Brandon H .

Louis
ECON3049 Econometrics : 1

FINAL

1. a) Yi =
Bi +
µi

Expectations E(Yi)=E( Bit )

:
ui

I =
{ it 0

Ñi=Y
The model 4. 1) is the true
•

• • relationship .

b) di = I -
iv.I calculate ai = I -

ait
Eloi ) .
=
El -

ElÉ¥¥✗ : E¥¥¥
a
=

[ (da) I
f
= -

Since Elxi -

4 =/ o Eloia) =/ 0

Eloi D= f- Etaa) I =/
p
•
: ai is biased

c) Including irrelevant variables when

doing OLS estimation
,
leads to a
larger standard error
of the

estimator .

heading to efficiency loss .

d)

e)
Brandon H .
Louis
ECON3049 Econometrics
: 1

FINAL

2 .
Brandon H .
Louis
ECON3049 Econometrics
: 1

FINAL

3 .
 Brandon H .
Louis
ECON3049 Econometrics : 1

FINAL

4 . If =
YO -
P
0.01 1- 0.24 -
1.55 , + 3.652 + 4.7 53

a)

b) for quarter 4 (reference) 51=52--53--0 ,

Quarter
Q ,
=
70 -0.01 Pt 0.24 -

1.54) -13.610) -14710)

=
To
-

0.014.1) -10.2400) -

1.5

=
88.489

Quarter2
Q ,
=
70 -0.01 Pt 0.24 -

1.5 6) + 3.64 ) -147 ( o )

=
70 -

0.014.1b) + 0.2402) + 36

= 93.988

Quarter 3

Q ,
=
no -0.01 Pt 0.24 -

1.5 (o ) -13.6 (o) -1474)

=
70 -

0.014¥ -1 0.2404) -1 4.7

95.488
=

Quarter 4
Q ,
=
yo -0.01 Pt 0.24 -

1.5 6) + 3.66 ) -147 (O )

=
70 -

0.014.14) -1 0.2403)

=
90.5886
Brandon H .
Louis
ECON3049 Econometrics
: 1

FINAL
4-
c)

d)
Brandon H .
Louis
ECON3049 Econometrics
: 1

FINAL

5. a) Ho :B Bz =P
,
=
,
=
o

H :@ least
,
one
§ =/ 0

In order to test this hypothesis ,

we need more information .

b) With DW
being 0742 , this is low and a cause for concern , another form of regression testing is needed .