Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
ECF480-ADVANCED ECONOMETRICS
Instructions to Candidates:
2. Check that you have the correct examination paper in front of you.
3. There are TWO sections in this question paper. Answer all questions in
section A and any one question in section B.
5. Write down the number of questions that you have answered on the
cover of the examination answer booklet provided.
Page 1 of 7
Attempt all questions
QUESTION ONE
Y i=β 1 i + β 2i X 1 i + β 3 i X 2 i +… … … .+ β k +1 X ki +U i
[TOTAL: 35
MARKS]
QUESTION TWO
Page 2 of 7
----------------------------------------------------------------------------------------------------------------------
Economic Growth* Coef. Std. Err. t P>|t| [95% Conf. Interval]
---------------------------------------------------------------------------------------------------------------------
Government Exp* 0.226 0.012 18.833 0.000 0.2007 0.249288
Gross Investment* 0.791 0.228 3.4698 0.010 0.3295 1.252472
Net Exports* -0.366 -0.47 0.7787 0.234 0.5853 -1.31728
_cons -1.462 0.75 1.9493 0.048 -2.98 0.056
*Variable in in log form
----------------------------------------------------------------------------------------------------------------------
Economic Growth* Coef. Std. Err. t P>|t| [95% Conf. Interval]
---------------------------------------------------------------------------------------------------------------------
Government Exp* 0.226 0.012 18.833 0.000 0.2007 0.249288
Gross Investment* 0.791 0.228 3.4698 0.010 0.3295 1.252472
Net Exports* -0.366 -0.47 0.7787 0.234 0.5853 -1.31728
D 0.382 0.109 3.5045 0.009 0.1613 0.602616
_cons -1.462 0.75 1.9493 0.048 -2.98 0.056
Page 3 of 7
a) Provide an economic interpretation for the dummy variable. Given the result of
the dummy, which model will you proceed with? [5
Marks]
[TOTAL: 25 MARKS]
QUESTION THREE
^= n θ
W=θ and W
n−1
^ is not an unbiased estimator of W but is asymptotically
Show that W
unbiased where θ is a constant [10
Marks]
Page 4 of 7
c) In cases of smaller samples, the classical linear regression model
assumptions are to hold if the OLS estimators are going to be BLUE
i. How will a non-constant variance affect the estimators? [3
Marks]
ii. How will a full rank of the regressors’ matrix affect the estimators?
[5
Marks]
iii. How will correlated residuals affect the estimators? [3
Marks]
[TOTAL: 40
MARKS]
QUESTION FOUR
a) Given the assumptions of the classical linear regression model, the least-squares
estimators, in the class of unbiased linear estimators, have minimum variance,
that is, they are BLUE.
i. Prove that indeed the OLS estimators are BLUE if the Gauss-Markov
assumptions hold [8 Marks]
b) You have been tasked to determine the relationship between consumption. and
income. You are provided with hypothetical data on weekly family consumption
expenditure and weekly family income as shown in the table below.
Where;
Page 5 of 7
iii. Interpret the estimated ^β s [4 Marks]
iv. Compute σ^ 2 [6 Marks]
v. It was estimated from the data above that the F statistic was 202.87. Test
the hypothesis that ^β 1= ^β 2=0 at 5% level of significance. [7 Marks]
[TOTAL: 40 MARKS]
FORMULAS
Page 6 of 7
2-Variable Model
2
(∑ x i y i)
∑ U^ 2i =∑ y 2i −
∑ x 2i
R2new −R2old
F=
( df )
1−R2new
( df )
( R 2UR−R2R )
m
F=
(1−R2UR )
n−k
3-Variable Model
^β 0=Ý − ^β 1 X́ 1− ^β2 X́ 2
y i x 1 i ∑ x 22 i−( ∑ y i x 2 i )( ∑ x 1i x 2i )
^β 1= ∑
2
∑ x 21 i ∑ x 22 i−( ∑ x1 i x2 i )
y i x 2 i ∑ x 21 i−( ∑ y i x 1 i )( ∑ x 1i x 2i )
^β 2= ∑
2
∑ x 21 i ∑ x 22 i−( ∑ x1 i x2 i )
2 2 2 2
1 X́ 1 i ∑ x 2 i + X́ 2 i ∑ x1 i−2 X́ 1 X́ 2 ∑ x 1i x 2i 2
^
var ( β 0) = +
n [ ∑ x 2 ∑ x 2 −( ∑ x x ) 1i 2i
2
1i
σi
2i
]
var ( ^β 1) =
∑ x 21 i σi
2
2
∑ x 21i ∑ x 22i −(∑ x 1 i x 2 i )
var ( ^β 2) =
∑ x 22 i σ
2
2 2 2 i
∑ x 1i ∑ x 2i −(∑ x 1 i x 2 i )
Page 7 of 7