qe-replication-and-data-policy

Replication & Data Policy (qe-replication-and-data-policy)

When to trigger

• You are preparing data and code for a QE submission or an accepted paper
• You need to know exactly what the ES Data Editor checks and when
• Some data are proprietary or restricted and you need to plan an exemption
• You want the package built so the pre-acceptance reproducibility check passes on the first pass

The QE / Econometric Society replication regime (source map refreshed 2026-06-20)

QE follows the Econometric Society Data and Code Availability Policy, shared across Econometrica, Quantitative Economics, and Theoretical Economics and compatible with DCAS (the Data and Code Availability Standard). Key facts:

• The Society publishes empirical / experimental / simulation papers only if data and code are clearly documented and non-exclusive to the authors.
• Before acceptance, authors must provide raw data, code, and documentation sufficient to replicate all results in the paper and approved online appendices.
• The ES Data Editor process conducts reproducibility checks before final acceptance. Partial-check scope must be documented in the README.
• Replication / supplementary materials are posted with the article.
• For long-running or hard-to-access computations, simplified/manageable versions and summary output files are encouraged.
• Any request for exemption or limits on data/code access must be stated at initial submission and is at editor discretion.

> Important: QE uses this centralized ES system and the ES Data Editor Website, NOT the JAE (Journal of Applied Econometrics) Data Archive. Do not prepare a JAE-style deposit.

Building the package

1. Raw data (or, for restricted data, the access pathway + the code that would run on it) plus all intermediate data-build steps.
2. Code that runs end to end: one master script (run_all) regenerating every table, figure, and number from raw inputs.
3. Documentation / README: data sources and licenses, software and exact versions, hardware/run-time notes, the mapping from scripts to exhibits, seeds, and any partial-check scope.
4. Environment pinning: renv.lock, requirements.txt/conda, Project.toml/Manifest.toml, recorded Stata ssc/net versions.
5. Heavy computations: include manageable versions and summary output files so the Data Editor can verify without re-running everything.
6. Experimental/own-data: include instructions/survey transcripts and the pre-registration reference (effective Jan 1, 2026).

Proprietary / restricted data

• State the exemption or access-limit request at initial submission (not at acceptance).
• Provide all code even when raw data cannot be shared, plus instructions to obtain access and synthetic or sample data where possible.
• Document exactly which results the Data Editor can and cannot reproduce, in the README.

Checklist

• [ ] Raw data + full build pipeline included (or restricted-data access path + code)
• [ ] One master script regenerates every result from raw inputs
• [ ] README maps scripts to exhibits; lists sources, versions, seeds, run times
• [ ] Environment pinned across all languages used
• [ ] Heavy computations have manageable versions + summary output files
• [ ] Any exemption / access limit stated at initial submission
• [ ] Partial-check scope documented in the README
• [ ] Built against the ES Data Editor regime (DCAS), not the JAE archive

Anti-patterns

• Preparing a JAE Data Archive deposit by mistake (QE uses the ES system)
• Leaving the package until acceptance — the reproducibility check is before acceptance
• Code that depends on absolute local paths or unpinned package versions
• Requesting a proprietary-data exemption only at acceptance instead of at submission
• A README that does not map scripts to the specific tables and figures

Output format

【Regime】ES Data and Code Availability Policy (DCAS) — NOT JAE archive
【Data】raw + build pipeline included? (or restricted-data path + code) [Y/N]
【Master script】regenerates all results from raw? [Y/N]
【README】sources/versions/seeds/script-to-exhibit map? [Y/N]
【Heavy computation】manageable version + summary outputs? [Y/N or N/A]
【Exemption】stated at initial submission? [Y/N or N/A]
【Next step】qe-review-process (pre-acceptance Data Editor check)