Proposed guidelines for the environmental fund

1. Introduction

In the Revised National Budget for 1999, the Government proposed the establishment of a NOK 1 billion fund that would be subject to specific environmental guidelines. Possible environmental criteria were discussed in the National Budget for 2000, and the conclusion was that initially, the most realistic alternative is to use criteria linked to environmental reporting and environmental certification. The regulation that governs the Petroleum Fund was changed with effect from 1 January 2000 to provide for the establishment of an environmental fund.

The Ministry of Finance is responsible for the management of the environmental fund. This means that the Ministry is responsible for defining the fund's objective, establishing an investment strategy that is consistent with this objective and drawing up guidelines for the management of the fund. In its letter of 16 March 1999, Norges Bank provided a general assessment of the consequences of basing an investment strategy on special environmental considerations. In this letter, Norges Bank will provide advice on how the guidelines for the environmental fund should be formulated.

2. Guidelines

The guidelines for the environmental fund shall provide a precise description of how the fund should be managed. This means that the guidelines must define an investment universe, a benchmark portfolio and establish limits for deviations from this portfolio. The guidelines must also establish rules for updating the investment universe and benchmark portfolio.

A thorough evaluation of the environmental fund's results will be conducted at the end of a three-year period. The National Budget for 2000 stressed that the guidelines for the environmental fund should remain virtually unchanged during this period in order to simplify the evaluation. In the opinion of Norges Bank, this should mean that there are no changes either in the environmental criteria used or in the countries included in the environmental fund during the evaluation period.

Investment universe

When the environmental criteria have been established, individual companies must be assessed in relation to these criteria. This requires knowledge of the individual company's operations. Gathering this type of information requires extensive resources. Therefore, it is necessary to limit the environmental fund's investment universe considerably compared with the existing investment universe of the Petroleum Fund (which comprises more than 20 000 companies in 28 countries). One way of doing this is to start with the companies that are included in the Petroleum Fund's equity benchmark portfolio. This portfolio contains less than 2 000 companies in 21 countries. In some of these countries, however, company information is currently insufficient to justify the application of environmental criteria. This is particularly true in Asian countries. This means that initial efforts at establishing environmental guidelines must be focused on the companies that make up the Petroleum Fund's benchmark indices in Europe, the US and Canada.

According to the National Budget for 2000, environmental criteria should not be applied to all industries, as priority must be assigned to increasing the environmental focus in those industries where environmental problems are most pronounced. In addition, environmental reporting and environmental certification are not especially common in some industries, such as service industries. As a consequence, these industries' share in the environmental fund would be small, since few companies would have satisfied the criteria. To solve this problem, the environmental criteria will only be applied to those industries that have the greatest environmental problems. All companies in other industries will be included in the environmental fund.

When a decision has been made about how the environmental criteria are to be interpreted and to which industries they should be applied, a list of companies that meet the environmental criteria can be prepared. This list will represent the environmental fund's investment universe in the countries concerned and will provide the basis for constructing the fund's benchmark portfolio.

Benchmark portfolio

The benchmark portfolio has two important functions in connection with the management of the environmental fund. In addition to reflecting the investment strategy, the benchmark portfolio shall also be used to assess the management of the fund. Therefore, it is important that this portfolio is constructed on the basis of clear, objective criteria. There are four main questions that should be raised in connection with the construction of the environmental fund's benchmark portfolio: 1) What weight should be assigned to each region? 2) What country distribution within each region should be used? 3)What should be the distribution of companies within each country? 4)How often should the benchmark portfolio be updated? We will now consider these questions in more detail.

Regional distribution

The regional distribution of the environmental fund's benchmark portfolio will depend on which countries are included. Efforts are currently under way to assess the possible use of environmental criteria in Asia and Oceania. So far, however, there is insufficient information to do this. As long as these countries are not included in the environmental fund, the regional distribution in the fund's benchmark portfolio may be adjusted in two ways. One way is to have the same relative increase in weight in Europe, the US and Canada. The other is to have the same absolute increase in weights. Table 1 presents the regional distribution resulting from the two alternatives.

Table 1: Regional distribution for the environmental fund, when Asia and Oceania are not included

Table 1 shows that the two methods result in regional distributions that are fairly similar. To simplify, we suggest using the method of equal absolute increases.

When the regional distribution for the environmental fund's benchmark portfolio has been established, the next question is how these weights should be changed over time. There are two main alternatives. The first is to employ the same principles used by the Petroleum Fund i.e. regional weights are reset to reflect the original distribution at the beginning of each quarter. The second alternative is not to rebalance the regional weights, but to allow them to change in step with market developments in the two regional portfolios. One argument in favour of the second alternative is that the costs involved in rebalancing the weights will probably be greater for the environmental fund than for the Petroleum Fund. There are three reasons for this.

First, it is not as practical to use futures contracts to change exposure in the environmental fund. This is because the use of environmental criteria will result in greater deviations between the market exposure achieved through futures and the market exposure actually preferred. The more major companies in each market that are excluded from the environmental fund's benchmark portfolio, the greater the problem will be.

Second, there will not be regular capital transfers into the environmental fund. This means that in connection with rebalancing, it will be necessary to sell some shares at the same time that others are purchased. This can lead to major transaction costs over time. In the Petroleum Fund, the transfer of new capital will make it easier to purchase the securities needed without necessarily selling other securities. This will reduce the number of transactions as well as the management costs.

Finally, the cost of rebalancing the environmental fund will be high because it will be more difficult to effect transactions without using brokerage firms. The Petroleum Fund managers can reduce transaction costs by means of internal trading of equity portfolios of index managers, since they are managing portfolios held by other major investors. This will be more difficult for the environmental fund if a number of the major companies do not meet the criteria, because we will be making transactions with a portfolio that is less liquid than the Petroleum Fund's equity portfolio.

It follows from the discussion above that in the interest of transaction costs, rebalancing of the environmental fund's benchmark portfolio should be limited. Historically, equity price movements in the two regions have been highly correlated, and if this covariation continues, only minor changes in regional distribution will occur over time, even without rebalancing. Therefore, Norges Bank advises against regular rebalancing of the regional distribution in the environmental fund. The fund should only be rebalanced in the event of significant changes in the regional distribution. We suggest that the environmental fund's benchmark portfolio be rebalanced if the regional distribution deviates more than 10 percentage points from the original weighting.1 Out of consideration for the measurement of returns and the assessment of the environmental fund, any rebalancing should be executed at the turn of the month.

Country distribution

Distribution within each country

The environmental fund's benchmark portfolio in each country can be constructed in two ways. The first method is to use market capitalisation weights. This means that when some companies are removed from the benchmark portfolio for a country, the remaining companies will be given new weights relative to their size. The other method is to weight the remaining companies in such way that the features of the environmental fund's benchmark portfolio in each country are as similar as possible to the features of the Petroleum Fund's benchmark portfolio. A key question is how to weight the different industries in the environmental fund's benchmark portfolio. Since it is not relevant to use environmental criteria in all industries, the composition of industries in the environmental fund may deviate from the composition of industries in the Petroleum Fund. It is possible, for example, to make a correction for this by setting the industry weights in the environmental fund's benchmark portfolio equal to the weights in the Petroleum Fund's equity benchmark portfolio. However, since the composition of companies in different industries will differ in the two portfolios, the industry weights will vary over time. Regular adjustment of the industry weights in the environmental fund's benchmark portfolio to correspond to the weights in the Petroleum Fund's benchmark portfolio will require considerable rebalancing in the environmental fund. We will not recommend such a solution since it will entail major transaction costs. Norges Bank will therefore recommend that in the environmental fund's benchmark portfolio, market capitalisation weights are used in each country and that these weights are calculated solely on the basis of companies that satisfy the environmental criteria.

Norges Bank has asked FT International, the supplier of the Petroleum Fund's benchmark portfolio, whether it is possible to construct and maintain a benchmark portfolio for the environmental fund based on a list of companies that satisfy the environmental criteria. The supplier has concluded that they are willing to do this. This benchmark portfolio can be supplied with the same frequency and detail as the official indices currently used by the Petroleum Fund. However, this will require greater resources on the part of the index supplier and will therefore be more expensive.

Updating and rebalancing the environmental fund's benchmark portfolio

A clarification is needed as to how frequently the benchmark portfolio shall be updated. In principle, frequent updates are desirable, so that the difference between the Petroleum Fund's equity benchmark portfolio (which is the basis for the selection of companies that satisfy environmental criteria) and the environmental fund's benchmark portfolio is only due to the use of environmental criteria. However, there are constant changes in the Petroleum Fund's equity benchmark portfolio. These changes may be due to the inclusion of new companies in the index or the removal of others. There may also be mergers between companies in the index or between a company in the index and another that is outside the index. During the second half of 1999, for example, there were 130 such changes in the Petroleum Fund's equity benchmark portfolio. In general, changes in this portfolio should be reflected as quickly as possible in the environmental fund's benchmark portfolio. This is because a key element in the assessment of the environmental investments will be to look at the difference in return on the two portfolios. Therefore, it is important that the difference between the two benchmark portfolios only reflects environmental guidelines and not differences in the frequency of updates.

The manner in which the environmental fund's benchmark portfolio is updated will depend, among other things, on the industry classification used when differentiating between industries that have environmental problems and those where these problems are so insignificant that all companies can be included in the fund's benchmark portfolio. In a letter dated 13 March 2000, the Ministry of Finance indicates that the industry classifications used in connection with selecting companies for the environmental fund's benchmark portfolio will differ from the industry classification used when companies are selected for the Petroleum Fund's benchmark portfolio. The Ministry of Finance requests Norges Bank's views on this matter. The Bank is of the opinion that it will be easier to update the environmental fund's benchmark portfolio if the same industry classification is used in constructing both the environmental fund's benchmark portfolio and the Petroleum Fund's benchmark portfolio. This will also simplify the assessment of the environmental fund since there will be fewer deviations between the environmental fund and the Petroleum Fund that are not due to the use of environmental criteria.

Assuming that a special industry classification is used in connection with the construction of the environmental fund, two methods may be employed to update the environmental fund's benchmark portfolio. The first is to automatically and immediately replicate all changes in the Petroleum Fund's benchmark portfolio. This will ensure that the only differences in the two indices will be due to active use of environmental criteria. It will take some time, however, to evaluate industry classification for new companies and whether they satisfy environmental criteria. There is a risk, therefore, that it may be necessary to sell equities that have been included in the environmental fund because it turns out that these companies do not belong in this fund. It may also be a disadvantage that for a limited period of time, the environmental fund contains equities that do not meet the environmental criteria.

The other method is not to replicate changes in the Petroleum Fund's benchmark portfolio until it is clear that the new companies satisfy the environmental criteria. One disadvantage of this method is that part of the difference in returns on the two benchmark portfolios will be due to different frequency rates for updating the indices and not to environmental criteria. The longer it takes to evaluate new companies, the greater this problem will be. Another disadvantage is that it may be necessary to sell equities in a company that merges with another company that is not in the environmental portfolio, and then, later, purchase equities in the new merged company if it meets the environmental criteria.

The discussion above shows that both methods have their weaknesses. Norges Bank recommends using the second method. This means that new companies or merged companies are not included in the environmental fund's benchmark portfolio until it is clear that they satisfy the environmental criteria. This will ensure that the environmental fund only invests in companies that meet the environmental criteria. If it is possible to clarify quickly whether new companies satisfy the environmental criteria, the difference in returns on the two benchmark portfolios will only be affected toa limited degree by different updating frequencies.

Deviations from the benchmark portfolio

The Ministry of Finance has imposed restrictions on Norges Bank's management of the Petroleum Fund. These restrictions limit the deviation allowed between the return on the Fund and the return on the benchmark portfolio. The Ministry has done this by setting a maximum limit for tracking error, which is defined as the standard deviation of the difference between the return on the Petroleum Fund and the return on the benchmark portfolio. When the permitted deviations from the benchmark portfolio are limited by setting an upper limit for tracking error, the actual return will vary within a range around the return on the benchmark portfolio. The lower the limit for tracking error, the narrower the range will be. The Ministry of Finance has set the limit to 150 basis points for the Petroleum Fund.

Different management methods with different degrees of deviation from the benchmark portfolio are used in managing the Petroleum Fund. The lowest deviation results in index management that is aimed at replicating the benchmark portfolio. A reasonable limit for tracking error here may be 50 basis points or lower. However, in the management of index portfolios, it may be appropriate to develop more advanced management methods that seek to minimise transaction costs at the same time that the portfolio's return can deviate slightly from the return on the benchmark portfolio. This may be accomplished by not including all companies in the actual portfolio and by making adaptations in the portfolio at slightly different times compared with changes made in the index portfolio. These methods require higher tracking error than pure replication of the index. The latter form of management is active management and often entails major deviations in relation to the benchmark portfolio. Norges Bank's mandates for external active portfolio management have set a target range of 300 to 700 basis points for tracking error.

It would not be appropriate to combine several forms of management in connection with the environmental fund since it is a relatively small portfolio. This is due in part to the fact that the investment universe for the environmental fund is considerably smaller than that of the Petroleum Fund. A smaller investment universe reduces the possibilities of selecting individual equities and thus also the probability of active management generating excess returns. Therefore, Norges Bank recommends that the fund should be managed on the basis of an indexing strategy that entails a fairly limited deviation from the benchmark portfolio. The limit for tracking error should permit the use of more advanced indexing strategies for the environmental fund. For example, index management of the environmental fund will probably be most cost-effective if the managers do not purchase all the equities in the benchmark portfolio. This is because the transaction costs will be disproportionately high in connection with the purchase of equities in the smallest companies. For this reason, Norges Bank recommends that the limit for tracking error between the environmental fund and its benchmark portfolio be set at 100 basis points.

3. Other conditions

Assessment of the environmental fund The Revised National Budget for 1999 stipulated that the results of the environmental fund shall be thoroughly assessed at the end of a three-year period. A number of aspects relating to the management of the environmental fund will be interesting to assess.

A key element in the assessment will be to investigate the impact of environmental criteria on the fund's return. Such an assessment cannot be made, however, by comparing the return on the Petroleum Fund and the return on the environmental fund. There are three reasons for this. First, the environmental fund's investments will be limited to equities, whereas the Petroleum Fund invests in both fixed income instruments and equities. Second, the environmental fund will not have equity investments in all the countries represented in the Petroleum Fund. Finally, the return on the two funds will be influenced by the deviations the managers choose to make in relation to the respective benchmark portfolios. Since the deviation limits are greater for the Petroleum Fund than for the environmental fund, the deviations will also be different. This means that an assessment of the return must be based on a comparison of the return on the environmental fund's benchmark portfolio in Europe and the US and Canada and the return on the Petroleum Fund's equity benchmark portfolio in the same regions. It may also be interesting to make more detailed analyses that compare the return on the environmental fund's benchmark portfolio and the Petroleum Fund's benchmark portfolio in the individual countries and the industries where environmental criteria are used. It is important to note that regardless of which differences in return are studied, there will be uncertainty as to whether we can draw any definite conclusions about these differences after the three-year period. This will depend on how great the differences are and how much they vary over time.

It will also be natural to evaluate the risk in the environmental fund in relation to the risk in a corresponding portion of the Petroleum Fund's equity portfolio. The risk differential between the two portfolios will depend on the number and size of the companies that are excluded from the environmental fund. The risk will also be affected by changes in industry composition when environmental criteria are used to exclude companies.

The quality of selection of companies in the environmental fund should also be assessed. Since the consulting company responsible for this selection will play a key role in working with the environmental fund, it is important to assess thoroughly the company's working methods. It is important to examine the company's sources, its analysis of information and the quality of the actual selection process. It will also be natural to review the relevance of the environmental criteria that are used and whether the composition of companies in the environmental fund correspond to expectations at the time the environmental criteria were selected. The Revised National Budget for 1999 stipulated that the guidelines for an environmental fund must be clear, consistent and practicable. It will also be natural to assess whether these requirements have been met.

A third aspect that should be evaluated is the costs associated with the management of the environmental fund. In the preceding, we have argued that the benchmark portfolio should be constructed in such a way that rebalancing costs are limited. However, there will also be other costs that should be assessed more thoroughly. First, the consulting company responsible for the selection of companies receives a fee. Second, there will be costs involved in preparing a tailor-made benchmark portfolio for the environmental fund. Third, it will be natural to look at the costs of buying equities for the environmental fund in relation to corresponding costs for the Petroleum Fund. Finally, the use of resources at Norges Bank in connection with the operational management of the environmental fund should be assessed.

Transfer of capital to the environmental fund
Norges Bank recommends that the environmental fund be established at the turn of a quarter. Since capital is transferred to the Petroleum Fund at this time, some of this capital can be allocated to the environmental fund. The simplest will probably be a cash allocation that Norges Bank uses to purchase the equities that will be included in the environmental fund. Since this may take some time and entail some transaction costs, it may be sensible to postpone the assessment of investment performance until the environmental fund's equity portfolio has been constructed.

Use of instruments

The Petroleum Fund uses equity futures to control the fund's exposure. To ensure effective operational management, the use of equity futures in the environmental fund should also be permitted. This is because the environmental fund will receive liquid funds on a regular basis through dividend payments or the sale of equities that are removed from the environmental fund's benchmark portfolio. It will be very costly to continuously invest this capital in physical equities. A better alternative will be to give this capital an equity exposure through futures contracts for a limited period. Since equity futures provide a return exposure to a different group of companies than those represented in the environmental fund, the use of futures will only be of interest for a small portion of the environmental fund for limited periods.

Management agreement

The management agreement currently in force between the Ministry of Finance and Norges Bank must be expanded to include the management of the environmental fund. Among other things, the agreement should stipulate how Norges Bank's costs in connection with the management of the environmental fund should be covered. In addition, the agreement must stipulate how often and in what way the bank shall report on the management of the environmental fund.

4. Summary and recommendation

Norges Bank recommends that 60 per cent of the environmental fund's benchmark portfolio be allocated to Europe and 40 per cent to the US and Canada. The regional distribution should not be rebalanced as long as the regional weights deviate less than 10 percentage points from the original weights. We recommend that market capitalisation weights be used within each region. If the benchmark portfolio is constructed in this manner, rebalancing costs will be low.

The limit for tracking error between the environmental fund and its benchmark portfolio should be set at 100 basis points to provide for the efficient management of the environmental fund by Norges Bank.

We further recommend that capital is transferred to the environmental fund in connection with the quarterly transfers to the Petroleum Fund, but that investment performance is not assessed until the environmental fund's equity portfolio has been constructed.

Svein Gjedrem

Harald Bøhn