During the past year, ATP, Denmark’s largest pension scheme which serves the public sector, with almost 4.9m members and more than Dkr300bn (€40bn) under management, has amended its risk management process in two important areas. It has developed a new asset-liability model in-house and then implemented what it calls a dynamic process for adjusting its asset allocation in response to the fund’s ability to bear risk. ATP believes that these two innovations combined form a unique phenomenon n the pension fund industry.
ATP says its overall investment strategy must always be tailored to suit its long-term pensions commitments and to achieve this objective, ATP has implemented its own stochastic asset-liability model that it is has been developing for the last two years. ATP had always worked in collaboration with a major US based bank in the past when undertaking its asset liability studies but the 30 year time scope of its new model means a vast number of possible balance sheet positions can now be viewed and simulated more readily. According to ATP, this means it is now more possible to weigh the returns against risk in both the long and short term. Moreover, the new model provides the basis for the fund’s long-term asset allocation targets with regard to foreign and domestic equities and fixed income, among others, as well as hedging of interest rate risk.
But there are other advantages to having an internally developed and maintained asset-liability modelling operation. The model can be used much more flexibly and be run more frequently than external models, which is particularly important in turbulent periods. It leads to a better understanding of asset-liability modelling issues within the fund and to improved decision-making, such as the dynamic changes, as explained below, made to the asset allocation policy, ATP says.
ATP’s risk management has two basic objectives: to ensure pensions preserve their long-term value and to ensure reserves are able to withstand short-term pressure resulting from adverse financial market conditions. Which of the two principles takes priority at any given time depends on the size of the reserves. Protecting long-time value is most important when reserves are high, but takes a back seat when they are low, in which case ATP will focus on limiting the risk of further losses in the short term.
ATP says asset-liability studies are traditionally based on the idea that benchmarks for the actual asset allocation policy remain static, or supplemented at times with limited annual revisions. However, the fund’s studies in the past three years have revealed this is rarely the case over prolonged periods. ATP believes asset-liability management is dynamic and so it is important to have well-structured guidelines in place to help in situations where the asset allocation needs to diverge somewhat from long-term targets.
ATP’s solution is to adopt a dynamic principle that allows ATP to tailor its investment risk to sustainable levels with respect to the levels of the reserves. The guidelines for the risk are drafted by the board and ATP says an adapted version of its new asset-liability model already forms the basis for the fund’s daily risk calculations. If risk is shown to be too high, the investment risk is reduced by decreasing the level of foreign equities in the benchmark. The key here is that risk is reduced well before it has a chance actually to threaten the level of reserves. However, if risk is too low, the fund increases its allocation to equities. Overall, this process ensures long-term asset allocation targets can eventually be achieved when reserves are at a high level but does not affect the possibility of having the portfolios managed on an active basis.
ATP believes the dilemma posed by the fact that the optimum investment strategy may differ in the short and long term is common to all pension funds and it has serious potential implications for risk and return ratios. Moreover, it says it is unsatisfactory to base decisions of such importance on ‘ad hoc’ considerations. ATP believes decisions about when and by how much asset allocation should change should be taken well in advance within a framework that is firmly linked to asset-liability studies. A structured approach raises the odds of making consistent and well measured asset allocation decisions when reserves are exposed to negative elements and hence it is important to be able to act swiftly without having to go so far as to reinvent the wheel.
Highlights and achievements
ATP’s new asset-liability model successfully solves the dilemma posed by the fact that a fund’s optimum investment strategy may differ over different time periods.
Bringing the ALM exercise in-house creates additional efficiency in terms of cost and alleviates the need for collaboration with external parties.
Adapting its ALM studies to take a more frequent, detailed and robust view of its positions gives ATP greater understanding and control over the overall asset allocation and risk management policies and ensures its retains its profitability and reserve levels adequately.
Basing asset allocation changes on ad hoc considerations is seen as poor management and ATP takes a proactive approach that it is a case of ‘when’ and ‘how much’, not ‘if’, allocation needs to be adjusted to suit market conditions. Its new asset-liability model in conjunction with its revised risk control policy is ideal to ensure it can pursue its goals using this new approach. This can be as simple, but effective, as revising various asset class weightings in the short term, depending on the levels of risk and the reserves.