The World's Smallest Festival - part 2

Part 1 conluded with some observations about Japan and that's where Part 2 continues. The debate then broke into the major theme of the weekend which was the gulf between the haves and the have nots (or the have less).

Japan, Korea and Indonesia feature in this section and then the conversation moves to the questionable economics of AI model builders, inflation, prospective shortages. Frances Dydasco disclosed her dry goods stores and, at the end of the section, Simon Ogus outlined his personal portfolio.

FD: So many Japanese companies exist after so long. Now, for the smaller and medium sized companies, we have the issue of succession, and some of them will hopefully just become corporatised, like many other large listed companies. But I would hope that Japan keeps doing this, because I think this is the reason Japan exists as an industrial power today, despite having so many problems and fewer people, it's because that's the way they think. If they start thinking like Americans, they're finished.

SO: But they also had very inefficient parts of the balance sheet, and at the margin the inefficient parts have become slightly less inefficient, so the delta on the ROE has improved. It's still a low ROE by global standards, but markets are made at the margin.

FD: Well, like a lot of private equity, but the part I'm worried about is if you get KKR and the like getting involved.

JS: They're trying, but it's interesting what happened is that the Japanese were very resistant to the American style buyout culture, and then they somewhat appropriated aspects of it and made it palatable to themselves. So that's there, but I don't think it completely drives corporate returns or share prices.

FD: I think that as a society they still have social cohesion. I was watching a TV show about the ageing problem, and about the population and ageing, and how they're actually effectively collectively accepting a lower social insurance payment. They mitigate this by doing a lot of volunteer work with each other, but the aim is actually to cut their withdrawals from social insurance, which is quite interesting as a very micro example of how this society is starting to think through how to manage as we get older, with fewer younger people and fewer resources, because the fewer resources point is very clearly understood by these elderly people.

SO: But can you have social cohesion and ethnic heterogeneity? My argument on Japan is that it will be able to have this grown-up discussion, where the assets of the elderly are transferred to the young, for precisely the micro reason you just talked about. But as we've seen somewhere like Sweden, a very generous welfare system is now coming under strain, because it was fine when that country was more homogenous but now when immigrants respond to exactly the same incentive structures, the system starts to fray. I'm not trying to be a Nigel Farage here, but I do think that heterogeneity in ethnic population mix is really going to complicate the ability to have that debate within societies.

IM: Is that happening in Japan?

SO: No. There is redistribution going on for exactly the reason Frances said. If you saw the Japan piece I circulated, called "Don't Try This at Home," it was based on exactly that idea, that you've got this pot that can be transferred over time, and I think it can be done voluntarily in the Japanese system, with a big netting off. Korea can probably do it as well.

FD: I don't think so.

SO: You don't think it's happening in Korea?

FD: No, they're very different societies. I invested in a lot of service companies in Korea, and if you look at their structure, temporary workers are really exploited. You know that movie “Parasite” is a metaphor about howe the people who work for Samsung, and the people who work for everyone else, are very different. Like Samsung Group, CJ Group, a lot of these companies, if you look at their structure, sometimes 70 or 80% of their workers are temporary. What that means is they're hired every single year, they have zero social security.

SO: But Korea also has the highest minimum wage relative to median wage in the world.

FD: But I don't think the Koreans have the same type of collective sacrifice mentality that the Japanese have. I think that's very unique to Japan.

SO: Good point. So, Japan's the only place that will survive it? Well, China won't be able to do it either, China is too much of a jealous society as well, I think.

IM: I was reading the paper you referenced from a while back, it's about ten years old, and I looked at the updated figure for the percentage of the government bond market owned by the Bank of Japan. It's forty nine percent. So have they effectively practised modern monetary theory?

SO: No, because you never had the broad money multiplier taking off. In all these places, if the credit multiplier really starts to work, and the government has to compete for funds, it gets very ugly very quickly. What Japan has done is manage to shift the government bond burden off the books of the banks and onto the central bank, especially the longer dated, ten year bonds, which is particularly where they were targeting.

I put in this piece that you could, in theory, issue a zero coupon perpetual just for the portion held by the central bank. You'd have to recapitalise the central bank, but in effect you're writing off that debt, it never has to be repaid. I think in Japan the markets could probably accept that, if it was accepted as a one off. Whereas if you've got foreign owners of the bonds, and you tried to do that in the US market or the UK market, it would be a rush for the door.

FD: I wanted to ask you this. Is it possible, in Japan, for the government or the Bank of Japan to just say, "Sorry, everything is now worth forty percent less, because we've just done a haircut"? They could basically just do that, right?

SO: Yes. And they probably wouldn't have to do capital controls to do it, whereas you'd have to do capital controls in advance in other countries. I've had this discussion with the Bank of Japan over the years, and it goes back to the balance sheet item of Social Security benefits. You have to basically term out pensions. What I put in this paper, and discussed with the BoJ, although they won't go public about it, is that you would say, if you've got a pension pot up to, say, 100,000 or 200,000 US dollars, that's guaranteed, we're not going to touch that. But if you've got a 5 million dollar pension pot, that pot is only worth 3 million now, because the bonds hypothecated against it are worth less.

FD: I could totally see Japan doing this.

SO: Yes, but I don't see any other country that can do it.

FD: I’ll tell you a story. On a company visit in Japan a few years ago, I visited a company on the day they announced they had defaulted on their pension fund, they just said, "No more pension, sorry, it's gone." I sat in front of the guy and said, "I'm so sorry, I just found out your pension has been cut and you've had to take a pay cut." The entire company took a 40% pay cut across the board, and they defaulted on the pension.

Do you know what he said to me? He said, "I have a job, we all have a job." That says so much about Japan. He was on a pay cut with no pension, and this guy was probably in his early sixties. But I think that's what everybody has to do, don't you think, Simon, that's what the ageing West has to do?

SO: I think it has to happen globally in some way, but I just don't think it's going to be polite in most places.

FD: It's not going to be polite, but it's de facto going to happen, isn't it, ultimately, we're just going to take a haircut.

SO: No, you can default in many ways, and the question is which route you go. You can get a left-wing government that just steals it, we've seen that in a few places. You can go for hyperinflation, which gets rid of the debt pretty quickly but has other consequences. You can invade other countries and steal their resources to pay back domestic bondholders, which is what Mr Trump likes to think he's going to do, that's what the British did after the Napoleonic Wars.

And then there's financial repression, Russell's speciality, the one that's been more commonly practised in modern times, but it involves capital controls. Frances, that's the paper you should write. It would carry a reputation risk for you, so it could be under a pen name. It's in Leviticus, chapter 25. ["And if thy brother be waxen poor... thou shalt relieve him... Take thou no usury of him, or increase... Thou shalt not give him thy money upon usury, nor lend him thy victuals for increase." – verses 35-37]

JS: You deserve credit for the width and depth of your references. Now the Bible.

IM: Didn't the FT do a podcast on this? The Mesopotamian king who wrote off debts when harvests failed?

SO: Hammurabi. [Podcast: https://www.ft.com/content/007c047e-7446-458a-8c59-f73e8f093283?syn-25a6b1a6=1] [Video: https://www.youtube.com/watch?v=L09nCf8X_d4] We should all listen to it. We need to write down some of these debts. This was very common in many ancient societies.

FD: Going back to your first question about where to put money, I worry about this, and I realise I can't do anything about it, but you know what, I'm going to be in great company, we're all going to get poor together.

JS: How does that view go down at dinner parties in Singapore?

FD: There are just certain things I don't talk about with certain people. I hoard, I've been a hoarder, I've been expecting it.

JS: How about, when we meet online for tomorrow morning’s session, you take us round your dry stores, a tour?

FD: I'm totally serious, under every piece of furniture.

JS: I know you're serious but Phil won't believe it until he sees it.

SO: But Frances, one thing for you before that, on the hoarding. I've been involved for about a decade in a New Zealand farming business, for many of the same reasons you're doing this. The chief executive is a very smart, part Maori, and we were on a tour once, we drove past Julian Robertson and Peter Thiel's bunkers. I asked him, "Do they think they're really going to be safe?" He said, "Mate, we were eating people a hundred years ago, they're going to be the first ones to go."

SO: I recommend a classic Japanese film "The Ballad of Narayama." Made in the fifties, by Kon Ichikawa. It sounds like it must be very sad, but it's actually a really happy film, I won't spoil it, watch it, it's wonderful.

This section of the conversation began with an exchange of views as to the role of Asian and Pacific sea-born peoples and the historical civilizations located in what is now Indonesia. It was noted that the entire region, spanning from Guam and Taiwan down to Java, the Philippines, and coastal Southeast Asia, historically functioned as an integrated maritime archipelago driven by the free exchange of trade and ideas prior to modern nationalistic boundaries.

SO: Frances, you have some Indonesian heritage as well, don't you?

FD: A little, from about 20,000 years ago. I'm from Guam, and our people originally came from Southwest Asia.

SO: As a Hong Kong resident, one might note the traditional view that the region was part of the Chinese sphere of influence since time immemorial.

JS: William Dalrymple points out that the most consequential route in Asia became the seaborne route down the east coast of India, not the Silk Road over land, as most China watchers posit.

FD: Probably even earlier than that, because people arrived around 20,000 years ago.

SO: But the Chinese dynasty is 30,000 years ago, you have to remember.

FD: If they were travelling by sea, probably much longer than we recognise. Can you imagine?

JS: Are you contesting what we're saying, Simon?

SO: You may say things I could not possibly comment on.

IM: He's representing his domicile.

SO: It's very easy to critique the Chinese, just as it's very easy to critique the Indonesians on these things.

JS: In my view, the Chinese have more of the airtime in laying down the narrative is because the archipelagic peoples are disparate, on different islands and now in different nationalities, it's much harder for them to pin down their own history.

SO: It's also analogous to what's happening in the EU, the Chinese are very good at playing one country off against another, and nobody collectively wants to push back. Look at the Philippines, it's put its head above the parapet on the South China Sea, and all the others have just let them take the flak. But China is also causing issues with Vietnam, with Indonesia at the moment. Unless there's a collective response to the Chinese narrative creating facts on the ground, and so on, killing the European car industry, and so on, China will end up winning this. Volkswagen has recently announced 100,000 layoffs in Europe.

I was going to say, in the West, if people don't follow Asia, because there's been so much coverage of the States, of Trump and the unpleasant things he wants to do with Greenland and other places. I have a friend who doesn't follow this stuff, just an average Brit, and he's saying, "Oh, the Chinese are so much better." I said, hang on a minute, they're terrible imperialists as well. But that just doesn't get the coverage in the West.

My Cantonese wife feels she's being invaded by the northerners, even though her surname is Mong, which means she was herself descended from a Mongol invader in the first instance.

JS: How many generations have to pass before you cease to be a Mongolian?

SO: You never forget you're a Mongolian.

The conversation then turned to Indonesian politics. It was noted that underlying public resentment in Indonesia remains high, with potential for future street unrest. President Prabowo's centralisation of presidential power reflects a movement towards the 1945 constitution framework. Younger voters, lacking direct memory of the 1997–98 human rights record, largely favoured a "strongman" leadership figure. However, economic policy challenges—such as the initial attempt to nationalise commodity export receipts through Danantara—have met with practical resistance and subsequent backtracks, while MSCI reviews enforce discipline on data disclosure without risking a broader bond market downgrade.

Prabowo’s candidacy was assisted by outgoing President Jokowi's late-term dynastic moves initially shifted support towards opposition figures. However, Prabowo subsequently allied with Jokowi—securing the inclusion of Jokowi's son—which generated a crucial late surge in electoral support.

The practice of elite politicking and the fact that Indonesian politics remains dominated by 1980s and 1990s elite structures. The ruling establishment remains focused on retaining power and dynastic succession, even at the expense of potential economic growth. Consequently, Indonesia is missing key opportunities to attract manufacturing shifting from China, failing to replicate Vietnam's investments in infrastructure and human capital.

IM: What interests me, and I don't know much about Indonesia, is that I read a paper by Simon a couple of months back on oil and Asia, and Indonesia took a hit in stock market terms from that, but when I looked at the numbers in your notes, Simon, they're actually pretty good on some metrics we worry about elsewhere, like public debt to GDP and deficits. Some of the macro stuff is actually pretty solid.

SO: The general consensus is that while Indonesia is growing below its potential, an Asia '97-style crisis remains unlikely without mass capital flight.
It was noted that while Indonesia benefits from location and key resources such as nickel and coal, its long-term economic advancement is capped by inadequate investment in physical and human capital.

It was observed that Vice President Gibran [ the son of Jokowi] appears to be maintaining a low profile while Prabowo tightly controls cabinet proceedings, which may be the pragmatically sound approach for a young politician building long-term standing.

SO: The Latin American analogy is very apt for Indonesia, the Philippines, and maybe even Thailand these days, in that people don't starve in these places, so the landed elites can keep things going by buying off near term unrest, using co-opted clerics, whether Buddhist monks or the Catholic Church in the Philippines, to keep everything in place. It has that Latin American feel, whereas the more Confucian societies in Asia seem to have moved beyond that, understanding that making everybody better off, even if it's unequal and corrupt, works better.

FD: How significant are remittances these days as a percentage of the economy? Is it a large figure like in the Philippines?

SO: It's about two to three percent of GDP versus ten to twelve percent in the Philippines.

FD: Is Indonesia self-sufficient in many commodities? I'm thinking about the inflation trigger for unrest, because it's usually food, energy and transport, that's a risk. Last time I looked, Indonesia ate more wheat flour than rice and that's the real risk, because it didn't really matter whether they became self-sufficient in rice if they ate more wheat.

SO: If you look at the FDI numbers, there's no non resource FDI going into the place, there's very little formal job creation. The Central Bank is generally much better than it used to be, and the fiscal side, even with Prabowo's changes, is still pretty good, so macro management is actually okay.

The lack of non-resource FDI stems from weak rule of law and skill shortages. Foreign tech firms (such as Apple) have cited difficulties in finding qualified local engineers, restricting high-tech manufacturing investments. Without structural reforms in education and legal frameworks, GDP growth remains capped around 4.5–5%, well below its potential 6–8%.

JS: Let’s move the discussion onto Korea.

The group noted that while South Korea's recent market surge is supported by genuine earnings growth and robust order books in key technology sectors, the rally exhibits high domestic retail leverage—including $30 billion in margin lending and leveraged ETFs—while foreign institutional investors face index concentration limits. Two key downside risks were highlighted: potential slowing in US AI data centre order flow as US tax incentives expire by year-end, and extreme market leverage.

SO: On that point, how do local pushbacks against data centres, water usage, and massive tech bonuses play into political and market risks in the US and Korea?
The conversation moved onto US data centres and allied requirements and draw on resources.

JS: Frances, you mentioned a fact to me earlier this morning about the AI build out in the US, do you want to share that?

FD: About 200 billion, I think, of capex committed, as in there's a loan for it, but out of that only five billion has actually started construction. So the remaining 195 billion is basically paying interest but no principal, for various reasons, whether it's building the water infrastructure or whatever.

JS: I find it very amusing that Jensen Huang travels around the world in a leather jacket, what's he signalling? It's a question of who gets to the front of the queue, and obviously Elon Musk got himself there, but so very much did Microsoft. Microsoft have met their orders, NVIDIA ships, they book revenues, but Microsoft haven't deployed it, so there's no depreciation.

IM: Well, as discussed in yesterday's regional overview, there was widespread skepticism over whether this level of capex can be sustained, whether as many large language models are needed, or whether resource inefficiency and token costs will force tighter customer spending...

FD: I did some reading on this overnight. Essentially, a token is simply a measure of size, like a byte, something like four words per token. But the output is the part you can't control, when there's reasoning or inference involved, you can't tell how much token output will come with the reasoning. From practical experiments it's multiples of even the answer itself, for example you could have a two-word answer but the reasoning token usage could be three hundred units. The challenge is, say you're an engineering firm and you want to let the AI work overnight, you have absolutely no idea how much reasoning it's going to use to complete a task.

Right now these companies are giving basically fixed price contracts, "all you can eat for two hundred dollars or two thousand dollars," whatever the unit is, for corporates or individuals. But if you look at the websites that track the highest users, the absolute amount they're using, if you gave them a fixed cost rate per token and multiplied it by the number of tokens they're actually using, do you know what the numbers are for single users for one night? Nine hundred thousand dollars, five hundred and fifty thousand dollars. There's a table of these people and their usage, calculated on the basis of a non fixed price contract. If you said to me, "Go and use whatever AI you want and build me a model," and I used five hundred and fifty thousand dollars worth, you'd only find that out when you came in on Wednesday.

So what's happening is companies are actually capping usage, because there's no way to calculate the ROI of a task given to an AI using this kind of reasoning. So if you look at it, the sophisticated AIs with reasoning capability are actually incredibly inefficient in their token usage, and if you moved from fixed pricing to unit pricing, at the end of the day, there's no economy of scale, is what I'm trying to say.

JS: And it comes down to just three major stocks. From an Asian portfolio management perspective, high index concentration creates a major tactical dilemma: managers must participate in extreme momentum rallies while preparing to go underweight when the cycle turns. Simon, you had a final question on Korea about the earnings profile of Korean companies outside the main technology stocks, and is Korea genuinely following Japan's path in corporate ROE reform?

It was observed that non-tech sectors in South Korea —such as shipbuilding, nuclear reactor construction, defense equipment, and luxury goods—show strong earnings, accompanied by increased dividend payouts and share buybacks. However, skepticism remains regarding the long-term "Value Up" initiative, as controlling family dynasties retain little appetite for dismantling complex chaebol cross-holdings, making structural reform more cosmetic than fundamental.

IM: Let me whistle through a few other things. Some of the other things I looked at in terms of the current boom relative to 2000. Everybody calls it the internet bubble now, but we called it the TMT bubble at the time, telecoms, media, technology, media was in there because it was expected to be a big beneficiary too. Just on valuations, for the record, the S&P 500 trailing PE today is 31 times [earnings], it was 30 times in Q1 2000, so that's very similar.

SO: What was the dispersion within that? Because in America my understanding is earnings are flat outside of about seven or eight companies.

IM: That's right, the concentration is much higher this time. The top 20 companies are 49% of the S&P 500 today, they were fairly high in 2000 too, but lower, at 38%. If you look at the Shiller PE, which we know has its flaws, it peaked at around 43 times in the 2000 bubble, it's been running at around 41, 42 today. It's wrong to say everything is much more profitable today, there was a bit of that "pets.com and not making any money" back then, but you've also had five years of very strong earnings growth to date, as you did in 2000, both were about 20% earnings growth. The forward earnings growth expectation for the S&P today is 20%, which is exactly what it was in March 2000, and it ended up being minus 30%. So, the fact forward expectations are robust doesn't tell you a lot, as we old hacks know.

One other thing, semiconductors, which we've talked about with Korea and NVIDIA, are now 18% of the US stock market, they were about half that in 2000, so if the hardware story slows it's going to be a reasonably big hit on the overall index. Are companies more profitable today? A bit, return on equity on the S&P is about 18%, it was about 15% back then, though it wasn't terrible back then either, it was decent. I've mentioned the capex cycle numbers, the actual numbers are a bit bigger now. What's interesting is fundraising this time is far bigger, if you look at IPOs plus venture capital raised in the last 12 months it's 3.5% of GDP, 860 billion dollars between venture capital and private investment, nearly a trillion dollars, Simon. Back in the 2000 boom, IPOs and VC were 1.8% of GDP, so the funding going into the bubble is actually twice what it was.

SO: What was the debt number though?

IM: That's just equity, I don't have a debt number, but I'd suspect debt was higher then too, the telcos would have been raising a lot more debt, we're only just getting going on debt in this cycle. What's incredible is the fundraising, one company, Anthropic, has raised 105 billion dollars in the last six months, an unprofitable company. And that's the other thing, where in 2000 we had pets.com and the like, this time the unprofitable companies have tended to be unlisted, including your favourite, Frances, SpaceX. There's more, but I'll leave it there.

JS: In terms of return on capital, since 2000 all the big US companies have been buying back a lot of shares, that wasn't as prevalent in the nineties, it's been super prevalent since.

IM: It tends to be ROCE enhancing, but it's financial engineering, and it's actually gone into reverse now, because you've got IPOs coming, that's equity supply, and now you've got Google even issuing equity, and Meta saying they might have to as well.

SO: I haven't run this chart for a while, but one I used to run through the last decade was the buyback element in US shares versus the real underlying capex going on. The US actually under invested systematically for a decade, because all the money was going into buybacks, one or two percent a year, almost equal to the flow of non-residential capex. Now I imagine that's suddenly reversed in the last couple of years.

FD: And remember they also have retained tax, so it's a combination of very corrupt, terrible practices that look completely unsustainable to me.

SO: As Leona Helmsley said, "Only little people pay tax."

IM: One final factoid, one big difference between the 2000 bubble and the AI boom now is the fiscal accounts.

SO: Dreadful.

IM: In 2000, I'd forgotten this, but I'd made a note of it in a book I read, Alan Greenspan, then chairman of the Fed, was publicly worrying that there'd be so many fiscal surpluses ahead, from the existing surplus and the benefits of the new economy, that they'd run out of Treasury bonds with which to manage the monetary system. The expectation was that the Fed would lose that tool by 2000.

JS: And now we've got these continual deficits instead, which is interesting, and Scott Bessent is funding at the short end.

SO: Greenspan died recently.

JS: Would you want to say something about his legacy in US macro management?

SO: I think he was overrated, for sure in retrospect. My long-term view has been that Greenspan put in motion asymmetric monetary policy expectations which have inflated this bubble for thirty years, because he basically said in '87, "I can't identify a bubble while it's inflating, but I always stand ready to help the market." Bernanke took that one step further, again didn't put rates up enough in 2002, and then cut too much.

JS: That's going to be your book, which I’m going to publish.

SO: I thought Sebastian Mallaby had already written that, he's quite good on that subject. [“The Man Who Knew: The Life and Times of Alan Greenspan" (2016)]

JS: In one of Adam Tooze's Substacks, he clipped in a chart from the FT showing the Bloomberg consensus for aggregate cash flow of the Magnificent Seven, historic and forward looking. Forecast cash generation from '27 or '28 going up massively to 2030, and if I was a betting man, I don't think that's going to happen, or at least not by the indicated quantum.

IM: No, that's not going to happen, the market is sniffing that, the multiples are starting to come down on the Mag 7.

SO: A very good point was made earlier that an Asian portfolio manager has to keep playing at this stage because the index is so concentrated. For us as individual investors, you don't have to play, you're still being paid four percent on dollar cash. I've personally de-risked quite a lot in the last year and obviously left quite a lot on the table.

JS: And your personal portfolio, Simon, are you happy to share it?

SO: I've got a lot of gold, a lot of commodities, some absolute return people who are long only, people I know. And I'm interested in Frances' view on this, you said the Singapore dollar is expensive, I've actually been buying more Singapore type assets, just for more currency diversification, and if you get a dollar to North Asia rebalancing, because the won, the yen, the Taiwan dollar and the RMB are just egregiously cheap on any valuation, the Singapore dollar basket will appreciate.

JS: And when you go into Singapore dollars, where do you place it, short term bond funds and things like that?

SO: Or things hedged back into Singapore dollars, there's quite a lot of product in Singapore now that offers that.

JS: And in commodities, are you holding ETFs or a portfolio?

SO: I've got gold, quite a lot of it, I've had gold since 2003. And I own a commodity manager, in quite large size, they're ex Cargill people, physical traders. It's called the Merchant Commodity Fund.

-end of part 2-