ecta-workflow

Econometrica Workflow (ecta-workflow)

Overview

This is the router. It does not replace any specialized skill — it tells you which ecta- skill to invoke at the current stage*.

Default assumption: unless the user says otherwise, the target is Econometrica, the journal of the Econometric Society (published by Wiley; founded 1933) — the flagship for econometric methods and rigorous economic theory (microeconomic, game, and decision theory, plus mathematical economics). Its charter is the unification of the theoretical-quantitative and empirical-quantitative approaches, penetrated by constructive and rigorous thinking. The bar is mathematical rigor: a new method or theorem with a complete, correct proof, generality, and a clear economic or inferential payoff.

This is the key re-slant from a sibling like AER / QJE / JPE / REStud: Econometrica's core product is the theorem or the estimator, not a clean causal estimate of a policy effect. A flawless applied difference-in-differences or RDD paper with an off-the-shelf method is off-fit here even if it would headline a general-interest top-5 journal. Lineage that fits: Heckman (1979, selection), Hansen (1982, GMM), Newey–West (1987, HAC), Rust (1987, NFXP), Kahneman–Tversky (1979, prospect theory) — all Econometrica.

Econometrica papers come in two broad shapes; route accordingly:

• Theory / methods (econometric theory, micro / game / decision theory): the contribution is a theorem, an estimator with its asymptotics, or an axiomatic result.
• Structural / empirical: a method applied with care, or a structural model whose identification and estimation are the contribution. These still need the methodological spine and a replication package.

When to trigger

• "What should I do next?" on an Econometrica-bound paper
• A draft arrives and you must diagnose the binding constraint
• You are oscillating between proofs, simulations, and writing and have lost the thread
• A decision letter / referee reports arrive and you need to switch to rebuttal mode

Routing table

Default order

1. ecta-topic-selection — fix the contribution: a general theorem or a method with real payoff
2. ecta-literature-positioning — locate the result precisely in the methods / theory lineage
3. ecta-identification — identification conditions + asymptotics (or axioms + existence/uniqueness)
4. ecta-theory-model — state the central theorem(s); build the proof strategy; check generality
5. ecta-robustness — Monte Carlo, finite-sample performance, regularity / edge cases
6. ecta-tables-figures — finalize simulation tables and any empirical exhibits
7. ecta-writing-style — terse formal polish; assumptions visible; theorem-numbered
8. ecta-replication-package — code (+ data) under the ES Data and Code Availability Policy (Data Editor reproducibility check at conditional acceptance; Zenodo deposit)
9. ecta-referee-strategy — red-team the proofs and the generality claim
10. ecta-submission — Editorial Express preflight, 45-page limit, Supplemental Material assembly
11. ecta-rebuttal — after the R&R / conditional-acceptance letter

ecta-writing-style is a late polish stage — do not polish prose while the central theorem or its proof is still unstable. Likewise, do not build tables before the identification / asymptotics are settled.

Submission-package tree

Keep the working directory shaped like the final Editorial Express upload from day one — the Zenodo deposit and Data Editor check arrive at conditional acceptance, when it is too late to reconstruct a replication folder from scratch:

econometrica-submission/
├── main.pdf                 # ≤45 pages incl. references and appendices
├── supplement.pdf           # Supplemental Appendix, ≤25 pages
├── cover-letter.pdf         # ES membership stated; fee tier noted
├── replication/             # ES Data and Code Availability Policy
│   ├── README.md            # exact commands regenerating every exhibit
│   ├── code/                # Monte Carlo + estimation scripts
│   └── data/                # raw data, or access instructions if proprietary
└── working-notes/           # not uploaded — internal gates
    ├── theorem-ledger.md    # per theorem: statement / proof / regularity conditions
    └── referee-redteam.md   # output of ecta-referee-strategy

A purely theoretical paper drops data/ but keeps code/ if any result is verified numerically; an estimator paper's "data" may be entirely its simulation code.

Decision mnemonics

• "I have a result but I'm not sure it's general enough" → ecta-topic-selection
• "A referee will ask how this differs from [classic estimator/theorem]" → ecta-literature-positioning
• "My estimator is consistent but I never derived its limiting distribution" → ecta-identification
• "My proof says 'it is easy to see that'" → ecta-theory-model
• "I report asymptotics but never simulated finite samples" → ecta-robustness
• "My size/power table has eight columns and no notes" → ecta-tables-figures
• "My intro is three pages of motivation before any result" → ecta-writing-style
• "I can't regenerate Table 2 from a clean checkout" → ecta-replication-package
• "Where will the referee find the hole?" → ecta-referee-strategy
• "Submitting tomorrow" → ecta-submission
• "Three reports back, all asking for more generality" → ecta-rebuttal

Differences vs. applied-economics packs (AER / QJE / JPE / REStud)

If the contribution is an application — a causal estimate of a policy effect, with the method off-the-shelf — an applied general-interest pack (e.g., AER / QJE / JPE / REStud skills) is the better fit. The core difference:

• Econometrica: the contribution is the method or theorem — generality, complete proofs, asymptotics, and finite-sample evidence are the product. A purely theoretical paper carries no replication package; an estimator paper's "data" may be entirely its Monte Carlo code.
• General-interest applied: the contribution is the answer to an economic question — identification of a specific effect carries the paper, and a clean empirical narrative matters more than a new limit theorem.

Concrete Econometrica-specific facts that are wrong for those siblings: at least one author must be a member of the Econometric Society to submit; the 45-page limit (incl. references and appendices) plus a ≤25-page Supplemental Appendix; submission via Editorial Express with an ES-member submission fee (US$125 regular / US$50 student from 2025); and replication is governed by the Econometric Society Data and Code Availability Policy with Data Editor checks and a Zenodo deposit — not the AEA Data Editor / openICPSR pipeline the AER/AEJ packs assume.

Anti-patterns

• Do not skip ecta-identification / ecta-theory-model and jump to simulations — referees read the proofs first.
• Do not let ecta-tables-figures polish exhibits before the asymptotics are settled.
• Do not let ecta-writing-style smooth prose over a proof gap — fix the math first.
• Do not let ecta-rebuttal draft a response before the revised theorems and code are done.