Photo by ChatGPT
We tell ourselves a reassuring story about technological change. Machines, software, and robots might eliminate some jobs, but they create far more than they destroy. Productivity rises, wealth expands, wages follow, and living standards improve. Disruption to the labor market is temporary. In the end, more jobs are created.
For most of modern economic history, that story described reality well enough to serve as a national creed. When we debate automation, we default to the same conclusion: it will be disruptive, but the economy will adjust.
But that narrative contains a blind spot — one that becomes visible once we stop assuming that employment naturally expands alongside output.
In the 20th and early 21st centuries, we lived in an industrial economy, and that assumption held. When companies were profitable, they built new factories, introduced new products and hired more people. The value of companies went up when they hired more workers, because investors saw it as a sign of future prosperity. Manufacturing boomed, and new sectors, such as healthcare and expanding government services, absorbed displaced labor. Even when technology eliminated specific occupations, overall employment growth kept pace with output.
There was always a demand for more labor, and new jobs were constantly being created.
That pattern began to shift in the late twentieth century. The recoveries following the 2001 recession and the 2008 financial crisis introduced the term “jobless recovery”. Output returned. Corporate profits recovered. But job creation didn’t. Businesses met rising demand with technology, while keeping hiring modest or non-existent. Labor expenses became something to avoid instead of a sign of financial health. [1]
Employment did not stop growing altogether. But the relationship between growth and hiring became uncoupled. Economic expansion continued, and the link between economic growth and rising wages weakened. Spells of unemployment became longer, while more workers left the labor force entirely.
Businesses adopt technology for one primary reason: to lower costs. If a new system does not reduce expenses, increase output, or ease a bottleneck, it is not adopted. In a in every economic sector, the highest controllable cost is usually labor. The most valuable innovations are those that allow firms to produce the same — or greater — output with fewer workers, reducing their bargaining power.
This challenges the assumption that productivity gains automatically translate into broad-based wage growth. A pattern once thought impossible is becoming familiar: Productivity rises while hiring stagnates. Output is increasing while payrolls remain flat. The link between growth and mass prosperity is weakening. The national economy is healthy and growing, but the working class’s share in that growth is declining.
For decades, productivity and rising wages moved broadly together. That relationship has weakened. [2]
Official unemployment statistics often suggest a resilient labor market. But the headlines ignore important trends: underemployment — working at a job below one’s education or abilities — persists, and discouraged workers fall entirely outside the headline unemployment measure. [3]
More important is a longer-term trend. In several recent recoveries, output has rebounded faster than employment. Productivity gains have outpaced hiring. Businesses have learned to extract more output from fewer employees.
That pattern matters. When productivity growth exceeds job creation, the economy expands without generating enough wage income to sustain demand. Workers can’t afford the products they are making. Growth continues — but it becomes increasingly reliant on profits, asset appreciation, and consumption concentrated at the top of the income distribution.
That is not a collapse, not at first. It begins as a gradual redistribution of wealth.
The effects of that redistribution are more easily seen at the household level. The real estate market is strained because many buyers cannot afford today’s prices and rates. Housing costs, including rent, have risen faster than many households’ incomes. Healthcare costs continue to rise. Student loan balances and credit-card debt — often used to build or maintain a middle-class lifestyle — have expanded even as headline economic indicators look healthy. [4] [5]
At the same time, consumption patterns have shifted. Buy-now-pay-later financing has moved from the margins into the mainstream, serving as a pressure valve when paychecks fail to keep pace with necessities. [6]
This is happening when productivity is rising, and corporate earnings are strong and healthy. It illustrates a divergence between the elite wealthy and everyone else. Output is expanding. Profits are rising. Asset prices are appreciating. Yet wage income, particularly for workers outside high-skill sectors, is constantly slipping behind the cost of living.
When that divergence persists, growth becomes disconnected from wage gains and more reliant on investment income and increased consumption of luxury goods and services. The system functions, but it functions differently than when productivity gains drove rising paychecks.
Wealth shifts toward investments and away from labor when large corporations expand output using automation, software integration, and AI. That means shareholders benefit. Asset holders benefit. Highly specialized workers benefit. However, wage growth lags behind technology-driven economic expansion, allowing higher-paid workers to buy investments such as real estate, equities, and bonds, thereby accelerating inequality and widening the gulf between rich and poor.
At the same time, labor demand decreases, pushing people to the margins of the economy.
Consider customer-service operations. For years, firms reduced costs by offshoring call-center work. AI systems now draft responses, resolve routine disputes, and triage complex cases before humans are needed. If software can perform a substantial share of a role at a lower cost than human workers require, employers hire fewer full-time employees.
When productivity gains no longer require mass employment, demand has to be sustained somehow — through credit, asset inflation, and high-income consumption.
That model can persist for years. It is not inherently unstable in the short run. But it changes the distribution of risk and opportunity in ways that standard employment statistics obscure. The consumer economy, however, is built on mass consumption. Without widespread purchasing power, consumer economies cannot exist.
The question is not whether technology will continue to increase productivity. It will. The question is whether the American economy can survive with a twentieth-century social contract in which productivity gains automatically support mass wage growth.
If technology continues to expand output while reducing the need for employment, the link between growth and shared prosperity decreases. Broad-based wealth will likely continue to rise, largely because of healthy investment returns for those who can afford them. But access to that growth will become increasingly limited.
The comforting story about automation raising all boats might still hold, but it is looking less and less likely. If it is to continue, productivity must be reconnected to paychecks. Absent that reconnection, the economy may not collapse. It might transform into something very different. Maybe a high-tech version of the feudalism of the Dark Ages.
The United States has experienced economic transformations before. The shift from agriculture to industry displaced millions of workers, yet new forms of employment emerged alongside rising mass demand. The transition from heavy manufacturing to services altered regional economies but did not sever the connection between productivity and broad employment, although it did result in lower wages.
What makes the present moment different is the growing possibility that advanced technologies may scale without proportionally scaling human labor. If output can rise indefinitely while payrolls remain restrained, the traditional mechanism through which prosperity spreads — wage growth tied to productivity — weakens. Growth continues, but its distribution narrows.
That narrowing does not announce itself dramatically. It appears gradually in labor-force participation rates, in the growing share of income flowing to capital, and in household balance sheets that feel tighter than headline statistics suggest. Over time, it reshapes expectations about what economic progress means — and who it is for.
And not everyone benefits equally from the transformation.
Sources and Notes
[1] Federal Reserve Bank of Boston, “Understanding the ‘Job-Loss Recovery’” (Research Review). https://www.bostonfed.org/publications/research-review/january-august-2004/1/understanding-the-job-loss-recovery.aspx; and Federal Reserve Bank of Chicago, Chicago Fed Letter (Dec. 2003) on low employment growth after 2001. https://www.chicagofed.org/publications/chicago-fed-letter/2003/december-197
[2] Economic Policy Institute, “The Productivity–Pay Gap” (updated Jan. 16, 2026). https://www.epi.org/productivity-pay-gap/
[3] U.S. Bureau of Labor Statistics, CPS concepts and definitions (including “part-time for economic reasons”). https://www.bls.gov/cps/definitions.htm; and BLS alternative measures of labor underutilization (U-6). https://www.bls.gov/news.release/empsit.t15.htm
[4] Federal Reserve Bank of New York, Quarterly Report on Household Debt and Credit (Feb. 10, 2026 release): credit card balances and student loans. https://www.newyorkfed.org/newsevents/news/research/2026/20260210
[5] Harvard Joint Center for Housing Studies, America’s Rental Housing 2024. https://www.jchs.harvard.edu/americas-rental-housing-2024; and U.S. Census Bureau (ACS 1-year estimates press release, Sept. 11, 2025) on owner costs. https://www.census.gov/newsroom/press-releases/2025/acs-1-year-estimates.html; and National Association of Realtors, Housing Affordability Index. https://www.nar.realtor/research-and-statistics/housing-statistics/housing-affordability-index
[6] Federal Reserve Bank of Richmond, “Buy Now, Pay Later: Recent Developments and Implications” (Economic Brief EB 26–05, 2026). https://www.richmondfed.org/publications/research/economic_brief/2026/eb_26-05; and Federal Reserve Board, Report on the Economic Well-Being of U.S. Households in 2024 (BNPL usage). https://www.federalreserve.gov/publications/2025-economic-well-being-of-us-households-in-2024-banking-and-credit.htm