The Capital Cycle Podcast - Capital Account Book

Episode Date: February 9, 2026

A discussion of Marathon’s first book, published in 2004.Edward Chancellor talks to Charles Carter, a European portfolio manager. For more information, or to access select articles from Marathon’s Global Investment Review publications which accompany this podcast series, please visit www.thecapitalcycle.co.uk Hosted on Acast. See acast.com/privacy for more information.

Welcome to another edition of the Capital Cycle podcast. This is a special edition. We're not going to talk about the global investment review pieces that we normally do, but we're going to look back at a book that Marathon published in 2004 called Capital Account. And I have with me Charles Carter, who's a European portfolio manager at Marathon, who helped put together this book, capital account. Capital account over the years has developed rather a occult following, and that's largely because it's very difficult to get hold of. The publisher we used went bust

during the publication and relatively few copies made it to market. The result is in the secondhand market, copies of capital account sell for upwards of $4,000. So there is this scarcity value to the book. But I don't think that the market's interest in the book is pure scarcity value. It also has to do with the fact that this is the book that first introduces capital cycle analysis, which was generated within marathon asset management by the founding partners. And it introduces this idea to a broader public. And I'll cite, for instance, say a recent tweet from Michael, the investor, Michael Berry of big short fame, who towards the end of the last year put out to his followers the fact that he'd read

capital account. He writes, I have found capital cycle theory to be a solid framework for analyzing mania-driven capital investment booms. As well, the book is an excellent firsthand source material for understanding the true history of the 1990s boom and bust. And later in the conversation Charles. We'll talk about the dynamics of the TMT boom and bust of the late 1990s. But first, can you tell me a bit about what you remember of the genesis of the book? We've been writing these articles for what we call our global investment review eight times a year. Each one had four or five articles in it. These were really sort of battlefield reports. I think they were once described as looking at the investment world through this lens of our investment.

in philosophy, the capital cycle philosophy, I think really after the bubble burst and we looked back at the reports that we'd written, we felt that actually it'd be quite interesting to put them together in a kind of compendium. Initially, the idea was that this would just be something we would give to our clients. I think you and I then landed up having a conversation and you had become the sort of dean of bubbleology through the book that you had written devil tape behind most, and you were interested in this particular period, we thought, well, actually, this probably might deserve a wider publication. Yes, and just add to that, what I felt reading the marathon reports at the time,

is, as Barry says, that they provided an excellent firsthand account, so investor in the foxhole account of what it was like to live through the bubble period. This was an era in which we were, if you remember, being deluged with reports, but these reports were the most incredible ephemeral items, none of which exists today, I imagine. And so here was a chance to actually preserve this count. Now, capital cycle analysis is really a bottoms up analysis, microanalysis, of individual companies operating within sectors. But it also turns out to be a brilliant tool for spotting bubbles as they're forming, because the great bubbles usually involve sizable amounts of capital spending.

And it is this investment spending or capex that tends to doom the bubble rather than the valuation itself. So Charles, what's interesting when you and I were both rereading capital account, we were then reminded of the actual origins of the capital cycle theory in an early tech bus. Yes, I think this was the experience. of one of the founders. So before he founded the company, he was working in the US on the West Coast dealing with the busted remains of what had been the PC boom of the early 1980s. He was coming to work and having to deal with the latest, what he called, high-tech horror of the day. There's an article in capital account which refers to this period and highlights one of the

problems, which was that the depreciable lives of the assets turned out to be much shorter than people believed, and that created obviously a lot of problems for that industry. And it also made reference to some of the kind of buzzwords that were flying around at that time. And they included many things which seemed very quaint and old-fashioned now, modems, floppy disk, irma boards, whatever they are. And actually, artificial intelligence was mentioned. The dirty secret of artificial intelligence is it's been around for 70 years going through occasional springs and summers followed by autumns and winter. We're now currently in one of the cyclical summers. But the thrust of the argument there was that new technologies don't always make good investments and often attract excessive competition.

But the capital cycle analysis was also used once Marathon was up. and running to analyze the boom in Southeast Asia amongst so-called tiger economies. Now, you remember at the time the early 1990s, the World Bank and others were saying that these tiger economies were in a period of endless benign growth. And Marathon, by analyzing individual companies operating in the region, had different views. So would you tell us about Siam Cement? Yes, so he features quite a lot in the book. And this company was a leader in a market, the Thai cement market, where there'd been this capacity boom. And the country landed up with twice the cement capacity of the US. And it was something like one ton of cement

capacity per Thai individual. And the company had also, in this euphoric period, had diversified into many other areas, often also suffering from excess capacity in the end. Steel bars, pulp, car, tires, what have you, and accumulated quite a lot of debt in the process. And I think in the summer of what was in the summer of 1996, Marathon in one of the GIRs wrote that the consequence of prolonged overvaluation is undisciplined asset expansion. And this has now reached such levels that a complete collapse of industrial profits in Southeast Asia stares investors in the face. And Thailand, as I say, developed this exercise.

capacity across a number of sectors and the macro call in a way from a capital cycle perspective was to avoid investments in that country. And no sooner had Marathon started to flag problems among the tiger economies and overinvestment, that its attention turned towards telecoms. Marathon gave a pretty early warning in 1994 about what was happening in the telecoms market. This was again looking at the supply demand dynamics within telecoms. And as early as 1994, there was a comment about how the laws of the capital cycle are such that in a deregulated environment, the price of a product will drop to the marginal cost of production and below for a while. Telecommunication services are moving in that direction at an accelerating speed, and the

elasticity of demand will soon be offset by the pace of price reductions. Eventually, this was the idea that telecoms had been this regulated monopoly business. It was then opened up to competition. And there was no real moat around the businesses. So they landed up competing against each other very aggressively and prices collapsed. Seven years later. Seven years later. So the contrarian message of capital account in the second half of the 1990s involved repeated criticism of investors.

obsession with short-term earnings. As I said earlier, capital cycle analysis is all about capital allocation and returns within industries and businesses. And yet in the second half of the 1990s, the market was completely obsessed with quarterly earnings. This was sort of, in a way, it felt like a game that developed at that time around quarterly earnings and companies were targeting EPS or earnings per share. And, the sell side would then formulate a kind of consensus expectation of that earnings result. And then the company would duly appear and deliver its results. And if it beat the expectations, that was great and the stock would go up. And if it failed, I think the expression was that it was

gaping down. The game evolved and became more sophisticated in terms of guidance, companies giving guidance. Which was later banned, I think, by the SEC. Yes. So it was a practice of selective disclosure where guidance was given to just the select few. And then the market responding to guidance, because companies would give guidance and perhaps underway the number so that they could get the pop on the day of the quarterly earnings report, the market then developed whisper numbers, and then you got into the whole world of pro forma numbers where companies wanted you to ignore certain things so that they could beat the number by excluding perhaps some unwelcome costs or restructuring charges or indeed stock options.

And the whole process of this bred a lot of short-termism. And I think in the book it refers to Jack Bogle commenting about how investor holding periods had shrunk from something like five years in the 1950s down to less than eight months. And I have to say, I mean, all of this stuff goes on today in spades. You pick up a typical analyst report these days. It'll be half a page of content mostly about whether a company beat the consensus earnings, or not and why not. Then you have 15 pages of disclaimers, which is the legacy of the sort of Eliot Spitzer era, if you remember, where these companies got into lots of trouble and then had to find a legal way out of their situation. But interesting, actually, one of the things

that Donald Trump has mentioned more recently is getting rid of the mandatory requirement for quarterly earnings, which strikes me as being eminently sensible on the basis that it creates a lot of regulatory cost, and of course, it fuels short-termism in markets. So the marathon line expressed in capital account really is also that of Warren Buffett, so the quarterly earnings per share tell you very little about the firm's intrinsic value. There's also an interesting piece on how earnings targets become corrupting, that they exemplify Goodhart's law of how, when a metric becomes a target, it ceases to become a good metric anymore. I mean, Charles Goodhart had been this economist at the Bank of England, and he had observed how any attempts they were making to tax or regulate banks when they were using a particular channel.

As soon as they did that, the banks would just shift their business to another channel to avoid it. So the introduction of the target affected the outcome and the same experience applied to the policies of the early Thatcher governments when they were trying to regulate the money supply and they would target a particular statistical measure. I think similarly with the EPS, it just encouraged various shenanigans, if you could call it that, of people looking to manipulate their earnings per share number in a favourable way. And of course, the prime example that people often give is Enron, which, in its annual report, in bold letters set out the statement that it was laser focused on earnings per share and that we expect the strong performance to continue. And of course, that got them into all this business of mark-to-market accounting, which they somehow managed to get the SEC to approve, and then, you know, a kind of accumulation

of larger and larger frauds in order to sustain the earnings per share number. So Enron wasn't the only company engaged in manipulate. its earnings in the late 90s. Coca-Cola, for instance, maintained an artificially high return on equity by keeping its assets among the bottler operations. General Electric, I mean, it's hard to remember at the time that Jack Welsh, the head of General Lecture, was really the most lionized executive of the era, but General Electric managed to increase its earnings for 13 years in a row. And in the... the book, this earnings manipulation, is linked to J.K. Galbrai's bezel. The bezel being,

as Galbrai calls it, the inventory of concealed embezzlement within the system that occurs during the boom period. So what was the bezel of the late 1990s, as far as you remember? Well, I think, yes. I mean, you had various practices going on at the peak of the market. You had vendor financing, a lot of circular deals. going on in the telecoms world, capacity swaps. Do you remember Bartah advertising, capitalization of R&D at WorldCom, off-Banchet vehicles you just mentioned? And then just the general use of derivatives, you know, so-called weapons of mass destruction,

at companies where they were bringing forward earnings and Green Tree Financial and Conceco were two examples of that. AIG later got involved, didn't it also? use of derivatives to bring forward earnings. Yes, and you had these, so these phenomenon in the peak period, and then, of course, when it all goes wrong, then you get the sort of gnashing of teeth and the slamming of prison doors. And let's not forget that during this period, US executive stock options schemes were not included in the profit and loss account. There'd been a huge forer in the mid-1990s, and corporate America put pressure.

of the US Congress to stop the wretched accounting standards board from bringing in a rule to include expenses. So at one stage, that was, I think that option grants came to roughly 10% of US earnings, and they were not being measured. By the late 1990s, as marathon identifies, there were three different earnings numbers out there. There was the pro forma earnings, sort of cooked up earnings that you mentioned. There were the audited earnings that had faults, let's say. And then there were the profits in national accounts, as in delivered through things like corporate tax returns. And what we find from 98 is a divergence between the earnings and profits that are reported by companies to the investing public and those that are reported by the same companies to the US tax

authorities and the official earnings are coming down. But the other aspect of another earnings metric that comes up at this time, still with us today, that Marathon took aim at, is the dreaded earnings before interest, tax, depreciation and amortization or EBITDA. Yes, and this was prevalent paradoxically at a time when you had a lot of investment going on. So you were excluding in a way one of the key features of the profitability. So, for instance, a Vodafone when it was in its pomp, that was all about EBITDA. And of course, that was at a time when interest rates were rising from the cost of their debt and there was higher depreciation and then huge amounts of amortization from some of these very large share financed acquisitions that it was doing. I remember there was a particular

abomination, a company I came across during that period, which I think had revenues of sort of 90 or so million. And then they showed a number below that L-bit Dasso, which was perhaps half of the revenue number. And I thought, well, that's encouraging because it was at least a positive number. Before I realized actually they were referring to loss before interest, tax, depreciation, amortization, and stock options. So earnings manipulation in the second half of the 90s was linked to the executive compensation. Marathon initially embraced shareholder that. in the 1980s and in the early 1990s, but started to worry that shareholder value was running awry as the boom period took hold. Yes. I mean, there was this measure of EVA or economic

value added that was used quite often by companies, particularly large companies. The temptation there was to cut good costs. So investments in things like R&D and marketing, which, you know, help you innovate and build your brands for the future in order to meet the earnings guidance. Another example of Goodhart's Law really of the corruption of the measure. Yeah. And there was a case I think where one of the pieces looked at Colgate Parmolive and found that between I think 1997 and 2000 reductions in their media spend accounted for over 20% of the profit growth during that piece. period. And it was during a time when actually advertising rates were going up at 10% a year because

of all the excess demand from the TMT bubble. Marathon were also concerned one aspect of shareholder value, which was to align the interests of senior executives with outside shareholders by giving them stock options, that this was creating an incentive for share buybacks, regardless of the stock price at the time. There's two parts to that, I think. One was that a lot of the share repurchases were actually being used to offset stock option dilution. So when you looked at the number of shares in issue for a company, that often didn't go down, despite the large share buybacks because much of the money was being used for stock options. I think there's a statistic in 1998, two-thirds of the shares purchased by S&P 400 companies were done to offset stock option exercise.

Merck, the pharmaceutical company, was spending 75% more on buybacks than it was on R&D. And yet, with all of those buybacks, the share count was actually going up because the level of stock option issuance was greater than the share buyback reduction. And then the other aspect is, I think, the point about where you are buying back shares and you are reducing the share account and counselling them, are you doing that at a sensible price? and there was a concern in the book that many of these companies were overpaying for their shares, and that was going to hurt their existing shareholders. But I have to say now when I look back at some of these companies and where the share prices are today, which are often multiples, you think, well, actually, probably many of these share purchases were really quite good investments at the time.

One, I'd call it a central tenet of capital cycle theory is a profound distrust of investment bankers. And the core of the idea there is that you have your buy side and your sell side, your fund manager and your investment banker. And the investment banker is interested in raising capital, whereas capital cycle analysis wants disciplined capital allocation and preferably capital to be withdrawn from an industry in order to boost returns. This is an era, if you remember, where there were problems of conflicts of interest between investment bank. banking part or the M&A part of a large investment bank and the brokerage research. There's a nice piece on the brokerage research conflicts of interest called how the game works, which actually you wrote and involved your personal experience. So this was before I joined Marathon, I'd been looking for a job and I'd been contacted by a headhunter

who was recruiting for a team of analysts to cover technology companies. is. Inadversantly, I was sent the notes of the headhunter from their conversation with their client who was expressing to the headhunter what it was that they were actually looking for. And I think the note which we publish in the book really shows how actually the banks were quite two-faced in a way in what they were doing. I think the line from the headhunter's note that they were looking for people who had no pretense as to how the analysis business works that corporate. Finance is a critical part of the group. The key is generating money through IPOs and using, in a way, the research analysts to promote those IPOs and that they didn't want analysts who were

in Verticomans prissy about their independence. They needed to be strong on marketing. They needed a reason to work hard, expensive tastes, large mortgages, and they must need the money. That note was 97, was it? Yes, that's right. And those problems and these conflicts of interest bubbled away, so to speak, until they blew up after the bubble burst. There was this big investigation led by the New York Attorney General, Elliot Spitzer, followed by the publication of other incriminating evidence

that famously the internet analyst at Merrill Lynch, Henry Blodgett, had been putting out five notes on stocks launched by Merrill Lynch, while at the same time saying to investors that they were what he called a POS. Anyone who was around at the time will remember what POS stands. Not a point of sale. Not a point of sale. And then I like this. There was the state of Massachusetts uncovered an email from two credit suites first Boston.

And it's from an investment banking colleague on the two unwritten rules of analysis. One, if you can't say anything positive, don't say anything at all. to go with the flow of the other analysts rather than try to be contrarian. One of the better pieces in Capital County describes how professional investors and corporate managers get entranced by investment bankers who are themselves often selected for their charm. It refers to sort of Stockholm syndrome. CEOs, I think, would be in a way seduced by investment banks to do deal. dealmaking and partly this reflected their lack of experience in capital allocation.

So they could be easily led. Warren Buffett neatly summarized this in his annual report from 1987 when he said that the heads of many large companies are not skilled in capital allocation. And it's not surprising because most bosses rise to the top because they've excelled in an area such as marketing, production, engineering, administration, or sometimes in institutional politics. Later in the book was a description of how the boss of one company that was involved in a large acquisition, BA.E. Systems, the chairman Richard Evans, talks about his experience of meeting with Bruce Vassesteen, who gave him a half-hour monologue on why they should

be buying the Marconi defense electronics assets. And Richard Evans described this experience as like having an intravenous injection with a 12-inch diameter hosepipe. He's such a great salesman that in the end, the only possible answer was yes, yes, yes. And on the subject of investment bankers, the capital count points out that almost the worst circumstances actually to have businesses run by investment banks. So at least that was the case in the late 1990s, early 2000s. That's the case of, I think, complete capture. the investment maker gets his hands on the controls, his or her hands on the controls.

And the case here that we follow quite closely and suffered from was Vivendi, where Jean-Marie Messier. I should say that Messier was the star banker of Lausanne, where Charles and I worked in the early 90s. So he took over what was a sort of stayed old-fashioned company, Gerald Desau, and turned it into this digital beermoth through a series of expensive, equity-funded acquisitions. Ultimately, that all fell apart. And, of course, I think there was a note we wrote at the time about how he appeared to be surrounding himself with yes men. And then subsequently, we discovered that his CFO was becoming increasingly worried about what was going on and wrote an email to Jean-Mauie messier saying that this reckless acquisitions,

strategy had given him the unpleasant feeling of being in a car whose driver is accelerating, and I am in the death seat. All that I ask is this not end in shame. Other examples in that era were people like Juan Villalonga at Telefonica. Again, in the book, it refers to how he was running seven or eight investment bankers at a time moving between meeting rooms, obviously in his element. And then the sad story of Marconi and John M. Mayo, again, where ultimately the companies fell apart as a result of poor acquisitions. Let's talk about capital spending and the TMT boom. We often look back on this era as the internet bubble or dot-com bubble.

But actually, if you read capital account, the real story is about what's going on in the telecoms and media and technology sectors. The dot-coms themselves, with perhaps the exception of Amazon, were largely a side-shay. The real story that this book focuses on is the telecoms capital spending and the rolling out of the information superhighway. It was really a almost perfect case study of the capital cycle in action. You had these high valuations which were leading to more investment that led to increased competition. supply, exceeding demand, and then bankers promoting excess. You had the examples in the UK, people digging holes in the ground when they're rewarded for doing that. There were these so-called alt-nets companies, there was energists and cult that were being valued at five to ten times

the invested capital, which was effectively the capitalised cost of building out these networks. And that was obviously an inducement for others to follow suit. KPN Quest in the Netherlands had spent 2 billion euros on its network and was being valued at 11. So it was being rewarded. We met with global crossing in 1999. That was valued eight times invested capital. Level 3 was talking about 10 billion of investment into a fiber optic network, had minimal revenues and yet was being valued at 30 billion.

So it was an extraordinary time. And in aggregate, just at the end of the bubble period, the US telecom stocks were trading. at six times book, which explains the incentive to invest. And aggregate telecom spending was vast. There's reference in the book to roughly $500 billion being spent by US telecom firms during the boom, half spent by new entrants and these alternative carriers. At the height of the boom, it was said that there were 150 alternative carriers in the US.

40 of those were quoted on the stock market. that led to eight national backbone networks being built from scratch and, of course, competing against each other. And Marathon was pointing out at the time that the CAPEX per household was running at roughly $2,000 per household and that it seemed remarkably unlikely that households were actually going to have enough money to validate that capital spending. But what drove the capital investment at the time, leaving aside the high share price, was highly erroneous forecasts of data traffic growth.

It's an axiom of the capital cycle that it's easier to identify oncoming supply than it is to identify demand. And here we have, this period has, I think, to my mind, the most beautiful example of that notion. Yes, there was this data point that was put out that internet traffic was doubling every, three months. And I don't know how many times we were hearing this from the CEOs of companies. And indeed, it actually crept into one of our own GIR articles so credible, it seemed. And it turned

out that it wasn't really growing at anything like that rate or not for a sustained period. We discovered that it had originated from a company called UUNet, which was owned by WorldCom. and it was being cited by the CEO of Level 3, and actually it turned out that internet traffic was doubling only once a year, so at a pace of eight times less than as it was thought. Right. When you say as it turns out, that was a piece of research put out by AT&T Labs researchers in 2000. So the data actually giving relatively accurate demand forecast or real-time demand change was available

but during the bubble no one seemed interested to get to the truth. But the upshot of all this was incredible over the capacity in the telecoms networks. Again, this was anticipated by Marathon, not just by its comments back in 1994, but a comment cited in the book from a piece towards the end of the bubble period, which I think is marvelous. It says the telecoms industry has the capital intensity of a railroad but enjoys the pricing power of a polyethylene plant. And events over the following couple of years were to justify that comment.

Yes, and you had bankruptcy of WorldCom. And I remember going to dinner at one of these big New York banqueting halls where Bernie Ebers was the guest speaker. And Ebers was the chairman's CEO of WorldCom. He came on stage and there was a large screen behind him and he switched it on. it just showed this upward parabola of his share price. And he stood there, paused, and then just said, any questions? Subsequently, after the share price collapsed,

would have been quite a few questions for him. And it reminds me of a quotation also in the book from Johann Rupert, the chairman and CEO of Richemont, who gave the advice to CEOs that if you talk up the share price and the share price comes down, the folks come looking for you. But the upshot of the overinvestment was that the US was left with 15 continental white telecoms networks and that wholesale telecoms prices fell by around 90% in 2001, 2002. And there being estimates of excess capacity of so-called dark fiber, unlit, unused fiber optic cables running to about 95%. When these companies went bust, their valuations in bankruptcy were negligible.

The book cites the case of global crossing, one of the US telecoms businesses that had a book value of $22 million, but was sold out by its administrators for $500 million, which was just over 2% of the book value, and less than the receipts of the share sales from the founder and chairman, Gary Winning. One of the lessons of this period was to just be very very very much. very wary of false data points, which reminds me of another article in the book about what we called investment Mcuffins after the Hitchcock plot device that was used to drive many of his films, which was often a sort of high-tech piece of equipment that was only slightly plausible, but could drive forward the narrative.

And that particular piece referred to the Y2K bug, which, of course, led to huge amounts of investment, but probably was a hoax. in the end. Well, we know it's because I think as a book points out, the Italian company has never bothered to rectify the Y2K problem and nothing happened to them. No, no, I think by the time the Italian regulator actually put out some guidelines as to what companies were doing, it was, it was in August 1999, so it didn't give companies very much time to respond in any case. So a late phase of the capital cycle is often marked by flotation of companies on the stop market and the late 1990s technology bubble saw, I think, it's fair to say, what was at the time

the greatest number of IPOs in history. And Marathon, with its capital cycle, discipline, wasn't buying into those. No, no, we weren't. I think we received 212 American prospectuses in 1999, 60% of which were tech and telecom companies. And the upshot of Marathon, not buying into the IPOs that they in turn were shunned by the investment bankers, great. Yes, you had all these practices going on, one of which was so-called spinning, which was the handing out of preferential allocations of shares to friends. And there was talk of laddering, supporting shares in the aftermarket. CSFB was involved with some of these practices and ultimately landed up paying a fine, I think, of $100 million to the US regulatory.

And actually even in our own case, there's an email that we received from a broker, which is referred to in the book where we are threatened with having business taken away from us, that is, the ability to meet with company management if we're not going to do more business with them on the so-called primary side. If you weren't going to play the game. The appendix of the book contains reports on a handful of companies that were just so egregiously overvalued. And I included those reports because I thought it was a good historical record of how the valuations, not of these flaky little dot-coms, but of companies, very big

companies, such as Nokia, for instance, and Nokia rose to be 65% of the Finnish hex index. Take Nokia, for example. Well, in that case, we did a note which looked at the valuation using a discounted cash flow analysis and came up with a share price, which I think was a third of the then share price, of course, which turned out to be wildly optimistic because of the subsequent collapse at a company. But we were thinking really about the longer term number of users, the likely pricing at the margin and so forth. And Charles, can I say looking back,

and this holds true of your analysis of Nokia and Vodafone and the other telecon services, you were using what in retrospect were very generous assumptions of what people were going to spend per month on their mobile connections and on their broadband, you know, some figures, you know, 100 pounds a month or whatever when, you know, now, when they're worth, you know, less than half that. We're spending less and a lot. In other words, by running, if you will, rational valuation models over these companies, you were coming up with valuations that were one-third. Just to end on Nokia, I mean, I think the CEO at the time had said to insiders that he felt that he said the tremendous growth cannot continue forever. But we don't want to mention this to outsiders because it wouldn't have been popular.

And so it actually years later, a later chairman of Nokia looking back on the history of the firm just made a quite succinct comment that success is toxic, which in a way is a kind of little reminder for capital cycle investors. And actually other aspect of Nokia is that, as we know in retrospect, Apple came along and took the mobile phone market along with the smartphones on the Android system, and Nokia completely lost its position in mobile phone market. So that points to an early insight in capital account that technology life cycles are often quite short. So Charles, to wrap up the conversation, I want to talk about what you see as parallels between what's going on today in the AI world and the TMT bubble of the late 1990s. If you remember, John Kenneth Carbraith writes, the memory of the investment world expands to 20 years or one generation.

I'm afraid to say that a generation has passed since we put together capital account. And I think there are some lessons in the book. Both us have commented and written about the AI boom in recent months. We see this tremendous capital spending going on in the AI space, trillions of dollars. You see these extraordinary valuations, open AI valuation now in private market up to latest $830 billion. dollars. We have what we saw in the TMT boom, particularly around the rollout of the internet and telecoms fiber, huge uncertainty as to what actually are going to be the profitable applications of the new technology. And of course, we have those overblown forecasts for growth. At least we know

they were overblown in terms of telecoms demand back in the late 1990s. So what do you see as the similarities today. Listening to you there reminds me of your friend Jim Grant's comment that in dentistry progress is secular, but in financial markets, it's cyclical. And I think that there are some lessons to be drawn from capital account. Obviously, we talked on our previous podcast about investment levels and the likelihood, at least in the short term of returns on those investments within the AI field. But I think rereading capital account, I felt the one of the points or one of the lessons was just this weariness around depreciation schedules, where you've got short asset lives

that may be getting shorter. You should be very wary from the perspective of profitability and returns, especially when there's a lot of competition. I think there's a point which ties in with what we talked about earlier with the telecom networks where if you don't have a moat around your business, then the likelihood is that pricing will tend towards marginal cost, if not below, and that can be a problem. So trying to figure out what businesses are doing that's really distinctive in a kind of capital arms race is very important. As you said earlier, IPOs often occur towards the end of bubbles. We're about to see the IPO season

within the AI world and it'll be obviously important to work out in that environment. What are the businesses they're going to survive for the long term? This issue of being very wary of the sort of fake news, the false data points. Of the hype. And always keeping in mind the sort of skepticism around whether what someone is telling you could just be a McGuffin or not. We've seen some elements of this circular financing going on, which are reminiscent of what some of the telcos were doing in that era,

off balance sheet, leasing of data centres. Yes, the SPV has made its comeback. Indeed, and capacity swaps. Ultimately, a lot of these deals are based upon bartering effectively. Where is the money going to come from, as you say, what is the sort of killer app or the consumer application or business application, which is actually going to generate the revenue? And then I think actually rereading some of the money,

some of the pieces in the appendix, particularly around the telcoes when they were bidding for the 3G licenses, if you remember. In Europe. In Europe, this was, this became a kind of existential issue. France Telecom had acquired Orange from Vodafone at a very high price. And if you then think, well, what is the value of staying in business, which was what the 3G auction was presenting you with, ultimately you should be prepared to pay up to that level. Because, if you didn't, you were going to lose your entire valuation. So I think that phomo fear of missing out aspect is also one to be conscious of. There's an entire chapter in capital account called Blind Capital,

which talks about both the growth of passive investing and then the development of these very large investment firms, which could only really invest in the largest companies, the so-called two-tier market that was developing with larger capitalization, and stocks becoming more and more highly valued. And of course, now looking back, this passive has grown exponentially since then. I think we're now up to 60% of the US market owned by passive. And then you've got obviously momentum strategies laying on top of that,

just buying the things that are going up. So those, I think, would be the principal parallels I would draw with that period. The two-tier market of the 1990s of a handful of stocks, moving the market and being a large, share of the market. You see these charts now of the top 10 stocks in the US by market capitalisation. And I'm saying off the top of my head, there's somewhere between 35 and 40% of total capitalisation. Again, off the top of my head, around twice the level they got to in 2000. So this is a much, much more concentrated market. The game of the investor or the professional investor, the value investor,

the capital cycle investor, as far as I see it, is to really be looking for the opportunities elsewhere. I mean, finally, I've read an article in the Wall Street Journal over the weekend about how we're now into a situation where we're running out of fibre optic capacity, which is ironic given the background from capital account. And it was interesting to read that the bottleneck in that industry at the moment is finding the labourers to dig up the roads, to lay the fibre. optic cable than having to pay the $40, $50 an hour for people to do that.

Yeah, but that and Charles, you know, the price of semiconductors is going through the roof and the energy is going through. So the capital spending boom in AI, we reckon to be somewhat larger than the telecoms capital spending boom of the late 1990s. So both positive and negative effects on the economy. So with that, Charles, thank you very much and speak to you soon. No, thank you, Eddie. Thank you for your time today.

I hope you will listen to the next edition of the Capital Cycle. This communication is provided for information purposes only. Please refer to Marathon's website and the Global Investment Reviews for further information, including important disclosures.