Falling interest rates, lagging equity markets and greater volatility in pension plan solvency ratios have created renewed awareness of the risks of pension plan assets and obligations, and the need to keep pension plans focused on providing for their beneficiaries. This has led to a more active debate in today’s investment community over how a pension plan can best manage its assets and its solvency ratio.

Liability driven investing (LDI) is not a specific investment product. Rather, it is a way of approaching pension plan management by combining the best elements of both growth-driven and immunization investment strategies. In addition to considering asset diversification, each of a pension plan’s investments is evaluated according to its potential for appreciation and its potential as a liability hedge, with the portfolio allocated to strike an appropriate balance between these two seemingly contradictory goals.

Advertisement

Due to differences in plan liability profiles, solvencies and fiduciary risk appetite, there is no one-size-fits-all LDI solution. Each pension plan must tailor its own solution to its particular circumstances. When evaluating LDI solutions, here are some key considerations for plan sponsors.

1 Surplus risk – How big should the hedge be and how much interest rate mismatch is appropriate?

There are four key drivers of surplus risk. One, the investment policy (the degree to which the plan takes asset allocation or beta risk with the expectation of growing plan assets in excess of liabilities). Two, the overall liability duration (the sensitivity of plan liabilities to changes in interest rates). Three, unhedgeable liability risks (including mortality, employee turnover, projected future wage inflation and other actuarial assumptions). And four, current funded status.

Figure 1 (see below) demonstrates that hedging greater percentages of the interest rate risk of liabilities leads to a reduction in surplus risk. However, the marginal surplus risk reduction beyond a 70% to 80% hedge is minimal for plans that are simultaneously investing in equities and other risky growth-oriented investments. Trustee risk tolerance and asset allocation conviction should drive the decision of how much mismatch should be retained.

2 Benchmark selection – What is the right benchmark?

An actuary’s liability cash flow projections are usually not investable and are typically only recalculated once per year. Therefore, they do not serve as a useful benchmark. In general, pension plan fiduciaries will want to adopt a benchmark that can be customized to the plan’s specific circumstances and that mimics, to the extent practicable, the interest rate risk of the plan’s liability cash flows. While there are a wide variety of benchmarks available, critical issues to consider include measurability, liquidity, transparency and investability.

Related Link

A Trustee's Guide to Alternative Investments

The answer need not necessarily be a customized benchmark. Proxy benchmarks such as the DEX Long Universe, which approximate a plan’s duration risk, are most useful when the intent is to hedge only a fraction of the liability exposure. These types of benchmarks may also be appropriate for plans with a shorter duration and/or greater uncertainty in the projected benefit stream. Proxy benchmarks have the advantages of being simple, publicly recognized, quoted daily and easy to explain. They have the disadvantages of embodying some duration, convexity and/or yield curve sensitivity mismatches to liabilities, which may not be present in a customized benchmark.

Customized benchmarks, such as a blend of government bond or interest rate swaps, improve on a proxy benchmark by attempting to more closely match the plan’s liability duration, convexity and yield curve exposures. Customized benchmarks may be appropriate when the objective is to hedge a larger fraction of liabilities and may be particularly suitable for a frozen or closed plan, as there is greater confidence in the projected liability stream. Possible disadvantages to a customized benchmark include complexity and the need to build and maintain such indexes.