Fed speeches: Barr ties AI to a record-low labor share of income
Governor Barr told a Yale audience that labor's share of U.S. income hit 52.8% in Q2 2026 and warned that concentrated AI could widen already high inequality.
In one of this week’s Fed speeches, Governor Michael Barr set out a set of new inequality figures at Yale University on Thursday and attached a warning to them. Speaking under the title “American Inequality and the Future of AI,” Mr. Barr drew on the Federal Reserve’s latest Survey of Consumer Finances to describe how unevenly wealth and income are distributed, and then argued that artificial intelligence could widen the gap if its returns accrue mainly to owners of capital. He said the views were his own and not necessarily those of his colleagues on the Board or the Federal Open Market Committee, and the speech did not address the near-term path of interest rates.
The state of inequality today
The starting point was the 2025 Survey of Consumer Finances, a triennial survey of nearly 4,500 households published last month. It confirms, Mr. Barr said, that both income and wealth are highly concentrated, with wealth more so. The top 10 percent of households by income receive almost half of all income, and the top 1 percent receive about one-fifth. On wealth, the top 10 percent hold nearly three-fourths of the total and the top 1 percent about one-third, which he put at nearly $60 trillion. The bottom half of households hold only 2 percent.
The gaps between households are stark when measured in dollars. A household at the lower edge of the top 1 percent has $16.3 million, which is 75 times the $215,000 of a household in the middle of the distribution. At the lower edge of the top 0.1 percent, the multiple is 300 times. That group, about 135,000 households, owns one-sixth of national wealth. Mr. Barr also cited a separate Fed survey, the Survey of Household Economics and Decisionmaking, which found that 31 percent of respondents in 2025 lacked the cash to cover an emergency expense of $500 or more.
Concentration at the very top has increased, he said. A rough estimate based on Forbes data and the survey puts the wealth share of the 400 richest individuals at 4 percent, up from 3 percent in 2022.
The composition of assets explains much of the pattern. Wealth at the top is held mainly in business assets, and the speech noted that private equity makes up 12 percent of the assets of the wealthiest 0.1 percent of households. Middle-class households rely on paychecks for most of their income, and their main assets are housing and stocks held through retirement accounts. Households between the 50th and 90th percentiles of wealth hold only 17 percent of publicly traded stock by value, even though about three-fourths of them own some. The bottom half averages $52,000 in wealth, with about 40 percent holding home equity and nearly 40 percent holding business equity.
A long view, and a falling labor share
Mr. Barr placed these figures in a historical arc. By many estimates, he said, inequality today is among the highest in the past 100 years, but it has widened and narrowed before. It rose in the Gilded Age, narrowed after the New Deal and again after World War II, when progressive taxes, public education and growing unionization helped, and began rising again in the late 1970s as top tax rates fell and business equity grew as a share of assets. His conclusion was that the level of inequality “is not immutable” and that policy choices play a large part, alongside growth, technology, education and globalization.
The sharpest new data point was the labor share of national income. After World War II it stayed between 60 and 65 percent. It began falling in fits and starts in 2001, fell faster after the pandemic, and reached a record low of 52.8 percent in the second quarter of 2026, a decline he described as especially steep over the past year. He acknowledged that some of the decline reflects measurement issues, but said a clear downward trend remains after adjusting for them.
He cited three causes from economic research: technology that reduces demand for labor, market concentration, and weaker bargaining power for workers. On concentration, he said researchers estimate it has increased significantly in three-quarters of U.S. industries since 2000, and that wages have grown more slowly than labor productivity over that time.
Why AI sharpens the question
The speech then turned to AI, which Mr. Barr said could be what some researchers call capital-biased innovation, in which returns flow mainly to capital. He sketched two outcomes. In one, competition keeps AI cheap and ubiquitous, and small businesses gain access to skills and information. In the other, AI’s dependence on data, model improvements and computing power favors scale, a few firms capture most of the investment returns, and the labor share falls further while concentration rises.
He was careful about the evidence to date. There has been relatively little displacement of workers by AI so far, he said, and there is evidence that it is raising the productivity of some workers and creating new occupations. Competition among AI firms has kept access fairly broad, and smaller models might meet many businesses’ needs. But he said there are signs AI may substitute for labor in some sectors, especially for younger, less-experienced workers, and that tasks that are easily automated with clear guardrails could see rapid substitution while those requiring judgment, coordination or creativity may be augmented.
What the Fed’s own data add to the debate
For readers who follow central bank policy, the speech is notable less for any policy signal than for the weight a Fed governor places on distribution. Mr. Barr argued that inequality bears on the economic outlook, because widening inequality can reflect a decline in economic opportunity that limits productive capacity. He also pointed to the pace of capital investment in the AI sector as something to watch closely.
His closing recommendation was to begin planning now. He said society should be prepared for serious short-term disruptions in the labor market, that the extent of any disruption will depend partly on investment in job creation, worker training and connecting workers to new jobs, and that “now is the time” to consider how to address these challenges while AI adoption is at a relatively early stage. The speech set out no specific policy proposals, and it was an individual governor’s assessment rather than a statement of Federal Reserve policy.
Prepared with AI assistance from public sources and reviewed under our editorial policy. Not investment advice.

