On Tuesday, the probability of a Federal Reserve interest rate hike in December rose to nearly 90%.

Written by: Yang Chen, Wall Street Insights

The Fed's "third in command," New York Fed President John Williams, recently stated that although the AI infrastructure boom and energy shocks constitute new inflation variables in the short term, the US economy is not overheating comprehensively. After one more rate hike this year, it is expected that no further hikes will be necessary until 2027.

On Tuesday, John Williams, who has permanent voting rights on the Federal Open Market Committee (FOMC) and is known as the Fed's "third in command," gave an in-depth interview to Bloomberg at the New York Fed headquarters.

Williams elaborated on the underlying logic supporting his support for the September rate hike and provided clear answers regarding the market's keenly watched interest rate path guidance, changes in the neutral interest rate (R-star), and the real impact of the AI infrastructure boom on macro inflation.

Before Williams spoke, the yield on the two-year U.S. Treasury bond rose by 2 basis points; after his speech, the yield declined, falling by 1 basis point to 4.792% as of this writing.

On the same day, Chicago Fed President Goolsbee expressed that price pressures have spread beyond data centers and directly related industries, stating, "I am increasingly worried that this is pure, old-fashioned demand overheating." Richmond Fed President Barkin pointed out that the Fed has been able to observe the impact of AI investment on U.S. inflation.

On Monday, Boston Fed President Collins stated that interest rates may need to be raised further before the end of this year, but specifics will depend on subsequent economic data. The key lies in observing whether inflation remains primarily concentrated in energy and AI infrastructure-related areas or will further broaden to other sectors. She also stated that the Fed's economic model has taken government debt levels into consideration.

On Tuesday, the probability of a Fed rate hike in December rose to near its highest point for this round, close to 90%; the probability of a rate hike in September also rebounded to around 25%.

One more rate hike this year, slight rise in neutral interest rate

Regarding the market's most concerned issue of the terminal interest rate, Williams provided a relatively clear roadmap during the interview. He reiterated his views on recent monetary policy but also clearly signaled a long-term "dovish" stance. Williams stated:

"I have not changed my view that one more rate hike might be appropriate. I see no reason to believe we will need to do more after that."

When asked about the rate hike path in the longer term (like in 2027), he asserted that further rate hikes would not be necessary then. However, he emphasized the "data dependency" of monetary policy:

"It's only mid-October. This economy is influenced by many different factors, and we just have to continue to monitor the data."

The neutral interest rate (the real interest rate level that neither stimulates nor suppresses the economy) that the market is closely watching was also central to this interview. As a co-author of classic models estimating neutral interest rates, Williams revealed that the neutral interest rate has risen moderately over the past few years by about 25 to 50 basis points due to fiscal policy and strong economic growth.

"I believe the long-term neutral interest rate remains fairly low, around 1% or slightly above that level. From this perspective, if the economy performs as I expect in the coming year, we are not far from the neutral interest rate."

AI infrastructure is pushing up short-term costs but will be a deflationary force in the long run

In this interview, Williams placed "AI" alongside tariffs and energy prices, viewing them as the three core variables currently affecting the path of U.S. inflation.

Williams pointed out that the trend of rising prices and costs associated with AI construction has already become apparent:

"We are indeed seeing rising goods and other costs associated with AI construction: data centers, huge demand for chips. Prices in AI-driven sub-industries have seen quite sudden and large increases. We are also seeing signs of some spillover into construction labor costs."

In response to market fears that overheating in AI investment might lead to widespread inflation, Williams provided a reassuring judgment. He believed that although the AI sub-industry is operating at full capacity, the U.S. has a massive economy exceeding $30 trillion, and monetary policy and tightening financial conditions are cooling other areas. Williams stated:

"This is an unusual situation; we have an economic sub-industry that seems to be running at full speed, and it will continue for a while. But I think what's really important is to look at the broader economy. I do not expect chip prices to decline. They may continue to rise. But we need to look not just at them, but at what is happening with overall inflation."

More critically, Williams is optimistic about the long-term positive effects of AI on the supply side. He believes that while demand is "definitely winning" in the short term, the price signals from extremely high profits will lead to capacity expansion, while the significant productivity improvements brought by AI "will actually become a deflationary force."

September rate hike logic: Energy shocks change short-term inflation outlook, economy is not overheating comprehensively

Williams firmly defended the Fed's decision to raise rates in September during the interview. He stated: "I strongly supported that decision. I believe it was the right decision—absolutely the right decision."

The core reason supporting his decision is the unexpected volatility in energy prices caused by recent geopolitical events and the widening refining margins. He noted that this has led to a significant increase in fuel costs as an "additional shock." Coupled with the extra demand driven by AI, it indicates that the economy can withstand slightly higher interest rates.

Despite this, he remains confident about inflation retreating to the 2% target. He illustrated the current health of the macro economy through several core indicators:

"This is not an overall overheating economy. This is not an economy that is increasing inflationary pressures beyond the specific shocks we are experiencing."

He pointed out that there are currently no signs indicating that imbalances in the labor market will bring widespread cost pressures, while the U.S. economy's potential output and GDP growth have both remained at a healthy level slightly above 2%, demonstrating strong resilience.

Disregarding the notion of "the market doing the Fed's work," his speech is not directed by Walsh and only reflects personal stance

Recently, with the yield on the 10-year U.S. Treasury nearing 5%, voices in the market have suggested that "the surge in long-term rates is effectively doing the Fed's tightening work," thereby reducing the necessity for subsequent rate hikes. In response, Williams gave a candid reply.

"I personally dislike the phrase 'the market is doing our work' because we need to do our own work."

Williams emphasized that the 10-year yield reflects not only monetary policy but also fiscal policy, productivity expectations brought by the AI boom, and uncertainties around energy prices.

Williams also addressed the market's interpretation of his comments on September 29. At that time, he stated that the Fed does not need to rush into action again after the September rate hike, a statement that, combined with similar remarks made by Fed Vice Chair Philip Jefferson thereafter, effectively lowered market expectations for an October rate hike. Some observers speculated that the two speeches had been coordinated and might have involved communication with Fed Chair Walsh.

In this regard, Williams clarified that this merely reflects his personal judgment:

"I speak for myself... I am expressing my own views. I stay in regular contact with him (Walsh). We meet often, but I won't say more about it."

Fed officials will hold a policy meeting from October 27 to 28, during which they will consider the September CPI data to be released on Wednesday—this is the last inflation report before the meeting.

Two Federal Reserve officials emphasize that AI is driving short-term inflation; Boston Fed President indicates a need for rate hikes this year

On Tuesday, Chicago Fed President Goolsbee stated that if the path for inflation to return to the 2% target becomes clearer and the job market maintains full employment, there will be room for lowering interest rates.

However, he also warned, "If service sector inflation continues to rise, and AI investment demand spills over and causes more widespread economic overheating, the situation will become more complicated."

Goolsbee further pointed out that price pressures have spread beyond data centers and directly related sectors, stating, "I am increasingly worried that this is pure, old-fashioned demand overheating."

Richmond Fed President Barkin indicated that the Fed has been able to observe the impact of AI investment on U.S. inflation.

On Monday, Boston Fed President Collins, speaking at a banking conference, stated that to ensure inflation returns to the Fed's 2% target in a timely and sustained manner, interest rates may still need to be raised further before the end of this year, but specific actions and timing will still depend on subsequent economic data.

Collins expected that the U.S. economy will continue to grow, the job market will remain stable, and productivity improvements will help balance supply and demand and lead to a decline in inflation.

She emphasized the importance of monitoring whether inflation remains primarily concentrated in energy and AI infrastructure-related areas or will further spread to other sectors of the economy.

The full translation of Bloomberg's interview with Williams on Tuesday:

Bloomberg: President Williams, thank you for hosting us at the New York Fed. From your public speeches, it seems that your inflation forecast this summer, at least for overall inflation, has shifted to above 3%. I would like to know what the main drivers are for this, and whether your core inflation forecast has also started to shift significantly?

Williams: Yes, I believe a series of developments have impacted our outlook on inflation. I mentioned tariffs, then the Middle East conflict and broader energy price trends, and thirdly AI. I believe over the past six months or so, we have seen that the effects of tariffs have played out essentially as predicted by many models and analyses. This means they have impacted prices of imported goods but have not had a spillover effect on broader prices in the economy. So we have seen that those effects have mostly faded, but they had a significant impact on core inflation over the past year or year and a half.

Regarding energy prices, I believe this is a more important story for the near-term inflation outlook this year, as energy prices have risen significantly and remain high. This is affecting inflation even within this month. We are seeing this effect transmitting from energy prices to core inflation as expected. The most obvious example is airfares, but there are many other examples, as energy is indeed a significant input for many goods and services in the economy, including food and others. So we are seeing these effects impacting my overall inflation forecast and also affecting core inflation to some extent, at least in the short term—considering 2026.

Finally, there's the AI story, which may be the most recent part. We are indeed seeing rising goods and other costs associated with AI construction: data centers, huge demand for chips. So we are seeing this reflected in specific categories, but prices in AI-driven sub-industries have seen quite sudden and large increases. Similarly, so far this has mainly affected those goods and goods using chips, including consumer products. We are also seeing some signs of spillover into construction labor costs. So this is the third area—AI demand-driven inflation—I'm closely monitoring.

However, I think there is also good news in the broader picture. Some of these factors have led to inflation being slightly higher than expected this year, but we see that inflation expectations remain well-anchored, and we see wage growth continuing to be moderate. Thus, the growth of unit labor costs is slowing, which aligns with inflation at 2% or lower, and some of the indicators we use to measure potential inflation have noticeably declined. If you look at the New York Fed's multivariate inflation index or trend inflation indicators, they are about at 2.3%. If you look at the trimmed mean from the Dallas Fed, it has also been roughly at the same level over the past six months.

So I believe these indicators tell us that over time we have indeed been influenced by tariffs, energy prices, and now AI, but some of these do not indicate that potential inflation is rising. In fact, those potential inflation indicators have decreased. No single indicator is perfect; these are just the indicators I am watching. So overall, this year there certainly have been some driving factors pushing inflation up, which has led my inflation outlook to be slightly higher. I think what I'm currently watching most closely is AI-driven demand and how it may evolve. But I do not believe there are other inflation driving factors emerging.

Bloomberg: We have been hearing from you for a few months now, and the main story seems to be: while there are risks, there is an anti-inflation story coming. In fact, after the July meeting until August, you were saying this, and we feel—also possibly we misread—your tendency to oppose a rate hike in September, but of course you ultimately supported the hike. So I want to know what data changed your mind to be more concerned about upside risks, or do you think the core of the committee is moving in that direction? Frankly, consensus in the FOMC is important. Was part of your support for the rate hike aligned with the core direction of the committee?

Williams: Well, I strongly supported that decision. I believe it was the right decision—absolutely the right decision. I believe the data you mentioned and how the data evolved led me to believe that decision was appropriate.

So I believe the two data points I have mentioned: one is the issue with energy prices. If you look earlier this year, futures markets were showing oil prices would decline significantly, which may have been overly optimistic. It was more than likely that the optimism at the time was thinking the conflict would be resolved, so I believe that has changed. Energy prices are still quite high now. Futures still expect energy prices or oil prices to decline, but the decrease is not as significant.

The second point I think is very important is that we are seeing the spreads from oil prices to other fuel costs (like gasoline, and of course diesel), which reflect that not only has the Middle East conflict reduced refining capacity, but also the conflict between Russia and Ukraine. So we are witnessing an additional shock, if you like to say, or additional manifestations of these developments, namely, the significant increase in fuel costs. So I think this is a set of data: higher energy prices, possibly lasting longer, consistent with my view that inflation will not decrease as quickly as I had hoped before. Secondly, I still believe energy prices will not push inflation higher next year and thereafter; I am just observing the current situation.

I think concerning AI, this transition in AI demand is real, sizable, and likely to persist, and that the supply side will take some time to catch up to the demand side, from my perspective, this has become clearer. For me, this is reminiscent of the traditional issue of using monetary policy to balance supply and demand. With more demand drivers from AI propelling the economy to grow faster at least in the short term, it indicates that adjusting monetary policy not only helps with inflation, but the economy is also clearly able to withstand slightly higher rates. So these are the factors influencing my thinking.

So to your premise, I still believe there are anti-inflation forces at work. I think these are not fundamentally changing. In fact, my forecast for inflation next year is that inflation will decline sharply—overall inflation will be close to our 2% target for the year, reflecting some of the factors I mentioned, namely, that tariffs will have become a thing of the past. Even looking at the changes over 12 months, energy prices are expected not to push inflation higher next year—they may become a deflationary force. I truly believe the overall balance of supply and demand in the economy is expected to remain well-balanced, and we will see inflation continue to decline.

Bloomberg: In that sense, the minutes from that meeting reflect that there was a set of people who believed rate hikes were indeed necessary, while another group viewed them more as a risk management decision. I wonder which group you belong to in your view. Do you think this was a genuinely necessary rate hike, or was it more about guarding against the upcoming inflation risks?

Williams: Fundamentally, I believe monetary policy is always about risk management. It is always about assessing the economic outlook, uncertainties around it, and the risks to achieving our dual mandate objectives (maximum employment, price stability). It’s always a probabilistic question and perspective. So my answer is yes. Yes, I think given that the downside risks to our maximum employment goal seem to be diminishing, a rate hike is appropriate. The economy has proven resilient. If there is any difference, it seems the economy has strengthened marginally, so it feels that the risks to maximum employment are slightly lower, but inflation risks, due to the factors I mentioned—particularly regarding energy prices and AI factors—indicate inflation risks are slightly higher. I would use the terms 'slightly higher' and 'slightly lower.'

I do not believe the data in any way indicates a fundamental change in the economic outlook; rather, it is more a qualitative reassessment of where the uncertainties and risks lie. So from my perspective, the answer is to start from risk management, look at the risks to achieving our objectives, and see what kind of monetary policy settings would best achieve that goal, and that is the right decision.

Bloomberg: Tomorrow, you will release the CPI, which is the last piece of inflation data you can see before the next meeting. I would like to know what you will focus on to assess whether your outlook still holds, and in your view, what kind of core data would meet your 2% expectations?

Williams: Well, I think the answer to your first question is that we will look at all the data. More importantly, my colleagues at the Fed will examine the data closely, understand all the details, seasonal factors, special factors—these are all reasons that the data in any given month may fluctuate for different reasons. The second point I want to make is that I firmly believe in relying on data, but it should not rely on any given month’s single data point. Data can be influenced by various factors, including seasonal factors, which are somewhat special. So what you really want to look at is potential trends. That’s my view.

You get another data point; it is a piece of the puzzle that, along with all other information we have, forms a whole. When I look at the potential trends in inflation, what am I looking for? Energy prices fluctuate due to reasons we understand—conflicts. Some other factors on the AI side, I expect this trend will continue. But we need to look more broadly—what are we seeing in services, housing, and goods, excluding some of those categories.

What it’s important for me is to continue observing whether potential inflation is moving towards the 2% target. I do not expect it to hit 2% immediately. But we are moving in a way that is potential trend-wise in the right direction. I think that’s actually what we've seen over the past few months. But we need more data to really try and interpret what it implies for next year. Monetary policy does not kick in within a day or a week. It takes a year or two to have a full effect on the economy. So what's really important is what this tells us about what could happen in the next year or so.

Bloomberg: But do you have a core data point that would make you more concerned than others?

Williams: Well, I prefer lower rather than higher, but I do not have a specific number. Again, after adjusting for some of the factors I discussed, I am looking at these potential trends.

Bloomberg: One more question about AI-driven inflation. You have said again that this is a primary driver you are focusing on. What makes you believe that the supply-demand imbalance caused by construction will be resolved in the short term? The pace of investment has not slowed. It is unlikely to be affected by higher rates. In the absence of a stock valuation adjustment (which could slow things down), why do you believe supply can catch up to prevent the broader inflation impulse that may continue to spread throughout the economy?

Williams: So that's precisely what I'm watching and observing. I think there are several reasons. First, a small adjustment in monetary policy won’t change the underlying trend of AI—the demand side—at all. I think investors there are relatively price insensitive. They view it as a tremendous opportunity that will bring substantial productivity improvements and transformative changes; I believe this is not entirely unaffected by interest rates, but I do not expect monetary policy to have a deep impact on this.

But in the rest of the economy, I mean, I know we’ve been talking about AI. But the U.S. economy is over $30 trillion. There’s much more than just AI. So I expect to see, and I think we are seeing, the impacts of changes in monetary policy and other financial conditions on the rest of the economy, whether it’s in housing or in other sectors of business investment outside of AI, other decisions that are more sensitive to interest rates, including financial conditions and the impact on long-term rates and the value of the dollar. So I think all these other channels are definitely at play.

This is an unusual situation; we have an economic sub-industry that seems to be operating at full capacity, and this will continue for a while. But I think what's critically important is to look at the broader economy. I do not expect chip prices to decline. They may continue to rise. But we need to consider not just them, but what is happening with overall inflation.

Let me give a few data points that I think might support this argument. First, there is indeed no indication that the overall labor market is imbalanced. The unemployment rate is just over 4%, and has remained quite steady. My personal view is that the labor market is slightly loose, but not by much. Others would say we are roughly in balance. So I think we are not seeing that imbalances in the labor market would bring additional cost pressures for businesses in the broader economy. We also see this from wage data. As I mentioned, if you look at ECI or hourly earnings or other indicators. We are not seeing inflationary pressures from there.

So when I step back and think about what problems might arise from inflation driven by AI in this industry, it’s if there are other factors in the economy that are also increasing inflationary pressures. Whether it’s high wage growth or pay increases in the labor market exceeding productivity or other sustained and large-scale import cost pressures.

Certainly we have already seen this in energy. That’s absolutely true. But outside of energy and AI, we have not seen these broader pressures that push business costs higher.

What we experienced in 2021 and 2022 is that businesses in the U.S. and elsewhere were essentially caught in a bind due to broad costs rising, whether it’s import costs, labor costs, or other input costs. They were unable to regain costs by substituting one thing for another or adjusting production methods. So they essentially had to raise prices. So we are not seeing the same broad inflation pressures now.

Another thing I want to highlight is GDP growth. As I mentioned the labor market, but GDP growth is growing, averaging slightly above what I think potential output has been over the past four quarters.

For example, growth slightly above 2%, I think potential output could also be slightly above 2%. So the overall economy, despite AI being such a strong driver and all that, is not overheating. So I think these are the things I'm looking at. Again, monetary policy will not change the relative prices or relative investments of AI versus other things; it will impact the broader economy.

Bloomberg: So you're not saying that AI will no longer be a driver of excess inflation, but rather that overall will remain balanced.

Williams: That’s precisely the goal I have when thinking about two things that are critical to our monetary policy. One is to maintain overall balance between supply and demand, and the second is to ensure that the price pressures we see—whether from energy, previous tariffs, or AI—do not spill over into broader or more persistent inflation.

So those are the things I am focusing on, knowing that chip prices will act according to their market forces.

Now, one thing you mentioned I should add is that I do see that ultimately AI construction will lead to supply effects in two major aspects. First, these are very, very strong price signals that will prompt an increase in supply. These prices are very high. Producing chips is extremely profitable. So obviously, significant investments are being made to build factories to produce more chips. This takes time. It will take a while, but ultimately, supply will come on board. Second, I ultimately think we will see bigger supply-side effects from AI construction leading to higher productivity. We see this in individual companies and enterprises using AI. They are gaining huge productivity boosts. I would say we are only starting to see hints of this in the broader economy. So I feel that in the future, the key challenge is: how long will it take to reach that point? In the future, we will gain positive supply-side effects from this, but what’s critical is the competition—in AI, it’s a competition between demand and supply. Currently, demand is certainly winning in the short term, but I do indeed expect supply-side responses in the future, and in fact, it will become a deflationary force.

Bloomberg: But even before you see those productivity boosts—which could be said to be more difficult to pinpoint in terms of time frame—in your model outlook, I believe inflation is still at 3.5% this year, then slows to 2% and reaches the target by 2028, is that correct? At what point within that time frame do you see AI’s supply and demand becoming more balanced?

Williams: Within AI?

Bloomberg: Yes, in AI. Is it within the next two-year time frame or longer?

Williams: I think it’s going to take longer. Assuming the demand impulses persist as expected, that’s a big assumption. It may change based on perceived returns on these investments. But given my baseline expectations, I expect the demand in the AI space will remain very strong. I believe supply will respond, but it will take some time to catch up. But I think at that point you would also expect to gain positive leverage in productivity, if you will, starting to see that.

Now productivity growth has been pretty good. With the latest revisions to the national income accounts, I mean, productivity since 2019 has certainly been above 2%. I am becoming more convinced that we are on a good or strong productivity path, and that this likely is not primarily AI. So I think this is another factor that helps me have some confidence that anti-inflation dynamics are forming. We have good productivity growth. As I said, unit labor costs are growing quite moderately. So in the medium term, not in the near term, I believe these are good indicators that inflation should return close to 2% by the end of next year.

Bloomberg: Let’s talk about interest rates. From your speech in Buffalo, you said you still believe there will be one more rate hike this year, which will bring the target benchmark rate to between 4% and 4.25%. Is that enough to get the job done? Given what you said yesterday, some effects you cannot really control with interest rates—chip prices, chip supply, oil.

Williams: Well, again, we cannot control those, nor should we. I mean, I think these price signals from AI investments, regarding AI, like chip prices or some others, we hope those price signals play out, prompting people to use them for the highest value, letting producers increase their output. But if you look at what’s happening to overall financial conditions, equity markets are still very active. People are very optimistic about the U.S. economy; I actually am as well, and very, very optimistic about AI investments. But we have seen long-term yields rising, so I think this means higher mortgage rates and higher borrowing costs for other businesses, including AI and other parts of the economy.

So do we feel we have done enough? Time will tell. We will get more data to assess. But my point is that this is not an overall overheating economy. This is not an economy that is increasing inflation pressures beyond the specific shocks we are experiencing. So I think this brings us back to this risk. It is always risk management, i.e., how do we maintain a robust economy that approaches trend growth without inflation spilling over into broader, more persistent inflation. So what's really important is to maintain the overall balance of supply and demand, which I believe relates to assessing whether the economy is overheating, and I don't believe it is.

Earlier this year, or even recently, up until the latest employment report, you might have heard that argument for renewed acceleration, that the labor market is heating up again. I think the labor market is performing quite well overall. It is robust. It has been stable. It has shown resilience, but we are not seeing widespread indicators of overheating. There have been indeed seen in certain categories, and employers I have spoken to have said certain types of skills or supplies are tight. But I think that just reflects different sectors of the economy performing very strongly while others are slowing down.

Bloomberg: So, again, trying to evaluate your outlook on interest rates. You still believe there will be one more rate hike later this year. Is that enough? What about 2027?

Williams: I want to return to data dependency. I feel that with strong demand, and I think the revisions to the national income accounts since then, they have reinforced that this is a growing economy that is quite decent. I think it also reinforces that this is a supply-side economy that is quite decent. So I have raised my view on potential growth in the economy by roughly a couple tenths of a percentage point, so that’s good news. I like a strong economy, especially a strong supply side. The labor market, I think, has again proven to be stable and robust. So the question is, as data accumulates, does the story we see of the economy having a good balance between supply and demand hold, with demand being slightly strong, frankly with supply also being slightly strong? This is an assessment of what level of interest rates can maintain that balance, and then obviously ensure inflation…

Bloomberg: I understand you still rely on data, but you must make an estimate for 2027…

Williams: If you fundamentally ask me whether I have changed my mind, whether the outlook is roughly in line with my expectations, I have not changed my view that one more rate hike might be appropriate. I see no reason at this moment to believe we need to do more. But you ask about 2027, we will have to see. I mean, it is only mid-October. This economy is influenced by many different factors, and we can only continue to observe the data. Again, the focus is not on a single data point or what you know, but really looking at the potential trends we see regarding labor, employment, and inflation.

Bloomberg: Many of your colleagues also see this anti-inflation story emerging, even some of the more hawkish FOMC members to some extent. But there are differences regarding the appropriate interest rate path. I am not asking you to speak for any individual. Do you think those differences boil down to different expectations for how the economy will evolve, or do they come from different estimates of the neutral interest rate? In other words, what level of rates is necessary to get us down to 2%? Is this really a neutral interest rate issue?

Williams: Well, thank you for mentioning r-star—it took a while to get to that. As you said, I won't speak for my colleagues, but I do look very carefully at the SEP and obviously listen to my colleagues' comments. I think the most important thing is that we absolutely share a commitment and determination to achieve the 2% target and do so in a timely manner. There is no disagreement on that point, and clearly, we are committed to both aspects of our dual mandate.

So I think when you gather 19 people together, you start to reach a consensus on the goals we are trying to achieve and the strategies for how to achieve them, that is a good starting point. Then comes the hard part. It’s about the execution of that strategy, how you assess the potential factors affecting inflation or inflation expectations, weighing different pieces of information. How do I view many of the frankly more hawkish arguments, as you said, there are scenarios in which the economy evolves in a way that you need a higher interest rate path to best achieve our goals? There are also, I would say, scenarios where the path of the economy going forward suggests no such need. We don't know which scenario we live in.

The major themes of global central banks and monetary policy are scenario analysis; we have done a lot of scenario analysis internally at the Fed. We do not publish these scenarios as part of our communication, but we do such analyses. But I believe my 18 colleagues embody this scenario analysis. They work with the team through all the briefs, through all the materials and data, and then arrive at, well, this is my baseline scenario; these are the other risks. So I feel this captures—my colleagues' comments capture this quite well, I think—this set of possible outcomes that we need to think about and be prepared for. I have expressed my personal views and my baseline scenario. I think, again, not speaking for my colleagues, but I think we all need to observe how the data evolves and how those different risks evolve and then make our best decisions.

Now, regarding the neutral interest rate, I can answer that question fairly easily, because my colleagues, or say most of my colleagues, filled out their terminal rates in the SEP, and you can see a range of views there, and they have evolved over time. They have evolved from 2012 to 2020. You know when we initially—they have gone down as evidence showed it was below our neutral rate. We saw a moderate to significant rise in the past few years as resilience in the economy and potential strength emerged, along with other factors. So I think the differences in the neutral interest rate have played a role in how people are thinking about the overall interest rate path needed to achieve our goals, because that’s the definition of neutral interest rate or R-star. For me, it tells you how you think about the eventual path of interest rates, where they need to be in 2027, 2028, 2029, and beyond; it does not tell you what to do in any given meeting.

Bloomberg: Do you still believe we are at a moderately restrictive level?

Williams: There are many different versions of that question, you can look at various indicators. I think if you specifically ask me about R less R-star, it’s not from all the factors that might be incorporated into assessing financial conditions. Based on my projection of inflation, assuming that holds true, I do believe we are in a—I would say a world where neutral rates are not as low as they were in 2019. I think the neutral rate has likely risen by 25 to 50 basis points, as fiscal policy or fiscal factors or stronger growth have pushed it up a bit, which is good news. But I think my point is that the long-term neutral interest rate may still be quite low, around 1% or slightly above that, so from this perspective, if the economy behaves as I expect in the coming year, we are not far from neutral.

Bloomberg: I want to briefly talk about your recent speeches. Now, you hold a special position in the FOMC. Historically, there has been a so-called "three-horse carriage" consisting of the New York Fed President, the Vice Chair, and the Chair, and most seasoned Fed observers and economists we spoke with, some of whom have worked at the Fed, rarely say they believe you and Vice Chair Jefferson got any signals telling you what you should say, but at the same time, most assume you would each inform the chair in advance of the messages you would convey, and those two speeches are two days apart. This is a pretty clear signal that you both believe there is no urgency to raise rates again after September. So the question is: Is it reasonable to conclude that the Chair received advance notice, and he clearly did not interfere to prevent those messages from being conveyed?

Williams: So I speak for myself. I won't speak for others, but as I have done over the years, I have given my own views on the economy, the risks to achieving our goals, and how I see monetary policy positioned in relation to that. Now, obviously, I agree with my colleagues. This is not the time for us to be trying to say in the FOMC statement, well, this is certainly the future path of interest rates. That kind of forward guidance. But for my remarks, my own view on where we are, and I think it is very important to truly look at the underlying trends in inflation, because we are being hit by various shocks, whether it’s energy prices, AI, or others.

So I am expressing my own views on how I think about data collection, and frankly, I am quite satisfied with how the committee has dealt with this in the last few meetings, collecting data, thinking carefully. We did not take action in July. We wanted to get more evidence. In September, we had two months of data. This was an opportunity to assess, evaluate what potential trends are telling us—not to rely on data points but to respond to those trends, and that’s exactly how I came out of the September meeting. Let’s do this again. Given our current situation, this is the right way forward. The economy is proving to be resilient. I think the potential anti-inflation trends are, let’s observe how these continue, but also to consider these other factors.

So this is how I approach the comments in my speech. Regarding the Chair, obviously, I stay in regular contact with him. We meet often, but I won’t say more about it.

Bloomberg: What does regularly mean? When was the last time?

Williams: We meet regularly; you can see this in my calendar and will meet whenever possible. This has been standard practice since I came here. It is really about me expressing as effectively as possible how I see the economy and where monetary policy is.

Bloomberg: But can we say that you are expressing your views, but at the same time, the wording you choose sends a clear signal to a market that seems inconsistent with its expectations? This is at least part of the point of your speech, which is to signal to the market, not about future meetings multiple times, but just one meeting, that you might have some variance?

Williams: Clearly, a significant part of my job is observing what the market is doing. When we talk about the market, we are of course discussing prices; there are many people trading and thinking through these issues. So clearly I observe what the market pricing is, what information is there, and I have learned not to try and guess the market. The market draws its own conclusions; it has its own view about what uncertainty exists regarding what the economy will do and what this means for monetary policy. So I look at this and observe. I find it useful information.

But when I give a speech, as I think about my remarks, they are not about the market. They are about my assessment of the economy, the risks to achieving our goals, and how monetary policy is positioned there. So I think the best thing for me is to once again say that this is my view and to be as clear as possible.

Bloomberg: So regarding the market, the 10-year yield is about 5%, I want to know how you see this. Is the market at this time doing at least part of the Fed's tightening work, and do these yield levels at least allow you to pause a bit in the rate hike cycle? If we indeed are in a rate hike cycle, perhaps we do not have to stack rate hikes one meeting after another.

Williams: So I personally dislike the phrase 'the market is doing our work' because we have to do our own work. The market, as I mentioned, clearly market participants are looking at all the data we are looking at. They are looking at all the factors and making their decisions. The 10-year yield is not just about monetary policy. The 10-year yield is about many long-term factors, about supply and demand and the AI boom and the resultant productivity. About fiscal policy, uncertainties around energy prices. So besides monetary policy, there are many factors affecting long-term yields.

From my perspective, these clearly impact real people’s decisions. If you are trying to buy a home, if you are trying to invest, if you are a business trying to expand, these are affecting your financing costs and other aspects of interest rates. So these are important for understanding how financial conditions affect different parts of the economy.

The stock market tells us a very bullish economic story, but at the same time, on the interest rate front, borrowing costs have risen. These are inputs I consider when thinking about the economic outlook and some of the risks associated with that outlook. But ultimately, we must come together to discuss how best to position monetary policy for the economy, which is part of the story, and manage risks. So I do not believe they are doing our work. I do believe when monetary policy is communicated well, consistently, and is focused on achieving goals, market participants understand that, so they will anticipate some of our actions.

I think this is a sign of how monetary policy works, but really it is not about them doing our work. At the end of the day, we have to make the right decisions to do that. I also dislike the term 'pause.' I feel like I have never paused. I am working every day. In my years doing this job, I have never felt like I have paused, but I return to what I said earlier, I see an economy that is in fact proving to be resilient. I feel we must understand the different factors affecting supply and demand and inflation, but this means carefully assessing and evaluating incoming data, what that data means based on all the information we have, gathering that information, having good discussions, and making decisions. This is how we do good monetary policy to set the stage for our success in the future and in the coming year.

Bloomberg: So regarding your position, this is your second term, you are nearly 65 years old, and have some time left. But I want to know if you have considered retiring early or if you plan to complete your term?

Williams: Well, you can gather from this conversation that I have a lot going on in my everyday job, with the economy and monetary policy. This is where I am 100% focused. Beyond all the things we do at the New York Fed regarding executing monetary policy and other responsibilities, I absolutely love this job. I enjoy working with those I work with. So when I started here in June 2018, my intention was to serve out my full term and continue to do so; that is my stance. I am not thinking about what comes next.

Bloomberg: To clarify, a full term refers to the 10 years following June 2018.

Williams: Again, when I took this job, that was my understanding. That is my commitment to the board of the bank, as well as my commitment when my family relocated to New York that we would be here for ten years.