Dear Fellow Investors and Friends,
Welcome to my newsletter, where I share my efforts to understand markets and the world around me.
I do appreciate you taking the time to read this. I love getting feedback; feel free to drop me a line. It’s great to start conversations.
Today is Thursday, June 25th, the 176th day of the year. There are 189 days left until the end of the year. We’re just about over the hump; it’s almost all downhill towards the end of the year.
Today in 1530, the Augsburg Confession, a 28-article document that served as the basic confession of the Lutheran churches, was presented to the emperor Charles V at the Diet of Augsburg.
75 years earlier, in 1450, Johannes Gutenberg had invented the mechanised printing press. The invention was arguably the “best idea” of the last 1,000 years, paving the way for the Renaissance and a revolution in human knowledge. The Augsburg Confession was a direct result of this radical new technology.
Gutenberg was a craftsman in his mid-40s when he began tinkering with the idea of the printing press. His previous entrepreneurial endeavour – a failed attempt to produce metal frames for mirrors – ended in bankruptcy. “The little that is known about Gutenberg,” reads one biography, “often comes from the many court cases in which he clashed with his business partners.”
Despite a series of failures, Gutenberg began experimenting with various alloys, drawing on his background as a metalworker. This cross-pollination of experiences (coupled with a deep sense of urgency to settle his debts) led Gutenberg to develop movable metal type, ideal for printing.
The invention of the printing press enabled the Catholic Church to mass-produce indulgences – documents promising the forgiveness of sins in exchange for a monetary donation. By mechanising this process, the Church could distribute these lucrative certificates to believers across Europe on an unprecedented scale, rapidly generating substantial funds. The technology also enabled the Church to mass-produce uniform Latin Bibles. This made standardised worship and instruction easier, reinforcing the Church’s central authority and institutional support.
The printing press rapidly spread mass-produced classical knowledge, scientific ideas, and literature across Europe. So much so that one could say it was ground zero for the Renaissance.
In due course, the cost of books fell by roughly 80%, making them affordable to a growing middle class of merchants and artisans and creating a newly literate, critical-thinking public. Education was no longer a luxury reserved exclusively for royalty and the high clergy.
This commercial use of the press in the 1450s eventually had significant unintended consequences. Decades later, critics of the Church (most notably Martin Luther in 1517) used the technology to mass-produce critiques of the indulgence trade, thereby fueling the Confessions at Augsburg, which were foundational to the Protestant Reformation.
Here’s another story about technology:
Thomas Newcomen, an ironmonger, built the first useful steam engine in 1712. However, the engine was slow and inefficient, preventing mass adoption. A few years later, the Scottish engineer James Watt was asked to repair a Newcomen engine. In the process, he noticed a design flaw. A simple change to the engine greatly increased its efficiency. It was now ready for mass commercial exploitation.
Before the steam engine, factories needed to be near rivers for power and transport. Now, a factory could operate anywhere coal, labour and markets were available, not only in areas with fast-flowing water.
Steam made power portable and controllable. The Industrial Revolution was born.
But Watt was not a great businessman. His invention required capital, manufacturing expertise, and customers. These were secured through his partnership with Matthew Boulton, a Birmingham manufacturer. Their firm, Boulton & Watt, became one of the most important engineering businesses of the age.
The commercialisation of Watt’s invention increased productivity and reduced costs across almost all areas of industry. Coal mining became more productive, lowering energy costs across the economy. It transformed manufacturing, increasing output per worker by orders of magnitude. It revolutionised transport, reducing the cost and time of moving people and goods around the world.
Eventually, the steam engine contributed to rapid urbanisation and the growth of wage labour. But this went hand in hand with harsh factory working conditions, child labour and increased mortality. Pollution from intensive coal burning became a major problem. As countries grew wealthier and economies developed, people demanded better conditions. Gas, oil and electricity eventually replaced coal and steam as the primary sources of energy.
These are two examples, but all new technologies generally follow a similar arc:
- A new technology is stumbled across and “accidentally” developed.
- Early adopters use it to make a lot of money.
- The technology is inherently deflationary because it reduces production costs.
- The resulting surplus is inevitably transferred to consumers through competition, leading to significant economic advancement.
- The technology provides widespread access to information and knowledge (“democratisation”).
- Unintended consequences of this access eventually undermine the business models of the early adopters.
The upshot is that it is hard to pick the long-term winners in a new technology, especially if it is highly capital-intensive.
I reviewed Chris Dixon’s framework book “Read, Write, Own” on the evolution of the internet last year. In his model, the internet democratised access to information (read), then democratised the ability to create and publish it (write). The next step is to democratise ownership by giving users direct economic rights in the networks and services they use. As an aside, the blockchain could just be the way in which the “own” era is expressed.
Be that as it may, LLMs, agents, and the entire AI superstructure might just possibly lead to the disintermediation of ownership of bits and bytes from the incumbent “Mag 7” grouping.
Wouldn’t that be a surprise? Or possibly just normal service, according to the regular arc of technology?
I’m sure we will find out.
In The Markets
1. Reinet / Remgro
I write about investment holding companies a lot. There are a few reasons for this:
- I listed, and managed one, not entirely successfully. So, I understand how hard it is to run such a business.
- These types of companies tend to evoke irrational reactions from investors, which is fun to write about.
- Their investments and investment strategies are diverse and interesting to analyse.
In the past, I have written about two very well-managed and successful investment holding companies, Sabvest Capital and HCI, both of which are significant holdings in the MWI Value Fund. Both are managed by skilled capital allocators (Chris Seabrooke and Johnny Copelyn, respectively), resulting in a NAV (net asset value) per share that has outperformed the All-Share Index handsomely over time.
Today, I want to discuss two less successful investment holding companies, Reinet and Remgro. These companies are effectively the Rupert family’s onshore (Remgro) and offshore (Reinet) investment vehicles.
As a result of their association with the Rupert family, both companies enjoy an understandable level of prestige amongst investors. The Rupert family has successfully built Richemont into one of the world’s pre-eminent luxury goods companies, a remarkable entrepreneurial achievement. This company has created significant wealth for many investors over time. Here’s a well-written history of the company of which the Rupert family can justifiably be proud.
A ten-bagger over the last 17 years (in Swiss Francs) is a significant achievement:
Personally, I think Richemont is right up there with the best companies in the world. Sadly, the same cannot be said of Remgro or Reinet.
Remgro’s year-end is June, and over the 15 years to June 2025, it has grown its NAV per share by just over 7% p.a. – after adding back the value of all investments it has unbundled to its shareholders over the years. It has also paid a dividend of between 1% and 2% of NAV in most years, so the total annual return generated by management has been less than 9% p.a. This compares with the All-Share Index’s (ALSI) annual total return (i.e. including dividends) of almost 13% p.a. over the same period.
Remgro is only up 4X over the same time frame that Richemont has ten-bagged. The 11 dead years from 2014 to 2025 really hurt their performance:
I think it’s fair to say Remgro has not fared well under current management. I should add here that management itself has made out like bandits. But who said life is fair?
Reinet has a March year-end and recently reported its results. It invests entirely offshore, and its balance sheet is denominated in Euros. So it has a “hard-currency tailwind” – or at least the perception of one – something for which South Africans have an irrational affinity. Yet its results have underwhelmed, despite this so-called tailwind.
Over the 16 years to March 2026, Reinet has grown its NAV per share – converted into Rands – by just over 11% p.a. Adding back a 1% or so annual dividend gives a total return of around 12% p.a. The ALSI has returned almost 13% p.a.
Reinet was originally listed in 2008 at R19.39 per share, but then underwent a 10-for-1 consolidation in 2017, implying an equivalent listing price of R193.90. Today’s share price of R473 is not much more than double that! (Dare I say it – beware of exciting IPOs?)
I can only get the history since the consolidation in 2017, but its share price has less than doubled over the past ten years, a poor performance:
Over the past 18 months, Reinet sold its two largest investments, BAT and Pensions Guarantee Corporation. Cash now accounts for over 80% of its NAV. The remainder is a mishmash of disparate, uninspiring investments. Understandably, given Reinet’s track record, investors are upset that no action is being taken to distribute this cash. You can see their current mood in the chart above.
My take: Any rational investor should prefer buying the index to buying shares in either of these two investment holding companies. Both currently trade at significant discounts to their NAVs, penalising shareholders who vote with their feet. Mr Market is telling the management of these companies to give us our money back – you’ve had your chance, and it hasn’t worked.
2. Astoria
This was the investment holding company that was listed and managed with my colleague Jan van Niekerk. It was unbundled from Goldrush Holdings in April 2021 and immediately began trading at a discount of almost 50% to its NAV. During its life as a listed company, it had one really bad investment in diamond miner Transhex, four mediocre investments and an outstanding one in Outdoor Investment Holdings, the parent company of, amongst others, Safari Outdoor.
Overall, Astoria’s NAV per share grew by 14.1% p.a. over the period it was listed, below the ALSI’s 16.8% p.a. return – but not dramatically so. In fact, the share price outperformed, growing by 19.6% p.a. from the unbundling price of R3.90 per share. This outperformance was mainly due to the massive discount closing (a bit). But trading in Astoria’s shares was highly volatile due to their illiquidity, so there came a point where being listed made no sense for anyone.
We got the market’s message and made an offer to shareholders – c.R9 per share, including the unbundling of its remaining holding in Goldrush. At the time, the NAV was around R12 per share, and it was trading at c.R6 on the JSE.
Two-thirds of Astoria’s shareholders (including Jan and I) elected to remain on. We will continue publishing Astoria’s results for those shareholders and offering them liquidity from time to time. This week, we published the December 2025 results, which you can read here.
Here is Astoria’s history as a listed company:
My take: We gave it a good go, and 4 and a bit years is arguably too short a period to draw meaningful conclusions. But the fact remains that it’s super hard to outperform the index, which is why I have so much sympathy for the management of Remgro and Reinet, and so much respect for the management of Sabvest and HCI.
3. Fiserv
John Kenneth Galbraith coined the term “bezzle”. It refers to the period between the time an embezzlement is committed and when the victim discovers the crime. The bezzle comes in various forms. One of them is when companies overstate their earnings and understate their expenses.
The FT reports that all eyes are currently on the developing AI bezzle – analysts at Morgan Stanley estimate that a total of around $1.8 trillion in liabilities in the AI domain have been kept off the hyperscalers’ balance sheets. No one really knows, but we will find out sooner or later.
Personally, I think this bezzle permeates the entire US stock market. One way to create a bezzle is to artificially inflate expectations and take short-term measures to fulfil them. Like a pyramid scheme, the bezzle eventually becomes impossible to perpetuate. At that point, the bezzle is uncovered, and the (negative) imbalances accrue to the shareholders.
Here is the share price of Fiserv, an erstwhile market darling:
10 years of bezzle shareholder value wiped out in less than a year. What happened?
Current management revealed that prior management’s growth forecasts were “objectively difficult to achieve” and were driven by short-term metrics rather than by sustainable, long-term investments. To meet short-term margins and hit targets, prior management allegedly cut necessary research and development (R&D) costs and delayed software upgrades, creating an expensive backlog of customer change requests (referred to as “software debt”).
When new leadership withdrew the inflated guidance and reset the company’s financial expectations, Fiserv shares plunged by more than 40%. That was in October of last year. The latest development is that “new leadership” has resigned, causing the share price to decline further. The resignation was prompted by a class action lawsuit against them, brought by disgruntled shareholders who preferred the bezzle over the reveal.
My take: I view Fiserv as a poster child for the speculative activity that has taken over American stock markets. There is much more of this to come in future. Caveat Emptor.
In the cockroach
I’m still busy implementing the trades in the “hard asset” portion of the fund* I discussed last week. So, I thought I would take a quick look at the equity portion today, which currently looks like this:
As a reminder, I aim to eventually own 10 stocks here, each at a 2.5% weight – my “10 stocks, forever”. So far, I have bought 7 of the 10, though not all in full positions yet. From my original list, I discarded J&J/Roche, Coca-Cola, and LVMH. I replaced them with LSEG, Hermes, and DSM-Firmenich. I have also not yet decided between Apple and Microsoft for the 10. To complicate matters further, I have added Constellation Software to the “tech competition” for a place in the ten.
I discussed the starter positions in Hermes and Nintendo when I reviewed this portion of the fund earlier this year. I’m waiting for either the speculative activity in the market to subside or for a significant break lower before buying them up to weight.
I remain optimistic about the returns to South African assets, hence the large weight in the MWI Value Fund – one of the top-performing local equity funds over the past 10 years. But the exposures to this fund and to the MSCI World and MSCI Emerging Markets ETFs are placeholders until I can allocate the entire 25% of the equity allocation to my 10-stock forever.
So far, I am at 13.6% of the fund after starting the process just over 2 years ago. The problem with the type of stocks I want to hold forever is that they are widely recognised as high quality, so they are seldom cheap. It’s not as if I have discovered the laws of gravity or anything. These are all well-known companies, so I have to wait for a temporary setback before acquiring them at a price that makes sense.
If you run the maths, it’s true to say that if you intend to hold a high-quality stock for a long period of time (like, say, forever), then the price you pay doesn’t really matter. But being a value investor, I just cannot help myself from trying to acquire them at a reasonable price. That’s why this is taking so long.
Right now, there is nothing to do here; just wait patiently. Oh, and do some work on Constellation Software, which might just become my 10th stock in the 10-stocks-forever classification. It needs to beat either Apple or Microsoft by my criteria.
Let’s see.
* I manage the fund on behalf of Merchant West Investments (Pty) Ltd (FSP 44508)
In The Media
1. Oscar’s Substack
My good friend Oscar Foulkes has started a Substack called – wait for it – Oscar’s Substack. He used to have a blog called Oscar’s Pleasure, which is much more descriptive of what he writes about, namely, the good things in life. Which includes, but is not limited to, wine, cycling, and food. It’s like the diary of an epicure.
More importantly, Oscar writes well. His style is relaxed and flowing – just like riding the flowline on a single track. You can read it (and subscribe) here.
2. Abdullah Ibrahim
I’m not on social media and avoid the news as much as possible, so I only learned of Abdullah Ibrahim’s passing when I received my regular “Friday Song” WhatsApp last week. What a sad day!
Everyone knows his song “Mannenburg”, but he had an amazingly deep and wide repertoire. He was truly one of the global jazz giants. Here is a Spotify playlist showcasing the breadth of his work. My friend Mark Rosin, who puts out the “Friday Song” every week, also compiled a “best of” list, which is a more tightly curated reflection of his musical spectrum; you can listen to it here.
I am lucky enough to have two friends – Oscar and Mark – who I can look up to when it comes to writing. You can do worse than subscribe to their work; reading them is guaranteed to bring you much joy.
3. What’s a Watt? Horsepower. What’s a Julie? Brainpower.
Over the course of a few lunches, spread out over the past 6 months, I have gotten to know Michael Power. Michael used to be the investment strategist at Investec Asset Management/N91. Our careers never quite overlapped – I had left the business just before he joined, but I was always aware of his outstanding work.
I’m happy to report that, despite “retiring”, Michael continues to research and write. This week, I read his piece on how to measure the value of AI correctly – by measuring output, or “Julie”, rather than input, or “tokens”. In AI engineering, a “Julie” is a proposed unit for evaluating how much cognitive work an AI model actually accomplishes, defined as 1 Julie = 200 joules.
A Toyota Prius and a Ford Mustang will both get you to Stellenbosch from Cape Town. The Prius will do so much more efficiently, which is why compact, efficient cars dominate the market, despite the widespread availability of Mustangs and other sports cars. And guess which manufacturer is more profitable?
Similarly, Michael concludes that the most efficient LLMs (Large Language Models) will win the race, not necessarily those with the most brainpower (Julies). It should not surprise you that, as in most industries, the Chinese are building the most efficient AI.
I’m not sure I can count Michael as a friend (yet?), but if I could, that makes three friends who write very, very well. In business, you should always appoint people who are better than you to work for you. In writing, it helps a lot to have friends who are better than you.
You can read Michael’s work on Medium here.
4. The MWI Value Fund
This fund is run by Rudi van Niekerk and Josh Viljoen – no relation, he just happens to have a great surname. I give the odd input that, fortunately for the fund’s investors, Rudi and Josh generally ignore. Tomorrow, they are presenting a webinar on the fund and its outstanding performance. Over 10 years, the fund has performed in line with the ALSI at 11.6% p.a., ranking it 10th out of 93 funds – with no offshore exposure at all!
These guys are the real deal.
You can register to watch the webinar here. It’s tomorrow (Friday) morning at 9 AM.
Last but not definitely not least, Bafana has brought us massive joy by beating South Korea and qualifying for the knockout rounds of the World Cup. It’s great to have a winning football team!
FIFA clearly does not care about your sleep quality, so be careful out there – watching football at 3 AM can impair cognitive function the next day.
Piet Viljoen
RECM
25 June 2026