Writing Style (wber-writing-style)

When to trigger

• The abstract states a topic, not a result, and never names the policy stake
• The introduction takes three pages to reach the question
• The paper reads either as a dry econometrics exercise or as an advocacy brief — not the WBER middle
• The bridge from "we estimate X" to "so development policy should consider Y" is missing
• A coauthor says the writing does not sound like a WBER paper

The WBER voice

WBER writing serves a dual audience: economists who demand precision and policymakers/practitioners who demand relevance. The house voice is rigorous but legible — technically exact, jargon-controlled, and always closing the loop to a development-policy implication. Two failure modes bracket it: prose so technical that a finance-ministry reader gives up, and prose so policy-breathless that an economist distrusts it. Aim for the middle: state the credible result plainly, then say what it means for a development decision without overclaiming and without implying alignment with any World Bank position (editorial independence is explicit). Write the body first; write the abstract and introduction last, once the result is final.

The introduction recipe (WBER variant)

A WBER introduction should, in order:

1. The development problem and its policy stakes — one paragraph: why this matters for poverty/growth/welfare in developing countries.
2. The question, sharply — the specific causal or measurement question.
3. The setting and data — country/region, the data spine (LSMS/DHS/admin/survey), the source of variation.
4. The headline result in policy units — the magnitude a practitioner would quote (a pp change, a dollar figure, a cost-effectiveness number), with its precision.
5. Identification in one paragraph — the design and the one assumption it rests on, stated honestly.
6. The contribution and the boundary — what is new vs. the frontier, and why this is a WBER (not JDE/World Development/Research Observer) paper.
7. The policy takeaway — what a development decision-maker should do differently, scoped to what the evidence supports.

The abstract

• ~100–150 words (confirm the current limit on the OUP page; 待核实).
• Sentence 1: the development question and why it matters.
• Sentence 2–3: the setting, data, and identification in plain terms.
• Sentence 4: the headline result in policy units, with magnitude.
• Sentence 5: the policy implication, scoped honestly.
• No undefined acronyms; no "we study"; lead with what you find.

Prose discipline

• Magnitudes over significance. "Enrollment rose 4 percentage points" beats "the effect is significant at 1%."
• Define development jargon once. Spell out program names, local institutions, and acronyms (PMT, CCT, LSMS) on first use — practitioners and economists from other fields read WBER.
• Active, claim-first sentences. Topic-sentence each paragraph with its point.
• Reference style is Chicago-style author-date as applied by OUP (verify current style on the author guidelines; 待核实). Keep citations engaging the best version of prior work, not a strawman.
• Honest hedging. Scope every policy claim to the estimand and external-validity argument; do not let the conclusion outrun the design.

Checklist

• Abstract states a result in policy units, not a topic, and names the policy stake
• Introduction reaches the question within the first page and the headline result soon after
• Identification stated honestly in one paragraph in the intro
• The "why WBER, not a sibling" boundary is visible
• Effects reported as magnitudes (with precision), not asterisk significance
• Development jargon/acronyms defined on first use
• Policy takeaway scoped to what the design supports; no implied Bank-policy alignment
• Abstract and intro written last, after results are final

Anti-patterns

• An abstract that describes the topic and method but never states the finding or its magnitude
• A meandering introduction that reaches the question on page three
• Pure-econometrics prose with no policy bridge (reads like the wrong journal)
• Advocacy prose that overclaims beyond the estimand
• Implying the paper endorses or follows a World Bank policy position
• Unexplained local acronyms and program names

Worked vignette (illustrative)

An abstract opens: "This paper studies the effects of a conditional cash transfer using a regression discontinuity design in [country]." It states method, not result. The WBER rewrite: "Means-tested cash transfers are a central anti-poverty tool, but it is unclear whether eligibility cutoffs improve schooling at the margin. Using a sharp eligibility threshold in [country]'s national program and administrative enrollment records, we find the transfer raises secondary enrollment by 5 percentage points (s.e. 1.4) at a cost of about $90 per additional enrollee. The effect concentrates among girls, suggesting that relaxing the cutoff modestly would be a cost-effective way to close the gender gap." It now leads with stakes, gives a magnitude in policy units, and scopes the implication.

Calibrating the technical register

The hardest WBER writing skill is pitching the technical level for a dual audience without dumbing down or losing rigor:

• Method in the body, intuition in the intro. State the estimator precisely where it is used; in the introduction, give the one-sentence intuition ("we compare eligible and just-ineligible households at the cutoff") a non-specialist can follow.
• Carry the magnitude through. Whatever the technical detail, the reader should never lose the policy number — repeat it (with its precision) in the abstract, intro, results, and conclusion.
• Quarantine the jargon. Push estimator derivations, identification proofs, and notation to a methods subsection or appendix; keep the narrative spine readable.
• Conclude with the decision, not a recap. End on what a development decision-maker should weigh, scoped to the evidence — not a paragraph that merely restates the findings.

Output format

【Abstract】states result in policy units + policy stake? [Y/N]
【Intro】question by page 1; headline result + honest identification soon after? [Y/N]
【Audience bridge】estimate → development-policy lesson present? [Y/N]
【Magnitudes】effects in policy units, not asterisks? [Y/N]
【Boundary】"why WBER not sibling" visible? [Y/N]
【Scope】policy claim within the estimand; no Bank-alignment implication? [Y/N]
【Next step】wber-replication-package