Quant Port Notes

Quantitative Methods
in Portfolio Management
C. Wagner
WS 2010/2011
Mathematisches Institut,
Ludwig-Maximilians-Universit
at M
unchen
Quantitative Methods in Portfolio Management Mathematisches Institut, LMU M
unchen
Markowitz
- efficient
Return
Efficient Frontier
Alpha
CAPM
BARRA
Sharpe
Shortfall
Information Ratio
C. Wagner 2 WS 2010/2011
Quantitative Methods in Portfolio Management Contents Mathematisches Institut, LMU M
unchen
Contents
1 Introduction 5
2 Utility Theory 8
3 Modeling the Market 16
4 Estimating the Distribution of Market Invariants 40
5 Evaluating Allocations 41
6 Optimizing Allocations 42
7 Estimating the Distribution of Market Invariants with Estimation Risk 43
8 Evaluating Allocations under Estimation Risk 44
9 Optimizing Allocations under Estimation Risk 45
C. Wagner 3 WS 2010/2011
Quantitative Methods in Portfolio Management Contents Mathematisches Institut, LMU M
unchen
Literature
introductory
Modern Portfolio Theory and Investment Analysis; Elton, Gruber, Brown, Goetzmann; Wi-
ley
Portfoliomanagement; Breuer, Guertler, Schumacher; Gabler
quantitative
Risk and Asset Allocation; Meucci; Springer
Quantitative Equity Portfolio Management; Qian, Hua, Sorensen; CRC
Robust Portfolio Optimization and Management; Fabozzi, Kolm, Pachamanova, Focardi;

Wiley
C. Wagner 4 WS 2010/2011
Quantitative Methods in Portfolio Management 1 Introduction Mathematisches Institut, LMU M
unchen
1 Introduction
Investor Choice Under Certainty
Investor will receive EUR 10000 with certainty in each of two periods
Only investment vailable is savings account (yield 5%)
Investor can borrow money at 5%
How much should the investor save or spend in each period?
Separate problem into two steps:
1. Specify options
2. Specify how to choose between options
C. Wagner 5 WS 2010/2011
unchen
Opportunity Set:
A save nothing, spend all when received, (10000, 10000)
B save first period and consume all in the second (0, 10000 (1 + 0.05) + 10000)
C consume all in the first, i.e borrow the maximum in from the second in the first period (10000 +
10000/(1 + 0.05), 0)
xi income in period i, yi consumption in period i
y2 = x2 + (x1 y1) 1.05
Indifference Curve:
Iso-Happiness Curve (see graph):
assumption: each additional euro of consumption forgone in periode 1 requires greater consumption
in period 2
ordering due to investor prefers more to less
C. Wagner 6 WS 2010/2011
unchen
Solution:
Opportunity set is tangent to indifference set
C. Wagner 7 WS 2010/2011
Quantitative Methods in Portfolio Management 2 Utility Theory Mathematisches Institut, LMU M
unchen
2 Utility Theory
Use utility function to formalise investors preferences to arrive at optimal portfolio
Base Modell
Investor has initial wealth W0 at t = 0 and an investment universe of n + 1 assets (n risky assets,
one riskless asset)
Investment horizon t = 1, short-selling, uncertain returns ri (rv) for risky assets i = 1, . . . , n

and deterministic r0 for bank account.
P
Describe portfolio through asset weights, i.e. P = (x0, x1, . . . , xn), W0 = W0 i xi
Uncertain wealth (i.e. rv) W1 at t = 1, W1P =

P
i xi W0 (1 + ri)
Assign preference through utility function to each possible opportunity set, U (W1P ), and prob-
ability to arrive at expected utility E[U (W1P )].
optimization problem: maxx0,...,xn E[U (W1P )]
C. Wagner 8 WS 2010/2011
unchen
Properties of Utility Function

determined only up to positiv linear transformations (ranking)
investor prefers more to less, W P < W Q then U (W P ) < U (W Q), U strictly increasing, if
differentiable then U 0 > 0
risk appetite: for W0 = E[W1]
risk averse: E[U (W0)] > E[U (W1)], U concave (Jensen)

risk neutral: E[U (W0)] = E[U (W1)], U linear
risk seeking: E[U (W0)] < E[U (W1)] U convex
C. Wagner 9 WS 2010/2011
unchen
Some Distributions
Uniform Distribution
X U (E,), E, elipsoid
( N2 + 1) 0
f (x) = N/2 1/2 1E, (x), U,() = ei ( 0)
||
Normal Distrubution
X N (, )
1 0 1 (x) 0 1 0
f (x) = (2)N/2||1/2e 2 (x) , () = ei 2
Student-t Distrubution
Z N (, ), W 2()
r

X Z St(, , )
W
C. Wagner 10 WS 2010/2011
unchen
Cauchy Distrubution
X Ca(, ) St(1, , )
Lognormal Distrubution
X LogN (, ) X = eY , Y N (, )
C. Wagner 11 WS 2010/2011
unchen
Distribution Classes
Elliptical Distribution
Definition: rv Y = (Y1, . . . , Yn) has a sperical distribution if, for every orthogonal matrix U
d
UY = Y
0
Properties: Y spherical function g (generator) such that Y (t) = E[eit Y ] = n(t0t)
Generator as function of a scalar variable uniquely describes sperical distribution
Y Sn()
equivalent representation:
Y = RU , where R = kY k is norm (i.e. univariate) and U = Y /kY k
Y Sn() R and U are independent rvs and U is uniformly distributed on the surface of the
unit ball
Definition: X has an elliptical distribution if
d
X = + AY
where Y Sm() and A Rnm, Rn. X El(, , ), = AA0.
C. Wagner 12 WS 2010/2011
unchen
positive definite, then isoprobability contours are surfaces of centered ellipsoides

More Properties:
affine transformation: BX + b EL()
marginal distributions: EL()
conditional distribution: EL()
convolution with same dispertion matrix : EL()
Examples: Uniform, Normal, Student-t, Cauchy highly symmetric and analytically tractable, yet
quite flexible
Stable Distribution
Definition: X, X1, X2 iid rv. X is called stable if for all non-negative c1, c2 and appropriate
numbers a = a(c1, c2), b = b(c1, c2) the following holds:
d
c1X1 + c2X2 = a + bX
i.e. closed under linear combinations
C. Wagner 13 WS 2010/2011
unchen
symmetric--stable (one dimension) iff
X (t) = E[eitX ] = exp{it c|t| }
location, c scaling, tail thickness

symmetric--stable (multivariate) iff
Z
it0 X it0
X (t) = E[e ]=e exp{ |t0s| m(s)ds}
R
function m is a symmetric measure, m(s) = m(s) for all s Rn and
m(s) 0 for all s such that s0s 6= 1
X SS(, , m)
Examples: Normal, Cauchy

Counterexamples: Lognormal, Student-t
C. Wagner 14 WS 2010/2011
unchen
Normal distribution as symmetric-alpha-stable
X N (, )
spectral decomposition of : = E1/21/2E
define n vectors {v (1), . . . , v (n)} = E1/2
define measure as n
1X
m = ((v (i) + (v (i)))
4 i=1
Remark: stability additvity, but reverse is not true in general, e.g. Wishart dist.
Infinitely Divisible Distributions

X is infinitely divisible if, for any integer M we can decompose it in law
d
X = Y 1 + + Y M
where (Y i)i=1,...,M are iid rvs with possibly different common distributions for different M . Exam-
ples: Normal, Lognormal, Chi2 Counterexamples: Wishart
C. Wagner 15 WS 2010/2011
Quantitative Methods in Portfolio Management 3 Modeling the Market Mathematisches Institut, LMU M
unchen
3 Modeling the Market

Market for an investor is represented by an N-dimensional price vector of traded securities, P t:
Investment decision (allocation) at T
Investment horizon
P T + N-dimensional random variable
Modeling the market means modeling P T + :
1. modeling market invariants
2. determining the dsitribution of market invariants
3. projecting invariants into the future T +
4. mapping of invariants to market prices
dimension of randomness numbers of securities dimension reduction
C. Wagner 16 WS 2010/2011
unchen
Market Invariants
Dt0, = {t0, t0 + , t0 + 2
, . . . } set of equally spaced observation dates
random variables Xt, t Dt, are called invariant if rv are iid and time homogeneous
simple tests: check histograms of two non-overlapping subsets of observations, scatter-plot of values
vs. lagged values
Equities
Pt, t Dt0, equally-spaced stock price observations

equity prices are not market invariants (exponential growth)
Total return
Pt
Ht, =
Pt
is a market invariant
g any function, then if Xt is invariant g(Xt) is also an invariant
C. Wagner 17 WS 2010/2011
unchen
hence
linear return Lt, = Ht, 1 and log-return Ct, = ln (Ht, )
are also market invariants.
equity invariants: compound returns C
compund returns can be easily projected to any horizon, distribution approximately symmetric
Example:
continuous-time finance, Black/Scholes, Merton

Pt
Ct, = ln N (, 2).
Pt
total return Ht, LogN(, 2).

Other Choices:
multivariate case
Ct, EL(, , g) or Ct, SS(, , m)
C. Wagner 18 WS 2010/2011
unchen
Fixed-Income Market
(E)
zero-coupon bonds as building blocks, Zt , E maturity
(E)
normalization ZE 1
consider set of bond prices

(E)
Zt , t Dt0,
pull-to-par effect bond prices are not market invariants
consider set of non-overlapping total returns
(E)
(E) Zt
Ht, (E)
, t Dt0,
Zt
pull-to-par also breaks time homogeneity of total return total returns are not market invariants
consider total return of bonds with same time to maturity
(t+)
() Zt
Rt, (t+
)
, t Dt0,
Zt
ratio of prices of two different securities
C. Wagner 19 WS 2010/2011
unchen
()
test shows that Rt, is acceptable as market invariant, hence also every function g of R.
define yield to maturity as (annualized return of bond)
() 1 (t+)
Yt ln Zt

consider now changes in yield to maturity

() () () 1 ()
Xt, Yt Yt = ln Rt,

fixed-income invariants: changes in yield to maturity X
invariant is specific to a given sector of the yield curve

Examples:
() () ()
Xt, N (, ), Xt, El(, , g), Xt, SS(, , m)
C. Wagner 20 WS 2010/2011
unchen
Derivatives
vanilla european call option

(K,E) BS (E) (K,E)
Ct =C E t, K, Ut, Zt , t
(K,E)
boundary condition CE = max (UE K, 0)
E expiry date, K strike
market variables:
(E) (K,E)
Ut underlying, Zt zero bond, t implied percentage volatility (vol surface)
for zero bond Z and underlying U market invariants are know

what about implied vol?
consider at-the-money-forward (ATMF) implied vol, i.e. implied vol at strike equals forward price
(Kt ,E) Ut
t , t Dt0, where Kt (E)
Zt
implied vol is not a market invariant as expiry convergence breack time-homogeneity
C. Wagner 21 WS 2010/2011
unchen
eliminate expiry date dependency through considering set of implied vols (rolling ATMF vols)
(Kt ,t+)
t , t Dt0,
(K ,t+)
r
(Kt ,t+) 2 Ct t
t
Ut
sill has time dependence
consider changes in ATMF implied vols
() (Kt ,t+) (K ,t
+)
Xt, = t tt , t Dt0,
derivatives invariants: changes in roll. ATMF implied vol
distribution of changes in roll.ATMF implied vol is symmetrical, hence modeling as

() () ()
Xt, N (, ), Xt, El(, , g), Xt, SS(, , m)
C. Wagner 22 WS 2010/2011
unchen
Projection of the Invariants to the Investment Horizon

invariants Xt, relative to estimation interval
representation of distribution in form of probability density fXt, or characteristic function Xt,
in general, investment horizon is different than estimation interval
estimation interval t
FX
T+t,t
FX
T+t,t
time series analysis investment investment horizon t

decision T
Figure 1: asset swap.
C. Wagner 23 WS 2010/2011
unchen
Projection to investment horizon:

determine distribution of XT +, , i.e. fXT +, or XT +, , from estimated distribution
assume is an integer multiple of
invariants are additive, hence
XT +, = XT +, + XT + , + + XT + ,
since invariants are in the form if differences:
equity return: Xt, = ln(Pt) ln(Pt )

yield to maturity: Xt, = Yt Yt
ATFM impl. vol: Xt, = t t
XT +n , are invariants wrt non-overlapping intervalls, i.e. iid.

Projection via convolution:
iX iX iid /
+X ++X

XT +, = E e T +, = E e T +, T + , T + =
, XT +,
representation involving the density fXT +, can be obtained via Fourier transformation
X = F[fX ], fx = F 1[X ]
C. Wagner 24 WS 2010/2011
unchen
Example: ! !
Ct, ln Pt ln Pt
Xt,
= N (, )
Xt,
Yt
Yt
Characteristic function:
0 1 0
Xt, () = ei 2
/ 0 1 0 / 0 1 0
XT +, = XT +, = ei 2 = ei 2

= XT +, N ( , )

Note: normal dist is infinitely divisible, hence /
need not to be an integer
Furthermore for moments:
r

E[XT +, ] = E[Xt, ], Cov[XT +, ] = Cov[Xt, ], Std[XT +, ] = Std[Xt, ]

Remarks:
- simplicity of projection formula due to specific formulation of market invariants
- projection formula hides estimation risk, distribution at horizon can only be estimated (estimation
error)
C. Wagner 25 WS 2010/2011
unchen
From Invariants to Market Prices

how to recover prices from invariants?
market prices of securities at horizon T + are functions of the the investment-horizon invariants
P T + = g(X T +, )
Equities
PT + = PT eX
(see choice of invariants, i.e. compound return)
Fixed-Income
(E) (E ) X (ET ) (ET )
ZT + = ZT e
in general (equities and fixed-income)
P = eY ,with Y + diag()X

ln(P ), if stock 1, if stock
T
n , n
ln(Z (E )) if bond. (E T ) if bond.
T
C. Wagner 26 WS 2010/2011
unchen
Y affine transformation of X distribution of Y , e.g. as characteristic function

0
Y () = ei X (diag())
Example:
two-security market (stock, bond), maturity E = T + +
! ! !
PT + +diag()X ln(PT ) 1
P = (E) = e , = (T +) , = .
ZT + ln(ZT )
characteristic function, X multi-normal
0 1 0
Y () = ei [+ diag()] 2 diag()diag()
Y multi-normal
Y N + diag(), diag()diag()

P log-normal
in most cases not possible to get distribution of future prices in closed form
usually sufficient to work with moments (Taylor)
E[Pn] = en X (n)
C. Wagner 27 WS 2010/2011
unchen
Example: simple stock

+ 2
E[PT + ] = PT e C 2
Derivatives
Derivative price is a nonlinear function of several investment-horizon invariants (e.g. call option)

(K,E) BS (E) (K,E)
P = g(X), e.g. CT + = C E T , K, UT + , ZT + , T +
with
(E) (E ) X2 (ET ) (K,E) (K,E )
UT + = UT eX1 , ZT + = ZT e , T + = T + X3
distribution assumptions for X1, X2, X3, but in general no closed form distribution for P , C.
= Taylor expansion of P :
1
P = g(m) + (X m)0xg|x=m + (X m)0xxg|x=m(X m) + . . .
2
1st order: delta-vega, duration
2nd order: gamma, convexity
C. Wagner 28 WS 2010/2011
unchen
Dimension Reduction
P T + = g(X T +, )
market includes a large number of securities, i.e. market invariant-vector X t, has a large dimension
dimension reduction to a vector F of few common factors
X t, h(F t, ) + U t,
with K = dim(F ) N = dim(X)

If X represents market invariant F , U must be invariants too.
common factors F should be responsible for most of the randomness, U should only be a residual,
h(X) X
X
measure goodness of approximation with generalized r-squared :

0

X) 1 E (X X) (X X)
R2(X,
tr(Cov(X))
C. Wagner 29 WS 2010/2011
unchen
restrict to linear factor model

X BF + U
K K-matrix B is called factor loadings
Ideally, F and U shouldbe independent variables, but too restrictive for pratical purposes, hence
impose only
Corr(F , U ) = 0
- explicit factor model: common factors are measurable market variable

- hidden factor model: common factors are synthetic variables
Explicite Factors
factor loadings for linear regression solve
B r arg max R2(X, BF )

B
from M E[(X BF )(X BF )0] and M/Bij = (0)ij follows
B r = E[XF 0]E[F F 0]1
C. Wagner 30 WS 2010/2011
unchen
which yields
r BrF ,
X r
Ur X X
In general, residuals U do not have zero expectation and are correlated with F (unless E[F ] = 0).
Enhance linear model with constant factor
X a + BF + U
minimizing M = E[(X (a + BF ))(X (a + BF ))0] yields
r E[X]Cov[X, F ]Cov[F ]1(F E[F ])

X
perturbartions U now have zero expectation and are uncorelated with F .
quality of regression:
- adding factors trivially improves quality but number should be kept at a minimum
- factors should be chosen as diversified as possible (avoid collinearity)
C. Wagner 31 WS 2010/2011
unchen
Hidden Factors
factors are not market invariants (not observable)
X q + BF (X) + U
Principal Component Analysis (PCA)

assume that hidden factors are affine transformations of invariants:
F d + A0X,
d is K-dim vector, A is K N -dim matrix

recovered invariants are affine transfomation of original invariants
m + BA0X,
X with m = q + Bd
PCA solution from

(B, A, m) arg max R2(X, m + BA0X)
B,A,m
Impose as additional condition E[F ] = 0
C. Wagner 32 WS 2010/2011
unchen
For this, consider spectral decomposition of covariance matrix
Cov[X] = EE 0
= diag(1, . . . , N ) diag matrix of decreasing positive eigenvalues,

eigenvectors E = (e1, . . . , eN ), EE 0 = 1N
one hidden factor, K = 1:

guess
F = (e1)0X,
i.e. orthogonal projection of X onto the direction of first eigenvector
recovered invariant is
= m + e1(e1)0X
X
= E[X]
impose E[X]
m = 1N e1(e1)0 E[X]

satisfying E[F ] = 0 yields

F = (e1)0X (e1)0E[X]
C. Wagner 33 WS 2010/2011
unchen
K hidden factors:

consider N K matrix E K = e1, . . . , eK
solution to PCA problem is
(B, A, m) = (E K , E k , (1N E K E 0K )E[X])
represent orthogonal projection of original invariants onto hyperplane spanned by the K longest
principal axes (i.e. contains the maximum information)
hidden factors:
F = E 0K (X E[X])
PCA-invariants:
= E[X] + E K E 0 (X E[X])
X K
X have zero expectation and zero correlation with factors F

residuals U = X
E[U ] = 0, Corr[U , F ] = 0
quality of approximation depends on number of hidden factors, with

PK
X) = Pn=1 n
R2(X, N
n=1 n
C. Wagner 34 WS 2010/2011
unchen
Kth eigenvalue is variance of the Kth hidden factor
V[Fn] = (en)0EE 0en = n
n-th eigenvalue is contribution to the total recovered randomness
Explicit vs. Hidden Factors
explicit factors models are interpretable, hidden factor models tend to have a higher explanatory
power
PCA: recovered invariants represent projections of the original invariant onto the K-th dimensional
hyperplane of maximum randmoness spanned by the first K principal axes
Explicit Regression: X = XX 1,...,K ; recovered invariants represent projections of the original

K+1,...,N
invariants onto the plane spanned by the K reference invariants
C. Wagner 35 WS 2010/2011
unchen
Examples
linear stock returns and index return (explicit)

Ptn Mt
Lnt, = n 1, n = 1, . . . , N, Ft, = 1
Pt Mt
Cov(Lnt, , Ft, )
n
L = E[Lnt, ] + n(Ft, E(Ft, )) with =
t,
V(Ft, )
suppose additinal constraint holds:
!
1
E[Lnt, ] = E[Ft, ] + (1 )Rt, with Rt, = (t)
1
Zt
n = Rt, + (Ft, Rt, ),
L CAPM
t,
market-size explicit factors:

(i) broad stock index return
(ii) difference in return between small-cap index and large-cap index (SmB)
(iii) difference in return between book-to-marlet value index and small-book-to-market value index
(HmL)
C. Wagner 36 WS 2010/2011
unchen
Case Study: Modeling the Swap Market

Swap:
plain vanilla interest rate swap (IRS), payer forward start (PFS): committment initiated at t0 to
exchange payments between two different legs, starting from a future time at times t1, . . . , tm,
(eight-year swap two years forward).
fixed leg pays
N iK,
N nominal, K fixed rate, i day-count fraction
floating leg pays
N iL(ti1, ti),
L(ti1, ti) floating rate between ti1 ti reset at ti1 paid in arrears.
discounted payoff of PFS at t < t1 is
m
X
D(t, ti)N i(L(ti1, ti) K)
i=1
C. Wagner 37 WS 2010/2011
unchen
present value of PFS at t:

m m
(t ) (t ) (t ) (ti )
X X
P V (P F S)t = N iZt i (F (t; ti1, ti) K) = N Zt 1 N Zt m NK iZt
i=1 i=1
(t )

1 Zt i1
F (t; ti1, ti) := i (t ) 1 forward rate
Zt i
swap market is completely priced by the set of zero coupon bond prices at all maturities
market invariants for FI:

changes in yield-to-maturity
() () () () 1 (t+)
Xt, = Yt Yt with yield curve 7 Yt := ln Zt

(E)
data set of zero-coupon bond prices Zt , = one week
C. Wagner 38 WS 2010/2011
unchen
dimension reduction:
convariance matrix h i
(v) (v+p)
C(v, p) = Cov X , X
properties
C(v, p + dt) C(v + dt, p) smooth
C(v, 0) C(v + , 0) diagonal elements are similar
C(v, p) C(v + , p)
C(v, p) h(p) approximate structure
with h(p) = h(p)
C. Wagner 39 WS 2010/2011
Quantitative Methods in Portfolio Management 4 Estimating the Distribution of Market Invariants Mathematisches Institut, LMU M
unchen
4 Estimating the Distribution of Market Invariants
C. Wagner 40 WS 2010/2011
Quantitative Methods in Portfolio Management 5 Evaluating Allocations Mathematisches Institut, LMU M
unchen
5 Evaluating Allocations
C. Wagner 41 WS 2010/2011
Quantitative Methods in Portfolio Management 6 Optimizing Allocations Mathematisches Institut, LMU M
unchen
6 Optimizing Allocations
C. Wagner 42 WS 2010/2011
Quantitative Methods in Portfolio Management 7 Estimating the Distribution of Market Invariants with Estimation
Mathematisches
Risk Institut, LMU M
unchen
7 Estimating the Distribution of Market Invariants

with Estimation Risk
C. Wagner 43 WS 2010/2011
Quantitative Methods in Portfolio Management 8 Evaluating Allocations under Estimation Risk Mathematisches Institut, LMU M
unchen
8 Evaluating Allocations under Estimation Risk
C. Wagner 44 WS 2010/2011
Quantitative Methods in Portfolio Management 9 Optimizing Allocations under Estimation Risk Mathematisches Institut, LMU M
unchen
9 Optimizing Allocations under Estimation Risk
C. Wagner 45 WS 2010/2011

Quant Port Notes

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Quant Port Notes

Uploaded by

Copyright:

Available Formats

Quantitative Methods

3 Modeling the Market 16

4 Estimating the Distribution of Market Invariants 40

7 Estimating the Distribution of Market Invariants with Estimation Risk 43

8 Evaluating Allocations under Estimation Risk 44

9 Optimizing Allocations under Estimation Risk 45

Portfoliomanagement; Breuer, Guertler, Schumacher; Gabler

Risk and Asset Allocation; Meucci; Springer

Quantitative Equity Portfolio Management; Qian, Hua, Sorensen; CRC

Robust Portfolio Optimization and Management; Fabozzi, Kolm, Pachamanova, Focardi;

Only investment vailable is savings account (yield 5%)

Investor can borrow money at 5%

How much should the investor save or spend in each period?

Separate problem into two steps:

2. Specify how to choose between options

A save nothing, spend all when received, (10000, 10000)

xi income in period i, yi consumption in period i

y2 = x2 + (x1 y1) 1.05

Investment horizon t = 1, short-selling, uncertain returns ri (rv) for risky assets i = 1, . . . , n

Uncertain wealth (i.e. rv) W1 at t = 1, W1P =

optimization problem: maxx0,...,xn E[U (W1P )]

Properties of Utility Function

risk appetite: for W0 = E[W1]

risk averse: E[U (W0)] > E[U (W1)], U concave (Jensen)

where Y Sm() and A Rnm, Rn. X El(, , ), = AA0.

positive definite, then isoprobability contours are surfaces of centered ellipsoides

i.e. closed under linear combinations

symmetric--stable (one dimension) iff

X (t) = E[eitX ] = exp{it c|t| }

location, c scaling, tail thickness

m(s) 0 for all s such that s0s 6= 1

Examples: Normal, Cauchy

Normal distribution as symmetric-alpha-stable

Infinitely Divisible Distributions

3 Modeling the Market

Investment decision (allocation) at T

P T + N-dimensional random variable

Modeling the market means modeling P T + :

1. modeling market invariants

2. determining the dsitribution of market invariants

3. projecting invariants into the future T +

4. mapping of invariants to market prices

dimension of randomness  numbers of securities dimension reduction

Pt, t Dt0, equally-spaced stock price observations

equity invariants: compound returns C

total return Ht, LogN(, 2).

consider set of bond prices

consider now changes in yield to maturity

fixed-income invariants: changes in yield to maturity X

invariant is specific to a given sector of the yield curve

vanilla european call option

for zero bond Z and underlying U market invariants are know

derivatives invariants: changes in roll. ATMF implied vol

distribution of changes in roll.ATMF implied vol is symmetrical, hence modeling as

Projection of the Invariants to the Investment Horizon

time series analysis investment investment horizon t

Figure 1: asset swap.

Projection to investment horizon:

since invariants are in the form if differences:

equity return: Xt, = ln(Pt) ln(Pt )

XT +n , are invariants wrt non-overlapping intervalls, i.e. iid.

From Invariants to Market Prices

Y affine transformation of X distribution of Y , e.g. as characteristic function

Example: simple stock

dimension of randomness numbers of securities dimension reduction

with K = dim(F ) N = dim(X)