mgsci-theory-development - SKILL.md Agent Skill

name: mgsci-theory-development description: Use when building the theoretical core of a Management Science (INFORMS) paper — either a formal analytical model (assumptions, equilibrium, propositions/theorems, comparative statics) or empirically testable hypotheses derived from a clear mechanism. Adapts to the paper's lane; it does not run the analysis (mgsci-data-analysis) or frame the contribution (mgsci-contribution-framing).

Theory Development (mgsci-theory-development)

When to trigger

You have an intuition but no model or no mechanism
Hypotheses are descriptive ("X relates to Y") with no stated mechanism
An analytical model has results but unclear/unstated assumptions
A Department Editor or reviewer asked "what is the theory here?" or "why is this the right model?"

Because Management Science is bimethodological, "theory development" means one of two disciplined things depending on your Department lane.

Lane A — Analytical / formal model

For Optimization and Decision Analytics, Stochastic Models and Simulation, Operations Management, Finance (theory), Business Strategy (theory), and economic-theory submissions:

State assumptions explicitly and defend them. Each assumption should be necessary, behaviorally or economically reasonable, and you should know which results break without it.
Define the decision problem. Players/decision-maker, actions, timing, information, objective. Make the optimization/equilibrium concept precise.
Derive results as propositions/theorems with proofs (in text or appendix). Reviewers check that proofs are correct and that results are not artifacts of a knife-edge assumption.
Comparative statics and intuition. The managerial payoff lives here: how the optimal policy / equilibrium moves with primitives, and why — stated in plain words, not only symbols.
Robustness of the model. Show which extensions (heterogeneity, alternative timing, relaxed assumption) preserve the qualitative insight.

Lane B — Empirical hypotheses from a mechanism

For Accounting, Finance (empirical), Information Systems, Behavioral Economics and Decision Analysis, Marketing, Data Science:

Name the mechanism (incentive, information, frictional, behavioral) that produces the predicted effect; the hypothesis is the mechanism's observable implication.
Derive predictions a priori, before looking at outcomes — HARKing (hypothesizing after results) is a credibility failure.
Moderation = boundary condition of the mechanism; mediation = the mechanism's intermediate step. Predict the sign and the conditions, not just "an effect exists."
Where a behavioral prediction comes from a model or from decision theory, state the formal source.

The unifying bar

Whichever lane, the theory must yield a decision-relevant, cross-department insight — Management Science rewards a mechanism whose managerial implication a reader in another Department still finds informative. Formal elegance without managerial reading, or hypotheses without a mechanism, will not clear the bar.

Anti-patterns

A model whose results depend on an undefended knife-edge assumption.
Propositions with no comparative statics and no plain-language intuition.
Empirical hypotheses with no named mechanism (pure correlation dressed as theory).
Post hoc hypotheses fitted to the data (HARKing).

Theory pass for Management Science

Run this as a concrete capability pass. First lock the decision problem, formal or empirical engine, managerial lever, and generality claim; then test whether the manuscript addresses OR/MS reviewers who expect a generalizable decision model, credible empirical leverage, or algorithmic insight with managerial consequence.

Primary move: Separate construct, mechanism, scope condition, and testable implication; refuse a theory section that only summarizes prior work.
Decision ledger: return claim / evidence / blocker / next edit rows so the next pass can patch the manuscript directly.
Sibling comparison: compare against Operations Research for method-first optimization, Marketing Science for marketing models, Organization Science for organization-theory mechanisms; if the neighboring outlet has the stronger audience claim, recommend re-routing before polishing.
Verification floor: before submission-ready advice, re-open resources/official-source-map.md for volatile rules and name the one unresolved fact that could change the recommendation.

Worked micro-example (illustrative) and theory pushback

A platform model assumes buyers single-home and derives that the platform always raises commissions as network effects grow. In the flagship's Operations Management / Business Strategy lane, a referee asks whether the result is mechanical — driven entirely by the single-homing assumption. The fix names the mechanism (network effects raise the platform's marginal value of a seller), relaxes single-homing as a scope condition, and shows the comparative static can reverse when multi-homing is cheap (commissions fall from 15% to about 9%, illustrative). The proposition now carries a managerial reading a finance reader also grasps — take rates are not monotone in network strength — the cross-department insight the flagship rewards.

"The model insight is mechanical." Show the result survives relaxing the assumption that drives it, or state precisely the scope condition under which it holds and reverses.
"What is the theory here?" Name the mechanism explicitly; a proposition or hypothesis without a mechanism reads as algebra or correlation, not theory.
"Elegant but no managerial reading." Add the comparative-static or boundary-condition intuition that tells a decision-maker in another department what changes.

Output format

【Lane】analytical model / empirical-mechanism
【Core mechanism】...
【Assumptions or constructs】necessary? defended?
【Results】propositions/theorems OR a priori hypotheses (sign + conditions)
【Comparative statics / boundary】managerial intuition ...
【Next step】mgsci-methods