Why CDC risk management breaks all the DB rules – Con Keating

con-keatingguestThe advent of CDC will offer an opportunity to introduce forms and techniques for risk management which are intrinsically more efficient than those currently advocated by the Pension Regulator and which are consistent with the true duties and obligations of trustees.
The issue of concern is the sufficiency of scheme assets to meet the “promises” made to members. The value of these promises depends upon the contribution made in respect of an award and the projected amounts of pensions payable, and the amount of time which has passed since the award was made.

Put another way, the amount of the contribution made and the projected pension payments define a rate of investment return on the contribution implicit in an award. We refer to this rate as the contractual accrual rate (CAR). It is the compounding of the contribution at this rate which determines the equitable interest of the scheme member at any point in time. This is illustrated in figure 1, below.
CDC risk management

In this illustration, mortality is deterministic. It is, of course, trivial to apply a mortality table which would change the shape of the decumulation phase but otherwise add nothing. All projected pension cash flows and their input parameters are best estimates.
This contrasts with current DB practice where the emphasis is upon prudence in technical provision derivation. However, CDC differs from DB or other financial contracts in two important ways. The first is that there are no external resources for the scheme to call upon – the asset fund is all there is. The second is that there is only one class of participant, a member, for whom their claims upon scheme assets all have similar priority.
In banking, the claims of depositors are senior to those of shareholders, in insurance it is policy-holders rather than depositors.  In essence the senior class has a call upon the resources attributable to shareholders. With DB the call is explicit and upon the sponsor. The overestimation of liabilities in a prudent technical provision context should serve merely to limit the frequency or magnitude of claims upon the sponsor employer. It is much abused in current practice.
It is pointless to partition the assets of a CDC scheme in any manner, though it is done in the Netherlands with their risk buffers; it is no different from moving money from one pocket to another. Total wealth is unchanged, though some may now ne less accessible. It does not enhance the security of the “promise” to members. However, it is possible to enhance member security by exploiting another property of pensions, their time dimension.
Balance sheet insolvency is not a binding condition. Even companies may continue to trade while balance sheet insolvent if the directors believe that such trading will rectify the insufficiency. It is cash flow delinquency which is usually binding; the non-payment of a sum now due constitutes an event of default in almost all circumstances. To an extent, it may be argued that the period allowed by the Pensions Regulator over which deficit repair contributions may be made is recognition of this.
Even schemes which are weakly funded, that is have substantial deficits, do have some assets, which may be used to pay current pensions. Cash flow insolvency is very remote in time. However, discharge of a current pension payment obligation is intrinsically inequitable to non-pensioner members of a scheme if that payment is made at a time when the scheme is in balance sheet deficit. All scheme members have claims upon the scheme, their equitable interests, but it is the net asset value of those claims which determines their quantum at any point in time. If the scheme pays pensioners in full, it is overpaying them relative to other members.
However, scheme risk-sharing rules may overcome these difficulties. The most obvious is to allow forbearance time limits, associated with an absolute amount of scheme equitable interests which may be committed to support of current pensions. In this, the scheme rules are substituting for the buffers and capital adequacy regimes, and indeed prudential valuation, seen elsewhere. The key to sustainability for such rules is that they should maintain the equitable balance among members.
A simple illustration is appropriate. Suppose we have a scheme which has pensions payment amounting to 5% of the total equitable interests of members, but that this is only covered as to 80% by scheme assets. Let us also suppose that the scheme membership is 60% non-pensioner and 40% pensioner. Then the support afforded to pensioners amounts to 1%.

To maintain equitable balance among members, it is necessary to increase the equitable interest of non-members in similar proportion.
The effect of intra-member support is illustrated, for pedagogic purposes,  in figure 2. This figure shows a scheme consisting of just three members who at time of the single award shown were aged 25, 45 and 64. We then consider the situation if the pension payment due (to the former 64 YO) in the fifteenth year is only covered by assets to the extent of 50%. Payment of the full pension in this case involves support as to 2.96% of the overall scheme equitable interest (3.7% scheme assets).

The equitable interest of the pensioner member is unchanged before and after support, the equitable interests coincide everywhere. The equitable interests of the non-pensioner members, and their pensions subsequently increase.
The amount of support needs to be supplemented by rules as to the term of forbearance. These may take the form of simple rules such as a 10% deficit must be recovered with ten years, a 20% deficit in five years and a 50% deficit in two years. Continuing deficits would trigger cuts to the equitable interests of all members, a rebalancing of the scheme.

There are many nuances which might be discussed but that would distract from the central thrust of this article, which is that these rules operate in the here and now; they are objective and operate only on the current pensions payable. Next year’s possible deficit or surplus or next year’s pensions are not considered.
This contrasts starkly with interventions used widely in the Netherlands and proposed by some here in the UK, with the prime example being variation in indexation. These interventions are invariably inequitable in their operation on the relative interests of members. In the case cite above, the quid pro quo for support of the pension payment is an increase in the relative claim of non-members on scheme assets.
CDC risk management 2

It is perhaps surprising that solitary intervention today may serve with respect long-term risk management, but it is, in fact, highly efficient. The statistical treatments of risk suffer from the disadvantage that they may include, and generate provisions for, many events which may happen but won’t actually eventuate.

The statistical approaches are also subject to the criticism that they disregard uncertainty dealing only with circumstances for which we may assign probabilities. Simply put, the method proposed eliminates the false positives of statistical testing, and their associated costs.
The amount of unutilised support available constitutes a new potential risk metric: it can be used to estimate the length of time during which pension payments are not likely to be cut. This also feeds into the investment management mandate. The contractual accrual rate provides asset managers with a target return to be achieved, and the unutilised support available provides an estimate of the term over which, on average, it is desirable to achieve that rate.

target pensions

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.

1 Response to Why CDC risk management breaks all the DB rules – Con Keating

  1. This deserves a longer study than I have given it – and (better still) input from my colleagues at Fowler Drew. What we tend to look for in suggested CDC solutions (as a firm with management of drawdown portfolios as a core competency) is the extent to which a collective risk-smoothing approach, subject to fairness rules, is superior to individual risk sharing, where fairness is satisfied without need of rules by the lack of third-party subsidies.

    We also look for the ability to satisfy individual client utility. We find this typically places significant value on the sustainability of real spending: avoiding unplanned cuts. Tradeoffs may be made, often involving bequest values (particular to individual pots and often applicable only where there is effectively a joint venture between two or more generations of one family) and usually between time periods (risk of cuts is best pushed out to later stages when easier to deal with – though this may be better dealt with by front-loading planned spending). The tradeoffs often involve creating optionality, which only works because the payoffs only benefit their creator. These idiosyncratic solutions for maximising utility being uniquely available to individual pots, there have to be other benefits in collective solutions to outweigh them.

    That is not to say there is not room for both, as the idiosyncratic features may not be that valuable to everyone, compared with (say) not having to think about how much to draw and spend. It is also theoretically possible that there is some level of risk where the individual’s utility is satisfied broadly equally by each, because any advantage depends on the mix of risky assets and risk free hedges. This after all has been one of the issues with insurance-company products that aim to smooth: the asset mix is too volatility averse.

    I am mindful that the more ‘worked’ the solution, the greater the model risk – as Con points out. (It is the avoidance of unplanned cuts and the importance of beating with very high confidence a tolerable minimum spending target that particularly call for precise estimates of the distribution of simulated outcomes of the portfolio approach.) But at least we look for clarity as to where the contributions come from in CDC that are not available to individual drawdown, as between investment risk sharing, longevity risk sharing, cost savings – and (yes) let’s add model risk too.

    Any insights from Con on this question would be much appreciated.

Leave a Reply