Jimf Tables Figures

Tables & Figures (jimf-tables-figures)

When to trigger

• The main result is settled and must be readable at a glance
• A cross-country panel table is wide and dense; the headline coefficient is hard to find
• An event-study / impulse-response / local-projection plot must carry the identification visually
• Country-level heterogeneity (AE vs. EM, by region) needs a map, scatter, or panel of plots
• You are preparing exhibits for an Elsevier submission and want them house-style clean

The JIMF exhibit bar

JIMF is an empirical finance/macro journal, so the headline international effect should be findable in seconds and the picture of the identification should appear before the reader reaches the table. Finance house style (unlike the Econometric Society journals) permits significance stars but requires standard errors in parentheses and self-contained notes. The distinctive JIMF exhibits are the cross-country panel regression table, the high-frequency event-study / impulse-response figure, and the country-heterogeneity display.

Exhibit craft for international finance

1. One table for the headline. The preferred specification should let a referee read the coefficient, its SE, the fixed-effect structure (country / time / two-way), N, and the dependent-variable mean without flipping pages. Demote the control sweep to the appendix.
2. Make the fixed effects and clustering legible. In a cross-country panel the FE structure is the identification; show "Country FE: Yes / Time FE: Yes" rows and state the clustering level in the note.
3. Figures carry the identification. Event-study windows, impulse responses, and pre-trend leads persuade more as figures with CI bands than as prose; mark t=0 and the window; state units (basis points vs. percent).
4. Show heterogeneity, don't hide it. When AE and EM differ, a forest plot of country/region coefficients or a scatter against ex-ante exposure is worth more than a pooled number.
5. Right precision and units. Two to three significant figures; state whether the exchange-rate effect is per 1% or per 1 s.d., per 25bp surprise, etc.
6. Self-contained notes. Sample, country set, period, frequency, FE, clustering, and star meaning all in the note.

Execution bridge (StatsPAI / Stata MCP)

Generate exhibits from the fitted result, not by retyping numbers. Full map: execution-with-mcp. JIMF is international macro-finance; cross-country panels + asset pricing — identification plus factor/Newey-West inference.

• Tables: etable (multi-model) or did_summary_to_latex straight from the result_id.
• Figures: plot_from_result / enhanced_event_study_plot / event_study_table — axis units and the SE/clustering note baked in.
• Every note names the estimator + clustering and states the magnitude in interpretable units.

See a full fitted-result → exhibit chain in the JF execution walkthrough.

Checklist

• Headline coefficient readable in one table: estimate, SE in parentheses, FE rows, N, dep-var mean
• Fixed-effect structure and clustering level stated explicitly (the identification, made visible)
• Identification figure present (event-study / IRF / local projection) with CI bands and a marked t=0/window
• Country/region heterogeneity displayed where it exists (forest plot / scatter vs. exposure / map)
• Units and the per-unit scaling stated (per 1%, per 1 s.d., per 25bp)
• Stars (if used) defined; SEs reported everywhere; precision 2–3 sig figs
• The underlying estimate is settled (identification + robustness) before the exhibit is polished
• Control sweep demoted to the appendix; the body carries only the preferred specification
• Event-study / IRF figures follow the canonical form (marked t=0, CI bands, normalized shock)
• Dollar vs. exchange-rate effect distinguishable from the exhibits (USD and NEER shown where relevant)
• Every exhibit note is self-contained (sample, period, frequency, FE, clustering)

Anti-patterns

• A main panel table with a dozen columns where the headline coefficient is buried mid-table
• An event-study or impulse-response figure with no confidence bands or an unmarked event date
• A single pooled coefficient when the country dispersion is the actual story
• Reporting t-stats or stars but omitting standard errors or the clustering level
• Failing to state the FE structure, so a reader cannot tell within-country from cross-country variation
• Exhibit notes that send the reader back to the text to learn the country set or the period