Con Keating’s prepares his thoughts on CDC


Like many of us, Con’s thoughts are on how we can best restore confidence in pensions using CDC , here we catch him thinking about the W&P inquiry. For those whose thoughts are tuned to pensions (as well as retirement saving) – this may give inspiration.

The Work and Pensions Select committee inquiry into Collective Defined Contribution pensions.

This inquiry is most welcome as this is a subject which is prone to much misconception and misunderstanding. It is evident in both the popular and trade press. Traces of this permeate the enabling legislation, the Pension Schemes Act 2015 (PSA 2015), and indeed even the introduction to this inquiry published on the Work and Pensions Committee’s website. It appears that not even the Minister is immune.

There is little to be gained from any analysis of this historic situation; many of the objections commonly and erroneously raised will be covered in the course of this exposition.
I am a financial analyst of long-standing, and would be considered by most as mathematically adept. For the avoidance of any doubt, I should rightly be called a supporter of the approaches now being considered; this is based upon a considerable body of research I have conducted myself, as well as critical reading of other people’s work – both practitioner and academic.

This work has been conducted over a period of two decades, and may fairly be dated back, at least, to 1997, when it became obvious to me that the Pensions Act 1995 and some other policies adopted following from the Goode Committee Report were encouraging perverse behaviours in the management of DB schemes. At that time, as my own time was committed to other activities, I commissioned a comparative and analytic study of the situation from Manfred Lam, an actuary from the University of Canterbury.
I am familiar with the practices in many of the countries which are often cited as having adopted CDC. I would urge caution in considering these in either support of, or opposition to the proposed developments here. These arrangements are the product of the history and culture of those countries and their pension systems. If they provide any useful guidance, it is of things that went wrong, and mistakes that were made. We can and should learn from those.
By way of preamble, I should like to make the point that private pensions play a much more important role in the UK than in many other countries. They are a core element of an adequate and sustainable retirement income, while in many other countries the equivalent to our basic state pension is far more generous, and private pensions play only a supplementary role. This greatly increases their importance, and the significance of the work of this committee, and this inquiry. The apparent breakdown, since the introduction of Freedom and Choice, of the non-partisan consensual compact which led to auto-enrolment is to my mind regrettable. Pensions are too important to be a party-political affair.
There is a tendency to consider the many possible variants of DA or CDC as lying on a continuum, or spectrum between individual DC (or money purchase) and traditional employer sponsored and supported defined benefit. This is not helpful. First, because there is no continuity across that span, and second, because CDC and most of its variants lie outside of, or beyond that range.
DC arrangements are really no more than tax-advantaged savings schemes; they generate a pool of savings, but those savings lack the insurance element of a pension as a lifetime retirement income. In traditional DB the pension contract is part savings plan and part longevity insurance. The employer sponsor guarantees the performance of both savings and longevity insurance.

The choice for the individual DC saver of either annuitisation or management by drawdown is a very difficult one. I did describe the difficulties with drawdown extensively some while ago in a series of four blogs published on ‘The vision of the pension playpen’ . Annuitisation is expensive, in large part because these are highly regulated insurance products.

Critical Characteristics


These are collective institutions, member mutual organisations. Given the familiarity of the pension world with trusts, and the desirability of protecting members’ interests from ordinary personal bankruptcy proceedings, this is my favoured institutional form. Note that we cannot meaningfully talk about insolvency of this body. This is more than a question of members “owing things to themselves”; there are no enforceable external liability claims. These “promises” are ambitions not guarantees.
With all such bodies, governance is important. With trustees elected by members and officers and administrative staff appointed by trustees, trust law should then be sufficient. We do note, also, that modernising trust law is a project which has been adopted by the Justice Department.
The basis of voting should be one member one vote. This favours younger members of the scheme relative to older, even though they have a smaller equitable interest in the scheme than older members and face greater uncertainty. It is consistent with the control rights solutions to intergenerational ‘commons’ problems. More discussion of these aspects and alternate possible voting rights allocations are available on request. In addition, members would have the unconditional right to transfer to another scheme at any time; they thus have both ‘exit and voice’ available to them.
There are, of course, issues of member engagement. While the proposals contained within this response address the subject of member communication, where the current position should be available to them in near real time, it does not specifically consider member engagement. But nor should any of the proposals made, other than true election, depend upon such participation.
The terms of the trust should introduce a duty on trustees to consider the continuation of the trust in time beyond the simple disbursement, when demanded or due, of all current members’ equitable interests. The precise meaning of a member’s equitable interest is described later; it would be legally enforceable in relative but not absolute terms. This duty places the trust firmly in the realm of the ‘going concern’.
It is difficult to see that further regulation of these institutions is either necessary or desirable in this governance regard.
Contrary to popular belief, there is no required role for any sponsor employer. Where a scheme is associated with a particular employer’s workforce, the role of the employer is limited to the payment of contributions for employee service rendered. Employees may also make contributions.
As with current arrangements trustees are charged with the collection of contributions. There are existing legal methods for the collection of such debts, and a senior status seems warranted in employer insolvencies. We shall return later to the question of setting contribution levels or amounts. There is absolutely no need for any of the sponsor debt legislation to be introduced. The reference to the s75 debt in PSA 2015 is misconceived. Indeed, as the body has no liabilities, there is no basis for the formation of any such a debt.
One of the great attractions of these arrangements is that a scheme may be associated with other bodies, such as professional associations and affinity groups. Indeed, they could accommodate those with relatively transient working patterns, such as ‘gig’ workers. By removing the explicit linkage with a particular employer, the duration of service with any employer is no longer a concern.
Nor is this an issue for employers, the pension can still be used to satisfy their recruitment, retention and reward ambitions. The schemes should also qualify for and meet auto-enrolment obligations.
There is a wide-spread belief, given their past experience of government intervention and regulation in traditional DB, that ‘no employer will touch CDC’. As is obvious, this concern is malfounded.

The employer involvement need be no more than payment of contributions, as they are doing with DC under arrangements such as auto-enrolment. In addition, it seems likely that government intervention in CDC arrangements will be far lower than the historic DB case. These are member mutuals, and there are votes at stake.
Some take ideological exception to anything collective, believing in individual responsibility. The approach is often sold with clarion calls such as “giving them (people) the tools to make informed decisions and take control of their finances.” This seems foolhardy to me.

First, individual DC arrangements are available for those who wish to pursue this approach.

Second, it seeks to deny the undoubted benefits of collective, co-operative solutions when these are the very fabric of both trust and civil society, and were the dominant process in our evolutionary development, both figuratively and literally.

Third, it fails to recognise that there are many who do not wish to have such responsibilities thrust upon them, and many who do not have the cognitive abilities to cope with them. Can we seriously expect individuals to manage their affairs in this way in older age, when chronic illness and cognitive decline are prevalent?
There are at least eight rigorous academic or professional studies of the comparative performance of individual and collective DC arrangements, both empirical and simulations. All indicate that collective solutions are superior to individual. Technically, this is second order stochastic dominance. This is a result that would be expected from the mathematics and economics of investment, insurance and risk. In these studies, there is considerable variation in the scale of these benefits, but that is to be expected given their sensitivity to the precise construction, parameterisation and calibration of these models. I have not seen any study, simulation or empirical analysis which runs counter to this.
These studies remove much of the uncertainty associated with a shift from individual DC to CDC. It is notable that the Royal Mail / Communications Workers Union proposal would, when enabled, substitute a CDC model for a money purchase or individual DC arrangement on closure of the existing DB scheme.

The introduction of a CDC arrangement, rather than money purchase, is also being quite widely discussed within the academic membership of USS, in response to UUK’s plans to close the DB section there. These developments also illustrate the fact that CDC arrangements are well suited to accommodate industry-wide membership. They do not introduce any of the problems and tensions associated with multi-employer defined benefit schemes.
It seems appropriate to recommend that the Royal Mail proposal, when fully developed, should be permitted immediately as a trial; this may need little more than the safe haven of a ‘no action’ comfort letter.
I have seen no rigorous academic or professional studies comparing the performance of traditional DB and arrangements which might be called CDB, DB-lite or DA, even though the enabling legislation of PSA 2015 was clearly written with this mind.

My own preliminary and incomplete analysis suggests that there may be circumstances in which these CDB arrangements are superior to traditional DB. The theory is simply that the employer guarantee is costly, particularly so now that there are substantial costs arising from the regulation of schemes containing these guarantees. Removing those costs should then produce better outcomes for members.

However, this is highly speculative at this stage, and the precautionary principle would support waiting until we have evidence of the actual performance of CDC schemes before addressing that question.

The ambition, target or “promise”

No voluntary contract can be sustained if it is inequitable among the parties to that contract. This is a lesson that should have been learned from the experience of traditional DB.

The inequity, the unfairness of active scheme members losing all of their pensions when close to retirement after a working lifetime of service while pensioners-in-payment continued to receive their full pension was obvious to all. It led to semi-naked protesters on the beach. However, the volumes of regulation which have been introduced since, have not fully resolved this issue.

The reductions in benefit payable by the Pension Protection Fund merely modify this. Those volumes have also introduced entirely new inequities, among the stakeholders of a sponsor firm. This has driven the trend to scheme closure and seemingly aberrant behaviour such as ‘pre-pack’ insolvencies.
The role of the “promise” is to define the equitable interest of a scheme member. It may also serve several other important purposes. It is endogenous, defining the relative magnitudes of claims of members on the assets of the scheme. It is clearly justiciable and enforceable.
Fuller descriptions of the concept, including its calculation are contained in a number of published articles appended to this submission. By expressing the “promise” in the manner of traditional DB, where a particular contribution confers the “right” to a particular defined retirement income for the member’s lifetime in retirement, the scheme is effectively determining an investment return on that contribution. The contribution is known and the pensions ultimately payable are projected using standard actuarial techniques.

This determines an implicit investment return, which I refer to as the contractual accrual rate (CAR). In turn these elements determine the trajectory of the value of the pension award over its lifetime. The aggregate of these annual awards constitutes the totality of the member’s equitable interest in the scheme.
At this point, I would recommend the reader to the four short articles in Appendix A. These elaborate and illustrate the equitable interest concept, and touch upon some of the benefits that accrue from this member risk-sharing. I would also note that, going forward, I will frequently draw a distinction between CDC and CDB. With CDB the member’s indicated retirement income is capped, limited ordinarily to the amount “promised”, with CDC it is not.
The article appended illustrate resolution of a few of the potential problems encountered with CDC type approaches.

The key point is that the concept of the equitable interest of a member may be used to maintain balance among members both instantaneously and over time, in accordance with the scheme rules. There are many possible sets of scheme rules and member preferences.

This means that there may be considerable diversity in the detail and behaviour of schemes. Contrary to the beliefs of those who choose to believe that this is harmful fragmentation, it is in fact diversity which strengthens the stability of the financial system, a very considerable public good.
Notwithstanding this diversity, it would be relatively simple to merge disparate schemes, reconciling the differences between them equitably, if that should be desired by members.
It also means that government should not attempt to write detailed prescriptive rules for the terms and management of such schemes. These schemes will not lie within the remit of the Pension Protection Fund, so there is no question of harm to others should a CDC scheme “fail” – though that concept is itself rather strange in a world of no guarantees or explicit liabilities. It was this potential diversity which rendered the concept of a model trust deed and rules redundant.
The interest rate which lies at the heart of member’s equitable interests is the target rate which it is necessary for scheme investments to achieve over time to pay all “promises” in full and on time. Its genesis is the negotiated contribution and award setting process. The rate for a scheme at any point is the product of many year’s individual awards; each of which will continue to have influence in the weighted average rate until the last member of that cohort dies. This means that the investment return target rate is consequentially very smooth over time.
This allows trustees the time to adjust investment policies as economic and financial circumstances unfold. It has moved the problem from the realm of meteorology, where forecasts are unreliable beyond a few days, to the land of climatology, of long-term trends with seasonal variation. Current market-consistent liability valuations are the equivalent of meteorology, which permits only hedging by way of management. Carry an umbrella at all times. Of course, hedging is itself problematic. The only perfect hedge of an asset is its sale – which is hardly consistent with investment. The only perfect hedge of a liability is its purchase from some entity acceptable to the owner of that liability.
Fund Management
The first and most important point is that these have far longer time investment horizons than individual DC; perhaps 75 years versus 40 years. The time scales are similar to open accruing DB arrangements. If we consider a CDC arrangement as an ongoing scheme, it is in fact sempiternal. This in turn means not just that it may collect the investment term premium but also that it admits indirect strategies – taking one step back to leap three forward.
In order to explain the changing nature of the investment policy of CDC schemes, it is necessary to step back and address the problem from first principles. A pension, an income in retirement, is economically simply a claim on future production. The assets we acquire in markets today are claims on some current production and for those that survive, some element of future production. Government may write such claims as it has and will retain the regalian power to tax.

However, the proportion of total production as far in the future as pensions payments that will be accounted for by today’s traded producers is imponderable, though likely to be proportionately much smaller than currently. The challenge for pension scheme investors is to identify those future producers or at least the factors of production that will be employed.
Land is a good example – though current investment practice is concerned principally with the structures built on that land, the offices, factories, houses and shopping malls. Much infrastructure is developed to these long time-scales; it can be expected to play a greater role than hitherto in scheme portfolios. Although it should be said that all too many of the business models currently employed, in for example, the PPP/PFI arena, are far too short-term in nature. Some funds will undoubtedly try to identify the emergent technologies, and “invest” there.
It is worth distinguishing between speculation and investment. Speculation is dependence upon a market for the cash flows which constitute the investment return. If a I buy an equity, the dividend is far less important than the price at which I may sell it. As the holding period length grows, the dividends received grow in overall significance, and usually also in annual amount. In this view, investment is the purchase of an asset, an instrument where the returns arise not from sale in a market but from the cash-flow generated by that asset and available to the investor.
It happens that these fundamentals are far more predictable than market prices. This is “puzzle” which has been known for some forty years – it is a feature of markets. We should expect far greater emphasis on such a fundamental form of investment. And when the long term is considered in this manner, the market price and its volatility are immaterial.

Undoubtedly many schemes will choose not to pay the liquidity premium of markets, preferring instead to invest in the non-traded illiquid equity of private companies. That said, the cash flows of emergent technology companies are highly uncertain, and investment in them is speculative in another sense, they are highly uncertain as to the generation of any distributable cash flows.
The requirement for schemes to invest principally in listed securities introduced by the European IoRP Directive was misguided. The cash flow claims on a scheme with respect to pensions payable are extremely predictable; they may be planned and provided for as long as decades in advance.

Claims arising from transfers, or drawdown are far less so. While this has the potential to disrupt strategies, it need not lead to distressed “fire sales” of assets. Stand-by lines of credit are one obvious precaution against that possibility.
It is obvious that the asset and liability matching studies, and the Markowitz mean-variance optimal portfolios of “modern” financial economics, which are the stuff of nearly all institutional pensions investment consultants’ advice, will be retired to find their place in history. The future portfolio should look very different from the past – it is also likely to be far less volatile.
It seems likely also that more schemes will operate with in-house management of their asset portfolios, as the trend towards greater cost and fee disclosure by asset managers gathers pace.

Risk Management

Appendix B contains a published articles concerned with the risk and risk management of CDC schemes.
There are some preliminaries to be understood when considering the risk and risk management of defined ambition or CDC schemes. The first of which is that the concept of prudence has no place in these schemes. This is rather more than the bias and volatility relative to the best estimate or expectation introduced by these practices, though those do make the funding situation look both worse and more volatile, riskier than is truly the case.

It is simply that it has no meaning when the “promise” is made by the collective to the collective, and does not creative a liability. The equitable interest of a single member serves to ensure solidarity between members and with that fair division of the pie, but it only marginally affects the size of that pie.
The second preliminary is to understand that both the assets and the pseudo-liabilities are likely to be far less variable or volatile than the assets and liabilities of a traditional DB, particularly so when the liabilities of the traditional scheme are valued in a market consistent fashion, or using the methods specified in the pensions acts.
The third is that it is always possible to devise catastrophic disaster scenarios. These are usually underspecified. Take the idea that we might see a doubling of life expectation at age 80. If instantaneous, this would indeed be financially horrendous. However, it would also be accompanied by a maintained surge in gross domestic product of unprecedented proportion. If it were to occur over the lifetime of those currently aged 80, it would be far less impressive.

These issues are well understood in the economics and game theoretic literatures. The use of partial equilibrium models or minor tweaks to game theoretic models can deliver results which are completely at odds with the holistic view. The correct question with such extreme events is to ask how likely they are to occur, and in this case, it is evident that my development over time scenario is far more likely.
Similarly, as we have learned from the US, it is always possible to make foundless accusations. Take the charge that CDC is just a resurrection of the insurance world’s “with profits” policy, and that, as that ended in tears, so must CDC; perhaps spectacularly so as with the Equitable Life instance.

CDC is not a “with profits” policy, nor are these trusts comparable to Equitable Life. As many reports and investigation have concluded, the Equitable Life situation developed as the result of a badly flawed commercially motivated business plan, the pursuit of new business, which saw all gains, including many unrealised, distributed to members.

It also systematically underprovided for the guarantees written, as the correct provision would have led to lower distributable bonuses, and a less attractive, less saleable proposition.
The final preliminary point is that the smoothness of the development of the equitable interest and also, to a lesser extent, of the asset portfolio means that the trustees would have time to adapt.
The primary risk management channel is the pricing of new awards, the determination of the required contribution by trustees. As we have noted elsewhere, the contribution, the projected pensions payable and the implicit investment return (CAR) are simultaneously determined.
Certainly, trustees should have regard with respect to the level of the implicit investment return (the CAR) in order to ensure that members are being offered value for money. In large part this is a question of what might be achieved by the member saver else wherein markets. However, while the rate obtainable by individuals is relevant in terms of value for money for them, that rate need not and usually is not relevant to the trustees. It should also be recognised that this value for money rate is not observable; its estimation is a matter of judgement.
For trustees, the concern lies with the returns that they may earn. In many circumstances, this is not a question of what may be earned by new money in markets but by what may the existing portfolio of assets earn. This would clearly be the case when new contributions are being used to pay pensions and the asset portfolio is unchanged. The framework of uncertainty for trustees in their determination is less than that applicable to markets. This is compounded by the earlier observation, in the section Investment, that the portfolio will likely be biased towards fundamental rather than market performance.
For simplicity, we will divide the risk factors influencing the projection values of pensions into three broad groupings: the accrual award, inflation and member longevity. Only the accrual award is a risk control variable.

Take the situation where the trustee believes that the correct CAR or investment rate is extremely low, and that the contribution rate should be very high, but the employer is unwilling to pay that high contribution. There may for instance be an agreement in place between the workforce and the employee members that the contribution may be no more than a fixed amount. In this instance, the trustees may lower the accrual rate for that year, so that the pension is lowered from that previously indicated in prior years to that which balances the contribution with the future pension sums payable.
Neither longevity nor inflation are true control variables, their outcomes will be what they prove to be. The trustee imperative with respect to these variables is with the best possible estimation of these variables, and hedging the exposures, using standard methods, when the trustee confidence in those estimates warrants this. It appears that modern techniques such machine learning or artificial intelligence can go far in improving the quality and precision these estimation processes.
When, deficits do arise between the scheme’s assets and the aggregate equitable interest (pseudo-liabilities), the trustees would be advised to consider the nature of the causal underperformance.

If this is transient in, for example, the sense of being a shift in market risk preferences, it will offer higher returns to the scheme (and new contributions in particular) going forward. In other words, this is a situation which will self-correct with time. If the impairment is permanent and irrecoverable, then action is warranted, a rebalancing of the risk and return profile of the asset portfolio.

If the “deficit” arises from the misestimation by trustees of the achievable portfolio returns which cannot be rectified by some rebalancing of the asset portfolio, then reductions in the equitable interests of scheme members may become necessary. There are many means by which the reduction may be implemented, including, most obviously, the suspension or limitation of indexation.
The most important thing to recognise is that there is time available to trustees to apply these remedial actions and in the interim, the formulaic risk-sharing described earlier, and in the articles of appendix A, allows equitable risk bearing to continue.
This is very different from the Dutch approach where the emphasis is upon prudence and risk buffers with all being contained within a market consistent framework; that approach is simply inappropriate if we wish to minimise costs and pension cuts. A range of the problems is illustrated by the article in Appendix C.
This is not to deny that CDC arrangements might fail. The last woman standing will have no one to share or pool risk with. Similarly, the sharing of risk among pensioners in payment in scheme which contained only pensioners could only be extremely limited. Indeed, we might even see some irresponsible trustees who were content to permit intergenerational inequity to develop, with the consequence that the scheme becomes unsustainable. However, this is a failing of the trustees and one of the reasons for the earlier suggestion that sustainability or continuity of the scheme should be a trustee responsibility. We can of course write laws and regulations which prohibit stealing, but some will still do so; the prescription is this circumstance is one of sanction.


Many of the practices common in the traditional DB world arose because they minimise the need for data and historical records. The setting of benefits in terms of final salary rather than some career average or other benchmark is a prime example. However, technology has advanced by leaps and bounds and many things which were previously not feasible or outrageously expensive are now mundane.

I have a separate paper in preparation (With Iain Clacher and David McKee of the University of Leeds) but not yet complete, on the subject of information technology and pensions. Everything described within these visions of CDC may be monitored and managed using today’s technologies (including distributed ledgers and smart contracts). The hold the prospects of reducing the administrative and management costs of such schemes to amounts in the low tens of pounds per member.
Such systems are capable of delivering status reports in real time and could also serve as the primary information conduit for members.


There appears to be a widespread belief that rafts of new regulation are needed. Even the Minister has made statements to the effect that significant legislation would be needed to enable and regulate this type of pension.
I do not believe this to be true. Indeed, on close reading of the PSA 2015, it appears that schemes of the CDC type described in this submission could now to created. However, the problem with going ahead and doing that is that it would run the risk of being exposed to subsequently introduced implementation legislation. As for CDB, it appears that most of the required changes are in fact dis-applications of existing traditional DB legislation to this class of scheme.

Responses to the Specific Questions Posed

The responses in this section are a straw man for the final – these are taken from my first pass reading and notes.
The Committee invites evidence from any interested parties on any or all of the following questions:

Benefits to savers and the wider economy:

• Would CDC deliver tangible benefits to savers compared with other models?
Today’s reality is that there really is only one model for new pension provision – individual DC, albeit with varying degrees of asset pooling. These are correctly just tax-advantaged savings schemes. The problem of conversion from a savings pot at retirement is a significant problem and source of risk which has not been satisfactorily resolved. Both CDC and CDB are holistic, addressing both accumulation and decumulation phases. In turn this means that schemes are likely to operate countercyclically rather than in the observed procyclical manner of DC schemes.
It is sufficient for relative stability of member outcome to serve as justification for the introduction of CDC/CDB schemes, and that is a near certainty.
• How would a continental-style collective approach work alongside individual freedom and choice?
The important lessons to be learned from the continental experience concern what not to do. Though there is the occasion counter-illustration, they do not constitute a good model for the UK, or even their own countries. Individual freedom and choice may operate fully under models such as those outlined in the attached files.
 Does this risk creating extra complexity and confusion? Would savers understand and trust the income ‘ambition’ offered by CDC?
With today’s technology, a saver would have available information as to their accumulated equitable interest and the level of funding or security associated with this – both in current and perhaps also projected terms, as a pension income and capital value.
Converting DB schemes to CDC:
• Could seriously underfunded DB pension schemes be resolved by changing their pension contract to CDC, along Dutch lines?
Perhaps, but this should not be done. DB schemes are not irrevocably broken. The attached file “Primer” covers the nature and extent of the problems that have been created for employer sponsored private sector schemes. The more relevant question given the state of decline of traditional DB provision is whether the time and effort involved in reversing this is warranted from a cost and benefit standpoint.
 How would this be regulated and how would the loss of DB pension promises to scheme members be addressed?
As we do not believe that traditional employer-supported DB schemes should be converted, this question is misplaced. The more relevant is the extent to which CDC/CDB should be regulated, if at all. With no explicitly defined liabilities, there is and can be no insurable interest, which means they would lie outside of PPF coverage.
It is possible that private sector insurance based on the investment performance of asset portfolios will develop, but this does not currently exist, and it is obvious that it would be extremely expensive if it did.

Regulation, governance and industry issues:

 How would CDCs be regulated?
As noted above, it is not at all obvious that CDC schemes should be regulated beyond the existing relevant law of their legal form, such as trust law. It may be prudent to explicitly regulate the basis on which transfers under freedom and choice may be made, but this could be managed under scheme rules.
• Is there appetite among employers and the UK pension industry to deliver CDC?
The appetite among employers is almost entirely irrelevant. There is no obligation on them beyond paying the contribution set and agreed with scheme trustees or management.

It is clear that there are some major groups which would like to see the option available to them. While there are some vociferous detractors of the idea in the pension industry, these are a minority whose views are based, in the main on misconceptions and misunderstandings. There are some staunch advocates who will pursue the approach at the earliest opportunity, and if they are successful, we may expect even the most ardent of objectors to then pursue their commercial self-interest.
• Would CDC funds have a clearer view towards investing for the long term?
It is clear that this would be the case. The combination of accumulation and decumulation almost doubles the term of the scheme. It is also clear that these schemes would not follow the (misguided) actuarial practice of subjecting these two phases to differing asset allocations (heavily bond based – and lower yielding – in payment).
It is to be expected that the investment of these funds will try to avoid the speculative volatility of financial markets focusing instead on the productive cash flow generation of the enterprises in which they participate.

One way of distinguishing between speculation and long-term investment is just this -speculation relies upon market sales for cash flow generation while investment relies upon the cash flow generated by the instrument held.
This will greatly lower the dependence of asset performance on market prices – and their unexplained volatility.


Appendix A

This appendix contains four published articles which explore and explain the central features of my vision of CDC and CDB pensions. Of course, other models may be implemented but all that I have examined closely have proved inferior in one or more aspects.

DB and CDC

This article was written before the inquiry was announced.
With Labour calling for completion of the unfinished legislation necessary to introduce collective defined contribution schemes, the usual suspects emerged to decry the concept. Their position really has to be one of wilful blindness given that the most elementary analysis of the institutional structure shows their criticisms to be baseless.
The critical characteristic of a DB scheme is that the employer sponsor underwrites an implicit return on the contributions made by both employer and employee. This is the rate which is needed to be earned to deliver the projected benefits promised. Shortfalls in asset performance relative to this rate result in scheme deficits.
The cure of such deficits is the sole responsibility of the sponsor – as indeed are surpluses. The vested benefits of members are unalterable. The scheme may call only upon the sponsor. The pricing of new awards, rare though they now are, should not include any element of past shortfall repair. It is possible to vary this basic position, but that would require explicit member consent – it appears never to have been explicitly done, outside of sponsor insolvency, in the UK.
Among the rhetoric and invective is the claim that collective schemes are inter-generationally unfair, and that younger members are profoundly disadvantaged. In both DB and CDC, the uniformity of the contribution rate, and the common accrual for a year of service are intra-generational risk-sharing. They are not systemic transfers from one class of member to another.

The diagram below shows the amount of transfer between the oldest and youngest active members of a collective scheme for a range of contribution rates, and a representative pension payment schedule. It shows clearly that the arrangement is risk sharing, not cross-subsidy. It also shows the average discount rate (or CAR) associated with a particular contribution rate.
Diagram 1diagram one

It may help to understand this phenomenon by considering the proportion of the pension ultimately payable that is simply a refund of the contribution made. In the case of a 10% contribution rate, this is 3.5% for the young member and 15.9% for the elder, but when the contribution rate is 50%, it is 17.7% for the young member and 79.5% for the elder.
In the sense of the workforce, this arrangement is intra-generational; in a time-serial setting it is a question of overlapping generations. It is not inter-generational in nature. It has decidedly positive incentive properties for the younger members of schemes. They give up value to elder members when returns are high, their future rosy and projected pension benefits substantial, in other words when they can best afford it, and they receive value when returns are low, the future bleak and they most need it. It is an ideal macro-risk management mechanism.
However, in a defined benefit arrangement this collective pooling and sharing arrangement is immaterial to the scheme member; their pension benefits are fixed. The sole risk that they face is that of sponsor employer performance under the underwriting guarantee, and practically this reduces to sponsor insolvency.
In a CDC arrangement, there is no sponsor underwriting guarantee; this role is assumed by the members collectively. There is no meaningful concept of sponsor insolvency here, as this is a case of members collectively owing a “debt” to themselves. The collective does now face investment, inflation and longevity risks. Unlike DB, CDC scheme trustees have a role in managing these risk exposures. The risk-pooling and sharing properties among members become relevant. Unlike DB, where the sponsor employer is usually the residual claimant, these trustees have a singular responsibility to beneficiary members, their heirs and successors.
Should deficits arise, it is not practically feasible to call upon members for repair contributions; collection of such calls would be profoundly problematic given the varying circumstances of members.

In addition to repair through the risk channels, investment, inflation and longevity, the benefits attributable to members may be varied. In the latter case, varying benefits in proportion to a member’s accrued claim would be equitable. It should also be realised that any such variation can usually be achieved, with considerable flexibility, by utilising the time dimensional properties of the scheme. After all, the scheme is simply a reflection of the collective consumption preference of members over their life-times.
The annual pricing of new award contributions is a matter for negotiation between the scheme and employers. Where a long-term contribution rate has been agreed with sponsor employers, this may be accommodated when considered necessary, by variation in the terms of award. New award pricing may contain elements of deficit repair, but it is important that this should be arranged in a manner which is equitable among all members. This applies equally to those entering the scheme solely for the purpose of using it as a decumulation arrangement.
The one thing which is patently obvious is that the mutuality of a collective defined contribution schemes is viable, and that we can expect from them delivery of attractive pension benefits over the long-term.
For more, see:

A Vision of CDC

Recently the TUC published a study, commissioned from and executed by the Pensions Policy Institute, of the variability of DC outcomes. They consider the fund performance and (annuitized) pension income for individuals retiring in each of the years 2000 – 2017. The results are rather stark. Pension pots have varied from £214,000 to £308,000, and the annuity income from £27,871 to as little as £14,484. The volatility of pot outcome was modest, just 11.2%, and that of annuity income 17.7%.

The situation faced by an individual is actually worse than this – the study assumed benchmark performance for all, while the reality is that there will be considerable dispersion of performance across the various funds they have chosen as their investment vehicles; volatility greater than 11% is commonplace.

As the TUC’s Tim Sharp observed:

“Retirees are being increasingly faced with a blizzard of risks that are bringing new sources of uncertainty and chance into financial arrangements for old age.”

Collective Defined Contribution schemes are an arrangement to mitigate this uncertainty by the pooling and sharing of risk among members. They are member mutual. They do not need any corporate sponsor – where these are workplace based, beyond making timely payment of contributions for new awards, there is no obligation or responsibility assumed by the employer. Though they may take a number of legal forms, the tradition of trust based arrangements in the pensions world makes this an obvious choice – with trustees elected by members and perhaps, from among them, and a small executive management group appointed by the trustees.
The member is “promised” a pension accrual in similar manner to DB – say, 1.5% of final salary inflation indexed, or 1.75% CARE, for each year of contributions. These contributions may be made by either or both employer and employee. There is risk-sharing among members in this arrangement. It subsidises older members pensions at the expense of younger, when investment returns are good, and younger members at the expense of older when expected returns are low. The contribution made and this “promise” define the contractual accrual rate (CAR) – the implicitly indicated investment return on contributions. The “promised” pension ultimately payable is projected using standard actuarial techniques. These three terms, contribution, projected pension and CAR, define the equitable interest of a member in the scheme.
The equitable interests of three members, for a single contribution, and their trajectory over time is shown below:
Diagram 2.concdc
The sum of the equitable interests of members are equivalent to the scheme liabilities under a defined benefit arrangement. Projections are updated with the passage of time to reflect survival, experienced inflation and other relevant risk factors, and with that the equitable interests of any and all members.
This terms of the “promise” for new awards are revisited from time to time by the trustees to ensure that it is in line with expected returns from financial markets – that it is offering value for money to members. The “promise” is not a guarantee; it is not inviolable. Indeed, the pension paid may be greater than this “promise”, not just lower. The pension paid in any year is determined by the value of the asset portfolio as a proportion of the equitable interest of the member. This is effectively a funding level. If assets have performed well this will result in higher than “promised” pensions. This equitable share of funding arrangement would also ensure that members will receive a pension over their entire lifetimes – it will also allow the distribution of the member’s complete savings.

There need not be any retentions for risk buffers or capital retentions, with their potentially stranded assets.
Members may transfer the asset equivalent of their equitable interest at any point in time, subject to the usual restrictions that they be into qualifying arrangements. Members may decide that they wish to enter a drawdown arrangement rather than the scheme default pension income arrangement. Taking a larger drawdown than that pension determined by the funding level will lower the member’s residual equitable interest in the scheme. Members may be admitted in either the accumulation or payment phase – such schemes could serve simply as decumulation vehicles.
The asset portfolio is subject to point in time price volatility as was shown by the TUC illustrations. The CDC arrangement exposes only the current year’s pension to that risk. If a decreased payment is unacceptable to a pensioner, he or she may opt to top up that year’s pension payment at the cost of a lowering of their residual equitable interest going forward. Active contributing members may voluntarily choose to make additional contributions when the asset funding is less than complete; this would increase their equitable interest and restore the amount of pension payable.
As a multi-period investment arrangement, the scheme benefits from pound cost averaging when net cash flow positive , but suffers the converse, sometimes colloquially referred to as “pound cost ravaging”, when net cash flow negative.
Many whistles and bells may be introduced – for example the scheme could operate as a provident fund, where borrowings are permitted from the fund for qualifying purposes, such as medical emergencies, house purchase and educational costs. This is also true of many of the other benefits associated with the traditional DB model, such as disability pensions.
Contrary to the assertions of some commentators, such schemes do not need huge numbers of members, though size does ensure risk-pooling has maximal effect. Nor do they need to commence with large numbers of members.
The complexity introduced through the concept of the member’s equitable interest is well within the capabilities of today’s technologies – and the absence of legacy systems is a positive in this regard. This technology could also serve as the primary information channel between trustees and scheme members.
There is one further way in which schemes might reduce asset volatility: by risk-pooling and risk-sharing the performance of part or all of their asset portfolios with other schemes.
CDC schemes are not exposed to sponsor insolvency risk. The concept of insolvency for a CDC scheme is vacuous in that these are merely obligations owed by members to themselves, and linked directly and equitably to the current asset value. There is absolutely no need for regulation, even though the proper concerns of trustees, unlike DB, would now include paying “promised” pensions as they fall due.
The initial public problems with DB arose from an inequity – pensioners in payment had priority over active members. All of the regulation and interventions we have seen since have not resolved this problem fully and have introduced another set of inequities – in this case between the stakeholders of a firm. CDC, as envisaged in this article, resolves all of those issues.

A Vision of Collective Defined Benefit (CDB) – (DA by another name)

An earlier article in this appendix presented a Vision of Collective Defined Contribution (CDC) pensions. This article presents my vision of the collective defined benefit counterpart. In common with the earlier blog, it uses the concept of the equitable interest of a member of the scheme in the scheme.

As with CDC, the organisation is collective among members; there is no sponsor or external guarantor of the scheme. The role of an employer is limited to the payment of contributions as agreed with trustees  for years of qualifying service. The employer has no liability beyond this.
As with CDC there is a “promised” pension for each year of contributions. This is not guaranteed; it is not inviolable. Unlike CDC, this “promised” amount is strictly capped – it is a defined benefit. Should returns exceed the implicit “promised” rate, these excess assets are orphan assets held by the trust, enhancing the security of its “promises”.
If the scheme is in deficit in a particular year, that is to say that scheme assets are less than the total equitable interest of members, then pensions are not ordinarily cut. A risk sharing mechanism among scheme members kicks in.

After payment of those pensions in full, the equitable interests of non-pensioner members are augmented (in proportion to their equitable interests) by an amount equal in total to excess payment. This is the quid pro quo of a risk-sharing arrangement within the differing classes of members. The equitable interest of pensioner members is unchanged. The terms of new awards for actives are unchanged from those prevailing prior.

This risk-sharing arrangement has the effect of raising the effective security of active and deferred members relative to pensioners after payment of the excess pension.
As with CDC, pensioners may opt for drawdown, but that will be strictly limited to the level of funding of the scheme relative to the equitable interest, capped at 100% of that. Subject to the usual provisos, members may transfer at any time: again, based on the funding level.

As with CDC, drawdown or partial transfer will lower the member’s equitable interest. This provides an incentive for members to remain in the scheme and to draw pensions as lifetime incomes.
As with CDC, the terms of the “promise” for new awards are revisited from time to time by the trustees to ensure that they are in line with expected returns from financial markets– that it is offering value for money to members.
Some have criticised the access to pension savings through drawdown or transfer as potential “moral hazards”. The problem in question being that members who know they are ill and unlikely to live to the population average lifespan, will draw their “pots” early, to the detriment of the collective. This certainly may be possible, but it is totally swamped by another behavioural anomaly, that we systematically underestimate the age to which we will live, with the consequence that many who should not cash in early do. The set-up of CDB does offer an incentive to remain and take the socially desirable pension payment option.
Some have argued that new members will not join a scheme which is in deficit. However, new awards, to both new members and ongoing actives, are made on the basis of the expected returns on those contributions – that they offer value for money. These define the member’s equitable interest and the value of the full pension. The level of scheme funding is only relevant with respect to transfers out of the scheme.
There is one scenario which deserves thought, that the scheme closes to new members (and possibly also further awards to existing actives), entering a long-term run-off. The first point to note is that members may withdraw the funded value of their equitable interest at any time – this is a situation in which they may be accepting lower benefits than promised. Risk sharing among the classes of member may continue, equitably, until there are only pensioners remaining. This is the point at which pensions may have to be reduced if investment performance has been inadequate. There are still a range of options other than liquidation and stand-alone run-off, such as transfer to another CDB scheme.
It is worth noting that the adjustments being made are rather small – say a scheme is 80% funded, with the pensions payable in that year being 3% of total scheme equitable interests, and that 40% of the scheme membership equitable interest is pensioner in payment, then the adjustment to active and deferreds’ equitable interests would be just: (1-0.8) * 0.03 * (0.40/0.6) = 0.4%.
While the mathematics of long versus short term investment returns may rapidly become complicated, it should be remembered that savers should prefer low prices and high associated returns. The declines in asset prices that trouble pensioners in payment add to the returns of new saving members. This is temporal smoothing which enhances compound returns.
Some have argued that CDB will require rafts of new regulation and legislation. My challenge to them is to ask them to identify the financial or economic failure that warrants any regulation.
This vision of CDB is not at all like that prevalent in the Netherlands. It really isn’t – but the rebranding of Dutch DB as CDC or CDB was just that a rebranding rather than a reconstruction of a flawed DB system. This design avoids the problems that have become evident there. One of the major lessons from the Dutch experience is that regulation of pension schemes as if they are insurance companies is a gross mistake.

The Design of Collective Schemes

The inquiry into Collective Defined Contribution pensions announced by the parliamentary work and pensions committee has led to a welter of objections to the concepts. Among these concerns, misconceptions abound. Indeed, some are even evident in the introduction to the inquiry offered by that committee.
The critical element to understand is the role of the “promise” in the design and construction of both CDC and CDB arrangements. This defines and quantifies the equitable interest of members in the scheme. It is determined by contributions made and the projected values of “promised” pensions that would ultimately be payable for each and every member.
The initial public problems with DB arose from an inequity – pensioners in payment had priority over active members. All of the regulation and interventions we have seen since have not resolved this problem fully, or satisfactorily, and have introduced another set of inequities – in this case between the stakeholders of a firm. CDC resolves all of those issues.
The equitable interest of a member, relative to the totality of member interests, is the proportion of scheme assets which may be transferred or withdrawn at any time. It enables all of the flexibilities of Freedom and Choice.
For CDC, as for ordinary DC, if the assets have outperformed (relative to the equitable value) then the pensioner may draw their entire equitable share. If the assets have underperformed, then again, they may draw their fair share. The “promise” made defines a default pension payment schedule, resolving the point in time uncertainty of annuity purchase on retirement.
If scheme assets are above the equitable value, the “promised” payment or more will be made, a member option. If scheme assets are below this equitable value, the pensioner may choose to ‘top up’ their current payment, though at the cost of lowering their residual equitable interest in the scheme. This may be attractive to the pensioner from a consumption smoothing standpoint. There are no transfers or subsidies among members, intergenerational or other, in this action.
By contrast, with CDB, as with DB, the “promise” defines the maximum which is expected to be paid. If scheme assets exceed the totality of members’ equitable interests, the excess is not distributable. These are orphan assets which serve to enhance the security of the scheme. The “promise” with CDB is, in a sense, harder. Here pensions will be paid even when the scheme assets are below the aggregate equitable interests of members. In this situation, in order to maintain equitable balance, the interests of non-pensioner members are raised to reflect the excess payment made to pensioners. The pensioners’ claim on the assets of the scheme are unchanged, while the non-pensioner members’ increase. There is no intergenerational inequity.
Such misplaced concerns are unfortunate, as they distract from the advantages and gains from the other risk-sharing and risk-pooling properties of collective mutual schemes. Some entirely new possibilities arise. For example, it is even possible for CDC and CDB to co-exist, with transfers between them taking place at the value of the member’s equitable interest. CDC is very much a concept for our times.




















Appendix B

This appendix contains a  published articles on the risk management and accounting of CDC schemes.

Defined Ambition – A Trap Avoided

Suppose that we are offered an investment promise  with some implicit compound rate of return (contractual accrual rate, CAR), say 6.47%, on a contribution of £10 – the pay-off to this promise is a defined income in retirement. Then obviously we expect this to be worth £10.65 after one year, and let us assume that this contribution has earned the requisite 6.47% in that year. The contract is performing as promised.
Now further suppose, that we wish to make a further investment with a similar pay-off, the next year. However, for whatever reason the implicit rate of return (CAR) that is offered is only 4.42%, then we need to find £30 as the contribution .
At this point, we have assets of £40.65 supporting the total claim. Now if we apply so called fair value, or market pricing to both contracts, then we arrive at a valuation of £60.00 for the total portfolio of two promises. The expected return on assets valuation method of pension regulation will deliver the same result. The “scheme”, the aggregate of these, is reported as 67.75% funded, even though the writer of these contracts has performed fully.
Of course, if this were a regulated pension, the writer of the contracts, the sponsor employer would now be pressured to make good this “deficit”.
Now let us assume that we have two different, but otherwise identical, investors in each year, and that an investor may make withdrawals from their contract, or transfers, and that this will take place at the aggregate net asset value. Then the initial investor who contributed £10 in the previous year, on withdrawing, will now get to take away £20.32. This investor has more than doubled the initial investment, making the remarkable return of 103.3%, when all that was initially promised was 6.47%. The second later investor has security only of £20.32, less even than their contribution of £30, and can only realise a loss.
The equitable interest of the first is £10.65 and of the second £30. The sponsor is facing a demand for an additional £9.68 contribution.
The moral of this tale is simple: the application of fair value, or mark to market accounting, or the methods specified in the Pensions Acts (which are counterfactuals) for valuation will result in inequitable outcomes. In the case of DB schemes, it is usually the employer sponsor who bears these costs, but “Freedom and Choice” extends this possibility to scheme members. In the member mutual collective world of Defined Ambition schemes, it is certainly the case.
The correct form of accounting for DA (or CDC, CDB or whatever) involves the calculation, monitoring and subsequent aggregation of the scheme members’ equitable interests, a matter of historical fact, and pseudo-solvency, the comparison of that with the market value of assets held. It really isn’t difficult.

Appendix C

Illustration of some problems with the Dutch approach

The €300m Dutch pension fund for the travel sector has said it was looking for a merger with a larger industry-wide scheme or a switch to defined contribution (DC) arrangements.

Reiswerk Pensioenen made clear that continuing independently was not an option because of its funding position and its predominantly young membership.

“The long duration as a consequence of our young population requires taking more investment risk,” said Frank Radstake, the scheme’s employer chairman. ”But the financial assessment framework (FTK) doesn’t allow us to do so because of our funding shortfall.”

The pension fund was therefore “stuck in the FTK trap.”

The small sector scheme has been in trouble since the introduction of a new and lower ultimate forward rate (UFR) in 2015, part of the discount mechanism for liabilities.

Despite a defensive investment policy, including a 90% interest hedge, the scheme’s funding level has since plummeted from 125% in 2014 to 99.5%.

In its annual report, Reiswerk Pensioenen said it was unable to hedge against the impact of the new UFR, which had caused a steep rise of its long-term liabilities.

Radstake said the UFR was expected to drop further and that the current asset mix would not solve the pension fund’s financial problems.

He indicated that the scheme’s small size – it has 9,000 active members – was another reason why it wasn’t deemed future-proof.

Joining another sector scheme or switching to DC would both give Reiswerk the option to increase the risk profile of its investments.

A merger would bring the added benefit of scale, but would also mean an instant rights cut for the scheme’s members.

“However, in the long term the results will improve,” said Radstake.

Based on its current position, the pension fund was headed for a rights discount in 2021 anyway, he noted.

The social partners in the travel industry said they wanted to make a decision about the scheme’s future next summer.

About henry tapper

Founder of the Pension PlayPen,, partner of Stella, father of Olly . I am the Pension Plowman
This entry was posted in CDC, pensions and tagged , , . Bookmark the permalink.

6 Responses to Con Keating’s prepares his thoughts on CDC

  1. Wow. I think, thanks to these postings, that I’m beginning to understand a bit more about the issues.

  2. Brian Gannon says:

    Really interesting, informative and enlightening explanation of CDC. In spite of this potentially being a stupid question, I will ask it anyway. Let us say the member retires at 65 and takes benefits. Let us say that for the sake of argument they take their pension at age 65 for one year and then die. What if any value does the members equitable interest in the scheme have for the members spouse/child/estate on death? And if any equitable interest remains, is there an option to continue taking a scheme pension for the survivor and/or transfer out to their own individual DC arrangements? Let us pretend instead of CDC the member had a drawdown pot of £300,000, and had taken £12,000 out, and had enjoyed post charge growth of 4% then the death benefit to his family/estate would be £300,000. How would the death benefits work if instead this person were a member of a CDC scheme?

  3. Brian Gannon says:

    And as a further question, would there be a theoretical possibility that a CDC scheme would be something that people with existing individual DC pots could transfer into with existing lump sum pots? And in theory could they do this post retirement/whilst already in income drawdown?

  4. henry tapper says:

    Brian , I am going to post the first question to the Friends, the second question is easy for me to answer, the transfer in of DC pots should be encouraged. You may remember that this used to be possible for DB – with pots exchanged for pension. In future , I’d hope that people would regard CDC as a viable replacement for that service, of course the pension would not be guaranteed, the difference between a CDC pension, an annuity and a DB pension,

    • Brian Gannon says:

      Hi Henry, that’s very kind of you and thank you for inviting me to join Friends of CDC, invitation gratefully accepted. As someone specialising in the benefits of transferring or not transferring DB schemes, and also being an at retirement specialist, it would be great to have a better solution than individual annuities and the current range of funds available for those with drawdown pots. There is a minority, but a sizeable one nonetheless, of people for whom transfer out of DB schemes makes sense, but there remains the challenge of how best to provide income for such individuals (although in practice most of those transferring with me are not usually looking or needing to use their pension for their own income). On a far far wider scale, for those with DC pots it would be great to have another solution than funds and/or annuities at a time when rates are pathetically low.

  5. Con Keating says:

    I responded to your first question on the linkedin site so you have doubtless seen that. As to whether you might transfer in a pot of money. There is certainly no reason other than the particular preferences of a scheme why you should not be able to do that. It would be moving into a hybrid situation where you would have options to annuitise on set terms or to follow a self-directed drawdown path, all while retaining the ability to switch to another scheme if you should so choose.
    If it turned out that people were churning among their options at excessive cost to the scheme it might be necessary to add an administrative charge for those chopping and changing – as those costs should not be borne by the others not doing so.
    I have an open mind as to whether a member in receipt of a pension might make additional contributions into the scheme, other than from another scheme. The concern there is potential tax relief recycling.

Leave a Reply