In recent years, liability-side issues have had a very substantial influence on Danish pension fund ATP’s investments. During the past year, therefore, ATP has radically changed its approach to liability-driven investments. Its work in splitting the investment portfolio into separate alpha, beta and liability hedge portfolios has helped ATP carry off the themed award for liability-driven investments.
The problems inherent in pension fund liabilities have been well-documented. Since the beginning of the new millennium, ATP and other pension funds have seen their liabilities rise rapidly because of falling interest rates (assets as well as liabilities are marked to market in Denmark) and because of increasing life expectancy. Furthermore, investment returns have been volatile.
In the light of these threats, over the years ATP has put in place a number of far-reaching investment policy measures. The most important of these was the decision taken in 2001 to hedge the interest rate risk on ATP’s liabilities, using mainly interest rate swaps. In 2003, another important decision was made — to implement a dynamic asset allocation rule. This rule links investment risk directly to ATP’s reserves and the risk tolerance of the board.
The changes were made easier by ATP’s strong focus on asset-liability modelling (ALM) including the development of an in-house model a few years ago.
This year, ATP has placed its asset management, actuaries and ALM under common control. This change in the organisational structure has led to much closer contact than before between the ALM analysts and the actuaries, and brought about important improvements in the modelling of liabilities. One example is that longevity effects have now been included in the ALM model.
Partly as a result of these measures, ATP’s funded ratio has improved considerably since 2003, and risks to the fund’s solvency have subsided. ATP’s investment returns stood at 18.9% in 2004 and at 14.5% in the first half of 2005, bolstered by high returns on interest rate swaps and Danish equities.
But ATP says that significant challenges remain. First, prospective returns on most asset classes are expected to be modest, at best. Secondly, life expectancy has risen in recent years, and may continue to rise in the future. Third, large negative investment returns need to be avoided, because the implied fall in reserves would reduce the risk budget and make it potentially very difficult to achieve the fund’s long-term goals.
ATP says that the traditional investment approach, which focuses on generating high returns compared with market benchmarks, cannot really cope with the challenges faced by its own portfolio. So instead, the pension fund has decided to emphasise liabilities and absolute risk and return.
The fund says that there are two main objectives in managing its investments. One is to ensure that its reserves – the bonus potential – can withstand adverse financial market developments so that the fund remains solvent. The other is to achieve a high return to protect the long-term purchasing power of benefits.
ATP’s new investment approach reflects these twin objectives. The basic idea is that the portfolio consists of three independent elements – liability hedge, beta and alpha. Each portfolio contributes in its own way to ATP’s objectives.
The starting-point for the new investment approach is the risk-minimising portfolio, which hedges liabilities to the greatest extent possible. This will ensure ATP’s reserves and its commitments to its clients.
However, this risk-minimising portfolio would normally not generate any excess returns. So to generate higher returns and achieve the second objective – protecting the purchasing power of benefits – it is necessary to accept some investment risk. Additional returns come from two sources: ‘beta’ (the excess return of asset classes over the risk-free return) and ‘alpha’ (investment skills). Beta and alpha are independent.

Normally, the asset class decision – beta – is made first. Next, a decison is made as to whether or not to try to generate alpha from the asset class. This “beta trumps alpha” approach is not efficient. So under ATP’s new approach, the amount of investment risk allocated to each is determined by ATP’s risk budget. The size of the risk budget depends on the level of reserves and the fund’s objectives and risk tolerance.
Separating the portfolio in this way has had a profound effect on the risk management of the portfolio, its investment processes and organisation design. The three portfolios – alpha, beta and risk minimising – are independent and will be managed independently by separate teams. ATP says it thinks its approach is more flexible and focused than traditional approaches, and that it fosters better risk management and more accountability.
ATP says its new approach is consistent with its vision of providing asset management services of the highest international standard.