25.06.2013 Views

statistique, théorie et gestion de portefeuille - Docs at ISFA

statistique, théorie et gestion de portefeuille - Docs at ISFA

statistique, théorie et gestion de portefeuille - Docs at ISFA

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

274 9. Mesure <strong>de</strong> la dépendance extrême entre <strong>de</strong>ux actifs financiers<br />

where ɛ is a random variable in<strong>de</strong>pen<strong>de</strong>nt of Y and α a non random positive coefficient. Assume th<strong>at</strong> Y and<br />

ɛ have a Stu<strong>de</strong>nt’s distribution with <strong>de</strong>nsity:<br />

Cν<br />

PY (y) = <br />

1 + y2<br />

ν<br />

Pɛ(ɛ) =<br />

Cν<br />

<br />

σ 1 + ɛ2<br />

ν σ2 ν+1<br />

2<br />

ν+1<br />

2<br />

, (E.97)<br />

. (E.98)<br />

We first give a general expression for the probability for X to be larger than F −1<br />

X (u) knowing th<strong>at</strong> Y is<br />

(u) :<br />

larger than F −1<br />

Y<br />

LEMMA 1<br />

The probability th<strong>at</strong> X is larger than F −1<br />

X<br />

with<br />

Pr[X > F −1<br />

−1<br />

X (u)|Y > FY (u)] = ¯ Fɛ(η) + α<br />

∞<br />

1 − u F −1<br />

Y (u)<br />

−1<br />

(u) knowing th<strong>at</strong> Y is larger than F (u) is given by :<br />

η = F −1<br />

X<br />

Y<br />

dy ¯ FY (y) · Pɛ[αF −1<br />

Y (u) + η − αy] , (E.99)<br />

−1<br />

(u) − αF (u). (E.100)<br />

The proof of this lemma relies on a simple integr<strong>at</strong>ion by part and a change of variable, which are d<strong>et</strong>ailed<br />

in appendix E.1.<br />

Introducing the not<strong>at</strong>ion<br />

we can show th<strong>at</strong><br />

η = α 1 +<br />

Y<br />

˜Yu = F −1<br />

Y (u) , (E.101)<br />

<br />

σ<br />

<br />

ν1/ν − 1 ˜Yu + O(<br />

α<br />

˜ Y −1<br />

u ), (E.102)<br />

which allows us to conclu<strong>de</strong> th<strong>at</strong> η goes to infinity as u goes to 1 (see appendix E.2 for the <strong>de</strong>riv<strong>at</strong>ion of this<br />

result). Thus, ¯ Fɛ(η) goes to zero as u goes to 1 and<br />

∞ α<br />

λ = lim dy<br />

u→1 1 − u<br />

¯ FY (y) · Pɛ(α ˜ Yu + η − αy) . (E.103)<br />

Now, using the following result :<br />

LEMMA 2<br />

Assuming ν > 0 and x0 > 1,<br />

1<br />

lim<br />

ɛ→0 ɛ<br />

∞<br />

1<br />

˜Yu<br />

dx 1<br />

x ν<br />

<br />

Cν<br />

1 + x−x0<br />

ɛ<br />

2 ν+1<br />

2<br />

whose proof is given in appendix E.3, it is straigthforward to show th<strong>at</strong><br />

The final steps of this calcul<strong>at</strong>ion are given in appendix E.4.<br />

= 1<br />

xν , (E.104)<br />

0<br />

1<br />

λ =<br />

1 + <br />

σ ν . (E.105)<br />

α<br />

36

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!