Judgement, Discretion, Risk and Predictability in CDC Pensions – Con Keating


People reading this blog often finding themselves reading this blog


In my vision of CDC pensions, there are numerous mechanisms for limiting the risk exposure and increasing the predictability of pension incomes. There are in fact far more than available under traditional DB, including several which enhance the solidarity of scheme members.

First there is the contribution setting mechanism. Here, in addition to varying the employer’s and employees’ contribution, the trustees may vary the annual award accrual. For example, when there are restrictions on the amount of employer contribution, the trustees may, if their return expectations are low, set a lower annual award amount. This form of risk management maintains the predictability of the pension income, even as it lowers the increase in active members’ new award equitable interests, the expected pensions receivable. Variations in the equitable interests of scheme members are not of immediate concern to scheme members, other than in the context of transfer values.

By offering a uniform annual award accrual rate regardless of age, there is an element of risk sharing among members embedded within the scheme. When expected returns are high, the old benefit relative to the old, and when expected returns are low, the young benefit relative to the old. This element is present in traditional DB but rendered redundant by the sponsor guarantee.

Of course, the trustees may also manage in traditional manners the expected risk and return of the asset portfolio, as well as exposures to exogenous risks, such as longevity and inflation. However, rather than managing the assets relative to a market-consistent present value of liabilities, the trustees are now managing to a target arising from the contractual accrual rate of the total scheme equitable interest, which, as it does not embed the ‘animal spirits’ of capital markets[1] is far smoother and more predictable. Scheme variability and risk tends to approach that of just the asset portfolio. The asset portfolio objective is unconflicted.

The equitable interest of a member together with its asset coverage value define the value of the pension, in both capital and income terms. This provides values for annual and lifetime allowance calculation and facilitates transfers. With one member one vote in the election of trustees, the member has both exit and voice in the governance of the member mutual.

When funding coverage of member equitable interests is less than whole, pensioner payments are at risk. If pensions in payment are cut, then the equitable interests of non-pensioner members are also cut in similar proportion – this maintains the equitable balance between members of differing status. It provides an incentive such that non-pensioner members should wish to support pensioners in payment. It also enhances the relative position of pensioners in payment after cuts have occurred.

The support mechanism for pensioners in payment when coverage is less than perfect, is simply to continue to pay pensions in full. This is accompanied by an increase, of similar magnitude to the support offered to pensioners, in the equitable interests of non-pensioner members. It provides them with an incentive to support those pensioner members.

With this combined incentive structure, solidarity among members might prove excessive, and it raises the possibility of the scheme entering an asset ‘death spiral’ from which recovery is impossible. With this in mind, there needs to be limitation upon the total risk sharing within a scheme.

This has two dimensions. First, there should be a forbearance period during which the risk-sharing will operate. This time should be related to the magnitude of the support. A shortfall of 10% may offer a ten-year forbearance period, and a 50% shortfall just two years. This allows trustees the time to distinguish between erroneous return expectations and the ‘animal spirits’ of asset prices. In addition, the aggregate amount of support should be limited. Support to the extent of ten percent of members’ equitable interests avoids excessive and unwarranted support. The rule is that if either of these rules is transgressed, it triggers cuts to pension payments and non-pensioner members’ equitable interests. This enhances the subsequent relative position of pensioners in payment.

Simulation, with empirical data, suggests that actual cuts may be an extremely rare event, less than one in one thousand.

There is some debate as to discretion in trust arrangements – and historically arrangements such as “with-profits” did confer great latitude on the managers of those arrangements. We should not be side-tracked by Maxwell. We can write a law or contract but there will always be some who choose to break that law or contract. The appropriate discussion for those is one of severity of sanction when exposed. We do note, also, that modernising trust law is a project which has been adopted by the Justice Department.

As with any trust, the scheme rules set the governance requirements, and trust-based CDC is no different. As has been noted in other articles, members have both voice and exit, votes and transfer, available for the enforcement of those rules. For this to work well as a governance mechanism, it requires some disclosure standards, with their associated timing and content requirements. There is no reason that these standards cannot themselves be contained with the scheme rules. With today’s technology, a member’s equitable interest and the asset portfolio, and related metrics, as well as historic developments, can be made available to members in near-real-time.

Inevitably, there are some matters which call for judgement on the part of trustees – the contribution setting process described earlier is a prime example, as also are the asset allocation and hedging strategies. The specific forbearance formula and the support limit discussed earlier were a matter of my judgement, in a trade-off where simplicity was considered a virtue. The need for judgement to play a role warrants the introduction of a “fit and proper person” test for trustees. I will add that I consider a trustee who waits or wants to be told by a regulator or other external agency, what to do would fail my “fit and proper” test.

If our notional regulator is capable of superior judgement in these matters than trustee groups, we should abandon the idea of independent CDC schemes and have them manage these schemes, and be accountable for their performance.

In fact, the fiduciary responsibility of trustees offers them little discretion by way of purpose, but considerable by way of means to achieve that purpose. It would be a shame, as well as inefficient, to lose this flexibility. Remember no contract or set of rules, no matter how extensive, can ever hope to cover all possible future developments, and there, judgement will be called for.

I worry greatly when I see “risk based” regulation being introduced. It seems almost inevitable that this becomes formulaic, excessive and ineffectual, when what is really called for is perception and judgement.


[1] This is an oversimplification. The equitable interest of a scheme member does embed capital markets opportunities through the contribution setting process, and it refreshes the aggregate contractual rate or expected investment return partially with every year’s pension payments and new awards. This is a far smoother process than mark to market. As a form of moving average, it has the property of lagging trends in the spot market, both up and down. It is also highly autocorrelated with its own prior values.

About henry tapper

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

Leave a Reply