brycewang-stanford Theory & Conceptual Frame (qje-theory-model)
When to trigger
- The empirics are clean but the paper reads as "a coefficient with no story"
- Referees would ask "what is the economic mechanism / model behind this?"
- There is a verbal mechanism but no formalization tying estimates to parameters
- The result is interesting but its broad implication is not articulated
Why QJE wants a frame, not necessarily heavy theory
QJE rewards a clear conceptual takeaway with broad implications — it is, after all, the journal that published Akerlof's "Market for 'Lemons'" (QJE 1970) and Spence's "Job Market Signaling" (QJE 1973), papers carried by a sharp idea rather than heavy machinery. That does not mean every paper needs a structural model. Many of QJE's most influential empirical-micro papers carry a light conceptual framework whose role is to (1) organize what is being estimated, (2) make the mechanism legible, and (3) map reduced-form estimates to an economically meaningful quantity (the sufficient-statistic style of Chetty-school QJE papers is the template). This is a real QJE-vs-sibling distinction: a paper whose contribution is a calibrated structural model often fits JPE or Econometrica better; QJE prizes the idea-plus-clean-evidence combination. Match the weight of theory to the question; do not bolt on a model the empirics cannot speak to.
Choosing the right level of formalization
| Level | Use when... | Risk if misused |
|---|---|---|
| Verbal conceptual frame | The mechanism is intuitive and the estimate is the contribution | Reads as undertheorized; "so what?" |
| Stylized model (1–3 results) | You need to show what parameter the design identifies | Best default for empirical-micro QJE |
| Sufficient-statistic frame | A few estimable objects map to welfare/policy | Powerful and parsimonious when it fits |
| Full structural model | Counterfactuals/welfare require out-of-sample extrapolation | Often a better fit for JPE/ECMA than QJE |
The framing moves
- Define the object. State precisely what economic quantity the empirical estimate corresponds to (an elasticity, a treatment effect on a welfare-relevant margin, a structural parameter).
- Make the mechanism a proposition. Even a two-equation model that yields "the sign/size of β reveals mechanism M" disciplines the interpretation and tells referees what the test is.
- Connect to welfare or policy. Where possible, use a sufficient-statistic or envelope argument to translate estimates into a quantity decision-makers care about.
- State the broad lesson. Close the loop with the topic-selection hook: the model should make the "beyond this setting" sentence rigorous and legible to a general-interest reader.
Checklist
- The estimated object is defined in terms of an economic primitive
- The model (verbal or formal) generates a testable prediction the data address
- Theory weight is matched to the question (no gratuitous structural model)
- Mapping from reduced-form estimates to the conceptual quantity is explicit
- A welfare / policy interpretation is offered where feasible
- The frame makes the "broad lesson" rigorous, not decorative
- Assumptions of the frame are stated and their plausibility discussed
Anti-patterns
- A standalone theory section disconnected from what the data can test
- A heavy structural model whose key parameters the design cannot identify (and that pushes the paper toward JPE/ECMA territory)
- "Mechanism" asserted verbally with no prediction that the empirics confirm/reject
- Theory that merely restates the regression rather than disciplining it
- A model so general it predicts any sign of the estimate (unfalsifiable)
Output format
【Estimated object】what economic primitive the coefficient maps to
【Frame level】verbal / stylized model / sufficient-statistic / structural
【Key prediction】the testable implication the data address
【Welfare/policy mapping】... (or "not applicable, by design")
【Broad lesson, made rigorous】...
【Next step】qje-robustness