FOURTH IN A SERIES Interest Rate Swap-Based Hedging ...

[Pages:8]Pyramis Global Advisors | Research Paper

By Michael J. Senoski, FSA, CFA Vice President and LDI Investment Director

Pyramis Global Advisors

FOURTH QUARTER 2008 FOURTH IN A SERIES

Interest Rate Swap-Based Hedging Strategies for Pension Plan Sponsors

November 30, 2008 --The use of interest rate swaps1 as part of a strategy to hedge the interest rate risk embedded in a pension plan's liability structure raises a number of issues, including the use of leverage, the valuation and posting of collateral, counterparty risk, and basis risk. Basis risk arises from the imperfect correlation between the cash bond market interest rates that are the regulatory standard for measuring pension liabilities, and swap market rates. The purpose of this paper is to review the regulatory framework for measuring pension liabilities, examine the historical record of cash bond market rates relative to swap rates, and evaluate the effectiveness of swapbased hedging strategies in reducing the tracking error between pension asset and liability returns of a sample group of pension plans. We will demonstrate that the potential benefits of swap-based hedging strategies are clear and measurable, notwithstanding the basis risk that is inherent in such strategies.

Regulatory Framework The move toward liability driven investing (LDI) is motivated largely by both the financial accounting and minimum funding requirements that apply to pension plans. In both cases, the basis for measuring pension liabilities is grounded in the cash bond market, not the interest rate swap market. This is an important distinction. Understanding its significance and taking it into consideration during the design phase will result in an LDI strategy that's better informed and more effective.

Financial Accounting The possibility of using a swap curve rather than the cash bond market as the basis for measuring pension liabilities was not an option available to the architects of Statement of Financial Accounting Standards No. 87 (FAS 87).2 The interest rate swap market was only in nascent form during the deliberation and discussion phase in the early

1980s that led to the issuance of the final version of FAS 87 in December 1985. The notional value of outstanding interest rate swaps was a mere $80 billion at that time (compared with $309 trillion in December 2007) and the process leading to the standardization of interest rate swap contracts had only just begun. The final version of FAS 87 adopted the concept of settlement with respect to the selection of the discount rate used to measure pension obligations. In estimating the cost of a settlement, plan sponsors could take into consideration the "rates of return on high-quality fixed-income investments currently available and expected to be available during the period to maturity of the pension benefits," as stated in FAS 87.

In 1993, the SEC issued guidelines with respect to the interpretation and application of the discount rate provisions of FAS 87. These guidelines referred to the rates available on AA-rated bonds as an appropriate basis

For Institutional Use Only

for FAS 87 discount rate determinations. In 1994, Salomon Brothers (now part of Citigroup) constructed and published a pension discount curve to help plan sponsors comply with the SEC guidelines. Currently updated on a monthly basis, the Citigroup Pension Discount Curve3 has been adopted by a large number of plan sponsors as the basis for pension liability measurements. In addition to the Citigroup curve, other service providers and plan sponsors have developed methods for creating an AA corporate bond yield curve for FAS 87 pension liability measurements.

Minimum Funding Requirements The federal minimum funding requirements applicable to defined benefit pension plans were amended by the Pension Protection Act of 2006 (PPA). An issue of particular concern to all the participants in the legislative process, including the Treasury Department, House, and Senate, was the methodology for determining the interest (discount) rate for measuring pension liabilities. Consequently, a considerable amount of time and effort was devoted to this particular issue during the deliberations leading up to passage of the final legislation. The outcome of the process was a yield curve approach for discounting pension liabilities that was similar in some critical respects to the financial accounting methodology. Like the rates used for financial

accounting, the discount rates used for minimum funding calculations are based on the cash bond market. In particular, discount rates for minimum funding are derived from cash market rates applicable to the top three quality levels, AAA, AA, and A. What is noteworthy here is that the size and significance of the interest rate swap market in 2006, when the PPA was passed, was obviously far greater than it was when FAS 87 was adopted. Yet there is no evidence to suggest that the use of a swap curve as a basis for discounting pension liabilities was ever a serious consideration. In fact, there is no mention at all of a swap curve in either a February 2005 white paper by the Treasury Department on the subject of pension discounting or a January 2006 update of the same.

A report on the subject of discount rates prepared by Ryan Labs, Inc., in 2001 as part of a research project sponsored by the Society of Actuaries addressed in some detail the pros and cons of the use of the swap curve as a basis for measuring pension liabilities. Prominent among the cons is the argument that the interest rate swap is primarily a tool for managing or hedging yield spread risk, but not suitable for managing total price or return risk. A second argument is that you cannot measure (market price) the value of a pension obligation using swap-based rates

because you cannot prefund or defease that obligation using a swap in association with any derivatives position. These arguments may or may not have reflected the thinking on this issue of the architects of the PPA.

Cash Market Rates Versus Swap Rates The starting point for developing a swapbased strategy for hedging the exposure of a pension plan's funded position to interest rate risk is an understanding of the relationship between cash market bond rates and swap rates. Focusing on the relationship between cash market spreads and swap spreads to Treasuries facilitates this understanding. Due to the linkage between swap rates and Libor, and because Libor is the borrowing rate across banks that typically carry an AA credit rating, we can control for differences in spread due to credit quality by comparing AA spreads in the cash market to swap spreads. The chart below compares historical 10-year AA spot rate spreads extracted from the aforementioned Citigroup curve to 10-year zero coupon swap-rate spreads.

Over the period from the inception of the Citigroup curve in 1995 through September 2008, 10-year AA spot rate corporate bond spreads exceeded swap spreads by an average of more than 50 basis points. The difference in basis points narrowed to less

Exhibit 1: 10-Year Spot Rate Spreads

500 400 300 200 100

0

AA Corporate Spreads Swap Spreads

Spread to Treasuries (bps) Sep-95 Mar-96 Sep-96 Mar-97 Sep-97 Mar-98 Sep-98 Mar-99 Sep-99 Mar-00 Sep-00 Mar-01 Sep-01 Mar-02 Sep-02 Mar-03 Sep-03 Mar-04 Sep-04 Mar-05 Sep-05 Mar-06 Sep-06 Mar-07 Sep-07 Mar-08 Sep-08

Date Sources: Lehman Live, Citigroup, Pyramis Global Advisors

2

For Institutional Use Only

than 20 basis points throughout much of 2005, while ballooning to well over 250 basis points in September 2008.

Similar results apply across the entire spot rate curve. Exhibit 2 shows the average of the difference between the AA corporate bond spreads and swap spreads over the entire (out to 30 years) spot rate curve over the September 1995 through September 2008 period. Generally increasing with maturity, the average difference between spot rate corporate bond and swap spreads ranges from roughly 20 basis points at the front end of the curve up to 80 basis points for longer maturities.

Excess Spread (bps)

Exhibit 2: Average Excess Spread--AA Corporate Spreads Over Swap Spreads

90 80 70 60 50 40 30 20 10

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Maturity

Sources: Lehman Live, Citigroup, Pyramis Global Advisors as of September 30, 2008

A number of reasons have been offered to explain why swap spreads are less than spreads on AA-rated bonds of the same maturity. Common among them is that the amount at risk to the counterparty on the receive fixed side of a swap transaction is limited to the mark-to-market value, if positive, of the swap based on the notional value of the swap. Because there is no exchange of principal in a swap transaction, this amount is limited to the present value of the difference between the remaining coupon payments and hypothetical coupon payments based on the receive fixed rate applicable to a current coupon swap of comparable tenor (maturity). This amount is typically far less than the amount at risk to the holder of an otherwise comparable cash market bond. Some of the other reasons relate to the level of hedging activity, the level and shape of the yield curve, and collateral and other credit enhancement features of swap contracts.

Of particular interest to plan sponsors considering a swap-based hedging strategy is the relationship between liability discount rates and swap rates. As noted above, the Citigroup Pension Discount Curve has gained widespread acceptance among plan sponsors as a basis for calculating pension liabilities in accordance with financial accounting standards. Exhibit 3 compares the Citigroup Pension Discount Curve as of September 30, 2008, with corresponding Lehman Brothers zero coupon swap rates.

For Institutional Use Only

Yield

Exhibit 3: Comparative Spot Rate Curves--September 30, 2008

Citigroup AA Curve Swap Curve 9 8 7 6 5 4 3 2 1 0

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

Maturity

Sources: Lehman Live, Citigroup, Pyramis Global Advisors as of September 30, 2008

Our review and discussion of cash market bond and swap spreads would lead us to believe that discount rates based on the Citigroup curve would be higher than swap rates. It turns out that this is exactly the case. Rates at every point across the Citigroup curve exceeded comparable swap rates with an average differential of over 290 basis points. This differential is unusually large and reflects the dramatic widening of credit spreads in the cash bond market that occurred as a result of the seizure of the credit markets in the wake of the bankruptcy or near collapse of several high-profile issuers, including Fannie Mae, Freddie Mac, AIG, Lehman Brothers,

and Washington Mutual. A comparison of Citigroup curve rates and swap rates over different time periods yields similar results, although the average differential is substantially narrower. It also seems reasonable to expect that a comparison of swap rates with the pension liability discount curves developed by other service providers and plan sponsors, and the monthly yield curve developed by the Treasury for federal minimum funding calculations, would yield similar results.

3

Since swap rates are lower than cash bond market rates, the fact that discount rates for financial accounting and minimum funding calculations are based on cash bond market rates and not swap rates should not be of any particular concern to a plan sponsor. On the contrary, it is easy to imagine that most plan sponsors welcome the near-term benefits, such as lower pension liabilities, pension expense, and contributions that result from the use of the higher rates. To plan sponsors considering a swap-based strategy for hedging the exposure of their plan's funded position to interest rate risk, the correlation between cash bond market rates and swap rates will be of much greater interest.

Exhibit 4 illustrates the correlation between month-end 2-, 5-, 10-, and 30-year Citigroup spot rates and the corresponding Lehman Brothers zero coupon swap rates over rolling 36-month periods from December 1998 through September 2008.

Correlations across the intermediate 2-, 5-, and 10-year maturities ran well above 0.9 for most of this period, while the 30-year maturity correlation ranged between 0.8 and 0.9 for an extended period. What is most striking is the dramatic decline in correlations across all maturities that began in mid-2007 as the impact of the subprime mortgage crisis on the availability and cost of credit intensified. In the summer of 2008, correlations across 10- and 30-year maturities actually dipped into negative territory. As has generally been the case during periods of financial stress, including the Russian debt crisis in 1998, the aftermath of 9/11 in 2001, and the Enron and WorldCom defaults in 2002, there was a significant widening of cash market spreads relative to swap spreads as these events unfolded.

Exhibit 4: Rate Correlations

1.000 0.800 0.600 0.400 0.200 0.000 -0.200 -0.400 -0.600

2-Year Zero 5-Year Zero

10-Year Zero 30-Year Zero

Correlation

Three Years Ending Sources: Lehman Live, Citigroup, Pyramis Global Advisors

Dec-98 Sep-99 Jun-00 Mar-01 Dec-01 Sep-02 Jun-03 Mar-04 Dec-04 Sep-05 Jun-06 Mar-07 Dec-07 Sep-08

4

For Institutional Use Only

Tracking Error At the end of the day, it is the performance of the swap-based hedging strategy relative to plan liabilities that is of greatest interest to the plan sponsor. More useful even than the correlation statistic in this regard is the tracking error of swap returns relative to liability returns. It is a relatively straightforward exercise to transform a time series of cash and swap market spot rates into their respective time series of spot rate returns, and the tracking error between them. Exhibit 5 illustrates the annualized tracking error of Lehman zero swap rate returns relative to Citigroup spot rate returns over rolling 36-month periods for the same time series of rates described above.

% Tracking Error

Exhibit 5: Tracking Error

25 20 15 10

5 0

2-Year Zero 5-Year Zero

10-Year Zero 30-Year Zero

Dec-98 Sep-99 Jun-00 Mar-01 Dec-01 Sep-02 Jun-03 Mar-04 Dec-04 Sep-05 Jun-06 Mar-07 Dec-07 Sep-08

The tracking errors across the two- and five-year maturities never rise above 1% and 2%, respectively, before spiking in September 2008 as the credit squeeze intensified. This is due to the generally high correlations noted above, as well as the relatively low durations at these maturities. For the 10-year maturity, tracking error periodically rises above the 3% level, and the upward drift since mid-2007 is especially pronounced. A totally different picture emerges for the 30-year maturity. After holding steady at around 10%, tracking error rose steadily for the next three years, peaking above 23% before descending to roughly 6% over the next three years. Since mid-2007, this figure has spiraled upward to over 15%.

Three Years Ending Sources: Lehman Live, Citigroup, Pyramis Global Advisors

Exhibit 6: Annualized Tracking Error

Plan A

Plan B

10

Plan C

8

6

4

Plan D

Plan E

% Tracking Error

Dec-98 Sep-99 Jun-00 Mar-01 Dec-01 Sep-02 Jun-03 Mar-04 Dec-04 Sep-05 Jun-06 Mar-07 Dec-07 Sep-08

While our focus until now has been on individual maturities, we are most interested in overall plan results. We can accomplish this by creating a portfolio of funded4 zero coupon swaps that is cash flow matched on an annual basis against the projected stream of benefit payments that represent plan liabilities. Just as above, we can then construct a historical time series of swap portfolio returns and plan liability returns (by assuming a stationary plan population), and observe the resulting tracking error. Exhibit 6 illustrates such annualized tracking error over rolling 36-month periods from January 1996 through September 2008 for five pension plans of increasing duration from 10 to 16 years.

2 0

Three Years Ending

Plan

Duration

A

10.1

B

10.6

C

12.7

D

12.8

E

16.0

Tracking Error

Minimum 1.4% 1.5% 1.8% 1.9% 2.5%

Sources: Lehman Live, Citigroup, Pyramis Global Advisors

Maximum 6.3% 6.5% 7.5% 7.8% 9.5%

Average 4.2% 4.4% 5.1% 5.4% 6.9%

For Institutional Use Only

5

These results appear reasonable based on the previous analysis and discussion. Tracking error persists, even with cash flow matching, a consequence of the less than perfect correlation between the cash market and swap rates. As we would further expect, there is an increase in tracking error with plan liability duration. And there is considerable variation in tracking error over time, a consequence of volatility in the relationship between cash market and swap rates.

The above results are indicative of the maximum potential benefits of a swap-based hedging strategy in terms of risk reduction. Such benefits should, of course, be weighed against the impact of the strategy on expected portfolio performance on a total return basis. Few, if any, plan sponsors

would be willing to allocate their entire portfolios to a pool of funded zero coupon swaps to achieve these results. The decrease in expected portfolio performance on a total return basis would be unacceptable. There is an almost infinite variety of hedging strategy possibilities. Many of these represent a clear and measurable reduction in risk without negatively affecting expected return. The following table illustrates the potential risk reduction benefits of several very basic hedging strategies for our five-plan universe. Each of the strategies maintains a 60% equity allocation, which is roughly equal to the average equity allocation for plan sponsors in the United States.

a conventional 60% equity/40% core (Lehman Aggregate) fixed-income strategy. A straightforward duration extension (Lehman Long Government/Credit) strategy reduces tracking error by 1.0% to 1.3% across the five plans. As an alternative to the duration extension strategy, a simple 40% overlay strategy using a 10-year bellwether swap reduces tracking error by 1.6% to 1.9%, while a 40% overlay using a 30-year bellwether swap reduces tracking error by 2.2% to 3.0%. None of these strategies should have a negative impact on expected return. In fact, some may argue that each would be additive to performance on a total-return basis.

Tracking error ranges from roughly 11% to 14% for our five-plan universe based on

Equity Core FI Long Duration FI Swaps

Asset Allocation 60% 40% 0% 0%

60% 0%

40% 0%

60% 40%

0% I

Plan (Duration) A (10.1) B (10.6) C (12.7) D (12.8) E (16.0)

Tracking Error 10.9% 11.1% 11.9% 12.2% 14.0%

9.9% 10.1% 10.8% 11.0% 12.7%

9.3% 9.4% 10.1% 10.4% 12.1%

I 40% notional 10-year bellwether swap II 40% notional 30-year bellwether swap

Sources: Lehman Live, Citigroup, Pyramis Global Advisors as of September 30

60% 40%

0% II

8.7% 8.8% 9.3% 9.5% 11.0%

0% 0% 0% 100%

4.2% 4.4% 5.1% 5.4% 6.9%

6

For Institutional Use Only

Conclusion Because of their unique features relative to cash market instruments, pension plan sponsors concerned about financial statement or contribution volatility should carefully consider the potential benefits of interest rate swaps. Over the short term, interest rate risk is the most significant risk embedded in a pension plan's liability structure. Unfunded or leveraged interest rate swaps can be used as a hedge against this risk while conserving plan capital for deployment in a variety of alpha-generating strategies such as absolute return or portable alpha strategies. Moreover,

transaction costs in the interest rate swap market compare favorably with cash bond market transaction costs, and transacting in the swap market is relatively straightforward once the appropriate documentation is in place. No other financial instruments can claim this unique blend of advantages within the LDI framework. At the same time, plan sponsors should be aware of the basis risk that arises when interest rate swaps are used as part of a strategy to hedge the funding level of a pension plan to interest rate risk. This risk arises because the basis

for measuring pension liabilities is grounded in the cash bond market, not the interest rate swap market. Yet despite this risk, swapbased hedging strategies offer the potential for a clear and meaningful reduction in risk. Moreover, this can be achieved with little if any negative impact on expected performance on a total-return basis.

1An interest rate swap is an agreement to exchange interest payments for a specific period of time on a specified amount of principal or notional value. The most common interest rate swap is a fixed-forfloating coupon swap. The notional principal is typically not exchanged.

2Statement of Financial Accounting Standards No. 87 establishes standards of financial accounting and reporting for an employer that offers pension benefits to its employees.

3The Citigroup Pension Discount Curve is intended as an AA corporate spot (zero coupon) rate curve at six-month intervals extending out to 30 years. It is not directly observable but is constructed based on a methodology developed by Salomon Brothers. This methodology adds option-adjusted AA spreads to an underlying Treasury curve.

4A swap paired with a cash investment that returns the payment of the floating rate leg of the swap.

Michael J. Senoski, FSA, CFA, is vice president and LDI investment director at Pyramis Global Advisors. In these roles he contributes to product development, analytical support, and overall marketing for LDI strategies. He has more than 30 years of experience as a pension actuary advising defined benefit plan sponsors on pension funding strategy and investment policy.

Note: You can access additional papers issued in the Pyramis series of white papers on contemporary investment topics, including "Liability Driven Investing: Setting the Benchmark," "Pension-Liability Analysis: An Application of Fixed-Income Analytics," and "Liability-Driven Investing: Risk Metrics and Strategy Evaluation," which each focus on liability driven investing, at .

For Institutional Use Only

7

About Pyramis Global Advisors Pyramis Global Advisors, a Fidelity Investments company, is an investment management firm focused on serving corporate and public retirement funds, endowments, foundations, other institutions, and non-U.S. investors. Pyramis offers active and risk-controlled domestic equity, international equity, fixed income, real estate, and alternative disciplines.

Pyramis Global Advisors, LLC, is a registered investment advisor, and Pyramis Global Advisors Trust Company is a New Hampshire?chartered trust company and an investment manager. Both are Fidelity Investments companies. Pyramis products and services are presented by Fidelity Investments Institutional Services Company, Inc., a nonexclusive financial intermediary that is an affiliate of Pyramis.

Information presented herein is for discussion and illustrative purposes only and is not a recommendation or an offer or solicitation to buy or sell any securities. Past performance is no guarantee of future results.

For more information about Pyramis and the custom LDI solutions being developed for defined benefit plans, please contact Michael Senoski, vice president and LDI investment director, Pyramis Global Advisors, at 401-292-4753, or Christian Pariseault, senior vice president and fixed-income investment director, Pyramis Global Advisors, at 401-292-4744.

? 2008 FMR LLC. All rights reserved.

All trademarks and service marks presented herein belong to FMR LLC or an affiliate, except for thirdparty trademarks and service marks, which belong to their respective owners. The views expressed herein are those of the individual contributors and do not necessarily represent the views of Pyramis Global Advisors.

Certain data and other information in this research paper were supplied by outside sources and are believed to be reliable as of the date presented. However, Pyramis has not and cannot verify the accuracy of such information. The information contained herein is subject to change without notice.

Pyramis does not provide legal or tax advice, and you are encouraged to consult your own lawyer, accountant, or other advisor before making any financial decision.

510548.1.0

For Institutional Use Only 1.882563.100

 ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download