There is room for improvement in the government’s plans to introduce variable annuities for DC pension funds, according to Wichert Hoeckert 

The Dutch government is proceeding with a proposal to introduce variable annuities for defined contribution (DC) plans. Legislation was sent to Parliament in November 2015. 

Variable annuities will allow members to hold risk-bearing assets beyond retirement, as well as share risks between groups of members. The target effective date is 1 July 2016. It is expected that the proposal should lead to improvements in DC arrangements. However, we expect it will create complexity for both members and providers. For these improvements to materialise, participants and consumers face challenges.

The complexity would be caused by the introduction of a one-time, irreversible choice at retirement between fixed and variable annuities. The member would have to choose between providers. Insurers and pension funds would be able to offer variable annuities, as would premium pension institutions (PPIs), which are currently not permitted to offer annuities because they are prevented from bearing insurance risk. Anyone with capital accrued in a DC plan would be free to switch provider. Providers would offer different types of variable annuities, such as those with or without: 

• Shared investment risk;
• Shared macro-longevity risk;
• Insured elements (most likely micro-longevity risk);
• Smoothed financial and non-financial results; and
• Fixed indexation.

The principles underlying risk sharing and smoothing mechanisms are included in the proposal. An important principle is to prevent ex-ante risk transfers. As a result, smoothed annuity adjustments can only be assigned to members of a collective arrangement at the time the profit, or loss, occurrs. For example: a 6% loss can be translated into an immediate adjustment of the annuity by 6%, but the adjustment could also be spread over five years. This would lead to an annual downward adjustment of about 1.4%. Anyone entering a collective arrangement during that five-year period should be excluded from any downward adjustments related to losses prior to their retirement. 

This is complicated as it is, but becomes increasingly complex if different sources of risk are treated separately and attributed to separate collective arrangements. The proposals neglect additional sources of risk, such as those related to provisioned future cost and marital status.

Although the measure primarily looks at choice at retirement age, providers would have to investigate the intended choice of members prior to that date, as lifecycle design should depend on that choice. This increases fiduciary responsibilities, based on the know-your-client and prudent-person principles. Consequently, the introduction of variable annuities would impact all existing DC arrangements and, in turn, may make these more attractive.

What is important is that the increased complexity, along with the risk, will be transferred to members, who will have to make choices that can have significant impact on their income beyond retirement. Among the questions the changes raise, a predominant one is which market demand is actually served by this much complexity and detail.

It is often observed in pensions discussions in the UK and the Netherlands that although people wish for more freedom of choice, they do not exercise it once they have it. 

In lifecycle design this implies that good defaults are all the more important, as a majority tend to follow it. In the proposals for variable annuities, it is left up to the institution to determine the default – which could be either a fixed or a variable annuity, or any combination of the two. In the explanatory notes to the proposed law, the ministry of social affairs indicates that it expects 90% of retirees in DC arrangements to opt for variable annuities in the short-term, and about 50% in the long-term – most likely based on an expectation that interest rates are to increase. This estimate seems unrealistic, unless a variable-annuity product is offered as default – which may be difficult to justify, in particular, for lower pension amounts.

In comparison with an earlier proposal, the legal draft initially broadened freedom of choice in a move that was particularly important for pension funds that offer DC arrangements – mainly alongside a defined benefit (DB) or collective defined contribution (CDC) plan up to a certain salary cap. This proposal meant that members with DC-accrual would be free to purchase their annuity elsewhere. The legislator seemed to assume this would work towards a more level playing field. One could argue that this is not the case, as for pension funds this freedom to move works in one direction only: members can purchase their annuity elsewhere, but annuities within the pension fund cannot be purchased out of capital that has been accrued at another provider.

A further consequence of this could have been that pension funds would have decided not to offer variable annuities, as they might have feared the size of their collective arrangement would become too small for effective risk sharing. As pension funds serve homogenous populations, this freedom may have a negative impact on the quality of the product if the consequence is that pension funds decide not to offer variable annuities.

We therefore welcome the amendment announced in January stating that pension funds offering both fixed and variable annuities will not be required to allow members to take their capital out of the plan.

If, however, the Dutch legislator really wants to increase the freedom of choice between providers to improve competition, it is unclear why banks should not be included as providers, in the same way as this is already possible in third-pillar products. This may force the pension sector to keep products lean, mean and cost-effective.

Another area where variable annuities could be of added value for funds is regarding net pensions. The number of pension funds offering annuities for net pensions – for EET (exempt, exempt, taxed) accrual for incomes above €100,000, introduced in 2015 as part of new fiscal restrictions – is relatively small. 

There are several reasons why pension funds do not offer these. Fiscal hygiene is one – additional rules warranting effective ring fencing between assets that have been taxed and those as yet untaxed are demanding, for example they are required to use mortality adjustments that are specific to high-income groups. Unattractive annuity pricing may be another one – regulation requires the inclusion of full funding of required own funds, which tends to be more expensive than the expected indexation in the contract, particularly at current coverage ratios.

With the possibility of the cap eventually reducing – which would lead to higher EET-accrual – we expect more pension funds to offer net pensions. Explicitly separating the arrangement in an APF would be one way to avoid fiscal hygiene regulations, but offering variable annuities for net pensions would be another.

It should go without saying that the proposals for variable annuities are at the risk of being overdesigned. Theoretically they seem an improvement over the current DC arrangements, although the freedom of choice may lead to complexities that could add cost and diminish appeal. This is important given the current direction of travel of the Dutch pension system. Overdesigned annuities may be one step away from no annuities at all – and this is an element of the Dutch pension system that we should not lose. 

Wichert Hoekert is a senior consultant at Willis Towers Watson