CHAPTER 16
INTRODUCTION TO BOND PORTFOLIO MANAGEMENT
I. INTRODUCTION
The products of the fixed-income market, the risks associated with investing in fixed-income securities, and the fundamentals of valuation and interest rate risk measurement were covered. In depth coverage of the valuation of fixed-income securities with embedded options, features of structured products (mortgage-backed securities and asset-backed securities), and the principles of credit analysis were covered. Now, we will put all these tools together to demonstrate how to construct portfolios in such a way as to increase the likelihood of meeting an investment objective.
In this chapter, we set forth the framework for the investment management process. Regardless of the asset class being managed (i.e., stocks, bonds, or real estate), the investment management process follows the same integrated activities. John Maginn and Donald Tuttle define these activities as follows:
1. An investor’s objectives, preferences, and constraints are identified and specified to develop explicit investment policies;
2. Strategies are developed and implemented through the choice of optimal combinations of financial and real assets in the marketplace;
3. Market conditions, relative asset values, and the investor’s circumstances are monitored; and
4. Portfolio adjustments are made as appropriate to reflect significant change in any or all of the relevant variables.
256
In this chapter, we will use the Maginn-Tuttle framework to describe the investment management process for fixed-income portfolios. We refer to these four activities in the investment management process as:
1. setting the investment objectives
2. developing and implementing a portfolio strategy
3. monitoring the portfolio
4. adjusting the portfolio
A discussion of the investment management process provides a context in which to appreciate the significance of the chapters to follow in this book. Our focus, of course, is on the management of fixed-income portfolios.
II. SETTING INVESTMENT OBJECTIVES FOR FIXED-INCOME INVESTORS
The investment objectives of a fixed-income investor are often specified in terms of return and risk. The investment objectives should be expressed quantitatively in terms of some benchmark. The benchmark varies by the type of investor.
In general, we can divide fixed-income investors into two categories based on the characteristic of the benchmark. The first category of investor specifies the benchmark in terms of the investor’s liability structure. The investment objective is to generate a cash flow from the fixed-income portfolio that, at a minimum, satisfies the liability structure. The second category of fixed-income investor specifies the benchmark as a particular bond market index. The investment objective may be to match the performance of the bond market index after management fees, or to outperform the bond market index after management fees by at least a predetermined number of basis points. Below we discuss each category of investor further in terms of its investment objectives.
A. Return Objectives
An investor identifies the benchmark to the portfolio manager. The manager may be employed by the investor or may be an external manager. Once the investment objective is specified in terms of a benchmark, the performance of a manager will be evaluated relative to that benchmark. It is important to note that, even if a manager is successful in terms of outperforming the benchmark, the client may not realize its investment objective. This can occur if the client does not correctly specify a benchmark to reflect its investment needs from a risk, return, and cash flow perspective. A good example is a defined benefit pension fund that specifies the investment objective in terms of a particular bond market index when, in fact, a more appropriate benchmark would be the liability structure of the fund. If the manager outperforms the benchmark, then the manager is successful. However, if sufficient cash flow is not generated to satisfy the pension liabilities, then the failure lies not with the manager but with the client’s inability to properly specify an appropriate benchmark.
1. Liabilities as the Investment Objective In general, investors who specify the benchmark in terms of a liability structure that must be satisfied fall into two categories. The first category consists of investors who borrow funds and then invest those funds. Here, the objective is to earn a return on the funds borrowed that is greater than the cost of borrowing. The difference between the return on the funds invested and the cost of borrowing is called the spread. These investors are referred to as funded investors.
Depository institutions (banks, savings and loan associations, and credit unions) are clearly funded investors. Insurance companies have a wide range of products. For some products, the insurance company is a funded investor. For example, an insurance company that issues a guaranteed investment contract (i.e., a policy where the insurance company guarantees a specified interest rate to policyholders for a specific time period) has basically borrowed money from policyholders and created a liability. Another example of a funded investor is a hedge fund, which is a highly leveraged entity that borrows on a short-term basis, typically using a repurchase agreement. (In Chapter 18 we will explain how a manager can use a repurchase agreement to borrow funds using the securities purchased as collateral.) The objective is to earn a return on the funds obtained through the repurchase agreement that is greater than the borrowing cost (i.e., repo rate).
The second category consists of institutional investors who must satisfy a liability structure but who did not borrow the funds that created the liability structure. One example is a pension sponsor that faces a liability structure based on defined benefits. A second example is a state that invests proceeds from a lottery so as to meet the state’s obligation to make payments to a lottery winner.
Earlier, the focus was on fixed-income products. Now we are concerned with bond portfolio management, so we must understand how the liability structure affects the selection of a portfolio strategy. In this section we take a closer look at the nature of liabilities.
a. Liabilities Defined A liability is a potential cash outlay to be made at a future date to satisfy the contractual terms of an obligation. An institutional investor is concerned with both the amount of the liability and the timing of the liability, because the investor’s assets must produce sufficient cash flow to meet promised payments.
b. Classification of Liabilities Liabilities are classified according to the degree of uncertainty with regard to amount and timing, as shown in
Exhibit 1. This exhibit assumes that the holder of the obligation will not elect to cancel the obligation prior to any actual or projected payout date.
The classification of cash outlays as either “known” or “uncertain” is undoubtedly broad. When we refer to a cash outlay as being uncertain, we do not mean that it cannot be predicted. For some liabilities, the “law of large numbers” makes it easier to predict the timing and/or the amount of the cash outlays. This work is typically done by actuaries. Below we illustrate each type of liability.
A Type-I liability is one for which both the amount and timing of the liability is known with certainty. An example is a liability for which an institution knows that it must pay $8 million six months from now. Depository institutions know the amount they are committed to pay (principal plus interest) on the maturity date of a fixed-rate deposit, assuming that the depositor does not withdraw funds prior to the maturity date. Type-l liabilities are not limited to depository institutions. A guaranteed investment contract is an example of this type of liability that is created by a life insurance company.
EXHIBIT 1 Classification of Liabilities of Institutional Investors
A Type-II liability is one for which the amount of the cash outlay is known, but the timing of the cash outlay is uncertain. The most obvious example of a Type-II liability is a typical life insurance policy. There are many types of life insurance policies, but the most basic provides that, for an annual premium, a life insurance company agrees to make a specified dollar payment to policy beneficiaries upon the death of the insured. Naturally, the timing of the insured’s death is uncertain.
A Type-III liability is one for which the timing of the cash outlay is known, but the amount is uncertain. An example is, a 2-year floating-rate certificate of deposit issued by a depository institution, with an interest rate that is reset quarterly based on a particular market interest rate.
For a Type-IV liability, there is uncertainty as to both the amount and the timing of the cash outlay. Numerous insurance products and pension obligations are in this category. Probably the most obvious examples are automobile and home insurance policies issued by property and casualty insurance companies. When, and if, a payment will be made to the policyholder is uncertain. Whenever an insured asset is damaged, the amount of the payment that must be made is uncertain.
The liabilities of pension plans can also be Type-IV liabilities. For defined benefit plans, retirement benefits depend on the participant’s income for a specified number of years before retirement and the total number of years the participant worked. This affects the amount of the cash outlay. The timing of the cash outlay depends on whether the employee remains with the sponsoring plan until retirement and when the employee elects to retire. Moreover, both the amount and the timing depend on how the employee elects to have payments made—over the employee’s lifetime or those of the employee and spouse.
2. Bond Market Index as the Investment Objective When there are no liabilities that must be met, the investment objective is often to either match or outperform a designated bond market index. It is important to note that some clients that have a liability structure choose to specify that the manager manage against a bond market index. This is particularly true of pension sponsors. The expectation of the plan sponsor is that the performance of the bond market index selected will generate sufficient cash flow to satisfy the liability structure.
In selecting a benchmark, there are several characteristics a client should consider so that the benchmark can serve as an effective tool for evaluating a portfolio manager. According to Bailey, Richards, and Tierney, the basic characteristics of any useful benchmark are:
257 • Unambiguous: The names and weights of the securities included in the benchmark are clearly identifiable.
• Investable: The client has the option to simply buy-and-hold the benchmark rather than have funds actively managed.
• Measurable: The benchmark’s return can be calculated on a reasonably frequent basis by either the manager, client, or some third party.
• Appropriate: The benchmark is consistent with the manager’s investment style.
•
Reflective of current investment opinions: The manager has current investment knowledge of the securities included in the benchmark and the way they are categorized within the benchmark.
258 • Specified in advance: The benchmark is constructed before the beginning of the period for which the manager is evaluated.
The investable characteristic of a benchmark is critical. A benchmark represents the return to a passive investment strategy. The establishment of a benchmark means that active management decisions made by a portfolio manager can be judged. For example, Bailey, Richards, and Tierney note that it is common for consultants and plan sponsors to use as a benchmark the median performance of a group of portfolio managers for the same asset class. Such a benchmark is not investable.
It is essential for a manager to understand the composition and risk profile of a bond market index. Here we review the composition of various bond market indexes. These indexes are classified as broad-based U.S. bond market indexes, specialized U.S. bond market indexes, and global and international bond market indexes. In Chapter 17, we explain the risk profile of a bond market index because the risk profile of a portfolio relative to that of the benchmark bond market index determines relative performance.
a. Broad-Based U.S. Bond Market Indexes The three broad-based U.S. bond market indexes most commonly used by institutional investors are the Lehman Brothers U.S. Aggregate Index, the Salomon Smith Barney (SSB) Broad Investment-Grade Bond Index (BIG), and the Merrill Lynch Domestic Market Index. There are more than 5,500 issues in each index. One study found that the correlation of annual returns between the broad-based bond market indexes is around 98%.4
The three broad-based U.S. bond market indexes are computed daily and are “market-value weighted.” This means that, for each issue, the ratio of the market value of an issue relative to the market value of all issues in the index is used as the weight for the issue in all calculations. The securities in the SSB BIG index are all trader priced. For the two other indexes, the securities are either trader priced or model priced. Each index handles intra-month cash flows that must be reinvested in a different way. For the SSB BIG index, these cash flows are assumed to be reinvested at the 1-month Treasury bill rate, while for the Merrill Lynch index, they are assumed to be reinvested in the specific issue. The Lehman index assumes no reinvestment of intra-month cash flows.
Each index is broken into sectors. The Lehman index, for example, is divided into the following six sectors: (1) Treasury sector, (2) agency sector, (3) mortgage passthrough sector, (4) commercial mortgage-backed securities sector, (5) asset-backed securities sector, and (6) credit sector.
Exhibit 2 shows the percentage composition of the index as of September 8, 2003.
The agency sector includes agency debentures, but not mortgage-backed or asset-backed securities issued by federal agencies. The mortgage passthrough sector includes agency passthrough securities—Ginnie Mae, Fannie Mae, and Freddie Mac passthrough securities. Recall that the mortgage passthrough securities issued by Ginnie Mae are referred to as agency passthroughs and those issued by Fannie Mae and Freddie Mac are called conventional passthroughs. However, we shall simply refer to all mortgage passthroughs of these three entities as agency passthroughs. Agency collateralized mortgage obligations and agency stripped mortgage-backed securities are not included in the index. These mortgage derivatives products are not included because inclusion would result in double counting since they are created from agency passthroughs.
In constructing the index for the mortgage sector, for example, the Lehman index groups more than 800,000 individual mortgage pools with fixed-rate coupons into generic aggregates. These generic aggregates are defined in terms of agency (i.e., Ginnie Mae, Fannie Mae, and Freddie Mac), program type (i.e., 30-year, 15-year, balloon mortgages, etc.), coupon rate for the passthrough, and the year the passthrough was originated (i.e., vintage). For an issue to be included, it must have a minimum of $100 million outstanding and a minimum weighted average maturity of one year. Agency passthroughs backed by pools of adjustable-rate mortgages are not included in the mortgage index. (We discuss the composition of this sector in the next chapter.)
The credit sector in the Lehman Brothers index includes corporate issues. In the other two U.S. broad-based bond market indexes, this sector is referred to as the corporate sector.
EXHIBIT 2 Percentage Composition of Lehman Brothers U.S. Aggregate Index as of September 8, 2003
Source: Global Relative Value, Lehman Brothers, Fixed Income Research, September 8, 2003.
Sector | Percent of market value |
---|
Treasury | 21.94% |
Agency | 11.96 |
Mortgage Passthroughs | 34.40 |
Commercial Mortgage-Backed Securities | 2.55 |
Accept-Backed Securities | 1.89 |
Credit | 27.27 |
Total | 100.00% |
b. Specialized U.S. Bond Market Indexes The specialized U.S. bond market indexes focus on one sector or sub-sector of the bond market. Indexes on sectors of the market are published by the same three firms that produce the broad-based U.S. bond market indexes. Nonbrokerage firms have created specialized indexes for sectors. For example, Ryan Labs produces a Treasury index. Since none of the broad-based U.S. bond market indexes include noninvestment-grade or high-yield issues, indexes for this sector have been created by the three firms that have created the broad-based indexes and the firms CS First Boston and Donaldson Lufkin and Jenrette. The number of issues included in each high-yield index varies from index to index. The types of issues permitted (e.g., convertible, floating-rate, payment-in-kind) also vary. The index creators treat interim income and default issues differently.
c. Global and International Bond Market Indexes The growth in non-U.S. bond investing has resulted in the proliferation of international bond market indexes. Three types of indexes that include non-U.S. bonds are available. The first is an index that includes both U.S. and non-U.S. bonds. Such indexes are referred to as global bond indexes or world bond indexes. The second type includes only non-U.S. bonds and is commonly referred to as international bond indexes or ex-U.S. bond indexes. Finally, there are specialized bond indexes for particular non-U.S. bond sectors. Indexes can be reported on a hedged currency basis and/or an unhedged currency basis.
There are two types of global bond indexes. The first type restricts each country sector to government bonds. For example, the Merrill Lynch Government Bond Index includes the government sectors of the following countries:
Europe-EMU: Austria, Belgium, Finland, France, Germany, Ireland, Italy, Netherlands, Portugal, and Spain
Europe-Non-EMU: Denmark, Sweden, Switzerland, and U.K.
North America: Canada and U.S.
Japan & Asia/Pacific: Australia, Japan, and New Zealand
Other similar indexes are the Salomon World Government Bond Index and the J.P. Morgan Global Bond Index.
An example of a global bond index that includes government and non-government sectors is the Lehman Brothers Global Index (which Lehman also refers to as its “Core Plus Plus” Portfolio). The index is divided into a U.S. dollar sector and a nondollar sector. The U.S. dollar sector includes all the sectors within the Lehman Brothers U.S. Aggregate Index plus high-yield bonds and dollar-denominated emerging market bonds. The nondollar sector includes the following countries: France, Germany, Italy, Spain, Sweden, and the United Kingdom. An index such as the Lehman Brothers Global Index is an appropriate benchmark for a manager who is permitted to invest in both U.S. and non-U.S. bonds.
An international bond index is an appropriate benchmark for a manager who invests only in non-U.S. bonds. There are two types of non-U.S. indexes. The first is an index that includes government and nongovernment sectors. Other indexes include just the government sector of an international bond index. For example, the Salomon Non-U.S. Government Index is a byproduct of the Salomon World Government Bond Index.
The third type of non-U.S. bond index is a specialized bond index. Examples of such indexes published by two investment banking firms are:
Lehman Brothers | Merrill Lynch |
---|
Eurodollar Index | Pan European Broad Market Index |
Emerging Markets Index | EMU Broad Market Index |
Pan-European Aggregate Index | European Currencies High Yield Index |
Pan-European High-Yield Index | Emerging Europe Index |
Euro-Aggregate Index | |
B. Risks
The inability to satisfy an investment objective is called performance risk. Let’s look at the risks associated with strategies for institutional investors managing funds against a bond market index and those managing funds against a liability structure. A more detailed discussion on quantifying performance risk appears in Chapter 17.
1. Risks Associated with Managing Relative to a Bond Market Index Earlier, the risks associated with investing in individual bonds were discussed. These risks are reviewed in the next chapter, and we also expand our discussion to include risks associated with a portfolio and a benchmark index. We shall refer to the bond market index as the “benchmark index.” It is essential that a manager understand the risk profile of a benchmark index. The portfolio’s relative performance is determined by the differences between the risk profile of the benchmark index and the risk profile of the portfolio.
It is essential that we describe the risk profile of a portfolio and a benchmark index, and that we are able to quantify that risk profile. We will see in the next chapter that there are several measures to quantify a portfolio’s risk. One measure that allows the manager to incorporate all of the major risks associated with a portfolio relative to a benchmark index is tracking error. (We postpone discussion of this important measure until the next chapter.) For now, it is important when constructing a portfolio to be able to predict how well the performance of the portfolio will track the future performance of a benchmark index. The larger the tracking error the greater the likelihood that the portfolio’s performance will differ from the performance of the benchmark index.
2. Risks Associated with Managing Against a Liability Structure Some institutional investors want their managers to construct portfolios to meet a liability structure. A liability structure can be a single future liability or multiple liabilities. The risk associated with managing portfolios when the benchmark is liabilities is that the managed portfolio might not generate sufficient cash flow to satisfy the liability structure. Managing relative to a liability structure is popularly referred to as asset-liability management. (Our focus here is on managing the assets to satisfy the liability structure, not in managing liabilities.)
As is the case for assets, there are risks associated with liabilities. Here are three examples that should give you a feel for some of these risks. We will see more examples—and more details—in other chapters.
a. Call Risk for Liabilities Consider a liability that can be terminated by the holder at his option. The simplest case is a certificate of deposit specifying that the depositor can withdraw funds prior to the maturity date, but must pay a penalty to do so. If, due to a rise in interest rates, one or more depositors terminate CDs prior to the maturity date, then the depository institution will have to fund itself with new deposits (or borrow in the market) at a higher rate. But, if the depository institution previously invested in a fixed-rate asset paying less than the new borrowing rate, the depository institution realizes a negative spread. This risk is very much like call risk but in this case, the depository institution is concerned with the premature withdrawal of funds when interest rates rise, not when rates fall.
b. Cap Risk Funded investors might invest in a floating-rate bond. Typically, the floater has a cap (i.e., a maximum interest rate). To fund the investment in the floater, the investor might borrow funds on a short-term basis. There is no cap on the borrowing cost. Thus, if rates rise above the cap specified for the floater and there is no cap on the liabilities, then, at some point the funding cost will exceed the rate earned on the floater. This risk is called cap risk.
c.
Interest Rate Risk Let’s look at the interest rate risk for an institution as a whole. The
economic surplus of an entity is the difference between the market value of its assets and the market value of its liabilities. That is,
economic surplus = market value of assets − market value of liabilities
While the concept that assets have a market value may not seem unusual, one might ask: What is the market value of liabilities? This value is simply the present value of the liabilities, where the liabilities are discounted at an appropriate interest rate. A rise in interest rates will therefore decrease the present value or market value of the liabilities; a decrease in interest rates will increase the present value or market value of liabilities. Thus, the economic surplus can be expressed as:
economic surplus = market value of assets − present value of liabilities
For example, consider an institution that has a portfolio comprised of only bonds and liabilities. Let’s look at what happens to the economic surplus if interest rates rise. This will cause the bonds to decline in value; but it will also cause the liabilities to decline in value. Since both the assets and liabilities decline, the economic surplus can either increase, decrease, or remain unchanged. The net effect depends on the relative interest-rate sensitivity of the assets and the liabilities.
We can define the duration of liabilities as the responsiveness of the value of the liabilities to a 100 basis point change in interest rates. We define the dollar duration of assets and the liabilities in terms of a 100 basis point change in interest rates per $100 present value. If the dollar duration of the assets is less than the dollar duration of the liabilities, the economic surplus will decrease if interest rates fall. For example, suppose that the current market value of the asset portfolio is equal to $100 million and the present value of liabilities is $90 million. Then the economic surplus is $10 million. Suppose that the duration of the assets is 3 and the duration of the liabilities is 5. This means that the dollar duration per 100 basis point change in interest rates for the assets per $100 of market value is $3 and the dollar duration of the liabilities per $100 present value is $5.
Consider the following two scenarios. In the first scenario, interest rates decline by 100 basis points. Because the dollar duration of the assets is $3, the market value of the assets will increase by approximately $3 million to $103 million. The present value of the liabilities will also increase. Since the dollar duration of the liabilities is $5, the present value of the liabilities will increase by approximately $4.5 million to $94.5 million. Thus, the economic surplus decreases from $10 million to $8.5 million as a result of a decline in interest rates.
In the second scenario, assume that interest rates rise by 100 basis points. Because the dollar duration of the assets is $3, the market value of the assets will decrease by approximately $3 million to $97 million. The value of the liabilities will also decrease. Since the dollar duration of the liabilities is $5, the present value of the liabilities will decrease by $4.5 million to $85.5 million. The economic surplus therefore increases to $11.5 million from $10 million as a result of the rise in interest rates.
Notice that if we looked only at the interest rate risk of the bond portfolio for this institution, our concern is that the assets increase in value if interest rates fall. The interest rate risk for the assets alone derives from a possible increase in interest rates. However, when we analyze the change in interest rates, considering both assets and liabilities, we see that a decline in interest rates is the source of the interest rate risk to this institution because it reduces the economic surplus.
This example is particularly important for corporate pension sponsors since financial accounting rules specify that assets and liabilities must be marked-to-market (FASB 87). Moreover, the accounting rules require that, if the surplus becomes negative, then the deficit must be reported as a liability on the corporate sponsor’s balance sheet.
While our focus has been on the duration of assets, recall that duration is only a first approximation of the sensitivity of an asset or a liability to a change in rates. In analyzing the interest rate sensitivity of assets relative to liabilities, the convexity of both must also be considered. Recall that assets can have negative convexity. This means that, even if the duration of the assets and liabilities is matched, a portfolio of assets with negative convexity will not increase by as much as the liabilities when interest rates decline. This would result in a decline in the economic surplus. Consequently, to determine the true impact on the value of the economic surplus, the convexity of the assets and the liabilities must be considered.
C. Constraints
Clients impose constraints on managers. Examples of constraints a client might impose are a maximum allocation of funds to particular issuer or industry, a minimum acceptable credit rating for issues eligible for purchase, the minimum and maximum duration for the portfolio, whether leverage is permitted, whether shorting is permitted, and limitations on the use of derivative instruments (i.e., futures, options, swaps, caps, and floors).
The constraints imposed should be realistic and consistent with the investment objective. For example, suppose an insurance company issues a 5-year guaranteed investment contract (GIC) with a rate guaranteed 200 basis points over the on-the-run 5-year Treasury issue which has a yield of 6%. The investment objective is then to earn 8% plus a spread for the risk the insurance company incurs. However, if the constraints imposed on the manager require investment only in AAA rated securities and maturities that do not exceed five years, then it will be extremely difficult for the manager to meet the investment objective without excessively trading the portfolio to try to generate short-run returns.
In addition to client-imposed constraints, regulators of state-regulated institutions such as insurance companies (both life and property and casualty companies) may restrict the amount of funds allocated to certain major asset classes. Even the amount allocated within a major asset class may be restricted, depending on the characteristics of the particular asset. Managers of pension funds must comply with ERISA requirements. In the case of investment companies, restrictions on asset allocation are set forth in the prospectus and may be changed only with approval of the fund’s board of directors.
Tax implications must also be considered. For example, life insurance companies enjoy certain tax advantages that make investing in tax-exempt municipal securities generally unappealing. Because pension funds too are exempt from taxes, they are not interested in tax-exempt municipal securities.
III. DEVELOPING AND IMPLEMENTING A PORTFOLIO STRATEGY
The second activity in the investment management process is developing and implementing a portfolio strategy. The strategy must be consistent with the investment objective and all constraints. This activity can be divided into the following tasks:
• writing an investment policy
• selecting the type of investment strategy
• formulating the inputs for portfolio construction
• constructing the portfolio
We discuss each of these activities in the following pages.
A. Writing an Investment Policy
The investment policy is a document that links the investor’s investment objectives and the types of strategies that the manager (internal and external) may employ in seeking to reach those objectives. The investment policy should specify the permissible risks and the manner in which performance risk is measured.
Typically, the investment policy is developed by the investor in conjunction with a consultant. Given the investment policy, investment guidelines are established for individual managers hired by the investor. For example, if the investment objective is to outperform an aggregate bond index, a different manager might be hired to manage each sector of the index. While at the investor level the investment objective may be to outperform the bond index, the investment objective of each manager hired would be to outperform a specific sector of the index.
The investment guidelines established for each manager are developed in conjunction with the investor, the investor’s consultant, and the individual manager. The investment guidelines must be consistent with both the investment policy and the investment philosophy of the manager being retained. The investment guidelines should also define how managers who are engaged will be evaluated.
B. Selecting the Type of Investment Strategy
In the broadest terms, portfolio strategies can be classified as either active strategies or passive strategies. (A finer breakdown of this classification is provided in Chapter 18.) Essential to all active bond portfolio strategies is the specification of expectations about the factors that influence the performance of an asset class. The risk factors that have historically driven the return on bond portfolios were reviewed, and more detail is provided in the next chapter. In the case of active bond management, this may involve forecasts of interest rates, changes in the term structure of interest rates, interest rate volatility, or yield spreads. Active portfolio strategies involving non-base currency bonds require forecasts of exchange rates as well as local market interest rates.
Passive strategies require minimal expectational input. One popular type of passive strategy is indexing, whose objective is to replicate the performance of a designated bond market index. While indexing has been employed extensively in the management of equity portfolios, indexing of bond portfolios is a relatively new practice. Indexing is covered more extensively in Chapter 18.
Several bond portfolio strategies classified as structured portfolio strategies have been commonly used. A structured portfolio strategy involves designing a portfolio so as to achieve the same performance as a designated benchmark. Such strategies are frequently followed when trying to satisfy liabilities. Immunization is a strategy designed to generate funds to satisfy a single liability regardless of the course of future interest rates. When the designated benchmark involves satisfying multiple future liabilities regardless of how interest rates change, strategies such as immunization and cash flow matching (or dedication) are being used. These strategies will be covered in Chapter 19.
1. Strategy Selection and Risk Strategy selection begins with a comprehensive analysis of the risk profile of the benchmark index. For that reason, in Chapter 17, we cover not only the various risk characteristics that make up the risk profile of a benchmark index but we also show how these risks can be quantified. Once we understand the risk characteristics of a benchmark index, we can differentiate between the two general types of investing—passive versus active management.
The most common form of passive strategy is bond indexing. In a bond indexing strategy, a portfolio is created to mirror the risk profile of the benchmark index. Typically, it is difficult to replicate the benchmark index precisely, at minimal cost, for the reasons explained in Chapter 18. As a result, the tracking error for an indexed portfolio should be small.
In active management the manager creates a portfolio that departs from the risk characteristics of the benchmark index. The types of departure determine the differences between the risk characteristics of the actively managed portfolio and the benchmark index. The manager makes a decision about the relative risks he wants to accept, compared to the benchmark, based on his expectations. Here tracking error relative to an indexing strategy is expected to be large.
Active management can be differentiated in terms of the “degree” of departure from the risk profile of the benchmark index. Some managers create portfolios with a risk profile that differs in a small way from the benchmark index. Such a management strategy is referred to as “enhanced indexing.” The problem is that “in a small way” is difficult to quantify, so knowing where enhanced indexing ends and active management begins is subjective. (Enhanced indexing is covered in Chapter 18.)
2. Role of Derivatives in Investment Strategies Regardless of whether an active or a passive strategy is selected, a manager must decide whether to employ derivatives in implementing a strategy. Derivatives include futures, forwards, swaps, options, caps, and floors. Of course, the use of derivatives and restrictions on their use are established by the client and/or regulators.
Derivative instruments can be used to control the interest rate risk of a portfolio. The advantages of using derivative instruments rather than cash market instruments are explained in Chapter 22. Bond portfolio strategies employing derivatives to control interest rate risk are also explained in Chapter 22.
C. Formulating the Inputs for Portfolio Construction
Formulating the inputs for portfolio construction in an active portfolio strategy requires two tasks. The first is a forecast by the manager of the inputs that are expected to impact the performance of a security and a portfolio. For many strategies this involves forecasting changes in interest rates, changes in interest rate volatility, changes in credit spreads, and, for international bond portfolios, changes in exchange rates. A discussion of forecasting models is beyond the scope of this book.
The second task is to extrapolate from market data the market’s “expectations.” Recall that a manager’s view is always relative to what is “priced” into the market. We illustrated this when we introduced forward rates. To illustrate the essential point, we can use a simple example of a manager who has a 1-year investment horizon and is deciding between the following two alternatives:
• buy a 1-year Treasury bill
• buy a 6-month Treasury bill, and, when it matures in six months, buy another 6-month Treasury bill
The alternative selected by the manager is not based solely on the manager’s forecast of the 6-month Treasury bill rate six months from now. Rather, it is based on that forecast relative to the 6-month rate that is “priced” into the 1-year Treasury bill rate. For example, suppose that the 6-month Treasury bill rate is 3.0% (on a bond-equivalent basis) and the 1-year Treasury bill rate is 3.3% (on a bond-equivalent basis).
259 Earlier, it was shown that the 6-month interest rate six months from now that would make the manager indifferent to investing in either alternative is 3.6%.
6 This 3.6% interest rate is sometimes referred to as the “breakeven rate,” or more commonly the “forward rate.” The manager would be indifferent between the two alternatives if her forecast is that the 6-month rate six months from now will be 3.6%. The manager would prefer the 1-year Treasury bill if her forecast is that the 6-month rate six months from now will be less than 3.6%. The manager would prefer the 6-month Treasury bill if her forecast is that the 6-month rate six months from now will be greater than 3.6%.
As explained in the discussion of forward rates, it is common to refer to forward rates as the market’s consensus of future rates; however, from the manager’s perspective it is irrelevant whether the rate is truly a “consensus” value. The manager is only concerned with her view of future rates relative to the rates built into today’s prices. Consequently, forward rates can be viewed as “hedgeable rates.” That is, if a manager purchases the 1-year Treasury bill, she is hedging the 6-month Treasury bill rate six months from now. By doing so, the manager has locked in a 6-month Treasury bill rate six months from now of 3.6%.
Forward rates are just one example of the way in which a portfolio manager can use market information. We will see other examples in some of the chapters that follow.
D. Constructing the Portfolio
Given the manager’s forecasts and market-derived information, the manager then assembles the portfolio with specific issues. In active bond portfolio management, asset selection involves identifying opportunities to enhance return relative to the benchmark. In doing so, the manager determines the relative value of the securities that are candidates for purchase in the portfolio and candidates for sale from the portfolio.
According to Jack Malvey, in the bond market the term “relative value” refers to “the ranking of fixed-income investments by sectors, structures, issuers, and issues in terms of their expected performance during some future interval.”
260 The various methodologies for performing relative value analysis are explained in Chapters 18 and 20.
IV. MONITORING THE PORTFOLIO
Once the portfolio has been constructed, it must be monitored. Monitoring involves two activities. The first is to assess whether there have been changes in the market that might suggest that any of the key inputs used in constructing the portfolio may not be realized. The second task is to monitor the performance of the portfolio.
Monitoring the performance of a portfolio involves two phases. The first is performance measurement. This involves the calculation of the return realized by a manager over a specified time interval (the evaluation period). Given a performance measurement over some evaluation period, the second task is performance evaluation. This task is concerned with two issues. The first issue is to determine whether the manager added value by outperforming the established benchmark. The second issue is to determine how the manager achieved the observed return. The decomposition of the performance results to explain why those results were achieved is called return attribution analysis. Performance evaluation is described in Chapter 18.
V. ADJUSTING THE PORTFOLIO
Investment management is an ongoing process. The activities involved in monitoring the portfolio indicate whether adjustments need be made to the portfolio. By monitoring developments in the capital market, a manager determines whether to revise the inputs used in the portfolio construction process. Based on the new inputs, a manager then constructs a new portfolio. In constructing the new portfolio, the cost of trading issues currently in the portfolio is evaluated. These costs include transaction costs and any adverse tax or regulatory consequences.
Performance measurement indicates how well the manager is performing relative to the investment objectives. Performance is used by the client in deciding whether to retain a manager. However, a client must understand that the time horizon must be adequate to assess the performance of the manager and the strategy selected by the manager. For example, if analysis of the first quarter for a newly hired active manager indicates that he has underperformed the benchmark by 30 basis points, this may not be sufficient time for a fair evaluation. Instead, suppose that a new manager is hired and given cash to invest using a bond indexing strategy. If the same 30 basis point underperformance is observed for this manager, a quarter may be adequate time to assess the manager’s ability to index a portfolio. In fact, even if the manager of an indexed portfolio outperforms the benchmark by 30 basis points, this may be sufficient evidence to question the manager’s ability. An indexing strategy should not have a 30 basis point underperformance or outperformance even for a newly retained manager.