Tables & Figures (jbes-tables-figures)
When to trigger
- Simulation tables are dense and a reader cannot read off whether the method controls size
- Size, power, coverage, bias, and RMSE are scattered instead of side-by-side
- Figures lack confidence bands, or compare against no baseline
- Exhibit notes do not state the DGP, sample size, replications, or nominal level
What JBES exhibits must communicate
A JBES paper is judged by method experts who read the tables as the evidence. Exhibits carry two distinct jobs: the Monte Carlo exhibits that establish the method's statistical properties, and the empirical exhibits that establish relevance on real data. Both must be readable by a statistician and an applied economist, since the journal bridges the two communities.
Monte Carlo exhibits
- Put size and power side-by-side across DGPs and sample sizes; mark the nominal level so over/under-rejection is obvious.
- For interval methods, report coverage and average length together; for point estimators, bias and RMSE together.
- Always show the incumbent baseline in the same table under identical DGPs.
- Figures: size-power curves, coverage-vs-n plots, empirical-vs-nominal QQ plots, sampling distributions with the asymptotic reference overlaid.
Empirical exhibits
- Report estimates with appropriate (HAC/cluster/dependence-robust) standard errors.
- Show the substantive payoff: what the new method changes relative to the standard approach.
Self-contained notes (every exhibit)
State the DGP or data source, sample size(s), number of Monte Carlo replications, nominal level, the estimator/test, the inference method, and units. A reader should not need the body to interpret the table.
Execution bridge (StatsPAI / Stata MCP)
Generate exhibits from the fitted result, not by retyping numbers (the usual source of
body-vs-appendix drift). Full map: execution-with-mcp.
- Tables:
etable (multi-model columns) or did_summary_to_latex straight from the
result_id — one variable definition, one set of numbers, body and appendix in sync.
- 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 (from the result's diagnostics) and
states the magnitude in interpretable units.
See a full fitted-result → exhibit chain in the JF execution walkthrough.
Checklist
Anti-patterns
- A wall of numbers where size control cannot be read off at a glance
- Reporting power without size (or coverage without length)
- Omitting the baseline, so "improvement" is unquantified in the exhibit
- Notes that force the reader back into the body to decode the table
- Figures with no uncertainty and no reference distribution
Worked vignette: a size-and-power table a referee can read
A hypothetical JBES paper reports a Monte Carlo size-power table for a new break test (numbers illustrative). The good version puts size and power side-by-side across n ∈ {120, 240, 480}, marks the 5% level in the caption, and shows the CUSUM benchmark in the same rows — so a reader sees the new test holds size at 5.2% while CUSUM over-rejects at 8.7%. The note states the DGP, replications, MC standard errors, level, and method.
Exhibit-pushback patterns (venue-specific fixes)
| JBES referee objection |
Fix this skill enforces |
| "I cannot read size control off this table." |
Put size and power side-by-side; mark the nominal level in the note |
| "No baseline, so 'improvement' is unquantified." |
Place the incumbent in the same exhibit under identical DGPs |
| "The note does not let me interpret the table alone." |
State DGP/data, n, replications, level, method, and units in every note |
Calibration anchor (hedged): JBES exhibits serve two audiences — a statistician reading Monte Carlo
properties and an applied economist reading the payoff — so every table must be legible to both. Exact
format specifics are live T&F preflight items.
Exhibit pass for Journal of Business & Economic Statistics
Run this as a concrete capability pass. First lock the statistical estimand, identification/simulation evidence, empirical illustration, and reproducibility path; then test whether the manuscript addresses econometrics/statistics reviewers who expect methodological credibility plus a business or economic use case.
- Primary move: For every table or figure, state the object, sample/case base, uncertainty display, and one sentence the exhibit proves for this venue.
- Decision ledger: return
claim / evidence / blocker / next edit rows so the next pass can patch the manuscript directly.
- Sibling comparison: compare against Journal of Econometrics for theory-heavy methods, Econometric Theory for proof-first work, Quantitative Economics for economics-theory methods; if the neighboring outlet has the stronger audience claim, recommend re-routing before polishing.
- Verification floor: before submission-ready advice, re-open
resources/official-source-map.md for volatile rules and name the one unresolved fact that could change the recommendation.
Output format
【Exhibit】Monte Carlo / empirical
【Size+power】side-by-side with level marked? [Y/N]
【Coverage/length or bias/RMSE】paired? [Y/N]
【Baseline】incumbent in same exhibit? [Y/N]
【Notes】DGP/data, n, reps, level, method, units present? [Y/N]
【Next step】jbes-writing-style
Source: brycewang-stanford/Awesome-Journal-Skills → Journal-of-Business-and-Economic-Statistics-Skills/skills/jbes-tables-figures/SKILL.md
1---2name: jbes-tables-figures3description: Use when designing the simulation tables and figures for a Journal of Business & Economic Statistics (JBES) methods paper so size, power, coverage, and the empirical results are legible to both statisticians and applied economists. Improves exhibits; it does not change results.4---567# Tables & Figures (jbes-tables-figures)89## When to trigger1011- Simulation tables are dense and a reader cannot read off whether the method controls size12- Size, power, coverage, bias, and RMSE are scattered instead of side-by-side13- Figures lack confidence bands, or compare against no baseline14- Exhibit notes do not state the DGP, sample size, replications, or nominal level1516## What JBES exhibits must communicate1718A JBES paper is judged by method experts who read the tables as the evidence. Exhibits carry two distinct jobs: the **Monte Carlo exhibits** that establish the method's statistical properties, and the **empirical exhibits** that establish relevance on real data. Both must be readable by a statistician *and* an applied economist, since the journal bridges the two communities.1920## Monte Carlo exhibits2122- Put **size and power side-by-side** across DGPs and sample sizes; mark the nominal level so over/under-rejection is obvious.23- For interval methods, report **coverage and average length** together; for point estimators, **bias and RMSE** together.24- Always show the **incumbent baseline** in the same table under identical DGPs.25- Figures: size-power curves, coverage-vs-n plots, empirical-vs-nominal QQ plots, sampling distributions with the asymptotic reference overlaid.2627## Empirical exhibits2829- Report estimates with appropriate (HAC/cluster/dependence-robust) standard errors.30- Show the substantive payoff: what the new method changes relative to the standard approach.3132## Self-contained notes (every exhibit)3334State the DGP or data source, sample size(s), number of Monte Carlo replications, nominal level, the estimator/test, the inference method, and units. A reader should not need the body to interpret the table.3536## Execution bridge (StatsPAI / Stata MCP)3738Generate exhibits from the fitted result, not by retyping numbers (the usual source of39body-vs-appendix drift). Full map: [`execution-with-mcp`](../../../shared-resources/empirical-methods/execution-with-mcp.md).4041- **Tables:** `etable` (multi-model columns) or `did_summary_to_latex` straight from the42 `result_id` — one variable definition, one set of numbers, body and appendix in sync.43- **Figures:** `plot_from_result` / `enhanced_event_study_plot` / `event_study_table` —44 axis units and the SE/clustering note baked in.45- **Every note** names the estimator + clustering (from the result's diagnostics) and46 states the magnitude in interpretable units.4748See a full fitted-result → exhibit chain in the [JF execution walkthrough](../../../Journal-of-Finance-Skills/resources/worked-examples/02-execution-walkthrough.md).49## Checklist5051- [ ] Size and power readable side-by-side, with nominal level marked52- [ ] Coverage shown with average length; bias shown with RMSE53- [ ] Incumbent baseline in the same exhibit under identical DGPs54- [ ] Figures have confidence bands / reference distributions55- [ ] Every note states DGP/data, n, replications, level, method, units56- [ ] Vector output (PDF/EPS); no chartjunk; legible in print57- [ ] Numbers match the manuscript text and the code output5859## Anti-patterns6061- A wall of numbers where size control cannot be read off at a glance62- Reporting power without size (or coverage without length)63- Omitting the baseline, so "improvement" is unquantified in the exhibit64- Notes that force the reader back into the body to decode the table65- Figures with no uncertainty and no reference distribution666768## Worked vignette: a size-and-power table a referee can read6970A hypothetical JBES paper reports a Monte Carlo size-power table for a new break test (numbers **illustrative**). The good version puts size and power side-by-side across n ∈ {120, 240, 480}, marks the 5% level in the caption, and shows the CUSUM benchmark in the same rows — so a reader sees the new test holds size at 5.2% while CUSUM over-rejects at 8.7%. The note states the DGP, replications, MC standard errors, level, and method.7172## Exhibit-pushback patterns (venue-specific fixes)7374| JBES referee objection | Fix this skill enforces |75|----|----|76| "I cannot read size control off this table." | Put size and power side-by-side; mark the nominal level in the note |77| "No baseline, so 'improvement' is unquantified." | Place the incumbent in the same exhibit under identical DGPs |78| "The note does not let me interpret the table alone." | State DGP/data, n, replications, level, method, and units in every note |7980Calibration anchor (hedged): JBES exhibits serve two audiences — a statistician reading Monte Carlo81properties and an applied economist reading the payoff — so every table must be legible to both. Exact82format specifics are live T&F preflight items.8384## Exhibit pass for Journal of Business & Economic Statistics8586Run this as a concrete capability pass. First lock the statistical estimand, identification/simulation evidence, empirical illustration, and reproducibility path; then test whether the manuscript addresses econometrics/statistics reviewers who expect methodological credibility plus a business or economic use case.8788- **Primary move:** For every table or figure, state the object, sample/case base, uncertainty display, and one sentence the exhibit proves for this venue.89- **Decision ledger:** return `claim / evidence / blocker / next edit` rows so the next pass can patch the manuscript directly.90- **Sibling comparison:** compare against Journal of Econometrics for theory-heavy methods, Econometric Theory for proof-first work, Quantitative Economics for economics-theory methods; if the neighboring outlet has the stronger audience claim, recommend re-routing before polishing.91- **Verification floor:** before submission-ready advice, re-open `resources/official-source-map.md` for volatile rules and name the one unresolved fact that could change the recommendation.9293## Output format9495```96【Exhibit】Monte Carlo / empirical97【Size+power】side-by-side with level marked? [Y/N]98【Coverage/length or bias/RMSE】paired? [Y/N]99【Baseline】incumbent in same exhibit? [Y/N]100【Notes】DGP/data, n, reps, level, method, units present? [Y/N]101【Next step】jbes-writing-style102```103104---105106**Source:** [`brycewang-stanford/Awesome-Journal-Skills`](https://github.com/brycewang-stanford/Awesome-Journal-Skills) → `Journal-of-Business-and-Economic-Statistics-Skills/skills/jbes-tables-figures/SKILL.md`
thedixitjain