The Dutch pension legislation that laid the foundations for the country’s solid pension system is undergoing significant changes. The Pension and Savings Act (PSW), which has been in existence for 50 years, will be replaced by the Pensioenwet (Pension Act), expected to be implemented at the beginning of 2006. See the box for a more detailed description of the Pension Act.
Parallel with this, the Pensions and Insurance Supervisory Authority (PVK) will introduce a new supervisory framework, the Financial Assessment Framework (FTK).
Under the Pension Act, the primary responsibility for pension arrangements will still be with social partners. On the sideline, the legislator is involved in establishing the framework and can use legislation where necessary. The Netherlands is quite different from other European countries in that the partners are legally required to have a professional pension entity setting up the arrangement. This can be either a pension fund or an insurance company. In this trilateral relationship, the pension fund carries out the pension arrangement at an employer’s behest for the advantage of employees and pensioners.
The new Pension Act needs to be more transparent regarding the responsibilities and rights of these parties. For example, in cases of underfunding it can be asked, to what extent is the employer responsible and to what extent will this be at the cost of certain participants? This, and other issues, such as indexation policy and premium (contribution) holidays, will be settled in the letter of assignment that will replace the current financing agreement.
Liabilities marked to market
Aligned with the preparations for the Pension Act, the PVK is working on the FTK, which will show some major differences compared to the current supervisory regime, the Actuarial Principles Pension (APP) funds .
One of the most important changes under the Pension Act and the FTK, will be the equal treatment of the valuation of assets and liabilities. Whereas assets are marked to market under the APP, liabilities are discounted against a fixed rate of 4%. When switching to market rates, there is still uncertainty about what interest rate curve to use. The starting point is a risk-free curve.
The PVK holds the view that this should be a government curve, whereas for long maturities (possibly defined as 10 years or more), this would be complemented with swap rates. Important arguments against this are that it should be an ‘investable’ curve and that government bonds function as a safe haven which can create distortions in this curve.
The swap curve seems better suited in this context. The PVK will publish the curve periodically and intends to also provide a real (inflation indexed) curve to value unconditionally indexed pension obligations. Because of the limited data for the latter, a pragmatic solution from the PVK may well be that inflation will be kept to the maximum 2% target inflation of the European Central Bank (ECB), whereas for wage inflation there will be an additional mark-up of 1.5%.
The Pension Act will provide a clear tool to separate legally unconditional and conditional indexation. If the rules of communication are adhered to, conditional pension obligations will be labelled as legally conditional. In any other case, they will be legally unconditional. After a lot of debating, it seems that legally conditional indexation will not need additional reserves.
Minimum and solvency tests
An important change concerns the minimum test and the solvency test. The Pension Act will contain a minimum test where the funding level should be above 105%, alternatively a one-year recovery plan has to be filed immediately. The new solvency test imposes a buffer, which with 97.5% probability should be sufficient to cushion market movements of current asset allocation during a period of one year. After this year the funding level should be 105% or more.
The solvency test should result in a 130% required target funding level for an average pension fund with a 50/50 asset mix. When below the target level, the pension fund has to file a 15-year recovery plan. One occurrence is that during this period the probability of underfunding at certain times surpasses this 97.5% level. The Pension Act does not clearly state how the 97.5% probability should be determined. Under the FTK, pension funds can either choose to use the standardised model, or their own internal model, which has to be approved by the PVK. The standardised model is likely to be easy to implement.
Equity, private equity, real estate and commodities are put to a k% stress test, where k% varies per asset class. Bonds and liabilities face a stress test with an r% change in interest rates, where r% varies per maturity. Effects of derivatives will be included. Furthermore, liabilities will undergo actuarial stress tests. When losses on assets and rises in liabilities are considered, a minimum buffer needed to cushion these scenarios can be calculated.
The k% and r% are still to be determined by the PVK, but will have to comply with the 97.5% probability and the required 130% target funding level, which are legally imposed. Two observations stand out. The first is that the standard model will probably use high correlations. It is likely that correlations will be equal to ‘one’ between asset classes, and a correlation between equities and interest rates will be very high. This implies a very low correlation between equity investments and liabilities, resulting in a very conservative approach. The second is that the 97.5% probability has to be translated into fixed values for the parameters k% and r%. The PVK can either estimate k% and r% with 97.5% historical confidence, or calibrate the k% and r% towards a situation where current funding levels for the average pension fund will match funding levels under the APP.
Because of the rigid and conservative contours of the standard model, it is likely that creating its own model will be of added value to the pension fund. However, because of the complexity, it will not be possible to develop such a model in-house for particularly small pension funds.
Funding level volatility
A consequence of the move from the fixed 4% discount rate to market interest rates (now approximately 5%) is that the present value of liabilities with an average duration of 16 years will in first instance decrease by 16%, resulting in a funding level increase under FTK of 19%. Another consequence is that, under market rates, reported liabilities will become extremely volatile. The odd situation happens, that whereas under the 4% regime it seemed to be risky to lengthen duration (liabilities increase with a steady 4%), under the FTK the duration of the bond portfolio cannot be long enough.
FundPartners’ analysis shows that average funding level volatility under the FTK will be around 1.7 times as high as under the APP with unchanged investment policy. By lengthening duration in cash, or via the use of derivatives, this risk can be significantly decreased. However, fully immunising this risk, lengthening duration will cause a funding level volatility of over 1.6 times current volatilities under APP. The question whether and how pension funds can anticipate the arrival of FTK in 2005 remains to be answered by the PVK.
Of course, we have to bear in mind that it is reported risk or reported funding level that changes, but the actual risk position doesn’t change with the new legislation. At the same time it is important to acknowledge that this change will be influential in the decision making process of pension funds, simply because pension funds are judged on these criteria.
Cost covering premiums
The height of premiums will be much more regulated. The Pension Act will prescribe the use of a cost-covering premium, plus a recovery surcharge (to grow to a sufficient funding level), plus a surcharge for target indexation (even when no reserve is required). Only when the funding level secures long-term indexation, premium holidays or return of contributions could be allowed, as determined in the letter of assignment.
The use of market rates for present value calculations of new liabilities also creates volatility in the requested premiums. For example, a 1% lower interest rate would require an approximate rise in premiums of 20%. Obviously, such a swing in premiums is not desirable for either the employer or the employee. The PVK is currently researching possibilities to dampen this volatility. For calculation of the funding level, the actual interest rate curve is expected to be used while for policy purposes, a moving average could be applied.
The stricter regulation, supervision and accounting standards come at a time when we have just experienced a three-year period of bear equity markets, with pension funds on average 40-50% invested in equities. However, the developments date back long before the start of this bearish period, and even before the bullish period before that.
In those bullish years, contributions were decreased, as in the case of ABP, from over 25% of gross salaries in the 1970s to less than 10% in the 1990s. Some pension funds even introduced contribution holidays or reimbursements. In the bearish years that followed, funding levels that on average surpassed 150% at the end of the 1990s, approached levels around 100% as a consequence of both the bearish markets and the far-below-cost price premiums received by the pension funds.
If investment management is not changed, current funding levels lack a buffer that is needed to cushion the increased volatility under the FTK. Therefore, many pension funds will want to decrease volatility and use duration lengthening investment techniques at the latest in 2006. Because of the size of the Dutch pension industry, this will cause, sooner or later, a run on long-term fixed income investments. As can be observed from similar situations in the UK and Denmark, this will influence the long end of the curve. One could argue that the pension funds that act soon have an edge compared to others.
An issue for many pension funds is that they do not know what the Pension Act and FTK will finally look like, and whether the target introduction date of 2006 will be met. Final texts for both the Pension Act and the FTK are expected to be published in the fourth quarter of this year. Most pension funds are already in the process of educating their boards on the many changes that take place, and this process will continue after publication of the final texts of both the Pension Act and the FTK. One challenge in this regard is to change the risk perception. Where the board used to steer on minimising asset volatility, under FTK, asset volatility needs to rise considerably to realise a lower funding level volatility.
A few larger pension funds have already anticipated the FTK in their investment portfolios. Other pension funds are still considering what to do, also in the light of historically very low interest rates. Most smaller pension funds await 2005, or even 2006, before acting. They may, however, follow shortly as large pension funds take action. The consequence could be that smaller pension funds have to lengthen duration in a market where the curve in the relevant segment has flattened significantly.
Mandatory sector pension funds may perceive an extra hurdle to take in implementing a duration lengthening strategy. By decree, they are forced to annually compare their performance to that of a standard portfolio, based on benchmarks (also referred to as the z-score). If the mandatory sector pension fund doesn’t pass the test, employers are no longer obliged to participate.
An issue they face ,if they have decided to change benchmarks, is what benchmark to choose. There is a lack of benchmarks with a duration of 10-25 years, and even if they are available, they are hard to use in combination with a duration lengthening strategy using derivatives. The inflexibility and lack of suitable benchmarks were reasons to have a look at the current decree. To allow mandatory sector pension funds for more flexibility in timing and use of benchmark, the decree should be adjusted accordingly.
For many corporate pension funds, it is an issue that the PVK doesn’t necessarily follow the path that is followed under IAS 19, for example, the type of yield curve to be used. If the two differ only slightly, this will mean separate reporting. Therefore, it will be welcomed if the PVK seeks alignment with IAS 19 for exchange-listed companies as of 2005, and it’s Dutch elaboration rule 271 for non-listed companies.
With regard to indexation, many pension funds with conditional arrangements already comply or intend to comply with the conditions within the Pension Act to have a legally conditional arrangement. However, as the case with Campina pension fund shows, by customary law, pensioners were able to legally enforce indexation of accumulated benefits, because the funding level was judged to be adequate. Therefore, a case can be made that accumulated benefits may not be influenced by complying with the conditions as under the Pension Act.
Changes in legislation, supervision and accounting are demanding much from the Dutch pension industry, especially after a relatively long period with negative equity returns. Some important questions remain to be answered.
How will politics and the PVK cooperate in the implementation process of the new Pension Act and the FTK? Will the introduction date be January 1, 2006? Will some sort of a transition regime be used? Many pension funds are awaiting final texts and some smaller pension funds will face a very difficult job to adapt to all these changes.
For this latter group, it may no longer be economical to operate on a stand-alone basis. They may need to reinsure, or have their arrangement executed by a third-party. Others may struggle with all administrative changes coming up, lacking the resources to adequately anticipate on adjustments needed within their investment portfolio.
Then there remains the question whether the pension funds should lengthen duration now, after several years of negative equity returns and historically low interest rates. Despite all these issues, few will disagree that, on balance, changes in pension regulations are a step forward. Pension funds, their sponsors and their participants, will all benefit from more clarity in the pension deal because of better communication, clearer promises and modernised risk management practice.
Menno Wiersma is head of structured products, FundPartners in Laren