Lesson 4 Review of Econometric Topics
4.1 Introduction
This lesson provides you with information (and links) on topics you should have encountered in
undergraduate econometric and statistics studies as well as some basic EViews operations.
Although they do not form part of the official study outline, they are, by implication, prior knowledge
and are thus examinable. For a more formal discussion on these topics, kindly consult your
textbook (Wooldridge).
Please note that this lesson can by no means replace the work you should have covered during
a typical undergraduate econometric course (typically the content of chapters 1-9 in your
textbook). The aim is merely to highlight some of the most important concepts we will be building
on.
Lesson 4.6 (recommended reading) highlights some sources you can use to review basic
statistical principles.
4.2 Review topics
Linear regression analysis:
http://www.youtube.com/watch?v=ccWkOUtiyFQ&list=PLRW9kMvtNZOgzWdkmkI3pYBYZmCw
FvPno
Hypothesis testing:
http://www.youtube.com/watch?annotation_id=annotation_575366&feature=iv&src_vid=UApFKi
K4Hi8&v=yTczWL7qJ-Y
Estimation methods (OLS, GLS, Maximum Likelihood, etc.):
http://www.youtube.com/watch?v=Wxi4Q5tYeq8&list=PLB2704711C3E47C90&index=1
BLUE (Minimum variance, linear, unbiased) properties of OLS estimators:
http://www.youtube.com/watch?v=jXU4-hYcNwc&list=PLB2704711C3E47C90&index=2
Assumptions of the classical linear regression model (CLRM):
http://www.youtube.com/watch?v=b2S6nAK9cZ8&index=3&list=PLB2704711C3E47C90
Dealing purposefully with econometric problems (relaxing the assumptions of the classical
model):
• multicollinearity
• heteroscedasticity
• autocorrelation
These three econometric problems are very important and it is assumed that you have a good
understanding of them. In each case its nature, consequences and how to detect and fix the
problems apply. [They will also feature prominently in lesson 1 (especially 1.8, the practical
activities and assignments), where the ECM is tested for various potential problems.]
4.3 Activity 4.1: A review of basic statistics and econometrics
This activity reviews the following:
• describing the scope of econometrics, and the ability to develop an appreciation for the role
of econometrics in applied economics
• understanding basic statistical formulae (mean, standard deviation, variance, covariance)
1. Define econometrics. (6)
2. Explain how econometricians usually analyse a problem (hint: see Wooldridge Section
1-2, and Chapter 19) (8)
3. List the objectives of regression analysis. (4)
4. Explain the following terms (hint: mathematical notation can be used):
4.1 the stochastic (random) error term
4.2 the residual
4.3 List four (possible) reasons why we include the stochastic error term in the
hypothesised population regression function (PRF). (10)
5. Distinguish between the z-distribution and the t-distribution. (4)
6. Explain each of the following terms: (6)
6.1 time series data
6.2 cross-sectional data
6.3 pooled data
7. Given the following model, Yi = ˆ 0 + ˆ1 X i + u i , state the theoretic formula for estimating
ˆ 0 and ˆ1 . (4)
8. Distinguish between models linear in variables and models linear in parameters. Write
down an example of each. (6)
9. Consider the following (hypothetical) data that show the average wage earned as a
function of an individual’s training:
Table 1: Average wage versus Training
Average wage (R per hour) Training (Years of schooling)
22.3 6
28.9 7
29.9 8
36.7 9
36.6 10
32.9 11
39.1 12
39.2 13
55.1 14
53.4 15
54.2 16
68.1 17
67.7 18
9.1 Write the regression model (not the estimated regression results) for “average wage”
as a function of “training” and explain your a priori (theoretical) expectations of the
relationship. (4)
9.2 Using the data for the “average wage” variable, calculate the mean, variance and
standard deviation. (6)
9.3 Calculate the covariance between the two variables. (4)
4.4 Activity 4.2: A review of basic statistics and econometrics (continued)
1. Given that X = 25, Sx = 9.5, n = 30 and k = 3, calculate the t-values and compare these to
the critical values for performing the t-test in the following cases:
1.1 H0: ux = 18.5 and H1: ux 18.5 at 5% and 10%
1.2 H0: ux = 18.5 and H1: ux > 18.5 at 5% and 10%
1.3 Interpret your results for each. (10)
2. Explain the properties of OLS estimators (hint: BLUE properties). (4)
3. Discuss the central limit theorem. (4)
4. List and briefly explain four properties of a correlation coefficient. (4)
5. Discuss the parts which make up the total sum of squares (TSS). (4)
4.5 Activity 4.3: An introduction to EViews
Consider the annual data below, which shows the number of new passenger cars sold in the
United States from 1971 to 1986. The variables used are:
NPC: new passenger cars sold (thousands)
NC: new cars consumer price index (1967 = 100)
CPI: consumer price index, all items, all urban consumers (1967 = 100)
PDI: real personal disposable income (billions of dollars)
INT: interest rate (percent)
ELF: employed civilian labour force (thousands)
Year NPC NC CPI PDI INT ELF
1971 10227 112 121.3 776.8 4.89 79367
1972 10872 111 125.3 839.6 4.55 82153
1973 11350 111.1 133.1 949.8 7.38 85064
1974 8775 117.5 147.7 1038.4 8.61 86794
1975 8539 127.6 161.2 1142.8 6.16 85846
1976 9994 135.7 170.5 1252.6 5.22 88752
1977 11046 142.9 181.5 1379.3 5.5 92017
1978 11164 153.8 195.3 1551.2 7.78 96048
1979 10559 166 217.7 1729.3 10.25 98824
1980 8979 179.3 247 1918 11.28 99303
1981 8535 190.2 272.3 2127.6 13.73 100397
1982 7980 197.6 286.6 2261.4 11.2 99526
1983 9179 202.6 297.4 2428.1 8.69 100834
1984 10394 208.5 307.6 2670.6 9.65 105005
1985 11039 215.2 318.5 2841.1 7.75 107150
1986 11450 224.2 323.4 3022.1 6.31 109597
Questions
1. According to economic theory, we expect the number of new cars sold to be negatively
correlated with vehicle prices (i.e. higher prices mean people (should) buy fewer cars).
What relationship do we expect to exist between the number of new cars sold (NPC) and
the other variables? Explain.
2. Estimate the following equations:
a) NPCₜ = 𝛃𝟎 + 𝛃𝟏NCₜ + 𝛃𝟐CPIₜ + 𝛃𝟑PDIₜ + 𝛃𝟒INTₜ + 𝛃𝟓ELFₜ + 𝜇𝑡
b) log(NPCₜ) = 𝛃𝟎 + 𝛃𝟏log(NCₜ) + 𝛃𝟐log(CPIₜ) + 𝛃𝟑log(PDIₜ) + 𝛃𝟒log(INTₜ) + 𝛃𝟓log(ELFₜ) + 𝜇𝑡
3. Give an economic and statistical interpretation of your output (note the difference in
interpretation between (a) and (b) above).
4.6 Recommended reading
Statistical principles:
Link to ECS3706 (Econometrics), study unit 17 (to be posted on myUnisa/Additional Resources)
Textbook:
Wooldridge, Math Refreshers A, B and C (not the detailed derivations, rather focus on the
concepts as mentioned in the ECS3706 guide).
Gujarati, DN & Porter, DC. 2009. Basic Econometrics. 5th edition. New York: McGraw-
Hill/Irwin.
