rfs-writing-style

Writing & Contribution Framing (rfs-writing-style)

When to trigger

• The abstract or introduction buries what is new
• Prose is dense, over-hedged, or jargon-heavy
• The contribution is implied but never stated as a crisp claim
• The paper is technically done and needs a final exposition pass

Run this only after identification, robustness, and exhibits hold up. Do not polish prose around an unsettled result.

How RFS papers read

The best RFS papers make the new question and the answer legible in the first two pages. An RFS Editor desk-screens before referees, so the contribution must land before page three or risk a desk reject. RFS welcomes theory + empirics, so the prose must integrate them, not present a model and then an unrelated regression — Brunnermeier and Pedersen (2009, RFS 22(6)) is a model of a tight theory→prediction→evidence arc. Two hard RFS house-style rules bind the writing: the abstract must be ≤ 100 words (RFS returns longer abstracts), and references follow Chicago Manual of Style author-date format. A long, literature-first abstract is both off-house-style and over the word limit.

The introduction formula (RFS-style)

1. Hook: the first-order question and why it matters now (one paragraph).
2. What we do: the design and data, in one paragraph — name the source of variation.
3. What we find: the headline result with magnitude, then the key robustness in a clause.
4. Why it is new: the explicit delta vs. the closest work (from rfs-literature-positioning).
5. Mechanism / interpretation: what the result means for theory or practice.
6. Roadmap: brief; one or two sentences.

Abstract

• Lead with the question and the finding, not the literature.
• State the design and the headline magnitude.
• End with the contribution / implication.
• Keep it tight — the RFS abstract is capped at 100 words (longer abstracts are returned). Every word must earn its place.
• No undefined acronyms; no "we study the relationship between."

Prose discipline

• Active voice; one idea per sentence; define each term once.
• Report economic magnitudes in interpretable units, not only significance.
• Integrate theory and evidence: state the prediction, then the test, then the result.
• Calibrate claims: "consistent with" vs. "demonstrates" vs. "suggests" — match the evidence.
• Cut hedging stacks ("may possibly suggest that perhaps").

Checklist

• Abstract leads with question + finding + magnitude, not literature
• Introduction states the delta vs. the closest paper explicitly
• Headline result appears with economic magnitude on page 1–2
• Theory and empirics are integrated, not sequential and disconnected
• Claims are calibrated to the strength of the evidence (no overselling)
• Jargon and acronyms defined on first use; no undefined symbols
• Hedging reduced; active voice; one idea per sentence
• Abstract ≤ 100 words; references in Chicago Manual of Style author-date format

Anti-patterns

• An abstract that opens "The literature has long debated ..." and never states the finding.
• Burying the contribution in the conclusion.
• A model in Section 2 and a regression in Section 4 with no logical bridge.
• Reporting "highly significant" effects whose magnitude is economically trivial.
• Over-claiming causality the design does not support.
• A 1.5-page roadmap paragraph.

Output format

【Abstract】leads with question+finding+magnitude? yes/no
【Intro delta】explicit vs. closest paper? yes/no
【Magnitude on p.1–2】yes/no
【Theory–empirics integration】tight / loose
【Overselling risk】[claims to recalibrate]
【Next step】rfs-submission