Jue Tables Figures

Spatial Exhibits: Maps, Tables & Figures (jue-tables-figures)

When to trigger

• The paper has a spatial result but no map — the reader cannot see where the variation is
• A boundary/RD or event-study design is in tables only; the figure that would convince is missing
• Tables are dense with controls and the headline coefficient is hard to find
• The geography of treatment vs control is described in words, not shown
• Reviewers cannot tell whether the spatial pattern is real or an artifact of the units chosen

Maps earn their place — but only when they carry information

JUE is one of the few economics journals where a map is often the central exhibit, because the identifying variation is geographic. But a map must do econometric work, not decorate:

• Show the identifying variation, not just the data. Map the treatment boundary, the corridor, the discontinuity, the shift-share exposure — the thing your design exploits — so the reader sees what generates the estimate.
• Make scale and units explicit. State the spatial unit (tract, block, grid cell), include a scale bar and north arrow, and pick a classification (quantile vs equal-interval) that does not manufacture a pattern.
• Pair the map with the design figure. A boundary map plus an RD plot; a treatment map plus an event-study; an exposure map plus the first stage.

Exhibit-by-design checklist

Table craft for JUE

1. Headline coefficient first, controls collapsed to "Yes/No" rows — the reader should find the estimate in two seconds.
2. Report spatial standard errors (Conley / spatial cluster) alongside or instead of naive; note the distance cutoff in the table notes.
3. Self-contained notes: sample, geography, FE level, clustering, units. A JUE table should be readable without the text.
4. House style: follow the journal's significance-reporting convention (检索于 2026-06；以官网为准); always show standard errors, and prefer reporting the magnitude in interpretable units (% capitalization, elasticity).
5. Figures legible in grayscale for print; do not encode the only signal in color.

Execution bridge (StatsPAI / Stata MCP)

Generate exhibits from the fitted result, not by retyping numbers. Full map: execution-with-mcp. JUE is urban/spatial economics — sorting and spatial dependence; identification + Conley/spatial-robust 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

• A map shows the identifying variation (boundary / corridor / exposure), not just raw data
• Map has spatial unit stated, scale bar, north arrow, and a non-misleading classification
• The design's convincing figure (RD plot / event-study / first-stage) is present
• Headline coefficient is immediately findable; controls collapsed to indicator rows
• Spatial SEs reported with the distance cutoff noted
• Table notes are self-contained (sample, geography, FE, clustering, units)
• Figures remain legible in grayscale

Anti-patterns

• A spatial paper with no map, or a map that shows the data but not the identifying variation
• A map whose classification (bins/colors) manufactures a visual pattern the estimate does not support
• Burying the headline coefficient under fifteen control rows
• Naive standard errors in the table for geographically clustered data
• An event-study or RD reported only as a table when the figure is the convincing object
• Color-only encoding that disappears in print or for color-blind readers