Académique Documents
Professionnel Documents
Culture Documents
123
steenbeek@few.eur.nl vanderlecq@few.eur.nl
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
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 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
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.
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.
4 S.G. van der Lecq and O.W. Steenbeek
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.
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.:
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
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.
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.
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
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.
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-
2 Solidarities in collective pension schemes 19
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
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.
2 Solidarities in collective pension schemes 21
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
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
2 Solidarities in collective pension schemes 23
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.
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
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.
2 Solidarities in collective pension schemes 25
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.
2 Solidarities in collective pension schemes 27
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.
2 Solidarities in collective pension schemes 29
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
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?
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
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.
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.
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.
3 Solidarity: who cares? 39
AWBZ (Exceptional
o ++ ++ +++
Medical Expenses Act)
Collective
State pension (AOW) 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).
40 P.P.T. Jeurissen and F.B.M. Sanders
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).
3 Solidarity: who cares? 43
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
3 Solidarity: who cares? 45
of disease. However, with this policy option, other risk factors such as
obesity can also be taken into account. People who generate extra health
profits and/or cost reductions by leading a healthy lifestyle could then
qualify for a premium discount or lower co-payment. Substantial wealth
gains can in principle be achieved in this way (Bhattacharya and Sood,
2005). The National Institute for Public Health and the Environment
(RIVM) recently concluded that higher excise on smoking is the most 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.
46 P.P.T. Jeurissen and F.B.M. Sanders
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.
48 P.P.T. Jeurissen and F.B.M. Sanders
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
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
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.
5 The data was collected for prudential supervision purposes, where costs only
play a minor role.
4 Operating costs of pension schemes 55
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 partici- pant 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.
Source: Bikker and De Dreu, 2007.
Economies of scale result from high fixed costs and other operating costs
that increase less than proportionally with pension fund size. Examples
include the costs arising from policy development, data management 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.
56 J.A. Bikker and J. de Dreu
The upper part of Table 2 presents administrative costs for different types
of pension funds. We distinguish three main types: company pension
funds, industry-wide pension funds and professional group pension
funds. Company pension funds provide pension schemes to employees of
the sponsor company. They are legally independent of the sponsor 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)
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.
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.
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
12 On average these funds have a balance sheet total that is only one tenth of that
of the other funds.
13 Outsourcing of pension activities to insurers is accompanied by additional
agency costs: the pension fund or the responsible employer needs to check
whether the insurer or pension provider fulfils all its obligations.
64 J.A. Bikker and J. de Dreu
Table 5. Administrative costs of life insurers and pension funds by size category
(2004)
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.
4 Operating costs of pension schemes 65
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.
66 J.A. Bikker and J. de Dreu
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
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
18 The quality and reliability of such advice is sometimes disputed (CPB, 2005).
19 This does not influence the administrative costs of insurers.
68 J.A. Bikker and J. de Dreu
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.
4 Operating costs of pension schemes 69
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.
70 J.A. Bikker and J. de Dreu
products and their policies can yield significant benefits for individuals,
partly due to the possibility of providing customised products. However,
collective pension schemes based on mandatory participation can be 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 Operating costs of pension schemes 71
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.
72 J.A. Bikker and J. de Dreu
Literature
Bateman, H., G. Kingston and J. Piggot, Forced saving: Mandating private
retirement incomes, Cambridge: Cambridge University Press, 2001.
Bateman, H., O.S. Mitchell, New evidence on pension plan design and 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.
4 Operating costs of pension schemes 73
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)
a In 2004 prices;
b Weighted averages.
Source: DNB.
74 J.A. Bikker and J. de Dreu
Table B.2. Technical provisions and gross premiums of life insurers and pension
funds (2004; billion)
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.
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.
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.
5 Optimal risk-sharing in private and collective pension contracts 79
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.
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
5 Optimal risk-sharing in private and collective pension contracts 81
capital to absorb risks and thus benefit from the reward for risk. In this
case risk-sharing takes place between non-overlapping generations: the
shocks are borne not only by the cohorts who are alive when the risks oc-
cur, but also by cohorts who must still enter the fund in the future.
4 A more extensive discussion of the basic model and the many ways in which it
can be refined is provided in Bovenberg et al. (2007). This paper also presents
additional empirical results on the value of risk sharing.
5 See e.g. Constantinides et al. (2002).
5 Optimal risk-sharing in private and collective pension contracts 83
young people are unable to take over the risks of elderly people in 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.
6 See e.g. Munnell and Sundn (2004) and also Van Els et al. (Chapter 9) in this
book.
5 Optimal risk-sharing in private and collective pension contracts 85
long term which cover the longevity risk) that are not yet (readily) 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
86 C.G.E. Boender, A.L. Bovenberg, S. van Hoogdalem, and Th.E. Nijman
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.
88 C.G.E. Boender, A.L. Bovenberg, S. van Hoogdalem, and Th.E. Nijman
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
5 Optimal risk-sharing in private and collective pension contracts 89
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.
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?
92 C.G.E. Boender, A.L. Bovenberg, S. van Hoogdalem, and Th.E. Nijman
Above all, this contribution sought to sum up the current state of affairs
regarding this central research question in the literature. The adopted 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.
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.
5 Optimal risk-sharing in private and collective pension contracts 93
Literature
Boender, C.G.E., S. van Hoogdalem, E. van Lochem and R.M.A. Jansweijer,
Intergenerationele solidariteit en individualiteit in de tweede pensioenpijler:
Een scenario-analyse, WRR (Scientific Council for Government Policy)
report 114, The Hague: WRR, 2000.
Bovenberg, A.L., R.S.J. Koijen, Th.E. Nijman and C. Teulings, Saving and
investing over the life cycle and the role of collective pension funds,
Netspar panel paper, Tilburg University, 2007.
Constantinides, G., J. Donaldson, and R. Mehra, Junior cannot borrow. A
new perspective on the equity premium puzzle, Quarterly Journal of
Economics, pp. 269-296, 2002.
Cui, J, F. de Jong and E. Ponds, The value of intergenerational transfers within
funded pension system, Working paper presented at the 4th RTN 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
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
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-
6 Intergenerational value transfers 99
retically justified. Instead, the appropriate discount rates would be those that
adjust for the riskiness of the stream in question. Since the riskiness of taxes,
spending and transfer payments presumably differ, the theoretically appropriate
risk-adjusted rates at which to discount taxes, spending and transfer payments
would also differ. () Unfortunately, the size of these risk adjustments remains
a topic for future research. In the meantime, generational accountants have 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
acceptance in the context of pension funds with the development of the so-
called value-based ALM, which we discuss below.
stochastic discount rate varies with the underlying stochastic risk factors such
as investment returns and inflation. This means that the deflator is dependent
on the risk in the payment out, and therefore more suitable for the valuation of
risky payments than a fixed discount rate.
6 Intergenerational value transfers 101
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.
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.
6 Intergenerational value transfers 103
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.
104 R.P.M.M. Hoevenaars and E.H.M. Ponds
Full
indexation
indexation
0%
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
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
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:
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.
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%.
6 Intergenerational value transfers 109
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
Fig. 3. Effects for generations on switch to 100% bonds as % of total liabilities 2006
110 R.P.M.M. Hoevenaars and E.H.M. Ponds
100% bonds
We begin by looking at the effect for the change in the generational accounts,
represented by the black line in figure 3. The picture shows that whilst young
employees contribute value, the older employees and those receiving 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,
6 Intergenerational value transfers 111
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
Fig. 4. Effects for generations in the case of a switch to 100% equities as a % of total
liabilities 2006
Contribution rate
Costprice
contribution
rate
indexation
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)
Change in value of claim to surplus 2025
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
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
114 R.P.M.M. Hoevenaars and E.H.M. Ponds
Full
indexation
indexation
0%
*FR=Funding Ratio
6 Intergenerational value transfers 115
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
6.4 Conclusion
Nowadays, many industry-wide pension funds in the Netherlands have a
financing structure characterised by a conditional indexation policy based
on an indexation ladder, a contribution rate that is stable and not used or
only restrictively used as a control mechanism, and an investment mix
that is roughly made up of 40% equities, 40% fixed income securities and
20% alternative investments.
5 This conclusion is important for the discussion concerning the substance of the
policy for indexation correction.
116 R.P.M.M. Hoevenaars and E.H.M. Ponds
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.
6 Intergenerational value transfers 117
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
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).
30%
contribution as a percentage of salary
25%
20%
15%
10%
5%
0%
25 35 45 55 65
age
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.
7 Intergenerational solidarity in the uniform contribution 123
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
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
124 T.A.H. Boeijen, C. Jansen, C.E. Kortleve, and J.H. Tamerus
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
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.
7 Intergenerational solidarity in the uniform contribution 125
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.
126 T.A.H. Boeijen, C. Jansen, C.E. Kortleve, and J.H. Tamerus
The subsidies in the uniform contribution and accrual system take place
not only between generations, but also within them. Consider, for 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.
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
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?
130 T.A.H. Boeijen, C. Jansen, C.E. Kortleve, and J.H. Tamerus
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.
7 Intergenerational solidarity in the uniform contribution 131
Sub e: because in the uniform contribution and accrual system and in the
decreasing accrual system each member pays an equal contribution, there
is no question of cost of labour competition within business sectors. Young
and old also have equal opportunities on the labour market. In the 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.
6 Not all members have paid contributions in advance. Take, for example, a
member who is now 60 and has only been a member from the age of 50 until
the age of 60. As figure 1 shows, this member has paid too little contribution all
that time, and therefore certainly has not paid contributions in advance.
7 Intergenerational solidarity in the uniform contribution 133
Case 6: What amount is the total compensation for the prepaid 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.
Ages that receive the compensation Level of the total compensation amount
in in % points of funding ratio
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
7 Intergenerational solidarity in the uniform contribution 135
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
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
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.
140 K. Aarssen and B.J. Kuipers
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
8 Everyone gains, but some more than others 141
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.
Return on Investment 6%
Investment costs %
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 Everyone gains, but some more than others 143
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-
144 K. Aarssen and B.J. Kuipers
pected to live another 17.5 years after the retirement age of 65, against
20.2 years for a female member. For the value of the 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 Everyone gains, but some more than others 145
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
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.
146 K. Aarssen and B.J. Kuipers
600
500
400
1000 euro
300
200
100
0 age
25 30 35 40 45 50 55 60 65
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.
25
20
% pension basis
15
10
Fig. 3. Value of the scheme with and without exchange of partner pension (Join-
ing at the age of 25 with average career development)
150 K. Aarssen and B.J. Kuipers
The value of her pension rights then increases by 1.4%-points. That the
single woman nevertheless still experiences a small disadvantage, is
caused by the fact that she has no benefit from the survivors insurance on
death before retirement age. So, single people benefit less from the pension
scheme than partnered employees. The difference, however, is not as big
as it used to be (see box).
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)
8 Everyone gains, but some more than others 151
For employees who have been members of a pension fund for their entire
working life, with a balanced and stable membership, this redistribution is
not a major problem. Employees then pay too much at a young age and
too little at an older age in relation to the actuarially fair contribution. It is
also considered as a benefit that this mechanism encourages old people to
continue working longer. Indeed, an extra year continuing to work 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.
Table 4. Actuarially fair contribution for a partnered man at a pension fund and an
insurera
Entry age
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
8 Everyone gains, but some more than others 155
schemes has led to more equal results between employees with much
and less career development. The largest redistribution within a pension
fund is undoubtedly between young and old people. However, with a
balanced and stable membership during the course of the years, young
people naturally receive the excessively paid contribution back again at a
later age.
The redistribution within a pension fund can certainly not be ignored.
With a uniform contribution rate of 21%, for example, single men annually
overpay 3% of the pension basis, and partnered men underpay 1%. These
differences, however, pale into insignificance when we compare the 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
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?
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
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-
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).
9 Why mandatory retirement saving? 163
the remainder are not compelled to save for a pension. They are expected
to supplement their state pension (AOW) through additional voluntary 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.
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
6 These plans are named after section 401(k) of the US Internal Revenue Code of
1978, which sets rules for the offering of savings plans.
9 Why mandatory retirement saving? 167
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.
9 Why mandatory retirement saving? 171
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 See www.uvt.nl/centerdata/dhs for more information about the DHS and the
CentERpanel on which it is based.
10 The results presented here have, moreover, been weighted for age, gender,
education, and income.
11 See also Prast, Van Rooij and Kool (2005).
172 P.J.A. van Els, M.C.J. van Rooij, and M.E.J. Schuit
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.
9 Why mandatory retirement saving? 173
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.
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.
9 Why mandatory retirement saving? 175
employed were even less occupied with their pension than the average
Dutch person: half of them said they did not worry about their pension
arrangements (compared with 43% for the average respondent). 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.
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
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.
Table 3. Knowledge of own pension arrangements, 2003 and 2005 results (Percent-
age of respondents)
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
9 Why mandatory retirement saving? 179
detailed analysis showed that men, higher educated people and those who
actively invest in stocks have better financial knowledge. People who score
well on financial knowledge questions, moreover, have a more accurate
picture of macroeconomic developments as measured by economic growth
and inflation (DNB, 2006). It can be assumed that a more accurate percep-
tion of economic growth and inflation contributes to the quality of the deci-
sions that households make concerning their financial future.
16 Saving for a sabbatical (15%) or parental leave (10%) are the other most impor-
tant reasons.
180 P.J.A. van Els, M.C.J. van Rooij, and M.E.J. Schuit
pants, the fear that they would otherwise use the contribution for other
purposes plays a role.
Reasons for not participating are many and various. One important 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
9 Why mandatory retirement saving? 181
Table 4. Reasons for whether or not participating in the life-course savings scheme
(Percentage of the number of employees)
Reasons for participating in life-course savings scheme instead of themselves saving for
compensated leave (n=62)1
Tax benefits 65
Contribution from my employer 32
Otherwise I 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.
182 P.J.A. van Els, M.C.J. van Rooij, and M.E.J. Schuit
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.
9 Why mandatory retirement saving? 183
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.
184 P.J.A. van Els, M.C.J. van Rooij, and M.E.J. Schuit
Literature
Ameriks, J., A. Caplin and J. Leahy, Wealth Accumulation and the 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.
9 Why mandatory retirement saving? 185
Lusardi, A., Planning and Saving for Retirement, Working paper, Dartmouth
College, 2003.
Lusardi, A., and O.S. Mitchell, Financial literacy and planning: implications
for retirement wellbeing, DNB Working Paper 78, 2005.
Madrian, B.C., and D.F. Shea, The power of suggestion: inertia in 401(k)
participation and savings behavior, Quarterly Journal of Economics 116,
2001, pp. 1149-1187.
Ministry of Social Affairs and Employment, Nationaal Strategierapport 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.
186 P.J.A. van Els, M.C.J. van Rooij, and M.E.J. Schuit
Van Els, P.J.A., W.A. van den End and M.C.J. van Rooij, Pensions and
public opinion: a survey among Dutch households, De Economist
152(1), pp.101-116, 2004.
Van Els, P.J.A., W.A. van den End and M.C.J. van Rooij, Financial 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
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
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.
Since 2000: including the ABP, privatised in 1996. Source: Lutjens, 1999 and CBS, 2006.
10 Mandatory participation for companies 191
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.
10 Mandatory participation for companies 193
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.
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).
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.
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%
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.
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
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
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.
208 J.P.M. Bonenkamp, M.E.A.J. van de Ven, and E.W.M.T. Westerhout
funding deficits;
the uniform contribution and accrual system;
deferred taxation system.
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.
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.
11 Macroeconomic aspects of intergenerational solidarity 211
COMPETITION
Although there is no explicit competition between pension funds, there is,
however, a question of implicit competition. Thus, for several years the
investment performance of mandatory industry-wide pension funds has
been mutually compared by means of the so-called z-scores. If a fund 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.
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).
214 J.P.M. Bonenkamp, M.E.A.J. van de Ven, and E.W.M.T. Westerhout
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
%
40
35
30
25
20
15
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
13 We have also examined the sensitivity of the results for alternative assumptions
with respect to interest, productivity and career path. The size of the subsidies
varies, the direction does not. In addition, the age of reversal remains virtually
unchanged.
11 Macroeconomic aspects of intergenerational solidarity 217
%
40
35
30
25
20
15
uniform contribution
5
actuarially fair uniform contribution
0
20 24 28 32 36 40 44 48 52 56 60 64
age
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).
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
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.
220 J.P.M. Bonenkamp, M.E.A.J. van de Ven, and E.W.M.T. Westerhout
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
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.
14 The contribution of Boender et al. goes deeper into this WRR calculation and
indicates which points of this calculation can be improved.
11 Macroeconomic aspects of intergenerational solidarity 223
15 A sensitivity analysis concerning the risk aversion and the time preference
leads to an added value of DB schemes between 8% and 12% of the capital.
224 J.P.M. Bonenkamp, M.E.A.J. van de Ven, and E.W.M.T. Westerhout
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 Macroeconomic aspects of intergenerational solidarity 225
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.
226 J.P.M. Bonenkamp, M.E.A.J. van de Ven, and E.W.M.T. Westerhout
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-
228 S.G. van der Lecq and O.W. Steenbeek
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 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.
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.
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.
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.
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.
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. (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.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.
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.
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.
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.
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.
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