What if the Bond Market Slows AI Even without the Fed?
In a recent newsletter, Greg Ip from the Wall Street Journal suggests that it’s a mistake to think AI is unaffected by monetary policy changes. The Federal Reserve doesn’t have to directly make AI less profitable to impact the investment boom in that sector. Instead, they could simply tighten financial conditions enough that financing this boom becomes unfeasible.
Ip is certainly onto something regarding how sensitive the AI sector is to rising interest rates and stricter financial conditions. However, it’s important to note that these conditions don’t necessarily have to come from the Fed. Capital markets themselves can tighten financial conditions. In fact, long-term interest rates are already on the rise, and according to Federal Reserve Chairman Kevin Warsh, the increasing investments in AI are partly to blame.
The Return of the Price of Capital
During a press conference on September 16, Warsh was asked about the climbing long-term Treasury yields. He attributed this rise to both solid economic growth and competition for capital, especially from tech giants investing in AI infrastructure.
“The so-called hyperscalers are out in the market raising funding,” he explained. “And so the competition for capital is real.”
Fed Governor Christopher Waller made a similar point last week at an event in New Delhi. He noted that increasing Treasury yields mainly reflect economic strength instead of rising inflation expectations. Waller pointed to AI investment as a key factor pushing borrowing costs higher, and he considered this a positive development: “I’d rather have higher growth and higher yields than low growth and low yields.”
This reflects how capital markets are supposed to function. When there are more profitable investment opportunities, demand for financing increases. Consequently, the price of capital rises, which encourages savings and discourages investments that no longer have sufficient returns to justify their costs.
For years, economists have been concerned that the so-called “natural rate” of interest had plummeted to unusually low levels, driven by an excess of savings and a lack of attractive investment choices, particularly in China. However, the emergence of AI might be shifting that dynamic.
An Emerging Consensus at the Fed
The notion that the natural interest rate could be on the rise is gaining traction among Federal Reserve officials.
In February, Fed Governor Lisa Cook pointed out that investments in AI-related data centers and semiconductors were surging, even with higher interest rates compared to the previous two decades.
“With investment contributing to strong aggregate demand, it is possible that the current neutral rate is higher than before the pandemic,” she noted in a speech to the National Association for Business Economics.
Fed Vice Chair Philip Jefferson expanded on this idea in a July address at Stanford University, arguing that accelerated productivity growth could boost expected investment returns and capital demand. Households anticipating higher incomes might also save less, potentially moving the natural interest rate upward.
Even though Jefferson and Cook are still Fed officials and their language reflects this, they acknowledge the complexities surrounding this discussion about a rising natural rate. But the consensus among them about the forces in play is noteworthy. Warsh, Waller, Jefferson, and Cook are all touching upon different facets of the same overarching change: AI is affecting the demand for capital, and in turn, interest rates are reacting accordingly.
Warsh’s Monetary Revolution
This notion aligns well with Warsh’s broader perspective on monetary policy.
At Jackson Hole in August, he described the AI revolution as a pivotal moment after years of concerns about stagnant growth and a global glut of savings. He pointed out that financial markets had become overly reliant on Fed signals.
His concern was that investors often focused too much on central bank predictions and too little on fundamental economic indicators, creating what he termed a “hall-of-mirrors problem,” where both markets and the Fed influenced each other.
Warsh is advocating for markets to generate independent signals about economic conditions. Rising long-term yields fueled by robust growth and competition for capital would be just such a signal.
This carries an important implication for monetary policy. If the AI boom has indeed raised the natural interest rate, the Fed might find it necessary to set a higher policy rate to achieve comparable levels of restraint. Yet, if the market rates are adjusting upwards on their own, the Fed might not need to increase its overnight rate as much.
Higher bond yields are already making mortgages pricier and pressuring other interest-sensitive investments. Some resources allocated to the AI boom can be freed up through this process without additional input from the Fed.
However, it’s crucial to note that the rise in yields isn’t solely a result of AI financing. There’s still a hall-of-mirrors effect; investors are demanding higher yields now due to expectations of a higher Fed policy rate. Moreover, the strength of the AI boom coupled with anticipated higher rates probably means that any increase in the Fed’s policy rate won’t severely impact employment or growth. As long as inflation is elevated and employment risks are low, it seems reasonable for investors to expect higher policy rates to persist.
Interestingly, this scenario doesn’t necessarily clash with Warsh’s goals. Expectations for higher future Fed rates shouldn’t be seen as distorting market signals. If investors foresee the Fed maintaining elevated rates due to a risen natural rate, they’re acting on economic fundamentals. This is different from a situation where the Fed artificially suppresses long-term rates through promises to keep rates lower than what fundamentals suggest. In this case, investors are anticipating the Fed’s appropriate response to economic conditions.
In summary, it seems that the bond market is already enacting Warsh’s vision for monetary policy. The market is re-establishing its role in determining the price of capital, and the Fed is recognizing the importance of allowing it to do so.
Ip is correct in pointing out that increased interest rates and tighter financial conditions could stifle the AI boom. Yet, these dynamics are already unfolding within the bond market, where the cost of capital is adapting to a technological revolution. The real question for the Fed is just how much additional tightening the economy truly requires.
