Collective buffers in a new Dutch pension system with individual pension pots – The impact of the non-negative constraint
The Dutch pension system is considered to be among the best pension systems in the world. For many years, the Netherlands has been second place in the ranking of the Melbourne Mercer global pension index (Mercer, 2007), just after Denmark. The Netherlands receives high scores on sustainability, adequacy and especially integrity compared to other countries. Despite praise words from Mercer, there is much room for improvement for the Dutch pensions industry. Due to the substantial losses on equity markets in the financial crisis of 2008, the low discount rates afterwards and an increasing life expectancy, most Dutch pension funds became underfunded. Few pension funds were able to index the pension rights of their participants and some even had to cut pensions. The financial crisis and its effects triggered both Dutch pension fund participants and policymakers to start a discussion about the future of the Dutch pension system.
In 2012, the previous government already increased the retirement age in the first pillar pension (AOW, the Dutch pension pillars are described in section 2.1) and, in 2015, it decided to speed up that increase. Afterwards, to also reform the second pillar, that same government asked the Social Economic Council (Sociaal Economische Raad, SER) to come up with possible new pension system and investigate its effect. The SER is the most important advisor of the government regarding social economic policy and consists of representatives of employers, employees and independent specialists. Partly based on reports written by the SER, the current government took a stand in the discussion. Specifically, in their coalition agreement of 10 October 2017, the government set guidelines for the new pension system (see section 2.3.1) and urged
the SER to come up with an agreement to reform the system. Moreover, the current government proposed a system with individual pension pots together with a collective buffer, described in the SER (2016) report as option IV-C-R, see section 2.2.4. However, the coalition remarks that this collective buffer should never be negative. As such, the largest Dutch labour union (FNV) said, in their reaction to the coalition agreement, that this `non-negative constraint’ is their biggest concern regarding the pension paragraph in the coalition agreement. The FNV advocated that this constraint is an obstruction for `Intergenerational Risk Sharing'(IGR), while the government is in favour of this constraint, to not forward deficits to future generations. This disagreement between the coalition and FNV is the primary motivation for this thesis.
In this thesis, the effect of the `non-negative constraint’ on the welfare of both the current and future generations is provided, assuming a new pension system described by the SER as IV-C-R. This thesis contains an analysis of the system proposed in the coalition agreement and the reports of the SER. Additionally, alternative buffer systems and the effects of these systems are provided.
In section 2 the background of the problem is defined. Specifically, a description of the current Dutch pension system is provided and the content is given of the new pension systems as proposed in the two SER reports. Additionally, the view on the pension discussion of both the current government and the FNV is further explained. The aim of the collective buffer is to share risks between generations, section 2.4 therefore contains a description of the positive effects of sharing risks between generations (IGR), based on previous research. Furthermore, in section 2.4.4 it is explained that accepting possible negative buffers, can lead to an unsustainable system if the in ow of new premium is lower than expected, known as the discontinuity risk.
In section 3, a Monte Carlo model, based on the IV-C-R option of the SER is described. In the simplified pension fund model we only consider stock market risk and abstract from all other forms of risks. In this section, the assumptions on the population, financial market, premium and livecycle are explained.
Furthermore, a description is given of how the pension benefits are calculated. Using the model described in section 3, three core buffer strategies are compared with each other;
• a system without a buffer
• a system with only positive buffers
• and a system with a buffer that can become both positive and negative
In section 4, the outcomes of the simulation are provided. The expectation and confidence intervals of the pension benefits are provided for the three different buffer strategies. Using utility theory, the welfare of the different strategies are compared. Furthermore, the welfare effects of two alternative buffer systems are
provided, and the robustness of the assumptions are tested.
Finally, in section 5 it is concluded that a buffer can be welfare improving for all generations (compared to a system without a buffer), but only if the buffer is allowed to be negative and a risk premium is provided for future generations. In addition to that, it is stated that there is a severe discontinuity risk when negative buffers are allowed.