As economists often remind us, productivity is not everything, but in the long run, it is almost everything. That sentiment, famously captured by Nobel laureate Paul Krugman, sits at the heart of how modern economies understand growth, prosperity, and the sustained ability to produce more output with the exact same amount of inputs. Economists measure this through "total factor productivity" (TFP), which quantifies the direct economic contribution of technological innovation.
Because the ongoing process of producing more with less is how human living standards steadily improve, TFP is arguably the most critical metric in macroeconomic analysis. As Krugman explains, a country’s ability to elevate its standard of living over time depends almost entirely on its capacity to raise output per worker. Technology is the engine that drives that capability, and TFP is the gauge that measures it.
To grasp just how monumental technology-generated productivity can be, consider a recent research paper published by the National Bureau of Economic Research (NBER). The study estimates that an additional 0.5% of annual TFP growth would be enough to completely stabilize the United States government’s finances, locking in the debt-to-GDP ratio right at today’s levels.
While 0.5% sounds modest on paper, the cumulative impact is staggering. If sustained over the next decade, the NBER estimates that a mere half-percent boost in TFP would slash the baseline forecast for US government debt by $2 trillion. Look out over a 30-year horizon, and that same sustained 0.5% increase would drive the federal government’s debt-to-GDP ratio down to a level 42 percentage points lower than the NBER’s baseline forecast—and an impressive 80 percentage points lower than its pessimistic scenario.
Given the seemingly hopeless state of modern federal finances, maintaining today’s level of indebtedness feels like a dream scenario, an outcome widely dismissed as too good to be true. Yet, new research from artificial intelligence pioneer Anthropic suggests that the economy might actually be capable of outperforming even those optimistic projections.
Anthropic recently conducted a comprehensive study examining 100,000 real-world conversations on Claude.ai. The explicit goal was to estimate how long specific tasks within those interactions would take with and without AI assistance, allowing researchers to study the broader productivity implications across the entire economy. Their conclusion was striking: large language models (LLMs) could potentially raise total factor productivity by a full 1.1 percentage points.
If a 0.5% increase is enough to stabilize US government finances for decades, the implications of a 1.1% boost are profound. Such an economic tailwind would likely resolve almost every major structural fiscal pressure currently facing the nation.
Naturally, there are plenty of reasons for economists and forecasters to view such optimistic projections with a healthy dose of skepticism. The Anthropic study highlights, for instance, that Claude can save teachers roughly four hours of labor by generating comprehensive curricula in a mere 11 minutes. However, translating those direct time savings into higher economic output requires complex macroeconomic modeling—a process inherently fraught with best-guess assumptions and the risk of false precision.
Even if Anthropic’s calculations regarding time savings are entirely accurate, the company’s model could miscalculate the ultimate productivity payoff. It is entirely possible that society will choose to channel all the time saved by artificial intelligence into pursuits that are economically unproductive, such as watching more short-form videos or reading newsletters. In that scenario, AI would undoubtedly enhance human welfare by providing more leisure and free time, but it would fail to expand aggregate wealth or economic output. While that outcome would be a net positive for everyday people, it would offer little relief to lawmakers and central planners hoping for a silver bullet to solve the national debt crisis.
Conversely, there are equally compelling arguments suggesting that Anthropic’s economic model might actually be overly pessimistic. As the study’s authors readily admit, their methodology does not account for the organic rate of technology adoption, nor does it factor in the exponentially larger productivity gains that will inevitably emerge from vastly more capable future AI systems.
In other words, the study’s baseline model operates on the assumption that society will continue utilizing artificial intelligence exactly as it does today, relying on today’s generation of language models without any systemic improvements for an entire decade. Because language models are improving visibly every few months—and humanity is only just beginning to learn how to effectively wield them—Anthropic acknowledges that its estimate likely represents an approximate lower bound on the true productivity effects of AI.
If a 1.1% boost is merely the lower bound of AI-induced productivity, the macroeconomic horizon shifts dramatically. Under that scenario, the nation might successfully pay down its sovereign debt while simultaneously securing more leisure time for entertainment. And that remarkable projection only accounts for AI’s impact on non-physical, digital labor—leaving out the eventual arrival and integration of humanoid and industrial robotics into the physical economy.
Dismissing this level of optimism entirely requires believing that the trillions of dollars corporations currently plan to pour into artificial intelligence capital expenditures, research, and development will be completely squandered. While that is certainly a possibility—major technological revolutions do not always arrive smoothly or on schedule—the most compelling reason for optimism lies elsewhere.
Anthropic’s 1.1% estimate is derived solely from AI making existing tasks faster to complete; the model does not attempt to account for AI’s capacity to completely reinvent how those tasks are executed. Historically, as Anthropic notes, the most transformative productivity leaps—from electrification and computing to the rise of the internet—did not stem merely from speeding up legacy tasks. Instead, they originated from the fundamental reorganization of production itself. While there is currently no reliable way to model these emergent workflows, history suggests their ultimate economic impact will far exceed what can be captured in preliminary measurements.
Researchers at Anthropic remain appropriately cautious, carefully enumerating the methodological limitations and underlying assumptions of their findings. Furthermore, even if all those assumptions prove correct and an AI-driven productivity boom successfully cures the US government’s debt troubles, fiscal history suggests lawmakers might simply spend their way right back into trouble.
Yet, given the pervasive, near-100% certainty that economists and investors place on an inevitable fiscal disaster, even a small probability that Anthropic’s estimates are correct provides a powerful reason to update long-held priors. The financial trajectory of the US government may not be as intractable as conventional wisdom suggests, and the long-term outlook for the US dollar may not be as bleak as feared. In the long run, productivity remains the ultimate economic force, and artificial intelligence may be standing on the precipice of making the global economy significantly more productive.
Comments (0)