You are on page 1of 3

Solution to Midterm Exam in Econometrics

1. Let X and Y have joint PDF


f(x; y) = ax+by for 0 x 1; 0 y 1; a > 0; b > 0; zero elsewhere
(a) (5) What restriction should a and b obey?
_
1
0
_
1
0
(ax+by)dxdy =
_
1
0
_
a
2
+by
_
dy =
a
2
+
b
2
= 1 =)a+b = 2
(b) (15) Let a = b = 1. Find
P(X 0:5jY 0:5)
P(X 0:5jY 0:5) =
P(X 0:5; Y 0:5)
P(Y 0:5)
=
_
0:5
0
_
1
0:5
(x +y)dxdy
_
0:5
0
_
1
0
(x +y)dxdy
=
1
4
3
8
=
2
3
(c) (15) Let a = b = 1. Find the PDF of Z = X
2
:
Note that
f
X
(x) =
_
1
0
(x +y)dy = x +
1
2
X =
p
Z and
dX
dZ
=
1
2
p
Z
f
Z
(z) = f
X
(
p
z)

dX
dZ

=
_
p
z +
1
2
_
1
2
p
2
=
1
2
+
1
4
p
z
for 0 z 1; zero elsewhere
An alternative proof is
F
Z
(z) = Pr (Z z) = Pr
_
X
2
z
_
= Pr
_
X
p
z
_
= F
X
_p
z
_
=
_
p
z
0
_
x +
1
2
_
dx
=
z
2
+
1
2
p
z
=) f
Z
(z) =
dF
Z
(z)
dz
=
1
2
+
1
4
p
z
1
2. (15) Let X = (X
1
; X
2
; X
3
) be a 1 3 vector random variable with
mean and covariance matrix
V =
_
_
1 2 3
2 1 4
3 4 1
_
_
Find the variance of Z = X
1
X
2
X
3
:
Let
a =
_
_
1
1
1
_
_
Then
Z = Xa
=) V (Z) = V (Xa) = a
0
V (X)a
=
_
1 1 1

_
_
1 2 3
2 1 4
3 4 1
_
_
_
_
1
1
1
_
_
= 1
3. Let X have PDF f(x) = e
(x)
for x < 1; 1 < < 1; zero
elsewhere (notice that x ).
(a) (15) Find the moment generating function M
X
(t) for X:
M(t) = E
_
e
tX
_
=
_
1

e
tx
e
(x)
dx
= e

_
1

e
(t1)x
dx = e

1
t 1
_
lim
x!1
e
(t1)x
e
(t1)
_
= e

1
t 1
_
0 e
(t1)
_
for t < 1
=
e
t
1 t
(b) (15) Find E(X) (hint: if you dont use (a) for your answer, you
may nd the following result useful,
_
1
a
xe
x
dx = e
a
(a + 1)):
dM(t)
dt
=
e
t
(1 t) +e
t
(1 t)
2
=) E(X) =
dM(t)
dt
j
t=0
= + 1
2
(c) (10) Use your answer to (b) to nd a method of moment estimator
for ; given a random sample X
1
; : : : ; X
n
.
Using the function m(X; ) = X 1 we get the moment
condition:
E(X 1) = 0
so that the MM estimator of is
1
n
n

i=1
_
X
i

1
_
= 0 =)

= X
n
1
(d) (10) Is the estimator from (c) consistent? Explain.
It is consistent because a Law of Large numbers implies plimX
n
=
E(X) = + 1 and therefore
p lim

= plimX
n
1 = + 1 1 =
(e) Bonus: Find the maximum likelihood estimator of : (hint: ex-
amine the function L() =
n

i=1
f(x
i
) and not its logarithm, and
recall that each x
i
)
L() =
n

i=1
e
(X
i
)
= e
n
e

P
n
i=1
X
i
Note that L() is increasing in so that we would like to make
as large as possible. BUT; is less than any of the X
0
s in the
sample so the highest it can go is as much as the lowest X
i
: Thus,

ML
= Min(X
1
; X
2
; : : : ; X
n
)
3

You might also like