You are on page 1of 15

34

I
n a paper wri t ten l ast year, a
group of hedge f und profes-
sionals and academics claimed
to have discovered how famed
investor Warren Buffett makes
his money. The outstanding returns expe-
rienced by Berkshire Hathaway (Buffetts
firm) can be explained by two main fac-
tors: (1) wise investments in underval-
ued, safe, blue-chip securit i es and (2)
extremely agreeable funding terms lead-
ing to economical leverage. By punting on
temporarily cheap assets with lots of bor-
rowed funds, and by being able to borrow
cheaply, Buffett has been able to reach
legendar y status among the investment
community. In the words of the papers
authors: Buffett has developed a unique
access to leverage that he has invested in
safe, high-quality, cheap stocks and these
characteristics largely explain his impres-
sive performance.
1
Here we focus on the f undi ng side of
the equat ion, l eavi ng the stock-picki ng
prowess analysis to others. Where is Berk-
shi re get t i ng t hat vast and af fordabl e
f unding f rom? How is Buffett being able
to erect the wal l of economical l everage
that makes his returns so mouth water-
i ng? Si mply stated: by bei ng wi l l i ng to
take on a l ot of risk. For a fee, of course.
Berkshi re sel l s i nsurance and rei n-
surance pol icies i nto the fi nanci al mar-
kets. It also sel ls derivatives. Al l of those
sales generate (for the most part, upfront)
premiums f rom those purchasi ng pro-
tect i on f rom Berkshi re. Those premi -
ums can amount to a ver y l arge sum.
Buffett then invests that money, an activ-
ity that should lead to interesting returns
given his t rack record. Given that a l ot
of the sold insurance pol icies and deriv-
at ives cont racts may take a l ong whi l e,
i f at al l , before Berkshi re has to make
any l oss payouts, Buffett can make good
use of t he premi um
col l ect ed for many,
many years. The hope
i s t hat any event ual
l oss payment is both
l ower t han t he pre-
mi um i ni t i al l y col -
l ected and l ong to come. If Berkshi re
breaks even, that is if the eventual insur-
ance claims and derivatives payouts equal
the amount of premium received, Berk-
shi re woul d have received the equiva-
l ent of zero-cost fi nanci ng for al l that
period of time (plus any returns obtained
f rom i nves t i ng t he premi ums ) . Were
Berkshi re to actual ly enj oy under writ-
i ng profits (payouts l ower than the pre-
mi ums), t he company woul d have, i n
ef fect , enj oyed negat i ve cost f undi ng.
This is what the papers authors mean
when they state that Buffett enj oys the
si gnificant advantage of havi ng unique
access to steady, cheap, l everage. In the
Sage of Omahas ver y own words:
If our premiums exceed the total of our expenses
and event ual l osses, we regi st er an under-
wr i t i ng prof i t t hat adds t o t he i nvest ment
income our float produces. When such a profit
is earned, we enj oy the use of f ree money
and, better yet, get paid for hol di ng it. Thats
like your taking out a loan and having the bank
pay you i nterest.
2
The difference between the premiums
col l ected and the l oss payments made
( i f any) i s cal l ed f l oat . Ber s khi res
prowess, t hus, woul d be based on t he
tremendous amounts of float it can gen-
erate. Accordi ng to the papers authors,
36 percent of Berkshires liabilities come
f rom i nsurance fl oat, on average. Berk-
shi re does not seem to i nclude deriva-
t ives-generated fl oat under the overal l
insurance float number, so the final num-
ber may be even greater. Exhibit 1 i l lus-
t r at es t he es t i mat ed annual cos t of
Berskhi res i nsurance f l oat si nce 1976
(2. 2 percent on average, 3 percent age
points below the average Treasury Bill rate;
not ice how the company seems to have
been gett i ng better at it as of l ate).
And Berkshi res fl oat has been grow-
i ng s pect acul ar l y t hrough t he years ,
mat chi ng t he companys s pect acul ar
growth. If float was $39 mi l l ion i n 1970,
it had jumped to $1. 6 bi l l ion by 1990, to
PABLO TRIANA is a professor at ESADE Business School and
the author of The Number That Kil led Us: A Stor y of Modern
Banking, Flawed Mathematics, and a Big Financial Crisis.
P
A
B
L
O

T
R
I
A
N
A
P
R
A
C
T
I
C
A
L

M
A
T
T
E
R
S
WHY DOES WARREN
BUFFETT MAKE MONEY?
CORPORATE FINANCE REVIEW NOVEMBER/DECEMBER 2013
.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.






.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.






35
PRACTICAL MATTERS NOVEMBER/DECEMBER 2013 CORPORATE FINANCE REVIEW
$27. 87 bi l l i on a decade l at er, and t o
$73.12 billion by 2012 (again, these num-
bers may not i nclude derivat ives-gen-
erated fl oat, onl y i nsurance-generated
fl oat). Thats a l ot of ver y cheap (even
negat ive cost) f undi ng.
Ver y few ot her i nsurers seem to be
able to achieve those t ypes of outcomes.
While Berkshire has generated an under-
wr i t i ng prof i t f or t he pas t 10 year s
st rai ght, compet itors dont appear to be
abl e t o boas t s i mi l ar l y ros y res ul t s
(report i ng, i n fact, under writ i ng l osses
as a whol e). Listen to Buffett expl ai n it:
Let me emphasize that cost-f ree float is not an
outcome to be expected for the [i nsurance]
i ndust r y as a whol e: There is ver y l ittl e Berk-
shi re-qual it y fl oat exist i ng i n the i nsurance
world. In 37 of the 45 years ending in 2011, the
i ndust r ys premiums have been i nadequate to
cover cl ai ms plus expenses.
3
Berkshi res st rategy has been l abel ed
as betting against Beta, after the famous
i nvest ment risk measurement vari abl e.
You buy low risk (low Beta) assets and
you sel l hi gh ri sk (hi gh Bet a) ones,
hoping that the former wil l do wel l while
the l atter do badly. Insurance and rei n-
surance policies, including on very exotic
underl yi ngs, are a way of maki ng that
bet. Derivatives are another. While many
would be expected to be familiar with Berk-
shires insurance forays, they may be less
so wi t h hi s deri vat i ves t rades. In t hi s
art icl e, we focus on this l ess-known l eg
of Warren Buffetts search for fl oat.
Derivatives games
Berkshire Hathaway began sel ling equit y
i ndex put opt ions and credit protect ion
through credit default obligations (credit
default swaps and the l i ke) i n 2004. At
t he t i me, Berkshi re Hat haway al ready
held a ver y substant ial derivat ives port-
fol io, l egacy of the acquisit ion of rei n-
surer General Re. Berkshire had embarked
on a st rategy to wi nd down the General
Re derivatives book, which included a myr-
i ad of products and underlyi ng assets,
more than 23, 000 cont racts outst and-
i ng. For i nst ance, as of December 31,
2003, Berkshi res derivat ives port fol io
included $11 bi l l ion in forei gn currency
EXHIBIT 1 Buffetts Cost of Leverage: The Case of His Insurance Float
Fraction Average
of years with cost of funds
negative cost (truncated)* Spread over benchmark rates
Fed Funds 1-Month 6-Month 10-year
T-Bill Rate Libor Libor Bond
19761980 0.79 1.67 4.59 5.65 5.76
19811985 0.20 10.95 1.10 0.27 1.28
19861990 0.00 3.07 3.56 4.61 4.80 4.90 5.30
19911995 0.60 2.21 2.00 2.24 2.46 2.71 4.64
19962000 0.60 2.36 2.70 3.10 3.33 3.48 3.56
20012005 0.60 1.29 0.82 0.96 1.05 1.19 3.11
20062011 1.00 -4.00 -5.84 6.06 6.29 6.59 7.67
Full Sample 0.60 2.20 3.09 3.81 3.69 3.88 4.80
* In years when cost of funds is reported as less than zero and no numerical value is available, cost of
funds is set to zero.
Data from: Frazzini, A., Kabiller, D., and Pedersen, L.H., "Buffett's Alpha," (May 3, 2012). The data are hand-
collected from Buffett's comment in Berkshire Hathaway's annual reports. Rates are annulaized, in percent.
36
CORPORATE FINANCE REVIEW NOVEMBER/DECEMBER 2013 PRACTICAL MATTERS
for wards, $333 bi l l i on i n i nterest rate
and currency swaps, and $102 bi l l ion i n
interest rate and currency options. These
cont ract s (bot h l ong and shor t expo-
sures) generated assets and l i abi l it ies in
similar amounts (about $15 bil lion each,
$10 bi l l ion if you al low for counterpart y
netting). A year later, the legacy portfolio
had al ready been wound down si gnifi -
cantly, with swaps notional (including now
credit products as wel l as i nterest rates-
currency) just at $153 bi l l ion and i nter-
est rates-currency opt i ons j ust at $35
bi l l ion. By December 31, 2005, with just
740 contracts left outstanding, the respec-
t ive numbers were $44 bi l l ion and $14
bi l l ion (currency for wards remai ned at
around $13 bi l l ion i n si ze); by Decem-
ber 2006 Berkshi res deri vat i ves book
had become domi nated by t he equi t y
index puts and credit default obligations
posit ions, with i nterest rates-currency
swaps at $10 bi l l ion, i nterest rates-cur-
rency opt ions at $4 bi l l ion, and forei gn
currency for wards at $1 bi l l ion. By the
begi nni ng of 2008, the l egacy portfol io
had been essentially liquidated and essen-
t i al ly al l of Berkshi res derivat ives book
consisted of the equity puts and the credit
obl i gat i ons. The unwi ndi ng had been
cost l y, wi t h l osses of more t han $400
mi l l ion by year-end 2005.
The equi t y put s and credi t def aul t
obl i gat ions posit ions were bui lt sl owly
at fi rst and more i ntensely l ater on. By
year-end 2004, the not ional si ze of the
put cont r act s was around $4 bi l l i on,
growi ng to $14 bi l l ion a year l ater and
.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.






EXHIBIT 2 The Evolution of the Notional Size of the Puts and the Credit Contracts
!
#!!!
$!!!!
$#!!!
%!!!!
%#!!!
&!!!!
&#!!!
'!!!!
'#!!!
()
%!!#
()
%!!*
()
%!!+
,$
%!!-
,%
%!!-
,&
%!!-
()
%!!-
,$
%!!.
,%
%!!.
,&
%!!.
()
%!!.
,$
%!$!
,%
%!$!
,&
%!$!
()
%!$!
,$
%!$$
,%
%!$$
,&
%!$$
()
%!$$
,$
%!$%
,%
%!$%
,&
%!$%
()
%!$%
,$
%!$&
,%
%!$&
,&
%!$&
!"#"$%& ()*+ ,- .)&&)"$/ )/0123 4025
678912 :8;<0=2 >?=1@<ABC5
!
#!!!
$!!!
%!!!
&!!!
'!!!!
'#!!!
'$!!!
'%!!!
'&!!!
#!!!!
()
#!!*
()
#!!%
()
#!!+
()
#!!&
,'
#!!-
,#
#!!-
,.
#!!-
()
#!!-
,'
#!'!
,#
#!'!
,.
#!'!
()
#!'!
,'
#!''
,#
#!''
,.
#!''
()
#!''
,'
#!'#
,#
#!'#
,.
#!'#
()
#!'#
,'
#!'.
,#
#!'.
,.
#!'.
!"# %&'&()* +, -.**.&(/ /012 (0345 6753839
6750:05;<4 =>?@>?<A3
/012 (0345 B =>?@>?<A3
CA<A39DE;70F0@<40G39
.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.






37
PRACTICAL MATTERS NOVEMBER/DECEMBER 2013 CORPORATE FINANCE REVIEW
to $21 bi l l ion by December 2006. The
position reached its pinnacle notional size
of $3540 billion in late 2007 to early 2008
and kept more or less constant at that level
from then on (save for a smallish unwind-
i ng several years l ater). Not ional si zes,
expressed i n dol l ars, take i nto account
currency exchange rates.
The notional size of the credit default
obl i gat ions was $2. 8 bi l l ion by year-end
2005, $2. 5 bi l l ion a year l ater, and $4. 6
bi l l i on i n December 2007. Up to t hat
point, Berkshire had sold protection only
on Amer i can hi gh- y i el d cor por at e
indexes (about 100 names per contract).
From 2008, the company not only si g-
nificantly ramped up such act ivit y, but
al so began to sel l protect i on on i ndi -
v i dual cor por at e names and on
st at e/ muni ci pal i t i es, wi t h t he conse-
quent drast ic i ncrease i n total not ional
amounts, which reached a high of $30 bil-
l ion by year-end 2008 (si zes decreased
f rom that point, due to a combination of
cont ract expi rat ions and cancel l at ions,
al l the way to l ess than $10 bi l l ion by
late 2013; essential ly al l exposures today
come exclusively f rom the state/munic-
ipal it ies cont racts).
As of year-end 2006, Berkshi re had
sol d 62 equit y puts and credit default
cont racts; this number went up to 94 a
year l ater and to 251 by year-end 2008
(down to 203 by year-end 2010, fol l ow-
i ng the expi rat ion of the fi rst cont racts,
some unwi ndi ngs, and the fact that no
new cont racts were bei ng written). The
l ast equi t y put cont racts were sol d i n
Februar y 2008. The last credit contracts
were written i n Februar y 2009 (just one
new cont ract).
Exhibit 2 displays the evolution of the
not ional si ze of the puts and the credit
cont racts. The equit y i ndex puts were
European (can only be exercised at matu-
ri t y), were st ruck at-t he-money (t hus
affording a ver y tast y premium for Berk-
shi re, as these opt ions are ver y cl ose to
havi ng posi t i ve i nt ri nsi c value), were
wr i t t en on f our i nt er nat i onal equi t y
indexes (S&P 500, FTSE 100, Euro Stoxx
50, and Ni kkei 225), and had t he fol -
l owi ng expi rat ion dates: between Sep-
t ember 2019 and Januar y 2028 ( t he
weighted average life of al l put contracts
was approxi mately 7. 5 years at Septem-
ber 30, 2013).
Originally, 47 put contracts were sold,
generating $4.9 billion upfront premium.
The maxi mum possibl e payment f rom
Berkshi re on these cont racts equal s the
put s not i onal si ze (cur rent l y around
$32 bi l l i on), but t hi s woul d onl y t ake
pl ace at cont ract expi rat ion and only if
al l the i ndexes reach a value of zero at
t hat t i me. The l i kel i hood of t hat i s
severely l i mited. Were Berkshi re forced
to make payments equal to $4. 9 bi l l ion,
it woul d break even on the puts (plus
any i nvest ment returns on the fl oat). A
25 percent drop in al l the equit y indexes
by expi rat i on date woul d yi el d a l oss
payout of around $8 bi l l ion, at current
forei gn exchange rates.
In Q2 2009, Berkshi re agreed wi t h
counterpart ies to amend si x equit y put
contracts, reducing maturities by between
3. 5 and 9. 5 years (br i ngi ng t he t ot al
weighted average maturity of the puts port-
fol io f rom 13 to 12 years). St ri ke prices
on those contracts were reduced between
29 percent and 39 percent. Fi nal ly, the
aggregate notional amount of three of those
cont r act s i ncreased by $160 mi l l i on.
These amendments were cost-f ree (no
money changed hands). In Q4 2010, eight
equit y i ndex put cont racts ($4. 3 bi l l ion
not ional , maturit ies between 2021 and
2028) were terminated (at the instigation
of t he count er par t y); t he unwi ndi ng
requi red Berkshi re to pay $425 mi l l ion,
for a net gai n of $222 mi l l ion as it
had ori gi nal ly received $647 mi l l ion i n
premium.
Al l corporate credit default contracts
(both hi gh-yield and invest ment grade)
expi re i n Q4 2013; a si gnificant port ion
of hi gh-yi el d cont racts expi red i n Q4
2012. Berkshi re i s l i abl e f or payout s
whenever a credi t event t akes pl ace.
Hi gh-yi el d cont ract expi rat i on dat es
ranged f rom September 2009 to Decem-
ber 2013. Individual corporate default con-
tracts had five-year maturit ies and were
referenced to about 40 different names.
Berkshire stopped selling individual cor-
porate credit default contracts from 2009
on, as dealers began asking for stringent
col l ateral .
BERKSHIRE
STOPPED
SELLING
INDIVIDUAL
CORPORATE
CREDIT DEFAULT
CONTRACTS
FROM 2009 ON,
AS DEALERS
BEGAN ASKING
FOR STRINGENT
COLLATERAL.
.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.






Premiums received upfront from the
high-yield credit contracts totaled $3.4
bil lion; payouts on those contracts had
totaled $2.6 bil lion by Q3 2013. Premi-
ums f rom i ndi vi dual cor porate credi t
default contracts are paid quarterly ($93
million a year). Berkshire assumes that the
final underwriting profit (premiums raised
mi nus payouts) f rom corporate credit
contracts wil l be around $1 bil lion when
t hey expi re by year-end 2013, havi ng
enjoyed on average annual $2 billion float.
State/municipal it ies credit cont racts
expi re bet ween 2019 and 2054 (about
500 underlying debt issues). Any poten-
tial loss payments cannot be settled until
expi rat i on. We found no speci f i c dat a
on the amount of the (upf ront) premi -
ums received by Berkshi re f rom sel l i ng
this risk, but we can make an i nformed
approximat ion. At year-end 2008, Berk-
shi re announced that the total premium
raised from selling the high-yield default
cont racts had been $3. 4 bi l l ion. A year
earlier, the announced number had been
$3. 2 bi l l ion. This impl ies new premiums
of $0. 2 bi l l ion in 2008. At the same time,
we know that the i ndividual corporate
cont ract s t hat were f i rst sol d i n 2008
i mpl i ed annual premiums of $93 mi l -
l ion. That makes it $293 mi l l ion i n new
premi ums f or 2008. Si nce Ber ks hi re
reported i n its 2008 annual report that
it had raised $633 million that year in new
premi ums f rom credi t def aul t obl i ga-
t i ons, we may be al l owed to conclude
that the st ate/municipal it ies cont racts
were sold for some $340 million (only one
credi t def aul t cont ract was sol d af ter
2008; I dont know which of the three
underlying risk categories that l ast con-
tract was referenced to). In August 2012,
$8. 25 bi l l ion of the state/municipal it ies
posit i on were termi nated (apparentl y,
t he or i gi nal count er par t y had been
Lehman Brot hers , and Lehmans l i q-
uidators were eager to unwind the trade,
which was heavily in their favor; thus, the
cont ract cancel l at ions may not neces-
s ar i l y i mpl y a negat i ve v i ew of t he
state/municipal it ies market on the part
of Berkshire). I very roughly assume that
the cost of this unwinding may have been
around $475 mi l l i on, or the approxi -
mate si ze of the l i abi l it ies generated by
these cont racts at the t i me.
Loss payout amount s on t he credi t
default obl i gat ions are subj ect to i ndi -
vidual and aggregate limits (for instance,
of around $5 billion in the case of the high-
yi el d def aul t cont racts), and payment
obl i gat ions are on a fi rst-l oss basis or
on an aggregate-deduct ibl e basis.
With limited exceptions, Berkshire has
not been required to post collateral. How-
ever, were it to suffer a rating downgrade,
it would have to post an additional $1.1
billion. See Exhibit 3 for a description of
collateral postings throughout. In 2011, Berk-
shire announced a big change in deriva-
38
CORPORATE FINANCE REVIEW NOVEMBER/DECEMBER 2013 PRACTICAL MATTERS
EXHIBIT 3 Description of Collateral Postings
!
#!!
$!!
%!!
&!!
'!!!
'#!!
() *+ ,-
#!!.
() *+ ,-
#!!&
() *+ /'
#!!0
() *+ ,-
#!!0
() *+ ,-
#!'!
() *+ ,-
#!''
() *+ ,-
#!'#
() *+ /1
#!'1
!"##$%&'$# )"*+,-* ./ 0,1
.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.






tives collateral requirements/policies by
the financial industry, making it now unac-
ceptable for the company to enter into
new major derivatives contracts.
Besides the attainment of a lot of extra
float (some $78 billion for several years,
some $6 billion for a lot of years), another
key reason for enteri ng i nto these par-
t i cul ar der i vat i ves pos i t i ons was t he
bel ief that they were vastly overpriced.
That is, Berkshi re was bei ng given the
chance to collect much more money from
sel l i ng the risk than it shoul d, accord-
ing to Berkshire. If you look at the equit y
puts, the premium, at way over 10 per-
cent of notional amount, certainly looks
tasty (not surprising given the long matu-
rities and the at-the-money strike). This
i s how Buf fet t , at year-end 2006, pre-
emptively tried to address any queries his
sharehol ders may have about t he f act
t hat he had been sel l i ng such a l arge
derivatives portfolio (which, as we know,
was only about to get even l arger):
The answer is that derivat ives, just l i ke stocks
and bonds , are s omet i mes wi l dl y mi s -
pr i ced. Though we wi l l exper i ence l osses
f rom t i me to t i me, we are l i kely to cont i nue to
earn overal l significant profits from mis-
priced derivat ives.
4
The port fol i o sel ected by Berkshi re
to short had one hi ghly intri guing char-
acteristic: It was, for the most part, devoid
of counter par t y ri sk. Si nce premi ums
on the equit y puts, the hi gh-yiel d cor-
por at e credi t def aul t s waps , and t he
state/municipalities credit default oblig-
at ions were received upf ront, Berkshi re
could not be stiffed any money on these
cont racts. Only i n the case of the i ndi -
vidual high-grade corporate credit default
obl i gat i ons was count er par t y r i s k
i nvol ved, as premi ums were recei ved
quar terl y, but t hi s posi t i on was j ust a
smal l f raction of the total trade. If Berk-
shi res t radi ng counterparts went broke
or moved to a far away isl and, the fi rm
woul d suffer al most no pai n. Given the
i ntense focus on counterpart y risk af ter
the fi nanci al crisis, this is no smal l feat.
Berkshires roller coaster
How did Buffetts derivatives play evolve?
Wel l, its been quite a ride, thats for sure.
Lots of ups and downs in gains and losses.
Given that Berkshire must account for the
changes in the market (fair) value of the
derivatives in its income statement, those
ups and downs have i mpacted reported
ear ni ngs on a cont i nuous bas i s. And
given that the value of derivat ives must
be accounted as either assets or l i abi l i -
t i es on t he bal ance sheet , Berkshi res
capital ratios and perception of the firm
as safe and sound (these derivatives hap-
pened to be l i abi l it ies essent i al ly al l the
t i me, si nce they mostly i mpl ied f uture
potent i al obl i gat i ons onl y f rom Berk-
shi re to its counterpart ies and not vice
versa) coul d al so be affected.
Such chute-the-chute is the unavoid-
able price to pay when one chooses to sell
a lot of long-term and varied derivatives
risk. However, i n this case, two factors
were present t hat made i t much more
bear abl e t han i t mi ght have been f or
other fi rms. Fi rst and cruci al ly, and as
we mentioned earlier, Berkshire got away
with ver y l i ght col l ateral terms at i nit i -
at ion of the cont racts. Many a fi rm has
been s unk becaus e t he mar ket went
against them, increasing their l iabi l it ies
and drastically enhancing margin require-
ments unt i l there was no more col l at-
eral available to post up, and liquidation
was t he next , sad step. Thanks to t he
preferential treatment obtained, Berkshire
coul d sel l al l that equit y, currency, and
credi t ri sk safe i n t he knowl edge t hat
any potent i al f uture margi n cal l woul d
be of a mi ni mal si ze. Col l ateral coul d
not si nk Berkshi re, maki ng the t rades
much more attractive and probably even
pl ai n feasi bl e (Berkshi re woul d qui te
probably not have sold the portfolio had
col lateral requirements been stringent).
Apparently, the most Berkshi re has had
to post duri ng the l ife of the t rade was
$1. 7 bi l l i on at some poi nt dur i ng t he
worst of the 2008 fi nanci al crisis. This
is money that st i l l cont i nues to produce
a return for Berkshi re whi l e it is bei ng
hel d as a guarantee.
Second, Warren Buffett doesnt seem
to care at al l about i nteri m earni ngs or
balance sheet volatility (the lack of strin-
gent col l ateral requi rement s possi bl y
pl ays a rol e here), repeatedly sayi ng so
to his sharehol ders. He fi rmly bel ieves
39
PRACTICAL MATTERS NOVEMBER/DECEMBER 2013 CORPORATE FINANCE REVIEW
THE PORTFOLIO
SELECTED BY
BERKSHIRE TO
SHORT HAD
ONE HIGHLY
INTRIGUING
CHARACTERISTIC:
IT WAS, FOR
THE MOST
PART, DEVOID OF
COUNTERPARTY
RISK.
.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.






that the t rades wi l l i n the end generate
positive float, and thats what truly counts.
Hey, thats why they were put on i n the
fi rst pl ace: f ree real money for the fi rm;
who cares about some collateral-light, unre-
al i zed turbul ence on the side? Not War-
ren Buffett certai nly:
Our derivat ive posit ion wi l l somet i mes cause
large swings in reported earnings, even though
[we] mi ght bel ieve the i nt ri nsic value of these
posit ions has changed l ittl e. [We] wi l l not be
bothered by these swi ngs even though they
coul d easi ly amount to $1 bi l l ion or more i n
a quarter and we hope [shareholders] wont
either. In our cat ast rophe i nsurance busi -
ness, we are always ready to t rade i ncreased
volatilit y in reported earnings in the short run
for greater gai ns i n net worth i n the l ong run.
This is our phi l osophy i n derivat ives as wel l .
5
The market value of Berskhires deriv-
at ives portfol io woul d be i mpacted by
several key vari abl es. In the case of the
puts, Berkshi re woul d suffer setbacks if
equit y prices fel l, if equit y volatilit y shot
up, if the dol l ar dropped i n value versus
the yen, euro, or pound, and if i nterest
rates went down. It woul d make gai ns if
the opposite moves took pl ace, and al so
just from the passage of time. In the case
of the credit default contracts, Berskhire
woul d suffer l osses if American corpo-
r at e and s t at e/ muni ci pal i t i es credi t
spreads shot up and (i n the case of a few
contracts) if the counterpart y defaulted
or l ooked cl ose to default i ng.
As these vari abl es fluctuated si gnifi -
cantly during the l ife of these cont racts,
Berkshi re experienced si gnificant tur-
bul ence i n mark-to-market derivat ives
gai ns and l osses, as wel l as on the port-
folios liabilities. And given that the port-
fol io was quite sizeable and that some of
those gai ns and l osses coul d be l arge,
t he t ur bul ence s omet i mes had a bi g
i mpact on Berkshi res overal l reported
earnings. Exhibit 4 detai ls the evolut ion
of t he gai ns / l os s es , l i abi l i t i es , and
notional amounts. The information comes
f rom Berkshi res quarterl y and annual
repor t s, and whi l e t horough care has
been taken to col lect the data accurately,
s ome er ror s or omi s s i ons may be
inevitable. Detai led info on the puts and
credit cont racts is only avai l abl e on an
annual basis f rom 2006 and on a quar-
terly basis f rom 2008.
Lets start with gai ns and l osses (part
of t he companys i ncome st at ement ).
Whi l e 2006 and 2007 were rel at i vel y
pl acid (with the except ion of Q4 2007),
2008 has by f ar been t he wor s t - per -
forming year, with a combined account-
i ng setback of $6. 8 bi l l ion. Si nce by that
t i me, unl i ke i n previous years, the puts
and the credit contracts comprised essen-
tial ly Berkshires entire derivatives port-
folio, from that point on, gains and losses
on the former al most exactl y matched
overal l derivatives gains and losses. 2009
was the best year ($3. 6 bi l l ion gai n), i n
spite of a horribl e Q1. 2010 saw a mod-
est gai n of $420 mi l l ion, notwithstand-
i ng a $2. 1 mi l l i on l os s i n Q2. 2011
was horrible a $2 billion setback. Mar-
ket s rebounded i n 2012, l eadi ng t o a
$1. 9 bi l l ion gai n. So far, 2013 has been
great with a cumul at ive $2 bi l l ion gai n.
Si nce 2008, the portfol io yiel ded gai ns
i n excess of $1 bi l l ion on seven quar-
ters, and l osses i n excess of $1 bi l l ion
al so on seven occasi ons. Gai ns above
$2 bi l l ion took pl ace three t i mes, l osses
above $2 bi l l ion al so three t i mes.
These gai ns/l osses had, on occasion,
a bi g impact on Berkshires overal l prof-
its. For i nstance, Berkshi re barely made
any money i n Q4 2008 (just a t i ny $140
mi l l ion i n pretax earni ngs, a 90 percent
decline with regards to the previous quar-
ter), and the massive $4. 5 bi l l ion deriv-
atives loss surely had something to do with
it. The fi rms ent i re pretax earni ngs for
2008, at just $7. 5 bi l l ion, were only 37
percent of 2007s fi gure; the $6. 8 bi l l ion
derivat ives debacl e cont ributed mi ght-
i l y to that sharp decl i ne (i . e. , without
the derivat ives, no such sharp decl i ne
i n profitabi l it y). To be fai r, derivat ives
gains have also contributed to significant
increases in profits and even to the mere
presence of such i ncrease, as i n Q4 2010
when overal l earni ngs grew by l ess than
$2 bi l l ion, coi ncidental with a deriva-
t i ves gai n of $2. 3 bi l l i on, or as i n Q4
2012 ($800 mi l l i on and $2. 1 bi l l i on,
respect ively).
The derivat ives portfol io coul d, with
mi nor except ions, represent only a l i a-
bi l it y for Berkshi re, given that any pay-
ments can only ori gi nate f rom the fi rm
(and not f rom its counterpart ies). Only
40
CORPORATE FINANCE REVIEW NOVEMBER/DECEMBER 2013 PRACTICAL MATTERS
BERKSHIRE
EXPERIENCED
SIGNIFICANT
TURBULENCE
IN MARK-TO-
MARKET
DERIVATIVES
GAINS AND
LOSSES, AS
WELL AS ON THE
PORTFOLIOS
LIABILITIES.
.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.






41
PRACTICAL MATTERS NOVEMBER/DECEMBER 2013 CORPORATE FINANCE REVIEW
EXHIBIT 4 Gains/Losses
!"###
!%###
!&###
!'###
!(###
!)###
#
)###
(###
'###
&###
*+
(##"
*+
(##,
-)
(##.
-(
(##.
-'
(##.
-&
(##.
*+
(##.
-)
(##/
-(
(##/
-'
(##/
-&
(##/
*+
(##/
-)
(#)#
-(
(#)#
-'
(#)#
-&
(#)#
*+
(#)#
-)
(#))
-(
(#))
-'
(#))
-&
(#))
*+
(#))
-)
(#)(
-(
(#)(
-'
(#)(
-&
(#)(
*+
(#)(
-)
(#)'
-(
(#)'
-'
(#)'
!"#$%&'(%%)% +, -#..#($/
+01234 5136 789:23 ;9<=1>3 ?@>2A=BCD6
!"###
!%###
!&###
!'###
#
'###
&###
%###
(
)
'
#
#
&

*
+
'
#
#
,

*
'
'
#
#
,

*
-
'
#
#
,

*
&
'
#
#
,

(
)
'
#
#
,

*
+
'
#
#
%

*
'
'
#
#
%

*
-
'
#
#
%

*
&
'
#
#
%

(
)
'
#
#
%

*
+
'
#
#
.

*
'
'
#
#
.

*
-
'
#
#
.

*
&
'
#
#
.

(
)
'
#
#
.

*
+
'
#
#
"

*
'
'
#
#
"

*
-
'
#
#
"

*
&
'
#
#
"

(
)
'
#
#
"

*
+
'
#
#
/

*
'
'
#
#
/

*
-
'
#
#
/

*
&
'
#
#
/

(
)
'
#
#
/

*
+
'
#
+
#

*
'
'
#
+
#

*
-
'
#
+
#

*
&
'
#
+
#

(
)
'
#
+
#

*
+
'
#
+
+

*
'
'
#
+
+

*
-
'
#
+
+

*
&
'
#
+
+

(
)
'
#
+
+

*
+
'
#
+
'

*
'
'
#
+
'

*
-
'
#
+
'

*
&
'
#
+
'

(
)
'
#
+
'

*
+
'
#
+
-

*
'
'
#
+
-

*
-
'
#
+
-

!"#$%&'(%%)% +, -#..#($/
!"###
!%###
!&###
!'###
'###
&###
!"#$%&'(%%)% +, -#..#($/
()*+ , -./01* 2/34)5*
67/.455 2/.17487/+
!"####
!%###
#
%###
"####
"%###
&####
&%###
'(
&##)
'(
&##*
+"
&##,
+&
&##,
+-
&##,
+.
&##,
'(
&##,
+"
&##/
+&
&##/
+-
&##/
+.
&##/
'(
&##/
+"
&#"#
+&
&#"#
+-
&#"#
+.
&#"#
'(
&#"#
+"
&#""
+&
&#""
+-
&#""
+.
&#""
'(
&#""
+"
&#"&
+&
&#"&
+-
&#"&
+.
&#"&
'(
&#"&
+"
&#"-
+&
&#"-
+-
&#"-
!"#$%&'(%%)% +, -#..#($/ 0123 4 567892 :7;<1=2
(<6>9>?3 @A67!2<BC
.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.






i n the case of the i ndividual corporate
credi t def aul t cont racts di d Berkshi re
face counterpart y risk, as the premiums
were paid quarterly rather than ent i rely
upfront. This explains that in some peri-
ods ( i n many, ac t ual l y) Ber ks hi re
recorded t hose hi gh-grade cor porat e
posi t i ons as assets and not l i abi l i t i es.
The expected value of the premiums to
be recei ved by Berkshi re was s i mpl y
hi gher than the expected value of any
default payments to be made by Berkshire.
Ber ks hi res der i vat i ves l i abi l i t i es
(recorded on the ri ght hand side of the
balance sheet) change for three reasons:
changes i n the fai r value of the deriva-
t ive (i . e. , gai ns or l osses), new premi -
ums col l ected, and new payouts made.
When no new premiums or new payouts
have taken pl ace, the change i n deriva-
t ives l iabi l it ies wi l l be equal to the gains
(l eadi ng to a decrease i n l i abi l it ies) or
l osses (l eadi ng to an i ncrease i n l i abi l i -
t i es ) i ncur red by t he pos i t i on. For
i nst ance, year-end 2008 credit default
contract liabilities increased by $2.3 bil-
l ion with respect to year-end 2007. This
was explained by pretax fair value losses
of $1. 8 bi l l ion, $633 mi l l ion i n new pre-
mi ums, and $152 mi l l i on of l oss pay-
ment s . Progres s i vel y, es s ent i al l y al l
changes i n deri vat i ves l i abi l i t i es were
expl ai ned by mark-to-market gai ns or
l osses on the puts and credit cont racts
portfolio, as no more premium money was
bei ng raised and as new l arge l oss pay-
ments vanished.
As can be seen i n Exhi bi t 5, equi t y
puts l i abi l it ies only reached $10 bi l l ion
a coupl e of t i mes, havi ng been between
$6 bi l l ion and $8 bi l l ion for most of the
t i me. Credi t def aul t l i abi l i t i es onl y
reached $4 billion a couple of times, hav-
i ng been l ess than $2 bi l l ion most of the
t i me. As the corporate credit cont racts
began to expi re, as i nvest ment -grade
exposures turned i nto net assets (f rom
l at e 2009 on) , and as hal f of t he
st ate/municipal it i es exposure was l iq-
uidated, credit default liabilities naturally
nosedived. As of Q3 2013, Berkshi res
credit l i abi l it ies stand at just $470 mi l -
lion. Thats how much it would cost War-
ren Buffett to buy back the contracts and
l iquidate the exposure once and for al l .
Was it worth it?
We are delighted that we hold the deriv-
at i ves cont ract s t hat we do, decl ared
Warren Buffett in his 2009 letter to share-
hol ders.
6
Comi ng as they did soon af ter
the fi nanci al crisis, which had l ed to bi g
l osses amid the worst performance ever
experienced by the portfolio, these words
are doubly reassuri ng as to Berkshi res
enthusi ast ic and st aunch commit ment
to the trade. The firm was looking for one
thing: substantial and long-lasting float.
As l ong as col l ected premi ums (bot h
upfront and quarterly) kept above any pay-
ments derived from the derivatives posi-
tion, Berkshire would be happy. That the
outcome was going to be favorable seems
to have never been i n serious doubt. In
the 2007 l etter, Buffett stated: I bel ieve
that on premium revenues al one, these
cont racts wi l l prove profitabl e, l eavi ng
aside what we can earn on the large sums
we hol d.
7
A year l ater, he rei terated:
Our expect at i oni s t hat we wi l l do
better than break even and that the sub-
stantial investment income we earnwill
be frosting on the cake.
8
As for longevity,
well, the equity puts and the state/munic-
i pal i t i es credi t def aul t s waps mat ure
between 10 and 45 years af ter premium
has been col lected, with no loss payment
by Berkshi re taki ng pl ace, if at al l , unt i l
those far away expi rat ion dates.
By year- end 2011, Ber ks hi re was
al ready danci ng the victor y l ap, at l east
when it came to one of the (sizable) com-
ponents of the t rade:
Our i ns ur ance- l i ke der i vat i ves cont r act s ,
whereby we pay if various issues i ncluded i n
hi gh-yiel d bond i ndexes default, are comi ng
to a cl ose. We are al most certai n to real i ze
a fi nal under writ i ng profit on this portfol io
because the premiums we received were $3. 4
bi l l i on, and our f ut ure l osses are apt to be
mi nor. This successf ul result duri ng a t i me
of great credit st ress underscores the i mpor-
t ance of obt ai ni ng a premi um t hat i s com-
mensurate with the risk.
9
There was al so l itt l e doubt t hat t he
equit y puts pl ay wi l l prove handsomely
profitable, as expressed in the 2012 mis-
sive to sharehol ders:
Though its no sure thing, [we] believe it likely
that the final l iabi l it y wi l l be considerably less
than the amount we currentl y carr y on our
books ($7. 5 bi l l ion). In the meant i me, we can
42
CORPORATE FINANCE REVIEW NOVEMBER/DECEMBER 2013 PRACTICAL MATTERS
BERKSHIRES
DERIVATIVES
LIABILITIES
(RECORDED ON
THE RIGHT HAND
SIDE OF THE
BALANCE SHEET)
CHANGE FOR
THREE REASONS:
CHANGES IN THE
FAIR VALUE OF
THE DERIVATIVE
(I.E., GAINS OR
LOSSES), NEW
PREMIUMS
COLLECTED,
AND NEW
PAYOUTS MADE.
.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.






invest the $4.2 billion of float derived from these
cont racts as we see fit.
10
The hi gh-yi el d cor porate bet was a
r i sky one, made even r i ski er by t he
unprecedented credit crisis of 20072008,
and yet it has delivered close to $1 bil-
lion in cash profits from premium alone.
How much did Berkshire make on top of
that, through reinvestment of the float?
It may be hard to say without direct knowl-
edge, but it could be reasonable to assume
that the return has been positive (data
seem to indicate that annual returns on
assets between 2004 and 2012 stayed in
the 1.805.10 percent range, being around
34 percent on average; historical returns
have been much higher as il lustrated in
Exhibit 6). Berkshire may be expected to
make several bil lions of dol lars, perhaps
tens of bi l lions, during the many years
that the positive float would last.
The i ndividual hi gh-grade corporate
credit default cont racts, about to expi re
for good, dont appear to have generated
43
PRACTICAL MATTERS NOVEMBER/DECEMBER 2013 CORPORATE FINANCE REVIEW
EXHIBIT 5 Liabilities
!
#!!!
$!!!
%!!!
&!!!
'!!!!
'#!!!
'$!!!
'%!!!
'&!!!
()
#!!*
()
#!!%
()
#!!+
,'
#!!&
,#
#!!&
,-
#!!&
()
#!!&
,'
#!!.
,#
#!!.
,-
#!!.
()
#!!.
,'
#!'!
,#
#!'!
,-
#!'!
()
#!'!
,'
#!''
,#
#!''
,-
#!''
()
#!''
,'
#!'#
,#
#!'#
,-
#!'#
()
#!'#
,'
#!'-
,#
#!'-
,-
#!'-
!"#$"%"&'( *+ ,"%%"-./ )/0123 4025 6178191:;5
<=;>12 ?;@7092 6178191:;5
<0A097:B;
?;=1B7:B;5 6178191:;5
!"##
#
"##
%###
%"##
&###
&"##
'###
'"##
(###
("##
)*
&##"
)*
&##+
)*
&##,
)*
&##-
.%
&##/
.&
&##/
.'
&##/
)*
&##/
.%
&#%#
.&
&#%#
.'
&#%#
)*
&#%#
.%
&#%%
.&
&#%%
.'
&#%%
)*
&#%%
.%
&#%&
.&
&#%&
.'
&#%&
)*
&#%&
.%
&#%'
.&
&#%'
.'
&#%'
!"# %&'(&)&*+, -. /&))&012 0123 )1456 786494:
7861;16<=5 >?@A?@=B4
0123 )1456 C >?@A?@=B4
DB=B4:EF<81G1A=51H4:











<5>>>
<6>>>
<8>>>
>
8>>>
6>>>
5>>>
AB 8>>5 AB 8>>: AB 8>>9 AB 8>>? AB 8>;> AB 8>;; AB 8>;8 @; 8>;= @8 8>;= @= 8>;=
+,-./012//3/ 4/ 1-,5-6-73/ 89,.:3/ ;< =.>
!"#$ &'()$*+,$$-$
./0 &'()$*+,$$-$
!"#$ +('1(2(C-$ .3')4-
./0 +('1(2(C-$ .3')4-
.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.






any loss payout, and there seem to be no
i ndicat ions of counterpart y default on
the quarterly premium payments. Those
$93 million annual fees (plus any upfront
fee) appear to have been ent i rel y f ree
money, for five l ong years.
The state/municipal it ies default con-
t racts cant generate a l oss payout unt i l
t hei r maturit y (far, far i n t he f uture),
and whi l e Berkshi re had to make good
on $8.5 billion of the position, the amount
paid to the counterpart y is not known
(a ver y rough and almost certainly inex-
act approxi mat ion may be i n the nei gh-
borhood of $400 mi l l ion). In any case,
it cant have been a prohibit ive amount
given that at the t i me of the unwi ndi ng,
the total mark-to-market liabilit y on the
ent i re $16 bi l l ion portfol io was around
$950 mi l l ion. Assumi ng that Berkshi re
raised around $340 mi l l ion through the
sal e of these cont racts (as per our cal -
cul at ion i n the prior sect ion), this par-
ticular trade may have been unprofitable.
Exhi bi t 7 provi des a descri pt i on of
the evolut ion of Berkshi res derivat ives
fl oat (agai n, drawi ng on the companys
quarterly and annual reports, and on our
analysis for the cost of the state/munic-
ipal it i es port fol io unwi ndi ng i n 2012,
this being the only relevant number seem-
ingly not having been publicly disclosed).
We see that the benefits have been prett y
substant ial so far. And the news get even
better when we take i nto account that,
barring any desperate request by a coun-
ter par t y to unwi nd a t rade (toget her
with Berkshi res acquiescence to do so
and i ncur the cost of buyi ng back the
exposure), no f urther l oss payouts can
happen before expiration of the only two
remaining positions, the equit y puts and
t he st ate/muni cipal i t i es credi t def aul t
obligations that mature in the pleasantly
distant 20192054 period.
Anot her way to anal yze t he perfor-
mance of the t rade is by compari ng the
fl oat obtai ned with the account i ng l i a-
bi l it ies generated. In other words, com-
pare premiums mi nus payouts with the
market cost of liquidating the exposures.
Had Berkshi re had or wanted to termi -
nate the puts and the credit default con-
tracts, would the raised premiums (minus
44
CORPORATE FINANCE REVIEW NOVEMBER/DECEMBER 2013 PRACTICAL MATTERS
EXHIBIT 6 Buffetts Performance
Public
U.S. stocks Overall
Berkshire (from Private stock market
Hathaway 13F filings) Holdings performance
Sample 19762011 19802011 19842011 19762011
Beta 0.68 0.77 0.28 1.00
Average excess return 19.00% 11.80% 9.60% 6.10%
Total volatility 24.80% 17.20% 22.30% 15.80%
Idiosyncratic volatility 22.40% 12.00% 21.80% 0.00%
Sharpe ration 0.76 0.69 0.43 0.39
Information ratio 0.66 0.56 0.36 0.00
Leverage 1.64 1.00 1.00 1.00
Sub period excess returns:
19761980 42.10% 31.40% 7.80%
19811985 28.60% 20.90% 18.50% 4.30%
19861990 17.30% 12.50% 9.70% 5.40%
19911995 29.70% 18.80% 22.90% 12.00%
19962000 14.90% 12.00% 8.80% 11.80%
20012005 3.20% 2.20% 1.70% 1.60%
20062011 3.30% 3.00% 2.30% 0.70%
Data from: Frazzini, A., Kabiller, D., and Pedersen, L.H., "Buffett's Alpha," (May 3, 2012).
.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.






any l oss payment s) have been abl e to
cope, or woul d the fi nal tal ly have been
way i n excess of that? Wel l , the evidence
i s somewhat mi xed, unl ess we assume
that the fl oat l ed to prett y i nterest i ng
investment returns. Exhibit 8 shows why.
While the float from the credit contracts
would have been in general enough to cover
their liabilities, that from the puts would
by itsel f have been widely i ncapabl e of
doi ng so.
Finally, we can look at the size of those
l i abi l i t i es wi t h regard t o Ber ks hi res
equit y capital . In other words, if those
mark-to-market l osses grew too l arge,
would the companys solvency be at stake?
Exhi bi t 9 provi des s ome gui dance. It
doesnt l ook as if the portfol io gravely
threatened Berkshi re.
Lack of strict col lateral requirements
was key for the trade to work and perform:
Berkshi re woul d ot her wi se not have
entered into the trades. Buffett has referred
to derivatives collateral as a lethal threat
that can sink companies. While it is true
that by avoiding stringent margin rules
Berkshire agreed to col lect less premi-
ums on the sol d cont racts than woul d
ot her wi se have been t he case, Buf fet t
declared at the end of 2010: That...lef t
us feeling comfortable during the finan-
cial crisis, allowing us in those days to com-
mi t to some advant ageous purchases.
Foregoi ng some addit ional derivat ives
premiums proved to be wel l worth it.
11
A year later, Buffett made clear that as a
consequence of the new post-crisis, much-
more-demandi ng i ndust r y pol i ci es on
col lateral, his firm would not be entering
into new positions:
Though our existing contracts have very minor
collateral requirements, the rules have changed
for new posit ions. We shun cont racts of any
t ype that coul d requi re the i nstant post i ng of
45
PRACTICAL MATTERS NOVEMBER/DECEMBER 2013 CORPORATE FINANCE REVIEW
EXHIBIT 8 Liabilities Versus Float ($ Million)
!
#!!!
$!!!
%!!!
&!!!
'!!!!
'#!!!
() #!!* () #!!& () #!!+ () #!'! () #!'' () #!'#
),-./0 1-/2 3.45.6.782
9:8;./ <8=4-6/ 3.45.6.782
),-./0 1-/2 >6?4/
9<@ >6?4/
EXHIBIT 7 The Evolution of Berkshires Derivatives Float
Equity Puts
($ mn) Premiums CDS Premiums Cumulative Put Payouts CDS Payouts Float
Up to YE 2007 4,500 3,200 7,700 0 472 7,228
Up to YE 2008 4,900 3,833 8,733 0 542 8,191
Up to YE 2009 4,900 3,926 8,826 0 2,442 6,384
Up to YE 2010 4,900 4,019 8,919 425 2,442 6,052
Up to YE 2011 4,900 4,112 9,012 425 2,528 6,059
Up to YE 2012 4,900 4,205 9,105 425 3,005 5,676
.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.






col lateral . The possibi l it y of some sudden and
huge posting requirement arising from some
out of -t he-bl ue event such as a worl dwi de
financial panic or a massive terrorist attack
is i nconsistent with our pri mar y obj ect ives of
redundant l iquidit y and unquest ioned fi nan-
ci al st rength.
12
Why did Warren Buffett select precisely
those underlying assets for his vast deriv-
atives bet? Was he wise and prescient, or
just fortunate? Why equit y indexes, and
why those indexes? Why credit default, and
why those credits? Why 2004? Why 2008?
Maybe Buffett was doing the mirror equiv-
alent of buying the dips (when people
buy assets ver y cheaply right after a mar-
ket col lapse), in essence sel ling the top
by sel ling options when their premiums
have exploded as a result of a market col-
l apse. We can see t hat t he equi t y puts
notional size increases substantial ly (it
almost doubles) in 2007 and early 2008,
which of course are dates when markets
began to unravel wi l dl y, t hus maki ng
downside protection very expensive, per-
haps in an irrational way. We can also see
that the credit default obligations notional
amounts explode in 2007 (double) and
2008 (seven times higher), periods when
credit spreads worsened severely. If you
bel ieve that people are overreact ing to
the panic, it can be a great opportunit y
to take on risk. Global equities and Amer-
ican credit happened to be two asset classes
that concurrently and suddenly went out
of control. And Buffett dutiful ly stepped
in to monetize that chaos.
In the end, Buffetts pl ay woul d not
have been possibl e had counterpart ies
not been available and wil ling to pay the
premiums. Why did they do it? Were they
insuring themselves against an existing
underlying risk profi le? Or where they
punt i ng on t anki ng market s, wi dened
credit spreads, outright defaults, overal l
tumult, and a falling dollar? Whatever the
case, Warren Buffett and others like him
(who seem to be the majorit y in the mar-
kets; people seem to prefer to be premium
receivers rather than premium payers)
should be thankful for the existence of
those eager to act as sources of the float
that has made the Berkshire Hathaway
miracle possible.
Holy triad
Our anal ysi s of Berkshi re Hat haways
derivat ives posit ion appears to confi rm
(for now at l east) the fi ndi ngs of Frazz-
ini, Kabi l ler, and Pedersen, at least when
it comes to the extremely favorable fund-
i ng t er ms t hat t he f i r m can achi eve
through its float business. If, as the hedge
f und t r i o posi t s, War ren Buf fet t t hen
i nvests the fl oat ver y wisely, it is easy to
appreciate that he has developed an edge.
But smart as generat ing so much pos-
itive float for so long surely is, and smart
as selecting profitable investments surely
is, the true secret sauce may lie in a third,
t ypical ly less discussed, factor at least
when it comes to the massive derivatives
46
CORPORATE FINANCE REVIEW NOVEMBER/DECEMBER 2013 PRACTICAL MATTERS
EXHIBIT 9 Derivatives Liabilities/Equity
!"!!#
%"!!#
&"!!#
'"!!#
("!!#
)!"!!#
)%"!!#
)&"!!#
)'"!!#
*+ %!!& *+ %!!, *+ %!!' *+ %!!- *+ %!!( *+ %!!. *+ %!)! *+ %!)) *+ %!)% /0 %!)0
.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.






portfol io analyzed i n this paper. Si mply
stated, Berkshire appears to have enjoyed
tremendous, and perhaps unique, advan-
tages when it came to sel l i ng the deriv-
atives f rom which the float (and thus the
edges foundation) comes. Without those
advantages in place, the whole thing may
not have been possibl e to begi n wi t h.
And t he t rue key i s t hat t hose advan-
tages may be reser ved for Buffett and,
maybe, just a handf ul of other peopl e.
Enj oy i ng t hos e advant ages , i n ot her
words, can lead to vast compet it ive ben-
efits.
Those three key factors that may not
have been avai l abl e to al l market pl ay-
ers are: (1) ver y sof t col l ateral requi re-
ments, (2) utter disregard for quarterly
earni ngs vol at i l it y, and (3) the abi l it y
to fi nd buyers of si zabl e and of ten het-
erodox contracts. Other players may have
f aced much more st r i ngent col l ateral
requi rement s. Ot her pl ayers may care
much more about cont i nuous earni ngs
t urbul ence. Ot her pl ayers may not be
able to sel l such contracts. Buffett is ver y
clear about it: If he had to face normal
collateral rules, he would not have entered
i nto the t rades. Did he get preferent i al
t reat ment because of who he is? Li kely.
Buffett was wi l l i ng to sel l cont racts for
l ess premi um j ust t o avoi d col l at er al
posting; since raising premium is al l that
mat ters, hi s concerns about col l ateral
are obvi ous. Wi t h st andard col l ateral
rul es, $4050 bi l l ion woul d have been
at i nstant risk (and coul d suddenly si nk
many a firm). Many other investors would
not have been able to sel l a similar deriv-
at ives portfol io for fear of those expo-
sures or for utter l ack of resources. That
is, a fl oat-generat i ng t rade that is pos-
sibl e and desi rabl e for Buffett to make
becomes impossible and undesirable for
many or most others. Only Buffett would
get to enjoy the float and thus the tremen-
dous i nvest ment edge.
The contracts Berkshire sold were not
al l orthodox. Buffett pl aced as much as
$40 bi l l i on of ver y l ong dated at-t he-
money equit y risk. Buffett col l ected $4
bi l l i on of hi gh- r i s k credi t premi um
upf ront, when quarterly payments tend
to be the norm. Some credit contracts had
1045 year maturi t i es, wi t h f ive years
being the norm. Loss payouts on several
cont racts can only take pl ace at matu-
rit y, with whenever a credit event takes
place being the norm. These unorthodox
cont racts were apparently wi l dly over-
pr i ced, af fordi ng Buf fet t l ot s of f l oat
f rom the get-go. Coul d anybody enter
i nto such t rades (i . e. , be abl e to f i nd
wi l l i ng buyers), or do you need to be
Warren Buffett?
As for quarterl y earni ngs vol at i l it y,
not ever yone may be abl e or wi l l i ng to
be so sanguine/complacent. Buffett enjoys
god-l i ke st ature with hi s sharehol ders
and has built a career on long-term focus.
Temporar y set backs, i ncludi ng ver y
l arge ones, may thus not turn them i nto
ferocious crit ics. Not ever yone may be
shielded from criticism in such a way. Of
course, this is di rectly l i nked to the col -
l ateral issue: If you dont face hard col -
l ateral penalt ies, you can afford to not
care about earni ngs turbul ence.
Legend or bust
Natural ly, Buffett coul d have borrowed
l i ke anybody el se, and then i nvested the
money. But he wants to build an edge. Its
hard to become a l egendar y i nvestor if
you do what ever yone can do. The fl oat
gives him that edge. Of course, the activ-
it ies that l ead to the generat ion of fl oat
contain the seeds of risks that may mate-
rialize into costs way above those of a sim-
ple loan. But Berkshire, and our analysis
seems to confi rm this, has been master-
ful at achieving underwriting profits and
thus negat ive cost f undi ng. Fl oat does-
nt have to be paid back, doesnt i mply
payment of interest, and is not debt. Had
Berkshi re, i n 2004, borrowed $6 bi l l ion
(approxi mately the average annual pos-
i t i ve f l oat f rom t he deri vat i ves t rade)
for 15 years at , say, 5 percent annual
i nterest, it woul d have been $10. 5 bi l -
l ion out of pocket by the loans maturit y
date. So far, its only l ost $3 bi l l ion on
t he deri vat i ves posi t i on, and, barri ng
some surprise, no ext ra cash disburse-
ments will take place until 2019if at all.
Thats a $7. 5 bi l l ion surplus. So the bi g
lesson f rom Berkshire Hathaway may be
that you must take chances if you want
to a superior i nvestor. n
47
PRACTICAL MATTERS NOVEMBER/DECEMBER 2013 CORPORATE FINANCE REVIEW
.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.






NOTES
1
Frazzi ni , A. , Kabi l l er, D. , and Pedersen, L. H. , Buf-
fetts Al pha, ( May 3, 2012) ( whi te paper) .
2
Berkshi re Hathaway I nc.s annual report ( 2012) : 7.
Avai l abl e at: http: / / www. berkshi rehathaway. com/
2012ar/ 2012ar. pdf.
3
I bi d. , p. 8.
4
Berkshi re Hathaway Inc.s sharehol der l etter (2006):
17. Avai l abl e at: http://www.berkshi rehathaway.com/
l etters/ 2006l tr. pdf.
5
Berkshi re Hathaway Inc. press rel ease (May 2, 2008).
Avai l abl e at: http: / / www. berkshi rehathaway. com/
news/ may0208. pdf.
6
Berkshi re Hathaway Inc.s sharehol der l etter (2009):
15. Avai l abl e at: http://www.berkshi rehathaway.com/
l etters/ 2009l tr. pdf.
7
Berkshi re Hathaway Inc.s sharehol der l etter (2007):
16. Avai l abl e at: http://www.berkshi rehathaway.com/
l etters/ 2007l tr. pdf.
8
Berkshi re Hathaway Inc.s sharehol der l etter (2008):
18. Avai l abl e at: http://www.berkshi rehathaway.com/
l etters/ 2008l tr. pdf.
9
Berkshi re Hathaway I nc.s sharehol der l etter ( 2011) :
17. Avai l abl e at: http://www.berkshi rehathaway.com/
l etters/ 2011l tr. pdf.
10
Berkshi re Hathaway I nc.s sharehol der l etter ( 2012) :
16. Avai l abl e at: http://www.berkshi rehathaway.com/
l etters/ 2012l tr. pdf.
11
Berkshi re Hathaway I nc.s annual report ( 2010) : 20.
Avai l abl e at: http: / / www. berkshi rehathaway. com/
2010ar/ 2010ar. pdf.
12
Op. ci t. note 9.
48
CORPORATE FINANCE REVIEW NOVEMBER/DECEMBER 2013 PRACTICAL MATTERS

You might also like