thedixitjain Literature Positioning & Contribution Framing (aejpol-literature-positioning)
When to trigger
- The contribution vs. existing policy evaluations is fuzzy, or sounds like "we add another estimate"
- A referee asks "what do we learn beyond [prior study of this or a similar policy]?"
- The literature review is a list of citations rather than a map of the open policy question
- You need to frame the contribution so the policy lesson, not the method, is the news
How AEJ: Policy contributions are staked
At AEJ: Policy the contribution is a better answer to a policy question, not a new estimate for its own sake. Position the paper on three axes and make the policy reading the headline:
- The policy-design gap. What does a policymaker still not know — the sign, the magnitude, the cost-effectiveness, the incidence, the optimal level, or the external validity of an existing estimate? State the gap as something a decision-maker would care about.
- The credibility upgrade. If prior evidence exists, say why yours is more credible (cleaner design, better data, a policy variation that breaks a confound, a setting that fixes external-validity worries) — but only as a means to a sharper policy answer.
- The welfare/cost-benefit advance. The distinctive AEJ: Policy move: convert estimates into a welfare or cost-benefit statement (MVPF, cost-per-outcome, incidence, deadweight loss) that prior work left implicit. This is often the real contribution.
Positioning patterns that land
- "Prior work estimates the effect; we show whether it is worth it" (adds the cost-benefit reading).
- "Estimates exist for setting A; we show they do/do not travel to the policy-relevant setting B" (external validity for policy).
- "The textbook prescription assumes X; we show the policy-relevant elasticity implies a different optimum" (re-optimizing the policy).
- "Reduced-form effects were known; we recover the sufficient statistic the welfare calculation needs" (links to
aejpol-theory-model).
Citation discipline (AEA house)
Author–year throughout; cite the canonical methods papers behind your design (e.g., modern DID, RDD) and the key policy-evaluation papers on this and adjacent policies — referees expect you to know the policy literature, not only the econometrics.
Checklist
- The open policy question is stated as a gap a decision-maker would recognize
- Contribution framed as a sharper policy answer, not "one more estimate"
- If prior estimates exist, the credibility or welfare upgrade is explicit
- At least one contribution bullet is a welfare/cost-benefit/distributional advance
- Both the relevant policy literature and the methods literature are cited (author–year)
- No overclaim of novelty that a referee can puncture with one citation
Anti-patterns
- "We provide new evidence on X" with no statement of what was unknown and why it matters for policy
- A literature review that is a taxonomy of papers, not a map to the open policy question
- Claiming to be first while ignoring a close policy-evaluation paper (fast desk-reject signal)
- Positioning purely on identification cleverness with no policy-design or welfare advance
- Numbered
[1][2]citations instead of AEA author–year
Referee pushback mapped to the fix
- "What do we learn beyond [prior study of this policy]?" → State the specific policy-design gap (sign / magnitude / cost-effectiveness / incidence / external validity) you close, not "new evidence."
- "This duplicates a known result." → Lead with the credibility or welfare upgrade; show the prior estimate could not deliver the cost-benefit reading you do.
- "Novelty is overstated." → Cite the closest paper explicitly and scope your claim to the margin you actually advance.
Worked vignette (illustrative)
A draft on a carbon tax says "we estimate the emissions response." Referees note three prior estimates exist. Repositioned: the prior estimates ignore the revenue-recycling margin; this paper recovers the response and the incidence of the recycled revenue, so it can state the net welfare cost per ton abated under alternative recycling rules (illustrative) — a cost-benefit advance no prior estimate delivered. The contribution is now a policy answer, not a fourth elasticity.
How the three contribution bullets read in the intro
A strong AEJ: Policy intro states the contribution as up to three numbered moves, at least one of them a welfare advance: (1) credibility — "we exploit [variation] that prior work could not, removing [confound]"; (2) magnitude / external validity — "we provide the policy-relevant estimate for [population/scale] where prior estimates do not travel"; (3) welfare — "we convert the effect into [MVPF / cost-per-outcome / incidence], the object a policymaker needs." Keep each to one sentence and make the welfare bullet the one a reader remembers.
Output format
【Open policy question / gap】what a decision-maker still does not know
【Closest prior work】1–3 papers + the specific limitation you address
【Contribution bullets】(a) credibility / (b) magnitude / (c) welfare-cost-benefit — ≥1 must be welfare
【Policy "so what"】one sentence a non-specialist policymaker would grasp
【Next step】aejpol-identification (design) or aejpol-theory-model (welfare mapping)
Source: brycewang-stanford/Awesome-Journal-Skills → AEJ-Economic-Policy-Skills/skills/aejpol-literature-positioning/SKILL.md