Norges Bank

Submission

Choice of a new benchmark for the Government Petroleum Fund's fixed income investments

The following letter was submitted to the Ministry of Finance on 23 August 2001

1. Background

In a letter to the Ministry of Finance dated 15 March 2001, Norges Bank recommended that the benchmark for the Government Petroleum Fund's fixed income investments should be expanded to include bonds with no government guarantee. In addition, the Bank recommended that the guidelines for the management of the Government Petroleum Fund be amended to ensure better adaptation to an expanded benchmark. Norges Bank proposed that the prevailing limits for investment in non-government-guaranteed bonds be changed from an absolute limit to a relative limit in relation to the portion of such bonds in a new benchmark. Furthermore, Norges Bank recommended that the Fund should be permitted to invest in fixed income instruments from all issuers satisfying the credit rating requirements and issued in any of the 22 accepted markets, irrespective of the issuer's home country.

In the Revised National Budget 2001, the Ministry of Finance endorsed the proposal to phase in an expanded benchmark from February 2002. The Ministry of Finance wished to consider the question of amendments to the guidelines in advance of the National Budget for 2002. The Ministry also wanted to assess whether the portion of bonds issued by two federal agencies in a market-weighted portfolio of fixed income securities in the US satisfied the Ministry's requirement regarding diversification of credit risk in the portfolio.

In this letter, Norges Bank will submit a recommendation regarding the choice of benchmark and strategy for phasing in a new benchmark. The proposals in the letter are contingent on the implementation of the proposed amendments to the guidelines. The proposal may, however, be adapted without difficulty to a reduction in the portion of bonds issued by US federal agencies, should the Ministry so require.

2. Current benchmark and alternative broader indices

Composition

The current benchmark for the Government Petroleum Fund's fixed income investments is based on Schroder Salomon Smith Barney's (hereafter Salomon) World Government Bond Index. This is a market-weighted index of government bonds with more than one year to maturity. In the Government Petroleum Fund's benchmark, indices for each country are weighted with the GDP weightings in each region, and the fixed weightings for each region (50% Europe, 30% Americas, 20% Asia). Between the quarterly rebalancing operations, market movements will change the country weightings in the benchmark continuously.

In order to ensure optimal transparency, one of the leading market indices should still be used as a basis for constructing the new Petroleum Fund benchmark. In recent years, several investment banks have introduced bond indices that aim to cover the entire spectrum of bonds with a credit rating of BBB or better (investment grade). Competing indices that are published by Lehman Brothers, Salomon and Merrill Lynch all cover a global market.

All three global index suppliers mentioned above have published broad indices for the US market for some time.1 The methodology of these indices has been used in the construction of the global indices.

Lehman Brothers has decided that for a bond issue to be included in the global index (Lehman Global Aggregate - LGA) the outstanding amounts in the various markets must be the equivalent of USD 300 million (Americas) / EUR 300 million (Europe) / JPY 35 billion (Asia). In addition to government bonds in developed markets, non-government-guaranteed bonds issued in the US, most developed European markets, Japan and Australia are also included in the index. Non-government-guaranteed bonds issued in the Canadian market may be included in the index in 2002. Lehman also publishes data for a bond index with a higher inclusion requirement.

When new bonds with credit risk are included in Lehman's indices, they are first priced at the offer price in the market, and then at bid price. Cash flows, e.g. from coupon payments, are kept as cash in the index until the start of the following month, and no interest is added in this period. At the start of each month, the index is rebalanced in accordance with existing rules.

Salomon has set more stringent inclusion requirements for its global index (World Broad Investment Grade Index - WBIG) than Lehman has for its global index. Consequently, the number of corporate bonds included in this index is lower than in the Lehman index. In addition to government bonds in developed markets, Salomon's index includes other bond segments in the US, the euro area, the UK and Japan.

Salomon prices bonds in the index at bid price, also when new bond issues are included in the index. In contrast to the practice at Lehman Brothers, any coupon payments and other cash flows that accrue in the index during the month receive interest at money market rates until the index is rebalanced at the start of the following month.

Merrill Lynch publishes data for a broad index with a low inclusion requirement for outstanding amounts (Broad Market Index- BMI) and a more liquid index (Large Cap Index - LCI) with a requirement that is more or less on a par with Salomon's WBIG. The broadest index includes both government bonds and non-government-guaranteed bonds issued in all developed markets.

Like Salomon, Merrill Lynch prices indices at the market bid price. The indices are rebalanced on a monthly basis and interest is paid on cash flows in the course of the month at money market rates until the index is rebalanced the following month.

A separate annex presents an overview of the market value and number of issues included in each index on a regional level at the end of July this year, for the market as a whole and for corporate bonds in particular.

All the index suppliers calculate a daily return on the index as a whole and for some sub-indices. Statistics are also published for market capitalisation, average yield, average duration, number of issues, etc., for the index as a whole and for the separate sub-indices that comprise the main index. For example, detailed return figures and statistics for each key sector in the market are published for each region (North America, Europe and Asia/Oceania):

  • Domestic government bonds
  • Other government-guaranteed bonds
  • Bonds with credit risk (corporate bonds and other issuers, e.g. international organisations)
  • Collateralized securities

Separate return figures are also published for bonds with various rating levels.

Norges Bank will, in any case, receive information about the benchmark, down to the individual bonds, and will therefore be able to adapt reports to the Ministry of Finance's information requirements, irrespective of which index supplier is chosen.

Yield differential between indices

Different minimum requirements for individual bonds will give rise to yield differentials between different indices on a regional level. Only in the US have bond indices that cover a broad market existed for any length of time. Changes in the yield differential between government bonds and other bonds have been the most important cause of return differentials observed in recent years. An index with a high minimum requirement for each issue (large cap) will have a higher share of government bonds than an index that includes a number of small issues, and has therefore recorded somewhat higher returns in recent years. In the period January 1997 to May 2001, the yield differential between market-capitalised indices with different minimum requirements for outstanding amounts has been modest. The yield differential between government bonds and other bonds in the US increased in the period 1998-2000, but has fallen slightly again since the start of the year. At end-2000, the yield differential between a representative large cap index and an index with a lower inclusion requirement2 was 68 basis points, equivalent to an annual yield differential of 17 basis points. At end-July 2001, the cumulative yield differential for the entire period was 26 basis points, equivalent to an annual yield differential of 6 basis points.

There is a high correlation between the different indices. From a purely strategic perspective, therefore, which of the broad global indices constructed by the different investment banks is chosen is of little importance. It is assumed that the choice of one particular index rather than another will have little effect on actual portfolio management. Irrespective, the replication of the part of the index that comprises non-government-guaranteed bonds will entail the purchase of a basket of bonds that are deemed to be most representative of the index as a whole. Nor will the performance measurement considerations provide a basis for recommending a particular index supplier. If reasonable notice is given before a new benchmark is phased in, it will be possible to measure management performance against any of the indices mentioned.

Since the differences between the indices are small, Norges Bank will consider it an advantage to have an index that represents the investment opportunities. One objection to indices with high inclusion requirements is that the market for corporate bonds, in particular, will not be adequately represented. For example, in Japan, a broad index with a high inclusion requirement would mean that the share of corporate bonds was very limited.

This points in favour of using the Lehman Global Aggregate as the new benchmark. Liquidity considerations have received some emphasis in this recommendation. As mentioned earlier, the Lehman Global Aggregate does not add interest on intra-month cash flows until the start of the following month. Depending on the average coupon rate in the portfolio and the return in the money market, this corresponds to a hidden excess return of approximately 1 basis point (0.01 percentage point) per year when indexing the portfolio. On the other hand, managers will have a disadvantage, when rebalancing the portfolio, because they must buy bonds at offer price in the market, while the bonds are priced at bid price in the index. In the Government Petroleum Fund's Annual Report for 1999, indexing costs were estimated at 2-6 basis points with the existing benchmark. This cost will increase with a new benchmark.

For the time being, Lehman Brothers has currently chosen to exclude the Swiss bond market from Lehman Global Aggregate. However, the investment bank publishes a separate index for this market, the Lehman Swiss Aggregate. As Swiss government bonds are already included in the existing benchmark, Norges Bank recommends that this should continue in the new benchmark. This may be achieved by supplementing Lehman Global Aggregate in Europe with the part of Lehman Swiss Aggregate that comprises domestic Swiss government bonds.

Country weightings in the benchmark

In the current benchmark, country weightings are determined in each region by the existing GDP weightings, which are updated once a year. GDP weightings were chosen, because pure market capitalisation weightings could mean that the Fund was increasingly exposed to countries with rising government debt (in relative terms).

GDP weighting for domestic government debt in each region may continue, even if a broader bond index is introduced. The allocation between domestic government bonds and non-government-guaranteed bonds in the benchmark will then be determined by the relative market weightings of these two segments in each region. The allocation between each country's domestic government bonds may, however, be determined by intra-regional GDP weightings. Norges Bank would propose, however, that when non-government-guaranteed bonds are included in the benchmark portfolio, country weightings should be determined by market capitalisation weightings alone. There are several reasons for this.

The share invested in each country's domestic government bond market is already limited by the fixed weightings for each region. This applies in particular in the Americas and Asia where investment is largely concentrated in one country (the US and Japan, respectively). In Europe, the fiscal obligations assumed by EMU member states through the Stability Pact will limit each member state's possibility of maintaining substantial government deficits. Government bonds issued in EMU member states account for 38% of the current benchmark for bonds. In isolation, the inclusion of non-government-guaranteed bonds will limit investment in domestic government bond markets in all regions.

In the period January 1999 to June 2001, a market-weighted government bond portfolio in each region, with existing regional weightings, would have achieved an average annual return that was 0.31 percentage point lower than the actual return with GDP weightings. The primary reason for this is that the UK has a higher weighting within Europe with GDP weightings than with market weightings, and the return on British government bonds was considerably higher than on other European government bonds in 1999. Over time, however, there is no reason to expect substantial differences in returns if market capitalisation weightings or GDP weightings are used in the regions.

One argument for using market capitalisation weightings is that it could reduce rebalancing costs, since country weightings will automatically be adjusted to relative market developments in the countries.

Based on the above discussion, Norges Bank recommends that market capitalisation weightings be used in each region.

The table in Annex 2 shows each country's domestic government debt as a share of the current benchmark (government bonds only) and as a share of broad bond indices with different minimum requirements for the size of outstanding debt. The tables in the annexes to this letter show the weightings for issuers from each country (irrespective of the currency in which the bonds are issued) for alternative indices.

3. Consequences for risk control and performance measurement

In a letter to the Ministry of Finance dated 15 March 2001, Norges Bank submitted a short presentation of the preparations to be undertaken by the Bank in 2001 to ensure that the Bank's systems for risk and performance measurement could deal with the management of a broad bond universe.

Before starting to manage non-government-guaranteed bonds on a large scale, it is necessary to establish systems which will ensure that the credit ratings of bonds and the credit risk in the portfolio as a whole are in compliance with the limits set by the Ministry of Finance. Norges Bank will ensure this by means of a credit monitor that collects updated credit information from the investment management's central database and sorts the portfolios in accordance with requirements for credit ratings and duration limits. A satisfactory credit monitoring system will be established by the end of the year.

The guidelines from the Ministry of Finance specify that the Government Petroleum Fund should be managed with an expected tracking error of less than 150 basis points (1.5 percentage points) against the Fund's benchmark, calculated using BARRA's analytical models. BARRA's analytical models cover the instruments included in a broad bond index. The risk measurement and reporting requirements stipulated in the guidelines are therefore satisfied.

When the market risk on non-government-guaranteed bonds is modelled in BARRA, calculations are based on the simplified assumption that changes in the yield differential between any non-government-guaranteed bond and the government bond yield curve can be fully explained by changes in the swap spread (yield differential between government bonds and the fixed interest rate offered by a bank with a high credit rating). Such an analysis is not suitable for bonds where changes in the yield differential are primarily ascribable to options inherent in the bond, for example, US mortgage-backed securities. Norges Bank will supplement BARRA's aggregate risk model with other analyses in the management of these instruments. Norges Bank continually assesses alternatives to the current risk measurement system that could more effectively model the interest rate risk involved in investment in non-government-guaranteed bonds.

In accordance with the authorisation specified in the Ministry of Finance's guidelines on credit risk, Norges Bank has set more detailed limits for exposure to banks. The rules currently apply to time deposits, exposure to custodian banks and margins furnished as a security for derivative trading. Bonds issued by banks comprise an important part of the credit segment in Lehman Global Aggregate, particularly in Europe and Asia. Norges Bank will assess the need to change the limits for bank risk on the basis of an expansion of the benchmark.

4. Phasing in a new benchmark

Liquidity varies considerably within the universe of bonds with BBB-rating or better. It will not be possible to make large transactions in less liquid bonds without substantially impacting the price of the securities to our own disadvantage. Given the size of the Government Petroleum Fund, this generates considerable challenges when phasing in a new benchmark. The choice of method and timing for phasing in a new benchmark will affect expected transaction costs.

Liquidity in each market is normally evaluated on the basis of the size of each bond segment and each issue. The larger an individual bond issue is, the more likely it is that individual blocks will be traded in the market at all times. New bond issues must also be expected to be more liquid than those that have been in the market longer.

Chart 1 shows average daily turnover in the different sectors of the US bond market in 2000.3

Domestic government bonds are clearly more liquid than other sectors, in particular compared with corporate bonds . Whereas turnover in the above chart is concentrated on just over 120 individual government bonds, turnover in corporate bonds will be spread over a considerably higher number of individual issues. The difference in average turnover per issue in each sector is therefore substantially larger than the difference in total turnover per sector.

Statistics for the turnover in bonds in European markets through Euroclear in 2000 show that, on average, the volume of corporate bonds was a good 11 per cent of the volume of government bonds. This may be a further indication of expected differences in liquidity between government bonds and non-government-guaranteed bonds.

Differences in liquidity are reflected in varying differences between the bid and offer price for bonds. Whereas this bid-offer spread typically constitutes 0.5-2 basis points in the yield for government bonds, it can be as much as 10-15 basis points for corporate bonds.4All in all, Norges Bank estimates that transaction costs in connection with phasing in a new benchmark that also includes non-government-guaranteed bonds will amount to 25-30 basis points (0.25-0.30 per cent) of the value of the non-government-guaranteed bond portfolio, or 12-17 basis points of the value of the entire fixed-income portfolio. This estimate does not take into account any negative impact that Norges Bank's transactions may have on the market.

It will be possible to a certain extent to achieve exposure in the new segments by participating in tenders, where transaction costs are lower than in the secondary market. In order to establish a broadly diversified portfolio with limited tracking error against the desired benchmark, it will still be necessary to purchase bonds in the ordinary market. Liquidity varies in the secondary market for non-government-guaranteed bonds. Less frequent trading in older issues and smaller bond issues in particular must be expected. In order to limit transaction costs, it is therefore important to establish a phasing-in strategy that provides the manager with sufficient flexibility to choose when to buy, depending on actual market opportunities. Norges Bank will have a disadvantage in that the market will know that the Bank regularly buys non-government-guaranteed bonds. This disadvantage could be reduced, if the phasing-in strategy was such that Norges Bank was only expected to be a minor participant in the market each day.

A new benchmark does not need to be phased in through the purchase and sale of bonds. Norges Bank will supplement transactions in the bond market with the purchase and sale of derivatives, within the specified limits. However, the use of derivatives does not reduce the need for a long and flexible phasing-in period.

Because liquidity varies in parts of the expanded investment universe, flexible transition rules are needed. The transition from the old to the new benchmark must be gradual, over a long period of time. Guidelines stipulating that larger transactions must be made on specific dates must be avoided. Norges Bank therefore proposes a longer transition period so that the new benchmark can be phased in gradually. At the same time, mechanisms to ensure that purchases of new bonds can be adapted to liquidity in the market and to limit the size of each transaction must be established. This will reduce the likelihood of negative effects in the market.

The Petroleum buffer portfolio, which is a part of Norges Bank's foreign exchange reserves, was established to contribute to the efficient phasing in of new capital to the Government Petroleum Fund. It would be natural to use the buffer portfolio to facilitate the transition to a new benchmark for the Fund's fixed income portfolio. The supply of new capital to the Fund via the Petroleum buffer portfolio is, however, not sufficient to ensure a smooth transition to a new benchmark. First, a phasing-in process that only uses capital from this source would take several years, even if 100 per cent of the new capital were used to purchase non-government-guaranteed bonds. Despite expectations of a large net supply of new capital to the Government Petroleum Fund, net sales of government bonds in both the US and Europe would be necessary to avoid an extended phasing-in period. Second, how much of this capital can be used to purchase non-government-guaranteed bonds is uncertain. If the stock market falls in relation to bond markets, funds from the Petroleum buffer portfolio will largely be used to purchase equities when the Government Petroleum Fund's benchmark is rebalanced. The Ministry of Finance has also announced that the equity share in the Government Petroleum Fund will be assessed in connection with the presentation of the National Budget for 2002. If the equity share is increased in 2002, a substantial portion of the capital transferred to the Government Petroleum Fund may be used to purchase equities. This further increases the need to supplement transactions in the Petroleum buffer portfolio with shifts from government bonds to non-government-guaranteed bonds in the actual Petroleum Fund, in order to adapt the portfolio to a new benchmark.

We presume that the Ministry of Finance will establish a plan for how quickly non-government-guaranteed bonds will be phased in to the Government Petroleum Fund, and that this will be accomplished by increasing over time the weightings of these bonds in the Fund's benchmark. Since there is considerable uncertainty about market liquidity and the amount that can be phased in reasonably cost-effectively, the Ministry of Finance and Norges Bank should maintain a close dialogue about the experience gained. The plan should be adjusted regularly in accordance with this experience.

The phasing in of non-government-guaranteed bonds poses considerable challenges for Norges Bank in terms of operational management. In order to achieve cost-effective transactions, such bonds must often be purchased at a time that differs from that stipulated for the benchmark, and the selection of bonds may also deviate. This generates an active risk (tracking error) which Norges Bank is prepared to take, see earlier discussions between Norges Bank and the Ministry of Finance regarding the background and limits for tracking error that the Bank can take in the operational management of the Government Petroleum Fund.5

Transaction costs in connection with the phasing-in of non-government-guaranteed bonds will have a negative effect on the Government Petroleum Fund's return, see earlier description of bid-offer spread. The same occurs when Norges Bank invests new capital in the stock market, see the Government Petroleum Fund's Annual Report for 2000. Norges Bank presumes that the Ministry of Finance will take such considerations into account when outperformance or under-performance is used as the basis for assessing the Bank's management performance. The Bank will give the Ministry a detailed description of returns and costs in each sub-portfolio in order to provide a sound basis for assessment.

5. Amendments to the guidelines

Norges Bank refers to the previously proposed amendments to the guidelines for the Government Petroleum Fund mentioned above.

In addition to these proposals, Norges Bank will recommend that a limited portion of the fixed income portfolio be invested in bonds with a Ba/BB rating (the highest rating in the speculative grade segment, which includes all bonds with Ba/BB rating or lower). The reason for this is a desire for flexibility in the management of fixed income instruments in connection with up or downgrading to/from the lowest acceptable credit rating. Experience has shown that when bonds are upgraded or downgraded to/from BBB, there can be substantial changes in value in the period immediately before or just after the change. Flexible rules will reduce the risk of under-performance in relation to the benchmark as a result of transactions that must be made at unfavourable times.

Bonds with BB ratings account for 1.2 per cent of the world's total bond market.6 Data for the US corporate bond market in the period 1980-1999 show that just over 4.6 per cent of bonds with BB rating were upgraded to BBB or higher in a single year.7In the same period, an average of 8.1 per cent of bonds with BBB ratings was downgraded to BB or lower each year.

Based on the above figures, a limit stipulating that up to 0.5 per cent of the fixed income portfolio's total value and duration may be invested in bonds with the highest rating of Ba/BB from Moody's and S&P would be acceptable.

In accordance with the Ministry of Finance's guidelines for credit risk, responsibility has been delegated to Norges Bank to establish more detailed rules for managing this type of risk. The Central Bank Governor has, among other things, used this authority to lay down rules for the time allowed for liquidating holdings of securities that have been downgraded below the Ministry of Finance's minimum requirement. Based on the experience of managing an expanded investment universe, the Central Bank Governor will assess whether these rules need to be changed.

6. Recommendations

New benchmark

Norges Bank recommends that after full phasing in, the weightings for the new benchmark index, should be:

  • 30% Lehman Global Aggregate US + Canada.
  • 50% Lehman Global Aggregate Europe, excluding NOK + domestic government bonds in the Lehman Swiss Aggregate
  • 20% Lehman Global Aggregate Asia/Oceania, excluding Malaysian ringgit, Korean won, Taiwan dollar and Thai baht.

The use of market-weighted benchmark indices in these three regions is recommended. If the Ministry of Finance chooses to reduce the portion of bonds issued by federal agencies in the portfolio, Norges Bank recommends that the weighting for the sub-index for collateralized securities8in the US be reduced by the desired ratio, and that the market weightings for the sub-indices for domestic government bonds and bonds with credit risk be increased correspondingly.

Changes to the guidelines

Norges Bank refers to earlier proposals submitted in the letter dated 15 March 2001. In addition, Norges Bank recommends that the Ministry of Finance allows up to 0.5 per cent of the fixed income portfolio's market value and duration to be invested in bonds with Ba/BB as the highest rating.

Phasing in a new benchmark

Norges Bank proposes therefore the establishment of a long transition period during which the new benchmark is phased in gradually.

 

ANNEX 1
Key figures for various bond indices that cover the whole investment grade segment

 

Lehman

Salomon

Merrill Lynch

North America

Global

9

Global

10

Broad

11

Large cap

12

Market value 31.7 (USD bn)

6121.2

5796.3

6933.8

5941.3

Total no. of issues in index

2985

1863

6645

1996

Share of corporate bonds in index

17.7%

15.0%

21.9%

12.8%

Lower limit for outstanding amount

$300m

$500m, CAD 500m

$150m

$500m

No. of corporate bonds in index

1635

912

4013

800

Europe

     

Market value 31.7 (USD bn)

3998.1

3702.1

4409.2

3764.8

Total no. of issues in index

1937

1359

5306

1516

Share of corporate bonds in index

12.2%

9.6%

17.2%

13.5%

Lower limit for outstanding amount

EUR 300m

EUR 500m, £300m

EUR 100m, £100m

EUR 500m, £250m

No. of corporate bonds in index

737

442

1736

661

Asia/Oceania

       

Market value 31.7 (USD bn)

2952.1

1768.4

2265.4

2013.2

Total no. of issues in index

1066

229

1589

671

Share of corporate bonds in index

11.4%

0.8%

20.5%

12.5%

Lower limit for outstanding amount

JPY 35bn

JPY 50bn

JPY 20bn

AUD 100m

JPY 60bn AUD 500m

No. of corporate bonds in index

545

25

1082

310

 

ANNEX 2
Domestic government debt as a share of various bond indices

Country

GDP weightings with fixed regional weightings

30-20-50

Domestic gov't debt as a share of the Large Cap Index13, with fixed regional weightings (30-20-50)

Domestic gov't debt as a share of the broad index14, with fixed regional weightings (30-20-50)

US

28.1%

8.1%

7.6%

Canada

1.9%

0.9%

0.8%

       

Japan

18.3%

19.1%

16.2%

Australia

1.7%

0.3%

0.3%

       

Germany

12.1%

6.9%

6.4%

Italy

6.7%

6.7%

6.5%

France

8.2%

6.3%

5.9%

UK

8.2%

4.1%

3.8%

Spain

3.4%

2.7%

2.6%

Belgium

1.4%

2.2%

2.0%

Netherlands

2.3%

1.8%

1.7%

Austria

1.2%

1.0%

1.0%

Denmark

1.0%

0.9%

0.8%

Greece

0.7%

0.9%

0.8%

Sweden

1.4%

0.6%

0.7%

Portugal

0.7%

0.5%

0.4%

Switzerland

1.5%

0.5%

0.0%

Finland

0.7%

0.4%

0.5%

Ireland

0.5%

0.2%

0.2%

 

ANNEX 3
Country weightings (issuer's home country) based on various weighting principles

3.1 Government bond indices (based on: Salomon World Government Bond Index)

Country

Global market-weighted gov't bond index

Global GDP weightings

Regional market-weighted (30-20-50) gov't bond index

GDP weightings with fixed regional weightings 30-20-50

US

25.8%

39.7%

27.1%

28.0%

Canada

2.8%

2.8%

2.9%

2.0%

         

Japan

27.7%

18.5%

19.7%

18.3%

Australia

0.5%

1.7%

0.4%

1.7%

         

Germany

8.3%

9.0%

9.6%

12.1%

Italy

8.2%

5.0%

9.4%

6.7%

France

7.7%

6.1%

8.9%

8.2%

UK

5.0%

6.2%

5.8%

8.3%

Spain

3.3%

2.5%

3.8%

3.4%

Belgium

2.6%

1.1%

3.0%

1.5%

Netherlands

2.2%

1.7%

2.5%

2.3%

Austria

1.2%

0.9%

1.4%

1.2%

Denmark

1.1%

0.7%

1.3%

0.9%

Greece

1.1%

0.5%

1.3%

0.7%

Sweden

0.8%

1.0%

0.9%

1.3%

Portugal

0.6%

0.5%

0.7%

0.7%

Switzerland

0.5%

1.1%

0.6%

1.5%

Finland

0.5%

0.6%

0.6%

0.8%

Ireland

0.3%

0.4%

0.3%

0.5%

         

 

3.2 Large Cap Index (based on: Salomon World BIG)

Country

Country weightings in global market-weighted Large Cap Index

Country weightings in market-weighted Large Cap Index, with fixed regional weightings (30-20-50)

Domestic gov't debt as a share of Large Cap Index, with regional fixed weightings (30-20-50)

US

46.7%

28.0%

8.1%

Canada

2.0%

1.2%

0.9%

       

Japan

15.3%

19.4%

19.1%

Australia

0.3%

0.4%

0.3%

       

Germany

9.4%

12.9%

6.9%

Italy

5.0%

7.2%

6.7%

France

5.6%

8.3%

6.3%

UK

3.8%

5.5%

4.1%

Spain

2.1%

3.0%

2.7%

Belgium

1.5%

2.2%

2.2%

Netherlands

1.2%

3.7%

1.8%

Austria

0.8%

1.1%

1.0%

Denmark

0.7%

1.0%

0.9%

Greece

0.6%

0.9%

0.9%

Sweden

0.6%

0.8%

0.6%

Portugal

0.4%

0.5%

0.5%

Switzerland

0.3%

0.4%

0.5%

Finland

0.3%

0.5%

0.4%

Ireland

0.2%

0.3%

0.2%

       

Other Asia

0.1%

0.2%

 

Other Europe

0.0%

0.4%

 

Other America

0.2%

0.5%

 

Africa

0.0%

0.0%

International

1.8%

1.8%

 

 

3.3 Broad index (based on: Lehman Global Aggregate)

Country

Country weightings in global market-weighted broad index

Country weightings in market-weighted broad index, with fixed regional weightings (30-20-50)

Domestic gov't debt as share of a broad index with fixed regional weightings

(30-20-50)

US

42.5%

28.6%

7.6%

Canada

1.9%

1.3%

0.8%

       

Japan

21.3%

19.5%

16.2%

Australia

0.4%

0.3%

0.3%

       

Germany

8.6%

13.8%

6.4%

Italy

4.4%

7.1%

6.5%

France

5.0%

8.2%

5.9%

UK

3.4%

5.2%

3.8%

Spain

1.9%

2.9%

2.6%

Belgium

1.2%

2.0%

2.0%

Netherlands

1.9%

2.8%

1.7%

Austria

0.8%

1.1%

1.0%

Denmark

0.5%

0.8%

0.8%

Greece

0.6%

0.9%

0.8%

Sweden

0.8%

1.2%

0.7%

Portugal

0.3%

0.5%

0.4%

Switzerland

0.1%

0.1%

Finland

0.4%

0.5%

0.5%

Ireland

0.1%

0.2%

0.2%

       

Other Asia

0.5%

0.3%

 

Other Europe

0.4%

0.4%

 

Other America

0.4%

0.3%

 

Africa

0.0%

0.0%

 

International

1.7%

1.7%

 

Footnotes:
1 Lehman Aggregate, Schroder Salomon Smith Barney Broad Investment Grade (BIG), Salomon Large Pension Fund Index, and Merrill Lynch Broad Market Index are examples of market capitalised indices that cover the US market
2 Merrill Lynch Large Cap Index and Merrill Lynch Broad Market Index in the US, respectively.
3 Source: Bond Market Association, 2000 data
4 This corresponds to a bid-offer spread of 50-75 basis points of the turnover value, with an average duration of
5 One alternative is that the Ministry of Finance establishes a separate portfolio for the shift to non-government-guaranteed bonds. It would then be natural that this portfolio was not measured against the Government Petroleum Fund's benchmark, but that the Ministry of Finance instead set strict limits for permissible transactions. The disadvantage of this alternative is that performance in parts of the Government Petroleum Fund (transition portfolio) would not be subject to comparison with the benchmark for the entire phasing-in period
6 Lehman Multiverse, July 2001
7 Source: Moodys Investor Service, January 2000
8 Mortgage-back securities and other asset-backed instruments
9 Lehman Global Aggregate, excluding NOK, Korean won, Taiwan dollar, Thai baht and Malaysian ringgit.
10 Schroder.Salomon Smith Barney World Broad Investment Grade Index
11 Merrill Lynch Global Broad Market Index, excluding NOK
12 Merrill Lynch Global Large Cap Index, excluding NOK
13 SSSB World BIG
14 Lehman LGA

Published 23 August 2001 11:02