Ben Harris on AI, Fiscal Sustainability, and the Resilience of the U.S. Economy

Episode Summary

Executive Summary: Ben Harris argues that Trump’s 2025 economic shocks—tariffs, immigration collapse, debt expansion, and Fed pressure—were smaller or more temporary than feared, while offsetting forces like data centers and stimulus blunted their impact. He then warns that oil-refining bottlenecks, not crude, are the real inflation risk and that AI is unlikely to “restore” fiscal sustainability because gains are offset by longer lifespans, higher rates, labor displacement, and defense costs.

Main Topics: Why the economy did not tank after Trump’s shocks (Priority: 5/5): Harris explains that the expected recession never arrived because the shocks were smaller in practice, markets anticipated reversals, and countervailing forces such as data-center investment and fiscal transfers supported growth. Tariffs, trade adaptation, and market resilience (Priority: 5/5): The conversation examines how headline tariff rates surged but actual collections were lower than feared because of exemptions, evasion, negotiation, and eventual legal/political rollback. Inflation and the real oil problem (Priority: 5/5): Harris distinguishes crude oil from refined products, arguing that refinery losses and crack spreads—not crude scarcity alone—pose the bigger and more persistent inflation threat, especially for diesel and jet fuel. Russia sanctions, shadow fleets, and price caps (Priority: 4/5): He reviews the novel oil price-cap sanctions regime he helped design, explaining how it constrained Russian revenue but was undermined by evasion, shadow tankers, and weak follow-through by the Trump administration. U.S. fiscal sustainability and rising interest costs (Priority: 5/5): Harris lays out the large structural deficit problem, rising Treasury risk premia, and the added burden from the One Big Beautiful Bill, arguing the fiscal outlook is still deteriorating. Can AI solve the fiscal problem? (Priority: 5/5): The paper’s core conclusion is that AI will help growth but not enough to fix the budget, because longer lives, higher equilibrium rates, more safety-net spending, and defense outlays absorb much of the upside. Policy responses: broaden capital ownership and update the social contract (Priority: 4/5): Harris recommends helping households own more capital through retirement-matching reforms rather than government equity stakes, while also rethinking taxes, benefits, and labor-market support.

Key Arguments: The tariff shock looked enormous on paper, but realized revenue and economic drag were much smaller than expected because rates fell, refunds were expected, and firms adapted. Markets and policymakers likely anticipated that the most extreme attacks on Fed independence would not fully materialize, limiting the shock to financial conditions. The economy absorbed shocks because offsetting stimuli—especially data-center investment and cash refunds—added significant growth support. Oil’s inflation impact is now concentrated in refined products, especially diesel, because global refining capacity has been damaged by war and strikes. The price-cap sanctions regime limited Russian revenue without causing the global recession many analysts feared, but Russia’s evasion strategy has reduced its effectiveness. AI will likely boost productivity, but its gains are partly offset in fiscal terms by longer lifespans, higher interest rates, labor displacement, lower labor-tax revenue, and higher defense spending. Rather than having government take ownership stakes in firms, policymakers should help more households own capital through expanded retirement matches and related savings incentives. If AI causes a shift from labor to capital income, the social contract may need redesigning around taxes, benefits, housing, and retirement ages.

Data Points: Average trade-weighted tariff rate at Trump inauguration: about 2.5% - Baseline tariff level before the 2025 trade shock Average trade-weighted tariff rate after Liberation Day: high 20s% (around 28%) - Peak tariff level discussed by Harris Average trade-weighted tariff rate after volatility and court ruling: around 15%, then around 12% after Supreme Court ruling - Tariff rate fell after market and legal reactions Net immigration in the 2010s: about 1 million net immigrants per year - Historical benchmark before 2025 slowdown Net immigration in 2025: a little bit negative, near zero - Best estimates of 2025 immigration flows Projected net immigration in 2026: even more negative - Expected continuation of labor-supply decline New debt from One Big Beautiful Bill: about $5 trillion - Estimated fiscal cost of the administration’s signature bill Tariff revenue collected in 2025: about $250 billion - Amount actually raised from tariffs Extra tariff revenue over a typical year: about $200 billion more than usual - Incremental revenue due to tariff increases Tariff refunds expected: about $170 billion - Portion of tariff collections expected to be refunded Data centers’ growth contribution: about 0.5 percentage point to growth - Estimated tailwind over the past year and a half Disposable-income boost from fiscal refunds: around 0.5 percentage point in 2026 - Near-term support from bill-related refunds Strait of Hormuz crude flow: about 15 million barrels per day - Volume at risk in the oil shock discussion Pipeline bypass capacity: 6 to 6.5 million barrels per day - Crude that can bypass the Strait via pipelines Worst-case crude shortfall estimate: about 6 million barrels per day - Post-bypass shortage estimate Worst-case Brent oil price estimate: $125 per barrel, possibly $150 - Modeled price response to severe supply disruption Global refining capacity destroyed: about 10% - Losses attributed to Russian and Gulf refinery damage Estimated inflation impact of Hormuz conflict: about 0.6 percentage point to headline inflation; 0.2 percentage point to core - Pass-through estimate for 2026 Pass-through rate from oil shock to inflation: about one-third - Share of shock transmitted to consumer prices Primary deficit: about 2% of GDP - Structural budget shortfall excluding interest costs Overall deficit: about 6% of GDP - Projected annual deficit over the next 10 years Interest payments: about 4% of GDP - Current annual debt-service burden Older population under optimistic AI longevity scenario: 74 million to 76 million Americans age 65+ - 10-year demographic effect in the paper Mortality improvement assumption: 3% average annual reductions in mortality - Extreme AI health scenario used in the model Interest-rate shock in extreme AI scenario: 35% increase across the yield curve - Potential effect of AI-driven productivity and capital demand AI-defense spending in model: about $35 billion per year - Baseline estimate for AI-related military spending

Pivotal Quotes: "if you locked 100 economists in a room or a cave and let them out at the beginning of 2026 and described all these shocks and said, What do you project for GDP growth? I think everyone would have a Negative in front of it." — Ben Harris: He is explaining why the lack of recession after multiple policy shocks was surprising "the problem really isn't crude. The problem is really refined product." — Ben Harris: He is distinguishing the oil inflation risk from crude supply to refinery bottlenecks "however much you think that AI will drive down deficits, just cut that in half" — Ben Harris: He summarizes his paper’s main caution about AI optimism and fiscal sustainability

Implications: The message is twofold: markets and institutions can absorb a lot, but inflation, debt, and AI are still reshaping the economy. For listeners, the bigger risks are refining bottlenecks, fiscal crowd-out, and a need to spread capital ownership more broadly.