Keynote Remarks by Mary C. Daly at SIEPR’s Policy Forum — Transcript
Full transcript
- 0:00Perfect. I'm going to going to show some
- 0:01slides. I can prove that I'm an
- 0:03economist when I show slides.
- 0:06Uh, so thank you Neil for that warm
- 0:08introduction and thank it's really
- 0:10lovely to be back at CPER. It's one of
- 0:11my favorite places to speak. It's right
- 0:13here uh close by San Francisco Fed, but
- 0:16also the home of so many bright and
- 0:18young people in particular and I always
- 0:20enjoy coming to the Stanford campus. So
- 0:23today I want to talk about AI in the
- 0:25economy and I want to focus on the
- 0:27expectations which Neil gave us a little
- 0:29preview of, the reality and I mean the
- 0:32realities today and maybe the realities
- 0:33tomorrow which could be different and
- 0:35how monetary policy makers in particular
- 0:37have to think through this because it's
- 0:39not you know something that we can just
- 0:41imagine will be great. We have to
- 0:43actually investigate what's happening
- 0:44today. So let me start with the promises
- 0:48of AI, the AI expectations. So this is a
- 0:51simple chart of or graph of labor
- 0:53productivity growth, 10-year moving
- 0:55average over the last you know century
- 0:58or more and it really shows that
- 1:01electrification was a big boom to
- 1:04productivity growth for a long time. The
- 1:06internet was a also a durable boom, but
- 1:09not quite as large as electrification
- 1:11and if you listen to most of the
- 1:12enthusiasts you know even three or four
- 1:14years ago, they were saying AI's impact
- 1:17is going to supersede or be greater than
- 1:19electrification and right now it doesn't
- 1:22look quite there. So the question is
- 1:24where are we going and Neil put this
- 1:26well. We all expect this to be terrific
- 1:30and transformative and disruptive and
- 1:32all the things in that category, but we
- 1:34don't know exactly when this will
- 1:35happen. So what I want to do is bring
- 1:38that vast question into a narrow window
- 1:41where monetary policy must look. So let
- 1:43me start with how we think about this,
- 1:46how I think about it as a monetary
- 1:47policy maker, also as an economist I
- 1:49think is that we have to focus on
- 1:51economic growth. There's four lenses,
- 1:53economic growth, productivity,
- 1:56inflation and the labor market and of
- 1:58course inflation and labor market's
- 2:00important for the economy, but it's also
- 2:02our dual mandate responsibilities at the
- 2:05Federal Reserve. So now I'm going to
- 2:06look through those four lenses and I'm
- 2:08going to think about what's happening
- 2:10today. What's What do we have so far? So
- 2:13I want to start with growth. These are
- 2:15I'm going to put two charts together
- 2:17here to to say one basic point. We have
- 2:20a technology-driven growth boom.
- 2:23The left-hand side of the charts shows
- 2:25real GDP growth, four-quarter change,
- 2:28and it shows that the blue line that we
- 2:30are growing 2.1% over that 12-month
- 2:33change. That is a really good growth
- 2:35rate right now, but if you took
- 2:36technology goods out of that equation,
- 2:39we'd be growing at 0.6%. So when you
- 2:41hear people say this is a technology-led
- 2:43boom, they're absolutely right. These
- 2:45are the the points for it. So then
- 2:47people get very nervous, and of course,
- 2:49we all want to think about this, about
- 2:51how long can this last? What really is
- 2:54going on? And especially when there's
- 2:55been some
- 2:57market volatility on AI valuations, the
- 2:59natural question is, is this something
- 3:01we can expect to continue? So I There's
- 3:03many ways I could show this, but I like
- 3:05to show this data center chart. And the
- 3:07reason I like to show the data center
- 3:09chart is it shows that you're having a a
- 3:11general upward trajectory. We have
- 3:13things already in play. They're under
- 3:15construction. That's the green bars in
- 3:1726, and then we have lots of things
- 3:19planned in the pipeline. So even if we
- 3:21get a pullback of data center enthusiasm
- 3:24or communities say, wait a minute, let's
- 3:26slow this train down, you still see
- 3:28there's so much in the pipeline that you
- 3:30can expect this to continue.
- 3:32Now data centers themselves are not
- 3:34They're just large buildings. The thing
- 3:36that really fuels the AI boom is the
- 3:39materials and goods and software that
- 3:41goes in the data centers. And so this is
- 3:43a small starting point, but you can just
- 3:46multiply that and you get something that
- 3:48looks like the picture on the left. So
- 3:50ultimately, when we think about growth,
- 3:52there's not only the growth we see, but
- 3:54some general appreciation for this
- 3:57growth will continue even if it's not at
- 3:59the same pace.
- 4:01So then let's look at productivity.
- 4:03So, productivity evidence or
- 4:05productivity growth evidence is mixed at
- 4:07this point. Let me tell you what I mean.
- 4:10So, if you think about labor
- 4:12productivity in 23 24 25 and 26 which
- 4:15are the middle uh set of bars, well then
- 4:18you see that it's it's above where we
- 4:20were in the previous decade. The
- 4:23historical average is roughly rule of
- 4:25thumb 1.5 %. That's what we had in the
- 4:27previous decade, but if you look at what
- 4:29we've been experiencing, it's much
- 4:31faster than that. So, there's no doubt
- 4:33that labor productivity growth is is is
- 4:36quick right now and that's a good thing
- 4:37for our economy. It means the economy
- 4:39expands faster, people can do uh the
- 4:42output grows. Those are all good things.
- 4:45But if you really want to think about
- 4:46the impact of AI and whether to
- 4:48attribute the labor productivity growth
- 4:50we've been seeing to AI and whether the
- 4:52transformation has begun, you can't just
- 4:55look at the blue bars. You have to look
- 4:57at what economists call total factor
- 4:59productivity. I think what people call
- 5:01the magic behind productivity growth,
- 5:03right? It's the the part of productivity
- 5:05growth that has that sustainable durable
- 5:08part. It's the part about
- 5:09electrification that it wasn't just you
- 5:12could turn on a light. It's that it
- 5:13transformed everything we knew about how
- 5:15to live and work in the US. So, that's
- 5:18the piece that has been less than
- 5:20positive. If you look at TFP, it was
- 5:22really strong for a little bit in 23 and
- 5:2524, but now it's it's at average or a
- 5:27little below average and so we haven't
- 5:30seen the magic there yet that is what
- 5:33the AI tools promise. That's quite
- 5:36consistent with the survey of of firms
- 5:39on the right. This is a survey by Nick
- 5:41Bloom and his team and it just basically
- 5:43shows that it over the past 3 years the
- 5:46firms that they survey haven't really
- 5:49said that they had an impact of AI on
- 5:51their productivity growth. There's some
- 5:53of course that have had a positive
- 5:54impact. If you go to certain industries,
- 5:56you'll see a positive impact. You can
- 5:57certainly find case studies. You can
- 5:59certainly find evidence of green shoots
- 6:02forming, but overall in the whole
- 6:04economy just haven't seen.
- 6:06Right now the big impact that many
- 6:09expect, including me.
- 6:11So now let's look at the the next thing
- 6:13about price pressures. I said I was
- 6:14going to look through the lens of the
- 6:15dual mandate. Let's start with
- 6:17inflation.
- 6:18There is no doubt that AI is fueling
- 6:21price pressures. You can see this in the
- 6:23left hand chart. I think that the left
- 6:25hand chart is particularly remarkable
- 6:27because it shows that after you know two
- 6:30decades, three decades of disinflation,
- 6:33deflation in technology goods, we now
- 6:36have 15.7% almost 16% inflation in these
- 6:40goods. And that is a huge reversal of
- 6:44price changes for a an important part of
- 6:46our economy. And it really is reflecting
- 6:49the demand for those goods relative to
- 6:51the supply.
- 6:53So then you want to ask if you're a
- 6:54policy maker, I think or all of us, how
- 6:57much of this relative price shock, this
- 6:59change in the relative price of
- 7:01technology goods is going to spill over
- 7:03widely into the economy. Well, I'm going
- 7:05to give you two ways to think about. The
- 7:07first one is in this chart. One is how
- 7:09much is it directly contributing to
- 7:12overall headline inflation. And you can
- 7:14see a modest amount, 0.2 percentage
- 7:17point onto the total inflation. That
- 7:20just tells you that right now headline
- 7:22inflation is being driven by other
- 7:24variables, not the AI goods.
- 7:26But we're also starting to hear and I'm
- 7:28sure you are and you probably
- 7:30experienced this if you bought a
- 7:31computer or a phone etc. The price of
- 7:33technology equipment is not just
- 7:35narrowly focused on those going up for
- 7:37data centers. It's going up everywhere.
- 7:40And so certainly firms are feeling that
- 7:42cost. The question is is are those
- 7:44indirect costs rising so sufficiently
- 7:47that we're going to see it in overall
- 7:48inflation? And so far we're not hearing
- 7:50that.
- 7:51Really hearing it's it's more directly
- 7:53associated with
- 7:55just that sector. Let me get my water
- 7:57just in case I need it.
- 8:00Okay. So then let's go to the labor
- 8:01market, which I think is really an
- 8:03important question that people have been
- 8:05asking.
- 8:06And so far we've seen a I would call a
- 8:08limited impact on the labor market. Now,
- 8:10that is not to say that there aren't
- 8:12impacted individuals.
- 8:14And we'd be uh remiss if we didn't say
- 8:17that there's certainly individuals who
- 8:19are experiencing that. So that is the
- 8:21case. But when you look through the
- 8:23aggregate economy, we are seeing a
- 8:25limited impact of AI at scale. You can
- 8:28see this in two ways. I like to show
- 8:30this hiring rate and separation rate,
- 8:32layoffs and and and quits. You just see
- 8:35that right now we're in a low hiring,
- 8:37low firing environment. You've probably
- 8:39heard that that phrase before. Low
- 8:41hiring, low firing. Firms aren't hiring
- 8:43a lot of people, but they're also not
- 8:45letting go of a lot of people. That
- 8:47doesn't mean certain firms aren't laying
- 8:48other off, but it does mean that there's
- 8:50a little bit of hiring, there's a little
- 8:51bit of separation, but mostly you're
- 8:54just in a low hiring, low firing
- 8:55environment.
- 8:57That means that if you look to the
- 8:58right-hand side of the chart, that the
- 9:00unemployment rate has remained really
- 9:02stable and at a level that's consistent
- 9:04with the natural rate of of unemployment
- 9:06that we expect, the one that just
- 9:09incorporates churning and other things,
- 9:10job search, people taking time to get
- 9:12from one place to another. But right
- 9:14now, this is labor market that's fairly
- 9:16stable and you haven't seen a big impact
- 9:19of AI on the labor market.
- 9:22So when I put all of this together, the
- 9:24the the point for monetary policy is
- 9:27we've got an inflation problem.
- 9:29Uh we have a labor market that's stable,
- 9:31growth that's good, but we definitely
- 9:33have an inflation problem. Right now, we
- 9:35don't have AI doing most of the driving
- 9:38of inflation, although there is a risk
- 9:39that it could contribute more. And we
- 9:42also don't have productivity growth in
- 9:44the TFP world, that magic part, in a way
- 9:47that's picking up steam. But you
- 9:48definitely see labor productivity
- 9:50helping us out.
- 9:52So then, what do I think about for the
- 9:54next several years that maybe the one to
- 9:57three uh impact? So now I'm going to
- 9:59look one to three years out, not that's
- 10:01the review of today, the health check on
- 10:03the economy. Now, what about one to
- 10:05three years out? What's possible? You
- 10:08can think of these as risks, you can
- 10:09think of these as scenarios, but
- 10:11basically I'm going to look through the
- 10:12four same four lenses, economic growth,
- 10:15productivity, inflation, labor market,
- 10:17and ask what's what's coming.
- 10:21So let's start with AI valuation
- 10:23movement. I think most of you read the
- 10:25news and you know that AI valuations
- 10:27have been moving around a lot. And I
- 10:30this chart on the left is really the
- 10:32meant to show that. The Magnificent
- 10:34Seven have dropped uh
- 10:37a lot
- 10:38relative to the S&P 493 and that's
- 10:42something that everybody's noticing. And
- 10:44there's a lot of articles out there
- 10:46about is this the the beginnings of the
- 10:48of the end, the unwinding of AI. And I
- 10:51wanted to put some perspective on it
- 10:53because I think of it
- 10:55I think context matters. So if you look
- 10:57next door on the right hand side of this
- 10:59chart, you see that yes, there has been
- 11:01some repricing, but it hasn't eroded the
- 11:05considerable gains that have been going
- 11:07on for the last three years. So yes,
- 11:10there's a repricing. No, it doesn't seem
- 11:13to be the unwinding of the technology,
- 11:16interest and enthusiasm and investment
- 11:19that we've been seeing.
- 11:20So how do I read this in terms of
- 11:22thinking about growth? I read this as
- 11:25the private sector, investors, all of
- 11:27the market participants, many of you in
- 11:29all likelihood, are actually asking good
- 11:32questions like what do we think this do
- 11:35we think the investment strategy that
- 11:37these companies are using, the borrowing
- 11:39strategy, etc., is going to be as
- 11:42risk-free as it seemed earlier on? Are
- 11:45we going to price in a little more risk?
- 11:47That is a good thing. I was on a podcast
- 11:49with Steve Davis and I said, you know,
- 11:51when the private he and I ended up
- 11:52talking about this, when the private
- 11:54sector responds, this creates more
- 11:56opportunity for there to be an orderly
- 11:59valuation to right-size
- 12:02uh returns with expectations. And so
- 12:04this is this is somewhat good news. But
- 12:07the second part of the story is that the
- 12:10underlying value
- 12:12isn't being eroded and ultimately that
- 12:15is consistent with there being some
- 12:16there there. You know, as Neil said,
- 12:18everybody expects this to be
- 12:20transformative. The question is how much
- 12:23and when? And you put those two things
- 12:25together and that single belief that
- 12:27this will be a transformative
- 12:28technology, it's useful to invest in,
- 12:31useful to think about, isn't going away.
- 12:33And that is contributing to some of the
- 12:36projections you see for growth and some
- 12:39of our own sense, my own sense, that
- 12:41growth will continue. There's not a risk
- 12:43to growth here.
- 12:46So what about um productivity?
- 12:49So if you stood there, you know, some
- 12:51people have been tempted to say, well,
- 12:52look, there's nothing happening yet. And
- 12:54I think yet is the important piece. We
- 12:57can't extrapolate from what's happened
- 12:58so far to talk about what's going to
- 13:01happen in the next 1 to 3 years. And one
- 13:03of the reasons is that the share of
- 13:05firms even using AI, just becoming
- 13:08familiar with it, employing it, thinking
- 13:11about it is growing. And I suspect that
- 13:14if firms fully understood what they were
- 13:16doing when they bought, you know, one of
- 13:17the big vendor programs and then AI is
- 13:19empowering a lot of this, you might even
- 13:21get a higher number. You know, people
- 13:22are turning on the tools, the AI tools
- 13:24in normal vendor programs they use for
- 13:27HR and payroll, etc., but in accounting,
- 13:30but they're also just getting instances
- 13:32of these models themselves and thinking
- 13:34about what they could do with it. So
- 13:35it's growing.
- 13:37Importantly, if we return to the survey
- 13:38that Nick Bloom runs, we see that if
- 13:41they look out, not past what happened in
- 13:43the last 3 years, but what do they
- 13:44expect to happen in the next 3 years,
- 13:46firms are much more enthusiastic about
- 13:49how it can have a positive impact on
- 13:51their um on their businesses. We had a
- 13:54round table
- 13:55recently of businesses. It was uh you
- 13:58know, just a panel of businesses, big
- 14:00businesses, global businesses. And we I
- 14:02asked them the speed question at the
- 14:04end, which is, do you expect
- 14:06productivity gains to be start to see
- 14:09transformation in 1 to 3 years or more
- 14:12like 5 to 10 years? And each person said
- 14:151 to
- 14:16because we've done all the hard work of
- 14:18investing, but now we're driving for
- 14:20ROI. And, you know, as one of the
- 14:22contacts said, when firms start driving
- 14:24for ROI, things happen fast. So I think
- 14:27this is a positive momentum for
- 14:29productivity growth. I can't say we're
- 14:30going to have it in 1 to 3 years, but I
- 14:33can say there's a lot of uh work being
- 14:35done to get it in 1 to 3 years.
- 14:40So
- 14:41you might worry that this inflation will
- 14:42never be bridled, that we just can't
- 14:44catch up fast enough and that this is
- 14:47something that the Federal Reserve and
- 14:49central banks across the globe are
- 14:50really going to have to deal with. And
- 14:52definitely we have to think about this
- 14:54as a risk and we have to deal with it.
- 14:55But the Federal Reserve, central banks
- 14:57are not the only game in town. One of
- 15:00the things you really rely on is the
- 15:01private sector of firms to say,
- 15:04wait a minute, maybe things are too
- 15:06costly. And so this chart short of
- 15:09reassuring in that way. So as new models
- 15:12became available, open weight, not just
- 15:14the closed weight models, as open weight
- 15:16models became more available, firms
- 15:18started using those. And this is a share
- 15:20of tokens processed
- 15:22on open router using open weight models
- 15:25versus closed weight models. And you can
- 15:27see that open weight models have gained
- 15:29a lot of ground. The share has increased
- 15:31quite a lot. They're much less
- 15:33expensive. And that's actually
- 15:36been correlated with the reduction in
- 15:38token prices for the OpenAI and
- 15:41Anthropic. So the private sector market
- 15:44seems to still be working. You have less
- 15:46demand for your service and you shift
- 15:48gears in terms of your pricing. Another
- 15:50thing we're hearing about is that firms
- 15:52are actively um think about memory and
- 15:55chips, chips, you know, regular chips
- 15:57and memory chips. Firms are paying
- 16:00higher prices, regular firms, not tech
- 16:02firms. They're paying higher prices as
- 16:04well for these chips. We're hearing more
- 16:06and more. The companies are saying, is
- 16:08it really true that the chip depreciates
- 16:10every 2 years or can I get more useful
- 16:12life out of this? Or maybe I can remodel
- 16:15my product line to use fewer chips or
- 16:17older chips and and not have to rely on
- 16:20this, you know, every two years I have
- 16:21to buy new and compete with the highest
- 16:23prices. So that's another what I would
- 16:25say competitive marketplace doing some
- 16:27of the work of disciplining this higher
- 16:29inflation. Doesn't mean that the Fed
- 16:31isn't can't think doesn't have to think
- 16:33about this and it's all going to work
- 16:35out beautifully, but I think an
- 16:36underappreciated point is that the
- 16:39private sector is responding to the ever
- 16:41rising price of these items and you're
- 16:43seeing prices start to respond.
- 16:48The final then thing I want to say
- 16:49before I turn to a little bit about what
- 16:50this means for policy is that
- 16:53the labor market is one of those places
- 16:55you can't extrapolate from either. You
- 16:57can't say nothing's happened so far, so
- 16:59nothing will happen. And I don't think
- 17:01we can also say that the worst will
- 17:03happen. So it's it's a very human
- 17:05characteristic to say to extrapolate
- 17:07from what's happening or what you fear
- 17:09is happening and say the worst will
- 17:10happen or the best will happen.
- 17:12Somewhere in the middle is what's likely
- 17:14to happen. Right now we see evidence
- 17:16that younger workers in particular, but
- 17:18mid-career workers as well are worried.
- 17:20This is a study out of Anthropic. They
- 17:23did a survey. You do see that workers
- 17:25are more worried at the younger part of
- 17:27the age distribution and earlier in
- 17:29career than at the more senior levels.
- 17:32You also see that unemployment rates for
- 17:34college recent college graduates are
- 17:36rising relative to the unemployment rate
- 17:38unemployment rate for all workers. And
- 17:40if I expanded this
- 17:43time frame, you'd see that this is very
- 17:44unusual. We haven't seen this in the
- 17:46United States for a long, long time. And
- 17:49it just speaks to the idea that, you
- 17:51know, the AI is changing the dynamic for
- 17:54younger workers. Now we spend a lot of
- 17:57time asking our our teams our our CEOs,
- 18:00our round tables, why is this? And I'll
- 18:03give you something that maybe we can we
- 18:05can think through during the course of
- 18:07the Q&A and that is basically maybe
- 18:09what's happening is they're waiting
- 18:11until they know who they want to hire.
- 18:13But if you're a younger worker, that
- 18:15doesn't feel very reassuring. So I think
- 18:17that is something we have to think
- 18:18about. So if now if I back up and ask,
- 18:20what does all this mean for policy?
- 18:23Uh, and I'll conclude my remarks there.
- 18:26So for monetary policy, we have to look
- 18:28through maybe really a one to three-year
- 18:30window. So right now, what we have, and
- 18:33I said this a moment ago, what we have
- 18:35is a stable labor market, really
- 18:37positive growth, some positive
- 18:39productivity growth for labor
- 18:40productivity, and an inflation rate
- 18:42that's been above target for six years.
- 18:44And we have to focus on price stability
- 18:47and bring that inflation down. It's not
- 18:49primarily driven by AI, but the AI price
- 18:53surges don't help our situation. So
- 18:55that's where the focus has to be.
- 18:57The second part of this though, what
- 18:59happens in the next one to three years
- 19:01just leads me to say this. Just because
- 19:03we have to fight the inflation problem
- 19:05now doesn't mean that we should
- 19:08completely ignore the possibilities on
- 19:10productivity growth, on how that could
- 19:12be disinflationary.
- 19:14Some people say the inflationary impulse
- 19:16will continue. That's a possibility, but
- 19:18I think it's also possible, and maybe
- 19:20it's because I came to the Fed in '96
- 19:22and I saw what computers and the
- 19:23internet did to the possibilities in the
- 19:26economy and how that added to the
- 19:28disinflationary
- 19:30impulse in the economy. We have to be
- 19:31open-minded. So planning out what
- 19:34policymakers will do for the next one to
- 19:36three years is a is kind of a fool's
- 19:38errand. I can tell you that what
- 19:39policymakers should be, what I think I
- 19:41will what I know I want to be focused
- 19:43on, what we all have to be focused on is
- 19:45what can productivity growth do for us?
- 19:47How will AI lead that? When do we start
- 19:49to see the transformation? And what is
- 19:52going to be the impact on the labor
- 19:53market, not 10 years from now, but in
- 19:56the next one to three years? So like
- 19:58others, I'm extremely
- 20:01bullish or enthusiastic about what this
- 20:02technology will mean eventually, but the
- 20:05thing that I've learned through studying
- 20:06all types of things, global trade, other
- 20:09technologies, all the way back to the
- 20:11industrial
- 20:12revolution, is that for policymakers the
- 20:14transition dynamic matters. And for
- 20:17humans living in the country, the
- 20:18transition dynamic happens. And that's
- 20:20where if we get 10 years from now and we
- 20:23find that
- 20:24it hasn't worked the way we wanted to, I
- 20:27will offer this. It won't be the
- 20:28technology's fault. It will be ours to
- 20:31the side. So intentional policy making,
- 20:33conversations like the one we have here
- 20:35going on today, these are what's
- 20:37essential to making sure we're proud of
- 20:3810 years from now as opposed to
- 20:40wondering why this happened.
- 20:42With that I'll hope that I'll be glad
- 20:44happy to talk to Neil.
- 20:45>> [applause]
- 20:50>> Amazing. I think this is uh a terrific
- 20:53presentation. Slides are really
- 20:55handsome. I I love the color scheme. I
- 20:57don't know if that's the the Fed's like
- 21:00uh
- 21:01schema.
- 21:02>> Francisco Fed's schema.
- 21:04>> It's it's great. Um
- 21:06>> We have a We have 12 reserve banks and a
- 21:07board of governors. Everybody has their
- 21:09own schema. That's ours.
- 21:10>> they're all equally good, but this one's
- 21:12the best. Um
- 21:13>> That's what I like to hear.
- 21:15>> So
- 21:16you set me up really well. Maybe I'm
- 21:18going to ask a couple questions which
- 21:20delve into the the topics you teed up.
- 21:24And then uh I have questions coming in
- 21:27uh on my iPad and so we can turn to
- 21:29those as well. And you can use the QR
- 21:32code to submit questions. Uh
- 21:36Let me
- 21:37Uh
- 21:40[snorts] So let's start with uh some
- 21:43questions you know that
- 21:46uh relate to the dual mandate and I
- 21:48think especially the inflation side,
- 21:50which is uh you know where the concerns
- 21:54are most heavy. So
- 21:57inflation inflation has been above
- 21:59target. You said six years?
- 22:02>> Almost six years.
- 22:03>> Almost six years. And you know it's
- 22:06it's probably worse than moving sideways
- 22:07now. It's it's moving in the wrong
- 22:09direction.
- 22:10Uh
- 22:11Our textbook
- 22:13sort of approach to monetary policy says
- 22:17we should look through a supply shock,
- 22:20right? And much of the inflation we've
- 22:23seen recently is supply shocks, tariffs,
- 22:27uh the war and energy prices.
- 22:29Uh
- 22:30Is that the way you think about the
- 22:33supply shocks we've been facing? That
- 22:35it's something that we should be looking
- 22:36through because a supply shock weighs on
- 22:39both sides of the mandate, right? It
- 22:41reduces growth and increases inflation.
- 22:43So if your job is to balance the two,
- 22:46then uh something that keeps these in
- 22:49balance uh is not something we should be
- 22:51over responding to. So is is that how
- 22:53you think about it?
- 22:54>> You know, I think that is the it's
- 22:55definitely the conventional wisdom as
- 22:57you said that that we would look through
- 22:59these types of shocks because it's also
- 23:01by the the idea that by the time our
- 23:04monetary policy tools would take effect,
- 23:06which is, you know, 12 to 18 months, the
- 23:08shock could be dissipated and then we
- 23:10would have overcorrected something that
- 23:12we didn't need to respond to at all or
- 23:15couldn't really affect. You can't affect
- 23:16the price of oil, but if you raise
- 23:18interest rates in response to the price
- 23:20of oil, this is the textbook example,
- 23:22you end up curbing an economy that's
- 23:23already struggling with lower growth.
- 23:25Okay, so that's that's the conventional
- 23:27wisdom.
- 23:28You know, the way I've been thinking
- 23:30about it since maybe the middle of the
- 23:32summer is a little bit different than
- 23:34that, which is that you know, we have
- 23:36inflation. We already start with an
- 23:37elevated level of inflation. It's
- 23:39already above target. Then we have three
- 23:41shocks in overlapping succession, right?
- 23:44So they're not just behind each other,
- 23:46they're actually overlapping. First we
- 23:48have tariffs, then we have the war and
- 23:50oil prices, energy prices, which is
- 23:52affected food prices, and then we have
- 23:54the AI shock. And I think it's a
- 23:56reasonable question for all all to ask
- 23:58and really investigate, do the fact that
- 24:00these shocks overlap and that they
- 24:02already come at a period of high
- 24:03inflation, do they compound on each
- 24:06other and start to change the underlying
- 24:08rate of inflation that really matters
- 24:11for the durability of inflation
- 24:12expectations, etc.? Our best check on
- 24:14that is inflation expectations. And at
- 24:18the medium and longer run, you haven't
- 24:20seen inflation expectations of markets,
- 24:22consumers, uh businesses drift. They're
- 24:24still consistent with 2% inflation, but
- 24:27we can't just be complacent and say well
- 24:30businesses and everybody's happy, so we
- 24:31shouldn't have to do anything. The way
- 24:33you keep inflation expectations well
- 24:35anchored is to bring inflation down to
- 24:37price stability. So for me, I was quite
- 24:40comfortable with the September rate
- 24:42increase. I think that's appropriate.
- 24:44And with, you know, my bias, as I've
- 24:46been saying, my bias is that further
- 24:48tightening will be needed because you
- 24:49just see this consistent pressure on
- 24:52inflation that isn't simply in the
- 24:55sectors that are most directly affected
- 24:57by tariffs or oil. They're actually
- 24:59starting to bleed into the services and
- 25:02that services inflation is once it gets
- 25:05up, it's extremely sticky. It has a hard
- 25:07time coming down. So that's how I think
- 25:09about it now and I don't think we can
- 25:11just look to the conventional wisdom and
- 25:13and hope for the best. We have to reckon
- 25:15with the realities we face, which is
- 25:17that inflation's been high for a while
- 25:19and the pressure is up, not down.
- 25:21>> Great. And And I think that basically
- 25:23covers my my follow-up, which was going
- 25:25to be, you know, how do we modify the
- 25:27textbook model when we have more
- 25:30frequent, more persistent supply side
- 25:33shocks? And your response was well we
- 25:35focus on inflation expectations
- 25:38of both the market-based measures and
- 25:41the survey-based measures. And you're
- 25:43also thinking about uh you know,
- 25:46services inflation ex housing as
- 25:48something that I think you're
- 25:50particularly concerned about or tracking
- 25:52closely in the current environment.
- 25:54>> If I may, we also ask, I mean, I think
- 25:56this is really important, it may be
- 25:57helpful. We also spend a lot of time
- 25:59asking businesses, what are you doing in
- 26:02the next year? And increasingly we're
- 26:03hearing businesses say I I I'm at the
- 26:06end of my my line. I'm going to have to
- 26:08pass some of these input costs through
- 26:10or they're going to eat into margins.
- 26:11Ultimately the disciplining device of
- 26:13margins is competitors who will say I'll
- 26:15take the margin hit and sell my goods
- 26:18and consumers who tell you whether
- 26:19they're going to buy your things or not.
- 26:21And you know, it seems like consumers
- 26:23should have run run out of gas a while
- 26:24ago, but they haven't. So you know,
- 26:26that's the dynamic that I think is
- 26:28really critical to watch as we plan for
- 26:30the, you know, think about policy
- 26:32response to the inflation impulse.
- 26:34>> No, that's and that's that's a great
- 26:36addition. So, let me
- 26:38talk about a another factor here, which
- 26:41is, you know, financial conditions. And
- 26:44I think for people who follow the bond
- 26:45markets, right, the 10 years at, I think
- 26:49it's over 5% this morning, in fact.
- 26:51>> morning.
- 26:51>> Um and, you know, then we there's also a
- 26:54term premium. So, I
- 26:57I guess the specific question is, you
- 26:59know, to what extent do you think the
- 27:01long end of the yield curve is doing
- 27:04sort of additional work for you in uh
- 27:08tightening financial
- 27:09conditions and providing a sort of
- 27:11headwind, appropriate headwind for the
- 27:13economy? How do you think about those
- 27:15factors?
- 27:15>> So, tighter financial conditions are
- 27:17most welcomed, I'd say, if you're
- 27:19thinking about bringing inflation down
- 27:20because we have that have tighter
- 27:22financial conditions to help with that
- 27:24that process.
- 27:25The the reason I hesitate on this
- 27:28question all the time is that I don't
- 27:30want to say markets do the work for the
- 27:31Fed because the Fed is given the
- 27:33responsibility. Congress gave the
- 27:35Federal Reserve the responsibility of
- 27:37restoring price stability, keeping
- 27:38prices low and stable, making sure we
- 27:41maintain full employment. And so that's
- 27:42not something we just acquiesce. But I
- 27:44do look at the market conditions and ask
- 27:46a very simple question. Are they out of
- 27:49alignment with where I think we need to
- 27:51be? And right now I see they're in
- 27:53alignment with where we need to be. And
- 27:55importantly, they're reacting to
- 27:56incoming information in a way that's
- 27:58consistent with the reaction function
- 28:01that I have. So, when data come in and
- 28:03say the pressure on inflation is up, you
- 28:05see market pricing for the Fed's uh
- 28:08policy response go up. And you also see
- 28:11uh yields go up. When you see oil prices
- 28:14rise, you know, you see these types of
- 28:15things. And that is a good sign that the
- 28:17economy, I said at the beginning talking
- 28:20about AI, when the private sector
- 28:22participates in their role, that's
- 28:24actually helpful. When consumers
- 28:26understand where we're going, that's
- 28:27helpful. So, I'm I'm I think that gives
- 28:30us a little more time to make really
- 28:33calibrated decisions. You know, we don't
- 28:35have a surgical tool, and so you want to
- 28:37be thoughtful about it. We don't also
- 28:39have a tool where you can go up and
- 28:40down, up and down. You know, it's hard
- 28:42on on businesses, consumers, etc. So
- 28:46what markets can adjust tomorrow? If the
- 28:48world changes, they can just reprice.
- 28:50They can move hour by hour, certainly
- 28:52day by day or week by week. The Federal
- 28:54Reserve has You know, we make policy and
- 28:56it stays for a while or it usually stays
- 28:59for a while, and importantly people make
- 29:01decisions based on it. So I think the
- 29:03markets are assisting in making sure
- 29:06that we're bringing inflation down, but
- 29:09the job is ours. And so, you know,
- 29:11that's why I have said that that my bias
- 29:13is we'll have to have additional
- 29:15tightening, but exactly how much and
- 29:18exactly when, that's a decision that we
- 29:20don't need to make today because we have
- 29:22many more pieces and days of information
- 29:24coming in over many many quarters to
- 29:27make the right decision that helps do
- 29:29those two mandated goals.
- 29:32>> Fantastic. No, being data dependent and,
- 29:35you know, working I guess you said the
- 29:38the markets are complementing.
- 29:41>> They're complementing. And they seem to
- 29:42understand what we have said we would
- 29:45deliver, price stability, and they think
- 29:48they look out and see a labor market is
- 29:49stable. If you want to deliver price
- 29:51stability and a stable labor market with
- 29:53good growth, that means
- 29:55tightening policy.
- 29:57>> So let's maybe I'll ask you a few
- 29:58questions about AI and then we'll open
- 30:00things up. You know, you had the the
- 30:02chart you showed before of of GDP growth
- 30:05with and without sort of the I think it
- 30:07was the you stripped out CapEx, AI CapEx
- 30:10from it. Um so, you know, really high
- 30:12level, uh
- 30:14how many of our economic eggs are in the
- 30:16AI basket? And
- 30:18if there is a correction,
- 30:21uh you know, pullback in CapEx,
- 30:24a correction in the stock market,
- 30:25how exposed are are we? What are the
- 30:29What are the channels? How big are the
- 30:31risks? How are you thinking about, you
- 30:32know, this scenario? And obviously
- 30:33nobody has a crystal ball, but but we do
- 30:37scenario planning, right? It's a plan is
- 30:39better than no plan. Uh so how do you
- 30:41think about?
- 30:41>> So I I mean I I think about it that's
- 30:43why I I you know I tried to get at this
- 30:44question. It's a great question by the
- 30:46way. It's excellent question because you
- 30:48look at that chart on GDP growth and you
- 30:50think all of our eggs are in one basket.
- 30:52But it's not really the experience I
- 30:54have when I talk to CEOs in industries
- 30:56that are not technology driven. They're
- 30:58feeling somewhat constrained by the It
- 31:00just there's a one possibility is that
- 31:02everybody else is silent and sitting on
- 31:05the sidelines and AI is the only thing
- 31:06saving us. The other explanation which
- 31:08you'll hear from other businesses is
- 31:10everything's going up and so I'm kind of
- 31:12feeling like I can't do everything I
- 31:14want to do because prices are so high.
- 31:16So I I do think there's that balance of
- 31:18the economy and right now there's a lot
- 31:20of investment in technology not just by
- 31:22technology companies but by regular
- 31:24companies that are non-tech that want to
- 31:27get in part of this game. So I think
- 31:29that's there. The other thing that I
- 31:30think is useful to remember and that's
- 31:32why I showed the stock market
- 31:33>> I actually pause on that? Like is that
- 31:36the sort of flip side of the coin of
- 31:38elevated rates from AI uh investment or
- 31:43sort of crowding out certain things that
- 31:45if there was a pullback then we would
- 31:46see an increase in other investment and
- 31:48it would be a stabilizing force.
- 31:49>> It just it I mean I it's hard to say
- 31:51that you could replace all the AI
- 31:52investment which is in a surge period
- 31:55with other investment. But it isn't the
- 31:58it is the case that other businesses
- 32:00that are non-tech are investing. And
- 32:02when I hear uh businesses across the
- 32:05district and even across the country
- 32:06talk they're they want to invest more.
- 32:09They feel the opportunity of the US
- 32:10economy. But that can't be completely
- 32:13separated from the surge in growth from
- 32:15technology. It's really just a sense
- 32:17that maybe we're on the
- 32:19I have several people tell me we're on
- 32:20the brink of uh a manufacturing
- 32:24revolution in the United States. We're
- 32:25on the brink of bringing AI to the
- 32:27physical infrastructure that we do.
- 32:29Isn't this great? What can I do with it?
- 32:31But we're not there yet. I think the
- 32:33other thing that is important is and
- 32:35this is where that stock market
- 32:36valuation chart comes in. Yes, we've
- 32:39taken some of the the enthusiasm off,
- 32:42right? But it hasn't taken the
- 32:44technology away. And ultimately that's
- 32:46because there is there there. It may be
- 32:50it may take a while to see the full
- 32:52impact, but even
- 32:5410% of the impact, probably less of what
- 32:57people imagine could be possible down
- 32:59the road, still contributes to the
- 33:01economy and doesn't leave it um
- 33:03completely zero. So the worst case
- 33:05scenarios are if you go from all of this
- 33:07to zero. I don't have a lot of weight on
- 33:10the zero hypothesis. I think you could
- 33:13slow the growth rate and still have a
- 33:14robust economy growing and a stable
- 33:18labor market and lower inflation.
- 33:21>> And that's it's actually similar. I just
- 33:22sort of did some of these numbers. I
- 33:24know my colleagues have done too like
- 33:26run a reduction in CAPEX, run the wealth
- 33:30effect of a correction, you know,
- 33:32through basic models. And you you get a
- 33:35slowdown, but I don't think you get a
- 33:37reversal in in GDP growth. But, you
- 33:39know, then there's every, you know, the
- 33:42current sort of theory of recessions,
- 33:44right? Everyone is different and, you
- 33:46know, maybe circular financing or
- 33:48something else
- 33:49uh it could create some sort of
- 33:52more uh deeper slowdown.
- 33:54>> You're absolutely right. I mean, that's
- 33:56always there's always the risks out
- 33:57there. One of the things that that I I
- 33:59think was the test case for whether that
- 34:02scenario is we're right on the brink of
- 34:04it is we've had a sentiment shock,
- 34:06right? On AI. And it hasn't unraveled.
- 34:09It's been orderly. So, you know, it's
- 34:11just not that doesn't I never am a you
- 34:13can't be a central banker and never be
- 34:14complacent. You can't say, oh, because I
- 34:16saw this piece of data, now I feel
- 34:17great. But I do think, you know, looking
- 34:19for signs of what the likely scenarios
- 34:22are are are what's playing out. You
- 34:25don't see really burgeoning signs of of
- 34:28fragility there.
- 34:29>> Right. So let me ask a couple labor
- 34:31market questions and we'll open things
- 34:32up. Um
- 34:34first one I want to go back to, you
- 34:36know, you had uh a slide on increasing
- 34:40unemployment rates for young college
- 34:42graduates. Uh, and I think to to
- 34:46summarize your slides and and I think
- 34:48broadly held views, you know, we're in a
- 34:50a low hire, low fire labor market where
- 34:54the most stress is, you know, on young
- 34:57college graduates, you know, including
- 34:59lots of students that, you know, I talk
- 35:02to, I I mentor. Uh,
- 35:04and
- 35:05there's a host of of theories out there
- 35:08for what's going on from work from home.
- 35:11I think Nick Bloom has talked about how
- 35:13it's harder to monitor a and uh
- 35:17and coach a young employee if you're
- 35:20more work from home uh to higher
- 35:23interest rates in a firms are feeling
- 35:25more pressure. This is where pulling
- 35:26back to AI,
- 35:29uh right that number of the tasks that
- 35:31uh young associates would do right now
- 35:34AI is doing.
- 35:35Uh,
- 35:37what is your review as to the different
- 35:40factors here? What are you giving the
- 35:41most weight? Uh,
- 35:43how concerned should we be?
- 35:44>> so I've done I've been a labor economist
- 35:45my whole career and one of the things
- 35:47I've learned in that period is there's
- 35:49no one factor that if you wait it you
- 35:51won't regret it the next day uh because
- 35:53ultimately all of these factors are
- 35:54playing out and probably the combination
- 35:56of them, right? Right now we have firms
- 35:58that feel cautiously optimistic,
- 36:00underlying cautiously, right? So they
- 36:02see this, but if you want to be if
- 36:04you're cautious, you're uncertain, you
- 36:06certainly don't want to get over your
- 36:08skis so to speak by hiring a lot of
- 36:09individuals that you then have to
- 36:11painfully, you know,
- 36:13let go. So that's one piece. The second
- 36:15piece is AI, you're spending so much
- 36:18time and energy thinking about AI and
- 36:21you really need people who can evaluate
- 36:23whether you're getting the right answers
- 36:25and newly minted people who just started
- 36:27your firm are unlikely to be those great
- 36:29evaluators, probably a little bit why
- 36:31the senior groups are feeling pretty
- 36:33okay. Like they're saying I I know that
- 36:35model's wrong, I know that output's
- 36:36wrong, but ultimately that's a
- 36:38transition dynamic. One we spend a lot
- 36:40of time because this is such a critical
- 36:42issue for the US and for policy making,
- 36:45uh, for the Federal Reserve. We spend a
- 36:47lot of time asking our, uh,
- 36:50vast network of employers,
- 36:52why are you hiring so little? And they
- 36:54give us a array of those answers, but
- 36:56one that you didn't mention that I think
- 36:58is potentially beneficial for younger
- 37:01workers in the transition is that they
- 37:03said we don't know who to hire.
- 37:05So here's an example. We're putting
- 37:07we're trying to rethink everything we do
- 37:09in accounting. If we make sure our
- 37:11customers are are getting the what they
- 37:13need, make sure we're getting what we
- 37:14need. We're revamping the entire
- 37:16accounting production chain for a a
- 37:20national company.
- 37:21If I hire a newly minted CPA and then I
- 37:24don't need as many newly minted CPAs, I
- 37:27need a CPA plus a person who does cyber
- 37:30or something, then I've hired the wrong
- 37:32person and I don't want to hire the
- 37:34wrong person because it's painful to let
- 37:36people go. So I'm going to go skinny.
- 37:38This is often how they say it. I'm going
- 37:39to go skinny until I figured that out.
- 37:41So then the next question we ask is when
- 37:43do you think you'll figure that out? And
- 37:45they say next year, so 27. So this was
- 37:48never going to be a year of rapid hiring
- 37:51because of the the different shocks that
- 37:53we've faced, the higher input costs,
- 37:55they're trying to get their budgets
- 37:56right. But I think that from the AI
- 37:58perspective, you know,
- 38:00it's very hard and this is maybe a theme
- 38:03of the whole conversation. It's
- 38:05extremely hard to take what we see today
- 38:07and extrapolate because if we
- 38:09extrapolate, we're likely to be wrong.
- 38:11And so instead it's like what could
- 38:13happen and what would what would change
- 38:15that equation? And I think right now
- 38:18we're in more or less a perfect storm
- 38:20for hiring. It's cautious,
- 38:23uh,
- 38:24probably trying to get some of the
- 38:25overhang from the post pandemic, you
- 38:27know, labor market hiring down, AI and
- 38:32younger workers are going to be the
- 38:33marginal
- 38:35tariffs,
- 38:36so much input, so many input costs
- 38:38rising and so it's just not going to be
- 38:40a great impetus for hiring. What has
- 38:42been
- 38:44noticeably, um, a sign of resilience in
- 38:47the labor market is there hasn't been a
- 38:48big increase in firing or separation. So
- 38:51I think that's the good news piece of
- 38:52this labor market.
- 38:54>> All right. So let me
- 38:56last question let's actually think
- 38:58through
- 38:59uh you know what may happen over the
- 39:01medium run in in the labor market. And
- 39:05there's two scenarios that I keep in my
- 39:07mind which I find
- 39:09uh clarifying, right? And like you said
- 39:11you know the truth is always a complex
- 39:12combination of different scenarios. So
- 39:14the sort of benign
- 39:17scenario I have for uh
- 39:20AI labor market disruption is
- 39:23you know we have
- 39:25uh
- 39:26job losses or you know
- 39:29changes in hiring mix that happen
- 39:32broadly across the country and happen
- 39:35slowly over time.
- 39:37And that sort of transition gets almost
- 39:41lost in the huge amount of churn we see
- 39:43in the labor market as it is, right? You
- 39:45know
- 39:46communities hire fewer people in this
- 39:49role, more people in this role uh to get
- 39:52lost in the churn and we can manage it.
- 39:53We have a very dynamic economy.
- 39:56You know the other scenario when I think
- 39:57about
- 39:59the hollowing out of the labor market
- 40:01and in particular the the loss of these
- 40:03sort of routine jobs are you think about
- 40:06Charlie Chapman in in modern times doing
- 40:08the same thing over and over and again.
- 40:10Like if you look in the data it's
- 40:12something like almost 90% of routine
- 40:14jobs were lost during recessions of the
- 40:17six months afterwards.
- 40:19When firms are cutting back and then
- 40:21they figure out how to do more with less
- 40:23uh and it's cheap to install capital
- 40:26because interest rates are low and wages
- 40:28are sticky so wages are high. Uh so you
- 40:30have
- 40:31job losses compounded
- 40:34structural job losses compounded with
- 40:37cyclical shocks.
- 40:38And I think that's the scenario I'm most
- 40:40worried about with AI that there is a
- 40:42recession and employers figure out how
- 40:45to do more with less and we don't hire
- 40:47back then the white collar service
- 40:50workers, and we have, you know, the
- 40:53longest jobless recovery that we've ever
- 40:56seen. So that's my worst case scenario.
- 40:58What what are the scenarios that you're
- 41:00thinking about?
- 41:02What do you think about the scenarios
- 41:03that I just laid out here? Uh, and then
- 41:05hopefully people will have uh more
- 41:07optimistic questions we can finish on.
- 41:09>> I I think those are two reasonable
- 41:12scenarios, but I think there's many more
- 41:13scenarios in the in between than I also
- 41:15think about. And and that's because, you
- 41:17know, ultimately we don't know, as you
- 41:19rightly said, those are two you know,
- 41:21kind of put them at the benchmark part.
- 41:23One thing that I will say since you
- 41:25brought up you a recession and this is
- 41:26when it happens, another time that you
- 41:28can see job transformation and employers
- 41:31really thinking about how do I automate
- 41:33is in response to a hot job market. So
- 41:36we had a very very tight labor market
- 41:40coming out of the pandemic, very tight.
- 41:42I mean, people were losing employees two
- 41:45weeks after they hired them or maybe
- 41:47they hired them and they didn't even
- 41:48show up. That's how remember ghosting
- 41:50and all this. So that's how tight it
- 41:52was.
- 41:53The a response to that type of tight
- 41:55labor market is I'm automating. I can't
- 41:58be left
- 42:00to
- 42:01wait for workers to come. I'm going to
- 42:02automate things I can. That happened to
- 42:05collide or correspond, I guess is a
- 42:07better word, with the AI boom. And so I
- 42:10think one of the reasons we're seeing a
- 42:12little bit of this right now without a
- 42:14recession, as you say, this is a
- 42:16classic, you know, historical example,
- 42:18but without that is because the firms
- 42:19already got the discipline of, you know,
- 42:21I better not rely on a labor force
- 42:24that's always available. And then you
- 42:26put onto that the fact that our labor
- 42:29force growth is really slow going
- 42:31forward in the absence of immigration it
- 42:33could get to zero by 2030 in terms So
- 42:35the employers like, okay, if that's what
- 42:38I've got, then how can this technology
- 42:40really help me continue output growth
- 42:44while I do more with fewer employees?
- 42:47And that's a response to those two
- 42:49things and I think that's a scenario
- 42:51that's probably top my mind right now is
- 42:54what is a transition dynamic look like
- 42:56when you have two accelerants, right?
- 42:58The slower labor force growth coming off
- 43:00a frothy labor market and you have the
- 43:03AI boom. How do employers combine those
- 43:06two things and how does that change the
- 43:08steady state demand for workers? But I
- 43:10always come back to something I was on
- 43:11stage with Patrick Collison of Stripe
- 43:13one about a two years ago at this point,
- 43:16but it's still it's still something he
- 43:17says. And he said, look, I'm hiring as
- 43:20many workers as I'm
- 43:22not as I'm letting go and the thing is
- 43:25they're just totally different people.
- 43:26So when you think about a labor market
- 43:28that is able to continue and thrive,
- 43:31it's about how we intentionally help all
- 43:34of our workforce learn the skills that
- 43:36are going to be necessary
- 43:38to embark on the new technology and the
- 43:40new jobs and the new tasks that come out
- 43:41of that and not you know, kind of stay
- 43:45behind on the old
- 43:47on the old
- 43:48like spreadsheet and say, well, that's a
- 43:50job, that's a job, that's a job. I think
- 43:52we're we're likely to see many many new
- 43:54jobs as we lose older jobs and I just
- 43:57want to make sure our workforce is
- 43:58ready.
- 43:59>> Yeah, no, that's that's a great scenario
- 44:02also to keep top of mind. Let me flip to
- 44:04questions and just seeing one from from
- 44:06Karen. I don't know oh, wait, there she
- 44:08is.
- 44:09She asks, what indicators will you be
- 44:12watching to assess whether AI indeed
- 44:14will support productivity and lower
- 44:16interest rates? And I think we didn't
- 44:17touch on that. So yeah, what what are
- 44:20the
- 44:21you know, we all know the productivity
- 44:22series is noisy and it's it's hard to
- 44:26just focus on that. What are you looking
- 44:27at?
- 44:28>> So you you know, Karen, thanks for the
- 44:29question.
- 44:31We talked we've talked about this a lot,
- 44:33but you know, one of the things that and
- 44:34I I was benefited by starting at the Fed
- 44:37in 1996. I was a a microeconomist, labor
- 44:39economist and then Chairman Greenspan
- 44:42would say, I think productivity growth
- 44:44is there and he'd set teams across the
- 44:46country to look for evidence or no
- 44:48evidence of this. And one of the things
- 44:51that I experienced was companies telling
- 44:54us that they were making these
- 44:55productivity gains and it was helping
- 44:57them protect margins while they
- 44:58continued to create output growth. And I
- 45:02the lesson I learned from that isn't the
- 45:03technology shocks are always great for
- 45:06productivity growth right away, but that
- 45:08you're never going to find it in the
- 45:09data until it's already happened. And so
- 45:12to be able to find it
- 45:15you got to look away from the published
- 45:16data on productivity growth, not to
- 45:18discount it. I think it's useful, but
- 45:20that is not going to give you a leading
- 45:22indicator of what's happening. To do to
- 45:24get that indicator, you have to talk to
- 45:26firms. You have to ask them what they're
- 45:28doing, how they're doing it, what's
- 45:29changed. The second thing that is really
- 45:32important that I look for is
- 45:34you know, I I go back I'm a I love
- 45:36history, so I go back I read about
- 45:37deeply about electrification. If you
- 45:39read any historical piece on
- 45:41electrification, they tell you that
- 45:43electricity is invented. It takes you
- 45:45know, 100 years to to see that transform
- 45:48the economy. The transformative part
- 45:51wasn't even when they put an electric
- 45:52motor at the end of the line to replace
- 45:54the steam engine. That saved money, but
- 45:56it did not change everything. What
- 45:58transformed everything is the invention
- 46:00of the unit motor and then you get a
- 46:02unit motor which can be moved anywhere
- 46:04and the unit motor itself isn't the
- 46:06transformative part. What's
- 46:07transformative is if businesses say,
- 46:09well, my gosh, if I don't have to have
- 46:10one continuous line, then I can move
- 46:13around, I can move machines, I can
- 46:15respond to output needs, I can do a
- 46:17variety of things and that transforms
- 46:20manufacturing. So that's the piece I
- 46:22look for. What are businesses doing to
- 46:24change not their technologies, but their
- 46:27processes? And when they change their
- 46:29processes, we will have the kind of boom
- 46:32that that people talk about.
- 46:35>> Uh so speaking of uh electrification, I
- 46:39have a question on
- 46:41uh energy in the here and now. So let me
- 46:43let me read it out maybe in a bridge
- 46:45form. Power hungry data centers are
- 46:47demanding uh immense amounts of energy.
- 46:50How's the Fed thinking about this demand
- 46:52reverberating through markets and
- 46:54contributing to broader long-term
- 46:55inflation in electricity markets, in
- 46:57other energy sectors, and in markets for
- 47:00energy sensitive goods.
- 47:03>> Now, we think about it a lot and and one
- 47:05of the things that we have to think
- 47:06about is what are communities doing in
- 47:07response. And, you know, we have we talk
- 47:10to we have round tables with the power
- 47:12companies. They tell us the same thing.
- 47:14It's hard to keep up with the demand for
- 47:16energy
- 47:17that communities have as we have growing
- 47:19communities, but data centers have. It
- 47:21was already hard to keep up and now we
- 47:22have a demand shock from from
- 47:25data centers. So, these are real
- 47:27problems. The infrastructure needs are
- 47:29present and the Fed can't solve those,
- 47:32but you're right. We have to think about
- 47:33them. So, one of the things I
- 47:35just to put a practical component on how
- 47:38I think of it as a policy maker is, you
- 47:40know, energy prices go up, but it really
- 47:43is all the people who pay for the energy
- 47:45that matters for the overall price
- 47:47inflation. If businesses are already
- 47:49paying higher prices for oil, now they
- 47:52pay higher prices for electricity.
- 47:54They're paying higher prices for other
- 47:55inputs due to tariffs or competing
- 47:58interest from AI, then this their input
- 48:00costs are going to rise and so that
- 48:02means their margins will fall unless
- 48:04they raise their prices. And so that's
- 48:06the dynamic you have to watch. So, we
- 48:08can't simply say, well, this is just
- 48:10electricity, so not a big deal. These
- 48:13are changing the price dynamics and if
- 48:15other prices don't adjust downward, then
- 48:18we're going to have a a higher pressure
- 48:21on the on the inflation rate than we
- 48:23would absent these shocks. And so this
- 48:26has got to be something we really think
- 48:27about. And, you know, if you really
- 48:29really geeky, people talk about this in
- 48:31the context of what's the long run
- 48:33interest rate that's going to be normal
- 48:35in an environment like this? The answer
- 48:37is nobody knows, but you can definitely
- 48:39put a table together and say things that
- 48:41move it up, things that move it down,
- 48:43and the the list of things that move it
- 48:45up are growing relative to the things
- 48:47that move it down.
- 48:48>> Yep.
- 48:49Um,
- 48:50we'll finish with with two questions
- 48:52sort of in the similar vein. First from
- 48:56Peter Blair here in the front. Oh, there
- 48:59are lots researchers and private sector
- 49:01firms in the room. What research and
- 49:03data would be most useful to the Fed and
- 49:05policy making in the age of AI uh to
- 49:08support the dual mandate? What's missing
- 49:10from the current research efforts of
- 49:12scholars and firms?
- 49:13>> So I'm going to I'm going to talk about
- 49:15something Peter, thanks for the
- 49:17question. I appreciate it. You know, one
- 49:19of the things I want to talk about on
- 49:20that which you mentioned is you know,
- 49:22the
- 49:23can and then I'll get to the AI part
- 49:24specifically, but the conventional
- 49:26wisdom in economics is that you can look
- 49:28through shocks, but in a world where the
- 49:31pressure on inflation we're going to be
- 49:33most central banks will be fighting
- 49:34inflation from above our target, pulling
- 49:36it down. And we've got these sequence of
- 49:38shocks, global shocks, national shocks
- 49:40that I think we just have to have a
- 49:42broader research hunger for how do I
- 49:45think of relative price shocks? What are
- 49:46the avenues that that it make them uh
- 49:50more durable, more sticky? How do they
- 49:52change the underlying level of inflation
- 49:55underlying rate of inflation? I think
- 49:56that's a ripe for this. On the AI
- 49:58specifically, you know, I've really
- 50:00appreciated all the research that's been
- 50:03done where you can't find it in the
- 50:05aggregate data. This is going to hook
- 50:07Karen's question to your question,
- 50:08Peter. Can't find it in the aggregate
- 50:10data, but we know a lot about where to
- 50:12look and all the studies that are
- 50:14looking at case studies and saying how
- 50:16does it happen? What's the mechanism
- 50:18that makes it just cost saving because
- 50:21AI is doing some work to transformative
- 50:24to your business? That is really great
- 50:26research to think about. And then the
- 50:28other part that I I really want would
- 50:30love the research community to embrace
- 50:32and think about is
- 50:35not what bad things can happen to the
- 50:37labor market eventually, but what can we
- 50:40do to prevent those things happening? I
- 50:44I do feel like it's it's easier
- 50:46oftentimes to itemize what could go
- 50:48wrong, but it's not that
- 50:51it's not an instinct of most humans
- 50:53really to say, well, what do we have to
- 50:55change in the models or in the data to
- 50:57have that go better? Like what would
- 50:59have to happen? That's not just a
- 51:01regulatory response, but it's actually
- 51:03like a training response. What would
- 51:05What how would it be different? And
- 51:06looking across the uh the globe would be
- 51:09helpful. Uh we had a a talk in our bank
- 51:11at the We have a Technology and Growth
- 51:13Center. We kicked it off a couple of
- 51:15weeks ago and Philip Lane from the ECB
- 51:18was giving a talk and he he talked about
- 51:20how different perceptions of AI are for
- 51:22workers here in the United States than
- 51:24they are in the EU. I think that's
- 51:26something to learn from. They're much
- 51:27less pessimistic about AI than our
- 51:29younger workers are. Well, why? And what
- 51:32what aspects of their countries and
- 51:34their policies and their programs do
- 51:35they have that might mitigate some of
- 51:38that fear?
- 51:38>> Yeah, you know, I I love that point. I
- 51:40you know, I've said to other people, we
- 51:41do too much predicting and not enough
- 51:44planning around AI.
- 51:45>> a great way to say it.
- 51:46>> Uh and you know, I think part of the
- 51:49uh
- 51:49purpose and you know, substance of
- 51:52today's event is to also do that sort of
- 51:55scenario planning and and start building
- 51:57that muscle ahead of when we need it. So
- 52:00let me finish with, you know, I think a
- 52:01very hard but important question. We
- 52:03have lots of young people, undergrads
- 52:06choosing majors, choosing where to get
- 52:08an internship, uh pre-docs, thinking
- 52:10about whether they're going to go to
- 52:12grad school in economics, go to the
- 52:14private sector.
- 52:15We're in a moment of huge uncertainty.
- 52:17Now what's your advice to to young
- 52:20people today?
- 52:21>> I'm happy to give it. Um and then you
- 52:23can tell me you can call me up 10 years
- 52:25from now and say that was terrible
- 52:27advice or that was excellent advice.
- 52:29>> you have the disclaimer, right? That you
- 52:31>> do have the disclaimer, but I'll I'll
- 52:33I'll take this on my on myself here and
- 52:36and say that, you know, I think the
- 52:38the instinct that many have when you see
- 52:41uncertainty is to stall out
- 52:43and do less. I think the instinct that I
- 52:46want you all to have is to lean in and
- 52:48do more.
- 52:49The economy's always going to be
- 52:51dynamic. You're you are coming out in a
- 52:54period of great uncertainty about what
- 52:57the world will look like, but we're not
- 52:58uncertain about where it is today. And
- 53:00today you know you have to have AI
- 53:02skills. You have to be able to
- 53:04interrogate things with great curiosity.
- 53:07I think there's a lot of human skills
- 53:10that AI hasn't been able to replicate
- 53:12yet that are going to be in high demand.
- 53:14In many ways, the scarcest skills are
- 53:16going to be leadership, uh, quality
- 53:18control, knowing when to use a model,
- 53:20when not to use it, knowing how to
- 53:22interrogate or investigate what is You
- 53:25know, you you started by talking about
- 53:26evidence-based. What's the standard for
- 53:28evidence-based? It's not that, you know,
- 53:30an LLM told you the answer. The standard
- 53:32for evidence-based is hypothesis
- 53:34testing, you know, having that
- 53:36curiosity, coming back with the data. So
- 53:39I think being positioning yourself so
- 53:41you're not mono-lined in a specific
- 53:43thing, but you're really open. You could
- 53:45be sitting
- 53:47right here at the beginnings of
- 53:49something that, you know, Neil and I
- 53:51will be jealous of because we weren't in
- 53:53the labor market. I don't think that's
- 53:55where you sit right now, but you want to
- 53:57be ready for that, right? You want to be
- 53:59ready for that. I also deeply believe
- 54:01that younger people are where all of our
- 54:04the idealism meets the the practicality.
- 54:07And if we want to look like a country
- 54:09and a a world we want 10 years from now,
- 54:11we're going to have to have each and
- 54:12every one of you participate. So I say
- 54:14lean in.
- 54:15>> Yeah. No,
- 54:17we all know as investors we should have
- 54:18a diversified portfolio. One thing I
- 54:20took from that comment is in terms of
- 54:22our human capital, right? We should lean
- 54:24in. We should be diversified, you know,
- 54:26not just the cognitive skills, the human
- 54:28skills,
- 54:29uh, and, you know, that leaves us best
- 54:31prepared for what opportunities are
- 54:34going to wait us in our uncertain
- 54:35future. Thank you so much.
- 54:37>> Can I say what one of my interns told me
- 54:38this year? He said, Mary, the thing that
- 54:39I learned by working at the Fed this
- 54:41year is that I'm not just what I do, I
- 54:44am how I do it. And I think if you take
- 54:46that to that out, you're going to find
- 54:49that how you do things matters as much
- 54:51as what you do. What you do can change
- 54:53every three years, certainly every 10
- 54:55years, but how you do it, how you come
- 54:57across in the labor market, how you work
- 54:59with others, how you collaborate, that's
- 55:00a durable skill.
- 55:01>> you you are a beacon in terms of what
- 55:04you do and how you you it. Uh, and it's
- 55:06been a delight to have you here. Thank
- 55:09you so much.
- 55:11>> Thank you.
About this transcript
This page contains the full transcript of Keynote Remarks by Mary C. Daly at SIEPR’s Policy Forum by Federal Reserve Bank of San Francisco, generated from the public captions YouTube serves with the video. The transcript has 10,393 words across 1,549 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.
What you can do with it
Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to choose a caption track available on the video.
Free YouTube transcript tool
YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, caption language selection, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.