Université Paris 1 Panthéon Sorbonne
MAEE, MMEF, QEM, IMAEF, 2022-2023
Retake Exam: Portfolio Theory and Asset Pricing
21th June 2023
" Documents, calculators and cell phone are prohibited.
" The duration of the exam 2h00.
" It is forbidden to leave without returning the coPy of your
exam.
" The exam iscomposed of 3 exercices which can be treated independently. In an exercise,
you can use the results from the previous questions. We freely use the notations introduced
in the course.
Exercice 1(Questions on the lectures)
1. An investor is said to satisfy the mean-variance preference assumption if for his utility
function U, it holds
EÊU(X"R)] -f(E;X"R), o(x"R)
where X is the vector of his position, R is the random vector of assets return and o(X"R)
stands for the standard deviation ofX"R.
What are the properties that the functionf must satisfy?
2. If 2 is the covariance matrix of the
random vector R, prove that
Var(X"R) = XEX.
3. For a given loss L and confidence level a e (0,1), recall the
at level a of L denoted by VaR.(L) and the Expected definitions
of the Value-at-Risk
by ES, (L). Shortfall at level a of L denoted
4. What are the four properties a risk measure p must satisfy to be called a coherent risk
measure?
5. Does the Value-at-Risk satisfy these four properties? Why?
Exercice 2
In this exercise, all investors are assumed to have
consists of N risky assets. We denote by R the random mean-variance preferences. The market
and by its covariance matrix. The matrix is
vector of the returns of the N assets
assumed to be invertible. The weights of any
portfolio is assumed to sum to 1.
We consider the minimization program
min ;xEx - AE;X" R]
s.t. 1"X= 1 (0.1)
where A>0.
1. Show that if X is a solution to this program and Y is another
portfolio different from X
satisfying E[X"R] = E[Y"R] then Var(X"R) <Var(Y"R). In other words, show that
any solution to this program is a portfolio on the minimum variance frontier.
2. Write the Lagrangian of the optimization program and the
optimality conditions, using
the notation y for the Lagrange multiplier associated to the constraint (0.1).
3. Show that the Lagrange multiplier satisfiesy= a where E[R]"E-'1 and c = 1'-'1.
pende
4. Determine the optimal portfolio weight X(A) as a function of A.
5. Deduce that any solution X(A) can be written as a linear combination of the GMV
portfolio and another portfolio X:
X =aX + (1 - a) XGMV.
What is the value of a, and the expression of the weights X?
6. Using the
characterization est:ablished
also a frontier in the course, show that the other
1. Study the
portfolio. portfolio X 1S
expected return of X(A), as
parameters a, b and c and explain whichgoes from 0to oo depending on the values of
portfolio we buy and short sell respectivelytheto
generate the efficient frontier.
8. Suppose now that the
market also contains a
optimization program analogous to (1), assumingriskfullfree asset with return r. Write an
investment.
Exercice 3
For a, we let z = -z if <0 and z- =0
p: L'(P) ’R by otherwise. For T ¬ 0, 1), we define the map
p(X) = E[X] +TE[(XE[X})-}.
1. Prove that for any c>0, it
holds
plcX) =cp(X). o agt a
2. Prove that for any d e R, it holds
p(X + d) = p(X)+d. m pr
What is the name of the above property in the language of risk measures? What is the
financial interpretation?
3. Prove that p is sub-additive, that is, for any X,Y e L' (P), one has
p(X + Y)< p(X) + p(Y).
4. We assume here that X<0. Prove that (X- EX])_ < -X a.s. and deduce that
p(X) <0.
5. Deduce from the two previous questions that p is monotone, that is, if X<Y a.s., then
p(X) <p(Y). meotoue
6. Is p a coherent risk measure?
(ost mcuo,iniarias'e par iaulationeat,onosone
