Norges Bank


Bonds with no government guarantee in the benchmark

The following letter was submitted to the Ministry of Finance on 15 March 2001

1. Background

The current benchmark indices for the Government Petroleum Fund's fixed income investments consist exclusively of government bonds in developed markets. The regulation and guidelines allow for investments in bonds that are not included in the benchmark index. Norges Bank may invest in bonds issued by both public and private bodies as long as the instruments have a rating higher than Baa/BBB1 from Moody's and Standard & Poors respectively. Bonds with this type of credit rating are often referred to as "investment grade" bonds or bonds with medium to low credit risk.

In this submission we will use the term "non-government-guaranteed bonds" to refer to all types of bonds that are not government bonds issued in the country's local currency and that are not included in the Petroleum Fund's benchmark index. This common term covers a number of different types of bonds, including corporate bonds. In the US, mortgaged-backed securities (MBSs) with marginal credit risk and a prepayment option represent another important segment of the investment universe. MBSs are not in the benchmark index. The term will also include bonds issued by international organisations, companies with a government guarantee and government bonds issued in another country's currency. Therefore, the term "non-government-guaranteed bonds" does not fully cover this universe. In this submission, the term "bond market" is defined as bonds with more than a certain minimum amount outstanding in the market. Issuers must have a credit rating of Baa/BBB or better from the large international rating companies. Bonds with only a small outstanding amount, private placements and bonds issued by enterprises with a rating lower than Baa/BBB will therefore be excluded.2

When the guidelines for the Government Petroleum Fund were expanded in the autumn of 1997 to include investments in equities, the existing benchmark for bonds (Salomon Smith Barney World Government Bond indices) remained unchanged. Norges Bank has stressed that it would not invest in corporate bonds until it had established satisfactory systems for assessing and measuring credit risk. The Ministry of Finance has endorsed this policy. Until now, it has been necessary to concentrate available resources on high priority tasks, such as developing an organisation that is capable of handling current investments and the management of large amounts in the international equity markets.

Expanding the Government Petroleum Fund's benchmark to include all types of bonds within the investment universe may be of interest for a number of reasons:

  • The benchmark will be more representative of the Fund's investment opportunities. Based on the Fund's universe, the current benchmark represents a good 27% of the investment universe in the US and a good 60% in Europe. In Japan, government bonds dominate the market.3
  • The expected long-term return on the benchmark will be somewhat higher if it is expanded. This is explained below.
  • Developments in the public sector's outstanding debt (not including Japan) indicate that government bonds will account for a steadily smaller share of bond markets in the future. The OECD estimates that gross public debt as a share of GDP will decline during the next few years throughout the OECD area, with Japan representing the exception. The chart below shows developments in gross outstanding debt in the public sector as a share of GDP from 1995 to 1999, with estimates for the period 2000-2002.4 In the OECD's reference scenario for the period 2002-2006, the share declines another 8% in the euro area and in the OECD area as a whole and by 18% in the US. Public debt will not contract as much as suggested in the OECD's reference scenario if the US tax cuts that are proposed for 2002-2006 are implemented.

2. Expected excess return and risk associated with investments in non-government-guaranteed bonds

The excess return from investing in non-government-guaranteed bonds relative to government bonds will be discussed in this section. One might ask whether equities should be discussed as an investment alternative in this connection. Although non-government-guaranteed bonds generate a somewhat higher expected return than government bonds, the expected excess return from investing in equities is considerably higher. Reference is made to the submission that discusses the consequences for expected return and risk if changes are made in the Government Petroleum Fund's equity allocation.

Constructing an optimal portfolio requires a decision about acceptable risk. Risk may be defined as the uncertainty of an estimate for expected future return or the probability that a return that will fall below a lower limit over a given period (shortfall risk). Risk depends on the volatility of the individual asset classes and the correlation between them. Shortfall risk also depends on expected return.

The correlation between government bonds and non-government-guaranteed bonds with middle to low credit risk (BBB or better) has traditionally been very high. In the US, the correlation coefficient between monthly returns on government bonds and other segments of the bond market during the period 1990-2000 was around 0.9. Changes in general interest rate levels affect prices on both government bonds and non-government-guaranteed bonds. Price changes depend primarily on the bond's interest rate sensitivity, or modified duration, and not on whether the bond has a government guarantee. The high correlation between two types of bonds indicates that general factors are decisive for total returns in both segments. The excess return, which depends on factors that only affect non-government-guaranteed bonds, accounts for a limited portion of the total return on these bonds. The correlation between bonds in general and equities is far lower but also varies greatly over time.

Assuming that the figures above are representative, the equity allocation alone should not significantly affect the optimal composition of the fixed income portfolio. It follows from this that changing the spread between government bonds and non-government-guaranteed bonds (within the "investment grade" range) will have minor impact on the portfolio's total risk. Changing the allocation between government bonds or private bonds and equities will have greater anticipated consequences for overall risk.

Non-government-guaranteed bonds will be priced so that the yield to maturity is higher than on government bonds with the same maturity. The yield differential may be seen as compensation to the investor for the extra risk involved in investing in an instrument with no government guarantee. Some of this is compensation for expected losses in the future. If the average investor were indifferent to risk as long as the expected return was the same, the yield differential between government bonds and non-government-guaranteed bonds would only have covered these expected losses. But since investors are normally risk averse, the risk premium inherent in the yield differential will be greater than this. The rest of the risk premium represents an expected excess return to the investor for investing in more risky, and somewhat less liquid bonds.

When investments are made in bonds with credit risk, there will be expected losses for two reasons:

a. Even in a universe consisting of bonds issued by companies with a high credit rating, some of the issuers must be expected to default in the future. In the event of insolvency, lenders will receive some of their claims from the estate, but some of the claims will be lost. In the universe comprised of corporate bonds with a BBB rating or better, the loss will be very limited. The annual average loss due to defaults on BBB-rated bonds in the US from 1970 to 1999 was 0.06%.5

b. Companies' credit ratings will change over time. Since a credit rating is the market's best estimate of actual credit risk, a downgrading will lead to a price fall on the company's bonds. Similarly, an upgrading will push prices up. Experience shows that within the universe of companies with BBB ratings or better and an average rating around AA/A, more companies will be downgraded than upgraded. When investors purchase a portfolio that reflects the universe and assume that they will maintain their investment in such a portfolio in the future, they must expect a loss in returns related to government bonds due to the uneven distribution of future changes in ratings.

The annual average excess return on corporate bonds in the US during the period 1989-976 was 0.5 percentage point. During the 1990s as a whole, excess return has been lower. During the period 1998-2000, corporate bonds have yielded lower returns than government bonds. The yield differential between government bonds and other bonds has widened due in part to the US government's substantial buy-back of government bonds. Viewed over a longer period, 1920-1999, long-term corporate bonds have generated 0.7 percentage point higher returns than comparable government bonds.7

In the other major segment in the US, outside government bonds, ie mortgage-backed securities, investors have credit risk exposure to three federal agencies. These agencies guarantee that individual homeowners will meet their loan obligations. One of the agencies, the Government National Mortgage Association (Ginnie Mae) provides an explicit government guarantee. The other two agencies are the Federal National Mortgage Association (Fannie Mae) and the Federal Home Loan Mortgage Corporation (Freddie Mac). These do not have explicit government guarantees. Due to these agencies' important and special role in meeting the political objectives of housing policy, political support is nevertheless assumed to be strong. Both agencies have a Aaa/AAA rating from Moody's and Standard & Poor's. Therefore, the agencies' credit rating must be regarded as very good. The yield differential between mortgage-backed securities and government bonds is therefore the result of factors other than compensation for credit risk.

The expected excess return on mortgage-backed securities compared with government bonds will be lower than the observed yield differential. The most important reason for this is homeowners' prepayment option. Homeowners must be expected to use this option to their advantage. When interest rates fall, they will apply to refinance their loans at a lower interest rate. Investors must then relinquish a current cash flow in cases where the running yield is higher than the market rate. This option has a value for homeowners. Compensation for this will be part of the yield differential. In addition, there will also be compensation for somewhat lower liquidity in the market.

When investors assess instruments of this kind, they are dependent on advanced models that can simulate future refinancing tendencies among homeowners. This is necessary in order to be able to price the implicit option and to estimate the bond's duration. It has been argued that existing models have underestimated refinancing tendencies and thus also overestimated the duration in a period with falling interest rate trends (second half of the 1990s). As a result, the excess return from investing in mortgage-backed securities has been lower than what the option-adjusted yield differential would indicate. During the period 1989-1997, a market-weighted portfolio of mortgage-backed securities has outperformed a government bond portfolio with the same (assumed) duration by about 0.3 percentage point. The excess return must be seen as a compensation for uncertainty (investors have a preference for secure cash flows compared with insecure cash flows, even if the actual option has been priced fairly) and for the extra administration connected with handling uncertain and incidental cash flows. If incorrect modelling of refinancing tendencies does not have the same negative effect in the future as during the period referred to above, the expected excess return will be higher than this.

The three large federal agencies in the US also issue ordinary bonds. This bond segment has outperformed government bonds during the period 1989-97 by about 0.3 percentage point.

In Europe and Japan, performance series for broad portfolios comprising non-government-guaranteed bonds are very short. In Japan, non-government-guaranteed bonds account for a very limited share of the bond market. Assuming that individual segment weights in each region do not deviate appreciably from market weights, differences in return between Japan's bond segments will have little influence on the benchmark portfolio's overall return. Non-government-guaranteed bonds issued by international organisations and other bodies with the highest credit ratings (AAA and AA) represent a larger portion of this market segment in Europe than in the US. The expected excess return on a market-weighted portfolio of non-government-guaranteed bonds will therefore be lower in Europe than in the US. In the long term, it is likely that the bond market in Europe will become a more important source of financing for European enterprises (outside the financial sector) than it has been. This will increase the expected return and improve the diversification properties of a market-weighted portfolio of non-government-guaranteed bonds in Europe.

The benchmark portfolios for the bond markets are changed once a month. New bonds are added and bonds with less than one year to maturity are removed. Country weights and asset class weights in Norges Bank's benchmark portfolio are changed on a quarterly basis in accordance with the prevailing rebalancing regime. When management of the actual portfolio results in only minor deviations from the benchmark portfolio, changes in the benchmark portfolio will lead to approximately the same changes in the actual portfolio. This results in transaction costs. Transaction costs connected with maintaining a broad bond index are higher than those connected with maintaining a government bond index. Whereas government bonds are generally quite liquid, with small differences between bid and offer prices, liquidity is more variable, particularly for corporate bonds.

Although transaction costs rise, it is unlikely that this will offset the excess return on the individual bond segments in question. Very high transaction costs must be expected, however, during a transition phase when a new benchmark index is phased in. Due to the size of the Government Petroleum Fund, it will be necessary to phase in any new benchmark over time.

3. Need for changes in guidelines

A. Requirement that there is extensive diversification of credit risk

The Ministry of Finance's guidelines on credit risk include an overriding rule which states that credit risk shall be limited by means of extensive diversification. Selecting a broad bond index as a benchmark should not create any problems in relation to exposure to private companies. The Government Petroleum Fund's share of an individual company's outstanding bond debt will depend on how the benchmark is defined and how closely the portfolio replicates the selected benchmark. Over time, growth in the Government Petroleum Fund and in the bond markets will determine the Fund's expected share of each company's debt.

If the Government Petroleum Fund had invested in a fixed income portfolio at the end of 2000 that was identical to the indices shown in Annex 1, the Fund would have owned 0.29% of the corporate bonds that comprise the European index and 0.12% of those that make up the US index.

Fannie Mae and Freddie Mac account for 75% of the bond issues from federal credit agencies. As mentioned, these agencies do not have explicit government guarantees. In its guidelines on credit risk, the Ministry of Finance has specified that credit risk shall be limited through extensive diversification. If the Ministry does not accept such a concentration, the alternative is to reduce the share of these agencies' mortgage-backed securities and ordinary bond issues in the benchmark index compared with full market weight. Both of these agencies have the highest possible rating (Aaa/AAA) from Moody's and S&P. The demand for diversification of credit risk is presumably most important for private companies which have somewhat lower credit ratings than these two agencies. Although these two agencies will account for a considerable share of the US portfolio, the exposure in the combined fixed income portfolio is limited. Norges Bank can therefore see no strong argument for reducing the share of these two agencies based on the Ministry of Finance's guidelines.

The table below presents Fannie Mae's and Freddie Mac's combined share of market value in the Government Petroleum Fund's fixed income portfolio and the US sub-portfolio if the share of mortgage-backed securities and ordinary bond issues from credit institutions is set at 100%, 75% and 50% of market weight respectively. Reducing these shares will lead to an increase in government bonds and corporate bonds in particular.

Share of mortgage-backed securities and ordinary bond issues in relation to market weights8

Share of Government Petroleum Fund's fixed income portfolio

Share of Government Petroleum Fund's investments in fixed income instruments in the US










B. Requirement that a maximum of 10% of the fixed income portfolio is invested in instruments with credit risk/maximum 5% is invested in instruments with the lowest acceptable credit rating (BBB).

The Ministry of Finance has decided that no more than 10% of the fixed income portfolio may be invested in securities with no government guarantee (measured both as a share of market value and as a share of the portfolio's duration). This rule must be seen in light of the fact that the share of such securities in the existing benchmark index is 0%, thus setting a limit for active risk. If the benchmark index is expanded, this rule should be changed so that the limit becomes +/- 10% in relation to the share in the benchmark. There is also a 5% limit for the share that may be invested in securities with the lowest acceptable credit rating (BBB). This rule should also be changed so that it applies in relation to the benchmark. Given the very low probability of payment problems, even in the BBB category, it is difficult to see the importance of limiting the absolute risk in this rating category.

C. Requirement that instruments must have an acceptable credit rating from both rating companies

The Ministry of Finance's guidelines state that investments may only be made in bonds that are classified as investment grade (does not apply to government bonds) by both S&P and Moody's. If only one of these agencies has assigned a rating to a bond, Norges Bank may purchase the bond after special evaluation. It must be established that the company that has not rated the instrument would not have given a rating below the minimum requirement. In the Ministry's guidelines, the lowest rating provides the basis for defining investment grade. About 3% of US corporate bonds have a split rating.9 A split rating occurs when one rating agency has given the company the lowest rating within investment grade while the other rating agency has given the company a lower rating, as a rule Ba/BB (one grade lower). Maintaining existing requirements is one possibility. However, changing the guidelines to permit the investment if one of the rating agencies has given a rating of BBB or better would be preferable based on operational considerations. This is due to the formulation of credit rating requirements in the most common bond indices that are meant to cover the entire bond market. Lehman Brothers, Salomon Smith Barney and Merrill Lynch publish these types of broad indices daily. They all aim at including only investment grade bonds, but the three index suppliers define "investment grade" differently. Salomon Smith Barney base their definition on the highest rating, whether it is provided by Moody's or S&P. Lehman Brothers uses Moody's rating and will only look at S&P's rating if Moody's has not given a rating. Merrill Lynch uses an average of the ratings but will attach greatest importance to the lowest rating. Although the approaches vary, all three index suppliers define more bonds as investment grade than is the case with the Government Petroleum Fund's existing guidelines. If requirements for inclusion in a benchmark index differ from requirements for inclusion in the most common bond indices published externally, the latter must be customised to be used as a benchmark for the Government Petroleum Fund. There is no problem involved in defining other country and regional weights in the benchmark as compared to the external indices. This is currently being done. However, having different rules about which bonds should be included in the actual index means more work. In Norges Bank's opinion, there are advantages to changing the credit rating requirement so that an investment grade rating from either Moody's or S&P is adequate. The increase in credit risk will be very small. In addition, such a change will improve the possibilities of diversifying the absolute credit risk since more companies from different industry segments will be included in the portfolio.

Rules already exist to cover Norges Bank's behaviour if an issuer that is represented in the portfolio is downgraded below the lowest acceptable rating. There is no need to change these rules if the benchmark is expanded.

D. Requirement about issuer's home country

In accordance with §6 in the Regulation on the Management of the Government Petroleum Fund, the issuer shall be registered in one of the 22 countries that are defined as permissible markets for investments in fixed income instruments or be an international organisation. According to the regulation, it would not be permissible to invest in a security denominated in US dollars issued by a country, organisation or company with a high credit rating, if the issuer is based in a country that is not included in the list of permissible countries. In Lehman Brothers' global bond index, which covers bond markets in all developed markets as well as South Korea, Taiwan and Malaysia, less than 1.6% of the bonds are issued by issuers that are not registered in one of the developed markets as at January 2001. Of these, more than 50% are denominated in dollars, euro, yen or other developed countries' currencies. The largest individual issuer in this group is the Mexican government, whose bonds (primarily in USD) account for 0.3% of Lehman's Global Aggregate Index.

Norges Bank sees no strong reasons for applying the country list limitations that pertain to the Fund's currency and market distribution to the portfolio's credit risk. Whereas the requirements regarding market size, stability, legislation and settlement systems are relevant for currency and market distribution, limitations on credit risk in the portfolio should be based on an assessment of each issuer's ability to pay. It should be assumed that the rating agencies have considered all known conditions relevant to giving a credit rating. Norges Bank will therefore recommend that the last sentence in the Regulation's §6 is annulled, so that it is permissible to invest in all fixed income instruments issued in the 22 allowed markets/currencies, assuming the issuer satisfies the credit rating requirement. Fixed income instruments issued in other currencies than these will still be excluded from the investment universe. Reference is also made to the previous discussion about the operational advantages of having the same requirement for inclusion of a fixed income instrument in the benchmark as that which applies for the most used externally published market indices.

At a later time, Norges Bank will provide an assessment about whether emerging bond markets should be included in the investment universe.

Annex 3 presents the percentage of issuers from emerging markets in the Lehman Global Aggregate Index as at January 2001. Government bonds issued by South Korea, Taiwan and Malaysia in the respective country's currencies are also included in this table. These bonds will not be included in an investment universe if the above recommendation is accepted.

4. Choice of benchmark

Before Norges Bank can begin managing non-government-guaranteed bonds on a large scale, a risk management system and effective settlement routines must be established for the new instruments. According to §9 in the Regulation on the Management of the Government Petroleum Fund, Norges Bank is required to ensure the existence of adequate risk management systems and control routines for the instruments under management.

The following work will be in progress in Norges Bank through 2001 to prepare for the management of non-government-guaranteed bonds:

  • An automated system will be established to measure credit exposure to borrowers.
  • The suitability of using BARRA and other risk models to measure market risk connected with different types of non-government-guaranteed bonds will be assessed. The acquisition of a new system for managing such instruments in a more satisfactory manner will be assessed.
  • Work on developing and testing routines for trading, settlement, accounting and measurement of returns on fixed income instruments with imbedded options has been initiated.

Norges Bank aims at phasing in a new benchmark from 1 February 2002. Whether preparations will have progressed so far that it will be possible to phase in all segments in a broad index at this time is still unclear. However, it should be possible to adapt the phasing-in profile to the complexity of the individual sub-index.

A benchmark index must be as representative as possible in relation to the universe it is intended to cover. In addition, the rules for defining the index should be transparent so that there is no suspicion that the benchmark index has been arbitrarily chosen. This makes a case for continuing to use an index that is provided and published in the market. Assuming that the representation requirement is met, operational considerations should also be taken into account. Different alternatives may vary as to ease of replication in actual management and ease of maintaining a good overview of index content at any given time. Norges Bank wishes to further assess which index best meets both representation requirements and operational considerations. If current guidelines for the investment universe are maintained, it will also be necessary to tailor individual country indices for the Government Petroleum Fund.

5. Conclusion and recommendations

Norges Bank recommends an expansion of the benchmark for the fixed income portion of the Government Petroleum Fund, with the planned start for phasing in a new benchmark as from 1 February 2002. This will provide a benchmark that is more representative of the Fund's investment universe than the current index. In the US in particular, large portions of the investment universe are omitted from the benchmark. In this region and in parts of Europe as well, non-government-guaranteed bonds will most probably account for an increasingly larger share of the bond market. An expanded benchmark will also have a somewhat higher expected excess return. At the same time, the correlation with government bonds is so high, that market risk will not increase significantly. An expanded benchmark will increase the portfolio's credit risk, but the increased expected excess return will compensate for this.

Norges Bank recommends that a new benchmark must be representative of the entire Government Petroleum Fund's investment universe, ie that all segments of the bond market within the investment universe are covered.

Since a purely market-weighted index in the US will result in a concentration towards two federal agencies without explicit government guarantees (Fannie Mae and Freddie Mac), the Ministry of Finance is requested to decide whether a market-weighted benchmark will satisfy the Ministry's requirement for extensive diversification of credit risk. These lending institutions have the highest possible credit ratings. The credit risk related to these two agencies is therefore very limited, although considerable amounts are invested in the bonds issued by them.

Norges Bank also recommends that the prevailing guidelines limiting credit risk in the fixed income portfolio should be changed from an absolute limit to a relative limit in relation to the benchmark index.

Norges Bank recommends that the last sentence in §6 of the Regulation on the Management of the Government Petroleum Fund be removed. The consequence of this is that it will be permissible to invest in all fixed income instruments that are issued in any of the 22 permitted markets, regardless of the issuer's country of registration.

If the Ministry of Finance approves the recommendations in this letter, Norges Bank will make a concrete recommendation regarding the benchmark prior to the presentation of the National Budget for 2002. In this letter, Norges Bank will also provide recommendations about the method for phasing in a new benchmark index.

Svein Gjedrem

Harald Bøhn


  • Presentation of the market for fixed income instruments in the US, Europe and Japan
  • Description of the three dominant issuers of mortgage-backed securities in the US
  • Table: Issuers from emerging markets as a share of the Lehman Global Aggregate Index, January 2001

Annex 1:

Presentation of the markets for fixed income instruments in the US, Europe and Japan

Non-government-guaranteed bonds represent a variable share of the bond market in the individual countries/regions. The charts below illustrate the composition of the markets in the US, Europe and Japan.10

Chart 1: US fixed income market - December 2000

Chart 2: Europe fixed income market - December 2000

Chart 3: Japan fixed income market - December 2000

It should be emphasised that the distribution between the different sectors depends greatly on minimum requirements for inclusion in the universe. In the indices used as a basis for these charts, the minimum requirement was low. If other indices with higher minimum requirements had been used, the sectors "corporates" and "other" in particular would have been reduced in favour of government bonds.

The sector "agencies" in the US represents ordinary bond issues from three large federal credit agencies that were established by separate legislation. These include the Government National Mortgage Association ("Ginnie Mae"), the Federal National Mortgage Association ("Fannie Mae") and the Federal Home Loan Mortgage Corporation ("Freddie Mac"). These agencies shall contribute to achieving the political objectives of housing policy. Ginnie Mae, in contrast to Fannie Mae and Freddie Mac, has explicit guarantees from the Federal Government for its mortgage-backed securities (MBS). Nevertheless, it is assumed that if problems should arise the federal authorities will support Fannie Mae and Freddie Mac in view of their important role in meeting political objectives. These securities have been represented in Norges Bank's portfolio for a long time.

The sector "public sector, international organisations" in the charts covering Europe and Japan comprises issues from companies with explicit or implicit government guarantees, regional or local authorities, foreign governments or international organisations. With the exception of instruments issued by regional and local authorities, these bond issues have been in Norges Bank's portfolio for a long time. These types of issues also account for a major portion of the "other" sector in the US. The remainder of the "other" sector in the US comprises asset backed securities.

All companies in the "corporates" sector in the three regions have credit ratings of Baa/BBB or better. In all three regions, the sector is dominated by companies with AA or A ratings, which account for 63% of companies in the US and 86% in Japan. The US is the only region with a greater share of companies with a BBB rating, more than 30%. Both in Europe and Japan, banks, insurance companies and other market participants within the financial sector are the largest bond issuers. Financial undertakings represent 50% of the Japanese market for corporate bonds and 61% of the European market. The American market is far more diversified. Here, corporate bonds issued by different industrial undertakings representing many industrial sectors account for 53% of the corporate bond market.

"Mortgage-backed securities" represent the largest individual segment in the US bond market and are expected to take over an even larger share in the future. MBSs are individual housing loans which the three major agencies Ginnie Mae, Fannie Mae and Freddie Mac buy in the secondary market, pool together and offer to investors. The agencies guarantee the loans. The investors' credit risk lies therefore with the issuers and not the homeowners. When individuals raise a loan, a pay-back plan and a fixed coupon rate on the loan are established, but the homeowners have the right to repay the entire loan or part of it more quickly than stipulated by the repayment plan. The three agencies will pass this capital directly through to the investors. In other words, the investors have written an option to the homeowners and this must be taken into consideration when these instruments' interest-rate risk is modelled.

"Collateralised securities" in Europe consist primarily of loans, secured by the buildings ("Pfandbriefe"), that are provided by different financial institutions with very high credit ratings for special public endeavours. Most institutions that provide this type of loan have a AAA or AA rating. Unlike the US mortgage-backed securities, there is no uncertainty, other than the marginal credit risk, about the timing of cash flows.

Annex 2:

Important issuers of mortgage-backed securities

Government National Mortgage Association (Ginnie Mae)

The US Congress established Ginnie Mae in 1968 as a new, separate agency under the Department of Housing and Urban Development. Ginnie Mae was separated from the Federal National Mortgage Association. One of the agency's aims is to finance federal housing projects. The housing loans on which Ginnie Mae is based are the only ones directly guaranteed by the US Federal Government, in other words, insured or guaranteed by the Federal Housing Administration, the Veterans' Administration or the Farmers' Home Administration. In addition, there is an upper limit on loans that are covered by the rules. All loans must be new and have a term of 15 or 30 years.

Federal National Mortgage Association (Fannie Mae)

First established by Congress as a government agency in 1938, Fannie Mae became a private corporation in 1968. Fannie Mae's stocks are listed on the New York Stock Exchange and are also part of Standard & Poor's 500 index. Fannie Mae, which is regulated by the Department of Housing and Urban Development, shall facilitate the channelling of capital from regions with excess capital to regions where there is demand for housing loans. Like Ginnie Mae, Fannie Mae may purchase mortgages that have been insured or guaranteed by the Federal Housing Administration, the Veterans' Administration or the Farmers' Home Administration. In addition, however, Fannie Mae may purchase other mortgages that are insured by the private sector. Otherwise, Fannie Mae is covered by rules similar to those that apply to Ginnie Mae, although more housing loans are covered by the rules at Fannie Mae than at Ginnie Mae.

Federal Home Loan Mortgage Corporation (Freddie Mac)

From the time of its establishment in 1970, Freddie Mac was owned by federal district banks, but following a restructuring in 1989, Freddie Mac was privatised and currently has the same status and function as Fannie Mae. The rules are currently the same for Fannie Mae and Freddie Mac.

Annex 3: Share of bond issuers from emerging markets in Lehman Global Aggregate Index, January 2001


The Americas



0.274 %

Cayman Islands

0.023 %


0.022 %


0.022 %


0.005 %





0.032 %


0.026 %


0.010 %


0.007 %


0.005 %

Czech Republic

0.003 %




South Korea

0.438 %


0.342 %


0.227 %


0.038 %


0.010 %


Africa and Middle East



0.021 %


0.030 %

South Africa

0.026 %


0.002 %


1)Moody's credit ratings rage from C (worst) to Aaa (best). S&P has a similar scale from C to AAA. Bonds with a rating of Ba/BB or worse are referred to as "speculative grade".

2)The different bond markets are described in greater detail in this submission's annexes. Three US issuers of mortgaged-backed securities are also presented, because an expansion of the benchmark in the US may imply higher exposure for American fixed income securities to these issuers.

3)The investment universe is defined here as bonds that are included in the Lehman Aggregate Index in the US and the European and Japanese part of Lehman Global Aggregate Index.

4)Source: OECD Economic Outlook, December 2000

5)Source: Moody's Investor Service: "Historical default rates of Corporate Bond Issuers, 1920-1999".

6)Source: Lehman Brothers

7)Source: Ibbotson Associates

8)In relation to the share in the Lehman Aggregate Index, US, December 2000

9)Source: Salomon Smith Barney Broad Investment Grade Index Profile, March 2001

10)The bond market is defined in the charts as all bonds that are included in Lehman Brother's broad bond indices in the individual regions. In the US, Lehman's Aggregate is the dominating benchmark index for institutional investors who invest

Published 15 March 2001 11:17