We Asked Vanguards Chief Economist Why Ai Has Two Huge Tails

There has never been a great technology that has not had a significant draw down in stock prices, which is code word for saying what some would say there’s a a bubble forms. We’re having economic growth projections that are 50% above the consensus. Our projections are AI is going to is going to affect 80% of the occupations at twice the rate of the personal computer in four years. The personal computer took 15. either the trend for growth is going material higher because of because of the innovation of AI uh that overcomes the demographics and the debt levels we have that is by far our most likely outcome. Um but if it if if it only if if AI only manifests along certain techn if it only automates which means it has not become a general purpose technology we use it all but it’s it’s like the farm tractor um that would be disappointing and then that sets into motion this eventually not not today not tomorrow but you get fiscal pressures because they’re already high. You’re watching Excess Returns, the channel that makes complex investing ideas simple enough to actually use, where better questions lead to better decisions. I’m Matt Ziggler. Justin Carbono is co-hosting with me today our guest, global chief economist and global head of the investment strategy group at Vanguard, down the street from me in Pennsylvania, author of the 2025 bestseller, coming into view, how AI and other mega trends will shape your investments. Joe Davis, welcome to Excess Returns.

Ah, thanks for having me. Pleasure to be here. I was telling you before we started recording, it’s worth saying it. If you haven’t seen a copy of this book, make sure you check it out. I think of the books that my team at Sunpoint or Raa like all held up and said, “This is really really cool. This book is one of them. Thank you for writing it.” Yeah. Well, again, we didn’t we didn’t aim to write a book. At least that wasn’t the goal. Our goal is to try to get a handle on where AI could go several years ago and how would that would compete with some other serious trends which we’re going to probably talk about today. And uh again all the proceeds from the book go to charity. Um but uh if if it’s helpful for for audiences really smart audiences but you know perhaps not reading economic white papers every day uh that was the goal is to make it you know as accessible as possible to a really smart and savvy investment audience. succeeding on all rounds. So, I’m taking you right here first, which is basically one thing I love about your work is the embracing of quantitative frameworks and not in the not in a way that loses touch with reality. So, how should investors think maybe differently about macro when they zoom out to that long-term lens? Because you see it kind of in a unique way and a lot of professionals, they just underweight the slowmoving nature of the way you approach this. And I I think that was, you know, was even a learning to myself and I’ve been business, you know, over 20 years. I think there’s a standard approach which looks at the near-term economic uh contours of the data. You can think of GDP or the inflation rate uh what the federal funds rate may be doing and then and then of course there’s a mapping you know implicit mapping or explicit to the bond market and the stock market. Uh we do all that. Uh but what we’ve added and what we spent some time which is behind the book and the analysis is is looking at the evolution of these longer term trends too. And and what what I found fascinating is that when those trends start to change, which which which can happen, you know, on a regular basis, they themselves affect the near-term, not just some long-term assumption that say, “Hey, I’ll worry about that 10 or 15 years from now.” It it it affects the business cycle. Um and and so you know what we what we concluded two years ago with AI is that that was going to have implications not just for the next 5 or 10 years but it was going to affect the economic growth projections you and everything we care about in 2027 2028. Um and so here we are. Uh so that’s that’s been the eye opening is that integrating them they’re tended they tend to be separated in a in an academic sense the near-term business cycles like the Federal Reserve uh asset prices you know which all of us care about as asset allocators and then this long-term trends are kind of sometimes kind of left to the side or just thought loosely. Uh what we did is just simply integrate them in a numerical way. uh but that’s affected our approach to not just our long-term assumptions but our near-term ones for the next several years. So let’s talk a little bit about the four structural drivers that are in this mega trends model you guys came up with technology demographics fiscal deficits and globalizations walk us through the framework. Well and again that’s taken some of these long-term trends which we all know exists. I mean we didn’t bring anything new we certainly didn’t make up these factors. They’ve been known for over a century uh that they can impact long-term standards of living, long-term economic growth. I mean, you can’t talk about long-term uh productivity and standards of living without talking about technology. And yet, uh they’re just again parked to the side when you think about, hey, what’s my outlook for the next year, the next two years? And so, uh we brought them in. uh those four it’s it’s effectly it’s a fancy way for saying that there’s supply factors and and demand uh supply the supply of of workers and and people so that’s demographic factors the aging of society immigration you can think about that globalization factors again can be in the headlines it’s the openness to trade it’s also tariff rates um and then you also have you know it’s the rise of China into the world trade organization they’re that’s what’s behind globalization so you can imagine they have cyclical effects. They can also affect the long run or mediumrun trajectory of growth in inflation and interest rates. And then finally, the biggest one by far is technology. And and what I’m proud of is that we look at through the lens of three factors of technology. Its ability to um to substitute for human work. So automation, the ability to augment to make us better. You can think of co-pilot much like the personal computer for some jobs. That’s augmentation. And then third which is really the magic uh and we don’t see that from every technology that is uh the technology becomes a platform to enable new products new industries electricity and the personal computer uh were great examples of that internal combustion engine and so again we we brought all those frameworks into the modern era and then look looking at AI and its ability to affect the world economy on all those three facets and that’s that is novel that’s generally not done even from central banks and uh practitioners and something that’s really integrated uh in our system. Is technology the proverbial one mega trend to rule them all or is it just right now? No, it it really is. I mean what I was fascinated for example is when you so what we are doing is you know all these factors are moving all the time along with GDP and inflation stock market interest rates they’re all coexisting I I think of a living breathing uh system. Uh but because of that we’re able to attribute or assign what’s pushing up and down growth at any point in time. We look at we look at back over well over a hundred years because great technologies only come around every so often. And what I was shocked to find is for example that that that growth has been low generally speaking or has been slowing over the past two decades in nearly every economy we looked at. Not just because we have fewer people entering the labor force, the aging of society. It’s it’s more importantly because we’ve had a lack of automation in a service-based economy. So, uh that that’s important because given what the work we did on AI, we we immediately thought, you know, before we brought data to bear that that AI could significantly lift growth over and above expectations because the lack of automation, uh if if AI could help automate, it would actually push up growth without needing new people or the same amount of people that we saw say in the ’ 50s and 1960s. So again it’s we we have to we but the data is keeping us honest uh rather than we did not want to be in the narrative business or hey it’s Joe’s personal opinion that AI could be transform uh transformational so it’s AI certainly is is technology is the by far the biggest one but it but it has to evolve in a certain way because there’s significant liabilities in nearly every economy we look at on the other side you got aging of societies look at the demographic patterns in China you got del globalization or the threat of deglobalization and you got the rise of debt levels in nearly every economy around the world. They those would all seem negative for growth and potentially push up uh interest rates and inflation and so that’s why the system is effectively doing a horse race and depend on the signals of AI could that offset uh and would it do it in a certain way that would be positive for growth and and labor markets or to be disruptive that that is what we’re quantifying in our framework. A fascinating, and when I say fascinating, I do mean I mean utterly fascinating point that you bring up is that changes in mega trends explain roughly half of the S&P 500’s quarterly movements. Yes, this is a wild long-term meets short-term. This is not the hairy dent demographics about 25 years ago. Well, and I I was shocked to find I did not know this. So like again what’s standard in macro and I’d say asset pricing is the short-term fluctuations in GDP for example with what all what what some central banks would argue is all called demand. So in other words if GDP is going up it must be consumers are spending more hence we got to raise rates. What we find is that that’s half that’s true half the time. Uh but the other half of the time it is it is the emergence of new ideas. it’s investments rates accelerating because the trend is likely to change in productivity in innovation uh and can drive earnings power. And so that’s what you’re saying. And so half the time it’s not necessarily and that’s had that has different implications for in if if if if growth and earnings are rising because what effectively is the supply of ideas is increasing that means that inflation ultimately will come down it means interest rates may not rise with that levels and you can support higher valuation. So again it’s that horse race of the two and we have them competing in real time all the time and that was eye opening to me. I thought all the near-term volatility in stocks in GDP was the so-called business cycle was just purely demand. And but if you think of commodity prices, I think any commodity trader would say, yeah, but it’s also the supply of oil that affects the price of oil, not just, you know, how many how how much we’re driving on the road. And so I it’s in one sense it’s natural, yet it’s not generally under done in in my practice. And again, we weren’t looking for this finding, but it is it is honed how we approach forecasting and and that’s why we published a lot of this work is to is to hopefully other researchers could build upon it um make it even better. Um but it has certainly changed not only our view but our approach to forecasting and I think that going into any conversation especially the end of the year when we have these about year- ahead forecasts or capital market assumptions. Yeah, this is a really valuable input and you stumbled across it. Well, and again, we’re looking at every horizon, right? I mean, you know, when we take when you think about capital market assumptions, at least historically in our publications, we’ve tended to focus on 10 years. Now, now why 10? We’re forecasting stock and bond returns, interest rates, currencies at every at every month, you know, uh in infinitely. I mean, in one sense, you can the computer can keep going. We focus on on 10 years in part because valuations can have some predictability. Um but but our our for our outlooks for capital markets uh they they do vary by the horizon and you know AI is going to play into that. Um you know it it suggests that AI um could very well rise you know increase earnings but it could depress valuations could still mean there could be a modest headwind over the next 5 to 10 years. However, because our economic growth projections are so high relative to consensus over the next year, it means that there’s, believe it or not, there’s upside risk to the equity market. So, I’m I’m I my I do not intend to be dissonant. It’s a matter of the horizon over which we’re planning. I think as an asset allocator, I think you have to take about the short term and the long run. There’s a stronger signal for stocks in the long run, but but but they’re not a market timing tool. And that’s why we say everything in moderation with respect to this. I I view our framework really as a scenario analysis and where are the risk that we’re trying to mitigate for clients rather than trying to you know look around every corner because no one even our even with our sophisticated models we will certainly not be able to do that. Joe, where do you think we are sort of in the innings here? I mean, in in some ways, you know, if it’s a mega trend, then you would expect there to be sort of some like long runway, but do you have any sense around like what inning we’re in? I mean, in some ways, you know, people look at the market and say we’re in a we’re in a bubble when they see some of these, you know, trillion dollar companies or prices, but on the other hand, you know, if it’s a mega trend, then that means there’s probably a long uh runway here. What what what are your opinions on that? Yeah, it’s and again, it’s a great question. Um it’s a it’s a it’s a tough question understandably I I’ll give you what our what all of our data suggests and that is we have investment rates going back 130 years across all the great technologies in the world and and that’s one of the reasons it took you know we really spent some time on this. We have the deepest data set to my knowledge on the evolution of technology um in macro at a high frequency. Um if you look at those projections uh and where we currently are in investment rates because investment rates is a good example. You can compare that to railroad to electricity you know a lot of transformative technologies. It suggests we are we’re certainly not done. Um, if I had to use a a a year, I I would put economically we’re in the 1996 1997, meaning the full buildout. If you use 1999, ergo is the peak. Um, the market seemed to be ahead of that. Um, but it suggests to me that the momentum in the market um could could really run. I mean, we’re having economic growth projections that are 50% above the consensus. We have 3% economic growth for 2027 with zero contributions from demographics and zero contributions from globalization. That shows you how much lift we’re expecting from automation and augmentation through AI. Um, but we haven’t seen it yet. It’s just what it’s anticipatory. So I I’d say we’re still in the buildout phase. At some point, every great transformative technologies has led to two things. one is an eventual rotation in the investment opportunities because we go from buying the the the technology to implementing it. Uh and then secondly, you have a rash of new entrance. Uh and then you have a you have a period of consolidation. There has never been a great technology that has not had a significant draw down in stock prices, which is code word for saying what some would say there’s a a bubble forms. Um I I always hesitate to use the word bubble and for this precise reason. It’s because you could lead to significant market overvaluation. The market could consolidate some point in the next two or three years. However, I that that would imply if I said the word bubble, I think some listeners would would would associate that potentially with tulips. In other words, there’s no intrinsic value other than looking at a at a picture of a flower. uh we’re talking about, you know, uh most of these technologies ultimately rewire the economy. So, in that sense, it wasn’t a bubble, yet you still have a market consolidation for a time in the stock market. And that’s where the economic transformation and the equity market performance diver divorce during great technology uh periods for a time. Uh and we’re going to see that very likely happen again. And that’s what our analytics strongly suggest. Um, and that’s our theme. You know, in the near term, we could have significant momentum, particularly in the mega cap US growth space. However, um, if I’m managing risks in client portfolios, as well as trying to harness AI’s payoff, it starts to push you outside of of the very area that’s getting the most attention now because it’s starting to you’re trying to get ahead of what that second phase of AI would be. And again, it’s tough. It’s but that’s that’s that’s what I talk a little bit in the book. Uh it starts to push you outside of the mega cap uh growth area. Not because we’re skeptical of AI, it’s it’s actually quite the contrary. Um but there is a timing element to this. So again, bottom line is if you had to pin me to the wall and say a year, uh it’s not 1999 because of the economic buildout. It’s still significant. We’re still in the learning by doing phase. Uh it’s moving quickly. Uh but I would say you know our projections are AI is going to is going to affect 80% of the occupations at twice the rate of the personal computer in four years. The personal computer took 15. So it is fast. We’ve done this at every occupation. We have 800 occupations. We’ve been looking at this for a decade and we’re monitoring it in real time. So we’re not seeing the disruptive both saving time and and some job loss. But that’s going to come. um it’ll benefit more jobs than it’ll eradicate which is why we don’t have a mass dystopia. Um but nevertheless we will see transformation uh and disruption that’s greater than personal computer. That’s how we get high growth. So we still have several probably two years at least of that. At some point however the market will continue if we’re right. My worry is that it continues to run on an explosive basis and that sews the seeds for eventually some uh market consolidation. I I just I can’t tell you. No, no one can say the year. Uh but I don’t use the phrase bubble yet because we’re still in this um buildout phase which probably has another year or two to go if our projections are right. What would be those second areas of the market that may be sort of some of the biggest beneficiaries of AI in phase two? like where would you be paying attention to what would you be looking at? So this is what I found fascinating with some of our research uh things I didn’t know we look back at all the technology great technology cycles what what what I what I found fascinating is that at some point during the great technologies as they as they truly transform what’s called general purpose technology at some point uh the the the benefits to stock investors acrew outside of the technology space itself for two reasons so the biggest beneficiaries are the users users are adopters of technology that have unmet need. Uh areas where there’s unmet needs, high cost to serve or new platforms emerge that unleash new revenue opportunities, effectively new fields or industries. For electricity, uh the example I talk about in the book is uh the entertainment industry. Without electricity, we don’t have movie theaters. We don’t have household products like vacuum cleaners. Electricity didn’t create any of that. But it we would not have had those products those so it pushed you outside of electricity and utilities. Uh the per the the the car did the same thing. Um uh you know it unleashed opportunities in retailing uh not just in auto manufacturing. Um and so uh what we are looking for again areas where there’s there’s potentially labor shortages, difficulty difficult difficulties in scaling, uh unmet needs, which means people would buy more of the product or service if the cost came down. And so areas that com immediately rise to the fold are in the service sector of the global economy. If you use the United States in leading in uh areas include healthcare, they include financial services including advice, you know, like RAIA services. Uh they would include education uh and then other business services. Um these are areas that are going to see some automation effects. However, there’s also going to be the potential for a wealth effect as consumers wish to buy more, businesses wish to buy more of those services, but at a lower price. Um, and so that they’re the areas and by the way, those sectors, those that are publicly traded are generally not in the growth or so-called tech space. Um, uh, the internet did the same thing. I mean some of the biggest beneficiaries were now consumer you know consumer products you know uh areas that sold things over the over the internet rather than internet service providing and so that that’s our in that’s what our our our analytics suggest I can’t prove to you it’s going to be healthcare but if I had to overweight one sector of what AI could potentially uh boost the most it would be things such as financial services not only just the co reduce the cost but the ability to unleash new products and services and then healthcare. Those would be the two leading candidates in my book. Which is why it tends to push you ultimately into valuebased orientations. Not because and the primary reason is not because AI is is overvalued. It’s because the benefit starts to transmit to the whole economy which is why it’s a general purpose technology. One of the things that you focus on in the book is sort of the fiscal challenges that we face as a country and that like you said earlier all over the world you know high levels of debt at the country level and I’m I’m kind of wondering like in some way you know if AI actually delivers it on on growth it kind of makes the kicking the can down the road just continuing to happen. You know, I it’s funny like I remember I don’t know if you guys remember like the Simpson BS like proposal where they were trying to do what was right. Yeah. 2011 2012. Yeah. And you know, we don’t hear about that really any of that stuff anymore. So in some ways, you know, AI is going to be possibly great in a lot of ways and good for our economy and good for growth, but maybe in some ways it’s not going to allow us to get at some of the root cause of these fiscal problems. I don’t know what you think of. Well, and that’s where we get these non-conensus outcomes. Our most likely outcome is that that AI is more transformative than the personal computer. I mean, that’s our baseline. Uh, and it’s coming from the analytics of our projections. It’s not my opinion as we mentioned before. So, we already have a very differentiated. I mean, our neck is out there, but I feel good because we’re using data to to bear. Uh, we have growth projections that are the highest in any consensus survey. Um um but um that allows us to your point to kick the can on fiscal deficits. What happens in our scenarios that structural deficits are so-called the sustainable fiscal deficit which is high in peace time. We’re close to 6% deficits to GDP. Um and this is with an economy expanding. You kind of hover in a four or 5% range. If growth goes from 2% in the US to 3% you get higher tax revenues, higher um uh and allows you to to forstall that. We saw this in the 90s. In fact, in the late 90s, we actually turned into a a surplus. Our projections generally don’t show that because of the aging of society, which we also have in our projections wanting to push the deficit up. And so you can see the importance of AI uh in trying to forstall these uh fiscal pressures which is why you know in one sense I’m cheering for AI to be as transformative as it is along all three me dimensions. Why why you know if if we’re wrong in our projections uh it’s it’s that um AI only automates. We don’t get augmentation. So we don’t become better as workers. We it just saves us time so much so that we have some job loss in some professions. uh we have seen technologies like that uh the assembly line was one the farm tractor was the other you know forgive the hundred-y old examples but it was true um if if AI only automates again that is not our likely that’s that’s the second most likely but it’s not nearly as likely as our baseline it’s half as likely um the fiscal deficit issues come in the come in the uh come into the four again in about two or three years because you have growth the AI buildout still occurs but the 3% GDP fades back after kind of the initial buildout. We don’t get the new industries that our projections, you know, generally anticipate and we get less benefit for us as workers. And so what happens is trend growth doesn’t ultimately change. We get this little, I’ll call it a sugar high for two or three years and we fade back down. Now you have a structural deficit that’s going to 6% to 8% to 10%. deficits to GDP because the aging of society and our fiscal commitments, social security, Medicare and Medicaid. And so what happens is now you start to get pressure uh on our currency. You start to get pressure in in in in the bond market and you have the Federal Reserve trying to fight those inflationary pressures um uh and and you know they’re forced to keep interest rates higher than they would to keep long-term interest rates kind of low, but there there’s fiscal pressures. And so that’s that’s this economic scenario I don’t like talking about. Um but that’s really in the book I call it deficitis dominate. I I if I was writing the book today I would just call AI only automates because it’s the same scenario. Um um and so that again it’s it’s deficits are a serious issue. So I wouldn’t want to portray our deficit issues are like there’s not a problem. It’s just that deficits are conditional on other things going on in in the world. And so um we could we could have five or 10 years where they don’t materialize in terms of higher interest rates and so forth. Um uh and we have seen it before but it does rest on AI becoming a general purpose technology like the personal computer and it’s got to augment our work. It can’t just save us time. I want to take you back onto some of the quant sides of what you just said because there’s a few more stats that really I think are eye opening. 2% growth and 2% in inflation going forward. You said a 10% probability of being correct. And then I’m lumping in too much, but you’re going to unpack these together on the deficit point. A 20% probability that the 10-year Treasury yield could reach over 9% in the next 5 to 10 years if AI disappoints. These are not your standard consensus takes here. No. And again, this again, I we were I was not and we were not at Vanguard looking for this economic diagnosis. I was very comfortable in my 2% growth, 2% inflation planning world that it always seems made sense to me. We may get a little bit of technology lift. Yeah. But we got all these negatives that we mentioned before on the other side. Like like I wasn’t like taking it not seriously. Like I always thought 2% growth, 2% inflation. Uh I I fill out these consensus surveys. Um that was our forecast until this work. Um what it’s been eye openening through the data is that it’s just very difficult to generate that sort of steadystate status quo forecast because of the push or pull either the trend for growth is going material higher because in because of the innovation of AI uh that overcomes the demographics and the debt levels we have that is by far our most likely outcome. Um but if it if if it only if if AI only manifests along certain techn if it only automates which means it has not become a general purpose technology we use it all but it’s it’s like the farm tractor um that would be disappointing and then that sets into motion this eventually not not today not tomorrow but you get fiscal pressures because they’re already high um and other things would have to occur um but we saw a little bit of this with so again I don’t that you know that’s that’s more scenario real planning and and like sort of um stress testing our portfolios um should should that scenario occur. It’s it’s not nearly as likely as our baseline which is you know AI transforms. Um but but this is client you know uh portfolios where they have fixed income exposure uh equity exposure. We want to look at that. Now the ironic thing is is that in our simulations for inst what’s the optimal investment strategy if we go down this disappointing AI only uh you know shaves jobs and doesn’t make us more productive. Um ironic you mentioned higher interest rates. However, it would push you into fixed income actually if you’re uh if you’re trying to optimize and avoid that world uh shorter duration fixed income. Why? Because you have given that interest rate reset at some point for for a temporary period. You have equity market that has significant draw down risk. You’re talking about a world where AI has not become a general purpose technology that has led to massive earnings disappointment in the US equity market, particularly in the large cap growth universe. Um, and so you have a higher discount rate or the pressure for that. You have higher short-term interest rates and you don’t have the earnings growth and you have the earnings multiples coming down in that world. And so that it’s trying to it would push you into fixed income only because it’s trying to mitigate some of the draw down in US stocks. So it may seem counterintuitive. Hey, in a fiscal pressure scenario, you actually want to go into short duration fixed income. It’s like, yeah, but most clients have exposure risk to equity market draw down. And so it’s going defensive in that way. So it may seem counterintuitive, but the broader risk is an equity market. Again, this is a world that AI flashes brilliance, but ultimately only automates uh because the higher we have higher unemployment without the growth lift from new from new stuff. And um again, it’s not a scenario I like talking about, but it is, you know, that’s that 20 25% probability. It’s it pales in comparison to our baseline, but it’s the second most likely outcome. And we we have the data to look in real time is to say which where is the where are the winds blowing? Right now, it’s still tracking as AI will will will transform along three dimensions, but two of them we haven’t seen yet. the augmentation and the and the new industries. We haven’t seen them yet emerge. What type of things would you be looking at to get a sense that that augmentation implementation is actually starting to bear fruit? Like where would what would you be paying attention to in the data? Uh some leading indicators uh that that I would be looking at that everyone could look at would be uh three come to mind. One would be um you know you’re going to see disruption in the labor market. However, if you’re going to see job loss, it would generally be older workers, not the new entrance college graduates. A lot of talk about why that would be is that you still have new tools being adopted. Um, and so you would have but if augmentation is still winning, it means that new entrance to the workforce, I mean, let let me be honest, they’re not as expensive and there’s an ROI to them embracing these AI tools as a compliment, not just as a substitute. you would see new business creation not just in the tech field not just in Silicon Valley you would start to see massive um job create or job or just startup creation outside of AI startups so which have been massive 4,000 or so the past three or four years you would start to see it in areas I mentioned before it could be in healthcare it could be in finance um because that’s where the diffusion now we’re starting to see that the sort of green shoots of these new fields emerging new platforms new applications that are revenue producing, not just cost saving. And that that’s that’s the key. And then finally, you would there would be a headlines in the news somewhere that a new product, a new solution has been um discovered by human beings using AI. So not just solving math problems, which is which is very nice, but say a new medical treatment for example, uh comes up in part because AI tools were powered with say medical professionals. And so again, I’m just trying to paint a brush of those would be consistent of us continuing down this AI transform path. You got to see the business startup creation um and as well as the job application outside of the tech field itself. The tech field will disrupt its own field the most. So the pressures we’ve seen on software company stocks is not surprising at all. We’ve seen in every technology cycle and we we could give examples. Um and we’re not just seen in the disruption of even software and IT jobs. Again, that that is that that is not a surprise to us given the work we did in the past. Um but what we got to see is these augmentation augmentation and new uh platform effects outside of the AI space. I’m not criticizing the excitement in the AI space, but tell me how it’s being used for new revenue purposes. That’s both the alpha opportunity if you’re actively inclin actively managed inclined as well as um indicators to watch if we’re going down that path. Zoom me out because a lot of what we’re talking about feels like a US centric story but this is a this is a truly global trend. Talk about yes globalization or del globalization how this fits in the context. Well and again there’s obviously pressures on del globalization. there’s a clear, you know, we’ve called it secular for some time, tensions in some ways between the United States and China from an economic perspective and and other dimensions just economic. Um, you know, however, when we look at like the the countries that could benefit the most from AI, I think again you have to look we have to I would urge us to look in two phases. One is the production of AI as a technology. That’s that’s that’s uh that’s that’s chips, that’s picks and shovels, that’s software. Um obviously that push, you know, that that’s as beneficial to China in some dimensions as is United States. Uh besides those two, and I talk about the book, there it’s a race for a distant third. Uh no no third country is even close. um you know maybe maybe when you look at Taiwan through that same lens um uh but but other countries no but other countries could benefit in the second half of the cycle if you look out beyond two years and that tends to again be on the consumption so there it’s it’s going to be economies that um as we go into the phase two it’s funny it starts to push you a little bit outside of the United States not because AI has been trans transformational it’s been the opposite it’s but it’s it’s countries that have very poor demographic profiles. Maybe they do have some debt headwinds. Uh, and they they have the need for greater automization because of poor demographics and a high service-based economy. So, as I talk about in the book, you know, that if that sounds familiar, that’s that’s some countries in Europe. Uh, that’s that’s areas such as Japan. Um, and and again, emerging markets, it’s mixed. Um, it could be beneficial for again for China, but it could be, you know, it may not do as much for Brazil. So, it’s it’s it’s country by country. But at least I try to provide a little bit of a framework of what those criteria are. Um, and so it’s this is a global phenomenon. You know, general purpose technologies, if we’re right in our baseline, they they are global by nature in the same way electricity was or or or the computer was. And and so AI won’t be anything different. We were always taught in economics, I’m not an economist, but I did take economics in college that, you know, free trade was good. You know, countries want to focus on where their comparative advantage is. Yes, tariffs are bad. Um, you know, dead weight loss, but it seems like it’s weird. It seems like like the market, maybe it’s because of this whole it’s like kind of shrugged off like all that trade related like economic uh theory. I guess it’s been a little perplexing. I mean, I can tell you what our projections we saw the I mean, the tariff increases we saw last year were were multiple standard deviations even in our long run data set. Um we never had recession as a baseline in part because trade although very important is a small share at least of the US economy five or perhaps 10% by certain measures higher from a from an S&P earnings uh footprint but nevertheless and but I would say the same thing with oil prices as well given recent events both tariffs and oil prices have not had a material dent um in the economy but again I think that depends they could have if we didn’t have AI investment continue to accelerate which is why I think the power of having all these factors in the same ecosystem you know we’re not going to be able to anticipate all these you know geopolitical shocks but when they occur you can at least say this is all around a magnitude question the direction of tariffs can be negative for growth negative for inflation but all else equal and and we know that the world we never live in that world and so how strong or weak are the other forces and um and so I this hasn’t been a surprise to us that growth has held up generally generally um um because because of our investment expectations on the AI front which are trillion dollars in growing. So but if we were if those shocks had occurred for tariffs and oil prices if those shocks had occurred I’d say five years ago uh we could have very well been talking about uh recession or or being very nervous about it. Um it’s not that those forces don’t matter. It’s just that what’s helping potentially to offset it. Um um but again some of those effects were delayed and some of the the fact is is that those two factors oil prices and tariffs they matter for the US economy but it is it is so diversified that it matters less to the United States than say the UK or parts of Europe which we’re seeing significant effects of those two shocks. And so it it it does vary where we are in the cycle and what else is offsetting potentially some of these headwinds. But I think it’s a very important point because as investors when we’re looking at the headlines, we’re watching, you know, it’s that point in time that a lot of times investors are paying attention to. So the tariffs or the spike in oil, but to your point, it’s like in the context of the entire economy and the market, there’s so many different moving pieces. Yes. and the magnitude of different things, you know, investors can kind of lose sight, I think, of of some of those things that can be different than what people think are how it’s going to play out. Yeah. And if someone’s trying to do again, I you know, I’ve been in the business 20 years. I And again, we have some, you know, some deep, you know, empirical frameworks, but I’d say let’s say we even have those frameworks. I what I and again, I’m a personal investor, too. Um, and talking with my adviser, you know, helped me uh get closer to retirement. Um, you know, I I I would argue we all have even mentally a multifactor scorecard. Really simple. So, you could have your oil prices that you’re concerned about. You could have geopolitical tensions that you care about or worry about. You could have our deficit issues that you worry about. And you have your other factors too. You have the AI threat as well as opportunity. And you go down the line. I why I think it’s important to put all those factors good and bad on one ledger is that and our our framework empirically is doing this like like naturally is that what that what that what I’ve always found when I do that is that it it doesn’t lead to drastic changes in one’s portfolio if I had just rather versus if I had just been looking at one of those factors in isolation right because if you see those headlines it could come out tomorrow oil prices hitting 120 not not $100 $ 1110 10 you say oh wow this is going to be really going to hit growth equity market I’m going to get defensive it’s like okay but but that factor versus those other ledgers now if you put that up against the AI investment in that six months which is unlikely to slow down now you may get to a different conclusion so I think at least having a heruristic that that tend well that that’ll tend to keep you closer to the policy portfolio it doesn’t mean uh at all by the way that one doesn’t change uh one’s you know uh investment exposure, but you you may view it from a riskmanagement perspective, more at the margin or less um near-term oriented. Um you may still be worried about fiscal deficits, for example, in that in that rubric regardless of what’s in the headlines today, but there would be ways that you could kind of mitigate that or I would say for listeners with your with your advisor, right? They would be able to walk them through that along with all the other goals uh that they would know better than anyone with you. I I think I would that’s how I would think about how macro meets the portfolio um is doing that in context also versus the market euphoria or or pessimism uh at any point in time. Again, generally when I do that, it it keeps me closer to my benchmark. It does not mean I’m not going to make some changes for opportunities or for risk management, but they’re not going to be as drastic if I had just looked at one factor uh in isolation. that that would lead me to really whipssaw portfolio more often than not. That being said, what’s it mean for the standard quote unquote 60/40 portfolio and what’s it mean for what an updated policy portfolio could look like reflecting all that you just shared? Well, I think that’s a you know that’s still that well that’s still a viable portfolio. I mean you can pick your sort of real return target as listeners. Um, like for example, for me personally, given my goals for the bequest for myself or my my retirement income, my wife and I, and then our goals in life, I’m closer to an 8020 period, regardless of my views on the market. Um, that’s just given our personal situation and my risk tolerance. Um, uh, but let’s say it’s 60/40, but I just picked that in the book as just it’s a fine representative benchmark. Um, uh, you know, there there there’s there’s great there there’s there’ll be a fine portfolio. there’s some risk uh to the 6040 in fact to to any it’s really because of the worst in the equity market but that’s only if AI only automates so it disappoints in some of its hope um for a time it’s it’s manageable um but there would be a period of disappointment that’s really on the 60 uh area not not so much on the 40 uh which is why you know I talk about in the book um if you’re going to worry about all states of the world you you would you would just you would you want to fade a little bit the euphoria that is in um the the mag and I’m not picking on these companies on the magnificent you know the the large cap growth companies not because they are not adding value it’s quite the contrary it’s because as it spreads opportunities will unfold so again you know the baseline is that it’ll be fine we may have some turbulence here but you know on the near term it’s upside risk on on the on the 60 uh segment but at some point there’s going be a market consolidation. So I I’d say just thinking about balance and risk mitigation. Um uh and I I think I think the biggest question for I I think value having with clients is two things. One is tell me talk to me about the two or three states of the world that have a significant odds or non non- dimminimous odds of happening and how would my portfolio weather them and is there modest sort of tilt in the portfolio that I could add for risk or or mitigation or alpha you know generation um um including active management um but that I think that’s like almost regardless of what AI does as a as a as a as consideration set but 6040 I don’t see changing as a reference portfolio the fact is most active managers uh don’t outperform it um but there are opportunities um uh for investors um regardless of how AI plays out so Joe we have two standard closing questions we like to ask all of our guests but before I get to that I do just want to point out uh to the listeners in the audience that in addition to the book Vanguard has an excellent what I would call like a mega trends research Arch hub where there’s data, there’s some actually other videos of Joe and there’s some, you know, interesting research and PDFs. So, that’s that’s just free on Vanguard site. So, we certainly encourage people go there to buy the book, too. Um, but of course, you know, if you want to start with the with the research hub, um, that’s a great place to get some really cool information. So, um, the first, uh, closing question we like to ask all of our guests is, what is one thing you believe about investing that most of your peers would disagree with you? Disagree with I do like that question. Um, I I would say that um that there’s a strong linkage all the time between what’s going on the economy and what’s going in the stock market. Uh, the fact is that those two can move in different in different uh different waves. um in part because the stock market can be anticipatory of what has not yet transpired in the economy. I think we’re seeing some of that today uh and quite and then it can go in the inverse at some point in time. So that would be something I would I would I would argue there the two are seem to occur one for one GDP with stock markets but I I think most you know most most readers ultimately would appreciate that those two don’t always move in lock step there one thing you could teach the average investor put it on a billboard fly it on a little plane on the sky over the Jersey shore. Um I would I tell you it comes from Ben Franklin uh but Jack Bogle would would uh big advocate for this too and you’ve talked about it yourselves and that’s the power of compounding. I mean even I could be the most savvy investor and and move around the market and time. The fact is how much money I put into my own portfolio and saving and letting it rest and let the capital markets do their job is going to dwarf any so-called alpha or out, you know, outsaving the average investor. And so that’s something I continue to remind myself. Put as much money in as you can and let it do the hard work for you. So compound interest. It is the biggest asset we have as long-term investors. One more thing about the book before we let you go. the proceeds going to charity. Yeah. What what was the thought there? What was the strategy? What was the idea? Well, again, no one even asked me. You only get one check at Vanguard. I’m a Vanguard employee. So, that was that that was the first or order. Secondly, uh you know, it was just as you know, Vanguard, we believe in uh our community as well as our clients. And so, you know, all the proceeds go for Vanguard Strongstar for Kids. This is young children. Uh they may not have you know the financial and their parents the financial resources for early childhood education. And so um if if that can help um you know you know some young children uh I couldn’t think of a better uh place for the proceeds to go. I think that’s an important point to end us on. Joe, thank you so much for your time today. Uh thank you for having me. Hey, if people want to bug you on the internet, find out more research on this. Can you tell them one more time where they can get this? Well, I think what you said on the Vanguard hub and then every year we’ll have our annual outlook, but you will see AI and these other factors prominently in there. We’re looking out not just for the next year, but the next several as we all are trying to navigate uh this universe. Advisors, allocators, uh do-it-yourselfers, you want to check out these resources. They are very, very cool. This is one of the reasons why Vanguard is such an interesting company is because of how much of this stuff they put out there. Joe, thank you so much for the time. You are watching Excess Returns. Like, comment, subscribe, all the things below. And we are out. Thank you for tuning into this episode. If you found this discussion interesting and valuable, [music] please subscribe on your favorite audio platform or on YouTube. You can also follow all the podcasts in the Excess Returns network at excess returnspod.com. [music] If you have any feedback or questions, you can contact us at excess returnturnspod@gmail.com. No information on this podcast should be construed as investment advice. Securities discussed in the podcast [music] may be holdings of the firms of the hosts or their clients.