To main content

Inclusion of real estate investments in the Government Pension Fund - Global

Norges Bank's letter to the Ministry of Finance 22 August 2007

22 August 2007

1.         Introduction

In its letter of 20 October 2006, Norges Bank advised on the long-term investment strategy for the Government Pension Fund – Global. One of the proposals was the inclusion of real estate and infrastructure as a separate asset class.[1] The letter also stated that the Bank would come back later with a more detailed account of how the Bank can manage the Fund’s investments in unlisted asset classes.

In the present letter, we provide information on how the Bank can carry out investments in real estate. We recommend that, with effect from 1 January 2008, the Ministry of Finance define a separate asset class for real estate and permit the inclusion of private (unlisted) investments in real estate and infrastructure in the investment universe. During the early years, the size of the portfolio can be limited to 3 per cent of the overall Fund. The idea of establishing this portfolio is partly to allow investments that can increase the return on the Fund in the coming years, and partly to build up the expertise needed in order gradually to be able to handle a large real estate portfolio.

It is also recommended that, in the early years, the measurement of the results of this management be based on specific requirements for absolute return set by the Ministry, and that no target be set for how quickly the portfolio should be built up. With this kind of set-up, the Bank will need to weigh up investments against the required rate of return, and the speed at which the portfolio is built up will be determined by the Bank’s assessments of the expected return on the individual investments.

The present letter also describes briefly how the Bank plans to handle the investment process, valuation and risk management, and proposes principles for the design of a management mandate and for follow-up. This letter does not discuss the issue of the size of the long-term strategic allocation to real estate and infrastructure, as this was discussed in the letter of 20 October 2006. The advice from the Bank was to set a long-term target of up to 10 per cent of the total value of the Fund. As this equates to a very large portfolio, and the bulk of the real estate market is unlisted, the Bank recommended expanding the investment universe to include private investments in real estate and infrastructure. Appendix 1 provides an overview of the global real estate market.

2.         The main choices for expanding exposure to real estate 

Real estate and infrastructure are already part of the investment universe defined for Norges Bank, and are included in the benchmark portfolio in various guises. Norges Bank Investment Management (NBIM) manages NOK 40-50 billion in real estate and infrastructure investments for the Government Pension Fund – Global. Equity instruments account for the bulk of this.[2]

The issue is not, therefore, whether investments in real estate should be included in the portfolio or whether NBIM should have real estate investment expertise. Instead, there are two main issues which seem to be of particular importance:

  • Should real estate be established as a separate asset class with a strategic target of a significant allocation?
  • Should the real estate investment universe be expanded to include private investments?

The Ministry of Finance discussed the introduction of real estate as an asset class in the Fund in Report to the Storting No. 24 (2006-2007). This refers to the recommendation from the Bank, and the Ministry presents several important arguments in favour of establishing a strategic allocation to real estate. A number of issues need to be resolved before a decision can be taken, including the tax situation, the control model, and the organisation of management.

The Ministry can adopt a set expansion schedule from the outset, with a gradual increase in the strategic allocation to real estate and infrastructure towards the desired long-term percentage of the Fund. However, it may be appropriate to have an interim solution in the early years, where the speed at which the portfolio is built up is determined by the opportunities spotted by the manager.

By adopting a set schedule for the expansion of exposure to real estate and infrastructure, the Ministry can ensure that the long-term weighting is achieved at the desired point in time. This may require the establishment of control systems which ensure that the Ministry itself bears the bulk of the market risk (as is currently the case in the management of investments in listed markets). As can be seen from Appendix 2 to this letter, benchmark indices for such investments have various shortcomings. 

It may take many years to build up a substantial strategic allocation to real estate and infrastructure. There are two main reasons for this. Firstly, the market has limited liquidity, and it would not be advisable to build up the portfolio over a period of just a few years on account of the transaction costs. Secondly, specialist real estate market expertise will be required in every single one of the countries in which it may be appropriate to invest. Although NBIM will choose to make the majority of these investments through external managers, advisers and partners, it will take time to build up contact networks, control systems and so on.

Instead of adopting a set expansion schedule, the Ministry can employ a somewhat simpler control model in the early years. A more limited allocation to real estate can be defined, and the yardstick for the Bank’s management can be based on total return rather than the assumed return on portfolios of broad real estate investments. An example of such a yardstick is the risk-free interest rate in each particular country, adjusted to reflect the specific risk premium to be associated with investments in real estate and infrastructure. Alternatively, the long-term return on government bonds can be used for this type of measurement.

A control model of this kind makes the Bank responsible for the results of investments. If the return on the individual investments in real estate and infrastructure is not at least as high as the return target, this will impact negatively on the Bank’s results. This will make it necessary to assess the individual investments not only relative to others in the same asset class but also in terms of whether the return is at least as high as the yardstick.

There are several reasons why this control model is preferable. As mentioned above, it will take time to build up more expertise in the management of real estate and infrastructure investments. In the absence of a minimum requirement for the size of the portfolio, the Bank can choose an expansion profile which is in step with the accumulation of expertise. Furthermore, it will probably be important for both the Ministry and the Bank to gain more experience of alternative benchmark portfolios before establishing a control system based on such.

One possible objection is that the current pricing of real estate in large developed markets suggests that we should wait before establishing the real estate portfolio. However, experience has shown that it is difficult to identify turning points in a market. Even if we believe the market to be highly priced, work on expanding the investment universe and building up expertise should still begin now. With the control model proposed here, only projects that meet the minimum requirement for expected return (relative to risk) will be candidates for investment.

It is natural for the Ministry to set requirements for how well the portfolio invested in real estate and infrastructure is diversified. In a control system with a benchmark portfolio and limits for expected deviation from this benchmark (tracking error), there will naturally be a high degree of diversification. Without this kind of control system, it may be necessary to establish more explicit criteria for the spread of risk in the portfolio. Examples of such are the breakdown between regions and main types of investment, and limits on the size of any one investment relative to the others. However, it will not be appropriate to have such diversification criteria for as long as the real estate portfolio remains a small part of the overall Fund. The Bank proposes that diversification criteria enter into force only when the portfolio reaches 3 per cent of the overall Fund.

As noted in the letter to the Ministry of 20 October 2006 and illustrated in Appendix 1, the bulk of the global real estate market consists of unlisted investments. A significant allocation to real estate in the Fund therefore requires it to be permitted to include unlisted investments in real estate and infrastructure in the investment universe. The absence of liquid marketplaces presents challenges in terms of the measurement of returns and risk, among other things. This is discussed in more detail below.

In the equity portfolio, there is an upper limit of 5 per cent for the ownership stake in any one company, based on voting rights. With investments in unlisted markets, it will be difficult in practice to have an upper limit for ownership stake. An investor needs to be more prepared to accept that an investment is not very liquid and to exercise active ownership, unless this takes place through a fund. The Bank recommends that no upper limit be set for the ownership stake in any one investment. When it comes to risk management, however, limits should be set for the size of any one investment.

A more precise definition needs to be formulated for the types of investment that are to be included in the separate portfolio for real estate and infrastructure. It is natural to have a general requirement that an investment’s most important return and risk parameters are associated with actual and future cash flows from real estate and infrastructure activities. A decision also needs to be taken on whether both equity and debt instruments are to be included. In the Bank’s previous letter, it was noted that it is equity which normally gives the desired portfolio characteristics. However, the guidelines should permit investment in debt instruments where there is a reasonable likelihood of subsequent conversion into equity.

Consideration must also be given to whether real estate and infrastructure should be removed from the current benchmark portfolio for listed equities. The Bank recommends that the Ministry defers a decision on this issue. During the early years, it will not be important for the Fund’s overall portfolio characteristics anyway. In the longer term, the best alternative will probably be for the Ministry, when setting the target for the strategic allocation to real estate, to take account of there already being some investments in real estate and infrastructure in the listed part of the strategic equity portfolio.

The establishment of the portfolio of real estate and infrastructure investments should be financed by a reduction in the allocation to the bond portfolio (cf. previous letter from the Bank).

3.         Norges Bank as an investor in real estate

Current investments in real estate and infrastructure

The Government Pension Fund – Global is already invested in the real estate market via listed equity and fixed income markets. At the end of May 2007, equity investments in listed property companies totalled approx. NOK 17 billion, or approx. 2 per cent of the Fund’s equity portfolio. Real estate’s share of the fixed income portfolio was only marginally lower. If we count the industrial transportation and airline sectors as infrastructure investments, investments in these sectors totalled approx. NOK 16 billion at the end of May 2007, or somewhat more than 1.5 per cent of the equity portfolio. Infrastructure’s share of the fixed income portfolio was only marginally lower.

Through residential mortgage-backed securities (RMBSs), the Fund has considerable exposure to one particular sector of the real estate market: the US home loan market. This market is not normally included in a real estate portfolio (cf. table in Appendix 1). Investments in commercial mortgage-backed securities (CMBSs) totalled approx. NOK 12 billion, or approx. 1 per cent of the fixed income portfolio, at the end of May 2007.

Possible investments in private markets

Based on the overview of the global real estate market in Appendix 1 and the special factors that apply when investing in this market, Norges Bank will, in the first instance, prioritise investments in the unlisted market. Investments will predominantly be made in various types of fund structure or take the form of joint ventures with property companies. Only in special cases will it be appropriate to own properties directly. The development and operation of such projects will, to a great extent, involve the use of external property specialists.

Use of NBIM’s expertise and investment culture

As mentioned above, NBIM already invests in real estate. There are significant movements between real estate and infrastructure which are listed (and so included in the investment universe) and the private market. When units in private funds are listed on an exchange, they are included in the current investment universe without the underlying investments necessarily changing. Conversely, there are instances of listed property companies being delisted and becoming private companies.

For a fund like the Government Pension Fund – Global, it is not a good idea to have rigid separation of markets between which there are continuous movements. This can mean having to sell off investments which may (still) be profitable. It also means not being able to exploit possible inefficiencies in the capital structure. Where information is imperfect and liquidity is limited, there is the potential for higher returns for investors who can compare the different investment options. In addition, clear separation of the investments permitted in the equity mandate and the investments permitted in the real estate mandate, say, will mean that expertise developed in one part of the market does not get to be used in other, related markets.

Access to information about potential transactions is of far greater importance in private markets than in public markets. NBIM already has an extensive global network of contacts with institutions which have information about potential transactions. These include large investment banks, which are often involved in preparing and executing transactions. NBIM has extensive relations with this type of player.

The development of an organisation for real estate investments will take time, and recruiting skilled specialists in the individual markets will be a challenge. Here, NBIM will exploit the benefits of being an established and increasingly well-known global management organisation. The importance of proximity to the market suggests that NBIM’s international offices in New York, London and Shanghai will be used.

NBIM will also draw on the central control and risk functions at head office in Oslo. Some of the established administrative functions can also be used relatively easily and cost-effectively for private real estate investments. But there will also be a need to expand the functions for independent measurement of risks and returns. Such functions are more complicated in private than in public markets, as estimates of value and risk are more uncertain. Here, NBIM will build further on environments already well-established. Legal expertise will also be strengthened; this will, for example, be important for clarifying and following up tax issues.

Control, transparency and valuation

Norges Bank will attach importance to establishing sound and comprehensive control procedures in line with best international practice for real estate managers. This means that we will establish investment processes for private markets where the individual investment decision is reviewed on the basis of a set of predetermined criteria (due diligence). Outside the real estate management environment itself, functions will be established under the chief financial officer to ensure reasonable valuation and independent performance measurement. The Bank will also ensure that its Internal Audit department develops or brings in the necessary expertise in this field so that the Executive Board’s control responsibilities for this part of the Bank’s operations are adequately discharged. Central Bank Audit has recently entered into a co-operation agreement with a large audit firm, whose global organisation with its extensive expertise in auditing real estate investments will ensure satisfactory auditing of the Fund’s investments in private real estate and infrastructure projects.

4.         Possible mandate from the Ministry of Finance

It will take many years to build up a strategic allocation to real estate and infrastructure of 10 per cent of the Government Pension Fund – Global. The markets have limited liquidity, and it will take time to develop management expertise. Norges Bank recommends that the Ministry of Finance permit a separate real estate portfolio to be built up from 1 January 2008, on the assumption that the Bank presents a satisfactory account of the investment process and independent analysis of risk and measurement of returns.

In the early years, the establishment of the real estate portfolio should be cautious and without any specific short-term targets for the size of the portfolio. This can be done by using a control model where the Bank’s management is measured against absolute requirements for returns in the individual markets. A ceiling should be set for how large the portfolio can grow before the control model is reviewed and formal criteria for the diversification of the portfolio are introduced. The Bank recommends setting this ceiling at 3 per cent of the Fund.

The Bank proposes that its management mandate have the following key features:

  1. A separate portfolio is established for real estate and infrastructure. Unlisted investments may also be included in this portfolio. A more precise definition of the types of investment that may be included is to be drawn up. The Bank will come up with proposals for such a definition. 
  2. The Ministry sets requirements for the return in the individual main markets based on market interest rates adjusted for specified risk premiums. The Bank will be able to draw up proposals for these return requirements based on information about what is customary in comparable funds.
  3. No requirement is made as to how quickly the Bank is to build up the portfolio. Once the portfolio reaches 3 per cent of the Fund, it will have to meet more specific criteria for diversification issued by the Ministry. The Bank is to come up with proposals for such criteria.
  4. Before making the first investment, the Bank is to account for how the investment processes for investments in unlisted markets will be organised, and how the Bank will implement independent measurement of returns and analysis of risk. The accounting and auditing set-up also needs to be reviewed.
  5. In its regular reporting to the Ministry, the Bank is to include a separate account of work on developing the real estate and infrastructure portfolio, and specify the return on this portfolio separately from the other portfolios.
  6. Investments are to be managed in accordance with the Fund’s Ethical Guidelines and must not include investments in Norway. There are no particular restrictions on ownership stakes.
  7. The Bank’s own costs for managing unlisted asset classes are to be covered through the overall cost budget set annually by the Ministry. Costs relating to the use of external managers are to be met outside this budget. The measurement of performance in terms of the return on the real estate portfolio is to be carried out net of such costs.


Yours faithfully

Svein Gjedrem

                    Knut N. Kjær

Appendix 1: The global property market

Norges Bank has provided an overview of the global real estate market in previous letters to the Ministry and in its annual strategy reports [3]. The following provides an updated presentation of the market for real estate investments and an overview of the types of investment that will be most relevant to Norges Bank in the early years.

We can divide the real estate investment universe along three axes:

  • Direct (investments in specific properties) and indirect (investments in units in property funds and/or shares in property companies)
  • Listed (public) and unlisted (private) real estate investments
  • Investments in equity instruments and debt instruments

The table below provides an overview of the global real estate market by main category:

Unlisted (private)

Listed (public)


The global unlisted real estate market (excluding owner-occupied property) is estimated to be worth USD 5,500 billion (1)

Direct investments

Single properties/portfolios of properties 

Indirect investments

Joint ventures with property companies

Property funds (commingled funds/co-investments)


Stub equity in debt securitisations

The global listed real estate market was worth USD 1,500 billion at the end of May 2007

Indirect investments (listed companies)

Property companies and REITs (2)

- Primary and secondary offerings
- Open-market purchases
- Block trades
- Funds of funds

Derivatives and exchange-traded funds (ETFs) 



Indirect investments

CMBSs (3) 
- investment grade/non-investment grade
- with or without “official” credit rating

Mezzanine debt

Privately placed convertibles from listed/unlisted property companies

Senior debt

Non-performing loans

Indirect investments

CMBSs (3) 
-  investment grade/non-vestment grade
- with or without “official” credit rating

Unsecured corporate debt

Perpetual preferred stock

(1)         The market can be broken down by region as follows: USA approx. USD 2,900 billion (53 per cent), Europe approx. USD 1,100 billion (20 per cent), and Asia approx. USD 1,500 billion (27 per cent).

(2)          Real Estate Investment Trusts. These are vehicles which pay dividends to investors that are tax-free for the REIT, but must also commit to investing the bulk of their assets in real estate and distributing the bulk of their net rental income to investors as dividends.

(3)          Commercial mortgage-backed securities.

When it comes to choosing investments, there are a number of special aspects of real estate investments which need to be considered separately. A few of these are presented in the following table:


Unlisted markets often feature low liquidity. High transaction costs and uncertainty about how quickly investments can be realised serve to increase the risk for investors and result in a risk premium being demanded for the lack of liquidity.

As a long-term investor, the Fund will be far more able to accept low liquidity in its investments than many other investors, and will therefore be in a position to exploit this risk premium.


The real estate market is generally under-analysed, and there is limited access to publicly available data in most markets.

Access to market information and high-quality investment opportunities will generally demand a regional/local presence, and management is often more resource-intensive than with listed equities and fixed income instruments.

One possible strategy, which ensures high-quality investment opportunities while minimising the need for resources, is to enter into joint ventures with local operators.


With unlisted investments, it is neither possible nor appropriate to carry out frequent valuations. Returns are therefore calculated on the basis of reasonable valuations performed by independent assessors, normally every six or 12 months.

Valuations will be performed in accordance with international standards.

Unlisted investments

Investments in unlisted assets often require a detailed review (due diligence) of a number of factors, including valuation, aspects of the property itself, legal matters, tax and accounting issues.

With each large investment in unlisted markets, formalised procedures are generally followed for assessing the business plan, earnings potential and risk, and for ensuring that the necessary investigations (due diligence) have been performed before the investment decision is taken.


Taxation will vary between countries and between different types of investment. This is therefore being considered separately by the Ministry of Finance, and Norges Bank will also clarify relevant tax issues before investment decisions are taken.

Appendix 2: Benchmark indices for real estate

In the listed real estate market, a number of benchmark indices have been developed which are investable. The three indices most commonly used are:

S&P/Citigroup BMI Property Index

  • Has the most extensive market coverage, as the inclusion criteria (minimum size and liquidity) are the least stringent
  • As at April 2007, the index covered 530 property companies in 21 developed and 13 emerging markets
  • Market capitalisation approx. USD 1,400 billion

FTSE EPRA/NAREIT Global Real Estate Index

  • Inclusion criteria are based on minimum size and liquidity
  • As at April 2007, the index covered 313 companies in 23 countries
  • Market capitalisation approx. USD 950 billion

Global Property Research 250 (GPR 250)

  • Covers the 250 most liquid property companies as measured by 12-month trailing trading volume
  • As at April 2007, these companies were based in 23 countries and had a market value of USD 810 billion

In the unlisted real estate market, there is no global benchmark index. The indices that have been developed often do not represent the underlying market in many countries and are therefore not normally investable. One exception is the unlisted real estate market in the UK, where investments can be replicated through a relatively illiquid derivative market. The indices most commonly used are:

IPD indices

  • These indices cover 14 European countries, Canada, Japan, South Africa and Australia
  • They cover approx. 40 per cent of the unlisted markets in Europe, approx. 50 per cent in Canada, approx. 60 per cent in South Africa, approx. 14 per cent in Japan and approx. 25 per cent in Australia

NCREIF Property Index (NPI)

  • This index covers less than 5 per cent of the unlisted market in the USA

Besides the challenges in terms of replicability, these indices also have a number of other shortcomings:

  • None of them has broad market coverage
  • Capital gains included in the indices are based on valuations that typically do not include the latest market movements
  • The indices do not take account of the effects of the redevelopment, renovation and refurbishment of properties
  • The indices are produced monthly, quarterly or annually (most European indices, for example, are published annually)

Since a substantial part of the Fund’s real estate portfolio needs to consist of unlisted investments, the usual requirements made of a benchmark portfolio (transparency, representativity, investability and so on) cannot be met. An absolute return requirement may therefore be more appropriate for this portfolio. Further work is required to establish a good management system and return requirements for the strategic real estate portfolio. Norges Bank will advise the Ministry of Finance on this at a later date. Market developments will probably also result in more suitable benchmark indices in the coming years.


[1] In the present letter, the emphasis is on real estate. Investments in infrastructure have many similarities, but it will probably be best for the management organisation to build up two separate competence centres. Real estate is by far the largest asset class, and Norges Bank is currently giving priority to developing real estate expertise. 

[2] In the fixed income portfolio, only commercial mortgage-backed securities (CMBSs) have been counted. The large market for residential mortgage-backed securities (RMBSs) is not normally included in definitions of real estate as an asset class.

[3] See letter to the Ministry of Finance of 20 October 2006 and Norges Bank Staff Memo No. 2007/1 (Strategy report - The Norwegian Government Pension Fund - Global).