jme-data-analysis - SKILL.md Agent Skill

name: jme-data-analysis description: Use when building or stress-testing the empirical/quantitative analysis for a Journal of Monetary Economics (JME) manuscript — VAR/SVAR, local projections, DSGE estimation, moment matching, IRFs, and FEVDs — to monetary-economics and macroeconomics norms. Covers estimation choices, inference, and robustness.

Data & Quantitative Analysis (jme-data-analysis)

When to trigger

The estimation runs but referees will question the specification or inference
You must decide between a VAR, a proxy-SVAR, and local projections
A DSGE is estimated/calibrated and needs convergence and fit diagnostics
The robustness battery for a macro paper is unclear

Macro-empirical norms at JME

JME analysis is aggregate and policy-relevant, so the workhorses are different from micro-econometrics. The core toolkit:

VAR / SVAR / proxy-SVAR for dynamic responses to identified shocks; report impulse responses with confidence/credible bands, lag selection, stability, and forecast-error variance decompositions (FEVDs).
Local projections (Jordà) as a robustness counterpart to VAR IRFs; show both when feasible, since LP trades variance for robustness to misspecification.
DSGE / quantitative models estimated by Bayesian methods (Dynare) or calibrated to micro moments; report prior/posterior plots, MCMC convergence, identification (Iskrev), and posterior predictive / second-moment fit.
Real-time data (FRED/ALFRED vintages, Greenbook/Tealbook) where the information set matters — using final-revised data to study a real-time policy decision is a known pitfall.

Inference must match the design: HAC / Newey–West or clustered standard errors for time-series regressions and local projections; credible intervals from the posterior for Bayesian DSGE; bootstrap or analytical bands for VAR IRFs. Report units consistently — e.g., responses to a 100-basis-point or one-standard-deviation policy shock.

Robustness battery (macro)

Alternative lag lengths, sample splits (e.g., pre/post-1984 Great Moderation, ZLB period), and sub-samples
Alternative identification (ordering, restriction set, instrument) and LP-vs-VAR comparison
Real-time vs. revised data; alternative shock series
For DSGE: prior sensitivity, alternative calibrations, and the mechanism on/off comparison
Zero-lower-bound / effective-lower-bound treatment where the sample spans it

Execution bridge (StatsPAI / Stata MCP)

Run the battery, don't just enumerate it. Full map: execution-with-mcp. JME is monetary macro — SVAR, local projections, high-frequency identification; local_projections/irf are in StatsPAI, DSGE/calibration is outside this toolchain.

Many outcomes / specifications: romano_wolf (step-down FWER) or benjamini_hochberg.
OVB sensitivity: oster_delta / sensemakr.
Inference: wild_cluster_bootstrap (few clusters), twoway_cluster / conley.
Re-fit off one handle: audit_result(result_id) lists missing checks + the exact suggest_function for each.
Exhibits: etable / did_summary_to_latex from the handle — no retyped numbers.

Decisive checks in the body, exhaustive battery in the appendix. JF execution walkthrough.

Checklist

IRFs reported with bands; FEVDs where informative
LP and VAR compared where the question allows
Inference matched to the design (HAC/cluster, posterior bands, bootstrap)
Real-time vs. revised data considered
DSGE convergence / identification / fit diagnostics reported
Shock units stated and consistent across exhibits
Robustness pushed to the online supplement to respect the 40-page / ≤10-exhibit cap

Anti-patterns

Recursive SVAR ordering presented as the only identification with no robustness
Using final-revised data to model a real-time policy choice
Reporting a single DSGE point estimate with no convergence or prior-sensitivity evidence
IRFs without bands, or with inconsistent shock units across figures

Evidence pass for Journal of Monetary Economics

Treat this skill as an executable review pass, not a prose hint. First lock the main macro object, the identifying variation, and the policy-relevant counterfactual; then judge whether the current manuscript answers the venue's real reader: macro and monetary economists who expect the shock, mechanism, and policy margin to be visible early.

Do the pass: Audit the research design before polishing prose: unit of analysis, comparison set, uncertainty, sensitivity, missingness, and reproducibility must be visible.
Return a ledger: give claim / evidence / risk / manuscript location rows, so the next agent can edit rather than rediscover the issue.
Sibling guard: compare against JIE for open-economy trade/finance emphasis, RED for dynamic macro theory, AEJ Macro for broader field positioning; if a sibling owns the contribution, recommend re-routing before polishing format.
Stop condition: do not give submission-ready advice until the pack's resources/official-source-map.md has been checked for volatile rules and the manuscript has one concrete fix for the largest venue-specific risk.

Output format

【Method】VAR / SVAR / proxy-SVAR / LP / DSGE / mixed
【Inference】HAC / cluster / posterior bands / bootstrap
【IRFs + FEVDs】reported? Y/N
【LP-vs-VAR】reported? Y/N/NA
【Real-time data】used where needed? Y/N
【Robustness done / missing】[...]
【Next step】jme-tables-figures