Thirty-three years after Milton Friedman made the case, it is time to run the experiment again.
In 1993, Milton Friedman published an essay with a deliberately provocative title:
Why Government Is the Problem.
He began with two tasks.
The easy one was to show the ways government had contributed to some of the country’s largest problems.
The difficult one was explaining why.
Why do intelligent, well-meaning, public-spirited people so often produce different results when operating through political institutions than they do through voluntary ones? Why do programs created to solve problems so frequently become permanent institutions for managing them?
Friedman proceeded to count the ways: education, crime, homelessness, family breakdown, housing, healthcare, finance, highways, airports and the accumulated weight of everything else government had decided it should manage.
Thirty-three years later, it seems only fair to rerun the experiment.
The names have changed.
The numbers have gotten larger.
The mechanism looks remarkably familiar. Friedman’s original essay divided the problem similarly: first identifying the outcomes he believed government had worsened, then turning to special interests, concentrated benefits, dispersed costs and institutional incentives to explain why.
Showing That Government Is Still the Problem
Education
Start where Friedman did.
America’s public elementary and secondary schools spent nearly $984 billion in fiscal year 2024. Current spending averaged $17,619 per student, up 6.6 percent in a single year.
What did we buy?
Not nothing.
That qualification matters. Fourth-grade math scores improved somewhat between 2022 and 2024.
But they remained below 2019 levels. Fourth-grade reading fell again in 2024 and sat five points below its 2019 level.
The argument is not that money does not matter in education.
Of course it does.
Teachers cost money. Buildings cost money. Books, buses, counselors, special education and technology cost money.
The question is why almost every conversation about struggling schools begins by asking how much more the system needs rather than whether the existing system is producing enough for what we already give it.
A business that raises its prices every year while its product deteriorates eventually loses customers.
A school district cannot lose its customers quite so easily.
That difference in incentives was Friedman’s point.
It remains ours.
Lawlessness and Crime
If there is one function almost everyone agrees belongs to government, it is protecting citizens from people who would hurt them.
And yet one of Friedman’s most controversial examples in 1993 looks less radical today: the War on Drugs.
After decades of prohibition, Americans have not been left with a drug-free country.
They were left with fentanyl.
The good news is that overdose deaths have recently fallen sharply. CDC provisional estimates put overdose deaths at about 80,000 in 2024, down nearly 27 percent from 2023. That is a remarkable and welcome improvement.
It is also still roughly 80,000 deaths in a single year.
Meanwhile, the federal criminal-justice system continues devoting enormous resources to narcotics enforcement. In fiscal year 2022, drug offenses accounted for 21 percent of federal arrests, and more than 19,000 people entered federal prison for drug offenses.
The government restricted the supply.
The market did not disappear.
It went underground.
There it became extraordinarily profitable to cartels willing to use violence, increasingly potent synthetic drugs and distribution networks that do not submit tax returns or ask the FDA for permission.
We criminalized addiction without eliminating addiction.
That is not order.
It is a reminder that outlawing a market does not repeal demand.
Homelessness
Friedman turned next to homelessness.
So should we.
HUD counted 771,480 Americans experiencing homelessness on a single night in January 2024, the highest number recorded in the modern federal count and an 18 percent increase from the year before.
There is no intellectually serious way to reduce every one of those people to a zoning ordinance.
Homelessness involves mental illness, addiction, domestic violence, poverty, family breakdown and individual circumstances that refuse to fit neatly inside anyone’s political philosophy.
But housing still has a price.
And government has spent decades making housing harder to build in many of the places where people most want to live.
Minimum lot sizes.
Density restrictions.
Lengthy permitting.
Parking requirements.
Development fees.
Environmental reviews.
Building restrictions.
Then, after restricting supply and increasing cost, government creates another program to make the resulting housing “affordable.”
The first intervention contributes to the problem.
The second intervention subsidizes the consequence.
The bureaucracy can then point to the continued existence of the problem as evidence that a third intervention is needed.
That cycle should at least make us uncomfortable.
Family Values
Friedman was careful about this section in 1993.
We should be too.
Government did not destroy the American family.
Culture matters. Economics matter. Technology matters. Individual choices matter.
But incentives matter too.
Census data continue to show a profound change in American family structure. In 2019, about 70 percent of children lived with two parents. The historical Census series shows a substantial decline from the family arrangements prevalent when Friedman was young.
Government does not have to order families apart to make forming a family more expensive.
Means-tested programs reduce assistance as household income rises. The tax code can produce both marriage bonuses and marriage penalties depending on a couple’s earnings. Federal researchers have long recognized that the design of programs such as TANF can create marriage disincentives in some circumstances.
That does not prove welfare caused the decline of marriage.
It proves something simpler:
Government can attach a price to behavior without intending to.
Design a program to help struggling families.
Create an income cliff.
Two people combine their household.
Their benefits disappear.
Then act surprised when incentives influence decisions.
Good intention does not repeal arithmetic.
Housing
Here, Friedman’s 1993 essay almost reads as though someone simply updated the dollar amounts.
He complained about zoning, permitting and building regulation increasing the cost of a house.
In 2026, the National Association of Home Builders estimated that regulations imposed by all levels of government account for 26.4 percent of the final price of an average new single-family home—about $131,734 in its model. NAHB is an industry group with an obvious deregulatory interest, but its methodology at minimum illustrates how significant builders believe the cumulative burden has become.
Some of those regulations have very good reasons behind them.
Buildings should not collapse.
Electrical systems should not start fires.
Developers should not be permitted to dump every cost they create onto their neighbors.
But each regulation still costs something.
And the family trying to buy its first house ultimately pays much of that cost.
We have spent years asking how government can help people afford housing.
Perhaps government should also ask what government did to the price.
Medical Care
In Friedman’s essay, healthcare consumed roughly 12 or 13 percent of national income.
Today, the United States spends $5.3 trillion a year on healthcare.
That is $15,474 for every person in the country.
Healthcare now consumes about 18 percent of GDP.
We have somehow constructed a system in which almost nobody knows what anything costs.
The patient does not know.
The doctor frequently does not know.
The insurer has one number.
The hospital has another.
Medicare has another.
The cash price may be something else entirely.
Employers buy insurance because federal tax policy helped cement health coverage to employment. Government programs then operate beside private insurers, each imposing different billing rules, documentation requirements and reimbursement systems.
The result is not a free market.
It is not a government-run system either.
It is a strange marriage of both in which everyone can blame everyone else.
And everyone does.
If $5.3 trillion has not made healthcare affordable, perhaps the first question should not be how to inject still more money into the same structure.
Perhaps we should ask what the structure is doing to the money already there.
Financial System
Friedman had a financial crisis to point toward.
We have debt.
CBO projects a federal deficit of $1.9 trillion in 2026.
Debt held by the public is projected to reach about 101 percent of GDP this year.
And the federal government is expected to spend roughly $1 trillion this year simply on net interest—more than mandatory spending on every individual federal program except Social Security and Medicare.
A trillion dollars.
Not to build a bridge.
Not to educate a child.
Not to equip a soldier.
Not to cure a disease.
To service yesterday’s promises.
Both parties have participated.
Both parties have discovered that spending money is politically easier than collecting enough taxes to pay for it.
The bill goes to someone who cannot vote yet.
Washington calls that fiscal policy.
A family would call it debt.
Highway Congestion
Friedman’s original comparison was almost comically simple.
Private industry could build cars faster than government could build enough roads for people to drive them on.
Thirty-three years later, America can build extraordinary things.
Our problem is increasingly how much they cost and how long they take.
The Transit Costs Project maintains data on hundreds of rail projects around the world. Its research has placed the United States among the world’s most expensive places to build rapid transit, with particularly poor comparisons once tunneling and project complexity are considered.
Its 2026 database reports a weighted average construction cost across the projects it tracks of about $252 million per kilometer.
The causes are not a single environmental law or one greedy contractor.
Procurement matters.
Design decisions matter.
Labor practices matter.
Consultants matter.
Project management matters.
Permitting matters.
Political interference matters.
That is almost the point.
When responsibility is divided among enough agencies, jurisdictions, consultants, contractors and approval processes, eventually everyone is responsible.
Which means no one is.
Airports
Friedman specifically complained about airports and air traffic control.
This week—thirty-three years later—the Government Accountability Office published another report on the Federal Aviation Administration’s attempt to modernize the air traffic control system.
GAO said the FAA has struggled for decades to modernize outdated systems and warned that the newest effort still lacks a comprehensive lifecycle cost estimate and master schedule. More than 75 percent of the FAA’s 138 air traffic control systems had previously been rated unsustainable or potentially unsustainable.
Think about that.
Boeing can design an airplane.
Airbus can design an airplane.
Space companies can land rocket boosters on autonomous ships.
A passenger can buy a ticket from Texas to Tokyo from a phone while sitting on the couch.
Yet parts of the national air traffic system have spent decades waiting to be modernized.
Friedman would recognize the paragraph.
He might only be surprised that we are still writing it.
Miscellaneous
We have not discussed agricultural subsidies.
Or tariffs.
Or student lending.
Or occupational licensing.
Or the Postal Service.
Or federal procurement.
Or the thousands of pages of rules agencies issue every year.
The Competitive Enterprise Institute—a deregulatory advocacy organization—counts 2,441 final federal regulations issued in 2025 against 133 laws enacted by Congress, though many of those rules were themselves deregulatory or administrative rather than new burdens.
The raw ratio does not prove that every regulation is illegitimate.
It illustrates the extraordinary amount of governing that now occurs outside the simple civics-book process of Congress debates, Congress votes, President signs.
Government has a legitimate role.
Friedman said so.
So should we.
Government should protect citizens from coercion, defend the country, enforce contracts, establish predictable rules and provide institutions through which disputes can be settled. His concern was that the more functions government accumulates beyond its core responsibilities, the less effectively it can perform even those tasks.
That brings us to the harder question.
Explaining Why Government Is Still the Problem
It would be comforting if the explanation were bad politicians. Vote them out. Problem solved.
It would be comforting if it were incompetent bureaucrats. Fire them. Problem solved.
Friedman’s argument was more troubling.
The people are not necessarily the problem. The incentives are.
Put a competent person into a system with bad incentives and eventually the incentives win.
A company that produces something nobody wants eventually runs out of customers.
A government program can produce disappointing results and use those disappointing results as evidence that its budget should increase.
A private organization that wastes its own money harms its owners.
Government wastes someone else’s.
Then there are special interests.
Suppose a federal program gives one industry $1 billion while spreading the cost across 340 million Americans.
To the average citizen, fighting it is worth roughly three dollars.
To the industry receiving it, keeping it is worth one billion dollars.
Who is going to hire the lobbyist?
Who is going to attend the hearing?
Who is going to commission the study?
Who is going to call the congressman every week?
The person with three dollars at stake?
Or the person with a billion?
That is the concentrated-benefits, dispersed-costs problem Friedman emphasized.
And the same logic appears everywhere.
The future homeowner who will pay more because of a zoning restriction does not know he is a future homeowner yet.
The existing homeowner who does not want an apartment building nearby knows exactly what he wants.
The taxpayer paying a few dollars toward a subsidy has little reason to organize.
The recipient whose business model depends on that subsidy has every reason in the world.
The government employee whose livelihood depends on a program knows precisely what will happen if that program disappears.
The citizen paying one ten-thousandth of its cost probably does not know the program exists.
None of these people need to be corrupt.
They are behaving rationally.
That is what makes the problem difficult.
The Reverse Incentive
Markets are not perfect.
Companies lie.
Executives make stupid decisions.
Consumers make stupid decisions.
Fraud happens.
Monopolies happen.
Markets produce externalities government sometimes must address.
The distinction is not that private actors are virtuous and public actors are stupid.
It is what happens next.
Failure in a competitive market at least contains a mechanism capable of killing the failed institution.
Customers leave.
Competitors arrive.
Capital disappears.
Bankruptcy exists.
Government possesses no equivalent mechanism nearly as powerful.
Instead, failure can become evidence of need.
The school is struggling, therefore it needs more money.
Housing is unaffordable, therefore it needs another subsidy.
Healthcare is expensive, therefore government must spend more on healthcare.
Infrastructure takes too long to build, therefore create another office to coordinate the offices coordinating the project.
The homeless program did not end homelessness, therefore the homeless program clearly remains necessary.
Every problem produces an institution.
Every institution produces employees.
Every employee produces a constituency.
Every constituency eventually develops an argument for why the institution cannot disappear.
That does not require conspiracy.
It requires incentives.
Thirty-Three Years Later
This essay is not an argument for abolishing government.
Neither was Friedman’s.
It is an argument for remembering what government is.
Government is force. Sometimes that force is necessary.
A murderer does not get to opt out of criminal law.
A foreign army does not get to invade because national defense would otherwise interfere with the market.
A corporation does not get to poison a river merely because preventing it constitutes regulation.
Civilization requires rules.
But rules should be judged by what they produce, not by the intentions of the people who write them.
That is where we have gone wrong.
We celebrate the appropriation rather than the outcome.
We announce the program rather than measure whether the problem disappeared.
We count dollars spent and call them investment.
We count regulations written and call them action.
And when the result disappoints us, we almost never entertain the possibility that the intervention itself deserves blame.
We ask what government should do next.
Friedman asked a more uncomfortable question:
What did government do already?
Thirty-three years later, the list has changed.
The mechanism has not.
We spend nearly a trillion dollars on public schools and wonder why too many children still cannot read proficiently.
We restrict housing and subsidize affordability.
We prohibit drugs and finance the consequences of the black market.
We built a healthcare system so complicated that navigating the system became an industry of its own.
We take years—or decades—to modernize infrastructure everyone agrees is obsolete.
We borrow trillions and then spend another trillion paying interest on what we already borrowed.
Then, when each intervention produces another problem, somebody arrives with another intervention.
Friedman’s warning was never really that government contains uniquely bad people.
It was something much harder to solve:
Bad incentives can defeat good intentions.
We had thirty-three years to prove him wrong.
Instead, we expanded the experiment.
And government is still the problem.