Economists urge 'fiscal insurance' policies ahead of AI disruption

Two prominent economists are calling for the United States to begin developing policy responses to potential AI-driven economic disruption now, rather than waiting until labor market upheaval becomes acute.

Douglas Elmendorf, former Congressional Budget Office director, and Louise Sheiner, a longtime Federal Reserve Board official, argue in The Hill that the uncertainty surrounding AI's economic impact makes advance preparation essential. While predictions range from broadly shared prosperity to mass job displacement and extreme inequality, the magnitude of potential disruption warrants what they call "fiscal insurance" — policy frameworks ready for deployment before they become urgently needed.

Why it matters

The economists point to the "China shock" as a cautionary example, where policy responses came only after concentrated economic and social damage had become entrenched. If AI concentrates income among capital owners, they warn, the resulting political power could make later policy interventions far more difficult. Developing credible options now, while political and economic conditions allow, could prove critical.

Worker displacement programs need modernization

The first risk Elmendorf and Sheiner identify is large-scale worker displacement. While AI may create new jobs, workers losing positions may struggle to find roles matching their skills or location, potentially facing prolonged unemployment or permanent earnings losses.

The existing Trade Adjustment Assistance program offers a starting point but requires workers to prove their job loss resulted from a specific cause — an approach the economists say makes little sense when AI, technological change, trade, and other forces interact. They propose developing a modernized adjustment assistance system available to most displaced workers regardless of cause.

Such a system might include temporary income support, occupational training, job-search assistance, and wage insurance to offset earnings decreases when workers accept lower-paying positions. Important questions about eligibility, generosity, and interactions with unemployment insurance remain unresolved, and evidence on training program effectiveness is mixed, requiring further experimentation.

Addressing concentrated capital ownership

The second major risk involves AI sharply increasing capital's share of national income. Because capital ownership is highly concentrated, this outcome could widen income and wealth disparities while increasing the political influence of asset owners.

Familiar responses include raising taxes on capital income, wealth, inheritances, or consumption. But Elmendorf and Sheiner also propose broadening financial asset ownership itself — an approach they suggest might prove more politically durable than tax increases.

One option would be a sovereign wealth fund acquiring assets on behalf of the public, though this raises difficult questions about financing, investment control, and preventing political interference. Another approach would place equity stakes in individual accounts, giving people direct claims on capital income, though design details would critically determine whether this preserves broad ownership or allows erosion through individual management decisions.

Timeline for action

Both policy directions require substantial development and testing before national deployment. The economists emphasize that prudent policy means beginning this work immediately — not because the need is certain, but because credible, well-crafted options cannot be designed overnight.

Collecting and interpreting evidence takes time, while developing, legislating, and implementing new policies takes longer still. Their central argument: better to have these tools ready and unused than to need them urgently without preparation.

These details were first reported by The Hill in an opinion piece by Elmendorf and Sheiner.