Yes, Markets Sometimes Fail, and Governments Fail Right Alongside Them

Let me say something that will get me uninvited from a few conservative dinner parties: markets fail.

There. I said it, and the roof did not fall in. Free enterprise is the most productive arrangement ordinary people have ever stumbled onto, and I will defend it with everything I have. But a defence that begins by pretending the other side has no argument at all is a defence that loses the first time a smart professor shows up. The professor has an argument. It is a good one. It has been taught in every introductory economics course for a century. And the way to beat it is not to wave it away. The way to beat it is to take it seriously, state it better than its own friends do, and then ask the question the professor never asks: compared to what?

Because the textbook case for intervention has a hidden second half. It describes a market that fails and then assumes a government that succeeds. Nobody checks whether the government in the story is made of angels. This post checks.

The case for intervention at its strongest

Here is the argument as a fair-minded economist would make it, and I am not going to water it down.

Externalities. Some costs and benefits land on people who were not part of the deal. A factory sells steel to a buyer at a price both accept, and the smoke drifts over a neighbourhood that was never asked. The price of the steel leaves out the cost of the smoke, so the market makes too much of it. The English economist Arthur Pigou made this argument famous in the 1920s, and economists still speak of a "Pigouvian" tax: charge the factory for the damage, and the price will tell the truth. The same logic runs in the other direction. A neighbour who keeps up his house raises the value of the whole street and is never paid for it, so there may be too little of that.

Public goods. Some things are costly to provide but impossible to sell one customer at a time. Nobody can build a national defence for you alone and leave your neighbour undefended. Because you cannot exclude the neighbour, he has every reason to let you pay. Multiply that by three hundred million people and the market delivers too little. The Constitution's Preamble lists "provide for the common defence" among its purposes for exactly this reason, and Article I, Section 8 gives Congress the power to establish post offices and post roads.

Monopoly. If one seller controls the market, he can raise prices and cut quality, because the customer has nowhere else to go. Competition is the discipline that makes the market work, and a monopolist has escaped it.

Information problems. The seller knows things the buyer cannot check. A used-car buyer cannot see inside the engine. A patient cannot grade a surgeon's work in advance. Where one side holds all the knowledge, the argument runs, honest dealers can be driven out by dishonest ones.

That is the case. It has four legs, it has real examples behind each, and anyone who tells you it is simply wrong has not read it. Now let me tell you where it leaves a gap you could drive a truck through.

Compared to what? The missing half of the textbook

Notice what every one of those four arguments does. It identifies a flaw in a market and then jumps to a conclusion: therefore, government should step in. It never pauses to ask whether the government doing the stepping has the knowledge, the incentives and the discipline to make things better.

That pause has a name, and it earned a Nobel Prize. It is called public choice theory, and its founders were James Buchanan and Gordon Tullock. Their book The Calculus of Consent, published in 1962, started from an idea so simple that it took a generation of economists to absorb it: the people who run government are people. They respond to incentives. They want to be re-elected, promoted, funded, praised and left alone. They are not saints who become saints when they take the oath of office, and the same professor who is sure that a corporate executive follows his own interest will assume a regulator follows only the public's. Buchanan received the Nobel Prize in Economic Sciences in 1986 for his work on the contractual and constitutional foundations of political decision-making.

Public choice does not say government is bad. It says government should be analysed the way we analyse everything else: as a set of people with incentives, limited knowledge and limited budgets. Once you do that, the picture changes.

1. Government has a knowledge problem. A regulator setting a pollution standard for ten thousand factories does not know what each one can do cheaply. The factory does. The price system carries that knowledge; a rulebook does not. Hayek Explained Why the Smartest Planner in the Room Always Loses makes this case at length.
2. Government has a concentrated-benefit problem. The people who gain from a rule know it and organise. The people who pay a few cents each do not. So rules drift toward the organised. Economists call the pursuit of such favours rent-seeking; Tullock described the cost of it in a 1967 paper on tariffs, monopolies and theft, and the idea has been in every serious textbook since.
3. Government has a capture problem. George Stigler, who won his own Nobel in 1982, argued that regulation is often acquired by the industry and designed for its benefit. A large company can afford a compliance department and a lobbyist; its small rival cannot. A rule written to protect the public can end up protecting the incumbent. That is why so many businesses ask for regulation, and it is a fact the other side's textbook skips.
4. Government has no bankruptcy. A failing business shuts down and its resources go to someone else. A failing program has a constituency, and the constituency hires a lobbyist. Failure becomes a reason for a bigger budget.

I'd add a fifth, and I mean it as an observation, not a joke: government fails with other people's money and calls the failure a pilot program.

Where the textbook case is right, and conservatives say so

Here is where I part company with the cartoon version of my own side. A principled conservative does not say government should do nothing. He says government should do what it must, and do it with eyes open. The record shows conservatives, again and again, accepting a role where the case is strong.

• National defence. The clearest public good there is, and the first duty named in the Preamble.
• Courts and contracts. A market needs someone to enforce a promise. Without a neutral referee there is no trade, only force.
• Antitrust. The Sherman Antitrust Act of 1890 was signed by Benjamin Harrison, a Republican, and Theodore Roosevelt's administration used it to break up the Northern Securities Company, a railroad holding company; the Supreme Court upheld the government in 1904. Whatever you think of Roosevelt's wider record, conservatives have never believed a monopoly protected by the state, or by a cartel, is the free market.
• Clear property rights in the air and water. Pollution is the standard externality. A neighbour's smoke is an invasion of your property, and the common law of nuisance understood that long before there was an alphabet soup of agencies.
• Using the market to fix the market. In the 1990 amendments to the Clean Air Act, signed by President George H. W. Bush, Congress created a cap-and-trade system for sulfur dioxide from power plants: a national cap on the pollutant, with the right to emit it bought and sold. Whatever you think of that program's later descendants, the principle was a conservative one: set the limit, let prices find the cheapest way to meet it, and stay out of how.

That last example matters, so let me draw the lesson. The question is not market versus government. It is which tool, by whom, at what level. The economist Ronald Coase, in his 1960 paper "The Problem of Social Cost," showed that when rights are clear and bargaining is cheap, neighbours can often solve an externality themselves without any agency. He won the Nobel in 1991. His point was not that government is never needed. It was that the choice is among imperfect arrangements, and the cheap, local, voluntary one deserves a look first.

The test: five questions before you reach for the lever

Here is a test I would ask every candidate, every think tank and every cousin at Thanksgiving to apply. It is not a magic formula. It is a discipline. Call it the intervention test.

1. Is the failure real, or is it just a result someone dislikes? A market that produces a price you resent has not failed. A market that charges a third party for a cost he never agreed to, or cannot supply a good nobody can be excluded from, has. Name the failure in one sentence. If you cannot, you do not have one.
2. Has a cheaper fix been tried? Clearer property rights, a lawsuit, a contract, a private standards group, a local ordinance. The lower the level of government, and the more voluntary the remedy, the better.
3. Does the government have the knowledge and the incentives to do better? Not "could an ideal government do better," but "will this agency, staffed by these people, answerable to this Congress, do better than the flawed market?" Use the real government, not the one in the diagram.
4. What does it cost when it goes wrong, and who pays? Every program has a failure mode: it grows, it is captured, it outlives its purpose. Ask what the failure mode of this one is and whether anyone can end it. A rule with a sunset clause is a different animal from a rule with none.
5. Can the benefit be measured? If nobody can say in advance what success looks like, nobody can say afterward that it failed. A program with unmeasurable goals will survive forever.

Run that test on a few real policies and you will find that some pass. Fraud laws pass. A national defence passes. A rule against dumping poison in a river passes with ease. Others do not: the ones that protect a protected industry from its competitors, or set prices below what it costs to supply the good and then wonder about the shortages. price gouging laws sound fair and then the shelves go empty walks through that last example, and Ten Questions to Ask Before You Support Any New Government Program is the long version of the test.

A worked example: the river

Let me label this clearly as a hypothetical. Picture a small town on a river, with a paper mill upstream and a fishing lodge downstream. The mill's discharge makes the water cloudy. The lodge loses customers.

The textbook says: externality, so regulate. Fine. Let me run the test.

Is the failure real? Yes. The mill is imposing a cost on the lodge that neither of them priced.

Has a cheaper fix been tried? Here is where the common law earns its keep. The lodge owner has property rights in the water running past his land, and in many places a nuisance suit has been available for centuries. Or the two could simply bargain. Perhaps the mill installs a filter, or the lodge accepts payment. Bargaining works well with two parties. It works badly with ten thousand, because nobody can organise them all. That is the honest limit of Coase's point, and he acknowledged it.

Can the government do better? With ten thousand affected households, probably yes: a standard enforced by a state agency, or a permit system with a cap. But here the public choice questions kick in. Who writes the permit? The mill's lawyers will be at the table every day. The lodge owner has a business to run. So the design matters: transparent limits, public monitoring, penalties that actually bite, and the right to sue if the agency sleeps. And a lower level of government may know the river better than a distant one.

That is the whole method. It does not announce a winner in advance. It asks what is the problem, who knows what, and who has the incentive to get it right. Sometimes the answer is a rule. Sometimes it is a lawsuit. Sometimes it is nothing, because the cure would cost more than the disease.

The strongest objection, answered

The best objection to everything I have said runs like this: "Government failure is a reason to fix government, not to give up on it. You are using the imperfection of the cure to excuse the disease. People got sick from food, workers were maimed, rivers burned, before regulation. You cannot just say the market will sort it out."

That is a serious argument, and I will give it two honest answers.

First, I agree with half of it. The remedy for bad government is better government, and conservatives have a long record of trying: the reforms of the 1970s and 1980s that deregulated airlines, trucking and railroads are among the great success stories of bipartisan economics. The Airline Deregulation Act of 1978 was signed by a Democratic president, Jimmy Carter, and it phased out the Civil Aeronautics Board, which had spent decades deciding which airline could fly which route. That is government failure being fixed, and it is the kind of reform public choice theory predicts will be hard and worth doing.

Second, the objection proves less than it appears to. The cases where the market clearly failed were also cases where the remedy was often available at a more modest level: contract and tort law, local rules, inspection by private certifiers, a reputation that a seller cannot afford to lose. The real argument is over which cases need a federal agency with the power to write binding rules for fifty states, and which do not. Reasonable people on my side and theirs can draw the line in different places. What I will not accept is the claim that the line is self-evident, or that anyone who draws it near the bottom is a heartless extremist. Regulation protects us, so why count the cost? Because someone pays it is my fuller answer to that last charge.

What this means for how you argue

When a relative tells you "the free market failed, so government must fix it," do not say "markets never fail." You will lose, and you should. Say instead: "You may be right that it failed. Tell me what exactly failed, who is going to fix it, how they will know whether it worked, and what happens when it doesn't." You have just asked for the other half of the argument, and you did it politely.

When a relative tells you "government is the problem, always," do the same thing to him. Ask what he wants done about a river on fire. He will have an answer, and it will turn out to involve a rule of some kind.

Adam Smith, who gets claimed by everyone, is a good guide here. He defended competition passionately and did not trust merchants who gathered together to fix prices; Adam Smith Was Never the Cheerleader for Big Business Liberals Claim explains why. Thomas Sowell, who has spent a career saying there are no solutions, only trade-offs, is another; Thomas Sowell Asks What Happens After the Policy Passes is a good place to start. Both would recognise the test above.

The bottom line

• Markets can fail in four real ways: externalities, public goods, monopoly and information gaps. Say so, and you earn the right to be heard on everything else.
• Governments fail in four equally real ways: poor knowledge, concentrated benefits, capture and no bankruptcy. Anyone proposing a cure must show that the cure is not worse than the disease.
• Conservatives accept a government role where the case is strong, from national defence to antitrust to market-based pollution limits. The fight is over the line, not over whether a line exists.
• Before you back any new program, ask five things: is the failure real, has a cheaper fix been tried, can this government do better, what happens when it goes wrong, and can success be measured?
• At the family table, ask the other side the question they skip: compared to what?