We Asked Vanguard’s Chief Economist Why AI Has Two Huge Tails — And Which One Wins

Episode Summary

Executive Summary: Joe Davis argues AI is likely a general-purpose technology that will boost growth, reshape markets, and eventually spread benefits beyond tech into sectors like healthcare and financial services. He sees the current phase as a build-out, not a bubble peak, while warning that if AI disappoints and only automates, fiscal deficits, rates, and equity drawdowns could become the dominant risks.

Main Topics: AI as a general-purpose technology: Davis frames AI not just as automation but as a platform capable of augmenting work and enabling new products and industries, similar to electricity or the personal computer. Megatrends framework and macro forecasting: He explains Vanguard’s model using four structural drivers—technology, demographics, fiscal deficits, and globalization—to integrate long-term trends with near-term market and growth forecasts. Market cycle, valuation, and 'where we are in the innings': Davis argues the AI investment cycle is still in a build-out phase, with markets potentially ahead of the real economy, but not yet at a dot-com-style peak. Sector winners beyond mega-cap tech: He expects AI benefits to diffuse into service-heavy sectors with unmet needs and high costs to serve, especially healthcare, financial services, education, and business services. Fiscal risk if AI underdelivers: If AI only automates and does not generate augmentation/new industries, growth could revert lower and deficits, rates, and equity valuation pressure could rise sharply. Global and geopolitical context: He places AI within a global race shaped by deglobalization, tariffs, debt burdens, and uneven national readiness, with some countries benefiting more in the second phase of adoption. Portfolio construction and risk management: Davis emphasizes scenario analysis over market timing, suggests 60/40 remains viable, and notes short-duration fixed income could be defensive in a disappointing AI scenario.

Key Arguments: Long-term structural trends affect near-term GDP, inflation, rates, and stock prices; they are not just 10-year abstractions. AI is the dominant megatrend because it can automate, augment, and create new industries, making it unlike most technologies. Roughly half of quarterly S&P 500 movement can be explained by changes in megatrends, showing macro and markets are tightly linked. The current AI cycle is likely in a build-out phase similar to the mid-1990s, not the 1999 peak, so further gains may still be ahead. AI could lift U.S. growth well above consensus, potentially to around 3% in 2027, despite zero contribution from demographics or globalization. Great technologies often produce both real economic transformation and stock-market drawdowns; a future consolidation does not necessarily mean the technology failed. If AI becomes only an automation tool, growth benefits fade, deficits reassert themselves, and bond yields could rise materially, hurting equities. The biggest equity beneficiaries of AI may ultimately be users/adopters in sectors with unmet needs rather than the AI vendors themselves. Investors should think in scenarios and probabilities rather than making binary calls based on headlines. Compounding and disciplined saving matter more than tactical market moves for most investors.

Data Points: AI impact on occupations: 80% of occupations - Davis says AI is projected to affect 80% of occupations faster than the personal computer did. Relative speed vs personal computer: 2x faster - AI is expected to affect occupations at twice the rate of the PC. Personal computer adoption timeframe: 15 years - Used as the benchmark for comparing AI diffusion speed. AI occupation impact timeframe: 4 years - Davis says AI could affect 80% of occupations within four years. Economic growth projection vs consensus: 50% above consensus - Vanguard’s growth outlook is substantially above market consensus due to AI. U.S. growth forecast for 2027: 3% - Their model projects 3% growth in 2027 with no contribution from demographics or globalization. Probability of 2% growth and 2% inflation: 10% - Davis says this once-standard steady-state forecast now has only about a 10% probability in their framework. Probability 10-year Treasury above 9%: 20% - If AI disappoints, he assigns a 20% chance the 10-year yield could exceed 9% in 5-10 years. Structural deficit level in peace time: ~6% of GDP - He says current sustainable deficits are already near 6% of GDP. Deficit range in weaker-growth scenario: 6%-10% of GDP - If AI only automates, deficits could widen materially over time. AI startup creation: ~4,000 startups - He cites massive startup creation in AI over the past three to four years. Trade share of U.S. economy: 5%-10% - Used to explain why tariff shocks have had limited macro effect in the U.S. so far. Likelihood ranking of AI scenarios: Baseline: transform along 3 dimensions; second-most likely: only automate - He repeatedly states the transformation scenario is most likely, with automation-only as the next most likely outcome.

Pivotal Quotes: "There has never been a great technology that has not had a significant drawdown in stock prices, which is code word for saying what some would say there's a bubble forms." — Joe Davis: On why great technologies can still create market corrections even when the underlying innovation is real. "I would say, you know, our projections are AI is going to affect 80% of the occupations at twice the rate of the personal computer in four years, the personal computer took 15." — Joe Davis: On the expected speed and scale of AI adoption across the labor market. "The fact is how much money I put into my own portfolio and saving and letting it rest and let the capital markets do their job is going to dwarf any so-called alpha or outsavvying the average investor." — Joe Davis: On the importance of compounding and disciplined long-term investing.

Implications: AI is likely to be a broad macro force, not just a tech story. Investors should watch for diffusion into non-tech sectors, rising productivity, and fiscal relief—or, if AI disappoints, prepare for higher rates, weaker equities, and greater need for defense.