{
  "id": "price-discrimination",
  "title": "Price Discrimination",
  "lang": "en",
  "source_lang": "en",
  "summary": "Price discrimination is the practice of charging different effective prices according to buyers’ characteristics, choices, or willingness to pay rather than differences in supply costs.",
  "category": "economics",
  "subcategory": "microeconomics",
  "type": "concept",
  "aliases": [
    "Differential pricing",
    "Discriminatory pricing"
  ],
  "generated_at": "2026-11-03",
  "model": "pro/gpt-6.1-sol",
  "url": "https://aiwiki.page/en/price-discrimination/",
  "translations": {
    "zh-Hans": {
      "url": "https://aiwiki.page/zh/price-discrimination/",
      "json": "https://aiwiki.page/zh/price-discrimination.json"
    }
  },
  "sections": [
    {
      "heading": "",
      "text": "Price discrimination is a pricing practice in which a seller charges different buyers different prices for the same or economically comparable goods or services, where the differences are not fully explained by differences in the costs of supplying them. It also includes arrangements in which the effective price varies with the quantity purchased or the contractual option selected. The central economic mechanism is the adaptation of prices to differences in buyers’ willingness to pay, rather than the application of a single price to all purchases.\n\nThe subject occupies an important place in microeconomics because it connects business pricing decisions with information, competition, income distribution, and access to goods and services. Student discounts, quantity tariffs, airline fare restrictions, negotiated business contracts, and personalized online offers can all involve price discrimination, although none can be identified as such solely from the observation that prices differ. Depending on its form and market setting, discrimination can expand consumption, transfer surplus from buyers to sellers, distort the allocation of output, or support services that would not otherwise cover their costs."
    },
    {
      "heading": "Economic meaning and conditions",
      "text": "The economically relevant comparison is between differences in prices and differences in costs. A higher delivered price for a distant customer may reflect transport expenses rather than discrimination. Similarly, a more expensive insurance policy may reflect a greater expected claim cost, and a higher-priced airline seat may provide additional flexibility or baggage services. Conversely, identical posted prices need not imply identical treatment: supplying customers with substantially different service costs at the same price can produce different price–cost margins. Empirical analysis therefore often examines effective prices, product attributes, and the costs of serving each transaction together.\n\nPrice discrimination generally requires some degree of market power. A seller must have discretion over its price rather than face a perfectly elastic demand curve at a competitively determined price. This does not require a single seller or an industry-wide monopoly. Restaurants, software suppliers, airlines, and other differentiated businesses may each possess limited pricing power despite facing rivals. Under the idealized conditions of perfect competition, identical products and freely available information make sustained discriminatory margins difficult: buyers can switch to another seller offering the same product at its competitive price.\n\nA second requirement is a way to identify or induce differences in demand. A business may observe characteristics associated with price sensitivity, such as student status, location, or purchasing history. Alternatively, it may let customers sort themselves by selecting among contracts. A discount available only after completing a rebate form, for example, can separate buyers willing to spend time obtaining a lower price from those who prefer convenience. Such sorting depends on differences in preferences or opportunity costs, not necessarily on the seller knowing each individual’s valuation.\n\nA third requirement is some limitation on arbitrage. If a buyer can purchase unlimited units at the low price and resell them cheaply to high-price customers, the price difference becomes difficult to sustain. Personal services are naturally resistant to resale: a discounted haircut cannot be transferred after consumption. Tickets may be tied to names, software licenses to accounts, and business supplies to contractual restrictions. Geographic segmentation may depend on transport costs, regulatory barriers, or distribution arrangements. Resale need not be impossible; its expected costs and risks need only be sufficient to prevent it from eliminating the relevant price differences.\n\nThese conditions explain why discriminatory pricing is especially prominent in services, digital products, and contract-based transactions. They also explain why changes in technology can work in opposite directions. Better customer identification makes segmentation easier, while price-comparison services and resale platforms may make it harder."
    },
    {
      "heading": "Historical development and the three-degree classification",
      "text": "Discriminatory pricing predates its formal economic analysis. Bargaining in markets, differentiated freight charges, and negotiated commercial contracts all allowed prices to vary across buyers long before modern pricing software. Nineteenth-century railways made the issue particularly visible because their services combined large infrastructure expenditures with relatively low incremental costs on some routes. Rail operators charged different rates across commodities, locations, and classes of travel, prompting debates over cost recovery, access, and the treatment of shippers dependent on particular lines.\n\nJules Dupuit, a French engineer, examined differentiated tolls and the benefits of public works in the 1840s. His work linked demand, willingness to pay, and the valuation of infrastructure. Railway pricing also illustrated how service quality could be used to separate customers. A carrier might offer basic accommodation at a low fare while reserving comfort and flexibility for higher-priced tickets. The relevant question was not simply why one passenger paid more than another, but how the available service categories induced passengers to reveal their preferences.\n\nArthur Cecil Pigou established the familiar classification into three degrees in The Economics of Welfare, first published in 1920. First-degree discrimination involves charging according to the valuation of each buyer or purchased unit. Second-degree discrimination varies the terms of purchase so that customers select among options, commonly through quantity schedules or product versions. Third-degree discrimination assigns different prices to identifiable groups. The classification remains a standard organizing device, but the categories are analytical ideals rather than mutually exclusive descriptions of actual businesses.\n\nLater research emphasized the informational foundations of pricing. Where sellers cannot directly observe valuations, contracts must elicit information through customers’ choices. This connects discrimination with information asymmetry, screening, and mechanism design. Work on nonlinear tariffs, including Michael Mussa and Sherwin Rosen’s 1978 analysis of quality choice, developed models in which sellers deliberately design a range of offers rather than merely attach different prices to an unchanged product.\n\nThe terminology can cause confusion outside economics. “Discrimination” here does not itself imply unlawful conduct, hostility toward a social group, or a judgment about fairness. Nor does “first degree” mean the most legally serious form. The terms distinguish how prices depend on information and purchasing arrangements. A single business may simultaneously use group discounts, quantity-based contracts, and individualized negotiations, so identifying its practices requires examining each pricing mechanism separately."
    },
    {
      "heading": "First-degree discrimination and individualized prices",
      "text": "First-degree, or perfect, price discrimination is the theoretical case in which a seller captures each unit’s willingness to pay. For a buyer willing to pay 100 monetary units for a first unit and 70 for a second, a fully informed seller could charge those amounts separately. The buyer accepts each purchase but retains no consumer surplus if the seller can extract the entire difference between valuation and payment. In continuous models, this outcome is represented by a total payment equal to the area under the buyer’s inverse demand curve up to the quantity purchased.\n\nLet \\(P(q)\\) denote willingness to pay for an additional unit and let production cost be \\(C(Q)\\). A perfectly discriminating monopolist chooses output to maximize\n\n\\[\\pi(Q)=\\int_0^Q P(q)\\,dq-C(Q).\\]\n\nFor an interior solution with differentiable functions, the condition is\n\n\\[P(Q)=C'(Q).\\]\n\nOutput therefore reaches the point where marginal willingness to pay equals marginal cost. Unlike a uniform-price monopolist, the seller need not lower the price on earlier units to sell an additional unit. Under the model’s assumptions, this removes the usual monopoly restriction of output.\n\nThe efficiency result is distinct from the distributional result. The seller receives the gains from exchange, while buyers may receive no net surplus. With no external effects and no other distortions, the allocation can be efficient even though the distribution of benefits is strongly favorable to the seller. The outcome also depends on strong assumptions: accurate information about valuations, enforceable individualized terms, negligible implementation costs, and an inability to purchase through another identity or intermediary. Valuations must also be understood in the context of buyers’ available resources; willingness to pay is not a measure of need independent of income.\n\nReal-world individualized pricing normally falls short of this ideal. Negotiated sales of business equipment, bespoke professional services, or targeted retention offers may adapt prices to perceived valuations, but sellers rarely know buyers’ maximum payments exactly. Bargaining power, competing offers, and strategic concealment affect the final price. A buyer who expects future prices to depend on current purchases may deliberately delay buying or avoid revealing a strong interest.\n\nData-driven personalization is therefore not synonymous with perfect discrimination. A prediction that a customer is likely to purchase at a particular price is probabilistic. Sellers may classify customers incorrectly, face legal restrictions on data use, or avoid personalization because of reputational costs. They may also personalize discounts rather than base prices. Economically, the resulting practice is usually partial discrimination based on imperfect signals, not the complete extraction of surplus described by the first-degree benchmark."
    },
    {
      "heading": "Second-degree discrimination and self-selection",
      "text": "Second-degree discrimination occurs when a seller offers a common menu of options and customers sort themselves by choosing among them. The seller need not identify a buyer’s type in advance. Instead, differences in demand are revealed through quantity, quality, timing, or contractual restrictions. Common examples include volume discounts, subscription tiers, data allowances, and advance-purchase fares. The defining feature is not merely that products differ, but that the menu is designed to exploit differences in valuations through self-selection.\n\nThe design problem is governed by incentive compatibility. Each customer must prefer the option intended for that customer to the alternatives. Suppose high-demand customers value both a basic service and a premium feature more highly than low-demand customers. The seller cannot price the premium option arbitrarily high: high-demand customers could switch to the basic plan. To preserve segmentation, the seller may reduce the quality or flexibility of the basic plan, thereby making it less attractive to customers with a strong preference for premium attributes. This can create deliberate distortions in the lower tier.\n\nA menu must also satisfy individual rationality: a customer must prefer participating to the available outside option. These constraints explain the concept of information rent. A seller unable to observe customer types may have to leave high-valuation buyers some surplus to keep them from selecting a cheaper contract. In standard screening models, the highest type often receives an undistorted allocation, while lower types receive reduced quantities or qualities. That pattern is model-dependent, but it illustrates why imperfect information prevents full surplus extraction.\n\nNonlinear pricing makes the total payment a nonlinear function of quantity. A simple two-part tariff takes the form\n\n\\[T(q)=A+pq,\\]\n\nwhere \\(A\\) is an access charge and \\(p\\) is a usage price. Average expenditure per unit is \\(A/q+p\\), so larger users pay a lower average price even when the marginal usage charge is constant. With identical consumers and suitable assumptions, a seller can set the usage price at marginal cost and use the access fee to collect the resulting surplus. With heterogeneous consumers, a high access fee may exclude low-demand users, creating a trade-off between extraction and participation.\n\nQuantity discounts are not automatically discriminatory. Large orders may reduce packaging, billing, storage, or distribution costs, making lower prices cost-justified. Likewise, a premium service may be substantially more expensive to provide. The empirical question is whether the price schedule reflects only these differences or also screens buyers according to demand. Bundling introduces an additional mechanism: combining products can reduce variation in total valuations when customers value the components differently. Its effects depend on the relationship between those valuations, the costs of the components, and the availability of competing products."
    },
    {
      "heading": "Third-degree discrimination and segmented demand",
      "text": "Third-degree discrimination charges different prices to identifiable customer groups. Eligibility may depend on age, student status, location, organizational affiliation, or another observable characteristic. Unlike second-degree discrimination, buyers do not obtain the relevant price solely by choosing from an unrestricted common menu. A student discount, for example, generally requires evidence of student status. The grouping variable matters economically because it predicts demand conditions, not because the category itself necessarily determines every member’s willingness to pay.\n\nFor a monopolist supplying separated groups, the standard pricing condition equates marginal revenue across groups to the marginal cost of total output. If group \\(i\\) has price \\(p_i\\), quantity \\(q_i\\), and an absolute price elasticity of demand \\(\\varepsilon_i\\), then an interior optimum with common marginal cost \\(c\\) satisfies\n\n\\[\\frac{p_i-c}{p_i}=\\frac{1}{\\varepsilon_i}.\\]\n\nThis is the inverse-elasticity rule. Subject to its assumptions, a group with less elastic demand receives a larger proportional markup. At a positive-cost interior monopoly optimum, the relevant demand is elastic; otherwise increasing price would raise revenue while reducing production costs.\n\nThe rule concerns elasticity at the chosen price, not a general label such as “rich,” “business,” or “urban.” A high-income group may have elastic demand because it has many substitutes, while a lower-income group may face inelastic demand for a particular essential purchase. In practice, student discounts may reflect limited budgets, numerous substitutes for leisure activities, and the value of attracting future customers. Business travel fares can reflect schedule constraints and flexibility requirements, although differences in ticket conditions mean that observed fares also contain a product-quality dimension.\n\nSegmented international pricing illustrates the importance of resale constraints. A supplier may charge different prices across countries because demand, alternatives, negotiation institutions, and purchasing power differ. If low-price products can be imported into high-price markets, parallel trade can erode those differences. Whether such trade occurs depends on transport costs, product compatibility, intellectual-property rules, and distribution contracts. Geographic price variation alone does not establish discrimination, because taxes, exchange rates, regulatory requirements, and supply costs can also differ.\n\nThe simple inverse-elasticity rule is not universal. It assumes separable group demands, a particular competitive setting, and no additional constraints. Capacity limits, relationships between products, bargaining, and competition can change the optimal schedule. A supplier facing a common capacity constraint must account for the opportunity cost of using scarce capacity in one group rather than another. Where customers can change groups or obtain another person’s discount, the segmentation itself becomes an additional constraint on feasible prices."
    },
    {
      "heading": "Output, allocation, and economic welfare",
      "text": "The effects of price discrimination cannot be inferred from the fact that some customers pay less and others pay more. Welfare economics distinguishes transfers between buyers and sellers from changes in total surplus. A higher payment for an unchanged quantity transfers surplus to the seller; it does not by itself destroy an equivalent amount of social value. Changes in consumption, production costs, product quality, and market participation determine whether total gains from exchange rise or fall.\n\nThird-degree discrimination has two principal effects. It may change total output, and it may change the allocation of a given output among buyers. Uniform pricing tends to equalize marginal willingness to pay among customers purchasing the same good. Discriminatory prices can instead lead one group to consume units with lower valuations while another group forgoes units with higher valuations. With constant marginal cost, this misallocation reduces total surplus for a fixed total quantity. An expansion of output can offset the loss, but an increase in output is not generally sufficient to establish an increase in welfare.\n\nA numerical example makes the distinction precise. Suppose two separated groups have demands\n\n\\[q_H=\\max\\{100-p_H,0\\},\\qquad q_L=\\max\\{60-p_L,0\\},\\]\n\nand marginal cost is 20. Under uniform monopoly pricing, the profit-maximizing price is 50, giving quantities of 50 and 10. Profit is 1,800, consumer surplus is 1,300, and total surplus is 3,100. Under third-degree discrimination, the optimal prices are 60 and 40, giving quantities of 40 and 20. Profit rises to 2,000, while consumer surplus falls to 1,000 and total surplus to 3,000. Total output remains 60, but its allocation becomes less efficient.\n\nDifferent results arise when discrimination opens a market that a uniform-price seller would leave unserved. If a low-demand group cannot support the profit-maximizing common price, a separate lower price may create additional mutually beneficial purchases. In some circumstances, both buyers in that newly served group and the seller gain without changing the terms offered elsewhere. By contrast, discrimination that principally increases charges to captive customers may raise profits while doing little to expand access. The relevant counterfactual is the feasible alternative pricing policy, not an assumption that every customer would otherwise receive the lowest observed price.\n\nLarge fixed costs add another consideration. A transport network, software product, or research-intensive service may be inexpensive to supply at the margin but costly to establish. Differentiated contributions can help cover those costs and affect whether the service exists at all. Nevertheless, cost recovery does not establish that every discriminatory schedule is efficient or necessary. Analysis must consider alternatives such as a different tariff, public funding, or competitive entry. Distributional judgments also remain separate: a policy can increase total surplus while reducing access for a disadvantaged group, or redistribute benefits without materially changing total output."
    },
    {
      "heading": "Competition, customer recognition, and pricing over time",
      "text": "Price discrimination is not restricted to monopoly, and competition can substantially alter its effects. In an oligopoly, firms may target different customer groups with discounts or personalized offers. A customer who is relatively captive to one firm may be especially attractive to another. When each seller can identify and undercut its rival’s customers, discriminatory offers can intensify competition. Prices and profits may fall relative to a situation in which every firm must quote one price to everyone.\n\nThe opposite outcome is also possible. If firms agree, without coordinating, on which groups have few alternatives, each may charge those groups relatively high prices. Market segmentation can then soften competition or redistribute its intensity across customers. The outcome depends on product differentiation, what firms know, and how customer preferences relate across suppliers. There is no general rule that improved targeting must increase industry profits or reduce consumer welfare. A practice profitable for one firm when rivals cannot imitate it may become unprofitable when all adopt it.\n\nCustomer recognition introduces a dynamic dimension. Firms may distinguish existing customers from new customers and offer different renewal or introductory prices. Switching costs—including the time needed to transfer records, learn a new system, or replace complementary equipment—can make established customers less price-sensitive. Rivals may counter with acquisition discounts. Anticipating these offers, consumers may switch strategically, bargain over renewal, or periodically create new accounts where permitted.\n\nPersonal information can also affect consumers’ behavior before prices are set. If purchasing at a high price signals strong future willingness to pay, a forward-looking buyer may postpone purchases or seek anonymity. Such behavior is sometimes described as a ratchet effect: revealing favorable private information today can lead to less favorable terms tomorrow. Sellers may respond by committing to uniform future terms or limiting how past transactions influence prices. Consequently, collecting more information need not increase the seller’s profit once customers’ responses are considered.\n\nPricing across time also raises commitment problems for durable goods. A monopolist selling a long-lived product may want a high price today but later prefer to discount remaining units. Buyers anticipating the discount may wait. Ronald Coase’s 1972 analysis of durable-goods monopoly showed how, under restrictive assumptions, the inability to commit to future prices can sharply constrain monopoly power. This is different from simply sorting impatient and patient consumers through planned discounts: future pricing incentives can undermine the initial segmentation. Product durability, resale, rental arrangements, and expectations all influence the result."
    },
    {
      "heading": "Industry practices, revenue management, and digital systems",
      "text": "Airlines, hotels, and other capacity-constrained services combine discrimination with revenue management. An unsold seat on a departed flight or an unused hotel room for a past night cannot be stored for later sale. Sellers therefore manage both prices and the availability of fare classes over time. Advance-purchase requirements, refund rules, minimum stays, and restricted booking categories can help separate travelers with different preferences. Inventory controls determine how much capacity remains available for each category as demand information develops.\n\nNot every changing fare is discriminatory. Dynamic pricing adjusts prices with time, demand, inventory, or other conditions. If every buyer at a particular moment receives the same offer, the practice may be a response to changing scarcity rather than customer-level discrimination. Nevertheless, dynamic pricing and discrimination can coexist. A seller can change all prices as capacity fills while also offering different terms to identifiable groups or maintaining a menu that induces self-selection. Distinguishing these mechanisms requires transaction-level information about timing, attributes, and availability.\n\nSoftware and digital media offer another setting because reproducing an additional copy often has a low incremental cost, while development and content creation involve substantial upfront expenditure. Suppliers can offer free, basic, and premium versions; restrict features; or charge differently for personal and organizational use. Usage limits may screen customers by demand intensity. Some differences also reflect genuine service costs, such as additional computing capacity, security obligations, technical support, or contractual guarantees. Analysis must separate these costs from the pricing role of the restrictions.\n\nIn e-commerce, sellers can use account records, location, browsing behavior, and past purchases to estimate demand. Machine learning can identify statistical relationships between observed characteristics and responses to offers. Experiments may then test alternative prices or discounts. But an algorithmically selected price is not necessarily personalized: many systems simply update a common price using aggregate market conditions. Search rankings, targeted advertisements, and individualized product recommendations can also change what customers buy without changing the prices of identical products.\n\nDigital implementation creates measurement difficulties. Two people may see different prices because they visit at different times, face different shipping charges, qualify for different promotions, or view distinct sellers on an online marketplace. A convincing test of personalized pricing must hold such factors constant or account for them explicitly. It must also distinguish price personalization from personalized ranking and from random price experiments. Questions of data privacy arise when behavioral information is collected or combined to infer valuations, while data governance determines who may access those records and how they may be used."
    },
    {
      "heading": "Legal treatment and regulatory questions",
      "text": "Economic price discrimination and unlawful discrimination are distinct concepts. Many differentiated prices are lawful, including numerous quantity discounts, negotiated prices, and reduced fares. Legal scrutiny depends on jurisdiction, the goods or services involved, the seller’s position, the identity of affected buyers, and the consequences for competition. Separate rules may address unfair commercial practices, protected personal characteristics, misuse of personal data, or contractual transparency. A price schedule can therefore be economically discriminatory without violating competition law, or legally problematic for reasons other than its effect on market competition.\n\nIn the United States, the Robinson–Patman Act of 1936 amended federal antitrust law to address certain discriminatory sales of commodities of like grade and quality. Its scope is narrower than the economic concept: it does not generally regulate every service price or every difference in consumer offers. Application involves statutory conditions concerning transactions, commerce, competing purchasers, and competitive injury. Cost justification and good-faith meeting of competition are important statutory defenses, subject to their specific requirements.\n\nUnited States case law also distinguishes injuries at different levels of commerce. Primary-line cases concern harm to competition involving the discriminating seller, while secondary-line cases concern competition among its purchasers. In Brooke Group Ltd. v. Brown & Williamson Tobacco Corp. in 1993, the Supreme Court articulated demanding requirements for a primary-line predatory-pricing claim, including below-cost pricing and a reasonable prospect of recoupment. In Volvo Trucks North America, Inc. v. Reeder-Simco GMC, Inc. in 2006, the Court emphasized the competitive relationship between favored and disfavored purchasers in the circumstances before it.\n\nIn the European Union, Article 102 of the Treaty on the Functioning of the European Union prohibits abuse of a dominant position. Article 102(c) addresses the application of dissimilar conditions to equivalent transactions with trading parties where this places them at a competitive disadvantage. The provision does not establish a general requirement that every seller charge identical prices. In its 2018 MEO judgment, the Court of Justice stressed the need to examine the relevant circumstances when assessing competitive disadvantage rather than treating any price difference as sufficient.\n\nDigital commerce adds disclosure and territorial-access rules. Amendments introduced by Directive (EU) 2019/2161 require information, where applicable, that a price has been personalized on the basis of automated decision-making in covered distance and off-premises contracts. This disclosure requirement is distinct from ordinary dynamic pricing responding to market conditions. The EU’s 2018 Geo-blocking Regulation addresses specified forms of unjustified discrimination based on nationality, residence, or establishment, while not imposing a general duty to maintain identical prices throughout the Union. Regulatory assessment consequently depends on the particular mechanism: whether the issue is exclusion, harm to downstream competition, opaque personalization, unlawful use of data, or unequal access to an offer."
    }
  ],
  "links": [
    "willingness-to-pay",
    "microeconomics",
    "market-power",
    "monopoly",
    "perfect-competition",
    "arbitrage",
    "arthur-cecil-pigou",
    "information-asymmetry",
    "mechanism-design",
    "consumer-surplus",
    "marginal-cost",
    "incentive-compatibility",
    "marginal-revenue",
    "price-elasticity-of-demand",
    "welfare-economics",
    "fixed-cost",
    "oligopoly",
    "e-commerce",
    "machine-learning",
    "online-marketplace",
    "data-privacy",
    "data-governance",
    "competition-law",
    "united-states",
    "european-union"
  ],
  "sources": [],
  "license": "https://creativecommons.org/licenses/by/4.0/"
}