In a market environment where the funding ratios of pension funds have suffered severely, asset liability management (ALM) is still a very helpful tool in understanding the various ways a pension fund might restructure its financial position and to repair its balance sheet. ALM is still particularly helpful and useful in testing funding ratios and in defining alternative investment portfolios.
Let us first have a look at the typical funding ratios of defined benefit pension funds in the Netherlands. I am referring to a typical Dutch pension fund where generally the schemes’ liabilities are linked to price or wage inflation. In most cases the indexation of the liabilities is on a conditional basis.
Average funding ratios declined from 150% at the end of the year 1999 to 140% in the year 2000, to 125% in 2001 and to around only 105% at the end of 2002 which is really in the danger zone. In the year 2003 we have seen some recovery up to now.
In general there are four instruments available to pension funds towards managing a recovery in the level of funding:
o simplifying the pension plan (for example, a shift from DB to DC);
o reduce the level of compensation given for price inflation or wages inflation;
o increase the level of the contributions;
o lower the downside risk of investments (a low funding ratio means there is little room for taking on high investment risk).
In this article, I focus on three instruments only: inflation, contributions and investment policy.
Looking back to 2000 (see chart 1) and assume a funding ratio of 140% as it was at that time on average. In our chart a reduction by means of a refund of contributions is assumed of 40%. In that year many pension funds were still calculating their contribution rate lower than the actuarial cost price. An estimate of those refunds is around 40% at that time.
Using this input the projection of funding in the next 20 years is as in chart 1. On the left hand scale, we show the level of funding and on the right hand the level of contributions.
We see then that the funding level sinks after nine years below the level of 100%. Contributions have to rise dramatically, in fact they have to double, (see chart 1, right hand side).
So we see that, even with a comfortable initial funding level of 140% as it was then on average, funding would not be sufficient over the years. The main reason for this is that pension funds in general have had a rate of contribution which was much lower than the actuarial cost price.
The first logical steps to solve these financial problems are to withdraw reductions of contributions and to postpone the compensation for inflation. In fact the Dutch Supervisory Board, the PVK, demands that pension funds calculate at least the actuarial cost price of their pensions liabilities.
So let us assume that the pension fund has arranged a recovery plan to increase the funding in a few years to 120%. Such a plan could include the two steps of withdrawing contributions reductions and postponing the inflation compensation (see chart 2).
Once we have reached the target level of 120%, then we are out of the danger zone of 105%, but the question then is if we definitely are out of the danger zone. We can see from Chart 3 that this is still not the case and the steps we have taken are not enough.
The main reason for the funding levels not holding is the negative real return on investments in this projection and the influence of aging participants in the scheme with the impact of the ‘baby boomers’. This is the case facing most pension funds in the Netherlands. Another issue is that a contribution rate only matching the level of the actuarial cost price is not simply enough; we need an extra contribution inflow to the pension fund to keep the surplus in shape.
Examples of what this pension fund in this case could do are as follows:
o to lower the target for indexation of the liabilities from wage inflation to price inflation (let us assume this makes a difference of 1% indexation):
o to add around 10% to the contribution level for several years, to create a sustainable surplus;
o to add some 1% yearly to the level of contribution to keep up with future demographical developments, regarding especially the impact that baby boomers will have on the early retirement part of our pension schemes.
If we take these measures we can see (see chart 4) that the funding will be sustainable and that the funding level even increases over time. This is very welcome because a funding higher than 120% lowers the future risk of underfunding to a more acceptable level (see chart 5).
Chart 5 shows the probabilities of underfunding for the next 20 years for various asset-mixes, and given an initial funding of 120%. It is clear that, especially in the first years, the probabilities of underfunding are too high. For example: with a proportion of 40% equities in the asset mix (the yellow line) the probability of underfunding is around 6% which is indeed too high. This makes it necessary to lower the risk on investments.
To lower the downside asset risk, it is possible to change the structure of the asset mix. There seems to be enough room for this. Up to now, Dutch pension funds have invested only to a small extent in alternatives (only 3 % or so on average). So in what way can investment policy help to reduce the average probability of underfunding?
It is possible to come up with a set of alternative portfolios including categories as private equity, high yield bonds, convertibles, commodities, real estate and even hedge funds and index-linked bonds. Since in general these assets have a low correlation with traditional assets, a lower standard deviation could be expected for the whole portfolio.
Chart 6 shows for various levels of funding that a lower portfolio risk (in terms of portfolio standard deviation) delivers a substantial lower probability of underfunding. So we can see the impact of the asset-only standard deviation on the pension fund risk as a whole.
Of course the probability of under funding is higher as the initial level of funding is lower. I have calculated this for a range of funding levels from 105 to 120%. Reducing the portfolio risk with 2 or 3 percentage points reduces the average probability of under funding with some 5 percentage points. Especially in the lower regions of funding, this risk reduction makes a lot of difference. For example: if we look to the level of 120% (the bottom line in the graph) a decrease of the portfolio standard deviation from 13% to 10% (asset only) will lower the probability of under funding for the pension fund as a whole from 10% to 5% which is a rather substantial improvement.
An important matter of course is what effect adding low correlated assets will have on the expected returns. If adding alternatives will lead to a significant lower portfolio return, on the long term the overall effect of reducing the downside risk on investments will be a worse financial situation and a lower funding ratio. Of course this is not what we want. The price for achieving a lower risk might be too high.
Dick Wenting is general manager with the Medewerkers Apotheken industry-wide pension fund in the Netherlands

Topics