Tables & Figures (rfs-tables-figures)

When to trigger

• The main table has too many columns or unlabeled specifications
• Standard errors, significance stars, or clustering are unclear or inconsistent
• A figure is a default software plot rather than a publication-grade exhibit
• A reader cannot tell from a table alone what the regression is

What an RFS exhibit must do

Every main table or figure should be self-contained: a referee skimming exhibits should grasp the design, the result, and the inference without the text. RFS articles are typeset by Oxford University Press for the SFS, so build exhibits to OUP production specs (vector figures, clean two-column-friendly tables). One RFS-specific constraint shapes every exhibit: because RFS requires public release of all code as a condition of publication, each table and figure must be regenerable by the released scripts — a number a referee cannot reproduce from your code is a liability. The portfolio-sort and spanning exhibits in Hou, Xue, and Zhang (2015) "Digesting Anomalies" (RFS 28(3)) are a good template for decile monotonicity and long–short t-stats.