RESEARCH DOSSIER
Monopoly Pricing and Deadweight Loss
Abstract. Monopoly pricing is investigated as the seller's constrained choice of prices and output and as a source of welfare differences relative to explicitly defined alternative allocations. This monograph derives the uniform-price monopoly condition, the elasticity relationship and the deadweight-loss integral, then examines demand curvature, cost recovery, price discrimination, access tariffs, capacity, multiproduct pricing, income effects, rent acquisition and dynamic investment. Transparent numerical comparisons isolate output restriction, cross-group allocation, market opening and fixed-cost viability. The historical and empirical inquiry covers the development of monopoly measurement, differentiated-product estimation, U.S. aggregate markups, California electricity, pharmaceutical competition and international price indexes, and UK network regulation. The original analytical synthesis distinguishes transfers from forgone resource gains and requires that each welfare conclusion specify the allocation, the welfare boundary and a feasible counterfactual. The work uses conditional formal analysis and documented source assessment; numerical illustrations are constructed, and published empirical estimates are attributed without claiming independent replication.
Keywords: monopoly pricing; deadweight loss; market power; elasticity; Lerner index; consumer surplus; producer surplus; price discrimination; natural monopoly; structural demand estimation; welfare economics; industrial organization
The scientific object and its boundaries
A monopolist controls the supply of a defined good or service within an economically relevant market. Monopoly pricing concerns the relationship between this control, the demand faced by the seller, the resources required to deliver an additional unit, and the institutional rules under which exchange occurs. Deadweight loss concerns the opportunities for mutually beneficial exchange or more efficient production that disappear relative to an explicitly specified feasible benchmark. The two objects belong together because the seller's optimization problem and society's resource allocation problem differ in their treatment of price. A seller receives the price paid on every unit it sells. A social surplus calculation treats that payment as a transfer between buyer and seller and values the underlying units and resources.
A market boundary is a substantive assumption. A supplier may be the only producer of a patented formulation while competing with other treatments. A town may have one distribution network but several retailers using it. A digital service may be free for users yet generate revenues from advertisers. A single seller of a narrowly described object cannot automatically be assigned the demand curve for a broader economic market. Similarly, a supplier with many competitors may still face a downward-sloping residual demand curve. The analysis therefore distinguishes literal single-seller monopoly from a monopoly model used as a benchmark for unilateral pricing power.
This dossier belongs to Microeconomics (EBS-002; JEL D40 and L11). Its central object is monopoly pricing, surplus and discrimination. Related subjects such as oligopoly equilibrium, auctions, vertical foreclosure, production duality, algorithmic collusion, and the law of individual jurisdictions enter only when they change this object's interpretation. The inquiry uses conditional mathematical derivations, historical and empirical source assessment, and comparative institutional analysis. No new named economic theory is introduced. The central analytical contribution is a consistent accounting of price, quantity, transfers, the missing transaction margin, and the feasibility of the alternative allocation.
Historical development of monopoly analysis
The intellectual movement from descriptive accounts of monopoly toward measurement depended on distinguishing ownership of a market position from the economic wedge imposed through its exercise. In the measurement of monopoly power, Lerner (1934) developed a ratio that relates price to marginal cost. Its enduring usefulness comes from the ability to describe a price-cost divergence in firms with differing scale and organization. The ratio addresses a decision margin, while a welfare assessment requires the consequences of that margin for quantities and alternatives.
Harberger (1954) turned the allocative problem into a question of empirical magnitude in Monopoly and Resource Allocation. In the calculation of forgone surplus, the dimensions of a price wedge and the missing quantity matter jointly. Measuring industrial concentration or elevated earnings alone cannot establish the missing area. The historical paper is an illustration of the research design and of the assumptions facing aggregate calculations; its tentative U.S. manufacturing results are neither universal constants nor current estimates.
Modern industrial organization widened the object of inquiry. Tirole's (1988) Theory of Industrial Organization places monopoly pricing alongside discrimination across consumers and time, quality, reputation, multiproduct decisions, and vertical restrictions. This wider organization clarifies the limits of the isolated triangle. A firm can restrict supply by deciding which products to introduce, choosing their quality, changing contractual terms, or shaping another firm's incentives.
The development of structural demand estimation also changed what an empirical study can test. Instead of relying exclusively on observable prices and accounting profits, the economist attempts to recover substitution patterns and marginal costs, solve an explicit benchmark equilibrium, and disclose the dependence of the result on estimated behavior. Historical continuity is strongest at the level of the economic question: which social gains would exist under an alternative feasible market arrangement, and which of those gains are absent in the observed arrangement?
Demand, willingness to pay, and the firm's opportunity set
Let q denote the quantity sold per period, p the net product price per unit, P(q) the inverse demand function, and C(q) the total resource cost of supplying q. The demand function is defined over a stated product quality, location, period, information environment, and set of substitutes. A differentiable inverse demand curve that decreases over the relevant domain represents a sequence of progressively lower marginal willingness to pay. In a simple quasilinear environment, P(q) also measures the marginal gross monetary benefit of consumption. Dollar measures require that the treatment of income and utility be specified.
The single-price firm chooses a point on its demand schedule. Its revenue is R(q) = qP(q). If it chooses price directly, quantity becomes q = D(p), and revenue is pD(p). The two formulations are equivalent when the relevant demand segment is invertible. Neither formulation gives the firm unrestricted control over both variables. A desired increase in price may lose customers, and an increase in quantity may require a lower price for all units.
The exposition in OpenStax (2022), How a Profit-Maximizing Monopoly Chooses Output and Price, starts from this distinction between the market-facing demand of a monopoly and the demand faced by a price-taking seller. The latter considers its own demand horizontal at the market price, although aggregate market demand can slope downward. In economic application the relevant object is the seller's residual demand, incorporating rival supply and the buyer's available substitutes. A monopolist over a narrow product version rarely faces perfectly inelastic demand over all prices.
The boundary conditions matter. Capacity, minimum order quantities, discontinuities of buyer participation, time-specific delivery requirements, and contractual ceilings can prevent the smooth interior optimum. An alleged seller-facing demand curve estimated from historical prices and sales also confounds changes in demand with changes in supply unless the research design provides independent variation. An economic model must describe the feasible set before differentiating the objective.
The uniform-price monopoly optimum
Consider profit π(q) = qP(q) − C(q), with q constrained to a nonnegative feasible interval. Where production occurs and the objective is differentiable, the first-order condition is P(q) + qP′(q) = C′(q). The left-hand side is marginal revenue, MR(q); the right-hand side is marginal cost, MC(q). The second term in marginal revenue reflects the reduction in the incumbent units' price when additional output is sold at a common price.
An interior stationary point requires more than the first derivative. The second derivative of profit is 2P′(q) + qP″(q) − C″(q). Strict negativity around the proposed solution gives a strict local maximum; global concavity over the feasible interval makes the stationary point global. In nonconcave cases one must compare all candidate optima and boundaries, including the option of no production. A capacity-bound optimum can have marginal revenue above marginal cost. At a kink, the appropriate one-sided inequalities replace the equality. In every case, profit at the chosen point must be compared with the correct shutdown payoff.
The fundamental pricing wedge follows at a regular interior solution:
P(q_m) − MC(q_m) = −q_m P′(q_m).
For positive quantity and strictly decreasing inverse demand the right-hand side is positive. Its magnitude depends on demand slope and scale, and it therefore cannot be inferred from the number of suppliers. A high marginal cost may produce a high absolute price and a small price-cost wedge; a very cheap product may have a large percentage markup and a small absolute wedge.
Fixed costs do not appear in the interior first-order condition when they are independent of output conditional on operating. They do enter the decision to operate, the investment decision before costs are sunk, and the design of feasible public alternatives. Confusing the marginal pricing condition with the profitability of the entire enterprise is a recurring reason for false welfare comparisons.
Elasticity, the Lerner index, and the meaning of a markup
Define the positive own-price elasticity ε(p) = −D′(p)p/D(p) for a downward-sloping demand curve at positive prices and quantities. Differentiating revenue with respect to price gives the equivalent condition D(p) + [p − MC(D(p))]D′(p) = 0. Hence the Lerner index at a regular single-product, single-price interior optimum satisfies
L = (p − MC)/p = 1/ε.
For strictly positive marginal cost, an unconstrained interior monopolist therefore operates on an elastic demand segment, ε > 1. Zero marginal cost admits the limiting unit-elastic condition. This result does not state that the demand schedule is everywhere elastic, and it does not apply without modification to capacity limits, interdependent product prices, binding controls or dynamic choice.
The gross markup ratio μ = p/MC is a different statistic. With positive marginal cost, L = 1 − 1/μ and μ = 1/(1 − L). A gross ratio of 1.50 corresponds to price 50 percent above cost and a Lerner index of one-third. Reporting either as a 50 percent Lerner index changes the denominator and the economic meaning.
Lerner's (1934) original monopoly index explicitly treated the price-cost relationship as a measure of exercised power. The index's measurement remains separate from the welfare question: without a quantity response, the inferred reduction in beneficial trades has no magnitude. In data, average variable cost or cost of goods sold is sometimes substituted for marginal cost, but these objects coincide only under additional technological and accounting assumptions. The ratio of operating profit to sales includes fixed operating expenditure, depreciation, overhead and other items that the marginal price condition does not directly include.
For a multiproduct seller the own-price elasticity alone is insufficient. Price changes reallocate consumers across products owned by the same firm. The portfolio margin depends on diversion between them and on their respective margins. An observed single-product Lerner ratio remains a descriptive price-cost ratio when these conditions fail, but interpreting its reciprocal as a structural elasticity then becomes invalid.
The welfare benchmark and the resource account
Under quasilinear preferences and a common monetary numeraire, gross benefit is B(q) = ∫₀^q P(x) dx. If C(q) measures resource cost, total surplus is W(q) = B(q) − C(q). Consumer surplus at the uniform price p is CS(q,p) = B(q) − pq. Producer surplus net of all costs included in C is PS(q,p) = pq − C(q). Adding the accounts cancels the payment pq. When C includes an avoidable fixed cost, this net producer amount is economic profit and should be distinguished from operating surplus before fixed cost.
For a differentiable interior welfare optimum q*, P(q*) = MC(q*). If the marginal willingness-to-pay schedule decreases and marginal resource cost does not decrease, the crossing maximizes W over a convex feasible domain, subject to any fixed-cost participation condition. A competitive allocation with price equal to marginal cost can implement the crossing when the market structure, information, capacity and financing make it feasible. The efficiency benchmark is not a prediction that an unregulated competitive industry will actually exist.
OpenStax (2022) illustrates the idea in Demand, Supply, and Efficiency: suppressed exchanges may destroy gains that could otherwise accrue to both sides. The aggregate monetary sum applies directly in quasilinear settings, or as a maintained approximation under modest income effects. When the income effect is material, welfare needs compensating or equivalent variation based on the appropriate expenditure functions, with careful treatment of income redistribution.
External costs and benefits change the condition. If an extra unit generates a pollution cost, its private MC understates social marginal cost. If it creates a benefit for others, private willingness to pay may understate social benefit. A change that reduces monopoly restriction can then worsen another distortion. The chosen benchmark must therefore specify which costs, values, agents and constraints are included before the welfare area is evaluated.
Deadweight loss as an integral
When q_m < q* and the same technology, quality, participation and social environment apply to both allocations, deadweight loss is the surplus difference:
DWL = W(q*) − W(q_m) = ∫_{q_m}^{q*} [P(x) − MC(x)] dx.
The integral evaluates the net value of the units forgone by the monopoly outcome. If both output levels use the same installed fixed resource, the fixed cost cancels. If one allocation requires construction, exit, or different fixed resources, the full difference in W must be calculated.
For linear inverse demand and constant marginal cost, the integrand decreases linearly from the monopoly wedge to zero at the efficient output. The integral becomes a triangle of height p_m − c and base q* − q_m. Its size is one-half of their product. This familiar special case is particularly useful because price and quantity dimensions are visible.
For nonlinear demand or increasing marginal cost, the area is not generally a geometric triangle. A local approximation to the quantity response can use a demand or supply derivative, but the extrapolation from a small price change to an entire counterfactual interval requires a maintained curvature assumption. The Harberger (1954) resource-allocation calculation illustrates the historical importance of making quantitative approximations explicit.
A welfare loss has monetary units per period. A markup has no units. Multiplying an observed price-cost percentage by revenues measures neither the missing quantity nor its value without an additional model. The resource account must also state whether measured surplus belongs to local residents, global consumers and producers, or a defined national welfare boundary. Payments to foreign owners may matter differently in a national distribution account than in a world total-surplus account.
A reproducible linear example
Take an illustrative inverse demand P(q) = 100 − q, with constant marginal cost c = 20, zero avoidable fixed cost, q ≥ 0 and money units per unit. These numbers are constructed for this study and are not observations about an industry. The competitive or efficient benchmark satisfies 100 − q* = 20, giving q* = 80 and p* = 20. The monopolist solves MR = 100 − 2q = 20, giving q_m = 40 and p_m = 60.
At q* consumer surplus is (100 − 20) × 80 / 2 = 3,200. Producer surplus is zero relative to constant cost. At q_m consumer surplus is (100 − 60) × 40 / 2 = 800; operating producer surplus is (60 − 20) × 40 = 1,600. Total monopoly surplus is 2,400, so the deadweight loss is 800. The price increase transfers 1,600 of surplus from buyers on units still purchased and destroys another 800 of buyer surplus associated with the restricted quantity.
| Allocation | Price | Quantity | Consumer surplus | Operating producer surplus | Total surplus |
|---|---|---|---|---|---|
| Efficient | 20 | 80 | 3,200 | 0 | 3,200 |
| Monopoly | 60 | 40 | 800 | 1,600 | 2,400 |
The correct accounting identity is ΔCS = −2,400, ΔPS = +1,600 and ΔW = −800. It is incorrect to add the consumer loss to the monopoly profit and call the sum lost social wealth. In this example the markup relative to cost is 200 percent, the gross markup ratio is 3 and the Lerner index is 2/3. Elasticity evaluated at the monopoly point is 60/40 = 1.5, verifying the optimum formula.
The example illustrates why market power may be important for distribution even if the deadweight area is smaller than the transfer. The same example also illustrates why the price effect alone cannot be interpreted without the quantity response. All amounts are period totals in the same monetary unit and all allocations use the same product.
General linear comparative statics and sensitivity
Let P(q) = a − bq, with b > 0 and a > c ≥ 0. Assume the entire economically relevant output interval is feasible. Algebra yields q* = (a − c)/b, q_m = (a − c)/(2b), p_m = (a + c)/2, and DWL = (a − c)²/(8b). Gross efficient surplus is (a − c)²/(2b), so under these exact assumptions the monopoly destroys one-quarter of the attainable surplus. The one-quarter fraction follows from the linear demand and constant-cost example; it is not a general monopoly parameter.
Holding the choke price and marginal cost fixed while doubling b halves both quantities and the absolute surplus loss. Holding the slope fixed while increasing a − c expands the loss quadratically. A price-cost wedge alone cannot establish the slope or the scale. For a given observed p_m and c, two different values of b can produce the same markup and different quantities, transmissions of price changes and absolute welfare losses.
The numerical example can be expressed as a sensitivity table with a = 100 and c = 20:
| Inverse demand slope b | Efficient quantity | Monopoly quantity | Monopoly price | Deadweight loss |
|---|---|---|---|---|
| 0.5 | 160 | 80 | 60 | 1,600 |
| 1 | 80 | 40 | 60 | 800 |
| 2 | 40 | 20 | 60 | 400 |
These results deliberately hold the intercept fixed. A study that instead holds observed volume fixed while varying elasticity changes the calibration problem and must re-estimate the intercept or curvature. Comparative statics describe a specified family of alternatives, not an unrestricted empirical inference. The arithmetic provides a transparent check on the dimensions of the welfare measure.
Curvature, pass-through, and the danger of extrapolation
Demand curvature affects both optimal pricing and the response to marginal cost. For a twice-differentiable direct demand D(p), constant marginal cost c and an interior optimum, differentiate D + (p − c)D′ = 0 with respect to c. The derivative of the chosen price is dp/dc = D′/[2D′ + (p − c)D″], provided the denominator is nonzero and the optimum remains regular. When direct demand is linear, this derivative equals one-half. The expression differs under another demand shape, even if the original elasticity is identical.
For isoelastic demand D(p) = A p^(−ε), ε > 1, the optimum is p_m = cε/(ε − 1) for c > 0, assuming constant marginal cost and no capacity bound. Marginal-cost pass-through equals ε/(ε − 1), above one. An illustrative D(p) = 1,600/p² with c = 20 gives p_m = 40 and q_m = 1, while the efficient price of 20 gives q* = 4. The inverse demand is 40/√q. Integrating the net benefit over q from 1 to 4 produces DWL = 80(2 − 1) − 20(4 − 1) = 20. The formula is valid over that interval and the result is not comparable in scale to the linear example without calibrating the units and baseline demand.
Pass-through is a local comparative-static property, not a direct estimate of monopoly profit. It can be altered by changing marginal cost, cost components that vary with output, strategic interaction, capacity or vertical contracting. A tax's statutory payer is also distinct from its economic incidence. In evaluating a tax, payments from firms or consumers to government are included as government receipts in a transfer-neutral welfare measure, with any administrative or externality effects accounted for separately.
The practical implication is that a demand system fitted accurately around one observed price may yield poorly identified welfare consequences if the proposed counterfactual lies outside the data's support. Shape restrictions, alternative functional forms and transparent uncertainty are economically substantive safeguards.
Fixed costs, entry, and natural monopoly
Consider C(q) = F + cq for q > 0 and C(0) = 0, where F is an avoidable cost of operating the network. Average cost falls toward c with scale. Splitting output across two suppliers would duplicate F when the facilities are substitutable, giving the single producer a cost advantage. More generally, natural monopoly concerns subadditivity of the cost function over the relevant output range: the cost of one supplier's aggregate capacity is no greater than the summed cost of a feasible partition. Falling average cost is a useful single-product illustration rather than a universal test.
The marginal-cost allocation may generate an operating deficit. Revisit P(q) = 100 − q and c = 20 with F = 1,200. Efficient gross surplus remains 3,200, and welfare after F is 2,000. Price equal to c yields zero revenue above variable costs and leaves F unfunded. The monopoly quantity 40 yields an operating surplus of 1,600 and net profit of 400. Welfare after F is 1,200. A uniform break-even price satisfies (p − 20)q = 1,200 and p = 100 − q. The roots are q = 20 and q = 60; the larger quantity yields p = 40 and welfare 1,800. Its loss relative to funded marginal-cost pricing is 200.
The high-volume break-even allocation has higher welfare than the lower-volume root. This illustrates a distinction between an economically feasible second-best tariff and an unconstrained marginal-cost counterfactual. If funds can be raised through a subsidy, the administrative and distortionary costs of that funding source must be included. If entry would duplicate fixed facilities, the resource cost of additional facilities may outweigh competitive pricing benefits. A natural monopoly analysis thus compares feasible cost recovery institutions and network configurations, rather than presuming that marginal-cost prices and many suppliers are jointly attainable.
Cost recovery is not a blanket explanation for any observed price. The magnitude of F, the capacity costs attributable to present and future users, the opportunity cost of capital, and the appropriate quality commitments remain matters for evidence.
Price discrimination and the sources of feasibility
Price discrimination occurs when a seller conditions prices or contractual terms on buyer characteristics, usage or purchase conditions in a way that cannot be fully explained by corresponding differences in costs. A price difference alone is insufficient: overnight delivery, differentiated quality, taxes, payment risk and service obligations can change the resource cost and the product purchased.
Varian (1989), in the handbook chapter Price Discrimination, organizes discriminatory pricing as a family of practices with differing welfare consequences. The feasibility conditions are a seller's pricing discretion, heterogeneity in willingness to pay or purchase behavior, an ability to segment or screen customers, and constraints on resale or imitation. These conditions may arise from identity verification, product versioning, nontransferable contracts, timing, or the nature of a service.
Perfect or first-degree discrimination is the limiting case where the seller extracts each buyer's available surplus by charging the buyer's marginal willingness to pay on successive units, assuming the seller observes it and can make enforceable individualized offers. Second-degree discrimination relies on self-selection across a menu of quantities, qualities, usage thresholds or fixed fees. Third-degree discrimination assigns observable groups different prices. Actual pricing systems may combine mechanisms, as when an airline offers fares conditional on cabin, booking date, refundability and passenger status.
This classification is about information and constraints rather than simply about the number of posted prices. A single tariff with a fixed fee plus marginal usage charge can allocate surplus differently from a schedule of personalized per-unit prices. A uniform price applied to quality variants may, in turn, mask meaningful variation in the margin. Relevant evidence includes the effective paid price and the conditions under which it is available, not only a public price list. Welfare comparison then requires specifying which buyers can purchase what quantity and at which resource cost under each mechanism.
Perfect discrimination, access, and distribution
Under complete information, contract enforcement, quasilinear willingness to pay and no incremental discrimination cost, a monopolist can in principle appropriate all available gains by charging each marginal unit's willingness to pay. Its gross revenue as a function of output becomes B(q) instead of qP(q). Profit is B(q) − C(q); hence the welfare maximizing output also maximizes profit, and the static allocative inefficiency of uniform pricing disappears.
For the linear example with a = 100, b = 1 and c = 20, the allocation expands to q = 80, total surplus equals 3,200, and the seller obtains the full 3,200 before fixed costs when every buyer's net surplus is extracted. Consumers collectively receive no net monetary surplus in this idealization. The uniform-price monopoly gives consumers 800. Thus the discriminatory allocation increases total surplus by 800 while reducing consumer surplus by 800. The change in producer surplus is +1,600. A policy objective maximizing unweighted total surplus and an objective emphasizing consumers' welfare can rank the two regimes differently.
Perfect discrimination does not automatically deliver efficient outcomes when willingness to pay differs from social benefit, when liquidity limits exclude otherwise valuable uses, or when collection and use of information entail privacy or administration costs. If the seller faces uncertainty about individual values, an offer that extracts all expected willingness to pay may violate some buyers' participation constraints. Resale, multiperson bargaining, repeat relationships and ethical restrictions further alter feasibility.
The idealization is useful because it identifies the source of the uniform-price inefficiency. With a common price, the seller sacrifices revenue on earlier units when offering a lower price to expand sales. Individualized charges can remove that trade-off. The distributional consequences and information costs remain separate components of a complete welfare judgment. Calling the outcome efficient answers a conditional allocation question; it does not establish an equitable distribution or desirable institutional practice.
Second-degree discrimination and two-part tariffs
A two-part tariff charges a fixed access fee A plus a unit price t. If every buyer has identical demand, quasilinear preferences and an enforceable participation choice, usage can be set at t = c. The access fee can then extract the consumer surplus from marginal-cost usage while preserving the efficient quantity for every participating buyer. A common fixed fee also funds fixed resources. The binding constraint is the buyer's participation condition: gross consumer gain from access must at least cover A.
Heterogeneity changes the result. Suppose one buyer's inverse demand is 100 − q and another's is 60 − q, with marginal cost 20. At t = 20 their consumer surpluses before access charges are 3,200 and 800. A fee of 800 can admit both while leaving the high-demand buyer 2,400. A fee of 3,200 admits the high-demand buyer only, subject to the usual convention about indifference. The firm compares access revenue and usage margins along with the cost of admitting each type. A high fixed price can exclude a group despite efficient usage prices for those who remain.
More general menus use incentive compatibility. A high-value buyer must prefer the contract intended for that buyer to one intended for a low-value buyer; each type must prefer participation to the outside option. When types are private information, the firm may deliberately restrict a low-type quantity or lower quality to preserve the separation of types. The efficiency loss then appears at both the participation and conditional consumption margins. The high type may retain informational rent that perfect discrimination would eliminate.
Observed subscription fees, software plans and utility tariffs require separate estimates of users who do not enter, users who enter but consume too little, and any resource cost of constructing the menu. The average payment per unit does not reveal the marginal usage incentive. A welfare analysis must represent the full contract.
Third-degree discrimination: segment pricing
Suppose groups i are economically independent and the seller can prevent arbitrage. Group i has inverse demand P_i(q_i), and total cost depends on Q = Σ_i q_i. Profit is Σ_i q_i P_i(q_i) − C(Q). At a regular interior optimum, MR_i(q_i) = C′(Q) for every active group. Under constant common marginal cost c, each price solves (p_i − c)/p_i = 1/ε_i. A less elastic group at its optimum has the larger proportional wedge.
The qualification concerning evaluated elasticities is important: elasticities depend on prices. An elasticity measured at a common historical price cannot simply be ranked and plugged into a discriminatory-price equation without solving both optimal prices. Costs may also differ between groups, and reciprocal elasticity comparisons then require the appropriate group-specific cost. Where resale is possible, a low-price group's product may reappear in a high-price market, collapsing the assumed separation.
Profit under unrestricted discrimination is at least as high as the optimal uniform-price profit when uniform prices remain feasible and the added pricing capability has no costs. The welfare result differs because discrimination can expand total sales and redistribute units across groups while changing the payments. In Price Discrimination and Social Welfare, Varian (1985) establishes conditions and bounds for this comparison; an expansion of aggregate output is necessary for a welfare improvement under the relevant surplus assumptions but is not generally sufficient.
To compare allocations in independent groups, calculate W = Σ_i ∫₀^q_i P_i(x) dx − C(Q). A common price equalizes group marginal willingness to pay among buyers served at the margin. Discriminatory prices can break this equalization. The resulting allocation may waste potential gains even when total output stays constant. A credible analysis must simultaneously identify aggregate output changes and how those units are distributed across groups.
A closed-market comparison with unchanged output
Consider two independent groups with inverse demands P_A = 100 − q_A and P_B = 60 − q_B; marginal cost is 20. At a common price below 60, total demand is Q = 160 − 2p. Profit is (p − 20)(160 − 2p), maximized at p = 50. Prices at or above 60 serve only A; their maximum profit is 1,600 at p = 60, below the common active-group optimum of 1,800. The global uniform solution is q_A = 50, q_B = 10.
Separate pricing yields p_A = 60, q_A = 40 and p_B = 40, q_B = 20. Both regimes have Q = 60. Under the common price, consumer surplus is 1,250 in A and 50 in B; total consumer surplus equals 1,300. Profit is 1,800, so total surplus is 3,100. Under discrimination, respective consumer surpluses become 800 and 200, while profits become 1,600 and 400; total surplus equals 3,000.
| Regime | Group A q | Group B q | Total q | Consumer surplus | Profit | Total surplus |
|---|---|---|---|---|---|---|
| Uniform | 50 | 10 | 60 | 1,300 | 1,800 | 3,100 |
| Discrimination | 40 | 20 | 60 | 1,000 | 2,000 | 3,000 |
| Efficient | 80 | 40 | 120 | 4,000 | 0 | 4,000 |
The efficient allocation prices both segments at cost. Welfare loss rises from 900 to 1,000 under discrimination. The result is driven by moving ten units from group A, where the foregone units have relatively high marginal values, to group B, where they have lower marginal values. Integrating the difference in marginal values over the transfer yields 100. Common pricing did not eliminate the total output distortion; it did allocate the existing 60 units efficiently across these two groups conditional on their total. This internally derived example is a concrete illustration of the distinction between output and allocation effects in Varian (1985).
Market opening and an opposite welfare result
Change only the smaller group's inverse demand to P_B = 35 − q_B, retaining P_A = 100 − q_A and c = 20. A common price above 35 serves A alone, with optimum p = 60 and q_A = 40. To serve both, a common price must be at or below 35; the profit function on this interval is (p − 20)(135 − 2p). Its unrestricted stationary price is 43.75, outside the active-group interval. At p = 35 the profit is 975, below the A-only optimum of 1,600. Lower prices cannot do better in the relevant interval. Thus B is excluded under uniform pricing.
With segmentation, A retains p_A = 60 and q_A = 40. B obtains p_B = 27.50, q_B = 7.50, a consumer surplus of 28.125 and profit of 56.25. Total welfare increases by 84.375 relative to the uniform-price outcome. Group A's allocation and surplus remain unchanged, while B gains access. The efficient quantity in B would be 15. Segmentation therefore improves welfare without eliminating B's residual inefficiency.
The paired examples isolate the analytical margins. When both groups already participate, discrimination can move a fixed total quantity toward a less valuable assignment. When a previously excluded group enters without withdrawing units from the incumbent group, its new gains can increase welfare. Whether an actual market belongs to either case depends on demands, costs, arbitrage restrictions and the feasible tariff options. The examples are internally consistent counterfactual constructions, not empirical estimates of airline, pharmaceutical or electricity pricing.
An intervention preventing discriminatory prices may protect a high-price group at the expense of a low-demand group's access, or it may remove a distortion in the allocation of a given output. A study should disclose which alternative restrictions it is imposing and solve the firm's actual feasible uniform-price problem, including the possibility of excluding one or more markets. Comparing a discriminatory optimum to an arbitrary average tariff produces a misleading benchmark.
Capacity, timing, inventories, and durable goods
Monopoly pricing unfolds across time when capacity is scarce, inventories expire or today's purchase affects tomorrow's opportunities. A seat on a departing aircraft has an opportunity cost that depends on remaining inventory, uncertainty and the possibility of selling to future passengers. A storage asset, by comparison, can transfer supply between periods with losses and costs. A durable product reduces future demand when sold today and may generate a secondhand market.
In a two-period allocation, the firm can maximize first-period profit plus the discounted expected profit of period two. If q₁ influences the demand or state variables in period two, the marginal condition in period one includes the derivative of continuation value. A current price-cost wedge may reflect the scarcity value of preserving stock or the anticipated consequence of changing the installed base. A static accounting of production cost alone can overstate the distortion at that decision margin. Social opportunity cost of capacity or inventory is likewise relevant in a welfare calculation.
Consider one reusable unit of capacity that can be allocated to one of two delivery times. With a known net resource benefit of 30 now and 50 tomorrow, preserving capacity is efficient when the discounted later value exceeds 30. If both values refer to per-unit willingness to pay minus operating costs, this comparison is distinct from the firm maximizing revenue given a common price schedule. The state variable and feasible intertemporal choices determine the appropriate comparison.
Durable-goods buyers may anticipate future price reductions if the seller cannot commit to future prices. The seller's ability to sustain a current markup then depends on strategic timing, product durability, resale and commitment. Even a hypothetical monopolist with complete control of the initial new-product supply may face competition from past output. These considerations are part of the broader intertemporal monopoly analysis described in Tirole (1988), The Theory of Industrial Organization. A static welfare triangle should be labeled as a period-specific partial-equilibrium result when the dynamic state has been held fixed.
Multiproduct pricing and connected markets
Let a seller offer J products with demands Q_j(p_1,…,p_J). For simplicity suppose constant product-specific marginal costs c_j and no production interactions. The derivative of portfolio profit with respect to p_k is Q_k + Σ_j (p_j − c_j)∂Q_j/∂p_k. In an unconstrained optimum it is zero. This equation includes the firm's own-product effect and the changes in profit on all other owned products. With substitutes, raising p_k can divert sales to another owned product. With complements, raising p_k can reduce demand for another source of profit.
One consequence is that a product-by-product price-cost wedge may substantially misdescribe the enterprise's pricing choice. A platform may subsidize a user-facing service to increase the value of an advertising service; a device seller may price hardware low and recover revenue through compatible services. Even where each product's marginal cost is known, the relevant welfare problem must include the connected demands, costs and cross-market externalities. An observed zero price does not establish absence of market power.
Berry, Levinsohn and Pakes (1995), in Automobile Prices in Market Equilibrium, provide a structural framework for differentiated product markets where own- and cross-price responses, firm ownership and marginal costs together determine equilibrium. This is particularly relevant when a monopolist is studied within a system of products rather than in an isolated market.
For an upstream and a downstream monopolist selling a single product, each successive markup can create a cumulative quantity distortion. In a simple linear example with upstream cost c, downstream demand a − bq and no downstream incremental cost, the integrated monopoly sets q = (a − c)/(2b). If the upstream firm sets a wholesale price w and the downstream firm independently chooses its retail markup, downstream output is (a − w)/(2b); upstream optimization gives w = (a + c)/2, so final output is (a − c)/(4b). Integration can increase output in this specific double-marginalization setting. Other vertical effects, including exclusion and quality incentives, require separate evidence and can reverse the net benefit.
Information, personalization, and data costs
A seller's information about a customer's willingness to pay can refine the feasible pricing mechanism. Information may reveal demand differences, reduce the risk of an offer being rejected, or identify an excluded lower-demand population. It can also permit the seller to extract surplus previously left with buyers. The economically relevant distinction is between information that changes the allocation and information that changes the division of gains.
Suppose a seller previously identifies groups only by a coarse observable characteristic. Better data allow it to separate two types. In the absence of information acquisition and implementation costs, the enlarged feasible set cannot lower the seller's optimum profit because the previous strategy remains available. The welfare consequence does not inherit that monotonicity. If information leads to price reductions for a previously excluded class, output may expand. If personalization increases payments on already-served units without changing resource use, the primary effect is distributional. If surveillance, computing, mistakes or privacy exposure generate resource or utility costs, they enter the welfare account separately.
Personalized posted prices are also distinct from personalized discounts administered through a common tariff. A researcher needs the exact offer set visible to each buyer, realized payments, eligibility, transaction quality and information about alternative offers. Comparing final prices alone may capture differences in timing or negotiated terms rather than discrimination caused by data.
An empirically credible study can exploit randomly assigned eligibility, experimentally varied prices or plausibly exogenous changes in the pricing rule, while protecting the subjects' privacy. It needs to measure abandonment and entry as well as purchases. Observation of completed transactions systematically omits the people excluded by the reservation price, precisely the population central to an output-based welfare analysis. The present dossier provides the comparative framework; it does not claim original access to individual pricing logs or experimental data.
Market power measurement and identification
The empirical issue is to distinguish the curve governing a seller's pricing decision from the realized price and sales points that result from simultaneous market determination. A high price paired with a high quantity may indicate an outward demand shift even if the underlying demand curve slopes downward. A price regression that treats observed price as exogenous can therefore misestimate willingness to pay and its slope.
Identification requires variation that affects the relevant economic object. An instrument for price must change it while remaining independent of the unobserved demand component after the maintained conditioning and equilibrium assumptions. A candidate input-price shock is useful only if it does not independently change buyer valuations, quality or product availability in a way excluded from the model. Natural experiments have analogous exclusion and parallel-trend requirements. Randomized prices can identify local demand responses if assignment is valid and interference, learning and stock constraints are accounted for.
Structural estimation takes the demand model, cost or production model and assumptions on firm conduct into an explicit equilibrium system. Berry, Levinsohn and Pakes (1995), Automobile Prices in Market Equilibrium, use product-level and aggregate consumer information to estimate differentiated-product demand and cost parameters under an oligopoly framework. A structural welfare calculation changes specified primitives or institutions, resolves the equilibrium, and compares predicted allocations.
An estimated demand elasticity can be well identified locally and poorly identified beyond the support of the observed variation. A counterfactual that substantially lowers price can introduce new buyers, new rival products, and new quality decisions. Researchers should report sensitivity to the functional form and the price range used for identification. The precision of a coefficient or standard error is not a measure of confidence in a remotely extrapolated market structure. Neither revealed preference nor a structural model by itself guarantees that the observed choices fully reveal welfare when information or behavioral constraints matter.
Estimating markups from production data
A production-based method can recover markups from a maintained model of cost minimization. Let a firm use a flexible input v at unit price w, producing y. The first-order condition for the flexible input in a frictionless interior cost-minimization problem is w = MC × ∂y/∂v. Multiply by v and divide by py to obtain the revenue-share identity s_v = wv/(py) = (MC/p) × θ_v, where θ_v = (∂y/∂v)(v/y) is the physical output elasticity of that input. The implied gross markup is μ = θ_v/s_v.
The identity needs the correct output elasticity, a variable input optimized at the relevant time, consistent measures of input expenditure and revenue, and an appropriate treatment of multi-product production, input price power and adjustment costs. A flexible-input assumption can fail under contracts, inventory constraints, procurement frictions or labor adjustment. If output is measured by revenue rather than physical units, unobserved price heterogeneity and endogenous demand can affect estimated elasticities.
De Loecker, Eeckhout and Unger (2020), The Rise of Market Power, report estimates of U.S. aggregate markups that moved from 1.21 in 1980 to 1.61 near their study's endpoint, with growth concentrated among higher-markup firms and a broadly stable median. These are the authors' reported historical model-based estimates. In Lerner units, the corresponding aggregate ratios would transform to approximately 0.174 and 0.379 if interpreted as ratios for a representative observation. Applying the nonlinear conversion to an aggregate statistic does not produce the average of firm-level Lerner indexes.
This distinction extends to welfare. Firm aggregation depends on weighting and product coverage. A change in sales-weighted markups can reflect changes in within-firm margins, shifts of sales toward high-markup firms, or measurement choices. Without corresponding demand responses and feasible counterfactual costs, the aggregate markup series cannot be converted into a national deadweight-loss estimate. The scientific value of the measure is greatest when kept tied to its identifying assumptions and compared with independent market-specific evidence.
Welfare measurement, income effects, and social weights
A total-surplus measure places one monetary unit of consumer gain and one monetary unit of producer gain on the same footing. Quasilinear preferences make this accounting particularly transparent. A common version of the model assumes utility u(x,m)=v(x)+m, where m is consumption of a composite numeraire. At a fixed price, utility maximization implies a marginal valuation schedule for x with a constant marginal utility of money. In such a setting the area under inverse demand has an exact monetary interpretation.
When demand exhibits material income effects, an uncompensated demand integral generally differs from compensating variation or equivalent variation. Let e(p,u) be the minimum expenditure required to attain utility u at price vector p. Comparing e(p_1,u_0) with e(p_0,u_0) quantifies the additional income needed at the new prices to maintain reference utility u_0. Comparing expenditure changes at the new utility level gives a different welfare measure. The sign convention must specify whether a positive number denotes a loss or a payment needed to offset it.
Distributional evaluation can assign social weights by income, consumer group, region or another publicly justified criterion. A transfer from a lower-income consumer to a higher-income shareholder may have social significance even when the sum of unweighted dollar surplus is unchanged. Conversely, losses and gains should not be silently counted twice by combining a weighted transfer account with an unweighted aggregate that already includes one side. The distribution of ownership, exposure to price increases and eligibility for lower prices are empirical inputs.
For internationally traded goods, the welfare boundary affects the treatment of profit transfers and tax revenue. Global analysis counts all parties under its chosen weights; a national analysis may separate domestic from foreign beneficiaries. For research reporting, presenting unweighted efficiency, consumer and producer incidence, and any alternative weighted calculation as distinct outputs maintains comparability. The choice of social objective must remain visible rather than hidden in a price-cost statistic.
Rent acquisition and costs beyond the conventional triangle
The static triangle asks how much net benefit is lost from a restricted quantity when cost and product characteristics are otherwise held fixed. A durable monopoly position can induce additional resource use. Firms may compete for exclusive rights, expend effort influencing regulation, defend barriers, or use resources to preserve a position. Posner (1974), in the working paper Social Costs of Monopoly and Regulation later published in 1975, examines the opportunity costs of seeking monopoly positions in addition to conventional allocation losses.
These resource costs must be distinguished from transfers. A fee paid to a government for a license reallocates money; the personnel, legal work, administrative delay or resources spent to secure the license may use scarce resources. An economic estimate should count the opportunity cost associated with wasteful activity, while preserving productive outputs and legitimate public administration where present. Research spending can generate knowledge even when its private reward is a monopoly margin. Treating all profit or all legal and commercial spending as wasted resources would confuse the mechanisms.
The full comparison can be represented by the differences between actual and counterfactual gross benefits, production costs, investment costs, costs of acquiring or protecting exclusive positions, external costs and other effects. Each component must use a coherent counterfactual. Costs already included in an estimated total cost function or measured resource account cannot be added again as separate rent-acquisition losses.
The distinction helps explain why assessments under different research designs differ. A narrow partial-equilibrium exercise may report a small missing-transaction area while another study finds substantial changes in productive efficiency, entry or innovation. These are compatible observations if they measure different margins. Combining them requires establishing that the margins do not overlap and that the benchmark changes can be jointly implemented.
Innovation, intellectual property, and dynamic welfare
Some exclusivity regimes create an intertemporal bargain: a firm invests resources before knowing whether a valuable product can be developed and later receives a limited opportunity to earn returns. The relevant welfare calculation includes the research cost, probability and timing of success, value of new products, dissemination, and the quantity restrictions in periods of exclusivity. It is a comparison of feasible institutional systems, not a comparison between the current monopoly price and a hypothetical free supply obtained without funding development.
Take two candidate regimes. One induces development of a product at a present resource cost K; another permits immediate price competition but changes the probability of development. The discounted expected total benefit must incorporate the product's availability probability under each regime, compare any difference in resource costs and include period-specific surplus. Positive realized patent profit is a transfer in a given period, but anticipated profit can affect whether resources are committed earlier. Distinguishing the incentive effect from the ex post distribution prevents inconsistency across the time horizon.
The dynamic effect of stronger market power has no universal sign. It can fund research or investment in particular conditions, and it can also reduce the incentive to improve a product, protect an obsolete position, raise downstream innovators' input costs, or delay entry. Productive investment and exclusionary expenditures must be separated with evidence. Policy options include changes to exclusivity duration, prize funding, procurement, licensing arrangements and other mechanisms, each with administration, selection and financing constraints.
A defensible study identifies what is known about research or entry responses rather than counting an asserted innovation benefit as a fixed offset to every measured monopoly loss. For pharmaceutical markets, in particular, demand reflects health conditions, prescribing, insurance reimbursement, available substitutes and affordability; the price effect of a patent's expiration cannot by itself identify the welfare effect of the patent system over the preceding development period.
Empirical case: California wholesale electricity
Wholesale electricity supplies a useful distinction between payment effects and resource costs because capacity scarcity, demand responsiveness, input prices and dispatch decisions affect market outcomes simultaneously. During the period June 1998 to October 2000, Borenstein, Bushnell and Wolak (2002) estimated the composition of California's wholesale electricity payments in Measuring Market Inefficiencies. They distinguished production costs, competitive rents associated with scarcity and payments attributed to the exercise of market power, while separately considering inefficient increases in variable production costs.
Their repository abstract reports summer expenditures of $2.04 billion in 1999 and $8.98 billion in 2000. The increase was $6.94 billion. The authors attribute 21 percent of that increase to higher production costs, 20 percent to higher competitive rents and 59 percent to market power. Applying those reported percentages gives approximately $1.46 billion, $1.39 billion and $4.09 billion respectively, subject to rounding. These amounts describe the authors' decomposition of the increase in expenditure, not a deadweight-loss estimate of $4.09 billion.
This distinction is decisive. Competitive scarcity rents can arise when scarce capacity has a high opportunity cost; payments above variable production cost do not by themselves exhibit monopoly conduct. Where electricity demand is highly inflexible in the short run, an increase in price can transfer very large amounts while producing a relatively small reduction in measured consumption. Inefficient dispatch, outages, rationing and delayed investment can create other resource or welfare losses. The counterfactual must represent plant availability, fuel input cost, transmission constraints, demand and operating rules.
The paper's period and market-specific structure preclude turning its decomposition into a contemporary estimate for California or an internationally general rule. Its transferable contribution is methodological: estimate a competitive resource-cost and scarcity benchmark conditioned on the system's constraints before assigning the residual payment component to market power. Then measure the net gains forgone, rather than interpreting transfers as resource destruction.
Empirical case: pharmaceutical entry and prices
The U.S. Food and Drug Administration's Generic Competition and Drug Prices research collection describes analyses of prices and savings following generic entry, including a November 2025 study concerning 773 generic drug applications fully approved in 2023. The agency finds evidence consistent with lower generic prices as competitor entry increases. Regulatory approvals are not identical to active marketed competitors, and a single application count cannot establish the number of independent economic suppliers or the intensity of competition for a specific molecule.
The case connects entry and the seller's residual demand. Patent protection, the availability of chemically or therapeutically substitutable products, insurance formularies, prescription behavior, wholesale and pharmacy contracting, and patient cost-sharing determine substitution. Generic entry can reduce the price charged on previously purchased quantities and permit additional treatment. The first effect is principally a distributional transfer between payer and seller when resource costs remain unchanged; the second may generate extra net health value. Reduced production costs or more expensive substitution can change the resource account as well.
A causal study of entry should establish the timing of approval and market availability, account for anticipatory pricing, separate generic and originator products, and address selection in which products attract entrants. Observed manufacturer sales revenue is not a patient out-of-pocket price; payer expenditure is not necessarily the net price received by the manufacturer; and discounts, rebates and intermediary fees may make list and net prices diverge.
The FDA collection explicitly discusses savings, while broader supply reliability and manufacturing policy are distinct questions. The appropriate health-economic welfare benchmark can include patient benefits beyond measured willingness to pay, medical expenditures and external health effects. The dossier relies on the dated institutional findings as evidence of the price-competition mechanism. It does not re-estimate the FDA's savings calculations, drug-level causal effects or clinical welfare.
International pharmaceutical comparisons
Cross-national prices depend on market arrangement, the composition of products, measurement conventions and policy institutions. Danzon and Chao (2000), Prices, Competition and Regulation in Pharmaceuticals, compare manufacturer-level drug prices in the United Kingdom, Canada, France, Germany, Italy, Japan and the United States using matched 1992 outpatient molecules and quantity-weighted price indexes. The authors' historical analysis directly illustrates how product matching, generic inclusion and weights affect comparative price conclusions.
Their published monograph description reports U.S.-weighted manufacturer-price indexes relative to the United States of −17 percent for the United Kingdom, +2 percent for Canada, +25 percent for Germany, −32 percent for France, −13 percent for Italy, and −12 percent for Japan. These values are relative indexes for the 1992 sample, using the reported weighting convention. They are not contemporary prices, comparable patient payments, or estimates of the absolute efficiency of the national systems.
A domestic-versus-foreign price difference can reflect a change in distribution margins, taxes, therapeutic mix, reimbursement rules, exchange rates, negotiated prices, volumes and product availability. A manufacturer-level comparison across matched molecules controls some composition differences but does not establish identical effective treatment options or an identical marginal cost in each jurisdiction. Because policy regimes are chosen within differing health systems, a cross-section alone has weak causal leverage on what would happen if one country transplanted another's price policy.
The welfare lesson derives from the nature of the index rather than from country rankings. Report the price concept, exact reference period, weights, product definitions, currency conversion and coverage. Then separately identify the quantities and treatment outcomes that a changed market rule would produce. A comparable price statistic is a necessary input for some welfare studies, not a complete comparative welfare result.
Empirical case: differentiated automobiles
The automobile market illustrates market power when buyers choose among products with many attributes and when firms sell portfolios. Berry, Levinsohn and Pakes (1995), Automobile Prices in Market Equilibrium, develop a framework for estimating consumer preferences and cost parameters using widely available product-level and aggregate consumer information in a structural model. Their application addresses the historical U.S. automobile market over approximately two decades of model offerings.
A car's posted price does not identify the buyer's valuation because its attributes, brand, perceived quality and choices among models matter. The availability of an alternative vehicle creates a substitution pattern. A single-product elasticity reports the demand response when that price changes and the other prices are fixed. A firm setting a portfolio of prices considers the destination of diverted buyers and their contribution margins. Recovering these patterns is central to the welfare effect of changed ownership or pricing conduct.
A counterfactual that imposes cost pricing on every product is not the same as a counterfactual introducing entry, changing dealer terms, or separating common ownership. Fixed product development costs, model introduction, dealer services and capacity may change under these alternatives. In the short run one can maintain the actual model set and product characteristics; in the long run that restriction needs explicit justification or relaxation.
The dossier does not recreate the underlying estimation data or parameters. The study gives a verified methodological example showing why demand, cost and equilibrium assumptions must be pursued together. Its findings on a particular market and observation period cannot supply a numerical monopoly-loss coefficient for other industries. A robust modern application would disclose ownership matrices, the price measure, excluded instruments, the sample, elasticity distribution and the uncertainty surrounding the simulated equilibrium.
Network regulation and the institutional benchmark
UK fixed telecommunications illustrates the challenge of shared infrastructure and downstream competition. Ofcom's historical Wholesale Fixed Telecoms Market Review 2021–26 describes a regulatory period from April 2021 to March 2026 designed to promote network investment and competition. The official statement page also describes a notice requirement for certain changes in wholesale access tariffs with conditional volumes or service ranges, intended to address concerns about the effects of such pricing on alternative network construction.
A vertically organized network provides a platform for downstream services. If the network is difficult to duplicate, wholesale access conditions affect retailers' marginal costs and ability to compete. The same wholesale charge also contributes to infrastructure funding and investment incentives. Cost-based access charges, price caps, quality standards, obligations to provide access, or reliance on competing facilities each imply a different allocation of risks and rewards. A theoretical monopoly triangle measured at the retail stage alone can omit network investment and access.
A network welfare comparison must consider the cost of maintaining installed infrastructure, the marginal cost of incremental traffic, congestion, reliability, coverage, incentives for modernization, and administration and monitoring. The price of service may include fixed access and variable usage charges. Lowering one component while allowing another to rise can change subscription and use differently. The relevant consumer decision may be adoption of a connection, the speed or reliability tier selected, or usage conditional on access.
This case serves as a historically documented institutional illustration rather than a statement about UK rules after March 2026. Current price controls and legal obligations would require their own effective-date verification. The central economic point survives the change of regime: use a feasible, financed network and service-quality counterfactual when describing the welfare effects of monopoly prices.
Competition policy, excess prices, and feasible remedies
The economic concept of monopoly power operates at a different level from a legal determination of liability. A positive Lerner index can characterize many sellers whose products are differentiated, whose business has substantial fixed costs, or whose conduct is otherwise lawful. Competition authorities evaluate specified legal questions and evidence under the governing statute and procedural rules, not a universal positive-markup prohibition.
The U.S. Department of Justice and Federal Trade Commission issued the joint 2023 Merger Guidelines on December 18, 2023. The document's existence identifies an institutional approach to prospective competition assessment, which has different evidentiary and legal purposes from a retrospective measurement of a monopoly wedge. In Europe the Official Journal notice of the 1978 United Brands judgment documents a historically important competition proceeding involving banana markets; the present work uses only its verified case identity and date rather than attributing a specific legal test on the basis of the notice.
An economic assessment of a remedy must specify the mechanism producing the observed distortion and the alternative the remedy can realistically achieve. Entry remedies target the constraints on alternative supply. Access regulation changes the conditions under which rivals or buyers can use a facility. Price caps limit the feasible tariff. Conduct remedies alter practices such as exclusionary contracting. Subsidized marginal-cost pricing addresses access and funding together. Acquisition or divestiture changes the ownership matrix and may affect production costs as well as incentives.
A binding ceiling below the unregulated price can increase traded output in a simple constant-cost monopoly model when it remains above cost and the seller willingly supplies the resulting demand. A ceiling below incremental cost may instead cause shortage or exit; a ceiling below average cost can make long-run operation unviable. The result depends on the shape of cost, obligation to serve, capacity and the enforceability of quality. An analyst should estimate the feasible allocation under the policy rather than assume that it produces the unconstrained marginal-cost benchmark. This scientific analysis supplies an economic method, not legal or investment advice.
Evidence design and a reproducible welfare study
A market-specific investigation begins with a written estimand. For example: the annual change in unweighted domestic consumer plus producer surplus under a specified lower-price tariff, holding product attributes, plant capacity and entry fixed. This estimand differs from the discounted global welfare effect of abolishing an entry barrier. The study must state the counterfactual, the parties counted, the horizon, the price concept and the available choices.
Data collection starts with prices net of relevant rebates and discounts, quantities and buyer participation, identifiable product and contract terms, own and cross-substitution information, marginal resource costs, capacity, taxes and other transfers. Establish units and timing. Monetary quantities from different years require an explicit price deflator when real values are compared. Multicountry price indexes require exchange-rate or purchasing-power conventions appropriate to the question. Buyer-level records should be handled under applicable confidentiality and data-minimization safeguards.
The next step is to identify demand and cost, check the assumed pricing behavior against observed choices, solve the alternative equilibrium, and calculate benefit and resource differences. Where only one observed price and one quantity are available, a flexible demand curve is not identified. One can report scenario bounds using explicit shape and elasticity restrictions, but should not label the result an empirical point estimate. Where microdata are inaccessible, a paper's abstract may substantiate a reported finding while being insufficient to reproduce the estimate. The evidence record for the present synthesis identifies this access distinction.
Robustness analysis should vary assumptions governing demand curvature, cost, participation margins, time horizon and market boundaries. Check that the observed allocation is feasible, that counterfactual quantities are physically possible, that fixed or sunk costs are treated consistently, and that price transfers cancel in the unweighted global account. Before drawing policy conclusions, inspect whether each welfare component comes from an identified causal or structural mechanism and whether a second component double-counts the same resource loss.
Original analytical synthesis: comparing pricing regimes
The research reviewed here supports a consistent interpretation of monopoly loss as conditional on the feasible alternatives. The central distinction is between the price-setting decision, the realized allocation and the counterfactual resource account. At the first level, a single-price monopolist faces a reduction in inframarginal revenue when output expands. At the second, actual transactions and the assignment of units across buyers determine who obtains the gains from exchange. At the third, a welfare evaluation compares all gross benefits and resource costs in the actual and alternative arrangements.
The constructed cases give this distinction quantitative content. In the linear single-market case, a 40-unit price wedge and 40-unit output gap create a loss of 800. In the split-market case, discrimination increases profit by 200 with aggregate output unchanged and lowers welfare by 100. With a previously excluded smaller segment, segmentation improves welfare by 84.375 while leaving the large market unchanged. With a fixed infrastructure cost of 1,200, break-even uniform pricing yields a feasible high-output allocation with residual inefficiency of 200 relative to externally funded marginal-cost pricing. These values are obtained directly from the stated demands and costs.
They demonstrate four logically distinct margins: total quantity; assignment of quantities across heterogeneous consumers; access to a contract or market; and the resource costs of financing and maintaining supply. Each margin can move while another stays fixed. Observed changes in price, concentration, producer profit or aggregate revenue cannot identify these movements simultaneously. A study should present output, allocation, access, cost and incidence in the sequence required by its substantive question.
The empirical cases reinforce the same interpretation without claiming uniform magnitudes. Electricity payment decompositions differentiate competitive scarcity rents and market power from resource costs. Generic-entry studies disclose how lower prices and new supply interact; international indexes establish the importance of measurement conventions; differentiated-product demand estimation makes substitution and ownership visible; network regulation foregrounds feasibility and funding. These independent literatures clarify the conditions governing the conditional derivations. The synthesis organizes established economic tools and the original worked comparisons as one research argument; it makes no claim to a new economic theory or independently estimated welfare series.
Limitations, unresolved questions, and extensions
The most important maintained assumptions of the basic analysis are the definition of the product and market, the use of a specified demand function, the cost concept, optimization by the seller, the feasible choice set, and the welfare metric. These assumptions are deliberately varied through the worked comparisons. The linear example's one-quarter loss share is conditional; the output criterion in the segment example is bounded by the assumptions of independent demands and a common resource-cost account; the perfect-discrimination result abstracts from information, privacy and participation frictions.
The literature evidence available to this study includes original papers in extracted text, publisher journal and book descriptions, institutional research summaries and official notices. Full raw datasets, code, confidential price logs and cost records were not obtained. Quantitative findings are attributed to the original studies, with period and method restrictions. This is a substantive literature synthesis, not independent empirical replication or a comprehensive legal review of every jurisdiction. Some paywalled articles and historical scans have limited direct access; the evidence ledger identifies the level used. The exact live publication status of some internal URLs could not be retrieved in this session; the related Science identity rests on the project registry.
Future research can extend these results through independently replicated sectoral models, better identification of marginal cost and demand curvature, credible estimation of participation and quality changes, measurement of privacy and other administrative costs, and explicit treatment of long-run entry and innovation. Cross-country designs must consider policy endogeneity, differences in accounting and quality, and the population represented in the welfare objective. Dynamic studies should distinguish anticipated investment incentives from realized transfers and use a compatible discount convention.
One open research problem has particular significance: current price-cost margins are easier to estimate than the distribution of willingness to pay among excluded buyers. Those buyers are absent from ordinary sales data, but their valuations govern the marginal gains from expanding output. Research that identifies this extensive margin, alongside cost and demand, can substantially improve the credibility of welfare counterfactuals. These are research directions rather than unresolved mathematical claims of the present worked examples.
Conclusion
Monopoly pricing creates an allocative distortion under the classical conditions because the marginal revenue of a seller using a uniform price falls below the buyer's marginal willingness to pay. The private optimum therefore restricts output relative to the surplus-maximizing allocation when demand decreases, marginal cost is appropriately specified, and the alternative is feasible. The missing net benefit is the integral of marginal valuation less marginal resource cost over the forgone quantity interval. The triangle is a useful geometry for linear demand and constant marginal cost; the integral and the complete surplus account govern more general cases.
A price-cost wedge measures a decision margin. It does not establish the absolute volume of missing transactions, the distribution of payments, the cost of alternative supply, or the dynamic consequences of market structure. The evidence examined here demonstrates the value of distinguishing these dimensions before quantifying welfare. Discrimination can reduce or increase the loss through changes in participation, output and assignment across buyers. A tariff that covers fixed network costs may dominate alternative feasible arrangements even when marginal-cost pricing is the first-best benchmark. Investment and rent-acquisition effects expand the account beyond static transactions and require independently supported estimates.
The lasting analytical result is a disciplined correspondence between the seller's decision, the realized pattern of exchange, and the resource differences under a feasible counterfactual. For policy evaluation, publish the price concept, quantity response, valuation measure, cost basis, transfers, affected agents, time horizon, and identifying assumptions. Evaluate observed outcomes against explicit alternatives and maintain separate consumer, producer, fiscal and resource accounts. This clarifies both the scope of formally established results and the evidence necessary to transport them across markets.
References
-
Berry, Steven T., James Levinsohn, and Ariel Pakes. (1995). Automobile Prices in Market Equilibrium. Econometrica, 63(4), 841–890. DOI: 10.2307/2171802.
-
Borenstein, Severin, James B. Bushnell, and Frank A. Wolak. (2002). Measuring Market Inefficiencies in California's Restructured Wholesale Electricity Market. American Economic Review, 92(5), 1376–1405. DOI: 10.1257/000282802762024557.
-
Court of Justice of the European Communities. (1978). United Brands Company and United Brands Continentaal B.V. v. Commission of the European Communities, Case 27/76, judgment of February 14, 1978. Official Journal C 76, March 30, 1978, p. 3 (notice).
-
Danzon, Patricia M., and Li-Wei Chao. (2000). Prices, Competition and Regulation in Pharmaceuticals: A Cross-National Comparison. London: Office of Health Economics.
-
De Loecker, Jan, Jan Eeckhout, and Gabriel Unger. (2020). The Rise of Market Power and the Macroeconomic Implications. Quarterly Journal of Economics, 135(2), 561–644. DOI: 10.1093/qje/qjz041.
-
Harberger, Arnold C. (1954). Monopoly and Resource Allocation. American Economic Review, 44(2), 77–87.
-
Lerner, Abba P. (1934). The Concept of Monopoly and the Measurement of Monopoly Power. Review of Economic Studies, 1(3), 157–175. DOI: 10.2307/2967480.
-
Ofcom. (2021). Promoting Investment and Competition in Fibre Networks: Wholesale Fixed Telecoms Market Review 2021–26. Statement, March 18, 2021. London: Ofcom.
-
OpenStax. (2022). Principles of Economics 3e, section 3.5: Demand, Supply, and Efficiency. Houston, TX: OpenStax, Rice University.
-
OpenStax. (2022). Principles of Economics 3e, section 9.2: How a Profit-Maximizing Monopoly Chooses Output and Price. Houston, TX: OpenStax, Rice University.
-
Posner, Richard A. (1974). Social Costs of Monopoly and Regulation. NBER Working Paper 0055. DOI: 10.3386/w0055. Published version: The Social Costs of Monopoly and Regulation (1975), Journal of Political Economy, 83(4), 807–827.
-
Tirole, Jean. (1988). The Theory of Industrial Organization. Cambridge, MA: MIT Press. ISBN 9780262200714.
-
U.S. Department of Justice and Federal Trade Commission. (2023). Merger Guidelines. Issued December 18, 2023. Washington, DC.
-
U.S. Food and Drug Administration. (n.d.). Generic Competition and Drug Prices. Center for Drug Evaluation and Research. Accessed November 17, 2026.
-
Varian, Hal R. (1985). Price Discrimination and Social Welfare. American Economic Review, 75(4), 870–875. JSTOR: 1821366.
-
Varian, Hal R. (1989). Price Discrimination. In Richard Schmalensee and Robert D. Willig (eds.), Handbook of Industrial Organization, vol. 1, 597–654. Elsevier. DOI: 10.1016/S1573-448X(89)01013-7.