Risk sharing and investment strategy

In this installment of our series of discussion papers, Jean Frijns concludes that a far-reaching reform of collective DC will lead to a more effective investment strategy by making it possible to harmonise risk frameworks and goals

(This series of articles, published in conjunction with our Netherlands sister title IPN, seeks to create a platform where various authors, academics and practitioners discuss possible solutions to solvency and risk sharing issues that have the common objective of making the DB pensions system more robust)

The traditional pension fund investment strategy is simply a stable mixture of debt and equity. With some imagination it is possible to justify this as a way of applying the capital allocation line model. The long-term is presented as a long short-term: it is a single period model. The only relevant risk is the equity risk that, as a result of mean reversion, decreases the longer the horizon. But accepting that at the end of the horizon the economic environment can have changed from that at the outset, the risk model becomes much more complicated. We need to build into our investment policy protection against unforeseen inflation or the investment risk of low returns. The risk model becomes even more complicated if we need to take into account the risks of negative outcomes during the journey as well as the final-value risk.

This article investigates the question of whether the following specific risk frameworks are a barrier to obtaining an attractive long-term return:

Simple end-value risk, risk borne by the participants; The same, except that final value is offset against the sum needed to buy real annuities at retirement; this acquisition cost varies depending on interest rates, inflation and life expectancy. In the case of DC schemes; A balance sheet approach that compares the sum accumulated in each period with the liabilities taken on in such a way that risk is borne by the pension fund sponsor. The sponsor can be an insurance company or an enterprise offering its employees a DB scheme; The same, but risk borne by fund participants; in the Netherlands this is referred to as collective DC.

The first framework is not an appropriate risk framework for pension saving. It is a simple extrapolation of the single period model, but it does not address the purpose of the investment. The second risk framework is very appropriate but significantly more complicated: the risk of discrepancies between final value and the necessary cost of purchasing annuities must be determined. Beyond simple investment risk, therefore, inflation and interest rate risk play a role. The investment strategy will include a mix of assets with an attractive yield and assets that protect against inflation and interest-rate risks. But it is still about final value, which makes it possible to make optimal use of the long-term characteristics of the different assets. Since we are not required to take into account intervening fluctuations, we can in principle make do with a fixed optimal portfolio, although in practice we will respond to important market shifts by making tactical changes to the strategic mix.

The balance sheet approach for pension funds has grown enormously in popularity over recent years under pressure from new regulations but also because participants have become more concerned about cover ratios. Here you look at the yearly development of the surplus. The relevant risk is the surplus risk; the risk measure is the maximum permitted surplus-at-risk (SAR). The optimal investment strategy involves matching obligations as precisely as possible. This requirement can be met by holding long-term debt filled with interest rate derivatives. This reduces the SAR to a significant extent. Depending on the sponsor's risk preferences, the surplus risk can be used to invest in riskier assets. But there is a warning to be made here: unfavourable returns reduce the surplus and reduce the maximum permitted SAR, and can force short-term selling of riskier assets. In the worst case, the surplus can be fully used and the only permissible strategy is complete matching of obligations to investments. This is an unattractive strategy from a return point of view.

The balance sheet approach is appropriate for DB pension funds; this is quite clear. But the approach has clear disadvantages. First, the approach leads places a sharp focus on the short term, meaning that the attractive long-term characteristics of equities, for example, can only be used in a very restricted fashion. Even more serious is that the balance is expressed in nominal terms: managing the balance sheet risk does not mesh with the fund's goal of paying real annuities. Furthermore, there is a certain crunch-time risk of a fund being forced to liquidate its riskier assets at a bad time on the market.

For many companies there is little reason to take on extra market risk as a sponsor of a DB pension fund alongside the existing risk of operating a company. The consequence is that companies have shifted the risk burden to participants or have chosen strategies that greatly reduce the risk and the required SAR in the pension fund, often at the cost of lower inflation protection for the participants (1).

In the Netherlands, the preference is often for an alternative solution referred to as collective DC. The term is misleading: they are not classic DC schemes, but schemes that are managed like classic DB pension funds, complete with a surplus and balance sheet liabilities. It is unclear who the sponsor is and to whom the surplus belongs. That problem is even greater if the composition of the participants is not homogenous and contains different age cohorts in different stages of their careers.

One solution is to agree explicit rules about who bears the surplus risks and to what extent, and to supplement these with rules about the parcelling out of the excess surplus, to the extent that this can be done (2). Modern option valuing techniques make it possible to specify the value of the rules that have been agreed so that solidarity between different groups of participants and future participants can be expressed in monetary terms (3). Given the complexity of the problem and especially the sensitivity of the discussion, there is some doubt over whether this will happen.

The investment strategy of this unsponsored DB pension fund is very similar to that of a DB pension fund with a sponsor. The balance sheet approach is central; in addition to that the fund's managers can, depending on the surplus and their risk appetite, invest in more or less risky assets. The same disadvantages also apply.

For some, including the author of this article, this amounts to an argument for a complete shift to collective DC (4). The can be done by investing the premiums of active participants in investment pools. These pools are run as investment funds. An annuity insurance scheme stands next to these funds, and there is an obligation to participate in this scheme from a certain age. The scheme's collective character lies in the obligatory participation and the limited choices for participants. The choice can be limited to just two funds: one with a risky asset mix and an explicitly long-term time horizon, and one fund with a low risk profile that matches the purchase costs of a long-term annuity. An individual participant when young will invest exclusively in risky funds and put a growing sum in the low-risk fund over time. When he or she retires, his or her investment in the fund can be used to purchase annuity insurance.

Instead of having one investment policy, there is now a fund-specific investment policy as well as balance sheet management for the annuity insurer. The big advantage of this approach is that the risk framework and the goals are in harmony, which makes it easy to define an adequate investment strategy for each fund. Risk is primarily taken in high-risk funds by the participant. The aim is to attain a high long-term yield while riding out intermediate shifts in the valuation of the accumulated assets. The low risk investment fund has the aim of allowing participants to accumulate assets in order to buy real annuities. This purchase sum changes according to the level of (expected) inflation and the interest rate. The low risk fund therefore offers protection against interest rate and inflation risks (5). The insurer's balance management does not differ much from the low-risk fund's management, although the risk boundaries are drawn more sharply and the insurer is subject to external prudential oversight. Risk frameworks and aims all point in the same direction, leading to an effective investment policy.

In conclusion, it can be stated that the effectiveness of an investment policy suffers when the risk environment does not match the investment goals. We have seen examples of an overly simplistic approach to risk which fails to address the specific purpose of pension saving. On the other hand, we have also seen that the balance sheet approach entails a very rigid risk framework that does not necessarily mesh with the desired goals. That applies if the balanced approach implies a nominal risk framework while the goals are stated in real terms. This applies in particular in the case of a DB pension scheme where all the risks are borne by scheme participants. An alternative approach - fully executed collective DC - appears to offer substantial advantages.

Footnotes:

1 This does not apply to all enterprises. For some companies, for example oil companies, the investment risk born in the pension fund comes as a welcome diversification of the company's own high risk levels. Other companies are confronted by the legacy of a rich occupational pension scheme. The investment strategy to be chosen is in such cases dependent on the extent to which it is possible to extract the surplus from the fund.

2 See for example E Ponds, ‘Naar meer jong en oud in collectieve pensioenen', Oratie Universiteit Tilburg, 2008

3 See for example T. Kocken, ‘Curious Contracts', Tutein Nolthenius, ‘s-Hertogenbosch, 2006

4 See J Frijns, ‘Collectief DC in huidige vorm niet duurzaam', Tijdschrift voor Pensioenvraagstukken, April 2006

5 Hedge fund would be a good name for this fund, were the term not already in use to describe funds with very different characteristics.

Jean Frijns, former director of asset management and member of the board of directors at ABP, is professor of investment studies at the Free University of Amsterdam