Before Donald Trump’s trade war gave it the Dusko Markovic shove, AI was the only macro-market story that really mattered. (Ah, the good old days.)
To be sure, the market’s still highly (hyper) sensitive to developments on the AI front. Indeed, the Trump administration scapegoated DeepSeek (and, implausibly, Joe Biden) for US stocks’ poor performance in 2025 right up until the violent reaction to “Liberation Day” made it impossible to deny the link between the administration’s policies and market turmoil.
To state the obvious, AI still matters tremendously, and there’s more than a little overlap between Jensen Huang’s “new industrial revolution” and the trade war.
I bring this up Friday because under-the-radar data released this week showed US labor productivity declined in Q1 for the first time in nearly three years. As I noted while documenting the release, the BLS’s productivity and unit labor costs report is important even as it virtually never makes headlines.
Simply put, the key to robust growth without uncomfortably high inflation is steady productivity gains, and according to many, AI adoption is the surest way to enhance productivity.
The figure above takes the four-quarter rolling average of the quarterly productivity prints. As you can see, it’s trending lower.
It’s important that the US reverses that trend, and as BofA’s Michael Hartnett noted, the AI revolution and the read-across for productivity “remains the one secular trend to buttress” stock valuations. If AI’s promise is realized, it’ll be “not a moment too soon,” he added, referencing the quarterly productivity decline mentioned above.
And yet, some still see a Devil’s bargain: Productivity gains in exchange for your jobs. Joblessness isn’t politically popular, which means “big AI adoption with productivity-enhancing unemployment” may ultimately force politicians “to protect US workers via wealth taxation,” Hartnett went on, noting that in the absence of AI-related layoffs (as robots render humans redundant), hundreds of billions spent on AI development just “means lower margins.”
It’s also possible, of course, that “AI = UBI = YCC,” to quote Hartnett one more time. In that scenario, displaced workers would be given a monthly government stipend funded by long-end issuance, yields on which would be capped by Fed-buying.
“AI = UBI = YCC . . .”
Lions and tigers and bears, oh my!
Didn’t you get the memo? Workers displaced by AI will happily assume new careers in the manufacturing arts and agricultural pastimes.
…or, given that Americans seem fine with letting some people starve and go homeless as opposed to taxing the wealthy, how about in future enhanced policing capabilities enabled by AI combined with the death of truth via deepfakes and targeted propaganda instead gradually disenfranchise everyone except the wealthy and political elite, forever, as human capital becomes obsolete and coordinated political opposition all but impossible, and robots can certainly control protesters with small arms in case people should riot.
The standard of living for many in US will go down and become closer to the rest of the world, particularly if others stop buying our debt. Alas, many Americans won’t be able to drive their big Chevy Suburbans to Costco for an amazing weekend of binge-shopping, followed by binging on fast food. The parking lot of the regional Costco is jammed on the weekends. Instead of raising income taxes which disincentivizes work, the most fair approach would be a flat wealth tax, which I realize will never be allowed to happen.
It seems that if LLMs offer a path to AGI and eventually lead to significant redundancies while policy is not altered to be significantly more redistributive (specifically wealth tax) then that seems to be a recipe for entrenched inequality, but policymakers do not generally have a great record of addressing inequality even though massive inequality already exists. I don’t know that current analysis of which jobs are exposed are terribly helpful when AI seems capable of entirely new things yearly. Simultaneously, AI offers great tools for keeping the masses distracted and under control and current trends seem to favor distracting people while maintaining inequality. That’s why I’m pessimistic. Doesn’t necessarily mean the newly redundant will be worse off in terms of purchasing power if productivity increases significantly, just the end of social mobility. It’s weird that this is a real scenario that needs to be considered now when it would have been sci-fi five years ago.
Productivity has long been a hobby horse for my riding pleasure. Look at the definition. Productivity is more output per hour worked (or more output for each unit of inputs used). The primary definition cannot be achieved without disadvantage to the the labor force. The last 50 years have seen labor force participation, especially among prime age men, decline significantly to 62%, while productivity has risen significantly. A primary factor driving these trends is the decline in low skill jobs, replaced by technology. Men’s participation in post-secondary education and training has fallen to less than 50%. As productivity has increased through advances in tech, the proportion of folks able to benefit from these trends has declined. Furthermore, we don’t pay those who produce that rising productivity for doing so. The decline in labor force participation down to 62% means that seven million people of prime working age are no longer in the workforce, nor are they seeking to return. Yet unemployment is nearly at the lower edge of practicality. Why? Men with the currently critical skills just aren’t needed. They have been replaced by tech, robots, automated machines, computers, AI, etc. just as predicted. Even if some of these millions are retrained to have great skills, the fact is, rising productivity has erased the demand for them. Productivity is great for growth in GDP … sort of … but it is creating an increasing underclass that tariffs won’t eliminate. What we need right now is a much smarter president … much smarter … with much smarter advisors who can lead the way to a real solution. Productivity is going to kill us, steadily, inexorably, until it stops and the good old American Dream is well and truly dead.
Thanks, Mr. Lucky…yet another outstanding comment and contribution…
Yes indeed. The future can be bleak but as long as conservatives stay in power nationally and in red states this will not happen.
I would like to think there is snark hidden here, but I fear there is none.
H-Man, for younger people this is truly a difficult time. The signposts for the future are not encouraging,
I recently spent a long weekend in San Francisco. My main form of transportation was Waymo. As this writer states: within one ride, it not only seemed comfortable, but was preferable to Uber. With Uber, I never know if I will get the occasional dirty car; bad driver; or bad, but insistent conversationalist.
Waymo is absolutely the future.
https://www.wsj.com/tech/waymo-cars-self-driving-robotaxi-tesla-uber-0777f570?st=2TSWzc&reflink=article_copyURL_share