Oligopoly

The assumption about rivals' reactions. Each firm believes that:

• if it cuts its price, the rivals will match the cut at once, so as not to lose customers. It therefore gains very little extra sales.
• if it raises its price, the rivals will not follow, and will be glad to pick up its customers. It therefore loses a great deal of sales.

The consequence. The firm's demand curve has two segments meeting at the current price: elastic above the current price, because a rise loses many customers, and inelastic below it, because a cut gains few. That produces a kink at the prevailing price.

Why the price then stays put. A kink in the demand curve produces a vertical gap in the marginal revenue curve at the current output. Marginal cost can rise or fall within that gap without altering the point where marginal cost equals marginal revenue. So costs can change appreciably and the profit maximising price does not move. That is price rigidity, derived rather than asserted.

The honest criticism, which a full answer includes. The model explains why a price, once established, stays where it is. It does not explain how that price came to be established in the first place, and it does not describe industries in which prices move together frequently, as they do where a cartel or a price leader is at work.

Price leadership and other non collusive patterns

Where firms do not agree formally, several patterns appear.

Price leadership. One firm sets the price and the others follow. The leader may be dominant, the largest firm; barometric, the firm best at reading market conditions; or low cost, the firm that can sustain the lowest price.

Cartels. A formal or informal agreement among rival firms on price, output, market shares or bidding. This is the subject of the law below.

Tacit collusion or conscious parallelism. Firms behave alike without any agreement, simply by watching each other. This is the hardest case for a competition authority, because parallel behaviour is not by itself unlawful; the authority must show more than the fact that prices moved together.

What the law says about oligopoly in India

An oligopoly is not unlawful. Agreeing is.

Section 3(1) prohibits any agreement in respect of production, supply, distribution, storage, acquisition or control of goods or provision of services which causes or is likely to cause an appreciable adverse effect on competition within India. Section 3(2) makes such an agreement void.

Section 3(3) is the provision that matters most here. Any agreement between enterprises or persons engaged in identical or similar trade, including cartels, which:

• (a) directly or indirectly determines purchase or sale prices;
• (b) limits or controls production, supply, markets, technical development, investment or provision of services;
• (c) shares the market or source of production by allocating a geographical area, a type of goods or services, a number of customers or in any other similar way;
• (d) directly or indirectly results in bid rigging or collusive bidding,

shall be presumed to have an appreciable adverse effect on competition. The proviso saves an agreement made by way of a joint venture that increases efficiency. A further proviso added by Act 9 of 2023 extends the presumption to a person not in the same trade who participates or intends to participate in furthering such an agreement, which reaches a hub and spoke arrangement organised by a trade association or a common intermediary. The Explanation defines bid rigging as an agreement which has the effect of eliminating or reducing competition for bids or of adversely affecting or manipulating the bidding process.

Why the word presumed matters. For the four practices in section 3(3), the authority does not have to prove harm to competition. The harm is presumed and the burden shifts to the firms to displace it. For other agreements, including vertical ones under section 3(4), harm must be established by reference to the factors in section 19(3): barriers to new entrants, driving existing competitors out, foreclosure of competition, accrual of benefits to consumers, improvements in production or distribution, and promotion of technical, scientific and economic development.

Section 46, the leniency provision. The Commission may impose a lesser penalty on a producer, seller, distributor or trader who is a party to a cartel and makes a full and true disclosure of the alleged violations, where the disclosure is vital. This is the mechanism that actually breaks cartels, and the economics behind it is worth stating: a cartel is unstable because each member gains by cheating on it, and leniency turns that instability into an incentive to confess first.

A worked example: four cement companies and a tender

The facts. Four companies supply almost all the cement in a State. A public works department invites tenders for 40,000 tonnes. The four quote 4,780, 4,790, 4,795 and 4,800 rupees a tonne. The lowest wins. Over the previous two years each of the four has won roughly one quarter of the department's tenders, and in each case the other three quoted within one per cent of the winner.

What the economics says. In a genuinely competitive tender with four bidders of differing costs, quotes should scatter, and the same firm should tend to win where its costs are lowest. Quotes clustered within a fraction of a per cent, combined with a rotation of winners, is the pattern a cartel produces, because the members must decide whose turn it is and the losers must bid just above the winner to make the auction look real.

What the law says. The conduct falls squarely within section 3(3)(d), bid rigging, and within section 3(3)(c), market sharing by allocating a number of customers. Under section 3(3) it is presumed to have an appreciable adverse effect on competition, so the department does not have to prove that it paid more. The four companies must displace the presumption. Under section 3(2) any agreement between them is void.

What breaks it. Section 46. The first of the four to make a full and true disclosure may obtain a reduced penalty. Each of the four knows the others may go first, which is exactly the instability the provision is designed to exploit.

The lawyer's caution. Similar prices are not by themselves an agreement. In a market for an identical product with similar costs, prices should be similar; that is competition working. What makes this example different is the combination of near identical quotes with a rotation of winners over time, which competition does not produce.

The four forms compared, completed

What beginners get wrong

"Oligopoly means two or three firms." It means few enough for interdependence. An industry with eight firms of which four are large can behave as an oligopoly.

"An oligopoly is illegal." No. Having few sellers is a fact about an industry. Agreeing on price, output, markets or bids is the wrong, and section 3(3) presumes its effect.

"Parallel prices prove a cartel." They do not. In a market for an identical product, similar prices are what competition produces. Something more is needed: rotation of winners, unexplained simultaneous increases, evidence of contact, or prices that move together against costs that do not.

"The kinked demand curve explains oligopoly prices." It explains why an established price is sticky. It does not explain what the price is.

Limits and criticism

There is no single theory of oligopoly. Because the outcome depends on what each firm believes the others will do, the models multiply: Cournot on quantities, Bertrand on prices, Stackelberg on leadership, Sweezy on the kink, and modern game theory on all of it. An examiner who asks "why is there no determinate solution under oligopoly" wants exactly this answer.

Cartels are unstable. Each member gains by secretly selling more than its quota, and the temptation grows as the cartel price rises. Most cartels either collapse or are betrayed.

Oligopoly is not simply bad. Industries with heavy fixed costs and continuing research, such as telecommunications, aircraft and pharmaceuticals, cannot support a large number of small firms. Some concentration is the price of the scale and the research, and Schumpeter's argument in [Why Trade Cycles Happen, and What Governments Do About Them] is that this is where innovation actually happens.

Quick revision

1. Oligopoly: a few sellers, each large enough that its decisions affect the others. Two sellers is a duopoly.
2. Interdependence is the defining feature and the source of every analytical difficulty.
3. Pure oligopoly sells an identical product; differentiated oligopoly sells branded versions.
4. The kinked demand curve (Sweezy, 1939): rivals match a price cut but not a price rise, so the curve is elastic above the current price and inelastic below it, marginal revenue has a vertical gap, and the price is rigid against changes in cost.
5. Non collusive patterns: price leadership, whether dominant, barometric or low cost; and tacit collusion.
6. Law: section 3(1) prohibits agreements with an appreciable adverse effect on competition and section 3(2) makes them void. Section 3(3) presumes that effect for price fixing, output limitation, market sharing and bid rigging, cartels included. Section 19(3) lists the factors where no presumption applies. Section 46 allows a lesser penalty for a cartel member who discloses.
7. Oligopoly is lawful; agreeing is not. Parallel pricing alone does not prove an agreement.
8. No determinate solution exists, because the outcome depends on assumed reactions.

Test yourself

1. Define oligopoly and state its features. Oligopoly is a market structure in which a few sellers supply the whole or most of the output of an industry, each with a share large enough that its price and output decisions materially affect the others. Its features are the small number of sellers; mutual interdependence, which is the defining characteristic; a product that may be homogeneous or differentiated; significant barriers to entry, usually of scale or licence; price rigidity; heavy non price competition and advertising; an indeterminate demand curve; and a constant temptation to collude.

2. Explain the kinked demand curve and what it is used to prove. The kinked demand curve, associated with Sweezy in 1939, rests on the assumption that rivals will match a price cut but will not follow a price rise. The firm's demand curve is therefore relatively elastic above the prevailing price, because a rise loses many customers to rivals who hold their price, and relatively inelastic below it, because a cut is matched and brings little extra custom. The two segments meet in a kink at the prevailing price, which produces a vertical discontinuity in the marginal revenue curve. Marginal cost may move up or down within that gap without changing the profit maximising output, so the price remains unchanged despite changes in cost, which is the price rigidity the model is used to explain.

3. Why is there no single determinate theory of oligopoly? Because each firm's best decision depends on what it expects its rivals to do, and its rivals are reasoning in exactly the same way about it. The outcome therefore depends on the reaction pattern assumed, and different assumptions yield different results: Cournot assumed rivals hold output constant, Bertrand assumed they hold price constant, Stackelberg modelled a leader and a follower, and Sweezy assumed asymmetric reactions to rises and cuts. Modern treatment uses game theory, which formalises the interdependence rather than removing it.

4. What is a cartel, and how does Indian law treat one? A cartel is an agreement among rival enterprises to fix prices, limit output or supply, share markets or rig bids. Under section 3(1) of the Competition Act 2002 an agreement causing or likely to cause an appreciable adverse effect on competition is prohibited, and section 3(2) makes it void. Section 3(3) expressly includes cartels and presumes such an effect where the agreement fixes prices, limits production or supply, shares markets or results in bid rigging, so that the burden shifts to the parties. The proviso protects genuine efficiency enhancing joint ventures, and an amendment of 2023 extends the presumption to a participant who is not in the same trade but furthers the agreement.

5. What is bid rigging, and how would you recognise it from bidding data? Bid rigging is defined in the Explanation to section 3(3) as an agreement between enterprises engaged in identical or similar production or trading which has the effect of eliminating or reducing competition for bids or of adversely affecting or manipulating the process of bidding. The signs in the data are quotations clustered within a very narrow band where costs differ; a rotation of winners across tenders; consistent losing bids by firms that never win but always participate; identical arithmetical errors or formats; and sudden withdrawal of bidders in favour of one another. None is conclusive on its own, and the presumption operates only once an agreement is established.

6. Why is section 46 of the Competition Act 2002 effective against cartels? Because a cartel is inherently unstable. Each member can gain by secretly selling more than its allotted share at slightly below the agreed price, and the higher the cartel price the greater that temptation. Section 46 allows the Commission to impose a lesser penalty on a party who makes a full and true disclosure of the violations, provided the disclosure is vital. That converts the members' mutual distrust into a race to confess first, and it is the mechanism by which most cartels are actually detected, since direct evidence of the agreement is otherwise very hard to obtain.

7. "Similar prices prove collusion." Comment. They do not. Where several firms sell an identical product with similar costs and can observe each other's prices, competition itself drives prices towards one another, and a market in which prices differed widely for the same good would be the surprising one. Parallel behaviour therefore has to be distinguished from agreement. What supports an inference of agreement is a pattern competition does not produce: a rotation of successful bidders, simultaneous increases unrelated to any change in cost, quotes clustered far more tightly than the firms' costs differ, or evidence of communication between them.