PART B : Quantitative Methods in Economics

1. Mathematical Methods in Economics:

Differentiation and Integration and their application in economics.
Optimisation techniques,
Sets, Matrices and their application in economics.
Linear algebra
Linear programming in economics
Input-output model of Leontief.

2. Statistical and Econometric Methods:

Measures of central tendency and dispersions,
Correlation and Regression.
Time series.
Index numbers.
Sampling of curves based on various linear and non-linear function.
Least square methods and other multivariate analysis (only concepts
and interpretation of results).
Analysis of Variance,
Factor analysis,
Principle component analysis,
Discriminant analysis.
Income distribution: Pareto law of Distribution,
longnormal distribution,
measurement of income inequality.
Lorenz curve and
Gini coefficient.
Univariate and multivariate regression analysis.
Problems and remedies of
Hetroscedasticity,
Autocorrelation and
Multicollnearity.

PART B : Quantitative Methods in Economics

1. Mathematical Methods in Economics:

Differentiation and Integration and their application in economics.

Q.Distinguish between Differentiation and Integration. Explain their

application in economics with suitable examples. .(15M)(2014)
Optimisation techniques,

Q. State the Kuhn- Tucker conditions. (2009)(7M)

Q. What is optimization problem in economics ? How does linear

programming technique help in assigning optimal solution in given
resource use ? Explain.(25M)(2014)

Q. Consider the optimisation problem : Maximize u(x1,x2)

Subject to M = p1x1+p2x2
Where M, p1 and p2 are positive constants. Write down the Lagrangian for
this problem and explain why you need to assume that an interior solution
exists before using the Lagrangian method to solve the
problem.(5M)(2018)

Sets, Matrices and their application in economics.

Q. Define a quadratic form, and state the conditions under which it is (i)
positive definite; (ii) positive semi-definite, and (iii) negative
definite.(15M)(2009)

Linear algebra and

Linear programming in economics and

Q.Using graphical approach, solve the following

linear programming problem :
Minimize : C = 80X + 60Y
Subject to : 2X + 2Y ≥ 3
2X + OY ≥ 1
X, Y ≥ 0 (5M)(2011)

Q.Write dual of the following linear programme and solve

the obtained dual graphically : (15M)(2013)
Minimise : Z = 3x1+3x2
subject to :
x1 + 2x2 ≥ 1
2x1 +x2 ≥ 1
x1 ≥ 0, x2 ≥ 0
Q. Four products A. B. C, D are to be bought to satisfy minimum
requirements of calories and Vitamin (which are 18 and 10 units
respectively) at miniMum cost. Formulate linear programming problem
using additional information given below : ..
Product type A B C D
Calorie content 2 0 1 3

Vitamin content 0 3 1 4

Price per unit 5 10 12 15

(15M)(2010)

Q. What is the dual problem in Linear programming ?Explain its use with
suitable examples. (2012)(12M)

Q. Distinguish between basic feasible solution, feasible solution and

optimal solution of a Linear Programming Problem(LPP). Solve the
following LPP graphically.
Maximize Y=q1+2q2
Subject to q1+ 3q2 ≤ 18
2q1+ q2 ≤ 14
q1,q2 ≥ 0

Input-output model of Leontief.

Q.Explain Leontief's static input-output model and describe its

limitations.(15M)(2009)

Q.What are the basic features and the limitations of Leontiefs input -
output model ?(5M)(2012)

Q. Describe Leontief's static input-output model. Calculate outputs of two

industries from following input-output table if final use for outputs of the
two industries increase to 15 and 20 units, respectively.
Industry I Industry II Final use Total
output
Industry I 20 9 11 40

Industry II 12 6 12 30

Figures are in physical units.(30M)(2010)

Q. Explain 'Leontief lnverse' in the input-output model suggested by

W.W. Leontief. (5M)(2013)
Q. Find out the total demand for industries 1, 2 and 3, if the coefficient
matrix A and final vector B are given as
0.3 0.4 0.1 20
A  0.5 0.2 0.6 B  10  (2015)(5M)
0.1 0.3 0.1 30

(2015)(5M)

Q. (a).Consider the Leontief static input-output model along with its

assumptions. How can you confirm that the model is either open or
closed?
State the fundamental objective of Leontief static open input-output
model.(8M)(2015)
(b) State the Hawkins-Simon condition and explain its economic meaning
and significance.(8M)(2015)
(c) Derive the consumption possibility locus.(9M)(2015)

Q. (a) Describe the Leontief static open input-output model along with its
assumptions. (5M)(2017)

(b) State the Hawkins.Simon conditions and explain their economic

meaning and significance. (5M)(2017)

(c) Find out the total demand for industries 1, 2 and 3 if the coefficient
matrix A and the final demand vector B are as follows:
0.3 0.4 0.1 20
A  0.5 0.2 0.6 B  10  (15M)(2017)
0.1 0.3 0.1 30

Q.Assume that there are three sectors .The input coefficient matrix A and
the final demand vector d is given as follows :

0.3 0.4 0.2 100

A  0.2 0 0.5 B   30  Would the amount of the primary input
 0.1 0.3 0.1  30 
required be consistent with what is available in the economy.?
(15M)(2018)

2. Statistical and Econometric Methods:

Measures of central tendency and dispersions,

Correlation and Regression.

a) In computing the correlation· coefficient between two variables X and

Y from 25 pairs of observations, the intermediate results are
n = 25, ΣXi·= 125 , ΣXi 2= 650, ΣYi =100 ΣYi 2 = 460, ΣXiYi = 508.
Later on at the time of checking it was found that two pairs of
observations which should be
X 8 12
Y 6 8
had been incorrectly recorded as
X 6 14
Y 8 6
Calculate the correct value of the correlation coefficient. (15M)(2011)

(b) Find out the variance of the numbers 1, 2, 3, .... , 50 and the
coefficient of variation. What is the advantage of computing the
coefficient of variation over the variance ? (5+3+2=10M)(2011)

Q. Consider two regression equations of y on x and x on y. S how that the

arithmetic mean of two rcgression coefficients is greater than the
correlation coeflicient. provided the correlation coefficient is pos itive.
Under which condition do the two regression lines coincide ? (12M)(2017)

(b ) In a bivariate di stribution, a researcher gets two lines of regression as

2x - y + 1 =0
3x - 2y+7=0
Identify the true r egression lines and find the mean of x and y. (6M)(2017)

Q. Heights of fathers (X) and Sons( Y) in inches are given in the

following table.
X 65 66 67 67 68 69 70 72
Y 67 68 65 68 72 72 69 71

(a) Calculate the correlation coefficent between the heights of fathers and
those of sons. (6M)(2018)
(b) Obtain the equations of lines of regression and estimate of X for
Y=70. (6M)(2018)
(c) Given that, X=4Y+5 and Y=kX+4 are the lines of regression of X on
Y and X respectively. Show that 0<4k<1. If k=1/16, what is the point of
intersection of two regression lines? (6M)(2018)

Time series.
Q. Define the terms ' white noise' and 'random walk' in time series
analysis.(5M)(2012)

Q. Explain the concept of co-integration in a timeseries analysis.

(5M)(2014)

Q.What is stationarity in time series analysis? Show that a random work

model is non stationary. Discuss the Dickey fuller test for stationarity.
(15M)(2016)

Index numbers

Q.What is a superlative index number ? How is it related to the theory of

aggregation ?(15M/5M)(2009/2015)

Sampling of curves based on various linear and non-linear function

Hypthesis testing

Q.Distinguish between parametric and non-parametric tests in testing of

hypotheses.(10M)(2009)

Q.What are type I and type II errors ? Why ts probability of type I error
fixed in a hypothesis testing problem ?(15M)(2010)

Q. Explain the terms as follows and their importance in the context of

inference analysis : Degrees of freedom, Level of significance, and Power
of the test. (4X3=12M )(2011)

Q. What are type I and type II errors in testing of a

hypothesis ? ·(5M)(2013)
Q.Define and distinguish between level of significance and confidence
in:erval. What do you mean by 'power of the test' (2015)(5M)

Q.Define level of significance. How is this level decided for a given

problem? Can we take it as 2 % or 6%? Explain. (2016)(5M)
Q.Distinguish between point estimation and interval estimation of
population parameter. State the small sample properties of a good
estimator. (15M)(2016)

Least square methods and other multivariate analysis (only concepts

and interpretation of results).

Q.State and explain the assumptions of two-variable linear regression

model.(2010)(7M)

Q.Explain the principle of least squares as a basis for multiple regression

analysis. Also state the underlying standard assumptions of ordinary least
squares estimation. Explain further the consequences of violation of such
assumptions.(30M)(2009)

Q. State the assumptions of classical linear regression model. Why are the
regressors(X) assumed to be non stochastic in repeated samples.
(5M)(2016)

Q. State and explain the assumptions for applying ordinary least squares
(OLS) method to two variable linear regression model :
Y = bo + b1Xt + ut t= 1, 2, ......... , n (15M)(2013)
Consider a simple model of classical regression as
Yi = βX i + u i, where ui stands for random disturbance term with the
standard assumptions and u i ~N(0, σ2), and X i is non-stochastic and
i = 1, 2, ... , n.
(a) Find out the OLS estimator for β, say ̂OLS .(5M)(2015)
(b) Show that the OLS estimator for β is BLUE. Prove ab-initio.
(10M)(2015
Y
(c) Prove that   , where Y and X are means X respectively, is
X
unbiased but less efficient estimator of β than ̂OLS

Q. Given the Classical Linear Regression model with usual assumptions

Yi=β0+β1Xi+Ui i=1,2,..n
A.Examine the goodness of fit of the model using
ANOVA .(10M)(2016)
2
B. If the value of R is low , how it can be improved.(10M)(2016)
Analysis of Variance,
Factor analysis,
Principle component analysis,
Q. Briefly discuss the principal component analysis and the rationale for
its use.(10+3=13M)(2011)

Discriminant analysis.
Income distribution:
Q. How is distributional inequality of various kinds measured with the help of
income as a resource?Name some common inequality measures and state their
properties.(20M)(2016)

Pareto law of Distribution,

Q. Explain how Pareto's law of distribution is useful in measuring income
distribution.(2009)(7M)

Q.Define Pareto’s law of income distribution and state its application.

How is the Pareto distribution reated to he Log-Normal Distribution? For
the Pareto distribution, calculate the Lorenz curve and Gini
coefficient.Explain their meanings.(25M)(2018)

lognormal distribution,

Q.What is log-normal distribution ? Where is it used in economic

analysis ?(2010)(7M)

measurement of income inequality.

Q.How does Lorenz curve explain income inequality ? Explain with one
suitable example. Define Gini coefficient with the help of Lorenz curve
and show
That Gini = [1- 2 x (Area below .Lorenz curve)]. (5+5+15=25)(2011)

Lorenz curve and

Q.How do you draw a Lorenz curve ? Explain its use? (2009)(7M)

Q.) How can you measure in ome inequality by using Lorenz curve
method ?(5M)(2012)
Gini coefficient.
Q.Describe Gini 's coefficient as a measure of inequality. (5M)(2013)

Univariate and multivariate regression analysis.

Problems and remedies of
Hetroscedasticity,

Q.a) Define heteroscedasticity..(5M)(2016)

(b) Explain:
I. Consequences of heteroscedasticity on OLS estimates.(5M)(2016)
II. Detection of heteroscedasticity in a model .(5M)(2016)
III. Estimation procedure in the presence of
heteroscedasticity. .(5M)(2016)

Q. (a) Explain what do you mean by heteroscedasticity. (5M)(2017)

(b) Given the model Yi = β1 + β2X2i+ui where E(ui2)=σ2Xi2 and i = 1, 2,
3, ... , n, find out the OLS and GLS variance of the regression
slope. (12M)(2017)
" (c) Show that if Xi takes the values 1 , 2, 3, 4, 5, Var( ˆ ) > Var(b2 ), where
ˆ is thE OLS estimator of β2 and b2 is the GLS estimator of β2.
(8M)(2017)

Autocorrelation

Q. What is autocorrelation ? How can we detect it ? How can it be

removed from a single equation model? .(15M)(2014)

Q. (A) Define autocorrelation and state what are the possible sources of
autocorrelation. (5M)(2018)
(b)Suppose that the time series data follows the auto regressive scheme of
order one, that is, AR(1). Show that an AR(1) process is simply an MA
(∞) process (that is, moving moving average scheme of order
infinity).(5M)(2018)
(C ) Find the mean and variance if the time series data are modelled by
the process
Y t=a+ Yt-1+ɛt
Where
Ɛt is a pure white noise. Find out also, the autocorrelation coeeficient of
the sth order. Interpret your results. How do you test stationarity in this
case.(15M)(2018)

Multicollnearity

Q. What is the problem of multicollinearity in a regression model ? What

is its plausibility ? Explain Farrar - Glauber method to detect it. How can
it be
removed?(25M)(2014)

Q.(a)What do you mean by multicollincarity ? (5M)(2017)

(b)How does it affect the precision of estimates? (5M)(2017)
(c) Consider a simple model
y i = β2x2i+ β3x3i+ui , ; i = 1, 2, ... , n
and the variables are in deviation fonn. The disturbance term ui
satisfies all the classical assumptions. Suppose x2i; and x3i are
multicollinear . Should you drop either x2i, or x3i, Lo have precise estimates
of the remaining parameters ? If so, under what condition are you
permitted to do so? (15M)(2017)

Spurious regression

Ql2. ( a ) Explain the meaning of spurious regression. (3M)(2017)

(b) How are the values of Durbin-Watson d statis tic and R2 indicative of
spurious regression ? (4M)(2017)

(c) Show that in the case of spurious regression between Yt and Xt,. where
both Yt and Xt are generated by random walks, (i) the errors have a
pe rmanent effect; (ii) the variance of the errors is infinitely large. What
should you interpret from your result ? (18M)(2017)