Vanguard Defined Benefit Perspectives LDI and two real-life plan sponsors: A study in contrasts The dilemma: To LDI or not to LDI? Two Vanguard defined benefit plan clients answered this question differently. The concerns of the client who answered no were the fear that rates would rise and harm bond portfolio returns, and that reducing equities would reduce the portfolio s overall expected return. For the client who chose LDI, these concerns were outweighed by a desire to increase and maintain funded status while reducing risk and pension plan volatility. The results Over a two-year period (January 1, 2013, to December 31, 2014) that included the 2013 bond market crash, the client who chose the LDI route saw its funded status increase from 73.3% to 90.4% (it also made plan contributions in 2014). The plan of the client reluctant to adopt LDI fell from 100.4% to 89.9%. Had this client taken the LDI route, the funding level drop would only have been to 98.5%. The implications If reducing funded status volatility is a plan objective, LDI should certainly be explored. Moreover, this strategy has preserved funding levels that have taken years to achieve. It has worked as expected in strong and weak equity markets and rising and declining interest rate environments. For those who haven t adopted LDI, there s no time better than now to rethink that decision.
Author Philip C. Daubney, CFA Vanguard Institutional Advisory Services October 2015 Liability-driven investing (LDI) has been part of the defined benefit (DB) plan investment landscape for many years. In its simplest terms, LDI strategies seek to manage the volatility of the plan s funded status by investing with a focus on the liability. The many DB clients with whom we partner reflect that LDI may be appealing to some plan sponsors but not to others. Many have adopted derisking strategies; others have decided that it just wasn t in line with their plan objectives. There are those clients who expressed an interest in LDI and found the concept reasonable, but in the end decided the time wasn t right to implement a long-duration bond strategy or to cut back on the equity allocation. LDI theory often sounds conceptually appealing but is still met with resistance when it comes to actual implementation. The primary reasons for plan sponsor reluctance to implement LDI include both the fear that rates will rise, harming bond portfolio returns, and the concern that cutting back on equities will reduce the overall expected return on assets. Both are legitimate fiduciary concerns, but they focus only on the potential performance of assets. Timing the implementation of an LDI strategy may cause worry. Some sponsors may be tempted to time markets predicting the future direction and magnitude of changes in interest rates a difficult task at best. But the motivation is based on the understandable concern that if rates rise, bond prices will fall, asset levels will decline, and the plan sponsor may have fallen short in its fiduciary duty. On the equity side, there is the reasonable concern that derisking the plan by reducing equity exposure will diminish future expected return. Over the long run, stocks have tended to outperform bonds, and we have no reason to doubt that this relationship will continue in the future again, over the long run. And that has both economic and financial reporting implications, because a lower expected return on assets (EROA) increases pension expenses. Here, we briefly compare the separate experiences of two DB clients one who chose to adopt LDI and another who considered it, but decided not to adopt it. We compare these experiences in both falling and rising rate environments, and in both weak and strong equity markets. We look at the effects of the actual market environment not just on assets, but on the funded status of the plans. Client 1: LDI adopter, and stability throughout rising and falling interest rates One of our early adopters of LDI, a corporate pension plan, has had a long partnership with Vanguard. After a decades-long defined contribution plan relationship, we began managing their DB plan in 2006 and implemented an LDI glide path strategy shortly thereafter. Each year we assess how assets have changed relative to liabilities and discuss the effectiveness of the LDI strategy with the client. This happens to be a frozen plan in which benefits are no longer accruing. At first one might think, with a frozen plan, that liabilities would be relatively stable year-to-year, with most increases stemming from the shorter discount period, which decreases with the passing of each year. This would be true if discount rates never changed. In reality, of course, discount rates can change dramatically from year to year, even from month to month. Ideally, if stability of funded status is the goal, we would like to see changes in values from both interest and from gains or losses because of interest rate changes to be about the same for both the plan s liabilities and its assets. However, with plans that are less than fully funded, a sponsor will often retain substantial equity exposure in hopes of equity growth contributing to funding. Concurrently, the sponsor will often use long bonds to hedge as much interest rate risk to liabilities as possible. As higher levels of funding are achieved, equity risk is further reduced, leaving more room for bonds and for hedging interest rate risk to liabilities. 2
Derisking the portfolio, matching duration, hedging interest rate risk Before the beginning of 2014, this plan had already achieved an 85% funding level, had derisked from 50% equity to 45% equity, and had utilized a long-bond portfolio with a composite duration of about 15 years. By the end of 2014, with assets close to 90% of liabilities and with 55% of the portfolio already in bonds, the composite total portfolio duration was about 7 years (15 years x 55% in bonds x 90% funded). With a liability duration of about 10 years, the theoretical hedge ratio at current funding was then around 70%. Figure 1 illustrates how our client s plan liabilities changed relative to its plan assets during 2014. (The numbers have been proportionally modified to preserve client confidentiality.) Figure 1. Client 1 LDI experience Liability Assets Surplus/Deficit Funded % January 1, 2014 $243,182,466 $212,522,249 $30,660,217 87.4% Normal costs/contributions $0 $3,433,073 Benefit payments $11,558,627 $11,558,627 Fixed income returns Interest $9,526,167 $7,017,249 Gain/loss due to interest rate changes $18,417,420 $15,335,265 Equity returns Equity dividends N/A $2,402,107 Equity gain/loss N/A $5,483,807 Total return $27,943,587 $30,238,428 December 31, 2014 $259,567,426 $234,635,123 $24,932,303 90.4% Change from January 1, 2014, to December 31, 2014 6.7% 10.4% 3.4% Economic estimates, not actuarial calculations. Source: Vanguard. Our client also made plan contributions during 2014, which increased plan assets and helped improve funded status. Regular disbursements to beneficiaries reduced both assets and liabilities by the same amount, of course, so funded status wasn t affected. In addition, it was a good year for equities especially domestic equities whose dividends and gains also increased plan assets and helped improve funded status. Interestingly, however, equities accounted for only 26% of the total portfolio s return that year. The crucial role of long bonds in preserving funding gains The contribution of long bonds in derisking strategies cannot be minimized: In a year (2014) when liabilities increased substantially because of the drop in discount rates, the funded status of our client s plan was preserved because of its allocation to long bonds. Our LDI strategy included achieving a yield on the bond portfolio that approximately matched the effective discount rate for the liability. We nearly matched the interest costs on liabilities with the interest earned on the bond assets. With only 55% of the portfolio allocated to bonds, the interest earned on the plan s assets covered about 74% of the interest cost of the liability. In 2014, the biggest contributor to the increase in liabilities for DB plans in general was falling discount rates, and our client s plan was no exception. Its liabilities increased when its discount rate dropped by 1.11%. However, its assets thanks to the plan s long-duration bond portfolio appreciated almost as much as the liabilities, hedging a healthy 83% of the liability increase. 3
Okay, but how did they do during the bond market crash of 2013? Early on, LDI tended to be a difficult concept for plan sponsors to embrace, especially if they had a bias to predicting rising interest rates when they seemed low. However, by 2006, more of our advisory clients began implementing LDI strategies. Since then we ve seen how these strategies perform in equity bear and bull markets, in rising and declining interest rate environments, and in years like 2013, when it was tough to be investing in bonds especially long-term Treasury bonds. As you would expect, with long-term corporate bonds (to provide credit match) and extended-duration Treasuries (to provide more duration matching), our client s bond portfolio took a real hit. At the same time, with liabilities plummeting and equity assets increasing, they hit a funded status trigger halfway through the year. They derisked equities by 5%, from 55% to 50%, increasing the bond allocation and the overall portfolio duration. Figure 2. Client 1 LDI client experience in 2013 Liability Assets Surplus/Deficit Funded % January 1, 2013 $272,319,256 $199,710,933 $72,608,323 73.3% Normal costs/contributions $0 $3,238,137 Benefit payments $11,614,150 $11,614,150 Fixed income returns Interest $11,152,866 $5,172,215 Gain/loss due to interest rate changes $28,675,506 $11,675,463 Equity returns Equity dividends N/A $2,409,865 Equity gain/loss N/A $25,280,711 Total return $17,522,640 $21,187,329 December 31, 2013 $243,182,466 $212,522,249 $30,660,217 87.4% Change from January 1, 2013, to December 31, 2013 10.7% 6.4% 19.2% Economic estimates, not actuarial calculations. Source: Vanguard. As Figure 2 illustrates, 2013 was also one of the best years on record for equities, whose gains more than made up for the bond losses. What s interesting, though, is that the loss in the liability because of rate increases was more than double the loss in the bond portfolio. Sponsors who may be considering LDI often forget that until their plan is close to 100% hedged with an all-bond portfolio, the loss in the liability will be far greater than the loss in bond assets from rising rates. Even at this early stage of a hedging strategy (with only half of the portfolio in bonds), it s important to consider the complexities of bond portfolio behavior and strike the right balance between duration matching and credit matching. While neither can be matched perfectly, a combination of long-term corporate bonds, augmented with some Treasury exposure, will often result in better liability matching. The client s goal achieved: Improving and maintaining funded status For this client, adopting LDI with a glide path that gradually decreased equity risk and increased interest rate hedging as funded status increased, ultimately provided the stability of funded status that it sought. 4
Client 2: And what might have been, when LDI was not chosen Now let s contrast this experience with that of another longtime client whose investment approach has always been focused on total return. Given that the size of its pension isn t that large relative to the size and revenues of the organization, the client rightly concluded that risk to its plan s funding level wasn t a significant risk factor to the financial health of the overall organization. Any required contributions could easily be made from cash flows. Still, an LDI strategy generally would have benefitted this client over the last decade. Again using 2014 as the most recent example, Figure 3 illustrates how the plan s assets performed relative to its liabilities (as before, the actual numbers have been proportionally modified to protect client confidentiality). This is an open plan, but the sponsor prefers to examine the liability in terms of the present value of benefits (PVB). Because normal costs are included in the PVB cash flows, normal cost accruals don t appear in the table. In addition, no contributions were made during this plan year. Figure 3. Client 2 who chose not to adopt LDI Liability Assets Surplus/Deficit Funded % January 1, 2014 $89,916,362 $90,313,857 $397,495 100.4% Benefit payments $2,134,465 $2,134,465 Fixed income returns Interest $3,882,707 $966,770 Gain/loss due to interest rate changes $14,974,297 $567,701 Equity returns Equity dividends N/A $1,195,588 Equity gain/loss N/A $4,931,997 Total return $18,857,004 $7,662,005 December 31, 2014 $106,638,901 $95,841,447 $10,797,454 89.9% Change from January 1, 2014, to December 31, 2014 18.6% 6.1% 10.5% Economic estimates, not actuarial calculations. Source: Vanguard. The most significant contributor to an increase in assets during 2014 a good year for stocks came from the portfolio s 65% allocation to equities. The intermediate-duration bond portfolio produced income and experienced gains as well, but not nearly as much as a long-duration LDI portfolio would have. However, the gain in liabilities was very substantial, and funded status decreased from about 100% to about 90% over the course of the year. This client was curious to see how the LDI portfolio we had recommended in 2013 might have performed. Figure 4 illustrates this hypothetical comparison, which shows that the recommended portfolio would have resulted in far less of a loss in funded status. 5
Figure 4. What might have been had client 2 chosen LDI Liability Assets Surplus/Deficit Funded % January 1, 2014 $89,916,362 $90,313,857 $397,495 100.4% Benefit payments $2,134,465 $2,134,465 Fixed income returns Interest $3,882,707 $1,557,011 Gain/loss due to interest rate changes $14,974,297 $9,168,211 Equity returns Equity dividends N/A $1,195,588 Equity gain/loss N/A $4,931,997 Total return $18,857,004 $16,852,806 December 31, 2014 $106,638,901 $105,032,198 $1,606,703 98.5% Change from January 1, 2014, to December 31, 2014 18.6% 16.3% 1.9% Economic estimates, not actuarial calculations. Source: Vanguard. While this client preferred not to reduce equities in hopes of maximizing returns over the long run, converting from intermediate-term bonds to long bonds would have made for a significantly more efficient fixed income allocation for this defined benefit plan as it would for most other plans. The equity return would have been the same, but the return on bonds would have been nearly seven times greater, preserving nearly all of the funded status that had been achieved. 6
What we ve learned About the fear that rising rates will always be with us There s no doubt that we ve been in a secular bull market for bonds for close to three decades. Most of plan sponsors concerns about LDI have been on the bond side, fearing that yields will go up and prices will come down, resulting in a bond bear market. It s important to remember that price change has little or no effect on bond returns over the long run and even in a rising rate environment where the interim price of the bond may have decreased, that decrease will have disappeared at maturity. The point is that while many investors expect interest rates to rise over time, the effect on returns may not be as dramatic as simple duration calculations would suggest. As importantly, current interest rates tell us nothing about their future direction. Unfortunately for many plan sponsors, the refrain has been that rates are so low, they can only go higher. Rates seem low, and the market expects them to rise especially at the short end of the yield curve yet still our outlook for bonds is positive. Indeed, it s very rare for a bond portfolio to lose money over a duration-matched holding period, or even on an annual basis. Investors should also remember that in the last 15 years, the aggregate U.S. bond market has had only one negative total-return year ( 1.97% in 2013). Going all the way back to 1980, just two other years showed negative returns from bonds ( 2.92% in 1994, and 0.82% in 1999).* About the fear that reducing equity allocations will result in lower returns Given that stocks have tended to outperform bonds, we ve no reason to doubt that will continue in the future over the long term. But the fact is that derisking strategies involve examining the trade-offs between risk and return, and acknowledging that lower risk almost always means accepting lower return. LDI: The benefit of experience and hindsight LDI is a liability-focused risk-mitigation strategy that might not appeal to every DB plan sponsor. A plan s risk posture can be influenced by the plan s funded status and demographics, whether it s open or frozen, its corporate cash flow, its ability to make contributions, and the size of its plan relative to the overall organization. If reducing funded status volatility is one of the plan s objectives, LDI should certainly be in the conversation. While very large plans sponsored by very large organizations have had LDI strategies for quite some time, smaller- and medium-size plan sponsors have been more reluctant to adopt them. These plan sponsors should consider that an LDI strategy can begin gradually: For example, even before derisking by decreasing the equity allocation, the plan can achieve more duration hedging simply by lengthening the duration of the bond portfolio a far more efficient strategy than using marketduration bonds in most cases. From there, a derisking glide path can be developed that adds duration only when predetermined funding-level triggers are met. We ve spoken with many plan sponsors whose initial thought was that LDI in concept is appealing, but who are hesitant to commit to it without some prior real-life experience. We hope that this small sampling of real-life LDI experiences provides some additional perspective. Our clients who have embraced LDI have seen it work as expected in both strong and weak equity markets, and in both rising and declining interest rate environments. Importantly, LDI has decreased the volatility of their plans funded status and preserved funding levels that have taken years to achieve. Our role is to help provide that broad perspective and experience, to help sponsors examine the trade-offs of different portfolio and strategy choices, and to help determine if LDI is appropriate for them to see if a Not now should be a Let s get started. Derisking, however, is commonly implemented as a phased-in strategy that s linked to improved funded status. And as funded status increases, pension expense is generally decreasing. This offsetting effect can make a derisking strategy more appealing when there are concerns about higher pension expenses resulting from lower return on asset expectations. *Source: Barclays U.S. Aggregate Bond Index. 7
For more information If you d like to find out more about LDI in theory and in practice, we suggest the following Vanguard research papers: Christopher B. Philips, Francis M. Kinniry, Brian Scott, Michael A. DiJoseph, David J. Walker, Risk of loss: Should the prospect of rising rates push investors from high-quality bonds, Vanguard, July 2013. Kimberly A. Stockton, Fundamentals of liability-driven investing, Vanguard, June 2014. Paul M. Bosse and Nathan Zahm, Derisking and pension expense: Not all bad news, Vanguard, June 2014. Kimberly A. Stockton and Nathan Zahm, A corporate finance approach to managing defined benefit plans, Vanguard, May 2015. Jeffrey Sparling, Pension derisking: Start with the end in mind, Vanguard, June 2014. Kimberly A. Stockton and Anatoly Shtekhman, Glide path ALM: A dynamic approach to derisking, Vanguard, May 2012. Paul M. Bosse, For better pension liability matching, consider adding Treasuries, Vanguard, December 2012. Kimberly A. Stockton, Pension risk: How much are you really taking?, Vanguard, June 2013. Investing essentials, ten short (about three minutes) videos about derisking available at institutional.vanguard.com. Connect with Vanguard > institutional.vanguard.com Important information All investing is subject to risk, including the possible loss of the money you invest. Bond funds are subject to the risk that an issuer will fail to make payments on time, and that bond prices will decline because of rising interest rates or negative perceptions of an issuer's ability to make payments. U.S. government backing of Treasury or agency securities applies only to the underlying securities and does not prevent share-price fluctuations. Unlike stocks and bonds, U.S. Treasury bills are guaranteed as to the timely payment of principal and interest. Vanguard Institutional Investor Group P.O. Box 2900 Valley Forge, PA 19482-2900 2015 The Vanguard Group, Inc. All rights reserved. LDIRLP 122015