Choice is the foundational construct of microeconomic analysis. It defines the decision-making world an individual agent inhabits—what objectives they pursue, what constraints they face, how tradeoffs are structured, and which assumptions about rationality, information, risk, and time govern their behavior. Before interactions with other agents can be modeled, before markets can be formed, and before aggregates can emerge, economics must first specify Choice: the agent’s preferences, the feasible set they confront, the rules by which options are evaluated, and the conditions under which an option becomes optimal. This section makes those elements explicit, providing the conceptual machinery that all higher layers of economic reasoning must respect. By setting the boundaries of solitary decision-making and the assumptions that render it tractable, Choice establishes the primitives from which strategic interaction and macroeconomic dynamics must be constructed.

Choice (Microeconomic Foundations) – Domain – SAT

Element	Choice (Microeconomic Foundations) – SAT – Domain
Scope Category	1.1 Scope of the Domain		1.2 Ontological Commitments			1.3 State-Variables		1.4 Admissible Idealizations		1.5 Domain Assumptions		1.6 Internal Coherence Requirements
Sub-Item	Choice (Microeconomic Foundations) – SAT – Domain – Boundaries	Choice (Microeconomic Foundations) – Scale	Choice (Microeconomic Foundations) – Entities	Choice (Microeconomic Foundations) – Properties	Choice (Microeconomic Foundations) – Categories	Choice (Microeconomic Foundations) – Variables	Choice (Microeconomic Foundations) – Parameterization	Choice (Microeconomic Foundations) – Simplifications	Choice (Microeconomic Foundations) – Validity Conditions	Choice (Microeconomic Foundations) – Structural Assumptions	Choice (Microeconomic Foundations) – Implicit Commitments	Choice (Microeconomic Foundations) – Consistency	Choice (Microeconomic Foundations) – Compatibility

---

1.1 Scope of the Domain — Choice

Scope of the Domain specifies the precise decision problems Choice is permitted to analyze and the level at which those problems are described. Boundaries determine which phenomena qualify as legitimate single-agent choices; scale determines the granularity at which the agent, their constraints, and their environment are represented. Together they delineate the terrain of Choice: the world of isolated optimization, defined by preferences, feasible sets, risk, time, and information, and insulated from the strategic or aggregate forces that belong to higher layers of economic theory.

Boundaries

What Boundaries Mean in Choice
Boundaries determine exactly which decision problems qualify as single-agent choice and which do not. They define the conceptual perimeter of the domain and prevent Choice from bleeding into Interaction or Aggregation.

I. Core Inclusion Criteria — A problem belongs to Choice if ALL of the following hold:

1. The agent’s payoff depends only on their own actions.
   No other actor’s behavior modifies the outcome.
2. All environmental variables are treated as parameters, not objects of influence.
   Prices, incomes, technologies, risks, and information states are given, not shaped by the agent.
3. The agent confronts a well-defined feasible set.
   Constraints do not depend on others’ decisions or system-level aggregates.
4. Optimization is internal to the agent.
   Decisions arise from the agent’s own preferences, beliefs, and constraints, not from strategic pressure or coordination.
5. No feedback loops exist between the agent and the environment.
   The agent’s choice does not change the environment in ways that feed back into its own decision.

II. Core Exclusion Criteria — A problem is NOT Choice if ANY of the following occur:

1. Another agent’s behavior changes the payoff or feasible set.
   Coordination, rivalry, bargaining, or competition belong to Interaction.
2. Prices or constraints adjust in response to the agent’s action.
   Market clearing, externalities, and equilibrium effects require Interaction.
3. System-level variables influence the agent’s environment.
   Inflation, unemployment, interest rates, and cycles belong to Aggregation & Dynamics.
4. Outcomes arise from combining or averaging many agents’ actions.
   Aggregation is outside the scope of Choice.
5. Institutions modify incentives depending on others’ behaviors.
   Contracts, auctions, mechanisms, and regulations belong to Interaction.

III. Purpose of Boundaries in the Domain of Choice

1. Preserve conceptual purity by limiting analysis to isolated optimization.
2. Maintain mathematical tractability for marginal analysis and dynamic recursion.
3. Clarify the seams between Choice, Interaction, and Aggregation & Dynamics.
4. Prevent misuse of single-agent tools in multi-agent or macro contexts.

IV. Boundary Statement
A decision problem belongs to Choice if and only if the agent’s outcome depends solely on their own preferences, beliefs, constraints, and actions, with no influence from other agents or aggregate conditions. Any violation of this independence condition shifts the problem into a higher domain.

Scale

What Scale Means in Choice
Scale specifies the level at which Choice describes decision-making and the temporal resolution at which an agent’s optimization is modeled. It determines how finely or coarsely the agent, their environment, and their tradeoffs are represented. For Choice, scale must remain strictly at the level of a single decision-making unit.

I. The Appropriate Level of Granularity for Choice

1. The agent is treated as a unified decision-maker.
   Choice does not subdivide the agent into interacting subagents or internal components.
2. The environment is represented only through parameters.
   Prices, incomes, risk exposures, and technologies appear at the scale relevant to one agent, not markets or systems.
3. Constraints operate at the individual level.
   Budget sets, time budgets, and technological capacities refer solely to this agent’s feasible set.
4. Preferences describe one person or one firm.
   No aggregation, averaging, or representative-agent smoothing is permitted at this scale.

II. Temporal Scale in Choice

1. Single-period decisions.
   Many choice problems treat decisions as instantaneous evaluations within a single moment.
2. Multi-period or finite-horizon decisions.
   Intertemporal consumption or savings decisions unfold over discrete time but remain internal to one agent.
3. Lifetime or infinite-horizon models.
   Dynamic programming can extend across long horizons so long as the only dynamics arise from this agent’s own future states.
4. No system-wide dynamics.
   Temporal evolution cannot arise from markets, interactions, or aggregates—those belong to higher domains.

III. Why Scale Matters in Choice

1. Ensures that the tools of single-agent optimization apply correctly.
   Marginal analysis, first-order conditions, and Bellman equations require the problem to remain at the individual scale.
2. Prevents unintended drift into Interaction or Aggregation.
   If scale expands to include strategic or emergent influences, the problem no longer belongs to Choice.
3. Clarifies the proper interpretation of variables.
   Prices, risks, and constraints are taken as exogenous to this agent, not as equilibrium objects or system feedback.
4. Maintains coherence with the domain boundary.
   Choice is only valid at the scale of isolated optimization; changing scale alters the nature of the problem itself.

IV. Scale Statement (canonical form)
Choice operates exclusively at the scale of a single, unified decision-maker whose environment is described through fixed parameters and whose decisions unfold across time only within their own constraint structure. Any expansion beyond this scale shifts the problem to Interaction or Aggregation & Dynamics.

---

1.2 Ontological Commitments

Ontological Commitments specify what the domain of Choice assumes is real: the agent whose decisions are modeled, the structures that define their feasible options, and the concepts used to describe their preferences, constraints, and evaluations of risk and time. These commitments establish the basic building blocks of individual decision-making and determine how optimization problems are formulated, how their components are interpreted, and how theoretical results are derived. By identifying the entities that populate the choice environment, the properties they possess, and the categories that group them into meaningful classes, this section provides the conceptual foundation for representing single-agent behavior in a clear and coherent way, ensuring that all later analysis aligns with the ontology that Choice presupposes.

Entities:

In the domain of Choice, entities are the fundamental components that make up the structure of an individual decision problem. They define what “exists” in the choice environment: the agent, the objects they evaluate, the constraints they face, and the parameters that shape their options. Identifying entities is essential because they determine what can enter a utility function, what forms the feasible set, and what additional elements—such as risk or time—must be represented to describe the agent’s situation. By specifying these elements, the domain establishes the conceptual objects from which all models of solitary optimization are built.

Properties:

Properties are the characteristics or qualities of the entities that are considered relevant within Choice. For the agent, these properties include the structure of their preferences, the way they evaluate tradeoffs, their attitudes toward risk, their discounting of future outcomes, and their beliefs about uncertain states. For feasible sets and constraints, properties include the shape of budget lines, the nature of technological possibilities, and the informational limits that restrict what the agent can do. Environmental parameters also carry properties—such as level, variability, and relevance to constraints—that determine how they influence the decision. These attributes matter because they are what Choice models quantify or represent when analyzing behavior. Identifying properties clearly ensures that the theory focuses on meaningful and measurable aspects of decision-making, provides the basis for quantitative modeling, and allows researchers to formulate hypotheses about how changes in one attribute—such as income, risk, or time pressure—affect the agent’s choice.

Categories:

This sub-item refers to the classification scheme used within Choice to organize entities and their properties into meaningful types. Choice distinguishes categories of agents (consumer, worker, firm), categories of preferences (ordinal, cardinal, expected-utility, behavioral), categories of decision environments (certainty, risk, uncertainty, intertemporal), and categories of constraints (budget, time, technology, information). These categories help structure the theory by grouping similar decision problems and guiding which modeling tools apply in each case. They clarify whether a given choice is best understood as a static optimization, a dynamic one, a risky one, or an intertemporal one, and they ensure that terms are used consistently across analyses. Clear categorization prevents confusion—such as treating a market-level constraint as an individual-level one—and maintains coherence by ensuring that every decision problem addressed within Choice is classified according to its relevant features.

---

1.3 State-Variables

State-Variables define how the domain of Choice represents the condition of an individual decision-maker at any given moment. Variables identify the measurable or inferable aspects of the agent’s situation—such as available resources, prices faced, information possessed, or the current point in time—that determine which options are feasible and how alternatives are evaluated. Parameterization specifies how these variables are encoded into the formal structure of the model, including the units, scales, functional representations, and levels of detail used to translate the agent’s situation into a solvable optimization problem. Together, state-variables and their parameterizations provide the bridge between the conceptual ontology of Choice and the mathematical models used to analyze it, ensuring that the agent’s decision environment is represented in a clear, consistent, and operational form.

Variables:

In the domain of Choice, variables represent the measurable or inferable features of an individual agent’s situation that determine which options are feasible and how alternatives are evaluated. They encode the agent’s resources, the external conditions they face, the structure of uncertainty and time in their environment, and the internal quantities generated by the optimization process. Identifying the correct variables is essential because they define the agent’s state at any moment and provide the inputs required for utility maximization, cost minimization, dynamic choice, and comparative statics. The following categories capture the full range of state-variables relevant to individual decision-making.

Parameterization:

Parameterization specifies the precise way in which the variables of an individual decision problem are encoded so that the agent’s situation can be represented mathematically. In Choice, this means determining how resources, prices, actions, uncertainty, time, and information are mapped into formal objects such as vectors, scalars, constraints, and functional relationships. Effective parameterization preserves the essential degrees of freedom of the decision problem while eliminating unnecessary complexity, ensuring that the model reflects the meaningful structure of the agent’s environment. By choosing how each category of variables is represented—whether as vectors, sequences, probability distributions, or state-dependent functions—parameterization translates the conceptual elements of Choice into a form suitable for optimization, comparative statics, and dynamic analysis.

Parameterization in Choice (By Variable Category)

Below, each variable category is paired with how it is parameterized when constructing a single-agent optimization model.

1. Resource Variables

(income, wealth, time, endowments)
Parameterization:

• Represented as scalars (e.g., m for income, T for time).
• Included directly in constraints (p·x ≤ m; ℓ + h ≤ T).
• Treated as exogenous constants unless modeling intertemporal wealth evolution.
• Time resources may be broken into continuous units (hours) or discrete units (days, periods).

2. Price and Cost Variables

(prices of goods, wages, interest rates, input costs)
Parameterization:

• Represented as vectors (p ∈ ℝⁿ₊).
• Treated as fixed parameters the agent cannot influence.
• Inserted into budget or profit constraints (p·x, wℓ, rW).
• Sometimes normalized (e.g., set p₁ = 1) to remove redundant degrees of freedom.

3. Choice-Outcome Variables

(consumption bundles, labor, effort, output/input levels)
Parameterization:

• Modeled as decision variables x chosen from a feasible set X.
• Represented as vectors (x ∈ ℝⁿ₊), discrete alternatives, or mixed types.
• Enter utility or profit functions directly (u(x), π(x)).
• Satisfy constraints parameterized by resources and prices.

4. Uncertainty Variables

(states of the world, probabilities, signals)
Parameterization:

• States indexed by finite discrete set S = {s₁,…,sₖ}.
• Probabilities represented as probability vector π ∈ Δ(S).
• Outcomes x(s) or incomes m(s) defined state-by-state.
• Expected utility formed as Σ π(s) u(x(s)).

5. Temporal Variables

(period index, horizon length, discount factor)
Parameterization:

• Time represented as discrete periods t = 0, 1, …, T or ∞.
• Discounting encoded by scalar parameter β ∈ (0,1] or time-varying β_t.
• Decision variables indexed by time, e.g., c_t, ℓ_t.
• Constraints connect periods (budget evolution, capital accumulation).

6. Information Variables

(information sets, signals, beliefs)
Parameterization:

• Information sets modeled as σ-fields or distinct partitions of states.
• Signals represented by random variables z with known distributions.
• Beliefs parameterized via conditional probabilities π(s|z).
• Decision rules x(z) become measurable functions of information.

7. Derived Optimization Variables

(shadow prices, marginal utilities, value functions, policy functions)
Parameterization:

• Shadow prices (λ) introduced via Lagrange multipliers.
• Marginal utilities as partial derivatives ∂u/∂xᵢ.
• Value functions in dynamic programming encoded as
  V(s) = max_a { u(a,s) + β V(s’) }.
• Policy rules parameterized as functions of state variables,
  a = μ(s), where μ is the optimal decision rule.

---

1.4 Admissible Idealization

Admissible Idealization specifies which simplifications Choice is permitted to use in order to analyze individual decision-making effectively. Simplified models reduce the richness of real behavior to core elements—such as stable preferences, smooth utility, convex feasible sets, or fully rational evaluation of risk and time—so that the decision problem can be represented and solved with clarity. Limit conditions determine when these abstractions remain valid and when they fail, such as in the presence of discontinuities, behavioral biases, severe informational frictions, or nonconvex constraints. Together, these idealizations formalize the acceptable distance between actual human behavior and the theoretical representation used in Choice, ensuring that models remain tractable while still capturing the essential structure of solitary optimization.

Simplified Models:

These are deliberate abstractions of individual decision-making that Choice adopts in order to make the analysis of solitary optimization mathematically and conceptually tractable. In this domain, the most common simplifications include assuming agents possess stable, complete, and transitive preferences; representing those preferences with smooth, differentiable utility functions; treating goods and actions as continuously divisible; and modeling information, prices, and probabilities as fixed, known parameters. Such idealizations strip away the psychological complexity, discreteness, and contextual factors that shape real human choices, allowing the theory to focus on the core mechanics of tradeoffs, marginal reasoning, and constrained optimization. Admissible idealization in Choice means these simplifications are permitted so long as they illuminate essential decision structures without distorting the underlying logic of how a single agent evaluates alternatives. They function as conceptual tools, not literal descriptions of human cognition, and must be used with the awareness that they approximate behavior in order to reveal the dynamics of choice more clearly.

Simplified Models Used in Choice:

1. Static Utility-Maximization Model
   • The agent selects a consumption bundle that maximizes utility subject to a budget constraint; this is the foundational model of isolated choice.
2. Cost/Expenditure-Minimization Model
   • The agent chooses the least costly bundle that achieves a given utility level; the dual representation of utility maximization.
3. Expected Utility Model
   • Under risk, the agent evaluates lotteries by weighting utilities with probabilities, producing a clean framework for risky choice.
4. Intertemporal Choice Model
   • The agent allocates consumption across time using discounted utility, capturing tradeoffs between present and future outcomes.
5. Dynamic Programming Model
   • The agent solves sequential decisions through value functions, representing choices that depend on evolving states.
6. Single-Agent Production or Cost Model
   • A firm chooses inputs or output levels to maximize profit or minimize cost, with technology and prices treated as fixed.
7. Single-Agent Portfolio Choice Model
   • An individual allocates wealth across assets to maximize expected utility of returns, without market interaction or equilibrium.

Limit Conditions:

Every idealized model of Choice has a domain of validity—circumstances under which its simplifying assumptions produce reliable insight into individual decision-making, and beyond which those assumptions no longer hold. Standard Choice models presume coherent preferences, smooth and convex utility, continuous goods, complete information, stable beliefs, and fully rational evaluation of tradeoffs; these assumptions work well when decisions are routine, stakes are moderate, uncertainty is well-characterized, and constraints vary smoothly. But they break down when preferences are unstable or context-dependent, when goods are discrete or indivisible, when information is incomplete or noisy, when liquidity constraints bind sharply, or when behavior is driven by heuristics rather than marginal comparisons. Recognizing these limit conditions is essential, as it distinguishes the environments in which solitary optimization faithfully represents actual choice from those requiring behavioral refinements, richer information structures, or alternative modeling frameworks. Stating these boundaries prevents the overextension of idealized Choice models into settings where they misdescribe behavior and helps clarify where new theoretical developments are needed.

1. Preferences

If preferences are not coherent → utility theory collapses → all choice models collapse.
This is one whole category by itself.

2. Feasible Sets / Constraints

If constraints are weird (nonconvex, discontinuous, discrete) → calculus-based optimization fails.
Different category entirely.

3. Information/Beliefs

If the agent cannot form expectations → uncertain/dynamic choice is impossible.
Another category.

4. Environmental Independence

If the environment reacts to the agent → the problem leaves Choice and becomes Interaction.
This is a domain boundary, not a preference/belief issue.

---

1.5 Domain Assumptions

Domain Assumptions identify the background commitments that Choice takes for granted when analyzing individual decision-making. Structural assumptions specify the basic stances of the domain—such as the existence of coherent preferences, well-defined feasible sets, and an optimization rule that governs selection—along with representational choices like continuity, convexity, or probabilistic descriptions of uncertainty. Implicit commitments include the expectation that individuals can evaluate tradeoffs meaningfully, that constraints capture all relevant limitations on behavior, and that preferences and technologies remain stable over the relevant horizon. Together, these assumptions form the conceptual scaffolding that shapes how Choice interprets behavior and what kinds of explanations it accepts as valid, ensuring consistency across models while restricting the domain to situations where solitary optimization provides an adequate description.

Structural Assumptions:

These are the broad, foundational commitments that Choice takes for granted about how individual decision-making operates. They are not tied to any specific model but permeate the entire framework of solitary optimization. At this level, the domain assumes that agents possess coherent preferences that can be represented consistently, that these preferences and constraints define a stable optimization problem, and that choices arise from maximizing or minimizing a well-defined objective. It also assumes a background structure for time, uncertainty, and feasibility—such as treating goods as divisible, constraints as continuous and well-behaved, and uncertainty as representable by probability distributions—because these stances determine the mathematical form that choice models can take. Such assumptions shape the entire architecture of decision theory: for example, assuming smoothness and convexity enables marginal analysis, while assuming probabilistic uncertainty allows expected utility to structure evaluation under risk. If these assumptions fail or only approximate reality, the resulting models may mischaracterize behavior or require more complex formulations. Making these structural commitments explicit clarifies the underlying stance of the domain and allows economists to question, refine, or replace them as evidence and theory evolve.

Implicit Commitments:

These refer to the tacit assumptions embedded in the practice of Choice that are rarely stated outright but are essential for the framework to function. Researchers often inherit these commitments from the dominant microeconomic paradigm, and because they are so familiar, they usually remain unexamined. For instance, Choice implicitly assumes that individuals can meaningfully compare alternatives; that preferences, while abstract, are sufficiently stable to support optimization; that utility is a legitimate representational tool rather than a literal psychological construct; and that agents behave as if they process tradeoffs coherently even when real cognition is more complex. It also assumes that environmental parameters such as prices or probabilities can be treated as fixed inputs rather than emergent phenomena. These implicit commitments shape how decision problems are framed and what counts as an explanation within the domain. Making them explicit allows economists to evaluate when they hold, when they serve merely as convenient modeling conventions, and when they need revision—especially in contexts where behavior systematically departs from the predictions of standard Choice models.

---

1.6 Internal Coherence Requirements

Internal Coherence Requirements ensure that a scientific domain forms a unified, non-contradictory whole. Consistency demands that its principles and definitions never conflict; compatibility requires that its entities, variables, assumptions, and laws integrate into a single workable framework. Together, they impose the logical discipline that allows a science to function as a coherent system rather than a collection of disconnected claims.

Consistency:

Consistency in the domain of Choice requires that all assumptions, definitions, and modeling components fit together without contradiction, ensuring that the theory of individual decision-making operates as a coherent whole. This means that the preference structures assumed in one part of the model must align with the utility representation used elsewhere; that the constraints defining the feasible set must be compatible with the optimization methods applied; and that the treatment of uncertainty or time must not conflict with how choices are evaluated. For example, assuming transitive preferences while simultaneously using a utility form that implies preference reversals would create an internal conflict, as would combining differentiable optimization tools with non-smooth or discrete feasible sets. Maintaining consistency involves careful attention to the logical and mathematical relationships among the model’s elements, especially when extending or modifying standard assumptions. A consistent Choice framework ensures that predictions remain interpretable and that new theoretical components integrate smoothly into the established structure rather than undermining it.

Compatibility:

Compatibility in the domain of Choice requires that all components of the theory—its entities, properties, variables, assumptions, and modeling tools—fit together into a unified and coherent framework for analyzing individual behavior. The preferences assumed in one part of the model must be compatible with the utility representation chosen; the feasible set must align with the constraints imposed; and the treatment of uncertainty or time must integrate smoothly with the optimization structure. For example, representing preferences as continuous and convex while modeling the feasible set as nonconvex would create an incompatible pairing of assumptions and undermine the applicability of marginal analysis. Similarly, adopting a particular discounting structure while using a decision rule that contradicts it would fragment the theory. Ensuring compatibility means that every concept introduced in Choice relates to others in a meaningful way and supports, rather than disrupts, the overall explanation of how the agent decides. A compatible Choice framework increases explanatory power by allowing results derived from one part of the model to reinforce interpretations elsewhere, whereas incompatibility creates isolated components that cannot be integrated into a coherent account of individual decision-making.