Pieter Slegers — Our Shopping List (Part II)

Fairfax is a Canadian insurance group built on the Berkshire Hathaway model. Insurance customers pay premiums now and claims are paid later, so the company always holds a large pile of other people's money in the meantime. That pile is called the float, and Fairfax invests it. When both sides work — the insurance itself makes a profit and the investments do too — you earn twice on the same capital.

Prem Watsa has run it since 1985. He is best known for correctly predicting the 2008 financial crisis and making about $3 billion from it. About a quarter of the business is outside North America, notably in India, which most insurers have no exposure to at all. Book value — the accounting measure of what the company is worth — has grown 12% a year for over a decade, and the shares change hands at only eight times earnings. His salary has been CAD 600,000 since the year 2000, so he is paid by owning the stock, not by drawing from it.

The verdict here is not yet. For an insurance holding company the sensible yardstick is a multiple of book value, and the stated willingness to buy is 1.2 times book — about CAD 1,777 a share against a market price of CAD 2,472. Worth following what happens next: the same company is eventually bought in August at CAD 2,300, well above this target and on a completely different valuation argument.

Fortinet sells the hardware that guards a company's network — firewalls — and then sells the software subscriptions and support that make the hardware useful. The hardware is the way in; the recurring software revenue is where the profit is.

Once installed, it is very hard to remove. Replacing a security vendor means retraining staff, ripping out equipment, and being exposed to attackers while you do it — a cost no rival can discount away. Meanwhile Fortinet sees live attack data from more than 800,000 customers, so it spots new threats faster, which attracts more customers, which improves the data again.

Two newer products, sold to customers it already has, now make up 38% of new business and are growing 40% a year, which is why profit margins are expected to rise from 30% to 36% by 2030 without needing more revenue.

On price, the same adjustment used on Fair Isaac is applied here. The reported multiple is 29 times next year's profits, but Fortinet pays staff in shares to the tune of 15% of profit; count that properly and the real multiple is 34. The stated buying level is 25 times on the adjusted basis, about $66 against a market price of $85.

Aircraft need replacement parts constantly, and normally you buy them from whoever built the engine. HEICO takes those parts apart, makes an identical version, gets it approved by the aviation regulator, and sells it for 30-40% less. Airlines have to maintain their planes whatever the economy is doing, so the demand does not go away.

The barrier to competing is time rather than money. It takes years of regulatory approval and hundreds of certified parts before anyone can be taken seriously as an alternative, and airlines will not gamble on a newcomer — HEICO has thirty years without a single in-flight failure of one of its parts, which is a record you cannot buy. It also has a large defence business, with components buried inside targeting systems, satellites and missile guidance, where nobody swaps a supplier to save a little money.

The business is doing well: revenue up 14% and operating profit up 15% last quarter, helped by older aircraft flying longer and the original manufacturers pushing their own prices up, which makes HEICO's discount worth more each year.

And it is still passed on, because it costs 47 times next year's profits and the stated willingness to pay is 30 times — about $178 against $270. As the write-up puts it, this is a company "that has always been expensive."

Hermès makes the Birkin and Kelly handbags and deliberately produces fewer of them than people want. You cannot walk in and buy one; you go on a waiting list. The shortage is not a supply problem, it is the strategy — and the proof is the resale market, where a bag bought for €6,500 can fetch €35,000 at auction.

It also owns its production end to end, from the tanneries to the workshops, so it controls quality in a way rivals cannot, and it operates only about 300 stores worldwide. The founding family owns 67% of the shares and 78% of the votes, which is what allows management to keep supply tight instead of chasing quarterly sales.

While the rest of luxury has struggled, Hermès has grown revenue 12.8% a year for a decade. The shares have nonetheless fallen 33% over the past year and are now at one of their cheapest valuations in ten years — and the judgment here is that 36 times profits is still too expensive. The stated buying level is 30 times, about €1,364 against €1,676.

Mastercard does not lend money or issue cards; it runs the network the payment travels across and takes a small fee each time. Together with Visa it is effectively a two-firm industry. Think of a toll booth on world consumer spending that owns no roads.

The network cannot realistically be rebuilt: shops accept it because shoppers carry it, and shoppers carry it because shops accept it — a loop that has been turning for over fifty years. The long-run driver is the slow death of cash; digital payments only overtook cash globally a few years ago, and billions of people in emerging markets have not switched yet.

It is astonishingly capital-light: $52 of every $100 of sales comes out the other end as cash, and most of that goes into buying back shares.

At 26 times next year's profits it is called "not very expensive for the amazing business Mastercard is", and — the useful comparison — only slightly dearer than Visa, which this portfolio already owns, at 23.5 times. The stated buying level is 24 times, about $470.

Installed Building Products does exactly what its name says: it installs insulation, and while its crews are on site it also fits gutters, fireplaces and shelving. Its customers are American homebuilders, so its revenue rises and falls with how many houses get started.

Two things make it more than a contractor. It buys up small local installers in a very fragmented market, which is the same roll-up logic used elsewhere in this portfolio. And the government keeps tightening energy-efficiency rules, so more insulation is required by law — regulation doing the selling for it. Profits per share are up more than tenfold since 2016.

The problem is the price and the cycle together. The shares cost 27.6 times next year's profits; in 2022 the same company could be bought at nine times. For a business tied to housing starts, that range is the cycle. The stated buying level is 15 times, roughly midway between the extremes — about $168 against a market price of $306, the largest gap on either shopping list.

Interactive Brokers is an online broker used mostly by professionals and serious private investors. It earns money from trading commissions, from interest on customers' idle cash and on money lent to them, and from currency conversion.

The whole case is one statistic. Of every dollar of revenue, Interactive Brokers keeps 77 cents as profit. Schwab keeps 48. Robinhood keeps 45. That gap has persisted for years, and the explanation is automation: it looks after $780 billion of client money with just 3,200 employees, work that would take a bank ten times the staff.

It is also still growing quickly — customer accounts rose almost a third last year, more than a million new ones, off a base of 3.6 million — and it reaches 160 exchanges in 36 countries, infrastructure that took decades to assemble. Nobody is likely to copy it: the big banks have no incentive, and Schwab needed five years just to digest one acquisition.

The shares cost about 20 times next year's profits; the stated buying level is 15 times, roughly $51. One thing to keep in mind when reading this: Interactive Brokers is also the broker this portfolio itself trades through, named at the foot of every issue.