a) Jelaskan perbezaan antara multikolineariti sempurna dengan multikolineariti tidak sempurna.

Explain the difference between a perfect multicollinearity with imperfect multicollinearity.

(5 markah/marks)

- Perfect and imperfect multicollinearity means there is perfect linear relationship between the
explanatory variables in a regression equation it is called as perfect multicollinearity.

- Perfect multicollinearity means that one explanatory variable is an exact linear function of one
or more explanatory variables with no error term. Imperfect multicollinearity means that there is a
linear relationship between the variables, but there is some error in that relationship.

b) Berikan DUA (2) sebab mengapa masalah multikolineariti berlaku dalam sesuatu model
ekonometrik.

Give TWO (2) reasons why the multicollinearity occur in econometrics model.
(5 markah/marks)
i) Insufficient data. In some cases, collecting more data can resolve the issue.

ii) It is caused by an inaccurate use of dummy variables. For example, the researcher may
fail to exclude one category, or add a dummy variable for every category.

iii) It is caused by the inclusion of a variable which is computed from other variables in
the data set.

iv) It is resulted from the repetition of the same kind of variable.

v) Generally occurs when the variables are highly correlated to each other.

c) Hasil pengiraan faktor inflasi varians (variance inflation factor – VIF) memberi jawapan 10
seperti di bawah.

The calculation of variance inflation factor (VIF) gives the answer of 10 as follows.

1 1 1
VIFYi = = = = 10
(1  RYi ) 1  0.90
2
0.10

Jelaskan sama ada wujud masalah multikolineariti.

Explain whether there is a problem of multicollinearity.
(5 markah/marks)
The interpretation of the variance inflation factor mirrors the interpretation of the coefficient of
multiple determination. As a rule of thumb, multicollinearity is a potential problem in this case
when VIFYi is 10 which is highly correlated. The more the VIF increases, the less reliable the
regression results are going to be. In general, a VIF above 10 indicates high correlation. If R2Yi is
greater than 0.9 or VIFk is greater than 10, it is likely that regression coefficients are poorly
estimated. And significance tests on those coefficients may be misleading.

2
Since the VIFYi is less than 10, there is no multicolinearity, therefore the 𝑅𝑌𝑖 must be less than
0.900.

d) Nyatakan DUA (2) cara untuk mengatasi masalah multikolineariti.

State TWO (2) ways to overcome the problem of multicollinearity.

(5 markah/marks)

i) Acquiring Additional Data or a New Sample

By increasing the sample size, the severity of the collinearity problem can be reduced.
Getting additional data on variables already in the sample may not be feasible because
of cost and other considerations. But if these constraints are not very prohibitive, by all
means this remedy is certainly feasible

ii) Rethinking the Model

Sometimes a model chosen for empirical analysis is not carefully thought out and
maybe some important variables are omitted, or maybe the functional form of the
model is incorrectly chosen

iii) Prior Information about Some Parameters

Sometimes a particular phenomenon, such as a demand function, is investigated time
and again
PEMBOLEH UBAH PATUNG/ DUMMY VARIABLE

a) Jelaskan perbezaan antara pemboleh ubah patung yang mempunyai hanya dua kategori dan
lebih daripada dua kategori dengan mengemukakan contoh sesuai.

Explain the difference between a dummy variable which has only two categories and more
than two categories with appropriate examples.
(5 markah/marks)

Dummy variable which has only two categories:

Yi = 0 + iDi + i

Yi is the use of female and male

Di is 1 if female and 0 if male
0 is the average use of male
0 + i is the average use of female
i is the difference between female and male

Dummy variable which has more than two categories:

𝑌𝑖 = 𝐵0 + 𝐵1 𝑋1 + 𝐵2 𝐷1𝑖 + 𝐵3 𝐷2𝑖 + 𝑒𝑖

𝐷1𝑖 = 1 if Malay
= 0 if other races

𝐷2𝑖 = 1 if Indian
= 0 if other races

b) Bina SATU (1) model pemboleh ubah patung ANOVA (analysis–of–variance) yang
mempunyai pemboleh ubah interaktif dengan mengemukakan contoh sesuai.

Build ONE (1) dummy variables of ANOVA model (analysis-of-variance) that has an
interactive variable with appropriate examples.
(5 markah/marks)
Yi = 0 + iDi + i

Where:
Yi is usage (female and male)
Di is 1 if female and 0 if male
0 is the average use of male
0 + i is the average use of female
i is the difference between female and male

Gradient or coefficient i is also the difference in intercept coefficient because it gives the
difference between average use of females and males.
 Use of female Use of male
Yi = 0 + i(1) + i Yi = 0 + i(0) + i
= 0 + i = 0

The result of the above estimation shows the average female use by substituting Di = 1.
= 3176.8333 - 503.1667 Di
= 3176.8333 - 503.1667 (1)
= 2673.67 (average female use = RM2673. 67)

Whereas the average male consumption is by substituting Di = 0.

= 3176.8333 - 503.1667Di
= 3176.8333 - 503.1667 (0)
= 3176.83 (average male use = RM3176.83)

c) Bina SATU (1) model pemboleh ubah patung ANCOVA (analysis–of– covariance) dengan
mengemukakan contoh sesuai.

Build ONE (1) ANCOVA model (analysis-of-covariance) of dummy variables with appropriate
examples.
(5 markah/marks)

Yi = 0 + 1X1 + 2X2 + 1D1 + 2D2 + i

Where:
Yi is quantity of chicken demanded (kg)
X1 is price of chicken (RM)
X2 is price of beef (RM)
D1 is 1 if urban area and 0 if rural area
D2 is 1 if in Selangor and 0 if in Kedah
0 is the quantity of chicken demanded (kg) when all variables are zero
1 is the coefficient of price of chicken to the quantity of chicken demanded
2 is the coefficient of price of beef to the quantity of chicken demanded
1 is the difference in the quantity of chicken demanded between urban area and rural area
2 is the difference in the quantity of chicken demanded between Selangor and Kedah

Yˆi = 0 + 1X1 + 2X2 + 1D1 + 2D2

= 15.8052 – 0.0258X1 + 0.6077X2 – 2.0176D1 + 1.3708D2

The value of 0 = 15.8052 means that the quantity of chicken demanded is 15.81kg when all
variables are zero.

The value of 1 = −0.0258 means that for every RM1.00 price of chicken increases, the
quantity of chicken demanded will drop by 0.0258kg assuming other factors are constant.
 The value of 2 = 0.6077 means that every RM1.00 price of beef increases, the quantity of
chicken meat demanded will increase by 0.6077kg assuming the other factors are constant
(substitutes).

The value of 1 = −2.0176 indicates that the quantity of chicken demanded in the urban area
is 2.0176kg lower than in the rural area assuming other factors are constant.

The value of 2 = 1.3708 indicates the quantity of chicken demanded in Selangor is 1.3708kg
higher than Kedah on the assumption that other factors are constant.

d) Kajian terhadap permintaan daging ayam dengan model Yi = 0 + 1D1 + 2D2 + 3(D1D2) + i
menghasilkan dapatan berikut:

Study on demand for chicken with a model Yi = 0 + 1D1 + 2D2 + 3(D1D2) + i produces the
following results:
Yˆi = 2.22 – 4.44D1 + 6.66D2 + 8.88(D1D2)

dengan;
where;

Yi adalah permintaan daging ayam (kilogram)

is demand for chicken (kilogram)

D1 adalah pemboleh ubah patung: 0 jika perempuan dan 1 jika lelaki

is dummy variable: 0 for woman and 1 for man

D2 adalah pemboleh ubah patung: 0 jika di bandar dan 1 jika kampung

is dummy variable: 0 for city and 1 for village

D1D2 adalah pemboleh ubah interaktif

is an interactive variable

Beri interpretasi terhadap pemboleh ubah interaktif dalam model permintaan ayam tersebut.

Give an interpretation of the interactive variables in the model of demand of chicken.

(5 markah/marks)

Interactive variables D1 and D2 can be interpret as indirect effect of gender (D1) or location
(D2). Given 3(D1D2) = 8.88, it means that on average, compare to man in village, woman in
Village consume more chicken by 8.88kg. This is indirect effect of gender through location.