Fairfax 2024

[Xerxes]

4 hours ago, John Hjorth said:

 

I'm kicking a small footnote in here, that there actually exist other insurance companies with similar business model, but they may be few. I am only aware of one though : Protector Forsikring ASA, Oslo, Norway [ PROTCT.OS][ CoBF topic ].

 

- - - o 0 o - - -

 

Now back to Fairfax 2024.

And to a much lesser degree, Manulife, the Canadian life insurance giant, and its investment in timberland. 

[Xerxes]

15 hours ago, 73 Reds said:

This has been one element of Berkshire's secret sauce as well; the question is if it works so well for companies like Berkshire, Fairfax and Markel, why don't other insurance companies adopt a similar model?

I think the right way to answer this comment is to refer back to the Bloomstran comment about Berkshire vs Swiss RE and Munich RE where he states the latter two “never known a business for which they didnt want to write insurance for it”

 

said differently, the traditional players chose to maximize their revenue for today (risking mispricing) by writing as much business as they can. Whereas, the “quasi-families” that run Berk, FFH and Markel, have no problem in hitting the brakes, which leaves more surplus, to invest in common equity and more recently in FFH case to return capital to owners. 
 

So the secret sauce (which is not so secret), as Viking pointed out requires a “long term” approach, that is hard to replicate with EPS driven companies at the mercy of Wall Street on a quarterly basis. 

[Hamburg Investor]

17 hours ago, Viking said:

 

@73 Reds , to better understand some of the variables at play, let's look a Travelers 2023 numbers and compare them to Fairfax.

 

When looking at Travelers 2023 numbers two things jump out:

• Underwriting income represents 57% of their income streams (like most P/C insurers, they only have 2 income streams).
• The yield (pre-tax) on their investment portfolio is about 3.3% (using the YE value of investment portfolio). 

Travelers is earning peanuts on its $88.5 billion investment portfolio - about 3.3%. It is expected to increase a small amount in 2024 ($200 million, which will bump the average yield to about 3.5%). This is because they are investing solely in fixed income. And it looks to me like they match the duration of their fixed income portfolio with their insurance liabilities.

 

Now compare Travelers to Fairfax. Fairfax is earning about 7% on its investment portfolio - double what Travelers is earning (see the chart at the bottom of the post). That is a massive gap. 

 

Fairfax is earning much, much more on their investment portfolio for a couple of reasons:

1.) They do not restrict their investments to bonds/fixed income.

2.) They are an active manager - they look to exploit dislocations/volatility (wherever it shows up).

3.) They are very good at what they do.

 

Comparing Fairfax and Travelers you really get some good insight into the power and value of the business model Fairfax that is successfully executing today. 

 

 

 

Below is a summary of investment returns for Fairfax. The returns have been smoothed over 2 year intervals to smooth out the annual volatility and make it easier to understand.

 

What companies do you see on a worldwode scale that do mimic Buffetts approach? I see BRK, MKL, FFH and Protektor Forsikring, a small Norwegian company

 

There are other good companies, but as far as I can see they miss one of the points:

- e.g. RLI is a very good underwriter; but they fail the value / active / stock / whole company approach. 

- Some do invest a tiny bit into stocks, but its so less, that it doesn’t move the needle.

Anyone with further suggestions?

[Luke]

1 minute ago, Hamburg Investor said:

What companies do you see on a worldwode scale that do mimic Buffetts approach? I see BRK, MKL, FFH and Protektor Forsikring, a small Norwegian company

 

There are other good companies, but as far as I can see they miss one of the points:

- e.g. RLI is a very good underwriter; but they fail the value / active / stock / whole company approach. 

- Some do invest a tiny bit into stocks, but its so less, that it doesn’t move the needle.

Anyone with further suggestions?

Does Protektor own a stock portfolio where they invest the float? Never heard of them, sounds interesting...

[MMM20]

2 hours ago, Luke said:

Does Protektor own a stock portfolio where they invest the float? Never heard of them, sounds interesting...

Yes. I passed around $2-3 a few years ago and it’s looking like a pretty bad error of omission. There’s a good amount of research floating around out there. I think I heard of it from Dave Kim at Scuttleblurb.

 

Edited August 23, 2024 by MMM20

[SafetyinNumbers]

3 hours ago, Hamburg Investor said:

What companies do you see on a worldwode scale that do mimic Buffetts approach? I see BRK, MKL, FFH and Protektor Forsikring, a small Norwegian company

 

There are other good companies, but as far as I can see they miss one of the points:

- e.g. RLI is a very good underwriter; but they fail the value / active / stock / whole company approach. 

- Some do invest a tiny bit into stocks, but its so less, that it doesn’t move the needle.

Anyone with further suggestions?

E-L Financial is 20% insurance (Empire Life) and the rest is index/global quality portfolio. 12%+ for 50+ years.

[petec]

On 8/23/2024 at 4:14 AM, SafetyinNumbers said:

If FFH can dividend out $4b from the insurance subsidiaries to the holdco that means they can buy a lot more equities at the subsidiaries if the opportunity presented itself. That could supercharge returns. They seem almost certain to buy in their insurance minority interests which have a 20% return based on trailing results. The cost as of December was $2.5b so that might add ~$20/sh to EPS.

 

I should know this, but is the $4bn number the current combined dividend capacity of the subs or is it your own estimate?

 

And what's the $2.5bn cover? Allied, Brit, Odyssey?

 

It's a pity we don't know enough about the terms of each of these deals to assess whether buying in the subs is better than buying back shares.

[SafetyinNumbers]

5 hours ago, petec said:

 

1. I should know this, but is the $4bn number the current combined dividend capacity of the subs or is it your own estimate?

 

2. And what's the $2.5bn cover? Allied, Brit, Odyssey?

 

3. It's a pity we don't know enough about the terms of each of these deals to assess whether buying in the subs is better than buying back shares.

1. I think it was $3b at the end of 2023 for the insurance subsidiaries they could dividend up. I might be too high thinking that’s up from year end. They have been slowing premium growth and sending dividends up for buybacks but I don’t think dividend capacity has diminished.

 

2. I’m just using the non-controlling interests note from last year which showed ~$2.5b as the balance for Allied, Brit and Odyssey. It’s probably higher now. 
 

3. The table shows ~$500m in earnings for 2023, so that gives us some idea. It seems like the P/B multiple FFH buys back the minority interest is the same multiple they sold at so maybe the returns are highest for Allied World, then Brit and finally Odyssey. I like buying the subs in vs buying back the stock because it increases durability. This is something, I think analysts could model in over the next few years to show earnings growth but they remain committed to conservatism as opposed to making the best guess.

 

 

[petec]

On 8/24/2024 at 12:29 PM, SafetyinNumbers said:

I like buying the subs in vs buying back the stock because it increases durability.

 

Why does it increase durability? I would have thought it was a straight financial decision (between buybacks vs buying back minorities) based on price and expected returns.

[SafetyinNumbers]

2 hours ago, petec said:

 

Why does it increase durability? I would have thought it was a straight financial decision (between buybacks vs buying back minorities) based on price and expected returns.

Because they can always sell a stake again if they need capital for any other opportunistic reason but perhaps more importantly a defensive reason. It also removes the minority interests which boosts the total nominal earnings.

[nwoodman]

The sub positions are an interesting thought exercise.  My take is that it creates some clever optionality and a sign that Fairfax is not short on ideas.  It creates a reasonably clear pathway to decent returns if they are light on for ideas in the future or there is a significant changing of the guard.  I guess the other side of the argument is the tipping point of running a highly leveraged book.  

 

I will never become complacent about their debt levels but I do give them the benefit of the doubt in terms of how they manage the risk and wherever possible make it non-recourse and bump it into “equity”.

Edited August 26, 2024 by nwoodman

[UK]

3 hours ago, SafetyinNumbers said:

Because they can always sell a stake again if they need capital for any other opportunistic reason but perhaps more importantly a defensive reason.

 

2 hours ago, nwoodman said:

I will never become complacent about their debt levels but I do give them the benefit of the doubt in terms of how they manage the risk and wherever possible make it non-recourse and bump it into “equity”.

 

These are really good observations. Thanks!

[SafetyinNumbers]

2 hours ago, nwoodman said:

 

I will never become complacent about their debt levels but I do give them the benefit of the doubt in terms of how they manage the risk and wherever possible make it non-recourse and bump it into “equity”.

I think keeping the excess liquidity at the insurance subsidiaries makes a lot more sense then sending it up to the holdco for a variety of reasons but principally the debt levels are less concerning  while that is true.

[Xerxes]

Is there a tax advantage/implication between say Allied or Odessey using its own surplus to buy back the minorities vs. dividend to the mothership, and it writing a check for the minority ?

 

or is that not relevant because these are corporate tax

 

 

[dartmonkey]

On 8/22/2024 at 4:55 PM, Viking said:

If we continue to get a few large asset sales in the coming years (likely, given what we have seen the past 10 years) then I think 7% is a reasonable baseline estimate to use looking out the next 3 to 5 years. and yes, the results will be volatile from year to year.

 

Makes sense to me.

 

This return would be on investment assets that are currently at $66b, whereas equity is $23b, so results are levered about 3 times. So I think it is fair to say that Fairfax is currently on track to earn 21% on equity, at least in the next few years. Thoughts?

[SafetyinNumbers]

49 minutes ago, dartmonkey said:

 

Makes sense to me.

 

This return would be on investment assets that are currently at $66b, whereas equity is $23b, so results are levered about 3 times. So I think it is fair to say that Fairfax is currently on track to earn 21% on equity, at least in the next few years. Thoughts?

The way I like to think about it is that the odds of earning north of 15% for the next three years could be as high as 90%. Also, the odds of earning 20%+ are much higher than earning 12% or less.
 

That makes buying at 1.2x BV seem like an incredible bargain but I appreciate many on this board think it’s fairly valued currently.

[Hamburg Investor]

41 minutes ago, SafetyinNumbers said:

The way I like to think about it is that the odds of earning north of 15% for the next three years could be as high as 90%. Also, the odds of earning 20%+ are much higher than earning 12% or less.
 

That makes buying at 1.2x BV seem like an incredible bargain but I appreciate many on this board think it’s fairly valued currently.

 

I like the way you think. Mine is a bit similar:

The rule of 72 tells us, that at 15% ROE equity doubles in around 5 years, at 18% it doubles in 4 and at 24% it needs 3 years. Of course ROE can be less or more in the upcoming 3 years (and thereafter hard to say), but I wouldn't bet on less than 15% or more than 24% for the next 3 years. My best guess is around 20%. So that's around a double in 3 1/2 years, isn't it?

Than I try to find a comparison of FFH to the market. It doesn't make a lot of sense to me to compare the book value of the market (think: S&P500) against book of FFH, as most companies of the index are better understood with pe. So how to come up with a normalized pe for FFH? I just pick a normalized ROE of FFH (my best guess is 15+% over the long run, so I take 15%). If FFH earns 15% on book and I can buy it at less than 1.2 book, than that's a normalized pe ratio of 8, compared to nearly 30 for the S&P500. Wow, that's cheap and I am pretty sure the average S&P500 company won't make 15%, not even 12%. That's my definition of Quality at a cheap price and of GARP.

Thinking one step further: Assuming a double of FFHs equity in 3 1/2 years and the price stays where it is, than FFH would be valued below a pe ratio of 4 (!). I would love that, as the buybacks would bring returns even higher. Assuming 15% after the next 3 1/2 years, would bring pe ratio down to 2 after 8 1/2 years. 

[dartmonkey]

 

22 minutes ago, Hamburg Investor said:

If FFH earns 15% on book and I can buy it at less than 1.2 book, than that's a normalized pe ratio of 8,

Just to show how this works, P/E=P/B*B/E = P/B÷E/B = 1.2÷15% = 8...

 

25 minutes ago, Hamburg Investor said:

Thinking one step further: Assuming a double of FFHs equity in 3 1/2 years and the price stays where it is, than FFH would be valued below a pe ratio of 4 (!). I would love that, as the buybacks would bring returns even higher. Assuming 15% after the next 3 1/2 years, would bring pe ratio down to 2 after 8 1/2 years. 

 

Realistically, if Fairfax continues to do well for 3 more years, the price is not going to be the same. But you got me wondering, what might a realistically repurchase scenario look like?

 

 

So I took current book of $23b, 22.48m shares outstanding at the end of Q2, share price of $1180, and $4b a year in earnings that management says we have visibility for for the next 3 years. Then I assume that half of those earnings ($2b/year) will be used for repurchases ($2b/year) and half will be retained (and invested). But since some of the earnings will be retained, book will increase over time, and there will be earnings on that extra book. So I assume that the current $23b in book will keep making $4b/year (a 17% ROE), but whatever is added to equity over the next 3 years will get a lower return (as interest rates fall, etc.), which I have very pessimistically put at 10%. 

 

Even with P/B staying at 1.15, that means share prices would almost double. Book would only go from $23b to $29.5b, but because of the 6.4m shares repurchased, book per share would almost double, as would the share price at a constant multiple of book. P/E would just come down a bit, from 7 to 6.

 

If the P/B multiple keeps rising, say to 1.3 in 3 years, there would be a little less repurchasing, but on the other hand, the share price would be a bit higher, just over a double. 

 

, 

 

[Munger_Disciple]

41 minutes ago, dartmonkey said:

 

Just to show how this works, P/E=P/B*B/E = P/B÷E/B = 1.2÷15% = 8...

 

 

Realistically, if Fairfax continues to do well for 3 more years, the price is not going to be the same. But you got me wondering, what might a realistically repurchase scenario look like?

 

 

So I took current book of $23b, 22.48m shares outstanding at the end of Q2, share price of $1180, and $4b a year in earnings that management says we have visibility for for the next 3 years. Then I assume that half of those earnings ($2b/year) will be used for repurchases ($2b/year) and half will be retained (and invested). But since some of the earnings will be retained, book will increase over time, and there will be earnings on that extra book. So I assume that the current $23b in book will keep making $4b/year (a 17% ROE), but whatever is added to equity over the next 3 years will get a lower return (as interest rates fall, etc.), which I have very pessimistically put at 10%. 

 

Even with P/B staying at 1.15, that means share prices would almost double. Book would only go from $23b to $29.5b, but because of the 6.4m shares repurchased, book per share would almost double, as would the share price at a constant multiple of book. P/E would just come down a bit, from 7 to 6.

 

If the P/B multiple keeps rising, say to 1.3 in 3 years, there would be a little less repurchasing, but on the other hand, the share price would be a bit higher, just over a double. 

 

, 

 

 

Doesn't one need to take into account the possibility of super CAT events and their effect on the "normalized ROE" and P/B?

Edited August 26, 2024 by Munger_Disciple

[dartmonkey]

5 minutes ago, Munger_Disciple said:

Doesn't one need to take into account the possibility of super CAT events and their effect on the "normalized ROE"?

Watsa's projected $4b in operating earnings in the next 3 years is based on:  "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."

 

Net premiums written in 2023 were $28.9b, so $1.25b is about a CR of 96. Last year the CR was 93%; in 2022 it was 96%; 2021 it was 88%. It is true that some day, there will be a super CAT that will knock a hole in one year's earnings, but perhaps a 96% average CR is realistic. In the catastrophic 2001, the total combined ratio was 121%, although only half of that was from megacats;  the CR would have been 110%, just from underreserving. What average CR do you think we should use?

[Munger_Disciple]

19 minutes ago, dartmonkey said:

Watsa's projected $4b in operating earnings in the next 3 years is based on:  "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."

 

Net premiums written in 2023 were $28.9b, so $1.25b is about a CR of 96. Last year the CR was 93%; in 2022 it was 96%; 2021 it was 88%. It is true that some day, there will be a super CAT that will knock a hole in one year's earnings, but perhaps a 96% average CR is realistic. In the catastrophic 2001, the total combined ratio was 121%, although only half of that was from megacats;  the CR would have been 110%, just from underreserving. What average CR do you think we should use?

 

Buffett said in 2017 that the probability of a $400B (USD) industry loss event in any given year (in his estimate) is 2%. Adjusting for inflation and increase in the cost of construction, that would be a $600B event today. I assume that that events that have lower loss may occur at greater probability than 2%. 

At Berkshire, they underwrite for a long term CR of 100% including CAT and super CAT years. I have no idea what CR is achievable for Fairfax over the next 3 years (which depends on whether there is a CAT/super CAT event in the next three years or not and its magnitude) but I would think that a 10-15 year average CR of 100 taking into account a couple of really bad CAT years is prudent. 

Edited August 26, 2024 by Munger_Disciple

[SafetyinNumbers]

2 hours ago, Munger_Disciple said:

 

Buffett said in 2017 that the probability of a $400B (USD) industry loss event in any given year (in his estimate) is 2%. Adjusting for inflation and increase in the cost of construction, that would be a $600B event today. I assume that that events that have lower loss may occur at greater probability than 2%. 

At Berkshire, they underwrite for a long term CR of 100% including CAT and super CAT years. I have no idea what CR is achievable for Fairfax over the next 3 years (which depends on whether there is a CAT/super CAT event in the next three years or not and its magnitude) but I would think that a 10-15 year average CR of 100 taking into account a couple of really bad CAT years is prudent. 

If we had to place odds (and in a way we are) what probability should we assign to a 93, 95, 100 or 105CR for the next 15 years. The fairest way is to probably measure in aggregate over a 15 year period and not an average because if there is a SuperCat in that time frame, there will probably a big increase in premiums for the years that follow.

Edited August 26, 2024 by SafetyinNumbers

[Munger_Disciple]

4 minutes ago, SafetyinNumbers said:

The fairest way is to probably measure in aggregate over a 15 year period and not an average because if there is a SuperCat in that time frame, there will probably a big increase in premiums for the years that follow.

 

Yeah aggregation makes sense. Most important is the ability of Fairfax to survive a super CAT event like a $600B loss event, which allows them to prosper in the hard market to follow the event.

Let us say in the worst case, Fairfax suffers 1.25% of 600B = 7.5B loss event pre-tax. They can survive that reasonably well but it would wipe out a couple of years worth of earnings.  

Edited August 26, 2024 by Munger_Disciple

[SafetyinNumbers]

11 minutes ago, Munger_Disciple said:

 

Yeah aggregation makes sense. Most important is the ability of Fairfax to survive a super CAT event like a $600B loss event, which allows them to prosper in the hard market to follow the event.

Let us say in the worst case, Fairfax suffers 1.25% of 600B = 7.5B loss event pre-tax. They can survive that reasonably well but it would wipe out a couple of years worth of earnings.  

How many years would it take to recover those earnings do you think? 

[Munger_Disciple]

23 minutes ago, SafetyinNumbers said:

How many years would it take to recover those earnings do you think? 

 

 I already provided an estimate in the earlier post based on the next three years earning visibility but after that I don't know.