Lorant Porkolab outlines the whys and wherefores of risk sharing in occupational pension schemes, an approach offering a compromise between DB and DC
The concept of risk sharing is not new, different variants of the approaches described in the document have been adopted in a number of other countries, most notably in the Netherlands, and it was also proposed by the Association of Consulting Actuaries for consideration in the UK as far back as 2006. Since legislation could change, the subject is likely to stay on the agenda of various interest groups for some time, and it is important for plan sponsors to understand the implications.
Why share risk?
Traditional defined benefit (DB) plans under current UK legislation leave scheme members with few risks. Interest rate, inflation, market, longevity and operational risks are placed firmly on the shoulders of the plan sponsor. In pure defined contribution (DC) schemes most of these risks are transferred to the scheme member.
The concept of sharing some of these risks between sponsors and members has become popular in some quarters, for a variety of reasons:
* To the extent that the government can persuade the private sector to share a larger part of the burden of providing pensions sufficient for people to actually live on, it will reduce its own costs in the future.
* Although the evidence is not incontrovertible, there seems to be general acceptance that 'better' pension arrangements (as again implied by most risk sharing concepts, compared to pure DC) are likely to reduce staff turnover and thus benefit sponsors.
* The adoption of risk sharing principles is likely to keep some schemes which would otherwise close open; this is generally considered attractive by scheme trustees. So is risk sharing good or bad, both in absolute terms and in relation to pure DC? The answer, of course, is that it depends on exactly what alternatives are being compared and the perspective that is adopted. Generalisation is difficult, but on a cost equalised basis, a risk sharing approach is likely to result in a lower, but more certain, level of benefits - whether this is appreciated by a typical employee is another question again.
The background to change
There can be no doubt that many of the recent UK legislative, regulatory and accounting changes for pensions were needed, as companies and all other market participants (e.g. investors, banks, rating agencies) realised that DB pension promises represented more significant liabilities than they previously thought and accounted for. However, they did much more than this - rightly or wrongly they changed the rules of the game sponsors thought they had been engaged in.
The reactions to these changes by companies have also been natural and understandable: closing final salary pension schemes, reducing benefits and moving to DC arrangements. The Netherlands is one of those countries where risk sharing in relation to pensions has been successfully introduced. Between 1998 and 2005 the share of active participants covered by career average schemes with conditional indexation there jumped from one quarter to three quarters. So, risk sharing solutions there have certainly been embraced and, perhaps, preferred to pure DC.
Risk sharing in the UK
While it was clear changes were needed, replacing one extreme approach (all risk on the employer) with another (all risk on the employee) may represent an over-correction; to many observers the pendulum has swung too far in the opposite direction. Approaching this issue by exploring more balanced risk sharing arrangements between employers and employees seems to be the broad objective of the new Risk Sharing Consultation exercise by the UK government.
Two of the potential approaches considered in the document are collective DC funds and career average schemes with conditional indexation. With respect to the latter (potentially the more interesting), indexation refers to the mechanism of adjusting or revaluing accrued benefits (including pensions in payment) by reference to a specific index of prices or earnings (e.g. the Consumer Price Index), while the conditional element indicates the above indexation is dependent on the funding position of the scheme.
As part of the design, the career average pension scheme with conditional indexation would need to specify a target rate for benefit revaluation and make a prudent allowance for this in the funding valuation of the scheme. As long as the scheme is fully funded, the benefits would be increased in line with the target index.
However, if the fund has a deficit in a particular year, the indexation rate for that year could be less than the target rate or the revaluation of the benefits could even be withheld, hence the risk sharing between employer and employee. In the latter instances, the resultant savings would reduce the funding deficit, allowing the scheme to recover the deficit sooner and be able to provide full indexation at the target rate again.
Various catch-up mechanisms can also be built into the indexation policy whereby the pension fund would use a rate for revaluation above its target to compensate for prior years when members received less than full indexation.
The important feature of the above approach is that - due to the conditional indexation - investment, inflation and longevity risks could be shared between sponsor and members, reducing the risk on sponsors and hence potentially making this type of scheme design an attractive alternative for them.
The benefit of the approach for corporate sponsors is clearly the extra control lever in the form of the indexation policy, which provides additional flexibility and a valuable safety valve. This would bring a new dimension to the corporate pension strategy.
Historically pension strategy was fairly one dimensional - at least on the corporate side - with a focus on funding. The recent changes in accounting policies (along with volatile market conditions) made companies realise there is also another important dimension they may need to give greater attention to; the investment strategy of the pension fund.
Then it was the Pensions Regulator who created another (or highlighted a previously hidden) dimension, the sponsor covenant, making the problem three dimensional. Now the indexation may add a fourth dimension to the picture. Indexation policy will need to be formulated in conjunction with the funding and investment strategy, taking also the covenant into account.
The expected and actual indexation will also depend on how the scheme is funded and how the fund is invested, and the level of risk sponsors and scheme trustees feel comfortable with.
Developing an integrated strategy framework where all these components are considered together is essential. Companies and pension funds often try to work out their best funding and investment strategy in isolation, but when the components are put together they realise the end product is hugely sub-optimal.
Clearly if indexation policy becomes part of the overall pension strategy, the importance of considering the various control levers and risk factors together will become even more relevant and important for corporate sponsors.
In order to determine the optimal integrated pension strategy, the Asset Liability Modelling frameworks used by many pension schemes will need to be extended to incorporate the new conditional indexation feature along with capturing the relevant interdependencies. The overall policy framework must be sufficiently robust, with stable funding requirements and the likelyhood of below full indexation kept under an agreed pre-specified threshold.
Communication is likely to be another area that would require careful consideration by companies that wish to introduce conditional indexation-based risk sharing arrangements. Pension funds in the Netherlands are required to disclose their indexation ambitions (targets) to their members, including a realistic estimate of the likelihood of success in pursuing this ambition, and the UK may follow a similar approach.
In addition, pension funds have to ensure consistency between expectations raised, contributions made, reserving and actual indexation decisions.
Companies now understand that there are significant financial costs and risks associated with the provision of traditional DB arrangements. Many of them also accept that, for some categories of employees, an offer of pure DC arrangements only may not be optimal. Risk sharing between sponsor and member may therefore be attractive in certain circumstances, but only as long as companies cost these arrangements properly and mitigate any remaining risks they are unwilling to shoulder.
The pendulum has swung firmly from DB to DC over the last few years, for understandable reasons. Whether it will swing back towards a more central position is too early to say, but there are certainly some interest groups that would like to see it do so, including the government. It may not be long, therefore, before the above practical aspects are being explored by advisers, companies and trustees.
This week's top stories include ITS' management buyout from Mercer, and The Pensions Regulator launching a probe into single-employer defined contribution schemes' default funds.
People retiring in the UK will on average outlive their pension savings by 10 years, according to research by the World Economic Forum (WEF).
Steps to improve auto-enrolment are uncontroversial and obvious, but the government is dawdling on introducing the necessary changes, argues Jack Jones.
Professional trustees will be expected to apply for accreditation as part of a framework intended to be launched on 1 July by the Professional Trustee Standards Working Group (PTSWG).