A Benchmark Approach To Quantitative Finance - Lecture Notes
Uploaded by
shuheng1108A Benchmark Approach To Quantitative Finance - Lecture Notes
Uploaded by
shuheng1108A Benchmark Approach to Quantitative Finance
Eckhard Platen
School of Finance and Economics and Department of Mathematical Sciences
University of Technology, Sydney
Lit: Platen, E. & Heath, D.: A Benchmark Approach to Quantitative Finance
Springer Finance, 700 pp., 199 illus., Hardcover, ISBN-10 3-540-26212-1 (2006).
Platen, E.: A benchmark approach to finance. Mathematical Finance 16(1), 131-151 (2006).
Bühlmann, H. & Platen, E.: A discrete time benchmark approach for insurance and finance
ASTIN Bulletin 33(2), 153-172 (2003).
Contents
1 Asset Pricing 1
2 Continuous Financial Markets 55
3 Portfolio Optimization 91
4 Modeling Stochastic Volatility 107
5 Minimal Market Model 154
6 Stylized Multi-Currency MMM 214
7 Valuing Guaranteed Minimum Death Benefits 219
8 Markets with Event Risk 273
References 337
1 Asset Pricing
Law of One Price
“All replicating portfolios of a payoff have the same price!”
Debreu (1959), Sharpe (1964), Lintner (1965),
Merton (1973a, 1973b), Ross (1976), Harrison & Kreps (1979),
Cochrane (2001), . . .
will be, in general, violated under the benchmark approach.
c Copyright Eckhard Platen 10 BA - Toronto C1 1
1
ln(savings bond)
ln(fair zero coupon bond)
0 ln(savings account)
-1
-2
-3
-4
-5
-6
-7
-8
1930 1940 1950 1960 1970 1980 1990 2000
time
Figure 1.1: Logarithms of savings bond, fair zero coupon bond and savings
account.
c Copyright Eckhard Platen 10 BA - Toronto C1 2
Financial Market
• jth primary security account
Stj
j ∈ {0, 1, . . . , d}
• savings account
St0
t≥0
c Copyright Eckhard Platen 10 BA - Toronto C1 3
• strategy
δ = {δt = (δt0 , δt1 , . . . , δtd )⊤ , t ≥ 0}
predictable
• portfolio
d
X
Stδ = δtj Stj
j=0
• self-financing
d
X
dStδ = δtj dStj
j=0
c Copyright Eckhard Platen 10 BA - Toronto C1 4
• strictly positive portfolio
S0δ = x > 0
Stδ ∈ (0, ∞) for all t ∈ [0, ∞)
=⇒ S δ ∈ Vx+
c Copyright Eckhard Platen 10 BA - Toronto C1 5
Numeraire Portfolio
Definition 1.1 S δ∗ ∈ Vx+ numeraire portfolio if
Sδ
t+h
δ∗
St+h
Et − 1 ≤ 0
Stδ
Stδ∗
for all nonnegative S δ and t, h ∈ [0, ∞).
• S δ∗ “best” performing portfolio
Long (1990), Becherer (2001), Pl. (2002, 2006),
Bühlmann & Pl. (2003), Goll & Kallsen (2003), Karatzas & Kardaras (2007)
c Copyright Eckhard Platen 10 BA - Toronto C1 6
Main Assumption of the Benchmark Approach
Assumption 1.2 There exists a numeraire portfolio S δ∗ ∈ Vx+ .
c Copyright Eckhard Platen 10 BA - Toronto C1 7
$US
WSI
8,000 EWI
7,000
6,000
5,000
DWI
4,000
MCI
3,000
2,000
1,000
0
1/1/73 28/9/76 25/6/80 22/3/84 18/12/87 15/9/91 12/6/95 9/3/99 4/12/02 31/8/06
Figure 1.2: Constructed MCI, DWI, EWI and WSI. In the long term the WSI
and EWI outperform all other indices, Le & Platen (2006).
c Copyright Eckhard Platen 10 BA - Toronto C1 8
• benchmarked value
Stδ
Ŝtδ =
Stδ∗
Supermartingale Property
Corollary 1.3 For nonnegative S δ
Ŝtδ ≥ Et Ŝsδ
0≤t≤s<∞
Ŝ δ supermartingale
c Copyright Eckhard Platen 10 BA - Toronto C1 9
• long term growth rate
!
δ 1 Stδ
g = lim sup ln
t→∞ t S0δ
Theorem 1.4 For S δ ∈ Vx+
g δ ≤ g δ∗ .
• “pathwise best” in the long run
Karatzas & Shreve (1998), Pl. (2004a),
Karatzas & Kardaras (2007)
c Copyright Eckhard Platen 10 BA - Toronto C1 10
Pl. (2004a)
Definition 1.5 Nonnegative portfolio S δ outperforms systematically
S δ̃ if
(i) S0δ = S0δ̃ ;
(ii) P Stδ ≥ Stδ̃ = 1
(iii) P Stδ > Stδ̃ > 0.
• “relative arbitrage” Fernholz & Karatzas (2005)
c Copyright Eckhard Platen 10 BA - Toronto C1 11
1
ln(fair zero coupon bond)
ln(savings account)
0
-1
-2
-3
-4
-5
-6
-7
-8
1930 1940 1950 1960 1970 1980 1990 2000
time
Figure 1.3: Logarithms of fair zero coupon bond and savings account.
c Copyright Eckhard Platen 10 BA - Toronto C1 12
Theorem 1.6 Numeraire portfolio cannot be outperformed systematically.
c Copyright Eckhard Platen 10 BA - Toronto C1 13
• portfolio ratio
δ
St+h
Aδt,h =
Stδ
• expected growth
δ
gt,h = Et ln Aδt,h
t, h ≥ 0
c Copyright Eckhard Platen 10 BA - Toronto C1 14
Definition 1.7 S δ growth optimal if
δ δ
gt,h ≤ gt,h
for all t, h ≥ 0 and nonnegative S δ .
• expected log-utility
Kelly (1956)
Hakansson (1971a)
Merton (1973a)
Roll (1973)
Markowitz (1976)
c Copyright Eckhard Platen 10 BA - Toronto C1 15
Theorem 1.8 The numeraire portfolio is growth optimal.
c Copyright Eckhard Platen 10 BA - Toronto C1 16
Strong Arbitrage
• market participants can only exploit arbitrage
• limited liability
=⇒ nonnegative total wealth of each market participant
Definition 1.9 A nonnegative S δ is a strong arbitrage if S0δ = 0 and
Stδ
P > 0 > 0.
Pl. (2002)-mathematical arguments (super-martingale property)
Loewenstein & Willard (2000)-economic arguments
c Copyright Eckhard Platen 10 BA - Toronto C1 17
Theorem 1.10 There is no strong arbitrage.
c Copyright Eckhard Platen 10 BA - Toronto C1 18
• Delbaen & Schachermayer (1998)
free lunches with vanishing risk (FLVR)
• Loewenstein & Willard (2000)
free snacks & cheap thrills
=⇒ no-arbitrage pricing makes no sense under BA
c Copyright Eckhard Platen 10 BA - Toronto C1 19
1
0.8
0.6
0.4
0.2
-0.2
1930 1940 1950 1960 1970 1980 1990 2000
time
Figure 1.4: P (t, T ) minus savings account.
c Copyright Eckhard Platen 10 BA - Toronto C1 20
0.0018
0.0016
0.0014
0.0012
0.001
0.0008
0.0006
0.0004
0.0002
-0.0002
-0.0004
1920 1922 1924 1926 1928
time
Figure 1.5: P (t, T ) minus savings account.
c Copyright Eckhard Platen 10 BA - Toronto C1 21
• Existence of equivalent risk neutral probability measure is
more a mathematical convenience than an economic necessity.
• Basing quantitative methods on classical risk neutral pricing is
restrictive.
• Trends are ignored under
Fundamental Theorem of Asset Pricing
Delbaen & Schachermayer (1998)
c Copyright Eckhard Platen 10 BA - Toronto C1 22
• Equivalent to risk neutral pricing are pricing via:
stochastic discount factor, Cochrane (2001);
deflator, Duffie (2001);
pricing kernel, Constantinides (1992);
state price density, Ingersoll (1987);
numeraire portfolio as in Long (1990)
All ignore trends !
c Copyright Eckhard Platen 10 BA - Toronto C1 23
• One needs consistent pricing concept that exploits trends
• All benchmarked nonnegative securities are supermartingales
(no upward trend)
c Copyright Eckhard Platen 10 BA - Toronto C1 24
Definition 1.11 Price is fair if, when benchmarked, forms martingale
(no trend)
Ŝtδ = Et Ŝsδ
0 ≤ t ≤ s < ∞.
c Copyright Eckhard Platen 10 BA - Toronto C1 25
Lemma 1.12 The minimal nonnegative supermartingale that reaches a
given benchmarked contingent claim is a martingale.
see Pl. & Heath (2006)
c Copyright Eckhard Platen 10 BA - Toronto C1 26
0.12
benchmarked savings bond
benchmarked fair zero coupon bond
0.1
0.08
0.06
0.04
0.02
0
1930 1940 1950 1960 1970 1980 1990 2000
time
Figure 1.6: Benchmarked savings bond and benchmarked fair zero coupon
bond.
c Copyright Eckhard Platen 10 BA - Toronto C1 27
Law of the Minimal Price
Pl. (2008)
Theorem 1.13 If a fair portfolio replicates a nonnegative payoff, then this
represents the minimal replicating portfolio.
• least expensive
• minimal hedge
• economically correct price in a competitive market
c Copyright Eckhard Platen 10 BA - Toronto C1 28
• contingent claim
HT !
HT
E0 <∞
STδ∗
• StδH fair if
!
StδH HT
ŜtδH = = Et
Stδ∗ STδ∗
Link with real economy !
c Copyright Eckhard Platen 10 BA - Toronto C1 29
Law of the Minimal Price =⇒
Corollary 1.14
Minimal price for replicable HT is fair and given by
real world pricing formula
!
HT
StδH = Stδ∗ Et .
STδ∗
c Copyright Eckhard Platen 10 BA - Toronto C1 30
• most pricing concepts are unified and generalized by real world pricing
=⇒
actuarial pricing
risk neutral pricing
pricing with stochastic discount factor
pricing with numeraire change
pricing with deflator
pricing with state pricing density
pricing with pricing kernel
pricing with numeraire portfolio
c Copyright Eckhard Platen 10 BA - Toronto C1 31
When HT independent of STδ∗
=⇒ rigorous derivation of
• actuarial pricing formula
StδH = P (t, T ) Et (HT )
with zero coupon bond
!
1
P (t, T ) = Stδ∗ Et
STδ∗
c Copyright Eckhard Platen 10 BA - Toronto C1 32
• real world pricing formula =⇒
S00
S0δH = E0 ΛT HT
ST0
Ŝt0
Λt = supermartingale
Ŝ00
1 = Λ0 ≥ E0 (ΛT )
=⇒
S00
E0 ΛT S 0 HT
S0δH ≤ T
E0 (ΛT )
similar for any numeraire (here S 0 savings account)
c Copyright Eckhard Platen 10 BA - Toronto C1 33
100
90
80
70
60
50
40
30
20
10
0
1930 1940 1950 1960 1970 1980 1990 2000
time
Figure 1.7: Discounted S&P500 total return index.
c Copyright Eckhard Platen 10 BA - Toronto C1 34
1.8
Radon-Nikodym derivative
Total RN-Mass
1.6
1.4
1.2
0.8
0.6
0.4
0.2
0
1930 1940 1950 1960 1970 1980 1990 2000
time
Figure 1.8: Radon-Nikodym derivative and total mass of putative risk neutral
measure.
c Copyright Eckhard Platen 10 BA - Toronto C1 35
• special case when savings account is fair:
dQ
=⇒ ΛT = dP
forms martingale; E0 (ΛT ) = 1;
equivalent risk neutral probability measure Q exists;
Bayes’ formula =⇒
risk neutral pricing formula
S00
S0δH = E0Q HT
ST0
Harrison & Kreps (1979), Ingersoll (1987),
Constantinides (1992), Duffie (2001), Cochrane (2001), . . .
• otherwise “risk neutral price” ≥ real world price
c Copyright Eckhard Platen 10 BA - Toronto C1 36
ln(index)
9 ln(savings account)
-1
1930 1940 1950 1960 1970 1980 1990 2000
time
Figure 1.9: ln(S&P500 accumulation index) and ln(savings account).
c Copyright Eckhard Platen 10 BA - Toronto C1 37
Extreme Maturity Bond
• short rate deterministic
• savings bond
St0
P ∗ (t, T ) =
ST0
• index S 1 as numeraire portfolio S δ∗
c Copyright Eckhard Platen 10 BA - Toronto C1 38
• fair zero coupon bond
Law of the Minimal Price
=⇒ real world pricing formula
!
1
P (t, T ) = Stδ∗ Et
STδ∗
• benchmarked fair zero coupon bond
P (t, T )
P̂ (t, T ) = martingale
Stδ∗
c Copyright Eckhard Platen 10 BA - Toronto C1 39
• discounted numeraire portfolio
Stδ∗
S̄tδ∗ =
St0
numeraire portfolio for continuous market:
q
dS̄tδ∗ = αt dt + S̄tδ∗ αt dWt
time transformed squared Bessel process
c Copyright Eckhard Platen 10 BA - Toronto C1 40
• model the drift of discounted index as
αt = α exp{η t}
=⇒
S̄tδ∗ time transformed squared Bessel process of dimension four
=⇒ Minimal Market Model
MMM, see Pl. & Heath (2006)
• net growth rate
η ≈ 0.054 with R2 of 0.89
c Copyright Eckhard Platen 10 BA - Toronto C1 41
5
ln(discounted index)
trendline
-1
-2
1930 1940 1950 1960 1970 1980 1990 2000
time
Figure 1.10: Logarithm of discounted S&P500 and trendline.
c Copyright Eckhard Platen 10 BA - Toronto C1 42
• normalized index
S̄tδ∗
Yt =
αt
p
dYt = (1 − η Yt ) dt + Yt dWt
√
• quadratic variation of Yt
p 1
d Yt = . . . + dWt
2
it h√ i
X p q 2 t
Ytℓ − Ytℓ−1 ≈ Y =
ℓ=1
t 4
• scaling parameter α ≈ 0.01468 with R2 of 0.995
c Copyright Eckhard Platen 10 BA - Toronto C1 43
30
[sqrt Y]
trendline
25
20
15
10
0
1930 1940 1950 1960 1970 1980 1990 2000
time
√
Figure 1.11: Quadratic Variation of Yt .
c Copyright Eckhard Platen 10 BA - Toronto C1 44
• fair zero coupon bond
( )!
∗ 2η S̄tδ∗
P (t, T ) = P (t, T ) 1 − exp −
α (exp{η T } − exp{η t})
• initial prices
P ∗ (0, T ) = 0.0335
P (0, T ) = 0.00077
P (0,T )
P ∗ (0,T )
< 0.023 =⇒ 2.3%
no strong arbitrage
c Copyright Eckhard Platen 10 BA - Toronto C1 45
0.16
benchmarked savings bond
benchmarked fair zero coupon bond
0.14
0.12
0.1
0.08
0.06
0.04
0.02
0
1930 1940 1950 1960 1970 1980 1990 2000
time
Figure 1.12: Benchmarked savings bond and benchmarked fair zero coupon
bond.
c Copyright Eckhard Platen 10 BA - Toronto C1 46
1
savings bond
fair zero coupon bond
0.9 savings account
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
1930 1940 1950 1960 1970 1980 1990 2000
time
Figure 1.13: Savings bond, fair zero coupon bond and savings account.
c Copyright Eckhard Platen 10 BA - Toronto C1 47
1
ln(savings bond)
ln(fair zero coupon bond)
0
-1
-2
-3
-4
-5
-6
-7
-8
1930 1940 1950 1960 1970 1980 1990 2000
time
Figure 1.14: Logarithms of fair zero coupon bond and savings account.
c Copyright Eckhard Platen 10 BA - Toronto C1 48
• savings account is outperformed systematically
by fair zero coupon bond
c Copyright Eckhard Platen 10 BA - Toronto C1 49
1
ln(fair zero coupon bond)
ln(savings account)
0
-1
-2
-3
-4
-5
-6
-7
-8
1930 1940 1950 1960 1970 1980 1990 2000
time
Figure 1.15: Logarithms of fair zero coupon bond and savings account.
c Copyright Eckhard Platen 10 BA - Toronto C1 50
• self-financing hedge portfolio hedge ratio
∂ P̄ (t, T )
δt∗ =
∂ S̄tδ∗
( )
∗ −2 η S̄tδ∗
= P (0, T ) exp
α (exp{η T } − exp{η t})
2η
×
α (exp{η T } − exp{η t})
c Copyright Eckhard Platen 10 BA - Toronto C1 51
1
ln(zero coupon bond)
ln(self financing hedge portfolio)
0
-1
-2
-3
-4
-5
-6
-7
-8
1930 1940 1950 1960 1970 1980 1990 2000
time
Figure 1.16: Logarithm of zero coupon bond and self-financing hedge port-
folio.
c Copyright Eckhard Platen 10 BA - Toronto C1 52
7e-005
6e-005
5e-005
4e-005
3e-005
2e-005
1e-005
-1e-005
-2e-005
1930 1940 1950 1960 1970 1980 1990 2000
time
Figure 1.17: Benchmarked profit and loss.
c Copyright Eckhard Platen 10 BA - Toronto C1 53
1
0.8
0.6
0.4
0.2
1930 1940 1950 1960 1970 1980 1990 2000
time
Figure 1.18: Fraction invested in the index.
c Copyright Eckhard Platen 10 BA - Toronto C1 54
2 Continuous Financial Markets
• d sources of continuous traded uncertainty
independent standard Wiener processes
W 1, W 2, . . . , W d, d ∈ {1, 2, . . .}
c Copyright Eckhard Platen 10 BA - Toronto C2 55
Primary Security Accounts
Stj - jth primary security account at time t,
j ∈ {0, 1, . . . , d}
cum-dividend share value or savings account value,
dividends or interest are accumulated, reinvested
• vector process of primary security accounts
S = {S t = (St0 , . . . , Std )⊤ , t ∈ [0, T ]}
c Copyright Eckhard Platen 10 BA - Toronto C2 56
• assume unique strong solution of SDE
d
!
X
dStj = Stj ajt dt + bj,k
t dWt
k
k=1
t ∈ [0, ∞), S0j > 0, j ∈ {1, 2, . . . , d}
• predictable, integrable appreciation rate processes aj
• predictable, square integrable volatility processes bj,k
c Copyright Eckhard Platen 10 BA - Toronto C2 57
• savings account
Z t
St0 = exp rs ds
0
predictable, integrable short rate process
r = {rt , t ∈ [0, ∞)}
c Copyright Eckhard Platen 10 BA - Toronto C2 58
• Market price of risk
The unique invariant of the market
θ t = (θt1 , θt2 , . . . , θtd )⊤
solution of relation
bt θ t = at − rt 1
where
at = (a1t , a2t , . . . , adt )⊤
1 = (1, 1, . . . , 1)⊤
c Copyright Eckhard Platen 10 BA - Toronto C2 59
Assumption 2.1
Volatility matrix bt = [bj,k d
t ]j,k=1 is invertible
=⇒
• market price of risk equals
θ t = b−1
t (at − rt 1)
c Copyright Eckhard Platen 10 BA - Toronto C2 60
• rewrite jth primary security account SDE
( d
)
X
dStj = Stj rt dt + bj,k k k
t (θt dt + dWt )
k=1
t ∈ [0, ∞), j ∈ {1, 2, . . . , d}
c Copyright Eckhard Platen 10 BA - Toronto C2 61
• strategy
δ = {δ t = (δt0 , . . . , δtd )⊤ , t ∈ [0, T ]}
predictable and S-integrable,
δtj number of units of jth primary security account
• portfolio
d
X
Stδ = δtj Stj
j=0
c Copyright Eckhard Platen 10 BA - Toronto C2 62
• S δ self-financing ⇐⇒
d
X
dStδ = δtj dStj
j=0
changes in the portfolio value only due to gains from trading,
self-financing in each denomination
c Copyright Eckhard Platen 10 BA - Toronto C2 63
• jth fraction
j Stj
πδ,t = δtj
Stδ
j ∈ {0, 1, . . . , d}
d
X j
πδ,t =1
j=0
c Copyright Eckhard Platen 10 BA - Toronto C2 64
• portfolio SDE
d
!
X
dStδ Stδ bkδ,t θtk dWtk
= rt dt + dt +
k=1
with kth portfolio volatility
d
X j
bkδ,t = πδ,t bj,k
t
j=1
c Copyright Eckhard Platen 10 BA - Toronto C2 65
• logarithm of strictly positive portfolio
d
X
d ln(Stδ ) = gtδ dt + bkδ,t dWtk
k=1
with growth rate
d
1
X
gtδ = rt + bkδ,t θtk − bkδ,t
k=1
2
c Copyright Eckhard Platen 10 BA - Toronto C2 66
Numeraire Portfolio S δ∗ = Growth Optimal Portfolio
=⇒ maximize the growth rate
d X
d d
!
X j 1X
gtδ = rt + πδ,t bj,k
t θtk − ℓ
πδ,t bℓ,k
t
k=1 j=1
2 ℓ=1
=⇒ for each j ∈ {1, 2, . . . , d}
d d
!
X X
0= bj,k
t θtk − πδℓ∗ ,t bℓ,k
t
k=1 ℓ=1
=⇒
θ⊤ ⊤
t = π δ∗ ,t bt
c Copyright Eckhard Platen 10 BA - Toronto C2 67
Since bt invertible =⇒
π δ∗ ,t = (πδ1∗ ,t , . . . , πδd∗ ,t )⊤
= (b−1 ⊤
t ) θt
=⇒ NP
" d
# d
!
X X
dStδ∗ = Stδ∗ rt + (θtk )2 dt + θtk dWtk
k=1 k=1
finite NP exists
c Copyright Eckhard Platen 10 BA - Toronto C2 68
Benchmarked Portfolios
Stδ
Ŝtδ =
Stδ∗
d X
X d
j
dStδ Stδ bj,k θtk dWtk
= rt dt + πδ,t t dt +
k=1 j=0
d
!
X
dStδ∗ Stδ∗ θtk θtk dWtk
= rt dt + dt +
k=1
d X
X d
j
dŜtδ = −Ŝtδ πδ,t sj,k k
t dWt
k=1 j=0
driftless =⇒ local martingale =⇒ supermartingale
when nonnegative
c Copyright Eckhard Platen 10 BA - Toronto C2 69
Theorem 2.2 Any nonnegative benchmarked portfolio is an
(A, P )-supermartingale (no upward trend).
c Copyright Eckhard Platen 10 BA - Toronto C2 70
=⇒ NP is numeraire portfolio
use NP as benchmark for investment and as numeraire for pricing
• markets have thousands of primary security accounts
c Copyright Eckhard Platen 10 BA - Toronto C2 71
Example of a Multi-Asset BS Model
• jth primary security account
" d
#
1 σ
X
dStj = Stj r+σ 2
1+ √ dt + √ dWtk + σ dWtj
d d k=1
t ∈ [0, T ], j ∈ {1, 2, . . . , d}, σ > 0
general market risk, specific market risk
c Copyright Eckhard Platen 10 BA - Toronto C2 72
• NP fractions
√ √ −1
πδj∗ ,t = d 1+ d
j ∈ {1, 2, . . . , d}
√ −1
πδ0∗ ,t = 1+ d
• NP SDE
d
!
2 σ X
dStδ∗ Stδ∗ dWtk
= r+σ dt + √
d k=1
• simulate over T = 32 years d = 50 risky primary security accounts
σ = 0.15, r = 0.05
c Copyright Eckhard Platen 10 BA - Toronto C2 73
25
20
15
10
0
0 5 10 15 20 25 30 35
time
Figure 2.1: Simulated primary security accounts.
c Copyright Eckhard Platen 10 BA - Toronto C2 74
10
GOP
B
9
0
0 5 10 15 20 25 30 35
time
Figure 2.2: Simulated NP and savings account.
c Copyright Eckhard Platen 10 BA - Toronto C2 75
Diversified Portfolios
Fundamental phenomenon of diversification leads naturally to NP:
diversified =⇒ fractions “small”
Definition 2.3 A strictly positive portfolio S δ is a diversified portfolio if
j K2
πδ,t ≤ 1
d 2 +K1
a.s. for all j ∈ {0, 1, . . . , d} and t ∈ [0, T ] for K1 , K2 ∈ (0, ∞) and
K3 ∈ {1, 2, . . .}, independent of d ∈ {K3 , K3 + 1, . . .}.
c Copyright Eckhard Platen 10 BA - Toronto C2 76
• jth benchmarked primary security account
Stj
Ŝtj =
Stδ∗
d
X
dŜtj = −Ŝtj sj,k k
t dWt
k=1
• (j, k)th specific volatility
sj,k
t = θt
k
− bj,k
t
measures specific market risk of S j with respect to W k
• θtk measures general market risk with respect to W k
c Copyright Eckhard Platen 10 BA - Toronto C2 77
• kth total specific absolute volatility
d
X
ŝkt = |sj,k
t |
j=0
regular =⇒ variance of specific returns “bounded”
Definition 2.4 A market is called regular if
2
k
E ŝt ≤ K5 .
c Copyright Eckhard Platen 10 BA - Toronto C2 78
Tracking Rate
Variance of benchmarked portfolio returns
2
d
X d
X j
Rδd (t) = πδ,t sj,k
t
k=1 j=0
Rδd (t) - quantifies distance between S δ and S δ∗
benchmarked NP constant =⇒
Rδd∗ (t) = 0
c Copyright Eckhard Platen 10 BA - Toronto C2 79
Approximate NP
tracking rate “small”
Definition 2.5 A strictly positive portfolio S δ is an approximate NP if
for all t ∈ [0, T ] and ε > 0
Rδd (t)
lim P > ε = 0.
d→∞
c Copyright Eckhard Platen 10 BA - Toronto C2 80
Diversification Theorem
Theorem 2.6 In a regular market
any diversified portfolio is an approximate NP.
• model independent
• similarity to CLT
c Copyright Eckhard Platen 10 BA - Toronto C2 81
Equal Weighted Index (EWI)
S δEWI holds equal fractions
1
πδj ,t =
EWI d+1
j ∈ {0, 1, . . . , d}
• tracking rate for multi-asset BS example:
d 2
X 1
Rδd (t) = θtk − bk,k
t
EWI
k=1
d+1
2
1 σ σ2
= d √ −σ ≤ →0
d+1 d (d + 1)
c Copyright Eckhard Platen 10 BA - Toronto C2 82
10
GOP
EWI
9
0
0 5 10 15 20 25 30 35
time
Figure 2.3: Simulated NP and EWI.
c Copyright Eckhard Platen 10 BA - Toronto C2 83
10
GOP
index
9
0
0 5 10 15 20 25 30 35
time
Figure 2.4: Simulated diversified accumulation index and NP.
c Copyright Eckhard Platen 10 BA - Toronto C2 84
120,000
100,000
80,000
60,000
40,000
20,000
0
1/1/73 28/9/76 25/6/80 22/3/84 18/12/87 15/9/91 12/6/95 9/3/99 4/12/02 31/8/06
Figure 2.5: World industry sector stock market indices as primary security
accounts.
c Copyright Eckhard Platen 10 BA - Toronto C2 85
$US
WSI
8,000 EWI
7,000
6,000
5,000
DWI
4,000
MCI
3,000
2,000
1,000
0
1/1/73 28/9/76 25/6/80 22/3/84 18/12/87 15/9/91 12/6/95 9/3/99 4/12/02 31/8/06
Figure 2.6: Constructed MCI, DWI, EWI and WSI. In the long term the WSI
and EWI outperform all other indices, Le & Platen (2006).
c Copyright Eckhard Platen 10 BA - Toronto C2 86
450
400
350
300
250
200
150
100
50
0
0 5 9 14 18 23 27 32
Figure 2.7: Primary security accounts under the MMM.
c Copyright Eckhard Platen 10 BA - Toronto C2 87
60
50
40
30
20
10
0
0 5 9 14 18 23 27 32
Figure 2.8: Benchmarked primary security accounts.
c Copyright Eckhard Platen 10 BA - Toronto C2 88
100
EWI
90 GOP
80
70
60
50
40
30
20
10
0
0 5 9 14 18 23 27 32
Figure 2.9: NP and EWI.
c Copyright Eckhard Platen 10 BA - Toronto C2 89
100
Market index
90
GOP
80
70
60
50
40
30
20
10
0
0 5 9 14 18 23 27 32
Figure 2.10: NP and market index.
c Copyright Eckhard Platen 10 BA - Toronto C2 90
3 Portfolio Optimization
• NP best long term investment
Kelly (1956), Latané (1959), Breiman (1961),
Hakansson (1971b), Thorp (1972)
• optimal portfolio separated into two funds
Tobin (1958), Sharpe (1964)
c Copyright Eckhard Platen 10 BA - Toronto C3 91
• mean-variance efficient portfolio
Markowitz (1959)
• intertemporal capital asset pricing model
Merton (1973a)
c Copyright Eckhard Platen 10 BA - Toronto C3 92
Expected Utility Maximization
• utility function U (·), U ′ (·) > 0 U ′′ (·) < 0
• examples
power utility
1
U (x) = xγ
γ
for γ 6= 0 and γ < 1
log-utility
U (x) = ln(x)
c Copyright Eckhard Platen 10 BA - Toronto C3 93
U
x
5 10 15 20
-5
-10
-15
Figure 3.1: Examples for power utility (upper graph) and log-utility (lower
graph).
c Copyright Eckhard Platen 10 BA - Toronto C3 94
• assume Ŝ 0 is scalar diffusion
• maximize expected utility
δ̃
S̄Tδ
v = max E0 U
+
S̄ δ ∈V̄S
0
V̄S+0 strictly positive, discounted, fair portfolios
S̄0δ = S0 > 0
c Copyright Eckhard Platen 10 BA - Toronto C3 95
Theorem 3.1
Benchmarked, expected utility maximizing portfolio:
Ŝtδ̃ = û(t, Ŝt0 ) = Et U ′−1 λ ŜT0 ŜT0 ,
λ s.t. S0 = Ŝ0δ̃ S0δ∗ = û(0, Ŝ0 ) S0δ∗
two fund separation with risk aversion coefficient
1
Jtδ̃ =
Ŝt0 ∂ û(t,Ŝt0 )
1− û(t,Ŝt0 ) ∂ Ŝt0
1
=⇒ invest of wealth in NP and remainder in savings account
Jtδ̃
c Copyright Eckhard Platen 10 BA - Toronto C3 96
• log-utility function U (x) = ln(x)
′ 1
U (x) =
x
′−1 1
U (y) =
y
′′ 1
U (x) = −
x2
utility concave
!
′−1
1 1
û(t, Ŝt0 ) = Et U λ ŜT0 ŜT0 = Et ŜT
0
=
λ ŜT0 λ
Lagrange multiplier
c Copyright Eckhard Platen 10 BA - Toronto C3 97
S0δ∗
λ=
S0
risk aversion coefficient
Jtδ̃ = 1
• expected log-utility
δ̃ δ∗
v = E0 ln S̄T
T
1
Z
δ∗ 2
= ln(λ) + ln(S0 ) + E0 θ(s, S̄s ) ds
2 0
if local martingale part forms a martingale
c Copyright Eckhard Platen 10 BA - Toronto C3 98
• power utility
1
U (x) = xγ
γ
for γ < 1, γ 6= 0
U ′ (x) = xγ−1
1
U ′−1 (y) = y γ−1
U ′′ (x) = (γ − 1) xγ−2
concave
1
! γ−1
γ
λ 1
1−γ
1
û(t, Ŝt0 ) = Et = λ γ−1 Et S̄Tδ∗
S̄Tδ∗ S̄T δ∗
c Copyright Eckhard Platen 10 BA - Toronto C3 99
S̄ δ∗ geometric Brownian motion
1 γ
û(t, Ŝt0 ) S̄tδ∗
1−γ
= λ γ−1
2
γ θ
× Et exp (T − t) + θ (WT − Wt )
1−γ 2
2
θ γ
1 γ
δ∗ 1−γ
= λ γ−1 S̄t exp (T − t)
2 (1 − γ)2
c Copyright Eckhard Platen 10 BA - Toronto C3 100
Lagrange multiplier
θ2 γ
λ = S0γ−1 S̄0δ∗ exp T
2 1−γ
Ŝt0 = (S̄tδ∗ )−1
∂ û(t, Ŝt0 ) û(t, Ŝt0 ) γ
=
∂ Ŝ 0 Ŝt0 γ−1
=⇒ risk aversion coefficient
Jtδ̃ = 1 − γ
expected utility
1 θ2 γ
γ
v δ̃ = E0 S̄Tδ̃ = exp − T (S0 )γ
γ 2 1−γ
c Copyright Eckhard Platen 10 BA - Toronto C3 101
Utility Indifference Pricing
H
discounted payoff H̄ = 0
ST
(not perfectly hedgable)
!!
δ̃ S̄Tδ̃
vε,V = E0 U (S0 − ε V ) + ε H̄
S0
• assume Ŝ 0 is scalar diffusion
Definition 3.2 utility indifference price V s.t.
δ̃ δ̃
vε,V − v0,V
lim =0
ε→0 ε
Davis (1997)
c Copyright Eckhard Platen 10 BA - Toronto C3 102
• Taylor expansion
!!
δ̃
′
V S̄Tδ̃
vε,V ≈ E0 U S̄Tδ̃ + εU S̄Tδ̃ H̄ −
S0
!
S̄ δ̃
δ̃
= v0,V + ε E0 U ′ S̄Tδ̃ H̄ − V ε E0 U ′ S̄T δ̃ T
S0
′−1
S̄Tδ̃ = ŜTδ̃ /ŜT0 =U λ ŜT0 =⇒
!!
δ̃
S̄T
δ̃
vε,V δ̃
− v0,V = ε E0 λ ŜT0 H̄ − V E0 λ ŜT0
S0
c Copyright Eckhard Platen 10 BA - Toronto C3 103
=⇒
• utility indifference price
E0 λδ∗ H̄ E0 H
δ∗
S̄T ST
V = = 1 S0
δ̃
λ S̄T
E0 δ∗ S
S0 S0δ∗
S̄T 0
independent of U and of given model =⇒
• real world pricing formula
!
H
V = S0δ∗ E0
STδ∗
c Copyright Eckhard Platen 10 BA - Toronto C3 104
Hedging
• benchmarked derivative price
!
H
ÛH (t) = Et
STδ∗
• real world martingale representation
d Z T
H X
= ÛH (t) + xkH (s) dWsk + MH (t)
STδ∗ k=1 t
MH -orthogonal martingale
k
E MH , W t = 0
Föllmer-Schweizer decomposition, Föllmer & Schweizer (1991)
least square projection
c Copyright Eckhard Platen 10 BA - Toronto C3 105
Theorem 3.3 Derivative price
!
H
StδH = Stδ∗ Et
STδ∗
fractions
⊤
π δH (t) = bδH (t) ⊤
b−1
t
with portfolio volatilities
d
1 X j xkH (t)
bkδH (t) = δH (t) Ŝtj bj,k
t = + θtk .
ÛH (t) j=0
ÛH (t)
c Copyright Eckhard Platen 10 BA - Toronto C3 106
4 Modeling Stochastic Volatility
• consider well diversified accumulation index ≈ NP
dStδ∗ = Stδ∗ rt dt + |θt | |θt | dt + dŴt
1
δ∗
|θt |2 dt + |θt | dŴt
d ln St = rt dt +
2
• volatility
r
d
|θt | = [ln (S δ∗ )]t
dt
c Copyright Eckhard Platen 10 BA - Toronto C4 107
0.45
0.4
0.35
0.3
0.25
0.2
0.15
0.1
0.05
1982 1992 2002
time
Figure 4.1: Estimated volatility of WSI from 1973–2004.
c Copyright Eckhard Platen 10 BA - Toronto C4 108
Log-Returns of a Stock Index
1.2
1.0
0.8
0.6
0.4
0.2
0.0
-10 -5 0 5 10
Figure 4.2: Histogram for S&P500 hourly deseasonalized log-returns.
fitted Gaussian density, much fatter tails
c Copyright Eckhard Platen 10 BA - Toronto C4 109
Implied Volatilities
• European call for Black-Scholes model
cτ,K (0, S, σ)
maturity τ ∈ [0, T ]
strike price K > 0
underlying index S > 0
volatility σ
c Copyright Eckhard Platen 10 BA - Toronto C4 110
• observed European call option price
Vc,τ,K (0, S0 )
• implied volatility
call
σBS (0, S0 , τ, K)
obtained such that
call
Vc,τ,K (0, S0 ) = cτ,K 0, S0 , σBS (0, S0 , τ, K)
by some root finding method
c Copyright Eckhard Platen 10 BA - Toronto C4 111
60
40 1.4
20 1.2
1
moneyness
100 0.8
time 200
0.6
300
Figure 4.3: Implied volatilities for S&P500 three month options.
c Copyright Eckhard Platen 10 BA - Toronto C4 112
implied volatility surface
T 1.6
1.4
1.2
1
0.8
0.6
0.4
0.2
0.28
0.26
0.24
0.22
0.5
0.75
1
1.25
K S
1.5
Figure 4.4: Average S&P500 implied volatility surface.
c Copyright Eckhard Platen 10 BA - Toronto C4 113
A Modified Constant Elasticity of Variance Model
• CEV Model
Cox (1975), Cox & Ross (1976)
Andersen & Andreasen (2000)
Lewis (2000)
Lo, Yuen & Hui (2000)
Brigo & Mercurio (2005)
Heath & Pl. (2005b)
c Copyright Eckhard Platen 10 BA - Toronto C4 114
• classical risk neutral CEV
Beckers (1980)
Schroder (1989)
Delbaen & Shirakawa (2002)
certain problems in risk neutral pricing of derivatives
c Copyright Eckhard Platen 10 BA - Toronto C4 115
Modified CEV Model
Heath & Pl. (2005b)
• savings account
dSt0 = r St0 dt
total market price for risk process
{θt ≥ 0, t ∈ [0, T ]}
• drifted Wiener process
dŴθ (t) = θt dt + dŴt
c Copyright Eckhard Platen 10 BA - Toronto C4 116
• NP
dStδ∗ = Stδ∗ (rt dt + θt dŴθ (t))
• NP volatility
θt = (Stδ∗ )a−1 ψ
exponent a ∈ (−∞, ∞)
scaling parameter ψ > 0
stochastic for a 6= 1
c Copyright Eckhard Platen 10 BA - Toronto C4 117
• hypothetical risk neutral CEV dynamics
dStδ∗ = Stδ∗ rt dt + (Stδ∗ )a ψ dŴθ (t)
Ŵθ - Wiener process under a hypothetical risk neutral measure Pθ
• If a = 1 =⇒ BS model
=⇒ equivalent risk neutral probability measure Pθ
with market price of risk θt = ψ
c Copyright Eckhard Platen 10 BA - Toronto C4 118
• real world CEV dynamics
δ∗ 2a−1
2 δ∗ a
dStδ∗ Stδ∗
= rt + St ψ dt + St ψ dŴt
modified CEV model remains for a < 1 strictly positive
this is not the case for its “risk neutral” dynamics
c Copyright Eckhard Platen 10 BA - Toronto C4 119
• squared Bessel process
δ∗ 2 (1−a)
Xt = St
for a 6= 1 with SDE
2
dXt = 2(1 − a) rt Xt + ψ (1 − a) (3 − 2a) dt
p
+2 ψ (1 − a) Xt dŴt
and dimension
3 − 2a
δ=
1−a
c Copyright Eckhard Platen 10 BA - Toronto C4 120
• Radon-Nikodym derivative process
Λθ = {Λθ (t), t ∈ [0, T ]}
(state price density)
• hypothetical risk neutral measure Pθ
dPθ
= Λθ (T )
dP
S0δ∗
q
X0
Λθ (t) = St0 = St0
Stδ∗ Xt
1
q=
2 (1 − a)
c Copyright Eckhard Platen 10 BA - Toronto C4 121
• squared Bessel process
2
dXt = 2 (1 − a) rt Xt + ψ (1 − a)(1 − 2 a) dt
p
+ 2 ψ (1 − a) Xt dŴθ (t)
dimension
1 − 2a
δθ =
1−a
c Copyright Eckhard Platen 10 BA - Toronto C4 122
• dimension
risk neutral
1 − 2a
δθ =
1−a
real world
3 − 2a
δ=
1−a
• if a < 1 =⇒ δ > 2 and δθ < 2
if a > 1 =⇒ δ < 2 and δθ > 2
c Copyright Eckhard Platen 10 BA - Toronto C4 123
8
Real World
6
Risk Neutral
4
-2
-4
-1 0 1 2 3
Figure 4.5: Dimensions δ and δθ as a function of the exponent a.
c Copyright Eckhard Platen 10 BA - Toronto C4 124
• hypothetical risk neutral probability
dPθ (ω)
Z Z
Pθ (A) = dPθ (ω) = dP (w)
A A dP (ω)
Z
= Λθ (T ) dP (ω)
A
=⇒
Z
Pθ (Ω) = Λθ (T ) dP (ω) = E0 (Λθ (T ))
Ω
c Copyright Eckhard Platen 10 BA - Toronto C4 125
• for a < 1 =⇒ δθ < 2
=⇒ Λθ is strict supermartingale =⇒
Pθ (Ω) < Λθ (0) = 1
=⇒ Pθ is not a probability measure
c Copyright Eckhard Platen 10 BA - Toronto C4 126
• for a 6= 1 P and Pθ are not equivalent
=⇒ modified CEV model does not have equivalent risk neutral
probability measure
• consider only a < 1 since for a > 1 NP explodes
c Copyright Eckhard Platen 10 BA - Toronto C4 127
Real World Pricing
• contingent claim
Hτ
Hτ = Hτ (Sτδ∗ ), E <∞
Sτδ∗
• fair price
uHτ (t, Stδ∗ ) = Stδ∗ ûHτ (t, Stδ∗ )
benchmarked price
!
Hτ (Sτδ∗ )
ûHτ (t, Stδ∗ ) = Et
Sτδ∗
c Copyright Eckhard Platen 10 BA - Toronto C4 128
Martingale Representation
Hτ
= ûHτ (τ, Sτδ∗ )
Sτδ∗
τ
∂ ûHτ (s, Ssδ∗ )
Z
δ∗ a
ûHτ (t, Stδ∗ )
= + Ss ψ dŴs
t ∂S δ∗
c Copyright Eckhard Platen 10 BA - Toronto C4 129
Stochastic Volatility Models
• volatility function of underlying diffusion
=⇒ LV model
• volatility as separate, possibly correlated process
Hull & White (1987, 1988)
Johnson & Shanno (1987)
Scott (1987)
Wiggins (1987)
Chesney & Scott (1989)
Melino & Turnbull (1990)
Stein & Stein (1991)
Hofmann, Pl. & Schweizer (1992)
Heston (1993), . . .
c Copyright Eckhard Platen 10 BA - Toronto C4 130
Empirical Evidence on Log-Returns
• Markowitz & Usmen (1996)
daily S&P500 → Student t(4.5)
• Hurst & Platen (1997)
daily stock market indices → Student t(4)
• Fergusson & Pl. (2006), Pl. & Rendek (2008)
daily WSI, EWI104s with high accuracy → Student t(4)
• Breymann, Dias & Embrechts (2003)
copula of joint distribution of log-returns of exchange rates →
Student t copula with roughly four degrees of freedom
c Copyright Eckhard Platen 10 BA - Toronto C4 131
−1
10
−2
10
−3
10
−4
10
−10 −5 0 5 10
Figure 4.6: Log-histogram of the EWI104s log-returns and Student t density
with four degrees of freedom.
c Copyright Eckhard Platen 10 BA - Toronto C4 132
Estimated LLF
× 105
−2.84
−2.86
−2.88
−2.9
2
−2.92
1.5
−2.94
−5
−4 1
−3
−2.15
−1
0 0.5
Estimated λ 1
2
3
4 0 α
5
λ
Figure 4.7: Log-likelihood function based on the EWI104s.
c Copyright Eckhard Platen 10 BA - Toronto C4 133
Country Student-t NIG Hyperbolic VG ν
Australia 0.000000 76.770817 150.202282 181.632971 4.281222
Austria 0.000000 39.289103 77.505683 102.979330 4.725907
Belgium 0.000000 31.581622 60.867570 83.648470 4.989912
Brazil 2.617693 5.687078 63.800349 60.078395 2.713036
Canada 0.000000 47.506215 79.917741 104.297607 5.316154
Denmark 0.000000 41.509921 87.199686 114.853658 4.512101
Finland 0.000000 28.852844 68.677271 88.553080 4.305638
France 0.000000 26.303544 57.639325 80.567283 4.722787
Germany 0.000000 27.290205 52.667918 71.120798 5.005856
Greece 0.000000 60.432172 104.789463 125.601499 4.674626
Hong.Kong 0.000000 42.066531 100.834255 122.965326 3.930473
India 0.000000 74.773701 163.594078 198.002956 3.998713
Ireland 0.000000 77.727856 136.505582 170.013644 4.761519
Italy 0.000000 25.196598 55.185625 75.481897 4.668983
Japan 0.000000 37.630363 77.163656 102.967380 4.649745
Korea.S. 0.000000 120.904983 304.829431 329.854620 3.289204
c Copyright Eckhard Platen 10 BA - Toronto C4 134
Malaysia 0.000000 79.714054 186.013963 221.061290 3.785195
Netherlands 0.000000 26.832761 51.625813 71.541627 5.084056
Norway 0.000000 42.243851 89.012090 115.059003 4.472349
Portugal 0.000000 61.177624 137.681039 165.689683 3.984860
Singapore 0.000000 36.379685 77.600590 98.124375 4.251472
Spain 0.000000 56.694545 109.533768 138.259224 4.517153
Sweden 0.000000 77.618384 143.420049 178.983373 4.546640
Taiwan 0.000000 41.162560 96.283628 115.186585 3.914719
Thailand 0.000000 78.250621 254.590254 267.508143 3.032038
UK 0.000000 26.693076 55.937248 80.678494 4.952843
USA 0.000000 40.678242 79.617362 100.901197 4.636661
Table 1: Ln test statistic of the EWI104s for different currency denominations
c Copyright Eckhard Platen 10 BA - Toronto C4 135
Index Log-returns
• Student t distributed with about 4 degrees of freedom
• clustering
Find corresponding stationary volatility processes !
Kessler (1997), Prakasa Rao (1999)
c Copyright Eckhard Platen 10 BA - Toronto C4 136
• normal mixing
log-return conditionally Gaussian distributed
with stochastic variance process
for small time step size h > 0
δ∗
!
St+h
∆ ln(Stδ∗ ) = ln
Stδ∗
θt2
∼ N h, θt2 h
2
when θt2 has inverse gamma stationary density
=⇒ estimated log-return appears to be Student t distributed
c Copyright Eckhard Platen 10 BA - Toronto C4 137
A Class of Continuous Stochastic Volatility Models
• discounted NP
Stδ∗
S̄tδ∗ =
St0
dS̄tδ∗ = S̄tδ∗ (θt2 dt + θt dŴt )
c Copyright Eckhard Platen 10 BA - Toronto C4 138
Particular Stochastic Volatility Models
• Hull and White’87 model
Hull & White (1987)
dθt2 = µ θt2 dt + ξ θt2 dW̄t
squared volatility does not have a stationary dynamics
c Copyright Eckhard Platen 10 BA - Toronto C4 139
• Heston model
Hull & White (1988), Heston (1993)
mean reverting dynamics
k, θ̄ and γ are positive constants
=⇒
dθt2 = k (θ̄ 2 − θt2 ) dt + γ |θt | dW̄t
κ 1
θ02 > 0, γ2
≥ 2
=⇒ has as stationary density gamma density
c Copyright Eckhard Platen 10 BA - Toronto C4 140
• Scott model
Scott (1987)
Stein & Stein (1991)
Ornstein-Uhlenbeck process
k, θ̄, γ positive constants
dθt = k (θ̄ − θt ) dt + γ dW̄t
θ0 ∈ ℜ
=⇒ θt2 has as stationary density a gamma density
c Copyright Eckhard Platen 10 BA - Toronto C4 141
• Wiggins model
Wiggins (1987)
Chesney & Scott (1989)
Melino & Turnbull (1990)
d ln(θt ) = k ln(θ̄) − ln(θt ) dt + γ dW̄t
θ0 > 0
θt2 has as stationary density a log-normal density
c Copyright Eckhard Platen 10 BA - Toronto C4 142
• inverted gamma distribution
1 1 − 12 ν−1 (
1 2
)
( 2 ν) 2 ν y νε
2
p̄ (y) =
θ2 exp −
ε2 Γ( 12 ν) ε2 y
y > 0, degrees of freedom ν > 0, scaling parameter ε
c Copyright Eckhard Platen 10 BA - Toronto C4 143
• SDE for squared volatility
4(ξ−1)
dθt2 = k θt (θ̄ 2 − θt2 ) dt + γ θt2 ξ dW̄t
degree of freedom
2 (2 ξ − 1)
ν= ε2
1− θ̄ 2
=⇒ stationary density inverted gamma
c Copyright Eckhard Platen 10 BA - Toronto C4 144
3
• 2
volatility model
3
ξ= 2
Pl. (1997), Lewis (2000), Pl. (2001)
dθt2 = k θt2 (θ̄ 2 − θt2 ) dt + γ θt3 dW̄t
c Copyright Eckhard Platen 10 BA - Toronto C4 145
• ARCH diffusion
ξ=1
Engle (1982), Nelson (1990) and Frey (1997)
dθt2 = k (θ̄ 2 − θt2 ) dt + γ θt2 dW̄t
continuous time limit of innovation process
GARCH(1,1) and NGARCH(1,1)
c Copyright Eckhard Platen 10 BA - Toronto C4 146
Implied Volatilities for the ARCH Diffusion Model
Hurst (1997), Lewis (2000), Heath, Hurst & Pl. (2001)
GARCH(1,1) of Bollerslev (1986) when Ŵ and W̄ are independent
continuous time limit of NGARCH(1, 1) model of Engle & Ng (1993)
two-factor model
c Copyright Eckhard Platen 10 BA - Toronto C4 147
0.22
1
0.21
0.8
0.2
80 0.6
T
90
0.4
100
K 110
0.2
120
Figure 4.8: Implied volatility surface for zero correlation.
c Copyright Eckhard Platen 10 BA - Toronto C4 148
0.22
0.2 0.4
0.18 0.2
80 0
rho
90
100 -0.2
K 110
-0.4
120
Figure 4.9: Effect of changing correlation on implied volatilities.
c Copyright Eckhard Platen 10 BA - Toronto C4 149
0.205
0.2025 20
0.2
0.1975 15
0.195
80 10 k
90
100
5
K 110
120
Figure 4.10: Effect of changing speed of adjustment on implied volatilities.
c Copyright Eckhard Platen 10 BA - Toronto C4 150
0.21 2
0.2
1.5
0.19
80 1 gamma
90
100
0.5
K 110
120
Figure 4.11: Effect of changing volatility of the squared volatility on implied
volatilities.
c Copyright Eckhard Platen 10 BA - Toronto C4 151
0.05 0.0
Option Price Difference
-0.05
GLMR
Heston
Black-Scholes
-0.10 -0.15
-0.3 -0.2 -0.1 0.0 0.1 0.2 0.3
Moneyness
Figure 4.12: Option price differences between the ARCH diffusion (GLMR),
Heston and Black-Scholes models, with time to maturity of three months.
c Copyright Eckhard Platen 10 BA - Toronto C4 152
0.20
Implied Volatility
0.16 0.18
GLMR
Heston
Black-Scholes
0.14
-0.3 -0.2 -0.1 0.0 0.1 0.2 0.3
Moneyness
Figure 4.13: Implied volatilities of the ARCH diffusion (GLMR), Heston
and Black-Scholes models, with time to maturity of three months.
c Copyright Eckhard Platen 10 BA - Toronto C4 153
5 Minimal Market Model
• search for parsimonious one-factor model
• Diversification Theorem =⇒
well diversified stock market index approximates NP
• discounted NP
dS̄tδ∗ = S̄tδ∗ |θt | (|θt | dt + dWt ),
where
d
1 X
dWt = θtk dWtk
|θt | k=1
c Copyright Eckhard Platen 10 BA - Toronto C5 154
100
90
80
70
60
50
40
30
20
10
0
1930 1940 1950 1960 1970 1980 1990 2000
time
Figure 5.1: Discounted S&P500 total return index.
c Copyright Eckhard Platen 10 BA - Toronto C5 155
5
-1
-2
1930 1940 1950 1960 1970 1980 1990 2000
time
Figure 5.2: Logarithm of discounted S&P500.
c Copyright Eckhard Platen 10 BA - Toronto C5 156
• logarithm of discounted NP
• is not a transformed Wiener process
• is not a process with independent increments
• has some feedback effect
c Copyright Eckhard Platen 10 BA - Toronto C5 157
Drift Parametrization
• discounted NP drift
αt = S̄tδ∗ |θt |2
strictly positive, predictable
=⇒
• volatility
s
αt
|θt | =
S̄tδ∗
leverage effect creates natural feedback
under independent drift
c Copyright Eckhard Platen 10 BA - Toronto C5 158
• discounted NP
q
dS̄tδ∗ = αt dt + S̄tδ∗ αt dWt
drift has economic meaning,
fundamental value
• transformed time
t
1
Z
ϕt = αs ds
4 0
c Copyright Eckhard Platen 10 BA - Toronto C5 159
• squared Bessel process of dimension four
Xϕt = S̄tδ∗
√
dW (ϕt ) = αt dWt
p
dXϕ = 4 dϕ + 2 Xϕ dW (ϕ)
Revuz & Yor (1999)
economically founded dynamics in ϕ-time
still no specific dynamics in t-time
c Copyright Eckhard Platen 10 BA - Toronto C5 160
Time Transformed Bessel Process
3 αt 1√
q
d S̄tδ∗ = q dt + αt dWt
8 S̄tδ∗ 2
• quadratic variation
Z t
hp i 1
S̄ δ∗ = αs ds
t 4 0
• transformed time hp i
ϕt = S̄ δ∗
t
observable
c Copyright Eckhard Platen 10 BA - Toronto C5 161
10
0
1930 1940 1950 1960 1970 1980 1990 2000
time
Figure 5.3: Empirical quadratic variation of the square root of the discounted
NP.
c Copyright Eckhard Platen 10 BA - Toronto C5 162
Stylized Minimal Market Model
Pl. (2001, 2002, 2006)
• assume discounted NP drift as
αt = α exp {η t}
• initial parameter α>0
• net growth rate η
c Copyright Eckhard Platen 10 BA - Toronto C5 163
• transformed time
t
α
Z
ϕt = exp {η z} dz
4 0
α
= (exp {η t} − 1)
4η
c Copyright Eckhard Platen 10 BA - Toronto C5 164
10
[sqrt(discounted index)]
theoretical variation
9
0
1930 1940 1950 1960 1970 1980 1990 2000
time
Figure 5.4: Fitted and observed transformed time.
c Copyright Eckhard Platen 10 BA - Toronto C5 165
• normalized NP
S̄tδ∗
Yt =
αt
p
dYt = (1 − η Yt ) dt + Yt dWt
square root process of dimension four
=⇒ parsimonious model, long term viability
c Copyright Eckhard Platen 10 BA - Toronto C5 166
=⇒
• discounted NP
S̄tδ∗ = Yt αt
• NP
Stδ∗ = St0 S̄tδ∗ = St0 Yt αt
• scaling parameter α0 = 0.01468
• net growth rate η = 0.054
1
• reference level η
= 18.5
• speed of adjustment η
ln(2)
• half life time of major displacement η
≈ 13 years
c Copyright Eckhard Platen 10 BA - Toronto C5 167
90
80
70
60
50
40
30
20
10
0
1930 1940 1950 1960 1970 1980 1990 2000
time
Figure 5.5: Normalized NP.
c Copyright Eckhard Platen 10 BA - Toronto C5 168
• logarithm of discounted NP
ln(S̄tδ∗ ) = ln(Yt ) + ln(α0 ) + η t
• no need for extra volatility process in MMM
• realistic long term dynamics
c Copyright Eckhard Platen 10 BA - Toronto C5 169
Volatility of NP under the stylized MMM
1
|θt | = √
Yt
• squared volatility
1
2 2 2
32
d|θt | = d = |θt | η dt − |θt | dWt
Yt
3/2 volatility model
Platen (1997), Lewis (2000)
c Copyright Eckhard Platen 10 BA - Toronto C5 170
0.35
0.3
0.25
0.2
0.15
0.1
1930 1940 1950 1960 1970 1980 1990 2000
time
Figure 5.6: Volatility of the NP under the stylized MMM.
c Copyright Eckhard Platen 10 BA - Toronto C5 171
Transition Density of Stylized MMM
• transition density of discounted NP S̄ δ∗
12
1 y x+y
p(s, x; t, y) = exp −
2 (ϕt − ϕs ) x 2 (ϕt − ϕs )
√
xy
× I1
ϕt − ϕ s
α
ϕt = (exp{η t} − 1)
4η
I1 (·) modified Bessel function of the first kind
c Copyright Eckhard Platen 10 BA - Toronto C5 172
• non-central chi-square distributed random variable:
y S̄tδ∗
=
ϕt − ϕ s ϕt − ϕ s
with δ = 4 degrees of freedom and non-centrality parameter:
x S̄sδ∗
=
ϕt − ϕ s ϕt − ϕ s
c Copyright Eckhard Platen 10 BA - Toronto C5 173
0.15
0.5
0.1
0.4
0.05
0 0.3
80
t
90 0.2
100
0.1
y 110
120
Figure 5.7: Transition density of squared Bessel process for δ = 4.
c Copyright Eckhard Platen 10 BA - Toronto C5 174
Zero Coupon Bond under the stylized MMM
• zero coupon bond
! !
1 S̄tδ∗
P (t, T ) = Stδ∗ Et = PT∗ (t) Et
STδ∗ S̄Tδ∗
with savings bond
( )
Z T
PT∗ (t) = exp − rs ds
t
c Copyright Eckhard Platen 10 BA - Toronto C5 175
δ=4 =⇒
( )!
S̄tδ∗
P (t, T ) = PT∗ (t) 1 − exp − < PT∗ (t)
2 (ϕT − ϕt )
for t ∈ [0, T ), Platen (2002)
with time transform
α
ϕt = (exp{η t} − 1)
4η
P (0, T ) = 0.00077
PT∗ (0) = 0.0335
P (0,T )
PT∗ (0)
≈ 0.023
c Copyright Eckhard Platen 10 BA - Toronto C5 176
1
savings bond
fair zero coupon bond
0.9 savings account
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
1930 1940 1950 1960 1970 1980 1990 2000
time
Figure 5.8: Savings bond, fair zero coupon bond and savings account.
c Copyright Eckhard Platen 10 BA - Toronto C5 177
1
ln(savings bond)
ln(fair zero coupon bond)
0
-1
-2
-3
-4
-5
-6
-7
-8
1930 1940 1950 1960 1970 1980 1990 2000
time
Figure 5.9: Logarithms of savings bond and fair zero coupon bond.
c Copyright Eckhard Platen 10 BA - Toronto C5 178
Forward Rates under the MMM
• forward rate
∂
f (t, T ) = − ln(P (t, T ))
∂T
= rT + n(t, T )
c Copyright Eckhard Platen 10 BA - Toronto C5 179
• market price of risk contribution
( )!
∂ S̄tδ∗
n(t, T ) = − ln 1 − exp −
∂T 2(ϕT − ϕt )
1 S̄tδ∗ αT
= 2
S̄tδ∗
n o
exp −1 (ϕ T − ϕ t ) 8
2(ϕT −ϕt )
lim n(t, T ) = η
T →∞
Pl. (2005a), Miller & Pl. (2005)
c Copyright Eckhard Platen 10 BA - Toronto C5 180
0.1
0.075 80
0.05
0.025 60
0
40
0.02 T
0.04
20
0.06
eta 0.08
0.1
Figure 5.10: Market price of risk contribution in dependence on η and T .
c Copyright Eckhard Platen 10 BA - Toronto C5 181
Free Snack from Savings Bond
• potential existence of weak form of arbitrage?
borrow amount P (0, T ) from savings account
Stδ = P (t, T ) − P (0, T ) exp{r t}
such that S0δ = 0
STδ = 1 − P (0, T ) exp{r T } > 0
lower bound
Stδ ≥ −P (0, T ) exp{r t}
c Copyright Eckhard Platen 10 BA - Toronto C5 182
• since Stδ may become negative
allowed by strong arbitrage concept
• cheap thrills may exist
Loewenstein & Willard (2000)
MMM does not admit an equivalent risk neutral probability measure
=⇒
there is a free lunch with vanishing risk
Delbaen & Schachermayer (2006)
c Copyright Eckhard Platen 10 BA - Toronto C5 183
Absence of an Equivalent Risk Neutral Probability Measure
• fair zero coupon bond has a lower price than the savings bond
( Z )
T 0
S
P (t, T ) < PT∗ (t) = exp − rs ds = 0 t
t ST
• Radon-Nikodym derivative
Ŝt0 S̄0δ∗
Λt = =
Ŝ00 S̄tδ∗
strict (A, P )-supermartingale under MMM
c Copyright Eckhard Platen 10 BA - Toronto C5 184
1.8
Radon-Nikodym derivative
Total RN-Mass
1.6
1.4
1.2
0.8
0.6
0.4
0.2
0
1930 1940 1950 1960 1970 1980 1990 2000
time
Figure 5.11: Radon-Nikodym derivative and total mass of putative risk neu-
tral measure.
c Copyright Eckhard Platen 10 BA - Toronto C5 185
• hypothetical risk neutral measure
total mass:
( )
S̄0δ∗
Pθ,T (Ω) = E0 (ΛT ) = 1 − exp − < Λ0 = 1
2 ϕT
Pθ is not a probability measure
c Copyright Eckhard Platen 10 BA - Toronto C5 186
European Call Options under the MMM
real world pricing formula =⇒
!
(STδ∗ − K) +
cT ,K (t, Stδ∗ ) = Stδ∗ Et
STδ∗
!+
K Stδ∗
= Et Stδ∗ −
STδ∗
2
Stδ∗
= 1 − χ (d1 ; 4, ℓ2 )
− K exp{−r (T − t)} (1 − χ2 (d1 ; 0, ℓ2 ))
χ2 (·; ·, ·) non-central chi-square distribution
c Copyright Eckhard Platen 10 BA - Toronto C5 187
with
4 η K exp{−r (T − t)}
d1 =
St0 αt (exp{η (T − t)} − 1)
and
2 η Stδ∗
ℓ2 =
St0 αt (exp{η (T − t)} − 1)
Hulley, Miller & Pl. (2005)
Hulley & Pl. (2008)
Miller & Pl. (2008)
c Copyright Eckhard Platen 10 BA - Toronto C5 188
0.275
0.25 10
0.225
8
0.2
6
0.6
T
0.8 4
1
2
K 1.2
Figure 5.12: Implied volatility surface for the stylized MMM.
c Copyright Eckhard Platen 10 BA - Toronto C5 189
• implied volatility
in BS-formula adjust short rate to
1
r̂ = − ln(P (t, T ))
T −t
otherwise put and call implied volatilities do not match
c Copyright Eckhard Platen 10 BA - Toronto C5 190
European Put Options under the MMM
• fair put-call parity relation
pT ,K (t, S̄tδ∗ ) = cT ,K (t, S̄tδ∗ ) − Stδ∗ + K P (t, T )
• European put option formula
2
pT ,K (t, Stδ∗ ) −Stδ∗
= χ (d1 ; 4, ℓ2 )
2
+ K exp{−r (T − t)} χ (d1 ; 0, ℓ2 ) − exp {−ℓ2 }
Hulley, Miller & Pl. (2005)
c Copyright Eckhard Platen 10 BA - Toronto C5 191
• put-call parity breaks down if one uses the savings bond PT∗ (t)
pT ,K (t, S̄tδ∗ ) < cT ,K (t, S̄tδ∗ ) − Stδ∗ + K exp{−r (T − t)}
for t ∈ [0, T )
c Copyright Eckhard Platen 10 BA - Toronto C5 192
• excellent hedge performance for extreme maturities
Hulley & Pl. (2008)
European calls
European puts
barrier options
c Copyright Eckhard Platen 10 BA - Toronto C5 193
Comparison to Hypothetical Risk Neutral Prices
• hypothetical risk neutral price c∗T ,K (t, Stδ∗ )
of a European call option on the NP
• benchmarked hypothetical risk neutral call price
c∗T ,K (t, Stδ∗ )
ĉ∗T ,K (t, Stδ∗ ) =
Stδ∗
local martingale
ĉ∗T ,K (·, ·) uniformly bounded
c Copyright Eckhard Platen 10 BA - Toronto C5 194
=⇒ ĉ∗T ,K martingale =⇒
ĉ∗T ,K (t, Stδ∗ ) = ĉT ,K (t, Stδ∗ )
=⇒
c∗T ,K (t, Stδ∗ ) = cT ,K (t, Stδ∗ )
c Copyright Eckhard Platen 10 BA - Toronto C5 195
• hypothetical risk neutral put-call parity
p∗T ,K (t, Stδ∗ ) = c∗T ,K (t, Stδ∗ ) − Stδ∗ + K PT∗ (t)
since PT∗ (t) > P (t, T ) =⇒
pT ,K (t, Stδ∗ ) < p∗T ,K (t, Stδ∗ )
t ∈ [0, T )
c Copyright Eckhard Platen 10 BA - Toronto C5 196
Difference in Asymptotic Put Prices
• hypothetical risk neutral prices can become extreme if NP value tends
towards zero
• asymptotic fair zero coupon bond
a.s.
lim PT (t, T ) = 0
S̄tδ∗ →0
for t ∈ [0, T )
• asymptotic fair European call
a.s.
lim cT ,K (t, Stδ∗ ) = lim Stδ∗ ĉT ,K (t, Stδ∗ ) = 0
S̄tδ∗ →0 S̄tδ∗ →0
c Copyright Eckhard Platen 10 BA - Toronto C5 197
• asymptotic fair put
fair put-call parity =⇒
a.s.
lim pT ,K (t, Stδ∗ ) = 0
S̄tδ∗ →0
• asymptotic hypothetical risk neutral put
hypothetical risk neutral put-call parity =⇒
lim p∗T ,K (t, Stδ∗ )
S̄tδ∗ →0
( )!
a.s. St0 S̄tδ∗
= lim pT ,K (t, Stδ∗ ) + K exp −
S̄tδ∗ →0 ST0 2(ϕT − ϕt )
a.s. St0
= K >0
ST0
dramatic differences can arise
c Copyright Eckhard Platen 10 BA - Toronto C5 198
ln(K*savings bond)
ln(risk neutral put)
8 ln(put)
1930 1940 1950 1960 1970 1980 1990 2000
time
Figure 5.13: Logarithms of savings bond times K, risk neutral put and fair
put.
c Copyright Eckhard Platen 10 BA - Toronto C5 199
MMM with Random Market Activity
• squared Bessel process with dimension δ = 4
S̄ δ∗ = {S̄tδ∗ , t ∈ [0, ∞)}
q
dS̄tδ∗ = αt dt + αt S̄tδ∗ dWt
c Copyright Eckhard Platen 10 BA - Toronto C5 200
• NP
Stδ∗ = St0 S̄tδ∗
• volatility
s
αt
|θt | =
S̄tδ∗
• discounted NP drift
αt to be modeled
c Copyright Eckhard Platen 10 BA - Toronto C5 201
0.275
0.25 10
0.225
8
0.2
6
0.6
T
0.8 4
1
2
K 1.2
Figure 5.14: Implied volatility surface for the stylized MMM.
c Copyright Eckhard Platen 10 BA - Toronto C5 202
T 1.6
1.4
1.2
1
0.8
0.6
0.4
0.2
0.28
0.26
0.24
0.22
0.5
0.75
1
1.25
K S
1.5
Figure 5.15: Average S&P500 implied volatility surface.
c Copyright Eckhard Platen 10 BA - Toronto C5 203
60
40 1.4
20 1.2
1
moneyness
100 0.8
time 200
0.6
300
Figure 5.16: Implied volatilities for S&P500 three month options.
c Copyright Eckhard Platen 10 BA - Toronto C5 204
50
40
1.4
30
20 1.2
10
1
moneyness
100 0.8
time 200
0.6
300
Figure 5.17: Implied volatilities for S&P500 one year options.
c Copyright Eckhard Platen 10 BA - Toronto C5 205
• market activity mt
Breymann, Kelly & Pl. (2006)
Pl. & Rendek (2008)
αt = ξt mt
with
ξt = ξ0 exp {η t}
p
2
dmt = k(mt ) β dt + β mt ̺ dWt + 1 − ̺2 dW̃t
̺ - correlation parameter
activity volatility β>0
W̃ - independent Wiener process
c Copyright Eckhard Platen 10 BA - Toronto C5 206
Example:
• drift
mt
k(mt ) = (p − g mt )
2
• stationary density
g p−1
pm (y) = y p−2 exp{−g y}
Γ(p − 1)
p−1 1
gamma density with mean g
and variance g
for p > 1 and g > 0
c Copyright Eckhard Platen 10 BA - Toronto C5 207
2 50
1 40
0 30
0.5 g
20
1
10
y
1.5
Figure 5.18: Stationary density of market activity mt = y as function of y
and speed of adjustment parameter g.
c Copyright Eckhard Platen 10 BA - Toronto C5 208
Zero Coupon Bond
• benchmarked zero coupon bond
! !
1 1
P̂T (t, S̄tδ∗ , mt ) = Et = Et
STδ∗ ST0 S̄Tδ∗
• zero coupon bond
PT (t, S̄tδ∗ , mt ) = Stδ∗ P̂T (t, S̄tδ∗ , mt ) = St0 S̄tδ∗ P̂T (t, S̄tδ∗ , mt )
c Copyright Eckhard Platen 10 BA - Toronto C5 209
European Options on a Market Index
• put option price
!+
K
pT ,K (t, S̄tδ∗ , mt ) = St0 S̄tδ∗ Et −1
ST0 S̄Tδ∗
effect of random market time on implied volatilities
curvature for the short dated implied volatility surface
Heath & Pl. (2005a)
c Copyright Eckhard Platen 10 BA - Toronto C5 210
0.36
0.34 1
0.32
0.8
0.3
80 0.6
T
90
0.4
100
K 110 0.2
120
Figure 5.19: Implied volatilities for put options on index as a function of
strike K and maturity T .
c Copyright Eckhard Platen 10 BA - Toronto C5 211
0.34 10
0.32
8
0.3
80 6
g
90
100 4
K 110
1202
Figure 5.20: Implied volatilities for put options as a function of strike K and
speed of adjustment g.
c Copyright Eckhard Platen 10 BA - Toronto C5 212
0.4
0.375 10
0.35
8
0.325
0.3 6
80 T
4
100
K 2
120
Figure 5.21: Implied volatilities for long dated put options as a function of
strike K and maturity T .
c Copyright Eckhard Platen 10 BA - Toronto C5 213
6 Stylized Multi-Currency MMM
Pl. (2001)
Heath & Pl. (2005a)
d + 1 currencies
Siδ∗ (t) NP in ith currency
rti - ith short rate
θik (t) - market price of risk
Siδ∗ (t) = αit Yti Sii (t)
c Copyright Eckhard Platen 10 BA - Toronto C6 214
αit = αi0 exp{η i t}
Sii (t) = exp{r i t}
• ith normalized NP
q d+1
X
dYti = 1 − η i Yt i i
q i,k dWtk
dt + Yt
k=1
scaling levels q i,k
d+1
X
(q i,k )2 = 1
k=1
c Copyright Eckhard Platen 10 BA - Toronto C6 215
• (i, j)th exchange rate
Siδ∗ (t) Yti αit Sii (t)
Xti,j = =
Sjδ∗ (t) Ytj αjt Sjj (t)
!
d+1 i,k j,k i,k
X q q q
dXti,j = Xti,j (r i − r j ) dt + p −q p dt + dWtk
k=1 Yti Ytj Yti
c Copyright Eckhard Platen 10 BA - Toronto C6 216
• jth savings account in ith currency
Sij (t) = Xti,j Sjj (t)
!
d+1
X q i,k q j,k q i,k
dSij (t) = Sij (t) r i dt + p −q p dt + dWtk
i i
k=1 Y t Yt j Y t
c Copyright Eckhard Platen 10 BA - Toronto C6 217
• stochastic market price of risk
i,k
q
θik (t) = p
Yti
• (j, k)th volatility
bj,k
i (t) = θi
k
(t) − θ k
j (t)
equity prices
commodity prices
c Copyright Eckhard Platen 10 BA - Toronto C6 218
7 Valuing Guaranteed Minimum Death Benefits
• variable annuities
fund-linked
tax-deferred
guarantees
• guaranteed minimum death benefits (GMDBs)
roll-ups:
original investment accrued at a pre-defined interest rate
ratchets:
death benefit based upon the highest anniversary account value
embedded options:
hedging against market downturn occurrence of death
c Copyright Eckhard Platen 10 BA - Toronto C7 219
• GMDB put, floating put and/or look-back put option
long maturities of contracts
standard option pricing theory ?
no obvious best choice for price:
IFRS Phase II
CFO-Forum (2008)
CRO-Forum (2008)
Solvency II
Verheugen & Hines (2008)
c Copyright Eckhard Platen 10 BA - Toronto C7 220
• pricing and hedging of GMDBs
benchmark approach in
Pl. (2002, 2004b), Pl. & Heath (2006)
Bühlmann & Pl. (2003)
best performing portfolio is benchmark
does not require
existence of an equivalent risk neutral probability measure
unique fair price is the minimal price
may provide significantly lower prices
c Copyright Eckhard Platen 10 BA - Toronto C7 221
Financial Model
• underlying risky security (unit)
dSt = (µt − γ)St dt + σt St dWt
γ ≥ 0 management fee rates
Wt - Brownian motion
• savings account
dBt = rt Bt dt
c Copyright Eckhard Platen 10 BA - Toronto C7 222
• market price of risk
µt − γ − rt
θt =
σt
• risky asset
dSt
= rt dt + σt θt dt + σt dWt
St
c Copyright Eckhard Platen 10 BA - Toronto C7 223
• self-financing portfolio
Vt = δt0 Bt + δt1 St
dVt = δt0 dBt + δt1 dSt
• fractions
0 Bt 1 St
πt0 = δt , πt1 = δt
Vt Vt
πt0 + πt1 = 1
c Copyright Eckhard Platen 10 BA - Toronto C7 224
• instantaneous portfolio return
dVt dBt dSt
= πt0 + πt1
Vt Bt St
= rt dt + πt1 σt (θt dt + dWt )
c Copyright Eckhard Platen 10 BA - Toronto C7 225
• numeraire portfolio (NP) as benchmark
Long (1990)
V ∗ is best performing in several ways
• is growth optimal portfolio
Kelly (1956)
V ∗ = max E(log VT )
µt − γ − rt θt
πt1∗ = =
σt2 σt
dVt∗
= rt dt + θt (θt dt + dWt )
Vt∗
Merton (1992)
c Copyright Eckhard Platen 10 BA - Toronto C7 226
• NP best performing portfolio
1 VT∗ 1 VT
lim sup log ≥ lim sup log
T →∞ T V0∗ T →∞ T V0
global well diversified index ≈ NP
Pl. (2005b)
c Copyright Eckhard Platen 10 BA - Toronto C7 227
100
90
80
70
60
50
40
30
20
10
0
1930 1940 1950 1960 1970 1980 1990 2000
time
Figure 7.1: Discounted S&P500 total return index.
c Copyright Eckhard Platen 10 BA - Toronto C7 228
5
-1
-2
1930 1940 1950 1960 1970 1980 1990 2000
time
Figure 7.2: Logarithm of discounted S&P500.
c Copyright Eckhard Platen 10 BA - Toronto C7 229
• benchmarked price
Ut
Ût = Vt∗
Û nonnegative =⇒
Ût ≥ Et Ûs
t≤s
supermartingales (no upward trend)
Pl. (2002)
no strong arbitrage
c Copyright Eckhard Platen 10 BA - Toronto C7 230
• benchmarked securities that form martingales are called fair.
Ût = Et Ûs
t≤s
c Copyright Eckhard Platen 10 BA - Toronto C7 231
• Law of the Minimal Price
Pl. (2008)
Fair prices are minimal prices
V nonnegative fair portfolio
V ′ nonnegative portfolio
VT = VT′
supermartingale property
=⇒
Vt ≤ Vt′
t ∈ [0, T ],
c Copyright Eckhard Platen 10 BA - Toronto C7 232
0.16
benchmarked savings bond
benchmarked fair zero coupon bond
0.14
0.12
0.1
0.08
0.06
0.04
0.02
0
1930 1940 1950 1960 1970 1980 1990 2000
time
Figure 7.3: Benchmarked savings bond and benchmarked fair zero coupon
bond.
c Copyright Eckhard Platen 10 BA - Toronto C7 233
• real world pricing formula
for
HT
Et <∞
VT∗
HT
UH (t) = Vt∗ Et
VT∗
t ∈ [0, T ]
no risk neutral probability needs to exist
c Copyright Eckhard Platen 10 BA - Toronto C7 234
• actuarial pricing formula
when HT is independent of VT∗
=⇒
UH (t) = P (t, T ) Et (HT )
zero coupon bond
1
P (t, T ) = Vt∗ Et
VT∗
c Copyright Eckhard Platen 10 BA - Toronto C7 235
• standard risk neutral pricing
Ross (1976), Harrison & Pliska (1983)
complete market
candidate Radon-Nikodym derivative
dQ Bt V0∗
Λt = =
dP At B0 Vt∗
supermartingale =⇒
1 = Λ0 ≥ E0 (ΛT )
c Copyright Eckhard Platen 10 BA - Toronto C7 236
=⇒
V0∗
UH (0) = E HT
VT∗
B0
B0
E ΛT BT
HT
= E ΛT HT ≤
BT E(ΛT )
c Copyright Eckhard Platen 10 BA - Toronto C7 237
Only in the special case when ΛT martingale
=⇒ risk neutral pricing formula:
B0
UH (0) = EQ HT
BT
Ignores any real trend !
c Copyright Eckhard Platen 10 BA - Toronto C7 238
• Trends ignored also when using:
stochastic discount factor, Cochrane (2001);
deflator, Duffie (2001);
pricing kernel, Constantinides (1992);
state price density, Ingersoll (1987);
NP as Long (1990)
c Copyright Eckhard Platen 10 BA - Toronto C7 239
• Example zero coupon bond
benchmarked fair zero coupon
P (t, T ) 1
P̂ (t, T ) = = Et
Vt∗ VT∗
martingale (no trend)
for rt - deterministic
T !
1 V̄t∗
Z
P (t, T ) = Vt∗ Et ∗
= exp − rs ds Et ,
VT V̄T∗
t
Vt∗
V̄t∗ = Bt
- discounted NP
c Copyright Eckhard Platen 10 BA - Toronto C7 240
0.16
benchmarked savings bond
benchmarked fair zero coupon bond
0.14
0.12
0.1
0.08
0.06
0.04
0.02
0
1930 1940 1950 1960 1970 1980 1990 2000
time
Figure 7.4: Benchmarked savings bond and benchmarked fair zero coupon
bond.
c Copyright Eckhard Platen 10 BA - Toronto C7 241
=⇒ downward trend reflects equity premium
=⇒ Λ strict supermartingale (downward trend), then
T
Bt
Z
P (t, T ) < exp − rs ds =
BT
t
c Copyright Eckhard Platen 10 BA - Toronto C7 242
1.8
Radon-Nikodym derivative
Total RN-Mass
1.6
1.4
1.2
0.8
0.6
0.4
0.2
0
1930 1940 1950 1960 1970 1980 1990 2000
time
Figure 7.5: Radon-Nikodym derivative and total mass of putative risk neutral
measure.
c Copyright Eckhard Platen 10 BA - Toronto C7 243
1
savings bond
fair zero coupon bond
0.9 savings account
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
1930 1940 1950 1960 1970 1980 1990 2000
time
Figure 7.6: Savings bond, fair zero coupon bond and savings account.
c Copyright Eckhard Platen 10 BA - Toronto C7 244
1
ln(savings bond)
ln(fair zero coupon bond)
0
-1
-2
-3
-4
-5
-6
-7
-8
1930 1940 1950 1960 1970 1980 1990 2000
time
Figure 7.7: Logarithms of savings bond and fair zero coupon bond.
c Copyright Eckhard Platen 10 BA - Toronto C7 245
5
-1
-2
1930 1940 1950 1960 1970 1980 1990 2000
time
Figure 7.8: Logarithm of discounted S&P500.
c Copyright Eckhard Platen 10 BA - Toronto C7 246
• discounted NP
Vt∗
V̄t∗ =
Bt
dV̄t∗ = V̄t∗ θt (θt dt + dWt )
q
= αt dt + αt V̄t∗ dWt
c Copyright Eckhard Platen 10 BA - Toronto C7 247
• discounted NP drift
αt := V̄t∗ θt2
=⇒ volatility
αt
r
θt =
V̄t∗
reflects leverage effect
c Copyright Eckhard Platen 10 BA - Toronto C7 248
• minimal market model
Pl. (2001, 2002)
MMM
discounted NP drift
assume
αt = α0 exp{ηt}
α0 > 0
net growth rate η > 0
=⇒ MMM
c Copyright Eckhard Platen 10 BA - Toronto C7 249
5
ln(discounted index)
trendline
-1
-2
1930 1940 1950 1960 1970 1980 1990 2000
time
Figure 7.9: Logarithm of discounted S&P500 and trendline.
c Copyright Eckhard Platen 10 BA - Toronto C7 250
• normalized NP
V̄t∗
Yt =
αt
p
dYt = (1 − ηYt ) dt + Yt dWt
square root process of dimension four
Only one parameter !
c Copyright Eckhard Platen 10 BA - Toronto C7 251
• volatility of NP
1 αt
r
θt = √ =
Yt V̄t∗
c Copyright Eckhard Platen 10 BA - Toronto C7 252
• Zero coupon bond under the MMM
rt - deterministic
T ( )!
∗
Z V̄t
P (t, T ) = exp − rs ds 1 − exp −
2(ϕ(T ) − ϕ(t))
t
Explicit formula !
c Copyright Eckhard Platen 10 BA - Toronto C7 253
1
ln(savings bond)
ln(fair zero coupon bond)
0
-1
-2
-3
-4
-5
-6
-7
-8
1930 1940 1950 1960 1970 1980 1990 2000
time
Figure 7.10: Logarithms of savings bond and fair zero coupon bond.
c Copyright Eckhard Platen 10 BA - Toronto C7 254
1
0.8
0.6
0.4
0.2
0
1930 1940 1950 1960 1970 1980 1990 2000
time
Figure 7.11: Fraction invested in the savings account.
c Copyright Eckhard Platen 10 BA - Toronto C7 255
7e-005
6e-005
5e-005
4e-005
3e-005
2e-005
1e-005
-1e-005
-2e-005
1930 1940 1950 1960 1970 1980 1990 2000
time
Figure 7.12: Benchmarked profit and loss.
c Copyright Eckhard Platen 10 BA - Toronto C7 256
since
! ( )!
ΛT V̄t∗ V̄t∗
Et = Et = 1 − exp −
Λt V̄T∗ 2(ϕ(T ) − ϕ(t))
Bt
=⇒ P (t, T ) < BT
Overpricing by savings bond
no equivalent risk neutral probability measure
c Copyright Eckhard Platen 10 BA - Toronto C7 257
1.8
Radon-Nikodym derivative
Total RN-Mass
1.6
1.4
1.2
0.8
0.6
0.4
0.2
0
1930 1940 1950 1960 1970 1980 1990 2000
time
Figure 7.13: Radon-Nikodym derivative and total mass of putative risk neu-
tral measure.
c Copyright Eckhard Platen 10 BA - Toronto C7 258
• European put option under the MMM
Hulley, Miller & Pl. (2005)
!
(K − Vτ∗ )+
p(t, Vt∗ , T, K, r) = Vt∗ Et
Vτ∗
p(t, Vt∗ , T, K, r) = −Vt∗ χ2 (d1 ; 4, l2 )
−r(T −t) 2
+ Ke χ (d1 ; 0, l2 ) − exp {−l2 /2}
c Copyright Eckhard Platen 10 BA - Toronto C7 259
with
4ηK exp{−r(T − t)}
d1 =
Bt αt (exp{η(T − t)} − 1)
and
4ηVt∗
l2 =
Bt αt (exp{η(T − t)} − 1)
c Copyright Eckhard Platen 10 BA - Toronto C7 260
χ2 (x; n, l) non-central chi-square distribution function
n ≥ 0 degrees of freedom
non-centrality parameter l > 0
∞
exp − 2l l k Γ x2 ; n+2k
X 2
χ2 (x; n, l) = 2 1 −
k=0
k! Γ n+2k2
c Copyright Eckhard Platen 10 BA - Toronto C7 261
• putative risk neutral price
( )
V̄t∗
p̃(t, Vt∗ , T, K, r) = p(t, Vt∗ , T, K, r)+Ke−r(T −t) exp −
2(ϕ(T ) − ϕ(t))
risk neutral is overpricing
for V̄t∗ → 0 =⇒ p̃(t, Vt∗ , T, K, r) → 0
and
p̃(t, Vt∗ , T, K, r) → K e−r(T −t) > 0
risk neutral ignores trends
c Copyright Eckhard Platen 10 BA - Toronto C7 262
ln(K*savings bond)
ln(risk neutral put)
8 ln(put)
1930 1940 1950 1960 1970 1980 1990 2000
time
Figure 7.14: Logarithms of savings bond times K, risk neutral put and fair
put.
c Copyright Eckhard Platen 10 BA - Toronto C7 263
Guaranteed Minimum Death Benefit (GMDB)
• payout to the policyholder
max(egτ V0 , Vτ )
time of death τ
g ≥ 0 is the guaranteed instantaneous growth rate
V0 is the initial account value
Vτ is the unit value of the policyholder’s account at time of death τ
c Copyright Eckhard Platen 10 BA - Toronto C7 264
• insurance charges ξ ≥ 0
=⇒ policyholder’s unit value
Vt = e−ξt Vt∗
=⇒
max(egτ V0 , Vτ ) = max(egτ V0 , e−ξτ Vτ∗ )
= e−ξτ max(e(g+ξ)τ V0 , Vτ∗ )
c Copyright Eckhard Platen 10 BA - Toronto C7 265
insurance company invests the entire fund value V in the NP V ∗
=⇒ payoff
h i
HT = GM DBT = e−ξT (e(g+ξ)T V0∗ − VT∗ )+ + VT∗
c Copyright Eckhard Platen 10 BA - Toronto C7 266
fair value GM DB0 of the total claim
real world pricing formula
GM DBT
GM DB0 = V0∗ E
VT∗
= e−ξT p(0, V0∗ , T, e(g+ξ)T V0∗ , r) + V0∗
c Copyright Eckhard Platen 10 BA - Toronto C7 267
MMM risk neutral Black Scholes
1.15
1.1 1.1 1.1
1.05 g=0 1.05 1.05
g=0
GMDB0
GMDB0
0
1 1 1
GMDB
g=0
0.95 0.95 0.95
0.9 0.9 0.9
0.85 0.85 0.85
0.8 0.8 0.8
0 5 10 15 20 25 30 0 5 10 15 20 25 30 0 5 10 15 20 25 30
T T T
MMM risk neutral Black Scholes
1.15 1.15 1.15
1.1 1.1 1.1 g=0.025
g=0.025
1.05 g=0.025 1.05 1.05
GMDB0
GMDB0
0
GMDB
1 1 1
0.95 0.95 0.95
0.9 0.9 0.9
0 5 10 15 20 25 30 0 5 10 15 20 25 30 0 5 10 15 20 25 30
T T T
Figure 7.15: Present value of the GMDB under the real world pricing formula (left),
the risk neutral pricing formula (middle) and the Black Scholes formula (right) for
η = 0.05, α0 = 0.05, r = 0.05, ξ = 0.01 and Y0 = 20.
c Copyright Eckhard Platen 10 BA - Toronto C7 268
• lifetime τ is stochastic
GM DBτ
GM DB0 = V0∗ E E Fτ
Vτ∗
Z T
GM DB0 = p(0, V0∗ , t, e(g+ξ)t V0∗ , r) + V0∗ e−ξt fτ (t) dt
0
fτ (·) - mortality density
c Copyright Eckhard Platen 10 BA - Toronto C7 269
0
10
male
−1
female
10
−2
10
log(qx)
−3
10
−4
10
−5
10
0 20 40 60 80 100 120
Age x
Figure 7.16: German mortality data
c Copyright Eckhard Platen 10 BA - Toronto C7 270
rational investors will lapse when embedded put-option out of the money
GMDBs have stochastic maturity
Titanic options, Milevsky & Posner (2001)
• roll-up GMDB payoff
(1 − β )V ∗ , if lapsed at time t,
t t
Ht =
max(egτ V0 , V ∗ ), if death occurs at time t = τ
τ
surrender charge βt
(8 − ⌈t⌉)%, t ≤ 7,
βt =
0, t>7
no credit risk
no accumulation phase
c Copyright Eckhard Platen 10 BA - Toronto C7 271
1.15
male
female
1.1
GMDB0
1.05
1
20 30 40 50 60 70 80
Age x
Figure 7.17: Value of the GMDB under the MMM for male and female pol-
icyholders aged x, assuming an irrational lapsation of l = 1%.
c Copyright Eckhard Platen 10 BA - Toronto C7 272
8 Markets with Event Risk
Jump Diffusion Markets
• filtered probability space (Ω, A, A, P )
• continuous trading uncertainty
m ∈ {1, 2, . . . , d}
Wiener processes W̃ k = {W̃tk , t ∈ [0, ∞)}, k ∈ {1, 2, . . . , m}
c Copyright Eckhard Platen 10 BA - Toronto C8 273
• events
counting process pk = {pkt , t ∈ [0, ∞)}
intensity hk = {hkt , t ∈ [0, ∞)}
hkt > 0
Z t
hks ds < ∞
0
• jump martingale
k − 12
dqtk = dpkt − hkt dt ht
k ∈ {1, 2, . . . , d−m}
c Copyright Eckhard Platen 10 BA - Toronto C8 274
2
k
E qt+∆ − qtk At =∆
• trading uncertainty
W = {W t = (W̃t1 , . . . , W̃tm , qt1 , . . . , qtd−m )⊤ , t ∈ [0, ∞)}
c Copyright Eckhard Platen 10 BA - Toronto C8 275
Primary Security Accounts
d
!
X
dStj = St−
j
ajt dt + bj,k
t dWt
k
k=1
j ∈ {1, 2, . . . , d}
c Copyright Eckhard Platen 10 BA - Toronto C8 276
Assumption 8.1
q
bj,k
t ≥− hk−m
t
for all t ∈ [0, ∞), j ∈ {1, 2, . . . , d} and k ∈ {m + 1, . . . , d}.
Assumption 8.2 bt is invertible
c Copyright Eckhard Platen 10 BA - Toronto C8 277
• market price of risk
θ t = (θt1 , . . . , θtd )⊤ = b−1
t [at − rt 1]
at = (a1t , . . . , adt )⊤
1 = (1, . . . , 1)⊤
=⇒
d
!
X
dStj = St−
j
rt dt + bj,k
t (θt
k
dt + dW k
t )
k=1
j ∈ {0, 1, . . . , d}
c Copyright Eckhard Platen 10 BA - Toronto C8 278
• strategy
δ = {δ t = (δt0 , . . . , δtd )⊤ , t ∈ [0, ∞)}
• portfolio
d
X
Stδ = δtj Stj
j=0
• self-financing if
d
X
dStδ = δtj dStj
j=0
c Copyright Eckhard Platen 10 BA - Toronto C8 279
Growth Optimal Portfolio
• fraction
j Stj
πδ,t = δtj
Stδ
1 d ⊤
π δ,t = (πδ,t , . . . , πδ,t )
n o
⊤
dStδ = δ
St− rt dt + π δ,t− bt (θ t dt + dW t )
strictly positive ⇐⇒
d
X q
j
πδ,t bj,k
t >− hk−m
t
j=1
k ∈ {m + 1, . . . , d}
c Copyright Eckhard Platen 10 BA - Toronto C8 280
m X
X d
j
d ln(Stδ ) = gtδ dt + πδ,t bj,k
t dWt
k
k=1 j=1
d d
bj,k
q
t
X X j
+ ln1 + πδ,t− q hk−m
t dWtk
k=m+1 j=1 hk−m
t
• growth rate
2
m d d
X X j 1 X
gtδ = rt + πδ,t bj,k
t θtk − j
πδ,t bj,k
t
2
k=1 j =1 j =1
d d d j,k
b
X X q X j
j j,k k k−m hk−m
+ πδ,t bt θt − ht +ln1+ πδ,t q t t
k=m+1 j =1 j =1 hk−m
t
c Copyright Eckhard Platen 10 BA - Toronto C8 281
Assumption 8.3
q
hk−m
t > θtk
k ∈ {m + 1, . . . , d}
c Copyright Eckhard Platen 10 BA - Toronto C8 282
• fractions of candidate benchmark
⊤
π δ∗ ,t = (πδ1∗ ,t , . . . , πδd∗ ,t )⊤ = c⊤
t b−1
t
θtk for k ∈ {1, 2, . . . , m}
ckt =
θtk
k−m − 1
for k ∈ {m + 1, . . . , d}
1−θtk (ht ) 2
θtk
if q
k−m
≪1
ht
=⇒ ckt ≈ θtk
c Copyright Eckhard Platen 10 BA - Toronto C8 283
• candidate benchmark
dStδ∗ = δ∗
St− rt dt + c⊤
t (θ t dt + dW t )
m
X
δ∗
= St− rt dt + θtk (θtk dt + dWtk )
k=1
d
X θtk
+ (θtk dt + dWtk )
−1
k=m+1 1 − θtk (hk−m
t ) 2
c Copyright Eckhard Platen 10 BA - Toronto C8 284
δ
Definition 8.4 S δ ∈ V + NP if gtδ ≤ gt for all S δ ∈ V + .
=⇒
S δ∗ is a NP.
=⇒
m d
1 X X θtk θtk
gtδ∗ = rt + (θtk )2 − hk−m
t
ln1 +q + q
2 k=1 k=m+1 hk−m − θtk hk−m
t t
c Copyright Eckhard Platen 10 BA - Toronto C8 285
θtk
for q
k−m
≪ 1 asymptotically
ht
d
1 X |θ t |2
gtδ∗ ≈ rt + (θtk )2 = rt +
2 k=1
2
d
X
d ln(Stδ∗ ) ≈ gtδ∗ dt + θtk dWtk
k=1
d
!
X
dStδ∗ Stδ∗ θtk θtk dWtk
≈ rt + dt +
k=1
c Copyright Eckhard Platen 10 BA - Toronto C8 286
JD
• jump diffusion market (JDM) S(d) = {S, a, b, ~r , h, A, P }
• benchmarked portfolio
Stδ
Ŝtδ =
Stδ∗
m
X d
X
dŜtδ = δtj Ŝtj bj,k
t − Ŝtδ θtk dWtk
k=1 j=1
d d
X X θtk
+ δtj j
Ŝt− bj,k
t
1 −q − Ŝ δ θ kdW k
t− t t
k=m+1 j=1 hk−m
t
driftless
c Copyright Eckhard Platen 10 BA - Toronto C8 287
• benchmarked savings account
d
X
dŜt0 = −Ŝt−
0
θtk dWtk
k=1
driftless
c Copyright Eckhard Platen 10 BA - Toronto C8 288
benchmarked portfolio Ŝ δ driftless =⇒ (A, P )-local martingale
Ansel & Stricker (1994)
=⇒
Theorem 8.5 In a JDM a nonnegative Ŝ δ is an (A, P )-supermartin-
gale
Ŝtδ ≥ E Ŝτδ At
τ ∈ [0, ∞) and t ∈ [0, τ ].
c Copyright Eckhard Platen 10 BA - Toronto C8 289
A nonnegative portfolio is a strong arbitrage if it can get out of zero.
Corollary 8.6
A JDM does not allow nonnegative portfolios that permit strong arbitrage.
c Copyright Eckhard Platen 10 BA - Toronto C8 290
Law of the Minimal Price
H
Corollary 8.7 Aτ -measurable payoff H with E( |A0 ) < ∞. If
Sτδ∗
there exists a fair nonnegative portfolio S δ that replicates the payoff
Sτδ = H,
then this is the minimal nonnegative replicating portfolio.
=⇒
• fair portfolios provide the cheapest hedge
c Copyright Eckhard Platen 10 BA - Toronto C8 291
H
H Aτ -measurable payoff, with E( )<∞
Sτδ∗
=⇒
real world pricing formula
H
UH (t) = Stδ∗ E At
Sτδ∗
• equivalent to risk neutral pricing as long as candidate Radon-Nikodym
Ŝt0
derivative Ŝ00
forms (A, P )-martingale
c Copyright Eckhard Platen 10 BA - Toronto C8 292
• zero coupon bond
!
Stδ∗
P (t, T ) = E At
STδ∗
• benchmarked zero coupon bond
P (t,T )
P̂ (t, T ) =
Stδ∗
d
X
dP̂ (t, T ) = −P̂ (t−, T ) σ k (t, T ) dWtk
k=1
c Copyright Eckhard Platen 10 BA - Toronto C8 293
=⇒
m Z t t
1
X Z
ln(P̂ (t, T )) = ln(P̂ (0, T )) − σ k
(s, T )dWsk + (σ k (s, T ))2 ds
k=1 0 2 0
d
! !
t t k
σ (s, T )
X Z q Z
+ σ k (s, T ) hk−m
s ds + ln 1 − p dpks
k=m+1 0 0 hk−m
s
c Copyright Eckhard Platen 10 BA - Toronto C8 294
• forward rate
∂ ∂
f (t, T ) = − ln(P (t, T )) = − ln(P̂ (t, T ))
∂T ∂T
=⇒
• forward rate equation
m Z t
∂
X
k k
dWsk
f (t, T ) = f (0, T ) + σ (s, T ) σ (s, T ) ds +
k=1 0 ∂T
d t
1 ∂
X Z
k k
dWsk
+ σ k (s,T )
σ (s, T ) σ (s, T ) ds +
k=m+1 0 1− √ ∂T
k−m
hs
similar to HJM equation
c Copyright Eckhard Platen 10 BA - Toronto C8 295
Sequences of Diversified Portfolios
δ
Sequence (S(d) )d∈N is a sequence of DPs if
j K2
πδ,t ≤ 1
d 2 +K1
for all j ∈ {0, 1, . . . , d}, t ∈ [0, ∞) and d ∈ {d0 , d0 + 1, . . .}.
c Copyright Eckhard Platen 10 BA - Toronto C8 296
• benchmarked primary security account
d
X
j j j,k
dŜ(d) (t) = −Ŝ(d) (t−) σ(d) (t) dWtk
k=1
• kth total specific volatility
d
X
k j,k
σ̂(d) (t) = |σ(d) (t)|
j=0
c Copyright Eckhard Platen 10 BA - Toronto C8 297
Sequence of JDMs regular if
2
k
E σ̂(d) (t) ≤ K5
for all t ∈ [0, ∞), d ∈ N and k ∈ {1, 2, . . . , d}.
c Copyright Eckhard Platen 10 BA - Toronto C8 298
Sequence of Approximate NPs
• tracking rate
2
d
X d
X
δ j j,k
R(d) (t) = πδ,t σ(d) (t)
k=1 j=0
δ δ
S(d) is NP ⇐⇒ Ŝ(d) ≡ constant
⇐⇒
δ
R(d) (t) = 0
c Copyright Eckhard Platen 10 BA - Toronto C8 299
δ
(S(d) )d∈N is a sequence of approximate NPs if
δ P
lim R(d) (t) = 0
d→∞
• expected tracking rate
eδ(d) (t) = E R(d)
δ
(t)
δ
(S(d) )d∈N has vanishing expected tracking rate, if
lim eδ(d) (t) = 0
d→∞
=⇒
c Copyright Eckhard Platen 10 BA - Toronto C8 300
δ
1
lim P R(d) (t) > ε ≤ lim eδ(d) (t) = 0
d→∞ d→∞ ε
Lemma 8.8 For a sequence of JDMs, any sequence of strictly positive
portfolios with vanishing expected tracking rate is a sequence of approximate
NPs.
c Copyright Eckhard Platen 10 BA - Toronto C8 301
Diversification Theorem
Platen (2005a)
JD δ
For a regular sequence of JDMs (S(d) )d∈N , each sequence (S(d) )d∈N
of DPs is a sequence of approximate NPs. Moreover
(K2 )2 K5
eδ(d) (t) ≤ .
d2K1
c Copyright Eckhard Platen 10 BA - Toronto C8 302
Diversification in an MMM Setting
• savings account
0
S(d) (t) = exp{r t}
• discounted NP drift
αδt ∗ = α0 exp{η t}
• jth benchmarked primary security account
j 1
Ŝ(d) (t) =
Ytj αδt ∗
c Copyright Eckhard Platen 10 BA - Toronto C8 303
• SR process
q
dYtj = 1 − η Ytj dt + Ytj dWtj
• NP
0
δ∗
S(d) (t)
S(d) (t) = 0
Ŝ(d) (t)
• jth primary security account
j j δ∗
S(d) (t) = Ŝ(d) (t) S(d) (t)
c Copyright Eckhard Platen 10 BA - Toronto C8 304
450
400
350
300
250
200
150
100
50
0
0 5 9 14 18 23 27 32
Figure 8.1: Primary security accounts under the MMM.
c Copyright Eckhard Platen 10 BA - Toronto C8 305
60
50
40
30
20
10
0
0 5 9 14 18 23 27 32
Figure 8.2: Benchmarked primary security accounts.
c Copyright Eckhard Platen 10 BA - Toronto C8 306
100
EWI
90 GOP
80
70
60
50
40
30
20
10
0
0 5 9 14 18 23 27 32
Figure 8.3: NP and EWI.
c Copyright Eckhard Platen 10 BA - Toronto C8 307
100
Market index
90
GOP
80
70
60
50
40
30
20
10
0
0 5 9 14 18 23 27 32
Figure 8.4: NP and market index.
c Copyright Eckhard Platen 10 BA - Toronto C8 308
Real World Pricing for Two Market Models
Specifying a Continuous NP
market prices of event risks are zero
θtk = 0
NP !
m
X
dStδ∗ Stδ∗ θtk θtk dWtk
= rt dt + dt +
k=1
S0δ∗ = 1
c Copyright Eckhard Platen 10 BA - Toronto C8 309
Benchmarked Primary Security Accounts
d
X
dŜtj = −Ŝt−
j
σtj,k dWtk
k=1
( m m
)
t X t
1
Z Z
j,k 2
X
Ŝtj S0j σsj,k dWsk
= exp − σs ds −
2 0 k=1 k=1 0
k−m
Z t d q d pt στj,k
k
l −
X Y Y
× exp σsj,k hk−m
s ds 1 − q
0 k=m+1
k=m+1 l=1 hk−m
τ k−
l
pivotal objects of study
c Copyright Eckhard Platen 10 BA - Toronto C8 310
• simplifying notation:
v
u m 2
X
|σtj | = t σtj,k
u
k=1
• aggregate continuous noise processes
m Z t
X σsj,k
Ŵtj = dWsk
k=1 0 |σsj |
c Copyright Eckhard Platen 10 BA - Toronto C8 311
• generalized volatility matrix bt = [bj,k d
t ]j,k=1 invertible
bj,k
t = θtk − σtj,k
for k ∈ {1, 2, . . . , m}
bj,k
t = −σtj,k
k ∈ {m + 1, . . . , d}
c Copyright Eckhard Platen 10 BA - Toronto C8 312
• constant intensities
hkt = hk > 0
√
σtj,k =σ j,k
≤ hk−m
• benchmarked jth primary security account:
Ŝtj = Ŝtj,c Stj,d
c Copyright Eckhard Platen 10 BA - Toronto C8 313
• continuous part
t t
1
Z Z
Ŝtj,c = S0j exp − |σsj |2 ds − |σsj | dŴsj
2 0 0
• compensated jump part
!pk−m
d
X √ d
Y σ j,k
t
Stj,d = exp σ j,k hk−m t 1− √
k=m+1
k=m+1 hk−m
c Copyright Eckhard Platen 10 BA - Toronto C8 314
Merton Model
rt = r, σtj,k = σ j,k
1
Ŝtj,c = S0j exp − |σ j |2 t − |σ j | Ŵtj
2
Girsanov’s theorem, see Section 9.5
c Copyright Eckhard Platen 10 BA - Toronto C8 315
A Minimal Market Model with Jumps
Hulley, Miller & Pl. (2005)
• discounted NP drift
αj (t) = αj0 exp{η j t}
η j - net growth rate
c Copyright Eckhard Platen 10 BA - Toronto C8 316
• jth square root process
q
dYtj = 1 − η j Ytj dt + Ytj dŴtj
• continuous part
1
Ŝtj,c =
αj (t)Ytj
c Copyright Eckhard Platen 10 BA - Toronto C8 317
• time transformation
t
1
Z
ϕj (t) = ϕj0 + αj (s) ds
4 0
• squared Bessel process of dimension four
j 1
Xϕ j (t) =α j
(t)Ytj =
Ŝtj,c
Ŝ 0 is a strict local martingale
c Copyright Eckhard Platen 10 BA - Toronto C8 318
Zero Coupon Bonds
! ( ) !
T
1 1
Z
P (t, T ) = Stδ∗ E At = E exp − rs ds ŜT0 At
STδ∗ Ŝt0 t
• MM case
1
P (t, T ) = exp{−r(T −t)} E ŜT0 At = exp{−r(T −t)}
Ŝt0
c Copyright Eckhard Platen 10 BA - Toronto C8 319
• MMM case
1
λjt =
Ŝtj (ϕj (t) − ϕj (T ))
ŜT0 = ŜT0,c
( ) !
T
1
Z
P (t, T ) = E exp − rs ds At E ŜT0 At
t Ŝt0
( ) !
T
1
Z
= E exp − rs ds At 1 − exp − λ0t
t 2
c Copyright Eckhard Platen 10 BA - Toronto C8 320
Forward Contracts
!
j
F (t, T ) − STj
Stδ∗ E At =0
STδ∗
forward price
j
S
j
Stδ∗ E ŜTj At t
P (t,T 1
) Ŝ j
E ŜT At if Stj > 0
j t
F (t, T ) = =
Stδ∗ E 1
δ∗ At
0 if Stj = 0
ST
c Copyright Eckhard Platen 10 BA - Toronto C8 321
• MM case
F j (t, T ) = Stj exp{r(T − t)}
standard expression
• MMM case
1 1 1
E ŜTj At = E ŜTj,c At E STj,d At
Ŝtj Ŝtj,c Stj,d
1
= 1 − exp − λjt
2
forward price
n o !!−1
1 j
1 − exp − 2 λt
( Z )
T
F j (t, T ) = Stj 1 0 E exp − rs ds At
1 − exp − 2 λt t
c Copyright Eckhard Platen 10 BA - Toronto C8 322
Asset-or-Nothing Binaries
Ingersoll (2000), Buchen (2004) and Buchen & Konstandatos (2005)
rt = r
c Copyright Eckhard Platen 10 BA - Toronto C8 323
!
STj
Aj,k (t, T, K) = Stδ∗ E 1{S j ≥K} δ∗ At
T
ST
Stj
= E 1{Ŝ j ≥K(S 0 )−1 Ŝ 0 } ŜTj At
Ŝtj T T T
Stj
= E 1{Ŝ j,c ≥g(pk−m −pk−m )Ŝ 0 }
Ŝtj,c T T t T
!pk−m k−m
−pt
n √ o σ j,k T
× exp σ j,k hk−m (T − t) 1 − √ ŜTj,c At
hk−m
c Copyright Eckhard Platen 10 BA - Toronto C8 324
Aj,k (t, T, K) =
∞ √
X k−m (hk (T − t))n n
j,k
o
= exp −h (T − t) exp σ hk−m (T − t)
n=0
n!
!n
σ j,k
Stj
× 1− √ E 1{Ŝ j,c ≥g(n)Ŝ 0 } ŜTj,c At
hk−m Ŝtj,c T T
for all t ∈ [0, T ], where
!−n
K n √ o σ j,k
g(n) = exp − r + σ j,k hk−m (T − t) 1 − √
St0 Stj,d hk−m
for all n ∈ N
c Copyright Eckhard Platen 10 BA - Toronto C8 325
MM case
∞
j,k
X k−m (hk−m (T − t))n
A (t, T, K) = exp −h (T − t)
n=0
n!
!n
n √ o σ j,k
× exp σj,k hk−m (T − t) 1 − √ Stj N (d1 (n))
hk−m
for all t ∈ [0, T ], where
σ j,k
Stj
√ ln 1− √
j,k hk−m 1 0,j 2
ln K
+r + σ hk−m + n T −t
+ 2
σ̂ (T − t)
d1 (n) = √
σ̂ 0,j T −t
c Copyright Eckhard Platen 10 BA - Toronto C8 326
q
σ̂ i,j = |σ i |2 − 2 ̺i,j |σ i | |σ j | + |σ j |2
̺i,j - correlation
c Copyright Eckhard Platen 10 BA - Toronto C8 327
MMM case
Ŝ 0 and Ŝ j,c are independent
∞
X (hk−m (T − t))n
Aj,k (t, T, K) = exp −hk−m (T − t)
n=0
n!
!n
n √ o σ j,k
j,k
× exp σ hk−m (T − t) 1− √
hk−m
ϕ0 (T ) − ϕ0 (t)
1 j
j ′′ j 0
× St G0,4 ; λt , λt − exp − λt
g(n) 2
c Copyright Eckhard Platen 10 BA - Toronto C8 328
G′′
0,4 (x; λ, λ ′
) equals probability P ( Z
Z′
≤ x)
non-central chi-square distributed random variable Z ∼ χ2 (0, λ)
non-central chi-square distributed random variable Z ′ ∼ χ2 (4, λ′ )
c Copyright Eckhard Platen 10 BA - Toronto C8 329
Bond-or-Nothing Binaries
MM case
∞
X (hk−m (T − t))n
B j,k (t, T, K) = exp{−hk−m (T − t)}
n=0
n!
× K exp{−r(T − t)} N (d2 (n))
for all t ∈ [0, T ], where
j,k
!
j √ ln(1− √σ )
St j,k hk−m 1 0,j 2
ln K
+ r+σ hk−m +n T −t
− 2
σ̂ (T − t)
d2 (n) = √
σ̂ 0,j T −t
p
0,j
= d1 (n) − σ̂ T −t
c Copyright Eckhard Platen 10 BA - Toronto C8 330
MMM case
B j,k (t, T, K)
∞
X (hk−m (T − t))n
= exp{−hk−m (T − t)}
n=0
n!
× K exp{−r(T − t)} 1 − G′′
0,4 (ϕ j
(T ) − ϕ j
(t))g(n); λ0
t , λ j
t
c Copyright Eckhard Platen 10 BA - Toronto C8 331
European Call Options
+
j
ST −K
j,k
cT ,K (t) = Stδ∗ E At
STδ∗
!
STj −K
= Stδ∗ E 1{S j ≥K} At
T
STδ∗
= Aj,k (t, T, K) − B j,k (t, T, K)
c Copyright Eckhard Platen 10 BA - Toronto C8 332
MM case
∞
X (hk−m (T − t))n
cj,k
T ,K (t) = exp{−hk−m (T − t)}
n=0
n!
!n
n √ o σ j,k
× exp σ j,k
hk−m (T − t) 1 − √ Stj N (d1 (n))
hk−m
!
− K exp{−r(T − t)}N (d2 (n))
c Copyright Eckhard Platen 10 BA - Toronto C8 333
MMM case
∞
X (hk−m (T − t))n
cj,k
T ,K (t) = exp{−h k−m
(T − t)}
n=0
n!
" !n
n √ o σ j,k
× exp σ j,k hk−m (T − t) 1 − √
hk−m
0 0 !
ϕ (T ) − ϕ (t) 1
× Stj G′′
0,4 ; λjt , λ0t − exp − λjt
g(n) 2
#
− K exp{−r(T − t)} 1 − G′′ j j 0 j
0,4 (ϕ (T ) − ϕ (t))g(n); λt , λt
c Copyright Eckhard Platen 10 BA - Toronto C8 334
Defaultable Zero Coupon Bonds
maturity T
defaults at the first jump time τ1k−m of pk−m
zero recovery
!
1{τ k−m >T }
k−m
P̃ (t, T ) = Stδ∗ E 1
At
STδ∗
!
1
= Stδ∗ E At E 1{τ k−m >T } At
STδ∗ 1
= P (t, T )P (pk−m
T = 0 At )
c Copyright Eckhard Platen 10 BA - Toronto C8 335
• conditional probability of survival
P pk−m
T = 0 At = E 1{pk−m =0} 1{pk−m −pk−m =0} At
t T t
= 1{pk−m =0} P pk−m
T − pk−m
t = 0 At
t
( ) !
Z T
= 1{pk−m =0} E exp − hk−m
s ds At
t
t
c Copyright Eckhard Platen 10 BA - Toronto C8 336
References
Andersen, L. & J. Andreasen (2000). Volatility skews and extensions of the LIBOR market model. Appl. Math. Finance 7, 1–32.
Ansel, J. P. & C. Stricker (1994). Couverture des actifs contingents. Ann. Inst. H. Poincaré Probab. Statist. 30, 303–315.
Becherer, D. (2001). The numeraire portfolio for unbounded semimartingales. Finance Stoch. 5, 327–341.
Beckers, S. (1980). The constant elasticity of variance model and its implications for option pricing. J. Finance 35(3), 661–673.
Bollerslev, T. (1986). Generalized autoregressive conditional heteroskedasticity. J. Econometrics 31, 307–327.
Breiman, L. (1961). Optimal gambling systems for favorable games. In Proceedings of the Fourth Berkeley Symposium on Mathematical Statistics
and Probability, Volume I, pp. 65–78.
Breymann, W., A. Dias, & P. Embrechts (2003). Dependence structures for multivariate high-frequency data in finance. Quant. Finance 3(1), 1–14.
Breymann, W., L. Kelly, & E. Platen (2006). Intraday empirical analysis and modeling of diversified world stock indices. Asia-Pacific Financial
Markets 12(1), 1–28.
Brigo, D. & F. Mercurio (2005). Interest Rate Models - Theory and Practice (2nd ed.). Springer Finance.
Buchen, P. W. (2004). The pricing of dual-expiry exotics. Quant. Finance 4(1), 101–108.
Buchen, P. W. & O. Konstandatos (2005). A new method of pricing lookback options. Math. Finance 15(2), 245–259.
Bühlmann, H. & E. Platen (2003). A discrete time benchmark approach for insurance and finance. ASTIN Bulletin 33(2), 153–172.
CFO-Forum (2008, June). MCEV principles. http://www.cfoforum.nl/pdf/mcev− principles− and− guidance.pdf.
Chesney, M. & L. O. Scott (1989). Pricing European currency options: A comparison of the modified Black-Scholes model and a random variance
model. J. Financial and Quantitative Analysis 24, 267–284.
Cochrane, J. H. (2001). Asset Pricing. Princeton University Press.
Constantinides, G. M. (1992). A theory of the nominal structure of interest rates. Rev. Financial Studies 5, 531–552.
Cox, J. C. (1975). Notes on option pricing I: constant elasticity of variance diffusions. Stanford University, (working paper, unpublished).
Cox, J. C. & S. A. Ross (1976). The valuation of options for alternative stochastic processes. J. Financial Economics 3, 145–166.
CRO-Forum (2008, July). Market value of liabilities for insurance firms: Implementing elements for solvency II.
http://www.croforum.org/publications/20082807− resource/File.ecr?
c Copyright Eckhard Platen 10 BA - Toronto References 337
Davis, M. H. A. (1997). Option pricing in incomplete markets. In M. A. H. Dempster and S. R. Pliska (Eds.), Mathematics of Derivative Securities,
pp. 227–254. Cambridge University Press.
Debreu, G. (1959). Theory of Value. Wiley, New York.
Delbaen, F. & W. Schachermayer (1998). The fundamental theorem of asset pricing for unbounded stochastic processes. Math. Ann. 312, 215–250.
Delbaen, F. & W. Schachermayer (2006). The Mathematics of Arbitrage. Springer Finance. Springer.
Delbaen, F. & H. Shirakawa (2002). A note on option pricing for the constant elasticity of variance model. Asia-Pacific Financial Markets 9(2),
85–99.
Duffie, D. (2001). Dynamic Asset Pricing Theory (3rd ed.). Princeton, University Press.
Engle, R. F. (1982). Autoregressive conditional heteroskedasticity with estimates of the variance of U.K. inflation. Econometrica 50(4), 987–1007.
Engle, R. F. & V. K. Ng (1993). Measuring and testing the impact of news on volatility. J. Finance 48, 1749–1801.
Fergusson, K. & E. Platen (2006). On the distributional characterization of log-returns of a world stock index. Appl. Math. Finance 13(1), 19–38.
Fernholz, E. R. & I. Karatzas (2005). Relative arbitrage in volatility-stabilized markets. Annals of Finance 1(2), 149–177.
Föllmer, H. & M. Schweizer (1991). Hedging of contingent claims under incomplete information. In M. H. A. Davis and R. J. Elliott (Eds.), Applied
Stochastic Analysis, Volume 5 of Stochastics Monogr., pp. 389–414. Gordon and Breach, London/New York.
Frey, R. (1997). Derivative asset analysis in models with level-dependent and stochastic volatility. Mathematics of Finance, Part II. CWI Quar-
terly 10(1), 1–34.
Goll, T. & J. Kallsen (2003). A complete explicit solution to the log-optimal portfolio problem. Ann. Appl. Probab. 13(2), 774–799.
Hakansson, N. H. (1971a). Capital growth and the mean-variance approach to portfolio selection. J. Financial and Quantitative Analysis 6(1),
517–557.
Hakansson, N. H. (1971b). Multi-period mean-variance analysis: towards a general theory of portfolio choice. J. Finance 26, 857–884.
Harrison, J. M. & D. M. Kreps (1979). Martingale and arbitrage in multiperiod securities markets. J. Economic Theory 20, 381–408.
Harrison, J. M. & S. R. Pliska (1983). A stochastic calculus model of continuous trading: Complete markets. Stochastic Process. Appl. 15(3),
313–316.
Heath, D., S. R. Hurst, & E. Platen (2001). Modelling the stochastic dynamics of volatility for equity indices. Asia-Pacific Financial Markets 8,
179–195.
Heath, D. & E. Platen (2005a). Currency derivatives under a minimal market model with random scaling. Int. J. Theor. Appl. Finance 8(8), 1157–
1177.
c Copyright Eckhard Platen 10 BA - Toronto References 338
Heath, D. & E. Platen (2005b). Understanding the implied volatility surface for options on a diversified index. Asia-Pacific Financial Markets 11(1),
55–77.
Heston, S. L. (1993). A closed-form solution for options with stochastic volatility with applications to bond and currency options. Rev. Financial
Studies 6(2), 327–343.
Hofmann, N., E. Platen, & M. Schweizer (1992). Option pricing under incompleteness and stochastic volatility. Math. Finance 2(3), 153–187.
Hull, J. & A. White (1987). The pricing of options on assets with stochastic volatilities. J. Finance 42, 281–300.
Hull, J. & A. White (1988). The use of control variate techniques in option pricing. J. Financial and Quantitative Analysis 23(3), 237–251.
Hulley, H., S. Miller, & E. Platen (2005). Benchmarking and fair pricing applied to two market models. Kyoto Economic Review 74(1), 85–118.
Hulley, H. & E. Platen (2008). Hedging for the long run. Technical report, University of Technology, Sydney. QFRC Research Paper 214.
Hurst, S. R. (1997). On the Stochastic Dynamics of Stock Market Volatility. Ph. D. thesis, ANU, Canberra.
Hurst, S. R. & E. Platen (1997). The marginal distributions of returns and volatility. In Y. Dodge (Ed.), L1 -Statistical Procedures and Related
Topics, Volume 31 of IMS Lecture Notes - Monograph Series, pp. 301–314. Institute of Mathematical Statistics Hayward, California.
Ingersoll, J. E. (1987). Theory of Financial Decision Making. Studies in Financial Economics. Rowman and Littlefield.
Ingersoll, J. E. (2000). Digital contracts: Simple tools for pricing complex derivatives. The Journal of Business 73(1), 67–88.
Johnson, H. & D. Shanno (1987). Option pricing when the variance is changing. J. Financial and Quantitative Analysis 22, 143–151.
Karatzas, I. & C. Kardaras (2007). The numeraire portfolio in semimartingale financial models. Finance Stoch. 11(4), 447–493.
Karatzas, I. & S. E. Shreve (1998). Methods of Mathematical Finance, Volume 39 of Appl. Math. Springer.
Kelly, J. R. (1956). A new interpretation of information rate. Bell Syst. Techn. J. 35, 917–926.
Kessler, M. (1997). Estimation of an ergodic diffusion from discrete observations. Scand. J. Statist. 24, 211–229.
Latané, H. (1959). Criteria for choice among risky ventures. J. Political Economy 38, 145–155.
Lewis, A. L. (2000). Option Valuation Under Stochastic Volatility. Finance Press, Newport Beach.
Lintner, J. (1965). The valuation of risk assets and the selection of risky investments in stock portfolios and capital budgets. Rev. Econom. Statist. 47,
13–37.
Lo, C. F., P. H. Yuen, & C. H. Hui (2000). Constant elasticity of variance option pricing model with time dependent parameters. Int. J. Theor. Appl.
Finance 3(4), 661–674.
c Copyright Eckhard Platen 10 BA - Toronto References 339
Loewenstein, M. & G. A. Willard (2000). Local martingales, arbitrage, and viability: Free snacks and cheap thrills. Econometric Theory 16(1),
135–161.
Long, J. B. (1990). The numeraire portfolio. J. Financial Economics 26, 29–69.
Markowitz, H. (1959). Portfolio Selection: Efficient Diversification of Investment. Wiley, New York.
Markowitz, H. (1976). Investment for the long run: New evidence for an old rule. J. Finance XXXI(5), 1273–1286.
Markowitz, H. & N. Usmen (1996). The likelihood of various stock market return distributions, Part 1: Principles of inference. J. Risk & Uncer-
tainty 13(3), 207–219.
Melino, A. & S. Turnbull (1990). Pricing foreign currency options with stochastic volatility. J. Econometrics 45, 239–265.
Merton, R. C. (1973a). An intertemporal capital asset pricing model. Econometrica 41, 867–888.
Merton, R. C. (1973b). Theory of rational option pricing. Bell J. Econ. Management Sci. 4, 141–183.
Merton, R. C. (1992). Continuous-Time Finance. Blackwell, Oxford.
Milevsky, M. A. & S. E. Posner (2001). The Titanic option: Valuation of the guaranteed minimum death benefit in variable annuities and mutual
funds. J. of Risk and Insurance 68(1), 93–128.
Miller, S. & E. Platen (2005). A two-factor model for low interest rate regimes. Asia-Pacific Financial Markets 11(1), 107–133.
Miller, S. & E. Platen (2008). Analytic pricing of contingent claims under the real world measure. Int. J. Theor. Appl. Finance 11(8), 841–867.
Nelson, D. B. (1990). ARCH models as diffusion approximations. J. Econometrics 45, 7–38.
Platen, E. (1997). A non-linear stochastic volatility model. Technical report, Australian National University, Canberra, Financial Mathematics Re-
search Reports. FMRR 005-97.
Platen, E. (2001). A minimal financial market model. In Trends in Mathematics, pp. 293–301. Birkhäuser.
Platen, E. (2002). Arbitrage in continuous complete markets. Adv. in Appl. Probab. 34(3), 540–558.
Platen, E. (2004a). A benchmark framework for risk management. In Stochastic Processes and Applications to Mathematical Finance, pp. 305–335.
Proceedings of the Ritsumeikan Intern. Symposium: World Scientific.
Platen, E. (2004b). Pricing and hedging for incomplete jump diffusion benchmark models. In Mathematics of Finance, Volume 351 of Contemporary
Mathematics, pp. 287–301. American Mathematical Society.
Platen, E. (2005a). An alternative interest rate term structure model. Int. J. Theor. Appl. Finance 8(6), 717–735.
c Copyright Eckhard Platen 10 BA - Toronto References 340
Platen, E. (2005b). Diversified portfolios with jumps in a benchmark framework. Asia-Pacific Financial Markets 11(1), 1–22.
Platen, E. (2006). A benchmark approach to finance. Math. Finance 16(1), 131–151.
Platen, E. (2008). Law of the minimal price. Technical report, University of Technology, Sydney. QFRC Research Paper 215.
Platen, E. & D. Heath (2006). A Benchmark Approach to Quantitative Finance. Springer Finance. Springer.
Platen, E. & R. Rendek (2008). Empirical evidence on Student-t log-returns of diversified world stock indices. J. of Statistical Theory and Prac-
tice 2(2), 233–251.
Prakasa Rao, B. L. S. (1999). Statistical Inference for Diffusion Type Processes. Arnold Publishers, Oxford University Press.
Revuz, D. & M. Yor (1999). Continuous Martingales and Brownian Motion (3rd ed.). Springer.
Roll, R. (1973). Evidence on the ”Growth-Optimum” model. J. Finance 28(3), 551–566.
Ross, S. A. (1976). The arbitrage theory of capital asset pricing. J. Economic Theory 13, 341–360.
Schroder, M. (1989). Computing the constant elasticity of variance option pricing formula. J. Finance 44(1), 211–219.
Scott, L. O. (1987). Option pricing when the variance changes randomly: Theory, estimation and an application. J. Financial and Quantitative
Analysis 22, 419–438.
Sharpe, W. F. (1964). Capital asset prices: A theory of market equilibrium under conditions of risk. J. Finance 19, 425–442.
Stein, E. M. & J. C. Stein (1991). Stock price distributions with stochastic volatility: An analytic approach. Rev. Financial Studies 4, 727–752.
Thorp, E. O. (1972). Portfolio choice and the Kelly criterion. In Proceedings of the 1971 Business and Economics Section of the American Statistical
Association, Volume 21, pp. 5–224.
Tobin, J. (1958). Liquidity preference as behaviour towards risk. Rev. of Economic Studies 25(2), 65–86.
Verheugen, H. & W. Hines (2008). Valuation methods for life insurers: Conversion or chaos. Milliman Research Reports Nov. 2008.
Wiggins, J. B. (1987). Option values under stochastic volatility. Theory and empirical estimates. J. Financial Economics 19, 351–372.
c Copyright Eckhard Platen 10 BA - Toronto References 341
You might also like
- Anglicanism In Australia A History 2nd Edition Kaye instant download100% (2)Anglicanism In Australia A History 2nd Edition Kaye instant download56 pages
- FSI - CE-560XL - PTM - Pilot Training Manaul Volume 1 XL and100% (2)FSI - CE-560XL - PTM - Pilot Training Manaul Volume 1 XL and402 pages
- Slides Aymeric KALIFE Derivatives As Hedge Instruments-2022No ratings yetSlides Aymeric KALIFE Derivatives As Hedge Instruments-2022234 pages
- Modelling The Price of A Credit Default Swap: Agnieszka ZalewskaNo ratings yetModelling The Price of A Credit Default Swap: Agnieszka Zalewska29 pages
- Numerical Solution of Stochastic Differential Equations With Jumps in FinanceNo ratings yetNumerical Solution of Stochastic Differential Equations With Jumps in Finance86 pages
- Mathematics of Finance: S USD Euro R R e Fe e F eNo ratings yetMathematics of Finance: S USD Euro R R e Fe e F e6 pages
- Central Limit Theorem For The Realized Volatility Based On Tick Time Sampling - SlidesNo ratings yetCentral Limit Theorem For The Realized Volatility Based On Tick Time Sampling - Slides21 pages
- American Option Pricing in A Tick-Calibration in A ClickNo ratings yetAmerican Option Pricing in A Tick-Calibration in A Click38 pages
- Estimating Discount Functions With Consumption Choices Over The LifecycleNo ratings yetEstimating Discount Functions With Consumption Choices Over The Lifecycle27 pages
- STCDO Model Calibration and Hedge Performance: QMF Sydney - December 13, 2006No ratings yetSTCDO Model Calibration and Hedge Performance: QMF Sydney - December 13, 200635 pages
- Mixing Models To Capture Stock Price Volatility Markov Lecture INFORMS Annual Meeting October 12-15, 2008 Washington, DCNo ratings yetMixing Models To Capture Stock Price Volatility Markov Lecture INFORMS Annual Meeting October 12-15, 2008 Washington, DC35 pages
- From Navier Stokes To Black Scholes - Numerical Methods in Computational FinanceNo ratings yetFrom Navier Stokes To Black Scholes - Numerical Methods in Computational Finance13 pages
- Examination: Subject CT6 Statistical Methods Core TechnicalNo ratings yetExamination: Subject CT6 Statistical Methods Core Technical13 pages
- Calibration and Data Analysis: in This Lecture. .No ratings yetCalibration and Data Analysis: in This Lecture. .73 pages
- Mathematical Modelling and Financial Engineering in The Worlds Stock MarketsNo ratings yetMathematical Modelling and Financial Engineering in The Worlds Stock Markets7 pages
- Calculating_Delta_Risks_and_Hedges_via_WavesNo ratings yetCalculating_Delta_Risks_and_Hedges_via_Waves7 pages
- CA-Inter-FM-SM-A-MTP-2-Jan-2025-Exam-castudynotes-comNo ratings yetCA-Inter-FM-SM-A-MTP-2-Jan-2025-Exam-castudynotes-com15 pages
- Moody KMV Explained UT Berlin Tetereva - 05042012 PDFNo ratings yetMoody KMV Explained UT Berlin Tetereva - 05042012 PDF32 pages
- SIQ3004 Mathematics of Financial Derivatives: Chapter 7: The Black-Scholes FormulaNo ratings yetSIQ3004 Mathematics of Financial Derivatives: Chapter 7: The Black-Scholes Formula38 pages
- Review: Arbitrage-Free Pricing and Stochastic Calculus: Leonid KoganNo ratings yetReview: Arbitrage-Free Pricing and Stochastic Calculus: Leonid Kogan16 pages
- 03fd39b380d3fffe6a809a75367a90e1_MIT15_450F10_assn02No ratings yet03fd39b380d3fffe6a809a75367a90e1_MIT15_450F10_assn023 pages
- Financial Engineering Homework 1 - Solution: MRM 8610, Spring 20160% (1)Financial Engineering Homework 1 - Solution: MRM 8610, Spring 20162 pages
- De Moiver's Theorem (Trigonometry) Mathematics Question BankFrom EverandDe Moiver's Theorem (Trigonometry) Mathematics Question BankNo ratings yet
- Dating Advice - A Practical, Modern GuideNo ratings yetDating Advice - A Practical, Modern Guide13 pages
- Intercooler Assembly & Tubing-Model 2340No ratings yetIntercooler Assembly & Tubing-Model 23402 pages
- Seismic Provisions in The Saudi Building CodeNo ratings yetSeismic Provisions in The Saudi Building Code13 pages
- The Fast Medium Wave Krelus Infrar Ed-HeatersNo ratings yetThe Fast Medium Wave Krelus Infrar Ed-Heaters8 pages
- CUSTOMER - SWITCHING - IN - TELECOM - INDUSTRY (1) Group 4No ratings yetCUSTOMER - SWITCHING - IN - TELECOM - INDUSTRY (1) Group 413 pages
- AD-120ES Parts Manual Phase 7: 88 Currant Road Fall River, MA 02720-4781 Telephone: (508) 678-9000 / Fax: (508) 678-9447No ratings yetAD-120ES Parts Manual Phase 7: 88 Currant Road Fall River, MA 02720-4781 Telephone: (508) 678-9000 / Fax: (508) 678-944760 pages
- COMP313 - Computer Systems IP Routing NotesNo ratings yetCOMP313 - Computer Systems IP Routing Notes13 pages
- Kami Export - FURTHER PRACTICE 5.2-USE OF ENGLISHNo ratings yetKami Export - FURTHER PRACTICE 5.2-USE OF ENGLISH3 pages
- Arcadis Quarterly Construction Cost Report 2023 1st QTRNo ratings yetArcadis Quarterly Construction Cost Report 2023 1st QTR18 pages
