The Organisation for Economic Co-operation and Development (OECD) has indicated that while a potential productivity surge from artificial intelligence (AI) could buy time for highly indebted developed economies, it is insufficient on its own to reverse the trend of rising debt ratios. Fiscal consolidation will still hinge critically on demographic trends, tax policies, and spending choices.
Preliminary estimates shared with Reuters by the OECD and three economists suggest that if AI boosts labor productivity and employment over the long term, it could slightly ease the debt burden for OECD nations compared to baseline forecasts, but the impact would be limited.
Several experts noted that AI exerts countervailing forces on both tax revenues and government expenditures. Distributional issues could lead to lower tax receipts, while potential increases in average wages might drive up social security spending.
For markets, the growth generated by AI might temporarily reduce the scrutiny bond investors place on public finances. However, rating agencies and multiple interviewees stressed that uncertainty remains high. They warned that if a recession occurs before productivity benefits materialize, rising financing costs could swiftly return debt concerns to the forefront.
Debt pressure is now a "hard constraint," with AI likely to delay rather than reverse the problem.
Economists cited by Reuters stated that if an AI-driven productivity boom materializes, it could indeed help major economies "manage their public finances better" and provide some buffer against the penalties of fiscal expansion, but it would not eliminate the challenges entirely.
Most wealthy economies already carry debt exceeding 100% of GDP and face multiple upward pressures: the costs of aging populations, interest payments, and demands for defense and climate-related spending.
Simultaneously, in an environment where government bond yields in developed economies have risen significantly post-pandemic, bond investors have less tolerance for fiscal "generosity."
The outlook for the United States shows divergent scenarios: a best-case scenario suggests a slower deterioration, while missteps could accelerate the problem.
In the U.S., two economists interviewed projected that under a "best-case scenario," the debt-to-GDP ratio, currently around 100%, might rise more slowly over the next decade, reaching approximately 120%. A third economist foresaw little change.
"Productivity is like magic; it would significantly improve fiscal dynamics," said Idanna Appio, a portfolio manager at First Eagle Investment Management. However, she emphasized, "Our fiscal problems are way beyond what productivity can fix."
Kevin Khang, Head of Global Economic Research at Vanguard, identified demographics as the root cause of debt issues, stating that the "root" of the problem lies in aging populations and the associated welfare promises. He suggested AI "just buys us time."
In his modeling, higher growth and tax revenues could slow the pace of U.S. debt accumulation, resulting in a debt ratio of about 120% of output by the late 2030s. However, if AI proves disappointing, growth slows, and market pressures increase borrowing costs, the debt ratio could climb to around 180%.
The OECD highlights employment, wage transmission, and government expenditure management as core variables.
An OECD economist emphasized that AI's impact on debt trajectories depends on several key conditions being met simultaneously: job losses from automation must be offset by subsequent job creation; increased corporate profits must be passed on to workers through higher wages; and governments must manage to control overall spending.
Scenario analysis provided shows that even if AI-driven productivity gains lead to an improvement—lowering the debt ratio by "10 percentage points"—debt would still be significantly higher than current levels. This implies AI acts more as a tool to "buy time" rather than an automatic fix for fiscal sustainability.
Countervailing forces exist on both tax and spending fronts, with interest rates and recession risks remaining crucial.
On the revenue side, while a boost in economy-wide productivity should theoretically expand the tax base, experts caution that if AI reduces employment or weakens competition, concentrating gains more on profits and capital—which typically bear a lower tax burden than labor—the improvement in government revenues might fall short of expectations.
On the spending side, efficiency gains in the public sector could reduce costs, but there is also a risk of expenditures rising alongside growth. Kent Smetters, Director of the Penn Wharton Budget Model at the University of Pennsylvania, projected that AI's impact on U.S. debt over the next decade would likely be "very small."
He noted that even growth exceeding current expectations would offer limited help in curbing social security spending, as benefit claims are linked to average wages. Furthermore, if productivity boosts wages in the private sector, other government-covered labor costs might also rise. The OECD economist also stressed the need to observe whether wages will actually increase, suggesting that if AI does not boost employment, wages are more likely to rise.
Additionally, economists believe debt costs will also depend on whether productivity gains push up real interest rates, a discussion already underway within the U.S. Federal Reserve. Christian Keller, Global Head of Economic Research at Barclays, warned of the possibility of a recession, stating, "The AI boom may not come fast enough."
Comments