dartmonkey

I think you are right. Say I buy Fairfax at 1.5x book and it is earning 15% ROE. And assume, as you suggest, that that 1.5 multiple stays the same for a long time. I buy Fairfax at $3000 with its $2000 book value per share and it makes $300 next year, 15% return on equity or 10% earnings yield. Now its book is $2300 and with the same ratio its shares will trade at 1.5*$2300=$3450, so I have my $450 return on $3000 which is 15%. You have convinced me.

You could add ORLA to that list. Buying shares at $5.20 from 2022 to 2024, then selling 25m at $17.60 at the end of 2025 for a C$310m gain, is almost as much of a gain as Stelco (US$344m), and for a much shorter holding period and much higher annual return. And hopefully lots more to come, as they still own 31.8m Orla shares and 17.5m warrants. And while Resolute was indeed a wonderful sale, with great timing, it only bailed them out of a bad situation: "Our journey with Resolute began in a significant way in April 2008 with the purchase of approximately $350 million of an 8% AbitibiBowater convertible bond (at $10 per share)– almost 14 years ago! We added to our investment in Resolute in common shares and bonds over the years with a net investment after dividends of $715 million. With the interest income received on our bonds, sale proceeds of $622 million and with a little bit of good fortune on our remaining holdings in the contingent value rights, we may end up breaking even over this long holding period– although clearly a very poor long-term return." (2023 AR)

As an investor, I can't buy Fairfax's equity, I can just buy shares. If FFH is getting 15% ROE, then at 1.5x BV, it is getting a 10% return on my investment. So if they buy their own shares for $1737, like right now, then they are essentially using their cash to give me more of an investment that is making 10%/year. If the alternative is a fairly valued acquistion that will return 15% a year, I think they should buy it, as it will be better for me as a shareholder AND better for their capital requirements. I just don't think there are many (any?) investments where they can be as sure of a 10% return, so this should be their hurdle. And as we are slowly learning, subsidiaries have a different logic, and probably have to buy something not called Fairfax for their regulatory capital requirements.

Yes, good point. Acquisition price about $5.20 (not $52.0, sorry for the typo), share price end of Q3 was $10.80, so while the pre-tax gain on sale of those 25m shares at $17.60 was C$319m, the gain since Q3 would only be $171mm. Still a big contribution, but thanks for the correction, almost half the gains would have already been booked by Q3. So if UW and interest and dividends and share of associates' earnings come in at the same level in Q4 as in Q3, that would put us close $1.2b for the quarter, of $5b for the year. I thought maybe Eurobank's earnings would have improved in Q3 over Q2, but this is not so: Q1: E314, Q2: E377, Q3: E345. Fairfax in Q4 should include its share of Eurobank's Q3 earnings and these will be strong but actually not quite as strong in Q3 as in Q2.

Through Q3 they had $1.25b in net earnings for a total of $3.80b for the past 3 quarters, compared to $2.94b for Q1-3 in 2024. Operating earnings (ignoring changes in the equity portfolio), before interest and taxes, was $2.01b, or $4.52b for Q1-3, compared to $4.59b for Q1-3 in 2024. Market cap at the end of Q3 was $37.3b, and book value was $25.7b, for a P/B of 1.45. Market cap is now about $38b, according to Yahoo, almost identical to end of Q3, although it might be down if there have been a lot of share repurchases. So what are we going to get for Q4? Do we hit $5b in earnings, with another $1.2b from Q4, and $6b in operating earnings? (Watsa said that $5b in operating earnings were foreseeable for the next few years, but I think this is a minimum, not a central estimate.*) Q4 was quite on the insurance/reinsurance side. The Eurolife $250m capital gain will likely be in Q1 2026, not 2025. But the sale of 25m shares of ORLA, acquired at an average price of about $52.0 and sold at $17.60, should have generated about C$310m pre-tax, say US$181m post-tax. If we had the same $540m in UW profit, $586m in interest and dividends, and $217m in share of earnings from associates, like in Q3, that would bring us to about $1.5b post-tax with the ORLA gain, so $12.b seems like an easy bar. *Actually, it was Peter Clarke: "We continue to benefit from a stable base of annual operating income of approximately $5 billion, and we expect, of course, no guarantees it is sustainable for the next three to four years, with $2.5 billion from interest and dividend income, $1.5 billion from underwriting profit with normalized catastrophe losses, and $1 billion from associates and non-insurance companies." (Q3 earnings call).

He was also Berkshire's largest shareholder and understood there is a time and place for share buybacks. When/if the share price is cheap enough it is OK to repurchase some shares but it should not be a strategic objective IMO. I think this is probably a fair description of Buffett's thinking - he has never embraced the idea of repurchasing Berkshire shares in size. Watsa comes closer, with a significant reduction in the diluted share count from about 28 million to 21.5 million, and more like 20 million if you include the TRS. As others have pointed out, this share count reduction has just brought Fairfax back to the number of shares it had before it issued shares to buy Brit and Allied World. But both Watsa and Buffett have expressed the view that this is not just something that it is OK to do when shares are very cheap, but a real strategic objective. Buffett: "Henry Singleton of Teledyne has the best operating and capital deployment record in American business." and: "When available funds exceed needs of those kinds, a company with a growth-oriented shareholder population can buy new businesses or repurchase shares. If a company’s stock is selling well below intrinsic value, repurchases usually make the most sense. In the mid-1970s, the wisdom of making these was virtually screaming at managements, but few responded. In most cases, those that did made their owners much wealthier than if alternative courses of action had been 16pursued. Indeed, during the 1970s (and, spasmodically, for some years thereafter) we searched for companies that were large repurchasers of their shares. This often was a tipoff that the company was both undervalued and run by a shareholder-oriented management." (AR, 1999) Watsa: "By the way, you may not know, but the Michael Jordan of stock buybacks was Henry Singleton at Teledyne. Henry began Teledyne in 1961 with approximately seven million shares outstanding and grew the company through acquisitions while shares outstanding peaked in 1972 at 88 million. From 1972 to 1987, long before stock buybacks became popular, Henry reduced the shares outstanding by 87% to 12 million. Book value per share and stock prices compounded in excess of 22% per year during Henry’s 27 year watch at Teledyne– one of the best track records in the business. We will always consider investing in our stock first (i.e. stock buyback) before making any acquisitions." (AR, 1997) and: "As far as investing goes, Berkshire was always the biggest influence. With Teledyne, it was Singleton’s talent for allocation in acquisitions but especially with share buybacks.” Teledyne and Berkshire were rarefied company to aspire to join." (The Fairfax Way) and: "Watsa started talking more about Henry Singleton, the legendary CEO of Teledyne who orchestrated the most aggressive and successful buyback strategy in corporate history. Watsa called him “the Michael Jordan of buybacks,” and Fairfax was going to lace up its Air Watsas and play some serious buyback ball now that it was taking a break from buying companies: “Our long-term focus was clear. We had a much higher quality bar on acquisitions going forward and, as long as cash wasn’t needed to shore up financial stability, the first investment we would consider was buying our own stock.” (The Fairfax Way) In Buffett's case, while he has applauded share repurchases at companies he has invested in, he has only repurchased Berkshire shares reluctantly, since 2011, sometimes absurdly claiming that this might be to the detriment of exiting shareholders. I think he just ended up having to logically concede that repurchases made more sense than buying the kind of businesses that were available at that time, and so he started doing, in a small way, what he had always said would make sense for others to do. In Watsa's case, there seems to be a fair bit more conviction, and in the past 5 years, he has done us proud. Viva! And may he continue!

I think your answer to your own question (the 2nd on) is exactly right. It is true that thel ow share price limits how much Fairfax can expand its business thru share issuance, but in any case I do not want Fairfax to morph into a trillion-dollar company like Berkshire in 25 years and to then be in a situation where there are no further acquisitions that move the needle. I would much prefer for it to be able to sop up all of its retained earnings with share repurchases, and for it to remain at roughly the same size indefinitely. As others have said, the important thing for investment returns is that the company have a high return on equity, and having the opportunity to buy shares of a company it knows intimately, at 8-9x earnings, means they can not only keep to a reasonable size, but also generate high returns. It also means they have a high bar for making other acquisitions - logically, why would they buy take the chance of buying another company at 15-20x earnings when they can buy their own at 9x? As for the first question, I have no idea, but I suppose it has traded at a high P:B in the past so it might well do that again, if and when investors ever forget the short sales, the inflation hedges, and the Blackberry-like equity investments of 10 years ago. If that happens, I trust that Watsa will shift gears and halt the repurchases, and maybe issue shares again. This is the Singleton playbook, and Watsa has said Singleton is the gold standard, and while I would like to see even more aggressive share repurchases, credit where credit is due, Watsa has done this much more effectively than the great Buffett. Just one more thing to like about Fairfax...

Strange that it is still not on their website, but ok, I guess that solves that one. A few days later than usual, but the Friday pre-announcement for the next Thursday, somewhere between the 10th and the 20th, I guess is the rule. Why it wasn't yesterday instead of next Thursday, I guess we will never know, but it doesn't really matter either. Update at 17:12 - ok, it's on the website now. I guess the put out the press release before updating the website, a few minutes later.

Still no news from the company about the expected earnings release next Thursday, Feb 19, nor any reason given for putting out the report later than the February 10-18 range...

I expected results this week - they always report on Thursday after market close, and in the last fourteen years results have been between February 10 and February 18, and never as late as February 19. That would seem to make Thursday Feb 12 (this week) more likely than Thursday Feb 19. But I am pretty sure Safety is right, since they have always pre-announced the conference call by at least the Friday of the prior week, and often by the Friday 13 days prior, and occasionally even earlier, and never in the same week they published earnings. So this year is likely to be the latest earnings report date in 15 years (by one day), unless they have just forgotten to pre-announce, which is possible but, I think, much less likely.

It looks like they had 63m shares at the end of January, and eyeballing the prices, these might be: 6m in September 2025 at about $5 37m in December 2025 at about $4.1 20m in January 2026 at about $6.10. If that's right, their cost basis is about $340m, for an average price of $4.82. At today's close of $7.33, up $1.18 on Q3 results, that would mean their stake is now worth $462m, for a paper gain of $122m. Not shabby. They are guiding to a forward year EPS of $0.11, meaning they are at 67x 2026 earnings, which sounds awful, for a company with declining revenues and shrinking margins. But that is a bit deceptive, because their year end is March 31st, so when they talk about 2026 earnings, they are talking about the known past 3 quarters plus the unknown future quarter (Q4) which ends on March 31st. So far in their year '2026', they have lost .31c per share, so to finish the year at +0.11, they would have to make 0.42/share in Q4, which would be back to previous years' levels, meaning they are making enough to be back on track for $0.50-$1 annual per share earnings again, and the price of $7.33 would be quite reasonable. So I guess Mr Market is giving them the benefit of the doubt, and believing they have already turned things around. I hope he is right, but so far, so good.

So by my calculation, pre-IPO, there are 29,050,368/0.80=36312960 shares outstanding, and if these are priced at $28 in the upcoming IPO, they would be worth $1.107b. So the company has gone from a $436m valuation when it was taken private in Dec. 2018. Assuming the IPO happens in about a month, this would mean that in the 7 1/4 years since the take-private, the value of the company has appreciated by 1107/468 - 1 = 137%, or 12.6% annualized. Given the leverage involved in Fairfax investments, that is an excellent return. I am not surprised they are keeping all their shares, although their percentage ownership will drop from 80% to 52%.

Last year Fairfax published their FY 2025 results on the second Thursday in February, i.e. Feb 13, after markets closed, and that seems to be their usual practice, although they don't announce the future date on their website more than a few days in advance. And BMO responded on Tuesday Feb 18 with an increase of their target price to $2400: "BMO Capital Markets maintained a positive outlook on Fairfax Financial Holdings (TSX: FFH), raising its target price to C$2,400 on February 18, 2025, ...." Its previous price target was $2300, so I guess they figured that $50.42 in earnings for 2024 Q4 (against their forecast of $21.32) and record full year earnings of $173.24 made it reasonable to project a price of $2400 for the end of 2025, 9.8 times trailing earnings or 1.4x book value. And in fairness, the end of year price ($2616) and the current price ($2317) are not far off that target. And it should be noted that as results improved through 2025, they increased the target from $2400 to $2500 (May), $2800 (July), and then chased the share price back down to $2600 (November) and now $2500 (January). With earnings likely to come in at >$200/share for 2025, next week, or >C$273, it will be interesting to see whether they still think 2500/>273= <9.1 is an appropriate multiple. No doubt they will cite the fact that 2025 turned out to be an unusually benign year for underwriting, the hard insurance market is softening, and interest rates have been dropping and are likely to continue dropping, and some new investments are puzzling (UA, the new Blackberry?), and whatever else they can drum up to justify keeping their target in the same ballpark. But I'm guessing it will get bumped up again, at least to $2500 and maybe right back to $2800 if they don't want to have to keep chasing it again.

In other breaking news, companies that pay less end up with more cash for their shareholders but companies that are bought out cheaply do less well for theirs…

I think they could buy back a lot more without affecting the price. Average daily volume, combining FFH and FRFHF, is about 100,000, or 25m shares a year. If they had the liquidity and wanted to buy back 10% of the company every year, it would still be less than 10% of the volume and shouldn’t affect the share price. The major limitation is insurance-regulatory approval with the cash they have on hand. But with $5b in annual earnings, they should be able to do this if they make it a priority, as the cash arrives. Buying back 2% of their company every year represents about $700m, it is currently only a small part of what they are investing in.

Would insurance regulators not allow them to buy FFH shares (or something economically equivalent, like TRSs on Fairfax shares) at the insurance subsidiary level? I could see why they might not, but it would seem like a safer bet than buying Under Armour…

Count me as one who is perplexed by this investment. First, it is a $400m investment, more than one percent of the whole company, and while it may be #9 among investments, it is not like Eurobank or Orla which has gotten there because of its performance, it’s that amount invested, comparable to the size of another investment in a struggling company with a fruity name that I try to forget. With an insurance company I love and which they know intimately, available at less than 8 times earnings, I just can’t understand why they are putting a penny into a struggling retailer. Watsa praises Henry Singleton, a CEO who bought back 90% of his company’s shares. At the current rate, how many decades will it take for Fairfax to do the same? Why not plough into this opportunity, instead of buying back 2-3% a year?

Do you think they really expect Recipe or Sleep Country or Peak to return 15% ? Leveraged, sure, but I would think FFH has a higher expected return than these acquisitions. I would love to see them really get serious about reducing share count. If Singleton is their North Star, it’s going to take many lifetimes to get there at the current rate.

This may be your belief but Fairfax begs to differ: Fairfax Financial Holdings Limited is a holding company which, through its subsidiaries, is primarily engaged in property and casualty insurance and reinsurance and the associated investment management. Anyways, I was responding to milu's question about whether price to book was as useful for looking at Fairfax as it is for looking at Berkshire. I agree there are a lot of differences between the 2, primarily the fact that Fairfax really is mostly an insurer, with huge leverage from float, whereas Berkshire has become more of a conglomerate with huge wholly and partially owned equity investments. And there are many other differences, size being one, of course (Berkshire has a market cap 30x as big, and so is much more constrained in its investment opportunities), the international nature of Fairfax whereas Berkshire's activities are 90+% in the USA, and many more. My major point is that Fairfax gets a much larger contribution from insurance float than Berkshire does, and this should be reflected in a higher P/B, in my opinion. Do you agree?

Yes. In Eastern Canada where 2/3 of the population lives (i.e. Ontario, Quebec and the maritime provinces), or add coastal BC to get to 3/4 of the population, it is just not cold enough to rely on outdoor hockey rinks, even though there are a lot of them (about 5000). Many of them only function for 4-8 weeks a year, mostly for recreational skating, with a little shinny hockey, but organized hockey leagues almost never use them. So in most of Canada, hockey is a sport that is conducted in one of our 2,500 arenas, i.e. indoor skating rinks, sometimes year-round, sometimes skipping the hot summer months. And yes, the equipment is expensive too (which is a good thing for Peak). As is the rental of ice time, and all the coaching, officiating, travel expenses for games, etc. etc - the median cost was $1900 in 2023. Despite all this, it remains a popular sport, with over 600,000 kids registered in leagues, a modest decline for boys (because of the expense, maybe, or the violence, or I think more probably because so many other sports are now available), but with more girls playing who have slowed the decline and now represent about a fifth of all players.

The big difference between Berkshire and Fairfax is that Fairfax has way more float, in relation to its equity. At the end of 2024, Berkshire had: -$171b in float and -$649b in equity, so they can invest $0.26 of insurance premium float alongside every $1 of equity. In Fairfax,s case, they had: -$35.4b in float, and -$28.3b in equity, so they can invest $1.25 of insurance premium float alongside every $1 of equity. Berkshire trades at 1.47x book, Fairfax at 1.42x, about the same. But Fairfax has a much higher return on its book (ROE), in large part because of the huge extra boost they get from having about 5x as much float as Berkshire, compared to their equity. So Berkshire has an ROE of 10%, and Fairfax has an ROE of 17%. Berkshire has some advantages: it is far less levered, far less exposed to a really bad insurance year, and it indisputably has a set of businesses (BNSF, BH Energy, stock portfolio) that is less risky than Fairfax's. But to my mind, the company with a 17% ROE is still worth a higher multiple of equity than the safer one with an ROE of 10%.

I look forward to that post. Here's how I think about Fairfax: As you said a couple of days ago, they are more than a conventional P/C insurance company, they are a capital-allocation system investing float and equity in fixed income, equities (minority stakes, associate stakes, or consolidated stakes, depending on the percentage) and wholly owned companies. You talked about three complementary business engines: Insurance operations that generate large, durable, growing amounts of low-cost (often negative-cost) float A flexible investment portfolio, spanning fixed income and public and private equities A growing collection of wholly owned operating companies across industries and geographies I think this could reasonably be expanded to a few more, maybe 6 as you say. I like the RBC ROE insurance model that SafetyinNumbers has posted on multiple occasions, which is a way of estimating what their total returns on equity will be, which for the sake of convenience I will reproduce here, although I don't have an updated table (maybe Safety does): Using this, I think you can think about Fairfax as getting its returns from underwriting (first blue line), investment return (second blue line), a little other income which I presume is the administrative cost of head office (3rd blue line; is this right?), and little financing drag from their borrowing (4th blue line). The second blue line, investment return, is a mix of a bunch of things which I think are worth teasing out, primarily fixed interest (largely representing insurance premiums which regulators require them to invest essentially in fixed income securities, largely treasuries but also some corporate debt and mortgages), public equity investments (including minority interests like Orla, associates like Eurobank, Poseidon and Exco Resources, and consolidated or fully owned equity investments like Recipe, Sleep Country, Fairfax India, AGT, Meadow Foods and Peak. I like to think of the insurance companies as providing value in 2 ways (i.e. 2 'groves'): underwriting and investment income from the float. So Grove 1 would be underwriting, with an estimate of their return being something like 4% of thier $27b in net premiums written (($20.5b in 2025 first 3 quarters). This represents a combined ratio of 96, which may be slightly on the conservative side based on their 95 average over the lasst 10 years, but one bad catastrophe year could easily put them back to a 5-year average of 96. The second grove would be their ~$50b fixed income portfolio, which is mostly but not entirely comprised of their $38.5b in float, with low returns typical of fixed income, say 4%. The third grove would be their equity portfolio, excluding of course their insurance subsidiaries (already counted). I suppose this could be broken down into minority stakes and associated and consolidated stakes. The drag from financing (interest expenses from their own debt) and from taxes has to be considered somewhere, although it is not really a grove. And then the final grove is head office, which involves some expense (RBC estimates this as -0.7% of equity), but also a major profit source. After all, it is head office that pulls rabbits out of hats by selling companies opportunistically, not just equities but also the bond portfolio and even, occasionally, an insurance business (like PetCo) when the price is really good. I don't think it's double counting to give them some credit for creating value above and beyond the intrinsic value of their holdings, by selling them at an above-market price when the opportunity arises.

Trailing P/E seems to be 25, is your hypothesis that foreign earnings will increase in value sufficiently, if the dollar is devalued from 96 (today) to 70-80, to bring the ratio down to less than 15?

I don’t think it’s relevant. The position is owned by the insurance subsidiaries Do you mean that each subsidiary has a small share, not enough to qualify for any tax reduction on dividends received? In any case, from what I can see, Canada doesn’t tax dividends received from other companies, regardless of the size of the holding, in order to avoid triple taxation, different from the US where some tax is paid, the amount depending on whether there is <20% ownership (50% deduction), 20-80% ownership (65% deduction) or >80% ownership (100% deduction.)

Eurobank continues its stunning run, now up from €3.425 at year end 2025 to €4.308, i.e. 26% in less than a month, or higher if you consider the fact that the Euro is up 2% against the US dollar. Fairfax's owned 1178m shares at the end of Q3, representing about 32.4% of Eurobank, so if they still own those shares, that would now be worth a stunning $6.1b (at €4.33; the price keeps moving up even as I'm writing this...) As Viking pointed out recently, this means that the 'hidden value', i.e. the difference between the market value of FFH's stake and the carrying value on Fairfax's books has gone from $US2b to over $3b. However, Fairfax does count its full share of Eurobank's earnings, which for Q3 (or possibly for Q2 - they may be reporting Eurobank's earnings with a one trimester lag) were $140.7m. Eurobank is likely to end up with about E1.4b of net income in 2025, and at the average exchange rate in 2025 that would mean Fairfax's share would be worth 1400*1.14*.324=$517m. So book value rules means there is a lot of hidden book value, but the earnings are in plain sight, so if you are looking at Fairfax from an earnings perspective, there is no extra hidden value. And Fairfax's share of Eurobank's net income by itself represents more than 10% of Fairfax's projected $4.7-5b or so in earnings. Questions for others that occurred to me while writing this: (i) is there a reporting lag for Eurobank's results? I presume there must be, since Eurobank reports in late February (Feb 27 last year) and Fairfax in mid-February (Feb 13 last year) (ii) what is the Canadian tax treatment of dividends received from an associate with 32% ownership? (iii) Fairfax made $3534m in the first 3 quarters of 2025 (net of minority interests); Q4 was quiet from an underwriting perspective, and there will probably be some additional gains from obligatory Eurobank sales and increased Eurobank earnings, the $316m Orla sale, and the mark to market gains, principally Orla and the total return swaps (about $500m between them.) Is it crazy to think that Fairfax might get to $5b in net income in 2025?