You are on page 1of 2

Financial Econometrics

Sogang University
In Choi
Fall 2008
Midterm Examination
Time: 4:30-5:45 PM
Total points: 120 points
Instructions: (i) Consulting an A4-sized sheet of paper on which you wrote down necessary
information is allowed.
(ii) Leave quietly if you nish the exam before others.
1. (10 points) An investor has held an asset for three years. The simple net returns of
the asset per annum for the three years are 0.02, 0.07 and 0.06. What is the simple
multiperiod return of the asset for the three year period? Provide a number based on
the exact formula.
2. (10 points) The simple nominal net return of an asset is 0.05 per annum. If the ination
rate is 0.03 during the last one year, what is the real net return of the asset?
3. (20 points) Consider the AR(1) model :
t
= c:
t1
+ c
t
. (t = 1. 2. .... 1). jcj < 1. c
t

\`(0. o
2
). Assume that the coecients c is known to us. (In practice, this is esti-
mated.)
(a) How can we predict :
T+1
? What is the variance of the forecast error?
(b) How can we predict :
T+2
? What is the variance of the forecast error?
4. (10 points) Derive the autocorance function of the MA(1) process
:
t
= c
t
+ 0.5c
t1
. c
t
iid(0. 1).
5. (10 points) Suppose that :
t
= 0.9:
t1
+c
t
+1.1c
t1
. where c
t
iid(0. 1). Is this process
stationary and/or invertible?
6. (10 points) Observed time series :
t
follows the process
:
t
= j + A
t
; A
t
= c
t
+ 0.1c
t1
; c
t
\`(0. 1); t = 1. ...1.
Does :
T
=
1
T
(:
1
+ :
2
+ + :
T
) become more and more accurate in estimating j as
sample size increases?
7. (10 points) Suppose that a time series :
t
can be decomposed as :
t
= c
0
+c
1
t +o
t
+A
t
.
where c
0
+ c
1
t denotes the trend component (c
0
and c
1
are unknown constant), o
t
the seasonal component and A
t
the random component. Suppose that the seasonal
component has the property o
t
= o
t4
. Show that a seasonal dierencing can eliminate
both the seasonal and trend components.
8. (20 points) Derive the capital market line and explain why its slope is steepest among
all the portfolios within and on the minimum variance frontier.
9. (10 points) Consider the predictive regression
:
t+k
= cr
t
+
t+1
(/ 1);
r
t+1
= jr
t
+ o
t+1
.
Explain why the predictive regression may provide higher coecient estimates for c as
/ grows.
10. (10 points) The variance ratio statistic is dened by
\
k
=
\ c: (:
t+k
[/])
/\ c: (:
t+1
)
.
Comment on the following statement.
If the variance ratio is less than one, it implies that stocks are safer for long-run in-
vestors who can tolerate ups and downs of the market.
2

You might also like