dartmonkey

You are being generous - I expressed myself poorly in referring to the notion that these $125 per share in after tax operating earnings could just be called 'net income' and put Fairfax at 9x earnings. Fairfax is actually a lot cheaper than that: net incoome is a lot higher than just after tax operating earnings. Operating earnings include a guess at underwriting income and interest income, but they don't account for retained earnings from mark to market stock holdings or realized and unrealized capital gains when the share price of some of those stock holdings starts reflecting their increased intrinsic value (I'm thinking particularly of Eurobank and Fairfax India.) Here's the quote from the annual letter (p.7): We can see sustaining our adjusted operating income for the next four years at $4.0 billion (no guarantees), consisting of: underwriting profit of $1.25 billion or more; interest and dividend income of at least $2.0 billion; and income from associates of $750 million, or about $125 per share after taxes, interest expense, corporate overhead and other costs. These figures are all, of course, before fluctuations in realized and unrealized gains in stocks and bonds! But what I was trying to express, awkwardly as it was, is that I think Watsa is just telegraphing that there are $4b per year in pre-tax operating earnings that already seem quite likely for the next 4 years, based on the earnings potential of present assets. Apart from gains from the stock portfolio, there are also a lot of assets piling up on the asset sheet for the next 4 years, and these will generate their own earnings. In other words, we presently have equity of $21.5b, and expect $4b annual operating in each of the next 4 years from those assets. But in just 3 years, we should have $29b in equity; those extra $7.5b in equity will produce its own return in year 4, apart from the $4b that we expect from current equity. I still don't think I've expressed this perfectly clearly, but the idea is that given the high returns on equity, 4 years of compounding should produce a lot more operating earnings than what we would get if all those earnings were being distributed...

Sort of the archetype of weak hands. It looks like a good company, but who knows whether Muddy Waters is right or not? Might as well sell, even after the 10% drop, since I’m still above my September buy price. I doubt many of us felt very threatened by the MW allegations, but if you don’t know better, how can you be sure enough to hang on and recoup your 10%?

========= Yeah I had to read that same sentence a few times. He needed an editor there. Now that we're down to 22.891m shares as of March 7th, using 20m shares outstanding in rough calculation is not too far off, and the annual operating income expected for the next 4 years, $4.0b, would be about $200 per share (ok, it would be $175/share usiing the exact number.) So it is $50 less from the combined effects of taxes, interest expense and corporate overhead. And yes, it would have been much clearer if he had given the 2 numbers, and a name for the thing you get when you subtract taxes, interest and corporate overhead: "$4.0b ... or $175 per share, which is $125 in net earnings per share after taxes, interest expense, corporate overhead and other costs.“ Two other things that I would have liked to see there: (1) He might as well have mentioned that this means the shares, currently trading at USD$1088, are at less than 9 times the anticipated earnings in each of the next 4 years; and (2) I think Watsa is really saying that he expects $4.0b in operating income for the next 4 years based on income from current holdings. Maybe he is just putting the bar very low, but when you are expecting to earn almost $3b for 4 years, while paying out $373m in dividends (at the current $15/share rate), and you are a company that has a book value of $21.5b at the beginning, I think it is reasonable to expect that you are have at least $29b in book value after 3 years. Is Watsa really saying he expects to still be making $4.0b in operating income in the 4th year, despite starting that year with a book value that is substantially higher? I don't think so, and I think the fact that he still says he thinks he can hit the 15% return on book target means that earnings in the 4th year would be 15% of 29b, or operating income of $6b, not $4b. I believe that $4b is the income he can already see coming, but there will be other income coming from things acquired with the $7.5b or so of retained earnings in the next 3 years (without even considering compounding...) What do you guys think?

Blackberry is now a $130 million position = 0.0021% of Fairfax’s $60 billion investment portfolio. Thank you so much for this summary, much appreciated. Minor quibble about a decimal place: Blackberry seems to be on the way to being 0.0021%, but it's not there yet. 130/60,000 is 0.0021 but that means it's still 0.21%. One other quibble - I would add one word to this sentence: This is another good example of Fairfax FINALLY exiting from a poorly performing legacy investment (financially and also in terms of involvement from the management team). Capital at Fairfax continues to shift to better opportunities.

Looking at the respective compositions of the S&P/TSX Composite (currently 225 Canadian companies) and the S&P/TSX 60 (the large cap subset of 60 big companies), it is clear that the sector balancing of the subset is pretty good, usually within one percentage point of the Composite, except for financials, which are already over-represented in the subset (34.5%,, instead of 31.1%), this being easily the largest discrepancy. Also, minimum turnover is preferred, and companies are not generally excluded unless they are acquired, go bankrupt, are restructured, etc., size not being one of the issues. It may well be that the 20 bps minimum has been a historical minimum, but if you delete a smalll non-financial, and include a big financial, you make the composition imbalance much worse. So my bet would be that Fairfax will not be added unless another financial is acquired or goes bankrupt, even if there are a number of much smaller non-financials in there. It might be a long wait.

Is there any compelling reason to believe he didn't cover almost the entire thing the day of the campaign, as he often does? Do we know for sure that they tend to cover quickly, or is this just a suspicion? As for Fairfax, before the conference call, on Feb 15 to be exact, they said they remained short, for what that's worth.

How do you get a comparable AY number for recent years without the benefit of knowing what the future reserve developments will be? You seem to be assuming that adjustments will follow historical trends, for instance taking this years CR as reported from 93.2 to an AY CR of 89.0 over time, is that right? And with average AY CR's from 2017-2023 almost 10 percentage points better than published CR numbers, as opposed to 3 points lower from 2007 to 2016, we should conclude that either they are have gotten much better at underwriting, or their standards for reserving have gotten much worse recently. Hopefully the former. But might it be fair to say that part of the pessimism about how much this company is worth is based on the suspicion that their reserving standards may have slipped and that the CRs (and AY CRs) will not be as good as they look? Along with skepticism that interest rates will not hold up and that the company might return to buying Blackberries?

ok, that's helpful. Your table does show that the volume of FFH.U that is traded is minimal, 0-20 trades in the last few months, with 0 and 1 being by far the most common volume in a given day. I have no idea what the columns after 'T-TSX' mean ('U-NEO-ATS', 'A-Alpha', etc.), but I probably don't need to know!

Yes, you're right, and twice for good measure. Corrected now. I think I have my answer, that no real trading is done on the TSX with the FFH.U ticker, and I am guessing it's just a way of referring to the US price of Fairfax, avoiding the hard-to-remember FRFHF OTC ticker that actually gets traded, but if I'm wrong, someone please correct me.

Yes, this is what I expected, except I can't find a quote for FFH.U on TD Direct Investing or Yahoo or Seeking Alpha ; well actually, yes, on TD Direct Investing there is a stale quote of $903.32 (bid $1013, ask $1019) so I suppose there are occasional trades. I am familiar with the TSX having .U shares denominated in USD, like FIH.U, which actually trades that way, and for FFH.U, I suppose this just means that the TSX will purchase FFH shares using USD funds, but given the absence of an active market on the Toronto exchange, buying or selling them that way would not seem optimal. Of course, once you have them, it makes no difference whether they are called FFH.to or FFH.U or FRFHF - a share is a share. But when people here refer to FFH.U, are they just referring to the USD value of FFH shares, usually obtained by looking at where they're trading over the counter in New York as FRFHF?

This FFH.U entity is new to me. I have usually bought FFH in CAD, but journalled it all over to my US account as FRFHF shares in order to avoid the automatic conversion of the USD dividend into CAD at disadvantageous exchange rates (TD Direct Investing likes to gouge its customers this way; one of these days I will get around to transferring these assets into Interactive Brokers which treats customers properly for commissions and especially forex and interest rates, but I have so far left it there because I have the same Fairfax shares there for >10 years...) Anyways, FFH.ca and FRFHF are the 2 share formats I am familiar with, and the ones I see quotes for on most financial sites. But what is FFH.U ? Is this just shorthand for the US FFH shares that trade OTC as FRFHF, or is it something else?

My feelings exactly. The difference between Berkshire's 1.5x book and Fairfax's 1.1x book is not so immense, until you consider that Fairfax has a huge float position and Berkshire has a relatively small one (about 130% of Fairfax's market cap, and 20% of Berkshire's.) Siince by definition float contributes nothing to book (it is essentially future insurance liabilities, along with present cash that can be invested), book is only part of the picture for Fairfax. Because of the huge float position, Fairfax is able to obtain a much higher earnings yield, which is why Fairfax is trading at an earnings yield of about 15% and Berkshire is more like 5% (when you back out the stock holdings and their income).

Float cuts both ways. Catastrophe losses will be magnified within FFH simply because of the asset to equity leverage. So good times...tons of income. Bad times...significant losses. There is no net tangible value of float other than it is a more useful version of debt. Leverage is leverage. I would say the leverage is investment leverage, not insurance leverage. Any pure insurer is at risk of having claims exceed premiums. But if you take Fairfax with its $30b of float , more than its market cap, and you compare it with Berkshire whose float is less than a quarter of its market cap, the leverage risk is on the investment side - at Fairfax, a negative return will be hugely negative because of the leverage, and a good return will be greatly amplified. But I don’t see a downside of this, if it’s mostly invested in treasuries, as it is at Fairfax. You can be pretty sure you’re not going to lose a lot of money on treasuries. I’d say that Fairfax might be safer with a lot more leverage on a bond portfolio than Berkshire with 20% of its assets in Apple.

Ok, sure, but can't they do the same thing with an equivalently-sized share repurchase?

It certainly makes it harder to understand what's going on than if they just bought back shares. I presume there are liquidity consequences or tax consequences (repurchases are now subject to a 2% tax in Canada) or versus taxes on gains on the TRS, or something else that motivates them to keep these positions, and I trust them to do whatever's best, but I would much appreciate some discussion of this by Fairfax management.

12. Why do they keep shamelessly manipulating book value with aggressive marks like the ones Muddy Waters has brought up? Ok, maybe we won't get that, but if they prefer, what are some marks where BV might reasonably be marked significantly HIGHER? And to what extent do the feel that BV useful for investors, anyway? 13. What's happening in some of the big holdings, like Eurobank, Digit, Atlas, Recipe, etc.? 14. What is the game plan with the TRSs, and how do they think about putting the choice between holding money in reserve against the TRSs vs using money to repurchase shares?

Wow, great pick-up! With hindsight, we should have noticed that the two missing words were the two words that were underlined.

Understood. Overtime, what credibility will MW have? Once they go after high profile companies and are exposed as nothing burgers, why would stocks crash on their report? I have lost most of my sympathy for Muddy Waters, whom I previously considered fairly courageous in revealing genuinely fraudulent companies. But yes, it is sad that they have squandered their reputation by firing this blank. In addition to harming lots of shareholders who will have taken fright and who will now regret selling their shares at a 10% discount, they have also squandered their future credibility for taking on the kinds of truly fraudulent firms the have targeted in the past, like Sino-Forest (a Chinese-Canadian pseudo-miner that turned out to be a complete fraud), JOYY, GSX Techedu, French retailer Casino, Luckin Coffee, etc. etc. Most of them genuine turds.

Part of the problem is the IFRS restatement. BV gain for the first 3 quarters of 2023 took it from $657.68 on Dec 31, 2022 to $762.28 at the end of Q3, for a 15% gain, not 33-37%. But all those historical numbers will have to be restated up, too, if we figure we can compare new IFRS BV to old GAAP BV. Anyway, I guess the numbers they stated were probably for Q3 so there's no scoop in this latest PR. We'll see if Q4 bumps us up any further.

That would make sense - although the PR doesn't say whether the 18.9% number is with or without dividends, the jump would be too big to go from 17.8% to 18.9%, but from 18.5% to 18.9%, it's doable. But then, we should also account for the 2023 $10 dividend. I don't know how they do this calculation exactly, but to go from $657.68 on Jan 1 to a 34.6% gain for the full year gain, that would presumably take us from $657.68 to 657.68*1.346-10 = $875/sh at year end, no? About the same as Q3...

Yes, this is intriguing. We already had, from 1985 to 2022, a 37-y annual growth rate of 16.1% for share value. Share price was $802.07 (CDN) at the end of 2022 and $1222.51 at the end of 2023, a 52.4% gain. If you add a 52.4% gain to 37 years of 16.1% gains, you get a 16.9% annualized gain, definitely an improvement over 16.1%, but still far from 18.0%. Perhaps this is adjusted for dividend payments ($10US in 2023), but it's not enough to get to 18.0%. Shares peaked at $1404 before the MW report, which would get us up to about a 17.4% gain, still not quite there, but close. What about book value per share? Using the same reasoning, with a 37-year annual growth rate of 17.8% at the end of 2022, we would obviously need a pretty big gain in 2023 to get that rate up to 18.9%; in fact, the number would be =(1,189^38)/(1,178^37) = 1.677-1 = 67.7%. Does anyone think that 2023 was THAT good? As a reality check, Jan1 2023 BV was 657.68 USD, and that was up to $876.55 on Sept 30 2023, for a 33.3% gain (non-annualized) over 9 months. That would require a 25.8% (non-annualized) gain in Q4, which is a bit hard to believe, but not unfathomable, if the changed the mark on something big. Thursday evening is getting harder to wait for...

OK, question for the board: Given the fact that Fairfax has a normal course issuer bid outstanding, enabling them to purchase up to 10% of outstandings shares between 30 Sept 2023 and 29 Sept 2024, am I correct in thinking this allows them to continue purchasing their maximum of 8219 shares a day? (i.e., 25% of the average daily volume of Subordinate Voting Shares last year.) Given the fact that shares are selling at about $100 less than the pre-MW price, this would mean that they are saving almost $1m a day, thanks to the waters having been muddied for a little while. Unfortunately, I don't expect this savings to last long enough for it to be material.

https://www.eurobankholdings.gr/en/grafeio-tupou/etairiki-anakoinosi-05-02-24 Shares up nicely to $1.80, with one agency's approval of their increased stake of Hellenic Bank (from 29.2% to 55.3%). This may now be Fairfax's biggest equity holding, up from $2.1b at year end last month to over $2.4b now.

That would be ideal, mabe I will send investor relations a note. Thinking about this further, they note that the 5% annual hurdle, under which no performance fee applies, is not compounded. For a 3 year period, this makes very little difference: uncompounded 3*5%=15%, compounded 1.05^3-1=15.8%. But obviously, compounding over a longer period would make a difference. So when they look at the hurdle over 9 years, are they taking a 45% return from $9.62 (or from $10, maybe) as the hurdle? Over 9 years, 9*5%=45% but 1.05^9-1=55%, so it is starting to be important to know whether the hurdle is $14.50 or $15.50 - an extra 20c/share. And the difference will keep increasing, obviously, from one period to the next. In the very long term, it might mean that that 5% hurdle ends up being a lot less than 5%. For instance, for 2024 (year 10 of the plan), a 5% compounded hurdle would mean that they only take the fee on BV gains beyond $0.775/share (5% of $1.55), whereas an uncompounded 5% of $10 means they only take their fee above a 50c/share gain in BV . And in 10 more years, a 5% gain on the original $10/share would end up being like a 1.5% hurdle...

Giulio quoted the relevant blurb in the 2022 AR: Fairfax Financial is entitled to a performance fee calculated at the end of each three-year period, of 20% of any increase in Fairfax India’s BVPS (including distributions) above a non-compounded 5% increase each year from the BVPS at inception in 2015. So I think the fee is not as high as you are suggesting. It is not 20% of the excess over 5% from the high water mark, it is 20% of the excess over 15% of the high water mark, ever 3 years. See if you agree with my example: If the book value per share (BPVS) goes from $10 to $12.50 in the first 3 year period, then we've had a 25% gain, which is 10% over the 15% hurdle (5% annual hurdle, times 3 years, equals 15%.) A 'compounded" 5% hurdle would just mean that it applies above 1.05^3-1= 15.8%, instead of applying above 15%. Not really much of a difference, but it's always better for it to be clear! So in my example, the 20% performance fee is applied to that excess 10%, meaning that the fee is 2% over those 3 years, or a little less than 1%/year. If in the second 3-year period the book value goes sideways, say down to $12, then there is no fee; the fee only kicks in again above the high water mark ($12.50 + 15%*12.50 = $14.375. To take a real example, the 1st period (2015-2018) highwater mark, i.e. the BVPS on Dec 31st, 2017, was $15.24 (before the fee, which would end up reducing book value a bit.) A 5% non-annualized hurdle would have been 15%, so a book value going from $9.62 (the actual book value after commissions were deducted from the $10 IPO price), to $11.06, so the 20% performance fee should be calculated based on 20% of $15.24-$11.06=$4.18. Book value AFTER the fee was $14.46, meaning they took $0.78 per share, which represents 18.7% of the excess $4.18 amount above the hurdle. (There may be some technicalities of whether they make any adjustment for buybacks below book value which would tend to increase BV...) From 2018 to 2021, BVPS went from $14.46 (adjusted for the 2017 performance fee!) to $16.37, a much more modest increase, or 13.2%. That is under 15%, but the relevant comparison is not Dec 31st, 2017, it is back to the $9.64 opening BV again, but subtracting performance fees alread paid. So $9.62 plus 6 years of non-compounded 5% would bring us to $12.51, and so the performance fee should be 20% of the excess of $3.86, or 0.77/share. But since $0.78 in performance fees have alread been paid, there should have been no performance fee at the end of 2020. In actual fact, they paid $0.03 per share ($5.2m), so there is something wrong with my calculation, but it is close. Now we have finished the 3rd period, but we don't yet have the BV for the end of 2023. We know it was up to $20.89 at Q3 end, so the fee will be much more substantial, and in the Q3 FIH report, they noted that a performance fee of $82.6m had been accrued (amounting to about $0.61 per share.) Because there had been 8.75 years since the IPO, the calculation of the performance fee, if it had applied as of Sept 30, would have been that day's book value, $20.89, minus the 8.75 5% hurdles, or $9.64*(1+8.75*0.05) = $13.86, so the fee would apply to the difference, $20.89-$13.86=$7.03, times 20%, or $1.41, less the performance fees already applied, i.e. 0.78 and 0.03, leaving a new performance fee of $0.60. Given the BV gains in Q4, it is going to be a bit higher than this, maybe around $0.70. Please feel free to pick apart my logic!