Fairfax 2026

[Viking]

Here is a short article that estimates the impact on BVPS of share repurchases above book value. The size of the impact surprised me.  Hat tip to @SafetyinNumbers for bringing this topic up in the past. 

 

Share Buybacks, Book Value and Intrinsic Value

 

Warren Buffett has long argued that the economics of a share repurchase depend primarily on price.

 

When a company repurchases shares below intrinsic value, continuing shareholders benefit. Their ownership interest increases without requiring them to invest additional capital, increasing intrinsic value per share.

 

Book value is a separate question.

 

Buybacks below book value increase BVPS. Buybacks above book value decrease BVPS. Neither outcome tells investors whether value was created or destroyed. That depends on whether the shares were purchased below intrinsic value.

 

A good share buyback can therefore reduce book value per share while increasing intrinsic value per share.

 

The Emerging Story at Fairfax

 

This framework is becoming increasingly relevant for Fairfax.

 

Since 2018, the company has aggressively repurchased its shares. In recent years, a meaningful number of those shares have been purchased at a premium to book value—but clearly below management's assessment of intrinsic value.

 

The result is a growing divergence: Fairfax's buybacks are reducing reported BVPS while increasing intrinsic value per share.

 

Consider Q2 2026.

 

Fairfax appears to have repurchased approximately 675,000 shares for $1.08 billion, or roughly $1,600 per share. With BVPS of approximately $1,250 during the quarter, Fairfax paid about $350 per share above book value.

 

Premium to book value:

• $1,600 − $1,250 = $350 per share

Total premium paid:

• $350 × 675,000 = $236 million

Assuming approximately 19.95 million effective shares outstanding at quarter-end:

 

Estimated reduction in BVPS:

• $236 million ÷ 19.95 million = approximately $12 per share

In other words, Fairfax's Q2 share repurchases are estimated to have reduced reported BVPS by approximately $12 per share.

 

The estimated impact was approximately $10 per share in Q4 2025 and $8 per share in Q1 2026. Including Q2, share repurchases above book value may have reduced reported BVPS by approximately $30 per share over the past three quarters.

 

Yet because Fairfax purchased those shares below intrinsic value, continuing shareholders are economically better off.

 

The Impact on Traditional P/C Insurance Valuation

 

For P/C insurers, investors often focus on two closely related measures of performance: BVPS growth and ROE. Share repurchases above book value affect both.

 

BVPS declines. Fairfax is paying more than book value for each share it retires, reducing book value per remaining share.

 

ROE increases, all else equal. The buybacks reduce common equity. If earnings do not decline proportionately, Fairfax generates a higher return on a smaller equity base.

 

Value-creating buybacks can therefore make BVPS growth look weaker and ROE look stronger. Investors need to understand how capital allocation is affecting both measures.

 

Why It Matters

 

Book value remains an important metric for Fairfax, but it is becoming a less complete measure of value creation.

 

As Fairfax repurchases more shares above book value but below intrinsic value, reported BVPS will increasingly understate the economic benefit of those repurchases. This also complicates valuation: Fairfax can increase intrinsic value per share while reducing the book value investors use to calculate its price-to-book multiple.

 

This is an emerging story for Fairfax investors. The pace of buybacks has accelerated, and the cumulative impact on BVPS is becoming meaningful.

 

Buffett's framework provides the right lens: book value measures the accounting impact of a buyback; intrinsic value determines whether it created value.

 

For Fairfax shareholders, the gap between the two is becoming increasingly important.

 

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

 

Appendix: A Partial Bridge from Book Value to Intrinsic Value

 

Book value is an accounting measure. Intrinsic value is an economic measure. The two are not the same.

 

One obvious adjustment for Fairfax is the excess of fair value over carrying value (FV over CV) of its market traded non-insurance associates and consolidated holdings. This is not a theoretical adjustment. The value is observable in publicly traded securities but is not fully reflected in Fairfax's reported common equity.

 

At June 30, 2026, the excess of FV over CV is estimated at approximately $4.1 billion, or $205 per share based on 19.95 million effective shares outstanding. Assuming a 15% tax rate (lower for capital gains), the after-tax value is approximately $175 per share.

 

If Fairfax reports BVPS of approximately $1,300 at June 30:

• Reported BVPS: $1,300
• After-tax excess of FV over CV: +$175
• Adjusted BVPS: $1,475

This is not an estimate of Fairfax's intrinsic value. It is simply one identifiable adjustment that provides a partial bridge from accounting book value toward economic value.

 

It also provides useful context for Fairfax's share repurchases. A buyback at $1,600 may represent a meaningful premium to reported book value, but only a modest premium to adjusted book value—and a discount to intrinsic value.

 

That is the key distinction. Fairfax is reducing book value per share to increase intrinsic value per share. As the pace of buybacks continues, understanding the difference between the two will become increasingly important for investors.

Edited Monday at 10:11 PM by Viking

[djokovic1]

3 hours ago, SafetyinNumbers said:

Nice job! Different methodology for calculating intrinsic value than I use but they all end up in the same ball park. 

Thanks!

[Txvestor]

I asked chatGPT to construct this going back to 2017 since when's they've taken out almost 30% of their outstanding shares via buybacks. 
I would argue this divergence goes back to ever since the year the buyback started happening in earnest about 9yrs ago, once their equity hedges came off. This is evidenced by the fact that their earnings whether measured by total net earnings or EPS has grown despite all the capital ~$6B cumulatively allocated towards these buybacks. 
 

Year	Shares Repurchased (millions)	Total Cost (US$ millions)	Avg. Purchase Price	Year-end Book Value/Share	Avg. Purchase Price as % of BV
2017	~0.35	~150	~$430	~$450	96%
2018	~0.55	~250	~$455	~$455	100%
2019	~0.38	~170	~$447	~$484	92%
2020	~0.56	~215	~$385	~$525	73%
2021	~2.31	~1,240	~$537	~$674	80%
2022	~0.61	~398	~$652	~$752	87%
2023	~0.31	~276	~$890	~$940	95%
2024	1.347	1,588	$1,179	$1,060	111%
2025	1.007	1,625	$1,614	$1,260	128% Â

Edited Monday at 10:50 PM by Txvestor

[Parsad]

23 minutes ago, Txvestor said:

I asked chatGPT to construct this going back to 2017 since when's they e taken out almost 30% of their outstanding shares via buybacks. 
I would argue this divergence goes back to ever since the year my back started happening in earnest about 9yrs ago, once their hedges came off. This is evidenced by the fact that their earnings whether measured by total net earnings or EPS has grown despite all the capital ~$6B cumulatively allocated towards these buybacks. 
 

Year	Shares Repurchased (millions)	Total Cost (US$ millions)	Avg. Purchase Price	Year-end Book Value/Share	Avg. Purchase Price as % of BV
2017	~0.35	~150	~$430	~$450	96%
2018	~0.55	~250	~$455	~$455	100%
2019	~0.38	~170	~$447	~$484	92%
2020	~0.56	~215	~$385	~$525	73%
2021	~2.31	~1,240	~$537	~$674	80%
2022	~0.61	~398	~$652	~$752	87%
2023	~0.31	~276	~$890	~$940	95%
2024	1.347	1,588	$1,179	$1,060	111%
2025	1.007	1,625	$1,614	$1,260	128% Â

 

Txinvestor, there is one more useful line you could add in there...percentage of total shares retired for that year.  So, that would indicate if they are paying up in terms of p/b, but buying in the same percentage of total shares outstanding. 

 

That would clearly indicate that they believe there is a long-term structural change in intrinsic value and a higher p/b is achieving the long-term return they want.  Cheers!

[LC]

3 hours ago, Parsad said:

Txinvestor, there is one more useful line you could add in there...percentage of total shares retired for that year. 

Chatgpt:

 

Year	Shares Repurchased (millions)	Total Cost (US$ millions)	Avg. Purchase Price	Year-end Book Value/Share	Avg. Purchase Price as % of BV	Estimated % of Shares Retired
2017	~0.35	~150	~$430	~$450	96%	~1.4%
2018	~0.55	~250	~$455	~$455	100%	~2.2%
2019	~0.38	~170	~$447	~$484	92%	~1.5%
2020	~0.56	~215	~$385	~$525	73%	~2.2%
2021	~2.31	~1,240	~$537	~$674	80%	~9.3%
2022	~0.61	~398	~$652	~$752	87%	~2.7%
2023	~0.31	~276	~$890	~$940	95%	~1.4%
2024	1.347	1,588	$1,179	$1,060	111%	~6.3%
2025	1.007	1,625	$1,614	$1,260	128%	~5.0%

[SafetyinNumbers]

31 minutes ago, LC said:

Chatgpt:

 

Year	Shares Repurchased (millions)	Total Cost (US$ millions)	Avg. Purchase Price	Year-end Book Value/Share	Avg. Purchase Price as % of BV	Estimated % of Shares Retired
2017	~0.35	~150	~$430	~$450	96%	~1.4%
2018	~0.55	~250	~$455	~$455	100%	~2.2%
2019	~0.38	~170	~$447	~$484	92%	~1.5%
2020	~0.56	~215	~$385	~$525	73%	~2.2%
2021	~2.31	~1,240	~$537	~$674	80%	~9.3%
2022	~0.61	~398	~$652	~$752	87%	~2.7%
2023	~0.31	~276	~$890	~$940	95%	~1.4%
2024	1.347	1,588	$1,179	$1,060	111%	~6.3%
2025	1.007	1,625	$1,614	$1,260	128%	~5.0%

Could also add the impact to BVPS as Viking calculated. Just need the effective shares outstanding at the end of 2024 and 2025. I don’t think the accretion to BVPS from buybacks at a discount are relevant. 

[Txvestor]

I think the key metric in assessing intrinsic value—while admittedly imperfect and particularly volatile in Fairfax’s case—is net earnings. Ultimately, a business’s intrinsic value is determined by its ability to generate earnings over time.

With that in mind, it may be informative to look at the trend in annual net earnings alongside earnings per share (EPS). Since Fairfax has also been aggressively repurchasing shares, the EPS trend provides additional insight into the value accruing to each remaining shareholder.

I also thought it would be useful to include return on equity (ROE). Although ROE is also subject to significant year-to-year volatility and accounting noise at Fairfax, it remains a helpful measure of how effectively the company has been compounding shareholders’ capital over time.

Year	Total Revenue (US$B)	Net Earnings (US$B)	Diluted EPS (US$)	ROE
2017	16.2	1.74	64.98	15.0%
2018	15.6	0.38	14.08	3.2%
2019	19.1	1.65	62.45	13.3%
2020	20.0	2.06	80.76	15.6%
2021	27.0	3.18	123.61	20.2%
2022	29.0	3.38	129.23	18.6%
2023	31.8	4.38	173.24	20.2%
2024	35.2	3.87	160.56	16.0%
2025	38.3	4.77	213.78	18.1%

Edited yesterday at 02:49 AM by Txvestor

[Viking]

4 hours ago, LC said:

Chatgpt:

 

Year	Shares Repurchased (millions)	Total Cost (US$ millions)	Avg. Purchase Price	Year-end Book Value/Share	Avg. Purchase Price as % of BV	Estimated % of Shares Retired
2017	~0.35	~150	~$430	~$450	96%	~1.4%
2018	~0.55	~250	~$455	~$455	100%	~2.2%
2019	~0.38	~170	~$447	~$484	92%	~1.5%
2020	~0.56	~215	~$385	~$525	73%	~2.2%
2021	~2.31	~1,240	~$537	~$674	80%	~9.3%
2022	~0.61	~398	~$652	~$752	87%	~2.7%
2023	~0.31	~276	~$890	~$940	95%	~1.4%
2024	1.347	1,588	$1,179	$1,060	111%	~6.3%
2025	1.007	1,625	$1,614	$1,260	128%	~5.0%

Another important ‘bucket’ is the annual change in excess of FV over CV for market traded non-insurance associate and consolidated companies. It was a loss of $660 million in 2020. And at June 30, 2026 it is likely +$4.1 billion. The ‘swing’ over the past 6.5 years is about $730 million per year pre-tax or about $620 million after-tax (assuming 15% tax rate on capital gains). That is a material amount per year - too big to ignore. We know Fairfax includes this metric in their analysis. 
 

When we include this metric, Fairfax’s buybacks make even more sense - they look even better. 

Edited yesterday at 05:29 AM by Viking

[Txvestor]

37 minutes ago, Viking said:

Another important ‘bucket’ is the annual change in excess of FV over CV for market traded non-insurance companies. It was a loss of $660 million in 2020. And at June 30, 2026 it is likely +$4.1 billion. The ‘swing’ over the past 6.5 years is about $730 million per year pre-tax or about $620 million after-tax (assuming 15% tax rate on capital gains). That is a material amount per year - too big to ignore. We know Fairfax includes this metric in their analysis. 
 

When we include this metric, Fairfax’s buybacks make even more sense - they look even better. 

I think like Berkshire, for a long time Fairfax taught their shareholders to consider book value when evaluating the share price.
I think the gap between intrinsic value and book value has diverged significantly overtime. In both of their cases and probably more so in the case of Fairfax. Considering Mr Market was assigning values as low as 0.6 BV in the past. 1.25 may appear like more reasonable. But in an environment where the insurance segment is much improved, interest rates are likely to stay materially higher for longer, and where excess fair value over carrying value is 4B+, and with where recent ROE numbers have been ie high teens compared to mid teens previously, I believe that is not much of a rerating. My view is at the recent average of 18% ROE, even accounting for the volatility and inherent risks of an insurance company, an adequately capitalized, well managed company that's been around nearly 50yrs  should merit 1.8x BV in this very expensive stock market. I'm happy for them to keep buying at these levels. 

, 

[Parsad]

3 hours ago, LC said:

Chatgpt:

 

Year	Shares Repurchased (millions)	Total Cost (US$ millions)	Avg. Purchase Price	Year-end Book Value/Share	Avg. Purchase Price as % of BV	Estimated % of Shares Retired
2017	~0.35	~150	~$430	~$450	96%	~1.4%
2018	~0.55	~250	~$455	~$455	100%	~2.2%
2019	~0.38	~170	~$447	~$484	92%	~1.5%
2020	~0.56	~215	~$385	~$525	73%	~2.2%
2021	~2.31	~1,240	~$537	~$674	80%	~9.3%
2022	~0.61	~398	~$652	~$752	87%	~2.7%
2023	~0.31	~276	~$890	~$940	95%	~1.4%
2024	1.347	1,588	$1,179	$1,060	111%	~6.3%
2025	1.007	1,625	$1,614	$1,260	128%	~5.0%

 

Thanks LC!  Clearly looks like there has been enough of a structural/economic change at Fairfax where buying back their own shares up to 1.3 times book value generates an after-tax return equivalent to or better than their target return for shareholders.  Cheers!

[thedanmancan]

From Bloomberg News -

 

India is close to accepting an offer from Fairfax Financial Holdings Ltd. for IDBI Bank Ltd., possibly at a slightly higher price, according to people with knowledge of the matter, in what could potentially be the biggest foreign investment in the country’s banking sector.

….

Fairfax’s purchase of a 60.7% stake in IDBI would be worth around $5.7 billion, at current prices

Edited yesterday at 12:47 PM by thedanmancan

[SafetyinNumbers]

24 minutes ago, thedanmancan said:

From Bloomberg News -

 

India is close to accepting an offer from Fairfax Financial Holdings Ltd. for IDBI Bank Ltd., possibly at a slightly higher price, according to people with knowledge of the matter, in what could potentially be the biggest foreign investment in the country’s banking sector.

….

Fairfax’s purchase of a 60.7% stake in IDBI would be worth around $5.7 billion, at current prices

Should be one of the biggest deals Fairfax has ever done. I expect FIH to get partners including FFH to get the deal done. It will be interesting how much of the ~$5b is owned by Fairfax companies vs outside investors. My guess is more than half. I’m also curious how the accounting will be done. FIH usually marks to market its positions that are listed but FFH might be deemed to have significant influence or control. 

[BenjiGraham33]

On 7/13/2026 at 11:31 AM, Viking said:

Here is a short article that estimates the impact on BVPS of share repurchases above book value. The size of the impact surprised me.  Hat tip to @SafetyinNumbers for bringing this topic up in the past. 

 

Share Buybacks, Book Value and Intrinsic Value

 

Warren Buffett has long argued that the economics of a share repurchase depend primarily on price.

 

When a company repurchases shares below intrinsic value, continuing shareholders benefit. Their ownership interest increases without requiring them to invest additional capital, increasing intrinsic value per share.

 

Book value is a separate question.

 

Buybacks below book value increase BVPS. Buybacks above book value decrease BVPS. Neither outcome tells investors whether value was created or destroyed. That depends on whether the shares were purchased below intrinsic value.

 

A good share buyback can therefore reduce book value per share while increasing intrinsic value per share.

 

The Emerging Story at Fairfax

 

This framework is becoming increasingly relevant for Fairfax.

 

Since 2018, the company has aggressively repurchased its shares. In recent years, a meaningful number of those shares have been purchased at a premium to book value—but clearly below management's assessment of intrinsic value.

 

The result is a growing divergence: Fairfax's buybacks are reducing reported BVPS while increasing intrinsic value per share.

 

Consider Q2 2026.

 

Fairfax appears to have repurchased approximately 675,000 shares for $1.08 billion, or roughly $1,600 per share. With BVPS of approximately $1,250 during the quarter, Fairfax paid about $350 per share above book value.

 

Premium to book value:

• $1,600 − $1,250 = $350 per share

Total premium paid:

• $350 × 675,000 = $236 million

Assuming approximately 19.95 million effective shares outstanding at quarter-end:

 

Estimated reduction in BVPS:

• $236 million ÷ 19.95 million = approximately $12 per share

In other words, Fairfax's Q2 share repurchases are estimated to have reduced reported BVPS by approximately $12 per share.

 

The estimated impact was approximately $10 per share in Q4 2025 and $8 per share in Q1 2026. Including Q2, share repurchases above book value may have reduced reported BVPS by approximately $30 per share over the past three quarters.

 

Yet because Fairfax purchased those shares below intrinsic value, continuing shareholders are economically better off.

 

The Impact on Traditional P/C Insurance Valuation

 

For P/C insurers, investors often focus on two closely related measures of performance: BVPS growth and ROE. Share repurchases above book value affect both.

 

BVPS declines. Fairfax is paying more than book value for each share it retires, reducing book value per remaining share.

 

ROE increases, all else equal. The buybacks reduce common equity. If earnings do not decline proportionately, Fairfax generates a higher return on a smaller equity base.

 

Value-creating buybacks can therefore make BVPS growth look weaker and ROE look stronger. Investors need to understand how capital allocation is affecting both measures.

 

Why It Matters

 

Book value remains an important metric for Fairfax, but it is becoming a less complete measure of value creation.

 

As Fairfax repurchases more shares above book value but below intrinsic value, reported BVPS will increasingly understate the economic benefit of those repurchases. This also complicates valuation: Fairfax can increase intrinsic value per share while reducing the book value investors use to calculate its price-to-book multiple.

 

This is an emerging story for Fairfax investors. The pace of buybacks has accelerated, and the cumulative impact on BVPS is becoming meaningful.

 

Buffett's framework provides the right lens: book value measures the accounting impact of a buyback; intrinsic value determines whether it created value.

 

For Fairfax shareholders, the gap between the two is becoming increasingly important.

 

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

 

Appendix: A Partial Bridge from Book Value to Intrinsic Value

 

Book value is an accounting measure. Intrinsic value is an economic measure. The two are not the same.

 

One obvious adjustment for Fairfax is the excess of fair value over carrying value (FV over CV) of its market traded non-insurance associates and consolidated holdings. This is not a theoretical adjustment. The value is observable in publicly traded securities but is not fully reflected in Fairfax's reported common equity.

 

At June 30, 2026, the excess of FV over CV is estimated at approximately $4.1 billion, or $205 per share based on 19.95 million effective shares outstanding. Assuming a 15% tax rate (lower for capital gains), the after-tax value is approximately $175 per share.

 

If Fairfax reports BVPS of approximately $1,300 at June 30:

• Reported BVPS: $1,300
• After-tax excess of FV over CV: +$175
• Adjusted BVPS: $1,475

This is not an estimate of Fairfax's intrinsic value. It is simply one identifiable adjustment that provides a partial bridge from accounting book value toward economic value.

 

It also provides useful context for Fairfax's share repurchases. A buyback at $1,600 may represent a meaningful premium to reported book value, but only a modest premium to adjusted book value—and a discount to intrinsic value.

 

That is the key distinction. Fairfax is reducing book value per share to increase intrinsic value per share. As the pace of buybacks continues, understanding the difference between the two will become increasingly important for investors.

Hi @Viking, enjoyed meeting you in April. On the topic of BVPS at Fairfax:

 

An asset is a controlled resource that is expected to provide future cash flows. If one assumes that in the long run:

1. Fairfax's float is likely to be greater than or equal to current

2. The insurance businesses are likely to approximately break even

 

Then Fairfax's float of $41B can be considered an asset that is at least as valuable as cash.

 

As @Packer16 has pointed out, the current combined cost of debt + float is less than zero. If this can possibly persist, Fairfax in 2026 might be simplified as being seen as a pile of ~$78B in T-Bills, yielding ~$3B per year after tax, buying itself for at a pace of 5+% per year at a valuation of $35B. Viewed from this lens, intrinsic value per share is >$3,500/share and buybacks are occurring in the $1,600s. This valuation seems to be in line with others' conclusions. I believe this is more or less the current situation, and approximately in line with how Prem views it. What do you think? See below for excerpts from the 2025 annual report.

 

 

In short, Book Value per Share has become irrelevant as it largely excludes important economics at Fairfax including:

1. Float

2. Excess economic goodwill vs reported goodwill of insurance subsidiaries

3. Excess fair value vs carrying value of non-insurance businesses, and

4. The growing pace of buybacks, and the incorrect effect effect they have on intrinsic value per share growth if success is measured in BVPS

 

From the 2025 annual report:

 

Page 27, float as "perhaps" an asset:

 

Page 22, goodwill accounting rules obscuring insurance subsidiary value:

 

Page 23, multiple-to-BVPS' irrelevance when approximating intrinsic value:

 

Page 7, likely after-tax income of ~$3B for the next 4 years:

 

 

[SafetyinNumbers]

3 hours ago, BenjiGraham33 said:

In short, Book Value per Share has become irrelevant as it largely excludes important economics at Fairfax including:

1. Float

2. Excess economic goodwill vs reported goodwill of insurance subsidiaries

3. Excess fair value vs carrying value of non-insurance businesses, and

4. The growing pace of buybacks, and the incorrect effect effect they have on intrinsic value per share growth if success is measured in BVPS

 

I don’t think book value is irrelevant it’s just not the same as intrinsic value. P/B can be a measure of intrinsic value but it has to be paired with a view on ROE. If assets are marked under fair value but they earn an acceptable return it should show up in the ROE. The relationship between P/B to ROE should be parabolic especially for companies with high returns on incremental capital and low dividends like Fairfax.

Edited 17 hours ago by SafetyinNumbers

[Maverick47]

23 minutes ago, SafetyinNumbers said:

P/B can be a measure of intrinsic value but it has to be paired with a view on ROE. If assets are marked under fair value but they earn an acceptable return it should show up in the ROE

Thanks for this!  It was super easy for me to understand the virtuous effect of share buybacks below book value of a few years ago.  I was struggling to model out in my mind the impacts of buybacks above book value, but below intrinsic value.

 

As a shorthand way of valuing shares, I also consider a multiple to normalized earnings per share.  This would also capture the impact of buybacks below intrinsic value but above book value, if the buybacks result in a higher EPS wouldn’t it?  Similar to seeing a higher ROE.

 

I’m probably going to have to be a bit more cautious about using book value per share of Fairfax as the basis for my personal estimate of intrinsic value unless, as you note, I also remember to consider the projected ROE.

[SafetyinNumbers]

6 hours ago, Maverick47 said:

As a shorthand way of valuing shares, I also consider a multiple to normalized earnings per share.  This would also capture the impact of buybacks below intrinsic value but above book value, if the buybacks result in a higher EPS wouldn’t it?  Similar to seeing a higher ROE.

 

I think that’s correct. If you are calculating a normalized EPS, it’s the same mechanics to get to a normalized ROE. I use 15x EPS for an estimate of intrinsic value which is a fair market multiple for an above average ROE so it has some margin of safety included.

[BenjiGraham33]

11 hours ago, SafetyinNumbers said:

I don’t think book value is irrelevant it’s just not the same as intrinsic value. P/B can be a measure of intrinsic value but it has to be paired with a view on ROE. If assets are marked under fair value but they earn an acceptable return it should show up in the ROE. The relationship between P/B to ROE should be parabolic especially for companies with high returns on incremental capital and low dividends like Fairfax.

I hear you - these methods work, and ROE as a sanity check on economic book value makes sense. Also, hello...hope to catch up more at the next conference. Need as many smart and generous people in my life as possible!

 

You'll have a perfectly decent day if you ignore the below. But if you get bored, humor this ridiculous & unrealistic thought experiment, if you will. How would you fill in the blanks? Is the premise too ridiculous to even consider?

 

 

Edited 6 hours ago by BenjiGraham33

[Hoodlum]

RBC reaffirmed their Buy rating on Fairfax this week, with target set at $2277US (3200 cdn)

[SafetyinNumbers]

1 hour ago, BenjiGraham33 said:

I hear you - these methods work, and ROE as a sanity check on economic book value makes sense. Also, hello...hope to catch up more at the next conference. Need as many smart and generous people in my life as possible!

 

You'll have a perfectly decent day if you ignore the below. But if you get bored, humor this ridiculous & unrealistic thought experiment, if you will. How would you fill in the blanks? Is the premise too ridiculous to even consider?

 

 

Scenario 1 is easier. P/b for a 17% ROE should be around 2.25-2.5x P/B. IVPS short cut of BVPS + float per share is also easy at $4750. Scenario 2 I don’t understand. 

[SafetyinNumbers]

12 minutes ago, Hoodlum said:

RBC reaffirmed their Buy rating on Fairfax this week, with target set at $2277US (3200 cdn)

Analyst estimates for Q2 look too high. They are ignoring bond losses again presumably because other insurance companies don’t include it in their adjusted EPS. 

[Viking]

Fairfax Q2 2026 Earnings Preview

 

My very rough guess is earnings will come in around $60/share when Fairfax reports Q2 results. Analyst consensus is about $80/share. I look forward to getting feedback from board members. 

 

This would put book value at June 30, 2026, at about $1,300/share.

• BVPS at March 31, 2026: $1,250
• Earnings: +$60
• Impact of buying back shares above BV: -$12 per share

Accounting results for Fairfax are an incomplete measure of value creation. If we include excess of FV over CV for market traded non-insurance associate and consolidated holdings, '"economic earnings" increases to $68/share.

 

6-Month "economic earnings" are $129 per share. Outstanding. 

 

Note, I am not doing this exercise to come up with a high conviction specific number (I know it will be wrong) for EPS or BVPS. Rather, I do it for the build of the individual items - to help prepare me for Fairfax’s earnings release. Below is the logic I used to come up with my forecast.

 

 

—————

 

Details

 

Underwriting profit:

 

Q2 2025

• Net premiums written: $7.17B
• CR: 93.3%
• Underwriting profit: $426.9M
• Net favourable prior year reserve development: $163.2M
• Catastrophe losses: $140.1M

Q2 2026 Estimate: similar to Q2 2025

• Net premiums written: modest growth (low single digit)
• CR: 93.5%
• Underwriting profit: $435M

—————

 

Interest income:

• Q4 2025: $646M
• Q1 2026: $662M

At March 31, 2026:

• Duration: 2.2 years
• Average maturity: 3 years
• Yield: approximately 5%.

Q2 2026 Estimate: similar to Q1 2026

• $665M

—————

 

Share of profit of associates

 

Waterous Energy III (Greenfire).

• Q1: $117 million tailwind
• Q2: $50 million headwind (loss)?

Poseidon sale closed May 29.

• Normal contribution: $75 million per quarter
• Adjusted for sale: $62 million ($13 million reduction)

Sanmar??? (Sale in March)

 

Q1 2026: $372M

 

Q2 2026 Estimate

• $200M

—————

 

Non-insurance consolidated companies

 

Q2 2025: $120M

Q1 2026: $3M (seasonality, AGT IPO costs)

 

My annual estimate is $450M.

 

Q2 2026 Estimate

• $120M

——————

 

Investment gains (losses)

 

Equities

 

From an accounting perspective, it is only the market to market equity holdings that impact this bucket each quarter.

 

Q2 2026 Estimate for equities

• loss of $50M

Fairfax has been slowly selling down their position in BlackBerry over the past year. BlackBerry was one of the big gainers in the quarter. If Fairfax continued to sell  BlackBerry shares in the quarter the gain may be much smaller than $242 million. As a result, the loss estimate for equities of $50 million would be larger.

 

 

Big movers:

 

 

Fixed Income

 

Interest rates were higher in Q1. Impact on fixed income?

• Q1 2026: loss of $364 million

Fixed income portfolio at end of Q1:

• Size: $49.8B
• Duration: 2.2 years
• Average maturity: 3 years
• Yield: approximately 5%
• “zero traditional private credit exposure”

Interest rates continued their move higher in Q2.

 

Q2 2026 Estimate for fixed income: Larger than loss in Q1

• loss of $450 million

 

 

Large one time investment gain

 

Poseidon: $837 million

 

I like to separate this number out in my model.

 

Total investment gains (losses)

 

Q2 2026 Estimate:

• Equities: -$50M
• Fixed income: -$450M
• Poseidon: $837M
• Total: $337M

There are couple of other items that impact investment gains each quarter:

1. Digit Insurance
2. Fairfax India - change in value of the individual holdings

Digits share price was flat on the quarter, so no impact of Fairfax.

I do not track the holdings in Fairfax India from quarter to quarter, so not sure of the impact here to Fairfax.

 

——————

 

IRFS 17 and Life Insurance and Runoff

 

This is a catchall bucket for our model.

 

IFRS 17 - Interest rate impact

 

Rising interest rates are a tailwind for IFRS 17. This offsets some of the loss in the fixed income portfolio.

 

Q1 2026:

• Gain of $180M (about half the loss in fixed income)

Q2 2026 Estimate

• Gain of $225M

Total IFRS + Life/Runoff

 

Q1 2026:

• $329M

Q2 2026 Estimate for this bucket

• $375M

This number is low conviction. 

—————-

 

Interest expense

 

Q1 2026: $212M

Q2 2026 estimate: $220M

 

Corporate expense and other

 

Q1 2026: $108M

Q2 2026 estimate: $120M

 

Tax rate

 

Q1 2026: 29%

Q2 2026 estimate: 20%

 

Q1 was elevated. Not sure why.

Q2: Significant gain ($837 million) from sale of Poseidon is likely taxed at lower capital gains rate.  

 

Minority interest

 

Estimate is 6%?

 

—————-

 

Share count

 

Shares outstanding Q1 2026;

• Effective: 20.62M
• Diluted:  $22.25

It appears Fairfax was very aggressive buying back stock in Q2:

• 675,000?
• Total: $1.08 billion?
• Per share: $1,600

Estimate of shares outstanding at June 30, 2026:

• Effective: 19.95M
• Diluted: 21.58M

——————

 

Impact on BVPS of share buybacks above book value

 

Shares were purchased above book value:

• BVPS: $1,250
• Difference: $350 per share or $236.25 million

Reduction in BVPS due to buybacks above book value:

• $236.25 million / 19.95 million effective shares = $12/share

—————-

 

Excess of FV over CV for associate and non-insurance consolidated holdings

 

March 31, 2026: $3.9B

 

June 30, 2026 Estimate:

• $4.1B, or $206/share pre-tax or $175/share after-tax (15% tax rate)

Quarter over quarter change:

• $200 million, or $8 per share after-tax

Headwind: Sale of 50% of Poseidon was a reduction of $837M.

Tailwind: Driven by Eurobank, remaining holdings had a strong Q2.

Net result in the quarter: an estimated increase of $200M to $4.1B.   

 

Economic EPS estimate: $61 + $8 = $69

 

Economic BVPS Estimate: $1,300 + $175 = $1,475/share

Fairfax share price (July 10): $1,660

P/BV: 1.13x

Edited 2 hours ago by Viking

[73 Reds]

32 minutes ago, Viking said:

Fairfax Q2 2026 Earnings Preview

 

My very rough guess is earnings will come in around $60/share when Fairfax reports Q2 results. Analyst consensus is about $80/share. I look forward to getting feedback from board members. 

 

This would put book value at June 30, 2026, at about $1,300/share.

• BVPS at March 31, 2026: $1,250
• Earnings: +$60
• Impact of buying back shares above BV: -$12 per share

Accounting results for Fairfax are an incomplete measure of value creation. If we include excess of FV over CV for market traded non-insurance associate and consolidated holdings, '"economic earnings" increases to $68/share.

 

6-Month "economic earnings" are $129 per share. Outstanding. 

 

Note, I am not doing this exercise to come up with a high conviction specific number (I know it will be wrong) for EPS or BVPS. Rather, I do it for the build of the individual items - to help prepare me for Fairfax’s earnings release. Below is the logic I used to come up with my forecast.

 

 

—————

 

Details

 

Underwriting profit:

 

Q2 2025

• Net premiums written: $7.17B
• CR: 93.3%
• Underwriting profit: $426.9M
• Net favourable prior year reserve development: $163.2M
• Catastrophe losses: $140.1M

Q2 2026 Estimate: similar to Q2 2025

• Net premiums written: modest growth (low single digit)
• CR: 93.5%
• Underwriting profit: $435M

—————

 

Interest income:

• Q4 2025: $646M
• Q1 2026: $662M

At March 31, 2026:

• Duration: 2.2 years
• Average maturity: 3 years
• Yield: approximately 5%.

Q2 2026 Estimate: similar to Q1 2026

• $665M

—————

 

Share of profit of associates

 

Waterous Energy III (Greenfire).

• Q1: $117 million tailwind
• Q2: $50 million headwind (loss)?

Poseidon sale closed May 29.

• Normal contribution: $75 million per quarter
• Adjusted for sale: $62 million ($13 million reduction)

Sanmar??? (Sale in March)

 

Q1 2026: $372M

 

Q2 2026 Estimate

• $200M

—————

 

Non-insurance consolidated companies

 

Q2 2025: $120M

Q1 2026: $3M (seasonality, AGT IPO costs)

 

My annual estimate is $450M.

 

Q2 2026 Estimate

• $120M

——————

 

Investment gains (losses)

 

Equities

 

From an accounting perspective, it is only the market to market equity holdings that impact this bucket each quarter.

 

Q2 2026 Estimate for equities

• loss of $50M

Fairfax has been slowly selling down their position in BlackBerry over the past year. BlackBerry was one of the big gainers in the quarter. If Fairfax continued to sell  BlackBerry shares in the quarter the gain may be much smaller than $242 million. As a result, the loss estimate for equities of $50 million would be larger.

 

 

Big movers:

 

 

Fixed Income

 

Interest rates were higher in Q1. Impact on fixed income?

• Q1 2026: loss of $364 million

Fixed income portfolio at end of Q1:

• Size: $49.8B
• Duration: 2.2 years
• Average maturity: 3 years
• Yield: approximately 5%
• “zero traditional private credit exposure”

Interest rates continued their move higher in Q2.

 

Q2 2026 Estimate for fixed income: Larger than loss in Q1

• loss of $450 million

 

 

Large one time investment gain

 

Poseidon: $837 million

 

I like to separate this number out in my model.

 

Total investment gains (losses)

 

Q2 2026 Estimate:

• Equities: -$50M
• Fixed income: -$450M
• Poseidon: $837M
• Total: $337M

There are couple of other items that impact investment gains each quarter:

1. Digit Insurance
2. Fairfax India - change in value of the individual holdings

Digits share price was flat on the quarter, so no impact of Fairfax.

I do not track the holdings in Fairfax India from quarter to quarter, so not sure of the impact here to Fairfax.

 

——————

 

IRFS 17 and Life Insurance and Runoff

 

This is a catchall bucket for our model.

 

IFRS 17 - Interest rate impact

 

Rising interest rates are a tailwind for IFRS 17. This offsets some of the loss in the fixed income portfolio.

 

Q1 2026:

• Gain of $180M (about half the loss in fixed income)

Q2 2026 Estimate

• Gain of $225M

Total IFRS + Life/Runoff

 

Q1 2026:

• $329M

Q2 2026 Estimate for this bucket

• $375M

This number is low conviction. 

—————-

 

Interest expense

 

Q1 2026: $212M

Q2 2026 estimate: $220M

 

Corporate expense and other

 

Q1 2026: $108M

Q2 2026 estimate: $120M

 

Tax rate

 

Q1 2026: 29%

Q2 2026 estimate: 20%

 

Q1 was elevated. Not sure why.

Q2: Significant gain ($837 million) from sale of Poseidon is likely taxed at lower capital gains rate.  

 

Minority interest

 

Estimate is 6%?

 

—————-

 

Share count

 

Shares outstanding Q1 2026;

• Effective: 20.62M
• Diluted:  $22.25

It appears Fairfax was very aggressive buying back stock in Q2:

• 675,000?
• Total: $1.08 billion?
• Per share: $1,600

Estimate of shares outstanding at June 30, 2026:

• Effective: 19.95M
• Diluted: 21.58M

——————

 

Impact on BVPS of share buybacks above book value

 

Shares were purchased above book value:

• BVPS: $1,250
• Difference: $350 per share or $236.25 million

Reduction in BVPS due to buybacks above book value:

• $236.25 million / 19.95 million effective shares = $12/share

—————-

 

Excess of FV over CV for associate and non-insurance consolidated holdings

 

March 31, 2026: $3.9B

 

June 30, 2026 Estimate:

• $4.1B, or $206/share pre-tax or $175/share after-tax (15% tax rate)

Quarter over quarter change:

• $200 million, or $8 per share after-tax

Headwind: Sale of 50% of Poseidon was a reduction of $837M.

Tailwind: Driven by Eurobank, remaining holdings had a strong Q2.

Net result in the quarter: an estimated increase of $200M to $4.1B.   

 

Economic EPS estimate: $61 + $8 = $69

 

Economic BVPS Estimate: $1,300 + $175 = $1,475/share

Fairfax share price (July 10): $1,660

P/BV: 1.13x

@Viking I love your work but have a lingering question for you and others:  Since most acknowledge the benefit of share buybacks even above BV which also go to reduce BV, why does everyone seem to focus on BV as a valuation metric?  Can't we find a better valuation metric that doesn't overtly reduce the estimated value that we are trying to measure?  Is it just because BV is so easy to measure or is there another reason why every time someone seeks to value Fairfax, book value is always considered?

[Viking]

21 minutes ago, 73 Reds said:

@Viking I love your work but have a lingering question for you and others:  Since most acknowledge the benefit of share buybacks even above BV which also go to reduce BV, why does everyone seem to focus on BV as a valuation metric?  Can't we find a better valuation metric that doesn't overtly reduce the estimated value that we are trying to measure?  Is it just because BV is so easy to measure or is there another reason why every time someone seeks to value Fairfax, book value is always considered?

 

@73 Reds, great question: "Why use BV as a valuation metric?" The primary reason for me is habit - it is built into my models/mental framework. Another important reason is it also the key metric that the investment community focusses on for P/C insurers (rightly or wrongly).

 

I do include "excess of FV over CV"  in my models - that provides an important improvement to accounting BV - getting us closer to "economic BV." But that is incomplete.

 

Is book value still relevant as a valuation measure for Fairfax? Great question. I need to think more about it. What do others think? 

 

PS: When I value Fairfax I like to use "normalized earnings" and PE. Much of their EPS is very stable. And for investment gains I use a three year average - which makes this part also very stable. As a result, PE works for me. Bottom line... stock is trading today at about 8.5 x "normalized earnings." Crazy cheap from my perspective. 

Edited 1 hour ago by Viking

[Malmqky]

Basically the Berkshire problem, no?

 

I know there are better ways to value it, but I have high confidence that when Berkshire trades at 1.3-1.4x book, it is trading below intrinsic value.

 

Likewise, I have high confidence Fairfax trades at a discount when it trades at 1-1.2x book. And I think intrinisic value is growing quickly.

 

Don't need to be exact to know when you got a deal.

 

I also like the idea of valuing float. Smarter people than I can do SOTP for a better valuation..

Edited 1 hour ago by Malmqky

[TwoCitiesCapital]

3 hours ago, SafetyinNumbers said:

Analyst estimates for Q2 look too high. They are ignoring bond losses again presumably because other insurance companies don’t include it in their adjusted EPS. 

 

Bond losses would be offset by the liability adjustment on the insurance if they ran equal duration. 

 

The bonds have typically been shy of duration of the insurance, so should actually be a sight net positive