June 2009

EDHEC RESEARCH: go Dutch with pension funds

The Netherlands provides an excellent example of how regulations are a driving force behind pension plan designs, says Samuel Sender of Edhec...

windmills.jpgThe Netherlands is home to the most highly developed pension industry in Europe, with the most professionally organised pension funds and, relative to GDP, more assets under management than any other country. Dutch regulation is aligned with Solvency II (similar structure, different calibration) and the IORP Directive (use of the accrued benefits measure).

With the Financieel Toetsingskader, or FTK, the Netherlands has chosen a regulatory framework with the same modular approach as that of Solvency II. Although it applies to both insurance companies and pension funds, the two industries are treated differently. The FTK for pension funds is less demanding in terms of capital, because of both liability valuation and lower calibration.

In the first pillar, the FTK requires that valuation principles are fair value accounting, with assets fully accounted for and marked to market. Also there are two levels of solvency requirements that must be tested every year.

One may note that both solvency tests apply only to guaranteed (in practice, nominal) pension rights. Also, there are no quantitative investment limits; by contrast, diversification is required by law.

In the second pillar, the FTK requires that, in addition to the two solvency tests of pillar I, a “continuity test” is required as part of pillar II. It involves a projection of capital requirements that serves to manage risk as well as to the set future contribution levels and likely indexation rates that are communicated to employees. Pension funds must provide evidence to the supervisor that their funding and investment strategies are consistent with their indexation ambitions.

The third pillar has resulted in increased transparency and financial communication for Dutch pension funds. It obliges pension funds to inform scheme members of their indexation policies. Pension schemes will have to state explicitly whether or not the pensions are indexed.

Driving force
The Netherlands provides an excellent example of how regulations are a driving force behind pension plan designs. The pension fund industry has reacted to these new regulations by forgoing formal indexation and replacing it with conditional indexation, a mechanism not unlike profit sharing in insurance companies.

What is unique to the Netherlands is that as a consequence of changes in indexation the pension fund industry faces a paradox that may have great consequences on asset-liability management: pension funds report liabilities to the regulator as if they were nominal fixed cash-flows.

The FTK prescribes stricter solvency requirements for guaranteed or nominal pension rights than does regulation in most other countries. By asking for full funding plus buffers against investment risks, it virtually ignores the role of the sponsor in providing or contributing over the long term to guarantees made to employees, and makes it very difficult to offer indexation guarantees at a reasonable price. On the other hand, much more flexibility is given regarding conditional rights (such as indexation), because these rights are not accounted for in the statutory value of the liability.

Though unconditional indexation of pension benefits used to be served unconditionally every year, guarantees were not always written into contracts. So indexation was often temporarily halted during the 2001-2003 stock market crisis. As required by the FTK, it was pointed out to employees in no uncertain terms that “the indexation of your pension is conditional; there is no entitlement to indexation and, over the longer term, it is also uncertain whether there will be indexation or its extent”.

With conditional indexation, the liability value diminishes markedly. In other words, making indexing conditional on the funding ratio of the pension plan naturally makes it possible to avoid announcements that the pension fund is unable to meet
its commitments.

Both indexation targets (which influence replicating assets) and indexation rules (which make clear that indexation depends on the funding ratio of the pension fund) are specific to each pension fund.

The move towards conditional indexation was instrumental to the prosperity and solidity of the Dutch pension fund industry. This situation was allowed by the regulator. However, it creates a divergence between the regulatory reporting and the internal views of pension funds. When reporting to the regulator and making statutory funding calculations, the liabilities are computed as ABO, as fixed cash flows. As a consequence, the statutory duration of the liabilities is very long.

As capital is required to cover interest rate risk, calculated in ‘S1’ as the change in net asset value (assets minus liabilities) subsequent to an interest rate shock, long dated nominal liabilities require a bond portfolio of very
long duration.

Indexation was made conditional mainly for reporting purposes, but most pension funds have an indexation target over the long run. As a consequence, their internal view is that they hold real liabilities.

So the internal view is that duration is in fact much shorter than that reported to the regulator, and that fully closing the regulatory duration gap creates risk from an economic standpoint, even though it reduces solvency capital requirements. With respect to indexation, the reverse is also true: hedging the sensitivity to future indexation would create additional capital requirements but would decrease economic risk.

Equities, which are a good fit for long-term economy-indexed real liabilities, are reported as mere risk-taking assets rather than as the replicating assets they actually are; as a consequence, investments in equities increase capital requirements.

The FTK, which can be summarised as a short-term funding constraint, is not always straightforward to implement because of the contradiction between the funding measure used for the long-term target (in the continuity analysis) and that used for the short-term funding constraint. Because of changes in pension plan design, risk as measured by the FTK is a misleading indication of the real risk embedded in the pension plan.

Wage indexation is generally achieved in the long run by investing in equities. However, under FTK rules a 40% investment in equities raises funding requirements by 10% to at least 120%. Pension funds with low funding ratios (and little capacity to raise contributions swiftly) cut their equity exposure, and swap it for long duration risk-free bond exposure. That, in turn, means employees lose the initial target indexation to wages and equity performance.

As a whole, the Dutch pension industry was better funded than, for instance, its counterpart in the UK, where it would have been downright impossible to enforce such stiff regulation.

Some of the largest multi-employer schemes had funding ratios sufficiently high to leave pre-FTK asset allocation unchanged. For a very restricted number a halt to indexing and an increase in contributions have always been sufficient tools, and post-2003 equity performance made it possible to make up for the halt and reach the pre-2003 indexation target. As long as reserves are sufficient, long-term goals can be safely pursued.

Naturally, most funds are somewhere between these two extremes, usually with sufficiently high funding ratios to withstand a pronounced allocation to risky assets, but used the standard formula for capital requirements, so partial closing of the regulatory duration gap is welcome. Swaps have been used to lengthen the duration of the portfolio, and risk exposure has been slightly reduced.

• This article was based on research carried out within the Edhec/Axa Investment Managers Regulation and Institutional Investment research chair.
Samuel Sender is applied research manager with the Edhec Risk and Asset Management Research Centre

©2009 funds europe