University of Dhaka

Department of Development Studies

Bachelor of Social Science (Honors)
Course DS 301: Basic Econometrics

1. Describe the following

a) Random variable
b) Mean, variance and their properties of a random variable
c) Joint distribution, marginal distribution, and conditional distribution
d) Covariance and correlation among random variables
2. If 𝑋 and 𝑌 are two random variables, and 𝑎 and 𝑏 are two constants, then prove
the followings:
a) 𝐸 (𝑎𝑋 ± 𝑏𝑌) = 𝑎𝜇𝑥 ± 𝑏𝜇𝑦
b) 𝑉𝑎𝑟(𝑎𝑋 ± 𝑏𝑌) = 𝑎2 𝜎𝑥2 + 𝑏 2 𝜎𝑦2 ± 2𝑎𝑏 𝑐𝑜𝑣(𝑋, 𝑌)
c) 𝐼𝑓 𝑋 and 𝑌 are independent, then 𝑐𝑜𝑣(𝑎𝑋, 𝑏𝑌) = 0 and hence,
𝑉𝑎𝑟(𝑎𝑋 ± 𝑏𝑌) = 𝑎2 𝜎𝑥2 + 𝑏2 𝜎𝑦2
3. The following table gives the joint PDF of the discrete variables X and Y.

a) Interpret the value of 0.16.

b) Find the marginal PDF of X, and the marginal PDF of Y.
c) Find the following:
i. 𝐸 (𝑋), 𝐸 (𝑌), 𝜎𝑥2 , 𝜎𝑦2 , 𝑐𝑜𝑣(𝑋, 𝑌), 𝑟𝑥𝑦
ii. 𝐸(𝑌|𝑋 = 2) and 𝑣𝑎𝑟(𝑌|𝑋 = 2)
4. Consider the following PDF
𝑓 (𝑥, 𝑦) = 2 − 𝑥 − 𝑦 0 ≤ 𝑥 < 1; 0 ≤ 𝑦 ≤ 1
a) Verify that it is a joint density function.
𝑓(𝑥,𝑦)
b) Find 𝑓 (𝑥 ), 𝑓 (𝑦), and 𝑓(𝑦|𝑥) [Note: 𝑓 (𝑦|𝑥 ) = ]
𝑓(𝑥)
c) Find 𝐸 (𝑥 ), 𝐸 (𝑦), 𝜎𝑥2 , 𝜎𝑦2 , and 𝑐𝑜𝑣(𝑥, 𝑦)
d) Find 𝐸(𝑥 + 𝑦) and verify that 𝐸 (𝑥 + 𝑦) = 𝐸(𝑥 ) + 𝐸(𝑦)
e) Are 𝑋 and 𝑌 independent
f) Find 𝑣𝑎𝑟(𝑥 + 𝑦)
5. Let assume that 𝑋 and 𝑌 have the following distribution:
𝑓 (𝑥, 𝑦) = 𝑘𝑥 (3𝑦 2 + 1), 𝑓𝑜𝑟 0 ≤ 𝑥 ≤ 2, 0 ≤ 𝑦 ≤ 1
= 0 elsewhere
a) Determine 𝑘 for which the given distribution will be a joint density
1 2 1
function. [𝑁𝑜𝑡𝑒 ∫0 ∫0 𝑓 (𝑥, 𝑦)𝑑𝑥𝑑𝑦 = 1 , 𝑐ℎ𝑒𝑐𝑘 𝑘 = ]
4
𝑓(𝑥,𝑦)
b) Find 𝑓 (𝑥 ), 𝑓 (𝑦), and 𝑓(𝑦|𝑥) [Note: 𝑓 (𝑦|𝑥 ) = ]
𝑓(𝑥)
c) Find 𝐸 (𝑥 ), 𝐸 (𝑦), 𝜎𝑥2 , 𝜎𝑦2 , and 𝑐𝑜𝑣(𝑥, 𝑦)
d) Find 𝐸(𝑥 + 𝑦) and verify that 𝐸 (𝑥 + 𝑦) = 𝐸(𝑥 ) + 𝐸(𝑦)
e) Are 𝑋 and 𝑌 independent
f) Find 𝑣𝑎𝑟(𝑥 + 𝑦)
1 1 1/2 1
g) Find 𝑃 (0 < 𝑥 < 1, 4 < 𝑦 < 2) [𝑁𝑜𝑡𝑒 ∫1/4 ∫0 𝑓 (𝑥, 𝑦)𝑑𝑥𝑑𝑦]

6. Consider the following population regression model:

𝑌𝑖 = 𝛽1 + 𝛽2 𝑋𝑖 + 𝑢𝑖
Here, 𝑌𝑖 is the dependent variable, 𝑋 is an explanatory variable, 𝛽′ 𝑠 are the two
unknown parameters, and 𝑢𝑖 is the stochastic disturbance term.
a) Write the sample regression function.
b) Write the meaning of the following assumptions:
i. 𝐸 (𝑢𝑖 |𝑋) = 0
ii. 𝐸 (𝑢𝑖2 |𝑋) = 𝜎 2
iii. 𝐸(𝑢𝑖 , 𝑢𝑗 |𝑋) = 0
iv. 𝐸 (𝑢𝑖 𝑋𝑖 ) = 0
v. There are variabilities in 𝑋 values but they are fixed in repeated
sampling
c) Briefly explain the assumptions of Classical Linear Regression Model.
d) Explain the principles of Ordinary Least Square method.
e) Derive the following OLS estimators:
i. 𝛽̂1
ii. 𝛽̂2
iii. 𝜎̂ 2
f) Find the following:
i. 𝑣𝑎𝑟(𝛽̂1 )
ii. 𝑣𝑎𝑟(𝛽̂2 )
iii. 𝑐𝑜𝑣𝑎𝑟(𝛽̂1 , 𝛽̂2 )
g) State the features of variances and covariance of OLS estimators.
h) Prove the followings:
i. Sum of residual is zero
 ii. Fitted values and residuals are uncorrelated
iii. Explanatory variable(s) and residuals are uncorrelated
iv. Average of predicted/ fitted values is equal to the average of
actual Y value.
v. The variance of the dependent variable is equal to the variance of
the stochastic disturbance term.
vi. If you regress 𝑌 on 𝑌̂ , the intercept and slope will be 0 and 1
respectively.
vii. The regression coefficient is independent of origins but dependent
on scale of measurement.
viii. OLS estimators are BLUE (Gauss-Markov Theorem).
ix. ∑𝑦𝑦̂ = ∑𝑦̂ 2
𝑌 = 𝛼 + 𝛽𝑌̂ + 𝑢
𝛼̂ = 𝑌̅ − 𝛽̂𝑌̅
∑𝑦𝑦̂ ∑(𝑦̂ + 𝑢̂)𝑦̂ ∑𝑦̂ 2 + ∑𝑦̂𝑢̂ ∑𝑦̂ 2
𝛽̂ = = = = =1
∑𝑦̂ 2 ∑𝑦̂ 2 ∑𝑦̂ 2 ∑𝑦̂ 2
i) State the statistical properties of OLS estimators.
j) Describe the coefficient of determination and elucidate the processes of
measuring coefficient of determination.
k) State the properties of coefficient of determination.
l) State the properties of coefficient of correlation.
m) If 𝑌𝑖 = 𝛽1 + 𝛽2 𝑋𝑖 + 𝑢𝑖 and 𝑋𝑖 = 𝛼1 + 𝛼2 𝑌𝑖 + 𝑣𝑖 , show that 𝛼̂2 𝛽̂2 = 𝑟 2 where 𝑟
is the correlation coefficient.

7. Consider the following simple population regression function:

𝑌𝑖 = 𝛽1 + 𝛽2 𝑋𝑖 + 𝑢𝑖 𝑓𝑜𝑟 𝑖 = 1,2, … , 𝑛
The corresponding sample regression function is given by:
𝑌̂𝑖 = 𝛽̂1 + 𝛽̂2 𝑋𝑖
Suppose that you have regressed 𝑌𝑖 on 𝑌̂𝑖 and your model is 𝑌𝑖 = 𝛼1 + 𝛼2 𝑌̂𝑖 + 𝑣𝑖 ,
here 𝑣𝑖 is a random variable having the properties of (i) 𝐸 (𝑣𝑖 ) = 0, (ii) 𝐸 (𝑣𝑖2 ) =
𝜎𝑣2 , and 𝐸(𝑣𝑖 𝑣𝑗 ) = 0. Find (i) 𝛼̂1 and compare it with 𝛽̂1 , (ii) 𝛼̂2 and compare it
with 𝛽̂2 , (iii) 𝑣𝑎𝑟(𝛼̂1 ) and compare it with 𝑣𝑎𝑟(𝛽̂1 ), (iv) 𝑣𝑎𝑟(𝛼̂2 ) and compare
it 𝑣𝑎𝑟(𝛽̂2 ), and (v) 𝑐𝑜𝑣(𝛼̂1 , 𝛼̂2 ) and compare it with 𝑐𝑜𝑣(𝛽̂1 , 𝛽̂2 )

8. Consider the following formulations of the two-variable PRF:

Model A: 𝑌𝑖 = 𝛽1 + 𝛽2 𝑋𝑖 + 𝑢𝑖 𝑓𝑜𝑟 𝑖 = 1,2, … , 𝑛
Model B: 𝑌𝑖 = 𝛼1 + 𝛽2 (𝑋𝑖 − 𝑋) + 𝑢𝑖
̅ 𝑓𝑜𝑟 𝑖 = 1,2, … , 𝑛
a) Find the estimators of 𝛽1 and 𝛼1 . Are they identical? Are their
variances identical?
 b) Find the estimators of 𝛽2 and 𝛼2. Are they identical? Are their
variances identical?

9. Consider the following formulations of the two-variable PRF:

Model A: 𝑌𝑖 = 𝛽1 + 𝛽2 𝑋𝑖 + 𝑢𝑖 𝑓𝑜𝑟 𝑖 = 1,2, … , 𝑛
Model B: 𝑦𝑖 = 𝛼1 + 𝛼2 𝑥𝑖 + 𝑣𝑖 𝑓𝑜𝑟 𝑖 = 1,2, … , 𝑛, 𝑥𝑖 = 𝑋𝑖 − 𝑋̅, 𝑦𝑖 = 𝑌𝑖 − 𝑌̅
a) Find the estimators of 𝛽1 and 𝛼1 . Are they identical? Are their
variances identical?
b) Find the estimators of 𝛽2 and 𝛼2. Are they identical? Are their
variances identical?

10. Regression on Standardized Variables: Consider the following formulations of

the two-variable PRF:
Model A: 𝑌𝑖 = 𝛽1 + 𝛽2 𝑋𝑖 + 𝑢𝑖 𝑓𝑜𝑟 𝑖 = 1,2, … , 𝑛
Model B: 𝑌𝑖 = 𝛼1 + 𝛼2 𝑋𝑖 + 𝑣𝑖
∗ ∗
𝑓𝑜𝑟 𝑖 = 1,2, … , 𝑛
𝑌𝑖 −𝑌̅ 𝑋𝑖 −𝑋̅
𝑌𝑖∗ = 𝑎𝑛𝑑 𝑋𝑖∗ =
𝑆𝑌 𝑆𝑋
a) Show that 𝑌̅ ∗ = 0 and 𝑆𝑌 ∗ = 1
b) Find the estimators of 𝛽1 and 𝛼1 . Are they identical? Are their
variances identical?
c) Find the estimators of 𝛽2 and 𝛼2. Are they identical? Are their variances
identical?

11. Consider the following simple population regression function:

𝑌𝑖 = 𝛽1 + 𝛽2 𝑋𝑖 + 𝑢𝑖 𝑓𝑜𝑟 𝑖 = 1,2, … , 𝑛
Assume that the model satisfies the assumptions of classical linear
regression model. Find the following matrices based on the above model:
a) The design matrix
b) The vector of parameters
c) The vector of errors
d) The vector of responses
e) Write the given model in compact form and specify the dimensions of
each of the matrices.
f) Derive the two normal equations and show that they give us (𝑋′𝑋)𝛽̂ =
𝑋′𝑌 where
𝑛 ∑𝑋𝑖
𝑋′𝑋 = [ ]
∑𝑋𝑖 ∑𝑋𝐼2
 ∑𝑌𝑖
𝑋′ 𝑌 = [ ]
∑𝑋𝑖 𝑌𝑖
g) Show that
𝛽̂ 1 ∑𝑋𝑖2 ∑𝑌𝑖 − ∑𝑋𝑖 ∑𝑋𝑖 𝑌𝑖
[ 1] = [ ]
𝛽̂2 𝑛∑𝑋𝑖2 − (∑𝑋𝑖 )2 −∑𝑋𝑖 ∑𝑌𝑖 + 𝑛∑𝑋𝑖 𝑌𝑖

12. Consider the following regression model:

𝑌𝑖 = 𝛽1 + 𝛽2 𝑋2𝑖 + 𝛽3 𝑋3𝑖 + ⋯ + 𝛽𝑘 𝑋𝑘𝑖 + 𝑢𝑖 𝑓𝑜𝑟 (𝑖 = 1,2, … 𝑛)
a) Expand the above equation into 𝑛 −simultaneous equations and write
the system of equations in matrix form.
b) State the dimensions of each of the matrices in the compact form of
system of equations: 𝑦 = X𝛃 + u
c) The strict exogeneity assumption of classical regression model (CRM)
𝑥11 … 𝑥1𝑘
is 𝐸(u|𝑿) = 𝟎, where 𝑿 = ⋮
[ … ⋮ ] is the data matrix.
𝑥𝑛1 … 𝑥𝑛𝑘
The residual for a hypothetical value of 𝜷, 𝛽̂ is given by

𝑢̂ = y − X𝛽̂

The sum of squared residuals (SSR)

𝑢̂′𝑢̂ = (y − X𝛽̂)′ (y − X𝛽̂)
The OLS estimate 𝛽̂ of 𝛽 can be obtained from min 𝑢̂′𝑢̂.
̂
𝛽

Show that 𝛽̂ = (𝑿′𝑿)−1 𝑿′𝒚

d) How does one know that the second order condition for minimization
holds?
e) Prove that under the following assumptions: I. linearity, II. strict
exogeneity, III. no multicollinearity, IV. spherical error variance
(homoskedasticity and no correlation between observations),
𝑽𝒂𝒓(𝛽̂ |𝑿) = 𝝈𝟐 (𝑿′𝑿)−1
f) Prove that under the assumptions I-IV, the OLS estimator is efficient in
the class of linear unbiased estimators.
13. The following table shows the weekly consumption and income of 60
households in an area.
Weekly Family Income (X)
80 100 120 140 160 180 200 220 240 260
55 65 79 80 102 110 120 135 137 150
60 70 84 93 107 115 136 137 145 152
Weekly
65 74 90 95 110 120 140 140 155 175
Family
70 80 94 103 116 130 144 152 165 178
Consumptio
75 85 98 108 118 135 145 157 175 180
n (Y)
88 113 125 140 160 189 185
115 162 191

a. Find 𝑃 (𝑌|𝑋𝑖 ), 𝐸(𝑌|𝑋 = 120) 𝑎𝑛𝑑 𝑉𝑎𝑟(𝑌|𝑋 = 200).

b. Draw one sample randomly from each income group and plot the
relationship between income and consumption expenditure.

c. Estimate the following regression model and interpret the results:

𝑌𝑖 = 𝛽1 + 𝛽2 𝑋𝑖 + 𝑒𝑖
where 𝑌𝑖 is annual income, 𝑋𝑖 is weekly family income and 𝑒𝑖 is the
stochastic disturbance term.
d. Compute the fitted values and residuals for each observation.
e. Predict the weekly consumption of the following family having income
of:
i. 110
ii. 210
iii. 290
f. Show the following:
i. ̅𝑌̂ = 𝑌̅
ii. ∑𝑒̂𝑖 = 0
iii. ∑𝑦̂𝑖 𝑒̂𝑖 = 0
iv. ∑𝑋𝑖 𝑒̂ 𝑖 = 0
g. Compute the following:
i. 𝜎̂ 2
ii. 𝑉𝑎𝑟(𝛽̂1 )
iii. 𝑉𝑎𝑟(𝛽̂2 )
iv. Covariance between 𝛽̂1 and 𝛽̂2 , that is, 𝐶𝑜𝑣(𝛽̂1 , 𝛽̂2 )
v. ESS, RSS, TSS, and 𝑅2
vi. 𝑟
14. Consider the following data related to expenditure (y) and income (x) [Note:
the variables are not in deviation form]
y x
70 80
65 100
90 120
95 140
110 160
115 180
120 200
140 220
155 240
150 260
a) Construct the vector of responses (Y), design matrix (X), vector of
parameters (𝛽), vector of errors (assume the error is given by 𝑢𝑖 ).
b) Find 𝑋 ′ 𝑋, 𝑎𝑛𝑑 𝑋′𝑌
c) Find 𝛽̂, 𝑣𝑎𝑟 − 𝑐𝑜𝑣(𝛽̂), 𝐸𝑆𝑆, 𝑇𝑆𝑆, 𝑅𝑆𝑆, 𝑅2
15. Interpret the following regression results.
a. A researcher obtained the following regression results:
𝑌̂𝑖 = −0.0144 + 0.724𝑋𝑖
𝑅2 = 0.9
Here, 𝑌𝑖 is the mean hourly wage and 𝑋𝑖 is the years of schooling.
b. The relationship between personal consumption expenditure (Y) and
Gross Domestic Product (GDP) is given below:
𝑌̂𝑖 = −299.59 + 0.722𝑋𝑖
𝑅2 = 0.99
c. The relationship between food expenditure (Y) and total expenditure (X)
is given below:
𝑌̂𝑖 = 94.21 + 0.437𝑋𝑖
𝑅2 = 0.37