Pieter Slegers — Adding To Our Portfolio (Compounding Quality)

One-line take: two days after publishing a five-name buy list, the money goes somewhere else entirely — $50,000 added to three positions the portfolio already owns, not one of them on the shortlist. That is the most informative thing about this issue: adding to conviction beat starting a new position, and no sale was needed, so the "you'll be notified in advance" switch promised on 28 April was never triggered. The three transactions are published in full: Topicus $20,000, 285 shares, limit CAD 97; HgCapital Trust $15,000, 3,075 shares, limit 365 pence; Brown & Brown $15,000, 160 shares, limit $63. Each is bought on a different, model-appropriate metric — Topicus on Free Cash Flow Attributable to Shareholders (FCFA2S grew 23% in 2025 against Constellation's 14%; assume +20% for 2026 and the projection is €262m, "a forward FCF Yield of over 5%… one of the cheapest valuation levels the company ever traded at"), HgCapital on NAV discount (560p NAV against a 350p price = 37.5% off, described as "Mr. Market being way too negative on software companies"), and Brown & Brown on a cycle-trough argument ("the majority of insurance markets Brown & Brown is involved in are currently softening… it gives us an opportunity to buy more at an attractive price", with E&S CAT property rates down 15-35% offset by higher contingent commissions, Accession contributing $445m of Q1 revenue and Q1 EPS +8% to $1.39). The sizing is explicitly conviction-weighted, with the smallest position getting a top-up for being small: HgCapital "is also one of the smaller positions in Our Portfolio. That's exactly why we'll be adding to it." Separately, this is where Fairfax gets its intrinsic-value estimate — CAD 3,000 against a CAD 2,340 price, "a discount of 22%… we would love to buy Fairfax on weakness" — written on a flight to the first Berkshire AGM since Buffett's retirement, and sourced to The Fairfax Way and Prem Watsa's shareholder letters. Watsa's record is given as CAD 3.25 in 1985 to CAD 2,346 today — "almost a 1.000-bagger (!)".

1. Stocks & names mentioned

Three purchases, one valued-but-not-bought name, and two context references. Research legend: QT Qualtrim · SA Seeking Alpha · STK Stock Analysis. Written post with no timestamps — the At link opens the article.

Ticker	Name	Research	View	What he said	At
TOI.V	Topicus.com	QT · SA · STK	Positive	BOUGHT — $20,000, 285 shares, limit CAD 97. The largest of the three. "Topicus is a younger, smaller, European version of Constellation Software. That means it has more runway ahead… Scale works against you in M&A. Constellation has grown so large that small acquisitions barely move the needle, while Topicus remains small enough to achieve rapid growth from the same deals." A second, less common argument: "European sovereignty concerns and stricter data regulation act as moats for local software companies, a tailwind that directly benefits Topicus's acquisition targets." Valued on FCFA2S — cash left after operating costs, maintenance capex and minority interests, "very similar to Warren Buffett's idea of Owner's Earnings": +23% in 2025 against Constellation's +14%; assume +20% for 2026 and the projection is €262m, "a forward FCF Yield of over 5%… one of the cheapest valuation levels the company ever traded at."	read ↗
HGT.L	HgCapital Trust	STK	Positive	BOUGHT — $15,000, 3,075 shares, limit 365 pence. Bought on the discount and on position size. The NAV method is spelled out step by step — "take the intrinsic value of all companies they own, add the cash they have in the bank, deduct all debt" — giving 560 pence at last year end against a 350 pence share price, a 37.5% discount. "I think Mr. Market is being way too negative on software companies like the ones HG Capital Trust invests in." The sizing rationale is stated openly: "it's also one of the smaller positions in Our Portfolio. That's exactly why we'll be adding to it." Described in the summary as "the best Private Equity company in Europe… owns great software companies that you normally can't access."	read ↗
BRO	Brown & Brown	QT · SA · STK · FA	Positive	BOUGHT — $15,000, 160 shares, limit $63. Bought into a soft market on purpose: "the majority of insurance markets Brown & Brown is involved in are currently softening and slowing. While this does slow down organic growth in the short-term, it gives us an opportunity to buy more at an attractive price." Four supporting points: E&S CAT property rates down 15-35%, but "an inverse correlation where lower rates often lead to higher contingent commissions, keeping margins from falling too far"; live AI agents automating over 25% of end-to-end submission processes in programs and wholesale; management expecting organic growth to improve through 2026 as Accession enters the organic comparison; and Accession contributing $445m of Q1 revenue. Q1 EPS +8% to $1.39. "To my knowledge, this is the cheapest valuation level they have ever traded at", with 38.5% EBITDAC margins. The diagnosis: "Mr. Market is focused on short-term headwinds and missing the massive earnings power of the integrated Accession business."	read ↗
FFH.TO	Fairfax Financial Holdings	QT · SA · STK · FA	Positive	Not bought here, but given its first intrinsic-value number. "When Prem Watsa took charge of Faifax in 1985, the company traded at a stock price of 3.25 CAD. Today Fairfax's stock price equals 2,346 CAD. That's almost a 1.000-bagger (!)." The valuation: "I estimate the intrinsic value of Fairfax equals 3,000 CAD. As the current stock price equals 2,340, this implies a discount of 22%. We would love to buy Fairfax on weakness." Sourced to The Fairfax Way ("It's a must read") and Watsa's shareholder letters, read on the flight to Omaha. Note the change of basis from a week earlier, where the entry was 1.2x book = CAD 1,777.5. Bought on 16 August at a CAD 2,300 limit.	read ↗
CSU.TO	Constellation Software	QT · SA · STK · FA	Neutral	The parent and the benchmark, used to argue for the smaller sibling: "Constellation has grown so large that small acquisitions barely move the needle." Its own FCFA2S grew 14% last year against Topicus's 23%. Held at Very Strong conviction elsewhere; no fresh view here, and no add.	read ↗
BRK.B	Berkshire Hathaway	QT · SA · STK · FA	Neutral	The occasion and the comparator. "As this will be the first AGM since Warren Buffett 'retired', the big question is how Berkshire Hathaway will evolve from here. I don't think there is a next Berkshire Hathaway out there. However, there are some companies that come close" — the setup for the Fairfax section. No stance. The AGM itself supplies the material for the 7 May and 10 May issues.	read ↗
private	Accession Risk Management	—	Positive	Private; the Brown & Brown acquisition, now contributing. "The recent Accession acquisition contributed $445M in Q1 revenue", and management expects organic growth to improve through 2026 "as the Accession business enters the organic comparison." The thesis is that the market "is focused on short-term headwinds and missing the massive earnings power of the integrated Accession business" — a notably more positive reading than the 16 April view, where the deal's size was the reason BRO sat at Strong+ rather than Very Strong.	read ↗

Two arithmetic checks. (1) The Topicus order. 285 shares at CAD 97 is CAD 27,645, which at prevailing rates is close to the stated US$20,000 — so the "$20.000" is dollars and the limit is Canadian, the same convention as the August Fairfax purchase. (2) The HgCapital order. The published quantity "Q 3.075" is European decimal notation for 3,075 shares; at a 365p limit that is £11,224, consistent with US$15,000. And a valuation note: the HGT NAV given here is 560 pence with a 37.5% discount, against "an NAV of £5.62 and a share price ~30% below it" in the 16 April review — the NAV is the same, so the discount widened by roughly seven points in a fortnight, which is the reason for the top-up.

2. Talking points

Adding to winners rather than starting new positions

Two days after naming KKR, 3i, FICO, Fairfax and MSCI as the buy candidates, the $50,000 goes to Topicus, HgCapital Trust and Brown & Brown — all existing holdings.
The advantage is that no sale was required, so the advance-notice switch commitment was never invoked. The cost is that the shortlist did not direct the capital, and only Fairfax is eventually bought, three and a half months later.
Sizing is conviction- and gap-weighted rather than equal: $20,000 to the highest-conviction serial acquirer, $15,000 each to the widest discount and the cheapest-ever valuation.

Topicus — smallness as the thesis

"Scale works against you in M&A." The argument for owning the spin-off rather than the parent is arithmetical: the same €20m acquisition moves a €5bn buyer and not a €50bn one.
The regulatory tailwind is the less obvious point and worth carrying forward: European data-sovereignty rules make local software vendors harder to displace, which raises the quality of the assets Topicus is buying.
FCFA2S is defined carefully — cash after operating costs, maintenance capex and money owed to minority partners — because the minority interests are exactly what makes reported earnings misleading for this structure.

HgCapital Trust — a discount that widened into the purchase

The NAV calculation is spelled out for the reader rather than assumed, which is the right instinct for a vehicle whose "value" is a manager's mark.
560p against 350p is a 37.5% discount, up from roughly 30% two weeks earlier on the same NAV. Buying more as a discount widens is either discipline or averaging down, and here it is justified with a view on the asset class: "Mr. Market is being way too negative on software companies."
The position-size rationale is refreshingly plain: it is one of the smallest positions, "that's exactly why we'll be adding to it."

Brown & Brown — buying a broker into a soft market

The cycle is acknowledged as a headwind and used as the entry: softening insurance markets slow organic growth, and that is what creates the price.
The margin defence is the technically interesting bit: E&S CAT property rates down 15-35%, but lower rates raise carriers' profitability and therefore the contingent commissions brokers earn — an inverse correlation that cushions the decline.
The Accession framing has flipped. In April's review the $9.83bn deal's size was the reason conviction was capped; here it is "massive earnings power" the market is missing, already worth $445m of quarterly revenue.
AI is mentioned as an operating fact rather than a theme: "live AI agents automating over 25% of end-to-end submission processes."

Fairfax — a second valuation, on a different basis

CAD 3,000 intrinsic value against CAD 2,340 = a 22% discount, "we would love to buy Fairfax on weakness." Written a week after the 1.2x-book target of CAD 1,777.5, which implied the stock was 28% too expensive.
Both numbers cannot be right at once. The change is not flagged, and the eventual purchase at CAD 2,300 carries no valuation at all — so the archive's most-tracked new position was underwritten on three different bases in four months.
The record cited is the strongest version available: CAD 3.25 in 1985 to CAD 2,346 now, "almost a 1.000-bagger."

The Berkshire AGM as an annual fixture

Fourth consecutive year in Omaha, with a subscriber meetup organised with TJ Terwilliger — the same TJ who writes the ten-stocks and ETF issues.
The framing question — "the first AGM since Warren Buffett 'retired'… how will Berkshire Hathaway evolve from here" — sets up two later issues: 7 May uses Buffett's "the market is behaving more and more like a casino", and 10 May builds an entire S&P 500 alternative around Berkshire from what was said there.
Reading as preparation: The Fairfax Way and Watsa's shareholder letters, on the plane.

Publishing the order before executing it

All three transactions give amount, quantity and limit, one day ahead of execution — the same discipline as 28 June, 26 July and 16 August.
The trade-off is stated nowhere but is real: subscribers buying first move thin stocks, and TOI.V and HGT.L are not liquid names.

3. In plain English

A jargon-free summary of the thesis behind each argued name. (Renders on each name's consolidated page.)

TOI.V — Topicus.com Positive

Topicus buys small European software companies that serve one specific industry — dentists, housing associations, local government — and then uses the cash they produce to buy more of them. It was spun out of Constellation Software, which invented the model, and it runs the identical playbook in Europe.

The reason for owning the smaller sibling rather than the parent is simple arithmetic. Constellation has become so large that buying a €20 million software company barely registers in its results. For Topicus the same purchase still matters. Being small is the advantage, and Europe's software market is unusually fragmented, so there is a long list of things to buy.

There is also a quieter tailwind: European governments increasingly want data and software kept locally and regulated, which makes local software vendors harder to replace — exactly the businesses Topicus buys.

How it is valued: not on reported profit, which is distorted by the minority partners who co-own many of the subsidiaries, but on the cash actually available to shareholders after everything is paid. That figure grew 23% last year against Constellation's 14%. Assuming 20% growth this year gives €262 million, which at the current share price is a cash yield above 5% — among the cheapest the company has ever been. The purchase: $20,000, 285 shares, at a limit of CAD 97.

HGT.L — HgCapital Trust Positive

HgCapital Trust is a London-listed fund that owns stakes in private European software companies — businesses an ordinary investor could not otherwise buy into. You make money as those companies grow in value and through dividends.

Because it is a fund, the way to value it is to add up what its holdings are worth, add the cash, subtract the debt, and divide by the shares. That figure — the net asset value — was 560 pence at the last year end. The shares change hands at 350 pence. So you are buying a pound of assets for about 62 pence, a discount of 37.5%.

The judgement being made is that the market has become too pessimistic about software companies generally, and this discount is the evidence. Two weeks earlier the same NAV implied a discount of about 30%, so it has got wider, and the response is to buy more rather than less.

There is also a portfolio-management reason, stated plainly: this is one of the smallest holdings, so adding brings it closer to the weight the conviction deserves. The purchase: $15,000, 3,075 shares, at a limit of 365 pence.

BRO — Brown & Brown Positive

Brown & Brown is an insurance broker. It does not take on insurance risk itself — it finds cover for customers and earns a commission from the insurer, plus bonus payments when the business it places turns out to be profitable for that insurer.

Insurance pricing runs in cycles, and right now it is falling. Premiums for catastrophe-exposed commercial property are down 15-35%, and since brokers earn a percentage, lower premiums mean lower revenue. That is why the shares are weak, and why the shares are being bought.

Two things soften the blow. When premiums fall, insurers usually make more money on the policies, and Brown & Brown gets a share of that through those bonus commissions — so the two move in opposite directions and partly cancel out. And the company now has software agents handling more than a quarter of the paperwork on new submissions automatically, which lowers its own costs.

The bigger point is Accession, the large brokerage it bought last year, which added $445 million of revenue in the first quarter. Because the acquisition has not been owned for a full year yet, its contribution does not count as "organic" growth, which makes the headline growth number look poor. That reverses through 2026. Meanwhile profits per share still rose 8% to $1.39, and the shares are at what is described as the cheapest valuation in the company's history. The purchase: $15,000, 160 shares, at a limit of $63.

FFH.TO — Fairfax Financial Holdings Positive

Written on a flight to the Berkshire Hathaway annual meeting, this is the section where Fairfax gets a value put on it for the first time. The context is Buffett's retirement and the question of what comes next: "I don't think there is a next Berkshire Hathaway out there. However, there are some companies that come close."

Fairfax is the closest. Prem Watsa took it over in 1985, copied the Berkshire model of using insurance premiums as investable capital, and the shares have gone from CAD 3.25 to CAD 2,346 — very nearly a thousandfold. People call him the Canadian Warren Buffett for good reason.

The number given is an intrinsic value of CAD 3,000 a share against a market price of CAD 2,340, so a 22% discount, and the stated intention is to "buy Fairfax on weakness."

Worth noting for anyone following the record: a week earlier the stated entry was 1.2 times book value, which worked out at CAD 1,777 — implying the shares were nearly 30% too expensive. Now they are 22% too cheap. The valuation basis changed without comment, and when Fairfax is finally bought in August at CAD 2,300, no valuation is published at all.

Summary derived from the archived Compounding Quality post (text in transcript.txt) for personal study. Not investment advice. © Compounding Quality / Pieter Slegers for source material.