aiCIO’s Year in Review: Daylight in the Garden of LDI

Estate owners wanting to dress up the landscape face crucial choices in both the strategic and tactical dimensions. Perennials, such as peonies, hosta, and chrysanthemums, require little tending and last for years, but bloom for only a few weeks. Annuals, such as marigolds, zinnias, and petunias, offer more and brighter colors for a longer season, but have to be pulled up in the fall and purchased again next year.

Late in 2008, after the bankruptcy of Lehman Brothers and Bear Stearns, and the rescues of AIG, Merrill Lynch, and other financial titans, US pension sponsors faced a similar choice between strategy and tactics. The liability-driven investing (LDI) programs which they had carefully chosen to hedge their pension liabilities against interest rate volatility—intended to be hardy long-term strategies that would require little more care than watering, feeding and an annual trimming—had in some cases suddenly shriveled, and in others overgrown monstrously. With a recessionary winter approaching, should they leave in place the LDI perennials, such as long-term interest rate swaps? Or should they uproot their strategies, planted just a few years ago, in favor of impulsive choices such as corporate bonds that were more colorful, of-the-moment, and selling at closeout prices?

The principle behind LDI is sensible enough: Invest the contributions to a pension plan in assets that will mirror the obligations to retirees. Assume the liabilities of your pension plan are such that, if interest rates drop by 2%, their present value will rise by 15%. Find assets with the same sensitivity to interest rates, and their value will respond in the same direction and amount, ensuring the plan can satisfy its obligations as they arise in the future. Rates can go up or down, and risk to the plan’s surplus is hedged by the assets.

Until recently, few U.S. corporate defined benefit sponsors saw the wisdom of the risk-controlling approach of LDI. Bonds, with inherently low returns, are an expensive way to invest for the long term. Instead, their strategy was to pile into assets with the highest returns—stocks and alternatives—as a way to minimize contributions.

About five years ago, however, LDI thinking was forced upon sponsors through new regulations on pension financial accounting and funding (respectively, SFAS 158 and the Pension Protection Act of 2006) that were a reaction to the “perfect storm” of 2000 through 2002.

The new rules require sponsors to address the effects of markets on the plan’s fair value. Both the PPA and SFAS 158 dictate that pension liabilities be valued with discount rates derived from the corporate yield curve. The regulation’s cash contribution requirements and earnings effects are, in turn, based on those valuations: SFAS 158 highlights pension shortfalls in sponsors’ financial statements, and the PPA requires sponsors to make up deficits over short periods of time.

These regulatory spurs to action, in addition to the sound economics underlying LDI, have focused sponsors’ attention on the volatility of their funds’ surplus. Among U.S. corporate pension managers, 91% consider managing the mismatch between pension assets and liabilities a priority issue for 2010, according to a recent survey by asset managers SEI Corporation.

The traditional U.S. pension portfolio confounds an implementation to 100% of a plan’s liabilities. Seeking high returns, many plans hold high proportions of assets, 60% and more, in equities. Because stocks don’t have the dependable relationship with interest rates that bonds exhibit, their market value can’t be counted on to match up with fluctuations in a plan’s liabilities. Holding lots of equities creates risk to the surplus.

Obligations to retired workers resemble bonds: a series of fixed payments stretching into the future until the retirees’ demise (in U.S. plans, benefits primarily are fixed and not indexed to inflation). Accordingly, pension liabilities can be closely matched with a well-chosen portfolio of bonds. Most important is that the duration of the bonds—the weighted average of their cash flows, which determines rate sensitivity—matches that of the liabilities.

The new regulations don’t actually mandate any particular sort of investment, but the risk control inherent in bonds makes them a logical choice and, within the bond universe, as the rules call for determining liability values with corporate yields, credit bonds are especially well-suited.

For LDI to be effective, however, a plan doesn’t have to hold 100% of its assets in bonds: An effective hedge applied to a 40% bond allocation can suffice, observes Jim Reidy, head of JPMorgan Asset Management’s long duration product team in New York and a managing director of the firm. “By implementing a duration strategy on the bond allocation, that can be enough to match a good part of the liabilities and, in the event of another ‘perfect storm,’ let your plan live to fight another day.”

Early implementations of LDI set out simply to extend the duration of bond portfolios. Liability durations tend to be quite long, at around 15 years, while that of the most prevalent bond benchmark, the Barclays Capital Aggregate Bond Index, is quite a bit shorter at about 5 years. Duration can be lengthened “physically,” by buying longer-lived bonds, such as those in the Barclays Long Government/Credit index, or synthetically, by adding longer-interest rate swaps or Treasury bond futures.

“You had to get the interest rate duration right, but whether you used longer bonds, swaps, or futures didn’t matter as much in early applications, because the spread between the instruments was small and pretty consistent,” notes Scott McDermott, managing director of Global Portfolio Solutions with Goldman Sachs Asset Management.

Figure 1 plots the history of yields on 30-year U.S. Treasury bonds, 30-year swaps based on Treasurys, and the Barclays Capital Long Government/ Credit index (LGC). (Fifty percent of the index is in government bonds.) For years, swaps and the LGC traded consistently above Treasurys in a range of 30 to 75 basis points, in up and down markets, until late 2007.

McDermott’s contention is echoed by Reidy of JPMorgan: “If you look at the left side of the chart (Figure 1), you would say it didn’t matter much which asset you used to hedge your liabilities, because they all moved in the same direction.”

Each type of application has its nuances, however. Physical long credit bonds, or shorter credit bonds extended with futures, are a more general form of risk reduction. What is available in the market can approximate the duration of a plan’s liabilities, say, within a year or two; for well-funded plans that may be adequate protection. Interest rate swaps, on the other hand, are custom-designed contracts that can match a plan’s liability duration very closely but, since they are based on Treasurys, they don’t have a credit component. Physical bonds are more conventional and easy to monitor, while a swap wrapped around 40% of the pension portfolio can require high maintenance on the staff’s part, and might be an idea too novel for many boards.

Whatever the reasons behind their choices of LDI methods circa 2007, few sponsors could have anticipated their results over the next two years.

First, consider 2008: Implementations based on credit bonds provided an effective hedge to the drama playing out in the markets. Unless there had been a rash of defaults in the portfolio, value of the bonds would have moved in line with liabilities, by definition. “Lately, it’s become clear that long credit bonds can offer a hedging advantage, keeping the value of plan assets in line with liability valuations,” notes Mark Ruloff Director of Asset Allocation with consultants Towers Watson, in the firm’s Arlington, Virginia, office.

That point was brought home in August 2007, when subprime became a household word and, as the markets became wary of credit risk, the spread between the LGC and Treasurys started to diverge. (The withdrawal of bond market liquidity of major dealers—Bear Stearns, Lehman Brothers and Merrill Lynch—exacerbated the effects of the fear of corporate failure.) By the end of 2008, credit spreads were at record wide levels, hurting the prices of bonds, but cutting the value of the liabilities as well. The hedges worked.

For LDI executed through swaps, however, the results of 2008 were weirdly mixed. As a hedge, swaps failed badly, because their yields moved in the opposite direction from those corporate bonds, the object of the hedge. In the right side on Figure 1, note how bond rates soared, while swap rates plummeted. (The swaps-Treasury spread even went into negative territory, an unusual condition that remains today.)

As an asset, however, swaps were phenomenal: A 30-year duration Treasury swap gained 67% during 2008. Sponsors were stunned by the sudden gains on such a large portion of their plan assets.

Asset managers and consultants rushed to meet with sponsors to convince them to unwind their swaps for two reasons: one, to get rid of the failed hedge, and two, to realize the windfall profit. Even though the swaps were intended as perennials—protection intended to remain in place for years—“Most clients were quite receptive to unwinding them,” recalls McDermott of Goldman Sachs. “We went in and said to them, ‘Frankly, the hedge didn’t work, but it’s one of those rare cases where it failed in your favor and, if I were in your chair, I’d be worried about having that movie play in reverse’,” that is, giving back the gain on the swap as the markets recovered a more normal view of counterparty risk. “That reasoning usually got their agreement, but many of them already had thought of it independently.”

Once the swaps were taken off, the plans needed to reinstate some sort of LDI. The solution was ready-made: long-term corporate bonds. “At the beginning of 2009, swap spreads were low, and corporate bond spreads were high,” notes Towers Watson’s Ruloff. Corporate bonds were an attractive investment choice for the tactical reason of their high yields, but also fit the strategic view, since corporates better match up with the new valuation regulations.

“When corporate yields hit their spike in October 2008,” recalls Reidy of JPMorgan, “we said to clients ‘If you’re trying to be close to your liabilities, you cannot ignore this opportunity,’” to buy the assets needed to reestablish the hedge at bargain pricing.

The payoff was immediate: Between November 2008 and October 2009, the Barclays Capital Long Credit index returned 44%. “There was enough time to react,” Reidy adds, “but, at the time of the spike in yields, with the world about to end, loading up on credit was not the thing that felt the best, but we told clients, ‘You can’t go with Treasurys—you’ve got to think about buying corporate bonds.’ Those clients who had nerve, and the ability to move quickly, are the ones that had the best experience.”

Attractive yields notwithstanding, Scott McDermott points out that unwinding swaps was problematic for some sponsors: Where swaps had been tailored to unusual situations, the move to physical bonds didn’t cover as much duration as the swap did. “It was a trade-off between picking up the credit spread and losing the unfavorable basis risk in the pricing of the swaps, but losing some of the interest rate risk coverage,” he explains. “So, for some sponsors, the right trade-off was to cut back the swap position, rather than eliminate it entirely.”

In early 2010, LDI practitioners still are favoring corporate bonds, both as to strategy and tactics. The swap market remains disrupted, with yields below Treasurys, so managers advise staying away. McDermott observes: “At some point in the future, derivatives will provide sponsors with long-term solutions but, right now, that market is just not favorably priced.”

“If a sponsor tells me that the plan needs to be very close to the liability, and they have a lot of money in fixed income, then, unless the plan is significantly overfunded, they still have to be looking at corporate bonds,” says Reidy. Fortunately, corporations have brought new bonds to the markets in record amounts in 2009 and 2010.

For those concerned that spreads have moved too far and corporate bonds have become expensive, he suggests a portfolio managed to the Barclays Capital Long Government/Credit index, about half of which is in U.S. government bonds. “The LGC has a favorable basis risk,” he notes. “If the markets were to blow up, the return on an LGC portfolio would be higher than that of the liability, because the Treasury portion would outperform and, if the discount rate is a corporate rate, that is the neutral portfolio anyway, and the sponsor can be indifferent to which way spreads go.”

Will broader implementation of LDI, and the reduction of risk it can bring to sponsors’ financial health, issue a reprieve to defined benefit plans in U.S. corporations? “It won’t shorten their lives—managing the assets better will help to keep them around,” says Goldman’s McDermott, “but the real headwind comes from other issues. In lots of cases, new employees don’t see a DB plan as a reason to go to work for a company. When such an expensive benefit isn’t appreciated, sponsors are more likely to close and end their DB plans.”

– John Keefe