BULE HORA UNIVERSITY

COLLEGE OF BUSINESS AND ECONOMICS

DEPARTMENT OF ACCOUNTING AND FINANCE

Course title:- Econometrics

Group Assignment :- Group 3

By

1. Tigistu Gizaw

2. Zegeye Tiruneh

3.Fasika Selemon

4. Eyerusalem sisay

5. Jemal Umer

6.Mihret Balcha

February 17,2022
Attempt all of the following questions

1. Heteroscedasticity

I. What is meant by heteroscedasticity? Why is heteroscedasticity a

problem?

Specifically, heteroscedasticity is a systematic change in the spread of the residuals

over the range of measured values.

Heteroscedasticity is a problem because ordinary least squares (OLS) regression

assumes that all residuals are drawn from a population that has a constant variance
(homoscedasticity).

ii. Draw a figure showing various forms of homosdastic and heteroscedastic

disturbances.

CLRM relies on the error term variance being constant. Homoskedasticity, refers
to a situation where the error has the same variance regardless of the value(s)
taken by the independent variable(s).

Homoskedasticity is expressed as :-

where Xi represents a vector of values for each individual and for all the
independent variables.

As you can see, when the error term is homoskedastic, the dispersion of the error
remains the same over the range of observations and regardless of functional form.
If the error term is heteroskedastic, the dispersion of the error changes over the
range of observations. The heteroskedasticity patterns depicted are only a couple
among many possible patterns. Any error variance that doesn’t resemble that in
the previous figure is likely to be heteroskedastic.
 2. Multicollinearity

i. What is meant by perfect multicollinearity?

Perfect multicollinearity occurs when two or more independent variables in a

regression model exhibit a deterministic (perfectly predictable or containing no
randomness) linear relationship.

If two or more independent variables have an exact linear relationship between

them then we have perfect multicollinearity.

ii. Show that the regression coefficients are indeterminate in the presence of
prefect MC?
Let us first consider the case of perfect multicollinearity, defining the deviations &
form sample mean:
y, = Y; - Y, x2, = X2, - X2 and x3, = X3,- X3

where Y, X2, X3 are the sample means of the variables respectively. By definition
the coefficient p2 in MLRM is calculated as,

For perfect multicollinearity in our earlier example, we have assumed X3=l OX,.
Let us make it more general and assume X3=UY2, where h is a constant as before.
This implies that in deviations form we can write x3 = k2.~substituting this in
equation we get,

which is an indeterminate expression. Thus the coefficient P2 is not determinate,

i.e., it cannot be estimated. Similarly we can show that the coefficient for P3 is also
indeterminate in the presence of perfect multicollinearity.

iii. List out the practical consequences of existence of multicollinearity

The consequences of multicollinearity can be statistical or numerical:

• Statistical consequences of multicollinearity include difficulties in testing

individual regression coefficients due to inflated standard errors. Thus, you
may be unable to declare an X variable significant even though (by itself) it
has a strong relationship with Y.
 • Numerical consequences of multicollinearity include difficulties in the
computer's calculations due to numerical instability. In extreme cases, the
computer may try to divide by zero and thus fail to complete the analysis.
Or, even worse, the computer may complete the analysis but then report
meaningless, wildly incorrect numbers

3. Autocorrelation
i. Give brief definition of autocorrelation? What problems would occur if
we continue to use the OLS estimators in the presence of autocorrelation?

Auto correlation is a characteristic of data which shows the degree of similarity

between the values of the same variables over successive time intervals

What problems would occur if we continue to use the OLS estimators in the
presence of autocorrelation?

The consequences of the OLS estimators in the presence of Autocorrelation can be

summarized as follows:

ii. Briefly explain the Cochrane – Orcutt transformation and state how it
would help to resolve the problem of autocorrelation

Cochrane – Orcutt is a procedure used to adjust a linear model for serial

correlation in the error term.
 • It is a procedure to transform the variables to reduce the autocorrelation of
the error terms
➢ Cochrane – Orcutt transformation Resolves the problem of Autocorrelation by
certain transformations on all variables can be performed for the AR (1) term.
These transformations aim at performing repeated iterative steps to minimize
the squared sum of errors in the regression model
4. A)Calculate the coefficient of parameters using OLS regression equation.
We want to fit a multiple linear regression model:-
𝑌𝑖 = 𝛽0 + 𝛽1 𝑋1𝑖 + 𝛽2 𝑋2𝑖 + 𝜀𝑖 𝑖 = 1,2, . . . ,15
𝑇𝑜 𝑠𝑖𝑚𝑝𝑙𝑖𝑦𝑓𝑦 𝑡ℎ𝑒 𝑐𝑎𝑙𝑐𝑢𝑙𝑎𝑡𝑖𝑜𝑛, 𝐼𝑡 𝑖𝑠 𝑏𝑒𝑡𝑡𝑒𝑟 𝑡𝑜 𝑤𝑜𝑟𝑘 𝑤𝑖𝑡ℎ 𝑑𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛:
𝑦𝑖 = 𝑌𝑖 − 𝑌, 𝑥1𝑖 = 𝑋1𝑖 − 𝑋1 𝑎𝑛𝑑𝑥2𝑖 = 𝑋2𝑖 − 𝑋2 .
Summary Statistics:
𝑛 = 15, ∑ 𝑥1 𝑦 = −505, ∑ 𝑥2 𝑦 = 107,500, ∑ 𝑥1 𝑥2 = −11,900,

∑ 𝑥1 2 = 60, ∑ 𝑥2 2 = 2,800,000, ∑ 𝑦 2 = 4,600, 𝑌 = 70,

𝑋1 = 6, 𝑋2 = 1,100
OLS estimates of the regression coefficients are:-
∧ [∑ 𝑥1 𝑦][∑ 𝑥2 2 ] − [∑ 𝑥2 𝑦][∑ 𝑥1 𝑥2 ]
𝛽1 =
[∑ 𝑥1 2 ][∑ 𝑥2 2 ] − [∑ 𝑥1 𝑥2 ]2
[−505][2,800,000] − [107,500][−11,900]
= = −𝟓. 𝟏𝟎𝟔𝟏
[60][2,800,000] − [−11,900]2
∧ [∑ 𝑥2 𝑦][∑ 𝑥1 2 ] − [∑ 𝑥1 𝑦][∑ 𝑥1 𝑥2 ]
𝛽2 =
[∑ 𝑥2 2 ][∑ 𝑥1 2 ] − [∑ 𝑥1 𝑥2 ]2
[107,500][60] − [−505][−11,900]
= = 𝟎. 𝟎𝟏𝟔𝟔𝟗𝟐
[2,800,000][60] − [−11,900]2
∧ ∧ ∧
𝛽0 = 𝑌 − 𝛽1 𝑋1 − 𝛽2 𝑋2
= 70 − (−5.1061)(6) − (0.016692)(1,100) = 𝟖𝟐. 𝟐𝟕𝟓𝟒

b)Write the estimated regression equation.

Hence, the estimated model is:
∧
𝒀 = 𝟖𝟐. 𝟐𝟕𝟓𝟒 − 𝟓. 𝟏𝟎𝟔𝟏𝑿𝟏 + 𝟎. 𝟎𝟏𝟔𝟔𝟗𝟐𝑿𝟐
c)Compute the standard error of the regression coefficients and conduct test of
significance at the 5% level of significance.

𝑇ℎ𝑒 𝑒𝑠𝑡𝑖𝑚𝑎𝑡𝑒𝑑 𝑒𝑟𝑟𝑜𝑟(𝑟𝑒𝑠𝑖𝑑𝑢𝑎𝑙𝑠)𝑎𝑟𝑒:

∧ ∧
𝜀𝑖 = 𝑌𝑖 − 𝑌𝑖 = 𝑌𝑖 − 𝟖𝟐. 𝟐𝟕𝟓𝟒 − 𝟓. 𝟏𝟎𝟔𝟏𝑿𝟏 + 𝟎. 𝟎𝟏𝟔𝟔𝟗𝟐𝑿𝟐
∧
𝑇ℎ𝑒𝐸 𝑟𝑟𝑜𝑟 𝑆𝑢𝑚 𝑜𝑓𝑆 𝑞𝑢𝑎𝑟𝑒(𝐸𝑆𝑆) = ∑ 𝜀𝑖 2 = 𝟒, 𝟑𝟕𝟐. 𝟗𝟔𝟑

∧ ∧ ∧
𝑜𝑟 𝐸𝑟𝑟𝑜𝑟𝑆𝑢𝑚𝑜𝑓𝑆𝑞𝑢𝑎𝑟𝑒(𝐸𝑆𝑆) = ∑ 𝜀𝑖 2 = 𝛽1 ∑ 𝑥1 𝑦 + 𝛽2 ∑ 𝑥2 𝑦
𝐴𝑛 𝑒𝑠𝑡𝑖𝑚𝑎𝑡𝑜𝑟 𝑜𝑓 𝑡ℎ𝑒 𝑒𝑟𝑟𝑜𝑟 𝑣𝑎𝑟 𝑖 𝑎𝑛𝑐𝑒𝜎 2 𝑖𝑠:
∧
∧ ∑ 𝜀𝑖 2
4,372.963
𝜎2 = = = 364.4135
𝑛−3 15 − 3
D)Compute the coefficient of determination (R2) and find the explained and
unexplained variation in the quantity demand.
Decomposition of the sample variation of Y
The variation in the independent variable Y can be decomposed in to:
Total Sum of Square=TSS ∑ 𝑦2 = 4,600
∧
𝐸𝑟𝑟𝑜𝑟𝑆𝑢𝑚𝑜𝑓𝑆𝑞𝑢𝑎𝑟𝑒(𝐸𝑆𝑆) = ∑ 𝜀𝑖 2 = 𝟐𝟐𝟕. 𝟎𝟑𝟕
⇒ 𝑅𝑒𝑔𝑟𝑒𝑠𝑠𝑖𝑜𝑛 𝑆 𝑢𝑚𝑜𝑓𝑆𝑞𝑢𝑎𝑟𝑒
𝑹𝑺𝑺 = 𝑻𝑺𝑺 − 𝑬𝑺𝑺 = 𝟒, 𝟑𝟕𝟐. 𝟗𝟔𝟑
𝑇ℎ𝑒𝑐𝑜𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡𝑜𝑓 𝑑𝑒𝑡𝑒𝑟𝑚𝑖𝑛𝑎𝑡𝑖𝑜𝑛 𝑖 𝑠𝑡ℎ𝑢𝑠:
𝑹𝑺𝑺
𝑹𝟐 = = 𝟎. 𝟗𝟓𝟎𝟔
𝑻𝑺𝑺
• 𝑹𝟐 =0.9506 indicates that 95% of the variation (change) in Demand is
attributed to the effect of price of the commodity & consumers income
• 1 - 𝑹𝟐 = 0.0494 This indicates that 4.94% of the variation in Demand is
due to factors( variables) not included in our specification.
Test of model adequacy
A test of model adequacy is accomplished by testing the null hypothesis:
𝐻0 : 𝛽1 = 𝛽2 = 0
𝐻𝐴 : 𝐻𝑜𝑖𝑠𝑛𝑜𝑡𝑡𝑟𝑢𝑒
 The test statistics for this test is given by
𝑅𝑆𝑆/(𝑘 − 1) 4,372.963/(3 − 1)
𝐹𝑐𝑎𝑙 = = = 115.5660
𝐸𝑆𝑆/(𝑛 − 𝑘) 227.037/(15 − 3)
𝑊𝑒𝑐𝑜𝑚𝑝𝑎𝑟𝑒𝑡ℎ𝑖𝑠𝐹 − 𝑟𝑎𝑡𝑖𝑜𝑤𝑖𝑡ℎ𝐹𝛼 (𝐾 − 1, 𝑛 − 𝐾)
= 𝐹𝛼 (2,12)𝑓𝑜𝑟𝑠𝑜𝑚𝑒𝑠𝑖𝑔𝑛𝑖𝑓𝑖𝑐𝑎𝑛𝑐𝑒𝑙𝑒𝑣𝑒𝑙𝛼.
• 𝐹𝑜𝑟𝛼 = 0.01, 𝐹𝛼 (𝐾 − 1, 𝑛 − 𝐾) = 𝐹0.01 (2,12) =
• 𝐹𝑜𝑟𝛼 = 0.05, 𝐹𝛼 (𝐾 − 1, 𝑛 − 𝐾) = 𝐹0.05 (2,12) =
Since the test statistics is greater than both tabulated values, the above ratio
is significant at the conventional levels of significance (1% and 5%). Thus,
we reject the null hypothesis and conclude that the model is adequate, that,
is, variation (change) in Quantity demand is significantly attributed to price
of the commodity & consumers income
e)If the interpret the results (regression coefficients).
f)Test for Multicollinearity with Variance Inflation Factor (VIF).
Using SPSS =0.9506 the VIF of is thus:
∧ 1 1
𝑉𝐼𝐹(𝛽2 ) = 2 = = 20.24
∧ 1 − 0.9506
1 − 𝑅2
g)Test for Autocorrelation with Durbin-Watson and with 5 % level of
significance.
The DW test statistics is computed as
∑𝑛𝑡=2(𝜀𝑡 − 𝜀𝑡−1 )2
𝑑𝑤 =
∑𝑛𝑡=1(𝜀𝑡 )2
• To test of 𝐻𝑜 : 𝜌 = 0𝑣𝑒𝑟𝑠𝑢𝑠𝐻𝐴 : 𝜌 ≻ 0, we can use the Durbin-
Watson lower (dL) and upper (du) bounds (critical values).
• Decision rule:
𝑅𝑒𝑗𝑒𝑐𝑡 𝐻𝑜 𝑖𝑓𝑑 < 𝑑𝐿
𝐷𝑜𝑛𝑜𝑡𝑟𝑒𝑗𝑒𝑐𝑡𝐻𝑜 𝑖𝑓𝑑 > 𝑑𝐿
𝑇ℎ𝑒𝑡𝑒𝑠𝑡𝑖𝑠𝑖𝑛𝑐𝑙𝑢𝑠𝑖𝑣𝑒𝑖𝑓𝑑𝐿 < 𝑑 < 𝑑𝑢
 5. A.Estimate the food function
  
Y =
i
 +  Xi i = 1, 2, ..., n

Summary statistics are:-

𝑥𝑖 = 𝑥𝑖 − 𝑥̅ and 𝑦𝑖 = 𝑦𝑖 − 𝑦̅
n=10 𝑥̅ = 59.6 𝑦̅ = 8.8 ∑𝑥𝑖2 = 394.4 ∑𝑦𝑖2 = 21.6 𝛴𝑥𝑖 𝑦𝑖=80.2
Estimation Of Regression Coefficients
𝛴×𝑦 80 ⋅ 2
𝛽̂ = = = 𝟎. 𝟐𝟎𝟑𝟑𝟒𝟕
∑𝑥 2 394 ⋅ 4
𝛼̂ = 𝑌̅ − 𝛽̂ 𝑋̅ = 8.8 − (0.203347)59.6 = −𝟑. 𝟑𝟏𝟗𝟒𝟖𝟏𝟐
Therefore the estimated regression equation is:-
𝑌̂ = 𝛼̂ + 𝛽̂ 𝑋
̂ = −𝟑. 𝟑𝟏𝟗𝟒𝟖𝟏𝟐 − 𝟎. 𝟐𝟎𝟑𝟑𝟒𝟕𝑿
𝒀
b. Compute the coefficient of determination and find the explained and
unexplained variation in the food expenditure.
Test of model adequacy
𝒏 𝒏

𝑻𝑺𝑺 = ∑(𝒀𝒊 − 𝒀)𝟐 = ∑ 𝒚𝒊 𝟐 = 𝟐𝟏. 𝟔

𝒊=𝟏 𝒊=𝟏
∧ ∧ ∧
𝑹𝑺𝑺 = ∑𝒏𝒊=𝟏(𝒀𝒊 − 𝒀)𝟐 = 𝜷𝟐 (∑𝒏𝒊=𝟏(𝑿𝒊 − 𝑿)𝟐 ) = 𝜷𝟐 ∑𝒏𝒊=𝟏 𝒙𝒊 𝟐
= (0.203347)2 (394.4) = 𝟏𝟔. 𝟑𝟎𝟖𝟒𝟒𝟎𝟗
ESS=TSS-RSS = 21.6 -16.3084409
= 5.2915591
R2=RSS/TSS = 16.3084409/21.6 = 0.75502
Thus we can conclude that
• About 75.5% of the variation in demand for food is due to GNP.
• The 24.5% of the variation in demand for food is not due to GNP but instead
due to other factors not included in the model
c. Compute the standard error of the regression coefficients and conduct test
of significance at the 5% level of significance.
𝟏 𝑬𝑺𝑺
̂𝟐 =
𝝈 ∑𝒊=𝟏 𝝐̂𝟐𝒊 = =5.2915591/10-2= 0.66144775
𝒏−𝟐 𝑵−𝟐
Thus an unbiased estimator of Var(𝛽̂ ) 𝑖𝑠 𝑔𝑖𝑣𝑒𝑛 𝑏𝑦: −
̂𝟐
𝝈
̂(𝛽̂ ) =
𝒗 =0.66144775/394.4= 0.001677098
∑𝒙𝟐𝒊

The standard error of 𝛽̂ is :-

s.e (𝛽 ̂(𝛽̂) = √0.001677098 = 0.04095238

̂ ) = √𝒗

ANOVA

df SS MS F Significance F
Regression 1 16.30842 16.30842 24.65564 0.001099363
Residual 8 5.291582 0.661448
Total 9 21.6

For 𝛼 = 0.05 the critical valuue from F- distribution is:

𝐹𝛼 (1, 𝑛 − 2)=𝐹0.05 (1,8) = 7.2
• Decision: Since the calculated variance ratio exceed the critical value, we
reject the null hypothesis of no linear R/ship b/n Demand for food and GNP
at the 5% level of significance. Thus, we then conclude that is significant, that
is, the linear regression model is adequate and is useful for prediction purposes.

6. Prove Guass-Markov theorem (Properties of OLS estimators)