ECM1002 - Econometrics

Workshop - Week 02
You are required to study Lecture notes – Week 02 prior to attending this workshop.

After attending this workshop, you would be able to:

- Understand simple linear regression model and Gauss-Markov theorem,

- Estimate the regression parameters and Assess the Least Square Estimators,
- Understand the probability distribution of the Least Square Estimators,
- Estimate the Variance of the error term,
- Obtain different plots from Eviews.

Section 1: Theory revision (30 minutes)

In this section, main points from Lecture notes - Week 02 will be revised.

Section 2: Working with Eviews

Download the file “capm5.wf1” from LMS.

Data visualisation is a powerful method to understand the properties of data and an effective
tool to communicate them with your audience. It can often be more informative and appealing
than presenting only regression statistics. In this section, we will learn how Eviews can be used
to visualise the data.

The file contains 180 monthly time series data for stock returns from 1998 to 2012. The stock
returns include Disney (dis), Exxon-Mobil (xom), General Electric (ge), IBM (ibm), Microsoft
(msft), and market index return (mkt, S&P500). Note that the return at time t is calculated as
(Pt − Pt-1)/Pt, where Pt is the price at time t.

Using Eviews to obtain the following:

1
a. A group of return series
b. Time plots, Box plots, Distribution plots, Scatter plots of return series

Eviews Instruction:

First, we want to generate a group for the above stocks.

Highlight these time series as above by pressing the Control key in your keyboard. When
highlighting, make sure you click the mkt first followed by the others. In this way, you can
position the series mkt in the first column as you can see above. Right-click the highlighted
area and click Open As Group to see the spreadsheet view as above.

1. Time plots

From the Spreadsheet menu, click View, Graph, Line & Symbol,

Choose Multiple graphs here

Click OK, then you will see the time plots as below:

2
You can see that all stock returns exhibit high degree of volatility around 0, indicative of high
risk involved with the mean return close to 0. Most of them show different degrees of volatility
over time; see, for example, GE stock return show high degree of volatility first, followed by
low and high periods. Most of them show a number of spikes over time, which again represents
highly risky nature of investment.
2. Boxplots

Generate the boxplots in a single graph as below, by clicking View, Graph, Boxplot and
choosing single graph.

3
The boxplots contain the following information:

IQR is the interval that contains

50% of the data points

All stock returns show the median values close to 0, indicating that their expected returns are
nearly 0. The market return shows the lowest degree of variability with the smallest IQR value,
while IBM and Microsoft showing a high degree of variability indicated by large IQR values
and a large number of outliers.

If you are an investor who wants to minimize the risk, it is best to invest in a well-diversified
portfolio such as market index.

3. Distributional Features (Histograms, Kernel Density)

To visualize distributional features, you may want to present histograms or kernel density
functions. The latter is a smoothed version of the former. To generate these plots,

Choose Kernel Density or

Histogram;
Single or Multiple Graphs

4
In this way, you can compare the returns based on the shape of their distributional features such
as asymmetry (skewness) and occurrence of extreme values. The returns are overall skewed to
the left, which means a higher occurrence of negative extreme values than the positive ones.

4. Scatter plots

To generate the scatter plots, click View, Graph, Scatter

Choose Multiple graphs

-First vs all, and then Fit
Lines, Regression Line

The above presents the scatter plots for all stock returns against the market return. The red
lines are the fitted regression lines. All stock returns are positively related with the market
return, in different degrees.

In the following weeks, you will estimate and test these regression lines in the context of an
economic model called the CAPM (capital asset pricing model).

5
Section 3: Calculation and Interpretation (textbook questions)

During the 2-hour workshop, you are not required to answer all the questions in this section.
Instead, your facilitator will choose random questions to demonstrate and help you
understand/apply related econometric theories mentioned in Section 1.

6
Section 4: Further self-practice

Choose a stock return from capm5.wf1 according to the last digit of your student ID as below:

Stock Last digit of your Student ID

IBM 0, 1
GE 2, 3
XOM 4, 5
Microsoft 6, 7
Disney 8, 9

For the stock that is allocated to you, generate its

• the kernel density,
• the Q-Q (Quantile-Quantile) plot, and
• For each plot, provide a one-sentence comment with your interpretations about the
shape of your return distribution, in comparison with a normal distribution.

7
Hints:
For information about the Q-Q plot, see here or refer to your statistics textbook.

Kernel density is a smoothed version of histogram. For an intuitive explanation of its

calculation, you may watch the first 3 minutes of this video.

For the market return, they are as below:

Comment: It appears that the market index return has an asymmetric distribution with longer tail
on the left. The distribution is showing a clear departure from a normal distribution, especially
in the left tail area, with high frequency of large negative returns.