Metallurgical Coal Exit List (MCEL) Data

It is the steelmaking half of the coal-exclusion universe, cut loose from its thermal twin. The Metallurgical Coal Exit List (MCEL), compiled by urgewald e.V. and launched in 2025 as a supplement to the Global Coal Exit List, exists because GCEL deliberately tracks thermal coal only - pure metallurgical coal developers were falling between two stools. Met coal here means coking coal and pulverized coal injection (PCI) coal, the fuels blast furnaces cannot easily replace.

The entry test is one sentence long: a company is listed if it is actively developing new met coal mines or expanding existing operations. No revenue thresholds, no capacity floors, no judgment calls - all 145 companies on the 2026 edition clear the same bar, together carrying more than 250 expansion projects across 20 countries. The stakes come quantified by the publisher itself: realized in full, that pipeline would lift global annual met coal production by 52%, in a segment that already accounts for roughly 13% of world coal consumption and feeds a steel industry responsible for about 11% of global CO2 emissions.

Within Datadory's catalog of 1,744 datasets across 159 viable industries, this record scores 7/10 on our rubric against a catalog average of 7.81.

Six columns hold each row - enough to identify the company, place it, count its ambitions and tie it to its group:

# one row per listed company - shape shown on the documented grain

company            : Shanxi Coking Coal Group Co Ltd
country_of_hq      : China
expansion_projects : 3
expansion_status   : active developer
parent_link        : <ultimate parent carried where mapped>
security_ids       : <ISIN/ticker where supplied>

# shape of the slice (2026 edition)
companies          : 145    every one an active developer
projects           : 250+   new mines and mine-life expansions
hq_countries       : 20     met coal mining jurisdictions

# why the denominator never drifts:
#   inclusion_test = actively developing new mines or expanding existing ones

Read what the shape implies. A list that fits on one screen still moves markets because it is exhaustive about intent rather than descriptive about output: there is no production-tonnage column to hide behind, only whether the company is adding capacity. The parent/subsidiary link is what turns 145 rows into group-level exposure - a subsidiary inherits its group's flag instead of posing as an unrelated name. And because every row passes the identical test, a count of listed companies is a count of developers, not a blended statistic.

Six verified fields, written against the published methodology rather than guessed from headers. Identification comes first: company_name and country_of_hq place each firm among the 20 covered mining jurisdictions. Intent is measured by expansion_projects, the count of new developments and extensions under way, and confirmed by expansion_status, an inclusion flag every row carries because every row earned it the same way. Corporate attribution rides on parent_link, consistent with GCEL's parent-subsidiary mapping, so both lists can be screened against one corporate tree. security_ids adds ISIN and ticker references where supplied, letting list rows land in a securities master without a fuzzy-matching detour.

The dictionary above is the verified core. Project-level detail behind each company's expansion count, edition-over-edition entry and exit flags, and any identifier coverage beyond the core file fold under additional fields on request: ask for them with your sample and they arrive documented in the same format.

• Geography: 20 countries - the working spread of met coal mining jurisdictions, from the producers that dominate seaborne coking trade to the domestic-market miners that never surface in port statistics
• Temporal: year-labelled editions since launch in 2025, with the MCEL 2026 edition as the current reference for this page; each edition supersedes the last rather than accumulating rows
• Granularity: one row per company, with underlying expansion projects counted beneath each company and a parent/subsidiary link for group-level rollups

That grain is the point. Production statistics tell you what was dug last year; this tells you which companies intend to dig more tomorrow - and names them.

API, files, or your warehouse. Daily, weekly, or hourly.

Pick the channel your stack already speaks. The rows arrive identical either way - one record per listed company with its headquarters country, expansion-project count and parent/subsidiary link intact - keyed so each pull joins to the last instead of opening a fresh reconciliation project.

Every delivery ships with the field dictionary above plus the sample rows for validation, so your first screen against the list happens on evidence rather than hope.

Exclusion and mandate screens. More than 150 financial institutions already work from MCEL, and the Science Based Targets initiative's Net-Zero Standard for Financial Institutions recommends excluding met coal developers outright - see esg & emissions analysis.

Forward supply views. Convert 250+ named projects into a bottom-up met coal outlook instead of extrapolating last year's production curve - see market sizing.

Counterparty and chain checks. Screen steel-value-chain counterparties before onboarding, and trace flagged miners up the corporate tree to the group that ultimately owns the decision - see supply-chain mapping.

Capital-at-risk reads. A company funding several greenfield expansions carries different exposure from one stretching existing pits; the project count separates them - see credit risk screening.

Competitive watches. Track which rivals committed to new met coal capacity between editions without reconstructing it from a year of trade press.

Scholarship and journalism. Year-labelled editions with a published, arithmetic entry test give steel-decarbonisation research a citation that survives review.

Investors & Quant Researchers top our relevance scoring for this record: expansion flags convert into transition-risk factors, and a list that 150+ institutions already treat as the reference makes the screen consensus rather than contrarian (investors & quants use cases). Market Researchers & Consultants get a 20-jurisdiction expansion map with every project attached to a named company (market researchers use cases). Data Scientists & ML Engineers get small, high-signal features - a boolean and a project count - that plug straight into coal-exposure models (data scientists use cases). Sales & Growth Teams screen accounts and supply chains against the roster before the logo lands in the pipeline. Journalists, Academics & Students cite the editions directly, entry test included (journalists & academics use cases).

Methodology note - the single-criterion design cuts both ways. Inclusion proves intent to expand, not scale: a one-project junior sits beside multi-project state groups, so read expansion_projects before ranking exposure.

Corporate-tree note - the parent/subsidiary link follows GCEL's mapping convention, so a subsidiary's flag inherits upward. Group-level screens should deduplicate on the parent column, not the company name.

Provenance note - every row traces to urgewald e.V.'s published research, the compiler whose thermal-coal list has become the reference point for institutional exclusion policy; edition-over-edition changes are attributable rather than folklore.

Where to go next:

• Global Coal Exit List (GCEL) - the thermal twin, roughly 3,000 companies under three tests; together the two lists cover the whole black side of the transition.
• Global Coal Plant Tracker (GCPT) - the consuming fleet, unit by unit.
• SteelHome Database - Coke/Coal & Steel Raw Materials - the price side of the same coking and PCI coals this list counts as expansion risk.
• Monthly Crude Steel Production (71 countries) - the demand these blast-furnace fuels ultimately serve.