berkshire - cheap?

[MarioP]

With regard to how to value the Berkshire insurance float, - popping up again here now -, I still think the best place I have read about it is rb's post #89 in this topic of July 19^th 2017. rb's angle & way of looking at it just makes so much sense to me.

Thank you for remembering that John. What I found interesting is that since the formula was posted

D=F*c*(1-t)/r where

D=Discount to float face value
F=Float face value
c=coupon BRK would have to pay on debt
t=BRK tax rate
r= hurdle rate

F is higher; t is lower and c is higher

So the discounted float is gaining value faster than the float face

[John Hjorth]

Thank you, Mario,

Actually, I didn't just "remember" here. rb's post is actually like some kind of script chiseled into my mind, as a - by basis - logic and quite simple solution to a question that I at that time had been struggling with for years. Simplifying it, yes, but still giving it some structure, that my brain can relate to. Up to us as readers of the post - based on our own expectations, to create the probability range.

It's like being hit over the head by a cub.

CoBF is actually stuffed up with beautiful things like this.  One just has to recognize it when one read it. It's all searchable - you just have to remember the bright person. By doing so, we all get less dumb.

Still H/T rb here.

[AdjustedEarnings]

Sell-side analysts (and 'Mr. Market'?) at work... if you are GOOG, you can get credit for undistributed cash in your valuation and your unproven 'moonshots'. But, if you are BRK, you get no credit for undistributed cash (thankfully, now being put towards the buyback) and only partial credit for your investments.

https://www.cnbc.com/2018/10/17/jp-morgan-berkshire-hathaway-shares-look-really-cheap-using-buffett-method.html

[The Investor]

With regard to how to value the Berkshire insurance float, - popping up again here now -, I still think the best place I have read about it is rb's post #89 in this topic of July 19^th 2017. rb's angle & way of looking at it just makes so much sense to me.

What do you think about this post on another thread?
http://www.cornerofberkshireandfairfax.ca/forum/berkshire-hathaway/buffett-buybacks-could-berkshire-tender-stock/msg350069/#msg350069

To me that suggests the formula given massively underestimates the value of the float, or am I missing something?

Imagine you get to choose one of the following two options:
a) You get $1m cash, no strings attached.
b) You get $1m in cost free float. This float grows at 5% per year, and while you never get to own it, you own all investment income derived from it for 50 years.

Assuming an 8% rate of return in the above scenario, $1m cash grows to $1.08m after 1 year if you choose "option a", and $0 grows to $80k if you choose "option b". You make $80k on the $1m float, and the float grows to $1.05m. So next year you make $84k from the float, plus 8% on the $80k you made in the previous year.

It takes 17 years for profit derived from cost-free float to exceed capital compounded at 8% starting with $1million cash.

After 50 years, $1m cash compounded at 8% has grown to $46.9m.
However, the investment income from $1m float growing at 5% per year, compounded at 8% per year, has grown to $94.5m, with the float standing at $11.5m.

So a few key points that to me seem to be missing from the formula:

1) The float is likely to grow over time (and I remember Warren or Charlie saying that if it contracts for any period of time, it would be very unlikely to be at faster than 3% per year, although I can't remember the source).
2) It's an extremely stable, long-term source of funds that is of far better quality than most types of borrowing, as there often tends to be an element of getting the rug pulled out from under your feet at the worst possible time. The more you need the money, the more likely it is the lending institution wants it back, or tries to charge you more when it's time to refinance. That's why float is not the Faustian bargain that a high level of 'normal' leverage is.
3) The longer your time horizon, the more the float is worth (assuming the float will tend to get larger over a long period of time).

Interested to hear your thoughts!

[rb]

Thank you John for the kind words and the flattery. I'm glad i was able to add some value. That's why most of us are here. But it's still nice to hear when people find something useful.

Now a few thoughts.

That formula is meant to attach a discount to the face value of flat liability at a point in time. It's not dynamic, i.e. it doesn't account for float growth. It can be made dynamic fairly easily by building a DDM sort of model around it for float.

Mario, while I think that BRK has a good story I don't know if it's as good as you think. t going down is not good. Also r went up which is also not good. c went up which is good. But with c it's a little more complicated. c can be positively correlated or negatively correlated with the investment portfolio. Furthermore, over time c is negatively correlated with underwriting profit. So while higher c is always good for float discount it's a lot more tricky when it comes to overall valuation. Context is key here.

The investor, I think your framework is flawed. You don't get to do what you want with the float and I think that 8% is really optimistic for return on float. There are regulations around what insurance companies can do with float - see Berkshire's huge bond portfolio --that's not an accident. Also you don't get to have fun with the float for 50 years. You pay it out and have to try to raise new one. There are thousands of people each day asking for their float money back from Berkshire.

What you say about float being a superior type of liability is true. I would account for that by adjusting c higher. But I would be careful about the size of the adjustment. Keep in mind. The vast majority of companies did not have a problem refinancing themselves in 2008. Berkshire wouldn't have had a problem either. The fact is just that Berkshire wasn't and still isn't very levered up.

[Cigarbutt]

Thank you John for the kind words and the flattery. I'm glad i was able to add some value. That's why most of us are here. But it's still nice to hear when people find something useful.

Now a few thoughts.

That formula is meant to attach a discount to the face value of flat liability at a point in time. It's not dynamic, i.e. it doesn't account for float growth. It can be made dynamic fairly easily by building a DDM sort of model around it for float.

Mario, while I think that BRK has a good story I don't know if it's as good as you think. t going down is not good. Also r went up which is also not good. c went up which is good. But with c it's a little more complicated. c can be positively correlated or negatively correlated with the investment portfolio. Furthermore, over time c is negatively correlated with underwriting profit. So while higher c is always good for float discount it's a lot more tricky when it comes to overall valuation. Context is key here.

The investor, I think your framework is flawed. You don't get to do what you want with the float and I think that 8% is really optimistic for return on float. There are regulations around what insurance companies can do with float - see Berkshire's huge bond portfolio --that's not an accident. Also you don't get to have fun with the float for 50 years. You pay it out and have to try to raise new one. There are thousands of people each day asking for their float money back from Berkshire.

What you say about float being a superior type of liability is true. I would account for that by adjusting c higher. But I would be careful about the size of the adjustment. Keep in mind. The vast majority of companies did not have a problem refinancing themselves in 2008. Berkshire wouldn't have had a problem either. The fact is just that Berkshire wasn't and still isn't very levered up.

BTW rb, I also appreciate your inputs, especially when I agree with you, something perhaps I should do more often.
The constructive aspect is to build on something else or to bring another perspective, which may have a component of disagreement.

I like the way you picure float. Another way to value float is to combine its dual nature. One can see the value of the float as a portfolio of set aside funds recorded on the asset side and adjusted for by a contra account of liability reserves which can be discounted to a large and varying degree because of time value and also because of the potential cost (negative) and growth of float (both the liability side and the asset side). Every float profile is different. For instance, looking at numbers from a certain perspective, the Geico acquisition (49% of what was left) in 1995 implied a discount of + or - 50% on the float liability. Since then, float at Geico has compounded at about 7-8% per year, in a profitable manner. Recently, BRK completed the acquisition of MLMIC, a professional liability insurer based in New York. Looking at numbers, it seems that the acquired liability float was only slightly discounted. Interestingly, it is reasonable to expect comparable returns going forward after acquisition in both scenarios.

Comment on BRK's "huge" bond portfolio in the "reserved" funds recorded on the asset side. I see what rb means (cash, bills and bonds + or - match the liability reserves) but technically the "bond" category is very small at BRK  (about 15-16% of liability float). In fact, apart from Fairfax, I don't see any insurer even close for comparison in terms of financial flexibility if or when needed. In terms of valuation of the float (which can be done many ways), a possibility is to include an option value for this aspect which, for instance, may mean the ability to grow profitable float inversely to what others could do under certain scenarios.

[The Investor]

The investor, I think your framework is flawed. You don't get to do what you want with the float and I think that 8% is really optimistic for return on float. There are regulations around what insurance companies can do with float - see Berkshire's huge bond portfolio --that's not an accident. Also you don't get to have fun with the float for 50 years. You pay it out and have to try to raise new one. There are thousands of people each day asking for their float money back from Berkshire.

What you say about float being a superior type of liability is true. I would account for that by adjusting c higher. But I would be careful about the size of the adjustment. Keep in mind. The vast majority of companies did not have a problem refinancing themselves in 2008. Berkshire wouldn't have had a problem either. The fact is just that Berkshire wasn't and still isn't very levered up.

In the 1995 annual meeting there is a question regarding this (the video should start on 1:42:40):
https://youtu.be/G2k0cYDWgI0?t=6160

Warren is asked how much flexibility Berkshire has in investing float, and the answer is 'a lot'. "We are not disadvantaged by that money being in float as opposed to equity, really in any significant way. If we had a very limited amount of equity, and a very large amount of float, we would impose a lot of restrictions on ourselves..."

I interpret that as meaning that Berkshire's existing equity is so much larger than it needs to be to support claims, that they can effectively do whatever they want with float. Most insurance companies need to invest in low volatility investments (bonds etc.) because they can't afford to lose much money, operating on wafer thin equity to cover claims.

Presumably the regulations you mentioned are far more restrictive for companies with less equity/more float than they are for Berkshire? Warren suggests he would restrict activity voluntarily if it was necessitated by low equity relative to float.

If the quote about float and equity above holds true today, I don't think 8% is particularly optimistic.

Berkshire float has grown over time, and I think it is reasonable to assume it will continue to grow over time at a slower rate. If that's true, you do continue to get to have fun with the (increasing) float. Sure that might take a lot of work, but the same could be said of all the other businesses that keep producing over time.

[Swedish_Compounder]

The Investor - I am happy to see that you have made your homework and studied Buffetts and Mungers words on how to look at Berkshires float, which can not be compared with most other major insurance companies float.

We must conclude that there are no restrictions on the future float increase for the foreseeable future. Therefore the normalized float increase should be added to the free cash flow generation capacity of BRK. What is the right number? USD 8 Billion was mentioned by Buffett relatively recently if I remember correctly (I think it was during the annual meeting 2018).

This is unique and that is probably why many people have difficulties grasping it. Thus no complicated formulas for how to value the float increases are needed 

[John Hjorth]

... We must conclude that there are no restrictions on the future float increase for the foreseeable future. ...

The discussion lately in this topic is covered at length in the Berkshire 2017 Annual Report, Item 1A "Risk factors", p. K-22 - K-25.

In note 19, p. K-91 some of it is quantified this way:

(17) Dividend restrictions – Insurance subsidiaries

Payments of dividends by our insurance subsidiaries are restricted by insurance statutes and regulations. Without prior regulatory approval, our principal insurance subsidiaries may declare up to approximately $16 billion as ordinary dividends during 2018.

Combined shareholders’ equity of U.S. based insurance subsidiaries determined pursuant to statutory accounting rules (Surplus as Regards Policyholders) was approximately $170 billion at December 31, 2017 and $136 billion at December 31, 2016. Statutory surplus differs from the corresponding amount based on GAAP due to differences in accounting for certain assets and liabilities. For instance, deferred charges reinsurance assumed, deferred policy acquisition costs, unrealized gains on certain investments and related deferred income taxes are recognized for GAAP but not for statutory reporting purposes. In addition, the carrying values of certain assets, such as goodwill and the carrying values of non-insurance entities owned by our insurance subsidiaries, are not fully recognized for statutory reporting purposes.

[rolling]

While there might be little restrictions to the use of float, the truth is that for decades (up until recently) Berkshire has kept an amount of cash available+fixed income above or very similar to the float. As such, float has, for years, offered almost no leverage at all and, at most, offered optionality.

When asked about it in this year annual report Warren said "I had never thought about it that way" (I couldn't believe it when I read the transcript, it has become progressively more obvious). But unless he changed that since then (and he might have, he is a learning machine and truth be told, he has been very agressive in the market since then), we should not expect more than a minor return on float.

question 17, afternoon session, for those interested