Multi-Sector Holdings Data: Institutional Positions, Ownership Chains and Utility-Scale Energy Inputs · Head-to-head
GLEIF Golden Copy and Delta Files (LEI Reference Data) vs EIA Consumption & Efficiency Surveys (RECS, CBECS, MECS)
Which multi-sector holdings data: institutional positions, ownership chains and utility-scale energy inputs data fits your job: GLEIF Golden Copy and Delta Files, or EIA Consumption & Efficiency Surveys. API, files, or your warehouse. Daily, weekly, or hourly.
GLEIF Golden Copy and Delta Files (LEI Reference Data)
EIA Consumption & Efficiency Surveys (RECS, CBECS, MECS)
Where the fields line up
No shared field names. These two answer different questions.
| Field | GLEIF Golden Copy and Delta Files | EIA Consumption & Efficiency Surveys |
|---|---|---|
LEI | documented | not in this set |
LegalName | documented | not in this set |
LegalAddress | documented | not in this set |
HeadquartersAddress | documented | not in this set |
EntityStatus | documented | not in this set |
Relationship.RelationshipType | documented | not in this set |
Relationship.StartNode.NodeID | documented | not in this set |
Relationship.EndNode.NodeID | documented | not in this set |
Relationship.RelationshipStatus | documented | not in this set |
Registration.ManagingLOU | documented | not in this set |
Registration.RegistrationStatus | documented | not in this set |
Registration.ValidationSources | documented | not in this set |
Coverage, side by side
| GLEIF Golden Copy and Delta Files | EIA Consumption & Efficiency Surveys | |
|---|---|---|
| Geographic | Global - all LEI-registered legal entities across every jurisdiction in the Global LEI System, with geocoded legal and headquarters addresses | United States only - national estimates, state-level tables for RECS, census-division floor for CBECS microdata, industry-by-region cuts for MECS |
| Granularity | One row per LEI record or per reported parent-child relationship | One row per surveyed household (RECS) or building (CBECS); aggregated industry tables for MECS |
What each contains
Pick by fit, not by loyalty.
| GLEIF Golden Copy and Delta Files | EIA Consumption & Efficiency Surveys | |
|---|---|---|
| Publisher | Global Legal Entity Identifier Foundation (GLEIF) | U.S. Energy Information Administration (EIA) |
| Subject lens | Who-is-who entity reference (Level 1) and who-owns-whom relationship records (Level 2), plus reporting exceptions for unresolvable parents | Demand-side end-use energy: residential consumption (RECS), commercial building stock (CBECS), manufacturing fuel and electricity tables (MECS) |
| Scale | Level 1: 3,408,526 LEI records; Level 2 relationships: 485,148 records; Reporting Exceptions: 6,310,616 records | RECS 2020: ~18,500 households x 799 columns; CBECS 2018: 6,436 buildings x 1,249 columns representing ~5.9 million US buildings; MECS 2018: ~100 summary tables |
| Geography | Global - all LEI-registered legal entities across every jurisdiction in the Global LEI System, with geocoded legal and headquarters addresses | United States only - national estimates, state-level tables for RECS, census-division floor for CBECS microdata, industry-by-region cuts for MECS |
| Granularity | One row per LEI record or per reported parent-child relationship | One row per surveyed household (RECS) or building (CBECS); aggregated industry tables for MECS |
| Best for | Entity resolution, ultimate-parent lookups, corporate hierarchy graphs, counterparty screening | Energy-intensity benchmarking, building-stock and housing-stock analysis, end-use load research, emissions-per-dollar estimation |
What each does better
GLEIF Golden Copy and Delta Files
The whole world, deduplicated. Level 1 holds 3,408,526 entity records in which each LEI appears exactly once, spanning every jurisdiction that participates in the Global LEI System. No EIA product reaches past the United States border; if your counterparty map includes a German manufacturer, a Singaporean fund manager or a Cayman subsidiary, this is the record where they live.
Ownership is first-class, not inferred. Level 2 carries 485,148 direct and ultimate-parent links typed by relationship kind - consolidation, fund management, sub-fund membership - each with accounting-period qualification, optional control quantifiers, and a reporting-exceptions companion of 6,310,616 rows that names the cases where a parent could not be resolved and why. That is an auditable chain of custody for corporate hierarchy, something the energy surveys simply have no analogue for.
Change tracking built into the file family. Every full copy travels with four delta views capturing eight-hour, 24-hour, seven-day and 31-day change windows, so the diff against last week's hierarchy arrives precomputed instead of inferred. When a counterparty's filing moves, the movement is a row you can select, not a rumour to chase.
EIA Consumption & Efficiency Surveys
Microdata depth on real consumption. CBECS 2018 publishes 1,249 columns for each of 6,436 sampled buildings - supplier-derived annual electricity, natural gas, fuel oil and district heat alongside square footage, shift-headcount and principal activity - and RECS 2020 ships 799 columns for nearly 18,500 households down to per-end-use breakdowns such as space heating, water heating and appliances. GLEIF records facts about organizations; these are measured behaviours of households and buildings, with nothing comparable anywhere in the identity estate.
Statistical machinery built in. Final weights and a hundred replicate-weight columns make standard errors a column operation rather than a modelling project, and imputation flags tell you which values were filled. For anyone publishing estimates rather than just joining entities, that is the difference between a defensible number and an anecdote.
Decades of benchmark cycles. CBECS has run since 1979, MECS every four years since 1994, RECS in repeated waves including 2009, 2015 and 2020. Fixed vintages mean a 2018 building can be tracked against earlier stock, and MECS's roughly 100 pre-aggregated tables hand you fuel use by NAICS industry and region without any aggregation work. The LEI register, by contrast, describes the present and near-past of filings, not a longitudinal sample.
Where they're equivalent
More than the subject matter suggests. Both are authoritative primary publications - a standards foundation on one side, a federal statistical agency on the other - and both ship machine-readable bulk formats: GLEIF in zipped CSV, JSON and XML with an RDF rendering, the EIA surveys in CSV and SAS microdata with XLSX codebooks and PDF/XLSX summary tables. Both document themselves unusually well: GLEIF's CDF format versions (LEI_3.1, RR_2.1) define every element, and the EIA microdata arrive with variable/response codebooks and RSE worksheet tabs beside the summary tables.
Both also share the same practical weakness: volume without curation. A 3.4-million-record Level 1 table and a 1,249-column building file are equally unfriendly to a spreadsheet-first analyst, and neither publisher normalizes their output into a warehouse-ready schema. That gap is precisely what Datadory's delivery layer closes - one consistent shape across both records, whatever the cadence your pipeline wants.
The verdict
Verdict: sample both, pick by fit - they score 10 and 9 out of 10, so the question decides, not the leaderboard.
Pick the GLEIF Golden Copy and Delta Files (LEI Reference Data) when your problem is identity-shaped: resolving which of eleven spelling variants of a group belongs to which parent, screening new counterparties before onboarding, building supply-chain or ownership graphs, or feeding an ESG scope-3 exercise with ultimate-parent attribution. It is the stronger record on freshness, on geographic span and on audit trail.
Pick the EIA Consumption & Efficiency Surveys (RECS, CBECS, MECS) when your problem is physics-shaped: benchmarking a portfolio's buildings against national energy intensity, sizing the market for heat pumps or efficiency retrofits, modelling residential load curves, or estimating emissions from actual metered consumption rather than floor-area guesses. It is the stronger record on measurement depth, on statistical rigour and on historical benchmark cycles.
If your work sits where those meet - say, attributing energy footprint to the corporate groups that own the buildings - neither substitutes for the other, and the honest answer is to hold both.
Sample both, pick by fit. See GLEIF Golden Copy and Delta Files · See EIA Consumption & Efficiency Surveys
Or take both in one feed
Yes, and the join runs through ownership rather than keys. A defensible workflow: resolve each CBECS or RECS geography's dominant building operators to their legal entities, walk the Level 2 graph to the ultimate parent, then roll surveyed consumption and expenditure up the hierarchy so a group-level energy footprint rests on measured buildings instead of sector averages. The replicate weights keep the error bars honest on the way up.
Daily, weekly, or hourly.** > > Name the cadence your decision actually needs - registry-fresh hourly pulls suit screening queues; annual survey vintages land once and stay good. Delivered through one schema, the LEI hierarchy lands next to the end-use microdata in the same warehouse without a second integration, and the sample arrives pre-cut to the jurisdictions, sectors and cycles you name.
Or take both in one feed.
API, files, or your warehouse. Daily, weekly, or hourly.
Fair questions
Is GLEIF Golden Copy and Delta Files (LEI Reference Data) better than EIA Consumption & Efficiency Surveys (RECS, CBECS, MECS)?
Better at different jobs. GLEIF wins on reach and freshness: 3,408,526 deduplicated entities across every Global LEI System jurisdiction, 485,148 typed ownership links with precomputed delta views. The EIA surveys win on measurement depth: 18,500 households and 6,436 buildings with hundreds of consumption, expenditure and weight columns apiece. Sample both and match each to the question.
Do the two datasets cover the same organizations?
Only incidentally. The GLEIF register lists any legal entity that obtained an LEI - banks, funds, corporates, subsidiaries worldwide - regardless of what energy it consumes. The EIA surveys cover US households, commercial buildings and manufacturing plants, which appear as respondents, not as identified registrants. There is no shared key; the connection runs through entity resolution, not through the files themselves.
Which of the two has more fields per record?
The EIA microdata, decisively. CBECS 2018 carries 1,249 columns per building and RECS 2020 has 799 per household, including supplier-derived consumption, expenditures, appliance inventories and replicate weights. The GLEIF Level 2 relationship header defines around sixteen documented elements per row; Level 1 entity records are broader but remain identity attributes rather than behavioural measurements.
Which one should back a corporate hierarchy model?
The GLEIF files, without competition. Their Level 2 records type each link - direct consolidation, ultimate consolidation, fund management, sub-fund - qualify it with accounting periods, grade corroboration through validation-source fields, and log 6.3 million exceptions where a parent could not be determined. The EIA surveys contain no ownership fields at all; their units are dwellings and buildings, not corporate structures.
Can Datadory deliver both datasets together?
Yes. Either record arrives alone or both land aligned on one calendar, delivered daily, weekly, or hourly - your call. Say which jurisdictions, sectors and survey cycles you need when you request the sample and it arrives pre-cut, with field definitions and coverage profiles attached. Or take both in one feed.