Jeem Theory Model

Theory and Model Craft (jeem-theory-model)

When to trigger

• A pollution/abatement or resource-extraction model is in the paper but a referee asks "what does the model buy us?"
• The empirical work needs a conceptual frame that maps estimates to a welfare or optimal-policy statement
• A dynamic-resource problem (fishery, forest, groundwater, exhaustible stock) needs its Euler/Hotelling logic made explicit
• The model's comparative statics are stated but never confronted with data or used to derive the efficient instrument

What a model must earn at JEEM

JEEM is empirical-leaning, so a model is rarely the whole paper; it must do work. A model earns its place if it (a) produces a sign or magnitude prediction the empirics then test, (b) delivers the welfare or optimal-instrument expression that turns reduced-form estimates into a policy number, or (c) reveals a mechanism — leakage, avoidance, dynamic adjustment — that changes the interpretation. A model that only restates intuition or that cannot be confronted with the data should be cut to an appendix or removed.

When the paper needs no formal model

Not every JEEM paper needs a model section, and forcing one weakens the paper. A cleanly identified reduced-form estimate — a regulation's effect on emissions, a capitalization of an amenity — can stand on a compact conceptual framework that states the externality, the welfare object, and the assumptions linking the estimate to welfare, without a full optimization. The test is whether a formal model would add a tested prediction, a welfare expression, or a mechanism insight. If it would only restate the framework in equations, a clear conceptual paragraph plus the welfare-mapping assumptions (handed to jeem-identification) serves better. Reserve formal modeling for where it earns its keep.

The canonical JEEM modeling toolkit

How to make the model load-bearing

1. State the externality or stock dynamics first — the environmental structure (marginal damage function, growth function, decay) is what distinguishes a JEEM model from a generic optimization.
2. Derive one comparative static you will test — e.g., "a stricter cap raises the permit price proportionally to the abatement-cost slope," then estimate that slope.
3. Get to the welfare/instrument expression — JEEM wants the efficient tax, the optimal quota, or the welfare loss, written in terms of objects the data identify.
4. Be explicit about second-best — real environmental policy is second-best (pre-existing distortions, incomplete coverage, monitoring costs); a first-best result that ignores leakage or enforcement will draw fire.
5. Match the model's parameters to the empirical estimands so the calibration/estimation in jeem-identification and jeem-robustness actually feeds it.

Price-vs-quantity and the role of uncertainty

A distinctively environmental theory contribution lives in the Weitzman prices-vs-quantities tradition: when marginal damages and marginal abatement costs are uncertain, the choice between a tax and a cap turns on the relative slopes. If a paper proposes a regulatory instrument, the model should say why that instrument — steep marginal damages favor quantity instruments (caps), flat marginal damages favor price instruments (taxes). Stating this comparison, and tying the slopes to objects the empirics estimate, is the kind of theory move JEEM rewards over a generic optimization restatement.

Calibration discipline

When the model is calibrated rather than estimated, JEEM referees expect the calibration to be disciplined and transparent: parameters tied to the paper's own data or to clearly cited field estimates (not borrowed wholesale from an unrelated macro paper), sensitivity of the headline welfare number to the key parameters reported, and the social cost of carbon or discount rate — when they drive results — varied across the range used in the literature rather than fixed at one convenient value. A counterfactual welfare number that hinges on an undefended discount rate or an off-the-shelf damage elasticity will draw fire.