Last week I noted that the impact of Federal Reserve policy, under new leadership, is likely to be the single most important driver of economic and stock market performance in 2026. Tell me the yield on the 10-year Treasury at the end of next year and I could likely draw a pretty accurate picture of economic and stock market performance over the same period.
But that doesn’t mean that interest rates are the only determinant of equity performance. Earnings are equally important. Recent economic data, perhaps distorted still from the government shutdown, suggests the economy is on a fairly steady path of modest-to-moderate growth and inflation still stubbornly above the Fed’s stated 2% target. While optimists believe that the upward pressure of tariffs on inflation will fade as 2026 progresses, many of the key factors keeping inflation stubbornly high are only minimally impacted by tariffs. These would include utility related costs, insurance premiums, auto repair services, the price of beef, and the cost of medical services. Slack demand for housing and an excess supply of oil worldwide are depressing inflation. But here again, tariffs aren’t a significant factor.
Trying to tie tariffs to economic activity, in rough numbers, given that imports are roughly 15% of GDP and applying effective tariffs rates, one can argue that tariffs could have as much as a one percentage point impact on GDP. But importers and exporters have been eating much of the cost. The best guess therefore is that tariffs may increase inflation of something around half a percentage point. That impact should fade once the full impact is felt, probably within the first half of 2026.
One might think that tariffs would squeeze profit margins. But there is no evidence that is happening as profit growth continues to exceed GDP growth by a healthy margin. What has happened is that companies have offset the impact of tariffs by reducing headcount. In economic terms that means increases in worker productivity, the output per manhour, is offsetting tariff impact.
There are some AI optimists who will quickly say that AI is the driver of the productivity increase. While it is possible that early AI adaptation has had some impact, most will say, and I believe rightfully so, that companies are simply finding that the cost to retain workers, given that the costs to hire them back or replace them is high, is a cost not to be borne in a period of slackening growth and increased pricing pressures. While the Trump administration wants us to believe that tariffs and other actions will lead to substantial reshoring of manufacturing activity, so far that hasn’t happened. Using data from the most recent earnings report, manufacturing employment so far this year is down by over 50,000 jobs. The industries that should be most advantaged by tariffs as high as 50%, notably steel manufacturing and aluminum processing, have seen no increase in hiring. On the other hand, construction employment is declining as contactors are having a difficult time replacing immigrant labor causing delays. This is despite the sharp rise in expenditures related to data center construction, now running more than $3 billion per month, triple the pace of just a few years ago.
Look at the graph below showing productivity all the way back to 1947.
One notices several points. First, while the graph appears jagged, any trend line would be pretty flat across more than 75 years. The gray areas are recession periods. You will see that productivity rises at the end of recessions because activity recovers faster than hiring. But those gains prove fleeting. For the last 25 years, productivity gains have fluctuated between zero and 5% with an average of about 2%.
One of the mantras AI enthusiasts voice vociferously is that AI is going to make our economy significantly more productive over an extended period of time. As with so many disruptive technologies, computers will displace workers while making those employed persistently more productive. On the surface that makes obvious sense. Lawyers will be able to produce documents more quickly. Drive-thru windows and call centers will become automated. Driverless autos and intelligent robots make the Jetsons come to life.
But wait a minute. Let’s think back over the past 75 years. Jets replaced propeller aircraft cutting travel time significantly. The 1950s and 60s brought us mainframe computers, clearly both disruptive and engines for better productivity. The 80s gave us personal computers bringing new power to the desktop. Goodbye typewriters and lots of paper-based transactions. At the start of this century, smartphones emerged giving us instant communication everywhere not to mention emails, texts, and navigation aids. The Internet blossomed around the same time instantly making information gathering easier and more efficient. Add in networking, cloud computing and now AI. Yet, do you see any inflection in the pace of productivity? Robots brought down the cost to manufacture. The Internet allowed us to order on line and have goods shipped to us. Streaming supplanted movie theatres. But how much did any of the above make us more productive? As consumers, we like the ease of having goods shipped to us, but is the process of getting those goods to your front door improving productivity? Uber and Lyft have supplanted taxis but the number of Uber and Lyft vehicles in New York City collectively is now much greater than the number of yellow cabs a decade ago. Streaming? Exactly one company so far has earned its cost of capital.
Look around you. The process of making a meal, whether at home or away, is largely the same as it was decades ago. Thanks to Amazon, etc. there are more trucks on the road. But has the process of getting goods from the retailer to the consumer become more productive? Homes are essentially built as they were decades ago.
That isn’t to say that there aren’t notable examples of greater productivity. Farming is one. Farms today are much larger as small farms prove economically less efficient. Technology has vastly increased the speed and accuracy of preparing tax returns. But, again, look at the chart. Technology clearly has changed our lives. In many ways it has made us smarter. It is easy to argue that the 2% average annual increase in productivity is technology-based. It is also clear that technology has lowered costs, in some cases significantly. But has technology moved the productivity needle?
Why is that important? Because one way to look at GDP growth is to multiply the changes in productivity by the growth rate of the labor force. If productivity advances at a variable rate averaging a bit over 2%, and population increases based on birth rates and net immigration by a bit over 0.5%, then one can compute a steady state growth rate of close to 3%. Over the near term, there is a fly in the ointment. Birth rates are down, the size of the labor force is growing at a slower rate than the past courtesy of aging demographics, and net immigration alone will cause a reduction of the work force of close to half a percentage point. In other words, if growth is going to accelerate as AI enthusiasts hope, given trends in the growth in the labor force, productivity is going to have to increase despite the lesson of the 75-year chart on productivity.
Technology advances are deflationary. They reduce costs. But productivity measures aren’t based on costs; they are based on sustainable increases in output per worker. In some industries, there will be major improvements. But I believe that for a decade, and perhaps decades to come, lawns will be mowed, goods will be delivered, homes will be built, dishes will be washed, and trains will roll mostly as they are today. I am not reaching any conclusion relative to the promised changes to be delivered by AI. But we study history as a guide to the future. That one chart suggests the burden of proof lays with those who believe we are at the cusp of a golden era of growth where despite aging demographics, lower birth rates and restrictions on immigration, technology this time will open a new path that will shift the trends of the past 75 years.
One final note. For a couple of years we have used the term Magnificent 7 to describe a group of stocks that have been market leaders and outstanding performers. In 2025 to date, Alphabet#, up over 64%, and Nvidia# up 32%, have lived up to their billing. But the other five (you know the names by now) are up on average 11%. By comparison, the equal weight S&P 500 is up just over 10%, the Dow Industrials are up over 13% and the overall S&P 500 is up 16%. Maybe we should use the term the Magnificent Two, at least as far as 2025 is concerned.
Today, Senator Ted Cruz is 55. Actor Ralph Fiennes is 63 while Phillie legend Steve Carlton turns 81.
James M. Meyer, CFA 610-260-2220