저자: ChristopherWaller/사회자SalihaMohsin제외 · 외국인
요약 보기
AI 생산성 효과는 시간이 걸리며 측정(GDP)에 잘 안 잡힘. AI 설비투자 붐이 반도체 가격을 올려 단기 인플레 압력을 만들고 2~3년 지속 가능. 장기금리 상승은 인플레 기대가 아닌 실질성장 요인이 크며, 미 국채는 안전자산 프리미엄이 줄고 있다는 연구가 있음. 금리 인상 여지는 단기금리보다 10년물 글로벌 수급이 좌우. 연준 독립성은 채권시장과 의회·법원이 지킨다는 입장.
다듬은 글
Now, Christopher Waller, the head of the Federal Reserve, is talking to our own Saliha Mohsin here at the Bloomberg New Economy Forum about monetary policy in the age of AI. So, let's listen. It's a great pleasure to be here with Chief Waller. We both flew in from Washington, D.C. to speak with you in person.
One thing I noticed is that just like in Washington, if you walk around Delhi, the main topic of conversation is AI, which is what we will continue to talk about. Leader Voller, over the past couple of years you have made many comments about the benefits of AI, including the productivity boom, contribution to GDP growth, and improved living standards.
And this all concerned the United States. I would like to hear from you how you see the impact of AI on our economy in the context of the rest of the world, especially emerging markets like India. Yes, I think about AI like this : first, there are those who create AI tools, models. Then there are those who supply all the resources to create them.
These are data centers, chips, all that stuff. And thirdly, these are the people who directly use these models. Of course, when it comes to creating models, the US is the dominant country. China is also present in this area . In terms of resource supply, most of the production comes from Taiwan and South Korea.
Data centers are located all over the world. They are not only located in the USA. And, of course, users are anyone who wants to use it in their work. There is a very high correlation between GDP per capita and the use of AI in human work. So that's one of the things that worries me a little: the use of AI in the world is uneven.
And this means that there will be a difference in performance if the technology is not used. So, because of this, you will see a potential increase in the gap in GDP per capita . But I think it's amazing technology . I really think so , I'm not a pessimist. I don't believe that AI will end humanity in the next 2 years or anything like that.
But it is a tool, and that is what I tell all of our employees at the Federal Reserve. We will teach you how to work with it because we want you to be more productive. If I were going to replace you with AI, I would n't spend a cent on your training. It would be a waste of money. So this is really a tool that I expect all of our employees to use.
So , this is a kind of basic global picture of how, in my opinion, the whole world of AI works. Europe, for example, doesn't produce much AI or import it, but they are big consumers of it. And this has already spread quite widely in the business sector and the community. Let's talk about the benefits.
You say you are an optimist. You don't worry about, you know , the human race. When will we see some of these benefits? A productivity boom, our wages rising, inflation curbed. These things have n't really emerged for the average person yet. Yes. And it's often the case that when there are big technological changes, economists know that you usually don't see productivity gains for years.
The big part is that companies have to figure out what am I using this for ? How am I going to actually improve productivity? They're not entirely sure, so there's a lot of trial and error and a lot of experimentation. But then you start to realize where it really has its impact. I mean, I'm telling you, if anyone is into programming, this thing is just amazing.
I mean, it's just shocking how good it is. So, in terms of programming, yes, I mean, if you're a programmer, your job is going to fundamentally change. This doesn't mean you will lose your job. You will simply be doing different work than you did before. Things you couldn't do because you were too busy with basic coding, you can now expand what you do in terms of producing your output.
So, uh, I think it's going to be a while before companies start figuring out how to actually do this. And you might. Listen, you might not see it from the perspective of. That's one of the things about measurement that I was thinking about. Let's say I can do the same amount of work faster. So, instead of 5 days, I can do it in 3 days.
Maybe I'll spend all this time just resting more. My output has not increased over a certain year. I simply have more time to devote to leisure and enjoying life. This still means it's good for welfare even if you don't see much growth in GDP per capita . So, you either have to rethink whether we measure GDP or productivity per unit of time, instead of quarters, because now it's a matter of days.
So you should think about it. And speaking of productivity, there are many things that this technology will simply make better, even if it doesn't translate into your work productivity. I always like to emphasize that Google Maps is an amazing invention. I don't use it to improve my own productivity.
It helps me navigate the cities I visit and I don't get lost. This is not reflected in the GDP figures, but it makes my life much easier . So I always try to explain to people that technology can simply make life better or really increase the productivity and volume of what you create. One indicator that is easy to measure is inflation.
And right now, in the United States and other parts of the world, the discourse is centered around an affordability crisis. There is certainly momentum coming from the impact of AI on the economy. In the short term, when do you think the impact of AI on US inflation will fade? Yeah, I mean, the whole build-out process—obviously, if you're putting a few trillion dollars of investment into AI and you need to buy all this equipment, and your demand is completely inelastic, you're going to pay whatever price you can if you're one of the hyperscalers.
You will raise the price of all types of semiconductors, and that will affect everyone else who uses those products. And this is exactly what we are starting to see in the US. We're starting to see AI causing widespread inflation because it simply increases the demand for everything else. Eventually, when all of this is completed and built out, this demand pressure will disappear, and then, hopefully, we will start to see supply effects, which will mean more production at a lower price, faster.
So, it's a two-step process. Demand will be significant, especially given the volume of investment. I mean, the amounts that are being spent on this, compared to any other point in history, are just shocking. Therefore, in the initial stage, you will observe this price pressure. How long will it last?
This could take two to three years, given how much they have already committed to spending in the future. If all of this comes to nothing , if there is a sudden pullback or you are unable to complete this development, it will hold down prices. But future inflation benefits, in my opinion, are not coming anytime soon .
They still have a few years to go. Several years of waiting. Is there a possibility that inflation caused by the development of AI could linger in the economy for longer? Yes, as I said, this is a very specific phenomenon. This is not a general effect. It applies to a narrower range of goods than other things.
So we know that someday this huge construction and huge demand will not stop completely, but will decrease significantly. And when that happens, we will see demand pressure ease. But judging by the hyperscalers' plans, they still plan to spend trillions of dollars over the next 3 years. So this demand pressure may persist.
And we just need to decide how we want to respond to it . Well , let's talk about the development of AI. Investments continue to come in large volumes. Much of this is issued in the form of corporate debt . We see a lot of funding to support the development of AI. And, of course, this debt depends on US Treasury bonds.
We have seen profitability increase dramatically. Global profitability indicators are high. Can you talk a little bit about what you see in US yields, whether that could be a long-term thing, and how much of the higher global yield picture is related to the situation in the US and what is separate? Yes, everyone suddenly started paying attention to high returns.
And if you look closely, the yield did not increase due to expectations of higher future inflation. This is not traceable. Inflation expectations are firmly anchored. So, all this growth is not coming from an inflation premium. That's not where it comes from. This is a real growth factor. If you have real growth, real productivity and that investment pays off, you are willing to pay more in interest costs.
And people know this, and they will demand appropriate compensation. So real growth affects this real return. Now, on top of that, there are general fiscal concerns around the world. Governments, at least in developed countries, are not in the best shape. And we're seeing it in France right now. So, in part, it's just increasing profitability in some countries, which is not necessarily related to the US or maybe other countries.
And competition from AI is a very, very serious problem. Because even if money is flowing into the US, meaning demand is growing, there is an inflow of capital from abroad , which offsets the price effect. This money comes from other countries. Their profitability will increase. So even if there is competition and no capital inflow, profitability in the US will increase.
If demand is higher, but there is an outflow of capital to other countries, you will see higher returns in those countries because funds will no longer be available for investment there. But the development of AI, I believe, definitely has its impact. I mean, if you look at it, Google issued what seems to be a 40- year bond with a yield that's 100 basis points higher than U.S.
Treasury bonds. What would you choose? A 100 basis point premium on Google bonds that are likely to be paid. It doesn't look like they're going bankrupt. So it affects how people look at things. So, in short , that's it. There are economists who say that we— at least the United States, and perhaps the entire world— are entering a phase of higher profitability to which we need to adapt.
Do you agree with this? Well, like I said, if it's all because of taller stature, then that's fine. I would rather choose higher growth and higher returns than low growth and low returns . For me it is obvious . What is worrying is whether this is driven by inflation expectations or fiscal concerns, which I think is a real problem in France right now.
This could happen in the US too. People are starting to realize that our fiscal situation is not that good. And then bondholders will naturally demand higher yields to buy that debt. So if you live with higher returns due to higher growth, that's fine with me. And if you live with a higher return because of fear of inflation, that's not the best outcome.
I want to ask you about the fiscal policy situation and how it affects your role and work at the Fed, but first: when yields rise, does it to some extent serve the function of raising rates? In the current situation, do you think this is enough, or does the Federal Reserve still have room to raise interest rates?
Yes, I have long argued that the 10-year rate is determined by global supply and demand. What I change in the overnight deposit rate doesn't have much of an impact on the 10- year yield. So, even if I do it, it only works for a short period, not 10 years ahead. So I've always believed that the 10- year rate is driven by global factors, global demand, as well as supply from the U.S.
Treasury. And it doesn't depend so much on my actions . So, as yields rise, will this affect interest rate-sensitive sectors like housing, autos, and others? Yes, of course it will have an impact. Will this lead to a decrease in demand for these goods and downward pressure on prices? It would be enough.
We do not set a goal to do this, but it will be one of the natural side effects of such a jump in yields on 10- year bonds and above. So, it does some of the work for... It really does some of the work. Let's return to the debt problem, which is global and has reached its highest level since World War II.
In France, we see that investors are now realizing that there may be a limit to how much debt France can issue. US debt, US Treasury bonds—that's something else entirely. They are considered a "safe haven" in times of turmoil and risk. We have...the US has excessive privilege. Do you think these privileges are running out?
Yes, I would say that there is a lot of academic research emerging in the US that suggests that the safety and liquidity premium on US Treasuries...It's... It's gone. So, it always meant the following. People are willing to pay more for U.S. Treasury bonds because they are safe, liquid, and easier to sell compared to other government debt.
So, the higher price they are willing to pay means a lower yield on that debt. But if you no longer consider it a safe asset or the best option to hold, then that premium disappears. People will not pay this higher price. And the aforementioned academic studies say that this has disappeared in the last couple of years.
This is another reason why your rates will increase. If you lose this advantage in profitability, which is not a small amount. This could be around 50-70 basis points. We are now seeing that emerging markets have managed their deficits a little better than many Western countries and developed economies.
Well, and again, regarding France, investors, " bond vigilantes" are showing that there is a certain limit. How do you think it would be possible to demonstrate in the US that a certain ceiling has finally been reached ? Well, this is... you'll see this sign, as I said, in the bond market. So, the bond markets are now worried that this debt is not as safe as I think or as I always thought.
They won't pay, uh, they'll pay a lower price to own it , and therefore a higher income. So you'll start to see that when profitability starts to increase. It doesn't have to be a crisis. I mean, sometimes it's a crisis, but otherwise it's just increasing profitability. This is roughly what you see in France.
The profitability just went up, but they can still release it. They just have to pay a significantly higher income to get people to buy it. So, that would be it. I think you would see this if you started having problems with auctions where demand was n't showing up, and that's a problem. Uh, if you're afraid that, you know, one way to get rid of debt is to just devalue it with inflation, then you get this inflation premium that suddenly accumulates.
This would be the most common way that fiscal policy affects rates if they believe it will lead to inflation. And that will push yields up, so I have to have a higher income to take on this debt. So, you know, auctions, rising bond yields, talk of a budget deficit that's being taken into account now.
We don't see a lot of this in the US per se, but like I said, France— man, everything just suddenly exploded there. Yes, US Treasury bonds had good auctions in August. There was one that was the best auction in 20 years, I think. Uh, what this really brings to light is talk about fiscal dominance and central bank independence.
We see some comments, again , going back to France, that the central bank there is under pressure. Do you think this " Chinese Wall" will hold up for world bankers who are faced with such fiscal policy issues that spill over into their monetary work? Well, we always have to take into account fiscal policy and how it affects the economy.
Not government funding, but simply how it flows and affects the economy. So, we always have to ask ourselves: if this tax or spending bill is passed, how will it affect consumption and investment? And then, you know, we have to adjust our actions according to the situation in the real economy. Fiscal dominance is when monetary policy has to be conducted to help the government with financing.
That's when you should start to get seriously worried . And so far, at least in the US, we haven't found ourselves in such a situation . Perhaps France is getting closer to this. Well, how are you going to get out of this mess other than by printing money? The only problem with the French is that they don't manage their own money, they don't print it.
This is what the ECB does. But in the US, let me give you an example: after World War II, the debt- to-GDP ratio in the US exceeded 100% during the war. Within two years of the war's end, inflation was so high that it wiped out 40% of this debt in real terms. It was massive, it was the largest depreciation in the real value of the national debt that we have ever seen in the United States.
Now, no, this did not lead to the collapse of the US . This didn't cause a lot of problems. This led to an agreement between the Fed and the Treasury to somewhat shield the Fed from participating in such financing. Uh, well , yeah, I'll just leave it at that. Well, I'm sure all the executives in this room are hearing questions and concerns from people in companies and the economy about job security with the advent of AI.
You are pessimistic about how AI will impact everyday life. But speaking of work, I have to ask, do you feel your job is secure , that you are not threatened by an AI that could take over your role as a member of the Fed's rate-setting committee? However, there are those in the White House who are unhappy with the FOMC's work, claiming that it is going too hard on rate cuts and creating problems for the new chairman, Kevin Warsh.
Do you feel your job is safe ? Uh, as far as I know. I mean, in terms of AI, yes, you could create a whole AI-FOMC. I know some researchers at Stanford have already done this. They imitated each of us through our speeches and statements, creating their own AI agents. So in this model, there's an AI agent by Chris Waller that sets interest rates.
I do n't know how effective it is, but I... I was just amazed that it was even possible to do it. On other things, yes, I think there is a threat to the Fed's independence from the administration. So far, it hasn't been very successful. I think our institutions have made it clear that they don't necessarily want this to happen.
I hope it will continue to be like this . And the main defender is simply the bond market. I mean, if the bond markets are afraid that you're going to create a high inflation environment, those interest rates aren't going to come down. They will grow a lot. So there are good reasons why it is worth having an independent central bank.
And it worked very well for 40 years in almost all countries that implemented it. I just think it's a mistake to even think about giving it up. So it is the bond market that helps protect members of the Board of Governors like you. And who else? Congress? Well, Congress and the courts. Our institutions are designed to…you know, they are a defense mechanism in the way they operate.
We've seen this several times in the courts over the past year. The bond market is so obsessed with AI, I don't know. They do n't even seem to be paying attention half the time, but I have faith in American institutions—Congress and the courts—that they guarantee that we don't lose our independence and that the president of the United States doesn't set monetary policy.
Well , that's optimistic. Let's end on this note , as you can see AI is saving our jobs, and your job is safe too. Guests, please join me in thanking Governor Waller.
Comments 0