Fairfax 2026

[SafetyinNumbers]

7 hours ago, This2ShallPass said:

So a double in BV in 5 years and 3x multiple, stock price of ~US$7500! I certainly won't be complaining if that happens:)

 

One risk to multiple rerate would be a once in 50 year (or worse 100) catastrophe year. How are people thinking about losses in a big cat event (is it still the 1% FF share of overall loss)? All these benign years mean the bad year is just around the corner. 

 

Last year wasn’t that benign, the California wildfires were a big hit in Q1. Cat losses were 4.8 points this year vs 4.5 last year. Cat has shrunk as part of total premiums so I don’t think the risk of a loss in underwriting on a full year basis is very high. If it does happen, they are very well positioned to take advantage of the hard market to follow. I’m looking forward to an update of the climate report graph.

 

[SafetyinNumbers]

8 hours ago, This2ShallPass said:

One risk to multiple rerate would be a once in 50 year (or worse 100) catastrophe year.

Actually, this might be the catalyst for a multiple rerate. The thinking is that a big cat year like that would harden the market and trigger big growth in gross premiums written. The same momentum investors that have been selling since June flood back in to the stock except, there are a lot less shares available as passive indices and the company have reduced supply while they were chasing returns elsewhere. 

[Viking]

9 hours ago, Maverick47 said:

Unless the mix of business is comparable, both by line of business and by geography, simply comparing comparing CRs between companies and assuming that the lower CR is proof of a better underwriter is fallacious.

 

Short tail lines such as auto typically have target ratios of about 95 or 96.  GEICO and Progressive are predominantly auto carriers, so outperforming their targets significantly is a signal that they overshot on their rate levels.  Allstate and Chubb have a decent amount of short tail (but cat exposed) homeowners business.  The target CR for that is probably around 90.  In a hurricane free year such as 2025, one would expect the CR for that portion of their business would be in the 80’s, since their hurricane losses would be nil.

 

Fairfax is going to be more commercially driven in Northbridge and Crum and Forster.  Casualty claims in commercial lines have a long tail so a somewhat higher CR for that is acceptable as a target compared to the shorter tailed commercial property claims.  A weighted target of 95 for that combined commercial casualty and property business is probably fine.  
 

The long tailed workers comp business written by Zenith can probably afford a target CR of 100 or even a bit higher and still make a 15% ROE given the ability to invest the substantial float produced by the loss reserves in that line.  Allied World,  Odyssey, Brit and Ki are reinsurers, not particularly comparable to the direct insurance companies in this chart.  Probably would be best to compare their CRs to reinsurers such as Ren Re, Munich Re, Gen Re, National Indemnity, etc.  But here too, the mix of reinsured lines of business will vary between all these companies so makes comparisons difficult.

 

Bottom line, I would suggest focusing just on the company’s own target CR.  Fairfax indicates they currently expect normalized underwriting profit of $1.5 billion a year.  Their 2025 underwriting profit of $1.8 billion suggests they outperformed their internal target, particularly in a year with higher cat losses than 2024.  That likely means that their risk adjusted worldwide target CR is closer to 94 or 95 than the 93.0 they reported.

 

In a heavy cat year, Fairfax believes they can handle losses within their underwriting profit margin, so a 100 CR for such a year is what we might expect.  
 

Personally, I think that for a large, globally diversified insurer such as Fairfax, a CR of 93 in an average Cat year is an indication that they are an excellent underwriter.  Something substantially higher than that would suggest they let their rate levels fall behind their loss trends, and will have a lot of rate increases to file for to fix the problem, as well as likely needing to restrict new business and cancel existing poor business to try to fix things.  None of these actions would bode well for the future stability and desired profitable growth of their insurance exposures.  Conversely, CRs much better than 93, unless explained by a light catastrophe year, could be a sign that the the business has been overpriced, and could mean that price levels are higher than competitors, and that future new business growth could be restricted.

Finally, consistent reserve releases shows that the reserving practices are biased toward the conservative end of the spectrum, which is another sign of a well run insurer.

 

Personally, comparing Fairfax’s current (past 10 years) insurance underwriting results to the ideal of what a well run insurer should produce, I can’t find anything to complain about.  And I say this as a fairly recently retired actuary for an insurer, that over my 30+ year career,  made all of the mistakes I allude to above, and more.  I’m much happier to own the current version of Fairfax in my retirement than I was to be employed by a company that wasn’t as well run during my working years.

@Maverick47, great overview. Thanks for posting. 

[Viking]

2 hours ago, SafetyinNumbers said:

Actually, this might be the catalyst for a multiple rerate. The thinking is that a big cat year like that would harden the market and trigger big growth in gross premiums written. The same momentum investors that have been selling since June flood back in to the stock except, there are a lot less shares available as passive indices and the company have reduced supply while they were chasing returns elsewhere. 

I agree. When trying to figure out the impact of something on Fairfax (bad cat year in this example) the timeframe used is key. If the timeframe is next quarter, there are lots of ‘risks’. If the timeframe used is the next 2 or 3 years the number of risks shrink dramatically. And some risks in the short term flip to opportunities in the longer term. It is very counterintuitive.

 

Of course, this then layers into whether Fairfax is a speculation (short term hold) or an investment (long term hold).

Edited February 21 by Viking

[73 Reds]

9 hours ago, Maverick47 said:

Unless the mix of business is comparable, both by line of business and by geography, simply comparing comparing CRs between companies and assuming that the lower CR is proof of a better underwriter is fallacious.

 

Short tail lines such as auto typically have target ratios of about 95 or 96.  GEICO and Progressive are predominantly auto carriers, so outperforming their targets significantly is a signal that they overshot on their rate levels.  Allstate and Chubb have a decent amount of short tail (but cat exposed) homeowners business.  The target CR for that is probably around 90.  In a hurricane free year such as 2025, one would expect the CR for that portion of their business would be in the 80’s, since their hurricane losses would be nil.

 

Fairfax is going to be more commercially driven in Northbridge and Crum and Forster.  Casualty claims in commercial lines have a long tail so a somewhat higher CR for that is acceptable as a target compared to the shorter tailed commercial property claims.  A weighted target of 95 for that combined commercial casualty and property business is probably fine.  
 

The long tailed workers comp business written by Zenith can probably afford a target CR of 100 or even a bit higher and still make a 15% ROE given the ability to invest the substantial float produced by the loss reserves in that line.  Allied World,  Odyssey, Brit and Ki are reinsurers, not particularly comparable to the direct insurance companies in this chart.  Probably would be best to compare their CRs to reinsurers such as Ren Re, Munich Re, Gen Re, National Indemnity, etc.  But here too, the mix of reinsured lines of business will vary between all these companies so makes comparisons difficult.

 

Bottom line, I would suggest focusing just on the company’s own target CR.  Fairfax indicates they currently expect normalized underwriting profit of $1.5 billion a year.  Their 2025 underwriting profit of $1.8 billion suggests they outperformed their internal target, particularly in a year with higher cat losses than 2024.  That likely means that their risk adjusted worldwide target CR is closer to 94 or 95 than the 93.0 they reported.

 

In a heavy cat year, Fairfax believes they can handle losses within their underwriting profit margin, so a 100 CR for such a year is what we might expect.  
 

Personally, I think that for a large, globally diversified insurer such as Fairfax, a CR of 93 in an average Cat year is an indication that they are an excellent underwriter.  Something substantially higher than that would suggest they let their rate levels fall behind their loss trends, and will have a lot of rate increases to file for to fix the problem, as well as likely needing to restrict new business and cancel existing poor business to try to fix things.  None of these actions would bode well for the future stability and desired profitable growth of their insurance exposures.  Conversely, CRs much better than 93, unless explained by a light catastrophe year, could be a sign that the the business has been overpriced, and could mean that price levels are higher than competitors, and that future new business growth could be restricted.

Finally, consistent reserve releases shows that the reserving practices are biased toward the conservative end of the spectrum, which is another sign of a well run insurer.

 

Personally, comparing Fairfax’s current (past 10 years) insurance underwriting results to the ideal of what a well run insurer should produce, I can’t find anything to complain about.  And I say this as a fairly recently retired actuary for an insurer, that over my 30+ year career,  made all of the mistakes I allude to above, and more.  I’m much happier to own the current version of Fairfax in my retirement than I was to be employed by a company that wasn’t as well run during my working years.

@Maverick47 That's quite a testament to Fairfax's underwriting - thanks for an insurance insider POV.  I never really thought about how there could be a CR "sweet spot" so appreciate you pointing that out.

[djokovic1]

15 hours ago, Maverick47 said:

Unless the mix of business is comparable, both by line of business and by geography, simply comparing comparing CRs between companies and assuming that the lower CR is proof of a better underwriter is fallacious.

 

Short tail lines such as auto typically have target ratios of about 95 or 96.  GEICO and Progressive are predominantly auto carriers, so outperforming their targets significantly is a signal that they overshot on their rate levels.  Allstate and Chubb have a decent amount of short tail (but cat exposed) homeowners business.  The target CR for that is probably around 90.  In a hurricane free year such as 2025, one would expect the CR for that portion of their business would be in the 80’s, since their hurricane losses would be nil.

 

Fairfax is going to be more commercially driven in Northbridge and Crum and Forster.  Casualty claims in commercial lines have a long tail so a somewhat higher CR for that is acceptable as a target compared to the shorter tailed commercial property claims.  A weighted target of 95 for that combined commercial casualty and property business is probably fine.  
 

The long tailed workers comp business written by Zenith can probably afford a target CR of 100 or even a bit higher and still make a 15% ROE given the ability to invest the substantial float produced by the loss reserves in that line.  Allied World,  Odyssey, Brit and Ki are reinsurers, not particularly comparable to the direct insurance companies in this chart.  Probably would be best to compare their CRs to reinsurers such as Ren Re, Munich Re, Gen Re, National Indemnity, etc.  But here too, the mix of reinsured lines of business will vary between all these companies so makes comparisons difficult.

 

Bottom line, I would suggest focusing just on the company’s own target CR.  Fairfax indicates they currently expect normalized underwriting profit of $1.5 billion a year.  Their 2025 underwriting profit of $1.8 billion suggests they outperformed their internal target, particularly in a year with higher cat losses than 2024.  That likely means that their risk adjusted worldwide target CR is closer to 94 or 95 than the 93.0 they reported.

 

In a heavy cat year, Fairfax believes they can handle losses within their underwriting profit margin, so a 100 CR for such a year is what we might expect.  
 

Personally, I think that for a large, globally diversified insurer such as Fairfax, a CR of 93 in an average Cat year is an indication that they are an excellent underwriter.  Something substantially higher than that would suggest they let their rate levels fall behind their loss trends, and will have a lot of rate increases to file for to fix the problem, as well as likely needing to restrict new business and cancel existing poor business to try to fix things.  None of these actions would bode well for the future stability and desired profitable growth of their insurance exposures.  Conversely, CRs much better than 93, unless explained by a light catastrophe year, could be a sign that the the business has been overpriced, and could mean that price levels are higher than competitors, and that future new business growth could be restricted.

Finally, consistent reserve releases shows that the reserving practices are biased toward the conservative end of the spectrum, which is another sign of a well run insurer.

 

Personally, comparing Fairfax’s current (past 10 years) insurance underwriting results to the ideal of what a well run insurer should produce, I can’t find anything to complain about.  And I say this as a fairly recently retired actuary for an insurer, that over my 30+ year career,  made all of the mistakes I allude to above, and more.  I’m much happier to own the current version of Fairfax in my retirement than I was to be employed by a company that wasn’t as well run during my working years.

 

Love this post. Thank you for your expert view. This has been my intuition too but you laid it out perfectly. I.e you rationally should target higher combined ratios the more float you get from the business and vice versa.

[Txvestor]

12 hours ago, SafetyinNumbers said:

Anecdotally, I have noticed the combined ratio of an acquisition goes up after Fairfax acquires them. I think this is because they are much more conservative on reserving so there is some padding required. The street seems to assume it means they did a bad deal but I think it’s classic Fairfax income deferral. 

More lately yes. Early on they did do some bad insurance acquisitions.  

[Munger_Disciple]

On 2/20/2026 at 12:44 PM, TwoCitiesCapital said:

That's a big if as multiples expand and contract regularly and Fairfax is still subject to an economic/interest rate/inflation cycle.

 

Which is why looking through to the earnings is correct approach IMO. 

 

 The investor earns a 5% return in year 1 by that measure. Over time, if Fairfax can continue to compound at 15% into perpetuity, the investor's return will asymptotically approach 15% regardless of entry/exit multiple....but we're talking multiple decades required to converge there. 

 

This is why starting valuation matters so much - because expansion/contraction of multiples WILL happen. And I'd rather expansion be a tailwind then contraction being a headwind. Buying at 1-1.5x probably provides for that upside optionality on expansion tailwinds. Buying at 3x is probably opening you up to contraction headwinds.

 

Even though both investors will approach 15% annualized on a century long time line, the guy buying at 1.5x has a lot more routes to get there and get there more quickly than the 3x buyer. 

 

+1

 

Good post. I believe it can be summarized as seeking "margin of safety". 

Edited February 22 by Munger_Disciple

[SafetyinNumbers]

3 hours ago, Txvestor said:

More lately yes. Early on they did do some bad insurance acquisitions.  

They didn’t do bad insurance acquisitions. They used expensive stock to buy cheap float. They knew losses were coming but thought they had enough margin of safety.

[djokovic1]

@Maverick47 why does short tail insurance like auto target a relatively high CR ~95-96 even though you get no float?

 

Is that because it’s more competitive and mostly a volume game?

[Maverick47]

4 hours ago, djokovic1 said:

@Maverick47 why does short tail insurance like auto target a relatively high CR ~95-96 even though you get no float?

 

Is that because it’s more competitive and mostly a volume game?

@djokovic1 Good question.  Most folks use 95 or 96 as an approximation for the CR target for personal auto insurance coverages as a whole.  In actuality, a typical auto policy is a combination of a number of sub-coverages which are often aggregated into two main buckets — liability and physical damage.  By the way, the two largest personal auto insurance carriers, Progressive and GEICO, both have publicly disclosed over the years that their own auto CR targets are 96.
 

The physical damages coverages are the ones that tend to be resolved very quickly, so produce little in the way of loss and adjustment reserve float (there will be some unearned premium that will add a bit to the investable float, but that’s not terribly significant ).   However, claims for these sorts of losses also tend to be quite predictable and stable over time, and regulators don’t typically require an insurer to allocate a lot of capital in proportion to physical damages premium.  Let’s say a dollar of shareholder equity is required to support $5 of physical damages premium.  The company pays tax on investment and underwriting income of 20%, and for simplicity sake, let’s say only the single dollar of allocated shareholder equity is available to be invested at a pretax rate of 5%.  Say also, that given the minimal float available, that a CR target for physical damage is 97.  The company earns 5 cents of pretax return on the invested allocated equity dollar and 15 cents of pretax underwriting profit on the $5 of premium.  The pretax return of $0.20 per dollar of equity ends up being $0.16 after tax, for an after tax return on equity of 16%.
 

Bodily injury liability claims do produce a much greater amount of longer tailed loss and adjustment reserves (and float) and those claims take much longer to resolve than the relatively simpler physical damage claims. But regulators and the company’s own assessment of risk will also result in companies choosing to hold a greater amount of equity to support these claims, which tend to be more volatile than the simpler physical damage losses.  Let’s say a dollar of shareholder equity is selected by the insurer to support every $2 of their auto liability premium, and that the investable float is equal to $1 for every $2 of premium.  If the company wants to calculate the pre-tax CR target required to produce a 16% after tax ROE for this portion of auto premium, they would first realize that investing the $2 of equity and float would generate 10 cents of pretax income, and 8 cents after tax.  Then they would require the other 8 cents of after tax income to come from the underwriting profit on the $2 of premium.  That would be 4 cents after tax on each of the two dollars of premium, and 4 cents / (1 - 0.20) = 5 cents of pre-tax underwriting profit for every dollar of auto liability premium.  That means a target CR of 95 for the auto liability coverages.

 

With the premium volumes of liability and physical damages coverages being fairly close to each other, many companies managements simply average the two separate CR targets of 97 and 95 together for a combined 96 CR target across all auto coverages instead.

Bottom line, there are more variables considered by management besides the amount of float produced by a line of insurance business that go into setting a CR target.  One very important one is how much equity is required to support a dollar of written premium.  Generally speaking, auto physical damage is considered one of the least risky lines of business, so requires the least amount of supporting equity.  Homeowners insurance, in my opinion, on balance, is riskier than auto liability.  Depending on the type and amount of catastrophe exposure, in some geographies it could range from a bit less risky than auto liability, to a whole lot riskier.  
 

Some areas exposed to hurricanes, tornadoes, wildfires, and fires following earhquakes (think San Francisco in 1906) could easily require a dollar of equity (and maybe even more) to support every single dollar of homeowners premium ( US states such as California or Florida are examples of geographies where this could be true).  
 

Since most homeowners claims don’t take as long to resolve as auto liability, the float per dollar of home premium is not as great as it is for auto liability.  The after tax investment return then is going to be less than the 8 cents it was for auto liability, and the pre-tax return on underwriting only comes from the single dollar of homeowners premium instead of the two dollars of premium as was the case with auto liability.  So instead of 5 cents of target pre-tax underwriting per dollar of auto liability premium, the target pre-tax underwriting profit per dollar of homeowners premium in a cat exposed state like California or Texas would be even greater than 10 cents, for a CR homeowners target in those states in the mid to high 80’s.

 

Finally, catastrophe lines of business such as earthquake are probably the riskiest of all. 
 

If I was pricing that line of business on a direct basis, ignoring the effects of catastrophe reinsurance that every prudent company would purchase,  I might assume that a dollar of equity would be required to support 30 cents of earthquake premium.  Most years there are no earthquake losses or reserves available to produce investable float, so a 16% after tax ROE would come from first investing the dollar of equity for 4 cents of after tax investment return.  Then we would need the remaining 12 cents of after tax return (15 cents pre-tax) to come from the $0.30 of premium, for a CR target of 50%.

 

In practice, every primary insurer is going to purchase a property catastrophe reinsurance treaty from reinsurers including any number of Fairfax reinsurers.  Then the primary company wouldn’t need to put as much of their own equity at risk, because it is much more cost effective to purchase cat reinsurance, and then the CR target for the cat exposed earthquake line wouldn’t need to be as low as the 50% noted above.

 

Sorry for the long-winded response!  But now you also have a flavor for why capital is required to be retained or even perhaps added to insurers when they would like to grow premium volumes substantially during hard markets, such as was the case with Fairfax’s insurance companies in the early part of this decade.  Regulators and rating agencies tend to look askance at companies whose rate of premium growth begins to outstrip the growth in their supporting equity. That’s what happened to a lot of companies that didn’t manage to protect their balance sheets as well as Fairfax did when interest rates began to rise.  Many of them had to either restrict growth, or find ways to shed premium in order to avoid ratings downgrades as their shareholder equity levels either stagnated or fell.

 

Pretty impressive that Fairfax managed their way through a period of substantial premium growth while also earning ratings upgrades for their insurance companies!

Edited February 22 by Maverick47

[UK]

1 hour ago, Maverick47 said:

Sorry for the long-winded response!

All this is very interesting! Please do not hesitate:)

[73 Reds]

2 hours ago, Maverick47 said:

@djokovic1 Good question.  Most folks use 95 or 96 as an approximation for the CR target for personal auto insurance coverages as a whole.  In actuality, a typical auto policy is a combination of a number of sub-coverages which are often aggregated into two main buckets — liability and physical damage.  By the way, the two largest personal auto insurance carriers, Progressive and GEICO, both have publicly disclosed over the years that their own auto CR targets are 96.
 

The physical damages coverages are the ones that tend to be resolved very quickly, so produce little in the way of loss and adjustment reserve float (there will be some unearned premium that will add a bit to the investable float, but that’s not terribly significant ).   However, claims for these sorts of losses also tend to be quite predictable and stable over time, and regulators don’t typically require an insurer to allocate a lot of capital in proportion to physical damages premium.  Let’s say a dollar of shareholder equity is required to support $5 of physical damages premium.  The company pays tax on investment and underwriting income of 20%, and for simplicity sake, let’s say only the single dollar of allocated shareholder equity is available to be invested at a pretax rate of 5%.  Say also, that given the minimal float available, that a CR target for physical damage is 97.  The company earns 5 cents of pretax return on the invested allocated equity dollar and 15 cents of pretax underwriting profit on the $5 of premium.  The pretax return of $0.20 per dollar of equity ends up being $0.16 after tax, for an after tax return on equity of 16%.
 

Bodily injury liability claims do produce a much greater amount of longer tailed loss and adjustment reserves (and float) and those claims take much longer to resolve than the relatively simpler physical damage claims. But regulators and the company’s own assessment of risk will also result in companies choosing to hold a greater amount of equity to support these claims, which tend to be more volatile than the simpler physical damage losses.  Let’s say a dollar of shareholder equity is selected by the insurer to support every $2 of their auto liability premium, and that the investable float is equal to $1 for every $2 of premium.  If the company wants to calculate the pre-tax CR target required to produce a 16% after tax ROE for this portion of auto premium, they would first realize that investing the $2 of equity and float would generate 10 cents of pretax income, and 8 cents after tax.  Then they would require the other 8 cents of after tax income to come from the underwriting profit on the $2 of premium.  That would be 4 cents after tax on each of the two dollars of premium, and 4 cents / (1 - 0.20) = 5 cents of pre-tax underwriting profit for every dollar of auto liability premium.  That means a target CR of 95 for the auto liability coverages.

 

With the premium volumes of liability and physical damages coverages being fairly close to each other, many companies managements simply average the two separate CR targets of 97 and 95 together for a combined 96 CR target across all auto coverages instead.

Bottom line, there are more variables considered by management besides the amount of float produced by a line of insurance business that go into setting a CR target.  One very important one is how much equity is required to support a dollar of written premium.  Generally speaking, auto physical damage is considered one of the least risky lines of business, so requires the least amount of supporting equity.  Homeowners insurance, in my opinion, on balance, is riskier than auto liability.  Depending on the type and amount of catastrophe exposure, in some geographies it could range from a bit less risky than auto liability, to a whole lot riskier.  
 

Some areas exposed to hurricanes, tornadoes, wildfires, and fires following earhquakes (think San Francisco in 1906) could easily require a dollar of equity (and maybe even more) to support every single dollar of homeowners premium ( US states such as California or Florida are examples of geographies where this could be true).  
 

Since most homeowners claims don’t take as long to resolve as auto liability, the float per dollar of home premium is not as great as it is for auto liability.  The after tax investment return then is going to be less than the 8 cents it was for auto liability, and the pre-tax return on underwriting only comes from the single dollar of homeowners premium instead of the two dollars of premium as was the case with auto liability.  So instead of 5 cents of target pre-tax underwriting per dollar of auto liability premium, the target pre-tax underwriting profit per dollar of homeowners premium in a cat exposed state like California or Texas would be even greater than 10 cents, for a CR homeowners target in those states in the mid to high 80’s.

 

Finally, catastrophe lines of business such as earthquake are probably the riskiest of all. 
 

If I was pricing that line of business on a direct basis, ignoring the effects of catastrophe reinsurance that every prudent company would purchase,  I might assume that a dollar of equity would be required to support 30 cents of earthquake premium.  Most years there are no earthquake losses or reserves available to produce investable float, so a 16% after tax ROE would come from first investing the dollar of equity for 4 cents of after tax investment return.  Then we would need the remaining 12 cents of after tax return (15 cents pre-tax) to come from the $0.30 of premium, for a CR target of 50%.

 

In practice, every primary insurer is going to purchase a property catastrophe reinsurance treaty from reinsurers including any number of Fairfax reinsurers.  Then the primary company wouldn’t need to put as much of their own equity at risk, because it is much more cost effective to purchase cat reinsurance, and then the CR target for the cat exposed earthquake line wouldn’t need to be as low as the 50% noted above.

 

Sorry for the long-winded response!  But now you also have a flavor for why capital is required to be retained or even perhaps added to insurers when they would like to grow premium volumes substantially during hard markets, such as was the case with Fairfax’s insurance companies in the early part of this decade.  Regulators and rating agencies tend to look askance at companies whose rate of premium growth begins to outstrip the growth in their supporting equity. That’s what happened to a lot of companies that didn’t manage to protect their balance sheets as well as Fairfax did when interest rates began to rise.  Many of them had to either restrict growth, or find ways to shed premium in order to avoid ratings downgrades as their shareholder equity levels either stagnated or fell.

 

Pretty impressive that Fairfax managed their way through a period of substantial premium growth while also earning ratings upgrades for their insurance companies!

Another thought-provoking post - really learning a lot.  If more capital reserves are needed for premium growth, why then do most insurers shy away from potentially higher investment returns through equity investments with at least some portion of their capital? 

[Hektor]

3 hours ago, Maverick47 said:

@djokovic1 Good question.  Most folks use 95 or 96 as an approximation for the CR target for personal auto insurance coverages as a whole.  In actuality, a typical auto policy is a combination of a number of sub-coverages which are often aggregated into two main buckets — liability and physical damage.  By the way, the two largest personal auto insurance carriers, Progressive and GEICO, both have publicly disclosed over the years that their own auto CR targets are 96.
 

The physical damages coverages are the ones that tend to be resolved very quickly, so produce little in the way of loss and adjustment reserve float (there will be some unearned premium that will add a bit to the investable float, but that’s not terribly significant ).   However, claims for these sorts of losses also tend to be quite predictable and stable over time, and regulators don’t typically require an insurer to allocate a lot of capital in proportion to physical damages premium.  Let’s say a dollar of shareholder equity is required to support $5 of physical damages premium.  The company pays tax on investment and underwriting income of 20%, and for simplicity sake, let’s say only the single dollar of allocated shareholder equity is available to be invested at a pretax rate of 5%.  Say also, that given the minimal float available, that a CR target for physical damage is 97.  The company earns 5 cents of pretax return on the invested allocated equity dollar and 15 cents of pretax underwriting profit on the $5 of premium.  The pretax return of $0.20 per dollar of equity ends up being $0.16 after tax, for an after tax return on equity of 16%.
 

Bodily injury liability claims do produce a much greater amount of longer tailed loss and adjustment reserves (and float) and those claims take much longer to resolve than the relatively simpler physical damage claims. But regulators and the company’s own assessment of risk will also result in companies choosing to hold a greater amount of equity to support these claims, which tend to be more volatile than the simpler physical damage losses.  Let’s say a dollar of shareholder equity is selected by the insurer to support every $2 of their auto liability premium, and that the investable float is equal to $1 for every $2 of premium.  If the company wants to calculate the pre-tax CR target required to produce a 16% after tax ROE for this portion of auto premium, they would first realize that investing the $2 of equity and float would generate 10 cents of pretax income, and 8 cents after tax.  Then they would require the other 8 cents of after tax income to come from the underwriting profit on the $2 of premium.  That would be 4 cents after tax on each of the two dollars of premium, and 4 cents / (1 - 0.20) = 5 cents of pre-tax underwriting profit for every dollar of auto liability premium.  That means a target CR of 95 for the auto liability coverages.

 

With the premium volumes of liability and physical damages coverages being fairly close to each other, many companies managements simply average the two separate CR targets of 97 and 95 together for a combined 96 CR target across all auto coverages instead.

Bottom line, there are more variables considered by management besides the amount of float produced by a line of insurance business that go into setting a CR target.  One very important one is how much equity is required to support a dollar of written premium.  Generally speaking, auto physical damage is considered one of the least risky lines of business, so requires the least amount of supporting equity.  Homeowners insurance, in my opinion, on balance, is riskier than auto liability.  Depending on the type and amount of catastrophe exposure, in some geographies it could range from a bit less risky than auto liability, to a whole lot riskier.  
 

Some areas exposed to hurricanes, tornadoes, wildfires, and fires following earhquakes (think San Francisco in 1906) could easily require a dollar of equity (and maybe even more) to support every single dollar of homeowners premium ( US states such as California or Florida are examples of geographies where this could be true).  
 

Since most homeowners claims don’t take as long to resolve as auto liability, the float per dollar of home premium is not as great as it is for auto liability.  The after tax investment return then is going to be less than the 8 cents it was for auto liability, and the pre-tax return on underwriting only comes from the single dollar of homeowners premium instead of the two dollars of premium as was the case with auto liability.  So instead of 5 cents of target pre-tax underwriting per dollar of auto liability premium, the target pre-tax underwriting profit per dollar of homeowners premium in a cat exposed state like California or Texas would be even greater than 10 cents, for a CR homeowners target in those states in the mid to high 80’s.

 

Finally, catastrophe lines of business such as earthquake are probably the riskiest of all. 
 

If I was pricing that line of business on a direct basis, ignoring the effects of catastrophe reinsurance that every prudent company would purchase,  I might assume that a dollar of equity would be required to support 30 cents of earthquake premium.  Most years there are no earthquake losses or reserves available to produce investable float, so a 16% after tax ROE would come from first investing the dollar of equity for 4 cents of after tax investment return.  Then we would need the remaining 12 cents of after tax return (15 cents pre-tax) to come from the $0.30 of premium, for a CR target of 50%.

 

In practice, every primary insurer is going to purchase a property catastrophe reinsurance treaty from reinsurers including any number of Fairfax reinsurers.  Then the primary company wouldn’t need to put as much of their own equity at risk, because it is much more cost effective to purchase cat reinsurance, and then the CR target for the cat exposed earthquake line wouldn’t need to be as low as the 50% noted above.

 

Sorry for the long-winded response!  But now you also have a flavor for why capital is required to be retained or even perhaps added to insurers when they would like to grow premium volumes substantially during hard markets, such as was the case with Fairfax’s insurance companies in the early part of this decade.  Regulators and rating agencies tend to look askance at companies whose rate of premium growth begins to outstrip the growth in their supporting equity. That’s what happened to a lot of companies that didn’t manage to protect their balance sheets as well as Fairfax did when interest rates began to rise.  Many of them had to either restrict growth, or find ways to shed premium in order to avoid ratings downgrades as their shareholder equity levels either stagnated or fell.

 

Pretty impressive that Fairfax managed their way through a period of substantial premium growth while also earning ratings upgrades for their insurance companies!

Thanks @Maverick47 for the posts. They add to the learning.

[SafetyinNumbers]

47 minutes ago, 73 Reds said:

Another thought-provoking post - really learning a lot.  If more capital reserves are needed for premium growth, why then do most insurers shy away from potentially higher investment returns through equity investments with at least some portion of their capital? 

I think it’s because equity returns are lumpy and less predictable. Most companies won’t even take risk on duration or composition of the fixed income portfolio because it introduces variability. 

[UK]

1 hour ago, 73 Reds said:

Another thought-provoking post - really learning a lot.  If more capital reserves are needed for premium growth, why then do most insurers shy away from potentially higher investment returns through equity investments with at least some portion of their capital? 

Dangerous for the quarterly guidance:))

[MMM20]

2 hours ago, SafetyinNumbers said:

I think it’s because equity returns are lumpy and less predictable. Most companies won’t even take risk on duration or composition of the fixed income portfolio because it introduces variability. 

Also the actuarial/ accountant mindset doesn’t translate directly to investing, and there are plenty of failed attempts to outsource the investing piece. It’s hard to get right and not blow up. It’s also been a pretty good call to just own bonds as interest rates declined for basically 4 decades, aka senior insurance execs whole careers.

 

Edited February 22 by MMM20

[roundball100]

11 hours ago, SafetyinNumbers said:

They didn’t do bad insurance acquisitions. They used expensive stock to buy cheap float. They knew losses were coming but thought they had enough margin of safety.

I got a different view from reading [The Fairfax Way], which suggested that for many of the early insurance companies bought, the liabilities they bought into were worse than expected.  Over the years, I also fooled myself somewhat: in the annual letters, I misinterpreted Watsa's optimism and support for the leaders of these individual units, as an indication that these units were already strong (whereas for many of them it took a long time, and losses, before they were turned around).  But I have not done any deep exploration myself, so I could be wrong.

[SafetyinNumbers]

25 minutes ago, MMM20 said:

Also the actuarial/ accountant mindset does not really translate to investing, and there are plenty of failed attempts to outsource the investing piece. It’s hard to get right and not blow up. It’s also been a pretty good call to just own bonds as interest rates declined for basically 4 decades, aka senior insurance execs whole careers.

It also increases the risk of getting fired in widely held companies. If Fairfax was widely held, there are a few moments in its history an activist could have ousted management so we see lucky control lies with Prem. I think it’s one of the motivating factors to keep buybacks going and to have employees own significant equity as well.

 

2 minutes ago, roundball100 said:

I got a different view from reading [The Fairfax Way], which suggested that for many of the early insurance companies bought, the liabilities they bought into were worse than expected.  Over the years, I also fooled myself somewhat: in the annual letters, I misinterpreted Watsa's optimism and support for the leaders of these individual units, as an indication that these units were already strong (whereas for many of them it took a long time, and losses, before they were turned around).  But I have not done any deep exploration myself, so I could be wrong.

I appreciate what they say much like Buffett with Gen Re. If they didn’t both use fully valued equity when making the purchase and had used all cash instead, the deals were worse.

[73 Reds]

2 hours ago, SafetyinNumbers said:

I think it’s because equity returns are lumpy and less predictable. Most companies won’t even take risk on duration or composition of the fixed income portfolio because it introduces variability. 

Yeah, I get that but over time if some (even small) % was allocated to equities, the long term results should be better both as to an increased ability to write more insurance and also nominal investment results, no?  As for activist investors, I've always liked them even when not always agreeing with their positions - they keep management on their toes.

[Buffett_Groupie]

On 1/29/2026 at 11:31 PM, Viking said:

@Buffett_Groupie, great question. Today, I include all associate holdings in engine 2. I think you could make a good case that they should be included in the third engine and make it: share of profit of associates + non-insurance consolidated holdings. What do you think? Is that how you think about Fairfax?

 

I am working on a post where I layer in Fairfax’s 6 income streams. I think that is another helpful way to think about Fairfax’s business model. 

@Viking

 

Will this be appropriate?

 

Engine 1: Insurance and Reinsurance

 

Engine 2: Investment: Fixed income, public and private equity (Seaspan, Eurobank, ...)

 

Engine 3: Non-Insurance: Wholly-owned businesses (Recipe, Sleep Country, Peak Performance, ...)

 

 

Edited February 22 by Buffett_Groupie

[Munger_Disciple]

6 hours ago, Maverick47 said:

@djokovic1 Good question.  Most folks use 95 or 96 as an approximation for the CR target for personal auto insurance coverages as a whole.  In actuality, a typical auto policy is a combination of a number of sub-coverages which are often aggregated into two main buckets — liability and physical damage.  By the way, the two largest personal auto insurance carriers, Progressive and GEICO, both have publicly disclosed over the years that their own auto CR targets are 96.
 

The physical damages coverages are the ones that tend to be resolved very quickly, so produce little in the way of loss and adjustment reserve float (there will be some unearned premium that will add a bit to the investable float, but that’s not terribly significant ).   However, claims for these sorts of losses also tend to be quite predictable and stable over time, and regulators don’t typically require an insurer to allocate a lot of capital in proportion to physical damages premium.  Let’s say a dollar of shareholder equity is required to support $5 of physical damages premium.  The company pays tax on investment and underwriting income of 20%, and for simplicity sake, let’s say only the single dollar of allocated shareholder equity is available to be invested at a pretax rate of 5%.  Say also, that given the minimal float available, that a CR target for physical damage is 97.  The company earns 5 cents of pretax return on the invested allocated equity dollar and 15 cents of pretax underwriting profit on the $5 of premium.  The pretax return of $0.20 per dollar of equity ends up being $0.16 after tax, for an after tax return on equity of 16%.
 

Bodily injury liability claims do produce a much greater amount of longer tailed loss and adjustment reserves (and float) and those claims take much longer to resolve than the relatively simpler physical damage claims. But regulators and the company’s own assessment of risk will also result in companies choosing to hold a greater amount of equity to support these claims, which tend to be more volatile than the simpler physical damage losses.  Let’s say a dollar of shareholder equity is selected by the insurer to support every $2 of their auto liability premium, and that the investable float is equal to $1 for every $2 of premium.  If the company wants to calculate the pre-tax CR target required to produce a 16% after tax ROE for this portion of auto premium, they would first realize that investing the $2 of equity and float would generate 10 cents of pretax income, and 8 cents after tax.  Then they would require the other 8 cents of after tax income to come from the underwriting profit on the $2 of premium.  That would be 4 cents after tax on each of the two dollars of premium, and 4 cents / (1 - 0.20) = 5 cents of pre-tax underwriting profit for every dollar of auto liability premium.  That means a target CR of 95 for the auto liability coverages.

 

With the premium volumes of liability and physical damages coverages being fairly close to each other, many companies managements simply average the two separate CR targets of 97 and 95 together for a combined 96 CR target across all auto coverages instead.

Bottom line, there are more variables considered by management besides the amount of float produced by a line of insurance business that go into setting a CR target.  One very important one is how much equity is required to support a dollar of written premium.  Generally speaking, auto physical damage is considered one of the least risky lines of business, so requires the least amount of supporting equity.  Homeowners insurance, in my opinion, on balance, is riskier than auto liability.  Depending on the type and amount of catastrophe exposure, in some geographies it could range from a bit less risky than auto liability, to a whole lot riskier.  
 

Some areas exposed to hurricanes, tornadoes, wildfires, and fires following earhquakes (think San Francisco in 1906) could easily require a dollar of equity (and maybe even more) to support every single dollar of homeowners premium ( US states such as California or Florida are examples of geographies where this could be true).  
 

Since most homeowners claims don’t take as long to resolve as auto liability, the float per dollar of home premium is not as great as it is for auto liability.  The after tax investment return then is going to be less than the 8 cents it was for auto liability, and the pre-tax return on underwriting only comes from the single dollar of homeowners premium instead of the two dollars of premium as was the case with auto liability.  So instead of 5 cents of target pre-tax underwriting per dollar of auto liability premium, the target pre-tax underwriting profit per dollar of homeowners premium in a cat exposed state like California or Texas would be even greater than 10 cents, for a CR homeowners target in those states in the mid to high 80’s.

 

Finally, catastrophe lines of business such as earthquake are probably the riskiest of all. 
 

If I was pricing that line of business on a direct basis, ignoring the effects of catastrophe reinsurance that every prudent company would purchase,  I might assume that a dollar of equity would be required to support 30 cents of earthquake premium.  Most years there are no earthquake losses or reserves available to produce investable float, so a 16% after tax ROE would come from first investing the dollar of equity for 4 cents of after tax investment return.  Then we would need the remaining 12 cents of after tax return (15 cents pre-tax) to come from the $0.30 of premium, for a CR target of 50%.

 

In practice, every primary insurer is going to purchase a property catastrophe reinsurance treaty from reinsurers including any number of Fairfax reinsurers.  Then the primary company wouldn’t need to put as much of their own equity at risk, because it is much more cost effective to purchase cat reinsurance, and then the CR target for the cat exposed earthquake line wouldn’t need to be as low as the 50% noted above.

 

Sorry for the long-winded response!  But now you also have a flavor for why capital is required to be retained or even perhaps added to insurers when they would like to grow premium volumes substantially during hard markets, such as was the case with Fairfax’s insurance companies in the early part of this decade.  Regulators and rating agencies tend to look askance at companies whose rate of premium growth begins to outstrip the growth in their supporting equity. That’s what happened to a lot of companies that didn’t manage to protect their balance sheets as well as Fairfax did when interest rates began to rise.  Many of them had to either restrict growth, or find ways to shed premium in order to avoid ratings downgrades as their shareholder equity levels either stagnated or fell.

 

Pretty impressive that Fairfax managed their way through a period of substantial premium growth while also earning ratings upgrades for their insurance companies!

 

Thanks @Maverick47 for the excellent post! Keep going with more long winded educational posts!!

[TwoCitiesCapital]

1 hour ago, 73 Reds said:

Yeah, I get that but over time if some (even small) % was allocated to equities, the long term results should be better both as to an increased ability to write more insurance and also nominal investment results, no?  As for activist investors, I've always liked them even when not always agreeing with their positions - they keep management on their toes.

 

I think the problem is you 1) know that there are going to be catastrophes that result in some portfolio liquidations and 2) that markets will go down. 

 

You can't really afford both to happen at the same time if using regulatory capital. A year like 2022 proved to be a disaster for the insurance industry even WITHOUT equity exposure which would have made it worse. 

 

Is my understanding that even Fairfax uses primarily fixed income for regulatory capital and only invest it's "equity" (or the equity of its subs) in equities. 

[Maverick47]

15 hours ago, 73 Reds said:

Yeah, I get that but over time if some (even small) % was allocated to equities, the long term results should be better both as to an increased ability to write more insurance and also nominal investment results, no?  As for activist investors, I've always liked them even when not always agreeing with their positions - they keep management on their toes.

I do think that you can find many insurers who have “small” percentages of their assets invested in equities, or other more volatile investments such as limited partnerships, private equity, etc.  It’s just that very few of them will consciously allocate a large enough percentage to really move the needle on the business model.

 

Berkshire is obviously the leader that was the shining example of the business model that would see a significant portion of assets invested in equities or non-insurance operating, but there aren’t that many insurers that have elected to follow them down that same path.  

Most companies are going to be run by management that doesn’t have significant ownership/control.  They may be somewhat more afraid of losing their high-paying jobs if they significantly underperform the industry in the short term.  Conversely, if their investments track most of the industry, their Boards will blame any poor performance on structural or economic forces outside their control and they’ll keep their jobs.  When interest rates finally rose after 40 years of declines, how many insurers kicked their CEO’s to the curb when large mark to market losses in their fixed income investments occurred?  I can’t think of any examples myself.

 

Additionally, regulators and rating agencies’ capital adequacy models take a much dimmer view of equities as investments than they do fixed income investments in government bonds, or high rated corporate bonds.  They want to measure or stress test how much supporting equity the insurer will have to support the amount of premium they are writing.  Their models allow insurers to get more credit for capital invested in bonds than in equities.  
 

So an insurer that wants to invest a lot of the equity supporting their premium in equities will either be encouraged or required by the rating agencies and regulators to hold a higher amount of total equity to receive the same rating (or to escape regulatory controls) than a company that invests most of their assets in government bonds would need to hold.

 

That can be a problem for a publicly traded insurer whose CEO or CFO (or both) have been trained to seek to optimize their reported return on equity.  To make that number as high as possible, they will try to run as “efficiently” as possible, with the equity in the denominator as low as possible.

 

They would actually see a need to hold a larger amount of shareholder equity if it is invested in stocks versus needing to hold a smaller amount if it is invested in bonds as a penalty to be avoided if they also want a certain rating agency claims paying ability rating.

 

To keep that equity denominator as low as possible, while still retaining their desired ratings, they will “logically” choose to avoid sizable investments in equities.

 

You’ll also often see their shareholder communications trumpet how much capital they return to shareholders via either dividends or share buybacks.  And unlike Fairfax or Berkshire, they won’t add the caveat that they will purchase their own stock only when it is either fairly, or preferably, undervalued, compared to management’s assessment of intrinsic value.  That’s how one can tell they are going to mindlessly repurchase their own stock along with investing mostly in fixed income in the effort to keep that equity denominator as low as possible.

 

Prem doesn’t have to worry about losing his job, and makes most of his money on shareholder return anyway, not in salary or stock options, so he’s comfortable targeting a “lumpy” 15% return promised by allocating large amounts to equities compared to much of the rest of the industry that targets a “steady” 12% return and will happily invest most of their assets in fixed income instruments in an effort to achieve that.

 

Personally, I’ll be interested in whether Chubb’s investment strategies will move more towards equities now that Berkshire has take a relatively large stake in that company.  

Edited February 23 by Maverick47

[Maverick47]

26 minutes ago, Maverick47 said:

Personally, I’ll be interested in whether Chubb’s investment strategies will move more towards equities now that Berkshire has take a relatively large stake in that company.  

Guess I should have checked first!  Chubb does have a decent amount of assets invested in private equity partnerships and the like (though I wasn’t seeing as much in the publicly traded equity category).

 

Much like Fairfax, their fixed income investments look to be sized to cover their loss reserves and a little bit extra.  They take the additional investment/volatility risk only with their own shareholder equity.