Tables & Figures (finman-tables-figures)

When to trigger

• The main result is settled and must be made readable at a glance
• Tables are dense, over-decimaled, or bury the headline coefficient among a dozen columns
• An event-study / first-stage / portfolio-sort plot needs to carry the design visually
• You are preparing exhibits for the FMA→ScholarOne submission and want them house-style clean

The FM exhibit bar

FM's brand is "articles that people actually read," and that taste extends to the tables. The main causal or economic estimate should be findable in seconds, and — because FM weighs practical relevance — every headline exhibit should make the economic magnitude legible, not just the t-statistic. Finance house style permits significance stars but requires standard errors in parentheses; report the dependent-variable mean so a reader can scale the coefficient into a managerial quantity. The journal's "less weight on trivial robustness" stance also shapes the exhibit set: do not pad the main paper with specification-sweep tables — promote the one headline table and the one identification figure, and demote the rest.

Exhibit craft

1. One table for the headline. A referee should read the main estimate, its SE, N, and dependent-variable mean without flipping pages.
2. Standard errors always; stars optional. SEs in parentheses are load-bearing; state the clustering level in the note; if stars are used, define them.
3. Make the magnitude speak. Add a one-line economic-significance statement or column — a one-SD change in X moves Y by Z percentage points / dollars — so the practical relevance FM prizes is visible.
4. Figures carry the design. Event-study leads and first-stage relationships persuade better as clean vector figures with CIs than as prose.
5. Right precision and a lean set. Two to three significant figures; keep the main paper to the exhibits that earn their place and push sweeps to the internet appendix.

Execution bridge (StatsPAI / Stata MCP)

Generate exhibits from the fitted result, not by retyping numbers. Full map: execution-with-mcp. Financial Management is empirical corporate finance + asset pricing; corporate-causal chain (DiD/IV/RDD) plus the factor-zoo haircut for cross-sectional pricing.

• 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

• Main estimate readable in one table: coefficient, SE in parentheses, N, dep-var mean, controls flagged
• Standard errors reported everywhere; clustering level in notes; stars (if used) defined
• An economic-magnitude statement accompanies the headline result
• Identification figure present (event-study with CIs / first-stage / sorts) where relevant
• Notes make each exhibit self-contained (sample, units, clustering, controls, period)
• Precision sensible (2–3 sig figs); specification sweeps demoted to the internet appendix

Making the magnitude speak (FM's signature exhibit move)

Because FM weights practical relevance, the single highest-return exhibit edit is making the economic size legible. Three reliable ways:

1. A dedicated economic-significance column or panel — alongside the coefficient, report the effect of a one-SD change in the regressor as a percentage of the dependent-variable mean.
2. A managerial-unit translation in the note — convert the coefficient into basis points, dollars per firm, or percentage of payout, so a CFO or investor reads the size directly.
3. A back-of-envelope aggregate — where credible, scale the per-firm effect to a portfolio or market total, with the assumptions stated. A table that reports only t-statistics asks the referee to do this work; FM referees often will not, and "is this big enough to matter?" becomes the rejection.

Anti-patterns

• A main table with 12 columns where the headline coefficient is buried
• Reporting t-stats or stars but omitting standard errors / the clustering level
• Statistical significance with no economic-magnitude translation — fails FM's relevance taste
• Over-precision (coefficients to 5 decimals) implying false accuracy
• An event-study plot with no confidence intervals or an unclear reference period
• Cramming robustness sweeps into the main paper instead of the internet appendix
• A figure that re-states a table without adding visual intuition (chartjunk that earns nothing)

Worked vignette (illustrative)

A draft's Table 4 sweeps every control combination across 12 columns, the headline is in column 9 with only t-stats, and nothing tells a reader the size of the effect. The FM fix: promote the preferred specification to a two-panel Table 3 (Panel A baseline, Panel B full controls), each with the coefficient 3.1 (s.e. 0.9 in parentheses), N, and the dependent-variable mean 0.44; add a one-line note that a one-SD rise in the regressor implies a ~7% increase relative to the mean (illustrative); move the sweep to the internet appendix; and add Figure 1, the event-study with CIs and a marked reference period. The result is now findable, identified, and economically legible.

Referee pushback mapped to the exhibit fix

• "I cannot find your main estimate." → One headline table with the coefficient, SE, N, and dep-var mean; everything else demoted.
• "Where are the standard errors / the clustering?" → SEs in parentheses everywhere; clustering in the self-contained note.
• "Is the effect economically meaningful?" → Add the economic-significance column or a managerial-unit translation.
• "This figure adds nothing." → Make the figure carry the identification (event-study leads, first stage), not re-plot a table.

Wiley production notes (verify before final upload)

A few presentation details are set by the publisher rather than the field, and they save a copyediting round if handled early:

• Figure resolution and format must meet Wiley's print specs; export figures as vector (PDF/EPS) where possible. (Exact specs 待核实 on the author guidelines.)
• Table layout should survive single-column print without truncation; very wide tables get reformatted or pushed to landscape.
• Colors that distinguish series in a figure should also be distinguishable in grayscale, since print and many readers are monochrome.
• Self-contained notes are not just good craft here; they are what the copyeditor relies on to typeset the exhibit correctly.

Output format

【Headline exhibit】one table carrying the main estimate? [Y/N]
【Inference shown】SEs in parentheses + clustering in notes? [Y/N]; stars defined if used
【Economic magnitude】effect scaled to a managerial/market unit? [Y/N]
【Identification figure】event-study / first-stage / sorts present with CIs? [Y/N]
【Self-contained notes】sample/units/clustering/controls/period in every note? [Y/N]
【Lean set】sweeps demoted to internet appendix? [Y/N]
【Next skill】finman-internet-appendix