Building confidence in indexation
In the Netherlands indexation of pension benefits to either wage or price increases has long been considered a guaranteed right. Although indexation used to be conditional on the pension fund’s position, in practice full indexation was virtually always given. In the communication to pension beneficiaries the possibility of indexation cuts was not given much attention.
The adverse stock market returns during 2000-2002 have radically changed this perception and practice of automatic indexation, as many pension funds became severely underfunded. In order to recover, not only were pension premiums increased firmly, but indexation was reduced as well.
Moreover, many pension arrangements were renegotiated and, as a result, changed from a conditional defined benefit (DB) final wage system to a conditional DB average earnings scheme. Consequently, indexation cuts cannot only be applied to pensioners but also to active workers.
Nowadays, the typical pension contract comprises an average earnings defined benefit pension in which only nominal benefits are guaranteed, but with the intention to provide wage indexation. Although index-linked benefits are of great importance to pension beneficiaries, almost no scheme in the Netherlands includes unconditional indexation commitments.
This has bred uncertainty concerning future pension benefits. Inadequate information on this subject could severely harm confidence in pension funds and encourage participants to accumulate non-optimal (i.e. excessive) savings.
In the FTK, a balance is sought by prescribing relatively strict solvency requirements for guaranteed (in practice, nominal) pension rights, whereas much more flexibility is given regarding conditional rights (such as indexation).
For guaranteed rights, the new regulatory regime is in line with modern finance views in the sense that market valuation is not only applied to assets but also to liabilities, and that more risk taking results in higher required buffers. In principle, guaranteed commitments should always be fully funded.
For conditional rights, provisioning is not required, though pension contributions and policies have to be consistent with communicated indexation ambitions.
Pension funds are required to disclose their indexation ambitions to their members, including a realistic estimate of the likelihood of success in pursuing this ambition. In addition, pension funds have to ensure consistency between expectations raised, finances, reserves and actual indexation decisions.
Thus, as a result of increasing consistency between expectations and outcomes, confidence in Dutch pension funds may be strengthened.
To what extent is this likely to happen? Our paper (‘Conditional indexation in defined benefit pension plans’ DNB Working Paper No 86/January 2006) had three aims: first, to investigate to what extent the new Dutch regulatory regime does indeed achieve greater clarity with respect to indexation ambitions; second, to analyse whether and to what degree FTK’s minimum requirements can provide reasonable guarantees that indexation ambitions will be realised; and third to assess whether the FTK regime is over-restrictive on funds that already pursue sound policies.
We analysed these questions with the help of the Pension Asset and Liability Model for the Netherlands (PALMNET). We used PALMNET to make illustrative simulations of indexation policies with varying degrees of ambition. These simulations provide an understanding of the average expected development of pension contributions, indexation cuts, funding ratios, etcetera, but also of the dispersion of these variables, thus sketching the outcomes of unfavourable scenarios.
Thereby, the simulations constitute a long-term feasibility analysis of indexation ambitions including the uncertainty surrounding the key variable results, just as the required continuity analysis should do.
Our conclusions are that, as regards transparency, a major improvement is that pension funds have to inform their participants in explicit terms about their indexation ambitions. The likelihood of indexation must be determined on the basis of a continuity analysis.
Yet, the content of that information would be inadequate if the emphasis is only on long-term average values of the key variables. In our view, differences between pension fund policies become most clear in the worst-case scenarios. Hence, adequate information should include – stylised facts of – the future distribution of the key variables.
The likelihood of indexation depends on factors, such as the earmarked reserves for indexation and mark-ups on the cost-covering premium level, which have not been fully regulated, as the supervisor does not presume to take over the responsibilities of a pension fund’s board.
Different funds will pursue different policies on, for instance, maximum contributions, recovery from a capital buffer shortfall or indexation cuts. An extended continuity analysis, including the full distribution of key variables, will have to provide conclusive evidence on the likelihood of indexation.
The question whether the minimum FTK requirements concerning conditional indexation offer sufficient assurance that a fund’s ambition will be realised can be answered in the affirmative.
Despite the fact that provisioning for conditional commitments has not been made mandatory, indexation cuts turn out to remain fairly limited, though still substantial in adverse circumstances.
On the one hand, this is because cost-covering contributions as defined under FTK are, in fact, more than cost-covering (as it includes a capital buffer building mark-up). On the other, it is because of the occasionally very high contributions due to one-off nominal underfunding (when, as in our simulations, benefit cutbacks other than indexation stops and other solutions are ignored).
It turns out that pension funds that intend to provide indexation with minimum effort (read: minimum contributions), so that voluntary indexing reserves fail to materialise, are more risk-prone in the sense that these funds run a much greater risk of nominal underfunding, incidentally leading to soaring contribution rises. If no surcharges to the premium are levied to finance indexation, cuts will be substantial, and moreover nominal underfunding resulting in extreme premiums is hard to avoid.
The question whether FTK implies over-restrictive requirements to a pension fund offering conditional indexation can be answered in the negative. The complaint, frequently voiced by the pension industry, that the FTK – and the 105% minimum funding ratio in particular – will impede indexation, because mismatching will become too risky, is not corroborated by our simulations.
A pension fund that has serious indexation ambitions, should build up earmarked indexation reserves. That will make the risk of under-funding very low, even assuming the current investment mix of 50% equities.
For pension funds offering an indexation guarantee, the simulation outcomes are unfavourable. Under FTK, realisation of unconditional indexation will be far more difficult, because very large implicit capital buffers need to be shored up. Further, such buffers are also extremely sensitive to movement in the real interest rate. For instance, the current 16-year interest rate of 3.8%, implies a discount rate of just 0.8%.
By consequence, as long as underfunding can only be dealt with by contribution rises, premiums may rise to extreme levels. The volatility in the funding ratio might be reduced by improving the match of investment portfolio and liabilities, but only to a limited extent, as wage-indexed bonds are yet unavailable and as the cost-covering premium is also sensitive to real interest rate changes.
All in all, our simulation analyses show that defined benefit pension plans can provide a reasonable insurance against wage or price inflation, even where full guarantees are fairly unattainable. Further, they illustrate the tenability of defined benefit pension plans under ageing, the new fair-value accounting regimes, and possible volatility on financial markets.
Jacob Bikker and Peter Vlaar work for, respectively, the Strategy Department of the Supervisory Policy Division and the Financial Research Department of the Economics and Research Division of De Nederlandsche Bank (DNB) Views expressed are those of the authors and do not necessarily reflect official positions of DNB.