Berkshire Hathaway: update after the annual report

On the previous post on Berkshire Hathaway, I pointed my view on BRK valuation under the two-column method and the return it implied. Under that calculation, the expected return is somewhere around 10% a year. Taking into account that I expected year end 2015 intrinsic value under this formula to be about 181,5$ per B share. Now we have definite results:

Earnings (excluding underwriting gains): 11.186$ per A share

Investments:  159.794$ per A share

Which result in 271.654 per A share or 181,1 per B share

It seems I wasn't that far.

But this year Warren Buffett proceeded to update on his view of intrinsic value:

1- Underwriting gains are now fairly stable and should be considered for intrinsic value purposes. He event told us that while this year underwriting result was 1.118$ per A share the average result for the last 10 years was 1.434$ per share (telling us the normalized underwriting gains)

2- As happened before he excluded some intangible write off from the earnings, but he also told us that since BNSF implied spending in excess of depreciation at all times, we should consider BNSF earnings to be lower than they actually are (6.775 million pre-tax earnings or 4123.6 per A share with depreciation of 1,932 and capex of 5,651)

About this:

1-  The purpose of excluding underwriting gains for me was

a) how stable are those in case of catastrophe? Since for 13 years they have been positive I was inclined to believe that they were to be expected on average (negative underwriting income in a year should give them the possibility of raising prices which would compensate on the following years). Warren Buffett just confirmed that and since he has a better understanding of their exposures I believe I should confirm my prior belief

b) Investment deferred taxes and float were not being discounted and BNSF earnings in excess of depreciation (point 2) were not being discounted also. So excluding underwriting gains was an overly simplified way to do that. However, that was (knowingly) incorrect because it is inaccurate. 

2- He mentions spending in excess of depreciation but doesn't mention:

a) BRK is able to defer taxes to compensate (at least in a big part) for that, as long as he keeps spending in excess of depreciation  (which he tells us is forever)

b) A big part of that spending is growth capex be it to: create new routes, to increase existing routes capacity or to maximize efficiency and thus get better pricing than competition and steal business (about this last point he does mention the difference in prices versus the competition: under 3c per ton mile versus 4.2-5.3c; a huge difference)

To make the calculation more accurate I should instead of ignoring underwriting gains:

1-Take them into account

2- Discount the excess value of float. However:

a) float is growing, which simultaneously further reduces the chance of it having to be paid back and if it grows a lot we should, instead of discounting it, increase its value due to its growing ability

b) Float implies excess cash (point 3) and implies fixed income investments (instead of more equity investments).

3- Discount the excess cash needed: float implies at least a 20 billion cash cushion (which Warren has stated he views as at least 25 billion since going under 20 would make him take emergency measures to sell). It actually doesn't. He proceeded to explain that part of it was because of the financing method at Clayton Homes. In addition at least part of it is needed for other businesses regular working. But these 20 billion whatever they are needed for they should be discounted. However.

a) They should not be discounted in full because: the cash still belongs to Berkshire and gives a cushion to the investor that most other equity investments don't by allowing survival to 1 in 100 year events;  

4- Discount excess debt: not applicable

5- Discount investment deferred taxes: a good part of it might never be repaid because he views it as businesses purchased (just like BNSF). However taking them as nonexistent is somewhat aggressive

6- Discount BNSF earnings (which as explained before is quite complicated, or impossible).

1- Plus 14.340 per A share (23.560 million or 9.56$ per B share)

2 and 3- Probably 2 a) nets with 2 b) and 3 on zero, but it could go to either side

4- not applicable, the business could take some more debt (in fact we could assign a positive value to the capacity to increase leverage but they do not intend to increase it much so it is more conservative not to)

5 and 6- Are very hard to discount. The appropriate discount might be much lower than the value in 1, but I always found it more conservative to simply ignore, but I am probably being too conservative

However by changing the provided values Warren in fact updated his estimate of current intrinsic value, stating it to be higher than the standard two column method would say. Most likely he is the one who is right. Anyway, the price is still much lower than the two-column price, why should I worry if the intrinsic value is higher 9,5$/B share? And anyways, even if I took that into account that would probably change the 10% price to the 9,5% price or something similar.

ps: the share repurchase value has increased to 124.4$ per B share (very near the price it was trading at the time of my first post a month ago)

Disclaimer: I own BRK-b shares. This is not investment recommendation. Always do your own research. Always read the introduction post.