Work and Pensions CommitteeWritten evidence submitted by David Pitt-Watson, Leader, Tomorrow’s Investor Project, Royal Society of Arts (RSA)

1. Three years ago, the RSA began an investigation into the efficacy of the investment system in the UK. After in-depth research using “citizen juries” we concluded it was not fit for purpose. Pension saving in particular was high cost, poorly structured, and patchy.

2. We therefore welcomed the government’s decision to adopt auto enrolment, but recommended it rethink the artificial restrictions it had placed on NEST, the default provider it had established to ensure the policy could work. Further it should encourage other providers who would offer similar terms to NEST. Longer term we argued that it should promote low cost collective investment structures like those in Holland.1

3. We therefore strongly endorse the conclusions of the last Work and Pensions Select Committee Report, that the restrictions on NEST (National Employment Savings Trust) should be removed.

4. Throughout our work, we have noted the need for “stakeholders” to reach consensus on how best we provide pensions. So we are delighted that the National Association of Pension Funds (NAPF), and the TUC (Trades Union Congress), have endorsed the proposals we have suggested, and that we have also received help from the CBI (Confederation of British Industrial). We are hugely encouraged by the fact that they are now working together to protect collective pensions in the UK from the unintended consequences of ill judged regulations.2 We would also like to thank the financial press for the exposure they have given to these issues of cost; not just the reports of the RSA but others work as well. The tide of opinion is changing.

5. If Britain is to enjoy a pension structure equal to the best sustainable system in the world, we need to “reinvent” collective pension provision, in a way which works both for employers and employees. We believe there is a huge opportunity to build on what has already been achieved to allow the people of Britain to save for their retirement in a way which is trustworthy, and vastly more efficient. In this paper we explain how this is possible, and the issues and the pitfalls which will need to be avoided.

6. The prize is huge. If today, a typical young Dutch person and a typical young British person were both to save the same amount for their pension; if they were to retire on the same day, and die at the same age, the Dutch person is likely to get a pension which is at least 50% higher than the British one. (A large proportion of this uplift can be attributed to collectivity.) Given that we spend 6.5% of the GNP (Gross National Product) on private pension savings, there would be a huge advantage to the economy if we could emulate the Dutch system in the UK.

7. In this submission, we seek to explain one of the important differences between the Dutch and the British system. That is that in Holland, pension saving is typically still done collectively. In Britain a typical saver will have an individual account, and, upon retirement the sum saved will be used to buy a pension. However, this individual method of saving is inefficient. Indeed, the government actuary has concluded that: “Collective [pensions are] expected to deliver a retirement income... [that] 39% higher than the corresponding [individual pension] outcome”.3

8. The actuary has also reported that a collective pension plan would be subject to less volatility than an individual one; that is that the pension paid would, on average be more predictable.

9. But why does collective saving make such a difference? This submission describes the difference between collective and individual savings; why the latter can be more efficient; what are the issues which need to be addressed in setting up a collective saving system, and critically, what are the pitfalls and how they can be avoided. Because, while collective pensions can offer better returns, they require a governance structure which ensures they are run in members’ interests, and not subject to misselling or mismanagement. So, for example, they need to avoid the problems which emerged in endowment and “with profit” policies.

Some History

10. Britons are well used to collective pension savings. When the country’s big employers set up their pension plans last century, they were organised as big collective pots into which employer and employee both contributed. The level of their contribution aimed, but did not promise, a pension at a level usually related to final salary. And the government offered handsome tax advantages to stimulate pension saving.

11. In the 1990’s pension schemes were more than well funded. Employers therefore asked to take “contribution holidays”, since it was felt that no more would be needed to meet the pension promise. And indeed the government encouraged them to do so in order to avoid the loss of tax revenue. Of course pension trustees were keen to ensure that the aims of the funds would be met, and so asked for an assurance from the sponsoring employer that they would guarantee the target pension would be paid; that it was a defined benefit (DB) which would be met under all circumstances.

12. However, over time, longevity increased, and returns from investment fell. Accounting treatments made explicit the scale of pension deficits, on a “mark to market” basis, which proved quite volatile. Many sponsoring companies felt they could no longer sustain the pension promise, and so closed their DB pension plans to new employees, and instead offered them a “defined contribution” pension. Employer and employee still made contributions, but these were made into an individual savings account, where the scale of the benefit depended on the amount saved, and the size of the pension it could buy at retirement age. These were known as individual defined contribution (IDC) pensions; the saver was responsible for his or her own saving.

13. Thus Britain ended up with a two tier pension system; of DB and IDC pensions. But there was a middle way, which was ignored. That was to go back to a system of target benefits, where all members of the pension plan pooled some of the risks, rather than leaving each to the individual. Such a system would be known as Collective DC. And there might be half way houses, where the employer also shared some risks; these are known as hybrid pensions. The characteristics of these pensions are illustrated in Exhibit 1.

Exhibit 1

14. There are therefore many different types of collective provision. In the rest of this paper we will focus on Collective DC; that is one where only the members of the plan share the risk, though we recognise that different forms of risk pooling may be appropriate in different circumstances.

Why does Collective DC provide high returns?

15. Collective investment provides better returns, in part because it is relatively low cost to administer, but more important, because it allows savers to pool their risk. Most of us are used to the investment advice which tells you not to put all your eggs in one basket. Of course one investment may turn into a gold mine, but equally it might be a dud; at the point when you invest you don’t know, the expected return from different investments may look the same. If the risk of making only one investment is high, all else equal, it makes sense to invest in a few, different opportunities. The expected return4 will be the same, but the risk will be lower. That investment theory is known as diversification. Used properly it is of fundamental importance in maximising the reward relative to the risk from investment management.

16. What is true for managing investment assets is also true for the management of pension liabilities. Imagine two young people about to save for their pension. They intend to retire at 65, and expect to live, on average until they are 80. But they know that it is likely that one will live until they are 70, the other until they are 90. They just don’t know who will be the lucky one. What should they do? If they both want to be on the safe side, they could start saving until they know they will have enough set aside for a life expectancy of 90, which means they will live for 25 years in retirement. But that will cost them both a great deal.

17. Alternatively, they could “insure” their lifetime income by buying an annuity when they retire; that is a promise of an income throughout their remaining life. But this annuity will be costly. According to government statistics, if someone who expects to have a normal life expectancy buys an annuity which will provide them with a real income in retirement, about 25 pence in the pound from their purchase will go on costs charges and reserves set aside by the annuity provider. So a quarter of their possible retirement income disappears.5

18. The sensible thing would be for our pension savers to save together. If one lives to 90, they will be provided for by the savings of the other. Both will have a secure income in retirement, but at a very much lower cost. And if you imagine hundreds of thousands of people all saving together, there are all sorts of risks that they can reasonably share, and benefit from. For example, as they reach retirement age, there would be less need to sell out of all risky investments and turn them into cash to buy an annuity. So returns could potentially be higher.

19. Several studies have been done on where the advantage of collective investment derives from. All conclude that it provides a huge uplift in benefit. We discuss these below, and have illustrated the approach taken, and the outcome of each of the studies in Exhibit 2.

Exhibit 2



Study Approach

Study Question and Method

What uplift in pension will collective provision provide?


Risk Sharing in defined contribution schemes6

De Haan, van der Lecq, Oerlemans, Van der Wurff

Compare DB and IDC.

Without annuitisation, how much more will need to be saved to be 97.5% that a DC outcome will cover a DB promise Monte Carlo simulation


This study method may exaggerate the benefit from CDC by assuming people have to “over save” to insure against longevity, rather than buy an annuity

Bang for the Buck7

Almeida and Fornia





Modelling Collective Defined Contribution Schemes8

Government Actuary

Compare CDC to IDC.
Uses appropriate assumptions on costs and investment policy to project outcomes

Monte Carlo simulation


This study assumed some cases where benefits were fixed. As a result in extreme cases the pension could go bankrupt. CDC schemes can never be designed with foolproof guarantees, though they should be able to hit targets.

Collective Pensions in the UK

David Pitt-Watson


Assuming different levels of returns and costs


(This paper)

Private Study10

Hamish Wilson




Risk Sharing Consultation, June 200811

Hewitt Associates




See footnote

What generates the costs of a pension?

20. However, before turning to this evidence, it is important to reflect on how the costs and returns from pension investment build up. Because very small difference in annual costs or investment returns make very big difference to pension outcomes.

21. Let’s image two 25 year olds who begins saving for their pension. Both invest the same amount in real terms every year. They both retire at 65 and die at 85. Inflation is 3% and both get a 6% return on their savings. The only difference is that one saver, Ms Canny, ensures that the pension charge is kept at 0.5%, Ms Hasty allows herself to be charged 1.5% per annum. So how much more pension does Ms Canny receive? The answer is that her pension will be nearly 50% higher. What seems like a small difference in charging, compounds over the years to give a substantial difference to pension outcomes.

22. What is true for costs is also true for returns. A 1% higher return, will over the sixty year lifetime of Ms Canny or Ms Hasty’s pension, lead to a 50% higher pension.

23. So what is the difference in costs and returns of collective and individual pensions? Well, research from the Dutch Central Bank showed that the costs for a collective pension in Holland were, on average, 0.15% of total assets. For a corresponding individual Dutch DC plan these costs are 1.27%.12

24. This is not to say that individual DC plans are not good. They are, for example, the only option for those who have no one with whom to share their pension risk, or who want a pension tailored to their specific needs. IDC offers a huge range of choice which may be preferred by those who are sophisticated investors.

25. But if the aim is to provide the highest income, for the same risk, to individuals who do not wish actively to be managing their pension plan, a properly constructed collective scheme will give a better outcome. Now it is difficult to make like for like comparisons between CDC (Collective Defined Contribution) and IDC pensions, because at some point, the IDC pension is used to buy an annuity. However, we can make some intelligent estimates of the differences.

26. Let’s start with the same assumptions about saving period, annual returns and retirement age that we used for Ms Canny and Ms Hasty. Let us assume that during the saving period, the IDC pension costs 0.3% per annum more in total charges, both declared and not declared during the saving. Let us assume that, in the five years running up to retirement, the IDC pension is invested more conservatively and as a result loses 1% of its return. And let us assume that it is used to purchase an annuity which, after all costs yields 80 pence in the pound; a number consistent with government figures. Note that this does not include any advantage that a collective pension might have from a more aggressive investment approach other than in the five years leading up to retirement. So these are fairly conservative assumptions, and we believe on the evidence that they compare good CDC schemes with good IDC schemes. In Exhibit 3 we show how each of these differences affect the pension outcome.

Exhibit 3


Lower Cost



(0.6% vs. 0.3%)

No Annuitisation


(Annuity Cost 80%)

Less Conservative Investment

+ 5%

(1% for 5 years)



27. The result is startling. The CDC pension offers around a 37% better outcome that the IDC. These figures are in the same ball park as the government actuary’s study, and the other studies shown in Exhibit 2.

28. Now of course, the assumptions we have made can be varied; perhaps there is a greater or a lower cost advantage, perhaps annuities can be more or less cost effectively purchased than government statistics suggest. A CDC provider may be able to make considerably higher returns. Further, it may be possible to run pensions which have some IDC and some CDC characteristics; for example by having collective annuity provision. But they clearly illustrate the huge advantage which collective investment can give.

What are the Pitfalls with CDC?

29. So, if CDC pensions are such a good idea, why doesn’t everyone have them? There are two reasons. The first is that, from the beneficiaries’ point of view, CDC schemes need to be particularly trustworthy. The reason is that everyone will be saving into a common pot, and after retirement, some will be taking money out. Who decides how much they can take out? If you set that number too low, older people will subsidise younger ones, too high and the opposite would happen. So beneficiaries have to accept that someone working on their behalf will make the best judgement which is possible. This of course can cause some tensions. If, for example, life expectancy were to fall relative to expectations, the younger generation will benefit, and will suffer if the opposite happens. How these issues are managed is a point of active debate in Holland. However the point is this. That if a system can generate such a level of trust, it will be 37% more efficient than one which only depends on contract. So, for CDC pensions to work, they require good governance and appropriate regulation.

30. And certainly, what you must avoid is that the person who is managing the pension is able to do so in a way which is to their benefit, but not to that of the pension savers. For example, a scheme sponsor cannot use unexpectedly good investment returns to offer advantageous terms which will attract new customers. It must keep charges low since, as we have seen very small changes in the terms and conditions of a pension can make huge differences to the outcome. Without protection from these sort of problems, collective pensions will generate the problems experienced by “with profit” insurance and endowment policies.

31. So, CDC pension schemes need trustee management. That is, they need to be managed by those whose interest is first and foremost the beneficiary, and not their own profits. Those who save with such funds need to be able to trust that they will be run only in the beneficiary’s interest. Indeed, this may be one reason that one hears so little about the advantages of collective provision, because they have little incentive to find new customers. Those who establish and run collective pensions must limit the profit they take from them. There is therefore little “market” incentive to promote collective provision. So collective pensions require a “sponsor”, such as an employer or employee organisation which is willing to act in the interests of beneficiaries.

32. The second is, from the sponsor’s point of view, CDC schemes need not to leave them with a liability. As we have seen, sponsoring employers have abandoned DB pensions because they feel they cannot accept the risk of taking on the liability of the pension promise. In the past, the legal situation made this particularly problematic. Many felt that the courts would interpret any collective provision as though it included a defined benefit. Therefore many experts assumed that Collective DC, where risk was shared amongst the beneficiary, but not by the sponsor would be illegal.

33. As we have already discussed, there are a range of different legal models which incorporate the characteristics of DC and DB pension provision in differing degrees. The key features of CDC pension provision include:

Targeting a particular level of benefit, but without any guarantees.

Investing contributions in a collective fund.

Smoothing or adjusting benefits on a discretionary basis to target benefit levels.

Providing an internal annuity in whole or in part, to maximise investment returns beyond the members’ retirement dates.

Who should be thinking about CDC Pensions

34. These conclusions are of profound importance to employers and employees. In offering workplace pensions, employers seek to provide a benefit to their employees. Designed within a collective framework, those pensions can be worth much more than if they are offered individually. Put another way, if an employer targets a certain pension benefit for their employee (dependant on salary and length of service), it will be able to do so at much lower cost through CDC than through IDC. Neither, of course, will crystallise a liability on the balance sheet. However, to establish a collective pension requires that there are adequate numbers of participants to begin the “collective”. That means that, until joint employer schemes can be established, small employers, may find less value from a collective approach. And those wanting to offer defined contribution pensions only to new employees, may also be constrained by lack of numbers. This does not preclude either of these groups from establishing CDC pensions, though it makes it more complex.

35. There is one group for whom Collective Defined Contribution should be of clear and immediate benefit. That is to those employers who are considering the closure of their DB pension plans to existing members. If they take this action, and move to an individual DC framework, then even if they pay exactly the same amount, the expected pension benefit they offer will be substantially less, because of the move from collective to individual provision. So employees not only lose their pension guarantee, they can also expect that their pension will be about 30% less.

36. However, that drop in the expected value of the pension can largely be avoided by continuing with collective provision, albeit without a defined benefit promise. CDC is therefore the natural replacement for Britain’s DB pension system, and the best long term solution for preserving adequate pensions while avoiding employers being faced with a potentially unsupportable pension liability. And collective DC may also have attractions to occupational individual DC plans which are of adequate scale, and where the members are concerned about the level of benefit which they offer.

12 April 2012

1 See Pitt-Watson, David, Tomorrow’s Investor: Building the Consensus for a People’s Pension in Britain, RSA, 2010

2 See

3 Modelling Collective Defined Contribution Schemes, Government Actuary, December 2009

4 That is the return expected before the event, calculated as the return from all outcomes multiplied by the probability of achieving those outcomes

5 See Cannon, Edmund and Tonks, Ian, Money’s Worth of Pension Annuities, DWP Research Report No 563, 2009. Table 6.5

6 Quoted in presentation by van der Lecq, to Conference on Risk sharing in Defined Contribution Schemes, University of Exeter Jan 2010

7 Almeida, Beth and Fornia, William, A Better Bang for the Buck, The Economic Efficiencies of Defined Benenfit Pension Plans, National Institute on Retirement Security, August 2008

8 Modelling Collective Defined Contribution Schemes, Department for Work and Pensions, December 2009

9 Pitt-Watson, David J, Collective Pensions in the UK, RSA, 2012

10 Quoted in article by Hamish Wilson, Collective Bargaining, Pensions World, November 2011

11 Risk Sharing Consultation, DWP, June 2008. Tables b.5 and B.6. Note both tables show significant upside and less risk from CDC. Table B.5. shows the advantage before modelling the lower costs of CDC. This gives a 15% premium, with lower costs table B.6. shows a 25% premium. If comparisons were made on an equal risk basis, the upside from CDC would be higher.

12 The study is discussed in Bikker and de Drue, Operating Costs of Pension Schemes, from Steenbeck and van der Lecq, Costs and Benefits of Collective Pension Systems. The figures quoted exclude a further 1.08% profit taken by the insurance company offering the individual pension. Caution should be used in interpreting these figures, since some of the insurance plans were relatively small and immature. Therefore in this study, a much narrower gap between IDC and CDC costs has been used.

Prepared 11th February 2013