America Needs an AI Boom to Grow Out of Our Debt Problem. There Is No Guarantee.
Why did dire headlines and market surprise accompany Moody’s recent downgrade of its U.S. credit rating, when the facts leading to the firm’s decision have been well-known for years? Like the proverbial slow-boiled frog, the challenges associated with America’s aging demographics and growing debt levels have been building for decades. Are markets only now waking up to the uncomfortable and uncertain fiscal outlook?
Some confusion is understandable, even if the stakes seem clear. We will need to solve the mismatch between government revenue and spending, but whether that is easier or more difficult is an open question. There are two highly probable paths ahead for the U.S., and we don’t yet know which one we will follow.
On one hand, my team at Vanguard estimates there is about a 40% chance that the unchecked fiscal dynamic will result in an extended period of stubborn inflation and higher interest rates, akin to a milder 1970s. In that scenario, stocks would lag behind bonds, and interest rates could climb above 7%. A slowing economy would leave millions of Americans poorer. On this path, we estimate 20% odds that the 10-year Treasury yield (a benchmark for mortgage rates) is above 9%, or twice the current rate.
On the other hand, we also estimate there is close to a 50% chance that technology, especially artificial intelligence, could unlock productivity at scale, enabling greater growth and tax revenue offsetting some of these headwinds. In either case, we are highly likely to see the end of the status quo of stable growth, inflation, and financial returns.
These probabilities are, of course, imprecise. But they draw on a unique quantitative framework that Vanguard has developed, based on decades of research, to provide a sense of the most likely economic scenarios through 2035, driven by an interplay of factors: technology, demographics, globalization, and fiscal debt—so-called megatrends. The goal is to help investors prepare for the most likely outcomes. Probability estimates help clarify how we think about the long-term outlook, so we don’t get bogged down in today’s headlines or paralysis of analysis.
Our framework suggests that the 2030s will be shaped by a tug of war between technology as a source of growth, and age-related spending deficits as a source of weakness. Two likely scenarios emerge.
Scenario one: Productivity surges (45% to 55% probability). AI continues to advance at an increasingly faster pace and becomes a general-purpose technology, like electricity. By the first half of the 2030s, AI has raised productivity more than the personal computer and the internet. It eventually sparks transformative knock-on effects, new industries, and significant job displacement. Technology keeps inflation in check, and deficits cease to climb given higher tax revenue from stronger growth. A surge in innovation occurs thanks to more work automation and task augmentation. Costs of living are somewhat more manageable. Standards of living for the average American household double every 30 years, not every 70.
Scenario two: Deficits drag (30% to 40% probability). AI disappoints, failing to live up to the hype. Worker productivity remains sluggish due to a continued lack of augmentation and automation. Meanwhile, government deficits keep climbing. Government spending remains indefinitely higher than revenue, requiring higher borrowing, and this structural deficit raises borrowing costs. Credit slows. Home prices either fall, homeownership moves more out of reach, or both. U.S. growth becomes less exceptional and more European. Based on current prices, financial markets are unprepared for this outcome—so declining prices and increasing volatility is a new trend. Deficits raise inflation expectations, making inflation stubborn. The acceleration of price increases makes the average worker’s standard of living lower than their parents’.
On one side is the possibility of technology-fueled economic growth. On the other, we have the near-certain weight of fiscal deficits, aging demographics, and rising borrowing costs. Which side wins will affect generations to come. Will AI deliver a surge in productivity even close to what electricity did 100 years ago?
Investors should plan for two future paths, not one. The “productivity surges” and “deficits drag” scenarios aren’t just possibilities, but meaningful probabilities. Together, our framework estimates the chance of one of these two forecasts occurring is greater than 80%. As investors, we need to consider portfolios that are prepared for both scenarios. The goal is to narrow the range of future outcomes and better manage risk.
If AI becomes a true general-purpose technology, the U.S. could experience a rare “best of both worlds” scenario where positive economic trends open a window for policymakers to enact pre-emptive and even gradually phased-in fiscal reform. But, critically, even if AI’s promise is realized, it only buys time. Without meaningful fiscal reform, economic growth will eventually decline.
If technology and AI disappoint, the era of tough choices will arrive sooner than markets might think. Investors can try to prepare, by shifting a greater allocation to bonds and a diversified set of value stocks, but policymakers are the key. They can control the path of the structural deficits far more than they can guarantee and harness transformative productivity gains. They have significant work to do because the water is getting warmer.