treasurehunt

I think this is roughly correct, although you are ignoring non-controlling interests. Non-controlling interests run around 10% of pre-tax income, I believe. But this is just for the first year, right? If Fairfax continues to trade at book, they can theoretically use all $3B in earnings to buy back shares and continue to generate a 15% ROE. But if Fairfax trades at a premium to book, it becomes more difficult to generate 15% year after year. For example, let's say Fairfax retains all $3B in net income so that equity is now $23B. With the same rate of return on investments and the same combined ratio, investments and float have to go up a lot - by around 15% each - to still generate a 15% ROE. How long can this be sustained if Fairfax trades consistently at a significant premium to book? Not very long, I suspect.

Viking, thanks for a great recap of the TRS investment. Given that FFH looks undervalued still, the TRS should provide value for a while. Do you know the mechanics of how the TRS works exactly? Is it settled in cash quarterly? I was wondering if it is possible that FFH may have to come up with a bunch of cash if the stock price goes down a lot for some reason. I'm sure management has considered any such possibility, so this is more curiosity than anything else. I have a vague recollection that this may have been discussed already, but a quick search of the site did not yield anything.

Berkshire is almost 25% of my portfolio and I have concluded that my foray into BABA was a mistake, so I am not fixated on screaming bargains. Just pointing out that Berkshire doesn't look particularly cheap. I won't be buying any at the current price, but I have no plans to sell any either.

After tax operating earnings are at a $40 billion annual run rate roughly. Add to that about $14 billion in normalized after-tax capital gains from the equity portfolio (6% pre-tax gains on $310 billion and a 25% tax rate). That's $54 billion of total earnings on a normalized basis. At a current market cap of $760 billion, Berkshire is trading at a P/E of 14 or so. Seems fairly cheap and a better value than the S&P 500, but the stock is no screaming bargain. I'm not adding any value for the optionality of holding a lot of cash, as cash is earning 5% right now.

@Viking, @UK and @Vinod1, thanks for your comments. They add some very useful context to how I was thinking about Evan Greenberg's comments. It makes sense that most insurers wouldn't benefit nearly as much from higher rates as Fairfax would.

I don't understand this comment by Evan Greenberg. If investment yields have gone up a lot (say by 3 or 4 percentage points), then it should be possible to raise the combined ratio by a lot more than 1 point and achieve the same risk-adjusted return, no? I realize that this depends on the size of the investment portfolio relative to earned premiums, but the math for Fairfax indicates that the CR can go up a lot and they can still earn a good risk-adjusted return. Is it different for most other insurance companies?

Morningstar's data for Fairfax does not look correct here, probably because there was no FRFHF ticker back in 1999. Here's the same chart for FFH.TO - in USD and with dividends reinvested - for the same period. The total return is quite a bit lower.

I bought my first shares over 23 years ago (Feb 2000). I remember that Warren's longevity was a concern even back then, but he has easily outperformed actuarial expectations since!

Fair enough. Thrifty3000 mentioned $10 to $50 in underwriting profits. I took the midpoint and added $30 to his base investment case to get about $170 in investment returns + underwriting profits. I think this will get whittled down to $100 after all expenses and taxes. But it may be that the investment base case and/or the underwriting profit estimate are on the conservative side.

Nice analysis. Thanks. To calculate the final return to common shareholders, however, shouldn't you deduct interest expenses, corporate overhead, income taxes and non-controlling interests? After all that, your base case looks more like $100 EPS attributable to common.

It's true that the comparison is not apples to apples, but GEICO by itself does not explain the difference. 2022 was the only bad year for GEICO with an underwriting loss of -$1.9 billion. In the four years prior, GEICO's underwriting profit totaled over $8.6 billion, or almost $2.2 billion per year.

Nice work! I checked the BRK data for the past ten years and your chart looks correct. Even accounting for differences in the lines of business, I am surprised that Fairfax's underwriting has been noticeably better than Berkshire's in recent years. I mean, Buffett is very proud of the quality of Berkshire's insurance business, but Fairfax has had better underwriting results for the last six years or so. Very impressive.

Data from FRED does not match up with this chart. Any thoughts on why there is a discrepancy? https://fred.stlouisfed.org/series/CPALTT01USQ659N

Does anyone know when Fairfax is scheduled to release Q2 results? I thought today might be the day, but there doesn't seem to be any notification regarding this. So maybe the earnings release is scheduled for Thursday of next week?

I'm curious about this sale as I have some USB in my portfolio and consider it quite undervalued still. Were you just getting your exposure down to a reasonable level or were there other reasons? My main concern with USB is that they may have to increase capital levels quite a bit over the next couple of years, but they showed with their Q2 results that they can build capital rapidly if needed.

Yes, the stock price has historically been very volatile. Perhaps it is better to look at changes in book value per share (plus dividends) to get a better idea of how the business has been doing over time.

Not sure what's going on with the yahoo charts, but they don't match up with yahoo's own historical price data. FFH.TO closed at 200.50 CAD on July 3, 2003. Thus the price return over the past 20 years is about +395% in CAD. The USD-CAD exchange rate was almost exactly the same 20 years ago as it is today, so that doesn't make a difference either. Maybe the chart has a starting date of March 2003; FFH.TO traded at below 80 CAD then according to yahoo's historical data.

What is the source for this chart? It doesn't look right to me. BRK.A closed at $56,100 on 12/31/1999 and at $519,460 today. That's a return of +826%. FFH.TO closed at $245.50 CAD on 12/31/1999 and at $992.29 CAD today. Ignoring dividends, that's a return of +304% in CAD. The return in USD is close to +350%. I don't think Fairfax has paid enough dividends to bring the total return in USD anywhere close to +826%, let alone exceed it.

The 13F-HR for Q1 is out. https://www.sec.gov/Archives/edgar/data/1067983/000095012323005270/0000950123-23-005270-index.htm Here are some changes that I noticed - added to OXY; reduced ATVI, CVX and GM; sold out of USB and BK but retained position in Citi.

I guess he was saying that online banking is much more common now than even a few years back. Here's a relevant excerpt from the recently published Financial Stability Report from the Fed. You can find the complete report at https://www.federalreserve.gov/publications/financial-stability-report.htm. "The runs on SVB and Signature Bank were of unprecedented speed compared with previous runs. During the run on Washington Mutual in 2008—to date, the run that caused the largest failure of an insured depository institution by inflation-adjusted total assets—depositors withdrew about $17 billion over the course of eight business days, with the largest deposit withdrawal in one day reaching just over 2 percent of pre-run deposits. By comparison, the highest one-day withdrawal rate was more than 20 percent in the case of SVB and Signature Bank, at the time the second- and third-largest depository institutions by inflation-adjusted total assets, respectively, that failed due to a bank run (figure B). At SVB, withdrawals would have been even larger had regulators not closed the bank on the morning of March 10."

Yeah, USB is taking it on the chin. One bit of good news is that two directors have bought shares recently. Scott Wine bought 30K at $32.85 per share; Alan Colberg bought 10K at $34.14 apiece. https://www.marketbeat.com/stocks/NYSE/USB/insider-trades/

Bought some more USB

I guess the evidence is that PGR had a combined ratio of 96% for 2022, while GEICO had a combined ratio of 104.8%. Since auto insurance is mostly a short-tail business, the combined ratio is probably a good indication of underwriting quality. PGR's combined ratio did rise to 99% in Q1 2023 though.

I agree with you that the energy transition will be a slow affair and that it makes sense to be bullish on oil. I am not so sure about metal prices going higher than people can imagine; maybe lithium, but I have a hard time seeing any other metal being in extreme undersupply. I think Mark Mills overstates the case for minerals. Here's the Economist on Cobalt (might be behind paywall): https://www.economist.com/finance-and-economics/2023/02/16/cobalt-a-crucial-battery-material-is-suddenly-superabundant Even the case for extreme shortages in lithium depends on sodium ion batteries taking a long time to ramp up, I feel. CATL is planning to start mass production of sodium ion this year, so this is not a slam dunk. https://www.nextbigfuture.com/2022/10/catl-will-mass-produce-sodium-ion-batteries-in-2023.html

I was down just over 11% in 2022. My two biggest positions - Berkshire and Fairfax - did well, but not well enough to offset the carnage in the rest of my holdings. US banks, Chinese stocks like Baba and Tencent, and GOOG all dragged my performance down.