A microeconomic concept defining how agents make choices and how markets allocate resources under constraints. It specifies relationships among incentives, prices, quantities, and strategic behavior used to predict outcomes. It does not guarantee predictive accuracy without assumptions about preferences, technology, and information available to participants. It supports pricing, regulation, and welfare analysis by clarifying tradeoffs and likely responses to changes in incentives. The concept is generally stable, though empirical methods and market design practices evolve over time.

Principle

Encode causal and behavioral relations (price, promotion, seasonality, cross-effects, customer heterogeneity) so the model can generalize beyond observed instances while quantifying uncertainty and bias.

Demonstration

A retailer fits a mixed-effects time-series model that estimates baseline demand, price elasticity per customer segment, and promotion lift; the model is used to simulate outcomes under proposed price and promotion schedules.

Misapplication

Using a black-box predictive model without testing counterfactuals or stability over policy changes, or applying a model trained on one market segment to another with different behavior without recalibration.

Consequence

Provides inputs for pricing optimization, inventory planning, and scenario analysis; when well-specified it supports counterfactual simulations and policy evaluation with quantified confidence.

Reversal

Ad hoc heuristics or purely descriptive summaries that are not structured to predict counterfactual outcomes or to incorporate causal assumptions.

Boundary

Includes econometric, machine-learning, structural, and hybrid specifications; excludes pure accounting of past sales without structural assumptions and models that ignore external drivers (weather, competition) when those drivers matter.

Semantic Tension

Tension between purely predictive ML models (optimized for out-of-sample accuracy) and structural models (optimized for causal interpretation and policy simulation); both contribute but trade off interpretability and robustness to distributional shifts.

Synthesis

A demand model is a purpose-built mathematical tool that formalizes how observable inputs and latent factors combine to produce demand, enabling prediction, simulation, and policy-driven decisions with explicit assumptions and uncertainty.