In recent months, Donald Trump’s administration has supplemented its globe-spanning tariffs with bespoke industrial policy deals with U.S. corporations. These include the federal government taking equity stakes in Intel, a “golden share” in U.S. Steel, and profit sharing on Nvidia and AMD’s chip sales in China. The deals have astonished and unnerved economic analysts across the political spectrum. Commentators from The Wall Street Journal editorial page to MSNBC have decried them as “unprecedented” and amounting to “socialism,” “corporatism,” or “state capitalism with American characteristics.”

In truth, Trump’s actions are rooted in a long history of presidents from both parties taking temporary control or ownership of systemically important companies. However, until now, most of these interventions have been temporary and involved either wartime emergencies or too-big-to-fail corporate collapse. Under Woodrow Wilson, the federal government took temporary ownership of the railroad industry to support troops during World War I. During World War II, Franklin D. Roosevelt’s administration directed much of the corporate industrial sector. In 1980, Jimmy Carter’s administration extended loan guarantees to Chrysler in exchange for stock warrants. After the company recovered and repaid its loans, the government earned $311 million in capital gains. During the financial crisis, Barack Obama’s administration provided the auto industry with tens of billions of dollars in exchange for equity stakes and commitments for substantial restructuring.

Trump’s moves also build on recent industrial policy interventions by his and Joe Biden’s administrations. Until now, those policies stopped short of the government seeking direct influence over corporate governance through golden shares or equity stakes. During his first term, Trump primarily conducted industrial policy through targeted tariffs and trade actions to promote American industry, and by jawboning corporations to boost domestic manufacturing. The Biden administration went further, providing computer chip and clean energy companies with hundreds of billions of dollars—and detailed requirements for how they could be spent—as part of a long-term post-neoliberal industrial strategy for national security. Those actions were consistent with, though a major step beyond, Trump’s first-term trade policies, just as Trump’s second-term corporate dealmaking is an advance on Biden’s (even if neither president would admit it).

So, while Trump’s dealmaking spree is not unprecedented, it ventures into new territory reflecting his strongman tendencies: He has claimed for himself concentrated power to oversee and influence corporate decision-making, authority thus far ceded to him by Congress and the courts. He has operated outside the law through handshake deals that corporations often cannot afford to refuse, endowing the presidency with powerful—and dangerous—economic tools for routine use.

This is unlikely to be a mere Trump-era aberration. Presidents typically don’t forego inherited powers. Today’s novelty policies may become tomorrow’s precedent, as the last decade has shown. Some elements of industrial policy with Trumpist characteristics may become long-term features of American governance. But Congress and the judiciary should assert democratic and constitutional control over American industrial policy now and in the future to prevent its abuse and, potentially, its collapse.

Here’s a rundown of Trump’s corporate dealmaking thus far:

U.S. Steel: In June, the White House greenlit Nippon Steel’s takeover bid for the American steelmaker U.S. Steel. Trump overturned Biden’s blocking of the deal—on the condition that the government receive a “golden share” of the company, allowing it to veto major corporate decisions, such as offshoring jobs and closing plants.

MP Materials: In July, the Department of Defense announced an investment in MP Materials, the country’s only active producer of certain rare earth minerals used for military equipment. The deal provided government support for the company, guaranteeing a minimum price for its products and providing a guaranteed buyer—the federal government—for all its output for the next 10 years. The feds also extended a $150 million loan and purchased $400 million in company stock, making it MP’s largest shareholder.

Japan: Also in July, the White House announced a trade deal with Japan that purportedly included Tokyo creating a $550 billion investment fund for the administration to use toward its industrial policy priorities, including energy infrastructure, semiconductor manufacturing, critical mineral supply chains, pharmaceutical production, and shipbuilding. Japan quickly contested the administration’s description of the deal, explaining that the investment fund would be nearly entirely composed of loans the U.S. would need to repay.

Nvidia and AMD: In August, the administration agreed to allow the semiconductor makers Nvidia and Advanced Micro Devices to sell certain modified chips in China, reversing national security restrictions. The companies also agreed to give 15 percent of all profits earned in China to the U.S. government.

Intel: Later that month, the administration took a 10 percent equity stake in the chipmaker Intel. The deal converted billions of dollars in grants the company was due under the CHIPS and Science Act into equity stakes, making the federal government Intel’s largest shareholder. That conversion also, without any legal authority, dispensed with CHIPS Act taxpayer safeguards that required Intel to reach milestones toward completing new domestic manufacturing sites to receive federal funding.

Trump’s economic interventions bear a number of distinctive qualities that expand on industrial policy precedent, but five stand out.

1. Unilateralism: Trump’s industrial policy has been personalist, flowing from and controlled by the president himself. Unlike the Biden administration, which shepherded industrial policy bills through Congress, Trump relies exclusively on executive action. His industrial policy has at times been colored by favoritism and personal relationships. The Nvidia export agreement, for instance, came after meetings between Trump and CEO Jensen Huang. The president himself asked Huang for a share of the company’s Chinese profits. In other cases, bargained-for benefits have been structured to accrue directly under Trump’s control. The Nippon Steel agreement gave control of the golden share not to the office of the president but to Trump himself; only after he leaves office does the share transfer to the secretaries of treasury and commerce. Finally, the (contested) $550 billion pot from Japan would be, according to Secretary of Commerce Howard Lutnick, a fund to be spent however Trump wished—part of “a national security and economic fund managed by Donald Trump.”

2. Public ownership in the absence of a fiscal crisis or national security justification: While direct government ownership of industry is not unheard of in American history, the Trump administration has been uniquely eager to wield it as a routine industrial policy tool. The MP Materials and Intel deals both gave the government equity ownership of private companies, and the U.S. Steel approval netted the government special, powerful veto rights over that company’s decision-making. After announcing the (contested) terms of the Japan trade deal, administration officials even contemplated using the deal’s (disputed) investment fund to build government-owned factories that could then be leased to private companies.

3. Bullying individual companies: Trump has a quid pro quo mentality of wanting a cut of any private gains he feels he has facilitated—he has openly compared his presidential dealmaking to waiving property rights on his real estate for the benefit of a third party, in exchange for a fee. When he agreed to grant Nvidia export licenses to sell modified chips in China, Trump told its chief executive, “If I’m going to do that, I want you to pay us something.” Or as Lutnick described the Intel equity stake, “If we’re going to give you the money, we want a piece of the action.”

4. Leveraging government power: Beyond just flexing the rhetorical might of the presidential bully pulpit, Trump has used the levers of government to offer firms deals they can’t refuse. The administration has created or exploited bargaining power to induce targeted firms to relent to its demands. Under export control regulations, Nvidia and AMD needed the administration’s permission to sell chips in China, forcing them to acquiesce to Trump’s profit-sharing demand. Nippon Steel could not take over U.S. Steel without approval from the government’s foreign investment review committee. Shortly before Intel agreed to the equity deal, Trump ratcheted up the pressure by calling for its executive to be fired and withholding its CHIPS Act grants.

Two other examples illustrate Trump’s leverage power plays. In May, he issued an executive order instructing pharmaceutical companies to agree to adhere to “most-favored-nation price targets … to bring prices for American patients in line with comparably developed nations.” The order threatened that the administration would put the squeeze on pharmaceutical companies that don’t comply, through newly authorized competition from drug imports, restrictions on their own exports, antitrust scrutiny, reconsideration of FDA approvals of their drugs, and loss of National Institutes of Health grants. And in August, the administration announced that it would impose high tariffs on semiconductor imports—but would exempt tech companies that made new domestic investment pledges. The administration thus imposed tariffs to create leverage and offered relief from those tariffs to advance its industrial policy goals.

5. Legally dubious boundary pushing: Much of Trump’s industrial policy has operated in legal gray zones. To bypass routine federal contracting and procurement laws that typically require competitive bidding, the administration’s investment in MP Materials relied on a little-used provision of the Defense Production Act authorizing transactions “without regard to the limitations of existing law.” Trump’s profit-sharing deal with Nvidia and AMD flouts prohibitions on export taxes and fees under both the Constitution and federal law. There is also no firm legal basis under the CHIPS Act to convert Intel’s grants into equity (nor was there to withhold its grants in the first place).

Trump has avoided legal constraints by relying on “regulation by deal,” rather than legislation or agency rulemaking. (He’s also been firing the regulators who would otherwise shape such government interventions.) Companies have acquiesced to White House demands that may not otherwise withstand legal scrutiny to avoid lost business opportunities, the cost and time of litigating against the federal government, and the general ire of the Trump administration. This has insulated much of Trump’s industrial policy from administrative procedure, due process, and judicial review.

Some elements of Trump’s approach are likely to endure. The wraparound support for MP Materials, for example, could well become a model for how the U.S. government cultivates strategically important domestic production. The type of golden share Trump obtained in U.S. Steel could be on the table for future mitigation agreements between the federal government and foreign companies seeking to invest in critical domestic sectors.

As I have written for this magazine, there are also compelling reasons for the government to insist on a direct return on its industrial policy investments. Taking equity stakes in semiconductor companies funded by the CHIPS Act was originally a Bernie Sanders proposal. When used in the right context, equity stakes ensure that industrial policy returns upside revenue to the public, instead of merely socializing profits for corporations.

External trends may also lend Trump’s industrial policy approach greater staying power. As policymakers seek new revenue sources beyond traditional taxation, they may increasingly look to levers like equity stakes, profit-sharing agreements, and royalties. Congress’s tendency toward gridlock and paralysis has long spurred administrations to seek creative executive workarounds, but the courts have increasingly limited the discretion given to the administrative state. That may steer administrations toward regulation-by-deal as a flexible means for the executive branch to regain power from the judiciary.

It’s not hard to imagine where Trumpist industrial policy may be headed. White House advisers have said they will seek equity stakes in other companies receiving federal grants and loans to seed a sovereign wealth fund. Moreover, if Intel cannot right its ship, it may be forced to revisit previous discussions of spinning off its semiconductor factories to a foreign firm like the Taiwan Semiconductor Manufacturing Company. The Trump administration, which has opposed such a move, could, as Intel’s largest shareholder, dictate the terms of such a deal. (Indeed, the Intel agreement included a poison pill to stymie the sale of its foundries.) Plus, any such sale must win the administration’s approval through the Committee on Foreign Investment in the United States. One could imagine the administration negotiating on economic and national security grounds for government ownership of Intel’s factories, which could then be leased to another chipmaker.

Trump’s industrial policy toolkit may empower future administrations with different priorities. For instance, federal policies—steep tariffs and national security exclusions—are sparing U.S. automakers from a flood of low-cost, cutting-edge Chinese electric vehicles. A future administration could condition that continued industry protection on American carmakers offering more affordable low-emissions vehicles and ramping up their research and development investments.

A future administration could use the specter of the government’s statutory powers to limit or override patents on federally funded technologies to negotiate long-term profit-sharing agreements with artificial intelligence companies. It could also use similar authority to negotiate price reductions from pharmaceutical companies and leverage the Medicare Advantage market to acquire golden shares in health insurance companies, with veto power over major corporate decisions.

A future Democratic administration could grant targeted tariff reductions to countries or companies that decarbonize their energy grids and production processes—and dangle extra reductions to those that use American-made clean energy technology. It could use the Defense Production Act’s preemption power to expedite clean energy and new housing development around other legal impediments. If the industrial policy arc from the first Trump administration to the Biden administration is any guide, the next Democratic president will make fulsome use of these kinds of Trump-inspired tools.

So sheer inertia may well propel Trumpist industrial policy forward. But there are several stark and urgent reasons Congress and the courts should pump the brakes and reconsider major elements of Trump’s approach. For one, this approach may simply yield worse industrial policy. Absent exceeding care and deliberation, go-it-alone executive branch industrial policy can overreach and implode without the safeguards and democratic legitimacy of operating through Congress. Even if Trump’s deals successfully evade the courts, a subsequent administration could deem some contrary to law and declare them null and void—regulation by deal has far less long-term durability than regulation by regulation, let alone legislation. Such whipsawing will make the U.S. government a less reliable, more tumultuous economic actor.

For another, elements of Trump’s industrial policy are a recipe for abuse and corruption. Pay-to-play regimes for exporting sensitive technologies to foreign adversaries jeopardize fair competition and national security. Personalist policymaking is a recipe for cronyism, corruption, and favoritism for well-connected dominant firms. As The New York Times reported, Trump’s family was directly benefiting from billions of dollars of investment in its crypto company by the United Arab Emirates—at the same time it agreed to provide the UAE with highly sensitive and valuable advanced semiconductor chips.

It is all too easy to foresee more future gambits redounding to the private benefit of the Trump family (like the Paramount merger with Skydance, which routed $16 million into the Trump presidential library). Moreover, while leverage can be useful for achieving policy aims, marshaling the vast regulatory state as one big pressure cooker to coerce concessions from targeted companies risks undermining other vital policy interests. We’re already seeing that regarding antitrust. Companies seeking merger approvals have won favorable treatment from the Trump administration by making side promises for domestic manufacturing investments. The potential for Trumpist misuse of industrial policy risks poisoning the whole enterprise.

That’s where Congress and the courts come in—or where they should. Congress was a true coequal branch during the Biden administration’s major industrial policy programs, shaping and passing the Bipartisan Infrastructure Law, CHIPS and Science Act, and Inflation Reduction Act through normal legislative processes. It could again instill democratic prerogatives and guardrails for industrial policy. Congress could determine when the government should take equity stakes in private companies. It could set ground rules around taking golden shares, like which corporate decisions may be subject to a government veto and which are off-limits. It could set standards around appropriate uses of revenue-sharing deals with the federal government. It could stipulate where the revenue generated from industry policy investments should go and how it should be spent. It could set rules preventing the government from unduly favoring companies where it has taken equity stakes in competitor firms. It could establish safeguards to ensure that industrial policy is not used to enrich public officials. Meanwhile, the courts may need to revisit legal doctrines like standing to ensure that some party can contest deals of questionable legality, like the Nvidia revenue-sharing agreement or the Intel equity swap.

If Congress fails to act, it could be understood as blessing Trump’s mode of industrial policy by one-man dealmaking. That doesn’t just invite corruption, but is also a recipe for managerial disaster. The Obama administration’s masterful overhaul of Detroit is arguably less remembered today than the failure of Solyndra. After receiving an Energy Department loan, the solar panel firm’s bankruptcy led to a series of high-profile investigations that ultimately uncovered no criminal wrongdoing or corruption but soured the public on Obama’s nascent clean energy industrial strategy.

With Trump’s state capitalism, the hands-off market fundamentalism era is over. The question is whether the government’s new terms of engagement with industry will be dominated by the president or deliberated on by Congress and the courts. Acquiescence by the other branches to Trump’s deals, warts and all, risks heralding the unchecked and unbalanced industrial policy of the unitary executive. Activation by them, however, could lead to a whole-of-government effort to sand away the roughest edges of Trump’s approach to craft an industrial policy that is both powerful and responsible to effectively guide the country forward.