Government Pension Fund Global Guidelines for Observation and Exclusion
25 October 2018
We refer to the Ministry’s letter of 28 June 2018 following up Report to the Storting No. 13 (2017-2018) “The Government Pension Fund 2018”. In its letter, the Ministry refers to the Storting’s consideration of the white paper, the Standing Committee on Finance and Economic Affairs’ Recommendation 270 S (2017-2018), and the Storting’s resolution concerning the coal criterion in the Guidelines for Observation and Exclusion: “The Storting asks the Government to present an assessment in next year’s white paper on the Government Pension Fund Global of whether the current criteria for the exclusion of coal companies from the GPFG are adequate with a view to excluding companies with significant coal-related operations”. The Ministry also refers to the Storting’s resolution concerning gambling companies in the fund: “The Storting asks the Government to explore the basis for divesting the Government Pension Fund Global (GPFG) from gambling companies with a view to an assessment in connection with the annual white paper on the fund”, cf. the Standing Committee on Family and Cultural Affairs’ Recommendation 242 S (2017-2018).
The Ministry asks Norges Bank to describe its work on assessing the observation and exclusion of companies under the coal criterion. The Bank is also asked to describe the extent of coal-related operations at companies excluded or placed under observation under this criterion and, as far as is possible, at the other companies in the fund. Finally, the Bank is asked to report the extent of gambling companies in the fund’s benchmark index and describe their operations, including differences between the companies.
The coal criterion in the Guidelines for Observation and Exclusion
The product-based coal criterion came about as a result of the Standing Committee on Finance and Economic Affairs’ comments in Recommendation No. 290 S (2014-2015) and the consideration of the National Budget for 2016, where the Storting endorsed the Ministry’s proposals for how the criterion could be operationalised. The Ministry introduced the coal criterion with effect from 1 February 2016.
The criterion states that mining companies and power producers which themselves or through entities they control derive 30 percent or more of their revenue from thermal coal, or base 30 percent or more of their operations on thermal coal, may be excluded from the fund. Besides a company’s current share of revenue or operations based on thermal coal, our assessment is required to attach importance to forward-looking assessments, including any plans to reduce the share of revenue or operations based on thermal coal and/or increase the share of revenue or operations based on renewable energy sources.
The Guidelines state that observation may be decided on “when there is doubt as to whether the conditions for exclusion are met or as to future developments, or where observation is deemed appropriate for other reasons”. Observation is a relevant option in situations where companies are considered to exceed the relevant thresholds, but where we have information on stated plans, initiatives or other material factors that make it likely that the company will fall below the thresholds within a reasonable period of time. Examples of such plans, initiatives and other factors include:
- Planned or recent purchases or sales of companies and assets
- Publicly communicated plans to shut down, start up or alter production capacity
- Droughts, accidents and other special events that may have affected the fuel mix
Work on the coal criterion at Norges Bank
After the criterion was introduced in 2016, Norges Bank established a process for identifying companies covered by it and preparing recommendations on observation and exclusion for consideration by its Executive Board.
The first step was to identify a universe of companies for further data collection and analysis. We began by taking the most relevant sectors in the Industry Classification Benchmark (ICB).1We also used other sources of information to identify further relevant companies. This enabled the relevant universe to be whittled down from the full 9,000 or so companies in the fund to 417 companies. Of these, 216 were selected for more detailed review.
After this narrowing of the universe, we obtained information about the companies. Information on mining activities is often more readily available than information on power production. With mining companies, accounting data is often sufficient for assessing thermal coal’s share of total production.
With power producers, further analysis is generally needed. We collect data from a variety of sources covering both revenue and the amount of coal used in power production. Sourcing information of sufficient quality and detail for the operationalisation of the criterion is a challenge. There are no single sources covering all relevant companies, and the information reported by companies themselves is not normally detailed enough for the analysis required. We have therefore contacted multiple suppliers. While the average levels from different data sources are comparable, the individual data points can vary substantially. Reporting across different time periods and metrics is a particular challenge. All of the data we use are systematically structured and stored. Our data sources include suppliers of market data, internal analyses and investment banks. Information and analyses from our external managers have also been important when assessing companies in emerging markets.
We also contact companies for more information. In some cases, companies themselves will be the only source of relevant information on both the current situation and their plans. In the forward-looking assessment, information provided by the companies – such as plans to change their fuel mix or develop new renewable energy capacity – will be particularly relevant.
After this data collection, analysis and company contact, Norges Bank Investment Management prepares recommendations for consideration by the Executive Board. To date, a total of 71 companies have been excluded under the coal criterion, and 15 placed under observation. In line with the Guidelines for Observation and Exclusion, the grounds for these decisions are made public. Norges Bank has published decisions on observation and exclusion under the product-based coal criterion in various tranches in 2016, 2017 and 2018.
Implementation has taken time and resources, but has also built on existing systems and expertise. Proximity to investment management has been important in this work. For example, it has been important to have up-to-date information on corporate actions. Purchases and sales of assets and subsidiaries can have a major bearing on whether or not a company is covered by the criterion.
The product-based coal criterion has accounted for around half of all decisions on observation and exclusion taken in the fund’s history. The design of the criterion, with the emphasis on thresholds and forward-looking assessments, requires extensive work on gathering and analysing information. The criterion also targets industries with high levels of corporate actions, such as purchases and sales of assets and subsidiaries, which necessitates continuous market surveillance, even after companies have been analysed and decisions taken. We also need to analyse any new companies entering the market. In this context, we would also point out that any reduction in the criterion’s thresholds would make the above challenges greater, and transaction costs higher, because more companies would move in and out of the investment universe.
One experience from work on the product-based coal criterion is that company contact has been essential for obtaining a sound basis for decisions, cf. the general requirement in Section 6(5) of the Guidelines to “ensure that sufficient information is available before making each individual observation, exclusion or revocation decision.”
In its letter of 28 June, the Ministry refers to a letter from the Council on Ethics of 12 February 2018 recommending “various changes to the Guidelines for Observation and Exclusion from the GPFG, partly concerning the need to invite comments from companies ahead of exclusion under the product-based criteria.” The Council argued that, when it comes to the product-based criteria, the draft recommendation “could often constitute confidential inside information” and so “the requirement to present it to the company for comment should not apply in these cases.”
In the 2018 white paper on the fund, the Ministry took on board the Council’s recommendation, and the Guidelines were amended from 28 June 2018 so that the requirement to present the draft recommendation for comment applies only to the conduct-based criteria. Section 5(4) now reads: “A company that is being considered for observation or exclusion shall be given an opportunity to present information and opinions to the Council at an early stage of the process. In this context, the Council shall clarify to the company what circumstances may form the basis for observation or exclusion. If the Council is considering recommending observation or exclusion pursuant to section 3, the draft recommendation shall be presented to the company for its comments; cf. section 7.”
The issue here is whether information about different steps in the process for observation and exclusion can be considered liable to impact materially on the prices of financial instruments. Norges Bank’s view is that the existence of a draft recommendation will not generally be liable to have such a price effect. In this context, we would also point out that the product-based criteria are publicly available. We are dependent on obtaining information from companies themselves in our implementation of the coal criterion, and we will continue to contact companies as before.
The Guidelines for Observation and Exclusion make special provision for the treatment of green bonds, stating that recommendations and decisions to exclude companies under the coal criterion do not apply to any green bonds issued by those companies that are included in specific indices for such bonds or verified by a recognised third party.
The technical implementation of this exception has proved complex and challenging. The systems at the fund’s bond index supplier identify individual companies’ instruments on the basis of a ticker code which covers all bond issuances within a company structure. These systems are not designed to handle the inclusion or exclusion of individual bonds issued by the same company.
When deciding on the exclusion of companies under the coal criterion, we have compared the list with the universe of green bonds published for the BarCap MSCI Green Bond Index. No overlap has been identified at the time of the decisions. A number of such bonds have, however, been issued after the decisions were taken. These green bonds have not been excluded from our managers’ investment universe, cf. the formulation of the criterion in the Guidelines. They are not, however, included in the benchmark index due to the technical challenges mentioned above. This applies to only a small number of bonds, with the result that the benchmark index’s reported return has not yet differed from a benchmark with these bonds included.
Extent of coal-related operations
In its letter of 28 June, the Ministry asks the Bank to describe the extent of coal-related operations at, and size of, companies excluded or placed under observation under this criterion. The extent of coal-related operations at mining companies and power producers that have been excluded or placed under observation can be described by looking at the production of thermal coal and coal-based power production capacity respectively.
We estimate that production of thermal coal at companies excluded or placed under observation was around 1,400 million tons in 2017, and that capacity for coal-based power at power producers excluded or placed under observation is around 574,000 MW. The figures for production capacity are based on the latest available data from the World Electric Power Plants database from data provider Platts Energy.
The Ministry also asks the Bank, as far as is possible, to report the extent of coal-related operations at, and size of, mining companies and power producers that themselves or through entities they control derive part of their revenue from, or base part of their operations on, thermal coal, but have not been excluded or placed under observation under the coal criterion. Information is further requested on the number of such companies, and on any plans at these companies to change their share of revenue and/or operations based on thermal coal.
Companies in the fund that have not been excluded or placed under observation may also have coal-related operations. It is difficult to describe or estimate the extent of such operations with any precision. One challenge is that companies use different metrics in their reporting, and we cannot compare relevant data without detailed analysis of each individual company. It should also be noted that when we look at companies that derive less than 30 percent of their revenue from thermal coal, or base less than 30 percent of their operations on thermal coal, this will include a large number of companies in a range of different sectors. These companies may have only a very small share of coal-related operations. They will also, for example, include one of the fund’s largest investments in the consumer goods sector.
We can nevertheless obtain an estimate of power production by comparing companies that have not been excluded or placed under observation with data from the World Electric Power Plants database. It can then be estimated that around 206,000 MW of coal-based production capacity has not been excluded or placed under observation. Based on these figures, we can estimate that companies accounting for 74 percent of total coal-based power production capacity have been excluded from the fund or placed under observation.
We can also make an estimate of production of thermal coal at companies in the fund that have not been excluded or placed under observation. As no centralised database is available here, we have looked at information disclosed by 39 companies we have identified as having coal production as part of their business. Companies in the fund that have not been excluded or placed under observation produced an estimated 430 million tons of thermal coal in 2017. Based on these figures, we can estimate that companies accounting for 77 percent of total thermal coal production have been excluded from the fund or placed under observation.
The Ministry asks the Bank to report on the extent of gambling companies in the fund’s benchmark index and describe their operations, including differences between the companies. This raises issues of definition and delimitation.
For equity investments, the Industry Classification Benchmark (ICB) defines the “gambling” subsector as “Providers of gambling and casino facilities. Includes online casinos, racetracks and the manufacturers of pachinko machines and casino and lottery equipment”. Companies that derive the bulk of their revenue from such activities will be assigned to this subsector.
When it comes to investments in corporate bonds, we have looked at the industry classification from the index supplier Bloomberg Barclays2. This includes the “gaming” subsector, without further definition. We have, however, had confirmation from the index provider that the subsector covers companies involved in gambling and casinos. Companies that derive the bulk of their revenue from such activities will be assigned to this subsector.
As at 30 June 2018, the fund’s benchmark index included 56 companies with a market value of 24 billion kroner in the ICB gambling subsector, and bonds from one company with a market value of 46 million kroner in the Bloomberg Barclays gaming subsector.
In addition to these companies in the benchmark index, the fund’s portfolio at the same date included shares in a further 13 companies with a market value of 749 million kroner in the ICB gambling subsector and bonds from one company with a market value of 384 million kroner in the Bloomberg Barclays gaming subsector.
There will probably be further companies in the benchmark index and the fund’s portfolio that derive a smaller share of their revenue from gambling as defined by the ICB or Barclays and have therefore been assigned to another sector. It is reasonable to assume that some companies in the ICB subsectors of toys, recreational services (including cruise lines), financial administration, electrical components & equipment, and hotels will derive small parts of their revenue from gambling.
The fund’s index suppliers have not further described the companies included in their gambling/gaming subsectors, nor the differences between them. From the list of companies included in the subsector, however, we can see that there is considerable variation. Some companies are operators of physical games, while others are more Internet-based. Others again are subcontractors supplying equipment, services or software to the industry.
We should also point out that these index suppliers’ classification of gambling companies does not take account of the companies’ purposes or customer base. For example, the ICB classification includes companies that supply state gambling companies and lotteries in Europe.
Øystein Olsen Yngve Slyngstad
Download the letter, with enclosure (PDF)
1ICB is the sector classification used by our index supplier FTSE.
2Bloomberg Barclays Global Sector Classification Scheme (BCLASS).