dartmonkey

Gemini: fairfax trs arrangement If you are referring to the corporate financial structure, Fairfax Financial Holdings Limited (TSX: FFH) has historically utilized Total Return Swap (TRS) arrangements to manage its stock portfolio, hedge risk, and execute share buybacks. [1, 2, 3] Fairfax's corporate TRS arrangements operate through the following mechanisms: How Fairfax’s Stock TRS Arrangements Work The Structure: Fairfax enters into derivative contracts with institutional counterparties who purchase Fairfax subordinate voting shares on the open market. [1, 2, 3] The Economics: Instead of physically purchasing all the shares itself (which requires massive upfront cash), Fairfax agrees to pay the counterparty a floating interest rate (e.g., SOFR plus a spread). In return, the counterparty passes all total returns (dividend payments and capital appreciation) back to Fairfax. [1, 2] Cash Preservation: Prem Watsa (Fairfax's CEO) uses these arrangements to gain the economic exposure of owning its own shares without heavily depleting the company's cash reserves at the holding level. If the share price drops, Fairfax must pay the counterparty for the loss; if the share price rises, the counterparty credits the gains to Fairfax. [1, 2]

I think there will always be a counterparty happy to do this trade with Fairfax. They buy the equivalent number of shares, and bill Fairfax every quarter based on how the shares have done. There’s no risk involved, since if the shares go up, they pay Fairfax the amount of the gain but they will have made the exact same amount on the shares they hold (presumably with no taxes, since they have offsetting losses on the TRS.) And inversely if the shares are down, they’ll have lost on the shares but they’ll have won an equal amount on the TRS. Then whenever Fairfax wants to end the arrangement, they sell the shares (probably but not necessarily to Fairfax) and pay out or take in whatever amount makes them break even, while keeping the small fees they’ll have earned every quarter. What’s not to like? I would think other banks would be jealous of whomever gets to do this deal. [I should have read to the end of the thread, this is all well explained by RichardGibbons and Djokovic.]

They will get more creative. Yes, the TRS trick could always be used again.

The $1.7b they have spent on share repurchases is a big investment, and I agree that it is reasonable to ask whether they didn't have other, better opportunities. I wouldn't want them to buy back shares if it meant they couldn't take out minority holdings like Allied. But how would you compare buying more Fairfax to buying Under Armour, for instance, or buying wine producer Andrew Peller? I have SafetyinNumbers on the conscience, and I am imagining him saying, right now, that we are failing to make the distinction between the insurance companies (who are buying Peller and UA) and Fairfax the holding company, buying Allied and doing repurchases. But to some extent, the insurance subs could be buying more things like Peller and UA instead of sending dividends to the holding company to repurchase shares and buy out preferred shares and buy out the minority interest in Allied. So it may be fair to compare acquisitions, whatever level they are at. I for one think that Fairfax repurchases are a very sensible investment, and if they can't sensibly expand their insurance holdings, I would just as soon see them repurchasing rather than increasing the dividend or, like Berkshire, just holding onto a mountain of cash. And I like to see them stay small and nimble, like Berkshire was at an equivalent point in its trajectory. By my calculation, that would be in 1992, when Berkshire had a market cap of $15b, which would be the equivalent of Fairfax's current market cap of $35 in 2026 dollars. I expect Fairfax will keep growing, but I selfishly want it to stay under $100b for the foreseeable future, and not get to Berkshire's $1.1t where small opportunities like the $1.4b Fairfax paid for Eurobank in 2014 don't move the needle.

That‘s an interesting observation, and another accounting difference between share repurchases and the TRS, since swings in share price don’t affect earnings if they have bought back shares but do affect earnings and book value if they have established TRS instead. This means that the TRS might tend to increase volatility, which is just what we should want, given the fact that Fairfax takes advantage of low share prices to repurchase more shares. Share repurchase limit is 10% of float, or 2.1m shares, yes, thanks for the correction. That means they still have about 0.7m shares they can repurchase this year. If they have spent $2.8b on shares this year, maybe they will get to about $3.7b by September, reducing share count by a stunning 10%, but also preventing them from getting that much bigger. For comparison, they had record profit of $4.8b last year, and paid out about $300m in dividends in January. Yes they have sold some equities (Orla, Poseidon and Eurolife) but they can also buy back Allied and preferred shares, so altogether, 2026 will be a year where they will have postponed their growth in market capitalization, allowing them to remain that much longer in the sweet spot where they are big, but not so big that small investments stop moving the needle, a problem that Berkshire has run into by not paying dividends and only repurchasing small numbers of shares. There‘s a lot to be happy about with Fairfax in 2026, despite (and, to some extent, because of) the lull in its share price.

BTW, in that table, what does vol mean? It doesn’t seem to mean volume, here, or at least, if it does, it would just be the volume of shares purchased by Fairfax. For instance, in the first line, $944,698,824 is clearly the amount paid (in $C), +416,600 vol seems to be the number of shares purchased, $2267.6400 is the price paid per share (in $C, surely as a block trade), and 500,140 seems to be the cumulative number of shares repurchased within some period of time - is it 30 days? Or quarter to date? Where is this table coming from? The 416,600 shares represent about 2% of shares outstanding, and the cumulative total of 500,140 is clearly higher than 2%. I think they can exceed the 2% 30-day limit, not because they are buying them on another exchange (these are in $C so I presume they are on the TSX) but rather because that 2% clause only applies to investment funds, not insurance companies or conglomerates. The 2 limits that apply to Fairfax are the 5% of all outstanding shares in a year and the 25% of average daily volume, but with the latter rule not applying to block trades like the 416,600 shares purchased in the week before June 30 and filed on July 9.

I believe this is the current TSX document: https://cdn-ceo-ca.s3.amazonaws.com/1in03pb-TSX_Company_Manual_-_April_20__2023__English_.pdf (ix) "normal course issuer bid" means an issuer bid by a listed issuer to acquire its listed securities where the purchases: (a) if the issuer is not an investment fund, do not, when aggregated with all other purchases by the listed issuer during the same trading day, aggregate more than the greater of: (i) 25% of the average daily trading volume of the listed securities of that class; and (ii) 1,000 securities; (b) if the issuer is an investment fund, do not, when aggregated with all other purchases by the listed issuer during the preceding 30 days, aggregate more than 2% of the listed securities of that class outstanding on the date of acceptance of the notice of normal course issuer bid by TSX; and (c) over a 12-month period, commencing on the date specified in the notice of the normal course issuer bid, do not exceed the greater of (i) 10% of the public float on the date of acceptance of the notice of normal course issuer bid by TSX, or (ii) 5% of such class of securities issued and outstanding on the date of acceptance of the notice of normal course issuer bid by TSX, excluding any securities held by or on behalf of the listed issuer on the date of acceptance of the notice of normal course issuer bid by TSX, and for the purposes of (b) and (c), whether such purchases are made through the facilities of a stock exchange or otherwise, but excluding purchases made under a circular bid; In other words, the 2% in 30 days rule seems to only apply to investment funds.

But this document https://www.fasken.com/en/knowledge/2022/07/toronto-stock-exchange-staff-notice-on-normal-course-issuer-bids does not mention the 2% restriction. OTOH, it does mention the fact that only 25% of the average daily volume in the last 6 months can be repurchased in a given day, but that one block trade per week is permitted even if it exceeds this limit, and that no trades are allowed at the open (how many minutes after the open?) nor in the last 30 minutes of the trading day, and that trades on other exchanges do not have these restrictions, except in as much as they contribute to the total allowable shares authorized to be repurchased in the year by the NCIB.

This document https://corporatefinanceinstitute.com/resources/wealth-management/normal-course-issuer-bid-ncib/ would seem to suggest that the 2% rule during a 30-day period is back.

I tried to look up how this ruile was framed and found this, which suggest that the 2% rule wouldn't apply to a company purchasing its own shares: "To prevent the company from manipulating the stock price, exchanges enforce daily purchasing restrictions. For TSX-listed companies, the daily limit is restricted to the greater of: [1, 2] 25% of the Average Daily Trading Volume (ADTV) of the shares. 1,000 securities per trading day. [1] (Note: There is a separate 2% rolling limit applied specifically to investment funds over a 30-day period, which doesn't apply to standard corporate equity). [1, 2]. The second note goes to a 2007 Stikeman Elliott document that summarizes the change from the prior rule which DID prevent a compan from buying more than 2% of its own shares over 30 days: "The rolling 2% restriction on the repurchase of shares within any 30 day period has been eliminated for issuers that are not investment funds. The new limit is now the greater of 25% of the average daily trading volume (ADTV) and 1000 securities per trading day. This limit will only apply to purchases made through the TSX. The rolling 2% limit for investment funds remains. " That was 20 years ago, but it would not surprise me that there is a more recent version. Does anyone know?

Nice write-up, thanks.

That is a pretty weak expression of Buffett's idea. If the spread between bond yields and cost of float is not a positive number, then the insurance business is worth nothing. Simply being positive, as in, 4% yield on bonds, 3% cost of float (or CR of 103) would be a pretty shoddy result, and a long way from 'adding up to the good over time'.

OK, I was barking up the wrong tree. One chapter on the 3:1 leverage then!

What do you mean, taking leverage from two to one? By my calculation, Fairfax invest an amount roughly equal to its equity in equity investments, and roughly twice its equity in fixed income investments. So I would have said thet have 3 to 1 leverage*, down from about 3.6 in 2021, but not 2 to 1 and certainly not 1 to 1. But perhaps you meant something else, and I don't see comments in this thread from sholland or patterson that might make it clear what you are referring to. *meaning they have $3 invested for every $1 of their own book value. But it depends how you define it. $1 invested for every $1 in book value, is that 1x leverage, or 0x leverage? Maybe the latter, in which case having 3 times as much investments as equity would be 2x leverage?

Unrealized equity gains of about $1.3b (not counting the tax liability.) On a $42b bond portfolio, it looks like losses might be about 0.8%, or $336m, if rates are up from 4.01% to 4.20% as the 5y treasuries are. Does that look about right to you? What I should have also mentioned is that there will also be the realized gains of $837m from the Poseidon sale that will be booked in Q2. Then there will probably be the $350m gain from the sale of Eurolife to Eurobank, anticipated for Q3. Our investment gains in 2025 were $3050m from equity investments ($841m from the TRSs), $385m from bonds, and -$284m from other investments. It looks like 2026 is shaping up to be as good as 2025, except for the TRSs of course which are currently well down but could recover by the end of the year.

Change of value in stock holdings for Q2 Clearly the authoritative source of info for this will be when Viking gets around to posting his table, but as a first guess, I took the 16 biggest publicly traded holdings from the end of Q1 and looked at their value at the end of Q2, assuming no changes in the number of shares hold (an assumption which is unlikely to prove to be true, especially for Blackberry which has gone to $32.6 to $12.65, and which we hopefully haven't sold just in time to miss out on a big gain..) Anyways, the total gains should be around $1.3b by my estimate, with almost all of that ($1.01b) coming from Eurobank, with an honourable mention for the unmentionable fruit stock ($329m) which will surely be smaller than that, and maybe much smaller if they have sold it early in the quarter. All the other changes pretty much cancel each other out, with the biggest in absolute terms being our gold miners Orla (-$190m) and Foran Mining, which has become Eldorado Gold (-$153m). Then come Metlen Energy (+$119m) and Egypt's Commercial International Bank (+$103m), with all the others smaller than $100m. Just for comparison, realized and unrealized losses from equity exposure in Q1 were $82m, so Q2 will be .. much better.

Yes, First Capital was the example I was looking for, with the huge gains they got from selling it to Mitsui Sumitomo in for $1.6b in 2017. Second place in terms of size would be the Pet insurance companies that they sold to JAB Holdings in 2022 for $1.4b. Eurolife that they sold to Eurobank for $945m last year would be the third. In contrast, I can’t think of any example of Berkshire selling an insurance asset. This is a major difference between the two insurance conglomerates, and I personally like the more flexible Fairfax approach better.

Your numbers show that all metrics, net premiums written, investments, insurance float and book value, all increased slightly more per share than they did in absolute terms, so your conclusion ("Much of the dilution had been offset through organic growth") actually seems understated. It seems to me it should be "All of the dilution, and more, had been offset through growth." Apart from that nitpick, one other point that perhaps bears mentioning is that Fairfax has shown flexibility not only by buying insurance assets when the market was soft, but also by selling insurance assets when the price was very good. The examples that come to mind are Vault Insurance (2021), the big Pethealth sale (2022) and the life insurance business Eurolife (2025). They also bought the 50% of Gulf Insurance they didn't already own in 2024 (40% from Kipco, and 10% in a squeeze-out of minority investors.) In other words, they have been opportunistic about insurance companies, buying a lot when the market was soft, but also selling some when the market was hard, and even buying some (Gulf) when the market was hard but the strategic reasons were good.

He clearly ruled out issuing equity, unsurprisingly, and he seemed to agree with your suggestion of having Fairfax India act as a general partner for 3rd party capital (“So that’s certainly a possibility.”) And by mentioning that (“Fairfax Financial wants to invest in India.“) and that 3rd parties might also invest alongside FFH (“There's a lot of people who'd want to be our partners.")

Yes. I hadn’t realized that FIH had such a big stake in IIFL Capital, but in May they announced their intention to go from 30.5% to at least 51%, so that would be possible, assuming they have acquired as many shares as they intended to. They own 49% of Go Digit so I presume they would need Kamesh Goyal’s approval to make Go Digit a subsidiary, but it’s possible.

Sounds like they would be merged: "offers to make GoDigit and IIFL Capital subsidiaries of the bank."

I respectfully disagree - it’s not really complicated. It just never made any sense, and even Buffett eventually abandoned that way of thinking.

I agree with your Fairfax India example. But I think it's a bit different from Atlas or Resolute. Fairfax set up Fairfax India so that shareholders could invest in Indian companies alongside Fairfax. I have no issue with Fairfax purchasing more shares of Fairfax India from shareholders who wish to sell, but I would have an issue with Fairfax taking out the company to the detriment of shareholders who would have wished to keep the holding. It would not be consistent with the initial promise, so in that way it would not be fair, and it would also make it difficult to trust management if they ever came up with another scheme (like Fairfax Africa, or Fairfax IDBI for instance...) Whereas if someone bought Resolute shares to coat-tail Fairfax's idea, and ended up being bought out at a low takeover price, that's just a risk everyone has to take with any investment and there would be no contradiction with any implicit or explicit prior engagement made by Fairfax, so I don't see anything unfair about it.

Whether the motto is marketing or not, I don't believe that it was intended to mean that they will pay more than they have to for acquisitions so that no outgoing shareholders in the acquired company are disappointed with the price. They wouldn't last long as a public company if that were what it meant.

I think you can even take realized gains on the bond portfolio, which is not part of the 5% return from the fixed income portfolio that Fairfax has quoted. From p.66 in the 2025 AR: (8) The effective interest rate is the rate that discounts estimated future cash payments or receipts through the expected life of the fixed income investment to its gross carrying amount at initial recognition. The effective interest rate does not reflect changes in market interest rates that affect the fair value of the fixed income investment over time.