Why AI Needs Humans

Jeffrey Roach [0:00] Hello, everyone. In this edition of the LPL Market Signals, Jeffrey Roach here, chief economist for LPL Financial. I am joined by Torsten Slok, partner chief economist at Apollo, and we're going to discuss a couple hot takes that we have heard from advisors really around the country, around the globe, and address some of those macro conditions. And at the end, we're going to tie it all together and say, okay, what does it mean? What does it mean with all these potential headwinds, tailwinds? What do we do about it? So we'll not get too heady, even though we are two PhD economists, that's not necessarily a good sale at this moment. I don't know, what do you think, Torsten?

Torsten Slok [0:39] No, I agree. [laughs]

Jeffrey Roach [0:41] Sometimes we don't highlight that. But Torsten is partner and chief economist at Apollo. He has been at Apollo since 2020, worked 15 years on the sell side, previously worked at the OECD in Paris in the money and finance division. Before that, was with the IMF, writing the world economic outlook. Very excited to have Torsten with us. I also studied University of Copenhagen and Princeton. I did find out that there was an overlap Torsten at Princeton on someone that we will discuss just in a little bit on the podcast. Mind sharing who that was?

Torsten Slok [1:22] Yeah. Well, I was there for one year in 95, 96. Ben Bernanke was the chair of the economics department. And he has, of course, become more important now with his savings glut and discussions that I'm sure we'll talk about in a minute here.

Jeffrey Roach [1:36] Well, that's right. So we have a couple hot takes and then we'll keep this under about 20-ish minutes. I want to spend some time on the so what of it all. And that's the key takeaway. So first hot topic, we get all the time, and that is this question. Is AI coming for your job, Torsten?

Torsten Slok [1:59] No. And the answer also is very important that it's not going to displace a dramatic amount of workers. It may have some impact on work processes. We may see over the next five years that some functions in financial world, in consulting, in legal services gets improved, gets automated. But that effect, which is negative on employment, is relatively small compared to the very significant positive effect on business creation.

If you look at the weekly data from the census for what is the number of new businesses that are created in the U.S. economy at the moment, you will see that today there has never been so many new businesses that are opening every single week. In other words, business creation is at the highest level ever in U.S. history. And if just a fraction of these new businesses create new jobs, of course, we should begin to see, and that's likely why the labor market continues to still do so well, we should begin to see that the positive effects of AI in the form of new businesses is going to outweigh the labor displacement effect, partly because the labor displacement effect comes much slower, but also because the labor displacement effect is also something that will be more complex and more challenging to implement for businesses.

Torsten Slok [3:15] So the short answer to your question is, I do not think that they'll come for my job at Apollo or your job at LPL because I still think that there will still be for investment professionals a significant need, especially as we all know on this podcast for relationships. A very common way of talking about this is that a lot of jobs are messy. Some of the tasks that you and I do, some of the tasks financial advisors do, they can be automated. But if we all do 20, 30 different tasks every day, then there's a lot of other tasks that cannot be automated, such as talking to clients, talking to management, talking to investors, talking, of course, ultimately to people in the investing world about what they should be doing.

So for that reason, I'm very, very optimistic that AI is actually both going to have a positive impact on productivity and it's also going to have a positive impact on employment because the displacement effect is being offset by business creation being so strong at the moment.

Jeffrey Roach [4:10] Yeah, and that's right. And that's one of the things that the new chair at the Federal Open Market Committee, Kevin Warsh has talked about, the benefits of productivity, how that influences the economy, and perhaps we get that boost from AI.

One of the things that I think our listeners should know if you're interested in reading a couple of really excellent articles on AI, thinking about it, what does it mean for the productivity side? What does it mean for the labor force? David Autor, A-U-T-O-R, I believe, does that sound right? David Autor, I think that's how you spell his last name, MIT, heard him speak at a economics conference. Excellent writer on explaining these things. So we do have a mixed bag of activities. You know, certain things are more routine, yes, that'll be automated.

Jeffrey Roach [5:03] Other things like the value of human judgment, clearly something that will allow the human side to flourish. So be more human is the answer, I think, to those nervous about the AI story.

You know, what do you think about the AI influence based on, almost based on your years of experience? You know, clearly you think about that uptick in job applications. I saw this right after COVID. Seemed like a lot of those 55 up folks that dropped out of the labor force wanted to go out and hang a shingle, right? So they had those uptick in the business applications. What's your take on the way that AI may impact someone with 15, 20 years work experience versus someone with two years work experience?

Torsten Slok [5:58] Yeah, no, it is important, of course, that there is a critical discussion exactly about this issue, namely that are entry level jobs being threatened by AI. So far, data from RAMP that you and I also have talked about is actually showing that companies that are adopting AI, they tend to have higher growth in entry level jobs. Maybe this is because companies that are adopting AI are generally higher growth companies and therefore generally grow faster. But at least up to this point, most of the evidence in the data, in particular from RAMP, but also some of the studies from the Fed have shown that entry level jobs is where most of that conversation sits, namely, do we need young workers or do we not need young workers? Yeah. The ultimate piece of evidence of this is, of course, the unemployment rate for people that are between 20 and 24 years old.

Torsten Slok [6:43] And the surprising news is that the unemployment rate for people that are between 20 and 24 years old has dropped more than the general unemployment rate. So yes, it may be that young people who come out of college cannot get the preferred job that they want, but they always actually end up getting a job. And at the moment, actually more of them have been getting jobs.

So yes, there's still some fears, and I also still worry, down the road, of course, about what are entry level jobs going to look like, but at least at this point, there is still, importantly, that high growth companies are having more entry level jobs, and at the same time, also the unemployment rate for young people is still significantly lower or has dropped more and lower than where it should be relative to the general unemployment. And it probably goes to the broader idea that if you and I came out of university and we had finished our degree and we couldn't find a job, it is now easier than ever before to start a new business.

Torsten Slok [7:34] There's a lot of dorm room entrepreneurs, there are even anecdotes about people dropping out of college to take great jobs doing AI. And if that's the case, that's of course a very dynamic labor market for people that are coming into the labor market. And I do think that this is a key reason why the latest non-farm payrolls has continued to be so strong, and it's a key reason why the unemployment rate continues to be so low at 4.1, namely because the labor market, especially for entry level jobs, is just much more dynamic now that people can sit at home in their basement or their parents' basement and do agents, graphs, loops, and simply build a business and get a job easier than before where you had to find a well-established firm in legal services, in consulting, or in finance.

Jeffrey Roach [8:15] Yeah, that's right. And so anyone who's in college listening to the podcast or our audience that knows anyone in college, I'll leave this hot take here with a book recommendation. I think it's highly relevant. And the title is Range. The subtitle is Why Generalists Flourish in a Specialized World. I think what this — Oh, very interesting. Book does, it's a fascinating read, and what it really highlights is the need for folks that are going through their training to think about cross-disciplinary research, right? Don't be a deep specialist—

Torsten Slok [8:55] That's really important. Because we all know that the main occupation that's been threatened by AI is, of course, coding and software programmers. And software, of course, programmers are characterized by they have one task and one task only, namely coding and programming software, which is different from someone who is a generalist that comes in and does sales or does marketing or does a back office or legal services where you have more different tasks.

So the more you have just one task, in this case, a programming or coding, the more at risk you are. And that's why I'm actually not surprised. That's a really interesting insight. Namely, the generalists should do better in this labor market because they, of course, have a broader range of skills and qualifications.

Jeffrey Roach [9:35] Yeah, it's a fascinating read, highly recommend. And it gets you thinking about really preparing for this AI story. So first take there. AI, is it coming for our jobs? No.

Second hot take here, why are rates rising and what does it have to do with Bernanke formally at Princeton? I think there's a couple of insights here that we can tease out, and what I'm highlighting, and then I'll give it to you, Torsten, to tell us if you have any additional insights here. So Greenspan, Bernanke, Bernanke probably wrote a little more on the academic side of things. This was probably, what, 2005-ish? And both of those gentlemen came up with a phrase, "The global savings glut." What does that mean? And perhaps, and I think this is where we'll make the case, perhaps we're on the flip side of that.

Torsten Slok [10:37] Absolutely. And this discussion has exactly come back on the radar screen, after having been somewhat dormant for several years, because let's remember back to what happened from essentially 2010 to 2020. Interest rates were zero, and interest rates were zero because central banks were trying to boost the economy after the financial crisis. It took a long time to clean up in the housing market. It took a long time to clean up in the banking sector. It took a long time to clean up on household balance sheets.

So as a result of that, when interest rates were zero, there was a savings glut that around the world, also China, remember we had a lot of globalization. We did not have trade wars obviously at the time. We had, China had a lot of money that they wanted to invest. They were exporting a lot to the U.S.

Torsten Slok [11:20] They had a lot of dollars they had to recycle into the U.S. or into financial markets. So as a result of that, China had a lot of savings, Europe had a lot of savings, Japan, Australia, Canada had a lot of savings, and there was simply not enough projects and things to invest in. So fast forward to today, now we have an AI miracle, an AI boom, and now we have the total opposite. Now we can't invest fast enough in data centers. We can't invest fast enough in energy. We can't invest fast enough in defense. We can't invest fast enough also in government bonds because government deficits are really, really problematic and very, very high.

And the bottom line for that is that we've gone from a world where we had a savings glut, meaning there was way more savings in the global economy, and this was all pushing interest rates down to very, very low levels.

Torsten Slok [12:07] Today, we now have a savings shortage. In other words, there are so many things that we need to build in data centers, in energy associated with data centers, in infrastructure, in defense. And for that reason, there is not enough money to invest in all these projects that suddenly needs to be built. And the consequence of that is the exact opposite, that if there were before, during the savings glut, downward pressure on interest rates because there was just so much money slushing around trying to look for interesting projects and things to do.

Now we have the opposite. Now these projects that are being built need to compete for the capital that's available. And the way that they compete is by offering a higher yield. So that's of course why yields on hyperscaler debt have been widening. That's why the yield on interest rates from the government has been going up because there's not enough money to buy all the Treasuries that are being issued.

Torsten Slok [12:55] So in short, simple language, we went from a savings glut, which was invented by Bernanke when he gave a speech, of course, after the financial crisis, to now we actually have a savings shortage. And that's a really important insight because it tells you that as long as there are these many projects that everyone wants to invest in, that means of course that interest rates are going to stay higher for longer. And by the way, we also have an inflation problem, which is also a reason why interest rates are going higher for longer. That's why recently we're now hitting 5% on the 10-year rate. And that's of course a very critical backdrop for this discussion that interest rates are likely going to stay higher for longer, and in my view, likely going to go higher because there's simply a shortage of savings to invest in all the projects and all the money that needs to be invested, both in data center AI built out, but also in all the government debt that needs to be financed around the world, but especially in the U.S.

Jeffrey Roach [13:45] Yeah. And those who've been around the markets long enough, remember that phrase, the conundrum that Greenspan was terming because all this. So if you think about that, the conundrum was, you have rates, well, at least short-term rates, staying surprisingly low, even in those mid-2000s, early 2000s and such, you know, the Fed was adjusting short-term rates very, very low because you're exactly right on the savings side, the flip side of that same coin is the investment story.

So yeah, the conundrum of low rates because, and the answer was we just had a surplus, a glut of savings. In other words, to say, we didn't have enough investment, that capital was sniffing out. You could say where we are 2026, there's this massive amount of investment that dollars are wanting to flow to.

Jeffrey Roach [14:40] Capital's going to flow to where it's used best, best returns. And so with all that investment, there's not enough savings to catch up with it. I think that's a real helpful explanation and I think a straightforward one for our listeners. You know, the term premium, reasons why that's getting larger, inflation, all the dynamics that are pushing yields up. I think one underappreciated story, perhaps one of many, but one underappreciated story is just this idea that we're the flip side of the Greenspan conundrum, as it were.

So one thing that's interesting to me, Torsten, you think about, you know, China had a number of dollars they had to reinvest, Asian economies. I think two things that still could be at play that may suppress rates is one, oil countries, you know, the oil exporters, you think where oil is right now, those petrodollars perhaps could be a factor.

Jeffrey Roach [15:41] And then the aging of the developed economies. One thing that I've been intrigued with is the idea that we're going to have a pretty quick drop off in prime age working force, say, in the next 25, 30 years. So perhaps the oil story is still there that added to the conundrum and the aging of these developed economies, could continue to generate those more savings. Maybe that offsets some of the ideas that we have all of this investment project attracting dollars.

Torsten Slok [16:17] You're right. As always, there's another side of this coin. So if we list the bullet points, why are rates going up? Number one is we went from a savings glut to now a savings shortage. Yeah. Number two is, of course, we have inflation at the Fed. Number three, and that's of course why the Fed is thinking about hiking rates here. And number three is that we also have, generally speaking, some fiscal challenges that are also putting some upward pressure on long rates, especially the term premium.

And number four, we also have this unique situation that hyperscalers are issuing a lot of debt for data centers, and that is crowding out demand for other investment grade type of debt, namely including also in U.S. Treasuries. So the list of reasons why rates should be going up is relatively long. Now, in the middle of that sits exactly the oil price, because the oil price going up is of course another bullet point of why rates should be going up.

Torsten Slok [17:03] But at the same time, as you're saying, of course, oil prices, if they do go up, that means that the Middle Eastern countries who are now benefiting from oil prices going up have more dollars to recycle into Treasuries potentially. So that could, of course, be arguing for why rates should be going down. And you're right, another bullet point on why rates should be going down is, of course, demographics that we have an aging population is also pushing in that direction.

And finally, there's very significant debate at the moment also about AI. It also has some tentacles into rates because if AI does not succeed, in other words, if the hyperscalers are not able to generate the revenue that the consensus is expecting at the moment over the next several years, if that becomes flatter, that also means that there will also be some issues about valuations of the equity of NASDAQ in particular.

Torsten Slok [17:49] And if that's the case, some investors, if we do get a correction in tech stocks, if the open source models are beginning to have a bigger market share relative to closed source models, that could run the risk that the equity will be selling off in tech and AI names. And if that's the case, some of that money will probably also be going into rates. So there are different scenarios when you look at the bullet points where yes, it's true, it's not only the rates are going up, up, up, as it has been for now quite some time, but there are some other arguments for what could be the reason why if we look six, 12 months ahead, that rates could begin to go down. We still think that rates, in summary, will stay higher for longer, so therefore investing in companies and in entities that have actual cash flows, investing in value over growth is a superior strategy in that type of environment, both for debt and for equity.

Torsten Slok [18:34] But the bottom line to your question is, let's just agree that rates have gone up a lot so far this year.

Jeffrey Roach [18:39] Yeah, that's right. So the global savings glut story certainly worth keeping your eye on in the near term and very helpful in terms of explaining macro conditions. All right, third here, let's talk about inflation briefly. So I have our projections that inflation will not get materially weaker until perhaps maybe Q1 of next year, end of Q1, early Q2. I agree with that.

And so part of the challenge has been, you know, pre-February, you know, we were going into the year, we were thinking, okay, maybe by Q4. I think our key takeaway has been inflation decelerating has been pushed out, or at least, you know, materially pushed out. We're somewhat in an inflation fog right now. But do you have any changes on your forecast that you've had going into, say, 2027, relative to what your expectations were, say, just two, three months ago?

Torsten Slok [19:41] Oh, absolutely. So, if we just back up, remember, as we all know on this, the conversation that in the pandemic, inflation basically peaked at 10%, and now we're basically down at three and a half. So we've gone a long way, but the problem is that the last mile is going to be very, very complicated and going to be very sticky. And why is it so sticky? Well, there's a number of reasons. Number one is, of course, that inflation is sticky, partly because oil prices have now gone up so far because of the conflict in the Middle East.

Other reasons why inflation is more sticky is that there's still some delayed effects of tariffs that are putting upward pressure also on inflation. And finally, we also have some upward pressure on inflation because of restrictions on immigration. Immigration restrictions, putting some upward pressure on wage growth in agriculture, construction, hotels, and restaurants.

Torsten Slok [20:27] And in those sectors where unauthorized immigrants normally work, have seen some shortages of labor, and that has also been helping in keeping inflation more sticky and higher for longer. So for that reason, exactly, as you're saying, the problem is we got to get probably all the way into the early or middle of 2027 before we begin to see inflation meaningfully begin to go down towards the Fed's target of 2%. So in very plain English, the consensus at the moment expects that in the middle of 2027, inflation would be 3%, both on headline and core, but we, of course, at 3% still are above the Fed's target, which is 2%. So for that reason, the FOMC, as we speak, is, of course, at meeting after meeting, still thinking hard about how do we deal with this risk that inflation is three, it's supposed to be two.

Torsten Slok [21:12] Some FOMC members think we should be hiking rates, and some think we should be hiking a lot. Other FOMC members think we don't need to hike that much, and some are even thinking we don't need to hike at all. So this is why the debate is, with that trajectory in mind that you exactly outlined in the question, I still think that we should be assuming that not only have we talked about long rates being higher for longer, but we also need to work under the assumption that short rates for that reason, because inflation is sticky, will also be higher for longer.

Jeffrey Roach [21:37] Yeah, that's right. And what, a couple of things that I've been looking at that really are quite interesting, and, you could say it's really a good thing that we're complaining about, but maybe we shouldn't be complaining about, the idea that, you know, you think about the amount of cash, that boomers have. So the baby boomers are traveling, they're going out to eat. You look at OpenTable, reservations up 10% from a year ago, this discretionary spending, the demand side is still pushing up the inflation story.

So you think about that question that we're discussing, can inflation ever get back to that 2% and where are we now in this fog? Part of the reason, in addition to the supply constraints, is that we still have pretty strong demand, consumer demand, and that's adding and contributing to inflation pressures.

Jeffrey Roach [22:26] OpenTable, I mentioned TSA throughput numbers still very, very high and, even relative to where we were last year. Again, discretionary spending, particularly from baby boomers are adding on the inflation side. Let me get to the next one. This we're recording here on Monday the 14th of September, and we still are a few days out before the Fed hikes. Not a surprise, I don't think, that the Fed will react, and tighten. Let's assume they do on Wednesday.

I think the real debate, I want to hear some commentary on this one, is this a one and done, and when are we going to see some repricing in October and December? Perhaps we're still going to be in inflation fog by the time we are in October, but maybe December repricing. At this point, markets are expecting more than just one hike.

Jeffrey Roach [23:24] It's not just a one and done. And that is, I guess, fundamentally driven by the fact that you have inflation pressures that are probably going to continue on in the near term, and it's not just a short-term, transitory, shall we say, inflation pressure. One and done or a couple hikes in the office?

Torsten Slok [23:45] It’s really, really unusual for the Fed to be one and done. So that's also likely the reason why the market, as we speak, as you know, and as you and I have talked about also, have really moved in a very, very significant way. Let's not forget, we all went into this year with the market pricing cuts, the dot plot from the Fed, meaning the expectations to show rates was that the Fed will be cutting rates in 2026. And here we are today, literally pricing that we're now having four hikes priced in, a hike here in September, a hike in December, a hike in March, and a hike again in June.

And that's, of course, a really unusual development. And that is exactly answering your question by basically emphasizing that the market is clearly seeing that inflation is just not under control. And that's the challenge, namely that if we still have inflation being sticky, if we still have a strong economy, and as you just mentioned, the consumer is still in aggregate doing fine.

Torsten Slok [24:35] The weekly data for consumer spending from Redbook is still very strong. And as you mentioned, throughput TSA, the daily data for travel still strong, OpenTable tables data for how many people go to restaurants is still strong. All those things are arguing for, well, maybe the economy is basically in better shape, than what we thought in the beginning of the year. And with that backdrop, I do think that we should begin to plan that the Fed will be hiking several times.

So what does that mean from an investing perspective? It simply means the main conclusion, namely the rates are going to stay higher for longer. And from a very, very simple perspective, that also means, of course, that there's still the whole issue around how do I think about debt, how do I think about equity, and how do I think about what types of assets and what type of risk do I take if the Fed is still engaged in trying to slow the economy down, engaged in trying to slow down risk-taking?

Torsten Slok [25:21] So the short answer to your question is I think the Fed will go several times and at the market currently pricing for hikes is a good guess and a good estimate at this point.

Jeffrey Roach [25:31] Yeah, yeah. So I think one of just tying this topic up before we get to our final question, that is this, again, let's harken back to Greenspan. We talked about his conundrum, as we may remember, he also talked about party goers complaining when the punchbowl is taken away. Yeah, exactly. And so, you know, it's interesting when you think about how markets react, you know, perhaps they complain for about two, three months. By the time four months passed after the first hike, markets have kind of gotten over that. We show a chart in our weekly market commentary to that effect.

But, you know, perhaps we're complaining too much. Why should we complain that, you know, a lot of households have healthy balance sheets, net wealth relative to disposable income, pretty high ratio there.

Jeffrey Roach [26:19] So we complain because perhaps, we don't like the punchbowl taken away if we're going to reference the Greenspan era. So last question, after four macro topics, the fifth and final here is the so what? Given the macro picture here, given the fact that rates are going to be higher for longer, yet we have low unemployment, we have a dynamic labor market, we have pretty healthy households, where do we allocate capital?

Torsten Slok [26:53] Well, if we just take our textbook out and ask the question, if interest rates are higher for longer, in other words, if interest rates are going to stay elevated because the Fed is still fighting inflation, what should we do in asset allocation? Well, one immediate conclusion is that, well, if interest rates are going to stay higher for longer, you need to make sure that you invest in companies that actually can pay the higher debt servicing cost that come along with interest rates staying higher for longer. In other words, this becomes a debate about value versus growth, because the risk with interest rates staying higher for longer is that growth is normally characterized by having earnings far, far out in the future. And if you have earnings and cash flows far, far out in the future, you invest in those companies today, you run the risk that those cash flows become much more sensitive to what interest rates are doing today.

Torsten Slok [27:41] In other words, when interest rates begin to go up, valuations of growth companies tend to move much faster down. And therefore, the short answer to your question is that value is preferred in a rising rates environment. In other words, companies that actually can service their debt and companies that actually have the ability to provide the payments that are needed when interest rates are higher for longer. And one way to express that view, of course, is in both debt and in equity, but in this case, in private debt or in private equity, and in private equity, of course, investing in companies that have actual earnings today.

Of course, one simple way of thinking about what are the alternatives to do, and especially, of course, in the private markets in an environment where rates are higher for longer. Because if you, for example, invest in venture capital, which is characterized by only having earnings and cash flows far, far out in the future.

Torsten Slok [28:34] So if rates are now going up and staying higher for longer, that of course means that those types of investments are going to be more vulnerable. So that's why the short answer to your question is, invest in firms, invest in assets that actually have cash flows. So that means invest in value and invest in this case in private equity and companies that actually have earnings and have the ability to pay the higher debt servicing cost when interest rates are higher for longer. Similarly, in fixed income, in private credit, in public credit, in rates, when the yield level is high, of course, that also is helpful for people with money that can be investing. But the short answer to your question is, in a rising rates environment, that is an environment where, in my view, value will be doing better or be a superior strategy relative to growth.

Jeffrey Roach [29:16] Yeah, very good. So tying it all together here, we talked about AI. It may prove to be a complement to workers boosting output productivity, thinking about how it's reshaping work, not necessarily eliminating, it's not coming for our jobs.

Second, rates higher, partially because we're not in the Greenspan conundrum of a global savings glut. We have a lot of global investment opportunities, taking cash and taking capital. So that's the second item. Third, of course, thinking about inflation. We are going to be pushing out our expectations and forecasts on when inflation can finally hit that 2% target.

Next, Fed will hike. I am certain of that. Well, I should say as certain as I can be, the debate, of course, that the markets are going to be working through. Is this a one and done or is this more than one?

Jeffrey Roach [30:13] And it really, I think, relates to what the Fed is thinking about inflation and the persistence of inflation. Next, of course, thinking about the stability of macro conditions, low unemployment, strong consumer health, thinking about the corporate side. What do we do with it all? The so what? Look for value in a rising rates environment. So a lot to unpack, a lot to think about. Thank you, Torsten, for joining. And that's all for now. Take care.

Torsten Slok [30:43] Thank you.