{
  "id": "consumer-choice",
  "title": "Consumer Choice",
  "lang": "en",
  "source_lang": "en",
  "summary": "Consumer choice examines how people allocate limited resources among goods, services, time, and uncertain future outcomes.",
  "category": "economics",
  "subcategory": "microeconomics",
  "type": "concept",
  "aliases": [
    "Consumer choice theory",
    "Theory of consumer behavior"
  ],
  "generated_at": "2026-11-03",
  "model": "pro/gpt-6.1-sol",
  "url": "https://aiwiki.page/en/consumer-choice/",
  "translations": {
    "zh-Hans": {
      "url": "https://aiwiki.page/zh/consumer-choice/",
      "json": "https://aiwiki.page/zh/consumer-choice.json"
    }
  },
  "sections": [
    {
      "heading": "",
      "text": "Consumer choice is the study of how individuals and households select among available alternatives when their resources are limited. In microeconomics, it provides the main framework for explaining purchases, the allocation of time, saving, and responses to changes in prices and income. Its central problem is to describe how preferences interact with constraints to produce observable decisions.\n\nThe standard theory represents consumers as choosing their most preferred affordable options. This approach connects individual decisions to supply and demand, market prices, and economic welfare. Extensions examine uncertainty, decisions over time, interactions within households, and the effects of imperfect information and limited attention. Consumer choice is therefore both a mathematical theory of allocation and an empirical field concerned with how actual decisions are made."
    },
    {
      "heading": "Preferences, utility, and the development of the theory",
      "text": "Modern consumer theory developed from nineteenth-century attempts to explain value through the benefits obtained from consumption. William Stanley Jevons, Carl Menger, and Léon Walras independently advanced marginal approaches in the early 1870s. Alfred Marshall subsequently integrated demand analysis into his Principles of Economics, first published in 1890. Early accounts often discussed utility as though it were a measurable quantity of satisfaction, with purchases determined by comparisons between prices and additional satisfaction.\n\nThe later ordinal approach required less. A consumer need only rank alternatives rather than measure the amount of satisfaction each provides. Vilfredo Pareto helped develop this perspective, while John Hicks and R. G. D. Allen’s 1934 work reformulated demand theory around rankings and substitution. A utility function consequently serves as a numerical representation of preferences, not necessarily as a psychological measurement. If one bundle receives a higher utility number than another, it is ranked above that alternative; the numerical difference has no intrinsic significance.\n\nFormally, a consumption bundle is a vector \\(x=(x_1,\\ldots,x_n)\\), where each component records an amount of a good or service. The notation \\(x\\succeq y\\) means that bundle \\(x\\) is at least as preferred as bundle \\(y\\). Standard models assume completeness, meaning that any two feasible bundles can be compared, and transitivity, meaning that rankings do not form cycles. Continuity is commonly added to exclude abrupt reversals under arbitrarily small changes and to support continuous utility representations in standard commodity spaces.\n\nAn indifference curve joins bundles that receive the same ranking. When goods are desirable and preferences are convex, these curves typically slope downward and favor mixtures over extreme allocations. Convexity is an assumption about trade-offs, not a requirement that every good provide diminishing satisfaction in isolation. The marginal rate of substitution measures the amount of one good a consumer would surrender for a small additional amount of another while remaining indifferent. Where utility is differentiable, it equals a ratio of marginal utilities.\n\nOrdinal utility is unchanged in its economic meaning by any strictly increasing transformation. Thus \\(u(x)\\) and \\(\\log u(x)\\), where defined, generate the same rankings and choices. This distinction prevents an interpretation of ordinary utility numbers as directly comparable across people. Standard consumer theory can establish that one person prefers one bundle to another, but it does not by itself establish that one person gains twice as much happiness as someone else. Interpersonal comparisons require additional ethical or empirical assumptions."
    },
    {
      "heading": "Budget sets, endowments, and other constraints",
      "text": "The simplest consumer problem combines preferences with a budget constraint. With positive prices \\(p_1,\\ldots,p_n\\), quantities \\(x_1,\\ldots,x_n\\), and expenditure resources \\(m\\), affordable bundles satisfy\n\n\\[\\sum_{i=1}^{n}p_i x_i\\leq m,\\qquad x_i\\geq 0.\\]\n\nThe budget set includes all such bundles. For two goods, its boundary is a straight line with slope \\(-p_1/p_2\\). This slope expresses the market opportunity cost of obtaining one good in terms of the other. A rise in income shifts the boundary outward without changing its slope, while a change in one price rotates it around the intercept for the other good.\n\nThe meaning of income depends on the model. In a short-period spending problem, \\(m\\) may represent money available for a particular category of purchases. In a broader model it can include labor earnings, transfers, and returns on assets. When consumers own goods or productive resources initially, their purchasing power depends on the market value of their endowment. A price increase can then raise wealth as well as make purchases more expensive. For example, a household that sells more of an agricultural product than it buys may benefit from a higher price even though a household purchasing the same product loses purchasing power.\n\nActual opportunities often differ from a single linear budget. Quantity discounts, subscription fees, increasing-block electricity tariffs, and means-tested benefits generate kinks or discontinuities. A transport pass may require an initial payment but make subsequent journeys inexpensive. A benefit withdrawn as earnings rise can alter the effective return to additional work. Such arrangements can create several locally attractive choices, so a tangency condition alone may not identify the best feasible alternative.\n\nTime is another scarce resource. A household must allocate a finite number of hours among paid work, unpaid work, travel, care, leisure, and sleep. Goods that appear inexpensive in money may be costly in time. Preparing a meal from basic ingredients and buying a prepared meal can therefore be alternatives involving different combinations of monetary expenditure and household labor. Gary Becker’s 1965 analysis of time allocation formalized the idea that households combine purchased inputs and time to produce activities or outcomes they value.\n\nLegal restrictions, product availability, credit access, and information also shape feasible choices. A household may prefer an energy-efficient appliance over its lifetime but be unable to finance the initial purchase. A rural consumer may face only one accessible retailer. These examples distinguish preferences from opportunities: choosing an inferior available option does not establish a preference for it over an unavailable one. In empirical work, accurately describing the feasible set is therefore as important as specifying what consumers value."
    },
    {
      "heading": "Optimal choice and the construction of demand",
      "text": "The standard optimization problem is\n\n\\[\\max_{x\\geq 0}u(x) \\quad\\text{subject to}\\quad p\\cdot x\\leq m.\\]\n\nA solution is a utility-maximizing bundle. If preferences are continuous and the budget set is nonempty and compact, a solution exists. Strictly convex preferences generally ensure uniqueness on a convex budget set. Local nonsatiation—the condition that some arbitrarily nearby bundle is preferred to any given bundle—implies that an optimum exhausts the budget in this model. Without that assumption, a consumer may optimally leave some purchasing power unspent.\n\nAt an interior optimum with differentiable utility, the first-order conditions imply\n\n\\[\\frac{\\partial u/\\partial x_i}{\\partial u/\\partial x_j} = \\frac{p_i}{p_j}.\\]\n\nThe consumer’s willingness to exchange goods at the margin then equals the exchange rate offered by the market. Equivalently, marginal utility per unit of expenditure is equalized across goods consumed in positive quantities. This condition is necessary under suitable regularity assumptions; concavity or related conditions are needed to make it sufficient. It is not a universal description of every purchase.\n\nCorner solutions illustrate the limitation. If two goods are perfect substitutes and one provides more utility per unit of expenditure, the consumer may buy only that good. With perfect complements, such as inputs useful only in fixed proportions, indifference curves have kinks and ordinary tangency conditions need not apply. Indivisible purchases—a refrigerator, a car, or a dwelling—also require comparisons among discrete alternatives rather than infinitesimal adjustments.\n\nThe resulting choice as a function of prices and income is called Marshallian demand, written \\(x(p,m)\\). For Cobb–Douglas utility,\n\n\\[u(x_1,x_2)=x_1^\\alpha x_2^{1-\\alpha}, \\qquad 0<\\alpha<1,\\]\n\noptimal demands are\n\n\\[x_1=\\frac{\\alpha m}{p_1}, \\qquad x_2=\\frac{(1-\\alpha)m}{p_2}.\\]\n\nThe consumer spends constant income shares on the two goods. If \\(m=120\\), \\(p_1=3\\), \\(p_2=2\\), and \\(\\alpha=1/2\\), the chosen quantities are 20 and 30. This functional form is analytically convenient, but its fixed expenditure shares are an empirical restriction rather than a general law.\n\nSeveral properties follow from the framework. Demand is homogeneous of degree zero in prices and income: multiplying every price and income by the same positive factor leaves the budget set unchanged. Under budget exhaustion, \\(p\\cdot x(p,m)=m\\). Substituting optimal choices into the utility function gives indirect utility, \\(v(p,m)\\), which records the best attainable ranking at each price-income combination. Individual demand can be added across consumers to obtain market demand, although the aggregate need not behave like the demand of a single representative person."
    },
    {
      "heading": "Price changes, income effects, and substitution",
      "text": "A price change does two things simultaneously: it changes relative opportunities and changes the purchasing power of a given money income. Consumer theory separates these channels into substitution and income effects. The substitution effect measures adjustment toward relatively cheaper goods while holding the consumer’s welfare constant in an appropriate sense. The income effect measures the additional adjustment caused by the change in real purchasing power.\n\nThe corresponding dual problem minimizes the cost of attaining a specified utility level:\n\n\\[e(p,\\bar u) = \\min_{x\\geq 0} \\left\\{p\\cdot x:u(x)\\geq\\bar u\\right\\}.\\]\n\nIts solution is Hicksian demand, denoted \\(h(p,\\bar u)\\). Unlike ordinary demand, it holds utility rather than money income fixed. The expenditure function \\(e(p,\\bar u)\\) shows how much money is required to reach a given level of welfare at different prices. Under differentiability, its derivative with respect to a price equals the compensated demand for that good.\n\nThe Slutsky equation connects ordinary and compensated responses:\n\n\\[\\frac{\\partial x_i}{\\partial p_j} = \\frac{\\partial h_i}{\\partial p_j} - x_j\\frac{\\partial x_i}{\\partial m},\\]\n\nwhere compensated demand is evaluated at the consumer’s initial utility. For an own-price increase, the compensated substitution effect is nonpositive under standard assumptions. If the good is normal, so that demand rises with income, the loss of purchasing power reinforces the reduction in demand. If it is inferior, the income effect works in the opposite direction.\n\nA Giffen good is an inferior good for which the opposing income effect is strong enough that ordinary demand rises when its own price rises. This requires a particular pattern of preferences and expenditure dependence; it does not mean that consumers disregard prices. Robert Jensen and Nolan Miller’s 2008 study of staple-food subsidies in China provided experimental evidence of Giffen behavior for rice among some poor households in Hunan. The result concerned particular households and circumstances, not an inherent property of rice everywhere.\n\nCross-price responses describe substitutes and complements. Compensated substitutes have positive compensated cross-price effects, whereas compensated complements have negative ones. Ordinary cross-price effects can differ because they incorporate income effects. This distinction matters when interpreting whether two products are competitors or are typically used together.\n\nEmpirical responses are often expressed as price elasticities of demand, which measure percentage quantity changes relative to percentage price changes. Elasticities can vary with income, available substitutes, adjustment time, and the share of spending devoted to a product. Engel curves instead relate consumption to income at fixed prices. They allow expenditure shares to change as households become richer, rather than imposing the fixed shares of a Cobb–Douglas model."
    },
    {
      "heading": "Revealed preference and empirical identification",
      "text": "Preferences are not directly observable. Economists typically observe purchases, prices, incomes, and characteristics, then ask what those observations imply about underlying choices. Revealed preference theory reverses the usual reasoning: rather than starting with a utility function and deriving demand, it asks whether observed decisions could have been generated by stable utility maximization.\n\nPaul Samuelson introduced a revealed-preference approach in 1938. Its central intuition is that choosing bundle \\(A\\) when bundle \\(B\\) was affordable reveals that \\(A\\) is at least as preferred as \\(B\\), assuming the relevant options and circumstances are properly specified. If a later choice establishes a contradictory strict ranking, the observations may violate the proposed model. Such tests concern consistency across choices, not whether a particular purchase appears sensible to an outside observer.\n\nThe generalized axiom of revealed preference, or GARP, extends this logic to chains of choices. Roughly, if a sequence of observations reveals that \\(A\\) is at least as preferred as \\(B\\), then \\(B\\) cannot be chosen when \\(A\\) is strictly cheaper. Sidney Afriat’s 1967 theorem established that, under standard conditions, a finite dataset satisfying the relevant consistency restrictions can be rationalized by a continuous, monotonic, concave utility function. This is powerful because testing does not require selecting a particular algebraic form for utility in advance.\n\nEmpirical demand estimation also uses explicitly specified models. Household expenditure surveys and retail scanner data can show how spending varies across prices, income groups, products, and time. Angus Deaton and John Muellbauer’s 1980 Almost Ideal Demand System provided a flexible way to estimate expenditure shares while imposing restrictions derived from consumer theory. Such systems are useful for analyzing food demand, tax changes, and differences in consumption patterns across households.\n\nDiscrete-choice methods address alternatives such as car models, transport modes, or mobile-phone plans. In a random utility model, an option’s utility contains an observed component and a component unobserved by the analyst. Under particular assumptions about the latter, choice probabilities take the multinomial logit form:\n\n\\[P_{ij} = \\frac{\\exp(V_{ij})} {\\sum_k\\exp(V_{ik})},\\]\n\nwhere \\(V_{ij}\\) is the modeled utility of option \\(j\\) for person \\(i\\). The simple model’s independence-of-irrelevant-alternatives restriction can be implausible when some alternatives are much closer substitutes than others; nested and mixed models relax aspects of that structure.\n\nEstablishing a causal price response requires more than correlating prices with sales. Prices may rise precisely when unobserved demand is strong, while retailers may discount products with weak expected demand. Experiments, instrumental variables, and natural experiments can help distinguish price effects from these confounding forces. Product availability, measurement error, changing quality, and unobserved household differences remain important limitations even in large datasets."
    },
    {
      "heading": "Choices over time and under uncertainty",
      "text": "Many decisions exchange present consumption for uncertain future benefits. Saving, education, durable purchases, and retirement planning cannot be adequately represented as isolated purchases from a current budget. Intertemporal models instead treat consumption at different dates as distinct alternatives, linked by borrowing, saving, asset returns, and expectations.\n\nIn a simple two-period model with unrestricted borrowing and lending at a common real interest rate \\(r\\), the lifetime budget is\n\n\\[c_1+\\frac{c_2}{1+r} = y_1+\\frac{y_2}{1+r}.\\]\n\nCurrent consumption \\(c_1\\) and future consumption \\(c_2\\) are financed by current and future resources \\(y_1\\) and \\(y_2\\). A common preference representation is \\(u(c_1)+\\beta u(c_2)\\), where \\(\\beta\\) measures the weight placed on future utility. At an interior optimum,\n\n\\[u'(c_1)=\\beta(1+r)u'(c_2).\\]\n\nThe consumer balances the marginal value of consuming now against the discounted marginal value of the future consumption obtained by saving.\n\nThis framework underlies explanations of saving and consumption smoothing. Milton Friedman’s permanent income hypothesis, developed in his 1957 book A Theory of the Consumption Function, distinguishes persistent resources from temporary income movements. Franco Modigliani and Richard Brumberg’s life-cycle approach, developed in the 1950s, emphasized the allocation of resources over working life and retirement. Neither approach implies that every household can smooth consumption freely: borrowing constraints, uncertain employment, and limited liquid assets can make current spending highly sensitive to current income.\n\nUnder uncertainty, alternatives can be represented as lotteries over outcomes. John von Neumann and Oskar Morgenstern’s 1944 Theory of Games and Economic Behavior established an axiomatic foundation for expected utility theory. With outcomes \\(c_s\\) occurring with probabilities \\(\\pi_s\\), expected utility is\n\n\\[EU=\\sum_s\\pi_s u(c_s).\\]\n\nUnlike utility over certain bundles, this representation is preserved by positive affine transformations, not by every strictly increasing transformation. The numerical spacing of utility values matters because it affects the evaluation of probabilities and risk.\n\nConcavity of utility over wealth or consumption implies risk aversion: a certain amount is preferred to a fair gamble with the same expected monetary value. This helps explain demand for insurance and diversification. The certainty equivalent of a risky prospect is the guaranteed amount regarded as equally desirable; the difference between expected monetary value and that certainty equivalent is the risk premium. Insurance choices nevertheless also depend on premiums, deductibles, coverage exclusions, and credit constraints.\n\nCombining time and uncertainty leads to models such as\n\n\\[\\max E\\left[\\sum_{t=0}^{T}\\beta^t u(c_t)\\right]\\]\n\nsubject to evolving resources and information. Precautionary saving can arise when uncertain future income interacts with preferences displaying prudence, commonly represented by a positive third derivative of utility. Expectations, survival probabilities, and the possibility of revising decisions all influence the resulting consumption path."
    },
    {
      "heading": "Households, social relationships, and the limits of aggregation",
      "text": "A consumer is not always a single person acting independently. Many purchases are made by households whose members have different preferences, resources, and degrees of influence. Treating the household as one utility-maximizing unit can be a useful simplification, but it suppresses the question of how individual interests are combined.\n\nThe unitary household model assumes that household behavior can be represented by a single preference ordering and a common budget. One important implication is income pooling: if total household income and relevant prices remain unchanged, reallocating income between members should not affect purchases. This implication is testable and often too restrictive. Evidence that the source or recipient of income affects expenditure has motivated models in which decision-making authority matters.\n\nCollective household models instead begin with individual preferences and an explicit or implicit allocation process. A common specification maximizes a weighted combination of members’ utilities, subject to household resources. The weights may depend on bargaining power, property ownership, outside opportunities, or institutional rules. Some models assume that the resulting allocation is Pareto efficient within the household, meaning that no feasible reallocation can improve one member’s welfare without harming another. Efficiency does not imply equal bargaining power or an equitable distribution.\n\nGoods also differ in how they are shared. Food portions and personal clothing are often individually consumed, while housing space, heating, and some durable goods provide benefits to several members. Joint consumption creates economies of household size. An equivalence scale adjusts household income or expenditure to account for composition, but its construction requires assumptions about needs, sharing, and living standards. Two households with equal spending need not provide the same welfare to each member.\n\nPreferences can also depend on other people’s behavior. Status goods may be valued partly because they signal relative position. Demand for a communication service may increase as more acquaintances adopt it, producing a network effect. Social norms can influence clothing, food, gifts, and ceremonial spending. Such dependencies do not automatically make choice irrational, but they mean that a person’s utility cannot always be modeled as depending only on their own physical consumption.\n\nAggregation creates a separate analytical problem. Individual demands can always be added arithmetically, but the resulting total generally depends on how income is distributed as well as on aggregate income. Conditions under which market demand behaves exactly like the choice of a single representative consumer are restrictive. Consequently, changes in income inequality can alter demand patterns even when total income is constant. Models of aggregate consumption must therefore distinguish a convenient representative-agent approximation from a literal description of all households’ preferences."
    },
    {
      "heading": "Behavioral departures from the standard model",
      "text": "Standard consumer theory supplies a benchmark, not a claim that every observed decision satisfies all its assumptions. Behavioral economics studies systematic departures and develops alternative models that retain explicit mechanisms. These include limits on attention and calculation, context-dependent valuations, inconsistent plans, and beliefs that differ from objective probabilities.\n\nHerbert A. Simon emphasized bounded rationality in the 1950s. Finding the best option may require information and computation that are costly or unavailable. A person may use satisficing—selecting an option that meets an acceptable threshold—rather than exhaustively ranking every alternative. Search itself can also be modeled as a costly activity: consumers may stop comparing prices when the expected benefit of another search falls below its cost. Limited search therefore need not always contradict an optimizing framework.\n\nDaniel Kahneman and Amos Tversky’s 1979 prospect theory described risky choice in terms of gains and losses relative to a reference point, rather than only final wealth. Its mechanisms include loss aversion and nonlinear weighting of probabilities. These can help explain why economically equivalent descriptions generate different choices and why small probabilities sometimes receive disproportionate weight. Reference dependence also contributes to explanations of the endowment effect, in which ownership changes the valuation of an object, although its magnitude varies with context and experimental procedures.\n\nPresent bias concerns a disproportionate preference for immediate rewards over delayed rewards. A consumer may plan to save next month yet reverse that decision when next month arrives. Quasi-hyperbolic discounting represents this using an extra discount applied to all future periods:\n\n\\[U_t=u(c_t)+\\beta\\sum_{k=1}^{\\infty}\\delta^k u(c_{t+k}), \\qquad 0<\\beta\\leq 1.\\]\n\nWhen \\(\\beta<1\\), immediate utility receives special weight. This differs from standard exponential discounting, in which relative weights between dates remain consistent as time passes.\n\nMental accounting describes the tendency to separate resources into categories such as rent money, holiday savings, or windfall income. Such categories can support budgeting but may also lead consumers to treat economically interchangeable funds differently. Defaults, display order, and the prominence of particular charges can likewise affect choices. Choice architecture examines how the presentation of alternatives influences decisions without necessarily changing the alternatives themselves.\n\nBehavioral findings create a welfare-identification problem. If choices change with framing, there may be no single observed ranking that can straightforwardly be treated as the person’s welfare ordering. Researchers therefore distinguish evidence of a behavioral mechanism from conclusions about which intervention improves welfare. Evaluations of defaults or a nudge may consider informed choices, stated preferences, subsequent satisfaction, and the costs of opting out. No one of these measures universally resolves the tension between respecting observed decisions and recognizing decision errors."
    },
    {
      "heading": "Welfare measurement and policy applications",
      "text": "Consumer choice theory supplies tools for evaluating price changes, taxes, product introductions, and changes in access. In welfare economics, a central question is how much money would compensate a consumer for a less favorable opportunity set, or how much could be taken away after an improvement while leaving that consumer no worse off. These are monetary measures of preference satisfaction under specified assumptions, not direct measures of happiness.\n\nFor a price change from \\(p^0\\) to \\(p^1\\), with initial and final utility levels \\(u^0\\) and \\(u^1\\), compensating variation for a price increase can be written as\n\n\\[CV=e(p^1,u^0)-e(p^1,u^1).\\]\n\nIt measures the additional expenditure needed at new prices to restore initial utility. Equivalent variation measures the corresponding welfare loss at old prices:\n\n\\[EV=e(p^0,u^0)-e(p^0,u^1).\\]\n\nThe two generally differ because willingness to exchange money for goods varies with purchasing power. Under unchanged nominal income and standard budget exhaustion, \\(e(p^1,u^1)\\) and \\(e(p^0,u^0)\\) both equal that income.\n\nConsumer surplus, often represented by the area below a demand curve and above price, offers a simpler approximation. It is exact for certain preference structures, notably quasi-linear utility when income effects for the relevant good are absent over the range considered. With substantial income effects or large changes, expenditure-function measures provide a more precise treatment. The distinction matters particularly for necessities that absorb a large share of a low-income household’s budget.\n\nTax analysis combines these welfare tools with behavioral responses. An excise tax changes relative prices and can distort consumption away from the untaxed optimum. In an otherwise undistorted market, the resulting deadweight loss depends on compensated responses rather than simply on tax revenue. Where an externality exists, however, a corrective tax can improve social welfare by aligning private choices with social costs. The welfare assessment must then include effects on people other than the purchaser.\n\nConsumer models also inform competition analysis. Estimated substitution between products helps define markets, evaluate market power, and predict the effects of mergers or product withdrawals. Demand for differentiated products is particularly important when consumers value characteristics such as location, reliability, brand, and service quality. A new product can benefit consumers through expanded choice even if it changes neither their nominal income nor the prices of existing products.\n\nPrice-index construction raises related questions. A cost-of-living index compares the minimum expenditure required to maintain a specified utility level across price systems. The consumer price index approximates changes in consumer prices using expenditure data and statistical methods, but it is not identical to a complete welfare measure. Substitution, quality change, new goods, and differences in household spending patterns complicate the comparison. A rise in energy prices, for example, affects households differently according to heating technology, climate, transport needs, and income; an average index cannot fully represent each household’s change in feasible consumption."
    }
  ],
  "links": [
    "microeconomics",
    "supply-and-demand",
    "alfred-marshall",
    "vilfredo-pareto",
    "utility-function",
    "marginal-rate-of-substitution",
    "opportunity-cost",
    "gary-becker",
    "marginal-utility",
    "income-effect",
    "giffen-good",
    "substitute-and-complementary-goods",
    "price-elasticity-of-demand",
    "saving",
    "milton-friedman",
    "expected-utility-theory",
    "risk-aversion",
    "insurance",
    "pareto-efficiency",
    "network-effect",
    "income-inequality",
    "behavioral-economics",
    "herbert-a-simon",
    "bounded-rationality",
    "satisficing",
    "prospect-theory",
    "loss-aversion",
    "endowment-effect",
    "present-bias",
    "mental-accounting",
    "nudge",
    "welfare-economics",
    "consumer-surplus",
    "excise-tax",
    "deadweight-loss",
    "externality",
    "market-power",
    "consumer-price-index"
  ],
  "sources": [],
  "license": "https://creativecommons.org/licenses/by/4.0/"
}