Light -touch regulation was , has (as the FT predicted) – been dead and buried since 2008.
Received wisdom suggests that auto-enrolment would need iron-fist regulation. Con Keating see thing’s differently. Here – in his words – is why.
Regulation of CDC
This blog is concerned by the many assertions that we need rafts of regulation for CDC. It is my contention that we do not, and indeed should not.
Defined contribution schemes consist of a fund into which the contributions of members and their employers are invested. There are rules, which usually take the form of units in the fund, which define the member’s interest in that fund. There are no liabilities, per se.
This is pooling of assets; investment in a common vehicle, but there is no risk-sharing. Nor is there risk pooling. The members, the unit holders, have merely accepted a common risk exposure. This differs from the pooling of, for example, mortality risk. Here members are exposed to differing life expectations (risks) but by pooling, may arrive at a reliable and more manageable average, which admits efficient management of the risk.
The DC member is exposed to the gyrations of markets between the acquisition of units and their redemption through sale, and most importantly to the value at realisation. This is a simple savings arrangement. The member is faced with the prospect then of managing the conversion of these assets into an income for their retirement lifetime. This process is not trivial, and addressed more fully later. The simplest solution is the purchase of an annuity but they are expensive. There are myriad options to be considered, single life, level, index-linked, with spouse’s benefits, and on, and on, and there are further, wider, exogenous issues such as taxes and the non-return of capital on early death, which are also relevant.
Of course, many different classes of unit may co-exist within a fund. We deal with this by disclosure through prospectuses and offering documents. We do not regulate what relative rights may be offered. The units define the relative interests of participants in the fund.
CDC is really no different to this. It is a fund, with a particular set of rules defining the members’ (equitable) interests, fulfilling the same purpose as the units previously. This is about division of the pie among members now and over time.
It happens that the CDC rules introduce both risk-pooling and risk-sharing among the members, and in the specific case of my vision of CDC, do so equitably among those members, instantaneously and over time. These rules serve to advance the coherence, sustainability and solidarity of the co-operative collective. No new risks are introduced to members; these are risks already faced, independently, by all members. The institutional form, the set of rules, merely serves to better manage these risks. It does so while maintaining the freedom of members to leave the scheme at any time, taking their equitable share of scheme assets with them. Along with one member one vote, this ability to transfer endows members with both exit and voice in the governance of the scheme. The alignment of their interests is reinforced by the risk pooling and risk sharing touched upon later, but described more fully in other notes.
Some have argued that these schemes hold the prospect of disappointing scheme members. That the targets set should be subject to regulation, perhaps even approved by a regulator. It may be that targets were set over-ambitiously. The most obvious question is what qualifies any regulator or supervisor to undertake such a role.
The identification of systematic error in trustees’ investment returns expectations, a serious concern, is facilitated by the forbearance period risk management rule. One of the properties of such a systematic error would be that its consequences should grow with time, though the assessment will be clouded by the volatility of market prices.
The rules for the payment of pensions and the management of deficits would serve ultimately to correct for such errors. The mechanism would be a cut in all member’s equitable interests.
However, there is, in any case, little room for disappointment. This variation to a market confidence objective is misplaced. The equitable interest of the member defines their relative interest in the assets of the scheme, and the assets of the scheme expressed as a proportion of the total equitable interest of the scheme defines progress achieved towards that goal. This may be expressed in capital or income terms. There is little room for disappointment in that shortfalls from the originally targeted income will not arise as a surprise to members, even if subsequently realised.
Indeed, it is even possible to project the degree of pension income additional support that may be expected to be available to pensioners from other members; the capacity for support, given some deficit and the scheme rules.
The accounting position is clear. There are no liabilities explicit, constructive, or otherwise, and consequently nothing to be reported by any sponsor or the scheme itself.
Notwithstanding this position, and particularly in light of the additional trustee responsibility with respect to sustainability, it is worth considering whether there is anything to be gained from further regulation. As this extension of trustee responsibility is prospective in nature, this might take the form of the Pensions Regulator’s “Guidance” for DB schemes.
That regime operates in what may be described as a “fair value”, mark to market world, that of financial economics. It involves valuing liabilities in some exogenous manner, using either bond yields or the expected returns on assets as a discount rate. This is a counterfactual. It answers the question what would these liabilities be worth if these rates were applied to all liabilities at the time of valuation, when, clearly, they don’t. This technique suffers the further disadvantage that it is not time continuous. While it may provide a valid, if particular, prospective view, it distorts the historical development of the fund’s liabilities.
Most importantly, it is a single rate applied to all liabilities regardless of their ownership. However, the ex-ante risk-sharing of the uniform “award” regardless of the age of the member means that there are a wide range of different discount rates applicable to the specific claims of individual members. The differences of these contractual accrual rates among members are not trivial. Given that the extent of this risk sharing has varied over time, with every contribution, and with that their weighted average contractual accrual rate, it is not possible to reverse engineer any member’s interest in any consistent and reliable manner. This method of valuation distorts the transfer values of the individual members’ “pots” and with that the equitable scheme of arrangement among members. As such, it would damage the sustainability and governance of the scheme.
The use of the contractual accrual rate of the scheme, which is a weighted average over time and membership, provides a precise valuation of the promises as made, and indeed perhaps modified. Any member’s contractual accrual rate is a sound basis for the calculation of annual and lifetime allowances for tax purposes. The contractual accrual rate of the scheme is the target rate needed for all benefits to be met on time and in full.
If we value assets at market prices and compare this with the scheme’s valuation made using the contractual accrual rate, we have an accurate and reliable indicator of the state of funding progress, as promised. If in balance, the scheme has performed as expected.
If in deficit, then the measure is an accurate guide to the proportion of the ultimate pension which may need to be cut. Pension payments are only at risk if the scheme is in deficit, and the amount at risk is just the proportion applied to that year’s pension payments.
There really are no gains to be made by the introduction of DB style regulation, and indeed that introduction would be very costly.
Finally, there is the issue that these are member mutual co-operations. The claims and obligations, such as they are, are from members to members. It is most unusual to intervene to protect an individual from him or her-self.
The problem with regulation in the pensions world is that it is costly. In some cases, in recent times, DB funds have even underperformed individual DC, notwithstanding the inherent disadvantages of individual DC. This is a direct consequence of DB regulation, and particularly its hectoring, minatory extensions delivered through “guidance” – de-risking and all that.
It is worth recalling some of the issues for which CDC is a proposed solution. The most notable beyond substitution for the costly sponsor guarantee centres on decumulation, as was discussed earlier. The problem of decumulation in individual DC is well known but its costs have not been widely discussed.
If we simply de-risk the asset portfolio and draw down over time – then we face a drag due to the difference in (realised) returns of the order of 3% pa. This is similar to or, more usually, slightly less than the cost of annuitisation. The effect of this is to halve the post-retirement income; it represents a lowering of about 25% of total lifetime investment income. The absence of any need for a CDC scheme to de-risk the asset portfolio in this way is the source of a significant part of the excess performance of CDC reported in simulation studies.
There is a further cost which arises from the variability of outcomes from year to year, which is effectively eliminated in CDC. If we take the PPI work for the TUC on individual DC outcomes, we may obtain a first order cost associated with the variability over time of annuitisation – the volatility of outcomes rises from 11.2% before annuitisation to 17.7% after. This is equivalent to an additional cost drag of 1% pa. This is added to the low return implicit in any annuity – it is the cost of variability from the average annuity rate over the period of the TUC/PPI study. It is an inter-cohort cost metric and as such rather important in maintaining inter-generational equity and solidarity. The elimination of this cost is a further attraction to CDC.
If we want to take, as a counter-example, a situation where regulation is comprehensive, we may take the Dutch example. Their schemes are effectively accounted and regulated as if they were DB. Indexation may only be paid if the scheme is 130% funded. Assets and liabilities are valued using market prices. The result is a disturbingly similar emphasis to UK DB on deficits derived in this manner and on de-risking, and with that excessive costs. Cuts to indexation are all too common.
 Notwithstanding the Pensions Regulator’s “guidance”, the duty of a trustee is to secure the benefits accrued to date, and not to consider possible futures.
Con Keating is a member of the steering committee of the financial econometrics research centre at the University of Warwick and of the Societe Universitaire Europeene de Recherche en Finance.
As a research fellow of the Finance Development Centre he published widely on the regulation of financial institutions and pension systems. He has been Chairman of the committee on methods and measures of the European Federation of Financial Analysts Societies and is currently a member of their Market Structure Commission.
Con has also served as an advisor and consultant to the OECD’s private pensions committee and a number of other international institutions, including the boards of a number of educational and charitable foundations and as a trustee of several pension schemes. He is currently Head of Research for the BrightonRock insurance group