red-identification-strategy

Identification & Model Logic for RED (red-identification-strategy)

When to trigger

• Establishing why the paper's central claim is credible, before robustness
• Unsure whether RED expects a causal-design argument or a model-assumptions argument
• A computational paper where "identification" means parameter discipline, not instruments

Branch by paper type (RED takes all three)

Theoretical / computational dynamic models

The credibility question is about assumptions, existence, and discipline, not instruments:

• State the model assumptions and regularity conditions explicitly (preferences, technology, stationarity, boundedness, transversality); flag where existence/uniqueness of equilibrium is proved or assumed.
• Make proof exposition clean: state results as propositions, separate assumptions from claims, and put long proofs in an appendix while keeping the intuition in the body.
• Show parameter discipline — which parameters are calibrated to data targets, which are estimated, and which are free; justify each so results are not an artifact of free parameters.
• Discuss generality: what survives relaxing key assumptions, and where the result is knife-edge.

Methodological / computational-method papers

• State the method's regularity conditions and where they bind; characterize accuracy and convergence of the numerical solution; report asymptotics where the method estimates parameters.
• Provide Monte Carlo / numerical experiments that show the method works under known data-generating processes.

Empirical dynamic papers

• Make the causal/identification design explicit (the source of variation, the exclusion logic, the dynamic structure being estimated — e.g., VAR identification, local projections, structural estimation).
• Tie the empirical object back to what it disciplines in the dynamic model.

Checklist

• The right branch is chosen for the paper type
• Assumptions/conditions (theory) or identifying variation (empirics) are explicit and defended
• Parameter discipline is documented; results are not driven by undisciplined free parameters
• Generality / accuracy / robustness of the core claim is characterized

Anti-patterns

• Importing reduced-form "identification" language into a calibrated model where it does not apply
• Hiding free parameters or equilibrium-existence gaps
• Asserting generality without showing what relaxing the assumptions does

Parameter-discipline table

For quantitative papers, create a table with one row per key parameter:

Parameter	Value	Source/target	Free or disciplined?	Sensitivity shown?

Any parameter that is free and influential needs a sensitivity check or a narrower claim.

Model-solution audit block

For computational claims, attach an audit record so a referee can see what the numbers rest on:

SOLUTION AUDIT — [model name]
  Method:       EGM on the household problem; sequence-space Jacobian for GE transitions
  State space:  assets 250 pts (log-spaced); productivity 7-state Rouwenhorst
  Convergence:  policy-function sup-norm < 1e-9; market clearing < 1e-7
  Accuracy:     max log10 |Euler error| = -4.3 (off-grid simulation, 100k agents)
  Refinement:   headline counterfactual moves < 0.5% when grids are doubled
  Existence:    stationary-equilibrium existence proved/cited in Appendix A

Any blank line means the matching claim in the text should be weakened until the line can be filled.

Worked discipline review: a search-and-matching draft

A draft calibrates a Diamond–Mortensen–Pissarides economy and claims wage rigidity explains unemployment volatility. Illustrative review of its parameter discipline:

• Matching elasticity 0.5, externally set from the literature — acceptable, but the volatility claim is sensitive to it, so a ±0.15 band belongs in the robustness section.
• Replacement rate 0.71, internally calibrated to market tightness — a RED referee will notice this sits near the Hagedorn–Manovskii region where small match surplus generates volatility mechanically; report the result at a conventional 0.4 as well.
• Rigidity parameter calibrated to the very volatility moment being explained — circular. Move that moment out of the target set, or downgrade "explains" to "is consistent with".

Credibility objections RED referees raise

Objection	Branch	Fix
"A free parameter drives the result"	quantitative	sensitivity table or a narrower claim
"Equilibrium existence is assumed silently"	theory	state it as an assumption or prove it
"Accuracy not stress-tested at the calibrated point"	computational	Euler/den Haan check at exactly that parameterization
"The reduced-form estimate maps to no model object"	empirical	name the structural parameter or moment the estimate disciplines

Supplementary resources

• ../../resources/external_tools.md — solvers and estimation toolkits
• ../../resources/official-source-map.md — scope sources