The US labour force’s share of national income has fallen to 53.8%, the lowest since records began, and those hoping that strong, accelerating productivity growth might improve workers’ lot are likely to be disappointed.
The benefits of cost savings and efficiency gains generated by any incipient AI boom are more likely to flow to companies and shareholders, with workers’ share of the country’s economic pie shrinking further.
That has potential knock-on consequences for economic activity, should weaker household income growth undermine consumer spending, which accunts for about 70% of US GDP.
That may not be felt immediately, though. Household balance sheets are still strong as rising asset prices have increased people’s overall net worth. That has helped sustain consumption even as company hiring has slowed and the unemployment rate has risen to a four-year high.
But consumers are drawing down savings to fund their spending. The personal savings rate is now 3.5%, the lowest in more than three years.
Further social tension
While the wealthiest, asset-rich Americans may account for the bulk of household spending, they aren’t responsible for all of it — and those consumers not in the top income decile may increasingly find their ability to spend curtailed.
More worrying, perhaps, is the political and societal impact of a continued decline in workers’ share of GDP. The political climate across the country is already febrile and social tensions are taut.
Affordability issues dominated mayoral and gubernatorial elections last year and will be pivotal to the mid-term elections in November too — and probably well beyond that.
“However ‘good’ the macro data looks, large parts of the population may not be feeling the benefits. That’s a major political challenge, one that’s likely to influence US election cycles for years to come,” Jim Reid, head of global macro research at Deutsche Bank, saidf in a note to clients.
Accelerating tide
These trends are not new, of course. Workers’ share of US national income has been steadily shrinking for years, a shift that accelerated notably around the turn of the millennium.
Analysts at the Economic Policy Institute say the tide really started to turn against labour and in favour of capital some 45 years ago, when the “productivity-pay gap” started to widen. Between 1980 and 2025, productivity rose more than 90% while hourly pay rose only 33%, meaning productivity grew 2.7 times faster than wages. That gap is now a chasm.
AI can either substitute or complement, depending on adoption patterns. Without supportive institutions, much of the productivity windfall could accrue to capital owners, reinforcing the trend of a declining labour share.
— Tiffany Wilding, MD and economist at Pimco
There are three key structural forces that help explain the shift: weaker employee bargaining power, technological changes, and globalisation. The last of those is losing steam, but not enough to reverse labour’s dwindling share of GDP.
There’s also little indication that trade union membership is about to rebound, especially with the labour market so fragmented. Physical offices and factories are on the decline, millions of people work remotely, and part-time, freelance, and contract work are on the rise.
Could an AI-driven productivity boom stop the rot?
In theory, yes. Rising productivity generates more income for companies, which can be funnelled into higher wages for employees. Tiffany Wilding, MD and economist at Pimco, says AI could lower prices in sectors such healthcare and business services, boosting real incomes even if nominal wages don’t surge.
But there’s another scenario, where tax incentives, government trade policy and the AI arms race push capital-intensive firms into investing more on technologies that reduce headcount and labour costs.
We may be seeing that play out right now: Amazon and other big firms are firing tens of thousands of workers. While it’s unclear how much of that is related to AI, it’s occurring as just four “megacap” tech firms, including Amazon, have committed to spend about $650bn on AI this year alone.
“AI can either substitute or complement, depending on adoption patterns. Without supportive institutions, much of the productivity windfall could accrue to capital owners, reinforcing the trend of a declining labour share,” Wilding says.
If that occurs, who will consume all the goods and services the AI-driven economy is set to create? That could be the trillion-dollar question.
Would you like to comment on this article?
Sign up (it's quick and free) or sign in now.
Please read our Comment Policy before commenting.