In early spring 2000, the report Continu op Koers (‘Keeping on Course’) was published jointly by the Foundation for Company Pension Funds (Opf), the Association of Industry-wide Occupational Pension Funds (VB) and PricewaterhouseCoopers. The aim of the report was to contribute to the debate on pension fund governance and solvency, a debate that has aroused a fair amount of interest given that the supervisory body for insurance companies and pension funds in the Netherlands, the Verzekeringskamer (VK), was preparing the Dutch new actuarial principles, or NAP.
The NAP are an updated version of the APP from 1997, which represented the (supervisory) situation at the time and were viewed as the first step towards more comprehensive and detailed actuarial principles for pension funds. The VK stated that all experiences, comments and criticism with respect to the APP would be given due consideration in the new version.
Although at the time the APP were seen as a clear step forwards, there was criticism. This was diverse, but primarily concerned three points:
q the prescribed basic interest rate of 4%;
q the inconsistency between the valuation of assets and liabilities, due in part to the prescribed basic interest rate mentioned above; and
q the static assessment of solvency.
The new report (as well as other publications) argues that any new actuarial principles should address these areas in particular. The report also pleads for a comprehensive solvency test that continuously assesses the position of a given pension fund. The driving thought behind the solvency test is the belief that any authority responsible for supervising solvency should focus on answering one question: will the financial position of the specific pension fund, given the policy and risk attitude of the fund’s management, guarantee that the fund achieves its objectives (ie, paying pensions) with a sufficient degree of certainty?
This comprehensive solvency test should include projections of the development of both pension payments and investments. The test applies methods that are used by many pension funds for asset liability management analyses, which are normally used as aids in setting optimal investment and premium policies. The solvency test, however, uses these investment and premium policies as input for the analysis, which results in a required capital. The test uses as starting points the current situation of the pension fund and various scenarios regarding the growth of investment performance, inflation, mortality and portfolio development, based on management’s expectations.
Using a fixed interest rate as a basis for calculating the capital in projections like these is less important since the test employs instead a margin for underinsurance or a ‘deficit allowance’. This margin for underinsurance largely determines the minimum position: a capital base is insufficient if the analyses indicate that this may lead to too great a deficit. A ‘deficit’ here is defined as a situation in which the capital base is (no longer) sufficient to cover the projected payments.
A compact test has also been proposed as an alternative for funds that are not (yet) in a position to perform the complex analysis of the comprehensive test. The compact test is an alternative approach that allows us to calculate a minimum required capital relatively easily. Because of the higher level of prudence built into the limited test, the test may in certain cases lead to a higher figure than that generated by the more detailed comprehensive test. The difference between the two is the price one pays for simplicity.
The outlines of the new actuarial principles were presented in March 2000. They represented the aim of the VK to bring its supervisory policy more in step with (international) financial developments in the area of risk management (higher degree of professionalism) and documentation (greater transparency, more explicit and real-value measurement).
A draft of the new principles was sent to a large number of parties concerned, such as organisations representing the interests of pension funds and insurers, as well as the Dutch Actuarial Association and the Royal NIVRA (the Association of Accountants).
The basic assumption for prudent risk management held to apply at all times has been set out as follows:
q consistent measurement of assets and liabilities on the basis of fair value
q explicit subdivision of the reserve into the following parts:
l best estimate;
l prudence; and
l mismatch.
q Emphasising risk analysis and risk management, while maintaining a statically measured solvency position on the balance sheet date.
Subdivision involves creating a clear distinction (transparency) between the best-estimate value of the pension cash-flows, the (actuarial) prudential provision for unfavourable mortality and cost development, and the mismatch provision.
To understand what is meant by ‘mismatch’, it is first necessary to understand how the best estimate is determined. The best estimate is equal to the value of future expected net cash flows, taking into account the forward rate structure for risk-free investments – that is, each cash flow is valued against a zero-coupon government bond of the same duration.
Opting for this type of investment would result in a perfect matching. An alternative mix inherently results in a mismatch with regard to yield, duration, convexity, credit-risk and valuta.
This mismatch provision should consequently be made. The VK has adopted an approach that closely resembles that proposed in Continu op Koers. Two alternatives are offered:
q using an internal model; and
q using a standardised method.
The internal model must meet a number of requirements. For example, it should not operate independently, but form an integral part of the process of planning, monitoring and managing the risk profile of the pension fund. Furthermore, the model must have a proven track record in accurately measuring risks. Proven accuracy may involve, for example, backtesting, whereby past estimated risks are compared with actual figures.
The internal model must meet a quantitative requirement in accordance with the solvency at risk principle, by which the provisions are such that the obligations in connection with the investments must be met to within a specific degree of probability (99%). Again this is in close agreement with the comprehensive model presented in Continu op Koers.
The standardised method offers an alternative. Here the mismatch risks are generously estimated on the basis of the method prescribed by the supervisory body. Using this method will result in higher prescribed provisions than with the internal method, so that pension funds are stimulated to develop a suitable and professional risk management model, and thereby qualify for the internal method.
The parties concerned have since had the opportunity to respond to the draft version of the memorandum containing the basic assumptions. In general, the parties have responded positively to the Verzekeringskamer’s policy proposals. They have pledged to work towards greater professionalism in risk management and to establish more commercially oriented supervisory structures. They have also recognised the wisdom of distinguishing three types of technical provisions. Indeed, this division makes clear which actuarial prudence the fund operates. Furthermore, the parties welcome the abandonment of a fixed basic interest rate and the possibility of using their own ALMs and risk models.
Not all responses have been positive, however. Pension funds face a specific problem with regard to the valuation of cash flows. Nearly all pensions are conditionally indexed, meaning that payments are generally adjusted to take inflation into account, though only if and in so far as allowed by the funds’ financial resources. This safety valve is an important dynamic management tool for pension funds, although it does hinder objective valuation of the obligations. In the case of unconditional (and, therefore, guaranteed) indexation, a ‘realistic forward rate structure’ would need to be employed. The problem is that the lack of indexed loans in the Netherlands limits the feasibility of this option. A solution for the conditional indexation is yet to be found. The VK takes the view that aiming for indexation is sufficient reason for build specials reserves for it. The pension funds, however, disagree with the VK on this point, stressing the point of conditionality.
Moreover, the pension funds feel that the VK is insufficiently aware that in practice pension funds in particular invest not only in bonds, but also in shares and property. In the memorandum containing the basic assumptions, the provision assumes that nominal obligations are completely covered by fixed-rate values, and that shares and property are used for indexed payments only after a period longer than 25 years. This indeed appears to have little bearing on the actual situation. One can allow for expected returns of shares (and the associated price risk) when determining the mismatch provision, but it can never lead to a reduction of the calculated technical provision.
In addition, the funds feel that the security of 99% stated in the draft memorandum is far too high; it would result in a drastic increase in the provisions and thus in high extra premiums (and consequently higher labour costs).
The most serious criticism from the pension funds, however, concerns the fact that the VK assesses the solvency position on the basis of the settlement scenario. The VK takes the view that the pension fund as an independent entity must always be able to either settle the obligations (accrued pension rights) or transfer them in the event the sponsor is no longer able to pay the premiums. Although this is defensible from the position of the participant, this situation is very uncommon. That premiums may be spread over a longer period is an important distinguishing feature of the way pension funds work. The sponsor risk is certainly real for some funds, but negligible for others. This makes it part of the specific risk profile of each pension fund.
This is a crucial point that may well start a debate in which the security of pension payments is seen in a broader context of employment conditions. A higher degree of security after all means higher levels of reserves and, consequently, a more expensive pension. Why should this not form part of employment condition negotiations?
It is not to be expected that this discussion will take place as part of the NAP. The VK bases its supervisory role on the law which still assumes the security of acquired pensions rights. Many outstanding issues remain. Pension funds and insurance companies are waiting for the amended memorandum later this spring following consultations. It is likely that implementation will not take place before 2003. Until then, the debate will continue.
Marco Vet is a senior manager in PricewaterhouseCoopers Actuarial Consultancy Services in the Netherlands