9.
Show the Correlation Matrix
(from Covariance Analysis) of the following natural
–
log versions of the
variables:
growth of the real GDP (at constant prices), the growth of labor force, the ratio of the
investment share of real GDP and the ratio between external debt and real GDP.
10.
By c
hecking the histograms that you created, c
omment on the
nature of the
distributions
(symmetry,
asymmetry, tails)
between
the regular variables and their natural
–
log versions.
Based on your answer,
natural
–
log functional form and choose
which one you are going to use.
*
Hint:
Another way to answer this question is to compare between the two models
by model
–
building tools.
The following questions are based on the model you
chose
at
Step 10
.
11.
Graph
the time series plots of each of the following variables: growth of the real GDP (at constant
prices), the growth of labor force, the ratio of the investment share of real GDP and the ratio between
external debt and real GDP.
Comment on each of the plots y
ou graphed. Can you make a connection
of the highs and lows that you observe with historical events?
12.
Graph the scatter plot
between the growth of the real GDP
and the ratio of the investment share of
real GDP
(with fit line: regression line).
13.
Graph the sca
tter plot between the growth of the real GDP and the growth of labor force (with fit
line: regression line)
14.
Graph the scatter plot between the growth of the real GDP and the ratio between external debt and
real GDP (with fit line: regression line).
15.
Commen
t on each of the scatter plots in Steps 12, 13, 14 regarding the correlation between each group
of variables.
16.
Regress the growth of the real GDP (at constant prices) on the ratio of the investment share of real
GDP, the growth of labor force and the ratio
between external debt and real GDP
.
17.
After you regress the model in step 15, plot its actual, fitted, residual graph.
18.
Plot the actual, fitted, residual graph for each of the following variables separately: the growth of labor
force, the ratio of the
investment share of real GDP and the ratio between external debt and real GDP.
19.
Test significance at a 5% significance level of each independent variable on the dependent variable of
your model in Step 15. Comment on the power of the significance of each in
dependent variable on the
dependent.
20.
Test the overall significance at a 5% significance level of independent variables on the dependent of
your model in Step 15.
21.
Detect
if multicollinearity
exists
and if it exists try to remedy it.
22.
Detect Heteroscedasticity
of your model if it exists, by making use of the Park Test, the Breusch
–
Pagan
–
Godfrey Test and the White Test with and without cross terms and comment on your results.
In case Heteroscedasticity exists, try to remedy it either by
using Generalized Least Squares or by using
Weighted Least Squares.
Why do you think heteroscedasticity exists/ not exists in your model?
23.
Detect if autocorrelation of your model exists
and in case it exists try to remedy it.
24.
In case autocorrelation exists
, plot the actual/fitted residuals of your new model and of your new
independent variables separately and compare with the plots in Steps 17 and 18. What do you observe?