Monopoly: Marginal Revenue, Deadweight Loss, Price Discrimination and the Three Real Costs

Monopoly — a market structure in which a single firm supplies the entire output of an industry, there are no close substitutes for its product, and entry is blocked by significant barriers.

The defining consequence: the firm’s demand curve is the market demand curve. It slopes downward. The monopolist is a price maker, not a price taker.

Baumol’s contestable markets theory (see Part VII) showed that what confers durable market power is not the cost of entering but the sunk portion of that cost — expenditure that cannot be recovered on leaving. A large but fully recoverable fixed cost (a leased aircraft) deters nobody. A modest but irrecoverable one (bespoke tooling, brand advertising) deters everybody. The barrier to entry is really a barrier to exit.

Read the equation. If demand is inelastic, |ε| < 1, so 1/ε < −1, and MR is negative.

A profit maximiser sets MR = MC, and marginal cost is never negative. Therefore a monopolist never operates on the inelastic portion of its demand curve. If it found itself there, it could raise price, sell less, spend less on production, and earn more — and would do so until demand became elastic. This is one of the most testable predictions in all of microeconomics.

With linear demand, the MR curve has the same intercept and exactly twice the slope. Graphically, MR bisects the horizontal distance between the vertical axis and the demand curve. If your MR line does not hit the quantity axis at precisely half the demand curve’s intercept, the diagram is wrong and every subsequent mark is at risk.

Students find Q_m where MR = MC and then read the price off the MR curve. Wrong. MR is an internal calculation used to choose quantity. The price is whatever consumers will pay for Q_m — read it vertically up to the demand curve. (This is structurally identical to the monopsony error, where students read the wage off MCL instead of the labour supply curve.)

This surprises almost everyone. A supply curve maps price → quantity supplied. But the monopolist does not observe a price and respond; it chooses the price.

Formally: two different demand curves passing through the same (P, Q) point but with different elasticities will generate different MR curves, and therefore different optimal quantities at the same marginal cost. The mapping from price to quantity is not a function. Only price takers have supply curves. This is worth a sentence in any evaluation answer.

Deadweight Loss (DWL) — the total surplus destroyed because output is restricted below the allocatively efficient level. It is the triangle bounded by the demand curve above, the MC curve below, and the vertical lines at Q_m and Q_c.

Between Q_m and Q_c there exist units for which consumers’ willingness to pay exceeds the cost of producing them. Those trades are mutually beneficial. They do not happen.

Harberger, A. C. (1954), “Monopoly and Resource Allocation,” American Economic Review 44(2)

What he did: Harberger took the deadweight-loss triangle seriously as an empirical object. Using data on rates of return across US manufacturing industries in the 1920s, he estimated the surplus lost to monopolistic output restriction.

What he found: the total welfare cost of monopoly in US manufacturing amounted to well under one percent of national income — a figure so small that, as one later commentator observed, eliminating monopoly entirely would have been worth less than a good harvest.

The crisis this created: economists had spent half a century condemning monopoly. Harberger appeared to show the condemnation was disproportionate to the harm. Either the theory mattered less than everyone thought, or the triangle was measuring the wrong thing.

The resolution occupied the next twenty years, and produced two ideas — rent-seeking and X-inefficiency — that between them recast the entire economics of market power.

Tullock, G. (1967), “The Welfare Costs of Tariffs, Monopolies, and Theft,” Western Economic Journal 5(3); extended by Posner, R. (1975), Journal of Political Economy 83(4)

The insight: Standard analysis treats monopoly profit — the rectangle (P_m − ATC)·Q_m — as a mere transfer from consumers to the producer. Nothing is destroyed; the surplus simply changes hands. Only the triangle is a genuine welfare loss.

Tullock’s objection: that rectangle is a prize. Firms will expend real resources to win it — lobbying legislators, litigating against rivals, capturing regulators, funding political campaigns, building excess capacity purely to deter entry. Those resources produce nothing.

Posner’s extension: if competition to acquire the monopoly is itself competitive, firms will collectively dissipate the entire monopoly rent in the pursuit of it. In the limit, the whole rectangle is a welfare loss, not a transfer.

The magnitude: the rectangle is typically an order of magnitude larger than the triangle. Harberger measured the small loss and missed the large one. This is why lobbying expenditure, not consumer prices, is where modern competition economists look first.

Leibenstein, H. (1966), “Allocative Efficiency vs. ‘X-Efficiency’,” American Economic Review 56(3)

The argument: The entire welfare analysis above assumes the monopolist minimises cost — that it sits on its cost curve. Why would it? A competitive firm that tolerates waste is destroyed. A monopolist that tolerates waste merely earns less than the maximum, and no one can punish it.

The consequence: managerial slack, overstaffing, gold-plated offices, an absence of cost discipline. Leibenstein called this X-inefficiency: the gap between actual cost and minimum attainable cost.

Why it dwarfs the triangle: the deadweight loss measures units not produced. X-inefficiency measures every unit that is produced being made too expensively. The first is a triangle; the second is a shift of the entire cost curve upward.

Empirical echo: Nickell (1996), studying UK manufacturing firms, found that greater product-market competition was associated with higher total factor productivity growth — consistent with competition disciplining managerial slack rather than merely reallocating output.

Cite all three in an evaluation answer. Students who draw only the triangle are reproducing the analysis that Harberger himself showed to be nearly irrelevant.

Price discrimination requires: (1) market power; (2) the ability to segment buyers by willingness to pay; (3) the prevention of arbitrage — resale from the low-price segment to the high-price one. Remove condition (3) and the whole scheme collapses.

A perfectly price-discriminating monopolist produces the same output as a competitive industry and generates zero deadweight loss. Total surplus is maximised. It simply all accrues to the firm; consumer surplus is zero.

Efficiency and equity are entirely separable here. An outcome can be Pareto-efficient and morally intolerable. If your instinct is that perfect price discrimination must be worse than ordinary monopoly, examine that instinct: on efficiency grounds it is strictly better. What you object to is distribution.

Natural monopoly — an industry in which the cost function is subadditive: a single firm can supply the entire market at lower total cost than any combination of two or more firms. Arises from very large fixed costs and low marginal costs, so ATC declines throughout the relevant range of output.

Efficiency demands P = MC. Solvency demands P ≥ ATC. Under subadditivity these are mathematically incompatible. Every real regulatory regime is a compromise between them, and the choice of compromise is a distributive decision about who pays: taxpayers (via subsidy) or consumers (via price above marginal cost). Regulatory capture — the tendency of the regulator to be gradually reconstructed in the image of the firm it regulates (Stigler, 1971) — is the ever-present failure mode.

The facts: By 1904 Standard Oil controlled roughly 90% of US refining, built through acquisition, preferential railroad rebates, and control of pipelines. In 1911 the Supreme Court ordered its dissolution into thirty-four companies under the Sherman Act.

The paradox: Rockefeller held shares in every successor. Freed from the constraint of internal cross-subsidy and permitted to compete, the pieces — which became Exxon, Mobil, Chevron, Amoco and others — were collectively worth far more apart than together. His fortune grew.

The economics: If the parts were more valuable separately, the conglomerate was destroying value — evidence of X-inefficiency on a heroic scale, not merely allocative distortion. Standard Oil’s inefficiency was invisible in the deadweight-loss triangle and enormous in Leibenstein’s terms.

The historical dispute: McGee (1958), in a famously revisionist article, examined the trial record and argued Standard Oil had rarely used predatory pricing at all — it bought rivals rather than starving them, because buying was cheaper. The debate over whether predatory pricing is ever a rational strategy has never fully closed.

The mechanism: For most of the twentieth century De Beers controlled the majority of world rough diamond supply, operating a single-channel marketing structure that released stones in strictly rationed quantities.

The economics: Diamonds are not geologically rare. Their price reflected managed supply meeting demand that De Beers itself had cultivated — the association of diamonds with engagement was largely the product of a sustained advertising campaign begun in 1938. The firm did not merely restrict output along a given demand curve; it shifted the demand curve outward and made it less elastic. By the Lerner index, that is the purest possible investment in market power.

The decline: Discoveries in Russia, Australia and Canada — outside the cartel — and later the arrival of chemically identical laboratory-grown stones eroded the structure. De Beers’s share fell from around 90% to a minority. No barrier to entry survives a large enough change in technology.

Microsoft: the US government alleged Microsoft maintained an operating-system monopoly by tying Internet Explorer to Windows. The appeals court found monopoly maintenance unlawful but overturned the ordered break-up, settling on conduct remedies.

Google: in August 2024, a US federal court found that Google had unlawfully maintained a monopoly in general search, principally through exclusive default-placement agreements with device manufacturers and browsers.

What is distinctive about network monopolies: the barrier is not a patent or a mine. It is that the product improves as more people use it, so the incumbent’s quality advantage is its market share. Search improves with query volume; the advantage compounds.

The genuine analytical difficulty: in a network industry, the monopoly outcome may be the efficient one — duplicating the network wastes resources, and a fragmented market delivers a worse product. This is close to natural monopoly. The regulator’s problem is therefore not to eliminate the monopoly but to prevent it exploiting its position, without destroying the network effects that made the product good. Nobody has solved this.

Schumpeter (1942) argued that the relevant competition is not between firms in a market but between the market and the technology that will destroy it — creative destruction. Temporary monopoly profit is the reward that induces the innovation. Eliminate it and you eliminate the innovation.

Against this, Arrow (1962) noted the replacement effect: a monopolist innovating merely cannibalises its own existing profits, whereas an entrant gains everything. The incumbent’s incentive to innovate is therefore weaker, not stronger. Aghion et al. (2005) found the truth is an inverted U — innovation peaks at intermediate competition. Neither Schumpeter nor Arrow was straightforwardly right.

Demand is P = 100 − 2Q. Total cost is TC = 20Q. Find (a) the monopoly price, quantity and profit; (b) the competitive outcome; (c) the deadweight loss.

A monopolist serves two separable markets: P₁ = 80 − 2Q₁ and P₂ = 60 − Q₂. MC = 20 (constant). Find the optimal price in each market. Which market pays more, and why?

A water utility has TC = 1,000 + 5Q. Demand is P = 105 − Q. (a) Find the unregulated monopoly outcome. (b) What happens under marginal cost pricing? (c) Under average cost pricing?

Prove it, and then explain the intuition without algebra.

• MR has the same intercept as D and twice the slope. Draw it bisecting the horizontal.
• Find Q_m where MR = MC. Go vertically up to D for the price. Never read price from MR.
• Profit rectangle = (P_m − ATC at Q_m) × Q_m.
• DWL triangle sits between D and MC, from Q_m to where MC crosses D.
• For a natural monopoly, draw ATC falling throughout with MC below it. Show that P = MC generates a loss.
• A lump-sum tax shifts ATC up but leaves MC unchanged: price and quantity do not change. A per-unit tax shifts MC up: price rises, quantity falls. Examiners love this contrast.

“Evaluate the view that monopoly is always against the consumer interest.”

1. Define monopoly; establish MR < P from the downward-sloping demand curve.
2. Derive P > MC at the profit-maximising output. Draw the deadweight loss. Analysis marks.
3. Note the failure of both allocative and productive efficiency, contrasting explicitly with perfect competition.
4. Then deepen it. Cite Harberger: the triangle is empirically small. Cite Tullock and Posner: rent-seeking dissipates the far larger rectangle. Cite Leibenstein: X-inefficiency shifts the whole cost curve.
5. Then invert it. Schumpeter: monopoly profit funds and rewards innovation. Arrow’s replacement effect and the Aghion inverted-U qualify this.
6. Economies of scale: a natural monopoly may deliver lower prices than a fragmented industry could.
7. Price discrimination: first-degree discrimination is allocatively efficient. The objection to it is distributive, not efficiency-based.
8. Contestability (Baumol): the threat of entry may discipline a monopolist even with one firm.
9. Conclude conditionally on barriers to entry, the pace of technological change, and whether the industry is naturally monopolistic.

• Reading the monopoly price off the MR curve. Price comes from the demand curve.
• Claiming a monopolist “charges whatever it likes.” It is constrained by the demand curve. It picks a point on that curve, not a point above it.
• Saying a monopolist always makes supernormal profit. Only if demand lies above ATC somewhere.
• Drawing a monopoly supply curve. There is none.
• Presenting the DWL triangle as the total cost of monopoly. Harberger showed it is the smallest of the three costs.
• Treating price discrimination as unambiguously bad. Perfect discrimination eliminates deadweight loss entirely.
• Claiming a lump-sum tax on a monopolist lowers prices. It does not touch MC, so Q and P are unchanged. Only profit falls.

Exercise 1 — If Rent-Seeking Dissipates the Rent, Why Does Anyone Seek It?

Posner argued that competitive rent-seeking dissipates the entire monopoly rectangle, so the expected profit from seeking a monopoly is zero. But firms manifestly do spend on lobbying, and manifestly do earn positive returns on it — some estimates of the return to corporate lobbying are extraordinarily high.

Reconcile these. Consider: is the rent-seeking “contest” actually competitive, or do incumbents possess a first-mover advantage in the political market that entrants lack? If rent-seeking is itself an imperfectly competitive activity, does Posner’s full-dissipation result survive — and if not, is the true social cost of monopoly greater or smaller than he claimed?

📄 Read: Posner, R. A. (1975). “The Social Costs of Monopoly and Regulation.” Journal of Political Economy, 83(4), 807–827. Read alongside Tullock, G. (1980), “Efficient Rent Seeking,” and note precisely which assumption delivers full dissipation.

Exercise 2 — Can a Network Monopoly Be Regulated Without Being Destroyed?

In a network industry, the product improves with market share. Fragmenting the monopolist makes the product worse for everyone. Yet leaving it intact permits the exercise of market power. This is the structure of the Microsoft and Google cases.

Design a remedy. Consider interoperability mandates, data portability, structural separation of the platform from the products sold on it, and behavioural conduct rules. For each, identify what economic mechanism it targets and what it destroys. Then ask the harder question: is a network monopoly simply a natural monopoly in information, and if so, should it be regulated like a utility rather than prosecuted like a cartel?

📄 Read: Baumol, W. J., Panzar, J. C., & Willig, R. D. (1982). Contestable Markets and the Theory of Industry Structure. Then read Stigler, G. J. (1971), “The Theory of Economic Regulation,” Bell Journal of Economics, 2(1), 3–21, and assess whether your proposed regulator would survive capture.