The Geography of GenAI

Episode Summary

Executive Summary: The episode opens with a macro check-in on Fed policy, tariffs, housing, and credit stress, then pivots to a deep dive with economists Scott Abrahams and Frank Levy on how generative AI may reshape labor markets, migration, education, and politics. Their core claim: AI’s employment shock may be comparable in scale to the manufacturing shock, but its effects will be more diffuse and concentrated in highly educated urban metros, potentially shifting people toward smaller, cheaper, still-urban cities like Savannah, Augusta, and Greenville.

Main Topics: Fed policy and economic resilience vs. weakening (Priority: 5/5): The hosts discuss the Fed’s decision to hold rates steady, interpreting it as consistent with a wait-and-see stance amid uncertainty from trade policy. They debate whether the economy is truly 'strong' or merely weakening without yet tipping into recession. Tariff negotiations and the China trade outlook (Priority: 5/5): The panel analyzes the U.S.-UK trade arrangement as a limited sign of a broader tariff off-ramp, but stresses that the central issue is China. They note that trade policy and tariff reductions will heavily influence recession risk. Signs of domestic economic softening (Priority: 4/5): House prices and credit delinquencies are cited as evidence of a softer economy. The hosts connect declining home prices, especially in overheated regions, and rising delinquency rates to weakening household finances and student loan repayment pressure. Manufacturing shock as the template for AI’s downstream effects (Priority: 5/5): Abrahams and Levy use the long-run effects of manufacturing decline—education shifts, migration, and political realignment—as the framework for analyzing AI. They emphasize that the key lesson is not just job loss, but how people and places adapt afterward. AI’s labor-market impact and scale (Priority: 5/5): The guests argue that AI exposure may be large enough to rival the manufacturing shock in employment terms, but unlike manufacturing it is more geographically dispersed and concentrated in white-collar, task-based work. Migration winners and losers across metro areas (Priority: 5/5): The paper predicts that highly exposed, expensive college-heavy metros may lose residents, while mid-sized, affordable, still-urban metros—especially in the South and Southeast—could gain. Savannah is highlighted as a leading beneficiary. Political and educational consequences of AI (Priority: 4/5): The guests foresee pressure on higher education returns, possible growth in nontraditional or skills-based training, and rising bipartisan demand for AI regulation as workers feel displacement.

Key Arguments: The Fed’s pause was expected, but the statement’s acknowledgment of uncertainty and dual mandate risks suggests the next meeting will be more consequential. The economy is not weak, but it is weakening; labor market indicators, house prices, and credit data all point to softening. The U.S.-UK arrangement is not a full trade deal and leaves the 10% baseline tariff in place; China remains the decisive issue for recession risk. Manufacturing’s long-run story was not just job loss but a shift from goods production to services, with fewer jobs in old manufacturing regions and major social/political adjustment. AI may produce a shock similar in employment magnitude to manufacturing, with around 3–4% of metro-area jobs highly vulnerable. Unlike manufacturing, AI exposure is concentrated in white-collar, college-educated occupations and major cities, so migration effects should favor smaller, affordable urban metros. The standard advice that everyone should get a four-year degree may weaken if AI reduces the returns to college for many occupations. Political backlash to AI will likely center on regulation and worker protection once the public sees concrete displacement, potentially reshaping partisan coalitions. The impact of AI on labor markets will depend heavily on whether government regulation slows adoption or mitigates the adjustment costs.

Data Points: Fed meeting action: No change in interest rates - The Federal Reserve held policy steady at the latest meeting. Fed rate-cut forecast: Three cuts in 2025 (July, September, December) - Mark Zandi states Moody’s baseline expects cuts beginning in July. U.S.-UK tariff change: Auto tariff lowered from 25% to 10% for first 100,000 vehicles - Part of the announced arrangement with the UK. Steel and aluminum tariffs: UK exempted - The U.S. said the UK would be exempt from these tariffs under the arrangement. China tariff discussion: Potential cut to 80% from 145% - A reported Trump statement discussed lowering China tariffs, though details were unclear. U.S.-China trade decline: About 25% - China trade with the U.S. fell sharply in the data discussed on the show. Monthly house price change: -0.3% month over month - Moody’s repeat-sales house price data for March showed an actual decline. Manufacturing shock employment loss: About 3% of total U.S. employment - Frank Levy compared manufacturing job losses from 1979–1985 to total employment. AI-exposed jobs (high risk): About 4.1 million jobs - Based on the OpenAI/Warden-style probability estimates for occupations at extreme risk. AI-exposed employment share: About 3–4% of metro-area employment - Levy and Abrahams compare AI risk to the manufacturing shock magnitude. Firms using AI tools: About 5% - Census-based estimate of U.S. firms using AI tools to produce output in the prior two weeks. Workers using AI: 20% to 40% - Fed survey estimate of workers using AI in some form at work. AI job postings, U.K.: 1.3% - Share of job postings flagged as AI-related in the UK by a labor-market data firm. AI job postings, U.S.: 1.8% - Estimated share of U.S. postings related to AI, according to the same methodology. Highest AI-posting country: Singapore at 3.3% - Among countries discussed, Singapore had the highest share of AI-related postings. Other high AI-posting countries: Hong Kong and Luxembourg - These were named as countries with higher AI-related job posting shares than the U.S. UK baseline tariff: 10% - The announcement suggested a 10% baseline tariff remained in place on most other goods. China tariff current level: 145% - The prevailing tariff level discussed before any possible reduction.

Pivotal Quotes: "the economy is strong" — Jerome Powell (as paraphrased by Chris/Mark): Debate over whether the Fed was understating weakness in the economy. "The economy is weakening" — Mark Zandi: Repeated assessment of macro conditions across housing, credit, and labor data. "There will be the AI haves and have nots" — Scott Abrahams: Explaining that AI will benefit some workers and hurt others even within the same education group.

Implications: AI may trigger major but uneven labor-market disruption, especially in white-collar cities. Expect pressure on colleges, more migration toward affordable mid-sized metros, and rising demand for AI regulation as workers and local economies adapt.