I was sold on this. Somewhere around the fourth time I heard that the market is the greatest engine of human flourishing ever devised, that prices are information, that the entrepreneur is the hero of the modern world and the bureaucrat its natural predator, I bought in. I read the books. I underlined the parts about spontaneous order. I nodded along at the funeral orations for every business that failed to serve the consumer.
So it has been a confusing few years.
Because it turns out that the people who taught me this catechism have been running, without ever putting it that way, the most extensive program of industrial planning, price supports, protective tariffs, and direct state ownership that Americans have seen outside of a war or a financial crisis. And they seem to feel no tension about it whatsoever. Which makes me think I misunderstood the assignment.
What follows is what I found, in the spirit of a man who suspects he's the one who's confused.
The Car Dealer
Start small.
In much of this country, it is illegal to buy a car from the company that made the car.
Not regulated. Not taxed. Illegal. Every state in the union has franchise laws that limit or forbid manufacturers from selling directly to you, and in more than a dozen of them the ban is total: a company like Tesla or Rivian cannot legally complete a sale inside the state at all. You order online, and the paperwork politely pretends you bought the car somewhere else. Tesla has spent well over a decade in courtrooms and statehouses fighting for the right to sell you a thing it built. In Connecticut, it finally opened a showroom at Mohegan Sun, on tribal land where the state's dealer laws do not reach, because that was easier than getting the legislature to let a company sell a product.
This is a law requiring the existence of a middleman, a legally mandated toll booth between a willing buyer and a willing seller. It is, in the most literal sense American law allows, a guild.
And it is defended most vigorously in Texas, where Tesla-backed repeal bills die on schedule every session. In South Carolina, where a direct-sales bill was shelved in subcommittee in 2025. In Louisiana, which banned direct sales by statute in 2017 and litigated against Tesla until a settlement last summer. In the states whose delegations will tell you, with a straight face and no visible strain, that government has no business getting between a man and his commerce.
I thought y'all loved capitalism. This is the freshman midterm.
The Farm
In December 2025, the administration announced twelve billion dollars in payments to American farmers. Eleven billion of it was allocated through something the Department of Agriculture named the Farmer Bridge Assistance Program; the checks started landing in late February.
The stated purpose was to compensate farmers for market disruptions and lost export sales, which in practice meant the collapse of Chinese purchasing after the tariffs. The President was blunter about it: "This money would not be possible without tariffs."
Follow the sequence. The government imposed tariffs. Foreign buyers retaliated. American farmers lost their customers. So the government wrote checks to farmers, credited to the tariff revenue, to compensate them for the damage the tariffs caused.
As policy, this is a man setting his own house on fire and then billing himself for the water.
And it isn't new. In the first trade war, the trade-aid packages built around the Market Facilitation Program totaled twenty-eight billion dollars, about twenty-three billion of it paid straight to farmers, in the same circular fashion. Farm-policy watchers now peg combined federal farm program spending across 2025 and 2026 at more than eighty-six billion dollars, record territory for the sector; USDA's own forecast of direct payments is about seventy-five billion for the two years, with the 2026 number up 45 percent in a single year. The Environmental Working Group's analysis of who actually receives the money found what everyone always finds: the largest operations take the largest share. Farms running more than a thousand acres of corn are 6.3 percent of corn farms and are projected to collect 39.9 percent of the corn payments. The bailout is an accelerant for consolidation. It is buying out the family farm in the name of saving it.
They named it a bridge, which is honest in one way: bridges are what you build when you plan to keep crossing.
And this is before we get to the rest of the architecture: crop insurance premiums that are 62 percent federally subsidized, a sugar program that props up domestic prices by restricting imports, and a renewable fuel mandate that solves the problem of insufficient corn demand by making corn-ethanol demand a matter of federal law: fifteen billion gallons a year, reaffirmed this March. If a Democrat proposed a statute requiring Americans to buy a specific Midwestern commodity, we would hear about Venezuela for a decade.
The Tariff
A tariff is a tax. This is not a controversial claim; it is a definition. It is a consumption tax, collected at the border, paid by the importer, and passed through almost entirely to the American side of the transaction. The New York Fed's tally of the 2025 tariffs found roughly ninety percent of the burden fell on U.S. firms and consumers. It is also among the most regressive taxes we have, because poor households spend a larger share of income on goods. Yale's Budget Lab put the short-run burden on the bottom decile at more than three times that of the top.
The people who taught me that every tax is theft presided over what the Tax Foundation called the largest tax increase, as a share of GDP, since 1993, and celebrated it as an act of economic nationalism.
Then, on February 20, 2026, the Supreme Court held that the emergency statute underpinning most of it never authorized tariffs at all. Roberts wrote that IEEPA "contains no reference to tariffs or duties," and that "until now no President has read IEEPA to confer such power." Six justices agreed on the result. Roughly $166 billion had already been collected from more than 330,000 importers. The administration terminated those tariffs within hours and immediately began rebuilding the same wall out of different bricks: a ten percent across-the-board Section 122 surcharge in force four days later, fresh Section 301 investigations, Section 232 cases already in the pipeline.
But my favorite part of the tariff regime was never the tariff. It was the exclusion process.
Under the exclusion processes attached to the steel and China tariffs, a company that needs an imported input files a petition with the federal government asking permission to buy it without the tax. Officials then evaluate the request and decide, firm by firm and product by product, who gets relief. USTR fielded about 53,000 exclusion requests and denied 87 percent of them, through a process GAO found was never fully documented even internally; Commerce ran a parallel operation for steel and aluminum that processed hundreds of thousands of petitions. And a peer-reviewed study of roughly 7,000 exemption applications found that campaign contributions to the party holding the White House raised a firm's odds of approval, and that contributions to the other party lowered them.
What that adds up to is an economy in which the price you pay depends on your ability to persuade a bureaucrat. We spent the entire twentieth century explaining to the world why allocation by petition produces bread lines, and we have now rebuilt it with better letterhead and a faster website.
The Portfolio
The portfolio is where the charitable reading gave out for me.
The federal government now holds equity or equity-like stakes in roughly thirty companies: about $27 billion worth, scattered across at least four agencies, with no consolidated public ledger of any of it. A 9.9 percent position in Intel, converted out of CHIPS Act and Secure Enclave awards. That is the same CHIPS Act that a great many of these same legislators denounced as corporate welfare before the plants broke ground in their districts. A fifteen percent as-converted stake in MP Materials that makes the Pentagon the company's largest shareholder. Warrants for five percent of Lithium Americas plus five percent of its Thacker Pass venture. Ten percent of USA Rare Earth, an agreed stake in Trilogy Metals, and a widening list of chipmakers. A golden share in U.S. Steel, giving Washington consent rights over plant closures and cuts to promised investment. A negotiated arrangement under which Nvidia and AMD agreed to pay the government fifteen percent of revenue on certain AI chips sold to China, as the price of their export licenses.
And in the MP Materials deal, the Pentagon guaranteed a floor price of $110 per kilogram on the company's rare-earth output for ten years. When the market price falls below the floor, the government pays the difference.
A guaranteed minimum price for a producer's output, backed by the state, has had a name since the 1930s. It is a price support, and until about eighteen months ago it was the thing we pointed at when we wanted to explain what was wrong with the other guys.
The state as regulator, customer, financier, and shareholder, all at once, in the same industries, with no meaningful public accounting of the portfolio. Even the Cato Institute, which is not exactly a redoubt of Marxist agitation, has been reduced to writing essay after essay about it.
I was told the problem with government picking winners is that government is bad at picking winners. I now understand the actual objection was that the wrong people were doing the picking.
The Bank
March 2023. Silicon Valley Bank fails. Federal deposit insurance covers a quarter million dollars per account. The uninsured depositors, sophisticated institutional investors who had made a calculated decision to hold enormous balances at a poorly hedged bank, were made whole anyway, under a systemic risk exception invoked on a Sunday night.
Many of those depositors were venture capitalists who had spent the preceding decade explaining creative destruction to everybody else. They spent one weekend explaining systemic risk to Washington.
This is the hypocrisy the whole structure rests on. Moral hazard is a term that only ever seems to apply downward. When a family can't make a mortgage payment, moral hazard is an iron law of human nature and forgiveness would corrupt the national character. When a bank can't manage duration risk, moral hazard is an abstraction we cannot afford to indulge in a moment of crisis.
Put the two big forgiveness programs of the last decade side by side. Seven hundred fifty-five billion dollars in Paycheck Protection Program loans had been forgiven by late 2022, with the SBA's own Inspector General putting potential fraud at $64 billion, and the forgiveness was designed into the program from the start. Student loan relief, a $430 billion plan barely half that size, was a moral catastrophe that would teach a generation that debts are optional, and the Supreme Court struck it down.
Nothing economic separates the two programs. What separates them is who was standing in line.
The Competitor
And then there is China, which has clarified everything.
The argument for free trade was never that trade is nice. It was that competition disciplines producers and that the consumer is sovereign. That if a foreign firm makes a better product at a lower price, the correct response is to make a better product at a lower price, and that shielding domestic industry from that pressure produces exactly the flabby, complacent, politically connected national champions we spent decades mocking in other countries.
BYD makes an electric car that sells at home for under ten thousand dollars, less than almost any used car in America, and abroad for less than the average American used car. The response has not been to compete. The response has been a 100 percent tariff so the vehicle cannot enter the country at all, a rule barring Chinese connected-vehicle software outright, and a growing list of Chinese firms barred from American markets by name.
I am not arguing there is no national security case here. There is a real one for some of it. What I'm noting is that the case is never made in national security terms alone. It is made in the language of protecting American jobs and American industry from unfair foreign competition, the same industrial-policy language we were told was economically illiterate when the United Auto Workers used it.
Competition, it turns out, is a magnificent disciplining force, right up until it is aimed at you. At which point it becomes an emergency.
The Honest Objection
Let me give the other side its best shot, because the weak version of this essay is a gotcha and the strong version has to survive contact.
The serious response is that pure free-market fundamentalism was always a debating position rather than a governing philosophy, and that adults facing an actual industrial competitor with an actual state-directed economy do not get to legislate out of a textbook. Semiconductors and rare earths are genuine strategic chokepoints; a country that cannot make its own magnets or its own logic chips has a foreign policy written for it in Beijing. Farm policy exists because food security is not a normal market and because a countryside emptied out in one bad decade does not repopulate in the next. Deposit insurance was extended at SVB because the alternative was a run on every regional bank in America by Monday morning, and being right about moral hazard is cold comfort in a bank panic. There is no interesting philosophical tension in acknowledging that markets are extraordinary at allocation and poor at resilience.
I think a good part of that is true. Some of it I would defend myself.
But the honest version has to be said out loud: markets are a tool, not a moral order; we will overrule them when the national interest demands it; here are the criteria; here is who decides; here is how we will know if it worked. That is a real position. It has a distinguished lineage running back through Hamilton and List. You can argue with it on the merits.
Nobody in power is saying that. The position on offer is that the market is sacred and self-correcting and that any interference is the road to serfdom, delivered from a podium, on the way to sign the price-support agreement.
What the Principle Actually Is
I no longer think there's an inconsistency here. I think I misread the doctrine.
The doctrine was never free markets. If you line up every one of these cases, a single rule explains all of them without a single exception:
Market discipline is for people who cannot afford a lobbyist.
Competition is a bracing tonic for schoolteachers, retail workers, and anyone seeking debt relief. It is an existential threat requiring immediate federal intervention for car dealers, row crop operations above a thousand acres, regional banks, chip fabricators, steel producers, and rare earth miners.
A market is a machine for telling you things you don't want to hear. Which is why no one who could afford to buy a different answer has ever wanted one.
None of this is unique to one party. Democrats subsidize their own constituencies with their own euphemisms, and I would write that essay too. But only one side is currently assembling the largest portfolio of federal corporate stakes ever amassed outside a war or a financial crisis while insisting the other guys are the socialists. That asymmetry goes beyond ordinary hypocrisy. Hypocrisy is failing to live up to a standard you hold. Holding a standard exclusively for other people is not a standard at all. It is a weapon.
So, sincerely, and with the residual affection of a man who actually did read Hayek, I would love to know when the free market is coming back. I'll wait. I've got a bridge payment to spend.
The Part That Isn't Funny
Everything above can be filed under hypocrisy, and hypocrisy is a comic offense. It costs money and it insults the intelligence, but the republic has survived worse and will survive this.
What follows is different: a country deliberately handicapping itself in the one domain that determines whether it stays rich and stays safe, on the basis of what appears to be a personal aesthetic preference about what a power plant should look like.
The Merit Order, Reversed
The numbers are not in dispute. Lazard's June 2025 levelized cost analysis puts new-build coal at roughly $122 per megawatt-hour, taking the midpoint of its range. New combined-cycle gas comes in near $78. Onshore wind is around $61. Utility-scale solar is around $58. That is the same cost accounting a utility uses to decide what to build, and it is why utilities have been retiring coal for fifteen straight years, under administrations of both parties. The market never attacked coal; it just kept filing the same report.
The administration has set out to overrule that report in both directions at once. Since May 2025, the Department of Energy has issued more than forty emergency orders and extensions under Section 202(c) of the Federal Power Act, and six aging fossil plants are now held past their scheduled retirements on rolling ninety-day renewals, with grid operators directed to keep them available. That provision was first used in June 1941, months before Pearl Harbor, to push electricity to aluminum plants for defense production. It was used twenty-two times in association with World War II, seven times in the following three decades, and zero times from 1977 to 2000; in this century, until now, it was reserved for hurricanes, blackouts, and heat waves: brief, acute, and requested by the grid operators themselves. It had never been an open-ended instrument for keeping retired plants alive. The Congressional Research Service calls the current practice a "seemingly new interpretation" of the authority.
The bill lands on ratepayers. The Sierra Club's tracker counted $235 million in additional costs across the first thirteen orders by March 2026; IEEFA clocked the burn at more than $30 million a month; by June the estimated net cost was about $1.5 million a day, roughly $550 million a year. At the Eddystone plant outside Philadelphia, FERC approved spreading the cost across electricity customers in the entire thirteen-state PJM region, and the tariff formula pays the owner its costs times a 110 percent multiplier that ratchets up ten points a year toward 150. A subsidy is at least honest about being a transfer. This is a transfer that arrives on your utility bill labeled as reliability.
And the reliability claim is the part that should worry a serious person. NERC's 2026 State of Reliability report found that the equivalent forced outage rate at coal plants rose from 11.2 percent in 2024 to 14.1 percent in 2025. The units became measurably less dependable, not more. Two of the plants operating under federal emergency order, Centralia Unit 2 in Washington and Craig Unit 1 in Colorado, produced zero electricity in the first quarter of 2026. Two more, both in Indiana, spent much of the spring offline for repairs. The defense is that the plants are being paid to be available rather than to run. The bill does not distinguish.
At the same time, the government is obstructing the cheapest generation available. At least 106 wind projects across twenty-one states, representing $47 billion in investment, were stalled by a Pentagon review freeze, per the economic analysis filed in the industry's lawsuit; the trade association's count has since passed 150. Roughly twenty-one gigawatts of clean generation and storage projects have been cancelled since the start of 2025. An already-issued air permit was pulled back; a stop-work order landed on a project that was eighty percent built; approval authority for wind and solar on federal land has been pulled up to the Interior Secretary's desk. The predictable happened: wind power purchase agreement prices rose nearly 24 percent year over year, to the highest level since the index began, driven substantially by scarcity the permitting regime itself created. Government intervention raised the price, and the higher price is then cited as evidence that the technology is uneconomic.
There is no coherent theory of energy abundance in which you simultaneously mandate the operation of your most expensive generation and prohibit the construction of your cheapest. "All of the above" is the slogan. The policy is a fuel preference.
The Part About Innovation
The expense is the small part, because manufacturing capability is not a stock you can draw down and replenish at will. It is a learning curve. Cost declines come from cumulative production volume, from engineers who have solved the same problem four hundred times, from a supplier three miles away who already makes the part. Lose the volume and you lose the curve, and the curve does not wait for you.
China produces about 97 percent of the world's solar wafers, 91 percent of its cell capacity, and 86 percent of finished modules. It holds 80 percent of lithium battery cell production and over 90 percent of anode capacity, and close to 90 percent of global rare earth processing. Two Chinese companies alone ship a majority of the world's solar inverters, and Chinese manufacturers took 78 percent of the 176 gigawatts of wind turbines installed worldwide in 2025. Its wind and solar fleet reached 1,840 gigawatts by the end of 2025, exceeding its coal fleet for the first time, with nearly 145 gigawatts of grid-scale battery storage behind it, up 85 percent in a year.
At those numbers China is no longer a competitor. It is the supplier, and we are the customer.
What does the current policy do to that position? It does not build American capacity in those categories. It suppresses American demand for them. A domestic manufacturer needs a domestic market to scale into, and the policy has spent two years demolishing that market. You cannot tariff your way into a supply chain you have forbidden anyone to build. Blocking the import and blocking the domestic buildout at the same time does not produce independence. It produces shortage, and a country that is short of something ends up buying it from whoever has it.
This is the innovation cost, and it is the one nobody bills you for directly. Every stalled interconnection queue, every revoked permit, every cancelled factory is an engineer who takes a different job and a process improvement that gets made somewhere else. Grid-scale storage, power electronics, transmission hardware, advanced manufacturing of every kind: these are the disciplines that determine what a country can build in ten years. We are choosing not to practice them.
And this is happening precisely as electricity becomes the binding constraint on everything else. Data centers, AI training capacity, semiconductor fabs, electrified industry, the defense industrial base itself. All of it is a function of how many megawatts you can add and how fast. A country adding capacity slowly and expensively, from a fleet with a rising forced outage rate, held together by rolling emergency orders and transformer lead times running two and a half to four years, is a country that has capped its own growth rate without ever announcing the decision.
The Security Argument, Taken Seriously
The people running this policy talk constantly about energy security. So let us take the phrase at face value and ask what would actually produce it.
Security in an energy system comes from diversity of supply, redundancy, domestic control of critical inputs, and spare capacity, so that no single failure and no single supplier can hold the grid hostage.
Measured against that standard, the policy fails on its own terms. Concentrating on a fleet of aging thermal units with deteriorating availability reduces redundancy. Suppressing the fastest-deploying generation reduces the rate at which you can add spare capacity. Leaving battery, inverter, and processed-mineral supply chains to a strategic rival while blocking the domestic demand that would fund alternatives to them tightens that rival's grip on American infrastructure. Every one of those is a security loss, and they compound.
There is a real reliability problem in the American grid. Intermittency is genuine. Dispatchable capacity matters. Interconnection queues are a disaster, transformers are backordered for years, and the market-friendly alternative is itself supply-constrained: gas turbine prices are up 195 percent with six-year lead times, and the all-in cost of a new gas plant went from about $2,000 a kilowatt last fall to north of $3,000 by March. A serious administration would be attacking those constraints with everything it has: permitting reform that applies to all generation equally, transmission buildout, domestic manufacture of grid hardware, storage at scale, and yes, new nuclear.
Instead we have emergency orders for coal plants, and a trophy presented to the President in the East Room this February by coal executives, engraved "Undisputed Champion of Beautiful Clean Coal."
Picking Winners at the Level of Whole Industries
What is left is the scale of the thing, because the intervention has changed in kind and not just degree.
Subsidies back firms and tariffs shelter sectors; the country has done both forever. What is happening now is a government issuing verdicts on entire technologies. Not "this company failed to compete," but "this branch of engineering is disfavored, and we will use emergency statutory authority to slow its deployment and prop up its replacement." The state has moved from adjusting outcomes at the margin to deciding, at the level of the whole economy, which physics is patriotic.
That is a claim of competence no government has ever had. It is the specific claim that free-market conservatism spent seventy years demolishing, and it was right to demolish it. The Hayekian argument was that no central authority, however well staffed, can possess the dispersed knowledge that prices aggregate, and that the ones who try will substitute their own preferences for the information and then be unable to tell the difference. It was never a claim that businessmen are virtuous.
That is exactly what has happened. Someone at the top of this government dislikes the look of a wind turbine, and that preference has been laundered through emergency authority into national industrial policy. There is no technical analysis at the bottom of it. There is a taste, and a great deal of statutory machinery pointed in its direction.
Americans will pay in the currencies the people responsible claim to guard: a more brittle grid whose components ship from Shenzhen, industrial capability that will have to be bought from someone who may not want to sell it, and a higher price on the input that sits underneath every other price in the economy.
And the deepest insult in it is this. All of that information was available. It was in the cost curves, in the interconnection queue, in the retirement schedules the utilities filed years ago, in the outage statistics, in the capital that kept flowing toward the cheaper technology no matter how much the rhetoric disapproved. The market was talking the entire time. That is the one thing a market is actually for.
They just didn't like what it said.