We recognise that corporations are manifestly involved in the progress of policy and legislation formation (in all areas, not only climate change) affecting their business and they regard the need to do this as part of their operating model. Research has indicated that this influence likely extends beyond the activities normally associated with the word "lobbying" (e.g. donations to clearly motivated political actors) and includes the domination of the public discourse on climate change science and policy via their hugely powerful and funded messaging tools (e.g. advertising, PR, social media, access to influential meetings) as well as the use of influencers like trade associations and advocacy groups. Details of our analysis of corporate influence in climate change policy can be found here.

The US Lobby Act of 1995 places certain disclosure requirements on lobbyists and corporations and provides for official databases, which are in part used by well respected NGOs such as OpenSecrets.org. We do not uses these extensively for two reasons. Most importantly, the motivation of the lobbying spend is not sufficiently clear for InfluenceMap to link to specific climate change regulatory strands. Most global multinationals fund a range of political targets for a range of reasons and we do not feel the links are solid enough for this to be one of our data sources. Secondly our analysis of corporate influence stretches far beyond traditional lobbying spend. CEO messaging, trade association capture of technical details of regulations, public facing capture of the climate change debate and other means may be equally or more effective tools. That said, we will use this data if carefully and convincingly analysed by the media or research group so the links to climate change policy are clear.

We recognise that China accounts for roughly a quarter of global GHG emissions and that policy and regulations from its governmental system, along with adherence by the corporate sector is extremely important to a low carbon future for the planet. Most of this corporate sector, especially the fossil fuels, chemicals and automotive sectors remain under government control if not majority ownership. So the meaning of corporate influence over climate policy requires additional analysis and detailed in-country research of China state-owned corporations.

We prioritize companies based on economic size as determined by the Forbes Global 2000 not including state owned enterprises (e.g. Sinopec). An analysis of corporate influence on policy by a state owned enterprise is not clear, hence our omission for the time being. In addition, we include more than 100 private sector trade associations, federations and advocacy groups that represent the interests of companies, based on a mix of criteria (a) their influence on policy in general and size and (b) how active they are on climate policy and legislation, both assessed by numerical metrics.

Climate change policy and regulations will increasingly affect all corporations regardless of sector, but clearly some will be affected more than others depending on region and regulation. Our performance bands are perhaps best used when corporations from the same sector are compared. A great example is the automotive sector, where manufacturers who are strategically positioning themselves for a low carbon future will support accelerated CO2 emission standards as bolstering their competitive position. Similarly comments can be made about energy intensive users like chemicals, building materials and utilities. Comparing the InfluenceMap performance band of BMW with that of Google, in isolation of information on how these companies perform with respect to their direct competitors may not be as meaningful.

Our scores may also be regarded as a measure of a corporation's readiness in the case of a shift towards a low carbon regulatory regime, based on the assumption that their support for this originates from a forward thinking competitive strategy. For example, a corporation who is actively supporting low carbon laws and measures may also be engaged in a strategic shift in its own activities in this direction, thus potentially giving it competitive advantage if legislation should move swiftly to disfavor GHG emissions. Conversely a corporation engaged in overt or covert obstruction of climate change policies may be unlikely to be strategically shifting its activities in a similar manner. This analysis may be pronounced in sectors which will be heavily affected by a fast shift to a low carbon regulatory regime but are not engaged in selling GHG emitting fuels directly, such as the automotive, electric utilities and chemicals processing related sectors. Moreover, some research suggests that political donations are followed by decreased shareholder returns.

For Corporations, the organizational scores and relationship scores are combined to result in an Overall Rating that places the Corporation in a performance band. There are 20 performance bands from A+ (representing a total score from 95-100%) through to E- (a score of 25-30%), with scores below 25% falling in the red "F" band. So it is a relative scale. Most companies fall within the 45-55% band that indicates they are neither actively obstructing nor supporting the climate policy process. A score of 60% or more indicates the company is actively supporting polices and regulations towards a low carbon future. Conversely lower than a 40% band generally indicates obstructing behavior. As different sectors face differing regulatory issues, it is most useful to compare the performance bands of corporations within the same sector.

A corporation, as well as its organizational score will have a relationship score. It is computed by aggregating the organizational scores of the Influencers (trade bodies etc.) it has relationships with, weighted by both the strength of these relationships and the relative importance of the Influencers towards climate change policy. A corporation will have a number of relationships with Influencers mapped in our database. In turn these Influencers will themselves have organizational scores as computed by exactly the same method as was done for the corporations. Such relationships vary in nature and strength and we account for this in our assessment by assigning relationship strengths. For example a trade association may have 2000 member corporations with 10 of them on its executive committee. The 10 executive committee members would have strength of 8 compared to 3 for the regular members, for example. The relationship score is expressed as a % with 100% representing very supportive influence on climate policy. In general if a corporation's organizational score is roughly equal to its relationship score the corporation's positions and interactions with climate policies are in line with those of the key influencers it has relationships with. If the organizational score is much higher than its relationship score, the trade bodies and other influencers the corporation is involved in are misaligned with the corporation on climate policy.

To measure and score corporate influence on climate change policy, we have developed a comprehensive process of examining a series of publicly available and reliable data sources (e.g. legislative consultations, social media, advertising, reliable media, top management messaging, financial disclosures, CDP responses) and querying them, for each corporation, against a set of climate change policy and legislation related queries (e.g. position on a carbon tax, energy efficiency standards). We score each data source/query intersection (or cell) on a 5-point scale, with clearly consistent evidence and guidelines for our scorers. There are 96 such scoring cells at present and the organizational score for each corporation is computed over these, by our proprietary algorithm that accounts for weightings and irrelevant data sources/queries. The organizational score is expressed as a % with 100% representing very supportive influence on climate policy. We note here that a corporation and an influencer (e.g. a trade body, chamber of commerce, advocacy group) are scored in exactly the same way to arrive at their respective organizational score.

If a particular query does not apply to a data source (e.g. external data sources do not apply to transparency queries), OR for a particular corporation the data is unavailable (e.g. due to our research limitations, or it does not disclose through a voluntary scheme like CDP), we mark the weighting cell with "NA" (not applicable), and the weighting is set to zero. Its original weighting is redistributed evenly through other data points for this query. If no evidence is available (after a search) for a particular cell, it is marked "NS" or not scored with the same impact on the weightings as NA. So each organization has its own specific weighting matrix depending on data availability and relevance. This means that a company will not be impacted by its lack of involvement in policies that do not affect it and conversely will be more accurately assessed for those that do.

Our analysis and scoring is focused on an organization's comments, interactions and influence on policy and legislation. We do not consider internal strategy, activities and performance of a company on climate change related issues, such as CO2 emissions, use of various energy forms and the company's business activities if these have no direct relevance to policy and legislation forming. There are numerous other research streams (e.g. Carbon Tracker, CDP) dealing with corporate performance on climate and we do not wish to widen our remit where adequate coverage exists.

To measure and score corporate influence on climate change policy, we have developed a comprehensive process of examining a series of publicly available and reliable data sources (e.g. legislative consultations, social media, advertising, reliable media, top management messaging, financial disclosures, CDP responses) and querying them, for each corporation, against a set of climate change policy and legislation related queries (e.g. position on a carbon tax, energy efficiency standards). We score each data source/query intersection (or cell) on a 5-point scale, with clearly consistent evidence and guidelines for our scorers. Historical evidence is weighted less than more recent evidence. Evidence more than three years old does not significantly contribute to an orgnaization's score. It is left archived on the profile to allow users to see how the narrative of the organization has evolved.

A process of global, regional and national consensus building, policy formulation and legislation/fiscal measures on climate change has been underway for the last two decades and is on-going. In our assessment of corporate influence, we consider the process from consensus forming on climate science to legislative interventions at various levels and take this process to be "climate change policy". We consider regulations/laws and guidelines as well as fiscal interventions such as carbon taxes and continuation of subsidisation of coal, for example. Importantly, we consider amendments to mainstream regulations and fiscal policy driven by or affecting climate change.

We stress here that our assessment and scoring platform does not claim to measure actual corporate influence over the climate change policy process. Such an assertion would need to involve vastly more cause-and-effect substantiation, which we do not attempt. Rather we assess and score corporations on a variety of activities that respected authorities (e.g. the UN's Caring for Climate) have asserted are highly likely to influence policy. On this site when we use the term "corporate influence" it should be regarded in this context.

Service sectors like retail and telecommunications do not have nearly as much engagement and influence (whether positive or negative) on climate policy as the energy and manufacturing sectors. So there are very few data points on which to base our query and score. Hence a single piece of evidence on a particular query (e.g. disclosure) could distort the organisation's score and performance band. We recognise ours in an imperfect system in this regard and encourage users to compare companies within sectors and also examine the pieces of evidence we have collated rather than focus too much on the overall score. We include an engagement intensity metric to provide a measure of the extent to which the company is engaging on climate change policy matters. It is a number from 0 (no engagement at all) to 100 (full engagement on all queries/data points). Clearly the fossil fuel value chain companies will have a far greater level of engagement that retailers.

The engagement intensity (EI) is a metric of the extent to which the company is engaging on climate change policy matters, whether positively or negatively. It is a number from 0 (no engagement at all) to 100 (full engagement on all queries/data points. Fossil fuel value chain companies are more affected by climate regulations and will have a higher EI than, for example retailers. So an organization’s score should be looked at in conjunction with this metric to gauge the amount of evidence we are using in each case as a basis for scoring. On our scale, an EI of more than 35 indicates a relatively large amount of climate policy engagement.

Our database contains over 200,000 pieces of evidence backed up by what we consider respected publicly available external sources relevant to our subject matter and the organizations we are ranking. We add to this database daily as new information is available and our scores change in real time. We welcome input from interested parties should any typographical errors, mistakes, perceived misinterpretation of data or other issues be noted, and we stand ready to review our information as comments are made available. We further point users to our Terms and Conditions for issues relating to the appearance of external data on our site and reliance on our platform.

We have a system of receiving alerts from hundreds of web sites with our preset searches and monitor these alerts daily, adding evidence and revising scores in real time. We also encourage interested and knowledgeable parties to comment on and contribute to our website and we will add to our database provided credible documentation exists.

FinanceMap is a platform generating metrics and analysis on commonly held investment portfolios (e.g. listed funds) through the lens of climate change, facilitating the inclusion of this key risk area in investment decision making. What makes this unique is that our metrics will be made publicly available. This will allow use by all parts of the investment chain, including small asset owners. It will also prompt better disclosure by the financial sector and drive competitive improvement among asset managers in the products they create and sell.

InfluenceMap will use the Sustainable Energy Investing Metrics and other relevant methodologies for assessing the portfolios of the world's listed funds and key asset managers. The data will be made public to allow fund buyers and asset owners to easily incorporate climate risk into their investment decisions. It will also be driven competitive improvement within the fund management industry. The resulting metrics will cover individual companies as well as portfolios, for example in this analysis of the fossil fuel production sector. InfluenceMap will also generate media and business focused reports deriving from our data platform to maintain awareness of climate risk in finance. Our launch report is entitled Who Owns the World's Fossil Fuels.

InfluenceMap is collaborating on this work with the 2 Degrees Investing Initiative, convenor of the Sustainable Energy Investing Metrics program that drives our assessment methodology. Another key collaborator is the WWF European Policy Office. We are also working closely with dozens of asset owners, financial data providers and financial consultants to ensure our work is accurate and taken up by users. The work is funded by the KR Foundation and Climate-KIC.

While our FinanceMap project will consider all industrial sectors sensitive to climate risk, the Who Owns the Fossil Fuels media release is an output based on the initial phase of this project. It demonstrates our mapping of the finance sector and considers a key component of portfolio climate risk - the fossil fuel production companies. We have mapped out fossil fuel ownership trends globally and the resultant back-end database has been in use by the climate and asset owners communities since May 2018, with the media report released in late November 2018.

The research considered a group of listed companies who control the largest amounts of fossil fuel (thermal coal, oil, gas) reserves and production. This includes large conglomerates such as Japan's Itochu, for whom fossil fuel production is only a minor business but whose size renders them a substantial player. This research defines these companies as the "largest" based on an aggregation of their sales, market cap and the amount of physical reserves and production they control.

These roughly 300 companies control more than 95% of all oil, gas and thermal coal reserves within listed companies and represent roughly $5 trillion in combined market capitalization as of Oct 2018. It should be stressed that this $5 trillion in market value is the aggregate of all the companies, not the value attributed to their fossil fuel production business. As noted, companies like Itochu and Berkshire Hathaway maintain a relatively small share of their business in fossil fuel production. When weighted for % of sales due to fossil fuels, the value within the fossil fuel business of these 300 companies is closer to $2 trillion.

Therefore, this group of companies is important for investors looking to utilize their power as shareholders to address climate change. All data on these 300 companies is based on the latest financial and annual report disclosures from the companies themselves (in most cases, as of end 2017).

The majority of proven reserves of oil and gas are controlled by non-public entities, such as Saudi Aramco. However, roughly half of all thermal coal reserves are controlled by publicly listed producers, and thus are held by the world's capital market investors. This research covers thermal coal only and excludes the significant amount of metallurgical coal mined for industrial use. This data will be updated annually, next in mid-2019.

The process of classifying and accounting for fossil fuel reserves varies considerably between companies, as well as between different authorities that require this disclosure. Where possible, this research takes oil/gas reserve data as classified under the 1P (proven), 2P (proven + probable) and 3P (proven + probable + possible) Security Exchange Council (SEC) or Society of Petroleum Engineers PRMS system. Where this has not been possible (e.g. for many of the Russia-based companies) the disclosure has been converted as accurately as possible to match the 1P, 2P, 3P system. While this research has collated data on 3P reserves where disclosed, this data is not currently available online nor does it feed into the computations. Where a company has not disclosed 2P data this has been set to zero, so that in these cases 2P is taken to be equivalent to 1P. This research maintains a single reserve number for proven coal reserves. Thus, 2P is used in allocating oil/gas reserves to shareholders while the single proven reserves number is used for thermal coal.

Data on the coal, oil and gas reserves and production for each company is taken from annual reports or financial disclosures. The quality and structure of disclosure by fossil fuel companies on their reserves varies by region and company. This research deploys a flexible tagging system to identify the location and characteristics of reserves (e.g. method of extraction, type of oil) for each company. We recognize that due to inadequate disclosure of reserve details by the sector, large investors and industry players may use proprietary databases (such as woodmac.com) which are not public. A secondary goal of collecting and publicizing this data is to challenge the fossil fuel sector to make more precise disclosure on its reserves.

To arrive at the potential CO2 emissions equivalent in gigatonnes (GtCO2), reserve totals are multiplied by an emissions factor specific to the type of reserve, which accounts for vented, flared and fugitive emissions (Heede et al, 2015). The IPPC 5th Assessment total 2C carbon budget represents the total emissions that can be released to give a 66% chance of remaining within a 2C temperature change relative to pre-industrial levels, while the IPPC total 1.5C carbon budget (as stated in the October release Global Warming of 1.5 °C) provides a current estimate for keeping temperatures within 1.5C. As well as oil and gas reserves, this research covers thermal coal only and excludes the significant amount of metallurgical coal mined for industrial use. It also recognizes that a portion of extracted oil and gas is used for non-fuel combustion but does not attempt to analyze this in depth. An October 2018 IEA report The Future of Petrochemicals estimates that in 2017 12% of crude oil was used for petrochemicals, a proportion expected to rise rapidly through to 2050 as transport fuel use declines.

Heede's methodology used for CO2 equivalent emissions factors from coal, oil, gas used in this repprt notes "Other estimates of potential emissions from reserves (IPCC, IEA, Carbon Tracker) assume that all of the carbon in the fuel reserves is combusted to the atmosphere. This study makes the more realistic assumption that not all carbon in fuel reserves is burned; we deduct for carbon in the products used for non-energy purposes, such as waxes, lubricants, petrochemicals, carbon fibers, pigments, fertilizers, steelmaking, and road oil. The methodology also accounts for emissions from subsequent combustion of non-energy products, such as tyres, waxes, lubricants, and plastics."

While the fossil fuel companies directly control fossil fuel reserves, these companies are owned by shareholders who in theory can impact the management of these reserves. The method this research deploys is to compute the reserves each shareholder effectively "owns" through their aggregated holdings. Two metrics are computed. One is simply expressed in tons of coal or BOE (barrels of oil equivalent) for oil/gas. This metric is useful as it can be directly correlated to greenhouse gas emissions should these reserves be combusted. The other metric is the monetary value of these aggregated reserves in US$. For each company, the value of its fossil fuel reserves is calculated by multiplying its market capitalization by the percentage of sales it accrues from the fuel. This percentage value is obtained from company disclosures. For example, this research estimates that ExxonMobil gains 30% of its sales from oil and hence the value of its reserves are 30% x $360 bn, or $108bn. A shareholder holding 2% of Exxon's shares will therefore be allocated $2.16bn in oil reserves value from its ExxonMobil holding.

At present, we do not track changes in reserves of the fossil fuel companies through time. Corrections in reserves attributed to shareholders as a result of any increase or decrease in reserve levels from 2016 to end 2018 are therefore not at present included in the system. The initial assessment of any resulting errors indicates that this correction is likely not material to the results (i.e less than 10%). The exception to this is thermal coal reserves, wherein we have tracked major changes in companies like Arch Coal, Rio Tinto and Glencore over the 2016-18 period and these adjustments are reflected in the reserves allocated to shareholders of the companies in the 2016-18 time span.

Clearly, the aggregated $ value of a shareholder's reserves in oil/gas/thermal coal will vary with both the shareholder's holdings and with the value of the underlying companies. However, the physical quantity of the shareholder's aggregated reserves (tons, or BOE) will only vary with changes in the shareholder's holdings. Therefore, the change in this metric is a useful indicator of how the shareholder is adjusting their portfolio in terms of thermal coal, gas or oil assets. Tracking these trends is useful for the Fossil Fuel Divestment movement, which is pressuring Asset Owners and Asset Managers to remove fossil fuel holdings from their portfolios.

Despite being commonly used, the terms 'oil' and 'gas' cover various types of hydrocarbons, as defined by different physical properties. It is therefore not always immediately clear what these terms specifically refer to. This research uses the following classifications:

Gas: Consisting primarily of methane, this refers to natural gas derived from both conventional and unconventional extraction techniques. This includes shale gas, tight gas, associated gas, non-associated gas, coal bed methane (coal seam gas) and Liquefied Natural Gas (LNG).

Oil: Companies frequently disclose crude oil reserves/production aggregated with the figures for natural gas liquids (NGLs) or condensate - often referred to as ‘liquids’. As such, this research uses the term 'oil' to encompass: crude oil extracted through conventional and unconventional methods, including tight oil and shale oil; NGLs, including liquefied petroleum gas, pentane-plus and condensate; and synthetic crude derived from bitumen, tar sands and oil shale. ‘Liquids’ might be a technically more accurate term but it is not widely recognised. The term 'oil' has therefore been used for communications purposes.

While all coal has similar geological origins, slight differences have resulted in varying properties. Depending on these properties, coal is used for one of two primary purposes: direct combustion for energy generation (thermal coal) or the creation of coke used in industrial processes such as iron and steel making (metallurgic coal). From a climate change perspective, thermal coal is the key concern due to the high greenhouse gas emissions resulting from the combustion process. Other uses of coal include coal gasification to produce Syngas which in turn may be used to generate power. A related, although at present test-scale process converts coal to hydrogen, with the hydrogen destined for energy generation for transport. The IEA states that in 2017 86% of global coal production was thermal with the other 14% accounted for by coking coal intended for metallurgical use. Coal destined for Syngas and other gasification uses is likely a minor proportion. BP's Statistical Review of Energy classifies coal as anthracite, bituminous, sub-bituminous and lignite, all of which to varying degrees are used in power generation, as noted by the US EIA in 2017 data on coal use for power. Bituminous coal also has metallurgical uses. Based on IEA estimates we assume 86% of coal production and reserves are thermal.

The research process looks at all coal production and reserves by the companies in our study. It then determines what portion of these reserves are thermal coal by the latest disclosures from the companies. Such disclosure varies and can be characterized by the type of coal or by its end use. If the company specifies only the type of coal in their production/reserves mix, we characterize lignite, anthracite and sub-bituminous coal as thermal coal. Where companies disclose by end use (which is usually the case) we categorize thermal coal as any coal tagged by companies as thermal coal, and also include coal gasification to generate syngas, which we assume is primarily used for power generation. The latter is significant in the case of coal producer Sasol of South Africa, for example.

This research maps out the finance sector (i.e. investment management) in a hierarchical manner as follows. At the top are "financial groups" (e.g. Blackrock), which are affiliations of commercial entities with cross-holding structures. Under this are nationally registered entities (e.g. Blackrock Ltd, Blackrock Inc.), which are "asset managers" that are the registered owners of shares on behalf of their clients. These asset managers may also operate "Listed Funds" (e.g. iShares ETFs), which are pools of capital market assets that can be traded on exchanges and are the registered owners of shares on behalf of the owners of the Listed Funds. Another class of funds, which are not traded on markets, are known as "pooled investment funds." These are special purpose investment companies managed by hedge funds and asset managers and sold to wealthy individuals and institutions. Our system at present does not identify these individual non-traded funds but it does assess the aggregated holdings of the asset managers who run them.

Outside of this structure are "asset owners" (e.g. pension funds, foundations and government funds) which may either own shares directly, employ asset managers to own shares on their behalf or invest in Listed Funds. In addition, there are also "other shareholders" such as corporations, wealthy individuals or government treasuries who own large portions of fossil fuel companies for strategic reasons.

The terms "financial groups" and "asset manager groups" are used interchangeably. Groups such as BlackRock are primarily engaged in asset management while groups like HSBC engage in asset management in addition to a variety of banking activities (hence these are better known as financial groups). Likewise, many financial groups - like AXA, Aviva and Legal and General - have extensive insurance branches as well as asset management businesses. Our system aggregates all companies, regardless of type of business, within their financial groups and aggregate all shareholdings we can track to these groups. We maintain a separate database of holdings by each of the companies within the financial groups.

Our data on shareholders in listed companies rely on a number of disclosure sources, which may be mandatory (e.g. US SEC 13-F filings applying to asset managers with more than $100mn under management) or voluntary (e.g. the Government Pension Fund of Norway's portfolio disclosure). These disclosures vary by region and shareholder type. Our data is most accurate for listed funds, followed by asset managers and then by asset owners. In terms of region, the US offers the most disclosure on shareholders. For example, roughly 70% of ExxonMobil's shareholders may be identified, while this figure is far lower for Chinese and Russian fossil fuel companies and investors. The gap in knowledge of any company's shareholders is primarily due to the lack of disclosure requirements for individual investors, special purpose companies, or small-scale asset managers to declare their holdings. We point users to our Terms and Conditions for issues relating to the use of and reliance on our data. In particular, our data should in no way be considered as guidance on investment activity.

We update our shareholding data bimonthly and the data on the fossil fuel companies each year as new Annual Reports become available. The system tracks ownership of shares via specific ticker symbols representing types of shares traded on particular markets. We aggregate all such ticker symbol-ownership attributed to a shareholder to arrive at the portion of the company own. The ownership % held by a shareholder is computed as the shares held divided by the total number of shares outstanding.

The system can thus state the minimum holdings held by an entity and measure any recent changes in these holdings. It should be noted that all such financial data is necessarily out of date in that it relies on past disclosures, generally up to several months prior to the date of data taken.

The project will cover all sectors where climate change presents business risk (e.g. chemicals, utilities, automotive). We will incorporate bond holding data into our system in the coming year, as part of our roll out of the full set of portfolio metrics.

This research considers "listed funds" as collective pools of capital, managed by investment professionals and traded on markets or offered to institutional/other investors in a regulated manner. The database used in this research is derived from the Thomson Reuters Lipper financial database, which states "Lipper includes mutual funds, closed-end funds, ETFs, hedge funds, retirement and pension funds, and insurance products."

The open-ended segment of this market is likely to contain up to 30% of all global market assets, according to the European Fund and Asset Management Association as of Q2 2018. The dynamics of these markets are thus highly important to short and long term market direction. Low cost funds which track external indices from index providers such as S&P or MSCI (and thus require minimal input from the asset manager providing the funds) are an increasingly popular investment mechanism, accounting for the strong recent growth exhibited by US asset manager giants Vanguard, Fidelity and BlackRock.

Listed funds are commonly described as passively or actively managed. In reality there is a spectrum of management strategies used by the fund management industry. To distinguish between different methods of fund management, this research employs the following designations and definitions from leading financial data provider Thomson Reuter's Lipper database. The database flags funds which track indexes, as well as identifies the specific method of index tracking.. The database identifies two methods of fund management in which the portfolio is fully allocated by tracking an index: 'Full Index-Tracking' (defined as funds which hold positions in all securities of the underlying index in proportion to their weightings in the index) and 'Swap Index-Tracking' (in which the fund manager does not have to physically hold the securities, but instead gains exposure to the underlying index through the use of derivatives, which serve as contracts obliging a 3rd party to pay the fund manager the exact performance of the index). In this study, 'Full' and 'Swap' index-tracking funds are grouped together due to their precise tracking of indices. Lipper identifies a separate category of partly index-tracking funds, labelled 'Optimized'; these funds involve some input from a fund manager, who aims to replicate exposure to an index by investing in a representative sample of securities. Funds not flagged by the database as employing an index-tracking method may or may not involve active management, and are therefore designated 'non-index tracking' in this research.

This research tracks 4000 Asset Managers, which are registered companies providing financial services in specific geographies. Our financial data tracks their shareholdings both directly, and through the listed funds they operate. It also tracks the total amount of capital market instrument (equities, bonds, other) assets they have under management (AUM). The research then "tags" each of these asset managers if they are part of larger financial groups (a term used interchangeably with "asset manager group"). For example, BlackRock UK Ltd will be tagged with "BlackRock". To arrive at the largest asset manager groups, the research aggregates all the operating companies under a given group's control. The 15 largest Asset Manager Groups manage a combined $40 trillion in AUM, representing more than 21% of all global capital markets value ($185 trillion in 2017, according to the IMF).

This method of categorizing the asset management sector may differ from others (for example the IPE Top 400 list does not aggregate asset managers to their highest corporate group level). In addition, our research also consider groups like Charles Schwab, which may provide transaction-only services for a large portion of the $3.5 trillion the company lists as AUM in our ranking. In reality, all asset managers may provide a spectrum of offerings - from low cost transaction services to bespoke institutional investment solutions covering multi-billion dollar portfolios - often within the same asset management group. All are in a position to provide some level of climate risk-related information or advice to their clients and hence are of interest to our project and platform.

In assessing the assets held by these large asset manager groups, our research effectively considers all the companies operating under the parent grouping. In some cases (e.g. AXA or Legal and General) the same parent grouping may contain: a company conducting insurance activities, which holds and invest premiums for citizens; a company providing asset management services for other institutional investors; and the parent grouping’s own corporate pension fund. Similar comments apply to UBS, which operates private and corporate banking, institutional asset management and also has its own corporate pension fund for its past and current employees. While our system separates out the holdings of each of these sub-group entities, they are aggregated together for the purposes of analyzing the 15 largest asset manager groups.

This research computes the change in effective ownership of fossil fuel reserves (coal, oil and gas) by shareholders in the time period 31/03/2016 to 30/06/2018 (covering their activities following the Paris Agreement). Measuring effective ownership of aggregate reserves is useful as it provides a metric for physical assets in the portfolio independent of share price movement. The presence of thermal coal in a portfolio will clearly be of interest to the coal divestment community, for example. With respect to effective ownership of oil/gas reserves, we consider 2P reserves (proven + probable). We temper the change metric for fossil fuels owned with outflows and inflows into the portfolio during this period. For listed funds this is available from financial data sources, and for asset managers and asset manager groups we obtain this from their financial filings). We express the change in thermal coal/oil/gas in terms of physical units (tons/BOE), as this is independent of share price variation and represents, in combination with the inflow/outflow factor, actual shifts in exposure to these asset classes.

We represent the percentage inflows/outflows into the portfolio over the two-year period as c, and the physical reserves at the start and end of two years by s and e, respectively. If the reserves held by a fund had changed in exact proportion to its size over the two-year period, then the reserves at the end would be e_0=s × (1+c). Our change factor f is defined as the relative deviation of the actual reserves e from e_0: f=(e-e_0)/e_0. This can be expressed as either a positive percentage for an increase or a negative percentage for a decrease.

Where we have information on the total Assets under Management (AUM) in a portfolio and the tons of thermal coal held by the companies in the portfolio, for comparison purposes we define a metric, "Thermal Coal Intensity", as the tonnage of thermal coal held per $mn AUM.

To define policy engagement, InfluenceMap relies on the 2013 Guide for Responsible Corporate Engagement in Climate Policy issued by the secretariat of the UNFCCC and the UNEP under the UN's Caring for Climate collaboration of the UN Global Compact. This document defines a list of corporate activities that constitute corporate climate engagement, covering direct and indirect tactics. These range from social media; public relations; sponsoring research; direct contact with regulators and elected officials; funding of campaigns and political parties; and participation in policy advisory committees.

Since this assessment focuses on EU policy, it is instructive to note that the EU’s institutions operate with varying levels of transparency. Therefore, coverage of interactions between organizations and the European Commission is higher than coverage of interactions with the European Parliament and Council of the European Union.

This analysis and scoring is focused on an organization's comments, interactions and influence on policy and legislation. It does not consider internal strategy, activities and performance of a company on sustainable finance or climate change related issues.

InfluenceMap's scoring process is policy neutral. It does not assess the quality of governmental policy but rather the positions of companies and industry groups relative to this policy. This is achieved by using the statements and ambitions of government-mandated bodies tasked to propose or implement sustainable finance policy as the benchmarks against which corporate and industry association policy positions are scored.

For this analysis, the recommendations of the European Commission’s High-Level Expert Group and its Action Plan on Sustainable Finance were relied on heavily as benchmarks. This was supplemented in places with the statements and recommendations of UN-backed scientific enquiry, such as the findings of the Intergovernmental Panel on Climate Change (IPCC).

Scored evidence is coded by InfluenceMap as: ‘strongly supporting’, ‘supporting’, ‘no position/mixed position’, ‘not supporting/supporting with exceptions’, or ‘opposing’ with reference to the benchmarks explained above. These categories correspond to a numerical five-point scale between +2 and -2, where +2 indicates strong support and -2 indicates opposition.

This research covers the period from 2017 to July 2020, following the High-Level Expert Group on Sustainable Finance and the first stage of the EU's Action Plan on Sustainable Finance. The weighting system prioritizes the most recent evidence pieces, so older evidence pieces have a smaller impact on the overall score. InfluenceMap's platform provides a snapshot of policy engagement so far and will be continually updated to cover future engagement, including on the EU's Renewed Strategy on Sustainable Finance.

This assessment has focused on the largest European financial groups’ engagement on sustainable finance policy, primarily EU-level policy with some engagement on policies in other European regions.

The universe of financial groups was selected by taking the top 25 European banks by assets, the top 25 European insurance groups by assets, and top 25 asset managers with a significant presence in Europe by AUM. These companies were scaled up to the financial group level, creating a final list of 63 financial groups after overlaps were excluded. 20 trade associations found to be active in this policy area were also selected.

The queries were selected to broadly cover the recommendations of the EU’s High Level Expert Group on Sustainable Finance and have been refined through consultation with experts in the field.

This is an emerging area of sustainable finance related policy which InfluenceMap plans to cover in the near future.

This is covered by InfluenceMap’s flagship platform on corporate influence and lobbying on climate policy.