It’s no surprise that optimism reigns on Wall Street. Both the economy and financial markets performed surprisingly well last year in the face of numerous obstacles:
- A historic increase in tariffs
- A shutdown in immigration, which accounted for nearly 90% of labor force growth over the previous five years
- A challenge to Fed independence
- A continued rise in our already historic level of public debt
- Geopolitical crises and the upending of traditional U.S. alliances
The optimism reflects more than just resilience:
Companies are investing hundreds of billions of dollars in Artificial Intelligence. Most projections suggest that the AI boom is far from over: Some $2.5 trillion in additional global spending is expected in 2026.
The President has backed off on a significant number of threatened tariffs since “Liberation Day” in April, and his ability to impose tariffs at will has been constrained by the recent Supreme Court ruling. Meanwhile, firms have mitigated their effects through inventory management (front loading imports ahead of the tariffs), transshipments (shipping through lower-tariff countries instead of directly to the U.S.), and importing directly from countries with lower tariffs.
The Big Beautiful Bill contains a significant amount of fiscal stimulus that could add as much as 1% to GDP. It includes:
- A one-off surge in tax refunds early in the year due to reduced tax liability in 2026
- Increases in the standard income-tax deduction and childcare tax credit
- Elimination of taxes on tips and overtime
- A bigger State and Local Income Tax (SALT) deduction
- 100% expensing for purchases of capital equipment (which will help maintain strength in AI spending)
- A higher limit on business interest deductions
- Corporate profits continue to increase at a solid pace; the latest figures show a 10% increase from a year earlier
Consumer spending has held up impressively despite a significant slowdown in employment that has generated weak income growth. That is due partly to lower energy prices and a substantial increase in household wealth.
But there are signs of trouble ahead:
AI spending accounted for over a third of GDP growth last year. No one knows how long AI spending will continue to increase, but it needs to if it is to contribute to GDP growth; if AI spending merely continues at its current (very high) level, it will not contribute to economic growth.
- Federal debt held by the public is at a record $30 trillion, up from a little over $5 trillion before the financial crisis. That amounts to more than 95% of GDP, compared with 35% previously. The Big Beautiful Bill is estimated to add some $4 trillion to the national debt over the next decade.
- Job creation has slowed dramatically. While January employment surprised on the high side, that was due primarily to a new model for estimating new business formation (the net number of new firms created minus those that failed). For all of 2025, job growth averaged only 15,000 per month, down from 122,000 in 2024. Much of this decline is due to greater restrictions on immigration that began to be implemented in June 2024.
- While economists and markets have been encouraged that consumer spending has held up in the face of weak job and income growth, that is due primarily to a continued drop in the savings rate, which has fallen from over 6% in early 2024 to 3.5% in December. Since the savings rate rarely falls below that level, the extent to which consumers can continue to eat into their savings to maintain consumption is limited, although it’s possible that the labor market will begin to improve.
But the biggest risk going forward is probably a correction in asset prices. Despite the surge in government debt and the passage of the BBB that will add to it, bond yields fell last year (i.e., bond prices rose). The price of both gold and silver has surged to levels that are well above what fundamentals would suggest and reflects at least in part concern around the value of paper assets like bonds and stocks. The stock market is at record levels and considered overvalued by every metric.
The AI boom looks very much like a classic asset bubble. AI is a very promising technology that is likely to change how companies and individuals across the economy operate, and no one wants to be left behind. The question has become whether companies can afford not to invest in AI, rather than whether they can afford to invest. As a result, a massive amount of debt is financing hundreds of billions of AI investment, and valuations of AI companies have soared, even for those with weak revenue and questionable business models.
What is happening is eerily similar to the experience of the late 90s during the dot-com bubble. The internet proved to be the game changer many expected, but this didn’t prevent stocks from becoming overvalued, and their subsequent collapse caused a recession. This all occurred despite clear indications of promising returns on investment through sales of software, mobile phones, computers, and other electronics, as well as a new gateway—the internet—to sell and advertise all sorts of products much more efficiently. AI also promises to be a game changer, but it’s currently cheap to use, and it’s far from clear that companies will be able to generate the revenue necessary to justify their huge investments. There are also questions regarding the intellectual property around AI, which could undermine the value of AI-created software and content.
The overall stock market continues to reach all-time highs and remains vulnerable to a significant correction or worse. The S&P rose 17% last year, after increasing 20-25% in each of the previous two years. Almost all stock market valuation metrics imply overvaluation:
- The so-called Buffet indicator—market cap as a percent of GDP—is well over 200%, an all-time high. For context, it peaked at approximately 150% during the dot-com boom
- The S&P P/E ratio is around 23% of forward earnings estimates, which is about 40% higher than its 20-year average
- The Schiller cyclically adjusted P/E is considered by many economists and analysts to be a superior valuation metric since the stock market is highly correlated with the business cycle, and this measure adjusts for that. Its current value indicates that stocks have been more expensive only once in the past 100 years: immediately before the dot-com bubble burst
A Valuation Signal Investors Can’t Ignore
S&P 500 Prices Are Decoupling From Earnings
It’s worth noting that while the bursting of asset bubbles associated with new technologies produces short-run economic damage, they have always been helpful in the long run. Lower prices enable more companies to adopt the new technology, thereby generating productivity gains that boost economic growth, income, and spending.
Job losses are a concern with every technological breakthrough, and many job losses do occur. While overall unemployment may rise in the short term, the number of jobs created by new technology has always exceeded the number of jobs lost. Investment in AI infrastructure is already creating jobs. Productivity gains from the new technology will also increase income, providing the wherewithal for spending on other things. In fact, technological advances have been a major driver of long-run job growth. Some 15 million workers are now employed in computer and internet-related jobs that didn’t exist 25-30 years ago. Consider Amazon; the consensus was that it would kill retail and associated jobs. Many retail outlets failed, but the net effect has been an increase in employment. Jobs in warehouses replaced those in retail stores, and there were massive additional increases in employment in the delivery industry.
While history has shown technological innovations to have overwhelmingly positive long-term impacts on the economy, it is possible that this time is different. AI technology seeks to emulate aspects of the human mind, such as learning, reasoning, and perception, and will inevitably replace many tasks currently performed by people. That said, it will also create new jobs that do not currently exist, although it is not certain that the net effect on employment will be positive.
What does all this mean for monetary policy going forward? The Fed cut rates three times late last year despite the surge in asset prices and a rebound in GDP growth in the second half of 2025. This reflected a view that the risks to employment were greater than that of rising inflation. Core CPI inflation ran at an annual rate of 2%—the Fed’s target—last fall, providing justification for that opinion. But in the three months prior to that, inflation ran close to 4% annually, and the pattern has been for inflation figures to alternate between high and low levels every few months. It appears that we are heading into a few months of higher inflation numbers due to several technical factors:
- Inflation was artificially depressed late last year by the government shutdown that caused the BLS to survey prices much later than usual in November, placing extra weight on the post-Thanksgiving sales period
- The BLS also penciled in a zero increase in housing costs, which account for around a third of core inflation, due to a lack of staff to collect survey data
- Finally, many firms that delayed price increases due to tariffs are expected to enact them early in the year
High CAPE Yields, Lower 10-Year Excess Gains
If economic growth remains solid as expected and monthly inflation readings are on the high side, the Fed will likely remain on hold for at least the next few months. The current policy rate of 3.5-3.75% is considered neutral by most economists; that is, neither restrictive nor stimulative for economic activity. The next rate-cutting cycle may not begin until after a sharp stock market correction and/or a significant weakening in economic activity, perhaps even a recession.
