The Government Pension Fund Global – analyses and assessments of the investment strategy for bonds

We refer to the Ministry of Finance's letter of 25 February 2026. In this letter, we answer the questions relating to the investment strategy for bonds in the Government Pension Fund Global (GPFG). The questions on geopolitical risk and concentration risk in the equity index are answered in a separate letter dated 1 September 2026. In both letters, Norges Bank emphasises the importance of broad diversification as a strategy in a changing world.

The Ministry raises two main questions about the investment strategy for bonds. First, the Bank is asked to assess whether the three defined roles of the bond investments – reducing fluctuations in the overall portfolio, providing liquidity and earning risk premiums – still appear appropriate, including how the various considerations should be weighted. Second, the Bank is asked to assess the composition of the benchmark index for bonds, including which markets and segments should be included and the associated weighting principles. The Ministry points to three matters that should form part of the assessments: whether GDP weighting still provides satisfactory diversification of risk with respect to sovereigns' debt-servicing capacity, whether the consideration of simplicity implies that large bond markets should be kept outside the index, and whether duration should follow automatically from the maturities determined by issuers.

The expert group appointed by the Ministry of Finance in 2025 has been given a similar, but somewhat broader, remit. The expert group is also to assess the fund's purpose and the owner's risk tolerance, and is to deliver its report by 25 January 2027. Norges Bank's advice in this letter takes today's framework as its starting point. Different assumptions regarding the owner's risk tolerance or the fund's purpose could affect how the bond index should be composed.

Summary of the advice

The Ministry of Finance has defined three roles for the fund's bond investments: to reduce fluctuations in the overall portfolio, to provide liquidity and to earn risk premiums in the bond market. We consider that the three roles remain appropriate, but that they can be weighted somewhat differently in the composition of the benchmark index than they are today.

The analyses in this letter indicate that a reduction in the share of government bonds does not necessarily materially weaken the bond portfolio's ability to reduce fluctuations in the fund. The split between government bonds and the other segments reflects the trade-off between the three roles and should therefore remain fixed over time. We recommend that the government subindex of the bond index be reduced from 70 to 50 percent. A government share of 50 percent will be sufficient to cover the liquidity needs, including in periods of turbulence in financial markets. We further recommend that the government subindex be weighted by market value instead of GDP, since high government debt is now a general feature of developed economies rather than a distinctive feature of a few countries.

The remaining part of the bond index should provide exposure to more sources of risk premiums in the bond market than it does today. The fund's long investment horizon makes it well suited to earning such risk premiums over time. Norges Bank's advice is that securitized bonds (including mortgage-backed securities, so-called agency MBS) and government-related bonds should be included in the bond index. A broad market index provides exposure to more risk premiums and gives a more diversified benchmark index than today.

Norges Bank recommends that the benchmark index for bonds be aligned more closely with the broader market represented by the Bloomberg Global Aggregate. The new benchmark index should be composed on the basis of the markets currently in the included segments and thereafter held fixed until the next major review of the bond index. We recommend that the duration of the benchmark index follow the market, as it does today, and that emerging markets continue to be kept outside.

The advice in this letter has a different orientation from Norges Bank's earlier advice on the bond index, in which the Bank recommended a substantial simplification of the index with fewer currencies and segments. In Report to the Storting No. 20 (2018-2019), the Ministry of Finance took as its premise that risk factors to which the fund wishes to be exposed should as a general rule be reflected in the benchmark index. The Bank's advice in this letter is based on that assessment.

Background

The benchmark index and the investment universe today

The management mandate sets out the fund's overall investment strategy. The mandate specifies what the fund may be invested in and defines a benchmark index against which the management of the fund is measured. The investment universe for the bond portfolio comprises tradable debt instruments, as well as depositary receipts for such debt instruments. The investment universe is broader than the bond index, which is tailor-made on the basis of sub-indices from Bloomberg. The index consists of a government subindex (70 percent) and a corporate subindex (30 percent).

The government subindex includes nominal and inflation-linked government bonds issued by developed markets, as well as bonds issued by supranationals.[1] The government subindex is weighted by gross domestic product (GDP), not market value. This means that countries with large economies relative to the size of their government debt receive a higher weight than market weights would imply, and vice versa. The corporate subindex includes corporate bonds and covered bonds. The corporate subindex is limited to seven currencies,[2] which are weighted by market value. The duration of the bond index is determined by the maturity structure of the market.

Emerging markets, high-yield bonds, mortgage-backed securities (agency MBS)[3], commercial mortgage-backed securities (CMBS), asset-backed securities (ABS) and government-related bonds, with the exception of bonds issued by supranationals, are not included in the bond index. The fund may nevertheless invest in these segments within the limits that apply under the management mandate.

The fixed income market

The Bloomberg Global Aggregate is often used to define the investable global bond market for institutional investors. The index includes government bonds, corporate bonds, securitized bonds and government-related bonds from 27 currencies.

The Bloomberg Global Aggregate represents by far the most significant part of the investable bond market (see Figure 1) despite certain exclusions. Inflation-linked government bonds, high-yield bonds and private credit are all outside it. Bonds with less than one year of remaining maturity are not included in the standard index. The rationale is that bonds close to maturity have close to zero interest-rate risk and behave more like money market instruments than like bonds.

Figure 1: Market shares in the global bond market

Figure 1.1 shows the market shares of the Bloomberg Global Aggregate, Bloomberg Global Treasury 0-1 Year, Bloomberg Global High Yield, Bloomberg Global Inflation-Linked and private credit. Data on assets under management in private credit are from Burgiss MSCI. The value represents the total amount of committed capital in the funds, including uninvested capital (so-called ‘dry powder’). Figure 1.2 shows the market shares of the sectors within the Bloomberg Global Aggregate.

Comparable funds

The bond indices of other large funds vary considerably, and several have not disclosed their approach. A review based on publicly available information is provided in Appendix D. This review shows, among other things, that the funds that build on broad market indices typically include mortgage-backed securities. 

The choice of bond index depends on the role bonds have in the portfolio. Defined-benefit pension funds, for example, use bonds to match future pension payments, so that the bond index is largely governed by the structure of the liabilities. The GPFG differs from these in that the fund has no explicit liabilities to be covered. For the GPFG, bonds are held for the three roles the Ministry of Finance has defined, and the benchmark index should reflect these.

On the three roles of the bond investments

The Ministry of Finance has defined three roles for the fund's bond investments: to reduce fluctuations in the overall portfolio, to provide liquidity and to increase the return through earning risk premiums in the bond market. Our assessment is that the three roles remain appropriate, but that they can be weighted somewhat differently in the composition of the benchmark index compared to today.

Reduce fluctuations

The first and most important role of bonds in the fund is to reduce fluctuations in the fund's overall return. Bonds' ability to reduce fluctuations in an overall portfolio has two sources. First, bonds have lower volatility than equities, so that any element of bonds reduces overall fluctuations regardless of which bonds are held. Second, bond prices tend to rise when equity prices fall, so that the bonds reduce the total fall in the value of the fund. All bonds with high credit quality help to reduce fluctuations, but to somewhat differing degrees.

In normal periods, the volatility of bonds is the most important source of reduced fluctuations in the fund's overall return, which means that the choice between government bonds and corporate bonds matters little for the fluctuations in a portfolio with 70 percent equities and 30 percent bonds. Our analyses show that a 70/30 portfolio had of between 10.7 and 11.5 percent in the period 1995–2025, depending on whether the bond component consisted of government bonds, corporate bonds or mortgage-backed securities.[4] A pure equity portfolio had annualised volatility of 15.3 percent over the same period. The large contribution to the reduction of volatility in normal periods comes from holding bonds at all, not from the choice between segments of the bond market.

This changes somewhat in crises. Over the period we examine, government bonds have had a negative co-movement with equities of 0.34 in crisis periods and have thus reduced fluctuations more when the need is greatest. Corporate bonds move in the same direction as equities in crises, with a positive co-movement of 0.42.[5] For mortgage-backed securities, the co-movement with equities is slightly positive in normal periods, but has historically turned negative in crises, so that they have provided an additional reduction of volatility in crisis periods which resembles government bonds more than corporate bonds.

The degree to which bonds will contribute to dampening volatility in future crises, will depend on the nature of the crisis. For example, one might expect government bonds to not have the same volatility-dampening effects in a government bond crisis.

Liquidity

The second role of bonds in the fund is to provide liquidity, mainly in connection with rebalancing.[6]  When the equity share falls more than two percentage points below the strategic share of 70 percent, the rebalancing rule requires the fund to sell bonds and buy equities. The more liquid and stable the bonds are, the more easily this can be executed without affecting market prices. Nominal government bonds issued by large developed markets are a natural source of liquidity for the fund, as trading volumes in these markets are stable, often also in periods of turbulence in financial markets. The size of the fund makes liquidity particularly important.

Simulations of the rebalancing rule over the period 1973–2025 indicate that the fund on average needs to sell bonds equivalent to around 5.5 percent of the value of the fund over a rebalancing episode. In the five percent most demanding periods, simulations show that the fund must sell bonds amounting to up to 11 percent of the value of the fund. In the most demanding periods, this corresponds to just under 40 percent of the bond portfolio given an equity share of 70 percent. Government bonds issued by developed markets currently account for around 70 percent of the bond index, or 21 percent of the fund's benchmark index. The simulations in Appendix B show that there is a comfortable margin to the estimated liquidity needs, including in periods of turbulence in financial markets.

Risk premiums in the bond market

The third role of bonds in the fund is to earn risk premiums in the bond market. Investors hold types of bonds other than government bonds to earn risk premiums. The most important segments are corporate bonds, government-related bonds and securitized bonds,[7] of which mortgage-backed securities are the largest part. Each of these segments offers distinct risk premiums with different characteristics.

Investors in corporate bonds expect a credit premium as compensation for the risk of default and for these bonds being less liquid than government bonds. In a similar manner, mortgage-backed securities are expected to provide a so-called prepayment premium. The premium arises because borrowers have the right to refinance their loans at a lower interest rate, which means that investors bear an option that is in the borrower's favour. Government-related bonds, that is, bonds issued by supranationals, federal institutions and municipalities, can be expected to provide a liquidity premium and a small credit premium over pure sovereign risk. These issuers have implicit or explicit government support and high credit quality, but are somewhat less liquid than government bonds from the largest markets.

In Appendix C, we show how all three of these segments have delivered higher average annual returns than government bonds over the period 1995–2025, as well as higher risk-adjusted returns.

The fund's long investment horizon makes it well suited to earning such risk premiums, which will vary over time. As described at the outset, we take as our premise that the risk premiums to which the fund wishes to be exposed should as a general rule be reflected in the benchmark index. The consideration of risk premiums must be weighed against the first two roles. Corporate bonds have historically moved in the same direction as equities in crisis periods, which weakens the ability to reduce fluctuations precisely when the need is greatest. Mortgage-backed securities, by contrast, have shown volatility-reducing properties in crisis periods, but at a lower level than government bonds.

Overall assessment

The review of the three roles indicates that they can be weighted somewhat differently in the composition of the benchmark index than they are today. The most important factor for reducing fluctuations in an overall portfolio is the inclusion of bond investments, not which segments are included in the index. In addition, our analyses show that the fund's liquidity needs may imply a lower government share than today. The fund's distinctive features also make it well suited to earning risk premiums over time. This may suggest that the bond index should be expanded to other segments. Mortgage-backed securities are of particular interest here, because they over time have provided a risk premium while at the same time contributing to reduced volatility in crisis periods.

Assessment of the composition of the benchmark index

A lower share of government bonds

Norges Bank recommends that the government subindex of the bond index be reduced from 70 to 50 percent. The analysis above shows that a reduction in the share of government bonds does not necessarily materially weaken the ability to reduce fluctuations. Simulations further indicate that a government subindex of 40 percent would be sufficient to cover the liquidity needs, including in periods of turbulence in financial markets.

A share of government bonds that is higher than the liquidity needs does, however, represent an implicit cost in the form of somewhat lower expected return. At the same time, it is important that government bonds account for a substantial share of the bond portfolio and help to reduce volatility also in the most demanding market periods. Norges Bank's assessment is that a government share of 50 percent provides a comfortable margin to the estimated upper limit for the liquidity needs and reduces the risk of significant market impact in the event of a liquidity event. Future periods of stress may deviate from historical patterns, for example if equities are more volatile, the value of the fund is considerably higher, or the rebalancing rule triggers sales over shorter periods than previously. A government share of 50 percent is also more in line with the market, represented by the Bloomberg Global Aggregate.[8]

We recommend that a fixed share of government bonds in the benchmark index be maintained, with monthly rebalancing as today. The split between government bonds and the other segments reflects a trade-off between the three roles. This trade-off should remain fixed over time, independently of market developments.

Inflation-linked bonds

Since 2005, the government subindex of the current bond index has included inflation-linked government bonds issued by developed markets. In 2012, Norges Bank proposed removing them.[9] The Ministry of Finance chose to retain them, on the grounds that diversification of risk weighed more heavily than the need to simplify the index, and that the fund as a long-term investor could earn a liquidity premium in this market.

Norges Bank recommends that inflation-linked government bonds be retained in the benchmark index. Unlike nominal bonds, they provide exposure to real interest rates and protection against unexpected inflation. In periods of unexpectedly high inflation, nominal bonds will lose real value, while inflation-linked bonds can help to preserve purchasing power. Since the Bloomberg Global Aggregate does not include this segment, it is recommended that inflation-linked government bonds be retained as an additional segment.

Emerging markets

It was decided in 2019 to remove emerging markets from the government subindex of the bond index.[10] The Ministry noted that upgrades and downgrades of issuing countries' credit quality appeared to be more of a challenge for emerging markets than for developed markets. Norges Bank still shares this assessment and recommends that emerging markets continue to be kept outside the benchmark index.

Weighting principle for the government subindex: market weights

The Ministry asks for an assessment of whether GDP weighting still provides satisfactory diversification of risk with respect to sovereigns' debt-servicing capacity, and of whether the consideration of fiscal strength should be addressed through active management or in the benchmark index.

The rationale for introducing GDP weights in 2012 was that GDP weights could provide better diversification of risk than market weights. Norges Bank's assessment is that GDP weights no longer provide diversification of risk with respect to sovereigns' debt-servicing capacity that is necessarily better than market weights. Since GDP weights were introduced, high government debt has gone from being a distinctive feature of individual countries, particularly Japan and some euro area countries, to becoming a more general characteristic of developed economies (see Figure 2). In Appendix F, we show that GDP weights have delivered higher returns than market weights over the period 2001 to 2025. The historical excess return is largely driven by the underweighting of Japan and the depreciation of the yen over recent years. That the underweighting of Japan has in hindsight delivered excess return is thus a result of one country's distinctive features over a particular period, and not an argument that GDP is a good measure of fiscal risk.

Figure 2: Government debt as a share of GDP

Government debt as a percentage of GDP. Annual data from 1950 to 2024.

The principle of market weighting means that borrowers issuing large volumes of bonds receive an increased weight in the benchmark index. The principle is self-adjusting in that the market continuously prices fiscal risk and other macroeconomic factors such as growth and inflation. Countries with weak public finances must pay a higher interest rate to borrow, which over time limits the weight in the index.

At the same time, there are features of the government bond market that raise questions about the pricing of government bonds. The investor composition and the regulation of institutional investors, central banks' large securities purchases and the role of government bonds as a safe haven are factors that may affect pricing in ways that deviate from what fundamentals would imply. Over long time horizons, our assessment is that government bond yields largely reflect macroeconomic trends, and NBIM's analyses show that most of the variation in long government bond yields can be explained by changes in long-term inflation prospects and the equilibrium real interest rate.[11] Any mispricing of government bonds over shorter horizons, for example related to buyers that are not price-sensitive, can be captured by active management.[12]

Norges Bank's assessment is that market weights are a natural starting point for the government bond index. This is also in line with the recommendation from the expert group that assessed the same question in 2018-2019.[13] The most widely used bond indices are based on market weights. Our assessment is that high government debt in several developed markets weakens the argument that GDP weights contribute to risk reduction by correcting for individual countries' high market weight. GDP weighting is in addition the most important source of complexity and operational risk in the composition of the benchmark index.

Regardless of whether government bonds are weighted by GDP or by market value, a situation could arise in which Norges Bank holds the government debt of a country with debt problems. As a creditor, Norges Bank could be expected to take a position on debt relief, restructuring or sanctions. These situations should be handled as part of the operational management.

A broader benchmark index that earns more risk premiums

The fund's distinctive features make it well suited to earning risk premiums in the bond market that vary over time. We recommend that the other half of the bond index provide exposure to more sources of risk premiums than it does today. In addition to the corporate bonds that are included in the index today, it is the Bank's assessment that securitized bonds and government-related bonds should be included in the bond index. Both of these segments are included in the broad bond index Bloomberg Global Aggregate. This would mean that the benchmark index moves from being dominated by the term premium on government bonds to providing a more balanced and diversified exposure to several sources of return.

Securitized bonds

Securitized bonds are a large segment of the Bloomberg Global Aggregate. The segment consists mainly of mortgage-backed securities (agency MBS). The segment was removed from the bond index in 2012. During the financial crisis, Norges Bank had exposure to private mortgage-backed securities through external mandates. These became illiquid and difficult to manage in that period.

Agency MBS is today a standardised and liquid segment with an outstanding amount of around 7,500 billion dollars at the end of 2025. This is around half of the outstanding corporate bonds in developed markets. Average daily turnover in agency MBS is around 350 billion dollars, considerably higher than for US corporate bonds. The bonds are guaranteed by Fannie Mae, Freddie Mac and Ginnie Mae, and the credit quality is close to that of US government bonds. The segment has historically provided a risk premium related to prepayment risk. This premium differs from the credit premium in corporate bonds and represents a source of additional diversification. As discussed earlier, this segment has had a negative co-movement with equities in periods of turbulence in financial markets. A more detailed analysis is provided in Appendix C.

Securitized bonds in the Bloomberg Global Aggregate also include commercial mortgage-backed securities (CMBS) and asset-backed securities (ABS). These segments are considerably smaller than agency MBS and together account for less than one percent of the Bloomberg Global Aggregate. The rationale for including them is not based on a separate analysis of these segments' characteristics, but that they form a natural part of a broad market index.

Government-related bonds

Government-related bonds, that is, bonds issued by federal institutions, municipalities, supranationals, and sovereigns in foreign currency, are included in the Bloomberg Global Aggregate, but largely kept outside the current benchmark index. These issuers have high credit quality and several operate with implicit or explicit government support. Historical return figures in Appendix C show that government-related bonds have delivered higher risk-adjusted returns than nominal government bonds, and thus represent a source of additional risk premiums that the benchmark index does not capture today. We recommend that these segments also be included in the benchmark index.[14] Bonds issued by supranationals, which are already included in the current index, are recommended to be moved from the government subindex to the non-government subindex in line with the classification in the Bloomberg Global Aggregate. These bonds account for 3.7 percent of the current benchmark index.

Duration

The Ministry asks for an assessment of whether the duration of the benchmark index should be set explicitly or follow automatically from the maturities determined by issuers, as it does today. Duration is a measure of sensitivity to interest rates and varies over time both as a result of the level of interest rates and of the maturity profile issuers choose. The average duration of the index is currently around six to seven years.

Norges Bank recommends that duration continue to follow the maturity structure of the market. Bonds enter the index when they are issued and leave it when the remaining maturity falls below one year. Although it could make sense for a long-term investor to have an index that holds bonds all the way to maturity, we are not proposing to include bonds with a maturity of less than one year at this time. This is because there is currently no index product that covers all the segments we propose to include in the bond index. Longer duration provides a higher term premium and stronger reduction of volatility in crises,[15] which is favourable seen from the perspective of the primary role of bonds in the portfolio. Shorter duration provides lower interest-rate risk and higher liquidity. Many pension funds will choose the duration of the bond portfolio on the basis of the duration of their liabilities. The fund has no explicit liabilities, and it therefore seems natural to take the market's pricing of interest-rate risk as the starting point. Changes to the assumptions regarding the owner's risk tolerance or the fund's purpose could affect how the bond index should be composed, and in particular what duration profile it should have.

A benchmark index based on the Bloomberg Global Aggregate

The analyses of the roles of bonds indicate that these should be weighted somewhat differently than today in the composition of the bond index.

Norges Bank recommends that the benchmark index for bonds be aligned more closely with a broad market index (Bloomberg Global Aggregate). Our analyses show that all investment-grade bonds dampen volatility in the overall portfolio, and that the liquidity needs are well covered with a government subindex of 50 percent. Moving closer to a broad market index will provide increased exposure to risk premiums in the bond market and a more diversified composition than today. Taken together, this means that the bond portfolio can deliver a somewhat higher expected risk-adjusted return without materially weakening the ability of the bond investments to reduce volatility or to cover the liquidity needs.

We recommend that all segments of the Bloomberg Global Aggregate be included, with the exception of bonds issued by emerging markets.[16] Inflation-linked government bonds should be included in the government subindex in line with current practice. The strategic split is proposed to be set at 50 percent government bonds and 50 percent other bonds. Within both parts, market weights should be used.

At the same time as the transition to market weights in both parts of the bond index, we recommend that the currency composition on the corporate side be expanded from a list of seven currencies to include all currencies in the Bloomberg Global Aggregate for developed markets[17] and that the freeze on new markets in the government subindex be lifted temporarily at the transition. This means that the new index will consist of the markets that are currently included in the Bloomberg Global Aggregate and that the market composition is thereafter held fixed until the next major review of the bond index. Duration should follow the maturity structure of the market, as it does today.

The Ministry asks for an assessment of whether the consideration of simplicity in itself implies that large bond markets should be kept outside the index, and mentions US mortgage-backed securities as an example. Today's index involves a high degree of customisation, with, for example, the use of GDP weights and fewer segments than the broad market index. Fewer segments make the management of the fund simpler, while customisation contributes to increased operational risk. Customisation also makes the index harder to verify. Aligning more closely with a standard index will make the fund's benchmark index more transparent and verifiable. It does, however, require somewhat more resources to manage since it includes more types of segments. Norges Bank's assessment is that this additional burden is manageable, and that the gain from reduced customisation weighs more heavily.

Change in risk in the recommended benchmark index

In Appendix G, we show the composition of the current benchmark index, the Bloomberg Global Aggregate for developed markets and the proposed new benchmark index broken down by segment, currency and country. The most important changes from today's index are a considerably lower share of government bonds and the inclusion of mortgage-backed securities, which account for around 13 percent of the recommended index against zero today. Government-related bonds increase from around 4 to around 11 percent. The currency distribution is close to unchanged, with the exception of the Japanese yen, which increases from around 5 to 8 percent. The US dollar is by far the largest currency, with a weight of just over 50 percent. The largest difference in the dollar market is that the share of US government bonds falls, which is offset by a roughly corresponding increase in other US bonds. In the appendix, we compare the return and risk characteristics of the current and the recommended benchmark index. The largest difference is that the recommended index provides better diversification of risk across more sources of return rather than being dominated by government bonds.

A broader benchmark index will mean that the fund takes on risk different from the risk it takes today. Mortgage-backed securities have prepayment risk, and although this has historically been compensated with a premium, the return may fluctuate or fail to materialise. Mortgage-backed securities are closely linked to the US government through an implicit or explicit government guarantee, but they are also secured by collateral. A lower government share reduces the liquidity of the portfolio. The effects of the proposal on expected return and risk for the fund are small, with a marginally higher expected return and marginally lower volatility. Overall, Norges Bank assesses that the advantages of a broader index that is closer to the market weigh more heavily than the changes in risk.

Conclusion

The Ministry of Finance asks for an assessment of how the recommended benchmark index should be reflected in the management mandate. Norges Bank will revert with a specific proposal for how the bond index should be specified in the mandate once the Ministry has taken a position on the advice. We consider that the other limits and requirements for the bond portfolio, including the limit on the exposure to high-yield bonds and emerging markets in the active management, should be continued as today.

A broader index does at the same time place somewhat greater demands on the ongoing management of the fund. Norges Bank already has experience with several of the segments we propose to include. At the same time, any inclusion of mortgage-backed securities requires that we develop expertise in this area with a few dedicated resources. A broader bond index will have very limited significance for the transaction costs of managing it.[18] The adjustment to any new benchmark index should be made gradually out of consideration for market impact and transaction costs.[19] Norges Bank assumes that we will be able to present a proposal for an implementation plan once the Ministry has taken a position on the advice.

Yours faithfully

Ida Wolden Bache
Nicolai Tangen

[1] Supranationals include, among others, bonds issued by the European Investment Bank, the World Bank and regional development banks.

[2] US dollars, euros, British pounds, Canadian dollars, Swiss francs, Danish kroner and Swedish kronor.

[3] By mortgage-backed securities we mean agency MBS in accordance with Bloomberg's classification in the remainder of this letter.

[4] See Appendix A.

[5] Co-movement is measured here as the correlation between monthly returns on bonds and equities respectively in crisis periods. In Appendix A, the same relationship is expressed through equity beta, which in addition to the correlation takes into account the ratio between the volatilities of the two asset classes. The two measures describe the same underlying property, but equity beta indicates how much the bond return changes on average when the equity market moves by one percent, and is therefore better suited to quantifying the contribution to the portfolio's overall fluctuations.

[6] Here and in Appendix B, we focus on rebalancing rather than on outflows from the fund. This is because outflows are taken from the whole fund, not only from the bond portfolio.

[7] By securitized bonds we mean the securitized segment in accordance with Bloomberg's classification.

[8] As of 30 June 2026, government bonds account for slightly more than half of the Bloomberg Global Aggregate, also if emerging markets are excluded.

[9] Letter to the Ministry of Finance of 9 August 2012.

[10] For more information, see Report to the Storting No. 20 (2018-2019).

[11] NBIM Discussion Note #1-2023.

[12] In Appendix F, we review the literature on factors that may give rise to mispricing in the government bond market.

[13] van Binsbergen, Jules H. and Ralph S.J. Koijen, “Benchmarking Global Fixed Income Portfolios”, report to the Ministry of Finance, November 2018.

[14] In the current benchmark index, bonds issued by supranationals are included in the government subindex. In the Bloomberg Global Aggregate, these are classified as government-related bonds. In the recommendation for a new benchmark index, supranational issuers will therefore not be included in the government subindex, but in the other half of the 50/50 split.

[15] Assuming negative co-movement between equities and bonds.

[16] Including nominal government bonds, government-related bonds, securitized bonds, and corporate bonds.

[17] In total, Japanese yen, Australian dollars, New Zealand dollars, and Singapore dollars will be added to the corporate sector.

[18] The recommended index gives an estimated additional cost of 6.6 million kroner per year, but the additional costs are smaller than the uncertainty in the model estimates. Overall, the ongoing costs of the two indices are so similar that the difference should not be given weight. 

[19] The one-off cost of the transition is estimated at around 750 million kroner as an upper limit. This can be reduced considerably through gradual implementation that makes use of the natural reinvestment of maturing bonds, netting against existing active positions and other capital flows in the fund.