By Curt Custard
Not too long ago, while in a client presentation for a medium sized defined benefit plan, the consultant asked how to value the bonds. I was shocked. Now, some five years later, I find the question less surprising, though sponsors and trustees are not approaching the question in the same way.
Why are bonds in a defined benefit plan’s portfolio? How should an investment in bonds be measured in terms of success or failure? The current discussions in the marketplace surrounding Liability Driven Investing attempt to tackle these fundamental issues and raises questions about the role of bonds in their portfolio. Gone are the discussions of tracking error relative to a market capitalization based benchmark. Now the focus is on the risk relative to liabilities. It seems that this is where the discussion should have started in the first place....
A plan’s projected benefit obligation is purely a series of cash flows subject to certain actuarial assumptions to be paid out at a future date. These cash flows can be matched with a moderate degree of precision by cash flows from fixed income assets, though depending on the assumptions used, the word moderate may be moderately stretched. Yet, if the cash flow profile of the bond benchmark you are using bears no similarity to the cash flow profile of the liabilities why is it there?
Some might argue that cash flows don’t matter but durations do – fine but the typical core bond index duration is years off a typical liability profile. Others say that the role of bonds is to bring down the overall volatility of a portfolio – however so does property or any other asset class with a low correlation to the equity markets.
There is also the argument – often from fund managers – that without a clearly defined broad market benchmark it is difficult to discern whether managers are adding value. However, the use of derivatives – typically interest rate swaps – means that matching durations can be done easily and without impacting the active management.
Still others argue that current nominal yields are too low – this is a market view and one that is greater in risk than any other risk their active managers are take. The common truth is that too much attention is focused on performance relative to a strategic benchmark and not enough on the relationship between the benchmark and the liabilities themselves.
Simply put, LDI is better bond investing. The bonds in a plan’s portfolio should match the risk characteristics of the liabilities, if not the cash flows. Even if a plan is under-funded relative to the PBO, the use of interest rate swaps can remove the interest rate risk in the portfolio, and, although there must be some cash kept back for collateral, there is still substantial scope for the generation of alpha.
The small reduction in the amount of alpha can be mitigated by managers adjusting their underlying tracking error and pales in comparison to the amount of interest risk reduction achieved. Better yet, the benefits of LDI are relatively low cost. At its simplest, a few interest rate swaps can be used to overlay the portfolio to remove the duration and yield curve risk. For those few plans that are fully funded and nearing their maturity, LDI can be used to substantially remove the risk from the sponsoring company’s balance sheet.
However, there are some challenges. Unlike a typical bond benchmark, there is no ‘one size fits all’ solution so the mandates are likely to be customized like they haven’t been in the recent past. Success or failure of the manager should be then measured in two parts, with the liabilities being the benchmark.
First, how bonds or swaps behave relative to the liabilities they are intended to match – this is the underlying tracking error, that unless intentional – should be minimized. Second, if the plan has opted to employ an active manager, the excess return should be measured –using a LIBOR basis is often the clearest expression of this success or failure and allows comparisons between managers.
Additionally, most managers will implement the use of derivatives to achieve an LDI solution and the trustees must be comfortable in their use. Interest rate swaps have been around for many years and have a deep and liquid market, and with their AA rating are safer than most of non-government holdings in traditional broad market benchmarks.
The use of derivatives may require the plan to sign an ISDA agreement or explicitly allow their fund managers to do so on their behalf. The manager or plan needs to be competent in collateral management, an administrative process that managers will increasingly need to become familiar with.
Finally, trustees need to start thinking in more concrete terms about how their assets behave relative to their liabilities. It is no longer sufficient to simply hope that the strategic benchmark will bail them out in the end. Sponsoring companies need to focus on how this unrewarded interest rate risk that they are taking will impact their balance sheet.
In the future, I foresee a world where the bond portion of plan’s portfolio is benchmarked against a plan specific liability benchmark, and outperformance is the relative return against that benchmark. Then if a client asks how to value the bonds, I will be able to give a different answer.
Contributor Curt Custard is head of Multi-Asset Solutions for Schroder Investment Management. He is based in London.