 46, 95, 107, 134, 137, 138, 145, 147, 149 – 151, 153 – 155, 188,
205, 206, 209, 214, 217 – 219, 233, 243 reinsurance 51, 59 – 61, 66, 70 representative participants 8, 233 risk aversion 35, 79, 91, 99, 102, 174, 224 risk sharing 2 – 6, 15, 18, 28, 29, 39, 75, 77, 78, 80, 82, 84, 85, 91, 96, 130, 154, 192, 200, 205, 210, 220, 222, 223, 226, 229, 230, 232, 234, 235, 242 risk solidarity (risk sharing) 5, 6, 18, 76, 78, 91, 119, 147, 230
S savings schemes 44 – 46, 166, 168, 169, 179 – 181 selection effects 191 social acceptance 14 social cohesion 13 – 15, 18, 34 social contract 30, 35, 36 solidarity transfers 33, 37, 233 subsidising solidarity 5, 18, 28, 37 – 41, 43, 45, 119, 120, 129, 130, 135, 230, 235, 236
T tailored pension solutions 183, 200, 235 transaction costs 83, 86
U uniform contribution and accrual system 25, 119 – 124, 126 – 135, 208, 209, 219, 232, 234, 236
Subject index
uniform contribution rate 7, 25, 95, 105, 119 – 124, 126 – 135, 137, 138, 142, 143, 145, 155, 192, 196, 207 – 209, 212 – 220, 224 – 226, 230, 232, 234, 236 utility 34, 77, 79, 82, 89, 106
V value transfers 2, 3, 5, 13 – 16, 18, 21, 27, 32, 37, 71, 95 – 97, 100, 101, 107, 116, 120, 229, 230, 233, 237 value-based ALM 95, 97, 100 – 102
247
value-based generational accounting 95 – 97, 101, 102, 104, 116 veil of ignorance 35, 36, 43, 46 vertical solidarity 14
W warm solidarity 4, 6, 147
Z zero-sum game 95, 97, 103, 107, 116, 220