brycewang-stanford Tables and Figures (ecopol-tables-figures)
When to trigger
- A minister or central-bank reader could not extract the policy takeaway from your main exhibit
- Tables are dense regression dumps with every coefficient and three sets of stars
- The key magnitude is buried in a coefficient cell rather than shown as a headline effect
- Figures lack the context (units, baseline, policy threshold) a non-economist needs
- You are choosing which exhibits go in the accessible main text vs. the technical appendix
EP exhibits serve two readers — design for the harder one
EP is read by economists and policymakers, and presented to two discussants. The main-text exhibits must be legible to the policy reader; the regression machinery belongs in the appendix. The governing principle: lead with figures that show the policy effect, demote dense tables to the appendix. A good EP figure answers the policy question by itself — an event-study plot of the reform's effect over time, a map of treated regions, a bar of the welfare/cost decomposition — with a caption a non-specialist can read standalone.
Note on house style: EP applies its template and copyediting at the commissioning/production stage, and authors receive style guidance on commissioning (检索于 2026-06；以官网为准). All figures should carry alternative text (alt text) for accessibility (检索于 2026-06；以官网为准). Confirm exact figure-format and table-style rules in the live author guidelines.
Main-text vs. appendix split
| Exhibit | Where it belongs |
|---|---|
| Headline policy effect (event study, treatment effect over time) | main-text figure |
| Magnitude in policy-meaningful units (€, jobs, % of GDP) | main-text figure/table |
| Welfare / fiscal-cost decomposition | main-text figure |
| Full regression table with all controls and specifications | appendix |
| Robustness grid across specifications | compact summary in main text; full grid in appendix |
| Balance / first-stage / density diagnostics | appendix |
Craft moves for the policy reader
- Translate units. Show effects as €, jobs, percentage-points of GDP, or per-capita — not just a standardized coefficient. The policymaker thinks in outcomes, not z-scores.
- Put the baseline in the figure. An effect of "−0.08" is meaningless without "relative to a baseline of 0.40." Annotate it.
- One message per figure. If a figure makes two points, split it. Caption states the single takeaway.
- Drop significance asterisks from the headline story. Show the point estimate and a visible confidence interval; let the policy reader see precision directly rather than decoding stars.
- Mark the policy object on the plot — the reform date, the eligibility threshold, the counterfactual path — so the design is visible.
- Tables: only the coefficients that matter in the main text; full controls in the appendix.
Execution bridge (StatsPAI / Stata MCP)
Generate exhibits from the fitted result, not by retyping numbers. Full map:
execution-with-mcp. Economic Policy is policy-facing applied economics; foreground a credible design and a policy-relevant magnitude.
- Tables:
etable(multi-model) ordid_summary_to_latexstraight from theresult_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
- At least one main-text figure that conveys the policy effect standalone
- Effects shown in policy-meaningful units, with the baseline annotated
- Confidence intervals visible; headline message not asterisk-dependent
- Policy object (reform date / threshold / counterfactual) marked on relevant figures
- Dense regression tables and diagnostics moved to the appendix
- Each figure has a standalone, non-technical caption (and alt text per accessibility policy — 待核实 on exact spec)
- One message per exhibit; no two-point figures
Anti-patterns
- A main-text table of 20 coefficients × 6 columns that no policymaker will read
- Reporting effects only as standardized/log coefficients with no real-world units
- Three tiers of significance stars carrying the headline instead of a shown CI
- Figures with no baseline, no units, and no marked policy event
- Leaving the single most policy-relevant exhibit in the appendix while the main text shows diagnostics
Output format
【Journal】Economic Policy (EP)
【Skill】ecopol-tables-figures
【Headline figure】does it convey the policy effect standalone? Y/N
【Units】effect expressed in € / jobs / %GDP / per-capita?
【Baseline + policy object】annotated on figure? Y/N
【Precision】CI shown (not asterisk-driven)? Y/N
【Main-text vs appendix】what moved to appendix
【Accessibility】alt text / format per guidelines (待核实 exact spec)
【Next skill】ecopol-writing-style