Replication Package & AEA Data Policy (aeja-replication-package)

Replication Package & AEA Data Policy (aeja-replication-package)

When to trigger

• The paper is empirical and you are heading toward acceptance (or want to build the package early)
• An R&R or conditional acceptance asks you to prepare the data + code deposit
• You need to write the README and Data Availability Statement to AEA standard
• Some data are restricted/proprietary and you must plan the deposit around that

Why this is the AEJ: Applied signature

The single most distinctive AEJ: Applied differentiator is the AEA Data and Code Availability Policy, administered by the AEA Data Editor (Lars Vilhuber). Papers offered a revise-and-resubmit are asked to submit a data replication package at resubmission, and accepted papers must clear the Data Editor's compliance and reproducibility checks before publication. The default home is the AEA Data and Code Repository on openICPSR; other trusted repositories require appropriate Data Editor access. Build this package as you analyze; do not treat it as an acceptance-day formality.

What the package must contain

Handling restricted or proprietary data

• You cannot deposit restricted data, but you must still deposit all code and document the exact access procedure (provider, application steps, cost, approximate wait).
• Provide a small synthetic or public extract so the code runs and the Data Editor can verify logic where possible.
• Declare any restricted-data or exemption request at the earliest opportunity — limited-access arrangements are at editor/Data-Editor discretion and must be flagged, not discovered at the check.

Reproducibility hygiene (build as you go)

• Pin versions: Stata version + recorded ssc/net package versions; requirements.txt/conda env (Python); renv.lock (R).
• Set and report seeds for every simulation, bootstrap, and randomization-inference step.
• No absolute paths — one root macro/variable; relative paths thereafter.
• Exhibit-to-code map in the README: Table 3 → code/05_main.do, Figure 2 → code/06_event_study.R, etc.
• Run it clean on a fresh checkout before depositing; the Data Editor will.

Adapt the vendored skeleton in ../../resources/code/ (master script → clean → descriptive → DID/IV/RD/DML → mechanism → robustness → tables) as the package backbone.