ECONOMETRICS
TUTORIAL 4 SOLUTIONS
Fall session
EXERCISE 1
1) 𝛽መ1 and 𝛽መ2 have elasticity interpretations. For example, for India a 1% increase in
per capita real GDP will lead to a 1.006% increase in per-capita energy demand. A 1%
increase in real energy price will lead to a 0.068% decrease in per-capita energy
demand.
2) For Korea:
RSS
We know that ˆ 2
so, RSS ˆ u2 n k
nk
u
0.031 17 3
²
0.0134
2
RSS RSS
We know that R 2 1 so, TSS
TSS 1 R2
0.0134
1 0.54
0.0291
ESS TSS RSS
0.0291 0.0134
0.0157
3
3) For Indonesa
H 0 : 2 0 H 0 : 3 0
H1 : 2 0 H1 : 3 0
ttab t0.025,14 2.145 ttab t0.025,14 2.145
ˆ2 1.564 ˆ3 0.488
tcal 6.684 tcal 2.5
se ˆ2 0.234
se ˆ3 0.195
tcal >ttab Reject H 0 tcal >ttab Reject H 0
4
EXERCISE 2
Yi 1 2 I i 3 Si U i
0.2782 0.00409 0.03859 yi 5475
X T Y yi Ii 195375
1
X T
X 0.00409 0.0002918 0.001955
0.03859 0.001955 0.04139 y S 15670
i i
1)
X X X TY
1
ˆ T
0.2782 0.00409 0.03859 5475 119.356
0.00409 0.0002918 0.001955 195375 3.982
0.03859 0.001955 0.04139 15670 55.343
5
Yˆi 119.356 3.982 I i 55.343Si
1)
ˆ1 119.356 If the annual family income and the number of family mumbers
were 0 then, the annual expenditure on fixed telephone by
a family is 119.356 Euros.
ˆ2 3.982 If the annual family income increases by one thousand Euro then,
the annual expenditure on fixed telephone by a family increases
by 3.982 Euros.
ˆ3 55.343 If the number of family members increases by one person,
the annual expenditure on fixed telephone increases by 55.343 Euros.
6
2) First method Second method
ESS
R
2
R 1
2 RSS
TSS
TSS
ESS ˆ T X T Y ny 2
5475 2 RSS yi2 ˆ T X T Y
5475
ESS 119.356 3.982 55.343 195375 15
15670 15 2340325 2298682.16 41642.84
2298682.16 -1998375 = 300307.16
41642.84
TSS y ny R 1 0.8782
2 2 2
i
341950
2340325 15 3652
341950
300307.16
R2 0.8782
341950
87.82% of the variation of annual expenditure on fixed telephone by a family
7
is explained by the regression.
RSS 41642.84
3) ˆ
2
3470.24
nk 15 3
u
var ˆ u X X
1
4) ˆ 2 T
0.2782 0.00409 0.03859
3470.24 0.00409 0.0002918 0.001955
0.03859 0.001955 0.04139
965.42 14.19 133.92
14.19 1.013 6.78
133.92 6.78 143.63
8
H 0 : 2 0
5)
H1 : 2 0
ttab t0.025,12 2.179
ˆ2 3.982
tcal 3.96
se ˆ2 1.013
tcal >ttab Reject H 0 , annual family income is significant at 5% level.
H 0 : 3 0
H1 : 3 0
ttab t0.025,12 2.179
ˆ3 55.343
tcal 4.62
se ˆ3 143.63
tcal >ttab Reject H 0 , the number of family members is significant at 5% level.
9
H 0 : 2 3 0
6)
H1 : 2 0 and/or 3 0
R2 / k 1 0.8782 / 2
Fcal 43.26
1 R / n k 1 0.8782 /12
2
Ftable Fk 1,n k F2,12 3.89
Fcal >Ftable
Reject H 0 , the model is globally significant at 5% level.
10
H 0 : 2 3 H 0 : 2 3 0
7)
H1 : 2 3 H1 : 2 3 0
ttab t0.025,12 2.179
ˆ2 ˆ3 ˆ2 ˆ3
tcal
se ˆ2 ˆ3
ˆ ˆ2 var
var
ˆ ˆ3 2cov
ˆ ˆ2 , ˆ3
3.982 55.343
1.013 143.63 2 6.78
4.08
tcal >ttab Reject H 0 .
11
H 0 : 3 10 2 H 0 : 3 10 2 0
8)
H1 : 3 10 2 H1 : 3 10 2 0
ttab t0.025,12 2.179
ˆ3 10 ˆ2 ˆ3 10 ˆ2
tcal
se ˆ3 10 ˆ2
ˆ ˆ3 var
var
ˆ 10 ˆ2 2cov
ˆ 10 ˆ3 , ˆ2
55.343 10 3.982
143.63 100 1.013 20 6.78
0.796
tcal <ttab Fail to reject H 0 .
12
Yˆi 110.372 4.095I i 46.642Si 17.36M i ; R 2 0.883
( t value ) (2.95) (3.52) (2.41) (0.4)
H 0 : 4 0
9)
H1 : 4 0
ttab t0.025,11 2.201
tcal 0.4
tcal <ttab Fail to reject H 0 .
We calculate the adjusted R squared for each model:
n 1 14
R1 1 1 R 2 1 1 0.8782 0.8579
nk 12
n 1 14
R 2 1 1 R 2 1 1 0.883 0.8510
nk 11
We choose the first model with the highest R 2 . 13
EXERCISE 3
ct 1 2 pt 3 rt ut
c
1) 2 0 ; if the price , the consumption
p
r
2 0 ; if the revenue , the consumption
p
RSS ct2 ˆ T X T C
24
t 1
0.06 0.03 0.06 25 1.02
X X X T C 0.03 0.2
1
ˆ T
0 2 0.35
0.06 0.35 7 0.95
0
25
33.34 1.02 0.35 0.95 2 33.34 31.45 1.89
7
1.89
So,ˆ
2
=0.09 14
21
ˆ ˆ ˆ u2 X T X
1
3) Var
0.06 0.03 0.06 0.0054 0.0027 0.0054
0.09 0.03 0.2 0 0.0027 0.018 0
0.06
0 0.35 0.0054 0 0.0315
se ˆ2 var
ˆ 0.018 0.134
ˆ ˆ2 var
se ˆ
3
ˆ ˆ
var 3
ˆ 0.0315 0.177
var
H 0 : 2 0
4) i)
H1 : 2 0
ttab t0.025,21 2.08
ˆ2 0.35
tcal 2.61
se ˆ2 0.134
tcal ttab Reject H 0 , the price is significant at 5% level. 15
H 0 : 3 0
ii)
H1 : 3 0
ttab t0.025,21 2.08
ˆ3 0.95
tcal 5.37
se ˆ3 0.177
tcal ttab Reject H 0 , the income is significant at 5% level.
H 0 : 2 3 0
iii)
H1 : 2 0 and/or3 0
25 2
ESS ˆ T X T C Tc 2 1.02 0.35 0.95 2 24 5.4
25
7 24
5.4 / 2
Fcal 30
1.89 / 21
Ftab F2,21 3.47
Fcal Ftab Reject H 0 , the revenue and the price affect both the consumption. 16
H 0 : 3 1
iv)
H1 : 3 1
ttab t0.025,21 2.08
ˆ3 1 0.95 1
tcal 0.28
se ˆ3 0.177
tcal ttab Fail to reject H 0 .
17
EXERCISE 4
The production function:
1) log PROD 1 2 log AREA 3 log LAB 4 log FERT u
The fitted regression equation is:
logˆ PROD =-1.548+0.3617log(AREA)+0.4328 log (LAB)+0.2095 log (FERT)
All estimates have elasticity interpretations. For example, a 1% increase in labor
will lead to a 0.4328% increase in rice output. A 1% increase in fertilizer will lead to a
0.2095% increase in rice output.
All p-values are less than 0.005 implying all coefficients are significantly different from
zero at 5% significance level.
18
2) CI ˆ t
4 4
2
,n k
ˆ
se 4
0.2095 1.96 0.0383
0.1344, 0.2846
H 0 : 2 0.5 H 0 : 2 0.5 0
3)
H1 : 2 0.5 H1 : 2 0.5 0
ttab t0.025,100 1.96
ˆ2 0.5 0.3617 0.5
tcal 2.16
se ˆ2 0.0640
tcal >ttab Reject H 0 .
19
H : 2 3
4) 0 H 0 : 2 3 0
H1 : 2 3 H1 : 2 3 0
The test can be performed with an F statistic or a t-statistic.
ttab 1.96
ˆ2 ˆ3 ˆ2 ˆ3 0.36174 0.43285
tcal 0.585
se ˆ2 ˆ3
ˆ ˆ2 ˆ3
var 0.0148
tcal ttab Fail to reject H 0 .
Using the F statistic we have:
Fcal 0.342 that is
RSSR RSSU / m
0.342
RSSU / n k
Ftab F1,100 3.94
Fcal Ftab Fail to reject H 0 . 20
5) H 0 : 2 3 4 1 H 0 : 2 3 4 1 0
H1 : 2 3 4 1 H1 : 2 3 4 1 0
ttab 1.96
ˆ2 ˆ3 ˆ4 1 ˆ2 ˆ3 ˆ4 1 0.3617 0.4328 0.2095 1
tcal 0.168
se ˆ2 ˆ3 ˆ4
ˆ ˆ2 ˆ3 ˆ4
var 0.00057
tcal ttab Fail to reject H 0 .
Using the F statistic we have:
Fcal 0.0295
Ftab F1,100 3.94
Fcal Ftab Fail to reject H 0 .
21
H 0 : 2 3 0 and 2 3 4 =1
6)
H1 : 2 3 0 and/ or 2 3 4 1
Fcal
RSSR RSSU / m RSSR RSSU / 2
RSSU / n k RSSU /100
Ftab F2,100 3.09
If Fcal Ftab Reject H 0 .
22
EXERCISE 5
MAˆi 8.5 0.3 NFi NSi 0.5 NTi 1.2Gi ; RSS 0.89
(4.5) (7.1) (2.6) (1.9) (0.3)
H 0 : 0
1)
H1 : 0
ttab t0.025,25 2.06
1.2
tcal 4
0.3
tcal ttab Reject H 0 .
2) is the average difference in average obtained in master between men and women.
23
3) MAi 1 2 NFi 3 NSi 4 NTi Gi ui (1)
MAi 1 NFi NSi 0.2 NTi ui (2)
There are four restrictions
1, 3 1, 4 0.2 and 0
2
H 0 : 2 1 and 3 1 and 4 0.2 and 0
H1 : 2 1 and/or 3 1 and/or 4 0.2 and/or 0
Fcal
RSSR RSSU / m 3.8 0.89 / 4
20.43
RSSU / n k 0.89 / 25
Ftab F4,25 2.76
Fcal Ftab Reject H 0 , we reject the restricted model.
24