There’s an 82 percent chance the 10-year Treasury yield will exceed 7 percent before the end of the decade, according to Joe Davis, global chief economist at Vanguard. The odds of yields topping 9 percent? One in five.
Those oddly specific projections are the starkest takeaway from Coming Into View, Davis’s new book on how megatrends like AI and structural deficits will shape markets. He presented the findings Wednesday at Nasdaq’s New York headquarters, ahead of Vanguard’s 50th anniversary bell-ringing.
The specificity of Davis’s projections has raised eyebrows, but he’s quick to stress that they aren’t guesswork. “These aren’t just numbers. I don’t personally make it up. This is coming scientifically from our framework.”
His warning comes as yields climb. The 10-year touched 4.6 percent this week, while the 30-year hit 5.09 percent after Moody’s downgraded the outlook on U.S. government debt. Investors are watching a proposed package of spending cuts and tax reforms, with concern mounting over Washington’s persistent reliance on bond markets to fund growing deficits.
Davis emphasized the projections are not short-term forecasts, but probability-based scenarios built on a new economic model that accounts for artificial intelligence, demographics, and fiscal trends. “We’re not saying this is imminent,” he told a group of journalists. “But we’re assigning probabilities to scenarios that have been vastly underappreciated by markets.”
Deficits vs. AI
Davis’s framework hinges on two competing forces: the drag from rising structural deficits and the potential boost from artificial intelligence.
“We are facing a tug-of-war,” he said. “On one side, structural deficits and aging demographics. On the other, the potential for AI to deliver a once-in-a-generation boost to productivity.”
Vanguard assigns a 50 percent probability to the AI-driven scenario, where automation lifts U.S. GDP growth to 3 percent or more. The chance of a deficit-dominated outcome, where persistent borrowing crowds out private investment and drives rates higher, is 35 percent. A continuation of the current low-growth, low-inflation environment gets just 17 percent.
The projected spike in yields comes directly from the deficit scenario. “It’s not debt levels, debt to GDP, per se,” Davis said during the Nasdaq presentation. “In fact, there’s no correlation between debt levels and future economic growth. It’s the structural deficit. It’s the fact that the deficits are rising each and every year.”
That factor, he added, is now “the highest since the 1970s.” According to Davis, it is “over 70 percent correlated with expected inflation over the past 100 years and with long-term interest rates” and is a key driver of the neutral rate.
Markets, he said, have only begun to react. “Right now, they’re small,” Davis said. “Recent headlines, though, are pointing to the fact that, yeah, this could be a growing issue in the future.”
“We are not talking about materially higher interest rates in our projections anywhere in the next one or two years,” he said. But if deficits continue rising and AI underdelivers, “in five years’ time, we can assess the probabilities of financial market outcomes. There’s a 22 percent probability that 10-year Treasury is over 9 percent and there’s an 82 percent [chance] the 10-year Treasury is over 7 percent.”
Portfolio implications
In a deficit-dominated world, fixed income outperforms equities. Higher real rates lift bond returns, while stocks face falling valuations, weaker earnings, and tighter financial conditions. “In a that scenario, equities underperform bonds,” Davis said. “That’s not something markets are priced for right now.”
In a deficit-dominated world, fixed income outperforms equities. Higher real rates lift bond returns, while stocks face falling valuations, weaker earnings, and tighter financial conditions. “In a that scenario, equities underperform bonds,” Davis said. “That’s not something markets are priced for right now.”
Vanguard’s model portfolio reduces exposure to U.S. growth stocks, especially large-cap tech, and favors global value, fixed income, and non-dollar assets. “You should underweight the growth part of the market and overweight, conversely, value stocks and non-U.S. dollar markets.”
Active management plays a role too, particularly in fixed income, where inefficiencies remain. “There’s room for active,” Davis said, “provided fees are low and skill is demonstrable.”
Value wins either way
Even if AI transforms the economy, growth stocks may not lead. “If AI turns out to be like electricity or the internet, it’s the second wave of adoption, in banks, factories, hospitals, that delivers the real productivity,” Davis writes. That dynamic tends to favor value stocks over early-cycle tech leaders.
Despite recent market enthusiasm for AI megacaps, Davis suggests their leadership may fade as broader macro trends take hold. “Even if you believe AI will save the day, that’s not a reason to be 80 percent equities,” Davis said. “These scenarios carry very different risk profiles, and investors need to prepare, not predict.”