Designing and Implementing a TE-Constrained LDI Strategy | Capital Group

  • ACCOUNTS
  • INVESTMENTS
  • CLIENT SOLUTIONS
  • INSIGHTS
  • EVENTS
  • ABOUT US

Investment Insights

September 2014

Designing and Implementing a TE-Constrained LDI Strategy

Wesley Phoa Portfolio Manager Los Angeles office 23 years of experience (as of 12/31/16)
E. Luke Farrell Investment Specialist Los Angeles office 25 years of experience (as of 12/31/16)

A well-designed LDI program with tracking error constraints can help plan sponsors strike a balance between expected excess return and risk controls.

Liability-driven investing has changed over the past decade. Defined benefit plan sponsors have gradually moved from broad mandates with considerable latitude to more tightly designed and benchmark-aware mandates. In recent years, some plans have taken this a step further and stipulated explicit tracking error constraints for LDI mandates. In our view, setting tracking error targets makes sense for pension plans that have substantially reduced plan risk, or the mismatch of assets and liabilities, also referred to as funded status volatility. In addition to tracking error targets, plan sponsors will benefit from also paying close attention to investment guidelines such as credit quality.

Plan sponsors will also benefit from evaluating LDI managers in the pension plan as a group, examining how their investment approaches complement each other. Moreover, plans should consider setting tracking error targets for the overall LDI program, in addition to targets for each manager. These steps should help plan sponsors strike a balance between expected manager alpha and adequate risk controls.

Pension De-Risking and LDI Index Awareness

Over the past decade, many defined benefit plans have steadily reduced allocation not just to equities, but also to core bonds, instead shifting assets to long duration strategies. Plan sponsors have developed a greater understanding that a core bond allocation — rather than playing a stabilizing role — actually can increase risk due to its duration mismatch with the plan liabilities. Furthermore, in contrast to equities or alternatives, core bonds offer no incremental expected return to compensate for their contribu­tion to plan risk, defined as the funded status volatility, or the mismatch between plan assets and liabilities.

This shift to long duration has been occur­ring alongside a more general trend of plan de-risking, i.e., an overall allocation out of risky asset classes as the plan’s funded status improves. The Pension Plan Volatility chart illustrates the historical risk profiles of different asset allocations and shows the ben­efits of a shift into long credit — the closest match to the liability discount curve— thus reducing funded status volatility.

Pension Plan Volatility*

Asset allocation can have a substantial impact on funded status volatility.

*From 2002 to 2012 based on asset class monthly index returns.

Core fixed income: Barclays U.S. Aggregate Index Equities: 70% S&P 500 and 30% MSCI EAFE Index. Long government credit: Barclays U.S. Long Government/Credit Index. Long credit: Barclays U.S. Long Credit Index.

Sources: Barclays, Capital Group.

One can see that the annualized funded status volatility reduced from 161 basis points for an allocation of 80% to equities and 20% to core bonds to 46 basis points for a plan with 90% of assets in long duration credit and 10% in equities. For the asset mix of 80% equities and 20% bonds, the worst year since 2000 was a 25.7% under-performance versus the liability with a 5/3 skew to the downside.

Both downside risk and downside skew was reduced as the asset allocation is shifted towards long bonds away from equities or a core bond allocation.

When Do Tracking Error Constraints Make Sense in the Glide Path?

In the early stages of a plan’s de-risking glide path, the greatest risk reduction comes simply from allocating out of equi­ties and other risky asset classes into long duration bonds in general. The benefits of finely calibrating the bond mandate are less important. Indeed, relatively uncon­strained fixed-income mandates can be optimal in the early stages of the glide path since some of the equity exposure can be replaced with asset classes such as emerging markets debt and high yield that have expected returns similar to equi­ties but with the added benefit of higher correlations with liability discount rates. (A glide path is an asset allocation method used to gradually de-risk a pen­sion plan over time.)

At higher funded status levels in the glide path, as overall duration exten­sion is complete or almost complete, the picture changes. At this stage, the bulk of plan risk comes not from the duration mismatch, but from the remaining alloca­tion to a few non-LDI assets, especially equities. There is also some funded status volatility from the mismatch of the corpo­rate credit in the LDI portfolio and the dis­count curve of the liability, as well as other exposures within the LDI portfolio.

At these higher allocations to long dura­tion fixed income, the most efficient way to reduce plan risk further is to impose constraints on the LDI portfolio, to make it a closer match to the liability discount curve. In practice, this means reducing investment flexibility, switch­ing from an unconstrained government/credit approach to credit-only long duration fixed income, and – for some plans – eventually imposing benchmark or liability-based tracking error targets. This is illustrated schematically in the Glide Path. (Tracking error is the divergence between the price behavior of a position or a portfolio and the price behavior of a benchmark. It is expressed as a standard deviation percentage difference.)

Glide Path

At higher funded ratios, fixed-income risk will drive funded status volatility. This is the stage in the glide path where tracking error constrained LDI strategies can be useful.

 

Source: Capital Group

There is no one-size-fits-all approach. The right solution depends on the circum­stances of the plan. Some plans will want to introduce tracking error constraints earlier in the glide path, while others may not want to introduce them at all. For example:

  • If the plan is retaining more risky exposures in its equities and alterna­tives allocation – for example, an unusu­ally high weight to private equity or emerging markets equity – then it may introduce constraints on the long dura­tion fixed- income allocation earlier in the process.
  • If the plan sponsor is subject to busi­ness risks that are highly correlated with the equity market – for example, if it operates in a highly cyclical industry – then its LDI implementation may need to be more index or liability aware from the start in order to reduce correlation between plan funded status and busi­ness conditions.
  • Conversely, relatively well-funded plans sponsored by firms operating in non-cyclical industries may find it suf­ficient to rotate into LDI while leaving investment guidelines relatively broad; this still delivers substantial reduction in funded status volatility.
  • Similarly, open plans will often manage risk within the LDI allocation less tightly than closed plans, since overall plan funded status risk is harder to manage anyway, making tracking error within the LDI allocation a second­ary consideration.

Designing Investment Guidelines

Once a plan sponsor has decided on a desired level of risk and a broad con­ceptual strategy for LDI implementation, the next step is to design investment guidelines that are consistent with these objectives while retaining optimal flexibil­ity to deliver higher returns. Appropriate flexibility is especially important for active strategies, since it helps ensure that they can deliver enough alpha to off­set their increased costs relative to passive vehicles.

The Returns and Correlations table gives an example of the con­siderations that come into play. While the liability discount curve is based on corporate credits rated A and above, a large part of the long duration invest­ment-grade corporate universe is rated BBB. The correlation between BBB bonds and bonds rated A and above is very high, implying that some flexibility to invest in BBB can be permitted with only a modest increase in potential tracking error, while achieving substantial diversification. The BBB-rated part of the market is composed of a wide group of credits. Electric utili­ties, cable and railroads companies, for example, have relatively stable credit profiles. However, this positive factor is offset to some extent by the fact that idio­syncratic risks for BBB issuers can be high. Therefore, in tracking error constrained mandates, plan sponsors need to be con­fident that their investment managers are actively managing this source of potential tracking error.

 

Returns and Correlations (%)

Investment guidelines on credit quality can be important. For instance, allowing investments in BBB corporates can provide excess return without commensurate volatility because of the low correlation to other credit quality buckets.

Total Returns

Long corporate A & above

8.50

5.59

Long non-corporate credit

7.90

6.74

Long corporate BBB

10.60

6.98

Excess Returns Vs. Duration-Neutral Treasuries

Long corporate A & above

0.88

−0.58

Long non-corporate credit

0.16

0.47

Long corporate BBB

2.76

0.64

Total Return Correlation

Long corporate A & above vs. long non-corporate credit

0.86

0.87

Long corporate A & above vs. long corporate BBB

0.92

0.93

Long non-corporate credit vs. long corporate BBB

0.80

0.82

Excess Return Correlation

Long corporate A & above vs. long non-corporate credit

0.80

0.79

Long corporate A & above vs. long corporate BBB

0.95

0.94

Long non-corporate credit vs. long corporate BBB

0.86

0.84

Source: Barclays.
Data as of June 30, 2013. Returns are in USD. Periods greater than one year are annualized. Each index is unmanaged.

Also important are the practical consider­ations for implementation. As an example, while AA- and A-rated corporate issuers have a high correlation with the discount rate for pension plan liabilities, these higher rated securities often have lower market float, poorer secondary market liquidity and a lower frequency of issu­ance when compared to lower-rated BBB credits or some non-corporate credits. By having plan guidelines allow for the inclu­sion of BBB credits and select non-cor­porate credits, plan sponsors will obtain greater diversification, liquidity and port­folio customization with only modestly lower correlation to the plan discount rate.

Additionally, the BBB-rated credits in particular can offer a significant total return benefit above the A-rated universe (6.98% return versus 5.59% return over the last 10 years). BBB-rated issuers have provided better excess returns above U.S. Treasuries, at 64 basis points for the last 10 years, compared to a negative excess return of 58 basis points for A-rated issu­ers. The high representation of banks in the A-rated universe resulted in its lack­luster performance relative to BBB issuers in the post-crisis period.

Manager Behavior and Manager Selection

Manager Median Tracking Error (3 year)

Tracking error remains high for many LDI investment managers, reaffirming

our view that active managers can add value in these mandates.

 

Sources: eVestment, Barclays, Capital Group

This chart illustrates how the nature of LDI investing has changed over time. It shows the median tracking error of the top 20 long government/credit managers – and, where available, of the top 20 long credit managers over the past 15 years.

The pre-crisis period 1998–2007 shows a steady decline in median tracking error as plan sponsors and their investment managers became more index aware and more focused on controlling risk beyond simply adding duration to the plan. The extreme market dislocation in 2008–2009 drove tracking error much higher again, but this was the result of market condi­tions rather than a deliberate decision; as markets emerged from crisis, track­ing error trended down again. It is worth noting that tracking error for long credit managers has always been lower than for long government/credit managers, partly because credit-only managers have tended to operate under more restrictive investment guidelines.

Long Duration Manager Dispersion

There is sufficient dispersion in manager excess returns for plan sponsors to build well-rounded LDI programs with select investment managers.

 

Sources: eVestment, Capital Group

This chart shows a measure of dispersion among active long duration managers. It is interesting to note that manager dispersion declined less than manager tracking error during the pre-crisis period; even as managers became more risk-aware, they continued to follow diverse investment approaches.

Given this diversity, if a plan sponsor has a tracking error target for its long duration assets, imposing this target on each indi­vidual investment manager is inefficient, since aggregated tracking error will then tend to be lower than the target, and thus alpha will not be as high as it could have been. To the extent that correlation of excess returns among managers is low or negative, tracking error budgets for individual managers can be raised, giving each of them more opportunity to add value.

Correlation of Excess Returns vs. Barclays U.S. Long Government/Credit Index

Correlations among manager excess returns are significantly lower than one. By selecting the right combination of managers, plan sponsors can improve the risk-return trade-off.

 

Over a five-year period ending June 30, 2013.
Source: Zephyr StyleADVISOR. Management returns supplied by eVestment Alliance for the 10 largest managers of long duration strategies by assets under management.

This correlation matrix shows that pairwise correlations among manager excess returns are, indeed, significantly lower than one. But it also shows that some managers are clustered while others offer better diversification from the pack. Thus, in selecting managers, plan spon­sors should not just consider investment results in isolation; by picking a stable of managers with a history of low correlation, they can improve the risk/return tradeoff in the LDI assets in aggregate.

Conclusion

Tracking error constrained LDI mandates make sense for many pension plans, especially those that are further along the glide path and have successfully reduced funded status volatility. Plan sponsors can design LDI mandates with tracking error constraints and a set of investment guidelines that allow for adequate flexibil­ity for active managers to generate alpha without giving up sufficient risk control. If implemented well, such plans can result in a good balance between control of funded status volatility and desired excess return from active managers.


Investments are not FDIC-insured, nor are they deposits of or guaranteed by a bank or any other entity, so they may lose value.

Investors should carefully consider investment objectives, risks, charges and expenses. This and other important information is contained in the fund prospectuses and summary prospectuses or the collective investment trust's Characteristics statement, which can be obtained from a financial professional, Capital or your relationship manager, and should be read carefully before investing. 

Statements attributed to an individual represent the opinions of that individual as of the date published and do not necessarily reflect the opinions of Capital Group or its affiliates. This information is intended to highlight issues and not to be comprehensive or to provide advice.