Can Democracy Survive the Next Financial Crash?
If the AI bubble bursts, Donald Trump will control a bailout apparatus with essentially limitless authority.
The bailout clock is ticking. In September, Oracle invoked “force majeure” to delay paying debt on a data center, which led to a drop in the corporation’s stock price. This month, OpenAI significantly cut its revenue projections; CoreWeave, a heavily leveraged data-center company, saw its share price plunge; and investors began racing to buy credit-default swaps to protect their portfolios. These are all signs of declining investor confidence. The risk that the heavily debt-fueled and interconnected AI sector could crash—taking the broader economy with it—is high and rising.
When the next crash happens, originating from whatever sector, the president will control a bailout apparatus with essentially limitless authority. Ever since the 2008 global financial crisis, the Federal Reserve and U.S. Treasury have claimed, with almost no resistance, more power to deal with each successive financial emergency. These bailouts were rationalized on the basis that they were in the public interest and at least theoretically subject to congressional oversight. But if the crash happens on Donald Trump’s watch, his administration would likely sidestep Congress entirely. A Trump bailout would proceed on whatever terms please the president and his allies, who would use it as a bonanza to enrich their favored companies and CEOs, even if they were the ones who caused the crisis to begin with.
The billionaires—and the trillionaire—in Trump’s inner circle have already usurped decision making about which federal agencies to decimate, which federal workers to fire, which scientific research to defund, and which industries to deem necessary for national security. They could also usurp the bailout apparatus, making it spin into action even before a crisis arrives. When they start feeling financially wobbly, they might go to Treasury Secretary Scott Bessent, Federal Reserve Chairman Kevin Warsh, or, most likely, Trump himself and request an emergency declaration to justify a bailout. They would know that the current Supreme Court majority would be unlikely to block an emergency declaration by the executive branch and central bank.
The bailout authority is a loaded gun. Once triggered, it would bring downsides that have become familiar since 2008: public deficits and debt that devour GDP, zombie companies that taxpayers are forced to carry, an expansion of wealth inequality, and a dramatic shift of losses from perpetrators to the public. In Trump’s hands, it would also feature a new and more frightening dynamic: the creation of a standing money printer that can be turned on for the president’s allies and turned off for his enemies, enabling state-supported favoritism and plunder on a staggering scale.
Historically, the rationale for a bailout was that central banks should play the role of what Walter Bagehot called the “lender of last resort,” extending loans as necessary to tide over a blameless firm—a firm that, but for circumstances beyond its control, would otherwise be solvent and able to post adequate collateral. In 1932, Congress put this function inside the Federal Reserve. Section 13(3) of the Federal Reserve Act gave the central bank the power to extend loans to banks and any corporate entity in “unusual and exigent circumstances.” During the Great Depression, the Fed made limited use of its new authority. Then, for seven decades, Section 13(3) sat idle.
The 2008 crash inaugurated the modern bailout era. When that crisis hit, its scale was unprecedented; financial markets had grown, becoming more intertwined and complex. The Federal Reserve, under Chairman Ben Bernanke, deemed the situation “urgent and exigent” and used Section 13(3) to establish special-purpose vehicles to buy up financial assets. Section 13(3) was also used to justify loans to AIG, an insurance company, and its affiliates, in exchange for collateral that consisted of nonperforming financial instruments with dubious market value. Working closely with the Treasury, which had related legal authority under the vague goal of dollar stability, the Federal Reserve began a program of extending loans to banks and other financial firms while purchasing huge quantities of mortgages and other shaky assets held by such entities as credit-card companies and private student-loan lenders.
I served as a Federal Reserve governor during the tail end of the financial crisis, and then as the deputy secretary of the Treasury in the Obama administration. I eventually came to see that the bailout of the financial sector was not doing much for the broader economy. Long after the financial sector had recovered, many Americans were out of work, unable to repay their mortgages, and contending with rapacious debt collectors and sloppy mortgage servicers. Ordinary people could not help themselves to the enormous sampling of bailout options that the Federal Reserve and Treasury had provided to the banks. Mortgage-relief programs were slow to get off the ground and undermined by arguments that the intended recipients never should have had access to subprime loans in the first place. Congress passed the Dodd-Frank Act, a financial-reform bill, in part to address the distributional consequences of what had just happened and to limit the Federal Reserve’s emergency powers. Ultimately, however, the law left largely intact the Fed’s discretion to choose bailout recipients, the interest rate charged on bailout loans, and the quality and value of the collateral taken to backstop those loans.
A decade later, when the coronavirus pandemic threatened an economic slowdown, the Federal Reserve mobilized preemptively. It initiated an even more aggressive version of large-scale asset purchases, with earlier money and more of it. Chairman Jerome Powell’s bailout opened up all of the faucets that spurt out continuous liquidity and created some new ones. Expanding its usage of Section 13(3) beyond mortgage assets, the Fed purchased the debts of various companies. In many cases, no requirements were imposed on these firms in exchange for cheap financing. By the Federal Reserve’s own estimates, the highly indebted fossil-fuel industry, for example, reaped more than $1.47 billion from the Main Street Lending Program—8.4 percent of the allocated funds—after lobbying the Fed to make the terms of its loans more generous. The Fed’s purchases of this debt “took all the risk overhang you normally associate with a risky company and took it off the table,” one chief investment officer told The Wall Street Journal at the time.
By the time Silicon Valley Bank suddenly failed, in 2023, the Federal Reserve had learned that it could bail out favored sectors without imposing any consequences on failed businesses. Even though no one would have described this midsize regional bank as a “too big to fail” institution, the Federal Reserve and Treasury drew on the techniques they had perfected during the financial crisis and the pandemic. They established the Bank Term Funding Program to lend against collateral with depressed values. Interest rates were set so low as to permit any eligible entity to borrow from the Fed rather than the private sector, even after the threat of any deposit run had subsided.
The result of these three bailouts was a collective recognition that the Federal Reserve had the power to choose winners and losers. Big banks and oil companies could be winners, the crypto sector and Silicon Valley–backed venture firms could be winners, and the CEOs and owners of these entities could be winners—even if they had acted recklessly. Families who had lost their homes and their incomes would be losers; so would entrepreneurs who aspired to displace companies that, but for a bailout, would have gone under.
This approach widened income and wealth inequality, further shrinking the middle class. The companies that had been bailed out continued to act in ways that had permitted them to shed their costs to the public. In each case, these bailouts were conducted hastily, because of the unchallenged view that allowing companies or sectors to collapse would deprive the economy of necessary public benefits. And they were conducted with little consideration of their long-term consequences.
The recent history of bailouts gives us a clear indication of what the next one will look like.
Enormous public money will be spent, in magnitudes approaching or exceeding the amount that Congress appropriates as part of its fiscal functions. The Federal Reserve and Treasury will once again choose the sectors that will benefit, and the terms on which bailouts will be offered. Only this time, the process will be under the aegis of an administration that brazenly trades political favors for money and support. The government could buy up nonperforming debt from favored firms; more nonperforming debt could be unloaded by data-center financiers and family foundations headed by allies of the president. Once the bailout machine is turned on, businesses will line up to buy favorable terms in exchange for political contributions. Those that refuse to play this game would be left to fail.
But this chaotic free-for-all is not inevitable. Congress created the government’s bailout powers, and Congress can constrain them. If it does not, then it will be implicated in the consequences. If no public benefit justifies reviving a failing firm, then Congress should allow it to fail. If failing companies are going to be rescued, then conditions must be attached to bailout funds to make sure that the public benefit will be achieved.
Congress should structure these terms before the next bailout. If the businesses being bailed out are data centers, for example, Congress could require them to limit their emissions or build their own renewable-energy sources. If the businesses being bailed out are the creators of frontier AI models, they could be required to prioritize investments in medical research. To make these judgments, Congress could establish a bailout manager to identify the public benefit, establish the conditions needed for the benefit to be delivered, and then oversee adherence to those conditions long after the crisis has passed.
Every bailout is a sign of economic failure. Without interventions by our elected representatives, future bailouts won’t just risk widening inequality—they stand to politicize our entire economic system. A president who wields the bailout gun on his own behalf could leave the country with societal damage that will take generations to remediate. It’s up to Congress to assess those dangers now and determine how to intervene, while it still has the chance.