In the field of economics, there is a specific type of frustration that develops subtly. It’s not the kind of dramatic crash, but rather the slow, persistent feeling that something promised hasn’t materialized. Any modern office will have screens all over the place, real-time cloud dashboard refreshes, and AI tools that summarize documents that used to take hours to read. However, the picture appears strangely flat when you look at the actual productivity figures. Technology advanced more quickly. In general, the output didn’t.
The first quarter of this year saw a significant decline in productivity, or the amount of value that employees produce each hour. An already faltering argument appeared to be put on hold by the government’s report. The pandemic-era spike in digital investment was seen by the optimists as the pivotal moment when years of cloud adoption and AI research finally appeared in the data. Whether that moment is delayed or just not happening as everyone had anticipated is still up in the air.
| Topic Overview: The Productivity Puzzle | |
|---|---|
| Concept | Productivity — value of goods/services produced per hour of work |
| Current Annual Growth Rate | ~1% (post-pandemic average, in line with rate since 2010) |
| Peak Productivity Era | 1996–2004, averaging over 3% annual growth |
| Key Technologies Under Scrutiny | Artificial Intelligence, Cloud Computing, Advanced Automation |
| Leading Skeptic | Robert J. Gordon, Northwestern University — calls AI “impressive but not transformational” |
| Leading Optimist | Erik Brynjolfsson, Stanford’s Digital Economy Lab — believes transformation is underway |
| Forecast (McKinsey) | Global productivity expected to nearly halve, from 3.8%/year (1950–2014) to ~2.1% (2014–2064) |
| Key Debate Resolution Date | End of 2029 — a formal “long bet” between Gordon and Brynjolfsson |
| Primary Adoption Gap | Three-quarters of U.S. businesses have fewer than 10 employees; advanced AI adoption remains limited |
| Structural Challenge | Aging populations, declining fertility rates, shrinking workforce |
There are actual names associated with the debate. For years, Northwestern’s Robert Gordon has maintained that modern AI is essentially a very advanced pattern-matching tool, impressive at what it does, but not the kind of fundamental change that electricity or the combustion engine represented. Stanford’s Erik Brynjolfsson retorts that a “tidal wave of transformation” is already underway. The outcome of the formal wager between the two economists will be determined at the end of 2029. Economists placing a personal stake on a macroeconomic prediction is uncommon. That in and of itself demonstrates how genuinely unresolved this issue is.
The discrepancy between the statistical reality and the boardroom narrative is difficult to ignore. Businesses have invested billions in AI platforms and cloud infrastructure. Every productivity-adjacent startup has been relentlessly pursued by venture capital.
However, since the pandemic, the economy has grown at a rate of about one percent per year, which is much slower than the nearly three-and-a-half percent average during the tech boom of the late 1990s. This is the same slow pace that has defined the economy since 2010. There is a problem with the translation.
Rather than being technological, a portion of the explanation might be structural. Less than ten employees work for 75% of American businesses. These include independent stores, bakeries, accounting firms, small clinics, and other establishments where using enterprise AI tools isn’t feasible, cost-effective, or even particularly pertinent. Large, well-funded, and forward-thinking businesses are typically the ones that truly profit from cutting-edge digital technology. It appears that the distance between the leaders and everyone else is growing rather than shrinking.
Additionally, there are demographic factors that subtly undermine the entire picture. Although longer life expectancy is generally positive, there are drawbacks. The ratio that once naturally fueled growth is changing: more older people are receiving pensions, and fewer young people are joining the workforce. Global productivity rates could almost cut in half over the next few decades, according to a McKinsey analysis. If that is true, there will be a tremendous burden on technological advancement. It can’t be incremental only. It must make up for a structurally contracting workforce.
Some businesses are doing it correctly. Because technological advantage is the only sustainable advantage in a high-cost economy, companies like Elmos Semiconductor in Germany invest about 15% of their sales back into product development, not because it looks good in an annual report. According to OECD data, a small number of truly innovative businesses are outperforming the industry as a whole. There is more to the insight than just R&D expenditures. It’s about integration, or whether a business can truly reorganize itself around new tools, retrain its employees, and alter decision-making processes. When technology is introduced into an organization that hasn’t changed, it usually sits there.
Observing all of this suggests that the productivity conundrum isn’t really about whether the technology is functional. In the right situation and with the right people, it most likely does. The more difficult question is whether the prerequisites for its spread—across industries, firm sizes, and income levels—actually exist. In the words of Laura Tyson, who served as Clinton’s chair of the Council of Economic Advisers, “even if the digital optimism proves correct, that doesn’t guarantee the benefits will be widely shared.” Simply put, a smaller economy is less productive. Additionally, smaller economies have less space to address the most pressing issues.