Professional Documents
Culture Documents
FACULTY OF
ECONOMICS AND
BUSINESS
Name
Matric Num.
1.
Kwan Hui Ne
47228
2.
46928
Sample
A sample is a set of data drawn from the population.
ii.
Population
Population is all items of interest in statistical problem such as mean and median. The
value for population may very large, and sometimes, infinitely large. In empirical
study, normally only samples of observations from population are available.
iii.
Variance
Variance is a measure of dispersion. It indicates how closely or widely individual X
values are spread, dispersed or distributed around the mean value.
iv.
Covariance
Covariance measures the linear association between X and Y measure how 2 random
variables vary or move together. The covariance between two random variables can be
either positive, zero or negative.
v.
Correlation
The correlation measures how strongly two variables related or associated with each
other. Correlation close to +1, means 2 variables are very strongly positively
correlated. Correlation close to -1, means 2 variables are very strongly negatively
correlated. Correlation close to zero, means 2 variables are not related.
vi.
Coefficient of Determination
Coefficient determination is a statistical meausres of how well the regression line
approximates real data points or gives information about model goodness of fit.
Question 2
From a sample of 209 firms, Wooldridge obtained the following regression results:
log ( ^
salary ) =4.32+0.280 log ( sales ) +0.0174 roe+0.00024 ros
se = (0.32)
R2 = 0.283
(0.035)
(0.0041)
(0.00054)
and where figures in the parentheses are the estimated standard errors.
i.
ii.
H0 : 4 = 0
Ha : 4 0
t = b4/se(b4)
= 0.00024/0.00054
= 0.4444
Since 0.4444 < 1.96, we do not reject H0.
Thus, the coefficient for ros is statistically significant at 5% level of significance.
iii.
iv.
What is the overall significance of the regression? Which test do you use? And why?
H0 : 1=2=3= 4=0
Ha : 1 2 3 40
F = [R2/k-1] / [(1-R2)/n-k]
= [0.283/4-1] / [(1-0.283)/209-4]
= 26.971
Critical F value = F(k-1,n-k) = F(3,205) = 3.32. Since Fcalculated=26.971 > Fcritical=3.32,
we reject H0.
Thus, not all coefficients are equal to zero.
Test used is F-test. This is because F-test can assess multiple coefficients
simultaneously unlike t-test.
Question 3
Below is the demand for money function for India for the periods 1948 to 1965:
^
ln M t=1.60270.1024 ln R t +0.6869 ln Y t +0.5284 ln M t 1
se = (1.4024)
(0.3678)
R2 = 0.9227
(0.3427)
(0.2007)
d = 1.8624
With the 5% lower and upper d values are 0.933 and 1.696, the estimated d value
of 1.8624 indicates that there is no positive autocorrelation. Can we trust the
computed d to find out whether there is serial correlation in our data? Why?
0.933
1.696
d = 1.8624
Yes. We can trust the computed d to find out whether there is serial correlation in
our data. This is because as we can see from the above figure, the value of d is near
to 2. Thus, there is no positive autocorrelation.
ii.
Durbinh= 1
1.8624
2
18
) 118(0.2007
)
2
0.5567
iv.
Question 4
The results from the Granger Causality test between Y and X for the period of 1960 1999, as
follows:
Yt = 0.3425Xt + 0.5046Xt-1 + 0.1485Xt-2 + 0.0911Xt-3 + 0.0255Yt-1
R2 = 0.8823
RSS = 210.24
Yt = 0.4515Yt-1 + 0.3272Xt
R2 = 0.9634
RSS = 265.13
Xt = 0.3397Yt + 0.8460Yt-1 + 0.6524Yt-2 + 0.4947Yt-3 + 0.0.0153Yt-4 + 0.0268Yt-5 + 0.0432Xt-1
R2 = 0.7724
RSS = 143.21
Xt = 0.3216Xt-1 + 0.0546Yt
R2 = 0.8956
RSS = 157.43
i.
The granger causality test is a statistical hypothesis test for determining whether one
time series is useful in forecasting another. According to the Granger causality, if a
one variable as the dependent variable (Y) and another as explanatory variable (X),
we made an implicit assumption that changes in the X will eventually induce the
changes in the Y. That is the notion of causality in which information about X is
expected to affect the conditional distribution of the future values of Y.
ii.
H a= X granger cause Y
RSS
( RRSSUR )
m
RSSUR
(nk )
Ftest =
(265.13210.24)
3
F=
210.24
( 405)
F 3.046
C.V = 2.81
We reject null hypothesis if F-test> C.V
Since F=3.046 > C.V = 2.81, so we reject null hypothesis
Hence, we can conclude that X granger cause Y.
iii.
157.43143.21
5
F=
143.21
407
F 0.655
C.V = 2.71
We reject null hypothesis if F-test> C.V
Since F=0.655 < C.V = 2.71, so we do not reject null hypothesis
Hence, we can conclude that Y does not granger cause X.