Response to DWP Consultation ‘Reshaping workplace pensions for future generations’ – Defined Ambition pensions.

Submitted by Dr. Ros Altmann

Independent Pensions Policy Expert

(All material on this page is subject to copyright and must not be reproduced without the author's permission.)

The DWPs consultation on Defined Ambition (DA) pensions aims to try to find a way to encourage employers to take on more of the risks of pension provision. DA is a chance to start again and design employer pensions that are more flexible and allow more leeway for change over time, as required to accommodate changes in demographic and financial variables.

Currently, the policy of auto-enrolment is spreading pension coverage across millions more workers in the UK, but the overwhelming majority of new pension provision is in Defined Contribution (DC) schemes. These schemes place almost all the risks of pension provision on the individual worker. This is in stark contrast to the ‘traditional’ British company final salary pension scheme, which used to offer Defined Benefit (DB) pensions. In DB schemes, the value of the future pension income is underwritten by employers, so the worker is only responsible for making the requested contributions, while the employer has to make up any shortfall in the funding of the promised future pension income in retirement.

Past mistakes and well-meaning legislation have resulted in the traditional DB schemes becoming unaffordable for most private sector employers. The increasingly draconian regulations and requirements placed on employers, without the flexibility to adjust to changing market, economic and demographic realities, has resulted in employers rushing to remove pension risk from their balance sheets. In the past, the law gradually tightened employer promises into immutable guarantees and added extra costs on top of the original pension promises, including inflation linking, spouse/partner pensions and revaluation of leavers’ benefits. These have added huge extra costs (around 50%) to the pensions promise and rising life expectancy without the ability to change pension start dates for past accruals has added further costs and risks. Every year that life expectancy rises while pension ages stay fixed, employer costs will increase. Even if the same level of pension is paid, the value of that pensions rises all the time, but workers do not understand that ‘no change’ actually means a change for the worse for the employer. There is so much misunderstanding of pensions and workers do not appreciate the additional costs that traditional DB schemes are imposing on employers as both demographics and market returns have been so different from previous forecasts. This has led employers desperate to get rid of the risks they face and to move to DC pensions, which leave the risks on employees instead. Firms merely promise to pay the specified contribution levels each month, but this merely builds up a pension fund for the worker, it does not provide a particular level of income.

In DB schemes, the employer must cover all the following risks:

  • Salary inflation
  • Price inflation
  • Interest rate risk
  • Longevity risk
  • Partners’ pension risk
  • Revaluation of deferred benefits
  • Financial market risks
  • Annuity risk
  • Regulatory change risk

In DC schemes, however, the employee carries those risks and the employers’ liability is finite.

This consultation tries to find a third way, with DA trying to push some of the risks back onto employers. Indeed, it is arguable that DA, as described in this consultation document, is really what the original DB schemes were intended to be. They were set up by paternalistic employers who wanted to help their loyal lifelong workforce enjoy a few years of retirement after they finished working for the company. The pension age of 65, set many decades ago, would have meant around 10 years of retirement on average, with many surviving far less long. Over the years, employers were encouraged to reduce normal pension ages (even as life expectancy increased) and then the Government continually tightened up the legal requirements so that pension promises from the past could never be reduced – only future accruals could be changed.

Having thought that these schemes were in surplus and could rely on future stock market returns to meet their liabilities, employers were saddled with additional obligations that turned out to be unaffordable, when it became clear that market movements could not be relied upon (even in the long-run) and that the surpluses were not really surpluses at all. They were buffers against bad markets and future demographic change, that would have ensured extra funds were available if required. Nevertheless, the DB system relies on employers always being able to make up any shortfalls and being willing to underwrite open-ended liabilities for many decades into the future. The way UK DB schemes have been prevented from making changes to reflect new circumstances has led to unrealistic expectations on the part of workers who do not realise how expensive pensions have become. Workers have almost been given the impression that pensions grow on ‘magic money trees’ and that the future pension income can be guaranteed with a fixed past contribution, even if the income has to be paid for far more years than originally budgeted for. Pensions cannot keep lasting ever longer at the same level but many workers do not realise this.

DB does not fit with 21st century capitalism. Corporate life has changed and employment patterns have changed, so that a DC model fits far better with new realities. Most companies will only exist in present form for 10-20 years nowadays and most workers change jobs on average 11 times in their working life. Therefore, the concept of a pension being paid to loyal lifelong workers in 30, 40 or 50 years’ time does not make sense any longer. The employer cannot be relied on to be around in the longer term and the worker is unlikely to stay employed until they retire.

By moving to pure DC, workers are in a much worse position than before. This is a problem for the Government, however, as the state pension is being cut and, whereas past state pension reductions could be somewhat offset (at least in theory) by relying on increased private or employer pension income, as Government shifted more burdens onto final salary schemes, the new-style pensions do not offer such generous benefits and cannot be relied on to provide good extra pension income.

DB schemes were a form of social welfare, underwritten by employers. This is not the natural role of an employer, it is usually a state role. The closure of most private sector DB schemes therefore poses a problem for pensions policy. The state pension cuts leave many people at risk of inadequate later life income and, if more and more pensioners are poor, the long-term economic outlook is damaged as they will not have money to spend and may increasingly require taxpayer support. Thus, the DA agenda is to try to persuade employers to take back responsibility for more of the pension risk that pure DC, albeit recognising that DB – as currently configured in the UK - is too draconian for employers to voluntarily accept.

The consultation suggests that the aim of DA is to find ways to offer more guarantees than are available in DC. There are two models – ‘DB minus’ (where the employer offers some extra guarantee or even discretionary increases) or ‘DC plus’ where either an employer or a pension fund offers a guarantee or discretionary additional benefit on some aspect of future pension.

I would question the wisdom of trying to offer too much in the way of guarantees on future pensions. As workers draw nearer to retirement, a guarantee could be provided but if the guarantee is meaningful it may be expensive and if the guarantee is low cost it may not be meaningful. A government underpin would be more attractive than a guarantee from a private firm or employer.

DA proposes various models of ‘risk-sharing’ between employer and employee. This is worth a try, although it is not going to be easy to persuade employers to take back significant amounts of risk. That is why it is sensible to propose ‘safety valves’ which will allow for discretionary increases in benefits year by year, if funding allows, and changes along the way as other factors change. This should always have been part of the DB landscape. Indeed, I would argue that DA should not aim for hard guarantees at all. It should help workers understand that pensions simply cannot be guaranteed by private sector firms many decades hence. The concept of guarantees should be replaced by a ‘best efforts’ promise, which cannot be relied on absolutely and which needs to be monitored and checked over time.

The model which proposes mandatory transfers of members out of DA schemes into DC schemes on leaving seems to me to be the wrong way round. Valuing DA benefits correctly will be a nightmare and I consider there will be significant difficulties with such a scheme design. I would suggest that this would work much better the other way round. After a certain number of years in a pension scheme, the benefits could be converted from DC to DA (or DB), so that employers will be rewarding loyal service, and adding guarantees (albeit less draconian than currently) to their long-serving workers’ pensions. Past accruals could be converted into an equivalent number of years worth of DA pension accrual. For example, after ten years’ service a worker could be offered a chance to convert their DC pension fund into a participation in a DA or DB scheme with ten years’ service at the salary they earned each year.

Problems with trying to move to DA

  1. Will employers be willing to once again shoulder pensions risks that they can pass onto workers in DC? If shareholders do not wish to be saddled with future risks, DA will not be adopted by many firms.

  2. Many employers may be fearful that the ‘ambition’ will not stay as just an ambition, and will be sceptical that DA will be converted into a legal guarantee in future, as happened with traditional final salary schemes.

  3. Guarantees may not be deliverable on a reliable basis – schemes must have flexibility to work more on the basis of adjusting to developments in people’s lives and financial markets

  4. There is bound to be a trade off between guarantees and potential returns. This needs to be understood. A ‘money back’ guarantee on contributions is not a very ambitious ‘Ambition’ and future pensions are bound to be uncertain.

  5. Even the state pension has turned out to be unaffordable for the state itself, and it has been cut several times, with further cuts coming. If taxpayers cannot afford unreformed pensions, then corporate UK cannot possibly be relied on to deliver future pensions and workers need to know this.

  6. Imposing conditional or discretionary benefits on employers or schemes could still prove too costly, especially in the face of EU EIPOA regulations that may come in and extra solvency or reserving requirements for pension funds.

  7. Ultimately, pensions risk cannot be eradicated. Passing risk from one party to another does not make it disappear and, if markets do not deliver, or if contributions are too low, pension provision will remain inadequate.

Specific Responses to Consultation Questions

1. Do you agree that a greater focus on providing members with more certainty about savings or preferably income in retirement may increase confidence in saving in a pension?

I agree that providing more certainty can help increase confidence in theory, but in practice not only will people find it difficult to trust pensions given past experience, but it may also prove impossible to actually offer meaningful guarantees anyway. Guaranteeing a future income many decades into the future is too risky for private companies to be relied on to deliver. It would be better to be honest with people and let them know that saving in a pension fund carries risks so that future pension income cannot be guaranteed (at least not till close to retirement). However, not saving is guaranteed to mean lower living standards than if one has saved.

2. As an employer, do you have experience of, or can you envisage any issues with, employees being unable to retire due to DC pension income levels or certainty about income levels?

Employers may have some concerns that workers cannot retire if they have inadequate incomes, but I don’t believe slight improvements to pension income outcomes are likely to be sufficient to address retirement affordability in a meaningful manner. Employers can best help older workers prepare for retirement by facilitating part-time work in later life, which can help them accrue higher incomes and continue some saving, in preparation for stopping work altogether at a later age, while cutting working hours for a period of time beforehand. Offering financial planning help to workers is likely to prove a more cost-effective method of dealing with retirement issues than trying to underwrite pension guarantees.

3. What are your views on the feasibility of this scheme design?

Flexible DB schemes are feasible and an improvement on inflexible DB, but there will still be concerns that employers may not be able to afford the core benefits themselves. Certainly, permitting discretionary year by year increases, as funding and economic conditions allow, would be necessary to attract employers to take on pension risks again. They would probably still be concerned that the flexibility in DA schemes could be taken away by future legislation.

5. Are employers likely to be interested in providing benefits in addition to a simplified flat-rate DB pension on a discretionary basis or otherwise?

Private sector employers may choose to provide additional benefits, such as life insurance which has disappeared from DC schemes, however additional pension benefits are likely to be less attractive.

6. What role do you see for scheme trustees in relation to discretionary payments? For example:

  • Should they be involved in deciding whether a discretionary payment is made at all?

Trustees should be involved in assessing the affordability of discretionary payments.

  • Should they be involved in setting out how these payments are apportioned to members or should this be down to the employer?

Trustees should be involved in apportionment decisions.

7. Do you agree that our starting point should be to keep regulatory requirements around discretionary benefits to a minimum?

Yes, definitely!

8. How do you see funding for the non-discretionary DB element being sufficiently protected while allowing for extra discretionary benefits? For example, is there a risk that paying discretionary benefits could threaten the funding for non-discretionary DB benefits for younger scheme members?

This all depends on investment returns to a large degree and there is always the risk that younger members’ benefits could be threatened by the amounts paid out to older members before the young ones reach retirement.

Flexible defined benefit – automatic conversion to DC

9. What are your views on the feasibility of this scheme design?

I consider there will be significant difficulties with such a scheme design. The formula for calculating transfer of benefits is unlikely to be easily agreed upon and is fraught with risks. I would suggest that this type of scheme design would work much better the other way round. After a certain number of years in a pension scheme, the benefits could be converted from DC to DA (or DB), so that employers will be rewarding loyal service, and adding guarantees (albeit less draconian than currently) to their long-serving workers’ pensions. For example, after ten years’ service a worker could be offered a chance to convert their DC pension fund into a participation in a DA or DB scheme.

10. If employers are able to use scheme designs 1 and 3, do you think it is still helpful for legislation to allow for this scheme design?

I am not in favour of this scheme design, since mandatory transfers will be difficult to value. If the design works the other way round, so that employee benefits are converted from DC after a set number of years, then those benefits would remain with the employer but would not be tiny amounts from just short periods of accrual.

11. Do you think this scheme design could be extended to permit employers to automatically transfer members out of the scheme at retirement?

I think it would be far better to leave it to workers to decide if they want to transfer out and if their accrual only becomes DA or DB after a certain period of time, they are less likely to want or need to transfer. Transferring between industry-wide employers could be another model to consider.

12. What would be the most suitable way for benefits to accrue under this model? And how might this best be communicated to ensure members understand the value of their pension benefits?

Benefits that could be automatically converted to DC would be difficult to explain and I do not believe it is sensible to try to do so.

13. Assuming a CETV would not represent ‘fair value’ for the accrued rights when the member leaves or retires, how might it best be calculated? Should the basis for calculation be different when the transfer is initiated by the employer (for example on redundancy)?

This question illustrates why I believe the proposed model is flawed. The value of the CETV would be open to manipulation and the complexity of different calculations depending on whether the employer or worker initiated the scheme exit would be costly, complex and unreliable.

14. For schemes providing a lump sum benefit, what are your views on how the cash value should be calculated for members who leave before retirement? Should the net present value of the lump sum be calculated on how many years away from pension age they are?

I am not in favour of this model

15. Could the accrual rate and pension value be along similar lines to existing cash balance arrangements?

Workers should be allowed to stay in current cash balance arrangements after they leave employment, not forced out.

16. What forms of regulatory requirements would be needed to:

prevent avoidance activity?

ensure the scheme has access to sufficient funds to enable a transfer when a member leaves?

Mandatory transfers could result in unfairness to members not transferring out, depending on assumptions used for the valuations. Scheme cash flow would need to be carefully managed but it is not clear that will be an easy task.

Changing scheme pension age:

17. What are your views on the feasibility of this scheme design?

Changing scheme pension age is sensible but again will be difficult to manage fairly. Certainly, the ability to say to younger workers that they cannot be sure of their scheme pension start date until nearer the time they may start to take benefits is vital. As life expectancy continues to rise, it will be necessary for employers to have the leeway to increase pension age in future. If future pensions have to start at different ages, depending on the original contribution date, the administrative complexity will be too challenging.

18. It could lead to more schemes having proportions of accrued pension payable at different pension ages. Would this further complexity outweigh the benefits?

Yes! Future pensions should start from a date to be determined in future, rather than pre-set many decades in advance.

19. What role do you see the scheme trustees playing? Should they be involved in setting a new NPA, or should this be down to the employer and the employer’s actuary?

Trustees need to be involved in order to protect members’ interests, but there should be negotiation between all the parties, with employer and actuary having to justify the increase.

20. What are your thoughts on how future pension ages are set?

For GAD to publish a standard index based on longevity assumptions?

Or do you prefer schemes linking their NPA with the State Pension age, so that when the latter changes, the scheme’s pension age automatically changes in line with this?

It is not up to Government to determine scheme design. Future pension age could be determined by industry, as some industries have lower life expectancy than others, or it could be tied to state pension age or another measure of longevity. This would all need to be justified in negotiation and would be complex in practice.

21. How might the decision to change the NPA work in multi-employer schemes?

I believe in trying to standardise benefits across employers – assuming multi-employer schemes are all in the same industry. If different employers have different life expectancy for their members, then it could prove difficult to adjust pension age a different scheme design may be needed.

22. As an alternative to opening a new scheme, do you agree it should be possible for an employer to modify the rules of an existing scheme so that it can be re-designed as a Flexible DB scheme in relation to new accruals, for example, it is possible to change the NPA and/or introduce automatic conversion to DC when a member leaves?


23. Do you agree that employers should not have the power to transfer or modify accruals built up under previous arrangements into a new arrangement, beyond what is allowed under current legislation?


24. Should there be a requirement to provide independent financial advice in all cases where an employer offers to transfer a member’s accrued rights from a traditional DB scheme to a new arrangement?


25. Do you think having more certainty than traditional DC would be welcomed by members, and help generate consumer confidence and persistency in saving?

See answer to Question 1. Members would probably welcome more certainty but need much more education about the value and cost of pensions.

26. As an employer, if these products mean there is no funding liability, only the requirement to contribute as for a traditional DC scheme, would you be interested in offering these products to employees?

If there is no funding liability, it is hard to see what kind of meaningful guarantee could actually be on offer. The costs of guarantees will vary, but a guarantee that is worthwhile in terms of delivering future certainty is likely to be very expensive and a guarantee that costs very little is likely to be worth very little.

27. In relation to medium- and long-term guarantees outlined in model 2 (capital and investment return guarantee), and model 3 (retirement income insurance), would removal of the legislative barriers be sufficient to stimulate the development of market-based solutions?

It is already possible to obtain capital and investment return guarantees, as well as current and deferred income guarantees. For example, MetLife offers such guarantees but they come at a cost. They are available on existing DC scheme arrangements or savings plans and do not require new DA legislation. The problem is that there are not enough providers of sufficient size and strength to underwrite mass-market guarantees. This is a niche area, only available to wealthier savers. Introducing a ‘fiduciary’ to negotiate cheaper guarantees seems a really nebulous concept and it is not clear that this is a practical option.

As regards retirement income insurance (option 3) this looks like buying deferred annuities, which are even more expensive than ordinary annuities, with an equity fund to provide a bonus on top of the deferred guaranteed income. This is likely to be a very expensive way to provide future pensions. Deferred annuities are exceptionally poor value due to the risk and profit margins built in. They may offer a guaranteed income, but it will be very low. It is far better to try to buy a guarantee much closer to retirement and leave members to keep on accruing additional savings during their working life.

28. As insufficient scale has been identified as a barrier to providing affordable guarantees, is there a role for the Government in facilitating different types of pension vehicles that would create greater scale for this purpose?

I do not believe the Government itself should design new pension vehicles, that should be up to the pensions industry. However, I think there could be a role for Government in providing longevity gilts, which could help offset some of the longevity risks of pensions and the Government itself could consider underwriting annuities, which would allow much better rates. The profit and risk margins on current annuities are clearly very high and there is room to pass significant cost savings onto consumers if the provision of annuities were taken over by the Government. By threatening to compete in the annuity market, the Government would probably then force annuity companies to improve the value they offer.

30. Do existing protection arrangements for DC products provide sufficient protection for members in the event of provider insolvency?

It is not clear that savers with more than the statutory minimum compensation limits in personal or group pensions with financial services firms are sufficiently covered by insurance protection in event of provider insolvency. It is not clear whether the PPF itself would be involved in this DA area and how it would value the promised benefits.

32. Are these models likely to be an attractive option for employers and members?

Such options could be made to sound attractive, however the long-term risks of these provisions and the probability of disappointment for current or future retirees makes them less attractive in the medium term.

33. On model 4 – pensions income builder – what are your views on this model in which members are in effect deploying their own capital to guarantee their own entitlements?

It is likely to prove difficult to explain this concept to members in a manner that they can understand and the likelihood of misrepresentation or misunderstanding is high. This model will not provide guaranteed pensions and, although employer contributions will be capped, this means workers’ pensions may not pay as much as they expect and they may have a false sense of security.

34. Do you agree that CDC schemes have the potential to provide more stable outcomes on average than traditional DC schemes?

I agree that CDC schemes have the ‘potential’ to provide more stable outcomes on average. That does not, however, mean they will definitely do so. CDC does mitigate risks, especially in the nearer term, rather than being fully exposed immediately to mark-to-market risks.

35. Given there is no tradition of risk sharing between pension scheme members in the UK, are individuals going to be willing to share the benefits of protection from downturns in the market and increased certainty of outcome, with the potential disadvantages of intergenerational risk transfer?

The question of whether members will be willing to share the protection benefits pre-supposes that they will understand how the scheme works. There is a notable lack of inter-generational solidarity in Britain today, with younger generations seeming to believe that older people have had too much of the share of national wealth in terms of pensions and housing. The CDC design certainly means younger generations could lose out relative to older members of the pension scheme, because future returns may not materialise as predicted and, therefore, funding may fall. This goes back to the point that we need to reduce the idea of ‘guarantees’ in pensions and help people understand that pensions cannot be relied on to be exactly as predicted many decades ago, due to changes in demographics and markets. If members are not led to believe that their pension income can only rise and if they are clearly told that there may be times when their pension income could have to reduce, then they may accept the principles, but it is important that they understand. People accepted with profits concepts at the time, but that was partly because they did not realise how much returns could fall during bad times. CDC has the potential to mislead in the same way as with-profits did, or indeed Equitable Life. It is important that a distinction is made between past mistakes and future best practice. CDC is likely to deliver more reliable outcomes for most people than pure DC, but this is not guaranteed. Pure DC is probably the most risky form of pension saving for individual workers.

36. Is a CDC scheme designed to manage funding deficits by cutting benefits in payment going to be acceptable in the UK where traditionally maintaining the value of benefits in payment has been an overriding priority?

We need to get away from the idea that pension benefits can never be cut under any circumstances – this is financially unsustainable. Even taxpayers have been unable to commit to maintaining state pensions, so private companies are even less able to do so.

37. What levels of funding do you consider would be appropriate to ensure that a CDC scheme has sufficient capital to meet the liabilities and minimise the risk of benefits in payment being cut?

It is impossible to answer this, there will always be times when assumptions about appropriate funding levels prove incorrect. A ‘best efforts’ basis is more reasonable.

38. Given the need for scale and an ongoing in-flow of new members to ensure the sustainability of a CDC scheme, will it be possible to set up a scheme without some form of Government intervention?

Industry-wide schemes could provide scale, but it will always be the case that new members carry more risk than older members. This is the nature of CDC. That does not mean younger members will necessarily get lower pensions, but they are more at risk of this.

39. As a mutual model, it has been suggested that CDC schemes might prove attractive to the trades unions and other social partners – might this be an option worth exploring?

Any ways in which we can increase the scale of pension saving schemes is worth exploring. Economies of scale are very important.

41. Do you have any comments on how to characterise the defining characteristics of DA pensions?

DA pensions should emphasise the word ‘ambition’ so that members know this is an aim, not a cast-iron guarantee. Members should be told that the benefits will be influenced by factors such as investment returns and changes in life expectancy and that the final pension received cannot be predicted until later life.

47. Do you think that setting up a CDC scheme should be subject to formal approval, for example licensing by a regulator?


48. Do you think that CDC schemes which do not provide a guarantee or promise should also be licensed?


49. Do you agree that such CDC schemes should also be subject to DA requirements on governance and member communications?


50. Should there also be an option for schemes that currently offer DC to convert to CDC?

This should surely be an option, but requirements for clear explanation and member choice are needed alongside this.

51. In the absence of both a guaranteed pension entitlement and an individually defined pool of assets, how should assets in a CDC scheme be apportioned such that pension accruals can be measured for tax purposes against the Annual Allowance and the Lifetime Allowance?

Restricting lifetime accruals does not make sense to me and the annual allowance should be based on the amount of money per member actually being paid into the scheme.

52. What specific areas should we address in relation to governance and member communications for DA schemes?

Governance needs to ensure best efforts to deliver the pension ‘ambition’ but communications need to make it clear that these are aspirations, rather than guarantees that can be relied on indefinitely.

Dr. Ros Altmann
December 2013

© Dr. Ros Altmann  |  Home  |  Profile  |  Disclaimer