Monopolistic Competition

The box that perfect competition and monopoly forgot

Open any introductory textbook and you will meet two stylised market structures right away. On one side, perfect competition: an infinite number of identical firms selling an identical good, every seller a price-taker at the market-clearing P = MC. On the other side, pure monopoly: one seller, no substitutes, demand sloping downward as steeply as the market itself. These are the two boxes the supply-and-demand machinery was originally built around.

The trouble is that almost nothing in the actual economy fits either box. A coffee shop is not a price-taker — it sets its own prices, and most customers do not walk out if a latte goes from $5.00 to $5.25. But it is also not a monopolist — there are three other coffee shops within five minutes' walk. A burrito restaurant, a hair salon, a craft-beer brewer, a clothing label, an iOS app: all of these sit awkwardly in the middle. They have some pricing power because their offering is differentiated from rivals, but the power is small because close substitutes are everywhere.

Monopolistic competition is the textbook structure that fills this gap. Two economists arrived at it independently in 1933 — Edward Chamberlin at Harvard with The Theory of Monopolistic Competition, and Joan Robinson at Cambridge with The Economics of Imperfect Competition. The model has been the workhorse for most of microeconomics' descriptive applications since.

The four market structures, side by side

Structure	Sellers	Product	Pricing power	Entry	Long-run profit	Examples
Perfect competition	∞	Identical	None (P = MC)	Free	Zero	Wheat, foreign exchange spot
Monopolistic competition	Many	Differentiated	Small (P > MC)	Free	Zero (entry erodes)	Restaurants, salons, apparel
Oligopoly	Few	Identical or differentiated	Strategic interdependence	Barriers	Positive (often)	Airlines, telecoms, cars
Monopoly	One	Unique, no substitutes	Full (P set by demand)	Blocked	Positive (sustained)	Patent holder, regulated utility

The two extremes — perfect competition and monopoly — are the polar cases. Monopolistic competition borrows the many firms + free entry half from perfect competition and the downward-sloping demand half from monopoly. The result is the awkward but realistic middle: every firm is a tiny monopolist over its own variety, but rivals close in on all sides.

Short run — each firm is a mini-monopoly

Take any firm in this market. Because its product is differentiated — a particular burger, a particular cut, a particular brand — a small price increase does not send every customer running. The firm's residual demand curve slopes downward. The firm maximises profit by choosing the quantity where its own marginal revenue equals its own marginal cost (MR = MC) and reading the price off the demand curve at that quantity. That price is necessarily above marginal cost, because MR < P whenever demand slopes downward.

Firm's choice: max π = P(q) · q − C(q)
First-order condition: MR = MC
Markup: (P − MC) / P = 1 / |ε_d|       (Lerner index)

If the demand curve sits above the average-cost curve at that quantity, the firm earns a positive economic profit (the rectangle (P − AC) · q). If it sits below, the firm makes a loss. The short run is whatever interval is too short for new firms to enter or for incumbents to leave; during that interval, profits and losses persist.

Long run — free entry erodes profit to zero

The crucial structural assumption is free entry: there are no patents, no regulatory barriers, no irreproducible inputs that keep newcomers out. If existing firms are earning positive economic profit, new firms enter. Each new variety pulls some customers away from existing firms, shifting their residual demand curves inward. Inward-shifting demand means the price at any given quantity falls, and so does the profit.

Entry continues as long as positive profit remains. The process stops only when each firm's demand curve has been pulled in just far enough to be tangent to its average-cost curve. At the tangency point, price equals average cost — economic profit is zero. Firms can no longer be tempted to enter, but they can survive (because P = AC means revenues cover all costs including normal return on capital). This is the long-run equilibrium of monopolistic competition.

Long-run equilibrium conditions:
   (1) MR = MC          (each firm profit-maximises)
   (2) P  = AC           (zero profit — free-entry condition)
   (3) Demand curve tangent to AC curve at (q*, P*)

Condition (3) follows from (1) and (2): tangency is the only way a downward-sloping demand curve can meet a U-shaped AC curve at exactly one point without crossing it.

The excess-capacity theorem

Here is the most-cited geometric consequence of the tangency condition. Average cost has a U-shape with a minimum (the efficient scale) at the bottom. A demand curve that slopes downward can only be tangent to a U-shaped curve on a portion where both curves have the same slope. The bottom of the U has zero slope; the demand curve has negative slope; therefore the tangency must occur strictly to the left of the AC minimum — on the downward-sloping branch of the cost curve.

Each firm in long-run equilibrium therefore produces less than the cost-minimising quantity. If the firm doubled its scale, average cost would fall — but doing so would require selling at a price below AC at the larger quantity, generating losses. The gap between actual output q* and the AC-minimising output q_eff is called excess capacity.

Chamberlin (1933) presented this as a critique: the industry has 'too many firms, each too small'. Critics of the model used the result to argue that monopolistic competition is wasteful — and at one level it plainly is, since the same total output could be produced more cheaply by fewer, larger firms. The defence, articulated decades later by Dixit and Stiglitz, is that the excess capacity is the price of variety, and that variety has consumer value.

The Dixit-Stiglitz framework — why variety pays

The breakthrough paper that gave monopolistic competition its modern tractability is Dixit and Stiglitz (1977), Monopolistic Competition and Optimum Product Diversity. They wrote down a constant-elasticity-of-substitution (CES) utility function in which consumer utility is increasing in the number of varieties consumed:

U = ( Σ_{i=1}^N q_i^ρ )^(1/ρ),     0 < ρ < 1
     ↑
elasticity of substitution σ = 1 / (1 − ρ) > 1

The key feature: holding total quantity Σ q_i constant, utility strictly rises when the same total is split across more varieties. Consumers love variety. A representative consumer is strictly better off with two slightly different burritos than with twice as much of one.

With this preference structure plugged in, equilibrium has clean, log-linear properties: every firm charges the same markup σ/(σ−1), demand for each variety has constant elasticity σ, and the number of varieties N is determined by the zero-profit condition. The framework is symmetric, tractable, and elegant — which is why it became the building block for both modern trade theory and modern macroeconomics.

The welfare calculus is now a tradeoff: more firms means more variety (good) but more excess capacity per firm (bad). Dixit and Stiglitz showed that the free-entry equilibrium produces too few firms compared with the social optimum if and only if the elasticity of substitution σ is high; for low σ, free entry actually produces too many. The model gives a quantitative grip on a tradeoff that Chamberlin could only sketch verbally.

The markup formula

Profit maximisation by a firm facing demand q(P) with elasticity ε_d < 0 implies (taking log and differentiating)

MR = P · (1 + 1/ε_d)
Setting MR = MC:
   P · (1 + 1/ε_d) = MC
   P / MC = 1 / (1 + 1/ε_d) = |ε_d| / (|ε_d| − 1)

This is the Lerner markup formula. The price-to-marginal-cost ratio is determined entirely by the elasticity of demand the firm faces. If |ε_d| is very large (close substitutes everywhere — competition is fierce), the markup is small and P → MC, recovering perfect competition. If |ε_d| is small (no close substitutes), the markup is large, approaching pure-monopoly behaviour.

Monopolistic competition sits in the intermediate regime. With Dixit-Stiglitz CES preferences, every firm has the same constant demand elasticity σ, so every firm charges the same markup σ/(σ−1). Estimated values of σ for differentiated products typically fall between 3 and 10, giving markups of 11% to 50% over marginal cost.

Real-world examples

• Restaurants. The classic example. Hundreds of restaurants in any city, each with a different menu, atmosphere, location, and chef. Each is a mini-monopoly over its own combination; rivals are everywhere. Entry is cheap (you lease a space, hire a few cooks, register an LLC), and exits are visible too — most independent restaurants close within five years. The tangency condition manifests as restaurants typically running at occupancy well below their kitchen's theoretical maximum: excess capacity in the flesh.
• Hair salons and barbershops. Same shape. Differentiation by skill, location, hours, and clientele. Each chair has a local monopoly over its specific stylist's work; entry barriers are essentially a chair and a licence.
• Clothing brands. A H&M jacket and a Uniqlo jacket are not identical even if functionally equivalent. Each brand has a sliver of consumer preference attached to its label. The fashion industry's constant churn of new brands and new lines is monopolistic competition in motion.
• App stores. Tens of thousands of meditation apps, photo-editing apps, to-do list apps. Each developer offers a slightly different UI, feature set, or pricing. Marginal cost of distribution is essentially zero, but development effort is the fixed cost; the equilibrium number of apps is set by free entry until expected revenue covers expected development outlay.
• Craft beer. Local breweries, each with their own IPAs and lagers. The American craft-beer explosion from ~100 breweries in 1980 to ~9,000 in 2024 is a textbook illustration of free entry into a monopolistically competitive market until profits were competed down.
• Independent coffee shops. The opening scene of every barista is a mini-monopoly story. Starbucks, Peet's, Blue Bottle, and the neighbourhood shop all sell coffee — but not the same coffee, not in the same atmosphere, not from the same barista.

Where does advertising fit?

Chamberlin's original 1933 framework devoted serious attention to selling costs — what we would today call advertising and promotion. In perfect competition there is no point advertising; the product is identical to every rival's, and any price below the market clears infinite demand anyway. In pure monopoly, advertising might shift the demand curve outward but adds little to a captive market. In monopolistic competition advertising matters: it is the firm's mechanism for differentiating its product in the mind of the consumer, shifting its residual demand outward and possibly reducing its elasticity (making demand less responsive, which raises the markup).

In equilibrium, firms spend on advertising up to the point where the marginal dollar of advertising shifts demand by an amount that just covers the dollar's cost. This is the rationale for the gigantic ad budgets of nominally competitive categories like soap, beer, and quick-service food: each firm individually has an incentive to keep advertising, and total expenditure stays high.

In macroeconomics — New Keynesian models

The whole apparatus of modern central-bank macroeconomics rests on Dixit-Stiglitz monopolistic competition. The reason is technical but important. A model in which all firms are perfect competitors gives them no margin of pricing power; the equilibrium price equals marginal cost, full stop. There is no slack for nominal rigidities — if changing a price is costly (a menu cost), the firm has no buffer to absorb that cost, so it changes prices anyway. Result: prices fully flexible, money fully neutral, monetary policy has no real effects.

Plug in monopolistic competition, and each firm prices above MC. Now there is a buffer. A small menu cost can rationally lead firms to keep their price unchanged for a while — they leave a few dollars on the table but avoid the menu cost. Aggregate this across millions of firms with staggered price adjustment (Calvo 1983) and you get sticky aggregate prices: a monetary expansion raises real output, not just nominal prices, in the short run. This is the engine of every New Keynesian DSGE model used by every central bank.

The math works because Dixit-Stiglitz preferences plus monopolistic competition give a closed-form, log-linearisable structure that can be embedded in stochastic general equilibrium without losing tractability. There is no equally clean substitute, which is why the structure has been near-universal in macro since the 1990s.

In international trade — Krugman's New Trade Theory

Pre-1979, the standard trade models (Ricardian, Heckscher-Ohlin) explained trade as countries specialising in goods according to comparative advantage and exchanging different products. The trouble: data from the postwar period showed enormous intra-industry trade — Germany exports cars to France while France exports cars to Germany; the United States exports chemicals to Europe while importing chemicals from Europe. Neither classical theory predicted that.

Paul Krugman's 1979 paper, Increasing Returns, Monopolistic Competition, and International Trade, used the Dixit-Stiglitz framework to build a trade model where firms produce differentiated varieties under increasing returns, and where consumers in every country love variety. The implications:

• Each country produces a subset of the world's varieties (because increasing returns favour concentration at the firm level).
• Consumers in every country want all the varieties.
• Therefore countries trade similar but differentiated goods with each other.
• Trade volume rises with country similarity (the so-called gravity equation in trade), opposite to what comparative advantage would predict.

The model fit the intra-industry trade data spectacularly. It also generated novel welfare implications: trade is welfare-improving even between identical countries because it expands the variety set available to consumers in both. Krugman shared the 2008 Nobel Memorial Prize 'for his analysis of trade patterns and location of economic activity'.

Critiques and limitations

• Excess-capacity ambiguity. The 'left of minimum AC' result is geometric — it tells you the direction but not the magnitude of the inefficiency. Empirically, the excess-capacity gap is small in some industries (≤ 5%) and large in others; the welfare cost of monopolistic competition is therefore not pinned down by the theory alone.
• Variety doesn't always benefit consumers. The Dixit-Stiglitz love-of-variety assumption is convenient but not innocuous. If consumers face cognitive costs of choosing among many similar options, more variety may reduce welfare (Iyengar & Lepper's 'jam study', 2000). The theory is silent on this.
• Symmetric varieties is a strong assumption. Dixit-Stiglitz assumes every variety enters utility the same way. Real product differentiation is asymmetric — some varieties are 'better' and command higher markups. The Melitz (2003) extension brings firm heterogeneity into the framework but at the cost of analytical tractability.
• Where do new varieties come from? The model treats N as an endogenous outcome but the type of new variety as exogenous. Real innovation involves choices about what kind of differentiation to introduce — a different flavour, a different feature, a different price point — and the model has nothing to say about those choices.
• Strategic interaction is ignored. In monopolistic competition, firms are assumed to ignore strategic interactions with rivals (because each one is small). In real markets with a few dozen players the assumption is dubious — oligopoly models (Cournot, Bertrand) handle that case but at the cost of dropping free entry.

Common pitfalls

• Confusing zero economic profit with zero accounting profit. The 'zero profit' in long-run equilibrium is economic profit — revenue minus the opportunity cost of all resources including normal return on invested capital. Accountants would call this a healthy ordinary profit. Firms in monopolistic competition do not all break even in accounting terms.
• Treating differentiation as exogenous. The textbook tangency picture takes the demand curve's shape as given. In reality firms spend on R&D, design, and advertising to shape their demand curve — to push it out and make it steeper. Modelling differentiation as a choice variable changes the equilibrium materially.
• Confusing monopolistic competition with oligopoly. Monopolistic competition has many firms each too small to influence rivals; oligopoly has few firms each large enough that rivals' responses matter. The strategic-interaction layer is the dividing line. Markets with ~10 to 20 sellers sit in a grey zone; whether to model them as monopolistic competition or oligopoly is partly a modelling-convenience call.
• Misreading 'excess capacity' as 'unused factories'. Excess capacity is a statement about the cost-minimising scale, not about idle physical capacity. A restaurant that fills every table every night still has 'excess capacity' in the technical sense if doubling its size would lower per-meal cost.
• Ignoring the markup-elasticity link. Saying 'firms have market power' is content-free unless you specify the elasticity. The Lerner formula gives the magnitude, and it is the elasticity — not the number of firms per se — that determines how distortionary monopolistic competition is.