petec

Thanks!

Do you have a source for this?

Yes. There are some good presentations on the investor website and you can make a case that Greece will start reflating soon which would be great news. However the ECB stress tests in May will be key. Even if capital levels are ok the ECB and Berlin may insist on dilutive capital raises - they want Greece to exit its bailout this year and they don't want it to collapse back into bailout in a couple more years. Now, you might argue that even if they issued 100% more shares you'd still only be at 0.5x book...but that's the risk.

+1 and a politer reply than I nearly wrote!

+1 Ultimately I think good, motivated underwriters can grow float by spotting opportunities less capable people don't. Berkshire is one of the best underwriters around - and Fairfax, thee days, is also superb.

Hi StevieV. My post was just an explanation of the theory in response to a question about the pref/warrant deals. I didn't actually meant to imply that leverage would grow. But it's a good question. First, there's every reason to believe that premiums will grow in line with nominal GDP or faster and therefore float ought to do the same. If you don't believe that then you must believe either that insurance premiums will shrink as a % of GDP or that Fairfax will lose share. Yet insurance is materially under-indexed in developing parts of the world so it is likely to grow faster than GDP globally - and Fairfax will capture more of that than many north American insurers because it does a material amount of business in those developing areas. Also, my base case would be that well run, disciplined underwriters might take share over time because they'll have equity when others are struggling. Fairfax has a good history of finding niches and growing in them, and I won't be surprised if Fairfax take share in places where existing business is small (LatAm, Eastern Europe) and in India, where Digit is a startup and could grow very fast. Second, we know they can write a LOT more premiums in a hard market - possibly as much as double. That obviously doesn't double float immediately, but it does grow it. Third, we know they will buy the stubs of Brit, Eurolife, and AWH. That will add to float. And while big acquisitions may be off the table, I fully expect continued tuck-ins. In the long run that may not be enough to maintain leverage - Fairfax's mix may shift towards operating buisnesses - but that hasn't been too bad for Berkshire!

They are using float. The holdco cash is probably going to be used to buy the OMERS stakes in Brit and Eurolife when those agreements come due. But I wouldn't necessarily differentiate between float and holdco cash. The point is that float levers your equity - you get to keep the investment returns on a far greater amount of money that what you invested in the business. Here's how I understand it: The leverage is derived from the fact that when you sell an insurance contract, the buyer effectively lends you money (premiums come in now, claims go out later, and you get to sit on the money in the meantime). The cost of the leverage is derived from the CR. The amount of the leverage is derived from the value of the premiums you write and the duration of the contracts. So for example, if you write $1bn in premiums every year on contracts that expire after a year (e.g. a typical house insurance contract) you'll have about $1bn in float, because you get to sit on one year's worth of premiums before the claims go out the door. If your CR is 100, then that leverage is cost free. Bit if you write $1bn in premiums every year on contracts that expire after 10 years, then (once the business is mature) you'll have about $10bn in float, because you get to sit on 10 years' worth of premiums. If the CR is 98%, then you're being paid to borrow that money. Of course, long tail is riskier than short tail - you can't predict the exposures so well and something you didn't foresee, like asbestos causing cancer, can come back to bite you - so you have to be very careful with underwriting. So, let's say you have $1bn in equity and you write $1bn in premiums every year on 3 year contracts. You'd have $4bn in investable assets ($1bn equity and $3bn float). If you can underwrite at 102% you're going to lose 2% of premiums a year in underwriting profit: $20m. If you can invest $4bn in a 7-year Seaspan debenture at 5.5% you're going to earn $220m in interest for a total of $200m in annual income, before holdco costs and tax. If you've also got warrants exercising at $6.50, and the Seaspan share price goes to $13, then towards the end of the debenture you're going to exercise the warrants for $8bn(!!), paying by forgiving the debenture. Not a bad return on your $1bn of equity. That's the impact of levering equity upside using float. A few points to note: - I have oversimplified the relationship between contract duration and float, but I think the basic principle holds. - Fairfax currently lever their equity about 2:1 using float, not 3:1. But they could probably double premiums in a hard insurance market. If you double premiums across the board it would take time for float to double, but logically you'd get there if the hard market lasted long enough. Unfortunately I suspect the regulator or the ratings agencies would panic long before they got to 4:1 levered, but there's scope for some growth. - underwriting profit in any given year is calculated off the $ of premiums, not float. That said, you can look at the float and know, if the average CR over all those contracts is 98%, that you're getting paid 2% to borrow. - float gives you huge investing leverage whether or not the CR is over 100% - it's just that the cost of that leverage is lower with a lower CR. That's why this model is so powerful for compounding if you can get both underwriting and investing right. - the debenture + warrant deals are great in theory because the downside is bondlike, so the regulator looks at these as bonds, but in most cases the warrant strike price is quite close to the share price at inception, and the shares look reasonably cheap, so there is near-full equity upside. That's powerful, if they can do it with a significant proportion of that big bond portfolio. If they can do $1bn a year on 5-7 year terms then they can basically convert $5-7bn of that bond portfolio into securities that have bond downside but equity upside. That more or less doubles their equity exposure but only on the upside. Get that right and lever it with float and the impact on shareholder's equity could be spectacular. I hope this helps but sorry if I am teaching grandma to suck eggs.

that would require fairfax india to list in india for the minority shareholders of these companies? not sure about the exact rules, but as an investor in the indian markets, i dont know of any foreign company being dual listed in india. the reverse exists where a few indian companies are listed in the US via ADRs I think he's saying that FFH would give the shares to FFH india in exchange for FFH india shares. Since they are both in Canada, I don't think there would be an issues? oh ok, got it. just change of control for thomas cook. Yeah I was merely wondering if they would "tidy up" the Indian holdings into FIH - partly to avoid the impression of a conflict of interest and partly to make my notes neater ;)

What are the odds of Fairfax's Thomas Cook and Quess stakes being swapped for Fairfax India shares at some point?

Reading all three letters I am really struck by the quality of managers these guys are able to attract. E.g. the guys they've hired to build Digit (who previously built Bajaj Allianz and then spent 5 years in Munich at senior levels in Allianz)) and the guy they have running Philafrica Foods (ex CEO Nestle China). There are impressive people just about everywhere.

Yes - extraordinary - and cheap!

Is there any evidence that they will sell it? I see this as a move to simplify and make the value obvious, but I don't see why they'd sell. It's a good business with a huge growth runway and they've been doing a lot of tuck-ins. I think this is a long term holding. Might be wrong.

The buyback history might also suggest it will be opportunistic. There is a very high chance of a material market selloff in the timeframe we are discussing (5-10y). That's when I'd expect to see material activity. Otherwise it will be an FCF-driven slow burn.

Agreed, but he hasn't made a commitment and they are value monkeys at heart. My guess is they will be very disciplined. I think of it more as a floor, like the Buffet 1.2x BV floor. That said he clearly thinks IV is well above book so it will be interesting to see where they stop!

I think the structure of the Brit deal with OMERS is: - They don't have to pay a dividend but if they do it all goes to OMERS up to a hurdle, and above that it goes to FFH holdco. - They can buy OMERS out over several years at cost plus an incentive which I seem to remember being about 7% a year. So if the BV compounds at >7% the p/bv falls and if not, it rises. AW has clearly had a rotten start to that but I suspect we are a few years from them buying the stub and it's got a great record of BV growth so time is on their side.

I think you're right and the same applies AWH eventually, I would think.

I think growler must mean something different in the US than it does here!

He's referenced Singleton and Teledyne before and in the letter. Teledyne issued to acquire for years and then turned the ship and bought back for years. The buyback is a long term thing. They can't put too much to work too fast - the stock isn't liquid enough, unless they do an SIB, but then they'd have to pay a premium - and they've never said they wanted to. Expect the share count to be much lower in 10 years but not necessarily in 10 months.

I think the culture is phenomenally valuable. It's very clear to me Fairfax attracts and retains superb people and is increasingly a company that people want to sell to. Changing that would be madness. But that's not the same as saying power should be heritable. Power should go to the person best able to protect the culture. I don't think I'd expect an operating profit (excl gains) in a big cat year, especially when interest rates are so low. I actually thought the underwriting results were phenomenal - if you look at the subsids FFH where FFH controlled the underwriting, i.e. excl AWH, FFH would have made an operating profit in a massive CAT year for the industry. That's extraordinary and not what I would have expected at all. AWH was awful but Prem was clear that they don't expect AWH's results to differ so substantially from FFH's going forward, which is a way of saying they'll influence the underwriting. The idea that one bad hurricane season somehow makes AWH - a company with a good history - a bad deal seems nuts to me. I suspect it'll be a home run. Could be wrong. As for the investing, far better that they focus on putting together a series of asymmetric outcome 1s and 2s than try to do something big. I love the details on those.

I'm fairly sure that as it is a >20% holding Fairfax would equity account it, reporting their share of Quess's net income and book value and ignoring the market price. That's why equity accounted associates contribute to the $1.2bn of unrealised gains that aren't in the book value currently. Quess would just add to that, I think. SJ - largely agree with you - not too bothered about Ben's $50m mandate but the appointment of a second child to the board worries me and the suggestion that it is done to defend the culture without any detail as to why she is the right person to defend the culture is insulting both to our collective intelligence as shareholders and, frankly, to the significant number of very bright people at FFH who are more deserving of that seat and whose experience of the culture is far greater. Very annoying. And if they are going into Singleton-style buyback mode as they say, at some point in the next decade those multivoters will have outright control.

+1 Very impressive.

...and even that part is fully rate-hedged, according to this letter.

Won't they equity account it and therefore still not record it in BV?

True, although the market clearly doesn't. AXA's market cap fell by about the market cap of XL when the deal was announced!

Thanks! I'd make the same guess. Here's hoping.