Archive 2Q 2019 Letter

Disclaimer: This letter is an archived letter to shareholders of the SaltLight Investment Holding company prior to the formation of the SaltLight SNN Worldwide Flexible Fund. The purpose of this archive to provide a historical narrative of our thinking and how it has evolved over time. We find that writing down our thoughts and allowing the passage of time to judge them are helpful in our learning and improving the art of investing. Results, past investments held and our thinking may have changed.

Dear Shareholders,

During the first half period, SaltLight Capital generated a growth in net worth per share of [redacted]. The JSE ALSI TRI index grew +12.21% – a formidable return from the JSE All-Share[1].

Our high exposure to SA Inc. has not rewarded us this period. Conditions have been some of the worst that I’ve experienced: a recession, a national election, property rights ambiguity and the lingering fear of a sovereign credit downgrade. This has resulted in a cloud of uncertainty and volatility in our market. The price-to-value gap continues to widen however I am still confident in the long-term prospects of the businesses that we own.

Much of the ALSI TRI’s performance in this half has come from a strong rebound in resource companies. SaltLight has zero exposure to this sector for one key reason.

In SaltLight’s June 2017 letter, I listed our simple criteria for businesses that we’d like to own:

  • A business that I am able to understand and have a reasonable assessment of long-term staying power;
  • A quality business with durable competitive advantages that compound invested capital at high rates of return;
  • Run by competent and ethical management who operate with all stakeholders in mind;
  • Available at the right price and therefore trading at a discount to intrinsic value.

Bullet two is most pertinent to this discussion. Commodity producers have a poor record of generating shareholder returns over sustained periods and I simply avoid them. In SaltLight’s June 2016 letter, I explained why:

“Theoretically, this means no growth in ‘value’ for a shareholder over the long term. The reason is because of what their primary product says on the can – it is a commodity. It is a product that has no brand and no customer captivity. A wedding ring manufactured with Russian platinum is no different to one made from Rustenburg platinum. A commodity’s value is created purely based on the vagaries of supply and demand. The lowest cost producer is usually the one who wins in the long-term”

Of course, from time to time miners can do very well for shareholders. It’s worth remembering that in 2015 when we started SaltLight, nobody wanted to touch resource companies. Their balance sheets were over-levered and commodity prices were low. For the patient shareholder, management has recaptured some lost value: miners took large write downs, closed non-performing mines, paid down debt and subsequently returned excess cash to shareholders. This has been a short-term windfall for investors who saw the cheap valuations, ignored the noise and made a quick buck. However, for long-term shareholders, returns have been on par with cash.

Sector3 Years CAGR5 Year CAGR10 Year CAGR
Market – All Share6.9%5.8%13.5%
SA Resources20.6%0.8%5.4%

Resource companies remain a significant component of the JSE All Share TRI (~27% of the index) and given our preference to avoid them, there will be ongoing times like these where we have a divergence to the main index.

Investment Update

In this period, my investment activity has been rather muted. Sometimes the best way to improve returns is to sit on your hands, avoid the noise and keep activity low (I talk about why later in this letter).

Despite the poor environment, our special situation positions (African Phoenix and Astoria) have started to bear fruit and are tracking their targeted IRR[2]s. Transaction Capital has turned out to be a wonderful addition to our portfolio and illustrates that betting on ‘daily necessities’ can generate returns irrespective of the economic cycle.

With new inflows and our first US shareholder, we now sit with sufficient cash to deploy. We also have a significant position in tax-efficient preference shares (yielding 13% on our capital) that could be recycled into new opportunities as they appear.

Why am I excited about the future? We own some wonderful businesses run by trustworthy managers who continue to perform admirably despite the difficult environment.

Discovery Holdings

Earlier in the year, Discovery launched their new banking product. The low-hanging fruit will be to migrate their 300,000 credit card customers to a transactional account. I’d encourage you to watch this video produced by the Discovery marketing-machine (here) that illustrates the millennial-friendly offering.

There have been a few new entrants entering the banking space over the last year, all are following a mobile phone, no branch strategy. The battleground in the banking sector has clearly morphed into a positioning of ‘lowest fees’ achieved through a low-cost IT infrastructure and a digital customer relationship. It is my view that Discovery is likely to avert the battlefield for ‘the lowest fees’ by reorienting its value proposition to a cleverly designed behavioural system that delights their customers.  It has one key advantage over new entrants — its integrated ‘customer aggregation’ platform that I wrote about in my June 2018 letter (link).

Using its integration with retailers and other suppliers, Discovery can provide a substantial incentive for prospective customers to switch banks. An example: customers can earn up to 75% cash back for designated healthy food items in a shopping basket if they shop at certain grocers (lettuce = positive points vs. chocolates = negative points). No other competitor has this kind of atomic basket-level data. Even if they did, Discovery has extremely favourable commercial arrangements with the most desired consumer companies that competitors have struggled to replicate. In return for these rewards, customers provide additional data, from which Discovery gathers insights to add additional suppliers and so on…The flywheel continues to turn.

Over the last few months we are starting to see the strength of the Vitality moat being tested. A competitor, frustrated with the inability to replicate Discovery’s Vitality product, has simply resorted to using a customer’s Vitality status as a major input in pricing their own life insurance. This is a unique battle for intellectual property as it asks a key question about the future of consumer aggregators whose competitive advantage is their deep understanding of their customers. Can another competitor simply use that IP for their own purposes? What data ownership rights does a customer really have? Does one own their raw data? Or does a customer also own the insights a platform has woven together about them? This new battle is now going to court (link) and I don’t have a firm view on the outcome. I have a niggle that this case is no different to a secret beverage formula and will remain Discovery’s IP.

Irrespective, Discovery’s earnings continue to be understated due to the investment in new growth vectors (~21% of earnings). Discovery’s other businesses are still growing at attractive rates.  Discovery Insure grew premiums by +21% and are now launching an adjacent commercial insurance offering. The China JV with Ping An grew premiums by +117% for the last half. Incidentally, the Ping An JV’s growth is outpacing all of Discovery’s existing markets. Attributable JV premium now exceed the SA Life business – despite only being eight years old.

Anti-fragile Businesses

I favour bets on timeless ideas that are relatively immutable across time and scale. One example of a theme that runs across many of our investees is a bet against poor public services. Invariably in developing economies, the most viable opportunities lie in areas where a government is doing a poor job of addressing a large need. A need that is usually provided as a public good in a developed market (healthcare, transport, or education).

Unfortunately, the reality for an emerging middle-income citizen in South Africa is that they have to pay for private healthcare, private security, private education and, increasingly, private power solutions. The public goods provided are simply too unbearable to endure and private enterprise finds cheaper, safer, more convenient solutions to those who can afford them. 

Some examples in our portfolio: Calgro produces social and private housing at half the average house price. Transaction Capital is funding largest private transport network covering 15m daily commuters. Curro, a new position, provides quality K-12 education at affordable fees.

Curro Holdings (COH)

I have been studying the SA and Brazilian education sector for over three years. Brazil is, at least, a decade ahead of South Africa and provides an exciting picture of where SA private education could be in a few years. I have been waiting for the right time to find an attractively-priced opportunity and with the decline in mid-cap shares, I have been slowly accumulating a position in Curro Holdings.

Curro was founded by Dr Chris van der Merwe in 1998. The first school was located in a church vestry with 28 children in Durbanville, Cape Town. Today, it has 138 schools and 51,000 students. It has been a remarkable story of doing well when government falters in delivering a public good.

The for-profit education business model is relatively simple. Fees are received in advance when a child starts the school year and most children staying through primary and secondary school will be a customer for twelve years. Schools are like toll roads: they require significant capital upfront but can earn high margin earnings over decades (and in many cases centuries).

At current prices, we are paying a slight premium to historic building cost of the current student capacity (excluding the value of land that has been banked and property price appreciation). I suspect that the market is likely to undervalue the duration of the earning’s power of the for-profit school business model. 

See the Appendix to this letter for a more detailed description of my Curro investment thesis.  



It is entirely not my sphere of competence to predict what the future economic outcomes of SA Inc. will be. Over the next year there are a number of crucial decisions that policy-makers need to make.   At this juncture, it feels like the country has hit a critical cross-road around structural reforms.

On a recent call with the Moody’s Country Analyst, the analyst noted a key peculiarity with South Africa. It has a well-capitalised, well-managed private sector contrasting with an over-indebted, poorly-managed public sector. The situation is not the typical context in which most countries run into trouble. SA has no housing bubble, no private credit bubble and no banking crisis. Unfortunately, it rests with policymaker decisions. A manager of capital has the unenviable job of trying to digest new information and weigh up how best to position a portfolio to protect against risks but also profit from a recovery. This kind of environment should be extremely favourable for the patient investor when uncertainty is high, risks abound, and a looming fear is palpable.

Recalling 2008

My professional career in the financial markets started in 2007 at the beginning of the top of a very pronounced bubble. I recall how in September 2008 we saw blue-chip South African companies being sold by market participants indiscriminately. At the time, the firm I was working for had two shareholders – FirstRand, the premier South African bank and Morgan Stanley, a top-tier US bank. In that September, FirstRand had declined 50% from its peak only a year before. Morgan Stanley faced a similar decline of 69%.

I will never forget going home on a Friday and wondering if Morgan Stanley would exist when I returned to work on the Monday. Thankfully the cavalry arrived at the last minute over that very weekend a Japanese company made a large investment in MS to recapitalise the bank. Hollywood movies would normally end here with a happy-ever-after scene and the share price would be up 30% after being rescued. However, this is not what happened: two weeks later, the share price declined by another 50%! Markets continued to be fearful and policymaker’s inertia was seen as the root cause of the problem.

What would have been the correct course of action for the shareholder of these two banks? Cut one’s losses and move into cash or hold on for better times? Without the benefit of hindsight, the future seemed ruinous with no end in sight. The investor who sold out based on this uncertainty made the correct decision…for another nine months.

As most readers would know, the market has since returned multiples on that capital. 

At the time, I specifically wrote down my valuable lessons that I wanted to remember in the future. There are two that I’d like to share with you:

  1. When the unthinkable happens and you think it can’t get any worse, even more ‘unthinkables’ can happen;
  2. Those who bought high-quality businesses in those very uncertain, fearful times created once-in-a-lifetime wealth.

Of course, during these fearful periods, it is unknowable when the ‘bottom’ will be. Timing them is exceptionally difficult. We must be realistic that in September 2008, there was no possible indicator that March 2009 would be the bottom. In a range of outcomes, there could have been worse outcomes and more uncertainty.

SaltLight’s portfolio

As we sit today, we have a few things on our side. We have an extremely discounted portfolio and therefore ‘price’ on our side. We also have a portfolio of high-quality investees run by trustworthy managers who provide products and services that will continue to be in demand for decades to come.

The ‘unthinkables’ may still yet happen. This may result in further ‘quotational’ share price declines in our investees. You will recall that we are forced to mark our positions to the last price on the last trading day of each period. These quotational prices, more often than not, do not reflect intrinsic value. Over a short period of time it may seem like the right decision would have been to move into heavy weightings of cash or buying what some suggest as safe-haven assets such as gold or offshore shares. However, I refer you back to my lesson #2. Patience and discipline can be extremely rewarding over the long term. History has taught the patient that ‘this too shall pass’.  

Once again, I would like to thank you for the privilege of partnering with you on this journey. It is an absolute joy to be a steward of your capital.

Appendix – Curro Holdings

Curro was founded by Dr Chris van der Merwe in 1998. The first school was located in a church vestry with 28 children in Durbanville, Cape Town. Today, it has 138 schools and 51,000 students. It has been a remarkable story of doing well when government falters in delivering a public good.

The for-profit education business model is relatively simple. Fees are received in advance when a child starts the school year and most children staying through primary and secondary school will be a customer for twelve years.

Demand Likely to Outweigh Supply for Some Time

We have some confidence that demand for affordable, quality education will likely outweigh supply for some time.

  • Abnormal increase in child population due to introduction of free anti-retroviral treatments: An abnormal population surge occurred with a large increase in birth survivorship (+13%) around 2003-2005 due to the national government being forced to offer free anti-retroviral drugs for HIV suffers. As these survivors reached school-going age, in 2009 capacity constraints were starting to appear in grade 1 enrolments (Incidentally, from 2021 these additional children will be entering into higher education)
  • A decline in per pupil government funding: Total public education expenditure has kept up with inflation however due to the ‘surge’, per pupil expenditure has declined 10% in nominal terms. Public teachers have continued to receive above-inflation increase which has put funding pressure on budgets to build schools. According to our data, Government has on a net basis closed public schools rather than creating new capacity (2010-2016: -735).

Class sizes have ballooned without additional school builds: 55% of public schools have class sizes of more than 30 students per teacher (averages: US – 16, UK – 21)

  • Middle-income public schools at over-capacity: the most detrimental legacy of Apartheid has been the two-tiered education system. Historically ‘white’ schools were given a special status to be able to charge parents additional fees to hire extra staff. Education standards in these public schools have remained high and children living in catchment areas around schools have a legal right to attend these schools. The challenge is that most of these areas have had considerable residential housing growth and school capacity has not met this increase. Student/teacher ratios have jumped.

These conditions result in parents having three options: (1) play the lottery to get a place for your child in a public school or (2) pick another public school further away or (3) look at private schooling options.

Government Not Building Schools in Middle-Income Areas

The Johannesburg northern suburbs are a great example. We obtained the national database and geo-located schools on a map (exhibit 1). For context, Johannesburg has been growing at 2x the national population growth rate. The government has not built a single new school (orange dots) in the Northern suburbs over the last six years. What is obvious is that private schools (blue dots), have had to fill the gap.

Exhibit 1- New Public and private schools built in Johannesburg (2010-2016)1[1]

For-Profit School Business Model is Long-Duration and Capital-Intensive

Exhibit 2: Curro: EBITDA Margins (by vintage)[2]

Schools are very similar to toll-roads. Their location is vitality important; upfront costs to build a campus are significant but future cash inflows can last over many decades. The business model follows classic ‘J-Curve’ economics. Once a school is mature, Curro’s schools can earn 30-35% EBITDA margins.

Post development is completed, maintenance costs a percentage of revenue are also rather limited. I recently visited my old high school (now 50 years old[3]) and much of it still looks the same after two decades since I matriculated. Comparatively, after a decade, a shopping mall or commercial building is out of date and every few years or so requires significant refurbishment to keep lease rentals up. As a matter of fact, older schools are actually become more valuable because a cross-generational culture and heritage is created over time.

New entrants fulfilling this unmet demand will require deep pockets to fund the upfront school development cost and start-up losses whilst the school fills up. New schools can take up to four years just to break even.  As each grade moves up, new grades come in and eventually the school is filled to capacity.

Supply and Competition

Competition in the school business is hyper-local. Proximity to a child’s school is an important decision factor for parents selecting a school (readers who are parents would be well aware of the ten trips a week they’re making to drop off and pick up a child at school).

Curro has been on an aggressive capex programme to build ‘strategic’ forts in growing catchment areas. There is always the risk that unfulfilled demand in catchment area is overestimated resulting in overcapacity. Larger school businesses with a geographic diversification will be able to absorb these capacity inefficiencies. For a mom-and-pop player, they could be fatal.

I have typically shied away from capital-intensive businesses because of their poor returns on capital after factoring in replacement capital. However, the education sector appears to be different to me because of the decades-long profile of fees from utilising the same asset, with minimal capital to maintain these cash flows.

To illustrate the point, it is worth inverting the problem to how many students it would take from today’s intake to cover the original cost of my newly-developed 1969 high school. In 1969, it would have cost R1.4m (the inflation-adjusted value of R1.4m in 1969 is R98m today[4]) to build a standard Curro school. Some quick math calculations would illustrate that the operating income from 233 students (a fifth of the total school) of this 2019’s intake, would have paid for that entire 1969 development cost. The school’s development cost, over time, has essentially become irrelevant. Importantly, time has not diminished the earnings power of the school over the last fifty years. The school was a well-regarded institution when I completed high school, it remains a quality school today and will likely continue to be able to keep its earnings power for decades to come.

There are risks to the thesis: a stagnant economy will make it difficult for parents to afford private schooling. Whilst parents will make every effort to keep their children in school, once a child is taken out of a school, it becomes very difficult for a school to replace that earnings stream. Given the high fixed costs, particularly development, the J-curve economics could work in the opposite direction.

At current prices, we are paying a slight premium to historic building cost of the current student capacity (excluding additional land that has been banked). Curro’s investment cycle is nearing an end and based on the EBITDA-margins by vintage (exhibit 2), cash generation over the next few years should materially increase.

Overall, I suspect that the market is likely to undervalue the duration of a school’s earnings power. On an EV/EBIT multiple, Curro may seem somewhat expensive today, but over time, I believe it to be rather cheap.

[1] Source: Department of Basic Education, SaltLight Research

[2] Source: Company

[3] For clarity, my high school is not a Curro school but is a government-run Model C public school

[4] source: