“Oh Lord, help me to be pure. But not yet!” – St. Augustine
Defining “purity” for a pension strategy is a tricky thing. Matching assets and liabilities, reducing risk – these might be seen as the equivalent of a pure and balanced approach. But for many pensions today – underfunded relative to liabilities, or with open and active plans – such an approach would condemn the sponsor to higher contributions for years to come. On the other hand, return-seeking, risk-taking strategies – these offer the prospect of building up capital without costly contributions. (Though we should not overlook the nasty habit they have of backfiring.)
Purity, in the form of full hedging of liabilities, may very well have to wait for those plans that need to outperform their liabilities today. But we can keep an eye on risk while we go about the difficult process of seeking returns. Let’s propose that a more realistic goal would be to achieve the highest risk-adjusted return on assets relative to liabilities. What follows is a simple framework to help pension investors achieve this goal.
Start from the liability
For a pension, the liability is the ultimate benchmark, both with respect to risk and return. Investments will have returns that may be higher or lower than the liability, and risk (volatility) that may be higher or lower than the liability. Figure 1 maps a variety of asset classes onto an asset-liability efficient frontier. The Y-axis shows returns; the X-axis, risk versus liabilities.
There are a few key takeaways. The first is that the least-risky investment will typically be a long-term bond portfolio with risk characteristics identical to the liability. A customized liability-driven investment (LDI) strategy can provide this solution. Intermediate bonds and cash generally exhibit both higher risk and lower returns relative to liabilities and are therefore unattractive from a strategic standpoint. Risk assets, such as absolute return (hedge funds) and equities, generally offer higher returns but potentially much higher volatility. Their use may be necessary insofar as a plan seeks to outperform. Very well-funded plans may have little or no need for outperformance, instead choosing to “lock-in” funding by shifting most or all assets to liability-matched bonds.
Seeking efficient pension portfolios
In practice, of course, pension investors don’t construct asset allocations in a vacuum, but rather relative to specific objectives such as targeted rates of return or levels of funded-status volatility. Given the multitude of investment options available, however, it is usually the case that there are many allocations that could be designed to achieve a single goal. It is therefore useful to identify a portfolio strategy that meets the objectives most efficiently.
Pension efficiency can be measured by adapting a basic investment concept: the Sharpe ratio. Traditionally, the Sharpe ratio is used to evaluate the risk-adjusted returns of individual investment strategies. Excess returns over a risk-free rate are divided by the volatility; the higher the ratio, the better.
Pension plans can apply the underlying logic of the Sharpe ratio to evaluate strategy at the total plan level. The transformation – into what is also known as the liability-adjusted Sharpe ratio – is straightforward:
The total plan has an expected rate of return that is the weighted average of the expected return of all assets in the portfolio. The risk-free rate is the natural growth rate of the liability, represented by the liability discount rate. The volatility of the total plan is the funded-status volatility. Each of these measures can be observed (or estimated) without much difficulty.
In our experience, the most significant efficiency gains in pension strategy have come from shifting from intermediate bonds to long-term bonds (higher return, lower risk) and introducing lower-volatility substitutes to equities in the return-seeking portfolio (similar return, lower risk).
Consider a hypothetical example: A plan starts out with a traditional 60/40 allocation of stocks and bonds, then shifts its fixed income to long-duration liability-matched bonds, and then diversifies its equities into alternative assets (see Figure 3).
The increase in the pension Sharpe ratio shows that both strategies may be more efficient: The first switch from core bonds to LDI reduces estimated volatility with a higher return potential, while the second switch from stocks to alternatives achieves the same level of estimated returns with significantly lower risk potential.
How efficient is further de-risking?
The two portfolio shifts we discussed above could fairly be described as “low-hanging fruit” because they produce material improvements in the Sharpe ratio with only limited changes in the structure of the asset allocation (the split between return-seeking assets and bonds remains at 60/40). Indeed, many plan sponsors have already made moves in this direction.
Going forward, however, most pension plans envision additional de-risking by shifting assets from the return-seeking category into LDI strategies. Given the loss of return this may entail, it’s reasonable to ask if these changes would actually make the strategy less efficient. Fortunately, it appears that this would not be the case – at least until the final stages of de-risking. As Figure 4 shows, expected return does decline, but the potential drop in volatility would be significant. (And frankly, plans that are taking the final step to completely de-risk are likely to be more focused on the absolute level of risk than the incremental efficiency of the plan.)
Here endeth the lesson
St. Augustine might well sympathize with the very real dilemma of balancing a short-term need for return with a long-term desire to de-risk. We hope that most pension plans will eventually reach a level of funding where the dilemma fades to insignificance, and where the transition to de-risking becomes relatively painless. Until then, however, the best option may be to choose a strategy that keeps close account of both risk and return.