Costs and Benefits of Collective Pension Systems

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
steenbeek@few.eur.nl vanderlecq@few.eur.nl
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 & Vckler 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 in-
terest 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 practi-
tioners, 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 be-
come 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 con-
tinued, 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 enthusias-
tic publisher Dr. Niels Peter Thomas. We are also grateful to the Pension
Science Trust (Stichting Pensioenwetenschap) for subsidizing the transla-
tion 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 pen-
sion sector has much to offer in terms of expertise and institutional experi-
ence, as the various distinguished authors in this volume show.
The debate on collective pension funds remains fierce, both in the Nether-
lands 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 dilemmas 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 com-
pared to that of the second pillar collective pension systems.
In part 2, several aspects of collective pension arrangements are quanti-
tatively assessed. It turns out that the costs of execution of pension agree-
ments 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 sub-
stantial ex ante value transfers between groups of participants. The value
transfers between generations are largest.
Part 3 discusses the mandatory participation by companies and indi-
viduals, 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 indi-
vidual 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. Kun
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 industry-

wide 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 Table of contents
8 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

About the authors................................................................................ 237
Subject Index........................................................................................ 243

1 Introduction
Employee solidarity is central to the second pillar of the Dutch pension system.
This solidarity is given shape in collective schemes implemented by industry-
wide, 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 collec-
tive 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 pre-
sent 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 per-
ceived to benefit more than others. This inequality in the distribution of
costs and benefits is one reason why opinions are divided as to the desir-
ability 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 S.G. van der Lecq and O.W. Steenbeek
1.1 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 (Scien-
tific 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 in-
dividual 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 sur-
rounding the transition from defined benefit system to defined contribu-
tion 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 soli-
darity 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 ad-
vent 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 pen-
sion 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 sub-
sequent 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 par-
ticipant in the pool suffers a damage, the other policy holders will com-
pensate the loss. This is usually risk sharing, which is an important
element of all collective arrangements. Less obvious is one-sided soli-
darity, which refers to elements in collective pension arrangements
where it is clear ex ante who will subsidize whom. For example, in cur-
rent collective pension systems, value is transferred from young par-
ticipants to old participants. The term value transfers is usually ap-
plied to these elements of collective pension arrangements.
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 solidarity-
based 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 sug-
gests that solidarity can be an aim in itself: the action is consciously de-
signed 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, proba-
bly 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 ser-
vices 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, institution-
alised (cold) two-sided solidarity is on the increase. This suggests that peo-
ple are becoming more calculating in their behaviour and want to guarantee
that their solidarity is returned in equal measure by means of formal ar-
rangements. 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 subsi-
dising 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 retro-
spectively.
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 collectiv-
ities. 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 re-
sult, 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. Subsidis-
ing solidarity goes beyond the sharing of risks that is inherent in collectiv-
ity-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 intrin-
sic motivation by extrinsic motivation (Frey, 1997). Warm solidarity remains as
much in evidence as ever.
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 ar-
rangements 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, peoples
deep-felt need for formal solidarity arrangements reflects uncertainty over
the future rather than a fundamental lack of solidarity as such. This is con-
sistent with the finding that warm solidarity, which manifests itself infor-
mally both in one- and two-sided relations, is not decreasing. In fact, secu-
rity 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 com-
plementary(De Beer, 2005).
SOLIDARITY IN PENSIONS
Within the pension sector mandatory solidarity takes shape via the collec-
tivity. 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 ex-
tent. These people with different risk profiles show solidarity towards one an-
other, thus giving rise to the term risk solidarity. In the context of pension
funds, by contrast, risk solidarity is equated with risk sharing because econo-
mists 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
dependants 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 mem-
bers).
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 collectiv-
ity. 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 insur-
ers, 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 pen-
sion funds, revealing that collective schemes also yield better investment
returns than individually tailored schemes. Chapter 6 views the genera-
tions 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 con-
tribution to promote solidarity and indicates the costs and benefits of this
instrument. Chapter 8 homes in on solidarity from the perspective of the
individual. On the basis of a number of representative participants it

shows in what situations an individual is a net contributor or net benefici-
ary of the solidarity arrangements in the current pension system. Collectiv-
ity 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, Amster-
dam: Aksant, 2004.
Boender, C.G.E., S. van Hoogdalem, E. van Lochem and R.M.A. Jansweijer,
Intergenerationele solidariteit en individualiteit in de tweede pensioen-
pijler: 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 trans-
itionele 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 ambte-
naren, 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 risi-
cobeheersing in de pensioensfeer, WRR (Scientific Council for Gov-
ernment Policy) Report 55, The Hague, 1999.
Part 1. The concept of solidarity
2 Solidarities in collective pension
schemes
J.B. Kun1
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 soli-
darity 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 collective-
ity 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 intro-
ducing pension contracts. These contracts specify the relationship between the
funds 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 an-
swer. 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 evolu-
tion 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. Kun
participants is an important condition for realising such a contract. Soli-

darity 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 pro-
tecting society by avoiding social unrest and crime, and promoting public
order, health and housing. The nation state plays a crucial role in this ver-
tical 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 pro-
vide welfare to its vulnerable citizens. A sense of shared community, val-
ues 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 de-
pends 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 individuals employment and salary history (Kun, 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 Eng-
bersen, 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, soli-
darity 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 ethno-

cultural lines, with a growing number of smaller homogeneous subpopu-
lations 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 homo-
geneous groups. This may further reduce the appetite for society-wide soli-
darity, 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 iden-
tifiable 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. Its worth
noting, incidentally, that the terms risk-sharing and risk-sharing arrange-
ments are increasingly being used to denote situations which actually in-
volve 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 supplemen-
tary pension system. Section 2.3 then tentatively formulates some principles
providing a basis for making a distinction between desirable and undesir-
able forms of solidarity. As will be clear, these principles are to some extent
personally coloured. The appendix finally provides a summary of the differ-
ent 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 trans-
fers are also known as solidarities.
16 J.B. Kun
THREEFOLD DISTRIBUTION MECHANISM

A pension system realises a threefold distribution of funds: two are simulta-
neous 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-as-
you-go, funded or any combination of the two. With pay-as-you-go financ-
ing, the pensions are paid out of the earned income of the working popula-
tion, 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 persons life. The distribu-
tion 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 over-
state 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 encom-
passes numerous redistribution mechanisms which lead in each time pe-
riod 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 pen-
sion 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-
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 signifi-
cant 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 pen-
sion), between those who die young and those who become old, between
employees with divergent career paths, between males and females, be-
tween 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 redistri-
bution 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 settle-
ment 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 redes-
igned 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 unde-
sirable 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. Kun
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 expec-
tancy 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 be-
tween various socio-economic groups and between men and women. Gen-
der solidarity is incorporated in European legislation, in this case the anti-
discrimination 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 ques-
tioned, 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 in-
vestment performance. The latter entails that when interest rates fall the obliga-
tions apart from perfectly matching investments usually increase more than
the fixed-income investments. Finally we should note that no mutual redistribu-
tion exists between pension schemes in the Netherlands. Evidently there is such
a close though unspoken cohesion and/or special bond between the partici-
also referred to as intergenerational solidarity. Lower-than-expected re-

turns 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 re-
tirement? 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 contribu-
tions, whilst accruing less or no pension rights for themselves.
Asking for this sacrifice from young people is often considered justi-
fied given that they may be able to benefit from higher-than-average
returns in the future. Whats 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 contribu-
tions 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 redistri-
bution between different occupational groups such as between the less well-off
cigar makers and rich professional groups such as notaries, (consulting) actuar-
ies, 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. Kun
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 solidar-
ity 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, intergen-
erational solidarity is a universal good that makes everyone more prosper-
ous. 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 tran-
sition period towards a fully funded pre-retirement pension, many will
make substantial contributions towards early-retirement and similar (tran-
sitional) schemes without ever being able to enjoy the fruits of these
schemes themselves. In the light of the ageing population, the Dutch gov-
ernment 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 unin-
tended (inverse solidarity) and still others are unnatural (perverse solidar-
ity).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 soci-
ety. 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.
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 unmar-
ried (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 person-
dependent (think e.g. of socio-economic class and age). The current ten-
dency, 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 corre-
spondingly 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. Kun
than abruptly. Building, maintaining, and releasing buffers involves nu-

merous 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 in-
quisitive participant from the fund. Participants also frequently want
advice. The Pension Act deals extensively with the provision of informa-
tion to participants.
One new phenomenon is the pension contract, which lays down ar-
rangements concerning the relationship between the recovery of the funds
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 in-
dexation of accrued rights (of retirees and the active participants) on the
other. This integrated contribution and indexation policy must be imple-
mented 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, con-
tribution 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 re-
sponsibility for the proper working of the national economy. This com-
mitment to society as a whole is the most abstract form of solidarity: it is
the furthest removed from peoples 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 pen-
sion policy. Therefore, it is up to the supervisor to ensure that the funds op-
erate 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 in-
strument, whilst leaving room for reducing or even stopping indexation.
When serious shortfalls occur, accrued rights may even be reduced. In soli-
darity terms, this policy line spares younger employees, new entrants in
particular, whilst demanding a greater sacrifice from retirees and older em-
ployees. 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 de-
gree 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 soli-
darity between the various risk groups. (Aalberse et al., 2004). The second quota-
tion 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 wont buy that any more. (C. Petersen, 2002).
24 J.B. Kun
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 soli-
darities 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 unnec-
essary, 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 ar-
gue that they should be redistributed in the same way as the current in-
come. If this is combined with the premise that income policy is a matter
for politics and government, then it follows that neither pension institu-
tions nor other social partners (e.g. employers or trade unions) should
have any influence in this case. Income policy is conducted through taxa-
tion, subsidies and so on. Consequently, any solidarity elements in pen-
sion 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 basi-
cally belong to, or are attributable to, those participants who paid for these
surpluses. Therefore surpluses only benefit these participants.8 New en-
trants 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 sacri-
fice 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.
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 set-
tlement of increased or reduced contribution rates as well as surpluses and
shortfalls may make sense. But this must always be based on sound argu-
ments 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 genera-
tions. 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 cri-
teria 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 in-
come (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 contribu-
tion levels always justified? The following case is instructive.
A 45-year-old participant A leaves his employers 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 self-
employed person he will in the future continue saving for an individual
pension by paying a cost-price contribution (which is relatively high com-
pared 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 employers pension scheme. In the previous 20 years this
26 J.B. Kun
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 partici-
pant 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 intergenera-
tional 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 retir-
ees, but only up to a certain degree. They, after all, also have their own fu-
ture 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 soli-
darity 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 par-
ticipants while 84% would be absorbed by the present participants
(CPB, 2004). Of the latter, the active population would pay the lions
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 obvi-
ously 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.
There are good arguments for persuading the younger generation to ac-
cept 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 solidar-
ity contribution and the rules to be observed in times of relative hardship
or prosperity. But these issues can no doubt be resolved provided the par-
ties 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, lets 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 proba-
bly also benefit from a low contribution rate? And conversely, sup-
pose 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 incom-
ing 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. Kun
2.4 Conclusion
The issue of redistributing or subsidising solidarity in supplementary pen-
sion schemes needs to be treated with great care. Valid arguments can be
put forward for risk sharing. The creation of consensual support for di-
verse 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 Bedrijfs-
informatie, 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, social-
ecohesie 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, Eco-
nomisch-Statistische Berichten, 89, 25 June 2004.
Entzinger H. and G. Engbersen, Immigratie en solidariteit, in: W. Arts et

al. (ed.), Verzorgingsstaat vaar wel, Van Gorcum, Assen, 2004.
Kun, J.B., Op weg naar n 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?, Neder-
landers 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. Gabrils (eds.), Sferen van integratie;
naar een gedifferentieerd allochtonenbeleid, Amsterdam: Boom, 1985.
Vorselen, L. van, Solidariteit en pensioen, denkbeelden over een solidair ouder-
domspensioen, 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 vari-
ous forms are summarised below.
1. Risk solidarity (risk sharing): this principle underlies all insurance con-
tracts, whether individual or collective. The participants bear one an-
others 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. Kun
nominal return less the indexation of pension rights and payments

is greater than the actuarial discount rate, then the population of work-
ers will benefit in the form of a lower contribution. The greater this dif-
ferential, 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 contribu-
tion 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 1% of the pensionable salary for a 24-year-old.
With a level premium of 9% this therefore entails an implicit addi-
tional charge of 7%. For a 64-year-old the actuarially fair premium
is about 17% and he therefore receives an implicit subsidy of 8%.
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 partici-
pants is no longer stable. In a rapidly shrinking sector, the young
people of yesterday have paid high premiums only to find out upon re-
tirement 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 ac-
crual costs are included in the level premium paid by the actives.
6. Divergent career paths: these result in what is frequently called per-
verse solidarity, i.e. those who make a relatively steep career, particu-
larly towards the end of their working life (known as pension pro-
motions), benefit considerably from participants with a flatter career.
This income redistribution effect occurs irrespective of the level of in-
come. In recent years many schemes have been transformed from fi-
nal 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 per-
verse 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 soli-
darity concerns the option of exchanging (usually towards the age of
65) the surviving dependants pension for a higher or earlier old-age
pension. The principle of equal rights demands gender-neutral ex-
change factors, leading to a weighted average of male and female fac-
tors that generally work out favourably for women and less favoura-
bly 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 equiva-
lent 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 dependants 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 be-
tween 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 pen-
sion 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 disadvan-
tage for any of the participants. However if we look at a pension
package on the basis of its different components, the subsidising ef-
fect 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 princi-
ple. No scheme or scheme component is ever entirely self-supporting.
32 J.B. Kun
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 gen-
eration receives. One crucial factor (with a strong influence on the out-
come) is the rate of time preference, the discount rate used to calculate
the present value of future cash flows. The method of financing a pen-
sion 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 neu-
tral 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 basi-
cally 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 op-
tions without it being clear which components belong to the pension-
able 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 employees
working life will assuming a corresponding increase in income re-
sults 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 soli-
darity 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 health-
care 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 ever-
more 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 soli-
darity is pivotal in achieving equitable and efficient healthcare provision.
The outcomes of this analysis are compared with the solidarity arrange-
ments 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. Col-
lective 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 solidar-

ity is given shape in the financing and implementation of healthcare (sec-
tion 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 con-
tribution ends with a brief comparison of the most important characteris-
tics distinguishing the insurance-based healthcare market from the collec-
tive pension sector (section 3.6).
3.2 Solidarity
Solidarity is a concept which we almost all endorse, but which is also sub-
ject 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 be-
tween culture-based solidarity which sprang from (shared) identity and
utility-based solidarity (Widdershoven, 2005). Schuyt posits that: Solidar-
ity, as a social phenomenon, means sharing of feelings, interests, risks and respon-
sibilities (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 defi-
nitions 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 soli-
darity, 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 hypo-
thetical original position, citizens agree on a social contract through a
process of negotiation. These negotiations take place behind a veil of ig-
norance. 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 better-
off. Conflict between different groups is thus avoided. In reality, this situa-
tion existed during the establishment of the welfare state when strong
economic growth diminished scarcity and greatly increased everyones
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 free-
doms, provided that others enjoy equal freedom. Second, social and eco-
nomic 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 everybodys 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
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. No-
tions 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 obliga-
tion acquired a more rational dimension as arguments such as enlightened
self-interest started to gain currency. In the literature this process is fre-
quently illustrated by the insurance world, where cold utilitarian inter-
ests progressively displaced the warm sense of social duty in the solidar-
ity framework (Widdershoven, 2005). In the healthcare sector this devel-
opment is evident from the fact that many now perceive health as an
enshrined right (Starr, 1982). This evolution from subjective duty to ob-
jective 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 sociol-
ogy, 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 collec-
tivities can conflict in all sorts of institutions. Health insurers, for instance,
have dealings with the insured, who want the lowest premium, and pa-
tients who want the best (read: most expensive) healthcare. In the course
of their lives individuals may switch between the roles of insured and pa-
tient any number of times. In the pension sector the switch between collec-
tivities 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-
munity work. The state, acting in its capacity as over-arching body, orches-
trates 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 non-
payers from their health schemes (Widdershoven, 2005) and private insur-
ers hardly ever charged totally risk-based premiums (Schut, 1995). When
these spontaneous mechanisms enhancing solidarity came under pres-
sure, the government took regulatory action to ensure that the system con-
tinued 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 propor-
tionate 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 oth-
ers (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 auto-
matically 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 dif-

ferently in the (health) insurance industry. As this book focuses on the pension
sector, we prefer using the term as used in the pension industry.
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 provi-
sion 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 tenta-
tively compared with pensions in Table 1. Pension solidarity was exten-
sively discussed in the previous chapter in this book.
The Exceptional Medical Expenses Act (AWBZ) provides a national so-
cial 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 im-
portant issue in the healthcare sector. Similarly, subsidising solidarity oc-
curs 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 in-
sured 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 no-
claim 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.
Table 1. Solidarity in the financing of healthcare and pensions
Risk Subsidising Income Intergenerational

solidarity solidarity solidarity solidarity
AWBZ (Exceptional
o ++ ++ +++
Medical Expenses Act)
Collective
State pension (AOW) o + ++ ++
Mandatory Insurance o +++ + ++

Regulated
Collective pensions + + o o
Supplementary Health
+ ++ o +
Insurance
Free market
Private pensions ++ + o o
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 gen-
eral premium differences are limited. The fund is filled with wage-
dependent 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 fre-
quently 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 supple-
mentary 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 supplemen-
tary health insurance schemes of the sickness funds fetched a high gross
margin of 20% (RVZ, 2005).
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 un-
certainty 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 undifferen-
tiated 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 soci-
ety is prepared to accept higher taxation, lower purchasing power or sub-
stantial spending cuts in other policy spheres. This has never been neces-
sary 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 spend-
ing 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
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 medicalisa-
tion 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 re-
garded 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 advanc-
ing 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 ill-
nesses 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 subsi-
dising 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 trans-
fers. 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
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 individualisa-
tion 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 insur-
ers. The launch in the Netherlands of private health insurance to replace
the former social health insurance scheme as the dominant system cor-
roborates 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 in-
sured 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 treat-
ment for certain diseases (Hansen, Arts and Muffels, 2005). Society at-
taches 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 in-
sured are prepared to pay a higher contribution to maintain solidarity, but
39% are no longer prepared to do this (RVZ, 2006).
The end of the veil of ignorance?

This trend is compounded by an entirely different phenomenon: new sci-
entific 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 sup-
port 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 (Rosanval-
lon, 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 concen-
trated 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 at-
tempts 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 sci-
entific insights into the causes of disease entail that the current solidarity
arrangements will not be left intact without discussion. All aspects of sub-
sidising 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 interest-
ing 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 coin-
cides 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
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. Fu-
ture 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 individuals 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 pay-
as-you-go system runs via the premium levy system. If insurers are per-
mitted 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 in-
creased 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 allow-
ance to compensate for this higher premium.
Yet another redistribution of the costs can be achieved within a pay-as-
you-go system. This, in fact, is exactly what Germany did a few years ago
with its Exceptional Medical Expenses Scheme (Pflegeversicherung). Inter-
generational solidarity was invoked in justification of the measure, the
argument being that people without children should pay a higher pre-
mium 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
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 effec-
tive 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 op-
tion. 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 insurers 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 sub-
sidising 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.
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 posi-
tioned 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 re-
distribution in pension systems is therefore mainly the (unintended) conse-
quence of such factors as the achieved return on investments and wage in-
flation (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 as-
sume that the strong healthcare solidarity we know today will remain in-
tact in the future. With health expenditures set to continue rising, the
health sector will make a structural claim on a large portion of the avail-
able economic growth. Whats 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, behav-
iour and lifestyle, is rapidly wearing thin. This is likely to influence the
degree of support for solidarity arrangements in society where differentia-
tion and individualisation are still gaining in significance. Views differ as
to the exact impact of these trends, but few expect them to strengthen sup-
port 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 supple-
mentary health insurance and certain recent measures, including the no-
claim 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 re-
lated premiums, supplementary savings schemes, and intergenerational
redistributions to make the (affluent) elderly pay more for healthcare.
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 collec-
tieve verantwoordelijkheid, Nijmegen: Valkhof Pers, 2005.
Berg, B. van den and F.T. Schut, Het einde van gratis mantelzorg?, Econo-
misch Statistische Berichten (88) 4413: pp. 420-422, 2003.
Battacharya, J. and N. Sood, Health insurance and the obesity externality, Work-
ing Paper 11 529, NBER, 2005.
Buijsen, M.A.J.M., Solidariteit, rechtvaardigheid en de zorg om gezond-
heid, 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, Mnchen: 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 patinten en
clinten 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 vergrij-
zing, Tijdschrift voor Openbare Financin (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 Re-
duce Global Savings, McKinsey, 2005.
NYFER, Blijvende Zorg: economische aspecten van ouderenzorg, Zoetermeer:

RVZ, 2005.
RIVM (National Institute for Public Health and the Environment), Risico-
solidariteit 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 onder-
linge 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 rein-
surance. 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 signifi-
cantly lower than those of company pension funds. Next, the costs of pension
schemes offered by pension funds and life insurers in the Netherlands are com-
pared 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 long-
term 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 con-
1 The Dutch central bank (DNB), Supervision Policy Division, Strategy Depart-
ment, j.a.bikker@dnb.nl. Jan de Dreu wrote this article when working for DNB;
currently he is employed by ABN AMRO, jan.de.dreu@nl.abnamro.com. Views
expressed are those of the individual authors and do not necessarily reflect of-
ficial 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 in-
crease 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 pri-
vate 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 pri-
vate schemes, unutilised economies of scale is the dominant factor explain-
ing 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 com-
plicated.
4 Operating costs of pension schemes 53
0%
-10%
Reduction of pension benefits
-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 as-
sume 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 exclud-
ing asset management, such as record keeping, communication with par-
ticipants, policy development and compliance with regulatory and super-
visory 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 re-
ported 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 statis-
tics on administrative costs in 2004 for, successively, different size catego-
ries, types of pension funds and types of pension plans. Size is measured
by the number of participants or total assets. Participants consist of con-

tributing employees, inactive participants and pensioners. Data from ear-
lier 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 underre-
porting does not occur among industry-wide pension funds, as these are
unable to shift costs to their sponsors. Later we will see that these imper-
fections 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 econo-
mies 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 actu-
ally even greater than shown in these figures. This is due to the underre-
porting 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 supervi-
sory 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.
Table 1. Annual administrative costs of pension funds by size category (2004)
Size categories of pension Administra- Administra- Total assets Funds that Total num- Number of
funds based on: tive costs/ tive costs per partici- do not report ber of funds
total assets per participant wage costs participants
(%) pant ( 1000) (%)a (1000)
()
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.
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 sys-
tems, 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 nega-
tively 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 operat-
ing costs on pension benefits. In summary, Table 1 shows that the operating
costs of pension funds are characterised by strong economies of scale, irre-
spective of whether the size of the institution is expressed in terms of par-
ticipant numbers or total assets.
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 com-
pany 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 em-
ployers and labour unions in this sector. There are two types of industry-
wide pension funds: mandatory and non-mandatory. Mandatory funds
are based on a binding CLA, making participation mandatory for all em-
ployers 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 practitio-
ners 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 Number of Average

costs/total costs per per participant of participants fundsa number of
assets participant ( 1000) (1000)a 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
Pension type
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
a The pension type of 21 pension funds is not known; four funds are savings funds.
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 num-
ber 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 be-
tween 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 partici-
pants 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 deter-
mined by scale effects. Overall, scale effects appear to be the dominant ex-
planation 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 organisa-
tional 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.

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 re-
gression 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 (Bate-
man and Mitchell, 2004), and in relation to DC pensions in sixteen coun-
tries 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 pen-
sion 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. Non-
mandatory industry-wide pension funds and company pension funds oc-
cupy 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.
Table 3. Estimates of the administrative and investment cost models (1992-2004)
Administrative costs Investment costs

Coefficients t-values Coefficients t-values
Number of participants (in logarithms) 0.63 105.1

Total assets (in logarithms) 0.83 76.4
Ditto, squared a 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
Number of observations 10 119 7.4 4 986
R 2 0.71 0.75
aRespectively, the number of participants (in logarithms) and total assets (in logarithms).
Note: Almost all coefficients are significant at the 99% level; italics indicate no signifi-
cance. All variables are expressed in 2004 prices.
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 admin-
istrative 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 pro-
vided with information and advice if they can choose to select their own in-
vestment mix for their pension.
least for company pension funds. It is found that both full and partial rein-
surance 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) por-
tion of the funds has evidently not prevented clear regression results. All
coefficients are significant at a very high reliability level. Also, if the re-
gression 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, there-
fore, 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 administra-
tive 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
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 pen-
sion funds for different size categories, expressed in numbers of partici-
pants. Investment costs as a percentage of total assets decrease as the num-
ber 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 par-
ticipant 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 low-
est 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 ear-
lier finding that scale has a strong impact on operating costs and that indus-
try-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 econo-
mies of scale are found to exist, though these are smaller than for administra-
tive costs (17% versus 37%). For investment costs it is found that after con-
trolling for other factors industry-wide pension funds again operate at sig-
nificantly 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 ex-

cess 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 in-
vestment 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.
Table 4. Annual investment costs of pension funds by size category (2004)
Size category of pension Administrative Administrative Total assets Total num- Funds that Number of
funds based on: costs/total costs per per partici- ber 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
2. total assets ( million)
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
4.5 Life insurers as providers of pension

schemes
The second part of this chapter examines the cost differences between pri-
vate 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 in-
surers and pension funds.
COLLECTIVE PENSION SCHEMES

Companies that do not belong to an industry with a mandatory industry-
wide 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.
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, in-
cluding 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 in-
surers or other institutions are better equipped to bear certain risks of pen-
sion funds or are able to perform certain activities more cost-effectively,
pension funds can increase their efficiency by reinsuring risks and out-
sourcing 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 employ-
ment and are not members of an professional group pension fund. In addi-
tion, many people choose to supplement their employee pension with ad-
ditional 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 in-
vestment 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.
income tax deduction, subject to certain conditions.14 In the case of en-

dowment 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 prod-
ucts, 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 oc-
cur 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 Pension funds
Size category Administrative Gross Administrative Number of Administrative Number of

based on total costs/gross profits/gross costs plus gross insurers costs/gross pension funds
assets premiums premiums (%) profits/gross premiums
( million) (%) 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 sav-
ings via blocked bank accounts.
15 The data are described in Bikker and Van Leuvensteijn (2007). Investment costs
are discussed later in this chapter.
12.4% for the largest life insurers to 37.9% for the smallest. In propor-
tional 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 lon-
gevity 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 av-
erage, 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 av-
erage costs of all life insurance products. We lack the data required for a
more refined analysis. It is plausible, however, that large cost discrepan-
cies will occur for different types of products, such as collective versus
private contracts. With collective contracts the costs will decrease rela-
tively 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 aggre-
gated 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 com-
parison, 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.
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 de-
liver, the same products as pension funds. This is not the case. Most pen-
sion funds provide DB pension plans where the size of the pension bene-
fits 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 contri-
butions). They cannot distribute the investment, inflation and longevity
risks over different generations by varying contributions. In general, in-
surers 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 em-
ployer undertakes to pay the additional contribution required for indexa-
tion and supplements up to a certain percentage of the final salary (so-
called back service in final pay schemes). Besides direct schemes and col-
lective contracts that are comparable with pension fund schemes, life in-
surers also provide reinsurance contracts and private policies for indi-
viduals, including both pension schemes as well as other types of insur-
ance. 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 par-
ticipation 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 intermediar-
ies. Insurers need to incur these costs to acquire customers, while pension
17 Sometimes partial indexation is guaranteed.

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 insur-
ers 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 partici-
pation, costs related to adverse selection play no role for pension funds. In
addition, buyers of annuities tend to be more highly educated and remu-
nerated people who, on average, are healthier and have a longer life expec-
tancy. 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.
DIFFERENT REGULATORY REGIME

Finally insurers must cover the risks on insurance contracts by maintain-
ing 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 li-
abilities for 105% and also to maintain a solvency buffer for investment
and longevity risks. The required buffer for an average fund is approxi-
mately 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 connec-
tion 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.
Table 6. Administrative costs of life insurers and pension funds (2000-2004)
Life insurers Pension funds

Year
Administrative costs
Administrative costs Gross profits Administrative costs
and gross profits
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 per-
centage 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, insur-
ers 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 mort-
gages) and annuities (immediate annuities and endowment policies with an
annuity clause). The latter include advice.
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 of-
fered 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. Unfortu-
nately, 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 be-
come 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. Evi-
dently 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 insur-
ers 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 mil-
lion insured persons.
4.7 Conclusions
This chapter shows that the administrative costs of collective pension
schemes offered by pension funds constitute only a fraction of the operat-
ing 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 dif-
ference becomes almost twice as large when the gross profit margin of
insurers is also taken into consideration: 4.4% versus 23.9%. These differ-
ences are explained by, among other things, scale effects, adverse selec-
tion, 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 pri-
vate 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 ques-
tion. More transparency to stakeholders about operating costs could help
in this respect. The conclusion here is that the consolidation of small pen-
sion 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 discus-
sion 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.
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 ad-
ministrative 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 pen-
sion 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 Econom-
ics and Finance, forthcoming, 2007.
Bikker, J.A., L. Spierdijk, P.J. van der Sluis, Market impact costs of institu-
tional equity trades, Journal of International Money and Finance, forth-
coming, 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 man-
datory 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 Econom-
ics, forthcoming, 2007.
Appendix: Key data of pension funds and life

insurers
Table B.1. Key data of pension funds (1992-2004)
Year Number Number Number of Number of Total Number of Total costs/ Total costs
of funds, of industry- company funds in assets, participants, total assets per partici-
total wide pen- pension 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.
Table B.2. Technical provisions and gross premiums of life insurers and pension
funds (2004; billion)
Private Collective Total
Gross premium
Endowment insurance - insurers 9.0 0.9 9.9
Pension and annuity insurance 0.1 3.0 3.1
Insurers
Total life insurersa 9.1 3.9 13.0
of which for pensionsb 1.2 3.9 5.1
Pension funds 22.8 22.8
Technical provision
Endowment insurance insurers 101.2 7.9 109.0
Pension and annuity insurance 21.4 84.8 106.2
Insurers
Total life insurers 122.6 92.6 215.2
Pension funds 446.9 446.9
a Excluding annual deposits in savings banks;
b 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.
Source: DNB, Financial data life insurance companies; and Statistics Netherlands
(CBS), National Accounts.
Table B.3. Gross profits of life insurers (2004)
Gross profits Gross profits /gross Gross profits / technical

( million) premiums (%) provisions (%)
Private insurance 1 192 12.3 1.0

Collective insurance 624 15.9 0.7
Totala 1 816 13.9 0.9
a 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 con-
tract? 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 aca-
demic 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 re-
garding the exact added value of pension solidarity and the precise form of opti-
mal 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 con-
tributions 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 pen-
sion 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 ade-
quate reward for their role as shock absorber.
This reward is all the more vital now that solidarity in the pension sys-
tem 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 struc-
tural decrease in annuity rates and returns in an ageing society and no
further increase in estimated longevity, the pension system will not be-
come 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 pen-
sion 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 systems
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 up-
shot, 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 ap-
pendix of Chapter 2 in this volume, by Jan Kun.
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 demon-
strated in a study carried out on behalf of the Scientific Council for Gov-
ernment Policy (WRR) (Boender et al., 2000). This report provides quanti-
tative 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 switch-
ing to defined-contribution (DC) systems where all pension risks are off-
loaded onto the individual participants.
However, the risk-sharing assumptions made in this report do not cor-
respond 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 assump-
tions. 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 aca-
demic 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 practi-
cal reality. The extent to which this is the case and the resulting conse-
quences 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 aca-
demic literature diverges significantly from the outcomes of the aforemen-
tioned WRR study. With the assumptions applied in the WRR study, it is
found that, notably, risk solidarity between generations delivers substan-

tial economic benefits and that buffers have great added value. This soli-
darity comes about because people in the workforce if necessary pay
extra contributions and temporarily accept underfunding in order to pro-
tect 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 confu-
sion over the actual importance of these two pillars within optimal pen-
sion 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 dif-
ferences for the resulting valuation of risk solidarity.
Quantification of the consequences of the differences in the applied as-
sumptions 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 partici-
pants. An optimal pension policy for the individual consists of a combi-
nation 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, indexa-
tion 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 ex-
treme 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.
individual maximises the utility of the expected consumption in each fu-

ture 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 con-
sumption decreases with the level of consumption. Another basic assump-
tion is that the more stable the development of this consumption, the
greater the positive utility awarded by the individual to the future devel-
opment 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 re-
tirement. 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 infor-
mation;
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 individ-
ual is surprisingly simple. At each moment the same portion of the total
assets must be invested in equities, where the total assets consist of finan-
cial capital and human capital (the discounted value of future earnings3).
The share of financial capital in the total assets increases with the individ-
uals 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 equi-
ties; and also that this young persons 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.

The young individual must then invest 500% in equities and the old indi-
vidual 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 ex-
pected 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 higher-
than-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 economys loss of capital is spread as equally as possible over all age
groups and also over each persons 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 theo-
retically 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 theoreti-
cal pension model by also including future generations in the risk-sharing
mechanism. In this context, the term future generations refers to genera-
tions who do not yet participate in the labour market, including genera-
tions who are not yet born when a shock on the financial markets occurs.
Basically these generations are then already investing in the financial mar-
kets before they start paying contributions. This increases the opportuni-
ties 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
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 oc-
cur, 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, imple-
mentable private and collective pension contracts relate to the recommen-
dations made by the simplest theoretical model. In the simplifying as-
sumptions underlying this optimal pension policy, equity returns have no
memory, i.e.: there is no question of mean reversion. Mean reversion en-
tails 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 ho-
rizon, because the returns average out to a certain extent over time (see, for
instance, Siegel, 2002; Steehouwer, 2006). The differences in optimal in-
vestment 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 ac-
customed 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 finan-
cial capital in equities is reasonably robust to changes in other assump-
tions 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 peo-
ple are unable to borrow against their human capital to adjust their con-
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 cy-
cle 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 con-
tracts that are optimal in practice may diverge from the optimal pension
contracts in section 5.2. In addition, more and more information is becom-
ing available (Van Rooij et al., 2004) about how people value possible fu-
ture developments concerning their pension savings and pension pay-
ments. 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 selec-
tion 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 addi-
tion, 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).
young people are unable to take over the risks of elderly people in ex-
change 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 longev-
ity 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 distri-
bution. Certain types of macro-economic shocks (such as political uncer-
tainties) 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. Ad-
verse 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 in-
dexes, 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 invest-

ment 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 strate-
gies that pension funds use to optimise performance). In addition, indi-
viduals often lack the expertise to save and invest rationally. People have
difficulty making complex decisions in uncertain circumstances. The re-
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 percent-
age point of the invested capital that an individual loses annually as a re-
sult 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 conse-
quently 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 over-
lapping and non-overlapping generations that is not (yet) possible in capi-
tal markets. This takes the form of young people paying catch-up contribu-
tions 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 inflation-
proof 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 capi-
tal 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 Sundn (2004) and also Van Els et al. (Chapter 9) in this
book.
long term which cover the longevity risk) that are not yet (readily) avail-
able 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 con-
tracts and the optimal pension contracts given the assumptions in section
5.2. Moreover, the funds provide employees with access to complex in-
vestment strategies that few individuals could use if left to their own de-
vices. 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 funds 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 pre-
vents adverse selection in the market for life insurance and annuities. All
this reduces the implementation costs. Due to their close ties with the so-
cial 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 capi-
tal. 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 retire-
ment 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 dif-
ferent sector or company or as a self-employed person. They can also de-
cide 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
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
participants individual circumstances, collective pension funds could
provide more customised pensions.
This, however, has a price tag in the form of higher transaction and in-
formation 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.
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 es-
tablish 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 de-
scribe 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 individ-
ual and the collective pay a single premium which, given a fixed ac-
tuarial interest rate, is sufficient for a nominal 70% average-pay old-
age pension and 49% survivors 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 pre-
mium mechanism means that the elderly can continue benefiting
from the equity risk premium via catch-up contributions. If neces-
sary, 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 stochas-
tic scenarios whose characteristics (uncertainties, correlations, mem-
ory) are based on historical figures (1966-1998) for the applied fac-
tors (inflation, interest and return). The central return expectations
in the long term are determined according to the insights applica-
ble at the time.
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 sur-
vivors pension. In plain language, this means that a twofold in-
crease 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 indi-
vidual 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 con-
tributed pension capital and the same expected pension result, the indi-
vidual 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 par-
ticipants 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 con-
tribution volatility. In the WRR approach this plays no role in the evalua-
tion, 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 be-
tween the envisaged and the realised pension. In the WRR study, the fact
that the actual pension undershoots the target pension is penalised quad-
ratically. 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 im-
plies 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
relatively more weight. If people also valued upside deviations in the pen-
sion 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 sys-
tem, shocks must be absorbed within one year. The WRR study assumes a
uniform investment mix for the collective fund, irrespective of the partici-
pants 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 mar-
ket 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 theoreti-
cal models. Thanks to the solidarity within collectives, risks such as infla-
tion risk and longevity risk, which are difficult if not impossible to trade
on financial markets, can be implicitly traded within the fund by the par-
ticipants. 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 con-
trast, 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 corre-
spond more closely to the most common assumptions in the academic
literature.
This is an important reason for carrying out a follow-up study. In addi-
tion, the WRR study needs to be deepened further even though it con-
tained 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.
5.6 Research agenda

The previous section showed that the WRR study must be enriched in cer-
tain 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. Re-
cent experiences suggest that contribution volatility can be costly for
the sponsor of a pension scheme. If this is the case, the optimal pen-
sion contract could shift from DB towards DC. By contrast, if par-
ticipants are less concerned about contribution volatility and contri-
bution pressure and attach more importance to keeping pensions in-
flation-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 opti-
mal pension contract will thus change.
In this case the participants take a risk in relation to their personal

health (at age 40 one doesnt 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 col-
lective 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, real-
ised pension accrual, etc. This can easily lead to wealth loss as com-
pared 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 fi-
nancial 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 analy-
sis. In this case, collective systems can add value as there are virtu-
ally 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 con-
tracts the most suitable and efficient way of achieving risk solidarity?
Above all, this contribution sought to sum up the current state of affairs
regarding this central research question in the literature. The adopted ap-
proach 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 soli-
darity. 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 Weaknesses
Private pension saving Individual customisation of invest- Suboptimal choices due to low
ment and contribution policy 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 mar- Continuity of solidarity not
kets guaranteed
Enables young people to take Not geared to heterogeneous
much more risk (elimination of participants
restrictive conditions) Ownership rights are not
Low costs 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 pen-
sion schemes both have strengths and weaknesses (see table) and because
the assumptions underlying the assertions about the optimal pension con-
tract 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.
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 Work-
shop on Financing Retirement in Europe: Public Sector Reform and
Financial Market Development, Louvain, Belgi, 2005.
Munnel, A. and M. Sundn, 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, Am-
sterdam: 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 oc-
curs. This method which we term value-based generational accounting is ideally
suited to investigating how far current policy itself, and changes to that policy, re-
sult in a balanced sharing of costs, benefits and risks across the generations partici-
pating 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 opera-
tion of the uniform contribution rate. We regard this practice as a given. The contri-
bution 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 Alex-
ander 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 evalua-
tion of current policy and policy variations. The approach of value-based genera-
tional 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 account-
ing. 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 pen-
sion fund. Recent studies reveal that intergenerational risk sharing within a
pension fund is welfare-improving for the plan members vis--vis an op-
timal 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 gen-
erational 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 un-
certainty. In formulating generational accounts we need to bear such un-
certainty 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 con-
tribution rate, indexation allowed and the development of the financial
position and the uncertainties involved. Value-based generational account-
ing makes clear how changes in the financing structure and the risk alloca-
tion 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 possi-
ble 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
funds 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 contribu-
tion methodology) or a change in the rules of risk allocation between mem-
bers (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 mo-
ment that a policy is altered does not change, the pension fund can there-
fore 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 be-
tween the members.
In a company pension scheme value transfers are mostly transfers be-
tween the companys 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 trans-
fers 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.
6.2.2 GENERATIONAL ACCOUNTS

Generational accounting is a method developed by public finance econo-
mists as a tool for investigating the intergenerational distributional effects
of fiscal policy (Auerbach et al., 1999, Kotlikoff 2002). Generational ac-
counting is based on the governments intertemporal budget constraint,
which requires that either current or future generations pay for govern-
ment spending via taxation. The governments net wealth (including debt)
plus the present value of the governments net receipts from all current
and future generations, must be sufficient to pay for the present value of
the governments current and future consumption. The generational ac-
counting 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 experi-
ence a decrease in net lifetime income. Planned increase or decrease in gov-
ernment 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 prom-
ised benefits must be backed by current and future contributions and re-
turns on paid contributions. Secondly, as the government uses the tax in-
strument 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 uncer-
tainty into consideration when valuing future benefits and costs. In apply-
ing 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-
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 sim-
ply chosen to estimate generational accounts for a range of discount rates. (Kot-
likoff, 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 exam-
ple 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 litera-
ture (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 en-
dogenous 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 underly-
ing 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
acceptance in the context of pension funds with the development of the so-
called 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 deter-
mining 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 euros at to-
days 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 solu-
tion to the problem of how to deal with uncertainty surrounding the re-
sults 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 val-
ues 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.
development of a financial position for the fund as a whole. Value-based

ALM allows us to see the economic value in euros of participation in the
fund by various members, and how large the possible value transfers be-
tween 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 nine-
ties, this led to a lot of discussion between stakeholders regarding the
allocation of what at the time was regarded as a high level of overfund-
ing. The discussion centred on whether this surplus should be used for
contribution holidays, extra indexation or the financing of early retire-
ment. 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 account-
ing 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 propor-
tion 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
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, value-
based 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 ap-
plied 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 projec-
tions 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 gen-
erational accounting.
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 ac-
crual 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 re-
ceives 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 actu-
arially.
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 allo-
cated 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 ex-
pense 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 indi-
cates 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 pen-
sion 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.
Table 1. Outline of generational accounts
Generational account Generational account Increase /decrease

standard variant X alternative variant Y
Age 2yrs. X-2 Y-2 X-2 Y-2
Age 30yrs. X30 Y30 X30 Y30
Age 60yrs. X60 Y60 X30 Y60
Age 105yrs X105 Y105 X105 Y105

Sum = 0 Sum = 0 Sum = 0
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 funds 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 in-
dexation 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 par-
tial 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 eco-
nomic 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
Costprice contribution rate
Full
indexation
indexation
0%
100% 140% Funding ratio
Fig. 1. Policy ladder basic variant
1.5%
1.0%
0.5%
% total liabilities
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
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 pic-
ture of the total contributions paid by the generations and the total benefits
being received by the generations. Secondly, the approach does not in-
clude welfare aspects regarding the pension scheme in the analysis. Mem-
bership of the pension funds pension scheme gives the prospect of pen-
sion 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 be-
fore 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 utility-
based ALM, but for the time being this is not useful yet for practical appli-
cations.4
4 The academic literature contains a number of initiatives for formulating an objec-

tive 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.
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 ef-
fects 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 pol-
icy 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 redistribu-
tion 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.
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* FR < 105% Contribution Indexation as % of wage

2024 2005 2024 2005 2024 2005 2024
Median 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 under-
funding (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 con-
tribution 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 stan-
dard variant to 11%.
The switch to 100% equities reveals a completely opposite picture to the

switch to 100% bonds. In the standard variant the contribution is here re-
duced 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 fund-
ing ratio is considerable. The risk of underfunding thereby increases dra-
matically, from 0.4% in the standard variant to over 8% for the 100% eq-
uity mix. Consequently, the spread of the indexation is also significantly
greater. The risk of the indexation being less than 80% of the full indexa-
tion 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 invest-
ment 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, be-
cause this cohort will be aged 18 in 2025 and then be in the pool of mem-
bers. The vertical axis represents the change in the value of the genera-
tional accounts as a percentage of the value of the liabilities in 2006.
2.5%
2.0%
1.5%
1.0%
% total liabilities
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
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 pen-
sion 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 signifi-
cance 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 ac-
crued benefits and change in value contributions paid in.
The lower contribution paid in leads to value gains for employees com-
pared 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,
0.20%
0.15%
0.10%
0.05%
% total liabilities
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 (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 ex-
isting 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 fluc-
tuates 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. re-
fund 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.
Contribution rate
Costprice
contribution
rate
indexation
140% Funding ratio
Fig. 5. Policy ladder for traditional DB variant.
Table 3. Results of classic ALM analysis for traditional DB variant
FR* FR < 105% Contribution Indexation as % of wage Volatility of con-

2024 2005-2024 2005-2024 2005-2024 tribution rates
median median median < 80% median

Standard variant 161% 0.4% 18% 98% 22% 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 av-
erage contribution is somewhat lower than for the standard variant. Natu-
rally, 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.

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
0.20%
0.15%
0.10%
0.05%
% total liabilities
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 (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 pay-
ments. 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 aver-
age, a smaller contribution rate. This variant therefore results in the pen-
sion fund providing greater value to its members than the standard vari-
ant. This is reflected by the direction of the interrupted line, which repre-
sents 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 employ-
ees 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 bal-
ance 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 collec-
tive DC variant. Here the indexation is always related to the financial
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 fund-
ing 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 in-
dexation is subject to greater volatility than in the standard variant.
Table 4. Results of classic ALM analysis for a collective DC
FR* FR < 105% Contribution Indexation as % of wage Indexation

2024 2005-2024 2005-2024 2005-2024 volatility
median 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

The results for the classic ALM analysis are reflected in the value analysis.
All members gain value through acquiring extra claims via higher indexa-
tion (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%
0.2%
% total liabilities
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 (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.
In this chapter we have analysed the generational effects of various pol-

icy 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 ad-
justment will therefore result in there being generations that benefit from
such changes and generations that lose out. For example, more risky in-
vestment (greater investment in equities) leads to value gains for younger
employees and value losses for older employees and people receiving pen-
sions. 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 allo-
cated (collective DC).
Future developments may give reason for implementing changes to the
financing structure. Such changes could lead to unintended and undesir-
able value transfers between generations. With the proposed method of
value-based generational accounting, we can identify such incidental gen-
erational effects where policy changes are made. The challenge is then to
search for such combination of policy parameters that the resulting gen-
erational effects can be regarded as acceptable and justifiable to the mem-
bers.
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): 605663, 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 Til-
burg, 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.
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. Steen-
kamp, 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 infla-
tion-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 Statis-
tische 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 - irrespec-
tive 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. Compensa-
tion 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 soli-
darity is increasing. This criticism originates both inside and outside the
sector. Within industry-wide pension funds, members as well as employ-
ers 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 solidar-
1 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 everybodys interest to
insure against these risks in a group scheme.
Things are different with subsidising solidarity. This solidarity goes fur-
ther 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 col-
lectivity 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 pen-
sion rights for the same price (both accrual and contribution are a calcu-
lated 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 solidar-

ity. 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 inter-
est 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 dif-
fers; 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 contribu-
tion 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 re-
lated. For old members, the required contribution rate is four times as high
30%
contribution as a percentage of salary
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 con-
tribution 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 cur-
rent 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 ele-
ment 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-year-
old reaching the age of 65.
80%
accumulated prepaid contribution as a
70%
percentage of annual salary
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 ac-
crues 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 ac-
crued 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. How-
ever, 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
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 subsi-
dies 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 contribu-
tion 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 con-
tribution, this amount would have been 350,000. Member A has there-
fore received a subsidy of 60,000 in the uniform contribution and ac-
crual 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 Kun explains several types of subsidizing solidarity in Chapter 2 of this volume.
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 with-
out a uniform contribution and accrual system. B has paid 36,500 in con-
tributions up to the age of 35. If he had paid the actuarially required con-
tribution, 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
funds board decides to economise on the scheme. The new accrual per-
centage 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 contribu-
tion, 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 aver-
age age of the contribution payers will become 5 years higher. This re-
sults 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 con-
tributions. 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.
The subsidies in the uniform contribution and accrual system take place
not only between generations, but also within them. Consider, for exam-
ple, 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 sub-
sidy to member Y.
Both members belong to the same generation. However, a mutual sub-
sidy 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 employ-
ees 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 contribu-
tion 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 per-
son 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.
NO COMPETITIVE DISADVANTAGE WITHIN INDUSTRIAL SECTORS

Thanks to the uniform contribution, a company with many older employ-
ees 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 sub-
sidies were not perceived as a problem, and, in view of the pattern of em-
ployment, 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 ac-
crual system in 1949, in spite of this possibility. In this section, we wonder
whether the mutual subsidies can lead to pressure on the uniform contri-
bution and accrual system, now or in the future.
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 con-
tribution 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 (Kun, 2005). Such soli-
darity 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 in-
crease. 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 vul-
nerable without harming the advantages of collectiveness and solidarity.
We examine two possible alternatives:
1. The progressive contribution system

In the progressive contribution system, each member annually ac-
crues 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.
2. The decreasing accrual system
In the decreasing accrual system, each member pays an equal contri-
bution rate, irrespective of his age. In return, he receives an age-
related 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%
accrual as a percentage of annual salary
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 col-
lective 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?
Table 1. Comparison of three pension systems
Uniform contribution Progressive Decreasing accrual

and accrual system contribution system system
Collective risk sharing? YES YES YES

Leeway for subsidising solidarity? YES YES YES
Mutual non-reciprocal subsidies
YES NO NO
(as in section 7.3)?
Does contribution burden depend on
YES YES NO
average age?
Competitive disadvantage within sector? NO YES NO
Does each individual receive the same
YES YES NO
accrual?
d. Does the contribution burden depend on the average age of the con-
tribution payers?
e. Does the system lead to competitive disadvantage within the busi-
ness sector for older employees and companies with a population of
older employees?
f. Does the system provide each individual member with the same ac-
crual 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 ac-
crual 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 contribu-
tions 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.
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 pro-
gressive contribution system, each member pays an age-related contribu-
tion. 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 gen-
der, 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 demo-
graphic 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 de-
creasing 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 ceil-
ings 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.
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 at-
tached 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 collec-
tive 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.
Case 6: What amount is the total compensation for the prepaid con-
tribution?
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 retrospec-
tively, so that every member receives a compensation for the contribu-
tion excess paid by him. But this is very difficult in practice, because
we then have to identify the precise history of every individual mem-
ber. For the purpose of illustration, we calculate the level of the com-
pensation by looking forward. For each member therefore we calcu-
late 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 ac-
crual 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 as-

sume that all the members continue working in the business sector
until the age of 65. That will probably not be the case in reality. How-
ever, it is noted that the compensation amount largely depends on the
composition of the membership. For a young membership, compensa-
tion is inexpensive, and for an old membership as well. For a pension
fund such as PGGM, a switch to the decreasing accrual system is rela-
tively expensive. The average age of the PGGM membership is actu-
ally 45, and many of the active members (30%) are aged between 40
and 50.
7.8 Structural effects

Suppose that all pension funds in the Netherlands were to abandon the
uniform contribution and accrual system to switch to the decreasing ac-
crual 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 be-
cause, 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 pen-
sion 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 pen-
sion 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
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 pen-
sion 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 soli-
darity will also lead to other forms of solidarity being put under pressure.
As a result, these forms are also removed, until the system ultimately con-
tains 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 vul-
nerable.
The loose thread is the pay-as-you-go element in the uniform contribu-
tion and accrual system. If the mutual subsidies resulting from that be-
come 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 pres-
sure on solidarity as a whole. A possible result of this is a flight to more in-
dividual 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 demo-
graphic developments and, just like the uniform contribution and accrual
system, prevents competition within the business sector.
However, a switch to the decreasing accrual system may entail undesir-

able 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 pensioen-
stelsel, Het Financieele Dagblad, 1 July 2005.
Kun, J.B., Billijkheid en doelmatigheid in het systeem van de (aanvullende)
pensioenvoorziening, Financile 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 consider-
able redistribution. Women benefit more from the retirement pension than men,
while benefiting less from the partners pension. Single people benefit less from
schemes than partnered members. The option of exchanging the accrued part-
ners 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 dis-
tinctions 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 ac-
crual. Every year, each employee in an enterprise or business sector ac-
crues the same percentage of pension for one and the same uniform con-
tribution. The uniform contribution rates in pension schemes lead to sub-
stantial 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 partners pension. Single
people do actually take part in paying for the partners pension, but until
recently they did not benefit from it. For a few years now, however, pen-
sion funds have been required to offer to exchange the accrued partners
138 K. Aarssen and B.J. Kuipers
pension for supplementary retirement pension. These are only a few ex-
amples of the many forms of redistribution within a collective pension
scheme. Elsewhere in this volume, Kun (Chapter 2) provides an extensive
summary of the different solidarities in collective pension schemes. Fur-
thermore, 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.
Hri, 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 earn-
ings scheme. In this typically Dutch pension scheme members at retire-
ment age have the option to exchange the accrued partners pension for
supplementary retirement pension. The scheme offers employees, during
their career, a survivors pension insurance in case of death, and contribu-
tion-free pension accrual in case of disability. We quantify the redistribu-
tion that is caused by the uniform contribution rate by comparing the ac-
tual price or the actuarially fair contribution for a number of representa-
tive 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 partners
pension must pay an actuarially fair contribution as well. Furthermore, for
private products, insurers can be selective with acceptance and may re-
quire a medical examination. However, it cannot be concluded from this
that members who pay the price for the redistributions in a collective pen-
sion 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 em-
ployees 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 139
8.2 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 pen-
sion and survivors pension. The exact elaboration of the schemes, however,
can diverge depending on the preferences of employers and employees in-
volved. 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 pen-
sion 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 three-
quarters 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 pen-
sion funds saw themselves forced to switch from a final wage scheme to
an economised average earnings scheme.
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
Partners pension 70% in case of death during participation (risk-based)
70% in case of death after retirement (accrual)
Exchange of partners pension 15% higher retirement pension
Orphans 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 per-
centage 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 earn-
ings. 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 usu-
ally indexed with the development of contractual wages. It is easier for a
pension fund with an average earnings scheme to absorb a sudden dete-
rioration 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).
SURVIVORS PENSION
In our stylised scheme, partnered employees have a survivors 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 partners pension during participation. When
the member leaves the company or retires, this insurance is cancelled. The
calculation of the partners 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 orphans pen-
sion 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
offer all members the option between partners pension and supplemen-
tary retirement pension. However, this only applies to accrual-based and
not to risk-based partners pension. In our stylised pension scheme, at re-
tirement, single and partnered members can opt to exchange the accrued
partners 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 re-
turn on the invested contributions. In our calculations, we assume a nomi-
nal 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 4% on government bonds and 7% 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 in-
crease 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 %
Net real return on investment 2%
Collection costs % of the pension basis
Pay- out costs 1% of the pension benefit

A pension fund incurs costs for asset management and administrative ex-
penses. 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 (1% of the pension bene-
fit) are related to the administration of the pension plan for non-actives
(including pension payments), whereas the collection costs (% 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 pen-
sion contribution. The uniform contribution is first and foremost depend-
ent on the composition of the membership. As we shall see below, women
are more expensive than men for the retirement pension and cheaper for
survivors 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 break-
even 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 break-
even 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 ex-
pected return by investing with a higher or lower degree of risk. The
higher the risk, the higher the expected return and the lower the break-
even contribution. At the same time, periods with solvency surcharges
and indexation shortfalls are more likely, or even periods with contribu-
tion cuts and catch-up indexations (see the contribution by Hoevenaars
and Ponds).
8.3 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 de-
pends on the probability of death before retirement, the remaining life-
span of the member after retirement, and the life expectancy of a possible
partner. A second component of the actuarially fair contribution con-
cerns the premium for the survivors and disability insurance. This pre-
mium 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 representa-
tive members. We distinguish between men and women, single and part-
nered 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 calcula-
tions 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-
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 partners 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 sur-
vive 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 ca-
reer 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 rea-
son, 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 full-
time employees in five-year age brackets in 2004 from the CBS)
8.4 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 ex-
amine 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 employ-
ees 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 addi-
tion, 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 represen-
tative members a
Man Woman
Benefit Actuarially fair Benefit Actuarially fair
from 65 b contribution c from 65 b 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%
Partnered, entry age 25
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%)
Single, entry age 45
no career development 15 200 23.0% 15 200 27.9%
average career development 16 100 23.1% 15 700 27.9%
Partnered, entry age 45
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%)
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 partners pension for supplementary retirement pension. The actuarially fair
contribution is shown in bracket, if the partnered member does choose supplementary re-
tirement pension.
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 bene-
fit 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 start-
ing 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 part-
ner. They will therefore exchange the accrued partners 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, be-
cause 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 pen-
sion basis. This is 1%-point lower than for a partnered man. The partnered
man has a lot of benefit from the partners pension. After his death, after
all, his partner receives 70% of the retirement pension until her death, sur-
viving on average for almost 8 years. A partnered woman has much less
benefit from the survivors 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 ex-
ample, if he already has a satisfactory retirement pension. If this is not the
case, the woman can also choose to personally insure an individual partners
pension separately. The behavioural finance literature shows, however, that
people often take the default option if they are faced with a choice (see the
contribution by Van Els et al. in this volume). Furthermore, it is cumber-

some 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 part-
ners 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 survivors 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 partners
pension of 70% of the pension that the employee would eventually have
accrued.
The redistribution via both insurances usually only takes place in ret-
rospect. Everyone runs a risk of becoming disabled or of dying prema-
turely. Only after such an event has taken place the redistribution oc-
curs. This redistribution with insurance is also referred to as risk solidar-
ity 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 fig-
ure 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 con-
tribution-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 contri-
bution-free period becomes dominant.
The value of the partners and orphans 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 partners pension. The long benefit
payment duration of the partners pension then no longer weighs
against the fact that the rights to retirement pension lapse.
600
500
400
1000 euro
300
200
100
0 age
25 30 35 40 45 50 55 60 65
dismissal or resignation disability 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 partners 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 partners pension, pension funds are not allowed to distinguish be-
tween men and women. Therefore they assume an average value of the
partners pension in calculating the exchange percentage. Because of this,
the 15% higher retirement pension does not compensate the partners pen-
sion, 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 partners 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 partners 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 partners pension could not yet be exchanged for sup-
plementary retirement pension had a benefit of only 15.6% of the pen-
sion 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 partners pension.
25
20
% pension basis
15
10
Man - Man - now Woman - Woman -

previously previously now
Single Partnered
Fig. 3. Value of the scheme with and without exchange of partner pension (Join-
ing at the age of 25 with average career development)
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 pen-
sion contribution and annually have the same pension accrual. However,
the actuarially fair contribution for a years 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 pen-
sion fund can still make on the assets. The value of one years pension ac-
crual in our stylised scheme for a member aged 25 amounts to approxi-
mately 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
35
% pension basis
30
25
20
15
10
5
25 30 35 40 45 50 55 60
age
man woman
Fig. 4. Actuarially fair contribution for one year of membership for partnered em-
ployees (Entry at the age of 25 with average career development)
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 pro-
duces 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 aver-
age career development is 1.2%-points higher than for the same man with-
out 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 educa-
tion 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. Hri, 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 re-
tirement pension.
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. Em-
ployees 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 ex-
pensive 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 exam-
ple, 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 associa-
tion for instance, show that the costs for individual pension products are
high. Around the middle of 2006, the top 10 providers of a guaranteed
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 an-
nuity (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 contri-
bution 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 contri-
bution. 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 1%
with a life insurer instead of the % 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 prac-
tice, 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
Entry age
25 years 22.0 29.7
45 years 27.9 34.5

a in percentages of the pension basis.
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 disap-
pointing return on investments is not entirely carried by the current gen-
erations, for example, by omitting indexation or imposing higher contribu-
tions. It is partially shifted to future generations and vice versa. The possi-
bility 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 earn-
ings 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 dif-
fers widely.
Single people benefit less from the pension scheme than partnered
employees. The option that pension funds offer nowadays for accrued
partners 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 partners 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
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 con-
tribution with the pension fund against the contribution with a life insurer.
Due to mandatory participation, pension funds can benefit from econo-
mies 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 pen-
sions, 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 pen-
sion fund at age 25. For the pension fund and the insurer, the weighted average of
these members is shown.)
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 pen-
sion schemes, unpublished paper, 2006.
Hri, N., R. Koijen and Th. E. Nijman, The determinants of the moneys 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 schok-
bestendig pensioenstelsel verkenning van enkele beleidopties op pensioen-
gebied, 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 pen-
sion savings and low returns on the accrued pension capital. Mandatory retire-
ment saving prevents these problems. The Dutch mandatory participation works
well and there is no reason for drastic modifications. What can be studied, how-
ever, 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 pen-
sion 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 intro-
duced for private pension schemes. In these cases, various mandatory par-
ticipation 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. Re-
search 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 pub-
lic in section 9.4. Section 9.4.1 deals with their preferences concerning re-
tirement 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 pen-
sion 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 partici-
pate 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 mecha-
nism 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 Pro-

fessional 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 em-
ployees 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 manda-
tory participation, employers are obliged to withhold pension contribu-
tions 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 indi-
viduals: 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 partici-
pate in pension schemes, therefore comes about by means of collective con-
tracts 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 - here-
after: 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 con-
taining undesired competition on the basis of wage costs. Besides this eco-
nomic 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 pen-
sion 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.
datory participation in an industry-wide pension fund (Wet verplichte deel-

neming 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 employers pension scheme. The
latter occurs on the basis of gender (women), working part-time, a tempo-
rary 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 supplemen-
tary pensions. Ten years later, in 1996, this share had decreased to 9%
(Ministry of Social Affairs and Employment, 2002). Of these 9%, 7 percent-
age 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 un-
doubtedly related to the fact that a number of grounds for exclusion have
meanwhile been prohibited by law (exclusion of women, part-time work-
ers and employees with a temporary contract).
Self-employed
The governments 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 persons 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-em-
ployed 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 self-
employed people participating in industry-wide pension funds for the con-
struction industry (BPF Bouwnijverheid) and for painters, finishers, and glass in-
stallers (BPF Schilders).
the remainder are not compelled to save for a pension. They are expected
to supplement their state pension (AOW) through additional voluntary sav-
ings 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 statu-
tory mandatory pension schemes exist for which the law also sets the con-
tent or the level of the contribution. Two different versions of direct man-
datory participation can be identified.
In the first alternative, the direct mandatory participation is arranged
via employers: employers are legally obliged to arrange a collective pen-
sion scheme for their employees, in which legal minimum requirements
(as regards the level of the pension contribution or pension result) are pre-
scribed. Employees are then compelled to participate in the employers
pension scheme. In the second alternative, the direct mandatory participa-
tion 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 pri-
vate pension schemes. This includes the three different mandatory partici-
pation alternatives mentioned above: direct statutory mandatory participa-
tion via the employer, direct statutory mandatory participation via the
employee, and quasi-mandatory participation by means of collective con-
tracts. Of the total of 30 OECD countries, nine countries have mandatory
private pension schemes of the types mentioned above. The other coun-
tries 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 employers contri-
bution 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
Table 1. Mandatory funded pension schemes
Statutory mandatory Statutory mandatory Mandatory participation via

participation via employer participation via employee collective contracts (Quasi-
mandatory participation)
Countries 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 there-

fore cannot be described as a defined benefit (DB) or defined contribution (DC)
scheme, but is a hybrid scheme. In Iceland, employers are required to pro-
vide a pension scheme with a high DB content, whereas, on the other
hand, the Australian pension schemes are DC in nature. Incidentally, em-
ployees 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 par-
ticipation 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 ac-
count. Employees can select themselves with which private pension admin-
istrator 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-you-
go state pensions.
Also, in other Latin American countries, including OECD member Mex-
ico, 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 pen-
sions 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 pen-
sions that are provided via the employer. These supplementary funded
schemes are achieved by means of collective contracts between employers
(or employers organisations) and employees (or unions). The Swedish

model is most similar to the Dutch model. In Sweden, the collective agree-
ments 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 supple-
mentary funded pension schemes. To what extent does mandatory par-
ticipation 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 par-
ticipation level, the majority has a system of statutory mandatory participa-
tion or quasi-mandatory participation. The Icelandic statutory mandatory
120
100
80
% of employees
60
40
20
0
ay
nd
o
k
nd
lia
en
ia
l
s
nd
in
n
om
es
ga
ar
nd
ic
ad
pa
ak
a
ra
at
la
ed
a
la
la
ex
Sp
rtu
m
d
rla
Ja
el
an
or
ov
er
st
Po
Ire
St
ng
Sw
en
M
Ic
Po
N
Au
he
itz
Sl
Ki
d
D
te
Sw
et
ni
N
te
U
ni
U
Fig. 1. Participation levels in tax-facilitated pension schemes (Percentage of the num-

ber 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).
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 suffi-
ciently 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 con-
tract 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, volun-
tary pension savings schemes (DC schemes) have largely replaced tradi-
tional 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
Sundn, 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 em-
ployees left the state pension or the employer pension scheme in the pe-
riod 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 par-
ticipating 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 Mex-
ico, 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.
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 par-
ticipation of the self-employed in pension schemes. Also, in countries
where mandatory participation has been implemented via the employ-
ees, 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 re-
sult of which fewer employees accrue a supplementary pension. This is the
reason that in several countries discussions are currently going on con-
cerning the possibility of introducing mandatory participation (Ireland,
UK), or concerning ways to provide more guidance to participants in pen-
sion schemes (US). This type of guidance (libertarian paternalism) can be
achieved by making participation in the employers pension plan auto-
matic, unless the employee explicitly requests otherwise. More generally,
this raises the issue of whether autonomy leads to lower or even inade-
quate 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 invest-
ment 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 them-
selves are responsible for the planning and adequacy of their pension sav-
ings 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-
periences in the US. The recent behavioural finance literature on households

financial planning and decision making shows that individuals have diffi-
culty with making consistent and prudent choices regarding saving and
investing. Underlying forces are inertia, procrastination, inconsistent sav-
ing 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 ex-
pression of a self-control problem. The theoretical basis of this has been
elaborated in more detail by Thaler and Shefrin (1981). Inertia and procras-
tination frequently lead to insufficient saving. It is illustrative that taking
part in saving for retirement strongly depends on the system used: is par-
ticipation 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 contribu-
tion and the allocation of the investments. Btler 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 eco-
nomic 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).
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 invest-
ment results. Benartzi and Thaler relate this behaviour to myopic loss aver-
sion, loss aversion combined with short-sightedness. For this reason, deci-
sions made are not always consistent with peoples actual individual pref-
erences. 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 (Mun-
nell and Sundn, 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 em-
ployers, 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 de-
vices, employers are now encouraged in the new Pension Protection Act to
choose a more guiding approach whereby, through offering standard
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 impor-
tance 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 posi-
tive relationship between financial education on the one hand and finan-
cial 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 tradi-
tionally 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 under-
standing of economic concepts such as equity risk, inflation and com-
pound interest.8 Moreover, it is found that people with more understand-
ing 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 con-
cluded from this that a higher level of financial literacy leads to better fi-
nancial planning of household finances and pension savings. Thus, finan-
cial 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 litera-
7 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 Sundn, 2006).
8 In particular, this applied to women, ethnic minorities and less educated people.
ture and experiences in the US particularly, question marks about the ca-
pacity 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 dis-
cussed here can be split into two questions of interest:
1. Do individuals prefer (more) autonomy in the pension domain (sec-
tion 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 individu-
als 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 approxi-
mately 1,500 households are questioned about their financial characteris-
tics 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).
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% pro-

nounced 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 sav-
ings 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 regard-
ing their pension, 43% say they dont worry about it now (well 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.
Fig. 3. Propositions concerning private pension provision (Percentage of the num-

ber of respondents)
Source: DNB Household Survey, 2003 and 2005.
pension being well organised, but dont want to know the details. The
changes compared to 2003 are minimal.
There has also been very little change in the publics preferences con-
cerning 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 con-
tribution 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 manage-
ment 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 in-
creased by 4 percentage points.
Van Rooij, Kool and Prast (2007) have also polled the preference for in-
dividual 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 contribu-
tions. Nevertheless, around half of the respondents would rather leave the
investment to a pension fund. More than a quarter opted for more influ-
ence on the investments, the rest were indifferent (approximately 10%) or
didnt know (15%). Employees who attribute good financial skills to them-
selves 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 pen-
sion that matches their own preferences better.
Pension preferences of the self-employed

The respondents underlying figure 3 also include approximately 75 self-
employed. 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 self-
13 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.
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). Self-
employed people have a stronger preference for a pension without guar-
antees against a lower contribution, although in absolute terms the num-
ber of self-employed with a stated preference for a certain pension against
a higher contribution is still larger. In addition, the self-employed com-
paratively 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 self-
employed has no pension scheme arrangement at all plays a role in this.
All in all, these findings of recent survey research indicate that a major-
ity of Dutch employees can happily live with the mandatory nature of
pension savings, appreciate the security of a DB system, and have no pref-
erence 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 Ill see to that when I come to that. More-
over, 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.
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 No, Ill see to that when No, I can fairly easily Dont know/no opinion
extra money myself for I come to that manage with my pension
my pension
2003 25 36 18 21
2005 29 33 15 23
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 in-
vestment 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 calcula-
tion of the pension benefits was based on the current income from em-
ployment, forty years contributions paid in, and realistic assumptions on
the level and volatility of real stock and bond returns. Without the respon-
dents 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 unattrac-
tive) to 5 (very attractive). Those people who initially had chosen a conser-
vative 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 cho-
sen 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 in-
different or chose the less risky portfolio. Thus the choice for the percent-
age of stocks in the investment portfolio in this experiment does not corre-
spond 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 auton-
omy that emerges prominently in the international literature is financial
illiteracy. Research into the Dutch situation shows that peoples knowl-
edge 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 citi-
zens 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 re-
peated 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 in-
dexation 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.
Table 3. Knowledge of own pension arrangements, 2003 and 2005 results (Percent-
age of respondents)
Proportion of respondents who know about:

Age (in years)
Type of pension scheme Indexation Accrued pension rights Level of pension
2003 2005 2003 2005 2003 2005 2003 2005

16-24 30 22 22 11 0 0 19 12
25-34 38 48 35 30 22 21 27 39
35-44 52 53 53 42 31 29 35 41
45-54 63 70 59 53 47 38 41 46
55-64 73 78 64 67 66 64 48 54
65 and older 81 84 81 81 n/a n/a n/a n/a
Total 60 64 56 53 40 36 37 43
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 be-
come more important in the future. Van Rooij, Lusardi and Alessie (2007)
conclude, however, that the general financial knowledge of the Dutch pub-
lic is limited. They derive this from the answers to five simple questions
concerning interest rates and inflation and eight questions concerning invest-
ing for testing the most basic financial knowledge and skills. No more than
40% of the respondents gave the correct answer to all the questions con-
cerning 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 ac-
counts. They also did not know that investing in the stock of a single com-
pany provides a less certain return compared to a mutual fund. A more
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 percep-
tion of economic growth and inflation contributes to the quality of the deci-
sions 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 consid-
ered 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/pre-
pensioen en introductie levensloopregeling VPL Act), many employees are
confronted with a retrenchment of the scheme for early retirement appli-
cable up until then (there is a transitional scheme for employees born be-
fore 1950). At the same time, the life-course savings scheme has been in-
troduced as an attractive possibility for saving individually in a tax-
friendly manner, in order to fund the intended early retirement. Never-
theless, 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 repre-
sentative 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 em-
ployer plays a large role, certainly when it is considered that a third of the
life-course participants receive no employers contribution. It should be
noted that the employers 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 partici-
16 Saving for a sabbatical (15%) or parental leave (10%) are the other most impor-
tant reasons.
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 rea-
son is the forced choice between either participation in the another tax-
friendly contractual savings scheme (Spaarloonregeling) or in the life-
course 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 can-
not easily be missed. The broad range of reasons of whether or not to par-
ticipate 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 litera-
ture 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 self-
employed to the question of whether they consider the level of their pen-
sion accrual (state pension (AOW) combined with a company pension)
adequate, or have made additional provisions. The majority of the self-
employed, 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 self-
employed 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
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 wouldnt save 11
I expect I will more easily receive permission for leave from my employer 6
Other reasons 6
Dont 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 schemes continued existence 10
Not considered it / forgot / too much trouble / didnt know it existed / didnt know how 10
Isnt necessary 8
Ill wait and see 7
Other reasons 2
Dont 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 dont 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 supplemen-
tary measures, mostly by means of annuities and single-premium policies.
More than half of the self-employed who have not arranged supplemen-
tary measures, however, state that they actually should save more for their
pension. These results indicate that people who are not under the manda-
tory 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.
Summarising the above, on the basis of empirical research, it can be ob-

served 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. Fi-
nally, research has shown that the financial knowledge of the Dutch leaves
a lot to be desired. This concerns not only the knowledge of pension provi-
sions, 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 re-
tirement savings, but a system of quasi-mandatory participation via collec-
tive contracts. Because of this, in practice, more than 90% of Dutch em-
ployees automatically save for their pension. Experiences in other coun-
tries 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 inade-
quate pension. The empirical research performed for the Netherlands
shows that there is no reason to assume that Dutch employees would ac-
tually behave differently.
The knowledge of basic financial concepts is limited and investment ex-
pertise and skills are poor. Moreover, adequate knowledge of own retire-
ment 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.
Can we conclude from all this that the current Dutch mandatory par-
ticipation for employees by means of collective contracts and the elabora-
tion 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 (quasi-
mandatory 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 im-
portant pension decisions out of peoples 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 participa-
tion 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 re-
search 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 improve-
ment (certainly if the aim is also to include the self-employed). A small
step towards more freedom of choice would possibly open the opportu-
nity 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.
Literature
Ameriks, J., A. Caplin and J. Leahy, Wealth Accumulation and the Pro-
pensity 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 Depart-
ment 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.
Btler, 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, Econo-
misch 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 Per-
formance in Britain in the 1980s and 1990s, NBER Working Paper 9556,
2003.
Engelhardt, G., and A. Kumar, Employer Matching and 401(k) Saving: Evi-
dence 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.
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 Pen-
sioenen 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: deter-
minants 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. Sundn, 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 expen-
diture, OECD, Employment and Migration Working Papers 23, 2005.
Prast, H.M., Psychology in Financial Markets: an Introduction to Behav-
ioural Finance, Financial Monetary Studies, Amsterdam: NIBE-SVV
publishers, 2004.
Prast, H.M., M.C.J. van Rooij and C.J.M. Kool, Werknemer kan n 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.
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 behav-
iour 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 Eco-
nomics 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 indus-
trial 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 advan-
tages and disadvantages. A comparison with other countries shows that the
mandatory participation for companies succeeds well in its objectives, as it en-
courages 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 in-
surance be publicly or privately implemented? In the Netherlands, em-
ployers 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 ex-
amples 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 industry-
wide 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 Nether-
lands. 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 participa-
tion and a jointly redistribution of the contributions. Other industrial en-
terprises 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 Coperatief Verzekeringsfonds (coopera-
tive insurance fund - 1917), the Algemeen Mijnwerkerspensioenfonds (general
mineworkers pension fund - 1918), two regional pension funds in the ci-
gar 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 govern-
ing 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 farm-
ing sector, with lots of small companies and relatively low incomes, is al-
ways difficult to regulate on a voluntary basis. The solidarity between em-
ployers was an explicit justification for mandatory participation. In addi-
tion, mandatory participation prevents competition on payroll costs or
conditions of employment within an industry. The mandatory participa-
tion also limits the blank spots (the number of employees without pension
facilities), because all employers in a business sector are compelled to pro-
vide 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 man-
datory 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 pen-
sion 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 supple-
mentary 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 contribu-
tions 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. Fur-
thermore, 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 pre-
vious chapter.
Equally important is the mandatory participation for companies: a pen-
sion 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 man-
datory participation of professionals like GPs, notaries and physiothera-
pists 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
Number Active Assets

of funds members (billions)
Mandatory participation for

Company pension fund 627 900 000 169.9
individuals
Mandatory participation for

Industry-wide pension fund 98 5 320 000 365.2
companies and individuals
Act on professional pension

Professional pension fund 15 48 000 19.0
funds
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 consid-
erably and the funds have also become major players on the capital mar-
kets. 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.
10.2 Why mandatory participation?

The most important reason for mandatory participation is that it dramati-
cally 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, eve-
ryone 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 participa-
tion is mandatory. Individuals are hardly capable of assessing their pen-
sion 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, il-
lustrates this problem. The costs of the private product were much too
high, as were the promised returns. Aggressive selling techniques led mil-
lions 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 pen-
sion funds. People with higher incomes tend to live longer and should
insure the most. Thus, those with the worst risks should insure them-
selves. Pension funds and insurers can and should not make any distinc-
tion 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 ex-
traordinarily 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 partici-
pation helps to mitigate this impact. After all, people must always be in-
sured 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 be-
come 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 ex-
ample, 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 be-
tween 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 impor-
tant 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 mem-
bers 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.
However, there are also risks that are not absorbed by an industry-
wide 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 pay-
ing 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 surren-
dered to the younger fund. If the returns fall short of expectations, an
increase of contribution with the younger fund ensures an extra contri-
bution 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 custom-
ise 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, stagna-
tion of the national economy or even war, still exist. In theory, Pan- Euro-
pean 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 sub-
ject to the free market, they have theoretically less incentives to efficient
operation. Yet in practice it is found that these funds do better than insur-
ers 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 in-
creasingly 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 com-
plete 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 com-

pulsory element is also used as an argument against the mandatory par-
ticipation, 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 perform-
ance 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 competi-
tors 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 Euro-
pean competition rules. Agreements that social partners have entered into
in the context of collective negotiations and that are meant for improve-
ment 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 man-
datory participation for individuals. Mandatory solidarity in a single
scheme leaves little room for individual freedom of choice. From a theo-
retic perspective, the pension accrual can also be too high. If someone is
single, moreover, the survivors 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, be-
cause 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 survivors pension for a higher old-age pension, must be offered
by a pension fund. Funds also offer to separate the cover for survivors
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.
10.5 Expanding the opportunities for exemption

In addition to voluntary exemption, the original Act on professional pen-
sion funds from 1949 provides one possible exemption from the industry-
wide 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 Rep-
resentatives 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 con-
sidered 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 Hooger-
vorst (VVD) posed the question are no alternatives conceivable, with which a
touch of the mandatory participation is settled, but that perhaps in another man-
ner 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 companys own
pension scheme, except when this has been agreed with collective labour
agreement (CAO), should be at least equivalent to the scheme of the indus-
try-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 enter-
prises with good risks taking their pension provisions elsewhere and plac-
ing 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 funds 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 adjust-
ing 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.
10.6 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 participa-
tion for companies? A paper concerning alternatives for the mandatory par-
ticipation 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 administra-
tor, as in Chile;
3. enforcement of the current mandatory participation to a maximum
salary, above which employees have the freedom to arrange a sup-
plement;
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. Em-
ployers 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 in-
crease. In all examples, solidarity will be reduced or disappear unless sup-
plementary measures are taken for this, although the freedom of choice for
the individual does indeed increase. The then coalition government there-
fore 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 sup-
plementary pensions. (House of Representatives of the Dutch Parliament,
2000-2001a). This actually returned to earlier suggestions concerning re-
striction 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 ar-
rangements. 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 in-
tended 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 re-
gime, 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
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 be-
tween the individual accounts: then there would be explicit cross-
subsidising 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 survivors 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 univer-
sal 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 pen-
sion 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 states 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 in-
stance 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 ma-
jority of schemes have been set up with a view to breadwinners who work
fulltime, the current population is often more varied. Part-time work, di-
vorces, remarriages and greater mobility on the labour market make a uni-
form scheme increasingly constrictive. For this reason the legislator in-
creased the number of exchange rights. Section 2b in the Pension and Sav-
ings Funds Act has provided the right to exchange survivors pension for
old-age pension since 2001. Section 55, an implementation of the Omtzigt-
Depla motion, provided the right to the reverse of this, exchanging old-age
pension for survivors pension. The first is very important for single peo-
ple, the second can be very valuable in case of remarrying or when the
pension scheme does not provide for survivors pensions.
Pension funds and social partners, however, do not need to wait for leg-
islation 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 com-
panies became effective and this contributed greatly to a supplementary
pension provision being arranged in the Netherlands for almost all em-
ployees. Also, without there necessarily being an empathetic solidarity
with an industrial sector, the economies of scale of a large collective, cou-
pled to mandatory risk sharing within an industrial sector, provide bene-
fits 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. Be-
cause these savings are outside the immediate sphere of government in-
tervention, creation of a funded scheme in the Netherlands has been a suc-
cess.
Literature
Bakker, R., E. Haaksma, T. Kool, K. Verhagen and C. de Wijs, Toezien of
toekijken?, verzekeringskamer 75 jaar, Apeldoorn: Stichting de Verzeke-
ringskamer, 1998.
Copeland, C., Employment-Based Retirement Plan Participation: Geo-
graphic 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 ver-
plichtstelling, 25 014 no. 1, 1996-1997a.
House of Representatives of the Dutch Parliament, Nota Werken aan zeker-
heid, 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 deel-
neming 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 Ver-
zekeringsarchief, 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 particu-
larly formed by a distortion of the labour market. Besides risk sharing, intergen-
erational 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 soli-
darity, 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 macroeco-
nomic side effects. It cannot be excluded in advance that the side effects
dominate and, on balance, solidarity does more harm than good. More-
over, 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 soli-
1 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 solidar-
ity.2 First of all there is solidarity which comes into operation after a per-
son for whatever reason suffers a loss and other people entirely or par-
tially compensate him or her for that. This form of solidarity is the basic
principle of insurance (e.g. fire insurance). Since this solidarity is condi-
tional 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 partici-
pants. 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 supplemen-
tary pension schemes (see Kun 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 be-
2 This distinction is based on De Laat et al. (2000).

3 De Laat et al. (2000) also distinguish a third type of solidarity: income solidar-
ity. This occurs, among other things, with the General Old Age Pensions Act
(state pension, AOW). The AOW benefit is independent of income, but the con-
tributions (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 gen-
erations, 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 uncer-
tain 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 pen-
sion 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.
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 shar-
ing between older, younger and future generations replaces a pension
funds explicit transactions on the capital market. In other words, the pen-
sion 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 sys-
tem of supplementary pensions. The articles of association and regulations
of pension funds can or do lead to intergenerational transfers. Moreover,
the governments legislation and regulations also play an important role.
In the second pillar, we can distinguish three sources of intergenera-
tional 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 contribu-
tion rates on the basis of the uniform contribution and accrual system. Be-
4 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 time-
weighted proportional pension accrual.
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 ob-
tain. 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 consider-
able tax allowance to retirement saving. Responsible for this are the lower
rate for people over 65, the progression in the tax system, and the exemp-
tion of the assets of pension funds for the capital yield tax. Whether these
subsidies imply redistribution, and in what form, is not clear. This de-
pends entirely on how the government finances these subsidies. If the fi-
nancing is a burden on the same generations as those who receive the sub-
sidy, 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 peo-
ple 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 gen-
erations without the mandatory participation. This would also mean that
the possibility of having deficits partly or wholly paid for by future gen-
erations 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 sys-
tem, 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.
the risk of living longer. Moreover, if the pension has been set up as wel-
fare-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 dtre of a pension fund.
With properly operating capital markets, there is very little to argue in fa-
vour 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, wage-
indexed 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 indus-
try-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 disadvan-
tages 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 con-
tribution of Bikker and De Dreu in this volume for a comparison of the
implementation costs of pension providers). Note, however, that this bene-
fit 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 an-
nuity insurance policies on the market in the US.
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 per-
forms worse than a standard portfolio over a period of five years, an em-
ployer 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 con-
tributions, 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 contribu-
tion 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 contribu-
tions, 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. Neverthe-
less, stopping working actually is a realistic option for people with a
relatively low income or a relatively high reservation wage. For old peo-
ple 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.
contribution can be a reason to retire early. Instead of no longer partici-

pating on the labour market, a member can also decide to seek employ-
ment 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 in-
creasing contributions, labour market distortions are also reduced. Indexa-
tion cuts always affect the rights already accrued. The loss because of this
cannot be reduced by working fewer hours. If it is expected that indexa-
tion cuts would be applied in future years, however, this can have a la-
bour 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 la-
bour 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 excep-
tional 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 mar-
ket elasticity and age.
Lastly, the favourable fiscal treatment of supplementary pensions en-
courages labour force participation. Although the subsidy is mainly in-
tended to encourage retirement saving, because of the mandatory partici-
pation, it is a subsidy on labour. This subsidy does not need to exist, effec-
tively. 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.
DISCONTINUITY RISKS
When contribution payers or sponsoring companies withdraw, the conti-
nuity of the pension scheme is threatened. This discontinuity risk mainly
occurs, as we will see below, when a fund is underfunded. The discontinu-
ity risk is higher the more intergenerational risk sharing is organised by a
larger number of small pension funds. For the larger the number of pen-
sion 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 re-
moved. 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 policy-
makers 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 pen-
sion capital for other purposes.9 This is an argument for not allowing int-
ergenerational risk sharing to be managed by the government, but by lo-
cally organised pension funds that have been placed at a distance from the
government. But discussions can also arise within pension funds, concern-
ing 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).
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 uni-
form contribution rate with an actuarially fair contribution rate. In the case
of an actuarially fair contribution rate, each member pays an annual con-
tribution that is equal to the cash value of the pension right. This actuari-
ally 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 condi-
tionally 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 quan-
titatively they are not similar because of differences including probabilities of
death, accrual percentage and membership. Furthermore, Boeijen et al. omit the
survivors pension and constantly take the average contribution instead of the
marginal contribution as used by us.
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 pen-
sion right. This also includes an allowance for the survivors 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, in-
creases not only because of the nominal wage growth, but also be-
cause 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 participa-
tion 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 employers part of the pension contribu-
tion is passed on to the employee.12
Figure 1 shows the uniform contribution rate and the actuarially fair con-
tribution rate, both expressed as the percentage of the contribution base
(marginal) and of the gross wage (average). Because of the constant popu-
lation, 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.
%
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 differ-
ence between the uniform contribution rate and the actuarially fair contri-
bution rate is large for the young and old members. For a 20-year-old en-
trant, 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 pay-
ment. 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 per-
centage 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.
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 members 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
10 actuarially fair contribution
uniform contribution
5
actuarially fair uniform contribution
0
20 24 28 32 36 40 44 48 52 56 60 64
age
Fig. 2. Life-cycle redistribution and intergenerational redistribution

arially fair contribution rate as seen so far. The difference between the ac-
tuarially 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 accu-
rately attribute the intergenerational redistribution of the uniform contri-
bution rate. A deviation between the uniform contribution rate and the
actuarially fair uniform contribution rate implies that the uniform contri-
bution rate is not only redistributive over the individual career, but also
between generations. More specifically, the difference between the actu-
arially fair uniform contribution rate and the actuarially fair contribution
rate is the transfer over the members 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% trans-
ferred 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 inter-
generational transfer amounts to 2% for all ages. In this regard, our analy-
sis 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 con-
tribution rate. This means that a member on a contract with a uniform con-
tribution 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 mem-
ber, the balance of all benefits received and contributions paid, is negative.
These extra contributions correspond to approximately 2% of all contribu-
tions 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 uni-
form contribution rate favourable for them, and did not need to make ad-
ditional payments during their young years. The account for this subsidy
was, probably unconsciously, left for the future generations.
x 1000
9
-1
-3
100 90 80 70 60 50 40 30 20 10 0 10 20 30 40 50 60 70 80 90 100 110 120
age of current generations 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 uni-
form 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 introduc-
tion of the uniform contribution in itself was a zero-sum game. This con-
clusion is closely in line with the observation of Sinn (2000) that each pen-
sion 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.
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 illus-
trating 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 contri-
bution has been charged, and the pension rights are fully indexed to the
nominal wage growth. We assume a conditionally index-linked average-
salary scheme. That means that the pension fund has two instruments to
work off funding ratio deficits: charging an increase on the break-even con-
tribution (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 As-
sessment 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
-12 DC scheme
-14 difference
-16
100 90 80 70 60 50 40 30 20 10 0 10 20 30 40 50 60 70 80 90 100 110 120
age current generations future generations
Fig. 4. Change in net benefit after an asset shock of 15%

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 as-
set 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. Mem-
bers 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 car-
ried 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 con-
tributions 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 gen-
erations (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 dtre of pension funds. Ultimately, the question is whether eve-
ryone is better off because of it. In other words: what welfare gain does it
achieve? Calculations of the benefits of intergenerational risk sharing are
very scarce in the literature, however. Moreover, the methods of approach

are so different that they make mutual comparisons difficult. For this rea-
son 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 sys-
tem (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 disre-
garded in this.
Teulings and De Vries (2006) approach the problem of intergenerational
risk sharing from optimal investment theory. Intergenerational risk shar-
ing is nothing more than the optimal insurance for risks on financial mar-
kets over generations. A generation that wants to take more risks, can in-
clude 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 portfo-
lio 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 ac-
count 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.
Gollier (2007) compares the situation of an individual DC scheme with a

collective DC scheme in which all current generations transfer their pen-
sion 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 capi-
tal. 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 contri-
butions paid in is guaranteed. Under this supplementary condition, a wel-
fare 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 collec-
tive DB scheme in which the benefits are defined and shocks are absorbed
by means of variable contributions. In a collective DC scheme the contribu-
tions 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 pe-
riod 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. Participa-
tion 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 addi-
tion, they are uncertain concerning the aggregated life expectancy. Fur-
thermore, 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 collec-
tive DB scheme. The result is an added value of DB schemes of approxi-
mately 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.
THE COSTS OF INTERGENERATIONAL RISK SHARING

As discussed previously, the repayment of underfundings, uniform con-
tribution rates and favourable fiscal treatment distorts the labour market.
Welfare losses that correspond with the labour market distortions identi-
fied above are hard to quantify exactly. The order of magnitude can be
assessed, however. We confine ourselves here to the case in which the la-
bour market elasticity is independent of the age of the workers and in
which the implicit government subsidies to supplementary pensions trans-
late into higher tax rates on labour, so that there is no question of subsidis-
ing. 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 cer-
tainty 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 elastic-
ity 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% in-
stead of the 0.8% that results with shocks of 30% of the wage total.
11.8 Concluding observations

The Dutch supplementary pensions have been impregnated with inter-
generational solidarity: because of the uniform contribution rate, the gov-
ernment subsidies to pensions and the way in which pension funds fi-
nance 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 pro-
motes 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 labo-
rious 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 Econo-
mic Policy Analysis), 2000.
The Economist, Pension Pork, The Economist, 15 April 2004.
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 Sys-
tem 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
In this volume a large number of renowned researchers and policy-makers dis-

cussed a broad range of topics. Together they offer as complete a picture as pos-
sible of the costs and benefits of collective pension systems. The concept of soli-
darity 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 re-
sponsible for the benefits. Where possible, these aspects are quantitatively sub-
stantiated.
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 trans-
fers are the two manifestations of solidarity. The authors discuss several
concepts of solidarity, the costs and benefits of solidarity and the macro-
economic 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-
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 pen-
sion 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 Kun provides an extensive summary of solidarity within
collective pension schemes. The most important distinction is that be-
tween, 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 dont know
in advance whether or not your house may burn down, but should it hap-
pen, 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 par-
ticipant benefits and the other one pays. Thus, by application of the uni-
form contribution rate or with final-salary schemes, value is transferred
e.g. from participants with a flat career pattern to the career makers. Par-
ticipants 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 indi-
vidual institutions with little mutual overlap, but in both systems risk
and uncertainty play an important role. Collective pensions reduce peo-
ples 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 pro-
visions 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 discus-
sions, for example, on the question of whether people who live an un-
healthy 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 be-
tween 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 implementa-
tion 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 sub-
stantially increase efficiency. When the comparison is made between col-
lective 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 in-
surance 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 Hoog-
dalem and Theo Nijman discuss the question of what an optimum pen-
sion contract theoretically implies from the participants 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 opti-
mum and what obstacles individuals encounter if they would want to im-
plement 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 as-
sumptions. 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
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 deci-
sion-making.
In chapter 7, Dick Boeijen, Corn Jansen, Niels Kortleve and Jan Tame-
rus go deeper into the intergenerational debate and investigate the uni-
form 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 recov-
ered later. On the other hand, however, it becomes an issue for a partici-
pant 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
lead to younger employees becoming cheaper and older employees rela-

tively expensive, as a result of which old people will encounter more diffi-
culty participating on the labour market. The authors indicate that adjust-
ment 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 redistri-
bution within collective pension funds on the basis of representative par-
ticipants. 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 partners 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 transi-
tion from final pay to average wage schemes.
Life insurers, however, when setting their premiums, can make distinc-
tions between characteristics of the participants. The costs of private insur-
ance 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 pos-
sible. 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 concern-
ing 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 con-
firm this picture. For example, it appears that the average Dutch persons
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
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 be-
tween 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 le-
gitimate 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 pen-
sion 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 dis-
courages labour participation at a young age, but it promotes a late retire-
ment. 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
gains and losses on financial markets are distributed among all genera-
tions, the collective is prepared for and able to accept more risks. The uni-
formity of the collective pension product, moreover, enables funds to pro-
vide 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. Be-
sides this, a collective scheme contains forms of subsidising solidarity, as a
result of which not all participants profit to the same extent from the bene-
fits 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 pen-
sions be reduced, without putting the benefits in danger? To this end we
discuss a number of dilemmas that sooner or later confront collective pen-
sion 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 re-
solved, 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 sys-
tem 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
retirement via a collective scheme. Although the investment mix may not
be optimal, in view of the higher costs that accompany investment free-
dom, 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 criti-
cally considered against the higher implementation costs, to avoid the risk
of throwing out the baby with the bath water. This concerns both the ad-
ministrative 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 average-
salary systems. There is also less subsidising from single persons to part-
nered participants now that many schemes allow the partners 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 ac-
crue 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 con-
tribution and the uniform pension accrual. This means in practice that
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 contribu-
tion for a 64-year-old person would become four times as high as for a 25-
year-old. In that case there would be less solidarity between generations.
If, on the other hand, an equal contribution is selected, but with age-
related pension accrual, an extra years work at the end of the career pro-
vides 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 Nether-
lands Authority for the Financial Markets (AFM) has been appointed as
the second supervisory body, specifically charged with assessing the com-
munication 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 pro-
vided 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 Depart-
ment of ABP Pension Fund, focusing on pension policy and risk manage-
ment. 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, pen-
sions, financial market competition, financial institution efficiency, regula-
tory 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 Eras-

mus 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 hous-
ing 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 Eco-
nomic 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 Uni-

versity 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 Direc-
tor 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 Supervi-
sory Policy Division of DNB (Dutch Central Bank). His research interests
include pension fund investment policy, lending technologies of micro-
credit institutions and the interaction between deposit insurance and mar-
ket 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 Univer-
sity of Tilburg and has worked since 1986 in various positions within the
research environment of DNB. In 1995 he obtained his PhD at the Univer-
sity 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 Opera-
tions 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. (Corn) 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 Coun-
cil for Public Health and Care. His work centres on financial-strategic is-
sues 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 Rot-
terdam, specialising in public finance.
C.E. (Niels) Kortleve (1965) studied Econometrics and is currently Actuar-

ial Projects & Special Accounts Manager at PGGM.
B.J. (Barthold) Kuipers (1972) works at the Financial & Risk Policy De-
partment of ABP Pension Fund. He focuses on the policy, strategy and risk
management of the pension fund. Previously he worked at CPB Nether-
lands Bureau for Economic Policy Analysis. He graduated in Econometrics
at the University of Amsterdam.
J.B. (Jan) Kun (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. Kun 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 method-
ology was long his main task. Since 2003 he has again been attached to the
University of Amsterdam as part-time Professor in Pension Actuarial Sci-
ence.
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 Fi-
nance 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 af-
filiation, 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 Am-
sterdam. 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 at-
tached to the Faculty of Economic Science and Econometrics of the Univer-
sity 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, succes-
sively, the Actuarial Department, Asset Research, and currently the Finan-
cial 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:
eduard.ponds@abp.nl
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, spe-
cialising 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 Advi-
sory 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 Econom-
ics 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 Mod-
ern 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 Depart-
ment 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 actu-
ary 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 Nether-
lands 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, dis-
ability and the labour market. In addition, he lectured for many years in
General Economics at the University of Amsterdam.
Subject index
A Collective Defined Contribution

(CDC) system 58, 107, 113
access to healthcare 46
116, 211, 223, 224
adverse selection 37, 39, 67, 69,
company pension fund 51, 54
71, 82, 85, 153, 210
58, 60, 73, 161, 188, 189, 192,
annuity 63, 64, 67, 69, 70, 74, 76, 199, 200, 212, 231
83 85, 153, 163, 168, 181, 191, contribution rate 25, 27, 30, 95,
210 96, 104, 108, 111 113, 115, 119,
Asset Liability Management 96, 121, 129, 137, 145, 146, 153,
100, 102, 104, 106 108, 112, 209, 212, 214, 217 220, 225
114, 115, 239, 240, 242, 243 actuarially fair 138, 143 146,
average-wage system 26, 31, 121, 150, 151, 155, 214, 216 218
215, 220 222, 233, 236 contribution volatility 75, 88 90
cost differences 52 54, 57, 61,
B 62, 71, 231
behavioural finance 146, 160, costs
168, 182 administration 53 61, 64 71
blank spots 162, 187, 188, 234 operation 51 53, 55, 57 61,
71, 152, 231
C coverage rate 183
catch-up contributions 22, 84

D
87, 90, 120, 121, 212, 215, 220,
222, 225 decreasing accrual system 119,
catch-up indexation 21, 97, 142 129 136
cold solidarity 4, 147 Defined Benefit (DB) system 2,
52, 56 59, 66, 84, 90, 107, 111
collective contracts 52, 63, 65
113, 164 167, 175, 207, 210,
67, 70, 89, 91, 161, 163 165,
223, 224
182, 183
244 Subject index
Defined Contribution (DC) financial illiteracy 170, 177

system 2, 26, 34, 52, 56 59, financial planning 160, 168, 170
77, 90, 113, 128, 164, 166, 167, first pillar pension 1, 14, 16, 19,
169, 174, 176, 207, 221 224 30, 37, 38, 82, 139, 162, 164,
deflators 95, 99 166, 180, 193, 206
direct schemes 62, 66 flexible retirement age 90
disability 7, 31, 91, 138, 141, 143, framing 169
144, 146, 147, 192, 194, 200, 244 frayed-edge theory 135
discontinuity risk 213 funded pension system 16, 33,
discount rate 21, 30, 32, 87, 90, 45, 122, 159, 164 166, 199, 220,
98 100, 102, 108, 214, 215 232
stochastic (deflators) 99, 100,
102 G
distribution mechanism 16
generational accounting 20, 96
100, 102 107, 109, 110, 112,
E 115
economies of scale 2, 36, 45, 46,
52, 54, 55, 58, 60, 61, 63, 70, 71, H
138, 152, 155, 193, 199, 200,
heterogeneity 91, 211
211, 231, 236
horizontal solidarity 13
employer matching
contributions 169 Household Survey 171, 173, 176,
178
endowment insurance policies
63, 64, 69, 70, 74 human capital 76, 79, 81 83, 85,
86, 89, 90, 235
equivalence principle 37
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, incomplete capital markets 80,
230 84, 210
indexation 13, 21 23, 30, 66, 67,
F 78, 83, 86, 88, 91, 95 106, 108
116, 121, 139 142, 154, 177,
fair value 78, 90, 215
192, 193, 207, 208, 212, 215,
final-wage system 32, 66, 107, 220, 222, 225
112, 139, 140, 152, 230
conditional 97, 115, 121
Financial Assessment Framework unconditional 107, 112 114
(FTK) 22, 68, 196, 221
Subject index 245
indexation policy 21, 22, 95, 104 life expectancy 18, 31, 45, 67,
individual arrangements 13, 32, 137, 143, 146, 151, 153, 154,
236 191, 199, 210, 224, 236, 237
individual pension scheme 52, life insurance policy 63
63, 71, 159, 163, 165 life-course savings scheme 160,
individualisation 42, 46 179 181
industry-wide pension fund 30, longevity bond 83 85
51, 54, 56 62, 71, 73, 95 97, longevity risk 66, 68, 83, 85, 89,
104, 115, 116, 119, 121, 139, 91, 199
161, 162, 188 190, 192, 193,
195, 199, 200, 211, 212, 231 M
inertia 168
mandatory participation 8, 28,
informal care 36
30, 66, 67, 70, 85, 152, 155,
insurance company 45, 60, 62, 159 167, 171, 172, 180 183,
63, 66, 74, 80, 84, 138, 153, 155, 187 195, 197, 199, 200, 209,
166, 190, 199, 231 211, 213, 223, 230, 233 236
intergenerational risk sharing companies 160, 161, 187, 189,
96, 154, 206, 208, 210, 211, 213, 197, 199, 200, 234
214, 222 224, 226 indirect 163
intergenerational risk trading 85 individuals 160, 161, 189, 190,
intergenerational solidarity 19, 194
20, 26, 38, 39, 41, 43 45, 102, mental accounting 168
120, 121, 126, 192, 205 211,
misselling scandal 166, 191
213, 214, 225
moral hazard 18, 32, 37, 82
investment policy 7, 78 81, 107,
mutual subsidies 124, 126 128,
109, 116, 167, 223, 231, 236, 240
135
myopia 168, 169, 182
J
John Rawls 35 N
net benefit 219 222
L
no-claim rebate 34, 38, 40, 43, 46
labour market distortions 59, 90,
205, 206, 211, 212, 224 226
O
labour mobility 85, 128, 200, 232,
235, 237 old age pension 140
libertarian paternalism 167 outsourcing 51, 52, 59, 60, 63
246 Subject index
P 205, 206, 209, 214, 217 219,

233, 243
pay-as-you-go 16, 38, 40, 44, 46,
95, 122, 135, 159, 164, 193, 198, reinsurance 51, 59 61, 66, 70
199, 220, 232, 237 representative participants 8,
pension basis 139, 141 146, 233
148 151, 153, 155 risk aversion 35, 79, 91, 99, 102,
defined 145 174, 224
pension benefit guarantee risk sharing 2 6, 15, 18, 28, 29,
corporation (PBGC) 198 39, 75, 77, 78, 80, 82, 84, 85, 91,
96, 130, 154, 192, 200, 205, 210,
pension contract 13, 22, 25, 75
220, 222, 223, 226, 229, 230,
78, 81, 82, 84 86, 90 92, 96,
232, 234, 235, 242
97, 99, 101, 103, 106, 207, 208,
222, 231 risk solidarity (risk sharing) 5, 6,
18, 76, 78, 91, 119, 147, 230
pension-cum-health policies 45
perverse solidarity 20, 30, 31
S
policy ladder 91, 101, 104, 111,
116 savings schemes 44 46, 166,
procrastination 168, 172, 175, 182 168, 169, 179 181
professional group pension fund selection effects 191
1, 56 58, 61, 63, 71, 162, 231 social acceptance 14
profit margin 65, 68, 70, 71, 231 social cohesion 13 15, 18, 34
progressive contribution system social contract 30, 35, 36
129 131 solidarity transfers 33, 37, 233
subsidising solidarity 5, 18, 28,
Q 37 41, 43, 45, 119, 120, 129,
130, 135, 230, 235, 236
quasi-mandatory participation
161, 163, 165, 182, 183
T
R tailored pension solutions 183,
200, 235
reciprocity 13, 14, 20, 41, 42, 44,
123 transaction costs 83, 86
recovery period 19, 22, 68
redistribution 16 18, 30 32, 44,
U
46, 95, 107, 134, 137, 138, 145, uniform contribution and accrual
147, 149 151, 153 155, 188, system 25, 119 124, 126 135,
208, 209, 219, 232, 234, 236
Subject index 247
uniform contribution rate 7, 25, value-based generational

95, 105, 119 124, 126 135, accounting 95 97, 101, 102,
137, 138, 142, 143, 145, 155, 104, 116
192, 196, 207 209, 212 220, veil of ignorance 35, 36, 43, 46
224 226, 230, 232, 234, 236 vertical solidarity 14
utility 34, 77, 79, 82, 89, 106
W
V
warm solidarity 4, 6, 147
value transfers 2, 3, 5, 13 16, 18,
21, 27, 32, 37, 71, 95 97, 100,
Z
101, 107, 116, 120, 229, 230,
233, 237 zero-sum game 95, 97, 103, 107,
value-based ALM 95, 97, 100 116, 220
102

Costs and Benefits of Collective Pension Systems

Hochgeladen von

Dokumentinformationen

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

Costs and Benefits of Collective Pension Systems

Hochgeladen von

Copyright:

Verfügbare Formate

Costs and Benefits

of Collective Pension Systems

Costs and Benefits

With 24 Figures and 28 Tables

Dr. Onno Steenbeek is financial economist

Library of Congress Control Number: 2007933404

ISBN 978-3-540-74373-6 Springer Berlin Heidelberg New York

In January 2006, the Dutch Association of Industry-wide Pension Funds

Part 1. The concept of solidarity

3 Solidarity: who cares? ................................................................... 33

Part 2. Quantifying solidarity

5 Optimal risk-sharing in private and collective pension

6 Intergenerational value transfers within an industry-

7 Intergenerational solidarity in the uniform contribution

8 Everyone gains, but some more than others........................... 137

Part 3. Mandatory participation

10 Mandatory participation for companies.................................. 187

12 Summary and conclusions ......................................................... 227

About the authors................................................................................ 237

Subject Index........................................................................................ 243

S.G. van der Lecq and O.W. Steenbeek

The concept of solidarity looms large in the public debate on pensions,

1.1 Backgrounds to this book

1.2 How are solidarity and collectivity related?

Voluntary solidarity can be informal and unorganised. This form is also

Empirical research shows that one-sided solidarity is retreating while

RISK SOLIDARITY AND SUBSIDISING SOLIDARITY

1 Note that there is no question of warm solidarity being substituted by cold

Different authors define and classify solidarity in different ways. So the

from employees to retirees and inactives;

from employees to early retirees;

1.3 Structure of the book

individual. On the basis of a number of representative participants it

2.1 Solidarity and social cohesion

1 This article was written in a personal capacity.

participants is an important condition for realising such a contract. Soli-

coming century. We therefore foresee more segregation along ethno-

2.2 Income and value transfers

THREEFOLD DISTRIBUTION MECHANISM

RISK SHARING OR SUBSIDIES

also referred to as intergenerational solidarity. Lower-than-expected re-

pants in a particular scheme that a certain degree of redistribution within the

Some generations may contribute substantially more than they receive

One example of an accumulation of (partly inverse, partly perverse)

5 A payment of 100 by future generations 30 years hence that is transformed to

than abruptly. Building, maintaining, and releasing buffers involves nu-

2.3 Setting the standard for solidarity

always paying the cost-price contribution in advance. But how strictly or

existing surpluses nor be burdened with existing shortfalls at the time of

AGE-RELATED PENSION ACCRUAL AND CONTRIBUTION PAYMENTS

participant paid a cost-price pension contribution (which was relatively

Entzinger H. and G. Engbersen, Immigratie en solidariteit, in: W. Arts et

Appendix: Different types of solidarity within

nominal return less the indexation of pension rights and payments

P.P.T. Jeurissen and F.B.M. Sanders

This contribution describes the principal forms of solidarity in the healthcare

relationship with collectivity (section 3.2)? Next we consider how solidar-

HOW MUCH SOLIDARITY?

social contract based on a veil of ignorance but a measure of the moral

SOLIDARITY VERSUS COLLECTIVITY

3.3 Health solidarity

1 As mentioned in Chapter 1, footnote 2, the concept of risk is used slightly dif-

erational solidarity component as will become increasingly evident in

Table 1. Solidarity in the financing of healthcare and pensions