Maverick47

Just an observation here from having worked in the US residential housing insurance business for a few decades. Average loss costs for a home insurance policy regularly increased faster than CPI inflation for that entire period of time. Part of the cost inflation had to do with some climate change and getting better at estimating/modeling the average annual costs of catastrophes such as storms, tornadoes, hurricanes and wildfires. But at least as much, if not more of the inflation arose from the newer housing stock being added each year being larger and more complex in terms of systems and features than the relatively modest, smaller, century old or older homes being demolished and removed from the overall housing stock. More bathrooms in newer homes, more complex plumbing in kitchen areas, more open concept designs, meaning if flooring gets damaged in a plumbing accident, it’s more difficult for a claims adjuster to determine that the acceptable repaired area could stop at an obvious visual break point such as a threshold doorway to a hall. I recall one water damage claim where a plumbing fixture in an upstairs bathroom broke and damaged bamboo flooring in and near the bathroom, but that same flooring continued uninterrupted throughout the rest of house, down the staircase, all throughout an open concept ground floor level, etc. All of it needed to be replaced at a cost of well over $100,000 compared to the cost of about $10,000 if similar damage had occurred in an older home with doors, hallways, etc. The average square footage, number of bathrooms, quality of construction, etc all went up year over year, while the average household size tended to get smaller. Without a significant multi generational sharing of housing costs, it’s fair to say that the cost to heat and cool a home, pay for other utilities, including internet, cable, regular maintenance and upkeep, property insurance, taxes, mortgage has risen much faster than inflation in pay over that same time frame. That’s just another reason that the housing supply and affordability issue has been so challenging to address.

I wonder the same thing about the possibility of issuing shares well above book value, and then buying them back at less of a premium to book value later, as Singleton did many times with Teledyne. I would never quarrel with the wisdom of such a transaction from the perspective of one who intends to hold Fairfax shares for the long term, but from the standpoint of the “other parties” to these sorts of transactions, the idea of “Fool me once, shame on you, fool me twice, shame on me” comes to mind. I’m still struggling with understanding the motivation of the counterparties to the current TRS. Is it as simple as finding financial institutions that wanted fixed or floating income instruments and the Fairfax side of the swap promised them a risk margin above the sort of interest rates they would otherwise be able to obtain at the time the swap was entered into? So they used the funds they wanted to invest in fixed income, and used them to buy Fairfax common stock instead, entering into the TRS agreement with Fairfax for them?

Good point. By year end, I imagine the share count reduction over the last 10 years or so will be in the range of 7 million. That’s a net reduction of over $100 million per year in required annual dividend distributions.

Wondering how treasury share purchases appear in these sorts of reports, if at all? Is it possible that some of the shares shown as being purchased by the company could have been set aside in connection with the company’s intended use of them to add to their treasury shares for future incentive compensation purposes? I’m guessing a few hundred thousand shares per year are issued from the corporate Treasury on five year vesting schedules to employees, and I believe the company indicated that they would purchase on the open market each year the number of shares needed for that program.

@Viking, Munger once said that he and Buffett were cautious of investing in financial companies that “were trying to do well”. Banks, for example, can appear to be more profitable than they actually are, if they are trying desperately to report profits, and in so doing, abdicate their responsibility to set prudent loan loss provisions…. An insurer I worked for once separated profit and growth goals by assigning the marketing department incentives based on growth, and the underwriting department incentives based on underwriting profit. You can imagine how these are in conflict. Tough to get a company pulling together when one part wants to write as many new policies as possible, while another part is motivated to restrict writing new business with its attendant new business penalty. Years later, a different company I worked for focused on growth some years and underwriting profit in others. Yet the underwriting and sales marketing departments continued to have individualized goals based on profit and growth separately. In years when the company as a whole emphasized underwriting profits, it became difficult to retain experience marketing reps as their bonus compensation, based on growth incentives, plummeted. And in one case I recall, when incentives were strongest for underwriting profitability, reserving actuaries for a large line of business apparently under-reserved that line significantly for several years in a row. Significant bonuses were paid for a few years in a row until the oversight could no longer be masked, and substantial loss reserve increases were booked instead. A few individuals were blamed and fired, but it was a painful few years thereafter to recover from that mistake. That is one of the reasons why it is so important to monitor reserving practices of an insurer. Fairfax is on the right track by focusing on making sure that reserving practices are conservative, and are more likely to be redundant than inadequate, as evidenced by a long track record of reserve releases from prior accident years. As you have so clearly outlined, Fairfax now has their insurance subsidiaries focused on steady underwriting profitability, prudent risk and catastrophe exposure management, with conservative reserving, at the same time that their investment and capital allocation decisions are virtually unmatched by competitors in the industry. From a value investor/shareholder perspective, their operational results are providing us with “margins of safety” in a number of areas, to say nothing of the excess of market over carrying value of their assets.

Really good topic—incentives do indeed drive behavior, and oddly enough, many companies fail to craft incentives that drive the optimal behavior and financial results they seek. Buffett used to focus on this in his letters back in the 90’s around the time he took control of Geico. He mentioned that incentive plans should be easy to understand, measurable…and Munger would add that they should also not be easily gamed. All great points. Oftentimes companies strive to have both profit and growth (the Holy Grail). But growth can be a challenge since it is so easy to grow if pricing is too low, and thus underwriting profit would suffer…. Buffett’s original GEICO incentive structure was a matrix that produced bonuses based on underwriting profits of policies written more than a year, and number of new policies written, so both growth and profit. He knew that if the underwriting profit goal took into account all policies, that he would be incentivizing employees not to write as many new policies, since in auto insurance, it is almost a given that the newest policies will be the most likely to produce underwriting losses (known as the new business penalty). If he wanted GEICO to grow, he was willing to let the company absorb the new business penalty, and not allow it to adversely affect employees decisions about whether or not to write new business policies. A company I once worked for thought they could motivate insurance agent policy production by granting agents bonus commissions for achieving year end policy count targets set for each agency. They thought that the measurement would be a foolproof way to drive growth, since it would motivate both retention of existing policies and production of new business. However, the incentive system was easily gamed because agents could cancel current auto policies for customers who insured more than one vehicle, and then write separate auto insurance policies for each separate vehicle. Presto! A single policy that previously insured four vehicles owned by a customer could canceled and become four separate new policies, each insuring one of the originally insured four vehicles. Do this for every Inforce customer with multiple vehicles in a household, and an agency’s policy count at year end could easily exceed their assigned growth targets, and the company would pay them a bonus. This could even be the case where the total number of vehicles insured and total premium volume declined for a given agency year over year. The company incurred increased mailing and printing costs to cancel and reissue all of the policies, and employees’ time was wasted. Clearly, in such cases, neither growth nor profit was the result of the agency incentive system. Fairfax appears to be a good deal smarter than the company I worked for. While we don’t have the details of their incentive systems, we are told that it is primarily based upon underwriting profit. That is definitely a good sign.

I think you’re comparing Trump’s $2.2 billion of income in 2025 to the volume of trades made by all the traders in Congress, totalling $665 million? Probably not apples to apples. Highly unlikely that the congressional trades resulted in the same volume of profit on those trades. If they made 20% profit on their trades, that would be $130 million. Trump’s grift alone is more than 15 times that.

Very enlightening. Thanks for sharing!

They did sell their stake in a life insurer they had owned part of with White Mountains, but that is different from selling a wholly owned insurance subsidiary, which Fairfax has been known to do. https://www.wsj.com/articles/BL-MBB-40289

Great point! Hadn’t thought about that before, but it makes great sense.

This (flexible and intelligent capital allocation) is a really important differentiator for top notch insurance companies/conglomerates @Viking. There are plenty of publicly traded insurers who attempt to maintain underwriting discipline through an insurance cycle, but they unnecessarily restrict their capital allocation options. Too many of the insurance CEO’s limit their capital allocation decisions (when underwriting opportunities become less attractive) to share repurchases, dividends and possibly acquisition of other insurance companies. They either don’t appear to be able to think outside those boxes, or perhaps they just don’t have an optimal holding company structure for insurance company subsidiaries that would also allow them to acquire subsidiaries outside of insurance. If they are focused on Return on Equity, when profitable underwriting growth for the numerator is difficult to achieve, they typically will turn towards returning capital to shareholders as a means of reducing the denominator. This can result in a relatively efficient insurer, addressing the issue of “excess capital” if it would otherwise be invested in low return fixed income instruments and damage the Return on Equity measurement. But the problem can be that in their urgency to “return” excess capital to shareholders, they forget to consider whether the price at which they are repurchasing shares is favorable or not. Both Fairfax and Berkshire have been (and are) proving themselves to be adept at rationally scanning a much wider global opportunity set of places to invest any excess capital that can accumulate during softer markets.

Good point!

Thank you both ( @SafetyinNumbers and @djokovic1) for pointing us to these charts (and I will agree that I happen to like the formatting with the green bars better as well)! It appears to me that the actual catastrophe dollars displayed on the charts are not inflation adjusted, and are relatively similar in dollar magnitude for the last 9 years, while the equity base on which the cat tolerance is calculated has been growing significantly. We’re seeing the power of the business model at work I believe. If we look out another decade or two, if we are fortunate enough to escape the big 1-in-250 year or even rarer type year of really bad catastrophe years, I can envision a future where this risk continues to moderate and becomes almost as much of an afterthought as it is currently for Berkshire Hathaway. Berkshire most recently appears to want to retain about $20 billion in cash against the possibility of a mega catastrophe year. That’s less than a half year of their normalized earnings, and much less than 15% of their shareholder equity. Similarly, management of Fairfax seems to have developed sufficient underwriting and risk management discipline to continue to reduce the probability that catastrophic insurance losses might represent a “company killing” event in the future. While the holding company is just over 40 years old, at least one of its subsidiaries (Crum & Forster) has been “alive” since it was incorporated over 200 years ago in 1822. That’s a good example for the holding company to want to emulate, and all indications are that they are on the right track.

This is an important question for any insurer, @Hoodlum. The company has set risk management guidelines for this risk that are disclosed each year in the annual report under note 22. From the 2025 Annual Report: Catastrophe risk Catastropheriskarisesfromexposuretolargelossescausedbyeitherman-madeornaturalcatastrophesthatcould result in significant underwriting losses. Weather-related catastrophe losses are also affected by climate change which increases the unpredictability of both frequency and severity of such losses. As the company does not establish reserves for catastrophes in advance of the occurrence of such events, these events may cause volatility in the levels of incurred losses and reserves,subject to the effects of reinsurance recoveries.This volatility may also be contingent upon political and legal developments after the occurrence of the event. The company evaluates potential catastrophic events and assesses the probability of occurrence and magnitude of these events predominantly through probable maximum loss (“PML”) modeling techniques and through the aggregation of limits exposed. A wide range of events are simulated using the company’s proprietary and commercial models, includingsinglelargeeventsandmultipleeventsspanningthenumerousgeographicregionsinwhichthecompany assumes insurance risk. Each operating company has developed and applies strict underwriting guidelines for the amount of catastrophe exposure it may assume as a standalone entity for any one risk and location, and those guidelines are regularly monitored and updated. Operating companies also manage catastrophe exposure by diversifying risk across geographic regions, catastrophe types and other lines of business, factoring in levels of reinsurance protection, adjusting the amount of business written based on capital levels and adhering to risk tolerance guidelines. The company’s head office aggregates catastrophe exposure company-wide and continually monitors the group’s aggregate exposure. Independent exposure limits for each entity in the group are aggregated to produce an exposure limit for the group as there is presently no model capable of simultaneously projecting the magnitude andprobabilityoflossinallgeographicregionsinwhichthecompanyoperates.Currentlythecompany’sobjective istolimititscompany-widecatastrophelossexposuresuchthatoneyear’saggregatepre-taxnetcatastrophelosses would not exceed one year’s normalized net earnings before income taxes. The company takes a long term view and generally considers a 15% return on common shareholders’ equity, adjusted to a pre-tax basis, to be representative of one year’s normalized net earnings.The modeled probability of aggregate catastrophe losses in any one year exceeding this amount is generally more than once in every 250 years. The bottom line is that the company models the aggregate amount of multiple potential catastrophe events occurring anywhere around the world in a given calendar year and currently estimates that the likelihood that this amount will exceed 15% of common shareholders equity (roughly $3.9 billion on an after tax basis, or about $5 billion on a pre-tax basis, given 12 2025 shareholder equity of $26.3 billion) is less than 0.4%. This risk management guideline then essentially would appear to have the goal of turning aggregate catastrophe losses, those expected to occur more rarely than once every 250 years, or with a probability of less than 0.4% in any calendar year, into an income statement event, not a balance sheet event. We probably should also keep in mind that the company regularly records catastrophe losses around the world which are being charged for in their annual premiums and which cover the average annual level of catastrophe losses expected over the long term. I don’t know what the average number is, but I think 2025 cat losses were $1.2 billion, and the company recorded a combined ratio of 93% with pre tax underwriting profit of $1.8 billion. So a $5 billion pre tax year for catastrophes would appear to be only $3.8 billion pre tax above the actual 2025 result. Since the pre tax earnings in 2025 exceeded $3.8 billion, this appears to indicate to me that if 2025 had been the year with $5 billion in pretax catastrophe losses, the company would still have reported positive earnings of around $1 billion or so. Between cash on hand and short term securities, the company looks to be situated well to handle such a year.

Interesting to see the asymmetry in inexpensive drones for attack versus expensive drone defense. Probably one of the reasons the US is interested in bringing the war with Iran to a close. US air defense missiles are quite expensive and the war has already likely depleted years worth of current production levels needed just to replace what has already been expended.