Econ Paper Writing

Key Principle: Be proactive. If the repo and results are clear, deliver a full draft. Don't block waiting for feedback on every section. Produce something concrete to react to, then iterate.

⚠️ CRITICAL: Never Hallucinate Citations

NEVER generate BibTeX entries from memory. ALWAYS fetch programmatically.

If you cannot verify a citation:

% EXPLICIT PLACEHOLDER - requires human verification
\cite{PLACEHOLDER_author2024_verify_this}  % TODO: Verify this citation exists

Always tell the author: "I've marked [X] citations as placeholders that need verification."

For citation APIs: Semantic Scholar, CrossRef (DOI→BibTeX), arXiv. See references/citation-workflow.md or 20-ml-paper-writing/references/citation-workflow.md.

Workflow 0: Starting from a Research Repository

Project Understanding:
- [ ] Step 1: Explore the repository structure
- [ ] Step 2: Read README, docs, key results (tables, figures)
- [ ] Step 3: Identify the main contribution with the author
- [ ] Step 4: Find papers already cited in code or notes
- [ ] Step 5: Search for additional relevant literature
- [ ] Step 6: Outline the paper structure
- [ ] Step 7: Draft sections iteratively with feedback

Step 1: Explore the Repository

ls -la
find . -name "*.do" -o -name "*.m" -o -name "*.py" | head -30
find . -name "*.bib" -o -name "*.tex"
find . -path "*/results/*" -o -path "*/output/*" -o -path "*/tables/*"

Look for: README, data/, code/, results/, tables/, .bib, .tex drafts, Stata/MATLAB/Python scripts.

Step 2: Identify Existing Citations

grep -r "doi\|cite\|@article" --include="*.bib" --include="*.tex" --include="*.do" .

Step 3: Clarify the Contribution

"Based on my understanding of the repo, the main contribution appears to be [X]. The key results show [Y]. Is this the framing you want, or should we emphasize different aspects?"

Step 4–5: Search literature; verify citations; deliver first draft. Draft first, ask with the draft (not before).

Balancing Proactivity and Collaboration

Draft first, ask with the draft:

Only block when: Target journal unclear; contradictory framings; results incomplete; explicit request to review before continuing.

The Narrative Principle (Economics Adaptation)

Your paper is not a collection of regressions—it's a story with one clear contribution supported by evidence.

If you cannot state your contribution in one sentence, you don't yet have a paper.

Paper Structure Workflow

Workflow 1: Writing a Complete Paper (Iterative)

Each step: draft → feedback → revision.

Paper Writing Progress:
- [ ] Step 1: Define the one-sentence contribution (with author)
- [ ] Step 2: Draft main figure (event-study or key result) → feedback → revise
- [ ] Step 3: Draft abstract → feedback → revise
- [ ] Step 4: Draft introduction (no "Introduction" heading in AER) → feedback → revise
- [ ] Step 5: Draft institutional background → feedback → revise
- [ ] Step 6: Draft data section → feedback → revise
- [ ] Step 7: Draft empirical strategy → feedback → revise
- [ ] Step 8: Draft results → feedback → revise
- [ ] Step 9: Draft mechanisms / heterogeneity → feedback → revise
- [ ] Step 10: Draft robustness → feedback → revise
- [ ] Step 11: Draft conclusion → feedback → revise
- [ ] Step 12: Simulate reviewer; final review cycle

Step 1: Define the One-Sentence Contribution (requires author confirmation)

"I propose framing the contribution as: '[one sentence]'. Does this capture the main takeaway?"

Step 3: Abstract (Economics: ~100–150 words)

1. Core question (1 sentence)
2. Main findings (2–3 sentences)
3. Mechanism (1 sentence)
4. Bottom line (1 sentence)

Remove from abstracts: methodological details, secondary findings, robustness. Delete weak openings like "This paper studies..."

Step 4: Introduction

• Contribution, setting, preview
• Identification framing (what variation, what assumption)
• Claims must map to identification assumptions
• See references/identification-language.md

Economics paper structure: See references/structure-map.md.

AER Template Workflow

Workflow 2: Starting a New Paper from AER Template

Template location: templates/aer/ (copied from Journal_Template_for_The_American_Economic_Review__AER_/)

- [ ] Step 1: Copy entire AER template directory to new project
- [ ] Step 2: Verify template compiles as-is (pdflatex + bibtex × 2)
- [ ] Step 3: Replace title, authors, abstract, sections
- [ ] Step 4: Add your .bib file; update \bibliography{}
- [ ] Step 5: Add figures/tables; use figurenotes/tablenotes

AER-specific:

• No "Introduction" heading; begin introductory material before first \section{}
• Captions below figures, above tables
• Use \begin{figurenotes} and \begin{tablenotes} for notes
• Appendix after bibliography; use \appendix
• Use aea bibliography style with aea.bst

Compilation:

pdflatex AER-Article.tex
bibtex AER-Article
pdflatex AER-Article.tex
pdflatex AER-Article.tex

Workflow 3: Creating Lecture / Seminar Slides from a Paper

Goal: Turn a working paper or published paper into a 45-min seminar talk (15–20 frames) or a course lecture.

Frame Budget

Key Principles

1. Results-forward. Show the main result by frame 7–8. Don't make the audience wait.
2. One idea per frame. If you need more space, split into two frames.
3. No overlay animations. See no-pause-beamer rule. Every frame shows complete content.
4. TikZ for intuition. A timeline diagram of your DiD beats a wall of assumptions.
5. Backup slides are free. Robustness checks, data details, proofs → backup after \appendix.
6. Figures adapted for slides → larger fonts, bolder lines, simpler legends than paper figures.

Beamer Template

\documentclass[aspectratio=169]{beamer}
\usetheme{default}
\usecolortheme{beaver}  % or Pittsburgh, Madrid, etc.
\setbeamertemplate{navigation symbols}{}  % remove nav bar

\title{Your Title}
\subtitle{Optional subtitle}
\author{Author Name}
\institute{University / Conference}
\date{\today}

\begin{document}

\frame{\titlepage}

\begin{frame}{Motivation}
  \begin{itemize}
    \item Key fact / puzzle that motivates the paper
    \item Why it matters for policy or theory
    \item What we do (one sentence)
  \end{itemize}
\end{frame}

\begin{frame}{Related Literature}
  \begin{itemize}
    \item \textbf{Strand 1:} Author (year) — what they do; \textbf{we add} X
    \item \textbf{Strand 2:} Author (year) — what they do; \textbf{we add} Y
    \item \textbf{Strand 3:} Author (year) — what they do; \textbf{we add} Z
  \end{itemize}
\end{frame}

% ... identification, results, mechanisms frames ...

\begin{frame}{Conclusion}
  \begin{itemize}
    \item \textbf{Main finding:} [one sentence]
    \item \textbf{Mechanism:} [one sentence]
    \item \textbf{Policy implication:} [one sentence]
  \end{itemize}
\end{frame}

\appendix

\begin{frame}{Backup: Robustness Checks}
  % Tables / figures showing robustness
\end{frame}

\end{document}

Slides from Paper Checklist

- [ ] Reduce paper abstract to 3 bullets on motivation frame
- [ ] Pick ONE main result figure or table (not all tables)
- [ ] Create TikZ diagram for identification intuition
- [ ] Add "Contribution to Literature" frame (3-5 papers)
- [ ] Add "Data" frame with summary statistics or map
- [ ] Backup: data construction details
- [ ] Backup: robustness checks and alternative specs
- [ ] Test compilation: latexmk -pdf slides.tex
- [ ] Run /slide-excellence for quality audit before presenting

Workflow 4: Deploying Slides

Option A: PDF (standard distribution)

latexmk -pdf slides.tex
# → slides.pdf ready for email / conference upload

Handout mode (2 slides per A4 page for printing):

% In preamble:
\usepackage{pgfpages}
\pgfpagesuselayout{2 on 1}[a4paper,border shrink=5mm]

Or compile with: pdflatex -jobname=slides_handout "\PassOptionsToClass{handout}{beamer}\input{slides.tex}"

Option B: GitHub Pages (shareable URL)

mkdir -p docs/slides
cp slides.pdf docs/slides/index.pdf
git add docs/slides && git commit -m "deploy slides"
git push

In GitHub: Settings → Pages → Source: main / docs/

GitHub Actions auto-build (.github/workflows/deploy-slides.yml):