Three Reasons Tariffs Don’t Work

He just likes tariffs. President Trump’s decades-long love of tariffs comes from the heart. The intellectual arguments Trump and his aides use to justify his tariffs are after-the-fact attempts to rationalize the president’s long-held anti-trade beliefs. Neither he nor his top aides are persuaded by economic arguments against tariffs. So why bother with this paper, which provides three such arguments against tariffs?

First, presidents come and go, while policy is a long game. The current political moment is just that: a moment. It will pass. The next administration, regardless of party, will have an opportunity to align trade policy with the nation’s economic and foreign policy interests. This paper is aimed at them. It provides them with policies and arguments they will need to succeed.

Second, America’s federal government has three branches, not one. The executive branch is not interested in reform, but there is plenty for Congress and the courts to do. This paper is aimed at them as well.

All of the Trump-era tariffs have been imposed unilaterally by the president himself. The last time Congress voted to raise tariffs was the Smoot-Hawley Tariff Act of 1930, which worsened the Great Depression. The electoral bloodbath of 1932 taught Congress that tariffs are political poison. Likewise, today’s opinion polls show that tariffs and re-election do not pair well. But members of Congress seem to have forgotten that trade policy is theirs to make. Trade policy is subject to the laws passed by Congress. The courts’ job is to uphold those laws.

Knowledge problems

How do policymakers know what level of tariffs is right? And how do they get them implemented? The short answers are that they can’t and they don’t. There are three reasons why an optimal tariff is unobtainable: knowledge problems, incentive problems, and an impossibility problem. First, knowledge problems.

Markets are a continuous process of trial and error. They consist of individual decisions made every day by billions of ordinary people, guided by decentralized and ever-changing prices. Nobody is in charge of markets, yet they feed nations, lift billions of people out of poverty, and foster innovation better than any other known social system. The knowledge problem is that successfully planning or directing markets requires more information than anyone could possibly handle.

Because of knowledge problems, Trump’s tariffs have resulted in unintended consequences and often unpredictable downstream effects.

Steel tariffs, for example, might help domestic steel producers by allowing them to charge higher prices without being undercut by competitors. But higher steel prices hurt steel-using industries like autos, construction, appliances, and more. One study found that Trump’s first-term steel tariffs likely saved about 1,000 jobs in the steel industry, but cost about 75,000 jobs in steel-using industries. That tradeoff was likely not the administration’s intent.

That study came out in 2020. Since then, the steel industry has fallen on hard times despite yet more tariff protection. The reason is that decades of government protection have turned America’s industry into a ward of the state—another unintended consequence of protectionism.

Cleveland-Cliffs, a major steel producer, announced 1,200 layoffs in March 2025. U.S. Steel might have gone out of business altogether if Japan-based Nippon Steel had not acquired it. This acquisition was a drawn-out, politicized process that spanned both the Biden and the second Trump administrations. It ended with Trump receiving a golden share of U.S. Steel stock, giving him the power to veto company decisions. Trump has already used this power to prevent the closure of a plant in Illinois.

The auto industry was an unintended victim of the steel tariffs. The administration tried to fix that damage by raising auto tariffs. That action, in turn, spawned still further unintended consequences. Trump’s 2025 auto tariffs raised the price of an average new car by between $3,000 and $6,000, depending on where its parts are sourced. By 2026, the average price of a new car was roughly $50,000, and automakers were considering removing their most affordable models from the US market altogether.

People who are priced out of new cars sometimes turn to the used market instead. With more people jockeying for the same fixed supply of used cars, used car prices went up, too, despite not being tariffed.

Trump’s auto tariffs and Liberation Day tariffs ended up hitting American car companies like Ford and GM harder than foreign brands. Tariffs on Mexican and Canadian parts for Fords and GMs, often at 25 percent, in some cases raised American car prices by more than tariffs on entire cars made in certain countries with 10 or 15 percent tariffs. In other words, American tariffs put American cars at a competitive disadvantage.

That is just the fallout from steel tariffs. Similar problems have played out in industry after industry across the economy. Manufacturing as a whole is down roughly 100,000 jobs since Liberation Day, and manufacturing output has yet to regain the all-time peak it reached in September 2018. That was the month before Trump’s first round of China tariffs.

Even the tinplate used in food and beverage cans is a victim of unintended consequences. These imports were hit by Trump’s 25 percent and then 50 percent tariffs on metals. Those were imposed under a national security rationale. There is no known security threat regarding tinplate.

Despite tariff protection, domestic tinplate production has declined since 2018. Twelve production lines are now down to three. Food and beverage producers now rely on imports for 70 percent of their cans, up from 30 percent before the national security tariffs.

Tariffs on imported canned food and beverages are typically lower than the 50 percent rate on tinplate. So, depending on the food costs for a given product, American tariffs can put American food producers at a competitive disadvantage against imports, just as they did to the auto industry.

Most tariff costs are passed along to consumers, who pay more and notice it. When they spend more on food, they have less remaining for other purchases, causing yet more unexpected effects on industries that have nothing to do with tinplate and may not be subject to tariffs at all.

The economy is more interconnected than most people realize. A deck of cards can help illustrate just how complex the economy is. Every time you shuffle a deck of cards, you almost certainly create a sequence that has never before appeared in the universe’s 14-billion-year history, even with just 52 cards. The number of possible ways to order a 52-card deck is 52!, or 52 factorial, which is 52 x 51 x 50 … all the way down to 1. Factorials grow very quickly. While 3! is 6 (3 x 2 x 1), 10! is 3,628,800. There are roughly 8 x 10⁶⁷ different ways to shuffle a deck of 52 cards. That is an eight followed by 67 zeroes.

Now compare this to a typical grocery store in the United States, which offers more than 45,000 different goods. This means there are more combinations of goods in an average grocery store than there are atoms in the universe, which scientists currently estimate to be between 10⁷⁸ and 10⁸².

There are thousands of grocery stores in the United States, each with a colossal variety of goods. Using sophisticated supply chains and logistics, they collect products from all over the world under one roof, ranging from food to medicine to school supplies. Their typical profit margin is roughly 3 percent, less than half the stock market’s average annual return.

Now throw into this tightly run system the highest tariffs since the Great Depression and impose them on just a few weeks’ notice. Then change those tariff rates more than 30 times in less than a year and see what happens. Wholesalers, retailers, shippers, and consumers are all facing rapidly changing conditions and are scrambling for ways to adapt. It’s trial and, often, error in double-quick time. Each person’s actions affect everyone else’s decisions in ways nobody can predict. Few of the consequences of tinplate and grocery tariffs will be intended.

While there is more to the knowledge problem than unintended consequences, these examples from the steel, sugar, tinplate, and grocery industries illustrate how difficult it is for policymakers to bend an economy to their will. This is the knowledge problem at work.

About the only consequences that can be predicted are that people will pay more and get less, and they will know that politicians are to blame.

Incentive problems

Now let’s assume the knowledge problem away. Even if policymakers could have the knowledge they would need to set tariff rates that work as intended, they would still lack the incentives to implement them.

Diffuse costs, concentrated benefits. Tariffs benefit the few and hurt the many. The entire American steel industry, for example, employs roughly 80,000 people. Pre-Liberation Day steel tariffs allowed the industry to collect more than $5 billion annually in higher prices for its products. (The federal government separately collects an additional $2.7 billion per year in tariff revenues, for a $7.7 billion combined hit to consumers and steel-using industries). That amounts to more than $600,000 of annual tariff benefits per steel employee. The stakes for the industry are high, while all other Americans each lose about $30 each year. The industry’s incentive is larger than that of consumers generally by a factor of roughly 20,000.

Nobody will organize political pressure campaigns over a loss of $30 per year. That is less than $3 per month. Even if people do succeed in removing steel tariffs, the benefits to each member of the public will be far less than the costs. Lobbyists typically charge more than $30 per year.

The steel industry has a much stronger incentive to lobby than do the masses, measured in hundreds of thousands of dollars per employee. Since there are relatively few steel companies, it is also easier for them to coordinate with each other than it is for 340 million unorganized consumers.

Meanwhile, the US sugar industry is so protected by tariffs that domestic producers can charge consumers between two and three times the world price. As a result, US sugar consumers pay an extra $2.9 billion to $3.5 billion per year. Most of the benefits go to sugar companies owned by the Florida-based Fanjul brothers. The Fanjuls are so unsubtle in their lobbying efforts that they hosted separate fundraisers for both the Trump and the Hillary Clinton campaigns during the 2016 election. One of the Fanjul brothers specializes in Republican donations, and the other specializes in Democratic donations.

Which industries should get help? Suppose that some industries would benefit from a well-designed tariff. What incentives do politicians have to choose the right industries? Well-meaning economic planners are usually less persuasive than the stories lobbyists tell about the political blowback of lost jobs in a congressman’s district.

Biden-era tariffs intended to help green industries were driven more by political fashion and interest group appeals than by their merits. Likewise, many national conservatives believe that the manufacturing sector deserves special protections, including tariffs, for reasons ranging from nostalgia to workers’ dignity arguments. The higher tariffs that follow are then countered by other countries’ efforts to convince Trump to lower tariffs on their industries’ products. Their inducements have included literal gold bars from Switzerland and a 747 jumbo jet from Qatar’s government. These efforts to persuade are based on something other than merit.

Infant industries. John Stuart Mill argued in his 1848 Principles of Political Economy textbook that some industries may benefit from protection while they are still establishing themselves. He also argued that these temporary tariffs should be removed when the infant industry matures.

This is called the infant industry argument, and even many hardcore free traders find it persuasive. The trouble, once again, is incentives. The protected infant industry has a very strong incentive to maintain its protection permanently. Each member of the public may pay only pennies per year in higher prices. Most people will probably not even be aware that the tariff exists. Few things are as permanent as a temporary tariff.

In a well-known study, Dartmouth College’s Douglas Irwin found that in the 1890s tariffs did indeed speed up development of America’s tinplate steel industry by about a decade. However, those tariffs also imposed costs on the rest of the economy that outweighed the tinplate industry’s benefits. The tariffs were a net loser for Americans.

Exemptions. Once new tariffs are in effect, some of the harmed companies will argue that they should be exempt. For example, tariff-related lobbyist spending increased 277 percent from Biden’s final year in office, 2024, to the first year of Trump’s second term, 2025.

Tariff exemptions typically cannot be known in advance. The Trump administration’s usual approach is to enact a widely applicable tariff and then hear requests for exemptions. Government officials might give exemptions because they see a need or because they want to pay off a political connection or because they are in a good mood that day. It is impossible to know in advance how many producers will be exempted from a tariff and how to adjust the rate so that it stays optimal. Tariff-related lobbying expenses should be understood as a deadweight loss to society that inevitably follows from a tariff policy.

The uncertainty around tariff exemptions is itself a source of harm. Companies have difficulty making long-term investments if they don’t know what their costs will be. From a policymaker’s perspective, it is hard to set a tariff rate that will accomplish his goals if exemptions will change the effective rate unpredictably.

Optimizing what? An impossibility theorem

President Trump and his advisers have put forward several justifications for their tariffs. Trump has long wanted to tilt America’s balance of trade toward exports. Former Trump economic adviser Stephen Miran, now on the Federal Reserve’s Board of Governors, favors using tariffs to devalue the dollar. There is bipartisan support for using tariffs to support certain industries, such as manufacturing and renewable energy. Trump and others promised that tariffs would raise trillions of dollars of tax revenue. That was also the goal of much American tariff policy before the income tax. Tariffs can also be a negotiating tool for trade-unrelated policy goals.

Trump has also cited fentanyl trafficking, the War of 1812, a television commercial that aired during baseball’s World Series, and the tone of the Swiss president’s voice on a phone call as justifications for raising tariffs.

The tariff impossibility theorem is rooted in every politician’s nemesis: tradeoffs. Different policy goals will likely have different optimal tariff rates. That means policymakers will have to choose which goals to accomplish and which to leave unfulfilled.

Suppose the government wants to protect domestic manufacturing while also raising substantial revenue. A tariff high enough to block most imports may protect an industry but will raise little revenue. A lower tariff that lets in more imports might raise more revenue, but it won’t protect domestic manufacturers from competition.

Sometimes the president raises an existing tariff as a negotiating tactic in a trade-unrelated dispute with another country. Examples from Trump’s second term include spats with Mexico over immigration and with Canada and China over drug trafficking. If Trump gets what he wants in a negotiation and withdraws the tariff, then it will raise no revenue at all and will not protect domestic industries.

For other objectives, such as reducing trade deficits, tariffs have indeterminate effects. There is no way to predict their net effect on the balance of trade. Tariffs reduce imports, which trade deficit hawks like. However, they also reduce exports, which some trade deficit hawks would rather increase. But it is impossible to predict by how much, or even in what direction, tariffs would change the difference between our imports and our exports, popularly called the trade deficit.

National security tariffs are a special case. On the one hand, national security is the strongest tariff justification there is. On the other hand, national security rationales are often fig leaves covering up other objectives. It is difficult to measure progress toward a goal that is unknown.

When Canada’s then-Prime Minister Justin Trudeau phoned Trump in 2018 to protest national security tariffs on steel and aluminum from Canada, Trump invoked the War of 1812. He justified the tariffs by claiming that Canadians burned the White House in that war. In fact, the culprits were British soldiers stationed in Canada, which did not yet exist as an independent country.

This justification satisfied the legal requirement to cite a national security rationale for tariffs enacted under Section 232 of the 1962 Trade Expansion Act. It was also obviously a cover story for objectives unrelated to security that were never made clear. Those objectives could include anything from trade deficits to protecting US steel and aluminum producers to Trump’s personal dislike of Trudeau. How does one calculate an optimal tariff for those policy objectives?

Policymakers have many goals. When some of them contradict each other, some or all of those goals will have to be left unmet. Tradeoffs are a fact of life, even in tariff policy.

Policy solutions

The common theme for most tariff reforms is the separation of powers. The US Constitution gives all taxing power to Congress and none to the president. For reasons of expediency, Congress has delegated to the president parts of its taxing powers in several bills from the 1930s to the 1970s. These powers were rarely or never used by previous presidents, so Congress was content to leave them in place.

President Trump has since used these statutes to enact America’s largest tax increase since 1992 without a congressional vote. Those actions are unconstitutional. One of Congress’ top priorities should be repealing the sections of various statutes that delegate Congress’ taxing authority to the president. These include:

Section 338 of the Smoot-Hawley Tariff Act of 1930, which allows the president to impose up to a 50 percent tariff by claiming foreign countries discriminate against American products.
Section 232 of the Trade Expansion Act of 1962, which President Trump used to enact steel and aluminum tariffs against allies on dubious national security grounds.
Sections 201 and 301 of the Trade Act of 1974, which Trump and Biden used to enact tariffs against China, Europe, and many other trading partners.
Section 122 of the Trade Act of 1974, which allows the president to enact a 15 percent universal tariff for 150 days if he declares a balance of payments emergency.

Courts also have a role to play. For too long, courts have oscillated between judicial passivism and judicial activism. A passive judiciary lets the other branches do almost anything they want, even when unconstitutional. An activist judiciary usurps the other branches’ powers. Neither is a sound judicial philosophy.

Recently, courts have instead moved toward judicial engagement. This philosophy is neither passive nor activist. An engaged judiciary strikes down the other branches’ unconstitutional actions. It also does not attempt to take their powers for itself.

The major questions and non-delegation doctrines that have emerged in the past decade are examples of judicial engagement. These were at the heart of the Supreme Court’s Learning Resources decision, which struck down the tariffs Trump had enacted under the International Emergency Economic Powers Act.

Trump quickly re-enacted most of those tariffs under Section 122 of the 1974 Trade Act. This usage is currently being challenged in court, in part because a balance-of-payments emergency cannot exist in a floating exchange rate system. Currently, that case is winding its way through the courts and may be decided before the Section 122 tariffs expire on July 24, 2026. Other legal challenges to other tariffs may follow.

Major questions and non-delegation cases could also strike down Trump’s other unconstitutional attempts to use taxing authority without Congress’ involvement. The administration is currently preparing Section 232 and Section 301 investigations.

Consumers, import-using industries, and anyone else affected by these likely tariffs should bring legal challenges to them, and the courts should recognize the plaintiffs’ standing.

Conclusion

Policymakers have no idea how to set a perfect tariff rate. They lack both the knowledge and the incentives to do so. Even if they did, it is impossible to optimize tariffs for contradictory policy goals. The best long-run policy solution is free trade. In the short run, Congress and the courts should take back the tariff-making powers that Congress delegated to the president.

While nearly any real-world tariff policy is suboptimal, the Trump administration’s haphazard tariff approach has caused trillions of dollars in economic damage. At the same time, it has harmed America’s foreign policy interests in Europe, Asia, and the Middle East. Until Congress and the courts restore a proper separation of powers, the door is open for future administrations to cause similar damage. It is well past time to close that door.

About the author

Ryan Young is Senior Economist and Director of Publications at the Competitive Enterprise Institute. He specializes in trade, regulation, monetary policy, and antitrust policy. He edited Adam Smith’s Guide to Life, Loveliness, and the Modern Economy. He has an M.A. in economics from George Mason University and is a member of the Mont Pelerin Society.

About CEI

All Policy Areas

All Products

All Events

Support CEI

Knowledge problems, incentive problems, and an impossibility theorem