Maverick47

@Thrifty3000 Back of the envelope math is helpful to aggregate the numbers to a state level as you’ve done. I would just extend the scenario to take into account expenses of the insurance companies, customer acquisition and retention costs, claims adjustment costs, and even the cost to purchase reinsurance from companies like Odyssey of Fairfax and National Indemnity of Berkshire so that the primary companies can stay in business after one or more hurricanes in a given year. First, it would not be unexpected for companies to only be able to pay out 50% of the premiums they collect in a state like Florida. Their expenses for claims and company management are likely to be near 20% of premiums, and they probably have to pay commission of 10% or more to the agents. They would probably require a profit load of 10 to 20% to compensate for the risk that they might have to pay many times the premiums they collect in a given year in a bad hurricane year. They might also have to pay 10 to 20% of premiums to purchase property reinsurance. In a best case view, they might only have 40 to 50% of premiums available to pay all homeowners losses. Some of the expected losses such as fires, theft and plumbing related water damage will happen every year with or without hurricanes. And in a total loss of a home, the $400,000 home would need to be rebuilt, but the policy would also pay to replace personal property (contents), damage to outbuildings, fences, sheds, etc, and also pay for loss of use until the home was rebuilt. In a large catastrophe, rebuilding costs can inflate, making a $400,000 home cost perhaps $100,000 more to replace. So let’s use close to million for the total loss to a home insured for $400,000, with $300,000 of personal property, $40,000 of outbuildings, $100,000 for loss of use and another $100,000 for increased building costs also needing to be paid. The balance to add up to $1 million may be covered by debris removal costs that are also covered, as well as increased costs to rebuild a home to current building codes, since many customers purchase coverage to pay for increased costs associated with these. Maybe half of the state’s loss budget is available for annual hurricanes, with the other half paying for all of the other causes of loss (fires, theft, water damage, etc). So a company can afford to pay annually perhaps 20 to 25% of premiums for hurricane losses. With 9 million $400,000 homes paying $11,000 annual premiums, let’s round the collected state premiums to $10 billion and the industry has a chance to achieve target profitability with annual hurricane losses of 20 to 25% of that, say $2.5 billion in hurricane losses allowable for target profitability to be achieved. Most homes damaged by hurricanes completely are likely to be more expensive homes with ocean views and proximity, so I think it’s fair to say they would cost on average well over $1 million for a total loss, but let’s stick with $1 million. How many homes can $2.5 billion replace in Florida if each home costs the industry $1 million? Now that we account for coverages beyond the home itself, and subtract annual expenses other than loss dollars alone, as well as losses that happen unrelated to hurricanes, it seems as if the adjusted back of the envelope math says that the industry can afford to replace only 2500 homes per year in Florida as a result of total hurricane losses and still make enough profit to choose to remain in business for the future. If they go 10 years without hurricanes, total hurricane related profit would be $2.5 billion for year or $25 billion in aggregate. But if in year 11, one or more hurricanes were to destroy only one half of 1 percent of all homes in the state (1 out of 200) that would be 45,000 homes costing $45 billion to replace, wiping out total profit from almost 20 years of no hurricanes. Now consider that a direct category 5 hurricane hit on Miami could cost the industry $160 billion or more and that single event could wipe out 64 hurricane free years of profits and you might begin to gain an appreciation why most national and multi state insurance companies are reluctant to write much business in the state, and why the state run carrier of last resort (Citizens Insurance) has the largest market share in the state with more than 1 million of the total 9 million homes in the state insured.

Personally I’m not too worried about corporate tax rate changes over the long term. Insurance companies will set rates targeting their desired after tax returns on allocated capital. So if tax rates on US based income increase, and if target after tax returns are unchanged, insurers’ pricing formulas will simply adjust to charge their customers rates that will be sufficient to generate those after tax returns. You are correct to be concerned about what would happen in the short term. Since rates are set in advance, to the extent that premiums in force are set based on lower tax rates, they will generate lower than desired profits between the time when tax rates are changed and when the rates can be adjusted upwards. This should affect all insurance companies in the taxing jurisdiction relatively equally as a one time upward or downward impact on after tax earnings, depending on whether corporate tax rates are increased or decreased.

@Viking I think you actually understand the forest of the P/C business as well as, if not better than, most folks who worked in a particular area of the business (such as myself) and who thus perhaps know more than you do only about a particular grove of trees in that forest! That said, humility is one of the better characteristics of true value investors, so far be it from me to ever ask you not to employ it…. As you alluded to, some of those groves are longer or shorter tailed, or are primary rather than reinsurance, or commercial vs personal, to say nothing of the various geographies and regulatory regimes. I do know enough about the surviving competitors in the business to feel optimistic about a soft cycle not necessarily hitting all lines at the same time, and that we’ve seen enough discipline from many participants in the business to not expect that this would necessarily mean a reversion to a 100 CR for Fairfax. If you search Buffett’s letters from first decade or so after he bought all of Geico, he would often bemoan the lack of underwriting discipline of GEICO’s competitor, State Farm, which as a mutual insurance company, had accumulated a massive amount of surplus and was apparently more than willing to expend that surplus in support of CR’s well above 100 to maintain their industry leading market share. This has changed over time, and as an example, State Farm is no longer willing to write business at a loss even in very large states such as Florida and California simply to maintain market share. I’ve seen similar examples of numerous US competitors also “getting religion” about the need to achieve CRs below 100. That is exactly the right question to focus on. They have exhibited this discipline in the recent past and Prem certainly highlights the CRs by subsidiary in his letters, which shows to us shareholders (and the leaders of the insurance subs) that he cares about this a great deal. I suppose the only thing that I might like to know in addition is whether the incentive compensation plans for employees value this as well. I do like it when everyone’s interests are aligned….

Sandy was a hurricane in 2012, but when it hit the Northeast it had weakened to Tropical Storm strength. A number of the northeastern states had coastal home insurance customers with hurricane deductibles of 1 to 5% of the replacement value of their homes. Insurance commissioners determined that the wind speed at landfall was less than hurricane strength, so did not allow the higher hurricane deductibles to apply. That probably helped contributed to the losses of about $93 billion. Since hurricane deductibles were not triggered, the industry tends to refer to the event as Superstorm Sandy. A category 3 or greater hurricane hitting the right portion of the Northeastern US could cause insured losses many times that of Sandy.

This is pretty typical of the cat risks faced by all reinsurers. The US tends to have the largest concentration of insured property exposures, and all the primary carriers exposed to that are interested in purchasing property cat reinsurance. While the frequency and severity of potential hurricanes hitting concentrated areas of Florida like Miami are of concern, the larger cat risk for the US tends to be Northeastern hurricane. These exposures are modeled, so that the impact of sample hurricanes hitting the current insured properties are estimated. One historical hurricane in 1938 hit Long Island and continued to cause significant damage inland. However, the density of insured dwellings in the area it hit was nowhere near as much as it is today. So a repeat of that exact hurricane would cost quite a bit more than a category 5 hurricane hitting Miami. Odyssey and Allied World and other of Fairfax’s reinsurers are careful to manage the property cat risk they write in aggregate.

You’re right. But perhaps we can learn something about how to prepare ourselves (and our investments) to respond to the sorts of dangers that we are bound to face sooner or later, both those that are unlikely to occur and those we simply cannot anticipate in terms of the timing and the magnitude when they do occur. That’s one of the reasons I was willing to hold on to my Fairfax shares after 2009. Even though I didn’t personally think global deflation was likely, I also didn’t mind paying(in the form of poor returns on the shares) for the insurance policy that investment provided me, in the event that I may have been wrong about the future.

Thank you for clearly adding this observation! I know little about corporate taxation. I had noticed the significant impact on Berkshire’s deferred tax liability driven by BHE and BNSF, but did not realize what exactly was driving this….. I don’t think that Fairfax’s own non insurance operating subsidiaries yet have the same sort of requirements or opportunities for capital investment and accelerated depreciation as Berkshire, but it might be worth looking out for this in the future. Right now though, Fairfax has plenty of opportunities to direct any operating earnings towards stock buybacks, purchasing minority interests, etc, so really have no need to invest in subsidiaries that would also require substantial capital investments.

@Munger_Disciple I like your idea of calling this our personal “float”! It does make sense to take this into account in taxable accounts. Speaking of which, Fairfax and Berkshire both have this additional type of float on their own balance sheets as deferred income tax liabilities. Since the cost of this float is always zero (an interest free loan from the government), it is basically just as valuable to shareholders as insurance float liabilities when combined ratios are at a 100 level. When combined ratios are less than 100, then insurance float is essentially like a loan with a negative interest rate, and thus is more valuable than a deferred tax liability “loan” of the same amount, while combined ratios above 100 imply a positive interest rate on the insurance liabilities, and then they would be less valuable than the interest free loan from the government. But as you note, we (and companies in similar deferred tax situations) can decide “if and when to incur it”). This optionality has some indeterminable but non-zero value which makes it somewhat more valuable than insurance liability float which itself depends upon the vicissitudes of weather and claims behavior as far as identifying when the insurance float “loan” (or a portion thereof) has to be repaid.

Thanks for the recommendation, @Saluki! I’m enjoying this greatly. One can easily draw lessons for investing from some of the stories. For example, the unanticipated results of roping climbers together could be compared with the dangers of tying your investments together, either by concentrating in a single industry, or, worse, by concentrating in the industry that your own employment is dependent upon. Instead of thinking about what could go right with an investment, I expect I’ll want to spend more time thinking of what could go wrong, so, as Charlie Munger notes, “I’ll never go there.” I appreciated his comments recommending writers on wildfire and airline disasters, and picked up some additional books to read on those particular topics…. I’d much rather learn these sorts of lessons vicariously than from personal experience…

Chap 4 of “The First 25 Years of Fairfax” talks about Prem’s controlling Fairfax and Markel owning 20% when Prem wanted to buy Commonwealth Insurance Company of Vancouver, BC. Steve Markel had little interest in expanding in Canada. ”There were only two options: either we put Markel Corporation and Fairfax together or we split them apart. We decided to disentangle them completely so that each of us would have greater freedom to pursue our growth.”

The Markel family was involved from the beginning, investing more than 25% of the capital needed to set up the Canadian operations of Markel insurance as being controlled by Prem. Look at the 1985 chairman’s letter on the Fairfax website and you can see that the initial name of the company was Markel Financial Holdings Limited, and changed to Fairfax a few years later. Stephen Markel apparently ran the Canadian operations from his base in Virginia until Prem could find a President to run it.

Thanks, @Viking, for pointing out the magnitude of the share count reduction. At 20.1%, it is becoming material. On a per share basis, each remaining share that we own now owns more than 20% more of the company’s earnings than we would have owned had the share count remain unchanged. All else being equal, the estimated increase in the portion of the company’s earnings controlled by surviving shares can be calculated as (1/(1-share count reduction)) -1. Assuming a 20% share count reduction, the estimated increase in per share earnings would be (1/0.8) -1 = 25%. A 10% reduction would lead to only an 11.1% increase. Just for fun, let’s examine a Henry Singleton Teledyne share count reduction of 88%: That produces an increase in per share earnings of (1/0.12) -1 = 733%. I’m gaining an appreciation for capital allocators who choose to opportunistically buy back their own company’s shares when they sell for below intrinsic value, because that just makes the above estimates higher!

I agree. At a 33, with not a whole lot of EPS growth the last few years, cash at 5% may well be more attractive, and the cash option has an unknown value, but likely to be greater than zero, especially in the hands of Buffett. I have a few small positions of my own that have had multiple expansions over the last several years and are now in the mid 30’s to 50 or so PE ratios. I think I’ll trade out of them and into some other positions with lower PE’s, including Berkshire. I like outsourcing my decisions about when to move to cash to Warren…

I was wondering about this too. All else being equal, I would think that after the stock buybacks in such a case, the resulting book value per share would be slightly less than if the buybacks hadn’t been made. But since we appear to believe that the fair market value per share ought to be in the range of 1.2 to 1.5 times book if Fairfax were valued similarly to competitors, then buybacks are still occurring below the intrinsic/expected fair value per share, and so they are still expected to be beneficial for existing shareholders. Berkshire used to have a set guideline that they would be allowed to buyback their shares when the price to book was 1.20 or less. Then as book value became less applicable to their business, they removed even this restriction and now allow buybacks whenever the price is less than the intrinsic value per share as determined by management (Buffett).

I have appreciated a rational capital allocator being in charge of companies like Berkshire and Fairfax. When they are engaged in buying back their own stock, it serves me as a small individual investor, because I interpret it as meaning that the stock is at least fairly, and in some cases, favorably, priced relative to intrinsic value. So during my capital accumulation phase, it gives me a level of comfort in continuing to add to a position, even though the market price per share continues to rise over time. I have a human tendency to anchor myself to valuations at which I first purchased a stock. Although one might occasionally be fortunate enough to see those prices again (see Fairfax a few years ago), a more typical situation is when the valuation continues to rise steadily over time. Stock buybacks at such times give me some comfort that I am not dramatically overpaying for my later purchases…

@73 Reds was on the right track with regard to current industry practice in pricing auto insurance customers. In order to get customers to switch to a new insurance company, almost all companies now offer prices or discounts to them which are not warranted as they are not based on their actual expected loss costs. Once they are in the door however, those discounts drop off over the next few years so that somewhere around years 2-4 or so, the company breaks even on them. Prices continue to rise at each subsequent renewal, while loss costs on longtime customers tend to improve. Most of the underwriting profit is generated by customers who’ve been with the company for over 4 years. In that sense, the longtime customers are indeed paying more than they should be and could likely get better prices if treated as a new customer with another company. Most of the underwriting loss is generated by the newer customers. If a company priced every customer to make a target underwriting profit, prices for new customers would be too high relative to other companies and so they wouldn’t sell any new policies. Their prices on existing customers would be better, so they’d be more likely to retain them, but eventually all their existing customers will die and they won’t have been able to replace them with new customers in the meantime. It really is sort of a catch-22 situation. For long term customers with good driving records, an optimal approach is likely to entail shopping for replacement coverage on a regular basis. The industry relies on actual customer behavior being much less rational. They rely on customers such as myself being too lazy or simply uninterested in shopping regularly for replacement coverage.

@Parsad I really appreciate what you have to put up with in managing this site. I’d been a longtime browser for free but as I neared retirement and was going to want to manage my retirement accounts myself, I discovered that the modest fee to join is worth it many times over, just for the insights others are willing to share here. So sorry you have to put up with this kind of crap and hope you realize how much the vast majority of us owe to you….

Thanks @Viking! As always, your posts are full of thought provoking observations and questions. The section highlighted above from the end of your post struck a chord with me. As a retired actuary who spent 33 years with the same P and C company, who also admired and personally invested in Berkshire, Markel, Fairfax and (before Berkshire acquired it), Alleghany, I spent a good portion of my own time trying to get the company I worked for to copy the business model of these other companies, with no success. Your points about why more competitors don’t copy them are pretty much the same as I eventually concluded. Exponential compounding connected with long term focused investing in equities really begins to pay off once it’s had a long runway to work its inexorable magic. Ten years, which is a long time for a typical CEO just isn’t long enough for the associated risk to pay off. Family run companies like Berkshire, Fairfax, Markel, are able to take the long view and see how powerful a few extra points of returns, albeit lumpy, can pay off once the timeframe approaches and even exceeds 30 years. Charlie Munger was once asked why more companies didn’t follow Berkshire’s approach of focused investing in a few companies. He responded with “It’s a good question. More companies should follow us. Look at our results and the fun we’re having. But Jack McDonald, the Stanford Business professor who teaches a course based on these value investment principles, says he feels like the Maytag repairman.” I used this same question and answer in a white paper I shared with a new CEO and CFO at the company I worked for. The CEO responded with a short email basically saying it was interesting, but attributing Berkshire’s success only to great stock picking. I got a brief audience with the CFO, during which I learned that the company I worked for had decided to outsource the investment function altogether to a company focused on index ETFs for both equities and fixed income. That’s when I gave up trying to advocate for a more intelligent and successful business model from the inside of a company, and now in retirement, I vote with my own assets and invest them only in companies like Fairfax that have a proven track record of managing float, investments and insurance underwriting in an intelligent fashion. I’ve seen how hard it is to get competitors to change to a similar approach, and am convinced that this holistic long term view of a successful insurance business model is a true moat in and of itself. Analysts like Brett Horn who don’t think Fairfax has a moat miss exactly what you outlined in your post — the proof in the pudding of a hugely successful long term compounder, one which still has a good amount of runway left ahead of it.

I don’t have an estimate of how much could be deployed from fixed income to equities, but I think it’s not uncommon for insurance company managements to choose to hold their insurance loss, and expense reserves (and unearned premium reserves) in the form of relatively secure fixed income instruments. The amount of float held by Fairfax might be a reasonable approximation for such a lower bound on fixed income — roughly $35 billion? I’d be surprised if they chose to hold fixed income investments in an aggregate amount less than that. I think they currently hold over $40 billion in bonds? in addition, there may well be some regulatory or rating agency constraints on investments in equities. There’s often an implicit trade off between underwriting and investment risk for a company. When a company expands its premiums written such that they become sizeable relative to supporting surplus or equity, then they generally have less of an ability to accept risk on the investment side of the house by moving into equities. A company such as Progressive, with premiums to surplus/equity ratios in the high 2+ area (sometimes close to 3.0) takes much of its risk on the underwriting side, so not surprisingly will not hold a sizable investment in equities. Fairfax has written premiums of about $29 billion, and equity including both preferred and common of about $24 billion. So they can be heavier into equities than a company like Progressive, and they are…with about $15 billion invested in equity-like instruments, associates, etc. If they anticipated opportunities to grow premiums dramatically, then they’d probably hold off on a further move from fixed to equities. But if they anticipated a soft market in which they even saw premiums shrinking, they might well choose to offset a reduction in insurance related risk with an increase in investment risk via a shift from bonds to equities.

From an economic standpoint, as long as Fairfax continues to prefer to hold equity investments for the long term and allow the gains to compound on the balance sheet unrealized, there should be no dramatic reduction in reported after tax income. But if as you’ve noted, the deferred tax liability would increase by $18 per share, the book value ought to drop by an equal amount. However, in my opinion, that is a somewhat misleading effect. A deferred tax liability is an interest free loan from the government which has to be paid back only when gains are realized. And a deferred tax asset is an interest free loan to the government that only is paid back to the company as future income is generated and able to be shielded from taxes because of the prepaid tax asset. Neither of these are present valued on the balance sheet. An asset that earns no interest is not worth as much in my opinion as a similar amount invested in 5% government bonds. Similarly, a no interest loan that might not have to be paid off until years in the future is not as large a liability as a similar amount borrowed by means of a 6% interest 10 year bond issue. Fairfax has about a $1 billion net deferred tax liability (interest free loan from the taxing authority) on the balance sheet at year end 2023. That is a sign of a savvy insurance company in my opinion, and if the liability grows to $1.4 billion after a tax change, it just makes the comparison with peer insurance companies that often record net deferred tax assets on their balance sheets that much more favorable.

I got a look at a prior report of Brett’s and noticed that he projects Earnings per diluted share…so presumably he converts actual book value per share at year end 2023 to a book value per diluted share instead. I think diluted shares are about 8% higher than actual shares. That may well explain the disconnect between his report and the reality that I care about personally.

Agreed! Thanks to @John Hjorth for reaching out to Andy!

I’m not an expert on surety bonds, but it can help to understand that these are probably quite different from a typical insurance risk transaction. I believe the provider of the appeal bond has to guarantee payment of the cash amount of the bond plus any accrued interest in the event that the appeal is lost. Generally speaking, the insurer wants a source of liquid collateral from the client at least equal to the amount they might have to pay. They basically just provide guaranteed liquidity to a client. It is the appellant who is supposed to pay the amount of any financial judgement, and an insurance company isn’t interested in providing coverage for the full amount of the award in exchange for a premium smaller than that. This is different from catastrophe reinsurance where the insurer accepts a relatively small premium compared to the total payout required against the risk of a catastrophe and so has to have an opinion on the probability of the event. As Greenberg’s letter noted, the surety bond provider does not actually assess the likelihood of the success of the appeal. Unfortunately, commercial real estate, which may or may not be unencumbered, is not viewed by surety providers as a desirable source of collateral. If an appeal fails, and the insurer has to pay the face value and accrued interest on the bond immediately, then they are stuck with having to collect on the collateral. Foreclosure on real estate is not high on the list of things that insurance companies want to deal with. The Chubb bond by contrast was secured with readily liquidated brokerage investments.

@Thrifty3000 : I’m not aware of how Fairfax’s share-based plans work. Is this something you can educate me on? Are they based on options grants or are shares given directly? Many companies do purchase sufficient shares to offset dilution resulting from share-based compensation, but Fairfax has been reducing share counts outstanding, and given the highlighting of Henry Singleton’s example of reducing Teledyne’s share count outstanding by something like 90% over his tenure, I think we can safely assume that the only way share counts might increase would be if additional shares were issued to help pay for an acquisition like Allied World in the future. Absent that possibility, I’m not personally worried about diluted per share earnings.