brycewang-stanford Topic Selection for JAE (jape-topic-selection)
When to trigger
- Choosing where to send an empirical or method-application paper and weighing JAE
- Unsure whether your idea is "applied enough" or "too theoretical" for JAE
- Deciding between a standard Research Article and JAE's Replication Article track
JAE's scope test
JAE publishes applied econometrics: papers that apply and develop econometric techniques on real data, with the focus on the application rather than pure econometric theory. A purely theoretical contribution (a new estimator with asymptotics but no real-data application) is off-fit — route it to a methods/theory outlet. A good JAE topic does one of:
- Applies an established or newly adapted method to a substantive economic question on real data, with credible inference;
- Develops a technique but anchors it in a real-data application that demonstrates and tests it; or
- Replicates previously published empirical results (the dedicated Replication Article category), reporting successes and failures.
The non-negotiable filter: replicability
Because accepted papers must deposit data and code in the JAE Data Archive (since 1994, now at ZBW, unless confidential), ask before you start: can the data behind every result be deposited as plain ASCII/CSV with a readme, or at minimum documented enough that others can apply for access? If the data can be neither shared nor described for access, the project is a poor JAE fit.
Fit matrix
Score the candidate on five gates before drafting:
- Economic question: the paper answers a substantive question, not only demonstrates an estimator.
- Econometric lesson: readers learn when or why a method changes an applied conclusion.
- Real-data anchor: the central evidence uses actual data, not only simulation or asymptotics.
- Replication path: every result can be regenerated from depositable or clearly documentable inputs.
- Article track: the paper is either a standard applied contribution or a Replication Article; do not mix the two without saying which promise is primary.
If any gate fails, redirect early. JAE fit is strongest when the data, method, and applied question are mutually necessary; it is weakest when one of the three can be removed without changing the paper.
Scope verdict table
Pattern-match the project against recurring candidate shapes:
| Project shape | JAE verdict | Why |
|---|---|---|
| New estimator + asymptotics + small empirical illustration | Borderline | Fit hinges on whether the application carries the paper; if the illustration is decorative, reroute to a theory/methods outlet |
| Established method, new real-data finding with credible inference | Strong fit | The venue's bread and butter — provided the deposit is feasible |
| Pure Monte Carlo comparison of estimators | Off-fit alone | Anchor the simulations to a real empirical problem or send to a methods journal |
| Re-examination of a prominent published result using its archived data | Strong fit (Replication Article) | The dedicated track exists exactly for this |
| Policy evaluation on administrative data that can never be shared or described | Poor fit | The mandatory archive deposit cannot be satisfied even via the confidential-data readme route |
Choosing a Replication Article target
The track rewards replications of prominent papers — results people teach, cite, or build policy on. Screen a target on four points: (i) the original's data are archived (ideally in the JAE Data Archive) or otherwise reconstructible; (ii) the claim is sharp enough that confirm/qualify/overturn is decidable; (iii) you can separate data-revision effects from coding differences from genuine fragility; (iv) a negative result would still be informative — JAE's track publishes failures as well as successes, which is rare and worth exploiting. Replicating an obscure paper, or one whose data are gone, fails the screen however careful the execution.
Worked fit assessment: two candidates (illustrative)
Candidate A: a new shrinkage estimator with proofs, demonstrated on one well-worn growth dataset where it barely changes the estimates. The real-data anchor is decorative — redirect to a methods venue, or find an application where shrinkage flips a conclusion. Candidate B: quarterly energy-demand elasticities for 14 countries, where switching from textbook HAC to a few-cluster-appropriate bootstrap moves the headline elasticity from "significant at 5%" to marginal — an applied question, an econometric lesson, public data exportable to CSV. B is the JAE paper despite A being technically deeper.
Output format
【Scope】applied on real data? [Y/N] | pure-theory risk? [Y/N]
【Replicable】data depositable or documentable? [Y/N]
【Lesson】method choice changes an applied conclusion? [Y/N]
【Track】Research Article / Replication Article
【Verdict】JAE fit / redirect → where
Supplementary resources
../../resources/official-source-map.md— JAE scope and Data Archive sources