Market Power II: Antitrust and Regulation

Part A built the price-searcher tool. Recall the two ideas we lean on here. First, most real firms are price-searchers, not price-takers: unlike a wheat farmer who must accept whatever the market pays, a price-searcher sets its own price and faces a downward-sloping demand, so it has some pricing discretion. That discretion is nearly universal and, on its own, mild, because real competition is a condition of a market, the pressure of who could enter, not a head count of how many sellers you can see at one moment.

Second, and this is the hinge of the whole topic, what makes market power durable and costly is not the number of sellers but whether the market is open or closed. A market is open when nothing but cost stands between a would-be competitor and entry; it is closed when entry is blocked by something else: a license, a patent, an exclusive franchise, a tariff, an acreage cap, a professional board. An open market keeps even a lone seller honest, because charging too much invites entry; a closed market lets the protected seller hold prices up and let costs drift, because the door behind it is locked. And closed doors are usually held shut by government, not by the market. We also saw that firms sometimes try to fake a closed market by colluding, forming a cartel to hold prices up, but a purely private cartel tends to cheat itself to death unless the law enforces it. Keep that open-versus-closed distinction in hand; everything in this post turns on it.

Antitrust Law Often Protects Competitors Rather Than Competition

The natural response to market power and cartels is a law against them. The United States built one, beginning with the Sherman Act of 1890 and extended by the Clayton Act and the Federal Trade Commission Act in 1914, enforced by triple damages and two federal agencies, aimed at stopping firms from “restraining trade,” “monopolizing,” and “fixing prices.” The trouble starts with the words: “Restraint of trade,” “monopolize,” and “price-fixing” are nowhere defined with any precision, and a share of a market, a concentration ratio, proves nothing by itself about whether buyers are harmed. Ten firms making most of the concrete blocks in a region might be colluding, or might simply be reading the same market conditions and arriving at similar prices independently, which is what competitors in any industry do.

The economist’s test is clean, even if the law’s is not: does the practice raise or lower the worth delivered to consumers? By that test antitrust enforcement has often punished the very behavior that helps consumers. Competition is a condition of the market, not a head count of firms, and it works precisely by eliminating competitors: the efficient win customers and the inefficient fold, so a law that fights to keep competitors alive can end up fighting competition itself. The prosecuted cases bear it out: the A&P grocery chain was pursued for its low prices and efficient buying at a time when four-fifths of American groceries were sold by other stores; Morton Salt, Standard Oil, and Borden milk were prosecuted for giving quantity discounts to large buyers, that is, for passing along the real savings of bulk; Microsoft was prosecuted in the United States and Europe for bundling a free browser into Windows on a “leveraging dominance” theory, with the market drawn narrowly enough to leave out Apple and Linux; and India once defined any firm above an asset threshold as a “monopoly” and blocked successful firms from expanding, driving entrepreneurs abroad, until repeal in the 1990s. Each firm was penalized for the things competition is supposed to produce: lower prices, bigger discounts, efficient scale. The clear benefit of antitrust is narrow: the flat, per se ban on outright price-fixing agreements, the one practice with no efficiency defense. Around that core the law uses a rule of reason weighing a practice’s actual effects, and it is in those judgment calls that protecting competitors gets mistaken for protecting competition.

Three distinctions keep the analysis sharp. A coalition is not the same as collusion: firms that form a partnership, set a common technical standard, or pool research can raise consumer value and do no harm, while collusion is the pretense of independent rivalry that secretly suppresses it and lowers consumer worth. A “tacit agreement,” likewise, is not a contract: if I cut my price expecting you to match it, I am reading the competitive landscape, not signing anything that binds us, yet “agreement” loosely used lets a lawsuit dress that up as conspiracy. And it is no accident that so many prosecuted collusion cases involve sales to government, which often buys through sealed public bids that make a cheater’s lower bid easy to spot, and whose purchasing agents, spending other people’s money, have weaker incentives to drive a hard bargain than a buyer spending his own.

The same confusion of a competitive pattern with a conspiracy shows up in how courts read prices that vary by location. A jury once awarded over a billion dollars against a group of southeastern plywood mills, persuaded that customers had been charged “phantom freight,” shipping costs on freight that was never hauled. The pricing it condemned is called basing-point pricing, and once you see the arithmetic it is nothing but competition. A mill ships to many cities, and each city’s price is its own market-clearing price. What ties those prices together is that the shipper nets one mill price everywhere it sells, so the price a buyer pays differs from city to city by exactly the freight it takes to reach him. If it costs $10 to reach one city and $6 to reach another, the two prices differ by $4, no more and no less. “Base price plus freight” is just the bookkeeping that describes this pattern; it is not a formula a group of sellers sat down together to fix.

The same misreading can convict a surgeon. Suppose one surgeon is as safe as any other but works twice as fast, and charges the same $500 standard fee for an operation that takes him half the time. Accuse him of billing “phantom time” and you have assumed that his hour is worth no more than anyone else’s, which is exactly what his speed disproves. He is paid for what he accomplishes, not for the time he spends, and the extra he earns over a slower colleague is a Ricardian rent.

The plywood mills were in that same position. Being nearer the timber, they save on freight, and competition bids that saving into the value of their land and their mills rather than into lower prices for local buyers, at least until the total supply of plywood grows enough to pass it along. The jury looked at a competitive pattern and saw a conspiracy. But the differential the mills earned is a Ricardian rent, the ordinary return to a better location, and it is not the monopoly rent we separated out in Part A, which comes only from an artificial restriction on entry. Courts misreading the economics is the running theme of this whole section.

Vertical Restraints Can Protect Service and Competition Between Brands

Agreements between a manufacturer and its dealers are vertical restraints, and their effects cannot be read as if rival manufacturers had formed a cartel. A minimum resale price can stop a bare-bones discounter from inviting customers to inspect a product at a full-service dealer, use its demonstrations and expert advice, and then buy cheaply from the dealer that supplied none of those services. If that free-riding drives the service out, the low sticker price can leave buyers worse off. A maximum resale price does the opposite job, preventing a protected dealer from adding a markup that shrinks sales of the manufacturer’s product.

Exclusive territories can likewise make dealers invest in local promotion without fearing that another dealer will wait nearby and capture the customers the first one educated. A brewer assigning territories may be organizing competition between its brand and rival brands, not ending competition among dealers. The manufacturer and dealer remain in tension, because each wants the other to provide service at a price that expands the joint stream of sales. Treating these arrangements as automatically illegal confuses coordination within a distribution chain with collusion among rival chains. The right question is the familiar one: does the restraint raise or lower the worth delivered to consumers?

Mergers admit an equally testable distinction. If a proposed merger mainly creates cartel power, firms left outside it gain from the higher industry price, so their stock prices should rise too. If the merger mainly creates efficiencies, the merging firms should rise relative to their rivals, whose competitive position has worsened. The rival-stock response is not a perfect verdict, but it forces the theory to make a prediction instead of merely attaching the word “anticompetitive” after the fact.

Few Firms and Network Effects Do Not by Themselves Prove Collusion

An oligopoly is a market in which each of a few large sellers anticipates how the others will react. That strategic interdependence is real, but the category alone predicts neither rigid prices nor successful collusion. A price cut may trigger matching, expansion may provoke entry, and a common price movement may reflect common costs or demand rather than an agreement. “Few firms” is an observation to explain, not a conclusion that the firms form a cartel.

Network effects add another reason a market may be concentrated without being closed. A communication system, payment network, or software format becomes more useful as more people join it, while learning a new system or moving data creates switching costs. Those facts give an incumbent an entry advantage, but an advantage is not a legal barrier and does not make displacement impossible. Users switch when a sufficiently better alternative, a compatibility bridge, or a coordinated migration makes the gain exceed the transition cost. The open-versus-closed test still matters: high switching costs explain why entry is difficult; they do not by themselves prove that exclusion has occurred or that one network is a permanent natural monopoly.

Regulators Created to Restrain Monopoly Tend to Get Captured by It

The other policy lever is direct economic regulation: an agency that sets the prices and terms a firm may offer, usually justified as protecting consumers from a monopoly. Take the rationale seriously, because that is what makes the failure legible. Some standardized services — electricity, water, gas, local phone — are supplied more cheaply by one firm running a single system than by several firms stringing duplicate wires and pipes down the same street. Economists call that a natural monopoly, and it is why a local government deliberately creates a closed-market utility and then regulates it: not only to cap the price a lone supplier could charge, but to assure dependent customers reliable delivery at predictable prices, so a household or business that sinks money into wiring and machines is not later expropriated. The rationale is real, but an agency built on it still drifts, and its core task — finding the “true” competitive price the missing market would have produced — is often impossible. The recurring result is the opposite of the intent, for a reason economists call regulatory capture.

The logic is about who pays attention. The public a regulator is meant to protect is large, diffuse, and lightly affected per person, and the crusaders who created the agency win their headline and move on; the regulated firms are few, intensely affected, and permanently engaged, so over time they dominate the agency’s hearings, staff, and rule-making, and the regulator drifts from restraining the incumbents toward shielding them from newer, more efficient rivals. The agency that was supposed to hold prices down for consumers ends up holding competitors out for producers.

The histories are blunt about it. The Interstate Commerce Commission was created to stop railroads from charging monopoly prices; when trucking arose and threatened them with genuine competition, the ICC got authority under the Motor Carrier Act of 1935 to hobble the more efficient truckers, and after Congress cut its power over trucking in 1980 freight charges fell substantially. The Civil Aeronautics Board did the same for airlines, keeping fares high and competitors out until it was abolished, after which fares fell and planes flew fuller, as Europe later saw too. The protected industry, not the consumer, was the regulator’s real client.

Sincere consumer protection still faces a two-error tradeoff. A drug regulator can approve a harmful medicine too soon or delay a beneficial medicine too long; reducing either error generally raises the risk of the other, so “more caution” is not costless. Safety rules can also induce offsetting behavior: when a device lowers the perceived danger of an activity, people may drive faster, pay less attention, or take other risks, giving back part of the engineering gain. The relevant comparison includes both kinds of error, their base rates, and how people change behavior after the rule. Counting only the visible accidents a rule might prevent is not enough.

The consumer-protection label also creates an incumbent advantage. Restrictions on margarine, Sunday sales, advertising, and entry into medicine were defended as shields for buyers or community standards, yet each also protected established sellers from a cheaper product, a more convenient seller, better information, or a new practitioner. Grandfathering incumbents is the clearest tell: a rule that exempts the people already supplying the allegedly dangerous service is poorly designed as a competence test and well designed as an entry barrier. This does not prove every protection is insincere; it means the protective purpose and the competitive effect must be analyzed separately.

Nor should regulation’s cost be measured by the agency’s budget alone. When a rule makes firms redesign products, keep records, wait for approval, or abandon an activity, those compliance costs are real resources absorbed by the policy even though they never appear as government spending. Regulation can therefore operate like an off-budget tax: lawmakers impose a costly obligation whose burden falls through higher prices, lower returns, or forgone output while the official budget records only the administrators’ salaries.

A close cousin of capture is the case for prohibiting particular marketing tactics, “predatory pricing” above all. The theory says a deep-pocketed giant prices below cost to bankrupt rivals, then jacks prices up once alone; it sounds plausible and almost never happens. The predator loses money heavier than its smaller prey while the war lasts; the bankrupt firm’s plant and workers do not vanish, they get bought cheaply and re-enter, often stronger with lower costs; and recoupment afterward requires barriers to entry an open market does not supply. Asked for a single documented case of predatory pricing succeeding, a Nobel laureate said he knew of none. A lower price from a genuinely lower-cost entrant is not predation but efficient competition, which is exactly why the “predatory” charge is so often aimed at it. And the cost the regulator must find to set a “fair” price often does not exist: the swing between an off-peak kilowatt-hour and a peak one is many times over, so there is no single “the” cost a commission can price against, as California learned in 2001 when average-cost rate-setting collided with above-average generation cost and the lights went out.

Durable Cartels and Producer Protections Are Political Arrangements

Private cartels are unstable because each member can gain by secretly cutting price and because a high cartel price invites entry. Government can solve both problems for producers: it can require participation, restrict imports, cap acreage or output, and deny licenses to potential entrants. The longest-lived cartels are therefore not private agreements hiding from the state; they are restrictions enforced by it.

That does not mean officials are outside the market correcting it from above. They control scarce permissions, contracts, subsidies, and legal barriers, and organized groups compete for those benefits. Suppress price competition and competition moves into lobbying, votes, campaign support, administrative hearings, and personal connections. Politics is a marketplace in that precise sense: incentives still operate, but the currency and the allocation rule change.

The distribution of stakes explains which restrictions persist. A sugar quota or rice tariff may cost each household only a few dollars while delivering a large gain to a small group of producers. The producers know exactly what is at stake and organize; millions of consumers each have little reason to learn the rule, much less campaign against it. Concentrated benefits and dispersed costs let a policy survive even when the total consumer loss exceeds the producer gain.

Farm programs make the mechanism visible. An above-market dairy support price creates the same surplus as any other price floor, so government must restrict herds, buy the excess, or let output go to waste. Import restrictions on sugar, peanuts, or rice use a different instrument but close the same door. The stated purpose may be food security or aid to farmers; the realized mechanism is restricted entry or supply, a higher price, and a transfer toward protected incumbents. The right diagnostic is not the policy’s label but what it does, who gains, and who pays.

The details of a farm restriction reveal distributive politics inside the protected group. Cotton acreage allotments based on a historical formula did not merely transfer wealth from consumers to growers; they determined which growers received the most valuable permissions, so regions fought over the base years and shares. The politics could turn circular: southeastern growers supported public irrigation that opened western land, then sought acreage restrictions to suppress the new rivals that the irrigation subsidy had helped create. A coalition can agree on restricting outsiders while fighting intensely over how the cartel rent is divided among insiders.

The citrus growers learned how fragile a private restriction is. When lemon growers tried to prop up their price by holding fruit off the market, the scheme collapsed into a scramble of holdouts, each grower better off letting the others withhold while he sold, the same cheating logic that unravels any cartel. So in 1941 they got the law to do what they could not: a statute let a majority of growers compel every grower to withhold, and government supplied the enforcement private collusion had lacked. That is this section’s opening claim made concrete, a durable producer restriction standing on public power rather than a private agreement.

What the growers won turned out to be less than it looked, for a reason worth a name. The propped-up price was a lure, and it drew each grower into planting bigger groves to claim a larger share of the limited sales that were allowed. The authorized fresh-fruit share fell from about nine-tenths of the crop in the program’s first year to under half, and the growers ended up no better off than before; the rent they had contrived was eaten by the excess capacity they built chasing it. That waste has a name, rent dissipation. The tobacco version hands the windfall a different way. An acreage license lodges the gain once, in the value of the land that held a license when the gate closed, so the original holder pockets it while every later entrant either buys a license or qualifies for one and earns only a normal return. That is why incumbents defend these restrictions so fiercely: repeal would hand them a capital loss. And the land barred from higher-valued crops is simply wasted.

The Same Price Jump Can Tell Three Different Stories

Put the tools to work on the most famous price jump of the last century. In 1973 and 1974 the cost of imported crude to American refiners multiplied several times over within a couple of years, from about $4 to $12.50 a barrel, and it happened precisely as the producing governments were tearing up decades-long concessions. Aramco’s original deal in Saudi Arabia had been a sixty-year grant; now Algeria took 51 percent of its industry in 1971, Iraq expropriated the Iraq Petroleum Company in June 1972, Libya seized foreign holdings and pressed for more, and the Gulf states negotiated “participation” climbing toward majority control. “OPEC is a cartel” became the reflex answer, and it is a possible one. But the cartel reading carries the Part-A problem we just watched sink the lemon growers: cartels leak, because every member gains by cheating. And the behavior at OPEC’s meetings fits a cartel poorly, since the small producers push for more restriction while Saudi Arabia plays the moderate holding output up, which is what a single large maximizer would do, not what a nervous cartel enforcer policing cheats would do.

Read the same jump a second way and no agreement is needed at all. Suppose one producer is so large that it sets its own output the way a single price-searcher would, while the smaller producers simply take the resulting price. The dominant firm’s demand is the market demand left over after subtracting what that competitive fringe will supply at each price. Then the big producer restrains its output to hold the price up and the small ones free-ride on that restraint, with no conspiracy to join and no cheating to police. Economists call such a seller a dominant firm.

There is a third reading, and it is the one most easily missed. The old concessions ran for fifty years and gave the companies no rights past expiry, and Iran had already shown, by cancelling a concession ahead of term in the early 1950s, that even the fifty years were not secure. A company pumping oil it may not own tomorrow has every reason to pump it fast today, so insecure rights meant over-rapid extraction, which is part of why oil was so cheap in the 1960s. When the producing governments took the fields back, just as the future value of oil was being revised sharply upward, the new owners held secure claims and did the opposite: they held oil back for a future now worth more, output fell, and the price rose with no collusion at all. Weighing a barrel sold today against a barrel kept for next decade is a present-value comparison, and the capital-values topic gives you the tool to make it precise.

So which story is true? All three fit the price path we actually observed, and honest analysis says so rather than pretending the data pick a winner. That is the useful lesson. Similar prices, a producer everyone watches for a lead, and a famous name are not evidence of collusion. What would be evidence is the costly enforcement machinery every durable cartel needs, the very thing we just saw the lemon growers reach for the law to supply. Absent that machinery, a price jump is a question, not a verdict. And when the price collapsed after 1981, falling by more than half within a few years, the durable-cartel-only story was the one that fit worst.

Technological Progress Displaces Some People, and Compensating Them Is Harder Than It Sounds

The same concentrated-benefits logic that shields a producer from foreign competition shields him from a machine. A new technology, take the rise of television, sorts people into three groups. Some earn higher wages working with the new thing. Some are untouched in their own work but gain as consumers of whatever the innovation makes cheaper or better. And some are displaced, pushed out of the old line of work and into their next-best jobs. That third group is not uniform. Many of the displaced land in jobs that leave them better off than before; others lose on this particular innovation yet still come out ahead of a world with no progress at all; and a few, usually older specialists whose skill was tied to the outmoded line, are left genuinely worse off.

Here is the tempting fix. Because the total gains from the innovation exceed the total losses, the winners could in principle compensate the losers and still come out ahead. That is logically airtight and practically treacherous. Nobody can say who lost how much, and the moment you offer to pay “the displaced” you invite everyone with a grievance to present himself as one of them. A principle that is clean on paper turns into a scramble the instant real money is attached to it.

The programs that actually exist compensate a visible, politically defined subset rather than the general run of losers. Retraining and relocation aid reach some of them. The Trade Expansion Act of 1962 created “trade adjustment assistance” for firms and workers hurt by imports, and it drew the line by cause rather than by harm: a worker qualified only if investigators traced his layoff “in major part” to trade concessions, so the worker displaced by a machine, or the trade-displaced worker who happened to live in a prosperous area, got nothing. Certifications swung from about 59,000 workers in 1975 to nearly 600,000 in 1980, the government’s own nine-year evaluation later found participants earning roughly $3,300 less than comparable workers who never enrolled, and since July 1, 2022 the program has been in phaseout, accepting no new petitions. And if we compensate workers for progress, why not the owners of the machines it renders obsolete? Compensation drifts from principle toward politics: the concentrated and visible losers get programs, the diffuse ones do not.

One caution before we move on. What displacement means for jobs and wages in general, why there is no fixed pool of work to be used up, and why a displaced worker’s search for his next job is itself a productive activity, is the labor topic’s story a few sessions from now. Here it is enough to notice that the politics of compensating losers looks just like the politics of protecting producers: the organized few are served, the unorganized many are not.

Why Sustained Private Monopoly Is Rare

The theoretical worry about monopoly is milder in practice than on the page. Economies of scale do make some industries the province of large firms, but diseconomies of scale eventually cap how big a single firm can efficiently get, as coordination and sluggishness past some size raise its costs. Beyond that, substitutes discipline even a sole producer, often coming from other industries entirely. Alcoa, for decades the only American maker of virgin aluminum, earned a modest return and watched its prices drop over the years, because it knew that pricing aluminum too high would send buyers to steel, tin, wood, and plastics. It was convicted under antitrust anyway. A market “share” overstates power whenever a buyer has somewhere else to go, as in an open economy buyers usually do, so genuine, lasting private monopoly with no government lock on the door is uncommon.

The very phrase “barriers to entry” deserves a second look, because a barrier to bad options is a service, not an injury. A reputation is a filter. A consumer leans on a seller’s proven record to predict quality precisely because trying every untested newcomer is costly, and a car every bit as good as a Cadillac, sold under a name nobody has heard of, would sell very few until it earned a record of its own. The filter would be worth tearing down only if every newcomer were exactly as good as the proven incumbent, and that premise is plainly false. What looks like a wall keeping good sellers out is mostly a screen keeping unproven ones from being mistaken for proven ones.

The same goes for the profits that entry has to earn. Trying to break into an industry means spending money you cannot get back if you fail: searching, testing, assembling a team, buying equipment with little resale value. Suppose a $1 attempt succeeds only half the time. Then a success has to bring in $2 just to cover the failures on average, which is why industries that require large forfeitable outlays show high profits among the firms that survive. That premium is the bait that induces anyone to try at all, not the mark of a locked door. Reading a survivor’s profits as proof of monopoly is the same error as reading the fast surgeon’s fee as an overcharge: in both cases a return to risk or to a superior resource is mistaken for a toll collected behind a closed gate.

The doors that genuinely matter are the ones locked by law, the open-versus-closed hinge from Part A. Cost, reputation, and risk are not locks; they are what competing has always meant.

Economics Explains These Practices Without Pronouncing on Them

One last reminder, the same one we have returned to all term. Everything in this topic is positive analysis. It explains why a firm discriminates, ties, bundles, colludes, or lobbies for a license, and it predicts the effects: this practice raises output, that one transfers a rent, this regulation protects an incumbent. It does not declare any of them fair or unfair, and it does not settle who has earned the gains from a trade. “Discrimination,” “dumping,” “predatory,” and “monopoly” all carry a sting in ordinary speech that the analysis itself does not supply. When you decide a practice is good or bad, that verdict comes from your values, laid on top of the economics, not out of it. The job here is to see clearly what each practice does; what you think ought to be done is yours to add.