StubbleJumper

You don't value insurance companies on a PE basis because the insurance cycle gives you a great deal of income variability. We are probably close to the top of that cycle, with peak underwriting income. FFH has an interesting advantage at the moment that it was running a short duration bond portfolio, which results in a bit of a lollapalooza effect at the moment. But, after a few years of solid underwriting profitability, new money floods into the sector and your peak income suddenly disappears. It's the same reason why a cyclical commodity stock is not considered a buy when it hits 6x EPS or 8x EPS at the top of the cycle. Valuing it with book works better. SJ

That is entirely possible. The fundamentals are very good for FFH. I have no trouble with that argument, as all it would take is a light cat year and/or a couple of transactions triggering major gains (like Digit). When was the last time FFH traded at 1.3 or 1.4x? It's been a while. We might ultimately get there, but it probably won't persist. That's a BRK-type valuation for assets that are not nearly as good as BRK's. I have no trouble imagining a 1.2x valuation, but it might take a few years to even get there. We will get good returns from FFH over the next few years, of that I am quite certain. There should be no trouble making the 15% ROE in 2023, and then if we are lucky, it might happen again in 2024 and then if we are super lucky, maybe again in 2025. But, the hard market must eventually turn, as it always does. In the mean time, growing book by 15+% and growing valuation from 1x to 1.1x or 1.2x will give a solid return...but if for some reason it does hit 1.4 (never say never) it's time to sell every last share. And, I would remind you once again, that Bradstreet elected to buy the preferreds a few weeks ago.... SJ

Not dirt cheap. Maybe BV will be more or less US$700 on Dec 31, 2023? So, choose your likely multiple of book that the stock will trade at: 0.8x, 0.9x, 1x, 1.1x, 1.2x? As you noted, currently it's trading roughly at 1x, and if it continues to do so, then we'd be looking at an end-year price of roughly US$700, which is 10% from where we currently sit. If we get lucky and have some multiple expansion, maybe we'll see it trade at 1.1x, which would be ~20% from where we currently sit? Those are good, healthy double-digit returns, AND they are not just plausible but rather probable. But, dirt cheap? Not quite. The interesting thing that nobody really addressed was @gfp's post from last week noting that Bradstreet has been buying preferreds, particularly the F and H series. Now, for the record, Bradstreet's forgotten more about investing than I'll ever know. He surely can see the same things that we can see, and probably more because he's an insider. Instead of doing the "obvious" and piling into the common shares, he's bought the floating preferreds. Curious, that. As you noted, it's a bizarre set up. But, let's try not to get ahead of ourselves. SJ

Sell.

No, the premium will probably end up being sticky to a large extent. It's possible that net written could decline if pricing is really bad, but in general volume trends up over time. What is extraordinary is the favourable CR at the same time as a favourable investment return. Usually you get either one or the other, but not both. What it means is if I gave Prem $1 to use in an insurance sub, he could use it to write $2 of premiums. Those $2 of premiums would give him 8 or 10 cents of underwriting profit, and probably another 10 cents of investment income. Seriously, a 18% or 20% return on incremental capital in an insurance sub? That's fabulous. Alternatively stated, it looks quite obvious that FFH will have little trouble making its target of 15% ROE during 2023....and maybe the real question is whether they are able to hit 20% or whether it's "just" 17% or 18% during 2023. This is not normal. This is Prem's "virtuous" part of the insurance cycle. In the end, all of this must end up attracting new capital to the industry which results in pricing pressure. This is the nature of the insurance cycle. We can hope that FFH can squeeze out 2 or 3 more really good years, but at a certain point new money jumps in. SJ

I don't disagree with any of that, but in general, "normalized" earnings would be an average taken across a whole insurance cycle. What we expect to see in 2023 is not a result that is normalized, but rather something that is extraordinarily favourable. We expect to see in 2023 is to once again be in a situation that Prem described in his letter to shareholders in the 2003 AR as the "virtuous" part of the insurance cycle, where we expect to get solid underwriting profits, solid investment income AND solid realized gains. That's not at all normal, and that's not at all the average (but I'll happily take it!). Other than nomenclature, I agree entirely with what you wrote. SJ

There is always a possibility that a US based fund is subject to the Government of Canada's withholding tax on dividends and has elected to temporarily sell a large position to some other fund with more favourable tax treatment, and is now re-purchasing those shares since it is trading XD. Usually, we should expect to see the share price drop about US$10 when it trades XD, and that actually happened on Wednesday, Thursday, and at the bell this morning. If somebody was doing that sale and repurchase trick, I hope they've already completed their transactions because they'd get screwed at this afternoon's prices. SJ

Yes, that seems to be the blockage. People do not seem to understand that a shareholder will be "stuck" with the CVRs, possibly for years. For clarity, this is from RFP's 8-K filing:

I hope you are correct, but I suspect you are wildly too enthusiastic. Current adjusted BV is roughly US$600. If all of the moons and stars align correctly, then perhaps the BV as at Dec 31, 2024 might be US$800. Then if all of the moons and stars once again align perfectly, the market might re-rate FFH and you might get a price to book ratio of 1.5, which would get you a share price of US$1,200. If I had to bet the over/under on that, I'd take the under because many things need to go perfectly for that scenario to come to pass. We are in a situation that promises good returns for the next few years, but the math can only take us so far. SJ

Yep. If you buy Resolute to get the CVR, you are stuck with it for some unknown period and cannot sell it or transfer it. Maybe you are stuck with it for 6 months, maybe a year, maybe 5 years. Nobody can say. Maybe it ends up being a zero. This is of particular importance if people elect to use @SharperDingaan's approach of holding the CVR in a Tax Free Savings Account or some other tax-advantaged account. Those CVRs will use up tax-advantaged "space" for some unknown period of time, so you need to apply a healthy discount rate to your estimate of the ultimate cashflow. So, maybe a 10% annual discount rate if you are going to use TFSA space, possibly for years on end? SJ

Nobody knows. Maybe $0, maybe as much as US$6/sh. The CVR is basically a claim on the potential refund of ~US$500m of duties that RFP has deposited with the US government, divided by about 76m shares. The theory goes that the US and Canada will eventually successfully negotiate a softwood lumber deal and some or all of the duties will be refunded. But, there is no guarantee that there will be a negotiated agreement and there is no guarantee that a duty refund will be part of an agreement. In the end, if 100% of the duties are refunded, the RFP CVR could be quite a profitable investment if a guy could buy it for 25 cents or 50 cents. But, the uncertainty of whether those CVR will be worth anything at all, or when the CVR will actually be disbursed means that I don't really want to pay much for them. As an example, suppose you believe that there is a 50% probability that the US government will ultimately refund 50% of that US$500m, but you suspect that they'll screw around for so long that it won't happen for three more years. In that case, you wouldn't want to pay more than ~$1 for the CVR ie, (50% likelihood x 50% refund x US$6)/ (1 + discount rate)^3 If they get really cheap at some point in the next few weeks, I'll probably buy some shares. But, I really don't want to pay very much as the likelihood, magnitude and timeframe of the payment are uncertain. SJ

Well, that's the question, right? We've seen several large-ish blocks over the past couple of weeks, and then that US$250m block yesterday. So who was it? Here are some of the large holders that I ripped off marketscreener.com : Southeastern Asset Management, Inc. 3,263,193 13.1% Mackenzie Cundill Investment Management Ltd. 2,595,071 10.4% Mackenzie Financial Corp. 1,552,806 6.24% Capital Research & Management Co. (Global Investors) 953,914 3.83% Capital Research & Management Co. (World Investors) 923,500 3.71% Vivian Watsa 794,534 3.19% Edgepoint Investment Group, Inc. 745,908 3.00% CI Investments, Inc. 624,985 2.51% BlackRock Fund Advisors 604,897 2.43% Davis Selected Advisers LP 596,671 2.40% The number of candidates who had 400,000+ shares to sell is pretty small. We can probably rule out "Vivian Watsa," better known to us as V. Prem Watsa. So, any ideas? You figure that guys like Mason Hawkins would be selling? When the Q4 numbers come out, what we will probably see is that adjusted BV (taking into account the deficiency of fair value over carrying value) will be roughly US$610. Slice off the $10 divvy, and we're talking a current BV of ~US$600/share. It seems overwhelmingly likely that FFH will make its 15% ROE target during 2023 as we know that there are significant gains to be booked, the fixed income port is finally kicking out some returns, and the underwriting market is fantastic. BV as at Dec 31, 2023 will likely be right around US$700. Even after the recent run-up, I just don't see the value guys selling at the current price of ~US$600. Anyway, by my rough arithmetic, @gfp has flagged a series of block sales totalling ~1.2m shares at the bell over the past week or so. Unless one of the larger holders both sold and bought back (ie, sell blocks totalling ~600k shares and then buy back the ~600k shares, thereby creating a total of 1.2m in block trades) it must be a very large holder. It will be interesting to see if all of this quiets down a few days after the XD date (the div is for shareholders of record on Jan 19, so XD should have been Jan 17). SJ

So, who has a quarter-billion dollars just laying around for a block trade like that? SJ

No apology required. I, for one, appreciate when you flag these strange trades as I rarely take a careful look at FFH's daily price quote. I usually just pull up the USD price to see what it's done during the day, so I wouldn't notice a large block trade on the Canadian market. SJ

It's always possible, I suppose. There might be some sort of time limit on their TRS contract, or there might be some sort of escalator in the carry cost that would provide an incentive to do the buyback. But failing that, we should consider some of the things that FFH could do with excess capital/cash and evaluate where the TRS might fit in the hierarchy of uses. I'm numbering some of these possible uses of capital/cash, but the sequence doesn't necessarily reflect where I see them in the capital allocation hierarchy: 1) Pay back debt 2) Keep/inject capital into the insurance subs to enhance underwriting volume 3) Buy back the minority interests from outfits like OMERS (eg the 10% Odyssey position) 4) Make an acquisition 5) Buyback shares on the open market through the NCIB or through another SIB 6) Close out the TRS There's a financial return to shareholders from each of these, but those returns vary drastically. So, paying back debt would provide a pre-tax return of perhaps 5% annually, which probably doesn't meet FFH's hurdle for excess capital at this time. Buying back some of the minority interests would provide a pre-tax return of perhaps 8% or 9% (because that's the ballpark dividend that FFH pays to OMERS), which is a bit of a better return on capital for shareholders (buying some Fairfax India would be an even better return). While Prem has promised that acquisitions will not be a priority, if something can be bought cheaply enough, that might provide a double-digit return on capital for shareholders. Buybacks through the NCIB or SIB might provide a one-time double-digit return for shareholders, depending on your views of FFH's intrinsic value compared to the prevailing market price. So, those are some of the alternative uses of capital. What do we get from closing out the TRS? What is FFH paying the counterparty annually? Personally I am guessing that it's more than the 5% that they could get from retiring debt, but less than the 8% or 9% that they could get by buying out OMERS's positions. Unless there is a contractual reason to do it immediately, I can't imagine it's a priority for FFH. That being said, I can't explain who else has been moving 250,000 shares at a time. Maybe it's time to take another look at the major holder list to see who might be getting out of FFH. The shares are currently trading at perhaps 1.01x or 1.02x adjusted BV. Would any of the large value guys dump FFH at that valuation when it seems obvious that FFH should have little trouble making its goal of 15% ROE during 2023? It seems to me like a no-brainer to hold FFH given the current prevailing share price and economic prospects. Unless there's some value guy out there who is selling because he sees a better opportunity than FFH during 2023? Anyway, just a bit of navel gazing... SJ

That's what is so silly about the current FFH situation. The "bad" scenario gives a 1.6% divvy plus a 10% unrealised capital gain, which is a perfectly good annual return in my book. You have to invent some sort of ridiculously catastrophic scenario to envision losing money. SJ

Valuations have expanded, but that doesn't mean that will be permanent or enduring. Price to book has oscillated in a very wide range over the life of FFH. There will be periods of gross over-valuation when it might trade at 1.4x and, as we've seen in the past couple of years, where are also periods when it trades at ridiculous multiples of 0.6x or 0.7x. But, as a baseline assumption, I'd say 0.9x to 1.1x is a reasonably likely range for the end year price. Don't get me wrong here! I'd be happy to see 2023 earnings come in at $150/sh, and I'd be happy to see a 1.5x multiple, but I certainly don't count on either occurring. But, overall, if the stock price came in at 1.0x the conservative book estimate of US$693, that would be a good return for 2023. SJ

No one can say what the market price might be, but you can make an educated guess about the range in which it might trade. To do this, start by developing a conservative estimate of book value on Dec 31, 2023: Current BV: BV on Sept 30, 2022: US$570/sh Deficiency of Fair value over carrying value: -$17 Adjusted BV as at Sept 30, 2022: $553 Add: Conservative estimate of EPS for Q4 2022: $50 Dividend to be paid in Feb 2023: -10 Conservative estimate of EPS for 2023: $100 Dec 31, 2023 BV: Adjusted BV as at Dec 31, 2023: US$693/sh The market price can vary wildly, but as a baseline, you should be thinking about a range from 0.9x BV to 1.1x BV. So call that a range from US$625 to US$760, assuming that there is nothing wild occurring in financial markets. And, I would say that the earnings number that I plugged in for Q4, 2022 and for the whole year 2023 will end up being considerably lower than the actual outcome, so there's probably some upside there too.... SJ

Yeah, and the arithmetic that I did was wrong even if the multiple voting shares had worked that way, so I deleted that post. SJ

If current trends continue, the super voting rights might become pretty irrelevant in a couple of years. Irrespective of whether some of us liked that change, the rationale underlying it was that the Watsa family had 41.8% of the voting power based on there being a total of ~23m shares outstanding. Prem seemed to know that he wanted to issue additional shares and seemed to be concerned that his control of the company would be threatened. In actual fact, the company did end up with about 28 million shares outstanding at one point. Clearly, the Watsa family's voting power would have dropped considerably due to FFH having increased the share-count by ~20%, so Prem's concerns were well founded. Fast-forward to today, and we are likely down to a little more than 23m shares outstanding due to the NCIB and SIB processes of the past few years (it was 23.4m shares on Sept 30). So we are approximately back down to the share-count that triggered Prem to have his multiple voting shares re-weighted. If we keep following this trend for another couple of years, perhaps FFH will be back down to 19m or 20m shares, which would provide the Watsa family with majority control *without* the benefit of the changes what were implemented in 2015. At that point, if Prem elects to walk away from his role as either CEO or Chairman, there would be no loss of family control. As I have suggested in the past, the buybacks are quite likely to continue for a few more years for the exact reason of maintaining family control. Prem knows very well that if he passed away, shareholders would have Ben and Christine on a very short leash. If he can get that share-count down to 19m or 20m, the kids would have 50%+ of the voting power. The power of incentives/motivation. SJ

@Viking Now you've done it. I was happily sitting on my ass thinking about making some supper before Monday Night Football and now my curiosity has gotten the better of me. So, I did a half-assed estimate of the change in fixed income based on a rough guess about the duration of the various fixed income maturity groups and the prevailing treasury rates for those groups: Assumed Duration Fair Value Sept 30 Interest Rates as at: Delta Y Delta V 12/09/22 09/30/22 Due in 1 year or less 1 7,571.80 0.0472 0.0405 0.0067 -50.73 Due after 1 year through 3 years 2 13,749.10 0.0433 0.0422 0.0011 -30.25 Due 3 though 5 years 5 5,335.30 0.0375 0.0406 -0.0031 82.70 Due 5 through 10 years 7 610.70 0.0369 0.0397 -0.0028 11.97 Due after 10 years 10 613.90 0.0357 0.0383 -0.0026 15.96 Total portfolio change 29.65 So, that's a shitty estimate of the change in the fixed income portfolio value based on a half-assed guess about the duration, and under the assumption of an instantaneous interest rate change (I've not considered Delta-t). The duration estimates for 3 through 5 years and for the after 10 years group are extra shitty, but that's life. The result of that shitty calculation would lead me to call it a M2M gain of ~$30m plus a bit of delta-t? Half-baked analysis like this is why they promoted me into management where I could do less harm.... SJ

What is your methodology for estimating the M2M bond gains of $250m? Treasury rates for 5-yr and 10-yr treasuries are definitely lower since Sept 30, but the shorter term rates such as the 1-yr and 2-yr have edged up a shade. I haven't attempted any arithmetic on the fixed income port, but my mental model was that from Q3 to Q4 it would be roughly a wash. My logic for that was the basic notion that dV/dy would likely be slightly negative because shorter treasury rates have edged higher, but that would likely be more or less offset by dV/dt being slightly positive. In any case, the M2M change for the fixed income port is pretty irrelevant when the bonds are intended to be held to maturity, but as you noted, it could make a bit of a difference for the EPS number that gets reported, and sometimes the market reacts favourably to a higher headline number (even if the earnings quality is dubious). SJ

You are being extremely charitable to even assert that predictability can be extended as far as six generations. Assuming a generation is 25 years, if you go back six generations in the United States (the year 1872), the civil war had just finished and the country had just begun to work its way through the consequences. If you go back six generations in Europe, the gunshot in Sarajevo that triggered the first world war hadn't even been fired, and the third reich and the second world war were still more than fifty years in the future. Almost no companies endure more than 150 years... SJ

Interesting timing. As I recall, FFH is scheduled to be paid a performance fee on December 31, 2023 if Fairfax India's book value grows sufficiently. So, is this IPO conveniently timed in a way that will enable a re-marking of Fairfax India's entire Bangalore Airport holding a few months before that performance fee calculation is made? SJ

If it actually does get implemented (lots of dumb ideas are conveniently forgotten), it won't be in effect until 2024. It is quite likely that the current government will be turfed in the next general election scheduled in 2025. In that context, it's probably not a big issue. SJ