I often start my monthlies with a little summary of what happened last month. But I suspect that won’t be necessary this month. But one thing that might be of interest, especially for those outside of Japan, is how much more volatile the markets are in Japan than in the US. Fundamentally, this makes sense given the greater sensitivity of Japan to middle eastern oil and currency markets.
 
For example, take the S&P 500. At least up to March, realized volatilities are not all that different from the last year when the tariff uncertainties had shaken the world (and, while we might have temporarily forgotten, it hasn’t gotten any clearer since). Admittedly, implied volatilities are definitely higher but compared to other events in the past decade, it isn’t all that high. Obviously, it was higher after COVID and Liberation Day. But it was also higher throughout most of 2022 as well as some bouts in 2018 from Volmageddon and Powell’s Fed pivot. Forecast volatility has been actually lower than the long-term average for most of March.
 
Now look at Japan’s Nikkei 225. Realized volatility has only been higher twice since the pandemic hit the world, once after the surprise BOJ rate hike in 2024 and the second after Liberation Day. And in both cases, the effect was extremely short whereas volatility has remained elevated this time. The same can be said about both implied volatility and forecast volatility. And while we don’t have a VVIX index in Japan (and I’m not sure if I entirely trust our VIX equivalent as Nikkei 225 options markets are not nearly as liquid nor do our markets have as much breadth or depth as S&P options), the realized volatility of the VIX equivalent is at the second highest level since it began in 2001. The highest level was at the surprise BOJ rate hike but that was a single day event. Excluding that one day, we are at the highest levels now. What this means is that the “volatility of expected volatility” is at its most extreme level since the dot com bubble. In layman’s terms, the uncertainty about how uncertain we are is really high. A similar argument can be made with the broader TOPIX index in terms of realized and forecast volatility (our TOPIX options market is terribly thin, making implied volatility curves based on listed options useless).
 
As such, both the Nikkei 225 and Topix had the worst monthly return since the Global Financial Crisis in 2008. The Takaichi-Trump summit in mid-March overcame many of Japan’s concerns leading up to the meeting and was largely lauded as a success, both in Japan and overseas and she saw a boost to her approval ratings. However, it did little to quell the markets and Japanese index volatility was one of the highest in the developed world as well as most of the emerging markets after heavily tech-weighted South Korea and Pakistan.
 
But what made March further interesting is the weekly consistency of how the markets moved. Almost like clockwork, the US markets were relatively strong during the first half of the week, only to fall during the second half. I understand this has to do with the relief rally during the start of the week after which investors took risk off as they approached the uncertain weekend as well as economic statistics that tend to be released toward the end of the week. While not quite as clean as the US, what this meant in Japan, being reactive to the US, is that Monday tended to be weak with the relief rally translating to a firm Tue~Wed, only to fade in the back half (this pattern broke down after the cease-fire was announced midweek last week). Being a long-only strategy, there is little that we can do with intramonth volatility. However, as we try to aim to be fully invested as much as possible, we tended to create some cash during those Tue~Wed (generally by reducing our tech hardware names) while buying into the weakness of other days (generally with software names). This was particularly acute with our diversified portfolio where factor/industry exposures are managed more closely, but the same can be said about our  concentrated portfolio.
 
As such, while March historically tends to be a difficult month for us on a relative basis (due to fewer March fiscal year end companies than the benchmark which means our dividend yield, which tends to already be lower on an annualized basis, is particularly lower in March while higher in December), the volatility of volatility proved to be a tailwind as well as the fact that the SaaS-pocalypse worries seems to have, at least temporarily, been put aside in place of broader inflationary and macroeconomic risks.
 
I mentioned several times from late last year that I believed 2026 to be volatile. But that comment was made due to Japan-specific politics and macroeconomic trends. On the margin, it hasn’t mattered quite as much and actions (and tweets) from the White House seem to matter much more. As we approach midterm elections, I suspect that will continue. All we can and will do is take advantage of this continued volatility to lean into our long-term trends specific to Japan. The inflationary pressures will more likely fan these trends even further.​

​

Masaki Gotoh​​​

​

===

 

​​“It is not the strongest of the species that survives, nor the most intelligent that survives. It is the one that is most adaptable to change.” – Leon C. Magginson, Professor of Management and Marketing at Louisiana State University at Baton Rouge, on Charles Darwin’s famous book, “On the Origin of Species”.
 
I suspect this is true of many families, but my two sons couldn’t be more different. My older son, Leo, is relatively reserved and dislikes standing out, whether it be himself or people around him (including us). He is particularly good at “reading the room” and thus gets along with everyone since he rarely expresses strong opinions or emotions about anything. He’s very mathematical but also likes to read a variety of subjects and is curious about everything. But he’s definitely not in the “cool” clique. Louis, 3 years his junior, is extremely outgoing and likes to sing and dance. His emotions are very strong (both in a good and bad way) so, while he has many friends, he also often gets into heated arguments with others. While they’re both on the swim team, Louis is also in the choir and is much more artistic. Unfortunately, his academics aren’t as good as Leo’s, probably because his attention span is so short. But he’s definitely the popular one in his grade and the other kids follow his lead. In either case, we’re blessed because, despite all that Leo puts up with his younger brother, he dotes on him, one of the few very strong emotions Leo does express regularly. And Louis looks up to his brother and tries to mimic what he does. After all of the frequent arguments and fights they have when playing together, they take care of each other when it matters.
 
One thing Louis has started to pick up is cooking. Leo has absolutely no interest in cooking food but eats pretty much anything except fried foods and, for some reason, mushrooms. Louis, on the other hand, has more particular tastes and knows what he likes and dislikes. Perhaps that’s why but, on weekends, he wants to cook together. I’m not much of a cook except for breakfast where I make a traditional American breakfast on weekends with eggs, meats, toast, and maybe some cottage cheese or fruits with a glass of tomato juice and coffee. Among breakfast, the egg is most important for me, but I only do over-easy, mainly because the others are too difficult (I’ll occasionally try to poach an egg, usually unsuccessfully). But after all of these years, Louis recently showed me something I never realized. I generally don’t boil eggs, mainly because I hate peeling the shell and I usually end up leaving the membrane while I peel, causing me to break the egg. And once you lose that membrane, it’s hard to get it back, sort of like when you don’t tear your Saran Wrap properly and it sticks to the entire roll, another failure I repeat frequently to my wife’s annoyance. Louis came up to me one day and told me I’m cracking it wrong. He simply cracked the bottom and peeled it upward. This is because there’s a small air pocket on the bottom so it’s easier to peel the shell with the membrane attached. Ever since then, I’ve added a whole variety of egg dishes to my menu using boiled eggs (mostly different types of salads). I asked him if someone taught him that and he said, no, and he just happened to notice the space on the bottom after peeling a boiled egg a few times.
 
I see AI a lot like Louis. It’s observing and adapting. I use it daily for work and home (and just random learning). I used to pride myself in being an “expert surfer” as I was able to find pretty much anything I need to find through the web. There used to be an art in surfing. Now, I just ask Gemini at home and Copilot at work. But I’ve noticed that, unlike my precision as a self-proclaimed expert surfer, AI is frequently incorrect. Still being a little suspicious about AI, I often double check via my own surfing skill, depending on how vital accuracy is for whatever it is I’m doing. When I find these errors or areas of incompleteness or biases of answers, I provide the context and why I think it might not be providing the information that I’m looking for, and it always corrects itself. What I find particularly helpful is that it remembers. So, a few days later, when I need to follow up again, I don’t need to repeat myself, unlike in the past where I had a dozen tabs opened in my browser plus another list of URL links that I’d email myself. Now, this has its negatives too. Sometimes I want to change the context but AI, not knowing this, will answer based on the context provided whereas I’m actually asking a similar question based on a different context that it obviously doesn’t know yet. So, I need to again teach what I know and frame the query. But as it continues to learn more about me (as well as the rest of the world through the infinite number of queries being processed), it continues to adapt and I’m sure at astronomical rates.
 
Still, despite all of the hype, it’s still nowhere near perfect. In fact, I recently had a problem with my wife’s new iPhone. I kept asking both Copilot and Gemini as well as using my own expert surfing capabilities, but I couldn’t resolve the problem. And each time I tried, I’d have to reset the phone, wasting yet another period of time. Ultimately, I had to call customer service, wait 40 minutes until my turn came up, and actually talk to a human! But, you know what? The operator fixed my problem! It wasn’t trivial and, like I do with AI, I had to keep providing the operator a little more context of what I did or did not do and in what order. But eventually, we licked it. I haven’t told Gemini or Copilot yet … I’ll test to see if they can eventually learn without my input.
 
But what is clear is that the world is changing. And we need to adapt during such momentous times. In the past 18~24 months, our portfolio has changed significantly. We are much more technology-weighted than we’ve ever been, even going back to my previous shop (before that, technology was one of my specialties so, by definition, I had high technology exposure, though not necessarily long). This is part of adaption. The philosophy hasn’t changed. We still focus on quality businesses (based on sustainable competitive advantage) where the value looks attractive. This naturally means high ROICs and high margins on the quality front, while contrarian on the value side, i.e. investing when the stock is out of favor whether for sector or idiosyncratic reasons, as long as it wasn’t structural. I’ve never been biased against technology and, if anything, I am biased positively since they naturally have high growth. There are many tech companies with strong moats producing high ROICs though, admittedly, understanding why those moats are sustainable takes more work (first mover is rarely ever a moat in technology, though I think people tend to give that more significance than it deserves). Unfortunately, they also are rarely cheap and I had trouble accepting such high multiples, especially since my core belief is that technological supremacy is overrated, at least in the long run. But in the last year or so, growth has become the new value. More recently, software has become the new value, as growth rates are being viewed as structurally slower, if not negative, due to AI. Clearly, my own actions would support that theory. I don’t search via Google or Yahoo Japan as much, which would obviously mean that, on the margin, search and banner advertising revenue must be slower.
 
But I still feel that we’re jumping the gun on that shift, especially in Japan. I still use Yahoo Japan fairly regularly. When I need something specific, I might ask Gemini. But when I’m just casually looking at news for entertainment purposes, I go to Yahoo Japan. And if Yahoo Japan utilizes AI well, the click-thru rate of those fewer banners I do notice should rise. I still prefer to read the newspaper (several, in fact, over the day) rather than have AI summarize it; my default browser setting opens 12 news feeds in addition to my Bloomberg-filled monitor. When I needed to outfit our new meeting room with video conferencing, I used Copilot to suggest different solutions based on the size of the room, location of the monitor, and the size of the table. It was great at suggesting the exact specs of the cables needed to optimize my setup. But I still had to find the best priced product using a combination of price comparison sites, Amazon or Rakuten, and my own field research. I even wasted a few dollars buying a cable I didn’t need because Copilot got it wrong. Of course, this is several magnitudes cheaper than hiring a professional IT consultant to design the room for me. But the time it took wasn’t trivial either.
 
When it comes to work, I’m even more sensitive about the results from Copilot such as when I’m researching new companies or industries. I treat Copilot like a junior analyst out of school. I may not need to spend days surfing the web and reading reports nor do I need to spend hours building a model from scratch. But I do need to check the output carefully. And because I can’t use AI to check AI, I need to do it the old-fashioned way and actually read reports. And when I need accurate, real-time data, I may have to buy that data. AI is great at summarizing what I need to know. But when the details are important, I end up having to read the source (or confirm via a second source). And that generally defines what I ask during my first meeting with the company confirming or denying the facts that I have researched with the help of AI. In a nutshell, AI made it easier to prep but it can’t run the meeting for me much less make an investment decision.
 
AI is adapting. And so are we. But I doubt I’ll let AI do everything for me. I’m not spending less hours at work. I’m spending less hours doing ancillary and low-leverage tasks and spending more time doing value-add work (which may include other software being used more heavily). I’m not spending less time researching things at home. I’m spending the same amount of time researching more things. AI is optimizing my time.
 
But in the case of my wife’s iPhone, I spent more time doing less because of AI. I should’ve just gone to the shop and paid the extra $50 or so to have it done. It would’ve saved me several hours of frustration.
​