Annex A. Methodological approach to tracking aspects of fossil fuel support

Isolating the output-based support to the electricity sector attributable to fossil fuels

The range of fuels and energy products in broad categories covered by the Inventory are coal, natural gas, petroleum products and end-use electricity. There are cases where some measures in the Inventory benefit more than one type of fossil fuels. In this case, the OECD Secretariat allocated support to particular fuels where official government sources do not provide such a breakdown. Measures benefiting more than one fuel or energy product were allocated according to the relative value of production or consumption using the calculated shares derived from the national balances published in the IEA’s World Energy Balances.

For fuels such as coal, natural gas and petroleum, the allocation concludes with a simple allocation of shares calculated using the IEA World Energy Balances data. However, for end-use electricity, the allocation presents an additional methodological challenge as raw amounts include embedded support for electricity generated from non-fossil-fuel origin (e.g., nuclear, renewables, biofuels and wastes) and electricity from traded (i.e., imported) sources, (which can make it difficult to trace the ultimate origin and consequently the generation type of the electricity once joining the national grid). Both the share of electricity from non-fossil-fuel sources and electricity from traded sources must therefore be removed to isolate end-use electricity support to fossil fuels alone.

To isolate the share of electricity domestically generated in the country (i.e., excluding imported electricity), the following formula was used:

Annex Figure A.1.

To isolate fossil-fuel from non-fossil-fuel generated electricity, shares from national electricity generation mix from the IEA Energy Balances were used. The following formula was used to calculate the share of fossil-fuel electricity generated per economy:

FF_share_x = (Total_gen - REN_gen – Nuclear_gen)/Total_gen

where:

FF_share_x – share of fossil fuel generated electricity in country x

Total_gen – total electricity generated by power plants in country x

Ren_gen - total electricity generated by renewable power plants in country x.

- consists of the sum of electricity generation from HYDRO, GEOTHERM, SOLARPV, SOLARTH, TIDE, WIND, MUNWASTER, PRIMSBIO, BIOGASES, BIOGASOL, BIODIESEL, OBIOLIQ, RENEWNS and CHARCOAL from the IEA World Energy Balances (IEA, 2020[1]).

Nuclear_gen – total electricity generated by nuclear power plants in country x

After obtaining these two coefficients for each country, the share of support attributable to end-use fossil-fuel electricity for a given measure y is then calculated as follows:

End-use electricity support from fossil fuels = (raw amounts for measure y) *(Ind_gen)* (FF_share)

In summary, the above formula only attributes a certain support amount to end-use electricity support using the share of electricity indigenously generated in the country and the share of electricity generation mix from fossil fuels. In particular, the above formula removes shares that are due to electricity imports (whose ultimate generation origin cannot be determined with the available information) as well as non-fossil-fuel generation sources.

Breaking down fossil-fuel support data to economic sector beneficiaries

Prior to the pilot sectoral disaggregation, the Inventory was organised following the OECD’s PSE-CSE framework. Under this classification framework, measures benefitting fossil-fuel producers are classified under Producer Support Estimate (PSE) while those that benefit individual fossil-fuel consumers fall under the Consumer Support Estimate (CSE). A third category, the General Services Support Estimate (GSSE) is assigned for measures that do not increase production or consumption of fossil fuels at present but may do so in the future.

Classification under the PSE-CSE classification framework is broad and does not allow further disaggregation of beneficiaries by economic sector. While it allows to isolate which measures benefit the upstream or midstream fossil-fuel sectors, it does not allow to isolate and pinpoint in greater detail the final end-user economic sectors (e.g. industrial, transport, residential, commercial etc.) targeted by fossil-fuel measures. Identifying and quantifying the benefit received by each economic sector in fossil-fuel support is key in order to evaluate the distributional impacts of proposed fossil-fuel reforms as well as in evaluating whether a particular targeted support programme is efficient in reaching their intended beneficiaries.

Sector tagging mechanics for fossil-fuel support measures

For each measure in the Inventory, two types of information are provided: (i) fiscal information on the budgetary transfers or tax expenditures (monetary value) and (ii) textual metadata with contents on a measure’s beneficiaries; eligibility criteria; historical background; and any relevant details on data procurement and data processing.

The information on the textual metadata is used to identify which economic sector is benefitted from each measure. In tagging each measure, the economic activity nomenclature follows the classification used in the IEA World Energy Balances flows. Measures can receive a single (in case only one economic sector is benefitted) or multiple sector tags. In case of a single sector tag, the attribution of values is straightforward and all the measure’s amounts get assigned to the single sector. However, there are cases where a measure is designed to benefit multiple sectors (e.g., preferential tax rates for natural gas targeting both residential and commercial sectors or some measures targeting both agricultural and road sector fuels). In this case, the amounts that each sector gets allocated is based on the calculated proportions as obtained from the energy consumption figures reported in the IEA World Energy Balances.

Finally, after the sector tagging exercise, the results are aggregated and mapped according to the following broad sectoral categories:

Annex Table A.1. Selected sector dimension tags to identify ocean-related FFS measures
Broad Sector	Remarks and Included Sectors (IEA short names)
Fossil-fuel production	This category comprises measures that benefits the upstream and midstream segment of fossil-fuel production. It includes measures targeted towards the exploration, production, trade (import or export), transportation or storage of fossil fuels. INDPROD, IMPORTS, EXPORTS, STOCKCHA or TPES (if no detail available).
Electricity generation	This category represents measures that provide support to fossil fuels used as inputs in power generation. Only input fuels fall under this category and does not include support for the consumption of generated electricity by end-use consumers. MAINELEC, AUTOELEC, MAINCHP, AUTOCHP, .MAINHEAT, AUTOHEAT
Transport	Measures benefitted under this category includes fuels used in the following transport activities: road vehicles, agricultural and industrial highway use, aircraft for domestic aviation, rail traffic including urban or suburban transport systems, energy used for pipelines transporting fossil-fuels, domestic maritime navigation (i.e., port of departure and arrival of the same country), and all transport not elsewhere specified. DOMESAIR, ROAD, RAIL, PIPELINE, DOMESNAV, TRNONSPE
Residential	Includes consumption by households (including households with employed persons), with fuels used for transport excepted.
Other sectors	Other sectors include measures that support the use of fossil fuels in energy transformation other than electricity and heat generation, industrial and manufacturing, commercial and public services, agriculture, forestry and fisheries and non-energy use. THEAT, TBOILER, TELE, TBLASTFUR, TGASWKS, TCOKEOVS, TPATFUEL, TBKB, TREFINER, TPETCHEM, LIQUEFAC, TNONSPEC, OWNUSE, EMINES, EOILGASEX, EBLASTFUR, EGASWKS, EBIOGAS, ECOKEOVS, EPATFUEL, EBKB, EREFINER, ECOALLIQ, ELNG, EGTL, EPOWERPLT, EPUMPST, ENUC, ECHARCOAL, ENONSPEC, TRANSFER IRONSTL, CHEMICAL, NONMET, PAPERPRO, NONFERR, TEXTILES, MINING, TRANSEQ, MACHINE, FOODPRO, PAPERPRO, WOODPRO, CONSTRUC, INONSPEC RESIDENT, AGRICULT, FISHING, COMMPUB, ONONSPEC, NONENUSE
Source: Adapted from IEA World Energy Balances, (IEA, 2020[1]).

The routines were implemented using Stata to automate the allocation of each measure. At the end of the tagging exercise, each measure is now classified according to three dimensions: (a) fuel(s) benefitted; (b) PSE-CSE indicator and (c) sectoral beneficiaries. For the sectoral aggregates, allocation involved dual dimensions (i.e. fuel and sector), with more than 100 fuel-sector combinations. This made the calculation to be computationally-intensive, with the dual dimension structure requiring O(n²) polynomial time complexity during the algorithm’s execution.

Identifying ocean-related government support for fossil fuels

The OECD Inventory of Fossil-fuel Support Measures is an online database that identifies, documents and estimates direct budgetary support and tax expenditures supporting the production or consumption of fossil fuels (http://www.oecd.org/fossil-fuels/data). The Inventory currently covers 37 OECD member countries, eight non-OECD G20 economies (Brazil, Colombia, the People’s Republic of China, India, Indonesia, the Russian Federation, and South Africa) and six EU Eastern Partnership countries (Armenia, Azerbaijan, Belarus, Georgia, Republic of Moldova and Ukraine) and has compiled more than 1 300 individual support measures (both active and terminated ones). In addition to national measures, sub-national support measures for selected economies are also covered (i.e., Australia, Canada, China, Germany and the US).

For each measure, two types of information are provided: (i) fiscal information on the budgetary transfers or tax expenditures (monetary value) and (ii) textual metadata about a measure’s beneficiaries, eligibility criteria, historical background, and any relevant information on data procurement and processing.

Following the OECD’s PSE-CSE framework, the measures benefitting fossil-fuel producers are classified as the producer support estimate (PSE) while those that benefit individual fossil-fuel consumers are classified under the consumer support estimate (CSE). A third category, the general services support estimate (GSSE), is assigned for measures that do not currently increase fossil-fuel production and consumption but may do so in the future.

The Inventory identifies the type of fossil fuels that benefit from each measure and presents a breakdown of the amount of support by assigning fuel type tags. In cases where this breakdown is not available in official government sources, the OECD performs data transformation procedure to allocate support to individual fuel tags according to the relative value of production or consumption as calculated from the IEA’s World Energy Balances database. Note that measures can benefit more than one type of fossil fuel at the same time and can thus receive multiple fuel tags in this respect. For example, a measure granting lower sales tax rates for road transport fuels will receive multiple fuel tags such as motor gasoline, diesel, LPG and natural gas.

Building on this methodology, an additional binary tag is developed for ocean-related government support for fossil fuels.

Search strategy

A search strategy is developed to identify measures directly relating to oceans. First, measures in countries not bounded by a coastline are removed.1 Second, a keyword search is conducted on both the programme name and description to pre-screen measures and identify potential candidates for ocean-related FFS. The list of keywords includes generic ocean terms, keywords related to off-shore oil and gas, maritime transport, maritime fisheries and the ocean economy more broadly (Annex Table A.2).2

Third, additional measures may be identified using the “sector” dimension which labels individual measures following the nomenclature used in the IEA World Energy Balances. Multiple sectors may be assigned to a single measure. Annex Table A.3 lists the sector tags that are used to pre-screen potential candidates for FFS ocean-related policies. Any measure bearing these tags are identified as potentially relating to the ocean.

Finally, following the pre-screening by the automated keyword-based searches, each candidate measure is then individually reviewed in order to eliminate false positives and to ascertain that measures inadvertently flagged as ‘negative’ have not been omitted.

Annex Table A.2. Keywords used to identify a candidate subset of ocean-related fossil-fuel support measures
ocean, [ ]sea[ ], marine, maritime, offshore, blue, reef, coral, deep-sea, deep-water, sea-bed, benthic, lagoon, mudflat, tidal, mangrove, coast*, coastal, wetland, coastal, marsh salt, marsh, salt-water, brackish
[ ]fish*, aquaculture, mariculture, cod[ ], [ ]tuna, coral*, mussels, oysters, crustacean, mollusc, pelagic, sea-food, aquatic plants
[ ]ship*, shipping, ship-building, boat, vessel, ferry, floating, naval, [ ]port[ ], sea-port, harbour, harbour, dock, sea transport, cargo, ballast, oil spill*, oceanogra*
cruise, tour, touris*, dredg*, sea salt, desali*, hurricane, cyclone, typhoon, flood, biotechnology, robotics, wave, underwater vehicle, continental shelf, EEZ, exclusive economic, arctic, piracy, submarine, undersea, [ ]cable*, seabed, hurricane, cyclone, blue water, cruise, ferry, international water*, [ ]rig[ ], territorial sea, high sea*

Annex Table A.3. Selected sector dimension tags to identify ocean-related fossil-fuel support measures
Sector (IEA shortname)	Remarks
Indigenous production (INDPROD)	This category comprises the production of primary energy such as various types of primary coal and natural gas. Among ocean-related measures falling under this category are measures benefitting oil and natural gas extraction in off-shore environments.
Oil and gas extraction (EOILGASEX)	This category represents the energy which is used for oil and gas extraction. Ocean-related measure under this category are those benefitting off-shore oil and gas extraction operations.
Domestic navigation (DOMESNAV)	Measures benefitted under this category includes fuels delivered to maritime vessels not engaged in domestic l navigation (i.e. determined in terms of the port of departure and port of arrival belonging to the same single country and not by the flag or nationality of the ship). Note that while the voyages considered are domestic, these may involve routes of considerable distance that may transit through foreign countries or international waters.
Fishing (FISHING)	Measures are classified under the fishing sector if the benefitted fuels are used for inland, coastal and deep-sea fishing as well as energy used in the fishing industry. It is recognised that this sector captures fuels used for inland fishing, which is not considered as ocean-related, but this approach is adapted in the absence of more specific breakdown that isolates fuels used for ocean-related purposes under the FISHING sector.
Source: Adapted from IEA (2019), “World energy balances”, IEA World Energy Statistics and Balances (database).

Annex Table A.4. Selected examples of ocean-related measures in the fossil-fuel support inventory
Sectors Benefitted	CSE	PSE	GSSE
Indigenous Production		Sales Tax Exemption for Exploration Equipment (Canada)	Norwegian Petroleum Directorate geological surveys (Norway)
Oil and gas extraction	Mineral oil tax exemption for offshore petroleum sector (Norway)	Sales-Tax Exemption for Repairs and Materials Used on Drilling Rigs (United States)
Domestic navigation	Fuel tax exemption for shipping (Italy)
Fishing	Fuel tax exemption for fisheries (Korea)

References

[1] IEA (2020), World Energy Balances, International Energy Agency, https://www.iea.org/data-and-statistics?country=WORLD&fuel=Energy%20supply&indicator=TPESbySource (accessed on 14 January 2021).

Notes

← 1. Austria, Czech Republic, Hungary, Luxembourg, Slovak Republic and Switzerland.

← 2. While entries in the programme name and description fields occasionally appear in their original language, all of these occurrences are consistently translated into English, thus removing the necessity to devise foreign language keywords in the dictionary.