Reconstructing Volatility
Transcript of Reconstructing Volatility
Finance Concepts January 08 2004
1
Reconstructing VolatilityNew techniques for understanding the implied
volatility of multi-asset options
Speaker: Marco Avellaneda
Avellaneda, Boyer-Olson, Busca and Friz:`Reconstructing Volatility’, RISK Oct 2002; `Large Deviations Methods and thePricing of Index Options in Finance’, CRAS Paris 2003
Outline• Major US indices and ETFs
• Implied volatility surfaces of single stocks and indices
• Marginalization
• `Reconstructing’ the implied volatility of index options
• Steepest-descent Approximation
• The most likely market configurations
• Multivariate stochastic volatility models
• Moment-matching technique: Lee, Wang and Karim
• Cross-currency options
Finance Concepts January 08 2004
2
U.S. Equities: Main Sectors & Their Indices
• Major Indices: SPX, DJX, NDXSPY, DIA, QQQ (Exchange-Traded Funds)
• Sector Indices & Index Trackers: Semiconductors: SMH, SOX
Biotech: BBH, BTKPharmaceuticals: PPH, DRG
Financials: BKX, XBD, XLF, RKHOil & Gas: XNG, XOI, OSX
High Tech, WWW, Boxes: MSH, HHH, XBD, XCIRetail: RTH
All these indices have options
Trading statistics (AMEX)
AVERAGE TRADED VOLUME IN DOLLARS LAST MONTH
0 1000 2000 3000 4000
SPY
QQQ
DIA
MM USD
Finance Concepts January 08 2004
3
AVERAGE TRADED VOLUME OVER THE PAST MONTH (IN USD)
0 100 200 300 400 500 600
SMH
BBH
RTH
OIH
XLF
PPH
XLK
MM USD
Sectors
COMS CMGI LGTO PSFTADPT CNET LVLT PMCSADCT CMCSK LLTC QLGCADLAC CPWR ERICY QCOMADBE CMVT LCOS QTRNALTR CEFT MXIM RNWKAMZN CNXT MCLD RFMDAPCC COST MEDI SANMAMGN DELL MFNX SDLIAPOL DLTR MCHP SEBLAAPL EBAY MSFT SIALAMAT DISH MOLX SSCCAMCC ERTS NTAP SPLSATHM FISV NETA SBUXATML GMST NXTL SUNWBBBY GENZ NXLK SNPSBGEN GBLX NWAC TLABBMET MLHR NOVL USAIBMCS ITWO NTLI VRSNBVSN IMNX ORCL VRTSCHIR INTC PCAR VTSSCIEN INTU PHSY VSTRCTAS JDSU SPOT WCOMCSCO JNPR PMTC XLNXCTXS KLAC PAYX YHOO
COMS CMGI LGTO PSFTADPT CNET LVLT PMCSADCT CMCSK LLTC QLGCADLAC CPWR ERICY QCOMADBE CMVT LCOS QTRNALTR CEFT MXIM RNWKAMZN CNXT MCLD RFMDAPCC COST MEDI SANMAMGN DELL MFNX SDLIAPOL DLTR MCHP SEBLAAPL EBAY MSFT SIALAMAT DISH MOLX SSCCAMCC ERTS NTAP SPLSATHM FISV NETA SBUXATML GMST NXTL SUNWBBBY GENZ NXLK SNPSBGEN GBLX NWAC TLABBMET MLHR NOVL USAIBMCS ITWO NTLI VRSNBVSN IMNX ORCL VRTSCHIR INTC PCAR VTSSCIEN INTU PHSY VSTRCTAS JDSU SPOT WCOMCSCO JNPR PMTC XLNXCTXS KLAC PAYX YHOO
Components of NASDAQ 100 Trust (AMEX:QQQ)
�QQQ trades as a stock
�QQQ index options are the mostheavily traded contract in AMEX
� Capitalization-weighted averageof 100 largest stocks in NASDAQ
Finance Concepts January 08 2004
4
Morgan Stanley 35 (MSH)
ADP JDSUAMAT JNPRAMZN LUAOL MOTBRCM MSFTCA MUCPQ NTCSCO ORCLDELL PALMEDS PMTCEMC PSFTERTS SLRFDC STMHWP SUNWIBM TLABINTC TXNINTU XLNX
YHOO
ADP JDSUAMAT JNPRAMZN LUAOL MOTBRCM MSFTCA MUCPQ NTCSCO ORCLDELL PALMEDS PMTCEMC PSFTERTS SLRFDC STMHWP SUNWIBM TLABINTC TXNINTU XLNX
YHOO
�35 Underlying Stocks
� Equal-dollar weighted index, adjustedannually
�Each stock has typically O(30) options over a 1yr horizon
SOX, XNG, XOI
XNG
APAAPCBRBRREEXENEEOGEPGKMINBLNFGOEIPPPSTRWMB
… & many others
SOX
ALTRAMATAMDINTCKLACLLTCLSCCLSIMOTMUNSMNVLSRMBSTERTXNXLNX
XOI
AHCBPCHVCOC.BXOMKMGOXYPREPRDSUNTXTOTUCLMRO
Finance Concepts January 08 2004
5
Example: Basket of 20 Biotechnology Stocks ( Components of BBH)
32-BBH5614GENZ
476.8271SHPGY813ENZN
846SEPR53.544DNA
645QLTI554CRA
9212MLNM3716CHIR
8215MEDI4113BGEN
7212IDPH4046AMGN
644ICOS1064ALKS
848HGSI644AFFX
468GILD5518ABI
ATM ImVolSharesTickerATM ImVolSharesTicker
75
80
85
90
95
100
ImpliedVol
VarSwap
ImpliedVol 96.6191 94.5071 88.4581 83.9929 81.7033 82.5468 81.4319 80.1212 78.6667 80.7064 78.8035
VarSw ap 87.1215 87.1215 87.1215 87.1215 87.1215 87.1215 87.1215 87.1215 87.1215 87.1215 87.1215
31.3 32.5 33.8 35 37.5 40 41.3 42.5 43.8 45 46.3
Strike
Vol.
AOL Jan 2001 Options:Implied volatility curve on Dec 20,2000
Market close
Stock probability is not lognormal
Finance Concepts January 08 2004
6
60
62
64
66
68
70
72
74
76
78
80
30.00 32.50 35.00 37.50 40.00 42.50 45.00 47.50 50.00
ImpliedVol
VarSwap
54
56
58
60
62
64
66
68
70
30.00
32.50
35.00
37.50
40.00
42.50
45.00
47.50
50.00
55.00
60.00
ImpliedVol
VarSwap
46
48
50
52
54
56
58
60
62
30.00
35.00
40.00
45.00
50.00
60.00
70.00
ImpliedVol
VarSwap40
42
44
46
48
50
52
54
56
58
27.50
35.00
42.50
50.00
57.50
65.00
72.50
80.00
87.50
100.0
0
ImpliedVol
VarSwap
Expiration2/17/01
Expiration4/21/01
Expiration7/21/01
Expiration1/19/02
The AOL ``volatility skew’’ for several expiration dates
30
35
40
45
50
55
75 80 85 90 95 100 105
strike
vol bid vol
ask vol
BBH March 2003 Implied VolsPricing Date: Jan 22 03 10:42 AM
Finance Concepts January 08 2004
7
DJX Mar 03 Pricing Date: 10/25/02
20
22
24
26
28
30
88 85 77 67 62 56 44 38 32 21 14
Delta
Vo
l BidVol
AskVol
Implied Volatility Curve for Options on Dow Jones Average
Stylized facts about equity volatility curves
� Implied volatility curves are typically downward sloping
� Counterexamples: precious metal stocks are upward sloping
� There is little curvature (or smile). Skew is important.
Finance Concepts January 08 2004
8
( )
( ) ( ) ( )( ) Fit Empirical/ln1,,
Volatility Stochastic
Volatility Local sDupire',
impliedimplied SKaTSTK
dZd
tS
dWS
dS
tt
t
t
t
+⋅=
=
=
=
σσ
κσσ
σσ
σ
Modeling single-stock volatility curves
Price dynamicsnot lognormal
Jumps …
What is the relation between index options and options on the
components?
Standard (log-normal) Volatility Formula for Index Options
ijjiji
jij
N
jjI ppp ρσσσσ ∑∑
≠=
+= 2
1
22 (*)
Does not apply when volatilities are strike-dependent
How can we incorporate volatility skew information into (*)?
Finance Concepts January 08 2004
9
“Marginalization”
( ) ( )dttYdWtYYd ,, κκ +=
( )
( ) ( )
( ) ( )
+=
=+=
∈
∑
∑
=
=
+
dttYXdWtYXdY
nidttYXdWtYXdX
RYX
m
j
jj
im
j
jij
i
n
,,,,
,,...,1 ,,,,
,
1
1
1
νκ
µσ
( )000 ,, tYX
Consider an n+1 dimensional diffusion:
Problem: given a starting point, , find a one-dimensional diffusion
such that and have the same probability distributions for all( )tY ( )tY 0tt >
( ) ( )
( ) ( )( ) ( )
( ) ( )( ) ( )
...,,...
...,,2
1,,
,,;,,,,
2
22
2
2
000
∑∑
∑
∂∂−
∂∂−
∂∂∂
+∂∂
∂+∂∂=
∂∂
≡
ii
ijji
ii
xtyx
yxx
xytyx
yt
tyx
tyxtYXtyx
νπ
πκπ
ππ transition probability function
…
Fokker-Planck Equation
Finance Concepts January 08 2004
10
( ) ( )
( ) ( )( ) ( )( )
( ) ( )( ) ( ) ( )
( ) ( )
⋅∂∂−
⋅∂∂=
∂∂
∴
⋅∂∂−⋅
∂∂=
∂∂
≡ ∫
tyty
ty
yty
ty
ty
yt
ty
tyy
tyyt
ty
xdtyxty n
,,
,,,
,
,,
2
1,
,,,,2
1,
,,,
2
2
2
22
2
ππ
νππππκπ
νππκπ
ππ
Computing the marginal distributions…
integration with respectto x eliminates terms withx-derivatives
new 1D diffusion diffusion equation
Equivalent 1-D diffusion
( ) ( ) ( )( )
( ) ( )( )
( ) ( ) ( )( )
( ) ( )( )
( ) ( ) ( )( ) ( ){ }
( ) ( ) ( )( ) ( ){ }ytYttYtXEty
ytYttYtXEty
xdtyx
xdtyxtyx
ty
tytyty
xdtyx
xdtyxtyx
ty
tytyty
n
n
n
n
==
==
=⋅⋅≡
=⋅⋅≡
∫∫
∫∫
,,,
,,,
,,
,,,,
,
,,,,,
,,
,,,,
,
,,,,,
22
222
νν
κκ
π
πνπ
πνν
π
πκπ
πκκ
Equivalent parameters = conditional expectations of local parameters
Finance Concepts January 08 2004
11
Application to Index Options
( )
( )
( ) ( )
( )( ) ( )
( )I
SwtS
I
SwSwtStS
tS
dttSdZtSI
dI
dtdWdWE
drdtdWtSS
dS
SwI
iiii
ijijjjiijjii
ijji
iiiiiii
i
i
n
ii
∑∑
∑
==
+=
=
−=+=
==
µµ
ρσσσ
µσ
ρ
µµσ
,
,,
,
,,
, ,,
loc2loc2
locloc
1
Index = weighted sum of stock prices (constant weights)
Diffusion eq.for each stockreflects vol skew
Characterization of the equivalent volatility for the index
( )( )( ) ( )( ) ( ) ( )
( )
( )( ) ( )( ) ( )( ) ( )( ) ( )
( ) .,...,1 ,
,,
,,
,2
2
niSw
SwSp
ItSwttSttStSptSpE
ItSwI
wwtStSttSttS
EtI
jjj
iii
iii
ijijjjiiji
iii
ijijjijijjii
==
==
==
∑
∑∑
∑∑
ρσσ
ρσσσ
• sigma_loc can be seen asa ‘stochastic vol’ driving the index
• sigma_bar is then the ``averaged vol”
Finance Concepts January 08 2004
12
Varadhan’s Formula
( ) { }( )
( ) ( ){ } ( ) ( ) 1 , 2
,0Prob. log
0
,
22
1
<<−≈==
=
==∑=
tt
yxdxXytX
xX
dtdWdWEdWtXdX
ii
n
jjkkjj
jii
σ
ρσ
( )( ) ( )
( )( ) ( ) ( )
( ) ( ) ( ) ( ) ( ) ijjiij
jin
ijij
yx
xxxaxaxg
dssssgyxd
ρσσ
γγγγγ
0,0,
inf,
1
1
0 11,0
2
==
=
−
⋅⋅
=== ∫∑ Riemannian metric
In practice: dimensionlesstime ~ 0.02
Dupire localvolatility model for each stock
Steepest-descent approximation
( ) ( ) .,...,2,1 log 0
log nitF
S
eS
Sx
i
it
i
ii i
=
=
≡ µ
( ) ( ) ( )
( ) ( )
==
=≅
∑i
xtii
xtii
IeeSwxdx
eeSStStI
ii
ii
µ
µσσ
0,0minarg
0 ,,
2*
**loc
22 *
Change to log-scale:
( ) ( )( ){ }( )( ){ }ItIE
ItIEtI
−−=
δδσσ
2loc
2,Formally,
Applying Varadhan’s Formula,
where
Finance Concepts January 08 2004
13
Steepest Descent=Most Likely Stock Price Configuration
IeF ix
ii =∑
Replace conditional distribution by “Dirac function” at most likely configuration
x*
Characterization of MLC
( ) ( ) ( ) ( ) ( ) ijjjiij
n
iji xxxpxptI ρσσσ ***
1
*2, ∑
=
=
( ) ( ) ( )
( )
=
==
∑
∫ ∑
=
+
=
n
i
txii
x
ijjj
n
jj
i
IeSw
nixxpu
du
ii
i
1
0
*
1
*
*
*
0
,...,1
µ
ρσλσ
Euler-Lagrange equations: find such that( )λ,*x
Finance Concepts January 08 2004
14
Solution of linearized systemin terms of the stock betas
( ) ( ) ( ) ( ) ( )
( )
( ) ( ) ( ) ( ) ( ) ( )
∆
∆∆==
=≅
≡
≡
∑
∑
=
=
I
IVar
I
I
S
SCovxx
xxCovx
px
x
eI
Ix
pp
ii
iI
jij
n
jij
Ii
t
ijjij
n
ijiI
/,
,0
0000
0ln
00000
^^*
21
2*
1
2
ββ
σσσρ
σ
ρσσσ
µ
Most likely config. : described by the risk-neutral regression coefficients of stock returns with the index return (``micro” CAPM)
Expression in terms of Implied Volatilities and Black- Scholes Deltas
( ) ( )
( ) ( ) ( )( )xx
u
du
xx
x
σσσ
σσ
+≈
≈
−
∫
02
1
1
impl.impl.
1
0
impl.
� Seek direct relation between implied volatilities of single-stock options and implied volatility of index options
� Tool: Berestycki-Busca-Florent large-deviations result for single-stock(“1/2 slope rule”)
303234363840424446
-2 -1.5 -1 -0
.5 00.
5 11.
5 2
log(K/S(0))
vol (
%)
local vol
implied vol
Finance Concepts January 08 2004
15
The Formula In Terms of BlackScholes Deltas
( )( )
( )( )
( )( )
( )( )
( ) ( )( )
×∆=∆
=
+
≡∆
∆∆⋅=
×≈
∑
∫
∑
=
−
∞−
−
=
f
jjn
jjijIi
y z
i
if
f
jjn
jjij
f
ii
ii
I
FpNN
dzeyNtS
F
tN
I
I
S
SCorr
I
II
I
Fp
I
II
S
FS
implI
impl.
1
1
2*
implI
implI
impl.
1implI
*impl.
*
2
2
1)( ,
2
1ln
1
,)/ln)/ln/ln
σσ
ρ
πσ
σ
σσσ
ρσσ
From Euler-Lagrange equations:
Translate into Black-Scholes deltas
Delta of eachstock is a function ofdelta of index option
Reconstruction Formula for the Index Volatility
( ) ( ) ( )
( ) ( ) ( )( ) ( ) ( )( )
5.025.02
2
5.0
1
impl.impl.impl.impl.
impl.impl.
=
−∆−∆=∆
∆+≅∆
∑=
i
n
ijjjjiiijiijI
IIII
p
pp σσσσρσ
σσσ
capitalization of stock in index (%)
Finance Concepts January 08 2004
16
Correlation & Optimal Configurations
( ) ( )( )
( )( )
( ) iI
II
S
FS
IFp
f
ii
ii
IifI
n
jjjj
∀≈
∆≡∆∴=∑=
)/ln/ln
implI
*impl.
*
impl
1
impl
σσ
σσ�If the stocks are perfectly correlated,
“ Equal-delta approximation “
�If the stocks are uncorrelated,
( )( )
( )( )
( )( )
( ) ( )( )
××∆=∆
××≈
−
fI
iiiIi
fI
jii
f
ii
ii
I
FpNN
I
Fp
I
II
S
FS
impl
impl.1
impl
impl.
implI
*impl.
* )/ln/ln
σσ
σσ
σσ Less volatile & smallstocks have smallerdeltas closer to 50%(their skew does notaffect the index as much)
DJX: Dow Jones Industrial Average
DJX Nov 02 Pricing Date: 10/25/02
20
25
30
35
40
89 86 81 78 72 67 61 54 47 40 34 27 20 15
Delta
Vo
l
BidVol
AskVol
SDA
T=1 month
Finance Concepts January 08 2004
17
DJX Dec 02 Pricing Date: 10/25/02
2022
2426
2830
3234
87 82 80 74 71 67 59 54 50 46 41 37 32 28 21 14
Delta
Vo
l
BidVol
AskVol
SDA
T= 2 months
DJX Jan 03 Pricing Date: 10/25/02
202224262830323436
88 84 79 75 69 62 55 48 40 33 26 15
Delta
Vo
l
BidVol
AskVol
SDA
T=3 months
Finance Concepts January 08 2004
18
DJX Mar 03 Pricing Date: 10/25/02
20
22
24
26
28
30
32
88 85 77 67 62 56 44 38 32 21 14
Delta
Vo
l
BidVol
AskVol
SDA
T= 5 months
DJX June 03 Pricing Date: 10/25/02
20
22
24
26
28
89 87 83 75 66 57 48 43 38 29 21 15
Delta
Vo
l
BidVol
AskVol
SDA
T=7 months
Finance Concepts January 08 2004
19
BBH Nov 02 Date: Oct 25 02
30
35
40
45
50
55
60
89 79 62 41 22
Delta
Vol
BidVol
AskVol
WKB
BBH: Biotechnology HLDRT = 1 month
BBH Dec 02 Date: Oct 25 02
30
35
40
45
50
55
60
88 81 72 60 47 34 23
Delta
Vo
l
BidVol
AskVol
SDA
T = 2 months
Finance Concepts January 08 2004
20
BBH Apr 03 Date: Oct 25 02
30
35
40
45
50
55
90 86 81 75 68 61 53 45 37
Delta
Vo
l BidVol
AskVol
SDA
T = 6 months
Is dimensionless time is too long? (Error bars: Juyoung Lim)
Is correlation causing the discrepancy?
Expiration: Sep 02
15
20
25
30
35
40
440
445
450
455
460
465
470
475
480
485
490
495
500
505
strike
impl
ied
vol
BidVol
AskVol
WMC vol
Steepest Desc
S&P 100 Index Options(Quote date: Aug 20, 2002)
Finance Concepts January 08 2004
21
Expiration: Oct 02
15
20
25
30
35
40
430
440
450
460
470
480
490
500
510
520
strike
imp
lied
vol BidVol
AskVol
WMC vol
Steepest Desc
S&P 100 Index Options(Quote date: Aug 20, 2002)
S&P 100 Index Options(Quote date: Aug 20, 2002)
Expiration: Nov 02
15
17
19
21
23
25
27
29
31
33
430 440 450 460 470 480 490 500 510 520
strike
imp
lied
vo
l
BidVol
AskVol
WMC vol
Steepest Desc
Finance Concepts January 08 2004
22
S&P 100 Index Options(Quote date: Aug 20, 2002)
Expiration: Dec 02
15
17
19
21
23
25
27
29
31
33
420 440 460 470 480 490 500 510 520 530 540
strike
imp
lied
vol BidVol
AskVol
WMC vol
Steepest Desc
General Stochastic Volatility Systems
( ) ( )
i
ii
i
ii
ijjiijji
iii
iii
i
i
dy
S
dSx
I
dIx
dtrdZdWEdtdWdWE
dZd
dWS
dS
σσ
ρ
κσσσ
===
==
==
,
( )nn yyxx ,...,,,..., 11
Look for most likely configuration of stocks and volscorresponding to a given index
displacement x
Finance Concepts January 08 2004
23
Most likely configuration forStochastic Volatility Systems
ijxx
ji
n
ijjiI
I
iIiiii
I
iIiiii
ji eettpptx
rxy
xx
ρσσσ
σκγγ
σρσββ
γγ),0(),0(),(1
2
*
*
∑=
≅
==
==Most likely configurationfor stocks moves and volatility moves, giventhe index move
SDA
Numerical Example
%40
5.0 %,301 %,20
5.02
2211
21
=−==−==
===
ργσγσ
ppn
10%
15%
20%
25%
30%
35%
40%
45%
50%
-1.0
-0.8
-0.6
-0.4
-0.2 0.
00.2 0.
40.
60.8 1.
0
dI/ILocal Index volImplied Index vol
Finance Concepts January 08 2004
24
Method I: Dupire & Most LikelyConfiguration for Stock Moves
N-dimensionalEquity market
( )tx ,11σ
( )txii ,σ
( )txnn ,σ
( )txI ,σ
• Step 1: Local volatilityfor each stock
•Step 2: Find most likelyconfiguration for stocks
Method II: Stochastic Volatility System and joint MLC for Stocks and Volatilities
N-dimensionalEquity market ( )txI ,σ
•Only one step: compute themost likely configuration ofstocks and volatilities at thesame time
Finance Concepts January 08 2004
25
Are methods I and II ``equivalent’’?
( ) ( )i
iiii
xiii
rettx ii
σκϖσσ ϖ =≈ ,0,
( ) ( ) ( ) xxijjij
ijiI
jjii eettpptxβϖβϖρσσσ ,0,0,2 ∑=
Answer is NO, in general.
( ) ( ) ( ) xxijjij
ijiI
ji eettpptx γγρσσσ ,0,0,2 ∑=
Dupire local vol. forsingle names
Index vol.,Method I
Index vol.,Method II
Equivalence holds only under additional assumptions on stock-volatility correlations
ijiiij
iIiiiI
I
iIii
I
iIiii
I
iIi
i
iiiii
rr
rr
r
rr
ρ
ρ
σκγ
σρκ
σρσ
σκβϖ
=
=
=
== Method I
Method II
Conditions underwhich both methods give equivalent valuations
Finance Concepts January 08 2004
26
The corresponding stock-volatility correlation structure is `diagonal’
( )
iiiiiiijiiij
kjk
ikij
lkiikk
iki
ddWrdzrr
mr
ddWEddWmdz
ςαρ
ρ
ςςα
+=⇒=
=∴
=+=
∑
∑ 0,
Examples: Dupire, CEV, uncoupled SV mode
%50 ,7.0
5.0
6.0
7.0
%40 %,30 %,20
21
21
==
−−
−−
=
===
κκ
ρσσ
r
10%
15%
20%
25%
30%
35%
40%
-0.5 -0.4 -0.3 -0.2 -0.1 -0 0.1 0.2 0.3 0.4 0.5
X
IVO
L(%
)
Method II
Method I
Numerical Example
Finance Concepts January 08 2004
27
Lee, Wang and KarimRISK, Dec 2003Gram-Charlier expansion
( ) ( )
( ) SmilexSkewxx
k
TSmile
Tks
TSkew
Tk
Ts
xEk
xEsxVar
F
Kx
⋅+⋅+=
=
+=
−++=
−===
=
2atm
2atm
4
4
3
32
!41
,!4
2!3
1
147
!4!31
3 , ),( ,log
σσ
σσ
σσσσ
σσσ
Calculate Index Moments
( )( ) ( )
( ) ( )lkjilkjilkjijk
i
kjikjikjijk
i
ijjijij
i
zzzzEppppxE
zzzEpppxE
ppxE
σσσσ
σσσ
ρσσ
∑
∑
∑
=
=
=
4
3
2
``Gaussian’’ closure for third and 4th moments
Finance Concepts January 08 2004
28
3rd Moments
( )
( ) ( )
( ) ( )
iikijkjikjijk
iI
iiiii
iikikikii
iikijikikikji
sppps
szEzzzE
szEzzCorrzzzE
kjisSymszzCorrzzCorrSymzzzE
ρρσσσσ
ρ
ρρ
∑=
==
==
≠≠==
3
3
3
1
),(
if ][]),(),([
Consistent with the SDA approximation formula!
Stochastic Correlation Model
• Lee, Wang and Karim: stochastic correlation
• Linear fit
34.0:
66.0 :
ln
−=−=
++=
ββ
εβαρ
BKX
OEX
I Rho_bar is the`average’ correlation
This model can be used to improve the accuracy of the SDA.
Finance Concepts January 08 2004
29
Cross-Currency OptionsMajor currency triad: USD/JPY, USD/EUR, JPY/EUR
( )
( )
( )
222112
21
2atm,
12
)(
2
222
1
111
2
0
ln/
0ln/
0ln/
12
2
1
σσσρσσ
µµ
µ
µ
+−=
−=
==
==
==
−
I
t
t
t
xxx
eI
IxJPYEURI
eS
SxUSDEURS
eS
SxUSDJPYS
-1
-0.8
-0.6
-0.4
-0.2
-1.3
878E
-16
0.2
0.4
0.6
0.8 1
-1-0.9-0.8-0.7-0.6-0.5-0.4-0.3-0.2-0.1-1.38778E-160.10.20.30.40.50.60.70.80.91
X-bar=-0.5
Sig(USD/JPY)=20 pctSig(USD=11.2 pctRho=50 pct
X-bar=+0.6
Formally, the answer is the sameas in the index case. The ``geometry’’is different.