A Claude Code skill that evaluates research ideas against standards distilled from 50+ papers published in top finance and economics journals.

Scores research ideas across 7 weighted dimensions before you commit months of work:

Outputs a structured report with scores, referee red flags, and actionable improvements.

• JFQA — 20 papers (political economy, corporate governance, innovation)
• JCF — 9 papers (executive compensation, labor investment, geopolitical risk)
• ESG/Sustainability venues — 21 papers (green finance, sustainable investing, climate policy)
• Review of Finance — bibliography coverage

• Edmans (2023) — "Learnings From 1,000 Rejections" — 998 rejection patterns from Review of Finance editor
• Mitton (2022, RFS) — "Methodological Variation in Empirical Corporate Finance" — p-hacking quantification
• Armstrong et al. (2022) — "Causality Redux" — causal inference evolution in accounting/finance
• Cochrane (2005) — "Writing Tips for PhD Students" — paper organization principles
• Jackwerth (2022) — "Tips for the PhD in Empirical Financial Economics"

6 deadly contribution sins from Edmans' 998 rejections:

• Convex combination — result predictable by chaining known findings
• "Just another" determinant — adding to a well-populated list
• Research by matrix — filling empty cells that are empty for a reason
• Context extension — same result in new country without institutional insight
• One-sided tradeoff — documenting only benefits without costs
• Curiosum — surprising but economically trivial

Recurring structures from top-journal papers:

1. Black Box Opening — identifying mechanisms in established relationships
2. Resolution Paper — introducing a dimension that resolves contradictions
3. Policy Shock — exploiting government policy for DiD identification
4. Unintended Consequences — documenting surprising spillover effects
5. Proxy Innovation — creating new ways to measure the unmeasurable
6. Cross-Country Variation — leveraging institutional differences
7. Mechanism Decomposition — breaking main effects into sub-channels

From Mitton (RFS 2022): with 10 routine methodological choices, 70%+ of random variables appear significant at 5%. The skill flags vulnerable dimensions.

Copy to your Claude Code skills directory:

cp -r . ~/.claude/skills/research-idea-benchmark/

• "evaluate my research idea"
• "benchmark my research"
• "score my paper idea"
• "assess research quality"
• "check if my idea is publishable"
• "review my research design"

research-idea-benchmark/
├── SKILL.md                              # Main skill (workflow + triggers)
└── references/
    ├── evaluation-rubric.md              # 7-dimension rubric with scoring anchors
    ├── paper-patterns.md                 # Research design patterns + templates
    └── editor-insights.md               # Editorial decision-making criteria

Hard rule: Identification Strategy must score ≥ 4 for any top-5 finance journal.

• top-journal-writing-mentor — Writing style mentor for the same journals (downstream: after idea is validated)
• research-ideation — Research startup workflow (upstream: generates candidate ideas)
• paper-self-review — Post-execution quality checklist (downstream: checks finished papers)

All rights reserved.