2018 Valuation

[cubsfan]

Pupil - wow, what a great explanation. Thank you!

[Dynamic]

Seconded. Thanks pupil

[thepupil]

thinking about it some more, the key is actually can each of those (BE and BNSF) outspend their tax depreciation, not their accounting depreciation....I think.

Anyways, I know why the difference exists, but am not rigorous enough to project how that will change over time, because I more or less choose to value those subs on comps at market value in base case NAV and haircut them in the bear case NAV.

[Okonomen]

The delta between the cash paid for taxes and income tax expense relates to the increase in the deferred tax liability.

the deferred tax liability has 2 major components:

1. Increase in DTL related to unrealized gains on investments: To the extent that Berkshire defers realization of gains indefinitely on large positions (I think this applies to some, but not all of the material equity positions), we can assume that this tax expense has very low present value.

2. Increase in DTL related to the companies, primarily  Burlington Northern and Berkshire Energy.

Page 96 of the most recent 10-K shows the breakdown between the two (obviously things will change in the subsequent quarters, but it gives you a feel). $18 billion of the $60 billion DTL ( about 30% again as of 12/2018) is related to unrealized gains on the equity portfolio. About 50% ($28/$60 billion) relates to property plant and equipment. This is associated with Berkshire Energy and Burlington Northern. This is helpful
https://ftalphaville.ft.com/2016/04/29/2160510/warren-buffett-a-dream-deferred/

Essentially, the key input for determining the "correct" cash tax rate for the railroad and utility is if they are able to keep increasing their ability to re-invest. Will they keep the asset base high such that they keep running on the treadmill of outspending depreciation expense.

Burlington Northern's D&A is about $2.3 billion / year. It's capex has been $2.6B (2010) - $5.6B (2015) and averaged above $3B or so, so a delta between the cash tax rate and income tax expense rate appears sustainable for now.

Berkshire Energy's depreciation is running at about $3 billion. It's capex is running at $6 billion (2010: $2.5 billion, 2014: $6.5 billion, LTM: $6 billion). Again there appears to be a sustainable difference between capex and depreciation.

I think that explains the high level, but I could be wrong on the exact particulars.

Honestly, I don't pay a whole lot of attention to it. I just use what the market values class 1 railroads and large high quality utilities as proxies for value for each of those and assume that Berkshire doesn't sell all of its unrealized investments at a gain very quickly. My rang of intrinsic value is squishy and my sizing is non-binary so being super precise on that hasn't been an emphasis for me.

Thanks for a great explanation. On (1) it seems like in a valuation one should not subtract the DTL on investments as it seems fair to assume that BRK will never ever sell their portfolio securities?

On page K-96 in the BRK AR 2018 we can see the detailed deferred tax net liability.

If I may ask, what would you assume in a BRK valuation? 25% tax on current pretax earnings + subtraction of the net tax liability? Some day it will reverse due to accelerated depreciation that cannot go on forever, but this will also boost accounting earnings. I just think this sounds somewhat too conservative

[thepupil]

I would value BNSF and BE separately and not have to deal with this issue outside of that. I value them at UNP and XLU multiples (earnings/revenue/tangible book/book etc.) and multiply by 0.5x and 0.75x for scary bear case and bear case. 1.0x for "the market's assessment for what this type of business is worth". Throughout my ownership of Berkshire (about 8 years) Berkshire has traded at a discount to NAV (when I mark up BE and BNSF) and been growing earnings faster than the market. I piggy back off others work for thoughts on the quantum of that discount.

I discount the DTL related to securities almost completely. KO and WFC won't be sold. losses from new stock picks can be realized to realize other winners). Berkshire is my and my family's largest postition, so I care a lot about the company, but I jsut don't spend time on figuring this type of stuff out (25% tax rate or 18% tax rate or 10% tax rate.

Sorry. the US corp rate is 21%, so I'd just take that and apply some haircut.

valuation is a range, not a point, as are earnings.

[Okonomen]

I would value BNSF and BE separately and not have to deal with this issue outside of that. I value them at UNP and XLU multiples (earnings/revenue/tangible book/book etc.) and multiply by 0.5x and 0.75x for scary bear case and bear case. 1.0x for "the market's assessment for what this type of business is worth". Throughout my ownership of Berkshire (about 8 years) Berkshire has traded at a discount to NAV (when I mark up BE and BNSF) and been growing earnings faster than the market. I piggy back off others work for thoughts on the quantum of that discount.

I discount the DTL related to securities almost completely. KO and WFC won't be sold. losses from new stock picks can be realized to realize other winners). Berkshire is my and my family's largest postition, so I care a lot about the company, but I jsut don't spend time on figuring this type of stuff out (25% tax rate or 18% tax rate or 10% tax rate.

Sorry. the US corp rate is 21%, so I'd just take that and apply some haircut.

valuation is a range, not a point, as are earnings.

Thanks for the insights!

What I don't get is why Pershing Square, who recently initiated a position, thinks that nearly half of BRK's IV is from their insurance ops. When I do the calculations I get no where near that value for the insurance ops. They generate around 2 bio. USD in pretax earnings (underwriting profits) through Geico, GenRe, BH etc and by looking at peers a 17x multiple seems fair on their after tax earnings so approx 25 bUSD assuming 25% tax. Last comes the float where I only look at it as a fixed income play and I do not consider dividends from their equity portfolio. So let's say: 123 bio. USD float generating 2% after tax interest and the float grows by 7% a year so that makes: (123 bUSD * 0,02)/0,07 = 35 bio. USD. So a combined 60 bio. USD and thus no way near half of the IV. Am I missing some major insights regarding their insurance ops?

[thepupil]

they are including the equity portfolio, fixed income portfolio and cash,  as part of the insurance operations.

Since the end of 2007, we estimate that Berkshire has averaged a nearly 7% annual rate of return on its insurance investment portfolio while holding an average of 20% of its portfolio in cash. Berkshire has been able to produce investment returns that significantly exceed its insurance company peers as the combination of the company’s long-duration float and significant shareholders’ equity allow it to invest the substantial majority of its insurance assets in publicly traded equities, while its peers are limited to invest primarily in fixed-income securities. We believe these structural competitive advantages of Berkshire’s insurance business are enduring and will likely further expand.

[Okonomen]

they are including the equity portfolio, fixed income portfolio and cash,  as part of the insurance operations.

Since the end of 2007, we estimate that Berkshire has averaged a nearly 7% annual rate of return on its insurance investment portfolio while holding an average of 20% of its portfolio in cash. Berkshire has been able to produce investment returns that significantly exceed its insurance company peers as the combination of the company’s long-duration float and significant shareholders’ equity allow it to invest the substantial majority of its insurance assets in publicly traded equities, while its peers are limited to invest primarily in fixed-income securities. We believe these structural competitive advantages of Berkshire’s insurance business are enduring and will likely further expand.

Thanks for the input! So it's insurance ops earnings (around 1,5 bUSD after tax with multiple X) + return on float (around 3 bUSD ex portfolio dividends with multiple X) + market portfolio value (215 bUSD) + cash (around 100 bUSD excess cash) and thus with no regards to the float value of 123 bUSD other than what it can produce in returns afaik