dartmonkey

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.

He's a very persuasive guy, and HHH is now my #4 position, after Fairfax, Fairfax India and Interactive Brokers. I don't know if it's a very good plug for Joe-like companies, though. When he mentions the HHH housing business that he intends to build his Berkshire-like company on, it is in comparison with the 'crappy textile company' that Berkshire Hathway is based on. In other words, a cheap, market-hated company whose decline can be managed while building something valuable on top. I don't think Ackman is really saying that HHH is a crappy company but it does have good assets that can be used for building something else. I'm not sure JOE can say the same thing, unless an Ackman-wannabe takes it over and uses its cash flow to buy things with higher returns.

My thinking is similar. They are unlikely to make 15% with this investment but if they establish themselves as reliable partners for retiring owners of good assets with steady returns, they might be able to do the kind of deals Buffett wished he could do but which Berkshire had become too big for. Fairfax can stay small for longer as Watsa has got the Singleton buyback religion far earlier than Buffett did, and this could be a competitive advantage for Fairfax.

It's a good question. I would venture that they are negative in the very short term, but no more so than for other insurance companies. In the medium term, they are good for all insurance companies, but in that sense, maybe not so good for Fairfax, since so much of its investments are NOT in fixed income. Higher bond rates would probably end up driving up combined ratios, as all insurance companies would be making more income investing their float. So low bond rates and good stock returns is probably a better environment for Fairfax.

I'm not sure of either these points. IFRS does require bonds to be marked to market, which means Fairfax would take a hit when interest rates increase, as they have (modestly). But as you say, the effect is not large, because of their short average duration, and in any case, its future liabilities also decrease because they are discounted at the new, higher rates, so my understanding is that it's mostly a wash. As for their fixed income portfolio, they typically invest their float in fixed income, but so do most insurance companies. The distinction with Fairfax is that they invest their equity in stocks and associate and consolidated non-insurance investments, providing for a higher return. Compared to other companies, they do not have more fixed income; float is 1.1x equity for Fairfax, for instance, and 1.3x at Impact, according to the RBC Capital Markets table that Safety has often posted here. So if anything, Fairfax is more diversified away from fixed income and should be less affected by bond rate increases, not more. But I would be curious to hear whether others here agree with this analysis.

I think most of us here are happy enough with the recent ~20% annual BV increases (or intrinsic value increases, if you will) and can settle for continuing 15-20% BV increases, regardless of the temper tantrums the market might take with its multiples. Multiple expansion is icing on the cake, but the cake is value expansion. Since we're at a low multiple already, multiple contraction looks unlikely, at least in the medium- to long-term, and if there is further multiple contraction, it will just make value expansion happen faster thanks to share repurchases. To get back to the original point about whether selling shares might be a good idea, I think this is what is setting this board ahum, as it goes against the instincts of value investors to sell low. It would be nice to have a small position that could be made bigger, but selling now is a horrifying concept for many of us, myself included. No wonder people are shaking their heads in dismay. I don't think it's denial - if the company were performing poorly, most of us would probably agree that sometimes it's better to take a loss than stick with a loser. But if the company is performing well, as is my belief, then it is naturally shocking to think about throwing in the towel at the exact wrong time, when multiples have just contracted.

Your are right, 15% is just average for me in the last 5 years or so because I have had half my investments in Fairfax, so I am spoiled. I don't really expect to get more than that in the long run. The 6% annualized from FIH has been a drag on my returns, and although is the average disappointing return from Indian shares, I really expected to do better, 11 years ago in 2015, especially given the unexpected bonus of having Modi continuously in power since then (beginning in 2014; now 2 more years in his 3rd 4-year term). Even better, considering that the book value return is much better than the FIH share price return, on paper, and that can't diverge forever.

I think the logic of holding this investment is not that another 75-100% might happen, just because it happened before. It is that 74% of BIAL might be worth about $5-10b, meaning Fairfax's 74% might be worth $22-$44/share including the after tax Anchorage gain. Assuming the other holdings are worth about $15, say $10 with the current discount applied, a successful Anchorage IPO might take FIH shares to $32-$54, a gain of 78%-200%. I should probably lighten up this holding which is way too big, with small gains, but I am terrified that the week after I do that, I will be looking at how many million % annualized I lost by selling half just before it goes from $18 to $45.

Yes, it's been a very long wait, but we may wake up one of these days to see that our shares have gone from $18 to $45, when the IPO gets priced, or maybe just when the IPO happens. Even that would only take this 11-year investment from being bad to being about average (I get 15% annualized if it went to $45 tomorrow), but it sure would make a big difference to my returns this year!

They certainly could do this. They have the cash to do whatever makes most sense, and getting rid of the TRS, while perhaps repurchasing an equivalent number of shares, would lessen the amplitude of swings in their earnings, which might meet with more market favour. On the other hand, bigger swings in share prices does have some advantages, for a company that wants and is able to repurchase a lot of its own shares, so maybe leaving the TRS alone makes sense, in a perverse way? If the TRS are contributing in any way to a negative feedback loop, then it’s more shares repurchased for the same dollar outlay, so we will end up in a better place. For a company with Henry Singleton-like ambitions, maybe they should have more TRS, not less?

I would be too, and so much the better. They have been very slowly reducing it: 5,940,000 shares sold in Q2 2025, 5,389,380 in Q3 and 415,100 in Q4, all at something like half today's price, so who knows what they will have done in Q1 2026 at similar prices and now in Q2 at much higher prices.

BB continues breaking out - closing today at $9.72, up from $3.79 at the end of 2025 and a triple from its $3.24 price at the end of Q1, or a $227m gain for Q2 so far. Still a long way from breaking even based on Watsa's description of a cost basis of $17.16 (although they made around $200m interest on the convertibles), but at least it will improve Q2 results if the current price holds up.

OK, let me make one last attempt to clear up where I disagree with the above, and then I will probably shut up. First, a quibble: the share price increase from €0.92 to €3.97 is about 330%, not 530%; I think you have a +1 in your table where it should be a -1. More importantly, when you say that Eurobank's demonstrates why book value and reported earnings are becoming less useful, and that much of the value creation never flowed through EPS or BVPS, I am agreeing with you about book value and book value per share, but not about reported earnings and earnings per share. This is because IFRS allows Fairfax to count its share of all the Eurobank profits, on its earnings statement, even if it doesn't get to count all that new value as book value on its balance sheet. Eurobank has increased in value by 330% largely because its earnings have gone from €15m in 2020 (9 months -shortened year) to €469m in 2021, €1388 in 2022, €794m in 2023, €1458m in 2024, and €1362m in 2025. Dividends paid out in those almost 6 years were 0, 0, 0, €410m, €240m and €556m, so assuming Fairfax's share was about 33% throughout this period, that means carrying value got marked up by about €5m, €156m, €496m, €184m, and €289m in those 6 years, or €1130m, which is roughly what happened to the carrying value of its Eurobank stake, plus or minus some forex adjustments. But the important point is that all those €1130m in earnings did get reported by Fairfax as they arrived, so if you are looking at reported earnings, or earnings per share (EPS), in other words the earnings statement, and if you are valuing Fairfax at some multiple of earnings (as I am, principally; currently at 8x last year's earnings, and now less than 9x this year's pessimistically assessed earnings, assuming average underwriting gains but only a 6% return on the non-fixed income portfolio, half their 6-year average), then there is really no distortion at all, and no hidden value. The distortion only happens with book value, where you have to take account of this extra hidden value that is not being counted on the balance sheet.

Because ROE has earnings in the numerator. For example, Eurobank has a ridiculous carrying value, so that its book value is far from intrinsic value, but Eurobank’s earnings are all reported properly, so earnings (whether it is eps or earnings as a percentage of book value) give a fair picture of intrinsic value. You can’t just add the CV-FV value creation as ‘additional earnings’, because then you would be double counting Eurobank’s earnings (and the earnings of the other associated/consolidated holdings.) I’m not saying you were doing this, just that someone might misunderstand it that way.

It hasn't been a great few years for value investors, that's for sure, and the odd time they took a stake in something that could have been brilliant, they have been quick to take small profits. But I wouldn't hang it in the hall of shame until the fat lady sings. A lot of investors sitting on huge unrealized gains may see them dissolve away in the next few years. Or not. But anyways, I'm quite happy with Fairfax if they can keep up anything like the recent pace, misses like APR and MU and INTC notwithstanding. I currently have FFH at 9x this year's earnings, assuming what I think is a very pessimistic 6% return on their non-fixed income portfolio, or about $1.5b, which would be half the average non-fixed income return (12%) of the last 6 years. With the Poseidon and Eurolife sales already in the can for about $1.2b, I think their chances are good of clearing that low bar.

I see your APR and raise Micron - they sold off their remaining shares at <$100 a year ago and the shares are now over $900. In Q1 2024, they had 3.9m shares, worth about $260m. Those shares would now be worth about $3.5b ...

Thanks for reviewing this. I think the last sentence I quoted is not quite right. The main point, I believe, is that accounting rules require the company to count towards book value only the historical cost of associate and consolidated interests, plus retained earnings and minus dividends and some other adjustments, as you pointed out. This is a good reason to distrust book value as a measure of value. However, earnings of associate and consolidated holdings are fully reflected in Fairfax's total earnings. This means that book value is off, but EPS and ROE are not affected in the same way. To take an example, Eurobank, which is responsible for almost half the difference between carrying value and fair value, is currently carried at $2.790b, whereas Fairfax's stake has a market value of $4.579b, as per your table. From Dec 31, 2021 to March 31, 2026, Fairfax's stake has gone up 6 times, from $800m to $4.6b, and this understates the rise since Fairfax has been selling shares to keep its stake below 33%. However, in the same time, total Eurobank earnings have gone from €328 to €1.35b, so Fairfax's share has gone from roughly €109m to €445m, a fourfold increase. If you are valuing Fairfax by taking some multiple of earnings (the current price is 8x last year's earnings), then your valuation of Eurobank's contribution has gone up by 4x using earnings, as opposed to 5x using market value or 2.4x using IFRS-mandated carrying value. Just one more reason to prefer using earnings (appropriately smoothed for underwriting and realized investment gains) rather than putting too much weight on book value. But my main point is that we should not be adding that $873m/year in FV-CV to the earnings of Fairfax, which would be double counting.

Who are the progressive, drunk, inexperienced FFH people here? Yes, many of us here like to think that we are conservative, sober and experienced-FFH people, so if petec is our leader, take us to him.

The ‚investment returns of 7.7% are only ‘subpar‘ because 2/3 of them are fixed income, mostly treasuries.

The total return swaps on 1.76m shares means that they lose $1.76m for every $1 decrease in their share price. The share price dropped US$206 in Q1, from 2025-12-31 to 2026-03-31, which is why they already lost $206*1.76m= $363m in Q1 (the company reported that the loss was $341.8m, which may be the effects of tax, or maybe there's something else that's wrong with my calculation.) But the point is, if we're counting in US dollars, the stock is off $206 from year end and about $300 from its peak (I believe this was Dec 30, 2025, at $1949 intraday on Jan 2, 2026) but that is already baked in as of the end of Q1. The share price was $1702 on March 31st, and is now slightly lower, at about $1650 as I write this, so that means there's another $50*1.76m= about $90m loss, pretty insignificant. Anything can happen to the share price, of course, but with $215 diluted EPS last year and very likely over $200 this year, I don't think there's much chance of the share price going lower than say $1000, and even that drop from the $1702 share price at the end of Q1 would be represent a manageable loss of about $702*1.76m = $1.2b pre-tax. And if they hold something like their current price, representing about 7-8x earnings, there would be no significant further loss in Q2-3-4.

I agree that $165 in EPS is unlikely for 2026 and $220-$240 seems more likely. Quick and dirty, Fairfax had insurance earnings from operations of $4.8b in 2024 and $4.6b in 2025. Adding losses from run-off and earnings from non-insurance businesses and backing out the effect of IFRS-mandated discounting of future claims, wiping out most of the gains from bonds, you get adjusted (my adjustment) operating earnings of $5.4b in 2024 and $4.3b in 2025. Then to get to pre-tax income, you have to subtract off interest paid and corporate overhead, which is pretty constant at about $1b, and add investment gains. Obviously 2025 was an exceptional year for gains on investments, going from $1.1b in gains in 2024 to $3.2b in 2025, so we ended up getting $5.6b of net earnings in 2024 and $6.4b in 2025. And barring a major insurance catastrophe, we are very likely to get something like $5b in operating earnings again in 2026 and 2027, as the company has reasserted many times, most recently at annual meeting a few weeks ago, and even extending the prediction (not a guarantee!) for another 2 years, to include 2028 and 2029. So what is a realistic guess for investment gains in 2026? We know we already have over $1.2b in gains from the Eurolife and Poseidon transactions, both in Q2, so I think a $2b gain for the year, about midway between 2024 and 2025, is a fair guess. Assuming that interest paid and corporate overhead remain at about $1b between them, that would put net pre-tax earnings at $5b +$2b -$1b = $6b, and assuming similar rate for taxes and non-controlling interests (25%), that would give us about $4.5b in net earnings, divided by about 20m shares at the end of the year, we should at $225/share. And every year that we buy 5% of the outstanding shares at less than 8x earnings (current share price is $1631) and accumulate more assets makes it easier to hit this $225/share target going forward after 2026. Possible positive deviations from this: realized investment gains of course, since I think my $2b is probably too low, but also Fairfax India has a lot of upside. On the downside, I didn't remove about $150m in earnings from the half of Poseidon that they sold, but I also didn't increase any of the earnings from their associated and consolidated companies nor from their acquisitions. Also, underwriting earnings may drop a bit as the soft insurance market develops, and then, there's always that next big megacap...

Maybe they want to keep it there, staring at them every year, so they don't get too cocky. Here's what they mentioned in the AR while gloating about the amazing Eurobank investment: "As the table shows, our annualized return since inception went from -7% in 2020 to 15% today as the Eurobank stock price increased. The long term is where it’s at and patience quite often (not always, remember BlackBerry!) is a virtue!" It looks like they have 34.98m shares left, right? With a cost basis of about $17, trading now at $5.74, or about $201m. Blackberry now disappears into the category 'Other' in the section of the AR where they talk about common stock holdings, but it may reappear in the 2026 AR if the share price holds up and if, as I expect, they don't sell it by the end of the year.