Fairfax 2025

[Hoodlum]

We need to keep in mind that they mainly only sold their very long 30 year bonds.  It doesn't take much selling of the very long duration to impact the average duration.  Now of course we could say that they should have bought more 3-5 duration treasuries in place of the long bonds.   I think what we found out during Q2 was that duration was already short before Q2, outside of the 30 year.   

 

Another thing to keep in mind is that if there is a slow down in the economy, then there will be more opportunities on the Corporate bond side.  We have already seen some good yields with KW and Vacatia, and Fairfax may be looking at others as well.  The yields during Q2 increased slightly to 5.1%, so the shorter duration doesn't seem to be impacting income from bonds, likely due to the increase in corporate bonds.

[gfp]

The average duration figures were also pulled down by the extremely short duration of the large KW construction loan portfolio.  Very high yielding but very short in official duration

[LC]

9 hours ago, TwoCitiesCapital said:

But I'm looking at buying small businesses in my local area - and a number of the bars that I've been looking through saw YoY declines of 20+% in 2024 sales. 2025 is tracking similarly for them. 

 

Good luck. Curious what attracts you to owning a bar/bars? How are you thinking about foot traffic decreasing? 

[LC]

9 hours ago, Txvestor said:

Once again, I'm not the expert.

They announced it fairly recently within the last Q or so. 
On the ground, I can sense a significant economic slowdown. If the economy enters a recession and long bond yields plummet, this will be a totally unforced error. Macro betting and occasional own goals seems to be a part of their DNA. 
I'm not even saying they should extend out too far out onto the yield curve. Matching their liabilities would be reasonable.
4yrs was perfect. In case of a recession it would factor in time for a recovery well ahead of the next presidential election cycle. 

 

I agree with this. I thought reducing duration was a mistake. Not just based on my idea of future rate movements (I think falling rates are more likely than rising rates), but also just practically - 4yr duration provides a lot more earnings transparency and stability, which 1) the market will reward with a stronger valuation and 2) will give them opportunity to reinvest that cash flow (vs reinvestment of the short duration principal)

[glider3834]

 

9 hours ago, Hoodlum said:

We need to keep in mind that they mainly only sold their very long 30 year bonds.  It doesn't take much selling of the very long duration to impact the average duration.  Now of course we could say that they should have bought more 3-5 duration treasuries in place of the long bonds.   I think what we found out during Q2 was that duration was already short before Q2, outside of the 30 year.   

 

Another thing to keep in mind is that if there is a slow down in the economy, then there will be more opportunities on the Corporate bond side.  We have already seen some good yields with KW and Vacatia, and Fairfax may be looking at others as well.  The yields during Q2 increased slightly to 5.1%, so the shorter duration doesn't seem to be impacting income from bonds, likely due to the increase in corporate bonds.

yes buying 28-30 year treasuries in Q3'24, helped bump up duration - I question whether this was more of a short term trade, given their insurance liabilities have historically been in the 3.8 year area and they ultimately sold these very long dated treasuries in around 12 mths. 

 

YoY yield remains at 5.1% & duration has gone from 2.7 yrs in Q2'24 to 2.4 yrs in Q2'25 on a YoY basis so there is a shortening of 0.3 yrs on duration - I don't see that as a wholesale shift in positioning (if we ignore the 28-30 year treasury moves) although we are on shorter side of duration at 2.4 yrs in Q2'25

 

Wade Burton, President and Chief Investment Officer, Hambla Watsa: 

'We continue to stay on the shorter side of duration as we watch inflation and the Fed actions.'

(Q2 2025 call transcript)

 

Maybe Q3'25 will give us more insight here - I do think they have been pretty consistent they want to protect the balance sheet for lower interest rates or higher interest rates & their thinking around inflation risk may well have also been impacted by tariffs in Q2'25.

 

Q2'24 - 2.7 yrs (5.1% yield)

Q3'24 - 3.5 yrs (4.7% yield) - 28-30 year treasuries purchased & duration up by 0.8 yrs

 

Q1'25 -  3.3 yrs (5.1% yield)

Q2'25 - 2.4 yrs (5.1% yield) - 28-30 year treasuries sold & duration down by 0.9 yrs

 

Given Fairfax has a very large equity portfolio that is basically a long duration asset plus they need liquidity to support TRS etc, I would expect they would generally want to take a shorter bias on fixed income, than say a peer that has minimal equities and a much larger weighting in fixed income. 

 

I agree on your corporate bond opportunity comments and Fairfax have been very constructive in their ability to structure debt investments with equity upside features and having dry powder in the form of short term liquidity for these sorts of opportunities makes sense.

 

 

Edited October 3, 2025 by glider3834

[djokovic1]

@glider3834 thanks for breaking down their actions over the past on the duration and yield side. Useful.

I generally agree with the sentiment that locking in these rates for 4 years (rather than 2) makes a lot of sense. In other words, I would much rather lock in currently yields for longer because it still produces amazing ROE and live with the potential risk of higher rates. Because the downside risk of exposing yourself to lower rates is much more, as the ROE and earnings power drops significantly. 

But its only a quarter and we will find out more in the Q3 results, if not will aim to ask them for their logic on the conference call.

[Hoodlum]

Northbridge and other Canadian insurers that are a part of Fairfax, have been upgraded by AM Best. 
 

https://news.ambest.com/pr/PressContent.aspx?refnum=36562&altsrc=2

[Hamburg Investor]

The Combined Ratios of Fairfax Insurance Business over time: From bad to good!

I finally managed to use AI to turn some tables into graphs (don't ask!). I gathered all the data for Fairfax, Markel from annual reports (no guarantee that it is always correct) and the rest from an extended internet research. Here are the graphs and the most important findings:

1. Combined ratio of Fairfax vs. Markel over the years

First, let's take a look at the CR comparison:

 

What stands out? Markel was significantly better than Fairfax for a long time, but that has changed over the years. Since around 2013, both lines have been very close to each other, and in recent years Fairfax has even become slightly more profitable than Markel. I bought both companies around 2011 first; my expectation was, that Markel would stay the winner maybe even as long as forever. So for me, this is spectacular, looking back 14 years later and watching such results. The overall trend is even more evident in the 5-year average:

 

Fine. So Fairfax has been about as good as Markel since around 2012 or 2015, after being about 5 percentage points worse than Markel in many years. The only convergence of the lines around 2001 (slightly later in the 5-year chart) is related to 11 September; so we can safely dismiss that as a one-off event. 

 

What does this tell us? Could it be that Fairfax has become significantly more profitable overall in structural terms? After all, we are not comparing Fairfax CRs with average longterm CR figures here; rather, we are comparing with another company, that is exposed to the same trends (e. g. soft market, hard market).

2. Let's take it one step further: Let's compare Fairfax with the P&C US industry

 

 

Here, too, we see that Fairfax only briefly outperformed the market after 11 September 2001; apart from that, Fairfax rarely matched the market's CR until 2011; often, the CR was much worse, sometimes even 7 percentage points. This changed around 2011. From 2011 to around 2016, Fairfax was significantly better in some cases, and has been since 2021 as well. In between, Fairfax roughly mirrors the market. This is even more evident when you add up how Fairfax has performed relative to the market over the years:

 

To explain the red dotted line: the right X-axis shows Fairfax's cumulative overserving over time in percentage points of the combined ratio.

So the red dots show, that Fairfax had underperformed the P&C US Market by a total of almost 30 percentage points between 2000 and 2011 (i.e. slightly more than 2.5% per year). Since then, the red dotted line has risen steeply, with a pause only between 2016 and 2020, ending at around 20 percentage points of outperformance. In the 13 years since 2011, Fairfax has thus made up a total of 50 percentage points (from minus 30 to plus 20), or just under 4% per year.
 

Bottom Line: What a difference! Until 2011, Fairfax lost 2.5% percentage points per year against the average US P&C company, but since then Fairfax has gained 4%. In other words, Fairfax's insurance business (which is not only US) has become about 6.5% percentage points more profitable in relation to the market after 2011 (compared with the timeframe 2000 until 2011 ).
 

Why? It has already been mentioned here many times: Andy Barnard has been president of Fairfax's insurance business since 2011. We can only be grateful to him!!!
 

Are Fairfax's good CRs here to stay? Compared to the market, I would say: it looks that way. Is Fairfax's insurance business independent of general market movements, of soft and hard markets? Absolutely not – the curves do not show this.

 

One more note: this comparison is flawed! Fairfax has (increasingly) insurance business outside the US. And yet I am comparing it with the P&C US market – which is methodologically completely inaccurate.

 

In my view, however, it is still the best comparison (after all, the US market is still the largest for Fairfax) if you don't want to go to enormous effort. My only goal is to gain a rough initial overview – I consider this approach to be legitimate for that (and only that!) purpose.

Edited October 3, 2025 by Hamburg Investor

[glider3834]

9 hours ago, djokovic1 said:

@glider3834 thanks for breaking down their actions over the past on the duration and yield side. Useful.

I generally agree with the sentiment that locking in these rates for 4 years (rather than 2) makes a lot of sense. In other words, I would much rather lock in currently yields for longer because it still produces amazing ROE and live with the potential risk of higher rates. Because the downside risk of exposing yourself to lower rates is much more, as the ROE and earnings power drops significantly. 

But its only a quarter and we will find out more in the Q3 results, if not will aim to ask them for their logic on the conference call.

[glider3834]

https://www.carriermanagement.com/news/2025/09/10/279329.htm

 

https://www.carriermanagement.com/news/2015/03/23/137123.htm

 

 

 

 

 

Edited October 4, 2025 by glider3834

[Santayana]

12 hours ago, glider3834 said:

In other words, I would much rather lock in currently yields for longer because it still produces amazing ROE and live with the potential risk of higher rates. Because the downside risk of exposing yourself to lower rates is much more, as the ROE and earnings power drops significantly. 

I think you are underestimating the impact that higher rates would have on the capital base if duration were longer.  Remember capital loss leads directly to not being able to write as much in premiums.   At least in my mind, capital loss is a much bigger risk than just having a lower ROE for a bit.

[djokovic1]

Fair point but let me phrase it differently then to make my point clearer.

 

At these interest rates (which provide for very strong ROE’s), I would try to match asset duration to liability duration, rather than making a macro bet by having a shorter duration of assets than liabilities.

[petec]

On 9/30/2025 at 9:15 PM, vinod1 said:

 

TIPS are widely traded and available in size. There is no real market constraint if Fairfax wants to hold a portfolio of inflation-linked bonds that match liability duration. The fixed income earnings would then fluctuate with inflation rates, compared to having a very predictable 2-3 year fixed income earnings now.

 

Vinod

 

Agreed. But the logical output of that argument is that FFH (and other insurecos) always take a macro view when they invest in nominal fixed income securities. Not that they're taking a macro view when duration spreads don't pay them to go long.

[petec]

On 9/30/2025 at 11:37 PM, TwoCitiesCapital said:

You'll have underperformed, yes. But in the world of Fairfax, where duration of liabilities is ~3-4 years, also an inaccurate example.

 

Yes of course. Sorry - I should have been clearer. I am not suggesting this example is/was reflective of FFH's position. I just find extreme examples useful for thinking through principles.

 

The point is you can seriously impair your balance sheet if you take duration risk without being paid for it. 

 

On 9/30/2025 at 11:37 PM, TwoCitiesCapital said:

If it was about price, why was it made in 2016, after an election, when rates were higher than they were in 2015 before the election? The election WAS the catalyst - ergo it was a macro bet. 

 

This is your strongest argument, imho.

[Santayana]

30 minutes ago, petec said:

The point is you can seriously impair your balance sheet if you take duration risk without being paid for it. 

+1.  And in the case of Fairfax, that impaired balance sheet would lead to a smaller insurance business right at the time when high yields would make that business even more profitable.

[djokovic1]

1 hour ago, petec said:

The point is you can seriously impair your balance sheet if you take duration risk without being paid for it. 

 

Sorry I don't understand. They are taking a macro bet and making a call about future interest rates by having their asset duration well below liability duration. I don't think there is a debate about that?

 

Whether they are right or not in having that interest rate view, we can debate about.

[Santayana]

So reducing duration is a macro bet, but extending it is not?  I think what they are trying to do is keep duration at a level where they are as agnostic as possible to whatever future changes may happen.

[TwoCitiesCapital]

2 hours ago, petec said:

 

The point is you can seriously impair your balance sheet if you take duration risk without being paid for it. 

 

Yes - in a "technical" sense. I would argue it depends on how we look at/measure it. 

 

Technically, if it matches the liability, the liability is also fluctuating with the rates and no real impairment occurs if perfectly matched. Now, there's risks to this because you're never perfectly matched and until recently the liability fluctuation wasn't captured on the balance sheet so you only had the asset side swinging wildly. 

 

But now that the liability impact also flows through, from a reporting standard little impairment would occur as you'd expect movement in the assets and liabilities to largely offset on another. 

 

There might be some debate on impacts to liquidity and the ability to sell the bonds vs realize a loss and etc. but from an asset/liability view (i.e. balance sheet), you should have little impairment matching your liability. The impairment comes for mismatches and being wrong which is where Fairfax was playing from 2016 - 2021 until 2022 - 2024 turned it around. 

[Munger_Disciple]

Good explanation: They are bond "value" investors. When the yields were puny (and hence risks were great), they reduced duration bigly. And then when ST yields rose, they slightly extended duration to 2-3 years to lock in what Bradstreet thought were good rates. 

 

Less generous explanation: They are just bond & macro trading junkies . 

 

I don't think they try to match assets to liabilities like MKL does but I could be wrong on that. 

Edited October 7, 2025 by Munger_Disciple

[djokovic1]

35 minutes ago, Santayana said:

So reducing duration is a macro bet, but extending it is not? 

I think you missed the part when I said asset duration is significantly below liability duration. Any big mismatch is a macro bet.

I am not arguing that they should always match asset duration to liabilities. They did a great job in 2021 going short duration when rates were 0. But right now I don't see risk reward of being short duration. We could just say, "In Bradstreet we trust..." and leave it there, but I would like to go deeper.

[Crip1]

1 hour ago, Munger_Disciple said:

I don't think they try to match assets to liabilities like MKL does but I could be wrong on that. 

You beat me to it...was going to add that.

 

-Crip

[Munger_Disciple]

2 hours ago, Crip1 said:

You beat me to it...was going to add that.

 

-Crip

 

[Hoodlum]

Here are a couple quotes from Q3 2023 and Q3 2024 Conference Call transcripts regarding Asset and Liability matching.   They is certainly not trying to match them up.

 

"you're right that if interest rates are flat, premium is flat, more or less, the discount you put on the current year is more or less offset on the unwinding of the discount from the previous years. And then as if by chance, the bond maturity is very similar to the duration on our liabilities, that matches as well. When those relationships change, you're going to see movements."

 

"And if you look at, you know, our liabilities, it's relatively close. We don't match on purpose, but where we sit today, our liability duration is close to our asset, our asset duration."

[Santayana]

4 hours ago, djokovic1 said:

I think you missed the part when I said asset duration is significantly below liability duration. Any big mismatch is a macro bet.

What exactly is the macro bet you think they are making?  I don't believe they are trying to make any prediction about future rates, and think they are trying to position themselves in a way that reflects that they don't know what rates will do in the future.  To me this is the opposite of a macro bet.

[petec]

12 hours ago, Santayana said:

+1.  And in the case of Fairfax, that impaired balance sheet would lead to a smaller insurance business right at the time when high yields would make that business even more profitable.

 

Absolutely.