Writing Style (jfi-writing-style)
When to trigger
- Polishing a JFI draft before submission or revision
- Aligning references, sections, and abstract with JFI/Elsevier conventions
JFI house conventions (verified 2026-06-20; re-confirm on the official page)
- Numbered sections (1, 1.1, 1.1.1) structure the manuscript.
- References: "your-paper-your-way" at submission — any internally consistent style is accepted;
the journal applies an author–date (name–year) Elsevier style at proof. In-text citations give
surname(s) and year, with "et al." for three or more authors; the list is alphabetical then
chronological, with a/b/c disambiguation for same-author/same-year items. Keep author–year fields clean.
- Abstract: concise, factual, and able to stand alone, with references avoided; keep it at 250 words
or fewer.
- Keywords: provide 1-7 English keywords for indexing.
- Optional Highlights: 3–5 bullets, max 85 characters each, strongly encouraged.
- Generative-AI disclosure: if AI writing tools were used, declare them in a dedicated section
before the References (see jfi-submission).
Make the mechanism land
JFI is read by intermediation specialists who reward a clearly motivated economic mechanism:
- Explain the intermediary friction and channel before the econometric specification or the formal
model — the reader should know why before how.
- Tie every result back to the intermediation contribution (see jfi-contribution-framing); avoid
coefficient-by-coefficient narration with no economic interpretation.
- Define institutional terms on first use; precision about banks, loans, and regulation signals fluency.
Mechanism-first rewrite, worked (illustrative)
- Before: "We regress loan growth on the capital shortfall interacted with bank size, controlling for
firm fixed effects. The coefficient is −2.1 and significant at the 1% level."
- After: "Banks that must rebuild capital face a choice between costly equity and shrinking assets.
If lending relationships embody information rivals cannot replicate, constrained banks can cut
relationship credit without losing the borrower — so recapitalization's burden falls on locked-in
firms. Consistent with this, a 1pp shortfall reduces credit to the same firm by 2.1pp, and twice that
for single-bank borrowers."
- The rewrite names friction → channel → prediction before the number, which is how this journal's
specialist readership parses a result; the regression mechanics move to the design section.
An intermediation lexicon check
Loose vocabulary signals a non-specialist to a single-blind referee who knows your name:
- liquidity creation (the Diamond–Dybvig transformation function) vs. liquidity provision
(supplying funds) — not interchangeable;
- capital requirement (the rule) vs. capital ratio (the outcome) vs. capital shock (the
change you exploit);
- say credit supply only after demand is absorbed — until then it is "lending growth";
- relationship lending implies repeated interaction and private information, not merely an existing
loan;
- a run is strategic withdrawal, not any deposit outflow;
- shadow bank denotes intermediation outside the regulated perimeter, not any nonbank financial firm.
Style pushbacks at this venue, and fixes
- "The introduction is a results list" → restructure as friction, mechanism, design, finding, lesson.
- "Notation drifts between model and empirics" → one symbol table; the model's friction parameter should
be the object the regression coefficient estimates or proxies.
- "Reads like a policy report" → replace recommendation language with the equilibrium or identified
margin the analysis actually delivers.
- "The abstract could describe ten papers" → one abstract sentence each for friction, design or model,
headline magnitude or proposition, and the intermediation lesson; cut everything else.
Anti-patterns
- A numbered/footnote citation style left in at submission with no internal consistency
- An abstract stuffed with references or with no stated finding
- Specifications introduced before the mechanism is explained
- Forgetting the AI-disclosure section when such tools were used
Output format
【Sections】numbered (1, 1.1)? [Y/N]
【References】internally consistent, author–year-ready? [Y/N]
【Abstract】concise, stand-alone, no refs, <=250 words? [Y/N]
【Mechanism-first】explained before specs? [Y/N]
【AI disclosure】present-if-used? [Y/N]
【Next skill】jfi-replication-and-data-policy
1---2name: jfi-writing-style3description: Use when revising a Journal of Financial Intermediation (JFI) manuscript for the journal's house conventions — numbered sections, author–date Elsevier referencing, a concise stand-alone abstract, specialist intermediation vocabulary, and prose that explains the mechanism before the specifications. It edits for style; it does not invent content.4---56# Writing Style (jfi-writing-style)78## When to trigger910- Polishing a JFI draft before submission or revision11- Aligning references, sections, and abstract with JFI/Elsevier conventions1213## JFI house conventions (verified 2026-06-20; re-confirm on the official page)1415- **Numbered sections** (1, 1.1, 1.1.1) structure the manuscript.16- **References:** "your-paper-your-way" at submission — any **internally consistent** style is accepted;17 the journal applies an **author–date (name–year) Elsevier** style at proof. In-text citations give18 surname(s) and year, with "et al." for three or more authors; the list is alphabetical then19 chronological, with a/b/c disambiguation for same-author/same-year items. Keep author–year fields clean.20- **Abstract:** concise, factual, and able to stand alone, with references avoided; keep it at **250 words21 or fewer**.22- **Keywords:** provide **1-7 English keywords** for indexing.23- **Optional Highlights:** 3–5 bullets, max 85 characters each, strongly encouraged.24- **Generative-AI disclosure:** if AI writing tools were used, declare them in a dedicated section25 **before the References** (see jfi-submission).2627## Make the mechanism land2829JFI is read by intermediation specialists who reward a clearly motivated **economic mechanism**:3031- Explain the intermediary friction and channel **before** the econometric specification or the formal32 model — the reader should know *why* before *how*.33- Tie every result back to the intermediation contribution (see jfi-contribution-framing); avoid34 coefficient-by-coefficient narration with no economic interpretation.35- Define institutional terms on first use; precision about banks, loans, and regulation signals fluency.3637## Mechanism-first rewrite, worked (illustrative)3839- **Before:** "We regress loan growth on the capital shortfall interacted with bank size, controlling for40 firm fixed effects. The coefficient is −2.1 and significant at the 1% level."41- **After:** "Banks that must rebuild capital face a choice between costly equity and shrinking assets.42 If lending relationships embody information rivals cannot replicate, constrained banks can cut43 relationship credit without losing the borrower — so recapitalization's burden falls on locked-in44 firms. Consistent with this, a 1pp shortfall reduces credit to the same firm by 2.1pp, and twice that45 for single-bank borrowers."46- The rewrite names friction → channel → prediction **before** the number, which is how this journal's47 specialist readership parses a result; the regression mechanics move to the design section.4849## An intermediation lexicon check5051Loose vocabulary signals a non-specialist to a single-blind referee who knows your name:5253- **liquidity creation** (the Diamond–Dybvig transformation function) vs. **liquidity provision**54 (supplying funds) — not interchangeable;55- **capital requirement** (the rule) vs. **capital ratio** (the outcome) vs. **capital shock** (the56 change you exploit);57- say **credit supply** only after demand is absorbed — until then it is "lending growth";58- **relationship lending** implies repeated interaction and private information, not merely an existing59 loan;60- a **run** is strategic withdrawal, not any deposit outflow;61- **shadow bank** denotes intermediation outside the regulated perimeter, not any nonbank financial firm.6263## Style pushbacks at this venue, and fixes6465- "The introduction is a results list" → restructure as friction, mechanism, design, finding, lesson.66- "Notation drifts between model and empirics" → one symbol table; the model's friction parameter should67 be the object the regression coefficient estimates or proxies.68- "Reads like a policy report" → replace recommendation language with the equilibrium or identified69 margin the analysis actually delivers.70- "The abstract could describe ten papers" → one abstract sentence each for friction, design or model,71 headline magnitude or proposition, and the intermediation lesson; cut everything else.7273## Anti-patterns7475- A numbered/footnote citation style left in at submission with no internal consistency76- An abstract stuffed with references or with no stated finding77- Specifications introduced before the mechanism is explained78- Forgetting the AI-disclosure section when such tools were used7980## Output format8182```83【Sections】numbered (1, 1.1)? [Y/N]84【References】internally consistent, author–year-ready? [Y/N]85【Abstract】concise, stand-alone, no refs, <=250 words? [Y/N]86【Mechanism-first】explained before specs? [Y/N]87【AI disclosure】present-if-used? [Y/N]88【Next skill】jfi-replication-and-data-policy89```