Market Structures

An onion farmer near Nashik can sell at the local mandi, to a trader who phones from Mumbai, or through an app. None of these places is "the onion market" on its own. Together, they are.

So in economics a market is not a place. It needs five things:

• Buyers and sellers of the good.
• A commodity: one good, such as onions or cement.
• Contact between them, through bargaining or exchange, in person or at a distance.
• Knowledge of market conditions, such as the prices others are paying.
• One price that tends to prevail at a given time.

Why one price tends to prevail. Suppose, as an illustration, onions sell for ₹20 a kg in one mandi and ₹26 in a mandi nearby, and moving them costs ₹2 a kg. Traders buy in the cheaper mandi and sell in the dearer one. That raises the low price and lowers the high one, until the gap is no bigger than the ₹2 cost of moving the onions.

What decides a market's structure. Economists classify markets by four things:

• the number and size of sellers and of buyers;
• the nature of the product: homogeneous (every seller's product is the same) or differentiated (each seller's is a little different);
• the conditions of entry and exit;
• economies of scale.

Most of this chapter counts sellers. The same counting works on the buyers' side.

In a bilateral monopoly the price cannot be worked out from demand and supply. The seller will not go below its lowest acceptable price, and the buyer will not go above its highest. Bargaining power decides where the price falls between the two.

Don't confuse monopsony with monopoly. Monopoly is one seller, with power to push prices up. Monopsony is one buyer, with power to push the prices it pays down, such as the wages of workers in a one-employer town.

Where it breaks. Real markets rarely reach one price. Transport costs, taxes and poor information keep prices apart. Cheap phones and online platforms have widened many markets, so the "region" in Cournot's definition can now be a whole country.

For exams. Remember Cournot's definition and that a market is not a place. Match each buyer-side term to its number of buyers. In bilateral monopoly, price is indeterminate and settled by bargaining.

A sweet shop sells 100 boxes of kaju katli a day at ₹500. To sell 110 it must cut the price to ₹480 for every customer, not only the extra ten. The ten extra boxes do not bring in 10 × ₹480.

Back to the sweet shop. At ₹500 it takes 100 × ₹500 = ₹50,000. At ₹480 it takes 110 × ₹480 = ₹52,800. The ten extra boxes add only ₹2,800, or ₹280 a box. The shop gains ₹4,800 on the new boxes but loses ₹20 on each of the first 100, which is ₹2,000.

Because average revenue equals price, the AR curve is the firm's demand curve. The schedule below (an illustration) shows how MR behaves when price must fall to sell more.

Price falls ₹10 a step, but MR falls ₹20 a step. TR stops rising where MR reaches zero, and falls once MR turns negative.

• Perfect competition: the firm sells any quantity at the market price, so AR = MR = P. Both are one horizontal line, and TR is a straight line from the origin.
• Imperfect competition: to sell one more unit the firm must lower the price on all units, so MR is below AR. For a straight-line AR, the MR line falls twice as steeply. It cuts the quantity axis halfway to where AR meets it.
• MR can be negative. TR is at its maximum where MR = 0.

Revenue and elasticity. The two are linked by one formula, where e is the price elasticity of demand with its minus sign ignored:

Take a price (AR) of ₹100. If e = 2, MR = 100 × (1 − ½) = ₹50. If e = 1, MR = 100 × 0 = ₹0. If e = 0.5, MR = 100 × (1 − 2) = −₹100. So MR is positive where demand is elastic (e > 1), zero where it is unitary (e = 1), and negative where it is inelastic (e < 1).

Don't confuse maximum revenue with maximum profit. TR is highest where MR = 0. Profit is highest where MR = MC: where the extra revenue from a unit equals its extra cost. Since extra units usually cost something, a firm maximising profit stops before revenue peaks.

Where it breaks. The gap between AR and MR assumes one price for every buyer. A seller who can charge each buyer a different price does not lose money on earlier units. For such a seller, MR equals the price of the extra unit (see price discrimination, below).

For exams. AR is the demand curve. Under perfect competition AR = MR. For a straight-line AR, MR has twice the slope and bisects the quantity axis. Know MR = AR(1 − 1/e) and the signs it gives.

A wheat farmer cannot charge a rupee above the going rate. A large cement maker can hold its price for months. The difference comes from how many sellers each market has, how alike their products are, and how easily newcomers can join.

Two kinds of profit. Normal profit is the minimum return needed to keep a firm in an industry. Supernormal profit is any return above it. Take a kirana store, as an illustration. Its owner could earn ₹6 lakh a year by working elsewhere and lending out her savings. Then ₹6 lakh is her normal profit. If the store makes ₹9 lakh, ₹3 lakh is supernormal profit. If it makes ₹6 lakh, she earns only normal profit, though her accounts show a profit.

Supernormal profit is a signal. Where entry is free, it draws in new sellers until it is gone. That is why the long-run row reads "normal" for the two structures with easy entry.

Don't confuse monopolistic competition with monopoly. A monopoly is the only seller of a product with no close substitute. In monopolistic competition there are many sellers; each has a "monopoly" over its own brand only. A restaurant owns its menu and name, but diners can walk next door.

Where it breaks. The four types are models; real markets sit between them. One firm can face different structures: a cement plant may be one of a few sellers in its region but one of many nationally. The number of sellers also misleads when one seller is far larger than the rest.

For exams. Expect "match the feature to the structure" questions from the table. Remember: price taker in perfect competition, price maker in monopoly, normal profit in the long run wherever entry is easy.

Count the sellers, look at the product, test the entry.

At a busy wheat mandi, a farmer who asks a little above the going rate sells nothing, because buyers go to the next farmer. A farmer who asks less simply gives money away. Each farmer takes the price the market sets.

Under perfect competition every firm is a price taker. The industry sets the price through demand and supply. Each firm faces a perfectly elastic (horizontal) demand curve at that price.

Conditions:

• a large number of buyers and sellers;
• a homogeneous product;
• free entry and exit;
• perfect knowledge of prices and conditions;
• perfect mobility of factors of production, such as labour and capital;
• no transport or selling costs.

Edward Chamberlin called the first three, many sellers, a homogeneous product and free entry, pure competition. Perfect competition adds the rest.

Time matters. Alfred Marshall showed that what sets the price depends on how long sellers have to adjust supply.

Equilibrium. The firm produces where MR = MC, with MC rising. Since MR = P, this means P = MC.

Short run. The firm may earn supernormal profit, normal profit or a loss. It shuts down if the price falls below average variable cost (AVC), the cost per unit of inputs that vary with output, such as raw material. Take a small rice mill, as an illustration, selling 1,000 quintals a month. Its AVC is ₹2,500 a quintal and its average total cost ₹3,200, so its fixed cost is ₹7 lakh a month.

• At a price of ₹3,000 it loses ₹200 a quintal, or ₹2 lakh. If it closed, it would still pay the ₹7 lakh fixed cost. Running loses less, so it keeps running.
• At a price of ₹2,400, below AVC, each quintal loses ₹100 even before fixed cost. Running would lose ₹8 lakh against ₹7 lakh for closing, so it shuts down.

Long run. Supernormal profit draws in new firms; losses drive firms out. Entry and exit stop only when price equals the lowest point of average cost. Then P = AR = MR = MC = minimum average cost, and every firm earns normal profit.

Don't confuse perfect competition with pure competition. Pure competition (Chamberlin) needs only many sellers, a homogeneous product and free entry. Perfect competition also needs perfect knowledge, perfect mobility and no transport costs.

Where it breaks. Few real markets meet every condition. Farm produce comes close on product and numbers, but buyers and sellers rarely have perfect knowledge, and government support prices can set a floor. The model is mainly a benchmark against which other structures are judged.

For exams. Know the conditions, Chamberlin's pure competition, Marshall's three periods, the shut-down rule (price below AVC), and the long-run equality P = AR = MR = MC = minimum AC.

A company with a patent on a new medicine has no close rival, and can price well above its cost. A restaurant also sets its own prices, but if it charges too much, diners walk to the place next door.

Monopoly. The firm is the industry. Barriers keep rivals out:

• patents, licences and franchises granted by law;
• control of a key input;
• economies of scale so large that one firm is the cheapest supplier. This is a natural monopoly, such as a local power grid, where a second set of wires would only add cost.

Monopoly equilibrium. The monopolist picks output where MR = MC, with MC cutting MR from below. It reads the price off the demand (AR) curve above that output. It can set the price or the quantity, not both, because buyers decide how much they take at each price.

A worked example (an illustration). A firm holds a patent on a water filter. Demand is P = 120 − Q, with price in rupees and Q in thousands of units. So MR = 120 − 2Q. The marginal cost of each filter is a flat ₹40, with no fixed cost.

• Set MR = MC: 120 − 2Q = 40, so 2Q = 80 and Q = 40 thousand.
• Price from demand: P = 120 − 40 = ₹80.
• Profit: (₹80 − ₹40) × 40,000 = ₹16 lakh.
• Under competition, price would fall to MC: 120 − Q = 40, so Q = 80 thousand. The monopolist sells half as much, at twice the price.

Because MR is positive only where demand is elastic, a profit-maximising monopolist always produces on the elastic part of its demand curve. At the filter's price and output, elasticity is 2: each ₹1 on the price loses one thousand units, and 1 × 80 ÷ 40 = 2. Supernormal profit can last into the long run, since entry is blocked, though short-run losses are possible.

Monopolistic competition was set out by Edward Chamberlin in The Theory of Monopolistic Competition (1933). Joan Robinson's The Economics of Imperfect Competition appeared the same year. Its features:

• many sellers;
• product differentiation by brand, quality, design and packaging;
• close but not perfect substitutes;
• free entry and exit;
• heavy selling costs and non-price competition;
• a product group of similar products, rather than an industry of identical ones.

Each firm has a monopoly over its own brand only, which gives it limited control over price. Examples: soaps, toothpaste, restaurants, clothing.

Equilibrium. In the short run the firm sets MR = MC and may earn supernormal profit. Think of chai stalls outside a college. If they earn well, new stalls open and draw customers away. Each stall's demand curve shifts left until it just touches the average cost curve on its falling part. Price then equals average cost, so profit is normal. Each stall works below its lowest-cost output, leaving excess capacity: it could serve more cups at a lower cost per cup.

Don't confuse "price maker" with "can charge any price". A monopolist chooses a point on its demand curve. Raise the price and it sells less. Its power is the power to choose, not to escape demand.

Where it breaks. Monopoly profit attracts substitutes, new technology and regulators, so few monopolies last forever. Patents expire. The excess capacity charge against monopolistic competition is often answered this way: the extra cost per unit is the price of variety, which buyers value.

For exams. Monopoly: MR = MC, price from AR, always on the elastic part, supernormal profit can persist. Monopolistic competition: Chamberlin (1933), product group, tangency of demand with AC, normal profit, excess capacity.

Monopoly keeps its profit; brands compete theirs away.

When one of three cement makers in a region cuts its price, the other two feel it within days. Before any of them acts, each must guess how the others will react.

An oligopoly's marks are interdependence, barriers to entry and heavy advertising. Its demand curve cannot be known without guessing rivals' reactions.

• A pure oligopoly sells a homogeneous product: steel, cement, aluminium.
• A differentiated oligopoly sells branded products: cars, televisions.
• A duopoly is an oligopoly of two sellers.

Each firm is torn between two motives. It can act alone to maximise its own profit, or cooperate with rivals to maximise their joint profit. Picture two bus operators on one route, as an illustration. If one cuts its fare from ₹600 to ₹500, the other loses passengers unless it follows. Both then earn less.

Models of oligopoly. Each model makes a different assumption about how rivals react.

The models side by side (an illustration). Market demand is P = 140 − Q, and each unit costs ₹20 to make, for both firms. The Cournot "one-third" result holds for straight-line demand and equal, constant costs like these.

Collusion. Firms may cooperate openly through a cartel, which fixes a joint price and allots output quotas; OPEC is the best-known example. Or they may cooperate tacitly through price leadership, where others follow the price of one firm. The leader may be the low-cost firm, the dominant (largest) firm, or a barometric firm, one whose price changes others trust to reflect market conditions.

Don't confuse oligopoly with monopolistic competition. Both can have branded products. In monopolistic competition there are so many sellers that each can ignore the others. In oligopoly there are so few that none can.

Where it breaks. There is no single theory of oligopoly; the result depends on what each firm assumes about its rivals. Cartels are unstable, because each member gains by quietly selling more than its quota. Competition laws in many countries ban them; in India the Competition Act, 2002 prohibits anti-competitive agreements such as cartels.

For exams. Match each model to its author, year and assumption. Cournot: output, two-thirds. Bertrand: price, P = MC. Kinked demand: price rigidity. Know the three kinds of price leader.

A wheat farmer, a restaurant, a cement maker and a town's only power supplier all face the same question: how much to produce. They answer it with the same rule, yet end up with very different prices.

Every profit-maximising firm in every structure sets output where MC = MR. What differs is the price it can charge at that output.

Why the rule works. Take the water-filter monopolist from the previous concept, with MR = 120 − 2Q and MC = ₹40.

• At 30 thousand units, MR = 120 − 60 = ₹60. Each extra filter adds ₹60 of revenue and ₹40 of cost. Producing more adds profit.
• At 50 thousand units, MR = 120 − 100 = ₹20. Each extra filter adds ₹20 and costs ₹40. Producing less adds profit.
• Only at 40 thousand, where MR = ₹40 = MC, does no change add profit.

The rule in each structure:

• Perfect competition: the firm cannot influence price, so P = MR, and it equals MC at equilibrium. Price taker.
• Monopolistic competition: P > MR, so the firm has partial price control. It relies on product differentiation and heavy advertising.
• Oligopoly: firms are interdependent. The kinked demand curve says rivals match a price cut but ignore a price rise. So demand is flat (elastic) above the prevailing price and steep (inelastic) below it. The kink leaves a gap in the MR curve, which explains price rigidity.
• Monopoly: P > MR, with the greatest control over price, limited only by demand. Supernormal profit can persist because entry is blocked.

Reading the kink. As long as MC passes through the gap, MR = MC holds at the same output. A firm whose costs rise a little keeps its price, which is why oligopoly prices tend to stay put.

Don't confuse MR = MC with P = MC. MR = MC is the profit rule for every firm. P = MC holds only in perfect competition, where price and MR are the same. Elsewhere price sits above MC, and the gap is a measure of market power.

Where it breaks. Few managers know their MR and MC curves. Many set prices by adding a margin to average cost, and adjust by trial. Some firms pursue sales or growth rather than profit. The marginal rule still describes where a firm that does maximise profit ends up.

For exams. MC = MR in every structure; P = MC only under perfect competition; the kinked demand curve and its MR gap explain price rigidity in oligopoly.

MC = MR everywhere; the price differs.

A doctor charges ₹500 a consultation at her city clinic and ₹125 at a weekly village camp. The doctor, the time and the advice are the same. Only the buyers differ.

Three conditions. Price discrimination works only if all three hold:

• the seller has monopoly power, so buyers cannot simply go elsewhere;
• the markets can be separated, and resale between them prevented;
• elasticity differs across the markets.

Three degrees, named by A. C. Pigou in The Economics of Welfare (1920):

How third-degree pricing works. The firm sets MR equal in every market and equal to MC. It charges the higher price where demand is less elastic. Use MR = P × (1 − 1 ÷ e) from revenue concepts. Suppose, as an illustration, a consultation costs the doctor ₹100 at the margin.

• Village camp, e = 5: P × (1 − 1/5) = 100, so P × 0.8 = 100 and P = ₹125.
• City clinic, e = 1.25: P × (1 − 1/1.25) = 100, so P × 0.2 = 100 and P = ₹500.

If elasticities were equal, the two prices would be equal, and discrimination would not pay.

Dumping is selling a good abroad at a lower price than at home, or below its cost, to win a foreign market. Under WTO rules, an importing country may impose an anti-dumping duty when dumped imports injure its own industry. Reverse dumping is international price discrimination the other way: a firm sells abroad at a higher price than at home, usually because demand abroad is less elastic.

Don't confuse price discrimination with prices that reflect cost. A higher fare for an air-conditioned coach pays partly for a costlier service, so it is not pure discrimination. Likewise, a low export price is not dumping if the exporter's costs really are that low.

Where it breaks. Resale breaks the separation: if village patients could resell appointments in the city, the two prices would converge. The effect on welfare is mixed. Discrimination transfers surplus from buyers to the seller, but it can also serve buyers who would be priced out by a single price.

For exams. Know Robinson's definition, the three conditions, Pigou's three degrees (1920), the rule MR₁ = MR₂ = MC with the higher price in the less elastic market, and the difference between dumping and reverse dumping.