Message Font: Serif | Sans-Serif
 
No. of Recommendations: 4
Note 16 on page 95 of last year’s 10k indicates Berkshire current Income tax expense for 2018 was $6,5. Billion. In 2015 it was $ 5.4 Billion, 2014 it was $3.3 Billion. This indicates an effective tax rate of 19.1%, in 2016, 8.6% in 2015, and 6.4% 2014. These figures if they are correct make BRK world class in its ability to use the tax code to the advantage of its shareholders.

For three years the average would be 11.3%. A 43% reduction would thus reduce Berkshire’s effective rate something in the area of 6.5

At the statutory rate of 35% Berkshire’s tax expense would have been $11.8 billion in 1016, $11.8 billion minus-$6.5 billion means there were $5.3 billion in deductions and credits. In 2015 it was $12.2 billion minus $5.4 billion or $6.8 billion in deductions and credits.
To make things even more obscure there is the table on page 85 that shows Cash paid for taxes in 2016 was $4.7 billion. I do not know how explain the differences between page 85 and 95 and hope there some one here who knows how to explain this. And of course, it is possible that Berkshire actual cash expense was more than either the figures on page 85, and 95. Am I being delusional in suspecting that Berkshire would like to make their real effective tax rat as obscure as possible?

Going forward the tax chance is certainly not raise Berkshire’s effective rate, and since Tax credits come off the top. The Drop in Berkshire’s Effective tax rate should be proportionally greater that the 57% drop implied by the drop from 35% to 20%. IF that average rate for the last three years was 11.37% then we can project the effective rate in future (or at lease till other party resumes control of the government) will be well below 4,9 %, but since Credits come off the top these credits could drop the company’ effective tax rate well below that more.
Print the post Back To Top
No. of Recommendations: 14
Note 16 on page 95 of last year’s 10k indicates Berkshire current Income tax expense for 2018 was $6,5. Billion. In 2015 it was $ 5.4 Billion, 2014 it was $3.3 Billion. This indicates an effective tax rate of 19.1%, in 2016, 8.6% in 2015, and 6.4% 2014. These figures if they are correct make BRK world class in its ability to use the tax code to the advantage of its shareholders.

To get the effective income tax rate you need to take provision for income taxes and not just
current taxes. Both of these are different from actual cash taxes paid.

Provision for taxes (for 2016) was: $9.2B (your $6.5B is simply the current provision; actual cash
taxes paid was actually just $4.7B). So the effective tax rate was (on $33.6B of pre-tax income)
27%.

Provision/pre-tax income (all GAAP of course) is the correct thing for effective tax rate because
the only difference between this and actual cash taxes paid is one of timing. Over a long (in
Berkshire's case very very long) period of time cash taxes paid will match up with the provision.

To make things even more obscure there is the table on page 85 that shows Cash paid for taxes in 2016 was $4.7 billion. I do not know how explain the differences between page 85 and 95 and hope there some one here who knows how to explain this. And of course, it is possible that Berkshire actual cash expense was more than either the figures on page 85, and 95. Am I being delusional in suspecting that Berkshire would like to make their real effective tax rat as obscure as possible?

No. Berkshire's effective tax rate in 2016 was 27%. It owed 27% of its pre-tax GAAP income to
the government. However, what Berkshire has been very adept at is issuing an interest free IOU to
the government with a very far out maturity and no interest. This is very different from say an
Apple that avoided having the tax liability in the first place.

It's float. It may not be coming due for a loong time, but it still is something owed.

Going forward the tax chance is certainly not raise Berkshire’s effective rate, and since Tax credits come off the top. The Drop in Berkshire’s Effective tax rate should be proportionally greater that the 57% drop implied by the drop from 35% to 20%. IF that average rate for the last three years was 11.37% then we can project the effective rate in future (or at lease till other party resumes control of the government) will be well below 4,9 %, but since Credits come off the top these credits could drop the company’ effective tax rate well below that more.

A more accurate, but still rough exercise, is to consider what the 2016 provision would have been at a 21% statutory rate:

Pre-tax income: $33.6B
Tax at 21%: $7B

Adjustments (this is why this is still rough, because these adjustments depend a bit on the statutory rate plus (deductions)):

Dividends received deduction: (.78B)
State taxes: .36B
Foreign tax rate differences: (.42B)
US tax credits: (.51B)
Non-taxable exchange: (1.1B)

For a provision of: 7 - .78 + .36 - .42 - .51 - 1.1 = $4.55B

I.e., an effective tax rate of 13.5%.

So the 21% tax rate dropped to 13.5%.

What the cash tax rate will be is a whole different ball game, since that is going to be heavily
affected by accelerated depreciation for taxes and the like. However, as before cash taxes vs the
GAAP provision is a matter of timing differences. The timing is only a big deal if there is a lot of
time value to money (which has to do with prevailing interest rates, availability of investments at
which capital can be deployed at and the like).
Print the post Back To Top
No. of Recommendations: 5
That's a whole lot of detail, and I laud you for the effort.
As you note, a lot of those credits are themselves perhaps a function of the nominal tax rate, so I
didn't bother doing the work as I figured it wouldn't be right enough to draw any conclusions. Bravo.

But you can learn a lot at the high altitude, too.
I like to start with KISS: on the top line income statement, look at income tax expense line divided by income before taxes line.
That's because I believe GAAP accruals usually give a good sense of the general economic reality,
and the gap between that and cash is best thought of as part of the float discussion.
It's not a good reflection of normal reality when there is a discontinuity in the deferred tax rate, of course, but we can look at that separately.

I think about these numbers first, then what I know that would change them materially.

2000    36.1%
2001 42.2% (anomalously high outlier)
2002 33.2%
2003 31.7%
2004 32.6%
2005 32.5%
2006 32.8%
2007 32.7%
2008 26.1% (anomalously low outlier)
2009 29.5%
2010 29.4%
2011 29.8%
2012 31.1%
2013 31.1%
2014 28.2%
2015 30.1%
2016 27.4%
Q3 28.8% (trailing four quarters)

Ignoring the two outliers noted, average 1998-2007 was 33.4% and average 2009 to date has been 29.5%.
I haven't delved into the reason for the one time drop, but I imagine it's rail related tax credits.

Jim
Print the post Back To Top
No. of Recommendations: 5
The basic point is that for financial reporting purposes (10-K/GAAP), expenses (such as taxes) must
be accrued for when they are incurred regardless of when the cash actually leaves.

So differences between cash taxes and GAAP provision will reflect timing differences.

For instance, say you spend $1B in constructing a new pipeline. For financial accounting purposes,
you can't just expense the $1B immediately. You gotta expense it over the revenue generating life of
the pipeline (depreciation schedule, which in this case might be something extending over 40 years).
The cash, however, went out of the door immediately. Accounting/10-K/GAAP depreciation is a non-cash
expense, but it's actually a terrible thing - it's reverse float!

So for financial reporting purposes, when you actually earn some revenue on the pipeline, the
depreciation (according to the mandated schedule) will be deducted and a provision for tax
calculated based on the prevailing statutory rate on the difference (modulo other credits and
deductions you might be eligible for). The recognition of revenue, depreciation expense and tax
expense might differ in timing by a substantial gap from actual cash inflows/outflows (already
demonstrated with the case of depreciation - think about that depreciation expense in year 39).

Side note: One often conflates "depreciation" with "maintenance expense". The two are often related,
but the accrual basis for "depreciation" isn't to account for maintenance expense (it is to match
the initial capital outlay on the asset with revenue generated by the asset).

Now the tax man might decide to let you expense the full capital outlay for the pipeline immediately
(accelerated depreciation). See financial reporting/GAAP is designed to match revenues with
expenses. The tax code is designed to often do completely different things. So you can get HUGE
timing differences.

Regardless, the timing differences do not change the effective amount of tax (measured by the GAAP
provision in the 10-K) owed, just when you actually have to pay them.
Print the post Back To Top
No. of Recommendations: 7
But you can learn a lot at the high altitude, too.
I like to start with KISS: on the top line income statement, look at income tax expense line divided by income before taxes line.
That's because I believe GAAP accruals usually give a good sense of the general economic reality,
and the gap between that and cash is best thought of as part of the float discussion.
It's not a good reflection of normal reality when there is a discontinuity in the deferred tax rate, of course, but we can look at that separately.


Yep. The same high level/quick analysis can even be done to assess the roughly right effect of tax
changes to book value (if you have net tax liabilities of $100 that were computed by assessing taxes
at 35% and now statutory rates drop to 21%, it's middle school arithmetic to figure out what the new
liabilities are).

I find it mind boggling at one level and disheartening at another level as to the amount of teeth
gnashing on effects of tax code changes to financial statements, book value, and the like. These are
easy to figure out (roughly enough to be good enough), all that one has to do is open the 10-K (the
requisite accounting knowledge is completely basic and something one should have if investing in
individual equities). The important question is to the actual economic/reality change in the
intrinsic value of the business which to some extent is a quite removed from the changes to the
financial statements.

Didn't Papa Buffett himself do this satirical exercise of accounting changes that changed statements
by night and day but left economic reality the same for a hypothetical (steel?) company many years
ago in an annual letter?
Print the post Back To Top
No. of Recommendations: 3
However, as before cash taxes vs the GAAP provision is a matter of timing differences. The timing is only a big deal if there is a lot of time value to money (which has to do with prevailing interest rates, availability of investments at which capital can be deployed at and the like).

Well stated TransverseSlice, and usually not appreciated.

Unless the tax difference between book due and current paid can earn good money, either through being invested in equities/businesses or by earning good interest rates, the impact - taking into account inflation and the time value of money - the impact is pretty small.

Especially when you already have over $100 billion to be invested, and more coming each month.

The issue to focus on with Berkshire is the ability to put money to work - not its availability.

That's where discussion needs to be focused.
Print the post Back To Top
No. of Recommendations: 1
Provision/pre-tax income (all GAAP of course) is the correct thing for effective tax rate because
the only difference between this and actual cash taxes paid is one of timing. Over a long (in
Berkshire's case very very long) period of time cash taxes paid will match up with the provision.


Thanks for your explanation this and your other comments very nicely explain the reason for the difference between note 13 and note 16.

Provision for taxes (for 2016) was: $9.2B (your $6.5B is simply the current provision; actual cash
taxes paid was actually just $4.7B). So, the effective tax rate was (on $33.6B of pre-tax income)
27%.


When I divide 33.6 by 4.78 I get an effective tax rate 14% not 27%
The figures for 2015 and 2014 14.6% and 13% this gives us a much average of 13.7%. Which makes more sense to me as I think I remember Buffett answering a question at an annual meeting to the effect that Berkshire effective rate was around 13%.

14% times a 57% tax cut (7.98%) still leaves us with an effective rate below 6% going forward.

Is there any way estimating how much of the difference between the statutory rate and the effective rate is deductions and how much is tax credits?
Print the post Back To Top
No. of Recommendations: 3
When I divide 33.6 by 4.78 I get an effective tax rate 14% not 27%
The figures for 2015 and 2014 14.6% and 13% this gives us a much average of 13.7%. Which makes more sense to me as I think I remember Buffett answering a question at an annual meeting to the effect that Berkshire effective rate was around 13%.


First, let me recount the figures (for my sake) for 2016:

Pre-tax income: $33.6B
Income tax expense: $9.2B

Of this $9.2B, about $6.5B is recorded as current.

Cash taxes paid during the period (the year 2016): $4.7B

The effective rate is $9.2B divided by $33.6B, i.e., 27%, not 14%.

The cash taxes paid number of $4.7B (which is what you want to use in the numerator to get the 14%) is the cash
outflow that occurred in 2016 for income tax payments. It has less to do with taxes on taxable
income in 2016 and more to do with taxable income in 2015 (but even that is not the full story; see below).

Now maybe what you are confused about is that the taxes paid in 2016 ($4.7B) should roughly be equal
to the current portion of taxes in the year end 2015 financial statements. This latter number is
$5.4B. This discrepancy is probably due to minor differences of when the cash actually left
the company (think about your own income tax expenses).

For instance, you might expect the total cash payments in 2016 and 2015 for income taxes to not
have much of a discrepancy between the current portion of tax expense recorded year end 2015 and
2014.

The former (cash taxes paid) is: $4.5B + $4.0B = $8.5B
The latter (current portion of income tax expense) is: $5.4B + $3.3B = $8.7B

However, this is not the full story. If you looked at very long periods in time, the cash taxes
paid number will (or would in the future) start exceeding the total current portion of income tax expense
(because the deferred amounts would eventually have to be paid).

The point is the same as before: the aggregate cash taxes paid will match the aggregate income tax
expense recorded in financial statements. It is just a matter of how long a period it takes for the
timing differences to manifest.

You are confusing effective tax rate (what Berkshire owes the government as a percentage of its
income, regardless of when the cash outflow occurs), with actual cash outflows occurring in any given period.

Think about it this way. Look at an "inverted" situation. Should unrealized capital gains not be
counted as increases in book value just because the cash inflow hasn't occurred (oh but maybe it
will occur relatively soon... really... look at the KO holding =D).

Is there any way estimating how much of the difference between the statutory rate and the effective rate is deductions and how much is tax credits?

Yes. I have to give a snarky response here: are you actually reading p.95 of the 10-K?
Because it is right there in the only table on the page, kinda hard to miss.
Print the post Back To Top