Investing in the age of AI

Christine Kashkari: Welcome to Season 1 of Better Vantage by Vanguard, a podcast series hosted by Custom Content from WSJ and Vanguard. In this series, we will be talking about the most pressing topics that are top of mind for financial advisors and investors today. In each episode, we'll be sharing some market insights, strategies, a little bit of history, and a lot of conviction.

I'm Christine Kashkari, editorial director for WSJ Custom Programming and your host for the series. And joining me today is my cohost, Joe Davis. Joe is the global chief economist for Vanguard and also the author of Coming Into View: How AI and Other Megatrends Will Shape Your Investments. And that is a topic that we will be talking about today. But Joe, today you’re on the hot seat. Before we do that, as you know, we'll have a lot of experts in this season. But for our inaugural one, you said it, you’re on the hot seat.

Joe Davis: Well, I’m really excited, and I think you teed it up really wonderfully, Christine. Again, you know, listeners are giving us their valuable time. Why should I care for the economy? Why should I care as an investor? And give it to me in an unvarnished way. Let me make my decision, but help me make my decision, giving me really unbiased information and perspective.

Christine: Your book talks about megatrends, AI among them, and the outsized role that they play in the global economy and financial markets. But talk to me about megatrends and why investors should care about them.

Joe: Now, what's not new, if you follow the markets, you care about the economy, is that over really long periods of time, there are forces that drive everything. Growth, inflation, stock and bond markets. And there are things such as technology. You mentioned AI. It's globalization, how we trade across borders. It's demographics, the aging of society, how many people coming in and out. Debt levels of countries. Even you can think about geopolitical risk and climate change.

So yes, they matter in the long run, but what the book was about and what's been eye-opening to me, is how much they matter, twists in those trends matter so much for the here and now. So this is not just a long term, and I'll worry about it in the year, I don't know, 2040, this matters for the here and now. It drives a business cycle, which means GDP up and down. It certainly drives the stock market—not solely. Things such as so-called demand and consumer spending matter, but underneath the surface are shocks to these that are affecting the economy today.

Christine: Your research, for example, on megatrends have led you to believe that the consensus view on the long-term growth trajectory of the economy as a continuation of the past decade, 2% growth, 2% inflation, is actually the least likely outcome, and AI plays a role in this. Can you talk about that?

Joe: This was an outcome or a forecast I wasn't looking for, Christine, to be honest. It makes you a little nervous when you're an outlier to the crowd. But, again, we weren't looking for it. And why is this emerging? It's because there’s a push and pull. There's a tug-of-war that doesn't always emerge, but I think we're starting to see the very early innings of it in the United States, and that is on the negative side, you have structural fiscal deficit. Structural is just a fancy word for saying they're ongoing. They come all the time. It's not due to recession or something. So they're always compounding and they're rising. And that's been well documented. So that's the negative side.

However, the other side you have where, in our system, our framework, AI starts to emerge as what’s called a general-purpose technology. And that's a fancy way for saying, it not only makes you and I better as employees, both from saving some time, and just making us better workers, better product.

Christine: And so you are betting on AI.

Joe: So that's what the probability. So, if you are looking for good news --

Christine: Yes, I am.

Joe: They are. So they're tilted. The roughly 30% to 40% scenario that we talk about in the book is called deficits dominate. It's more of a pessimistic, it's not end of the world, I want to be very clear, but it's more of a drag, and certainly the bond and the stock market for sure is not prepared for that scenario. 30%, 40% is pretty elevated. There's not a little tail risk. However, closer to 50% is where AI transforms. It lifts productivity to the levels, think of like the late ’90s, different period of time, but that's a thing you can have in your head. That was with the personal computer and the internet really starting to supercharge some innovation and some productivity in the economy. In fact, Chairman Greenspan, at the time, from the Federal Reserve, called it the New Economy. It points to, that's where it's tilting.

Christine: I think the consensus probably still is that this technology, AI, is more innovative versus transformative, which is what you're looking at to include that as a megatrend.

Joe: Well, yes, we have three … What we tried to do is operationalize. So again, we've created no new theories. We take the leading minds in the world, we borrow their framework, and we've just put in a statistical way, the three dimensions of technology. Again, it either saves us time, and it can be all the way to automation, and then we lose a job. It can actually make our jobs better. PC software did this for the accountants, as an example, in the 1990s. That leads to higher incomes, and perhaps an even higher number of jobs. And then finally, it's that third dimension, that's the transformative ability, is what new products, new industries? The internet led to the whole online shopping and retail. So that that's a form of a new, if it's AI, I don't know what it would be. It points to a likelihood of that happening. Which new product or service? I have no idea. Our system won't know.

It's funny, I get two reactions to our forecast. For those that worry about our debt levels, they’ll say, “Forget AI. Why isn't the deficit dominates scenario 100%?” Of course, when I go to Silicon Valley, they're pushing me, “Why is the AI win scenario not 100%?” But either signal, right now, it's still undetermined how far advanced AI will become. If it plateaus right now, we'll get some lift from growth, but not enough to overcome some of the deficits that ultimately will start to build, say five years from now. What I was struggling with is, we're talking about all these trends and narratives out there, but give me probabilities and magnitudes, because I can't construct a portfolio if I don't know the probabilities, because some of these things sound serious, some of these things sound actually very interesting, but is this a small chance? Is this kind of like a hope and a prayer?

Christine: I think what I like about the framework that you're describing, too, is it is dynamic, right? It's not static.

Joe: It's not static. As signals come in, so we've had, you know, we'll have continued focus on the balance of spending and tax rates in the United States. You obviously will have variation in the geopolitical tensions, potentially. We even have climate factors in here, and you can have changes in those. We're looking at labor and investment rates at the aggregate, and looking at discrete patterns in them along with the ROI, or return on investment, from companies. So it's the fact that all the decisions CEOs all around the world are making, and we're watching them in real time, and then we're projecting.

What's really interesting, I think our one novel contribution is that we've put all these forces together, and that's what was eye-opening for me is that we're quantifying the relative importance of these factors. And then based upon those signals, they vary by country, and they vary over periods of time. And so I think it's going to give us a better framework over the next three, five, seven years in terms of where the risks are, knowing that this is inherently the future we're talking about.

Christine: And so, but when we talk about AI, automatically people, the first thing that people think about are their jobs, right? Jobs and responsibilities. How are you reframing this AI-work conversation?

Joe: We've looked at every job description in the world. 800 occupations in the United States. We've looked at every task of every job description against the AI, its ability to do that task today, and in five years from now. And it says, more often than not, we're going to probably have a lift in productivity because of the augmentative. We will get better at our jobs. The headline is, it's not all doom and gloom, but much like that we saw with the personal computer in the ’80s, it went from, “Do I have to worry about this thing? Do I use it?” to, “No, actually it could help.” But yet it's change. That's the sort of environment where we are already entering and will accelerate over the next five years.

Christine: So I guess, Joe, the question also becomes who would be the first affected?

Joe: What I try to think about is, as an economist, sometimes I make the mistake, I may have made it on this episode, of just talking about numbers. Talk to me as a human being. So if I'm sitting around the dinner table … I have two children. They’re entering the workforce. Is their life, when they’re 25 or 35, is it going to be better than mine? Because when you throw all the numbers away, that's the answer I want. Now, the numbers will say, “It depends.” But let's talk about as real people, Christine. More people will benefit than hurt. The entire economy, if you talk to the average person, there's always been a battle, always been a race, I should say, between education and technology. And it's not going to be any different in the next 10 years than the previous 100 years, right?

So the importance of education in the next 10 years going up, even with AI. Those that say education is not important, they’re out of my field. They don't know the history of the workforce. So it’s going to be, generally speaking, yes. But AI’s got, that's why if I'm an investor for sure, and if I'm the average citizen, I'm cheering for AI done the right way. But with that territory of technological advance certainly comes change and disruption.

Christine: So, Joe, for investors that are –

Joe: So let’s talk money.

Christine: Yes. Let's see what we investors can do that are interested in AI and are looking to capitalize. Most of the time they flock to tech stocks of all stripes, overweight tech sector. And you're saying that, eh, not a good idea.

Joe: Loosely speaking, in every technological phase, there's two phases to the investment cycle. The first phase, the technology is being produced. It's starting to be on social consciousness. I mean, talking to the average person, not someone who knows it. Well, the second phase, that's where the companies and indices, think of a broad basket of an ETF or a mutual fund, a broad basket of securities outside of that tech index actually tends to outperform. And that sounds counterintuitive.

Why? It's because the very definition, we're talking about a technology that's lifting efficiency across companies that are nowhere near Silicon Valley. If this is that transformative, tell me how a hospital is getting more efficient, delivering care. Tell me how a utility company, or tell me how an asset manager or a bank is processing records or so forth, way more efficiently than in the past. The workers, you and I in our jobs, in our fields, we're getting better. If that's happening, then that's where those industries and those companies in, I’ll call it the value space. Our framework comes to a fairly strong conclusion is that whether AI is disappointing for the next five or seven years, or AI truly does start to prove transformational, you're talking about a broadening in equity returns, or convergence. Convergence either way. So it means the end of U.S. exceptionalism, but that sounds like that's a pessimistic statement. That's not my intent. Yes, if AI is a dud, we have some deficit issues.

Christine: Joe, there are some smart investors that have real concern about a fiscal crisis happening in the next two, three years. What do you say to that?

Joe: Our framework is imperfect, but it's certainly data-driven and it's putting probabilities on these scenarios, and there's a 5% chance that the scenario happens. I give credit to those in, say, that view of the world that are raising these concerns, because it's been a bipartisan issue.

But why are they low at the moment? It's because three things. To have a high conviction that scenario’s going to play out, and some investors positioned their portfolios like that. I've heard of a high allocation of gold and other really defensive assets. If you have that view of the world, I just want to underscore that you have three assumptions that are going to play out in the next three or four years, and that's what our framework helps to show. One is that AI is going to have no impact. It's completely marginal, much like social media. We all use it, but it hasn't done anything for economic growth.

Second assumption is, there's alternative stores of reserve currencies they call it. Christine, it's a fancy way of saying, there's other bond markets that I can invest billions of dollars at scale with depth and liquidity like the U.S. Treasury market. That would mean that the euro market, say, euro bonds would emerge, or China becomes much more investable, like say for me in the United States. Assumption three is what I think many leave out, and that is the important role that the Federal Reserve has played, and I believe, and more likely than not, will continue to play to keep expected inflation under control over time, which would fight against those fiscal pressures.

Christine: What if AI doesn't transform, that AI actually disappoints?

Joe: There's a 30% to 40% chance that AI disappoints along the areas that matter a lot for the investment community, which is growth, keeping inflation down, and expanding the opportunity set, keeping interest rates stable. Really the greater risk, if AI disappoints in our framework, the greater risk is for the equity market in the U.S. for a time, because they have high expectations for earnings growth.

Christine: If there's one key takeaway?

Joe: 80% chance, not to be alarmist, 80% chance that we have, we just have a different-than-expected economic environment. Not radically different, but they're going to be different from what most expect, and that presents some risks for us as investors. It also presents some opportunities. And think about it that way, as you navigate your portfolio for the next five to seven years, and stay invested as you think through that process.

This is around risk management. It's not about being smarter than the market. It's about, no, I actually, I don't want us to be smarter than the market. I'm going to think about these scenarios in a prudent way and balance that so I'm not so concentrated in only one state of the world. That's what we're talking about, and that's what I think investors could think about for the next several years. And we have a little time on this. We have the next several years to think about this.

Christine: I am confident you will not be replaced by AI.

Thank you, Joe. That was terrific.

Joe: Yeah, maybe for you. You weren't on a hot seat, but looking forward to our next episode.

Christine: Me too. Thanks for joining us. If you enjoyed this episode and found it helpful, subscribe and share.

Disclosure:

All investing is subject to risk, including the possible loss of the money you invest.