LDI Portfolios: Keep the Focus on Robust Liability-Hedging Assets

Six drawbacks of STRIPS-heavy approaches

With these core principles in mind, we take issue with some recent proposals for LDI implementation – especially those concerning heavy usage of Treasury STRIPS. For example, one notion that has arisen recently is to sell or de-emphasize long credit positions and redeploy the proceeds into a combination of STRIPS and equities (or other return-seeking investments). On the surface, this may sound interesting, but digging deeper, we see a number of issues. So, as practitioners of LDI, we need to caution plan sponsors and their advisors that

1) While STRIPS are a capital-efficient way to get duration exposure in fixed income portfolios, long credit tends to be the better cornerstone of LDI programs. As such, there is a high bar to sell long credit positions. After all, corporate bonds are the natural liability-hedging instrument, as they are able to provide duration, credit spread exposure, and yield in line with liability discount rates (see Figure 2).

2) Long credit portfolios are currently yielding 5.0%–5.5% versus STRIPS offering less than 4%. In general, starting yields correlate strongly with future returns, so new dollars going into LDI should benefit more from a long credit allocation (which closely aligns with liability discount rates) than from long-dated STRIPS.

3) Pairing STRIPS and equity to create a substitute for long credit is a barbelled approach to managing against a liability. To be sure, one can model scenarios where, if past correlations hold, this could create a similar hedging outcome as a physical bond portfolio of corporate credit. However, as soon as correlations break down or even move away from those assumed, this approach is much riskier from an asset-liability management standpoint. Essentially these are diversification trades masked as liability hedges. Looking over time at the consistency of equity/spread correlations (or lack thereof), it’s easy to see that in certain environments this barbelled approach could underdeliver compared to owning credit bonds as a hedge (see Figure 3).

4) In addition to the fragility of certain correlation assumptions, volatility driven by yield curve mismatches can transpire in hedging strategies that are unstable and deviate from their intended targets. When assessing the risk of hedge-ratio volatility between a portfolio that was holistically designed to the liabilities across key risk factors versus a STRIPS-only portfolio, the results are impactful. STRIPS-only strategies tend to create much more volatile hedging outcomes. This will result in a more frequent need for rebalancing and trading – thus, additional transaction costs (see Figure 4).

5) Transaction costs are another consideration when making such an adjustment. As mentioned, corporate bonds are the preferred hedging assets. At the end of the day, to have the tightest hedge versus the liability, plans should consider a meaningful allocation to corporate bonds. If trading from corporates to a barbelled STRIPS/equity portfolio is temporary in nature (most likely this is the case as plans follow their glide paths), when it’s time to move back into long credit bonds, it’s hard not to think about transaction costs of a round-trip move – which could exceed 1% of the size of the allocation.

6) What’s more, blunt usage of STRIPS may result in significant yield curve mismatches within a plan’s LDI program. Such an overweight to the back end of the curve may be tolerable in a normal yield curve environment (i.e., the shape of the curve is upward sloping). But today’s yield curve is quite flat (and even inverted at certain points - see Figure 5 for today’s curve shape versus history). Thus, if yield curves steepen from here, realized hedge ratios will come in lower than what the headline hedge number may suggest because assets at the long end of the curve will underperform their liabilities – and the result would be a lower funded ratio.

A DEEPER DIVE INTO LIABILITY-HEADGING PORTFOLIOS

CREDIT VS. EQUITIES AND STRIPS

To test whether equities when paired with Treasuries/STRIPS can robustly provide a hedge to liability risk, we examined two portfolios during 17 periods of market stress going back 50 years (ranging from the 1973-74 oil crisis to the 2020 COVID-19 drawdown) and the associated tracking error of the portfolio versus the liability. Allocation 1: 100% LDI (credit + government to hedge 100% on both rates and spreads). Allocation 2: 40% equities / 60% LDI (Treasuries / STRIPS only). While both portfolios target the same liability hedge ratio (i.e., duration neutral), there are two key differences:

1. In the LDI Strategy, one holds corporate bonds and Treasuries that are aligned with the plan's liability risk characteristics (Allocation 1), while the other portfolio targets only the liability duration with Treasuries and STRIPS (Allocation 2) - resulting in no explicit liability credit spread duration hedge.
2. In allocation 2, the 40% increase in equity exposure aims to simulate a "proxy of credit spread hedge." To be clear, PIMCO is not a proponent of this as we believe that equities do not provide a stable hedge to liability credit spreads. Additionally, moving fixed income assets from actively managed lo0ng credit into more passively managed STRIPS mandates increases the need to take risk elsewhere (e.g., equities) to achieve the same level of return expectations - further increasing funding ratio volatility. 

The results confirm our initial assessment - Allocation 2 ( the portfolio with the Treasury/STRIPS-only LDI strategy) demonstrated significant volatility in returns versus the liability relative to the holistically built Allocation 1. In fact, the standard deviation of relative returns of Allocation 2 versus the liability during these select high-volatility periods was about 6% (versus Allocation 1 of less than 1%). Said another way, re-risking the allocation with equities to make up for a poor-quality liability hedge 9of Treasuries and STRIPS) is an exercise in futility. See Figure 6 for more details on the Select scenarios analyzed.

A more LDI-friendly way to re-risk

For plan sponsors that seek additional returns from equities or want to take a tactical exposure to equities without compromising their liability hedge, there are more capital-efficient ways to accomplish this than holding equities and STRIPS. Sponsors can enhance their asset allocations by pairing equities and long bonds in one portfolio – akin to a liability-friendly equity strategy. A long duration strategy that incorporates an overlay of equity exposure fits this objective well, in our view.

The preferred hedge – corporate bonds

Without doubt, when preparing for a variety of market environments, it should be clear that large STRIPS allocations may address duration risk within the liability, but they are not a silver bullet. In fact, they can introduce a subjective call on the shape of the yield curve. Instead, we believe plan sponsors should set their sights on a more robust liability hedge – especially given how close many plans are to being fully funded.

As the economy weakens, it will be important to consider emphasizing true LDI hedges driven by robust LDI portfolios that target all liability risk factors. Strategies that are less reliant on historical correlations have been shown to improve funding ratio stability across most environments.

In this context, plan sponsors may be better served by using a toolkit of diversified credit and government bonds targeting the desired liability hedge.^Footnote1 Strategies such as these seek to stay aligned with liability changes in a resilient fashion – after all, the goal of hedging strategies is to recognize and prepare for uncertainty in markets.

¹ If needed, an embedded overlay comprised of Treasury futures can help tighten the liability hedge. Return to content↩