Fairfax 2026

[Viking]

5 minutes ago, TwoCitiesCapital said:

 

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

 

@TwoCitiesCapital, is that what happened in Q1?

 

From the Q1 conference call:

 

"In the first quarter, net earnings included a $184 million unrealized loss due to increasing interest rates in the quarter. This consisted of unrealized losses on our bonds of $364 million, as I previously mentioned, offset by the increase in discount under IFRS 17 on our insurance and reinsurance reserves of $180 million. For the first quarter of 2025, this number was a net gain of $120 million."

Edited 1 hour ago by Viking

[CharlesMunger]

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. 

So after earnings release potentially another big down day and Fairfax with the opportunity to buy back more shares? I’ll take it

[TwoCitiesCapital]

31 minutes ago, Viking said:

 

@TwoCitiesCapital, is that what happened in Q1?

 

From the Q1 conference call:

 

"In the first quarter, net earnings included a $184 million unrealized loss due to increasing interest rates in the quarter. This consisted of unrealized losses on our bonds of $364 million, as I previously mentioned, offset by the increase in discount under IFRS 17 on our insurance and reinsurance reserves of $180 million. For the first quarter of 2025, this number was a net gain of $120 million."

 

Interesting.

 

I'm speaking in generalizations of raw interest  rate moves, but the generalization didn't seem to hold true for that quarter (when it has been true in others). 

 

There's probably some component of what credit and mortgage spreads did too since Fairfax doesn't just own straight treasuries and that may explain the difference in Q1. 

 

But as a ballpark, if they're under in duration, they'll outperform their liability when rates rise. 

 

 

[SafetyinNumbers]

10 minutes ago, CharlesMunger said:

So after earnings release potentially another big down day and Fairfax with the opportunity to buy back more shares? I’ll take it

Maybe! BVPS growing again and a 100% increase in sequential earnings might offset the sellers though.

[SafetyinNumbers]

Just now, TwoCitiesCapital said:

 

Interesting.

 

I'm speaking in generalizations of raw interest  rate moves, but the generalization didn't seem to hold true for that quarter (when it has been true in others). 

 

There's probably some component of what credit and mortgage spreads did too since Fairfax doesn't just own straight treasuries and that may explain the difference in Q1. 

 

But as a ballpark, if they're under in duration, they'll outperform their liability when rates rise. 

 

 

It’s possible Fairfax has shortened the duration of its reserves which would dampen the impact of higher rates on discounting. Perhaps another way to defer earnings. 

[Maverick47]

38 minutes ago, Viking said:

 

@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.

 

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

 

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

Personally, as sort of an armchair novice investor, I have often focused in a first pass view of investments, on multiples of earnings per share, the P/E ratio.  That works fairly well for investments that have relatively stable, predictable earnings, similar to interest paying bonds, and is less valuable when earnings are volatile, and occasionally even negative, or when earnings growth prospects are extremely difficult to estimate.

 

But for a mature insurance company, or even a conglomerate like Berkshire, it can be a helpful starting point.  For example, a hybrid view of valuing Berkshire might look at the market value of the equity portfolio, the cash, and then add to those two pieces a multiple of normalized earnings.  Say that equities are worth $300 billion, cash is at $380 billion, and the normalized projected operating earnings of all of the subsidiaries are estimated as $48 billion per year, or $4 billion per month.  What multiple would you be willing to pay for those earnings is something you have to determine individually.  Compared to risk free bond rates of say 5%, you’d typically want something higher than that, perhaps 8%, for a PE multiplier of the inverse, or 12.5.

 

That would value the operating subs at $600 billion, and added to the equity and cash, you’d get  an estimated total intrinsic value of $1.28 trillion compared to a recent market cap of $1.06 trillion.  So a bit of a margin of safety at current prices.  Of course you might divide the operating subs into categories, and assign typical PE multipliers used by the market for similar industries…say about 30 for the railroad and something less than 10 for the electric utilities, but that’s too far into the weeds for me personally on a quick first pass.

 

Looking at the history of earnings per share for Fairfax prior to about 2018, this was all over the place, with not a very predictable pattern.  But since then, we’ve had a good track record of earnings per share well above $100, and my own expectation would be closer to the $200 level and above going forward, particularly in light of the impact of share buybacks on the earnings per share going forward and the interest rates locked in for the bond portfolio.

 

Ignoring the cash and market value of equity piece of the Berkshire approach, and using the same 12.5 multiplier for an estimated earnings per share value of $200 gives me an intrinsic value per share of $2500 US.  If we were to add in cash and equity, I would not be at all surprised to see a result closer to $3,000 per share US, similar to @SafetyinNumbers preference for using a 15 multiplier to an estimated EPS of $200.

 

With management retiring shares at less than US $1700 a share recently, I am confident that they are buying back well below any reasonable estimate of intrinsic value.