Jump to content

Recommended Posts

Posted (edited)

As I work my way through Prem's letter in the 2025AR I will be positing some random thoughts. 

 

Decentralization. Is this important to an investor in Fairfax? Yes. How important? Very. 

 

Fairfax is walking the talk when it comes to decentralization - this is one of the core building blocks of its organizational structure, competitive advantages and moat.

 

The number of profit centres in its insurance business continues to increase every year:

  • 2025: 275
  • 2024: 250
  • 2023: 225
  • 2022: 200

We see a similar dynamic at play in its investment management business. In 2025, Quess was split into three companies: Quess, Digitide and Bluspring. Additionally, the Keg was spun out of Recipe (both are still private holdings).    

 

Benefits: transparency, accountability and empowerment, resulting in strong employee retention and better business results. This isn't a sexy topic. But it very important for long term shareholders.

 

2025AR: “We have over 275 profit centres across our group. Each profit centre is focused on a unique set of customers, geographies or products that benefit from market leadership, product knowledge and the ability to provide excellent customer service. These profit centres facilitate transparency, enabling Andy Barnard, Brian Young and Peter Clarke to effectively monitor the insurance operations. Empowerment thrives at Fairfax. We are always working on making our companies more indispensable to our customers.”

 

2024AR: “We have over 250 profit centres across our group.”

 

2023AR: “With the addition of added business from GIG, the $32 billion in consolidated gross premiums written is generated through our more than 225 profit centres across the group.”

 

2022AR: “The $27.6 billion of our consolidated gross premiums is generated through approximately 200 profit centres across the group.”

Edited by Viking
Posted (edited)
16 minutes ago, Viking said:

As I work my way through Prem's letter in the AR I will be positing some random thoughts. 

 

Decentralization. Is this important to an investor in Fairfax? Yes. How important? Very. 

 

Fairfax is walking the talk when it comes to decentralization - this is one of the core building blocks of its organizational structure, competitive advantages and moat.

 

The number of profit centres in its insurance business continues to increase every year:

  • 2025: 275
  • 2024: 250
  • 2023: 225
  • 2022: 200

We see a similar dynamic at play in its investment management business. In 2025, Quess was split into three companies: Quess, Digitide and Bluspring. Additionally, the Keg was spun out of Recipe (both are still private holdings).    

 

Benefits: transparency, accountability and empowerment, resulting in strong employee retention and better business results. This isn't a sexy topic. But it very important for long term investors.

 

2025AR: “We have over 275 profit centres across our group. Each profit centre is focused on a unique set of customers, geographies or products that benefit from market leadership, product knowledge and the ability to provide excellent customer service. These profit centres facilitate transparency, enabling Andy Barnard, Brian Young and Peter Clarke to effectively monitor the insurance operations. Empowerment thrives at Fairfax. We are always working on making our companies more indispensable to our customers.”

 

2024AR: “We have over 250 profit centres across our group.”

 

2023AR: “With the addition of added business from GIG, the $32 billion in consolidated gross premiums written is generated through our more than 225 profit centres across the group.”

 

2022AR: “The $27.6 billion of our consolidated gross premiums is generated through approximately 200 profit centres across the group.”


thanks @Viking for bringing attention to this as I had not realized how much the number of insurance profit centres have grown over the past 4 years.  It would be interesting to see this broken down by sector and region. 
 

Edited by Hoodlum
Posted (edited)

We know Fairfax's book value is grossly understating its economic/intrinsic value. 

 

In the 2025AR, Prem was unusually blunt in highlighting this growing discrepancy: "In the past, we used book value per share as a first measure of intrinsic value. Those days are long gone! Our intrinsic value is way above our book value now!"

 

I don't think Fairfax has ever been this clear in its communication on this topic. Investors need to pay attention. 

 

In his letter, Prem provided a couple of good examples of where the discrepancy is growing the most:

  • Excess of FV over CV for the non-insurance associate and consolidated holdings, which was $3.1B ($151 per share) at December 31, 2025. This is a very transparent example.
  • In the AR, Prem hints at another example: the MV "could be" understated for many of the private companies in the consolidated non-insurance companies bucket. The fact that Prem felt it necessary to point this out suggests to me that this is likely a material amount.

And the disconnect is not limited to the equity holdings. It also involves the insurance subsidiaries. 

  • Prem points out that Allied World earned $1.2B in 2025 and had a 2025YE book value of $7.7B.
  • Prem is uncharacteristically blunt in his conclusion: "You don’t need to be a genius to see that Allied World is worth much more today than its book value of $7.7 billion.

This is not an example of Prem being promotional. He is simply stating the obvious. I would describe his communication as being more incredulous than anything. 

 

Bottom line, what we are seeing play in recent years is Fairfax's intrinsic value per share is increasingly becoming unmoored from its accounting book value per share. Investors (and analysts) who decide to only use BVPS will be analyzing Fairfax with one arm tied behind their back - good luck with that approach. 

 

------------

 

Screenshot2026-03-08at12_24_04PM.thumb.png.26d4594ee439aa328fa2e91c2dee8f77.png

 

-----------

 

image.thumb.png.5ad48fc5b9b359870888b510d222af95.png

Edited by Viking
Posted
1 minute ago, Haryana said:

I was just going to point that out but now the x-axis is too short on the 2015 - 2020 range.


haha!! True! Well there is the bull case for Adobe and other SAAS companies, Can't win with AI sometimes! 

Posted (edited)

On page 34 of his 2025 letter to shareholders Prem writes

Quote

Despite its unique and attractive attributes, Orla continues to trade at an attractive ~8x multiple of free cash flow. In December of 2025, Fairfax sold 25 million shares of Orla Mining, approximately a quarter of our position if you include our convertibles and warrants as part of a rebalancing of our investment portfolios.

 

Does anyone know how Prem conceptionalizes rebalancing? I don't understand why he would sell a position that's relatively small and still trading at an attractive valuation.

Edited by Wanderer
Posted

Does anyone have an idea how Prem gets to $150 per share in the 2025 shareholder letter please? Applying his figures I get:

 

Underwriting Profit - $1,500,000,000

Interest and Dividend income - $2,500,000,000

Share of profits of Associates + Non-insurance subsidiaries - $1,000,000,000

Total Operating Income - $5,000,000,000

FY2025 interest expense - $(822,000,000)

FY2025 corporate overheads - $(480,500,000) per page 166 of FY2025 AR

Profit before tax - $3,698,500,000

Taxes @ 22% - $(813,670,000)

Profit after tax - $2,884,830,000

less: Series K Preference Share dividend - $(8,844,000) - Notional CAD$237,500,000 @ 5.05%, USD/CAD rate ~ 1.356

less: Minority Interest (assume 10%) - $(288,483,000)

Profit to FFH shareholders - $2,587,503,000

Number of shares outstanding @ 31 Dec 2025 - 20,856,086

 

Earnings per share - $124.06

Difference - $(540,909,900)

Posted (edited)
4 hours ago, Wanderer said:

On page 34 of his 2025 letter to shareholders Prem writes

 

Does anyone know how Prem conceptionalizes rebalancing? I don't understand why he would sell a position that's relatively small and still trading at an attractive valuation.


I think it matters what the investment is. Orla is a gold producer. Commodity stocks are not buy and hold type investments. You have to sell them when they go up. 
 

Fairfax’s average cost on the common shares was C$4.80/share. They sold 25 million shares at $17.64. This was a 268% gain in ~2.5 years. Outstanding. 
 

They still have exposure to 76 million shares. And now that Foran is being taken out by El Dorado, they will have even more exposure to gold (when that deal closes). 
 

Fairfax sold RFP at the top of the lumber cycle. Stelco was sold at a crazy high price. Fairfax sold 25% of Orla and locked in a massive return. They have been making obscene money with their resource investments. I hope they keep monetizing some of them - locking in big gains. And redeploying the proceeds into other good opportunities - which is what they have been doing. 
 

Also, importantly, Orla has quickly become Fairfax’s 4th largest equity holding - it has become a very large position. Especially for a commodity play. Even after selling 25% of the position it is still the 4th largest holding (at about US$1.4B).

Edited by Viking
  • Like 1
Posted
2 hours ago, zhuanquan said:

Does anyone have an idea how Prem gets to $150 per share in the 2025 shareholder letter please? Applying his figures I get:

 

Underwriting Profit - $1,500,000,000

Interest and Dividend income - $2,500,000,000

Share of profits of Associates + Non-insurance subsidiaries - $1,000,000,000

Total Operating Income - $5,000,000,000

FY2025 interest expense - $(822,000,000)

FY2025 corporate overheads - $(480,500,000) per page 166 of FY2025 AR

Profit before tax - $3,698,500,000

Taxes @ 22% - $(813,670,000)

Profit after tax - $2,884,830,000

less: Series K Preference Share dividend - $(8,844,000) - Notional CAD$237,500,000 @ 5.05%, USD/CAD rate ~ 1.356

less: Minority Interest (assume 10%) - $(288,483,000)

Profit to FFH shareholders - $2,587,503,000

Number of shares outstanding @ 31 Dec 2025 - 20,856,086

 

Earnings per share - $124.06

Difference - $(540,909,900)

 

You are missing about $500M or so by double-counting subsidiary corporate overhead and interest expense on non-insurance businesses.  Prem is relaying that the holding company earns about $150 per share excluding investment gains.  Earnings/dividends to the holding company are after those expenses are already deducted at the subsidiary level.  So holding company expenses is below $200M on an annual basis, not close to $500M...$300M there.  Interest expense and interest expense on lease liabilities for non-insurance businesses is around $200M.  Total around $500M.  Cheers!

 

Posted
9 hours ago, Parsad said:

 

You are missing about $500M or so by double-counting subsidiary corporate overhead and interest expense on non-insurance businesses.  Prem is relaying that the holding company earns about $150 per share excluding investment gains.  Earnings/dividends to the holding company are after those expenses are already deducted at the subsidiary level.  So holding company expenses is below $200M on an annual basis, not close to $500M...$300M there.  Interest expense and interest expense on lease liabilities for non-insurance businesses is around $200M.  Total around $500M.  Cheers!

 

I am having the same problem as @zhuanquan

Prem said “…about $150 per share after taxes, interest expense, corporate overhead, and other costs.”

 

$6440 pre-tax income minus $3151 of investment gains minus the 2025 tax rate of 18% is $129 per share, not $150.  (I am pretty sure that tax rate is different between operating income and investment income, but I don’t have a better number and it doesn’t matter that much anyways.)

 

@Parsad I don’t understand how you are saying that we are double counting corporate overhead and interest expenses. Perhaps I can understand not counting corporate overhead dedicated to investment functions (don’t think that is what Prem is saying, but maybe).  I do not understand how interest expenses on non-insurance companies is double counting.

IMG_1092.jpeg

Posted
2 hours ago, sholland said:

I am having the same problem as @zhuanquan

Prem said “…about $150 per share after taxes, interest expense, corporate overhead, and other costs.”

 

$6440 pre-tax income minus $3151 of investment gains minus the 2025 tax rate of 18% is $129 per share, not $150.  (I am pretty sure that tax rate is different between operating income and investment income, but I don’t have a better number and it doesn’t matter that much anyways.)

 

@Parsad I don’t understand how you are saying that we are double counting corporate overhead and interest expenses. Perhaps I can understand not counting corporate overhead dedicated to investment functions (don’t think that is what Prem is saying, but maybe).  I do not understand how interest expenses on non-insurance companies is double counting.

IMG_1092.jpeg

 

Maybe work back from what he said in 2023 and 2025.  There is a difference of about $1B from underwriting profit of increases of $250M, interest and dividend income increases of $500M, and income from associates and non-insurance consolidated income of $250M.  In 2023, the recurring income before gains was $125 per share...today it is $150 per share. 

 

The text from 2023 Letter states:

 

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 $750million, 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!

 

The text from 2025 Letter states:

 

We can see sustaining our consolidated operating income for the next four years at $5 billion (again, no guarantees) consisting of underwriting profit of $1.5 billion or more, interest and dividends of $2.5 billion and income from associates and non-insurance consolidated income of $1 billion. That would represent about $150 per share after taxes, interest expense, corporate overhead and other costs. These figures are all, of course, before fluctuations in realized and unrealized gains and losses in stocks and bonds!  Lately, our earnings are less lumpy than they used to be, but investment gains and losses remain lumpy by nature.

 

Cheers!

Posted
3 hours ago, sholland said:

I am having the same problem as @zhuanquan

Prem said “…about $150 per share after taxes, interest expense, corporate overhead, and other costs.”

 

$6440 pre-tax income minus $3151 of investment gains minus the 2025 tax rate of 18% is $129 per share, not $150.  (I am pretty sure that tax rate is different between operating income and investment income, but I don’t have a better number and it doesn’t matter that much anyways.)

 

@Parsad I don’t understand how you are saying that we are double counting corporate overhead and interest expenses. Perhaps I can understand not counting corporate overhead dedicated to investment functions (don’t think that is what Prem is saying, but maybe).  I do not understand how interest expenses on non-insurance companies is double counting.

IMG_1092.jpeg

 

I think he’s leaving out all the IFRS 17 stuff. 

Posted
16 hours ago, zhuanquan said:

Does anyone have an idea how Prem gets to $150 per share in the 2025 shareholder letter please? Applying his figures I get:

 

 

If you are using the lens of adjusted operating income ($1.5bn + $2.5bn + $1bn = $5bn) then the central costs to use is $157m not $480m (see below). This is because (I think) the delta of central costs within operating companies ($323m) is already captured within the $5bn operating profit calculation.

So if you add this $16/share ($323/20.7) to your calculation, then you get to $140+.

 

Below from Pg 28 of shareholder letter:

image.thumb.png.1c49a613e67a43664232698ca4ce2f52.png

And the reconciliation to @sholland numbers, is yes excluding the IFRS17 impact.

Posted

Looking at sedi filings it looks like Fairfax bought ~167K shares over Jan-Feb-26 and then looking at AR below, looks like they may have bought another ~59K in the first week of March.

 

'Subsequent to December 31, 2025, the company purchased for cancellation 226,694 subordinate voting shares under the terms of its normal course issuer bids at a cost of $384.0' AR2025

Posted
4 minutes ago, glider3834 said:

Looking at sedi filings it looks like Fairfax bought ~167K shares over Jan-Feb-26 and then looking at AR below, looks like they may have bought another ~59K in the first week of March.

 

'Subsequent to December 31, 2025, the company purchased for cancellation 226,694 subordinate voting shares under the terms of its normal course issuer bids at a cost of $384.0' AR2025


Great catch. They might have done a block. Will be interesting to see when they file March buybacks in a month. 

Posted
13 minutes ago, SafetyinNumbers said:


Great catch. They might have done a block. Will be interesting to see when they file March buybacks in a month. 


220k shares were traded today including a 50k block on the CBOE exchange where Fairfax as done large blocks previously, although I can’t see who was the buyer.  I suspect Fairfax was very active today.  They will likely buyback well over 300k shares before end of Q1, possibly a lot more if the share price stays near this level. 

Posted (edited)
On 3/8/2026 at 1:48 AM, nwoodman said:

Thanks for this, burnt a few tokens with the usual vendors to generate some background notes. Pros and cons but this is where Prem and team shines, the look thru.  

Erkan Appointment Analysis.pdf 213.21 kB · 89 downloads

 

A quick follow-up. It has been reported overnight in multiple Turkish mainstream outlets, including Sözcü, Cumhuriyet, Yeni Akit and Medyascope, that Tom Barrack, Trump’s US Ambassador to Turkey, Special Envoy for Syria, Special Envoy for Iraq and one of the most geopolitically connected individuals in the current US administration, served as the personal reference behind Hafize Gaye Erkan’s appointment as President of Fairfax Banking & FinTech.

 

The reporting traces a relationship that began at First Republic Bank, where Barrack is said to have championed Erkan’s advancement against internal board resistance, continued with a personal visit to her in Ankara during her tenure at the central bank, and culminated in the Fairfax introduction. On first read, that looks like a personnel footnote. It may be considerably more than that.

 

The attached note explains why the Barrack thread may materially change the analytical frame on the Erkan appointment and what it suggests about what Fairfax is actually building. These things are easy to over-read, but ignoring them risks missing the moment entirely. Either way, it is fascinating to watch unfold and worth a few tokens to pull together a narrative that can be weighted even at 0.1%.

 

Edit:

 

I am sure we all love an analogy even if its on the hyperbolic side:
 

“By hiring Hafize Gaye Erkan with the reported sponsorship of Tom Barrack, the firm has effectively installed a “geopolitical router” inside its corporate headquarters. This router connects Fairfax’s capital to:
1. The US Administration’s Inner Circle: Direct, relational access to the architects of US policy in Turkey and the Middle East.
2. Sovereign Wealth Networks: A four-decade link to the capital of the Gulf, where Fairfax already has a significant insurance footprint.
3. Regulated Banking Transitions: A peer-level relationship with sovereign banking regulators, and the institutional credibility that comes with it, at the precise moment Fairfax is attempting to close an $8 billion privatisation with the Government of India and the RBI.”

 

 

Erkan - Barrack Supplementary March 2026.pdf

Edited by nwoodman
Posted (edited)
7 hours ago, Parsad said:

 

Maybe work back from what he said in 2023 and 2025.  There is a difference of about $1B from underwriting profit of increases of $250M, interest and dividend income increases of $500M, and income from associates and non-insurance consolidated income of $250M.  In 2023, the recurring income before gains was $125 per share...today it is $150 per share. 

 

The text from 2023 Letter states:

 

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 $750million, 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!

 

The text from 2025 Letter states:

 

We can see sustaining our consolidated operating income for the next four years at $5 billion (again, no guarantees) consisting of underwriting profit of $1.5 billion or more, interest and dividends of $2.5 billion and income from associates and non-insurance consolidated income of $1 billion. That would represent about $150 per share after taxes, interest expense, corporate overhead and other costs. These figures are all, of course, before fluctuations in realized and unrealized gains and losses in stocks and bonds!  Lately, our earnings are less lumpy than they used to be, but investment gains and losses remain lumpy by nature.

 

Cheers!

And yet their actual EPS was 2023 $174, 2024 $173 and 2025 $214. 
The take away message being they're regularly generating over $50 a share a year in investment gains. Even if 2025 was an outlier year, there is atleast 3.5B+ more in the tank, and more will inevitably be produced as they monetize their now wide variety of investment assets. Taken over a 5-10yr average $50(about $1-1.25B a year doesn't seem too onerous to me. 
A bigger question is what will they do with these rivers of cash. Share buybacks, insurance sub. minority buyouts, funding internal growth opportunities(insurance or otherwise). Equity minority interests, highly selective CRE opportunities, seeding selected venture opportunities with experienced, proven and visionary leaders, all seem like wonderful uses of capital to me. they don't seem to have issue with finding opportunities with a 15% plus hurdle rate. 
I'm not sure how they have acquired such a broad global network of contacts. Perhaps via their insurance subs? Either way it's impressive having that many options. I don't think that superpower ie contact network got much focus in the book "the Fairfax way". 

Edited by Txvestor
Posted
9 hours ago, glider3834 said:

Looking at sedi filings it looks like Fairfax bought ~167K shares over Jan-Feb-26 and then looking at AR below, looks like they may have bought another ~59K in the first week of March.

 

'Subsequent to December 31, 2025, the company purchased for cancellation 226,694 subordinate voting shares under the terms of its normal course issuer bids at a cost of $384.0' AR2025


Thanks this is super insightful! It’s only one week but it’s 4-5x the pace seen in Jan/Feb at a pace of 15% annualized shares out. Great to see them take advantage of the opportunity. (Similar to Eurobank but Fairfax is even more aggressive on the buyback!)

 

I took the liberty to share this insight on VIC.

Posted
9 hours ago, thedanmancan said:

fyi please, price of ~$28 almost 2x of CV of $15

 

Feels a bit of rubbing salt in the wounds but having recycled the old Atlas Corp proceeds back into the mothership, I guess it's not too painful. There's a lesson in there somewhere, I'm sure.

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
×
×
  • Create New...