Will the K-Shaped Economy End on Its Own?

Technological Progress, Wealth Distribution, and Capitalism’s Next Adjustment

The term “K-shaped economy” originally described what happens when different groups emerge from an economic crisis along sharply divergent paths. Some see their incomes and assets recover quickly and continue to rise, while others remain stuck with slow wage growth, limited savings, and mounting pressure from the cost of living. One line moves upward while the other falls behind, forming something that resembles the letter K.

If this divergence were simply a temporary consequence of the business cycle, the problem would be relatively straightforward. As employment recovered, wages rose, and economic growth resumed, the two lines should eventually begin moving closer together.

But what if they do not?

Over the past several decades, a more persistent pattern has become increasingly visible in advanced economies. People who own stocks, real estate, businesses, and other capital assets can accumulate wealth through appreciation and investment returns. People who depend primarily on wages must first use their income to pay for housing, food, healthcare, transportation, education, and other basic expenses. Only what remains can be saved or invested.

If wages, living costs, and asset prices continue to grow at very different rates, that gap can begin to reinforce itself.

At that point, a K-shaped economy is no longer simply the shape of a recovery. It raises a much deeper question: As technological progress creates more wealth, through what channels does that wealth reach different households? Will today’s divergence eventually correct itself? And if it does not, what mechanisms can restore a sustainable balance?

Is the K-Shaped Economy Cyclical or Structural?

Economic growth has never benefited everyone at exactly the same time. Every recession hits some industries harder than others, and every recovery favors certain sectors before others. The existence of unequal outcomes alone therefore does not prove that the economic system has fundamentally changed.

The more important question is whether those differences can correct themselves.

A household that depends mainly on wages can still accumulate wealth if earnings rise over time with productivity. Homeownership, retirement accounts, savings, and long-term investment can gradually turn workers into asset owners as well. Even if wealthy households accumulate assets faster, inequality does not necessarily become structurally destabilizing as long as ordinary households continue to make meaningful economic progress.

The situation changes when housing, healthcare, education, and other basic costs absorb an increasing share of wage gains while stocks, real estate, and business assets appreciate much faster.

Those who already own substantial assets can use investment returns to acquire even more assets. Those with few assets must first meet current expenses, making it increasingly difficult to enter the asset market at all.

An income gap then becomes a wealth gap, and the wealth gap generates additional income through capital returns. Capital produces more capital, while labor income must first pay for everyday life.

Once that feedback loop becomes established, K-shaped divergence begins to look structural rather than temporary.

Why Technology Can Widen the Divide

Technological progress itself is not the problem. Much of the extraordinary improvement in living standards over the past two centuries has come from rising productivity.

Steam power, electricity, automobiles, computers, and the internet all eliminated old jobs while creating new industries and occupations. There is no reason to assume that artificial intelligence must automatically be an exception to that historical pattern.

The more important question is not whether technology raises productivity. It is how the additional value created by higher productivity is distributed.

Imagine a company that once required 1,000 employees to produce a certain amount of output. With AI and automation, it can now produce even more with 600 employees. From the standpoint of the economy as a whole, productivity has increased.

But the value created by that improvement can go in very different directions. The company could lower prices. It could raise wages. Employees could work fewer hours. The company could invest in new businesses and create new jobs. Or a larger share of the gains could appear as higher profits, dividends, share buybacks, and company valuations.

Technology itself does not determine the balance among those outcomes.

Research by the International Labour Organization and the OECD has pointed to long-term pressure on labor’s share of total income, with automation, globalization, changes in industrial structure, and workers’ bargaining power all playing potential roles. The fact that outcomes differ considerably across countries is itself important: technology cannot be the only variable.

A more accurate way to put it is that technology influences how much wealth an economy can create, but it does not independently determine who ultimately receives that wealth. Ownership, competition, corporate structures, labor institutions, and public policy all affect the result.

What AI may accelerate, therefore, is not simply inequality itself, but a change in the relative economic importance of capital and labor.

Will Time Eventually Solve the Problem?

This may be the easiest part of the debate to misunderstand.

One optimistic argument holds that every major technological revolution initially disrupts employment, but workers eventually acquire new skills, companies create new occupations, wages recover, and the economy finds a new equilibrium. The Industrial Revolution, agricultural mechanization, and the computer revolution all seem to offer historical support for this view.

There is considerable truth in that argument. So far, there is not enough evidence to conclude that artificial intelligence will inevitably produce permanent mass unemployment.

But history has another side.

Looking back at industrialization, it is easy to say that technological progress eventually raised living standards for ordinary people. What is easier to overlook is that this transition occurred alongside enormous institutional change. Labor unions and collective bargaining expanded. Minimum-wage and worker-protection laws developed. Public education became widespread. Social insurance systems emerged. Tax systems changed, and governments gradually developed new approaches to monopoly and market competition.

In other words, history was not simply a story in which technology created problems and time solved them.

A more accurate description is that technology changed the economic structure, those changes created new social pressures, and political and economic institutions gradually adapted to the new reality.

There is therefore little reason to assume that another ten or twenty years will automatically cause today’s K-shaped divergence to disappear.

Time is neither an economic policy nor a distribution mechanism.

The Deeper Problem May Begin Before Redistribution

Discussions about inequality often move almost immediately toward familiar political questions: Should the wealthy pay more taxes? Should governments provide more benefits? How extensive should social programs become?

Those are legitimate questions, but they largely concern what happens after income and wealth have already been generated.

If K-shaped divergence is structural, we also need to ask an earlier question: How are the gains from economic activity distributed in the first place?

Suppose AI allows a company to increase productivity by 30 percent. How does that additional value enter the economy? How much becomes higher wages? How much reaches consumers through lower prices? How much finances additional investment? How much becomes corporate profit, and how much ultimately appears as capital gains for shareholders?

Economists sometimes describe this as the primary distribution of income—the distribution that occurs through wages, profits, and returns to capital before taxes and government transfers substantially alter the result.

If an increasing share of productivity gains flows toward capital while labor income grows more slowly, the burden placed on redistribution after the fact naturally becomes larger. Political debate can then become trapped between two increasingly polarized positions: one side demanding more taxation and redistribution, the other arguing that government is penalizing investment and success.

But those are not the only possibilities.

Greater worker bargaining power, stronger market competition, reduced monopoly rents, profit sharing, employee ownership, broader retirement-asset participation, lower housing and healthcare costs, and education and training that help workers adapt to new technology can all affect how the initial gains from economic growth are distributed.

The common question behind these approaches is not simply how much government should take away after wealth has been created.

It is how ordinary households participate in productivity growth in the first place.

AI Could Turn an Income Divide Into an Ownership Divide

Artificial intelligence introduces another complication that was less pronounced in earlier waves of automation.

Traditional machinery initially replaced large amounts of repetitive physical labor. AI is increasingly entering writing, translation, programming, legal research, financial analysis, design, customer service, and even some management functions.

This technological shock therefore may not fall first—or exclusively—on the lowest-paid workers.

More importantly, AI could affect inequality in two opposing directions at the same time.

Some research suggests that AI tools can help less experienced or lower-productivity workers perform complex tasks more effectively. If that effect becomes widespread, AI could actually reduce productivity differences among workers within the same occupation.

But there is another possibility. If companies use AI to reduce demand for certain kinds of labor while owners of capital receive higher profits and investment returns, the wealth gap between asset owners and those without substantial assets could widen even if wage inequality among workers does not increase dramatically.

The result could appear paradoxical: AI might make ordinary employees substantially more productive without increasing their incomes by the same proportion. Skill gaps among workers could narrow while wealth gaps between capital owners and non-owners expand.

That is why future debates about the K-shaped economy may increasingly need to look beyond wages.

The relevant questions will also include who owns companies, stocks, real estate, intellectual property, and the other assets that capture the returns generated by technological progress.

Does the Answer Have to Be Government Redistribution?

Not necessarily.

If the initial distribution of economic gains becomes increasingly uneven, taxes and social programs can certainly offset part of the difference. Modern capitalist economies have long done this to varying degrees.

But relying almost entirely on taxes and transfers after the fact can produce its own political tensions. One group may believe redistribution remains insufficient, while another may see government as increasingly interfering with private wealth and investment.

A broader question is how economic growth itself can involve a larger share of the population.

One approach is to strengthen labor income so that wages capture a greater share of productivity growth. Another is to reduce the burden of housing, healthcare, and education so households retain more income with which to build assets. Still another is to broaden access to retirement accounts, employee ownership, profit sharing, or other forms of capital ownership, allowing more workers to participate directly in investment returns.

Competition policy matters as well. If the gains from productivity become concentrated in a small number of firms with enormous market power, technological progress can continue while its economic benefits remain highly concentrated.

There is no single formula for combining these approaches, and different societies may make very different choices.

The real choice, therefore, is not simply between “free markets” and “government redistribution.” The more practical question is this: In an economy where capital and technology are becoming increasingly important, how can enough people continue to share in productivity growth?

What Happens If Institutions Fail to Adjust?

Changes in economic structure eventually enter politics.

When a substantial part of the population feels that steady work no longer provides a realistic path toward homeownership, savings, or asset accumulation, while another group becomes wealthier largely through rising asset values, that divergence is unlikely to remain an exclusively economic issue.

It may produce demands for higher wages or expanded social protection. It may contribute to renewed labor organizing, stronger antitrust enforcement, tax reform, or changes in housing policy. It can also generate growing dissatisfaction with globalization, free trade, immigration, financial markets, or large corporations.

Different political movements will offer very different explanations for those frustrations and very different solutions.

If political and economic institutions can absorb these pressures, the economy may gradually move toward a new equilibrium through reform.

If they cannot, social pressure can eventually take more disruptive forms.

Wars, revolutions, mass strikes, and periods of severe social conflict have historically altered distributions of wealth and political power. But that does not make social upheaval an effective or desirable mechanism of redistribution. Upheaval can destroy an old system without creating a fairer one. It can just as easily produce economic collapse, capital flight, authoritarianism, or even greater inequality.

Social instability is therefore better understood as a possible consequence of failed adjustment than as a solution.

When a society loses the ability to adapt through normal institutions, adjustment does not necessarily stop. It simply becomes more costly.

How Long Can the Work-Wage-Consumption Cycle Continue?

At this point, the K-shaped economy becomes more than a question of inequality.

Industrial society developed around a powerful economic cycle. Businesses invest capital. Workers provide labor. Businesses pay wages. Workers use those wages to purchase goods and services. Businesses receive revenue and invest again.

Capital, labor, and consumption reinforce one another.

Modern capitalism over the past two centuries has depended heavily on this relationship.

But suppose an increasing share of production can be described as:

Capital + AI + automation → goods and services

while the amount of labor required to produce the same output continues to decline.

A new question eventually arises: If a declining share of income reaches households through labor, who will purchase the expanding quantity of goods and services the economy is capable of producing?

This does not mean automation must inevitably cause insufficient demand. New industries can create new jobs. Lower prices can increase real purchasing power. Investment returns can reach ordinary households through retirement funds and other forms of ownership.

But it does raise a long-term question that cannot simply be dismissed: If capital becomes increasingly important in production while labor income becomes relatively less important, the ways ordinary households obtain purchasing power may also have to evolve.

At that point, wealth distribution is no longer only a question of fairness.

It also becomes a question about the sustainability of the economic cycle itself.

A K-Shaped Economy Will Not Necessarily Disappear If We Simply Wait

The K-shaped economy should therefore be understood neither as the inevitable final stage of capitalism nor simply as a temporary side effect of a technological revolution.

It is better understood as the product of several forces interacting at once: technology, ownership of capital, labor institutions, market competition, and public policy.

Technology can make the economic pie larger, but it does not automatically determine how the additional pie will be divided. Markets can allocate resources with remarkable efficiency, but they do not guarantee that productivity gains will flow into wages in any fixed proportion. Governments can redistribute income through taxes and social programs, but redistribution after the fact may not indefinitely compensate for growing imbalances in how economic gains are initially distributed.

The fundamental question for the future may therefore be larger than how much more the wealthy should pay in taxes.

It is this:

If technology becomes increasingly powerful and capital increasingly important to production, how can a society ensure that most ordinary people continue to participate in economic growth and share in the wealth it creates?

The answer could involve stronger labor income, broader asset ownership, greater market competition, lower basic living costs, more effective public services, and some degree of redistribution. Different societies will choose different combinations, and there is little reason to assume that we already know the final answer.

But one point matters more than the precise policy mix: time alone will not make the adjustment for us.

If institutions evolve while the tensions remain manageable, the two arms of the K may still begin to move closer together. If they fail to adjust for too long, economic divergence can gradually become political polarization and, eventually, a question of social stability.

History does not tell us where every such adjustment will end.

It tells us something more modest, but perhaps more important:

When the underlying structure of an economy changes, society eventually has to find a new balance that can live with it.

By Voice in Between