What version 0.1.0 specifies
Version 0.1.0 specifies a prospective cross-dataset measurement layer that separates candidates from decisions and records organizational level, temporal validity, confidence semantics, and unresolved cases.
Open Accounting–Labor Data Linkage Hub · Sequence 02
Lightcast ↔ Revelio company crosswalk
A planned company-year crosswalk for reconciling employer entities across job-posting and employee-history data while preserving parent, subsidiary, and temporal ambiguity.
01 · Research design
Why this linkage is a measurement problem
Combining job-posting and employee-history data requires more than a company-name join. Corporate groups may appear at different organizational levels, websites and locations can conflict, and the economically relevant link can change over time. The resource therefore treats organizational level and temporal validity as measured attributes.
Intended uses after validation
- Connect recruiting measures to employee-history outcomes using a documented company-year mapping.
- Distinguish parent-level and operating-subsidiary interpretations instead of silently collapsing them.
- Evaluate sensitivity to organizational level, temporal continuity, and unresolved common-name cases once benchmark evidence is released.
02 · Measurement object
Entities, decision unit, and output structure
The definitions below describe the intended analytical object. The scope statement is deliberately separated from released coverage so that a design target cannot be mistaken for an observed sample.
- Source entities
- Lightcast company entities represented by employer names and authorized matching features.
- Target entities
- Planned accepted target: a Revelio company identifier with an explicit target level—operating company, ultimate parent, or other documented level—rather than an implicit parent substitution.
- Decision unit
- One Lightcast company entity × calendar year decision, with non-overlapping intra-year validity intervals only when a documented corporate event changes the accepted company or organizational level.
- Candidate-audit row
- One Lightcast company entity × calendar year × candidate Revelio company row, retaining target level, alternatives, and evidence components.
- Accepted-release row
- One accepted Lightcast entity × validity interval × Revelio company identifier × target-level row; unresolved and no-candidate decisions remain in a separate decision table.
- Cardinality policy
- At most one co-valid accepted company-level interpretation per source decision. Non-overlapping event intervals and many-to-one mappings are retained and diagnosed; unresolved co-valid one-to-many cases are not forced to a winner.
- Target scope
- Planned U.S. company-year scope, annual versions, 2010–2025. Actual released coverage is not yet established.
03 · Disclosure boundary
What is public—and what remains restricted
The public contribution is the measurement design, reproducible reference logic, release metadata, and eventually benchmark-supported aggregate diagnostics. Possession of licensed records is not treated as redistribution authority.
Available in version 0.1.0
- A public-safe methodology and field contract.
- A generic TypeScript reference implementation using synthetic inputs.
- A synthetic example with reserved identifiers and .example domains.
- A machine-readable schema, release manifest, checksums, citation file, license, and changelog.
- An explicit evidence register showing which validation objects are absent, private, or not yet measured.
Not redistributed
- Row-level Lightcast or Revelio records.
- Full private crosswalks or licensed source identifiers.
- Private bucket locations, signed URLs, credentials, or internal absolute paths.
- Unpublished research outcomes or coauthor-controlled records.
- Quantitative reliability claims that are not supported by a released adjudication benchmark.
04 · Matching method
A staged, inspectable decision process
Candidate generation, scoring, acceptance, and adjudication are separate stages. The public implementation demonstrates mechanics with synthetic data; it does not reveal private thresholds or substitute model scores for labeled evidence.
Normalize names and preserve raw evidence
AvailableApply transparent normalization while retaining original source strings and field availability for adjudication.
Generate candidates from multiple public-safe signals
AvailableUse normalized names, authorized domains, geography, and token blocks to create alternatives before ranking.
Separate company and ultimate-parent candidates
Not releasedRecord the organizational level of each candidate and do not treat a parent match as equivalent to an operating-company match.
Score and retain competing alternatives
AvailableExpose score components, candidate rank, and the margin over the next candidate. Missing geography or validity evidence receives no agreement credit, and scores remain separate from final decisions.
Check geography, domain, and adjacent-year consistency
Not releasedFlag contradictions rather than letting a strong name score override organizational or temporal evidence.
Adjudicate a development set and frozen holdout
Not yet measuredRecord accepted, rejected, unresolved, and no-candidate outcomes separately from candidates; stratify labeled source decisions across confidence and error classes; and evaluate the frozen holdout against predeclared criteria without retuning.
Acceptance principle
A link is not accepted solely because it increases coverage. Competing candidates, unresolved cases, and the reason for each decision remain part of the measurement record.
05 · Validation evidence
What has—and has not—been validated
Reliability labels require adjudicated evidence. This prototype reports no precision, recall, unresolved-rate, or temporal-consistency estimate because no released frozen holdout supports those numbers.
Inspect the shared benchmark protocol and evidence register.
| Evidence object | Public state | What the current release supports |
|---|---|---|
| Method and field contract | Available | The shared public methodology, separate candidate/decision schema, and runnable tested package cover this linkage class. |
| Production linkage rows | Private | No Lightcast or Revelio row-level mapping is redistributed. |
| Authoritative private source chain | Not released | Candidate mapping objects exist, but public provenance has not resolved one canonical production chain. |
| Parent/subsidiary adjudication | Not released | No public labeled set supports claims about organizational-level accuracy. |
| Frozen holdout | Not released | No released holdout supports an out-of-sample reliability claim. |
| Precision, recall, and unresolved rate | Not yet measured | No quantitative metric is reported in this prototype. |
| Temporal consistency | Not yet measured | Adjacent-year continuity has not been reconciled in a public evaluation output. |
06 · Error taxonomy
Where the linkage can fail
The relevant errors are not limited to spelling distance. Organizational level, time, duplicated source entities, common-name collisions, and missing candidate features can change the economic interpretation of a match.
Legal-name variationNames
Suffixes, punctuation, abbreviations, and branding differ while the underlying entity remains the same.
Abbreviation and token lossNames
Short names or acronyms remove the tokens needed to distinguish firms.
Parent versus subsidiaryCorporate structure
The source employer and target company refer to different levels of the same corporate group.
Merger, acquisition, or spin-offCorporate structure
A relationship is valid only for part of the time series or changes organizational meaning.
Historical name changeTime
A valid name at one date becomes stale after a rename or reorganization.
Stale temporal linkTime
A candidate matches on name but falls outside the target identifier's valid interval.
Duplicate source entitiesCollisions
Multiple source records represent the same employer or business unit.
Common-name collisionCollisions
Several unrelated entities share a short or generic name.
Many-to-one concentrationCollisions
Multiple source entities map to one target and may represent either valid subsidiaries or over-collapsing.
One-to-many ambiguityCollisions
One source entity has several plausible targets without enough evidence for deterministic acceptance.
Geography or domain conflictSource coverage
Name similarity conflicts with location or website evidence.
Source missingnessSource coverage
A required blocking or validation field is unavailable, producing no candidate or lower-confidence review.
07 · Reproducibility
Reproduce the method with licensed inputs
The downloadable package uses generic adapters and synthetic examples. A researcher supplies independently licensed source data, maps fields to the public schema, runs the reference stages, and evaluates the resulting candidates against their own adjudication evidence.
Markdown
Methodology
Measurement scope, matching stages, evidence ceiling, validation gates, and disclosure boundary.
Download MethodologyCSV
Synthetic example
Reserved synthetic identifiers and .example domains for testing the reference implementation.
Download Synthetic exampleJSON Schema
Public schema
Generic input and candidate-output fields without licensed values.
Download Public schemaRunnable npm package (.tgz)
Reference implementation
Self-contained TypeScript source, package metadata, tests, license, and synthetic inputs; extract it, run npm install, then npm test.
Download Reference implementationJSON
Release manifest
Allowlisted files, byte sizes, media types, and SHA-256 checksums.
Download Release manifestCFF
Citation file
Machine-readable citation metadata for the prototype method release.
Download Citation fileText
License
License for original public-safe code and artifacts only; source-dataset rights are excluded.
Download LicenseMarkdown
Changelog
Version history and explicit statements of what each release does not contain.
Download ChangelogCitation
Yang, Hongye Chris. 2026. Open Accounting–Labor Data Linkage Hub: Public-Safe Entity-Linkage Methods and Validation Protocol, version 0.1.0.
Release history
Version 0.1.0 · Prototype method release
Version 0.1.0 publishes the common method and validation protocol while identifying source authority and parent/subsidiary adjudication as unresolved evidence requirements.
Accepted limitations
- This resource follows the Lightcast↔CCM sequence and remains an earlier-stage prototype.
- The authoritative mapping lineage, organizational-level policy, and versioned production output are not public evidence.
- No adjudicated benchmark or reliability estimate is released.
- Licensed identifiers and source records remain private even when their field roles are described.