Ch 20 Hull Fundamentals 8 the d
Transcript of Ch 20 Hull Fundamentals 8 the d
-
7/25/2019 Ch 20 Hull Fundamentals 8 the d
1/48
Fundamentals of Futures and Options Markets, 8th Ed, Ch 20, Copyright John C. ull 20!"
Value at Risk
Chapter 20
1
-
7/25/2019 Ch 20 Hull Fundamentals 8 the d
2/48
Fundamentals of Futures and Options Markets, 8th Ed, Ch 20, Copyright John C. ull20!"
The Question Being Asked in VaR
What loss level is such that we areX%
confident it will not be exceeded inNbusiness days?
2
-
7/25/2019 Ch 20 Hull Fundamentals 8 the d
3/48
Fundamentals of Futures and Options Markets, 8th Ed, Ch 20, Copyright John C. ull20!"
VaR and Regulatory Capital
Regulators base the capital they reuireban!s to !eep on "aR
#or $ar!et ris! they use a 0&day ti$ehori'on and a ((% confidence level
#or credit ris! they use a (()(%
confidence level and a year ti$ehori'on
3
-
7/25/2019 Ch 20 Hull Fundamentals 8 the d
4/48
Fundamentals of Futures and Options Markets, 8th Ed, Ch 20, Copyright John C. ull20!"
VaR vs. Expected ShortfallSee !igures "#.$ and "#."% page &'()
"aR is the loss level that will not beexceeded with a specified probability
*xpected shortfall is the expected lossgiven that the loss is greater than the "aRlevel
+lthough expected shortfall is theoretically$ore appealing than "aR, it is not aswidely used
4
-
7/25/2019 Ch 20 Hull Fundamentals 8 the d
5/48
Fundamentals of Futures and Options Markets, 8th Ed, Ch 20, Copyright John C. ull20!"
Advantages of VaR
-t captures an i$portant aspect of ris!
in a single nu$ber
-t is easy to understand -t as!s the si$ple uestion. /ow bad can
things get?
5
-
7/25/2019 Ch 20 Hull Fundamentals 8 the d
6/48
Fundamentals of Futures and Options Markets, 8th Ed, Ch 20, Copyright John C. ull20!"
*istorical Si+ulation
Create a database of the daily $ove$ents in all$ar!et variables)
he first si$ulation trial assu$es that the
percentage changes in all $ar!et variables areas on the first day
he second si$ulation trial assu$es that the
percentage changes in all $ar!et variables areas on the second day and so on
6
-
7/25/2019 Ch 20 Hull Fundamentals 8 the d
7/48
Fundamentals of Futures and Options Markets, 8th Ed, Ch 20, Copyright John C. ull20!"
*istorical Si+ulation continued
1uppose we use 0 days of historical data34ay 0 to 4ay 005
6et vibe the value of a $ar!et variable on day i
here are 00 si$ulation trials he ith trial assu$es that the value of the
$ar!et variable to$orrow is
7
1
500
i
i
v
vv
-
7/25/2019 Ch 20 Hull Fundamentals 8 the d
8/48
*istorical Si+ulation continued
he portfolio7s value to$orrow iscalculated for each si$ulation trial
he loss between today and to$orrow isthen calculated for each trial 3gains arenegative losses5
he losses are ran!ed and the one&day((% "aR is set eual to the thworst loss
Fundamentals of Futures and Options Markets, 8th Ed, Ch 20, Copyright John C. ull20!"
8
-
7/25/2019 Ch 20 Hull Fundamentals 8 the d
9/48
Exa+ple , Calculation of $-day% /
VaR for a 0ortfolio on Sept "(% "##1 Ta2le"$.$% page &'3)
Fundamentals of Futures and Options Markets, 8th Ed, Ch 20, Copyright John C. ull20!"
9
Index Value ($000s)
48-+ 9,000
#1* 00 :,000
C+C 90 ,000
;i!!ei 22 2,000
-
7/25/2019 Ch 20 Hull Fundamentals 8 the d
10/48
4ata After Ad5usting for Exchange
Rates Ta2le "#."% page &'1)
Fundamentals of Futures and Options Markets, 8th Ed, Ch 20, Copyright John C. ull20!"
10
Day Date DJIA FTSE 100 CAC 40 Niei !!"
0 +ug ,:)>9 =,:( :), 200= , =,:)= :9):>
2 +ug (, 200= ,0 ,>): =,9)> =,0::)(: 9)2=
00 1ep 2, 200> ,022)0= (,(()(0 =,200)90 2)>2
-
7/25/2019 Ch 20 Hull Fundamentals 8 the d
11/48
Scenarios 6enerated Ta2le "#.'% page &'1)
Fundamentals of Futures and Options Markets, 8th Ed, Ch 20, Copyright John C. ull20!"
11
S#enai% DJIA FTSE 100 CAC 40 Niei !!" &%t'%li%Value ($000s)
%ss($000s)
0,( (,>
-
7/25/2019 Ch 20 Hull Fundamentals 8 the d
12/48
Ranked 7osses Ta2le "#.&% page &')
Fundamentals of Futures and Options Markets, 8th Ed, Ch 20, Copyright John C. ull20!"
12
S#enai% Nu*e %ss ($000s)
9(9 92)209
:2( 2
22< 2
-
7/25/2019 Ch 20 Hull Fundamentals 8 the d
13/48
The 8-day VaR
heN&day "aR for $ar!et ris! is usuallyassu$ed to be ti$es the one&day "aR
-n our exa$ple the 0&day "aR would becalculated as
his assu$ption is in theory only perfectlycorrect if daily changes are nor$allydistributed and independent
Fundamentals of Futures and Options Markets, 8th Ed, Ch 20, Copyright John C. ull20!"
13
N
274,801385,25310 =
-
7/25/2019 Ch 20 Hull Fundamentals 8 the d
14/48
The 9odel-Building Approach
he $ain alternative to historicalsi$ulation is to $a!e assu$ptions aboutthe probability distributions of the return on
the $ar!et variables and calculate theprobability distribution of the change in thevalue of the portfolio analytically
his is !nown as the $odel buildingapproach or the variance&covarianceapproach
Fundamentals of Futures and Options Markets, 8th Ed, Ch 20, Copyright John C. ull20!"
14
-
7/25/2019 Ch 20 Hull Fundamentals 8 the d
15/48
Fundamentals of Futures and Options Markets, 8th Ed, Ch 20, Copyright John C. ull20!"
4aily Volatilities
-n option pricing we express volatility asvolatility per year
-n "aR calculations we express volatilityas volatility per day
15
day
yea$
= 252
-
7/25/2019 Ch 20 Hull Fundamentals 8 the d
16/48
Fundamentals of Futures and Options Markets, 8th Ed, Ch 20, Copyright John C. ull20!"
4aily Volatility continued
1trictly spea!ing we should define dayas
the standard deviation of the continuously
co$pounded return in one day -n practice we assu$e that it is the
standard deviation of the percentagechange in one day
16
-
7/25/2019 Ch 20 Hull Fundamentals 8 the d
17/48
Fundamentals of Futures and Options Markets, 8th Ed, Ch 20, Copyright John C. ull20!"
9icrosoft Exa+ple
We have a position worth A0 $illion inBicrosoft shares
he volatility of Bicrosoft is 2% per day3about :2% per year5We useN 0 andX ((
17
-
7/25/2019 Ch 20 Hull Fundamentals 8 the d
18/48
Fundamentals of Futures and Options Markets, 8th Ed, Ch 20, Copyright John C. ull20!"
9icrosoft Exa+ple continued
he standard deviation of the change inthe portfolio in day is A200,000
he standard deviation of the change in0 days is
18
200 000 10 456, $632,=
-
7/25/2019 Ch 20 Hull Fundamentals 8 the d
19/48
Fundamentals of Futures and Options Markets, 8th Ed, Ch 20, Copyright John C. ull20!"
9icrosoft Exa+ple continued
We assu$e that the expected change inthe value of the portfolio is 'ero 3his isDE for short ti$e periods5
We assu$e that the change in the valueof the portfolio is nor$ally distributed
1ince ;3F2):2=50)0, the "aR is
19
300,471,1$456,632326.2 =
-
7/25/2019 Ch 20 Hull Fundamentals 8 the d
20/48
Fundamentals of Futures and Options Markets, 8th Ed, Ch 20, Copyright John C. ull20!"
AT:T Exa+ple
Consider a position of A $illion in+Ghe daily volatility of +G is %
3approx =% per year5he 1)4 per 0 days is
he "aR is
20
114,158$10000,50 =
800,367$326.2114,158 =
-
7/25/2019 Ch 20 Hull Fundamentals 8 the d
21/48
0ortfolio
;ow consider a portfolio consisting of bothBicrosoft and +G
+ssu$e that the returns of +G andBicrosoft are bivariate nor$al and that thecorrelation between the returns is 0):
Fundamentals of Futures and Options Markets, 8th Ed, Ch 20, Copyright John C. ull20!"
21
-
7/25/2019 Ch 20 Hull Fundamentals 8 the d
22/48
Fundamentals of Futures and Options Markets, 8th Ed, Ch 20, Copyright John C. ull20!"
S.4. of 0ortfolio
+ standard result in statistics states that
-n this case X 200,000 and+ 0,000
and 0):) he standard deviation of the
change in the portfolio value in one day istherefore 220,200
22
+X+X+X
++= +
222
-
7/25/2019 Ch 20 Hull Fundamentals 8 the d
23/48
Fundamentals of Futures and Options Markets, 8th Ed, Ch 20, Copyright John C. ull20!"
VaR for 0ortfolio
he 0&day ((% "aR for the portfolio is
he benefits of diversification are
3,9
-
7/25/2019 Ch 20 Hull Fundamentals 8 the d
24/48
Fundamentals of Futures and Options Markets, 8th Ed, Ch 20, Copyright John C. ull20!"
The 7inear 9odel
We assu$ehe daily change in the value of a portfolio
is linearly related to the daily returns fro$$ar!et variableshe returns fro$ the $ar!et variables are
nor$ally distributed
24
-
7/25/2019 Ch 20 Hull Fundamentals 8 the d
25/48
9arko;it< Result for Variance of
Return on 0ortfolio
Fundamentals of Futures and Options Markets, 8th Ed, Ch 20, Copyright John C. ull20!"
25
sinstru$entthand
thofreturnsbetweenncorrelatiois
portfolioin
instru$entthonreturnofvarianceis
portfolioininstru$entthofweightis
ReturnIortfolioof"ariance
iJ
2
i
,
i
i
i-
--
i
n
i
n
,
,i,ii,
= = =1 1
-
7/25/2019 Ch 20 Hull Fundamentals 8 the d
26/48
Fundamentals of Futures and Options Markets, 8th Ed, Ch 20, Copyright John C. ull20!"
VaR Result for Variance of
0ortfolio Value i= w
iP)
26
))&i=<
-
7/25/2019 Ch 20 Hull Fundamentals 8 the d
27/48
Covariance 9atrix vari= cov
ii)
Ta2le "#.>% page &&()
Fundamentals of Futures and Options Markets, 8th Ed, Ch 20, Copyright John C. ull20!" 27
=
nnnn
n
n
n
C
varcovcovcov
covvarcovcovcovcovvarcov
covcovcovvar
321
333231
223221
113121
-
7/25/2019 Ch 20 Hull Fundamentals 8 the d
28/48
Alternative Expressions for P
"
page &&>
Fundamentals of Futures and Options Markets, 8th Ed, Ch 20, Copyright John C. ull20!" 28
transposeitsisandisele$ent
thwhosevectorcolu$ntheiswhereT
T
?
?
??
i
&
,
n
i
n
,
ii,&
.
i
C=
= = =
2
1 1
2 cov
-
7/25/2019 Ch 20 Hull Fundamentals 8 the d
29/48
Fundamentals of Futures and Options Markets, 8th Ed, Ch 20, Copyright John C. ull20!"
*andling @nterest Rates
We do not want to define every bond as adifferent $ar!et variable
We therefore choose as assets 'ero&coupon
bonds with standard $aturities. &$onth, :$onths, year, 2 years, years, < years, 0years, and :0 years
Cash flows fro$ instru$ents in the portfolio are$apped to bonds with the standard $aturities
29
-
7/25/2019 Ch 20 Hull Fundamentals 8 the d
30/48
Fundamentals of Futures and Options Markets, 8th Ed, Ch 20, Copyright John C. ull20!"
hen 7inear 9odel Can 2e sed
Iortfolio of stoc!sIortfolio of bonds
#orward contract on foreign currency -nterest&rate swap
30
-
7/25/2019 Ch 20 Hull Fundamentals 8 the d
31/48
Fundamentals of Futures and Options Markets, 8th Ed, Ch 20, Copyright John C. ull20!"
The 7inear 9odel and ptions
Consider a portfolio of options dependenton a single stoc! price, S) 4efine
and
31
S&
=
S
Sx
=
-
7/25/2019 Ch 20 Hull Fundamentals 8 the d
32/48
Fundamentals of Futures and Options Markets, 8th Ed, Ch 20, Copyright John C. ull20!"
7inear 9odel and ptions
continued eDuations "#.& and "#.(% page &&3)
+s an approxi$ation
1i$ilar when there are $any underlying$ar!et variables
where iis the delta of the portfolio withrespect to the ith asset
32
xSS& ==
=i
iii xS&
-
7/25/2019 Ch 20 Hull Fundamentals 8 the d
33/48
Fundamentals of Futures and Options Markets, 8th Ed, Ch 20, Copyright John C. ull20!"
Exa+ple
Consider an invest$ent in options onBicrosoft and +G) 1uppose the stoc!prices are 20 and :0 respectively and the
deltas of the portfolio with respect to the twostoc! prices are ,000 and 20,000respectively
+s an approxi$ation
where xand x2are the percentagechanges in the two stoc! prices
33
21 000,2030000,1120 xx& +=
-
7/25/2019 Ch 20 Hull Fundamentals 8 the d
34/48
But the distri2ution of the daily
return on an option is not nor+al
See !igure "#.&% page &&1)
Fundamentals of Futures and Options Markets, 8th Ed, Ch 20, Copyright John C. ull20!"
Iositive Ka$$a ;egative Ka$$a
34
-
7/25/2019 Ch 20 Hull Fundamentals 8 the d
35/48
Translation of Asset 0rice Change
to 0rice Change for 7ong Call
!igure "#.(% page &&)
Fundamentals of Futures and Options Markets, 8th Ed, Ch 20, Copyright John C. ull20!"
6ong Call
+sset Irice
35
T l ti f A t 0 i Ch
-
7/25/2019 Ch 20 Hull Fundamentals 8 the d
36/48
Translation of Asset 0rice Change
to 0rice Change for Short Call!igure "#.>% page &&)
Fundamentals of Futures and Options Markets, 8th Ed, Ch 20, Copyright John C. ull20!"
1hortCall
+sset Irice
36
-
7/25/2019 Ch 20 Hull Fundamentals 8 the d
37/48
Fundamentals of Futures and Options Markets, 8th Ed, Ch 20, Copyright John C. ull20!"
Quadratic 9odel
#or a portfolio dependent on a singlestoc! price
where is the ga$$a of the portfolio)his beco$es
37
2)(
2
1SS& +=
22 )(21 xSxS& +=
-
7/25/2019 Ch 20 Hull Fundamentals 8 the d
38/48
se of Quadratic 9odel
+nalytic results are not as readily available/istorical si$ulation can be used in
conJunction with the uadratic $odel 3hisavoids the need to revalue the portfolio foreach si$ulation trial5
he uadratic $odel is also so$eti$esused in conJunction with a Bonte Carlosi$ulation
Fundamentals of Futures and Options Markets, 8th Ed, Ch 20, Copyright John C. ull20!"
38
-
7/25/2019 Ch 20 Hull Fundamentals 8 the d
39/48
Fundamentals of Futures and Options Markets, 8th Ed, Ch 20, Copyright John C. ull20!"
Esti+ating Volatility for 9odel
Building Approach eDuation "#.3% page &($)
4efine nas the volatility per day between day n#
and day n, as esti$ated at end of dayn#
4efine Sias the value of $ar!et variable at end of
day i 4efine ui$ ln3Si/Si15
he usual esti$ate of volatility fro$ observationsis.
39
=
=
=
=
i
in
i
inn
u)
u
uu)
1
1
22
1
)(1
1
-
7/25/2019 Ch 20 Hull Fundamentals 8 the d
40/48
Fundamentals of Futures and Options Markets, 8th Ed, Ch 20, Copyright John C. ull20!"
Si+plifications eDuations "#.1 and "#.% page &($)
4efineuias (SiSi1)/Si1+ssu$e that the $ean value of uiis 'ero
Replace 1by
his gives
40
n n ii)
)u2 2
1
1= =
-
7/25/2019 Ch 20 Hull Fundamentals 8 the d
41/48
Fundamentals of Futures and Options Markets, 8th Ed, Ch 20, Copyright John C. ull20!"
eighting Sche+e
-nstead of assigning eual weights to theobservations we can set
41
n i n ii
)
ii
)
u2 21
1
1
=
=
=
=
where
-
7/25/2019 Ch 20 Hull Fundamentals 8 the d
42/48
Fundamentals of Futures and Options Markets, 8th Ed, Ch 20, Copyright John C. ull20!"
E9A 9odel eDuation "#.$$% page &(")
-n an exponentially weighted $ovingaverage $odel, the weights assigned tothe u2decline exponentially as we $ovebac! through ti$e
his leads to
42
21
21
2 )1( += nnn u
-
7/25/2019 Ch 20 Hull Fundamentals 8 the d
43/48
Fundamentals of Futures and Options Markets, 8th Ed, Ch 20, Copyright John C. ull20!"
Attractions of E9A
Relatively little data needs to be storedWe need only re$e$ber the current
esti$ate of the variance rate and the $ostrecent observation on the $ar!et variable
rac!s volatility changes 0)(9 is a popular choice for daily
volatility forecasting
43
-
7/25/2019 Ch 20 Hull Fundamentals 8 the d
44/48
Fundamentals of Futures and Options Markets, 8th Ed, Ch 20, Copyright John C. ull20!"
Correlations
4efine ui2(3i3i1)/3i1and vi2(ViVi1)/Vi1+lso
un5daily vol of 3calculated on day n&vn. daily vol of Vcalculated on day n&
covn. covariance calculated on day n&
covn$ nunvn
where nis the correlation between 3and V
44
-
7/25/2019 Ch 20 Hull Fundamentals 8 the d
45/48
Fundamentals of Futures and Options Markets, 8th Ed, Ch 20, Copyright John C. ull20!"
Correlations continuedeDuation "#.$'% page &(&)
Lsing the *WB+
covn covn&H3&5un&vn&
45
9 d l B ildi *i t i l
-
7/25/2019 Ch 20 Hull Fundamentals 8 the d
46/48
Fundamentals of Futures and Options Markets, 8th Ed, Ch 20, Copyright John C. ull
20!"
9odel Building vs *istorical
Si+ulation Approaches
Bodel building approach has thedisadvantage that it assu$es that $ar!etvariables have a $ultivariate nor$aldistribution
/istorical si$ulation is co$putationallyslower and cannot easily incorporate
volatility updating sche$es
46
B k T ti
-
7/25/2019 Ch 20 Hull Fundamentals 8 the d
47/48
Fundamentals of Futures and Options Markets, 8th Ed, Ch 20, Copyright John C. ull
20!"
Back-Testing
ests how well "aR esti$ates wouldhave perfor$ed in the past
We could as! the uestion. /ow oftenwas the loss greater than the "aRlevel
47
St T ti
-
7/25/2019 Ch 20 Hull Fundamentals 8 the d
48/48
Fundamentals of Futures and Options Markets 8th Ed Ch 20 Copyright John C ull
Stress Testing
his involves testing how well aportfolio would perfor$ under so$eof the $ost extre$e $ar!et $oves
seen in the last 0 to 20 years
48