Back to Newsroom

Warsh outlines an AI scenario with faster growth and higher unemployment

Fed Chair Kevin Warsh outlined three possible AI outcomes. One pairs faster growth with labor disruption—a distinction that matters for earnings and demand.

Federal Reserve Chair Kevin Warsh outlined a possibility in prepared remarks published Monday that complicates the simplest AI investment story: economic growth could run above its long-run average while unemployment rises above levels consistent with full employment.

That is one of three illustrative outcomes the Fed’s staff has begun examining, not a forecast. The distinction matters because an improvement in what businesses can produce does not establish how quickly displaced workers find new jobs, how widely income gains spread or which companies earn a return on their investment.

The scenario between a boom and a bust

In his November 16 opening remarks for the Productivity and Jobs Task Force Conference in San Francisco, Warsh described staff work using simplified, model-based frameworks. He said the scenarios are a starting point for discussion and that many other outcomes are possible.

The most favorable case combines real GDP growth persistently above its long-run average with gains shared broadly by workers and owners of capital. Employment stays close to its full-employment level.

The middle case keeps the stronger growth but changes how it reaches people. Automation replaces some jobs rather than helping people do them, while obstacles to moving into new work leave unemployment above full-employment levels.

The third case starts with substantial investment and stock-market gains, followed by disappointing expected returns. Valuations fall, economic activity slows and unemployment rises more sharply. In that scenario, the eventual recovery arrives without the anticipated productivity gains.

These are different mechanisms, not three probability-weighted predictions. The published remarks do not provide model equations, numerical results or likelihoods. They also announce no interest-rate decision.

Higher output does not settle the income question

The middle scenario is the important analytical distinction. It allows the technology to deliver an economic benefit while the adjustment remains difficult for workers. An AI boom would not have to fail technically for labor-market disruption to matter.

For a business, producing more with the same resources can create room for higher margins, lower prices or expansion. Which of those occurs determines who captures the gain. A supplier’s improved efficiency, a customer’s lower bill and an employee’s income are related outcomes, but they are not interchangeable measures.

That creates a separate question for companies selling to households. Stronger aggregate output alone would not establish that their particular customers have more income to spend. Under the disruptive scenario, the distribution and timing of the gains would matter alongside the headline growth rate.

Warsh left the productivity breakthrough unconfirmed

Warsh argued that advanced AI could accelerate the generation of ideas and sustain faster growth. He also said recent productivity growth has been reasonably strong without yet showing a definitive, structural break upward. That qualification limits how far investors can take the optimistic argument today.

There is a substantial counterargument to treating the middle scenario as the likely destination. Technology can complement workers, create new tasks and support expansion. Better matching between workers and jobs could spread the benefits. Warsh’s favorable scenario explicitly leaves room for broad gains and near-full employment; examining disruption does not mean the Fed has selected it as its central outlook.

Monday’s Morning Brief flagged the conference before these remarks were available. The new information is Warsh’s presentation of three illustrative staff scenarios for productivity, employment and investment returns—not a policy signal about the next rate move.

For investors, that separation is more useful than treating every AI efficiency claim as evidence of the same economy-wide boom. Company results can show whether spending produces cash returns. Productivity, employment and income data can show whether the gains extend beyond the initial beneficiaries. A convincing investment case needs to explain which of those claims it is making.

More from the Newsroom