An onslaught of new regulations is prompting the Dutch pension fund industry to reevaluate the way plans are both managed and monitored. To industry outsiders, the move to value liabilities on a fair-value basis may seem to be nothing more than a smart accounting change, yet those intimately familiar with our industry understand its wide-reaching implications.
First, this move will increase the volatility within
the plan as the duration gap between assets and liabilities becomes an important factor of risk. Second, inflation becomes a major factor in examining the plan’s current investment strategy and future solvency. And finally, plan sponsors will have to focus more on the pension fund’s surplus, shifting their risk management approach from “value at risk” to
“surplus at risk”.
All of this has prompted many plan sponsors to ask, “How do we translate this into best practice?”

The demise of static benchmarks
One answer to that question involves the use of benchmarks. In the Netherlands and around the world, the pension fund community needs to consider shifting away from a reliance on static strategic benchmarks, moving instead toward liability benchmarking and strategic policy setting.
In the past, Dutch pension funds have relied upon static strategic benchmarks to measure the performance of their pension funds. Strategic benchmarks have been put forward as the practical implementation of capital theory which suggests the optimal portfolio is made up of all the wealth in the world leveraged up or down to reflect a given investor’s risk tolerance. In this context, “optimal” applies to expected return, and not in terms of minimising risk relative to the liabilities of the plan. Pension plans have often translated the theoretical optimal portfolio into a manageable strategic benchmark of a mix of equities and bonds. The average Dutch pension fund allocates roughly 55% of its assets to fixed income investments, and the majority of that is held in euro fixed-income securities. The average duration of the fixed income portfolio is around five years.
Until the mark-to-market rules of the Financieel Toetsingskader (FTK) go into effect in 2006, liabilities are discounted annually at a fixed rate
of 4%.
With the advent of FTK however, the liability side of a pension fund will be as volatile as the asset side, and funds will need to focus on the liabilities as much as they have on assets. The average duration of the liabilities is around 15 years, creating a duration gap of approximately 10 years between the current fixed income portfolio and the liabilities. The impact of mark-to-market valuation of the liabilities is therefore significant.
Under the old guidelines, periods of sustained interest rate decreases like we’ve experienced since the beginning of the 1990s led to an increase in the valuation of the bond portfolio and therefore - other things being equal - to an improvement of the coverage ratio. Under the new guidelines however, declining interest rates will bring higher liability valuations well beyond any increase in value of the bond portfolio, thereby shrinking the coverage ratio.
With the implementation date of the FTK nearing, Dutch funds are now beginning to look at the duration gap between assets and liabilities. Increasing the duration of securities on the asset side is one way to solve the gap dilemma, if only it were that simple. What’s more, there’s no bona fide way to assess the duration of equity instruments. We’ve seen plenty of debate around how the calculation should be made, but no consensus has been reached.
Other ideas to solve the duration gap issue have included shifting the asset mix of pension funds toward bonds, an attempt to kill two birds with the same stone by decreasing the duration gap and lowering buffer requirements. Unfortunately, many of these solutions limit a pension fund’s diversification and ability to gain alpha.

Liability benchmarking and strategic policy setting are the answers
The optimal solution is to institute some form of liability benchmarking. In order for this to be put into effect, pension funds first need to articulate their risk preferences as the surpluses in their plans grow or retract. When funding ratios are low, some plans opt for a conservative, risk-averse approach. Other plans have the opposite response, deciding to take risk off the table and lock in funding when
a surplus is
present.
For corporate plans, which are in a slightly different position than industry sector plans, the size of the pension fund relative to the sponsor´s balance sheet in relation to the risk of underfunding will also be a determining factor in the move to liability benchmarking. Given this, it is unrealistic to predict a mass sell-off in equities by Dutch pension funds as the FTK regulation is implemented.
However, considering the dynamic relationship between risk aversion and funding is essential when setting benchmarks for a pension fund. Doing so allows pension funds to shift from relying upon a simplistic strategic benchmark to a sound, strategic policy that will guide the trustees of the plan as market- and fund-specific events unfold.
In addition, pension funds need to take into account the varying levels of alpha they expect to gain from their investments. Through the use of utility functions and intertemporal optimisations, the consultant and fund management community can assist pension funds in identifying the risk premia and risk tolerance which are unique to each plan.
World-renowned economist Robert C Merton, 1997 winner of the Nobel Laureate in Economics, pioneered the theory of how to allocate assets to meet changing liabilities through time in the 1970s. In practice, though, only very simple problems could be addressed using his research. However, a number of recent developments have made it possible to solve asset allocation problems as complex as those faced by pension funds. A recent paper in the Journal of Finance by Detemple, Garcia and Rindisbacher set out a technique that allows such problems to be solved in practice. State Street is currently working with them to develop a version of their work tailored to pension funds. There are others, of course, who are treading similar paths.

Embracing liability
benchmarking
Once a plan has embraced the concept of managing both the asset and liability side of the pension fund, it must move to the next level. In today’s world, managers need to be selected and then monitored. However, in the world of liability benchmarks, there is an added level of complexity.
A manager can no longer be measured simply against a static benchmark. Instead, pension funds need to be assured that the manager has complied with their risk preferences, taking into account the present funding ratio of the plan and the current level of expected returns. Once a pension fund is satisfied that its chosen manager has employed an appropriate risk budget, it needs to make sure the manager has employed that risk budget well.
This in and of itself becomes a complex task, as the risk budget is changing in two dimensions, both with the plan’s surplus or shortfall and with expected returns. Now, however, it is possible to take into account the correlation between the plan’s assets and its liabilities.
For example, Dutch pension
plans often implicitly or explicitly guarantee pension payments indexed to a sector wage inflation indicator. Although it may be difficult to bring these factors in with precision, and even more difficult to find proxies in the bond market to generate returns in line with these rather specific indices, it is better to recognize some of these correlations than to ignore them.

New investment
solutions needed
Liability benchmarking presents a whole new problem to Dutch pension plans. Attempts to create a portfolio with the same profile as the liabilities create practical problems for pension plans because of the current composition of bond markets. Bonds with long-term maturities are in limited supply. In fact, only about E335bn worth of fixed-income securities in the EMU government bond market have maturities longer than 20 years. Moreover, Dutch funds planning to keep providing inflation indexation after 2006 will face an even more limited choice of bonds in the inflation-linked bond market.
Currently the euro inflation-
linked bond market consists of only 10 sovereign issues of France, Italy and Greece, with a total market value of approximately E95bn. Although the market is expected to grow significantly in size and more sovereign issuers are expected to come to the market, it does not provide the depth and breadth Dutch pension funds would need when trying to manage risk relative to their liabilities.
The solution may lie in the derivatives market, which has relatively good liquidity both in nominal and real terms, and provides the ability to fine tune cash flows to specific maturities. Moreover, it allows pension funds to add alpha overlay strategies in order to enhance the return of the fund and thereby reduce fund costs. To go this route, however, Dutch pension funds and specifically their boards of trustees will need to revisit their investment guidelines and adopt much more flexible rules.

The winds of change
While some may bristle at the challenges that lie ahead, the winds of change are already blowing. The Dutch and UK regulators have already taken steps to ensure that pension funds look at both their assets and their liabilities, and manage both accordingly. Globally, pension funds are beginning to see how dangerous static benchmarks can be, as more and more companies find themselves facing plans in deficit. And worldwide, industry-leading funds, consultants, and managers are embracing the concept of liability benchmarking.
Allowing asset allocations to change as wealth, expected returns and liability profiles move seems to be the most sensible approach. A revelation such as this could have saved our industry much trouble had it been made a decade ago. This change will enfold slowly but inevitably across the global pension fund community.
Liability benchmarking is the answer to help pension funds gain a greater understanding and control
of funds’ surpluses. Static benchmarking will leave us all twisting at windmills.
Susanne van Dootingh is senior fixed income product engineer of State Street Global Advisors
in Brussels

Topics