The Warsh Plan: The Unasked Question

The New Fed's Gamble

Kevin Warsh, the new Chairman of the Board of Governors of the Federal Reserve System, is presiding over the emergence of a seductive new orthodoxy in Washington and financial markets.

Its details vary from one commentator to another, but the underlying assumptions are remarkably similar. Artificial intelligence will unleash a new productivity revolution. Economic growth will accelerate. A period of relatively low interest rates and moderate financial repression will gradually erode the burden of public debt. Stronger growth, accompanied by modest wage gains, will do the rest.

For investors, the story is irresistible. It offers a way out of America's fiscal dilemmas without painful austerity, higher taxes, or large-scale institutional reforms. At its core lies a deceptively simple proposition. Productivity will do for twenty-first-century America what it did for postwar America.

Its greatest vulnerability lies in the assumption that productivity and broad-based prosperity naturally move together.

The economy that artificial intelligence is helping to create may prove extraordinarily efficient. It may also be characterized by a widening separation between production and purchasing power.

The central question of the coming decade is therefore not whether artificial intelligence can increase output. It almost certainly can.

The more consequential question is whether the incomes required to absorb that output will grow alongside it.

Financial markets have been seduced into wagering that they will.

The answer may be less straightforward than today's optimists assume.

The Illusion of the Postwar Miracle

The postwar experience occupies an almost mythical place in the American imagination. Historians of sovereign debt such as Carmen Reinhart and Kenneth Rogoff have documented how the United States emerged from the Second World War with public debt exceeding 100 percent of GDP and gradually reduced that burden over the following decades. Meanwhile, Ronald McKinnon and Edward Shaw identified the mechanisms of financial repression that helped make such an adjustment possible.

For many observers associated with the New Fed under Kevin Warsh, the lesson appears straightforward. If America managed to grow its way out of debt once, there is no reason it cannot do so again.

But history should not be mistaken for destiny.

The America of 1950 was both the factory and the creditor of a world devastated by war. Europe and East Asia were rebuilding. Much of the world depended on American capital, American technology, and American industry. Strong demographics, rising wages, and expanding global demand allowed productivity gains to translate into broad-based prosperity. The conditions that made the postwar miracle possible were exceptional.

The world that sustained that postwar miracle no longer exists. The Global North is largely self-sufficient. American households are already heavily indebted. The extraordinary rise of China has transformed the geography of production and trade. Artificial intelligence may increase output, but it cannot recreate the geopolitical and demographic conditions that underpinned the postwar era.

History offers lessons.

It does not offer templates.

The AI Efficiency Trap

This brings us to the central assumption underlying the emerging consensus. Artificial intelligence will undoubtedly increase productivity. Few serious observers dispute that. The real question is whether rising output will be accompanied by rising incomes. As Daron Acemoglu has repeatedly argued, technological progress does not automatically translate into broadly shared prosperity. Productivity and wages can diverge. Corporate profits can rise while household purchasing power stagnates.

In truth, this divergence is not a future possibility. It has already been unfolding for decades. Rising productivity and rising incomes no longer move together. The gains from technological progress have become increasingly concentrated, while households have relied more heavily on debt to sustain living standards. Artificial intelligence threatens to accelerate this trend by increasing output without necessarily increasing the purchasing power required to absorb it.

Economies do not run on production alone. They also run on demand. An economy can become extraordinarily efficient while simultaneously generating too little income for households to purchase what it produces.

The obvious question then becomes: what happens when output grows faster than purchasing power?

The Distribution-Leverage Cycle

Modern economies have developed an answer to that question.

They substitute credit for income.

Households borrow to maintain consumption. Businesses borrow to sustain investment. Governments borrow to support demand. Debt temporarily fills the growing gap between production and purchasing power.

To understand this dynamic, I formalize this framework as the Distribution-Leverage Cycle (DLC). Drawing on the financial instability insights of Hyman Minsky and the leverage cycle framework developed by John Geanakoplos, the DLC argues that credit expansion can postpone the consequences of insufficient income growth, but it cannot eliminate them. As leverage rises, debt service absorbs an increasing share of income, the effectiveness of new borrowing diminishes, and the economy becomes progressively more fragile.

Eventually, the system encounters its limits. What begins as a productivity boom gives way to slowing consumption, mounting financial stress, and a painful correction. The economy does not fail because it produces too little. It falters because it produces more than its underlying distribution of income can sustain.

In that sense, the greatest threat to prosperity in the AI age may not be scarcity.

It may be abundance without purchasing power.

The Deflationary Delusion

Optimists will object that the problem I describe is largely imaginary. In their view, artificial intelligence will unleash such an enormous supply shock that prices will collapse, making stagnant wages largely irrelevant. If everything becomes cheap enough, abundance itself will solve the demand problem.

This argument overlooks a fundamental feature of modern economies: debts do not automatically fall alongside prices. As Irving Fisher explained nearly a century ago, heavily indebted societies are particularly vulnerable to debt-deflation dynamics. AI may reduce the cost of software, digital services, and automated production. It cannot deflate a mortgage, a car loan, or America's vast stock of public debt. When prices and incomes fall while nominal debt remains fixed, households and governments are forced to devote an ever-larger share of their income to servicing obligations inherited from the past. Deflation does not eliminate the burden of debt. It magnifies it.

Nor does abundance affect all sectors equally. William Baumol's work on cost disease reminds us that some activities remain stubbornly resistant to productivity gains. Artificial intelligence may make entertainment and information extraordinarily cheap. It cannot easily reduce the cost of land, housing, physical infrastructure, or many forms of healthcare. We have already lived through a preview of this world. Televisions, software, and digital services became progressively cheaper, while housing, education, and other essential costs moved in the opposite direction. A society in which entertainment is free but rent and debt service absorb the entire paycheck is not a utopia. It is a recipe for financial fragility.

Deflation does not rescue the consumer from the Distribution-Leverage Cycle.

It accelerates the cycle's collapse.

The Battle for the Consumers of the Future

If American households cannot absorb the abundance created by artificial intelligence, the obvious question becomes: who will? The answer is unlikely to be found within the Global North. Most advanced economies face many of the same demographic and debt constraints confronting the United States. As trade barriers rise and advanced economies become increasingly self-sufficient, the remaining frontier of demand lies elsewhere.

That frontier is the Global South. Over the last two decades, Beijing has pursued a strategy very different from that of Washington. As analysts such as Michael Pettis and Brad Setser, along with institutions like the Rhodium Group, have documented, China has systematically built trade networks, infrastructure, and financial relationships across Asia, Africa, and Latin America. Through the Belt and Road Initiative and a vast ecosystem of subsidized manufacturing, China has transformed much of the developing world into an outlet for its excess capacity.

This may prove to be one of the defining geopolitical developments of the twenty-first century. The struggle among great powers is no longer simply about resources, technology, or military capabilities. It is increasingly about consumers. In an age of artificial intelligence and abundance, the decisive advantage may belong not to the power that produces the most, but to the one that succeeds in cultivating and financing the purchasing power required to sustain that abundance.

U.S. Comparative Advantages

If Michael Pettis is right that surplus economies must ultimately find consumers for their excess production, then the challenge facing the United States becomes increasingly clear. A hyper-efficient AI economy without sufficient domestic demand cannot indefinitely rely on ever-rising credit expansion at home. It must either cultivate new consumers abroad or accept slower growth and recurring financial instability.

Fortunately, the United States possesses advantages unlike those of any other nation. Robert Triffin recognized decades ago that the issuer of the world's reserve currency occupies a unique position within the international economy. As economic historians such as Michael Hudson have observed, the dollar sits at the center of the global monetary system, giving the United States forms of financial flexibility unavailable to most countries.

The twenty-first century may require a different understanding of this privilege. Instead of viewing dollar hegemony primarily as a mechanism for attracting foreign savings or imposing financial discipline, the United States may increasingly need to employ its privileged position to expand purchasing power abroad. The strategic imperative is no longer simply to globalize American supply. It is to globalize American demand.

This requires rethinking the purpose of the international financial architecture. Instead of treating the Global South primarily as a source of raw materials or a destination for austerity programs, the United States should seek to transform the global system from one of extraction into one of demand creation.

Such a strategy need not remain abstract. It could involve productivity-linked swap lines that support domestic credit creation in emerging economies. It could include investments in digital infrastructure and public goods that lower the cost of participating in the global economy. It could also create new mechanisms capable of converting the natural resource wealth of developing countries into productive purchasing power.

Joseph Nye's concept of soft power reminds us that enduring influence extends far beyond factories and military power. America's comparative advantage lies not only in the global preference for the dollar, but also in the trust associated with its institutions, the prestige of its universities, the recognition of its brands, and the enduring appeal of its culture and way of life.

Consumers throughout the world are often willing to pay premiums for American products and services because they associate them with quality, innovation, and status. Comparative advantage is not determined by costs alone. Trust, reputation, and aspiration matter as well. These forms of intangible capital may prove every bit as consequential as natural resources, manufacturing capacity, or technological leadership in shaping the direction of the twenty-first-century world economy.

The Blind Spot

The twentieth century was defined by competition over scarce physical resources, industrial capacity, and human labor. The age of artificial intelligence presents a fundamentally different challenge. In a world increasingly characterized by digital and automated abundance, the central economic question is no longer simply how to produce more. It is how to sustain the purchasing power required to absorb what modern economies are capable of creating.

This is the question largely absent from the emerging consensus surrounding Kevin Warsh. His wager rests on the belief that artificial intelligence, moderate wage growth, and financial repression can reproduce the broad contours of the postwar experience. Implicit in that vision is the assumption that rising output and rising purchasing power will continue to move together.

History offers little reason for such confidence.

The Distribution-Leverage Cycle suggests that when incomes fail to keep pace with production, credit can postpone the consequences, but it cannot eliminate the underlying imbalance. A highly productive economy without a sufficiently broad consumer base eventually encounters the limits of leverage.

That blind spot may prove to be the defining economic gamble of the twenty-first century. If Warsh's wager succeeds, the United States may experience another era of prosperity. If it fails, the consequences may extend far beyond slower growth. They may usher in a new cycle of leverage, financial instability, and recurring crises.

Silicon Valley, Wall Street, and Washington may well succeed in engineering the most blindingly efficient engine of production in human history. The problem is that they have devoted far more attention to creating output than to cultivating the purchasing power required to sustain it. Their gaze remains fixed on mature domestic economies and established Western partners, even as the consumers of the future increasingly emerge across the Global South.

Unless Washington recognizes the changing geography of demand and integrates it into its broader economic vision, the relationship between productive power and broad prosperity may prove far less automatic than today's optimists assume.

The American AI strategy envisioned by Kevin Warsh may prove to be an extraordinarily powerful engine with no transmission.

Jo M. Sekimonyo
Political Economist, Université Lumumba

The framework discussed in this essay is developed further in the working paper:

@2024 Venus Flytrap Press | All Right Reserved..