Addis Ababa Medical & Business College
INTRODUCTION TO ECONOMETRICS
FOR
DEPARTMENT OF ACCOUNTING & FINANCE AND
BUSINESS MANAGEMENT
1-1
CHAPTER 4: Violations of Econometric Assumptions
(Topics in Multiple regression)
Introduction
Recall that in the classical model we have
assumed
Zero mean of the random term
Constant variance of the error term
(i.e., the assumption of homoscedasticity)
No autocorrelation of the error term
Normality of the error term
No multicollinearity among the explanatory variable. 1-2
Question
1-3
Reasons for Heteroscedasticity
The presence of outliers.
That is, the variance of Ui’s increase as income
increases. This means high-income families
show a much greater variability in their
saving behavior than do low income
families.
Violating the assumption that the
regression model is correctly specified.
Heteroscedasticity often occurs in
cross-sectional data.
1-4
The consequence of Heteroscedasticity
I f t h e a s s u m p t ion of h om os c e d a s t ic
disturbance is not fulf il led we have the
following consequences:
i. If U is heteroscedastic, the OLS estimates
do not have the minimum variance property
in the class of unbiased estimators.
ii. The coeff ic ient estimates would still be
statistically unbiased.
iii. The prediction (of Y for a given value of X)
wou l d be ine ff ic ie nt be ca u s e of high
variance.
1-5
Detecting the problem of Heteroscedasticity
(Tests of heteroscedasticity)
1-6
1-7
1-8
Decision rule:
1-9
2. Breusch-Pagan –Godfrey (BPG) test
ESS
cal
2
2
1-10
Decision rule:
Reject the null hypothesis of homoscedasticity:
H o
: ... 0
1 2 k
if : cal K
2 2
Where K is the critical value from the
2
Chi-square distribution with K degrees of freedom
for a given value of α .
Illustrative example: Consider the following data on
consumption expenditure (Y) and income (X) for 20
households (both in thousands of Dollars):
1-11
Test the hypothesis of homoscedasticity
using Goldfeld-Quandt and
Breusch-Pagan –Godfrey (BPG) tests.
1-12
1-13
1-14
ESS 12 . 132
2
cal
6 . 066
2 2
K
2 2
0 . 05
1 3 . 841 .
1-15
Decision:
Since 6 . 066
2
cal
is greater than the tabulated
value, we reject the null hypothesis of
homoscedasticity at the 5% significance level.
1-16
GROUP ASSIGNMENT
Discuss and give illustrative examples for the
following Violations of Econometric Assumptions.
1. AUTOCORRELATION
2. MULTICOLLINEARITY
1-17
Wednesday, July 17, 2024 1-18
