jfe-internet-appendix

Internet Appendix (jfe-internet-appendix)

When to trigger

• Robustness tables, proofs, and extra derivations are crowding the main text
• The main text exceeds a readable length and you must triage exhibits
• A referee asks for many additional tests you cannot fit in the body
• You need every main-text claim that says "untabulated" or "in the appendix" to actually land somewhere

Why the Internet Appendix matters at JFE

JFE hosts an Internet Appendix where exhaustive robustness, proofs, additional tests, and detailed methodology live. Given JFE's robustness-heavy culture, a deep, well-organized appendix is expected — it is where the "every alternative ruled out" work is documented without making the main paper unreadable.

JFE specifics (verify on jfinec.com): the journal's submission instructions ask that any online/internet appendix be attached to the end of the main manuscript file at submission, rather than uploaded as a wholly separate document. Number it as its own IA series regardless. Keep the Internet Appendix conceptually distinct from the replication package (code + non-proprietary data deposited to Mendeley Data at acceptance, mandatory for papers submitted on/after 1 July 2021) — the appendix holds results and proofs, the deposit holds code and data.

What belongs in the Internet Appendix

• Proofs and derivations for any theoretical results stated in the main text.
• Secondary robustness — alternative measures, additional subsamples, extra cluster levels, wild bootstrap — beyond the headline checks kept in the body.
• Variable construction details too granular for the main text (filters, matching algorithms, data cleaning).
• Additional results referenced as "untabulated" or "see Internet Appendix."
• Supplementary figures and diagnostic plots.
• Replication notes that complement (not replace) the required code/data deposit.

What stays in the main text

• The headline result and the 3–6 most decisive robustness checks.
• The identification evidence a reader needs to believe the result (parallel trends, first-stage strength).
• The core variable-definition table.

Organization rules

• Number appendix exhibits in their own series (e.g., Table IA.1, Figure IA.1).
• Every main-text pointer ("untabulated," "Internet Appendix Table IA.3") must resolve to a real exhibit.
• The appendix should be self-contained: a reader who never opens the main paper can still read each appendix table from its title and notes.
• Mirror the main text's reporting conventions (SEs, stars, cluster level) so cross-reading is frictionless.

Checklist

• Every "untabulated"/"see appendix" claim resolves to a real exhibit
• Proofs and granular derivations are in the appendix, not the body
• Headline robustness kept in main text; the long tail moved out
• Appendix exhibits numbered in their own IA series
• Appendix appended to the end of the main manuscript file (JFE practice)
• Appendix exhibits are self-contained (titles + notes)
• Reporting conventions match the main text
• Mendeley Data code/data deposit prepared separately (mandatory for post-2021 submissions)

Anti-patterns

• Main-text pointers to appendix tables that do not exist
• Treating the appendix as a dumping ground with no ordering or notes
• Moving the identification evidence to the appendix to shorten the body
• An appendix whose tables cannot be read without the main text
• Confusing the Internet Appendix with the mandatory code/data replication package

Output format

【Main-text kept】headline result + decisive checks
【Moved to IA】proofs / robustness / construction / extra results
【Pointer audit】all "untabulated"/"see IA" claims resolve? yes/no
【IA numbering】IA series applied? yes/no
【IA placement】appended to main file (JFE practice)? yes/no
【Code/data deposit】Mendeley Data package prepared separately? yes/no
【Next】jfe-writing-style