Transcription
Merryn Talks Money listeners. Are you a Bloomberg subscriber? If you're not, here's why you should be. You'll get ad-free episodes of this podcast and access to my Merryn Talks Money newsletter, as well as access to John's award-winning newsletter Money Distilled. And you'll get access to subscriber-only events, such as the one we'll be hosting on March 17th. See the link in the show notes below to sign up to that. And of course, you will get unlimited access to Bloomberg.com and the Bloomberg app, including exclusive stories and premium market tools. Subscribe now at Bloomberg.com/PodcastOffer.
Welcome to Merryn Talks Money, the podcast in which people who know the markets explain the markets. I am Merryn Somerset Webb, editor at large for Bloomberg Wealth. And this week, I am speaking with Edward Chancellor, investment strategist and financial historian. He is the author of The Price of Time, which I'm sure all of you will have read by now. And if you haven't, now is the time to go get it and start reading. He recently collaborated with Jeremy Grantham on the book The Making of a Perma Bear: The Perils of Long-Term Investing in a Short-Term World. We interviewed Jeremy a few weeks ago, so do go back and listen to that one as well. And we thought this was a good moment to have a financial historian on the show, someone to help us pinpoint the market moves we've seen in the first quarter of the year into context, both the volatility around AI, but also more recently, obviously, the reaction to the escalating conflict in the Middle East. What does that mean for markets going forward? And more importantly, from Eddie, what does it mean for the price of money? Edward Chancellor, welcome to Merryn Talks Money. Nice to see you again, Merryn.
Now, we've got quite a lot to talk about, actually. Originally last week, or a couple of weeks ago, when we thought about having you back on, we thought, well, we'll just talk about bubbles. We'll talk about AI, we'll talk about the great rotation and stock markets, etc. But there's quite a lot more to talk about now. But let's start where we originally planned to start, shall we? Let's start with before the conflict, the new conflict in the Middle East started at the weekend. We were looking at markets and going, there is something going on here. We're seeing a shift out of AI, AI-adjacent stocks and into something that connects more to real assets. We talk about the shift from YOLO and FOMO to HALO. All these things. Do read columns that John and I have written on those things. So I wonder if I could just ask you to give us your your thoughts on that, what you think has been happening in the market over the last few months.
Well, and the story about the market's response to AI, as with every sort of boom or bubble period, there are feedback effects and and one narrative runs for a while and then another narrative takes over. So I'd say at the beginning of this year, and I wrote a paper with Jeremy Grantham looking at AI from the perspective of previous technology booms, and it looked at the time, you see that, you know, Oracle stock could come down sharply. CoreWeave, which is one of these so-called new cloud operators, that stock was off sharply then. You know, the big hyperscalers, the market was responding negatively to the amount of money they were spending and that collapsing free cash flow. So you, it looked, you could think quite reasonably at the time that that the things looked as if AI enthusiasm was waning. And then, but then in the last couple of weeks, we've seen really quite extraordinarily powerful viral AI stories, narratives. And there was one one report or note by Matt Schumer, who's some Californian tech guy, which was the story of AI is coming for your job, very conventional story, but immediately got, you know, 50 million views on day one. But more interesting to my mind was the research that came out, a piece of research that came out I've had called Citrini Research towards the end of last week. And that was that was mooting the idea that if AI were ready to realize its potential, there would be a collapse in all white-collar jobs followed by the mother of all recessions. And that she is definitely worth reading as a as a story of, of what would happen if AI were actually to achieve what the likes of Sam Altman and OpenAI claim that it will achieve. I personally do think it will will get there, but and we can discuss that.
And why don't we discuss that right away? So we know, so we can set this up. I mean, the idea being that if AI can really do what its intense believers believe it can and effectively take over pretty much all white-collar jobs, massively improve productivity, then that gives us a huge level of unemployment and a recession as a result. So the productivity revolution destroys itself.
No, but it's further than that, is that and this is the point, is back to Schumpeter's theory of the business cycle, that the Great Depression is created when a new technology comes along, interrupts established profit lines, and then the debt built upon that profit or cash flow collapses. So in fact, the Citrini research is really a sort of Schumpeterian dark picture and very well expressed. So why do I think I, humble I, who can't code to save their life, I think it's not going. So why do I think that it's it's we're not going to see the Citrini research scenario played out? And I think this goes back to, I think it goes back to every every bubble. If you think about bubbles, there's there's always a there's a core of truth to every bubble, that something is revolutions, something's going to change. What. And then there is an element of reality that is being overlooked. So the market, if you will, chases the narrative and it chases the narrative uncritically and it doesn't examine the niggling elements of reality, which, and that is the nature, I'm afraid, that is the nature of a bubble. So you call it groupthink, or you can call it sort of people's resistance to dissonant information, cognitive dissonance, call it what you will.
So what is the job or what's the reality that the AI enthusiasts seem to be overlooking? Well, that's namely that, as far as I can understand, bear in mind, you know, that I'm picking this up at second hand. But if you listen, if you listen to the certain of the specialists, and I'm thinking of Turing Award winners, John, the woman who left Meta last year, that echoed Richard Sutton, also another Turing Award winner. And then I listened, and I recommend that your listeners look, listen to it too, the Faraday Lecture, given the role started last week by someone called Michael Wooldridge, who's professor of AI and Computing at Oxford. And all all of them really say the same thing, that the large language model is not a model that is actually engaged in reasoning, is not engaged in seeking the truth, it's not engaged in logical processing. It uses, and this is where I get a bit fuzzy, it uses really a sort of probabilistic framework for guessing, so to speak, what comes next. And and it does an incredible, incredible good job a lot of the time. But it also produces these pesky hallucinations. It makes mistakes and it doesn't correct those mistakes. And it can't, so far, how I see it, is not learning from having made a mistake. Now, people like LeCun, Sutton, Wooldridge, they all say that this flaw is inherent to the large language model architecture. And so far, the advances with the LLMs have come from using more and more compute, more energy, and scaling everything. And that led to advances. But there are, to my mind, quite compelling arguments that we are that we're reaching the limits of scaling. And that's just that's not me saying that. The person who was saying that towards the end of last year was Elias Gaither, who's one of the founders of the LLM model. So think about this way. If these machines are not really reasoning, are not really intelligent, and have and whose tendency to create errors or hallucinations are built into the model, how far can you go with that technology?
Okay. So the key point here is that, you know, is the new wind nice to have around the edges, makes a good story, but won't actually change our world?
Well, it's difficult to know where to draw a line. I mean, the other thing I'd say, Merryn, is this, that I've been doing quite, I'm writing a new chapter for my old book, Devil Take the Hindmost, the history of speculation. So I'm going over the whole Internet bubble of the TMT bubble of the late 1990s. And one thing what I find interesting there is that of course, the huge amount of hype around around the Internet from the mid-1990s onwards. And if you, if you look back actually, that hype was justified in the end. If anything, the internet exceeded the expectations of all except the most feverish exponents of the technology. And yet, you know, the Nasdaq was down 80% after 2000 and took 12 years to regain its peak. And even Amazon, the eventual winner of the early dot-coms, dropped, you know, 93% before coming back. So the timeline, even when the technology actually delivers, the timeline tends to be wrong. Markets get sort of too feverish about something.
Yeah. But can I just stop there and say to you that one of the things that you and I have talked about before is that you have to have a bubble to build a new infrastructure. There isn't any other way to do it. You need overexcitement. You need that that that fever to actually get the money together to go out and build these very unlikely infrastructures. I would, I think, as we get into the modern era of technology, like the Internet, I mean, think of the Internet. They paid these companies later, hundreds of thousands of miles of fiber cable. Now, the businesses that did it, you know, Global Crossing and and other companies, they all went bust and there was massive capacity, I don't know, 95% overcapacity. But yes, then you had the infrastructure for, you know, you know, for YouTube and and video and, you know, video conferencing and whatever.
You and I like to see things from the point of view of investors rather than society at large. So the my view is that if the benefits are delivered for for consumers or society at large, there's no need or requirement for any individual investor to provide the capital with the expectation of loss. That's my, you know, that's my caveat to, too, that, you know, the bubbles being necessary. I don't think one, I don't think there is, let's say, a necessity to partake. In fact, actually, I was mentioning earlier to you the Alisdair Nairn's book, The Engines That Move Markets, which is really a tremendously good book, which I would recommend to anyone who is interested in the inception of how markets respond to new technologies. And it's really quite clear that in the early stages of these new technologies, when there is, so to speak, the bubble enthusiasm and that investors don't make money, there tends to be overinvestment, excessive competition, and then eventually you get a a fallout and a shakeout. And and then you get dominant companies appearing at that later stage. And those dominant companies often make pretty good investments. You think of AT&T dominating telephone or and we think, you know, obviously more recently of the likes of Facebook and Google. But but Google and Facebook were not on offer in in the late 1990s. It doesn't pay. You know, frankly, when a new technology comes along, the investor, almost invariably in the past, would have done better just to stay in bed for ten years and get out of bed and say, What is this thing? What is this railroad I'm going to buy some stock in? It's just it's just better to hold back. There is, I was thinking that Warren Buffett and Charlie Munger, they all they say about technology, that they had a trade thing too difficult. And really, when one looks at the outcome for a given these uncertainties that I mentioned earlier about what its capacities might be, and also given the huge amount of investment that a number of extremely well-funded companies are engaged in, and the prospect then that probably quite realistic prospect that, you know, that the services they offer will be particularly differentiated and therefore will, will, will probably their marginal, you know, marginal cost, which might not even pay back on the capital they've invested, then, yeah, why not? Why not stay? Why not. Why not go for a halo trade instead?
Is that where we are now? Where we were, let's say on Friday, on the idea of investing in in AI and AI-related stocks or not? Which stage of this bubble cycle are we in?
This is definitely a different bubble or inception of new technology than the ones we've seen in the past, because, as you know, a lot of it is being driven in the private markets. I mean, we've never, I mean, last week or a couple of weeks ago, OpenAI raised, I put it in inverted commas, $110 billion.
Why are we having inverted commas here?
Oh, we're having inverted commas because the, well, first of all, there was $50 million, $50 billion coming from Amazon, of which of which $35 billion was contingent on OpenAI achieving artificial general intelligence. Well, I told you, there are a lot of people who seem to know about AI who say you're not going to get there. Okay. So that's an inverse. The other thing is that, you know, that that OpenAI is the world's largest vendor financing operation because and Nvidia was putting up, I can't remember how much it's $35 billion, whatever, and and and OpenAI was committing to using Amazon Web Services and and buying three gigawatts or whatever or gigawatts of of Nvidia compute. Now, if you remember, you know, back in the late 1990s, the likes of Cisco and the other telecoms operators, you know, Nortel telecoms equipment suppliers, Nortel and and Lucent, they were all heavily engaged in vendor financing. Now, the vendor, I mean, I was reading yesterday, the vendor financing of these companies, which we thought we all thought was an outrage when it was revealed, if you remember. Yes, we did, actually, total $15 billion. I mean, it's this it's small change to compare.
Billion is small change. Now.
It is. But look, I'm just saying that that, you know, that OpenAI loan seems to be, you know, it is lost funding. It's roughly 8% of it, roughly $80 odd billion. What was was a sort of vendor financing operation. But wouldn't we love to see the valuations of these companies that are still private tested at an IPO? That's the other thing is that I should have said that Amazon, the $35 billion was to OpenAI was contingent on either achieving AGI or IPO. Now, these big companies, you know, they, I think OpenAI's funding, you see different figures, but I think that around $730 billion, some say higher, of the of the valuation. So the question then, you know, what happens when, if and when the likes of Anthropic and OpenAI and SpaceX, if they all come to the market and get it in free in a couple of other businesses, and you're you're getting $3 trillion all of of new in a market cap coming to the markets? Whether I mean, another scenario, another source, Citrini type scenario, is you actually have a sell-off as all these as these companies come and, you know, and push into the public market. But on the other hand, I was thinking that, you know, with the world, in particular the US markets, so dominated by passive investment, they say roughly 60% of the US market stock market is passive nowadays, of course, the passive. Investors, the passive funds just open up their arms. There's nothing you can't sell them in some ways. And perhaps, you know, perhaps that some of the thinking is to build up to these huge trillion-dollar valuations and then pass them into the market. And so once they get to that size, the IPO automatically works.
Yeah, I'm wondering and assigned to think about and the other, you know, again, my my view from an individual investor's perspective is that you're already picking up AI exposure. According to JPMorgan, last year, roughly 45-47% of the US market had exposure to AI. It would it would go up considerably, probably to 55 or 60% if these if these AI companies were actually to achieve their IPO. I think my view is that the individual investor has other things to think about because there are, as you know, other opportunities in the world. And frankly, the the the the passive blind capital of the index funds will absorb the AI for better or for worse.
Yeah. Okay. Let's move on to talking about what's happened over the weekend. I mean, we've talked about AI in the past. I mean, I'm interested that your your view of the limiting factor of the future of AI is, in fact, the models themselves that are inherently flawed. So would do what the biggest optimists think they will do. But the other limiting factor on AI has always been energy use, right? And that's been discussed a lot. And now we're finding that there are massive constraints on energy globally, which I think most people knew before, but those you hadn't recognized before. The the energy fragility of our world are, of course, learning that now. And you can see that fear reflected in markets. I mean, for example, the fairly massive fall in the Korean market this week is about the fact that it's very heavily dominated by memory chip companies, two big companies, and the majority of their energy is imported. And that now becomes a major constraining factor on that manufacturing. So we now see this constraint across the board, and that's going to have quite an impact. Right. Brings us back to HALO, of course, on the energy usage of of AI. It's that's that's a political concern. But before the Iran war breakout, so the US electricity prices have been have been going up strongly, that's causing political ructions. And there are, you know, the question is, will the, you know, the American populace put up with much higher electric, domestic, you know, domestic electricity costs out of the AI rollout, the data centers? And the answer is probably not. And, you know, last week, President Trump said that the data centers are going to have to start generating their own electricity. And that that, you know, that should be a good thing, I think, and particularly if it encourages the development of, you know, small-scale nuclear and and other, you know, reliable alternative sources of energy. We've heavily underinvested in conventional energy and, if you will, overinvested in the unreliable alternative energy. And and therefore there's a bit of know fragility in the system. And when that, you know, that fragility, then you get exposed by, you know, by a war in the Gulf. And it's interesting that been two responses haven't made to the to the war in the Gulf when it comes to energy. One response in the UK is, oh my goodness, I wish that we were doing more to exploit our own fossil fuel energy resources. And the other response is, we'll be needing more wind and solar to deal with this.
Yeah, I got I got an email from my nephew last night saying that due to that effect, we need energy security. And the answer is yes, we do. But the the unfortunate fact, and again, goes back to the bubble, that, you know, the important salient facts being missing from the bubble report is that, you know, you're not going to get energy security from from wind and turbines. And we, in fact, Britain is dependent on importing ever larger amounts of of electricity from the continent, you know, often from at gas turbine power stations. I agree with President Trump. I think we need to utilize our our own fossil fuel resources if we're going to have energy security, at least given current technology.
Mm hmm. Okay. Now, what does all this mean, do you think, for the interest rate cycle? So before the war in Iran, there was much discussion about easing cycles across the board. We were expecting to see how some people were expecting to see rates come down fairly significantly in the UK. That now seems at risk.
Well, let's get right to the AI discussion. Citrini, Citrini Research says that you're going to unleash the mother of all depressions, and that'll be a deflationary depression. And that would be a deflationary depression that was resistant to central bank impetus. So under that scenario, if AI were to realize its promise, then you might say, well, perhaps interest rate, you know, inflation is coming down, deflation will replace it, and interest rates would come down. Now, that's a sort of theoretical position. I think, you know, if we have energy and commodity inflation, then you would expect interest rates to rise on the back of, you know, in the old days, they never referred to inflation as a consumer product, they always referred to it as commodity price inflation. So I think if one, if one is going to get higher energy prices and high commodity prices, then one would expect the interest rate cycle to to pick up. The only thing I would say that the thrust of my book, The Price of Time, has always, you know, no one knows how to predict interest rates. What we do know. The only thing I like to say, the only thing we know is that interest rates run on very long cycles. And I'm talking about long-term interest rates, bond yields. And we had a a good 40-year downward trend from 1982 to 2022, and we're now in the fourth year of the uptrend. So my, my, my view has always been, regardless of what anyone tells you, is we're on we are on the uptrend of interest rates over the decades, and there will be oscillations around it. And those oscillations are inherently unpredictable.
Mm hmm. Okay. I'm feeling stagflation and UK public debt crisis.
Yeah, I've been, you know, of that view. And incidentally, no, another, I, you know, you and I are both sort of gold bugs, correct?
Afraid so. And as you know, I'm not afraid. So I'm thrilled to say I'm a gold bug. This is our time, Eddie.
But Merryn, you know, I, I actually, I sort of wake up every morning. I, I think all these different scenarios of the world. I think actually yes, gold fits. Take the Citrini model, that oh, everything collapses and the government at government finances would collapse at the same time, then you'd want gold for that. Take the inflation model. Take the tape to take the inflation commodity, you'd want the gold for that. Government debt crisis, you want the gold for that. And so I do think, and I wrote a piece towards the end of last year saying that we are in a new paradigm for gold. And I think that I think that for asset allocation going forward, you people have just got to sort of wake up and have. You now have gold as the sort of, you know, a sizable gold position as the anchor for their portfolio. And when I say sizable, I'm talking of somewhere in the range of 10 to 25% of the portfolio. You know, our friend Russell Napier has 50%, which is quite high.
And does that mean that sovereign bonds no longer have a place in a portfolio?
I think some place, but a much smaller one. I mean, if you think about the 60/40 being 60% equities, 40% bonds being the traditional portfolio. My my I think you can keep to the 40 non-equities would go even higher, 50 non-equities, and then whatever isn't your gold position is left over for your sovereign. And then, of course, you know, you're not obliged to buy, you know, sovereign bonds of of near-bankrupt states like France or Britain. So you can actually buy buy sovereign bonds elsewhere. You can buy. And then you you can also choose to get. You can also choose to get inflation-protected bonds. But I would be in a I think, the long term. Yeah. I suppose my long-term present with the bonds is that we are unique heading towards a sovereign bond crisis in the among the advanced economies. And even if you own the inflation-protected bonds, there's no guarantee that, you know, that when government is slashing their pensions and and and slashing other forms of discretionary expenditure, they they turn on the wrong tier and say you should take a cut too. So I'm very wary. And I over the last few years, I've been cutting down my, my, my inflation-protected bond.
Yeah. And as and you say we're heading towards debt crisis in various developed economies. Do you feel that getting closer?
And we've been discussing this public debt problem for some time now. We've both, I think, pretty much everyone now sees that countries such as the UK really do have a very severe debt problem and no, no ability or apparently no incentive to attempt to cut spending to deal with that sovereign debt problem. So how close is it and what might what might make it go away?
Well. Um, I think I think it would appear much closer in France than than in Britain. Not just because, you know, France's tax take is is much higher than in Britain, and therefore for this less room to to to raise it. Political paralysis in in in France would appear stronger. And then you've got this the the Eurozone problem. The fact is we're now already in a position and you can see that interest rates in in Europe are too high for France and too low for the periphery. So you've got strong house price inflation in Portugal. Whether it was we have to see what what the outcome of the French elections are. But the I think the crisis first comes to France.
Yeah. But let's go back to the equity part of their portfolio. We talked about how much you should have in gold and being slightly wary of the bond element of your portfolio. And that. Well, I'm just on that. I wanted to ask you about one of the things that you've talked about when we've been discussing this kind of thing before is the cracks in the system and how super low interest rates get into the cracks and you don't see where the problems are until significantly later. I was wondering if you were worried about private credit as the source of one of our future crises.
I'm worried, and everyone else seems to be worried too. I mean, the private credit story, as far as I see it, is, is that it was engendered by the low interest rates and the we knew we had this period where bank deposits were yielding nothing. And there were these a variety of different ways in which you could achieve higher returns. And private credit grew and private credit grew to fund private equity in large parts. And the and the private equity, as you know, took over life insurers in order to have a, you know, to have a ready market to buy their. The private credit issued to fund the LBOs and and so, yes, I think that I think something like 80% of private credit is what they call sponsor-backed or related to private equity. And we know that in the new up to the end of the private equity boom, that that debt multiples were rising. In short, I. Yeah. I it wouldn't surprise you to know that I would not be, and I've never been advocating private credit exposure. And with regards to my, you know, to the thesis of The Price of Time, that it was, it was always that there would be a crisis eventually coming out from the interest rates, the ultra-low interest rates getting into the cracks. The only surprise, I mean, it's a big surprise to my mind, is that, you know, it's taken quite a long time for these markets to open up that we are.
Yeah, everything I would say is longer than you think this kind of thing, doesn't it? Let's go back briefly then, because I've had you far too long and I must let you go, but to the equity part of a portfolio. So, you know, here we are, tricky times with that 50% of your portfolio that you're holding in equities one way or another. How do you manage that today?
Well, the the one the other idea of of The Price of Time was with interest rates very low and huge amount of suits, big financialization, private, private credit that one. And with with with money going into these ethereal assets.
Can I use inverted commas again?
Yes, you can. You can see you're right. You can keep using them. And so these that theory that one wanted, one wanted investment had to get real. And so I told in the 1930s, in the 1920s, tonight plantations of Germans talked about the deep, deep, deep looked in in exact that and took the flight into things of things, the real values. I was I was always in favor. I have for a while been in favor of it, commodities, traditional energy companies, and in particular, this not the themes that which you'll be aware of, the whole sort of capital cycle story of underinvestment, huge boom in both, you know, in energy investment and mining up to about 2013, 2014, and then this, you know, this ten, 12 year famine. And I think so for me, that has been a core part. It's the core part of my portfolio. It's the reason is, and it's done almost slightly too well. It's sort of I'm now a bit worried about it. And then you're like, you. I'm a sort of Japan bug. I like the story in Japan. And, you know, it took it took what? That it took 30, 30 years, you know, for the for the bubble of the late 19 of the late 1980s to to wear off in Japan. And where we're now a few years into the recovery of Japanese equities. And it's not expensive, and they're restructuring, and and so I'm I'm I have, you know, pretty big holdings in in Japan. And I like emerging markets, too. I mean, I've been in the last couple of years, probably most money I new money I've invested largely in in emerging, you know, not attractive 15 years ago. But, you know, one things change. And and I think, again, you know, we're we're a few years into an emerging, you know, I mean, who needs I mean, to get up and down. Yeah, but but a few years into an emerging cycle. And go back to what I was saying earlier about energy is I don't trust economies that have, you know, hugely uncompetitive energy costs. I know the US, UK equities are cheap, and I know that European equity is cheap, and I'm just like, you know, our friend Andy's macro strategy, I think that, you know, economy is the transformation of energy. And if you, if you push energy costs, I'd be I'm beyond control, it's not it's not the only thing, but if energy costs are too high, you cripple an economy. So I'm that I'm just where I, I actually like to look at a chart, see who has the lowest energy costs and just invest there. Not just invest there, but, you know, then think through it, you know, and and emerging has, needless to say, right now, energy costs.
How interesting. Slightly brings us full circle, doesn't it? Good. Thank you so much. That was really interesting. We loved having you with us today. Thank you, Merryn.