This note has been written to accompany a Pensions Management Institute virtual panel taking place on April 22nd on the topic of CDC pensions. It takes, as a straw man for some of the issues, a recent blog by Stefan Lundbergh. (See: (7) Miracles and mirages of CDC | LinkedIn )
Let us say, expressly, that we found this blog to be balanced and reasoned – a breath of fresh air in what is otherwise an all too often heated and boorishly bad-mannered debate. A particularly pleasing feature of it is that it is framed in general terms and not fixated upon the specifics of the proposed Royal Mail scheme.
The blog does us a great service by describing accurately the history of the transition from DB to CDC in Holland, though for completeness, add that there were also a handful of schemes created in the early 2000s which were from inception and by design CDC. It is obvious from this historiography that the problems which did emerge could have been predicted to emerge – and probably were by some. There is much contained within the blog with which we concur, notably on the power of diversification. However, there is one expansion of that which deserves mention.
If we reduce portfolio volatility by diversification, then the long-term realised compound returns will be higher for a given set of arithmetic security returns than the expected value of those returns otherwise would be. Put another way, the geometric compound mean return of a portfolio approaches the arithmetic mean as volatility is reduced.
The blog contains the following: “The idea is to base today’s pension payments on future expected investment returns. This means that we need ‘objective’ assumptions on longevity, future interest rates and stock market returns. Unfortunately, the track record of ‘objectively’ predicting longevity and interest rates has been extraordinary poor. In addition, when a CDC scheme comes under financial pressure, so does the ‘objectiveness’ of the assumptions.”
It is clear that new awards of pensions should be based upon the trustee expectations of investment returns and best estimates of the factors which determine the amount of those benefits. Any other arrangement would be unlikely to meet tests of fairness to members and would render a voluntary arrangement unstable. As we have stated on numerous occasions, it is inappropriate to value the liabilities of scheme using discount rates chosen on the basis of the expected returns on their assets. (See for example, “Discount Rates, Defined Benefit Pension Schemes and their Sponsors” which can be found Here). We recognise that CDC liabilities are best efforts promises, not the contractual obligations of DB schemes, but the valuation technique and in particular the discount rate should be the Contractual Accrual Rate (CAR). This rate has its roots in the implicit rates of return embedded in awards. We agree that the track-record of prediction has been extremely poor, but as the CAR is updated to reflect the experience of these factors to the date of valuation, and their variation from the prior values, this does not become a problem. The CAR will converge over time, over the life of an award, to that which has been achieved.
Intergenerational risk sharing
In a paragraph entitled “The mirage of intergenerational risk sharing” Dr Lundbergh avers “Any attempt to provide a coherent line of reasoning requires the introduction of several highly hypothetical assumptions. The problem is that even if our logic is impeccable, it is only valid if the assumptions hold. The two critical assumptions behind intergenerational risk sharing are that participation is mandatory and that the arrangement will continue perpetually.”
Unfortunately, this argument is unsound. Voluntary arrangements which deliver this risk-sharing can be devised and perpetuity is not necessary. It is sufficient that the scheme operates sine die, with no pre-determined end date. The necessary condition for such a voluntary arrangement to succeed and continue is that it should be equitable i.e. that is fair among members. This applies to the current members and those as yet unborn. The motivation for members to join and remain is not philanthropy or benevolence, but rather self-interest and the common good.
An important precursor for success is that the terms of award should be competitive with alternative possible savings arrangements. This means that expectations of future returns will be an element.
We agree that given that “there are no clear definitions of what Collective DC actually is.”, so in this note we will offer several straw men to give substance to our view.
A Defined Straw Man
We offer as a straw man, the traditional defined benefit award as a suitable structure. This has a uniform contribution rate as a proportion of pensionable salary together with a uniform rate of award at retirement regardless of the age of the contributing member. The benefits are typically linked to some measure of price inflation. Here we take contributing member to encompass the contributions made by an employer for the benefit of the member. This arrangement has been widely criticised as being unfair among members. The idea that a 64-year-old should receive the same award as a 25-year-old runs counter to the time value of money.
It should be noted that so long as we do not breach the age discrimination legislation, as an employer we can devise very flexible remuneration packages and many companies do just that e.g. one which rewards loyalty and experience or one that encourages younger employees to join because of their more up to date skills in certain fields.
The traditional argument was that the young member could in time and turn expect to benefit in that same way as today’s 64-year-old. This argument has fallen from favour with the often-cited view that labour mobility has increased; few of today’s 25-year-olds will be employed by the same sponsor firm and members of the same scheme in their fifties and sixties. The source of some of the statistics e.g. that the average millennial will have 12-15 jobs over their lifetime seems to come from the US, but more recent evidence from the US suggests that labour market flexibility has reduced.[i] While we are more than willing to accept the idea that increasing job mobility is present, this would need much more careful decomposition in the UK as it is not possible to extrapolate from one country to another and there are measurement issues e.g. When does switching occur, under 30? What counts as a switch, is it firm or industry?
Regardless, it is perfectly possible to have the CDC scheme independent of any sponsor. It could be the member’s scheme in much the same way that members have bank accounts. As the member changed jobs, it would continue to have contributions paid to this account.
We will now introduce a few numbers to aid the exposition. Let us suppose that the young member has a salary of £25,000 and that the senior’s salary is £67,126 with a contribution rate of 20%. If investment returns are 5%, the future value of the young member’s invested contribution over the next 40 years would be £35,200 while the future value of the older workers would be just £14,097 after 1-year at which point they retire. This disparity in the pricing of this one year of award is the basis on the ‘intergenerational unfairness” criticisms. The older member appears to be 2.5 times as well treated as the younger. However, it is far from complete in that the life expectation of the young member at retirement is far longer than that of the senior member. If we take the life expectations at age 65 to be 23 years for the older member and 29 years for the younger, with inflation at 2%, the younger member’s total lifetime pensions received will be 1.35 times the seniors. This would still leave the senior better treated than the young – we can quantify this as being 1.85 times as well treated as the younger member.
But this is still an incomplete analysis, the younger member faces and constitutes a far greater uncertainty and risk to the scheme than the older member. This is uncertainty is its broadest sense and risk is simply the quantifiable subset of that (and a very small part of it at that). It is present in all of the variables that contribute to the pension amount and duration once awarded. If this risk were uniformly distributed across pre-retirement and post retirement periods, the younger would be 1.7 times as uncertain and risky as the older. The young and old are now approaching relative parity in relative treatment terms. Of course, the levels of uncertainty and risk are not uniformly distributed, the pre-retirement period is significantly more uncertain ex ante than the post-retirement. This tips the balance in favour of the younger member.
We must, of course, observe that these relative treatment ratios are highly uncertain. However, this arrangement clearly conforms with the Rawlsian concept of fairness[ii] as arrangements which are constructed and accepted behind a “Veil of Ignorance”. As such, a willingness to be part of the system is due to the fact that their likelihood of a better outcome while not costless i.e. they may have to give something up, is much more secure than going it alone. Member motivation is therefore simply that they want to do well for themselves and are prepared to conform to reasonable terms of cooperation, provided others also do so.
Scheme Design and Operation
The question then moves to the design of the structures and operations of the scheme which will maintain fairness among members including incoming member and is a question of how to achieve equitable treatment of all. This is a collection of rules for risk-sharing among members, which incidentally could be hard-coded as a set of computer based ‘smart ledgers and contracts’[iii], eliminating the possibility of time inconsistencies or human intervention by well-meaning trustees, rendering redundant Dr Lundbergh’s concerns with the ‘objectiveness’ of assumptions in times of stress.
This need for consistency and certainty is concerned with what we have and the current circumstances. It does not conflict with the blog’s concluding paragraph: “In a complex world of many unknowns, the only thing that I am certain of is that things will continue to change – necessitating pension solutions that are adaptive as the conditions changes. … My recommendation in the UK debate is to keep our system simple, flexible and adaptable to the rapid changes challenging and shaping society today and in the future.”
The decision criterion for alterations to ‘promised’ benefits is the funding, or solvency ratio. This compares assets valued at market prices with projected benefits discounted at the contractual accrual rate. This ‘liability’ value represents the total equitable interest of all members in the assets held.
The most important general criterion for the maintenance of equity among members is that all alterations to pensions are applied equally to all. The mechanism for doing this is alteration of the overall equitable interest and those of individual. One way of thinking of the equitable interest of a member is as this being a number of units held by the member, just that these units are expressed in notional pounds. A member’s equitable interest, expressed as a share of the total interests of all members, defines their claim upon the assets of the scheme – and would also define its transfer value.
It is also important that, in the absence of operable risk-sharing rules (as we describe later), a funding deficit should be rectified immediately on occurrence by the cutting of all benefits (promised and in-payment). The general principle is that if the scheme is in deficit, then all pension entitlements will be cut equally. The mechanism for doing this is a similar proportional cut to the equitable interests of all members unless the risk-sharing rules are operative. This leaves the scheme attractive to new members and to existing members for future accrual (as there is no deficit for the past to be subsidised).
We will again offer a straw man; this time for the design of a set of risk sharing rules among the members of a scheme. The purpose to these risk-sharing rules is to smooth the variation in pension payments and values that might otherwise arise from transient volatility in the value of the scheme’s asset portfolio (and with that the funding ratio).
With one exception, the risk-sharing rules only come into operation if there are pensions in payment and the total scheme assets are less than the totality of member’s equitable interests, that is the scheme, in traditional terms, would be considered in deficit. In the absence of any pension payments, the only outgoing cash flow to members would be transfers out, and these would take place at the net asset value of the equitable interest of the departing member.
The exception is the case of new awards and contributions made when the scheme is in deficit. These new members would ordinarily be unwilling to contribute to a scheme which was in deficit when their contribution and benefits are in balance – it would give rise to dilution of their benefits. It is in any case a poor idea to use new awards and contributions to subsidise existing members. By equal part, the distribution of surpluses through over-generous awards is to be avoided. The solution is simple. It is to increase the notional equitable interests of new members such that these produce a similar deficit to that prevailing before their introduction.
One of the frequently observed and unfortunate aspects of mandatory systems is that they do tend to use new award contributions to repair prior deficits. This would be an abuse of new members and is clearly intergenerationally unfair. Mandatory systems really should come with an anti-abuse rule.
So, to the straw man rules. There are two dimensions to these; they define the maximum amount of support that may be offered to pensioners in payment from all other members of the scheme, and the maximum time that may elapse before a deficit is rectified by an overall cut in the equitable interests of all members. They operate only on current payments in the sense of determining if a pension due may be paid in full.
The two rules are:
- A deficit should be repaired by cutting benefits if the deficit has not been cured within a period defined inversely to the magnitude of the deficit. A 50% deficit would have a cure period of 2 years, a 20% deficit 5 years, and so on. In the case of multi-year deficit sequences, the rule operates on the maximum deficit.
- A pension may be paid in full only if there is available risk-sharing capacity. The capacity is defined as a cumulative 10%[iv] of the equitable interests of non-pensioner members.
The amount of support given to pensioner members in any year would actually be quite small. To have 5% of a scheme payable as that year’s pensions would be quite normal and a deficit of 10% would imply a ‘subsidy’ of 0.5% of the scheme. If we suppose that pensioners and non-pensioners are equal in their shares of the overall scheme equitable interest, this support represents 1% of the equitable interests of non-pensioner members.
The key to maintaining fairness or equity among scheme members is to increase the notional equitable interests of non-pensioner members by a similar amount to the ‘subsidy’ they have afforded to pensioners by full payment of the pension. In the illustration above, this would be 1% of the equitable interests on all non-pensioners. This rebalances the extent of their claims on the assets of the scheme. In other words, they are rewarded and compensated for their risk-sharing.
There are many further details and minutiae which might be considered, but the purpose of this article is only to illustrate that CDC schemes which are fair among members and over generations can be constructed. Of course, schemes which rely on variation of a benefit factor, such as inflation indexation, to restore solvency balance will not, in general, be intergenerationally fair. In this case, with cuts, those with the longest life expectations will suffer the largest reductions.
Incidentally, simulation of the above rules using the performance of financial markets since 1921 suggest they are, as stated, sufficient to avoid any cuts, though there are periods in the 1930s and early 2000s when they would have come close to requiring cuts.
An incidental to this system of risk-sharing among members is that it permits an ongoing growth investment strategy, and therein lies one of the major advantages of CDC over individual DC. The risk-sharing described implies that CDC schemes are inherently mutual bodies and that they should be governed as such. These are pensions, but not as we know them.
[i] Labor Market Fluidity and Economic Performance, S. J. Davis & J. Haltiwanger, National Bureau of Economic Research, 2014.
[ii] John Rawls (1999) A Theory of Justice: Revised Edition, Cambridge, MA: Harvard University Press
[iii] See: Clacher, Keating and McKee (2018), Smart_Ledgers_And_CDC_Pensions.pdf (longfinance.net)
[iv] This amount could be some other agreed value.