OPTIMIZATION

METHODS FOR
ECONOMIC ANALYSIS
B.A. (PROGRAMME) ECONOMICS
SEMESTER III
DISCIPLINE SPECIFIC COURSE
DSC (MAJOR PAPER)

DEPARTMENT OF DISTANCE AND CONTINUING EDUCATION

UNIVERSITY OF DELHI
Optimisation Methods for Economic Analysis

Editors
Prof. J. Khuntia, Devender
Content Writers
Swarup Santra, Sarabjeet Kaur
Content Reviewer from the DDCE/COL/SOL
N. Shradha Varma
Academic Coordinator
Deekshant Awasthi

© Department of Distance and Continuing Education

ISBN: 978-81-19417-89-6
1st edition: 2023
E-mail: [email protected]
[email protected]

Published by:
Department of Distance and Continuing Education,
Campus of Open Learning/School of Open Learning,
University of Delhi, Delhi-110 007
Printed by:
School of Open Learning, University of Delhi

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

This Study Material is duly recommended and approved in Academic Council

meeting held on 11/08/2023 Vide item no. 1015 and subsequently Executive
Council Meeting held on 25/08/2023 vide item no. 1267.

Corrections/Modifications/Suggestions proposed by Statutory Body,

DU/Stakeholder/s in the Self Learning Material (SLM) will be incorporated in the
next edition. However, these corrections/modifications/suggestions will be
uploaded on the website https://sol.du.ac.in. Any feedback or suggestions may
be sent at the email- [email protected]

Printed at: Vikas Publishing House Pvt. Ltd. Plot 20/4, Site-IV, Industrial Area Sahibabad, Ghaziabad - 201 010 (1500 Copies)

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Optimisation Methods for Economic Analysis

SYLLABUS
Optimisation Methods for Economic Analysis
Syllabus Mapping

Unit – I : Comparative Statics Analysis Lesson 1: Comparative Statics

Derivatives, Slopes, Limit Theorem Analysis
(Pages 3-29)

Unit – II : Differentials and its Role in Comparative Statics Analysis Lesson 2: Differential and its
Role in Comparative
Statics Analysis
(Pages 33-66)

Unit – III : Optimisation Problems Lesson 3: Unconstrained

Unconstrained and Constrained Optimisation with Single and Multiple Optimisation
Variables, Lagrangian Functions, Quasi-Concavity and Convexity, Envelope Lesson 4: Constrained
Theorem Optimisation
Lesson 5: Convex and
Concave Functions
(Pages 69-156)

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Optimisation Methods for Economic Analysis

CONTENTS
UNIT I: COMPARATIVE STATICS ANALYSIS
LESSON 1 COMPARATIVE STATICS ANALYSIS 3-29

1.1 Introduction
1.2 Rate of Change and the Derivative
1.3 Derivatives and the Slopes
1.4 Concept of Limit
1.5 Limit Theorem
1.5.1 Theorems Involving a Single Function
1.5.2 Theorems Involving Two Functions
1.5.3 Limit of Polynomial Function
1.6 Continuity and Differentiability of a Function
1.6.1 Continuity of a Function
1.6.2 Differentiability of a Function
1.7 Summary
1.8 Glossary
1.9 Answers to In-Text Questions
1.10 Self-Assessment Questions
1.11 References

UNIT II: DIFFERENTIALS AND ITS ROLE IN

COMPARATIVE STATICS ANALYSIS
LESSON 2 DIFFERENTIAL AND ITS ROLE IN COMPARATIVE
STATICS ANALYSIS 33-66

2.1 Introduction
2.2 Rules of Differentiation for a Function of One Variable
2.2.1 Constant-Function Rule

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

2.2.2 Power Function

2.2.3 Power-Function Rule Generalised
2.2.4 Linear Function
2.2.5 Derivative of the Product of Two Functions
2.2.6 Derivative of the Quotient of Two Functions
2.2.7 Finding Marginal-Revenue Function from Average-Revenue Function
2.2.8 Elasticity of Demand
2.2.9 Relation between AC and MC
2.2.10 Derivative of a Composite Function (Chain Rule)
2.2.11 Inverse-Function Rule
2.3 Partial Differentiation
2.4 Applications to Comparative Statics Analysis
2.5 Jacobian Determinant
2.6 Summary
2.7 Glossary
2.8 Answers to In-Text Questions
2.9 References

UNIT III: OPTIMISATION PROBLEMS

LESSON 3 UNCONSTRAINED OPTIMISATION 69-100

3.1 Introduction
3.2 Optimal Values and Extreme Values
3.3 1st Derivative Test
3.4 Second and Higher Derivatives
3.5 Optimisation in Case of Two Variables
3.6 Differential Conditions for Optimisation
3.7 Convex and Concave Functions and Inflection Points
3.8 Summary
3.9 Glossary

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Optimisation Methods for Economic Analysis

3.10 Answers to In-Text Questions

3.11 Self-Assessment Questions
3.12 Reference
LESSON 4 CONSTRAINED OPTIMISATION 101-128

4.1 Introduction
4.2 Effects of a Constraint
4.3 Finding the Stationary Values
4.4 Constrained Optimisation by Substitution
4.5 The Lagrange Multiplier: Constrained Maximisation with Two Variables
4.6 Second-Order Condition
4.7 The Bordered Hessian
4.8 Summary
4.9 Glossary
4.10 Answers to In-Text Questions
4.11 Self-Assessment Questions
LESSON 5 CONVEX AND CONCAVE FUNCTIONS 129-156

5.1 Introduction
5.2 Convex Set
5.3 Convex Functions
5.4 Concave Functions
5.5 Test of Concavity and Convexity
5.6 Quasi-Concavity and Quasi-Convexity
5.7 Envelope Theorem
5.8 Summary
5.9 Glossary
5.10 Answers to In-Text Questions
5.11 Self-Assessment Questions
5.12 References

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 UNIT I: COMPARATIVE STATICS ANALYSIS

LESSON 1 COMPARATIVE STATICS ANALYSIS

Comparative Statics Analysis

LESSON 1 NOTES

COMPARATIVE STATICS ANALYSIS

Swarup Santra
Assistant Professor,
Department of Economics,
Satyawati College, University of Delhi

Structure
1.1 Introduction
1.2 Rate of Change and the Derivative
1.3 Derivatives and the Slopes
1.4 Concept of Limit
1.5 Limit Theorem
1.5.1 Theorems Involving a Single Function
1.5.2 Theorems Involving Two Functions
1.5.3 Limit of Polynomial Function
1.6 Continuity and Differentiability of a Function
1.6.1 Continuity of a Function
1.6.2 Differentiability of a Function
1.7 Summary
1.8 Glossary
1.9 Answers to In-Text Questions
1.10 Self-Assessment Questions
1.11 References

1.1 INTRODUCTION

Comparative statics is concerned with the comparison of different equilibrium states that
are associated with the comparison of different equilibrium states that are associated
with different sets of values of parameters and exogenous variables. We always start by
assuming the given initial equilibrium state. The initial equilibrium state can be changed if
Self-Instructional
the values of the parameter and/or exogenous variable change. Then we can get another Material 3

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES equilibrium state. The comparison of these two equilibrium states, i.e., initial and final due
to a change in parameters or exogenous variables is called comparative statics.
However, it must be noted that the comparative statics does not deal with the
adjustment process between the two equilibrium states. It actually compares the two
equilibrium states, the initial equilibrium (pre-change) and the final equilibrium (post-
change) state. It is also possible to have an unstable equation state. However, for
comparative statics, we assume that the new (final) equilibrium is defined and attainable.

Qualitative and quantitative

A comparative statics analysis can be either qualitative or quantitative in nature. For

example, if income increases, then consumption will also increase, it is a qualitative analysis.
Here, we are interested to know only the magnitude of changes. On the other hand, if we
want to determine the amount of changes in consumption with the one-unit change in
consumption with the one-unit change in income of an individual, then, we are intended
to calculate the magnitude of the change of the variable. This is a quantitative analysis.
One important point to remember is that if we get a quantitative result, we can
also get the qualitative result (direction of change) by obtaining the algebraic sign. The
rate of change of the equilibrium value of an indigenous variable with respect to change
in a particular parameter of an exogenous variable is called the ‘rate of change’. The
mathematics ‘derivative’ is helpful to get the rate of change. So the concept of a
derivative has an important role in a comparative statics analysis.

1.2 RATE OF CHANGE AND THE DERIVATIVE

Suppose, there are two variables, x and y. The variable, y is the dependent variable of
our interest. It depends on another variable, x. The functional form is presented as:
y = f(x).
In the comparative statics analysis, the variable, y will represent the equilibrium
value of an endogenous variable; and the variable, x will represent some parameter or
exogenous variables. For the simplicity of the discussion, we can start with a single
parameter, x.
Self-Instructional
4 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Comparative Statics Analysis

It involves the rate of change of the equilibrium values of y with respect to the NOTES
change in the parameter or exogenous variable, x. The rate of change can be measured
by the difference quotient.

Difference Quotient

When the value of x changes from x0 to x1, then the change of the value of x is
represented by the difference between the two values, x0 and x1. Technically, the
difference between two values of a variable can be represented by the Greek letter,
capital delta, x = (x1 – x0). When the value, x changes from x0 to x1, the equilibrium
value of y will also change from y0 = f (x0). to y1 = f (x1); where, f (x0) and f (x1) are the
values of the function f (x) at the points x = x0 and x = x1, respectively.
In other words, at the point, x = x1, the value of x is changed by x; hence,
x1 = x0+ x.
Therefore, the change of the equilibrium value of y is:
y  y1  y0
y  f  x1   f  x0 
After replacing the value of x1 by (x0 + x), we obtain;
y  f ( x0  x)  f ( x0 ) ...(1.1)
The rate of change is measured by the ratio of the change of y to the change of
x. The ratio is a quotient of the two differences. This is called the difference quotient.
It is to be noted that this is the average rate of change.
So, the difference quotient is measured by:
y f  x0  x   f  x0 
 ...(1.2)
x Δx
This quotient measures the average rate of change of y due to a change in x
between x0 and x1.
y
From Equation (1.2), we can say that the difference quotient, is the function
x
of x0 and x Such that;
y
 f  x0 , x  ...(1.3) Self-Instructional
x Material 5

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES The Derivative

If we need to know the rate of change of y when x is very small (a very small change
in x), we can take the help of the concept of derivative. When, x is very small
(approximately equal to zero, but not equal to zero), i.e., x  0, then it is possible to
y
have an approximation value of by dropping all the terms in the difference quotient,
x
involving the expression, x.
y
The smaller the value of x, the closer is the approximation of to the true
x
y
value of . The symbol lim can be read as ‘the limit of …. as x tends to zero’. If,
x x 0

y
as x  0, the limit of the difference quotient indeed exists, and that limit is called
x
the ‘derivative’ of the function, y = f (x).
Some important aspects of the derivative are discussed as follows:
a) A derivative is a derived function: In fact, a derivative is the derived function
of its original function. The original function, y = f (x) is called a primitive
function. The derivate function is derived from the primary function. The
difference quotient is a function of x0 and x. But, the derivative is a function of
x0 only, as x tends to zero in the derivative. Hence, it can be said that the
derivative is a function of x only. That means, for each value of x, there is a
unique corresponding value for the derivative function.
b) A derivative measures some rate of change: The derivative is a limit of the
difference quotient, and this difference quotient measures the rate of change of
y. Therefore, it can be said that the derivative must be a measure of some rate of
change. The change in x is infinitesimal, i.e., x  0, and the rate of change of
y, in derivative, is in the nature of an instantaneous rate of change.
c) Notation of the derivative: The notations of the derivative are of two types:
i) Notation by Lagrange: If the primitive function is = f (x), then the function
derived from it, the derivative function, can be written as f (x), or simply
Self-Instructional
f .
6 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Comparative Statics Analysis

dy NOTES
ii) Notation by Leibniz: The other type of notion of derivative is .
dx
The Lagrange notation of derivative, f (x) has an advantage that one can say
that this derivative is derived from the primitive function, f (x) and the derivative also
dy
depends on x. On the other hand, the Leibniz notation of derivative, , says that the
dx
derivate measures a rate of change.
Using these two notations, we may define the derivative of a given function,
y = f (x), as follows:
dy Δy
 f   x   lim ....(1.4)
dx  x  0 Δx

Example 1.1:
Given the function, y = 3x2 – 4;
a) Find the difference quotient as a function of x and x.

dy
b) Find the derivatives, .
dx
c) Find and f (3) and f (4).
Solution:

a) f  x   3 x 2  4 ; and f  x  Δx   3  x  Δx   4
2


Hence, Δy  f  x  Δx   f  x   3  x  Δx   4  3 x 2  4
2

 3x 2  6 xΔx   Δx   4  3x 2  4
2

 6 xΔx   Δx 
2

Hence, the difference quotient,

Δy 6 xΔx   Δx 
2
  6 x  Δx
Δx Δx
Self-Instructional
Material 7

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES The difference quotient is a function of x and

b) The derivative:

Δy
f   x   lim  lim  6 x  Δx   6 x
 x  0 Δ x x  0

c) When, x = 3, then f (3) = 18

And when, x = 4, then f (4) = 24

In-Text Questions
1. Given the function, y = 4x2 + 9;
(a) Find the difference quotient as a function of x and x.
Δy
(b) Find the derivative, .
Δx
(c) Find f (3) and f (4)
2. Given the function,
(a) Find the difference quotient as a function of x and x .
Δy
(b) Find the derivative, .
Δx
(c) Find f (2) and f (3).

1.3 DERIVATIVES AND THE SLOPES

Economic analysis deals with many equations and curves. The slopes of a curve are
also important to know the curvature and the shape of the curve. It can be seen
whether a curve is increasing or decreasing from the sign of the slope of the curve.
Another important aspect is the rate of change. From the slopes of a curve, we can
determine the rate of change of the function. Therefore, in the analysis of determining
the rate of change, the study of the slope of a curve is also important.

Self-Instructional
8 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Comparative Statics Analysis

In the previous section, we have discussed the association between the rate of NOTES
change and the derivate of any function. Therefore, it is also important to observe the
association between the derivate of a function and the slope of a curve, which represents
the function graphically.
To understand the association between the derivatives and the slope of a curve,
we have considered an example of a total cost function of a firm. The C = f (Q) as
shown in Fig. 1.1, represents the total cost function; it is a primitive function of the
curve. It implies that cost, C, depends on the quantity of the output, Q. The total cost,
C, is presented along the y-axis and the quantity of output is presented along the
x-axis.
Suppose that we consider Q0 as the initial level of output; and the point, A on the
cost curve provides the corresponding initial level of cost, C0. The point, A is the
relevant point at which the slope of the curve will be measured and hence the derivative.

Fig. 1.1

Consider that there is an increase in the level of output by Q from Q0 to Q2,

such that Q2 = Q0 + Q. Similarly, the total cost is also increased by C from C0 to
C2, such that C2 = C0 + C. The corresponding point of (Q2, C2) is B on the total cost
curve. Hence, the difference quotient is measured by:

Self-Instructional
Material 9

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES ΔC  C 2  C 0 
 ...(1.5)
ΔQ  Q2  Q0 
Geometrically, this difference quotient can be presented by the ratio of the two
line segments, BE AE, or the slope of the line, AB. This ratio measures the average
rate of change in C, due to the change in Q. This difference quotient in Equation (5) is
a function of Q0 and Q. The difference quotient is measured by the slope of the line,
AB.
If the increase in Q is Q1, instead of Q2, then we can get the corresponding cost,
C1. Then we could measure the difference quotient by the slope of the line, AD, such
that the slope of AD is equal to the ratio of line segments, DF AF .

Geometrically speaking, the ratio, DF AF is less than the ratio, BE AE.

Therefore, the slope of the line, AD is less than the slope of the line, AB. Hence, it can
be said that the smaller the increase in Q, the flatter will be the line. Consider a very
small change in Q, such that C  0, we may obtain a much flatter line, KG, which is
the tangent of the cost curve at the point, A. The slope of the line, KG is = GH KH .

ΔC
The point, A represents the limit of , as C  0, when the initial output is at
ΔQ
Q = Q0.
Now, consider the concept of marginal cost (MC). The marginal cost is defined
by the change in total cost, resulting from a unit change in output, such that;
ΔC
MC  ...(1.6)
ΔQ
Now, consider that there is a small change in Q. If the variable, Q is a continuous
variable, then a small change in Q, i.e., Q, can refer to an infinitesimal change. The
marginal cost can be measured by the slope of the total cost curve. However, the
ΔC
slope of the total cost curve is nothing but the limit of the ratio, , when C  0.
ΔQ
Self-Instructional
10 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Comparative Statics Analysis

The concept of the slope of a curve is merely the geometric counterpart of the concept NOTES
of the derivative. Therefore, in terms of the derivative, the slope of the curve,
C = f (Q), at the point A corresponds to the particular derivative value f (Q0).

1.4 CONCEPT OF LIMIT

dy Δy
The derivative, has been defined as the limit of the difference quotient, as
dx Δx
x  0. If we adopt the shorthand symbol q, then

dy Δy
 lim  lim q ...(1.7)
dx x 0 Δx v0
The derivative concept heavily relies on the notion of limit. Hence, it is imperative
that we get a clear idea about that notion.

Left-Side Limit and Right-Side Limit

The concept of limit says that if a variable (say, v) approaches a specific value (say,
zero), then it tells the value the another variable (say, q) approaches to. Here as the
value of q depends on the value of v; we can say, q = g (v). Now consider that N is
any finite real number. The variable, v approaches to N, such that v  N. Now
consider the two types of limit; the limit of q as approaches to minus infinity (v  –),
and the limit of q as v approaches to plus infinity (v  ). Also consider that when
v  N, then q  L, where L is the finite number.
 Left-Side limit (v  –): When, approaches to N from the left-side of q
lim q .
(from values less than N), then the limit of can be written as: v N

 Right-Side limit (v  +): ): When, approaches to N from the right-side of

lim q .
(from values more than N), then the limit of q can be written as: v N

Self-Instructional
Material 11

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES  Existence of limit of q: If L is a finite number, and the left-side limit and the
right-side limit approach to the common finite number, L, then we consider that
the limit of q does exist. We can write that limit as; vlim q  L . Therefore,
N

lim q  lim q  L ...(1.8)

v N  v N 

Graphical Illustration

The concept of limit can also be presented graphically. In Fig. 1.2 and Fig. 1.3, there
are four panels, two panels in each figures, which are showing four different types of a
function, q = g(v). In the panel (a) of Fig. 1.2, we can see that the function q = g(v) is
a smooth curve. As the variable, v tends to the value N from either sides on the
horizontal axis, the variable, q tends to the value, L. Therefore, in this case, the left-
side limit and the right-side limit are equal; and lim q  L. It implies that the limit q
v N
exists in panel (a).

Figure 1.2

In panel (b) of Fig. 1.2, however, the curve q = g(v) is not smooth. There is a
sharp turning point just above the point, N. However, as v  N from either side, then
q  L, from both sides. Hence, the limit of q exists at N, and q is equal to L.
In panel (c) of Fig. 1.3, below, the function q = g(v) is a step function. In this
case, s tends to N, the left-side of q is L1, that is lim q  L1 , and the right-side of q
v N 
Self-Instructional
12 Material lim q  L2 . As L  L , then q does not have a limit as v  N.
is L2, that is v  N 1 2

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Comparative Statics Analysis

NOTES

Figure 1.3

In the panel (d) of Fig. 1.3, the function q = g(v) has two parts of the hyperbolic
curve. As, v approaches to N from the left-side, i.e., v  N  , then approaches to
‘minus infinity’, i.e., q  –. In this case, the hyperbolic curve will fall indefinitely
while approaching the broken vertical line as an asymptote. Also, as v approaches to
N from the right-side, i.e., v  N+, then q approaches to ‘plus infinity’, i.e., q  +.
In this case, the hyperbolic curve will rise indefinitely while approaching the broken
vertical line as an asymptote.

In-Text Questions

3. Given the function, q 

v 2
 v  56 
 v  7  , where, (v 7), find the left-side
limit and the right-side limit of q as v approaches 7. Can we conclude from
these answers that q has a limit as v approaches to 7?

4. Given q  
5  1 lim q .
, where (v 0), then find: a) lim q and a) v
v v  

Self-Instructional
Material 13

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES
1.5 LIMIT THEOREM

Our interest in the rate of change led us to the consideration of the concept of derivative,
which, being in the nature of the limit of a difference quotient, in turn, prompted us to
study questions of the existence and evaluation of a limit. The basic process of limit
evaluation involves letting the variable, v, approaching a particular number (say, N)
and observing the value that q approaches. When evaluating the limit of a function, we
may draw upon certain established limit theorems.

1.5.1 Theorems Involving a Single Function

When a single function, q = g(v) is involved, the following theorems are applicable:
Theorem 1:
If, q = av + b, (a and b are constant)

Then, vlim q  aN  b
N

Theorem 2:
If, q = g(v) = b,

Then, vlim qb

N

The theorem 2 says that the limit of a constant function is constant in that function.
This is a special case of the theorem 1, with a = 0.
Theorem 3:

If, q = v, then, vlim qN

N

If, q = vk, then vlim q  Nk

N

Self-Instructional
14 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Comparative Statics Analysis

1.5.2 Theorems Involving Two Functions NOTES

If we have two functions of the same independent variable, v, such that:

q1 = g(v) and q2 = h(v)
And if, both the functions possess limits as follows:

lim q1  L1 , and lim q2  L2

v N v N

Where, L1 and L2 are two finite numbers. The following theorems are applicable:
Theorem 4: (Sum-difference limit theorem)
lim (q1  q2 )  L1  L2
v N

The limit of a sum (difference) of two functions is the sum (difference) of their
respective limits. In particular, we note that,

lim 2 q1  lim  q1  q1   L1  L1  2 L1
v N v N

Which is in line with the theorem 1.

Theorem 5: (Product limit theorem)
lim (q1 q2 )  L1 L2
v N

The limit of a product of two functions is the product of their limits.

Applied to the square of a function, this gives;

lim (q1 q1 )  L1 L1  L12

v N

Which is in line with the theorem 3.

Theorem 6: (Quotient limit theorem)
q  L
lim  1   1  L2  0 
v  N  q2  L2

The limit of a quotient of two functions is the quotient of their limits. Naturally,
the limit is restricted to be non-zero; otherwise the quotient is undefined.
Self-Instructional
Material 15

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES 1.5.3 Limit of Polynomial Function

With the given limit theorems, we can easily evaluate the limit of any polynomial function,

q  g  v   a0  a1v  a2v 2 ..  an v n ...(1.9)

As v tends to the number N, vlim

N

Since, the limits of the separate terms are, respectively;

lim a0  a0
v N
lim a1v  a1 N
v N

lim a2v 2  a2 N 2
v N

lim an v n  an N n
v N
The limit of the polynomial function is (by the sum limit theorem):

lim q  a0  a1 N  a2 N 2  ..  an N n ...(1.10)

v N

This limit is also actually equal to g(N), that is, equal to the value of the function
in Equation (1.9), when v = N. This particular result is significant in discussing the
concept of continuity of the polynomial function.

In-Text Questions

5. Find the limits of the function, q  7  9v  v 2 ;

(a) As v  0 (b) As v  3 (c) As v  –1

6. Find the limits of the function, q  

3v  5 
v  2
(a) As v  0 (b) As v  5 (c) As v  –1

Self-Instructional
16 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Comparative Statics Analysis

NOTES
1.6 CONTINUITY AND DIFFERENTIABILITY OF A
FUNCTION

The concept of limit and the limit theorem can now be used to define the continuity and
differentiability of a function. The continuity and differentiability of a function are
important to define the derivative of a function.

1.6.1 Continuity of a Function

When a function, q = g(v), possesses a limit as v  N, and when this limit is also
equal to g(N)which is also equal to a finite number, L, then the function is said to be
continuous at N. The term ‘continuity’ involves no less than three requirements, such
as:
a) The point must be in the domain of the function;
i.e., g(N) is defined;
b) The function must have a limit as v  N;

i.e., vlim g  v  exists;

N

c) The limit must be equal in value to g(N);

i.e., vlim g v  g  N 

N

When the function, q = g(v), is continuous at all values of in the interval (a, b),
it is said to be continuous in that interval. If the function is continuous at all points in its
domain, we say that it is continuous in its domain. In panel (a) of Fig. 1.2, all three
requirements are met at the point, N. The point, N is in the domain; q has the limit L as
v  N; and the limit, L happens also to be the value of the function at N. Thus, the
function represented by that curve is continuous at . In panel (b) of Fig. 1.2, since L
is the limit of the function as in the domain, and since is also the value of the function at
N. Therefore, although the function is not smooth at , the function is still continuous at
N.
Self-Instructional
Material 17

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES Again in panel (c) of Fig. 1.3, the function is discontinuous N at because a limit
does not exist at that point, in violation of the second requirement of continuity. In
panel (d) of Fig. 1.3, the function is discontinuous at v = N. The discontinuity comes
from the fact that N is excluded from the domain, in violation of the first requirement of
continuity.

Definition of continuous function

A function g(v) defined on an open interval including the point v = N is continuous at

that point if;

i) v lim g  v   lim g  v 
N  v N 

This means, vlim g  v  exists.

N

ii) vlim g v  g  N 

N

Another Definition

A function g  v  is continuous at the point, v  N , if g  v  is arbitrarily close to g  N 

for all values of v close to v  N . That means, if there is some   0 such that
g  v   g  N    , where v  N   for any   0 .
Example 1.2:

Show that the linear function f  x   2 x is continuous.

Solution:
Consider, the point x = 3.

The function value at x = 3 is f  3  6 .

If,  is very small and   0 , then x  a   , then we can write,

a     x  a    .
Self-Instructional
18 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Comparative Statics Analysis

NOTES
At the point, x = 3, it becomes,  3     x   3    . Therefore, the values of
x are defined within the interval, ( 3   , 3   ).

And,   0 is also very small, then f  x   f  a    , then we can write,

 f a    f  x   f a   . At the point, x = 3, it becomes

 f  3      f  x    f  3    . Therefore, the function values set as

 f  3   , f  3    , i.e.,  6   , 6    .
Let,   0.01 and   0.002 .

We find that all function values f  x  defined on the set

x   3  0.002, 3  0.002  , or (2.998, 3.002) lie within the distance  of f  3 .

Properties of continuous function

If f and g are continuous at a, then,

i) f  g and f  g are also continuous at a

ii) f .g is also continuous at a

iii) f / g is continuous at a, if g  a   0

iv)  f  x   is continuous at a, if  f  a  
p/q p/q
is defined.

Composite of continuous functions are also continuous

Consider a composite function, f  g  x   , where f and g are assumed to be

continuous. If x is close to a, then by the continuity of g at a, g  x  is close to g  a  .

In turn, f  g  x   becomes close to f  g  a   because f is continuous at g  a  , and

Self-Instructional
Material 19

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES
thus f  g  x   is continuous at a. Therefore, in general, any function that can be
constructed from continuous functions by combining one or more operations of addition,
subtraction, multiplication, division (except by zero, of course), and composition is
continuous at all points where it is defined.
Example 1.3:
For what values of a is the following function continuous everywhere?

ax 2  4 x  1 if x  1
f  x  
 x  3 if x  1

Solution:
The function is obviously continuous at all x  1 .

For x = 1, the function is given by the upper formula, so f 1  a  3 .

If x is slightly larger than 1, then f  x    x  3 is close to (-1 + 3) = 2,

and f  x   2 as x  1 .

In order to have f continuous at x = 1, we must have f  a   a  3  2 , which

requires a = -1.
But if a  1 , the function is discontinuous at x = 1, but continuous at all other
points.

1.6.2 Differentiability of a Function

Now, we are required to use the concept of limit to verify if a function is differentiable
or not. We can apply the concept of limit at a different level and take the limit of the
Δy
difference quotient of that function, , as Δx approaches zero. Before, starting
Δx
discussing the differentiability of a function, let us recall the concept of continuity in a
nutshell.
Self-Instructional
20 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Comparative Statics Analysis

Continuity condition NOTES

To be a function y  f  x  as continuous at x  x0 , the function needs to fulfil three

conditions, which are given as follows:

i) x  x0 must be in the domain of the function, f  x  ;

ii) y must have a limit as x  x0 ; and

iii) The said limit must be equal to f  x0  .

If a function satisfies all the three conditions at x  x0 , the it will be called that
the function is continuous at x  x0 .

Therefore, xlim f  x   f  x0  .. 11 ...(1.11)

 x0

Next it is to verify the differentiability condition of the function at x  x0 .

Differentiability condition

dy
The question whether the derivative exists at x  x0 , or f   x0  exists will be dealt
dx
Δy
here. To get the derivative from the difference quotient, , we use the concept of
Δx
limit. However, before having the derivative, it is necessary to check the differentiability
condition of the function at that relevant point, such as at x  x0 . That means, if the
function is differentiable at the point, x  x0 , only then we can get the derivative at

x  x0 , i.e. f   x0  .

Self-Instructional
Material 21

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES
Δy
As x  0 , if the limit of exists at x  x0 , then the derivative f   x0  also
Δx
exists. Therefore, the symbolic expression of the differentiability of the function, f  x 
is as;
Δy
f   x0   lim
 x  0 Δx

f  x0  x   f  x0 
f   x0   lim (1.12)
x 0 Δx
These two properties of a function, such as continuity and differentiability, are
interrelated. The continuity of a function is the necessary condition for its differentiability.
However, the continuity cannot ensure the property of differentiability. Hence, the
continuity property is not a sufficient condition for differentiability. This can be proved
by using the Fig. 1.4 and by applying the limit theorem.

Fig. 1.4

In Fig. 1.4, consider the function y  f  x  and the initial level of x is at x  x0 .

Self-Instructional
22 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Comparative Statics Analysis

NOTES
The function f  x  is continuous at x  x0 . Now let the value of x increases from x0
to (x0 + x) by x. Due to this increase in x, the function f x) also changes from
f x0) to f (x0 + x) by y.
Now let us change the notation and replace the value of x0 by N, and (x0 + x)
by x.
Therefore, x = (x – N).
Now, as x  0, then it implies x  N .
If the continuity conditions hold at N, then we obtain;

lim f  x   f  N  ...(1.13)
x N

If the differentiability conditions at N, then we obtain;

f  x  f  N 
f   N   lim ...(1.14)
x N x  N 

Here, as x  N , implies that x  N . Therefore,  x  N   0 .

Now, start with an identity to verify whether a differentiable function is also

continuous or not;
f  x   f  N   f  x   f  N  .. 15 ...(1.15)

Then, after diving and then multiplying the right-hand side by  x  N  , we obtain;

f  x  f  N 
f  x  f  N    x  N  16 ...(1.16)
x  N 
Now, taking the limit of the left-hand side as x  N , we obtain;

Left-side  lim[ f  x   f  N ]
x N

 lim f  x   lim f  N  [by applying the limit theorem 4; difference

x N x N

limit theorem] Self-Instructional

Material 23

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES
 lim f  x   f  N  [as f  N  is a constant, by applying the
x N

limit theorem 2]

 f N f N  0

f  x  f  N 
Right-side  lim x  N 
x N x N

f  x  f  N 
 lim lim  x  N  [by applying the product limit
x N x  N  x N

theorem, limit theorem 5)

 f   N   lim  x  N   [by using equation (1.14)]

 x N 

 f   N   lim x  lim N  [by applying difference limit theorem,

 x N x N 
limit theorem 4]

 
 f   N   lim x  N  [as f  N  is a constant, by applying the
 x  N 
limit theorem 2]

 f   N  N  N   0

If the function f  x  is not differentiable at N, then f   N  does not exist, as in

Equation (1.14). Hence, the right-side of Equation (1.16) would not possess a limit.
Therefore, the function is not continuous at N.

And, if the function f  x  is differentiable at N, then f   N  does not exist, as

in Equation (1.14). Hence, both the left-side and right-side of Equation (1.16) possess
a limit and are the same. This means, for both sides, lim f  x   f  N   0 . It proves
x N

that the function f  x  , is continuous at N.

Self-Instructional
24 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Comparative Statics Analysis

It implies that continuity follows from differentiability. If a function is differentiable NOTES

at every point in its domain, we may conclude that it must be continuous in its domain.
Differentiability implies continuity. However, continuity does not ensure differentiability.
That is, continuity is a necessary condition, but not sufficient for differentiability.
Differentiability is a more restrictive condition than continuity. Since differentiability
requires something more beyond continuity, continuity, at a point, rules out only the
presence of a gap, whereas differentiability rules out the ‘sharpness’, also. Therefore,
differentiability calls for smoothness, along with continuity of the function.

In-Text Questions
7. Taking the set of all finite real numbers as the domain of the function,
q  g  v   v 2  5v  2 ;
(a) Find the limit of q as tends to N (a finite real number).

(b) Check whether this limit is equal to g  N  .

(c) Check whether the function is continuous at N and continuous in its domain.

v  2
8. Given the function, q  g  v   v 2

2 ;

(a) Use the limit theorems to find vlim

N
q , N being a finite real number..

(b) Check whether this limit is equal to g  N  .

(c) Check the continuity of the function g  v  at N and its domain  ,   .

1.7 SUMMARY

The comparative statics analysis requires the concept of derivatives. It shows the
change of the equilibrium value of one variable due to the change in the values of a
parameter. It can be measured roughly from the rate of change of one variable due to
a change of the other variable. However, this is the average rate of change. If we want Self-Instructional
Material 25

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES to study the rates of changes along the function, then we need to study the rate of
changes, when the change in the parameter is very small.
The difference quotient is the ratio of the differences of the two variables, when
one variable changes due to a change in another variable. With the limit case, the
difference quotient gives the derivative. Therefore, a derivative is the rate of change,
when a parameter changes by a very small amount. Moreover, the slope of a curve
also gives the rate of change; hence, the concept of derivative becomes important.
Therefore, from the study of slopes of a curve, we can associate the result with the
comparative statics analysis.
It is important to understand the concept of limit. The slopes of a curve can
provide a derivative, but only in the limit case. The concepts of continuity and
differentiability of a function have a crucial role in the existence of the derivate of a
function. The concept of limit is important for continuity and differentiability. Therefore,
if a function is continuous and differentiable, then the derivative of the function exists
with the domain. Hence, we can get the comparative statics analysis.

1.8 GLOSSARY

 Continuity: It refers to the continuity of a function which is a property of a

function. It follows three conditions to be continuous at a point with the domain.
Continuity is the necessary condition for the differentiability of a function.
 Derived function: It is the derived function which is actually derived from the
primitive function. The derivative function is an example of a derived function.
 Derivative: It is limit case of the difference quotient.
 Difference quotient: It is the ratio of the two differences.
 Differentiability: It is a condition to obtain derivatives. To be differentiable,
one function needs to be continuous. However, the continuity of a function does
not ensure the differentiability of a function.
 Limit: It is the value that a function or sequence approaches when the input gets
closer to a specific value.
Self-Instructional
26 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Comparative Statics Analysis

 Marginal cost (MC): It is the incremental total cost due to a one-unit increase NOTES
in output. The marginal cost can be measured from the slope of the total cost
function.

 Primitive function: It is the original function as y  f  x  .

1.9 ANSWERS TO IN-TEXT QUESTIONS

Δy dy
1. (a)  8 x  x (b) f   x    8 x (c) f   3  24 , f   4   32
Δx dx

Δy dy
2. (a)  10 x  x  4 (b) f   x    10 x  4 (c) f   2   16 , f   3  26
Δx dx
3. left-side limit = right-side limit = 15; the limit is 15
4. (a) 5 and (b) 5
5. (a) 7 and (c) 17

1
6. (a) 2 and (c) 2
2
7. (a) N2 – 5N – 2 (b) Yes (c) Yes
8. (a) (N + 2) / (N2+2) (b) Yes (c) Continuous in its domain

1.10 SELF-ASSESSMENT QUESTIONS

1. Given the function, y  5 x  2 ;

Δy
a) Find the difference quotient . What type of function is it?
Δx

Self-Instructional
Material 27

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES Δy
b) Since the expression Δx does not appear in the function in part (a),
Δx
Δy
does it make any difference to the value of , where Δx is large or
Δx
small? Consequently, what is the limit of the difference quotient as Δx
approaches zero?

 v  2  3  8 
2. Given q   
v ; where v  0 , then find;

a) vlim
0
q b) vlim
2
q and c) vlim
a
q

3. Find the limits of the function, q   v  2  v  3 ;

a) As v  1 ; b) As v  0 ; c) As v  5 ;

4. Under what condition does the limit of f  x  as x tends to a exist?

5. Give the conditions of continuity of a function f  x  at the point x = a.

6. Give two examples of a function that are not continuous.

7. Define and explain left-side and right-side derivatives of a function f  x  at a

point x = x0.
8. Use the concept of left-side and right-side derivatives to indicate when a function
f  x  is differentiable at a point x = x0.

9. Why is it the case that if a function f  x  is differentiable at a point x = a, then

it must also be continuous at that point?

Self-Instructional
28 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Comparative Statics Analysis

NOTES
1.11 REFERENCES

 Chiang, A. and K. Wainwright. 2005. Fundamental Methods of Mathematical

Economics. Boston: McGraw-Hill.
 Sydsaeter, Knut and Peter J. Hammond. 2002. Mathematics for Economic
Analysis, Pearson Educational.
 Hoy, M., J. Livernois, C. McKenna, R. Rees and T. Stengos. 2001.
Mathematics for Economics. Prentice-Hall India.

Self-Instructional
Material 29

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 UNIT II: DIFFERENTIALS AND ITS ROLE IN
COMPARATIVE STATICS ANALYSIS

LESSON 2 DIFFERENTIAL AND ITS ROLE IN

COMPARATIVE STATICS ANALYSIS
Differential and its Role in Comparative Statics Analysis

LESSON 2 NOTES

DIFFERENTIAL AND ITS ROLE IN

COMPARATIVE STATICS ANALYSIS
Sarabjeet Kaur
Assistant Professor,
Department of Economics,
Zakir Husain College (E), University of Delhi
Structure
2.1 Introduction
2.2 Rules of Differentiation for a Function of One Variable
2.2.1 Constant-Function Rule
2.2.2 Power Function
2.2.3 Power-Function Rule Generalised
2.2.4 Linear Function
2.2.5 Derivative of the Product of Two Functions
2.2.6 Derivative of the Quotient of Two Functions
2.2.7 Finding Marginal-Revenue Function from Average-Revenue Function
2.2.8 Elasticity of Demand
2.2.9 Relation between AC and MC
2.2.10 Derivative of a Composite Function (Chain Rule)
2.2.11 Inverse-Function Rule
2.3 Partial Differentiation
2.4 Applications to Comparative Statics Analysis
2.5 Jacobian Determinant
2.6 Summary
2.7 Glossary
2.8 Answers to In-Text Questions
2.9 References

Self-Instructional
Material 33

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES
2.1 INTRODUCTION

In comparative statics analysis, the role of differentials is crucial for understanding how
changes in independent variables affect the values of dependent variables in an economic
model. Differentials provide a mathematical framework for measuring the sensitivity of
a variable to small changes in another variable.
Differentials are used to express the concept of marginal changes in variables.
They capture the incremental change in a variable resulting from a small change in
another variable, while keeping all other variables constant. By taking differentials of
equations representing economic relationships, economists can analyse the effects of
changes in variables on equilibrium values.
For example, in the context of supply and demand analysis, differentials help
economists determine the impact of changes in factors like price, income, or input
costs on the quantity demanded or supplied. By taking the differential of the demand
or supply equation with respect to a particular variable, economists can calculate the
marginal effect of that variable on the quantity.
In short, differentials play a fundamental role in comparative static analysis by
quantifying the marginal effects of independent variables on dependent variables. They
provide a mathematical framework for understanding and analysing the sensitivity of
economic models to changes in variables, allowing economists to make predictions
and policy recommendations based on the comparative analysis of different scenarios.

2.2 RULES OF DIFFERENTIATION FOR A

FUNCTION OF ONE VARIABLE

Let us discuss the various rules in detail.

2.2.1 Constant-Function Rule

If y = f(x) = c, where c is a constant, then

dy
Self-Instructional Then  y  f   x   f'  0
34 Material dx

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Differential and its Role in Comparative Statics Analysis

We may also write these in the form: NOTES

d d
y f x  0
dx dx
where d/dx and y [or f(x)] are the two independent elements of the derivative
symbol. The first component, d/dx, can be interpreted as an operator symbol indicating
mathematical operation. The sign d/dx stands for an instruction to take the derivative
of, or to differentiate, (some function) with respect to the variable, x. The second
portion specifies the function that needs to be changed (differentiated); in this case, it
is y = f(x).
The following is the rule’s proof. We know that y = f(x) = c
dy f ( x)  f ( x ) cc
 lim  lim  x  x 0  0 .
dx x x x  x x  x x  x

dy df
We can also write  as
dx dx

d d
y f
dx dx

d
Here, will be considered as an operator symbol.
dx

In-Text Questions
1. Differentials are used to express the concept of _______in variables.
dy
2. If y = 6 then  _______ .
dx
Example 2.1:
Y = 40

dy d  40 
 0
dx dx
Hence, the derivative of a constant is always zero. Self-Instructional
Material 35

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES 2.2.2 Power Function

The derivatives of a power function y = f(x) = xn is nxn-1,

Symbolically,
d d d n
y f x  x  nx n 1
dx dx dx
where n is any real number – < n < ,
We prove this rule,
In the simplest case, let n = l, the function is f(x) = x, and according to the rule,
the derivative is
d d d 1
y f x  f x  x  1x11  x 0  1
dx dx dx
The proof of this result follows easily from the definition of f’( x0) . Given f(x) =
x, the derivative value at any value of x, say, x = N, is
f  x   f  x0  x  x0
f'(x 0 )  lim  lim  lim 1  1
x  x0 x  x0 x x 0 x  x
0
x x0

As, ‘x0’ represents any value of x, it is permissible to write f’( x) = 1. This

proves the rule for the case of x0 = 1.
Let’s consider the larger case; x0= 2, 3, ... , consider the following identities:

x n  x0n   x  x0  x n 1  x0 x n 2  x 02 x n3    x0n 1 
Then,
x n  x0n
 x n 1  x0 x n 2  x02 x n3    x0n 1
x  x0
Thus,
f ( x )  f ( x0 ) x n  x0n
f  ( x0 )  lim  lim
x  x0 x  x0 x x0 x  x0

 lim x n 1  x0 x n 2  x02 x n 3    x0n 1

x  x0

 x0n1  x0n1  x0n1    x0n1

Self-Instructional  nx0n1 .
36 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Differential and its Role in Comparative Statics Analysis

Thus, NOTES
f ’ (x) = nx n-1
which proves the rule for n, any positive integer.

In-Text Questions
Fill in the blanks.
d n
3. x  _______ x n 1
dx
d 8
4. x  8x _____
dx
Example 2.2:
1. If y = x5

d n
x = n.xn-1
dx
Here, x = x and n = 5

dy d 5
= x
dx dx
= 5.x5-1 = 5x4
2. If y = x1

d n
x = n.xn-1
dx
Here, x = x and n = 1

dy d 1
= x
dx dx
= 1.x1-1 = x0 = 1
3. If y = x-8
Self-Instructional
Material 37

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES d n
x = n.xn-1
dx
Here, x = x and n = -8

dy d 8
= x
dx dx
= -8.x-8-1 = -8x-9

8
=
x9

1
4. If y =
x7
Rewritten as,
y= x-7

d n
x = n.xn-1
dx
Here, x = x and n = -7

dy d -7
= x
dx dx
= -7.x-7-1 = -7x-8

7
=
x8

5. y = x
Y = x1/2

d n
x = n.xn-1
dx

Self-Instructional Here, x = x and n = ½

38 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Differential and its Role in Comparative Statics Analysis

1
NOTES
1
dx 1
2 1
 x 2
dx 2

1  12
 x
2

1
 1
2x 2

1

2 x

2.2.3 Power-Function Rule Generalised

If the function is given by y = f(x) = cxn , then

dy d
 f  x   n.c x n 1 .
dx dx

Here, n is the product of constant.

To prove, this rule consider, x = N, the value of the derivative of f(x) = cxn is

f ( x)  f ( N ) cx n  cN n  xn  N n 
f ( N )  lim  lim  lim c  
x N xN x N xN x N  x  N 

xn  N n
 lim c lim [product limit theorem]
x N x N xN
xn  N n
 c lim [limit of a constant]
x N xN
 cnN n 1
In view that N is any value of x, this last result can be generalized immediately to
f (x)= cnxn–1, which prove the rule.
Self-Instructional
Material 39

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES Example 2.3:

1. Suppose y = 2x. Then

dy
 2 x0  2
dx
2. Suppose y = 4x3. Then

dy
 4  3 x 31  12 x 2
dx

dy
3. If y = –10x–5 find
dx
dy
 a  n  x n 1
dx
a = –10 n = –5

d  10x5 
 10   5 x51
dx

dy
 50 x 6
dx

2.2.4 Linear Function

If y = mx + c
Here, m and c are constant.

dy d ( mx  c )
 m
dx dc
Example 2.4:

dy
1. If y = 5x + 6 find
dx
Self-Instructional
40 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Differential and its Role in Comparative Statics Analysis

d (5 x  6) NOTES
 5  1 x11 0
dx
= 5x0 + 0
= 5 × (1) = 5
dy
=5
dx
dy
2. If y = 7x –3 find
dx
d (7 x  3)
 7  1 x11 0
dx
= 7x0 + 0
= 7 × (1)= 7

dy
7
dx

Derivative of a sum

If y = u + v

dy d (u ) d (v)
 
dx dx dx

Example 2.5:
1. If y = x3 + x8 find dy/dx
dy d (u ) d (v)
 
dx dx dx
u = x3
v = x8

dy d ( x3 ) d ( x8 )
  Self-Instructional
dx dx dx Material 41

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES
dy
 3 x 31  8 x81
dx

dy
 3x 2  8x7
dx

Derivative of a difference:

If y = u – v

dy d (u ) d (v)
 
dx dx dx

Example 2.6:

dy
If y = x3 – x8 find
dx
dy d (u ) d (v)
 
dx dx dx
u = x3
v = x8

dy d ( x3 ) d ( x8 )
 
dx dx dx

x n  nx n1

= 3x3–1 – 8x8–1
dy
 3x 2  8x 7
dx

dy
 3x 2  8x7
dx
Self-Instructional
42 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Differential and its Role in Comparative Statics Analysis

NOTES
In-Text Questions
True or false.
dy
5. If y = 6x+8 then  14
dx

dy
6. If y = 6x–8 then  2
dx

dy
7. If y = 8x+10 then 8
dx

dy
8. If y = 8x–10 then 8
dx
Example 2.7: y = 5x2 + 10x +6
Then the derivative of y is given by

dy d

dx dx
 
5x 2  10x  6  5*2 x  10  10x  10

Example 2.8:
Suppose a short-run total-cost function is given by c = 2Q3 – 2Q2 + 20Q + 125. Then
the marginal-cost function is the limit of the quotient C/Q, or the derivative of the C
function:
dC
 6Q 2  4Q  20
dx
In general, if a primitive function, y = f(x) represents a total function, then the
derivative function, dy/dx is its marginal function. Since the derivative of a function is
the slope of its curve, the marginal function should show the slope of the curve of the
total function at each point, x. Sometimes, we say a function is smooth if its derivative
is continuous.
Example 2.9:
Given a consumption function C = 100 + 0.5 Y, find MPC and MPS. Self-Instructional
Material 43

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES Marginal propensity to consume measures the change in consumption due to a

change in the income of the consumer. Mathematically, MPC is the first derivative of
the consumption function. Marginal propensity to save measures the change in saving
due to a change in income of the consumer.
Or
MPC+MPS = 1; hence, MPS = 1- MPC

dC d
Solution: MPC =  100  0.5Y   0.5
dY dy

MPS = 1- MPC = 1-0.5 = 0.5

Hence, MPC = MPS = 0.5.

2.2.5 Derivative of the Product of Two Functions

If y = u.v, where u and v are functions of x, then

dy dv du
 u.  v
dx dx dx
This rule says that the derivative of the product of two functions is equal to the
first function times the derivative of the second function plus the second function times
the derivative of the first function. As a corollary to this rule, if y = cu, where c is a
constant, we have

dy du dc du
 c.  u c
dx dx dx dx
As an extension of the rule to the case of three functions, we have

d
[ f ( x ) g ( x )h( x)]  f ( x) g ( x)h ( x)  f ( x ) g ( x )h( x )  f ( x) g ( x)h( x)
dx
Example 2.10: Find for the function y = (x - 1) (x +3).

dy d d
Solution:   x  1  x  3   x  3   x  1
dx dx dx
Self-Instructional
44 Material =(x - 1) *1 +(x+3) *1 = 2x+2

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Differential and its Role in Comparative Statics Analysis

2.2.6 Derivative of the Quotient of Two Functions NOTES

If y = u/v, here u and v are functions of x and v  0, then

du dv
v. u
dy dx dx
 2
dx v
dy
Example 2.11: Find of the function y = (1+x2)/(1 - x2 ).
dx

1  x  2

Solution: Given, y =
1  x 
2
2

dy 
 1 x2   2x   1  x   2x 
2

=
dx 1  x  2 2

dy 4x

 
2
dx 1  x 2

2.2.7 Finding Marginal-Revenue Function from Average-Revenue

Function

Suppose that the average-revenue (AR) function is specified by

AR = 15 – Q.
The total-revenue (TR) function is
TR =AR  Q
TR = (15 – Q) Q
TR = 15Q – Q2
Then, the marginal-revenue (MR) function is given by
Self-Instructional
Material 45

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES
d d
MR =
dQ
TR 
dQ
 
15Q  Q2  15  2Q

2.2.8 Elasticity of Demand

Example 2.12:
Given the demand function q = –5p + 100, find price elasticity of demand when the
price is equal to 5.
Solution: Elasticity of demand can be calculated as:

dP P
Ed = *
dQ Q

given q = –5p + 100,

= –5
when P = 5, then,
Q = –5(5) + 100
= 75
substituting
Hence, Ed = –5* = –0.33
= 0.33 (since we discard the sign in measurement of price elasticity).

2.2.9 Relation between AC and MC

Suppose a company produces a certain quantity of widgets. The total cost (TC) of
producing ‘q’ widgets is given by the equation:
TC = 100q2 + 500q + 5000
To find the average cost (AC), we divide the total cost by the quantity produced:
AC = TC / q = (100q2 + 500q + 5000)/q = 100q + 500 +5000/q
The marginal cost (MC) represents the additional cost incurred by producing
Self-Instructional
46 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Differential and its Role in Comparative Statics Analysis

one more unit, so it is the derivative of the total cost with respect to the quantity NOTES
produced:
MC = dTC / dQ
Let’s calculate the average cost and marginal cost for a specific quantity, say
q = 10.
Calculate the total cost (TC) at q = 10:
TC = 100q2 + 500q + 5000
TC = 100(10)2 + 500(10) + 5000
TC = 1000 + 5000 + 5000
TC = 11000
Calculate the average cost (AC) at q = 10:
AC = TC / q
AC = 11000 / 10
AC = 1100
Step 3: Calculate the marginal cost (MC) at q = 10:
MC = dTC / dq
MC = d/dq(100q2 + 500q + 5000)
MC = 200q + 500
MC = 200(10) + 500
MC = 2000 + 500
MC = 2500
Therefore, for the quantity q = 10, the average cost is $1100, and the marginal
cost is $2500.
The relationship between the average cost and marginal cost can vary depending
on the cost function. In general, if the marginal cost is below the average cost, the
average cost will decrease. If the marginal cost is above the average cost, the average
cost will increase. If the marginal cost is equal to the average cost, the average cost
will remain constant.
Self-Instructional
Material 47

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES Diagrammatically, it is explained below:

2.2.10 Derivative of a Composite Function (Chain Rule)

If z is a function of y; say f(y) and y if a function of x; say g(x) then the derivative of z
with respect to x is given by:

dz dz dy
 *
dx dy dx

Given ax, there must result in a corresponding y via the function y = g(x), but
this y will in turn being about a z via the function z = f(y).
Proof. Note that

dz z z y
 lim  lim .
dx x 0 x x 0 y x

Since x  0 implies y  0 which in turn implies z  0, we then we have,

dz dz dy
 *  f   y  g'  x 
dx dy dx

In view of the function y = g(x), we can express the function z = f(y) as z =

f(g(x)), where the contiguous appearance of the two function symbols, f and g indicates
that this is a compose function (function of a function). So sometimes, the chain rule is
Self-Instructional also called the composite function rule.
48 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Differential and its Role in Comparative Statics Analysis

As an application of this rule, we use it to prove the quotient rule. NOTES

For z =1/ g(x) let z = 1/y = y–1 and y = g(x). Then we have
dz dz dy 1 g ( x)
   2 g ( x)   2 .
dx dy dx y g ( x)
Thus,
d  f ( x)  d  1
    f ( x)  g ( x) 
dx  g ( x)  dx
d  1 
 f ( x) g 1 ( x)  f ( x)  g ( x) 
dx
 g ( x) 
 f ( x) g 1 ( x)  f ( x)   2 
 g ( x) 
f ( x) g ( x)  f ( x) g ( x)
 .
g 2 ( x)
Example 2.13:

dz
1. If z = 7y + 3 when y = 5x2 find .
dx
Solution:

dz dz dy
 *
dx dy dx
Z = 7y + 3

dz
7
dy
y = 5x2

dy
 10 x
dx
z = 7(5x2) + 3
Self-Instructional
Material 49

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES
dz
= 7 *10x
dx
= 70x

dz
= 70x
dx

dy
Example 2.14: If y = x2 and x = 1 + t2, find .
dt
Solution:

dy dy dx
 *
dt dx dt

dy
 2x
dx

dx
 2t
dt
Therefore,

dy
 2 x * 2t
dt
Putting x = 1 + t2

dy
= 2 (1 + t2) . 2t
dt
= (2+2t2) . (2t)
= 4t + 4t3
Example 2.15: Suppose TR = f(Q), where the output, Q is a function of labour input
L, or Q = g(L). Then, by the chain rule, the marginal revenue product of labour (MRPL)
is
Self-Instructional
50 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Differential and its Role in Comparative Statics Analysis

NOTES
dR dR dQ
MRPL    f (Q ) g ( L)  MR  MPL ,
dL dQ dL

Where MRPL is marginal physical product of labour. Thus the result shown
above constitutes the mathematical statement of the well-known result in economics
that MRPL = MR·MPL.

In-Text Questions
9. If the marginal cost is below the average cost, the average cost will________.
dz dz   
10.  *
dx dy dx

2.2.11 Inverse-Function Rule

If a function y = f(x) represents a one-to-one mapping, i.e., if the function is such that
a different value of x will always yield a different value of y, then the function, f will have
an inverse function x = f”1(y). Here, the symbol f”1 is a function symbol which signiûes
a function related to the function, f; it does not mean the reciprocal of the function f(x).
When x and y refer speciûcally to numbers, the property of one-to-one mapping is
seen to be unique to the class of function known as a monotonic function.
Definition: A function, f is said to be monotonically increasing (decreasing) if x1 > x2
implies f(x1) > f(x2) (f(x1) < f(x2)).
In either of these cases, an inverse function f”1 exists.
A practical way of ascertaining the monotonicity of a given function y = f(x) is to
check whether the f (x) always adheres to the same algebraic sign for all values.
Geometrically, this means that its slope is either always upward or always downward.
Example 2.16: Suppose y = 5x + 25. Since y = 5 for all x, the function is monotonic
1
and thus the inverse function exists. In fact, it is given by x = .
5y  5
If an inverse function exists, the original and the inverse functions must be both
monotonic. Moreover, if f–1 is the inverse function off, then f must be the inverse
function of f–1. Self-Instructional
Material 51

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES In general, we may not have the explicit inverse function. However, we can
easily find the derivative of an inverse function by the following inverse function rule:

dx 1

dy dy / dx
Proof:

dx x 1 1
 lim  lim 
dy y 0 y x0 y y
x

by noting that y  0 implies x  0.

Example 2.17: If y = 6x2 + 8x +9, then find dx/dy.
Solution:
dy
 12 x  8
dx
dx 1 1
 
dy dy / dx 12 x  8

Example 2.18:
Given y = ln x, its inverse is x = ey. Therefore, by the inverse-function rule, we have

dx 1 1
   x  ey .
dy dy dx 1 x

2.3 PARTIAL DIFFERENTIATION

In comparative statics analysis, however, we are likely to encounter a situation in

which several parameters appear in a model, so that the equilibrium value of each
endogenous variable may be a function of more than one parameter. Because of this,
Self-Instructional we now consider the derivative of a function of more than one variable.
52 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Differential and its Role in Comparative Statics Analysis

Partial differentiation is used to calculate the rate of change of a function with NOTES
respect to one of its variables while holding all other variables constant.
The symbol used to represent partial differentiation is  (the lowercase Greek
letter which means ‘partial’). The partial derivative of a function f(x, y) with respect to
x is denoted as f/x, read as ‘the partial derivative of f with respect to x’.
Consider a function:
y = f(x1, x2, ··· , xn),
where the variables xi (i = 1, 2, ···, n) are all independent of one another so that
each can vary by itself without affecting the others. If the variable xi changes, xi while
the other variables remain ûxed, there will be a corresponding change in y, namely, y.
The difference quotient in this case can be expressed as:

y f ( x1 , x2 ,  , xi 1 , xi  xi , xi  , xn )  f ( x1 , x2  , xn )

x xi

If we take the limit of y/x, that limit will constitute a derivative. We call it the
partial derivative of y with respect to xi. The process of taking partial derivatives is
called partial differentiation. Denote the partial derivative of y with respect to xi by
y
xi

y y
 lim .
xi xi 0 xi

In-Text Questions
Fill in the blanks.
11. A function, f is said to be monotonically increasing if x1 ____ x2 implies f(x1) >
f(x2).
12. Partial differentiation is used to calculate the rate of change of a function with
respect to one of its variables while holding all other variables ________.

Self-Instructional
Material 53

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES Example 2.19:

z
Z = 3x2+8y+6x. Find z/x and .
y

z      
Solution: 
x x
 
3x 2  8y  6x   3x 2 )  8y  (6x 
x  x x 

z
 6x  0  6
x

z
 6x  6
x

z      

y y
 
3x 2  8y  6x   3x 2 )  8y  (6x 
y  y y 

z
 080
y

z
8
y

Marginal Product

The marginal product (MP) of a factor of production refers to the addition to the total
product due to the use of an additional unit of that factor. The marginal product of
labour (MPL) or marginal physical product of labour (MPPL) is given by the change in
the total product (TP) due to a one-unit change in the quantity of labour used. MP is
derived by finding the derivative of the TP.
MPL = where L is labour, MPK = where K is capital
Example 2.20: For the production function Q = 20K0.5L0.5
(i) derive a function for MPL, and
Self-Instructional
54 Material
(ii) show that MPL decreases as one moves along an isoquant by using more L.

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Differential and its Role in Comparative Statics Analysis

Solution: NOTES
(i) MPL is found by partially differentiating the production function Q = 20K0.5L0.5
with respect to L. Thus

Q 10 K 0.5
MPL   10 K 0.5 L0.5  0.5
L L
(ii) If the function for MP above is multiplied top and bottom by 2L0.5, then we get

 2 L0.5   10 K 0.5  20 K 0.5 L0.5 Q

MPL   0.5    0.5   
 2L   L  2L 2L

An isoquant joins combinations of K and L that yield the same output level.
Thus if Q is held constant and L is increased then the function (1) shows us that MPL
will decrease.

2.4 APPLICATIONS TO COMPARATIVE STATICS

ANALYSIS

Comparative statics analysis is a method used in economics to analyse the effects of

changes in exogenous variables on the equilibrium values of endogenous variables in a
model. It allows us to understand how the system responds to different shocks or
policy changes.
Example 2.21:
Suppose we have a basic supply and demand model for a product, where the
equilibrium price (P) and quantity (Q) are determined by the following equations:
Demand: Q = 100 - 2P
Supply: Q = 10 + 3P
To perform comparative static analysis, we will analyse the effects of a change
in an exogenous variable, such as a shift in demand due to a change in consumer
preferences.
Self-Instructional
Material 55

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES Step 1: Determine the initial equilibrium values.

To find the initial equilibrium, we set the demand equal to supply:
100 - 2P = 10 + 3P
Solving for P:
5P = 90
P = 18
Substituting P = 18 into the demand equation, we find:
Q = 100 - 2(18)
Q = 100 - 36
Q = 64
So, the initial equilibrium is P = 18 and Q = 64.
Step 2: Introduce the change in the exogenous variable.
Let’s say, consumer preferences change, causing the demand equation to
shift to:
Q = 120 - 2P
Step 3: Determine the new equilibrium values.
We set the new demand equal to supply:
120 - 2P = 10 + 3P
Solving for P:
5P = 110
P = 22
Substituting P = 22 into the demand equation, we find:
Q = 120 - 2(22)
Q = 120 - 44
Q = 76
So, the new equilibrium is P = 22 and Q = 76.

Self-Instructional
56 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Differential and its Role in Comparative Statics Analysis

Step 4: Compare the initial and new equilibrium values. NOTES

Comparing the initial and new equilibrium values, we can see that the
change in demand from 100 - 2P to 120 - 2P led to an increase in the
equilibrium price from 18 to 22 and an increase in the equilibrium quantity
from 64 to 76.
Example 2.22:
What is point elasticity when price is 12 for the demand function p = 60 – 3q?
Solution:
Given
p = 60 – 3q

dp
 –3
dq
3q = 60 – p

60  p
q
3
If p =12; then

60  12 48
q   16
3 3
Hence,

p 1 12 1
e  ( 1)  (1)   0.25
q dp dq 16 3

Example 2.23:
Let’s consider a simple economic model where the quantity supplied of a product
(Qs) is determined by its price (P) and the cost of production (C) according to the
following equation:
Qs = aP – bC
Self-Instructional
Material 57

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES where ‘a’ and ‘b’ are constants representing the sensitivity of quantity supplied
to changes in price and cost, respectively.
Suppose the current price of the product is $10 and the cost of production is
$5. We want to analyse the impact of a 10% increase in both price and cost on the
quantity supplied.
To do this, we need to calculate the partial derivatives of Qs with respect to P
and C, and then use these derivatives to determine the percentage change in quantity
supplied.
Partial derivative of Qs with respect to P:
Qs/P = a
Partial derivative of Qs with respect to C:
Qs/C = -b
Given that a = 2 and b = 1, we can substitute these values into the partial
derivatives:
Qs/P = 2
Qs/C = -1
Now, let’s calculate the percentage change in quantity supplied (Qs/Qs)
resulting from a 10% increase in both price and cost.
Percentage change in price (P/P) = 10%
Percentage change in cost (C/C) = 10%
Using the partial derivatives, we can calculate the percentage change in quantity
supplied as follows:
Qs/Qs = (Qs/P) * (P/P) + (Qs/C) * (C/C)
= (2) * (10%) + (-1) * (10%)
= 0.2 - 0.1
= 0.1
Therefore, the percentage change in quantity supplied is 0.1 or 10%.
This means that a simultaneous 10% increase in both price and cost leads to a
Self-Instructional 10% increase in the quantity supplied according to the given economic model.
58 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Differential and its Role in Comparative Statics Analysis

Example 2.24: NOTES

The Keynesian cross model represents the relationship between aggregate demand
(AD) and national income (Y). It is given by the following equation:
Y=C+I+G
Where:
Y represents national income (output),
C represents consumption spending,
I represents investment spending, and
G represents government spending.
Solution:
Assuming a simple consumption function, let’s say the consumption function is
given by:
C = 100 + 0.75Yd
Where:
Yd represents disposable income.
Let’s assume investment spending (I) is constant at 200, and government spending
(G) is initially set at 1000.
Step 1: Determine the equilibrium level of national income (Y) at the initial
government spending level (G = 1000).
In equilibrium, aggregate demand (AD) is equal to national income (Y).
Therefore, we can set AD equal to Y and solve for Y.
AD = C + I + G
Substituting the consumption function, investment spending, and government
spending:
Y = (100 + 0.75Yd) + 200 + 1000
Since Yd = Y - T, where T represents taxes, let’s assume T = 500:
Y = (100 + 0.75(Y - 500)) + 200 + 1000
Simplifying the equation:
Self-Instructional
Y = 1800 + 0.75Y - 375 Material 59

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES 0.25Y = 1425

Y = 1425 / 0.25
Y = 5700
Therefore, at an initial government spending level of G = 1000, the equilibrium
level of national income is Y = 5700.
Step 2: Determine the new equilibrium level of national income when government
spending increases to G = 1500.
Using the same equation:
Y = (100 + 0.75Yd) + 200 + 1500
Substituting Yd = Y - T, with T = 500:
Y = (100 + 0.75(Y - 500)) + 200 + 1500
Simplifying the equation:
Y = 1800 + 0.75Y - 375
0.25Y = 1925
Y = 1925 / 0.25
Y = 7700
Therefore, when government spending increases to G = 1500, the new
equilibrium level of national income is Y = 7700.

Comparative statics analysis

To determine the change in national income resulting from the change in government
spending, we can calculate the difference between the new equilibrium level of national
income and the initial equilibrium level:
Y = Y_new - Y_initial
Y = 7700 - 5700
Y = 2000
Therefore, the increase in government spending from G = 1000 to G = 1500
leads to an increase in the equilibrium level of national income by Y = 2000.
Self-Instructional
60 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Differential and its Role in Comparative Statics Analysis

NOTES
2.5 JACOBIAN DETERMINANT

The Jacobian determinant is a mathematical concept used in calculus and linear algebra.
It is associated with a system of partial derivative equations and represents how the
variables in the system change with respect to each other.
To understand the Jacobian determinant, let’s consider a system of equations
with n variables:
f (x , x‚ , ..., x™ ) = 0
f‚ (x , x‚ , ..., x™ ) = 0
...
f™ (x , x‚ , ..., x™ ) = 0
The Jacobian determinant is a square matrix that contains the partial derivatives
of each equation with respect to all the variables:
y1 y1 y1

x1 x2 xn
y2 y2 y2
 ( y1 , y2  , yn ) 
J   x1 x2 xn
 ( x1 , x2  , xn )
   
yn yn yn

x1 x2 xn
The Jacobian determinant, denoted as det(J) or |J|, is calculated by taking the
determinant of the Jacobian matrix.
To numerically compute the Jacobian determinant, we need to evaluate the
partial derivatives of a given function with respect to each variable and then calculate
the determinant of the resulting Jacobian matrix. Here’s a step-by-step numerical
explanation:
1. Identify the function and variables: Let’s consider a simple example
with a function f(x, y) = x2 + 2xy + y2 and two variables, x and y.
Self-Instructional
Material 61

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES 2. Calculate the partial derivatives: Compute the partial derivatives of

the function with respect to each variable. In our example, we have:
f/x = 2x + 2y
f/y = 2x + 2y
3. Evaluate the partial derivatives at a specific point: Choose a point
(x€ , y€ ) at which you want to evaluate the Jacobian determinant. For
instance, let’s use the point (2, 3).
Evaluate the partial derivatives at this point:
f/x = 2(2) + 2(3) = 10
f/y = 2(2) + 2(3) = 10
4. Construct the Jacobian matrix: Create a square matrix using the partial
derivatives. In this case, the Jacobian matrix is:

10 10
J=
10 10

5. Calculate the determinant: Compute the determinant of the Jacobian

matrix. In our example, the determinant is:
det(J) = (10)(10) - (10)(10) = 0
The Jacobian determinant in this case is 0.

Interpretation

When the Jacobian determinant is 0, it indicates that the function is not invertible or
there is degeneracy in the system. Geometrically, it means that the transformation
associated with the function does not preserve the full dimensionality of the space.
The Jacobian determinant has several important applications, including:
 Change of variables: In multivariable calculus, when transforming integrals
from one coordinate system to another, the Jacobian determinant is used to
account for the change in variables. It helps to convert integrals in one
coordinate system to integrals in another coordinate system.

Self-Instructional
62 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Differential and its Role in Comparative Statics Analysis

 Implicit function theorem: The Jacobian determinant is used in the implicit NOTES
function theorem, which provides conditions for determining when an equation
can be solved for one variable as a function of the others. The determinant is
used to verify the existence and uniqueness of solutions.
 Linearisation: The Jacobian determinant plays a crucial role in linearising
non-linear systems. By evaluating the Jacobian matrix at a particular point,
we can get information about the local behaviour of the system near that
point and determine stability and equilibrium points.
 Optimisation: The Jacobian determinant is utilised in optimisation problems
involving multivariable functions. It is used to compute critical points, determine
the nature of extrema, and verify conditions for constrained optimisation.
In a nutshell, the Jacobian determinant is a mathematical tool that captures the
interrelationships between variables in a system of equations. It has numerous
applications in calculus, linear algebra, and optimisation, providing valuable insights
into the behaviour of systems and transformations of variables.

In-Text Questions
13. The Jacobian determinant is utilised in optimiSation problems involving
multivariable functions. a. True b. False
14. The Jacobian determinant is a matrix of order m*n that contains the partial
derivatives of each equation with respect to all the variables. a. True b. False

2.6 SUMMARY

The lesson discusses the role of differentials in comparative static analysis, which involves
analysing the effects of changes in independent variables on dependent variables in
economic models. It discusses the constant-function rule, power functions, the power-
function rule generalised for any real number, linear functions, derivatives of the product
and quotient of two functions, and the chain rule for composite functions. It also explains
the derivative of composite functions, inverse functions, partial derivatives, and
applications of differential analysis in comparative statics analysis. Self-Instructional
Material 63

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES In short, this lesson covers various rules of differentiation for functions of one
variable and their applications in economic analysis, specifically in comparative statics
analysis. It provides a mathematical framework for understanding the sensitivity of
economic models to changes in variables and economists to analyse the effects of
these changes on equilibrium values and make informed decisions and recommendations.

2.7 GLOSSARY

 Differential: It is a mathematical concept used to measure the sensitivity of a

variable to small changes in another variable.
 Comparative static analysis: It includes analysing how changes in independent
variables affect the values of dependent variables in an economic model.
 Derivative: It is the rate at which a function changes with respect to its
independent variable.
 Constant-function rule: It is the derivative of a constant function that is always
zero.
 Power function: It is a function in the form of f(x) = xn, where n is any real
number.
 Linear function: It is a function in the form of f(x) = mx + c, where m and c are
constants.
 Marginal revenue: It is the derivative of the total revenue function with respect
to quantity, representing the additional revenue from selling one additional unit.
 Elasticity of demand: It is a measure of the responsiveness of quantity demanded
to changes in price.
 Average cost: It is the total cost divided by the quantity produced.
 Marginal cost: It is the additional cost incurred by producing one more unit.
 Chain rule: It is a rule for finding the derivative of a composite function by
multiplying the derivatives of the individual functions.

Self-Instructional
64 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Differential and its Role in Comparative Statics Analysis

 Composite function: It is a function that is formed by taking the output of one NOTES
function and using it as the input for another function.
 Inverse function: It is a function that undoes the effect of another function.
 Partial derivative: It is the derivative of a function with respect to one of its
variables, keeping the other variables constant.
 Sensitivity: It is the degree to which a variable or system is affected by changes
in another variable.
 Equilibrium: It is a state of balance or stability in a system.

2.8 ANSWERS TO IN-TEXT QUESTIONS

1. Marginal changes
2. 0 (zero)
3. N
4. 7
5. (b) False
6. (b) False
7. (a) True
8. (a) True
9. Decrease
10. dy
11. ‘>’ (greater than)
12. Constant
13. (a) True
14. (b) False

Self-Instructional
Material 65

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES
2.9 REFERENCES

 Chiang, A. and Wainwright, K. 2005. Fundamental Methods of Mathematical

Economics. Boston. McGraw-Hill.
 Sydsaeter, Knut and Peter J. Hammond 2002. Mathematics for Economic
Analysis, Pearson Educational.
 Hoy, M., J. Livernois, C. McKenna, R. Rees and T. Stengos. 2001.
Mathematics for Economics, Prentice-Hall India.

Self-Instructional
66 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 UNIT III: OPTIMISATION PROBLEMS

LESSON 3 UNCONSTRAINED OPTIMISATION

LESSON 4 CONSTRAINED OPTIMISATION

LESSON 5 CONVEX AND CONCAVE FUNCTIONS

Unconstrained Optimisation

LESSON 3 NOTES

UNCONSTRAINED OPTIMISATION
Sarabjeet Kaur
Assistant Professor,
Department of Economics,
Zakir Husain College (E), University of Delhi

Structure
3.1 Introduction
3.2 Optimal Values and Extreme Values
3.3 1st Derivative Test
3.4 Second and Higher Derivatives
3.5 Optimisation in Case of Two Variables
3.6 Differential Conditions for Optimisation
3.7 Convex and Concave Functions and Inflection Points
3.8 Summary
3.9 Glossary
3.10 Answers to In-Text Questions
3.11 Self-Assessment Questions
3.12 Reference

3.1 INTRODUCTION

Unconstrained optimisation is a problem-solving technique that seeks to find the

maximum or minimum value of a function, given a set of constraints. In the case of
single-variable optimisation, there is only one variable that can be changed to affect
the value of the function.
For example, we might want to find the value of x that maximises or minimises
the function f(x) = x2.

Self-Instructional
Material 69

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES To solve an unconstrained optimisation problem, we can use the following steps:
1. Define the function: The first step is to define the function that we want
to optimise. This function can be any mathematical expression, such as a
polynomial, a trigonometric function, or an exponential function.
2. Find the critical points: The next step is to find the critical points of the
function. A critical point is a point in the domain of the function where the
derivative is either equal to zero or undefined.
3. Classify the critical points: Once we have found the critical points, we
need to classify them as either maxima, minima, or saddle points. A
maximum is a point where the value of the function is at its highest point. A
minimum is a point where the value of the function is at its lowest point. A
saddle point is a point where the value of the function changes from
increasing to decreasing or vice versa.
4. Choose the optimal solution: Once we have classified all of the critical
points, we can choose the optimal solution. The optimal solution is the
point in the domain of the function that has the highest or lowest value,
depending on whether we are maximising or minimising the function.

3.2 OPTIMAL VALUES AND EXTREME VALUES

Economics is a science of choice. When an economic project is to be carried out,

there are normally a number of alternative ways of accomplishing it. One (or more) of
these alternatives will be more desirable than others from the standpoint of some
criterion, and it is the essence of the optimisation problem to choose.
The most common criterion of choice among alternatives in economics is the
objective of maximising something (e.g., utility maximisation and profit maximisation)
or minimising something (e.g., cost minimisation). Economically, we may categorise
such maximisation and minimisation problems under the general heading of optimisation.
From a purely mathematical point of view, the collective term for maximum and minimum
is the more matter-of-fact designation extremum, meaning an extreme value.
In formulating an optimisation problem, the first order of business is to delineate
Self-Instructional
70 Material an objective function in which the dependent variable represents the objects whose

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Unconstrained Optimisation

magnitudes, the economic units, can pick and choose. We shall, therefore, refer to the NOTES
independent variables as choice variables.
Consider a general-form objective function
y = f(x)
Where
Y  dependent variable
f  function
x  Independent variable
of ‘Y’
Fig. 3.1 shows three examples of functions. The point, E in (c) is a relative
maximum, and the point, F is a relative minimum. A relative maximum is a point on a
function where the value of the function is greater than or equal to the values of the
function at all points in a small neighbourhood of the point. A relative minimum is a
point on a function where the value of the function is less than or equal to the values of
the function at all points in a small neighbourhood of the point.

Fig. 3.1 Extremum for various functions: (a) constant function; (b) monotonic
function, (3) non-monotonic function.

We will focus our discussion on finding relative maxima because an absolute

maximum (or global maximum) must be either a relative maximum or one of the
endpoints of the function. Therefore, to find the absolute maximum, we can find all of
the relative maxima and then select the largest one. Hereafter, the extreme values
considered will be relative or local ones, unless indicated otherwise. Self-Instructional
Material 71

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES
In-Text Questions
1. What is the process of finding the maximum or minimum value of a function
without any constraints?
a) Constrained optimisation b) Single-variable optimisation
c) Unconstrained optimisation d) Objective function
2. Which term refers to the highest value attained by a function?
a) Saddle point b) Minimum
c) Maximum d) Extremum
For this we have following criteria:
In optimisation, we find out the value of ‘x’ that yield desire value:
 1st derivative test
 2nd derivative test

3.3 1ST DERIVATIVE TEST

Before we can discuss how to find the value of x that optimises the value of y, we need
to understand the concepts of local maxima and local minima. A local maximum is the
highest point in a function’s graph within a certain interval. In Fig. 3.2, the point, A is a
local maximum. A local minimum is the lowest point in a function’s graph within a
certain interval. In Fig. 3.2, the point, B is a local minimum.

Self-Instructional
72 Material Fig: 3.2

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Unconstrained Optimisation

In Fig. 3.2, the vertical axis represents the dependent variable, Y, and the NOTES
horizontal axis represents the independent variable, X.
The local maximum point ‘A’ is at X1. This means that if we move to the left of
X1, i.e., at X0, the value of Y is less than A. Similarly, if we move to the right of X1, i.e.,
at X2, the value of Y is less than A. In other words, the value of Y is less than A at any
point to the left or right of X1.
The local minimum point ‘B’ is at X4. This means that if we move to the left of
X4, i.e., at X3, the value of Y is greater than B. Similarly, if we move to the right of X4,
i.e., at X5, the value of Y is greater than B. In other words, the value of Y is greater than
B at any point to the left or right of X4.
According to the first derivative test, we can find out whether the dependent
variable is a maximum or minimum by finding the derivative of the function and equating
it to zero. This will give us the extreme value of the function, assuming that the function
is differentiable and continuous.

dy
The first derivative of the function y = f(x) is  f x
dx
The above equation gives an extreme value of the dependent variable, which
can either be a maximum or a minimum. To determine whether it is a maximum or a
minimum, we need to find out the change in the sign of the derivative from left to right.
If the sign changes from positive to negative as we move from left to right, then we can
conclude that it is a maximum. Similarly, if the sign changes from negative to positive,
then we can conclude that it is a minimum. This can be explained in Fig. 3.3.
As we know, the slope of a function can be determined by drawing a tangent to
the function at that point. The slope of the tangent also represents the first derivative of
the function.
 If the slope of the tangent is positive, then we can say that the dependent
variable is rising with the increase in the independent variable.
 If the slope of the tangent is zero, then the dependent variable is at a maximum
or minimum point.
 If the slope of the tangent is negative, then we can say that the dependent
variable is decreasing as the independent variable increases. Self-Instructional
Material 73

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES

Fig. 3.3

In Fig. 3.3(i), the slope of the tangent is negative to the left of the point, B and
positive to the right of the point, B. This means that the sign of the derivative changes
from negative to positive as we move from left to right. This implies that the point, B
represents a minimum, or local minimum.
In Fig. 3.3(ii), the slope of the tangent is positive to the left of the point, E and
negative to the right of the point, E. This means that the sign of the derivative changes
from positive to negative as we move from left to right. This implies that the point, E
represents a maximum, or local maximum.
In Fig. 3.3(iii), the slope of the tangent is zero at the point, H. This means that
the derivative is zero at the point, H. Since the derivative is zero, we cannot say
whether the point, H is a maximum, minimum, or neither. This is a stationary point.

In-Text Questions
3. What are the points where the derivative of a function is either zero or undefined
called?
a) Optimal solutions b) Critical points
c) Inflection points d) Tangent points
4. Which test is used to determine the relative extrema of a function using the first
derivative?
a) First derivative test b) Second derivative test
c) Necessary condition test d) Sufficient condition test
Self-Instructional
74 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Unconstrained Optimisation

All the three cases, as shown in can be explained with the help of the following NOTES
example.
Let the function be
Y = x2 -10x +15
Diff w.r.t.‘x’

dy
 2x  10  0
dx

2x  10
Or x= 5
To determine whether the function attains a maximum or minimum at x = 5, we
can substitute x < 5 into the equation, 2x  10  0 to get a negative value, and substitute
x > 5 into the equation to get a positive value. This means that the sign of the derivative
changes from negative to positive as we move from left to right, which implies that the
value of Y is minimised at x = 5.
To determine whether the function attains a maximum or minimum at x = 5, we
can substitute x < 5 into Equation (3) to get a negative value, and substitute x > 5 into
Equation (3) to get a positive value. This means that the sign of the derivative changes
from negative to positive as we move from left to right, which implies that the value of
Y is minimised at x = 5.
In other words, the function reaches its lowest point at x = 5.
Let’s consider another function.
Y = 15+10x - x2
Diff w.r.t.‘x’

dy
 10  2x  0
dx

2x  10
Or x= 5
Here, Y is the maximum at x = 5. This can be proved by substituting x < 5 in the Self-Instructional
equation, 10-2x = 0 and x > 5 in the same equation. If we substitute x < 5, then the Material 75

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES equation will be positive. This means that Y is increasing when x is less than 5. If we
substitute x > 5, then the equation will be negative. This means that Y is decreasing
when x is greater than 5. The fact that the sign of the derivative changes from positive
to negative as we move from left to right in the domain of the function implies that Y is
maximum at x = 5. In other words, Y is increasing until x = 5, and then it starts
decreasing. This means that Y is at its highest point at x = 5.
Note:

First-Derivative Test Relative Extremum:

Suppose that f (x0) = 0. Then the value of the function at x0, f(x0), is:
a) A relative maximum occurs when the sign of f’(x) changes from positive to
negative as you move from the immediate left to the immediate right of the point
x0.
(b) A relative minimum occurs when the sign of f(x) changes from negative to
positive as you move from the immediate left to the immediate right of the point
x0.
(c) An inflection point, which is not an extreme point, occurs when f(x) has the
same sign within a certain neighbourhood.
Example 3.1:
Find the relative extremum of the
y = f(x) = x3 – 12x2 + 36x + 8.
Solution:
Since f (x) = 3x2 – 24x + 36, to get the critical values, i.e., the values of the function
at x satisfying the condition f (x) = 0, we set f (x) = 0, and thus
3x2 – 24x + 36 = 0.
Its roots are x 1 = 2 and x 2 = 6. It is easy to verify that f (x) > 0 for x < 2 and
f (x) < 0 for x > 2. Thus x = 2 is a maximal point and the corresponding maximum
value of the function f(2) = 40. Similarly, we can verify that x = 6 is a minimal point and
f (6) = 8.
Example 3.2:
Self-Instructional
76 Material Find the relative extremum of the average-cost function.

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Unconstrained Optimisation

AC = f(Q) = Q2 – 5Q + 8 NOTES
f’(Q) = 2Q – 5 = 0
Q = 5/2 = 2.5
Since f(2.5) = 0, f(Q) < 0 for Q < 2.5, and f(Q) > 0 for Q > 2.5, so the given
average cost function is minimum at Q = 2.5.

3.4 SECOND AND HIGHER DERIVATIVES

Since the first derivative f (x) of a function y = f(x) is also a function of x, we can
consider the derivative of f (x), which is called the second derivative. Similarly, we
can find derivatives of even higher orders. These derivatives will enable us to develop
alternative criteria for locating the relative extremum of a function.
The second derivative of the function, f is denoted by f (x) or d2y/dx2. If the
second derivative f(x) exists for all x values, f(x) is said to be twice differentiable; if,
in addition, f(x) is continuous, f(x) is said to be twice continuously differentiable.
Before explaining whether the function attains maximum or minimum at x = x0,
we should be aware of the nature of the curve of a function which can be explained
with the help of the 1st and 2nd derivatives.

dy d2 y
(1) If the 1st derivative  0 the second order derivative >0 then the
dx dx 2
curve of a function is increasing at increasing rate shown in Fig. 3.4.

Self-Instructional
Material 77
Fig. 3.4

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES
dy
(2) If the 1st derivative  0, then the curve of a function increases at the
dx
diminishing rate, as shown in Fig. 3.5.

Fig. 3.5

dy
(3) If the 1st derivative  0, then curve of a function decreases at an increasing
dx
rate, as shown in Fig. 3.6.

Fig. 3.6

dy d2 y
(4) If 1st derivative  0 and 2nd derivative 2 >0 then the curve of a function
dx dx
Self-Instructional
y = f(x) decreases at decreasing rate.
78 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Unconstrained Optimisation

NOTES

Fig. 3.7

Through the 2nd derivative tests a function that attains maxima under the following
conditions:
Let the function be Y = f (x)

dy
 0 is the 1st order necessary condition
dx

d2 y
0 is the second order necessary condition
dx 2

d2 y
<0 is the second order sufficient condition
dx 2
A function y = f (x) attains minima under the following conditions:

dy
 0 is the 1st order necessary condition
dx

d2 y
0 is the second order necessary condition
dx 2

d2 y
>0 is the second order sufficient condition
dx 2

Self-Instructional
Material 79

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES
In-Text Questions
5. What is the term for a point on a curve where the curve changes concavity?
a) Root b) Stationary point
c) Inflection point d) Local extremum
6. What is the condition that must be satisfied for a particular result to be possible
called?
a) Necessary condition b) Sufficient condition
c) Objective condition d) Constrained condition

Example 3.3:
Y = x3 –10 x +15
For 1st order necessary condition,

dy
 3x 2  10  0
dx
Or x2 = 10/3
X =  10 / 3

X =  3.333
For 2nd order necessary condition,

d2 y
 6x
dx 2
Substituting the values of x; we get
When x = + 3.333 then the second order derivative is -1.89; hence, the given
function attains maximum at x = 3.333 .

When x = - 3.333 then the second order derivative is 1.89; hence, the given
Self-Instructional
80 Material function attains minimum at x = - 3.333.

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Unconstrained Optimisation

Note that when f (x0) = 0, f (x0) < 0 (f (x0) > 0) is a sufficient condition for NOTES
a relative maximum (resp. minimum) but not a necessary condition. However, the
condition f(x0)  0 (f(x0)  0) is a necessary (even though not sufficient) for a
relative maximum (resp. minimum).
Example 3.4:
Show that TR is a maximum when q is 18 for the non-linear demand schedule.
p = 194.4 – 0.2q2
Solution:
TR = pq
= (194.4 – 0.2q2) q
= 194.4q – 0.2q3
For a stationary point on this cubic function, the slope must be zero and so
dTR/dq = 194.4 – 0.6q2 = 0
194.4 = 0.6q2
324 = q2
±18 = q
When q is +18 then the second-order derivative is

d 2 TR
= –1.2q = –1.2(18) = –21.6 < 0
dQ2

Therefore, the second-order condition for a maximum is satisfied and TR is a

maximum when q is 18.
Note:
When q is -18 then the second-order derivative is

d 2TR
= –1.2q = –1.2(-18) = 21.6 > 0
dQ 2

Therefore, second-order condition for a minimum is satisfied and TR is a maximum when

q is -18.
Self-Instructional
Material 81

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES Example 3.5:

A firm faces the demand schedule p = 200 – 2q and the total cost function
TC = 2/3q3 – 14q2 + 222q + 50
Derive expressions for the following functions and find out whether they have
maximum or minimum points. If they do, say what value of q this occurs at and calculate
the actual value of the function at this output.
(a) Marginal cost
(b) Average variable cost
(c) Average fixed cost
(d) Total revenue
(e) Marginal revenue
(f) Profit
Solution:
(a) The marginal cost is the derivative of the total cost with respect to the quantity.
The total cost function is given by TC=32 q3–14q2+222q+50, so the marginal
cost function is

dTC d
MC= = (32 q3–14q2+222q+50)
dq dq
= 32 *3q2–28q+222
=2q2–28q+222.
(b) The average variable cost is the total variable cost divided by the quantity. The
total variable cost is the total cost minus fixed cost, and the fixed cost is the cost
that does not change with the level of output. In this case, the fixed cost is $50,
so the total variable cost is TVC=TC–50=32 q3–14q2+222q.
The average variable cost is then AVC=TVC/q=32 q2–14q+222.
(c) The average fixed cost is fixed cost divided by quantity. The fixed cost is $50,
so the average fixed cost is AFC=50/q.
(d) The total revenue is price multiplied by quantity. The demand schedule is given
Self-Instructional by p=200–2q, so the total revenue is
82 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Unconstrained Optimisation

TR=p*q NOTES
= (200–2q)*q
= 200q–2q2.
(e) The marginal revenue is the derivative of the total revenue with respect to quantity.
The total revenue function is given by
TR=200q–2q2, so the marginal revenue function is
MR=200–4q.
(f) Profit is the total revenue minus total cost. The total revenue is
TR=200q–2q2 and total cost is TC=32 q3–14q2+222q+50, so profit is
=TR–TC
=200q–2q2– (32 q3–14q2+222q+50)=31 q3+12q2–222q–50.
To find the maximum or minimum points of each function, we can take the
derivative and set it equal to zero. For the marginal cost function,
MC=2q2–28q+222,
the derivative is MC =4q–28.
Setting this equal to zero,
4q–28 = 0
q = 7.
This is the point at which marginal cost is minimised.
For the average variable cost function,
AVC = 32 q2–14q+222, the derivative is
AVC = 34 q–14.
Setting this equal to zero,
34 q–14=0
q=221 .
This is the point at which the average variable cost is minimised.
For the average fixed cost function, AFC=50/q, the derivative is
Self-Instructional
AFC =–50/q2. Material 83

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES This is always negative, so the average fixed cost is always decreasing.
For the total revenue function,
TR = 200q–2q2, the derivative is
TR2 = 200–4q.
Setting this equal to zero,
200–4q = 0
q = 50.
This is the point at which the total revenue is maximised.
For the marginal revenue function, MR=200–4q, the derivative is MR =–4.
This is always negative, so the marginal revenue is always decreasing.
For the profit function,
 = 31 q3+12q2–222q–50,
the derivative is  =q2+24q–222
Setting this equal to zero,
q2+24q–222= 0
q=11 or q=–20. Since quantity cannot be negative, the only relevant point is
q=11. This is the point at which profit is maximised.

3.5 OPTIMISATION IN CASE OF TWO VARIABLES

Unconstrained optimisation refers to the process of finding the minimum or maximum

of a function without any restrictions on the values of the variables. In the case of two
variables, we typically aim to find the values of the variables that optimise the function.
To solve an unconstrained optimisation problem with two variables, one can
follow the following steps:
1. Define the objective function: Determine the function you want to optimise.
Let’s denote it as z= f(x, y), where x and y are the variables.
Self-Instructional
84 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Unconstrained Optimisation

2. Determine the optimisation goal: Decide whether you want to maximise or NOTES
minimise the objective function. This will depend on the specific problem you’re
trying to solve.
3. Find critical points: Critical points are the locations where the gradient of the
function is zero or undefined. To find them, calculate the partial derivatives of
f(x, y) with respect to x and y, and set them equal to zero. Solve the resulting
equations to find the critical points.
4. Determine the nature of critical points: Evaluate the second partial derivatives
of f(x, y) at the critical points to determine whether they correspond to a minimum,
maximum, or a saddle point. The second derivative test can be used to analyse
the nature of the critical points.
5. Check boundary points: If the domain of the function is bounded, evaluate
the objective function at the boundaries to check if any of them correspond to
the minimum or maximum.
6. Analyse and compare results: Compare the values of the objective function
at the critical points and boundary points to determine the global minimum or
maximum, if it exists.

Geometrical Characteristics
Techniques of optimisation and geometrical characteristics could be analogously
constructed from one variable to two variables. For a function of one variable, an
extreme value is represented by the peak of a hill or bottom of a valley in a two-
dimensional graph as in Fig. 3.2. When z becomes a function of x and y (z = f(x, y)),
now plotting this graph in 3D space these hills and valleys will appear as domes and
bowls like in Fig. 3.8 and Fig. 3.9

Self-Instructional
Material 85
Fig. 3.8

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES

Fig. 3.9

In 2-D space, when a function of one variable is graphed at points, where it

attains minimum or maximum, the tangent is parallel to the x-axis, i.e., the axis of a
choice variable. At the point, A in Fig. 3.2, f (x) attains a maximum and the tangent is
parallel to the x-axis. Likewise, in Fig. 3.8; at the point A, the tangent, Tx is parallel to
the x- axis in the xz plane and Ty is parallel to the y-axis in the yz plane. Alternatively,
like in Fig. 3.10, we could have a plane tangent to z at the point, A. The ‘hill’ in 2D is
equivalent to dome in 3D.

Fig. 3.10

Self-Instructional
86 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Unconstrained Optimisation

Similarly, at the minimum value at the point, B in Fig.3.2, we have a tangent NOTES
parallel to the x-axis. At the point, B’, in Fig. 3.9, we have Tx: a tangent parallel to the
x-axis in the xz plane and Ty is parallel to the y-axis in the yz plane. The valley in 2D
(for minimum) is transformed as a bowl in 3D.
In Fig. 3.2, at the point, C, the tangent is parallel to the x-axis but x is neither a
maximum nor minimum but rather a point where the curvature of the curve changes,
i.e., inflation point. Similarly, in Fig. 3.11 and 3.12, points C and C, respectively, are
not optimum points even when tangents, Tx and Ty are parallel to the x-axis in the xz
plane and the y-axis in the yz plane, respectively.

Fig. 3.11 Fig.3.12

At the point, C’ in Fig. 3.11, it is a minimum when viewed in background of the

yz plane and a maximum when viewed in the background of the xz plane. A point with
such a dual personality is known as ‘saddle’ point. At C, in Fig. 3.12, then the
surface gets twisted so it is an inflexion point like the point, C in Fig. 3.2.

3.6 DIFFERENTIAL CONDITIONS FOR

OPTIMISATION

Let’s consider the functional form as:

z = f (x, y) .
The first-order condition for an extremum is dz = 0 . With reference to z = f (x,
y); dz is total differential given by:
Self-Instructional
Material 87

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES dz = fxdx + fydy

And dz = 0 implies that an extremum point must be a stationary point and
simultaneously z must be constant for changes of two variables, x and y. This implies
that both dx and dy are not zero.
Then for dz to be zero, it is necessary that partial derivatives, fx and fy must be
zero. So, the first-order condition becomes:
fx = fy = 0
This condition was satisfied by the point, A in Fig. 3.8. B in Fig. 3. 9, C in Fig.
3.11 and C in Fig. 3.12. But at points, C and C the function did not attain an
extreme value so, this condition though necessary but is not sufficient.
As in one variable case second order, total differential will determine whether
any point is extremum or not. For the function:
z = f (x, y); d2z is calculated as follows:
d2z = d(dz)

 (dz )  (dz )
 dx  .dy
x y

 
 ( f x dx  f y dy )dx  ( f x dx  f y dy )dy
x y

= fxxdx2 + fyxdydx + fxydxdy + fyydy2

By young’s theorem, all cross partial derivatives are identical, i.e.,
fyx= fxy. So d2z = fxxdx2 + 2fxydxdy + fyydy2.
For maximum of z, d2z < 0 becomes a sufficient condition for maximum of z for
dx and dy both not being zero and likewise for minimum of z. For any values of dx and
dy, not both zero.

 0 if f xx  0; f yy  0 and f xx f yy  f xy2

2
d z
 0 if f xx  0; f yy  0 and f xx f yy  f xy2

Self-Instructional
88 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Unconstrained Optimisation

NOTES
In-Text Questions
7. What is the term for a line that touches a curve at a single point, representing
the slope of the curve at that point?
a) Maximum b) Minimum
c) Tangent d) Saddle point
8. What is the term for the transition from positive to negative or vice versa in a
function?
a) Sign change b) Curve nature
c) Slope change d) Maximum shift
Example 3.6: Consider the function z = e2x – 2x +2y2 + 3. To find extreme values of
z, following are relevant partial derivatives:
fx= ze2x – 2 fxx = 4e2x
fy = 4y fyy = 4 fxy = 0
Solution:
First-order necessary conditions become:
fx = 0
 2e2x – 2 = 0
 e2x = 1
 2x = 0
 x=0
and fy = 0
 4y = 0
 y=0
So, the stationary point for z is (0, 0). To know whether z attains minimum or
maximum at (0, 0), let’s check second order conditions.
fxx(0, 0) = 4
fyy(0, 0) = 4
Self-Instructional
Material 89

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES fxy(0, 0) = 0
fxx fyy = 16
 fxx > 0; fyy > 0 and fxx fyy > fxy2
hence z attains minimum at point (0, 0).
z at (0, 0) is e0 – 0 + 0 + 3 = 4 so th maximum point is denoted by (0, 0, 4)
Example 3.7:
A firm produces two products which are sold in two separate markets with demand
schedules.
p1 = 600 – 0.3q1
p2 = 500 – 0.2q2
Production costs are related and the firm faces the total cost function
TC = 16 + 1.2q1 + 1.5q2 + 0.2q1q2
If the firm wishes to maximise total profits, how much of each product should it
sell? What will be the maximum profit level?
Solution:
The total revenue is
TR = TR1 + TR2
= p1q1 + p2q2
= (600 – 0.3q1) q1 + (500 – 0.2q2)q2
= 600q1 – 0.3q12 + 500q2 – 0.2q22
Therefore, the profit is
 = TR – TC
= 600q1 – 0.3q21 + 500q2 – 0.2q22 – (16 + 1.2q1 + 1.5q2 + 0.2q1q2)
= 600q1 – 0.3q21 + 500q2 – 0.2q22– 16 – 1.2q1 – 1.5q2 – 0.2q1q2
= –16 + 598.8q1 – 0.3q21+ 498.5q2 – 0.2q22 – 0.2q1q2
First-order conditions for maximisation of this profit function are
/ q1 = 598.8 – 0.6q1 – 0.2q2 = 0 …(3.1)
Self-Instructional
90 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Unconstrained Optimisation

and NOTES
/ q2 = 498.5 – 0.4q2 – 0.2q1 = 0 …(3.2)
Simultaneous equations (3.1) and (3.2) can now be solved to find the optimal
values of q1 and q2.
Multiplying (3.2) by 3
1,495.5 – 1.2q2 – 0.6q1 = 0
Rearranging (3.1)
598.8 – 0.2q2 – 0.6q1 = 0
Subtracting gives 896.7 – q2 = 0
Giving the optimal value 896.7 = q2
Substituting this value for q2 into (3.1)
598.8 – 0.6q1 – 0.2(896.7) = 0
598.8 – 179.34 = 0.6q1
419.46 = 0.6q1
699.1 = q1
Checking second-order conditions by differentiating (3.1) and (3.2) again:

2 2
  0.6  0   0.4  0
q12 q22

This satisfies one set of second-order conditions for a maximum. The cross
partial derivative will be:
2
  0.2
q1q2
Therefore,
2
  2    2  2  2 
 2   2    0.6   0.4   0.24  0.04   0.2    
 q1   q2   q1 q2 

Self-Instructional
Material 91

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES and so the remaining second-order condition for a maximum is satisfied.

The actual profit is found by substituting the optimum values q1 = 699.1 and q2
= 896.7 into the profit function. Thus,
 = –16 + 598.8q1 – 0.3q21 + 498.5q2 – 0.2q22– 0.2q1q2
= –16 + 598.8(699.1) – 0.3(699.1) 2 + 498.5(896.7) – 0.2(896.7) 2
– 0.2(699.1)(896.7)
= £432, 797.02
Example 3.8:
A firm sells its output in a perfectly competitive market at a fixed price of £200 per
unit. It buys the two inputs, K and L at prices of £42 per unit and £5 per unit, respectively,
and faces the production function.
q = 3.1K0.3 L0.25
What combination of K and L should it use to maximise profit?
Solution:
TR = pq = 200(3.1K0.3 L0.25) = 620K0.3L0.25
TC = 42K + 5L
Therefore, the profit function the firm wishes to maximise is
 = TR –TC = 620K0.3 L0.25 – 620K0.3L0.25 – 42K – 5L
First-order conditions for a maximum require

 
 186 K 0.7 L0.25  42  0  155 K 0.3 L0.75  5  0
K L
Giving
186L0.25 = 42K0.7 and 155K0.3 = 5L0.75
42 0.7
L0.25  K (3.3)
186

31K0.3 = L0.75 (3.4)

Taking (3.3) to the power of (3.5)
Self-Instructional
92 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Unconstrained Optimisation

3
NOTES
 42 0.7  42 2.1
L0.75  K   K (3.5)
 186  1863
Setting (3.5) equal to (3.4)
42 2.1
K  31K
186
3
31186 
1.8
K   2, 692.481
423
K  80.471179 (3.6)
Substituting (3.6) into (3.3)

42
L0.25  (80.471179)0.7  4.8717455
186
4
L  ( L0.25 ) 4   4.8717455   563.29822

Therefore, first-order conditions suggest that the optimum values are L = 563.3
and K = 80.47
Checking second-order conditions:

2 
 0.7  186 K 1.7 L0.25
K 2
0.25
  130.2(80.47) 1.7  563.3
  0.3653576  0
2 
2
 0.75  155K 0.3 L1.75
L
1.75
  116.25(80.47)0.3  563.3
  0.0066572  0
2
  0.25  186 K 0.7 L0.75
K L
0.75
  46.5(80.47)0.7  563.3 Self-Instructional
Material 93
  0.0186404

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES
  2    2  
 2  2    0.3653576   0.0066576   0.0024323
 K  L 
2
 2 
    0.0186404   0.0003475
2

 K L 
Therefore,
2
 2  2   2 
 
K 2 L2  K L 
and so all second-order conditions for maximum profit are satisfied when K =
80.47 and
L = 563.3. The actual profit will be
 = 620K0.3L0.25 – 42K – 5L
= 620(80.47)0.3(563.3)0.25 – 42(80.47) – 5(563.3)
= 11,265.924 – 3,379.74 – 2,816.5
= £5,069.68

3.7 CONVEX AND CONCAVE FUNCTIONS AND

INFLECTION POINTS

As discussed above, the first derivative determines whether a function is increasing or

decreasing:
f (x)  0 on (a, b)  f(x) is increasing on (a, b) (3.7)
f (x)  0 on (a, b)  f(x) is decreasing on (a, b) (3.8)
And the second order derivative:
f (x)  0 on (a, b)  f (x) is increasing on (a, b) (3.9)
f (x)  0 on (a, b)  f (x) is decreasing on (a, b) (3.10)
Self-Instructional
94 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Unconstrained Optimisation

The equivalence in Equation (3.9) is illustrated in Fig. 3.8. The slope of the NOTES
tangent, f’(x), is increasing as x increases. On the other hand, the slope of the tangent
to the graph in Fig. 3.14 is decreasing as x increases.
The test for concavity and convexity of a function: f  test.
Intuitively, if the graph of f lies above all of its tangents or an interval I, then it is
called concave upward or convex function on I. If the graph of f lies below all of its
tangents or I, it is called concave downward or simply concave function.

Fig. 3.13 The slope of the tangent Fig. 3.14 The slope of the tangent
increases as x increases. decreases a x increases.
f(x) in increasing. g¢(x) is decreasing
The new definition is; assuming that f is continuous in the interval I and twice
differentiable in the interior of I, denoted by I0:
f is convex on I  f (x)  0 for all x in I0
f is convex on I  f (x)  0 for all x in I0
Example 3.9:
Check the convexity/concavity of the following:
(a) f(x) = x2 – 2.x + 2 and
(b) f(x) = ax2+ bx c
(Example from Sydsaeter and Hammond)
Solution:
(a) Here f (x) = 2.x–2 so f (x) = 2. Because f (x) > 0 for all x, f is convex.
(b) Here f (x) = 2ax + b, so f (x) = 2a.
If a = 0, then f is linear and f is convex as well as concave.
Self-Instructional
If a > 0, then f  (x) > 0, so f is convex. If a < 0, then f (x) < 0, so f is concave. Material 95

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES

3.8 SUMMARY

Unconstrained optimisation involves finding the maximum or minimum value of a function

without any restrictions. It is a problem-solving technique commonly used in economics
and other fields. The process of solving an unconstrained optimisation problem includes
several steps. First, the function to be optimised is defined, which can be any
mathematical expression. Then, the critical points of the function are identified, where
the derivative is either zero or undefined. These critical points are then classified as
maxima, minima, or saddle points. Afterward, the optimal solution is chosen, which is
the point with the highest or lowest value depending on whether the function is being
maximised or minimised.
Optimisation problems are often encountered in economics, where the objective
is to maximise or minimise some criteria such as utility or profit. In such problems, an
objective function is formulated with dependent and independent variables. The first
step is to delineate the objective function and identify the independent variables as
choice variables.
To determine whether a function attains a maximum or minimum, various tests
are used. The first derivative test involves finding the derivative of the function and
equating it to zero. The sign change of the derivative indicates whether the point is a
maximum or minimum. A positive to negative sign change implies a maximum, while a
negative to positive sign change implies a minimum. The second derivative test considers
the second derivative of the function and determines whether it is positive or negative.
A negative second derivative corresponds to a maximum, while a positive second
derivative corresponds to a minimum.
Self-Instructional
96 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Unconstrained Optimisation

The first and second derivative tests help identify the relative extrema of a NOTES
function, which are points where the function reaches its highest or lowest value within
a certain interval. These tests involve analysing the slope of the tangent line to the
function at a particular point and examining the sign change of the derivative. By finding
the critical values and evaluating the derivatives, one can determine whether a function
attains a maximum or minimum at a specific point.

3.9 GLOSSARY

 Unconstrained optimisation: It is the process of finding the maximum or

minimum value of a function without any constraints.
 Single-variable optimisation: It is the optimisation involving a function of a
single variable.
 Constraints: They are limitations or conditions that restrict the feasible solutions
in an optimisation- problem.
 Critical points: These are the points where the derivative of a function is either
zero or undefined.
 Maxima: It is the plural of maximum, representing multiple highest points of
a function.
 Minima: It is the plural of minimum, representing multiple lowest points of a
Function.
 Saddle points: These points exist on a surface where the function has neither a
maximum nor a minimum value, but instead, the surface curves up and down.
 Objective function: It is the function that is to be optimised in an optimisation
problem.
 Extremum: It is a maximum or minimum value of a function.
 Local maximum: It is the highest point in a specific region of a function.
 Local minimum: It is the lowest point in a specific region of a function.
 Stationary point: It is a point where the derivative of a function is zero.
Self-Instructional
Material 97

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES  Relative extremum: It is a maximum or minimum value that occurs within a

specific range or interval.
 Inflection point: It is a point on a curve where the curve changes concavity.
 Even higher derivatives: They include successive derivatives beyond the
second derivative of a function.
 Necessary condition: It is a condition that must be satisfied for a particular
result to be possible.
 Sufficient condition: It is a condition that guarantees a particular result.

3.10 ANSWERS TO IN-TEXT QUESTIONS

1. (c) Unconstrained optimisation

2. (c) Maximum
3. (b) Critical points
4. (a) First derivative test
5. (c) Inflection point
6. (a) Necessary condition
7. (c) Tangent
8. (a) Sign change

3.11 SELF-ASSESSMENT QUESTIONS

1. A monopolist produces two commodities that are substitutes and having demand
functions: X1=8-P1+P2 and X2=9+P1-5P2, where 1,000 X1 units of first
commodity are demanded if its price is Rs P1 per unit and 1,000X2 Units of the
Second commodity are demanded if its price is Rs P2 per units. It costs Rs 4 to
produce each unit of the first commodity and Rs 2 to produces each unit of the
Self-Instructional second. Find the output levels and prices of the two commodities in order to
98 Material
have maximum profit.

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Unconstrained Optimisation

2. A monopolist sells two products x and y for which the demands are: x=50- NOTES
0.5px and y-76-py
The combined cost function is C=3X2+2XY+2Y2+55.Find:
(i) The profit maximising levels of output and price for each product.
(ii) The maximum profit.
3. A scooter manufacturer produces the same model of a scooter at two different
production plants. The cost of production of x1 scooters at plant I is given by
C1= +1,000X1+2,500, and the cost of production of X2 scooters at plant II is
given by, C2=1.5 +2000X2+1800, Where x1 and x2 are the annual outputs of
plant I and II respectively.
(i) If each scooter is sold at a uniform price of Rs. 20,000, find the levels of
production of each plant so that profits are maximised.
(ii) If the annual demand of scooters follows the demand law x=30,000- p,
where p is the price of a scooter, find the levels of production of each
plant for maximum profits and the price of a scooter.
4. (a) A discriminating monopolist can separate his consumers into two distinct
markets with the following demand functions:
Market I: Q1=16-0.2P1 Market II: Q2=180-2P2
Assume that the monopolist’s total cost function takes the form TC+20Q-20=0,
Where Q (=Q1+Q2) is the total output. Obtain the total profit function and
determine the prices he would charge in the two markets to maximise profits.
What is the total profit? Do you agree that the price charged in the market with
a higher elasticity of demand would be higher? Show it by calculations.
(b) Calculate the ratio of prices charged by a discriminating monopolist in the
two markets with price elasticities of demand equal to 3.0 and 1.5.
5. Let there be two sellers of a homogeneous product with market demand given
by
P=B-ax,(x=x1+x2)a and B are positive constants. The cost functions of each of
the firm is given by C=axi, i=1, 2 and a is positive constant. Assuming that the
conjectural variations are zero,
Self-Instructional
Material 99

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES (i) Find the reaction curve of each seller.

(ii) Find equilibrium output of each seller and the industry.
(iii) Find equilibrium output if the industry becomes a monopoly.
6. The production function of a firm is given by q=12 and the prices (in Rs) of q, L
and K are 9, 2 and 4, respectively.
(i) Find the profit maximising values of q, L and K.
(ii) Find the amount of maximum profit.
(iii) Verify the second-order condition.

3.12 REFERENCE

 K. Sydsaeter and P. Hammond. 2002. Mathematics for Economic Analysis.

Pearson Educational Asia, Delhi.

Self-Instructional
100 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Constrained Optimisation

LESSON 4 NOTES

CONSTRAINED OPTIMISATION
Sarabjeet Kaur
Assistant Professor,
Department of Economics,
Zakir Husain College (E), University of Delhi

Structure
4.1 Introduction
4.2 Effects of a Constraint
4.3 Finding the Stationary Values
4.4 Constrained Optimisation by Substitution
4.5 The Lagrange Multiplier: Constrained Maximisation with Two Variables
4.6 Second-Order Condition
4.7 The Bordered Hessian
4.8 Summary
4.9 Glossary
4.10 Answers to In-Text Questions
4.11 Self-Assessment Questions

4.1 INTRODUCTION

Constrained optimisation is a mathematical optimisation problem where the objective

function is optimised subject to a set of constraints. In this type of optimisation, you
not only aim to find the maximum or minimum value of the objective function, but also
consider the limitations or conditions imposed by the constraints.
Constraints in constrained optimisation can be inequalities or equalities that restrict
feasible solutions. These constraints define the feasible region or feasible set, which is
the subset of the solution space that satisfies all the given constraints. The goal is to find
the optimal solution that simultaneously satisfies all the constraints and optimises the
objective function. Self-Instructional
Material 101

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES
4.2 EFFECTS OF A CONSTRAINT

In general, for a function, say z = f (x, y), the difference between a constrained extremum
and a free extremum may be illustrated in Fig. 4.1.

Fig. 4.1: Difference between a constrained extremum and a free extremum

In Fig. 4.1, the free extremum is the point at the top of the entire domain, while
the constrained extremum is the point at the top of the inverse U-shaped curve that sits
on the top of the constraint line. In general, a constrained maximum (which has less
freedom) can be expected to have a lower value than a free maximum (which has
more freedom). However, the two maxima may happen to have the same value by
coincidence. In no case, the constrained maximum can exceed the free maximum. To
have some degree of freedom of choice, the number of constraints should generally be
less than the number of choice variables.
 A free extremum is a point in the domain of a function where the function
reaches its maximum or minimum value. A constrained extremum is a point
in the domain of a function where the function reaches its maximum or
minimum value, subject to a set of constraints.
 In the figure, the constraint line is a line that represents the set of all points
Self-Instructional
that satisfy the constraint. The inverse U-shaped curve is the graph of a
102 Material function that is constrained to lie below the constraint line.

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Constrained Optimisation

 The free extremum is at the top of the entire domain because it is the point NOTES
where the function reaches its maximum value without any constraints. The
constrained extremum is at the top of the inverse U-shaped curve because it
is the point where the function reaches its maximum value while still satisfying
the constraint.
 In general, a constrained maximum will have a lower value than a free
maximum because the constrained maximum is limited by the constraint.
However, the two maxima may happen to have the same value if the
constraint is not very restrictive.

In-Text Questions
1. What is constrained optimisation?
a) Maximising a function without any constraints
b) Minimising a function without any constraints
c) Maximising or minimising a function subject to constraints
d) Finding the derivative of a function
2. The function to be maximised or minimised in a constrained optimisation problem
is called the _______________.
a) Feasible region b) Objective function
c) Constrained extremum d) Lagrangian function

4.3 FINDING THE STATIONARY VALUES

Let us consider a problem.

Example 4.1:
Find the stationary values of the function f(x, y) = x2 + y2, subject to the constraint g(x,
y) = x + y = 10.
Solution:
To find the stationary values, we need to find the critical points of the function f(x, y)
Self-Instructional
while satisfying the constraint g(x, y). Material 103

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES Step 1: Solve the constraint equation for one variable. In this case, we can solve
g(x, y) = x + y = 10 for y:
y = 10 - x.
Step 2: Substitute the constraint solution into the objective function f(x, y):
f(x) = x2 + (10 - x)2
= x2 + 100 - 20x + x2
= 2x2 - 20x + 100
Step 3: Differentiate f(x) with respect to x to find the critical points:
f’(x) = 4x - 20
Step 4: Set f’(x) = 0 and solve for x:
4x - 20 = 0
4x = 20
x=5
Step 5: Substitute the value of x back into the constraint equation to find y:
y = 10 - x
y = 10 - 5
y=5
Therefore, the critical point is (x, y) = (5, 5).
Step 6: Determine whether the critical point is a maximum, minimum, or neither.
To do this, we can use the second derivative test. Calculate the second derivative of
f(x):
f(x) = 4
Since the second derivative is positive (f (x) > 0), the critical point is a minimum.
So, the stationary value of the function f(x, y) = x2 + y2, subject to the constraint
g(x, y) = x + y = 10, is (x, y) = (5, 5), and it corresponds to a minimum.
Let’s consider an objective function which is either to be maximised or minimised.
Like

Self-Instructional
104 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Constrained Optimisation

Utility function, denoted by NOTES

U = f(x,y)
Output function, denoted by
Q= f (k,l)
Profit function, denoted by
 = f(r,c)
 Utility function of a consumer is constraint by household income.
 Output function is constraint by resources available (generally cost) with a
producer.
 Profit function is constraint by both cost and output.
We shall consider two methods:
(i) Constrained optimisation by substitution, and
(ii) The Lagrange multiplier method.
The Lagrange multiplier method can be used for most types of constrained
optimisation problems. The substitution method is mainly suitable for problems where
a function with only two variables is maximised or minimised subject to one constraint.
We shall consider this simpler substitution method first.

4.4 CONSTRAINED OPTIMISATION BY

SUBSTITUTION

Consider the example of a firm that wishes to maximise output Q = f(K, L), with a
fixed budget M for purchasing inputs, K and L at set prices, PK and PL. This problem
is illustrated in Fig. 4.2. The firm needs to find the combination of K and L that will
allow it to reach the optimum point, X which is on the highest possible isoquant within
the budget constraint with intercepts, M/PK and M/PL.

Self-Instructional
Material 105

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES

Fig. 4.2

To determine a solution for this type of economic resource allocation problem,

we have to reformulate it as a mathematical constrained optimisation problem.

In-Text Questions
3. Feasible solutions in constrained optimisation are _____________________.
a) Solutions that satisfy all the given constraints
b) Solutions that don’t satisfy any constraints
c) Solutions that maximise the objective function
d) Solutions that minimise the objective function
4. Which method is used to solve optimisation problems with a single constraint
by substituting one variable from the constraint into the objective function?
a) Feasible region b) Marginal product
c) Substitution method d) Budget constraint

The following examples suggest ways in which this can be done.

Example 4.2:
A firm faces the production function, Q = 12K0.4L0.4 and can buy the inputs, K and L
at prices per unit of £40 and £5, respectively. If it has a budget of £800, what
combination of K and L should it use in order to produce the maximum possible
output?
Self-Instructional
106 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Constrained Optimisation

Solution: NOTES
The problem is to maximise the function Q = 12K0.4L0.4 subject to the budget constraint
40K + 5L = 800 (4.1)

The theory of the firm tells us that a firm is optimally allocating a fixed budget if
the last £1 spent on each input adds the same amount to output, i.e., marginal product
over price should be equal for all inputs. This optimisation condition can be written as:

MPK MPL
 (4.2)
PK PL

The marginal products can be determined by partial differentiation:

Q
MPK = = 4.8K –0.6L0.4 (4.3)
K

Q
MPL = = 4.8K 0.4L-0.6 (4.4)
L
Substituting (4.3) and (4.4) and the given prices for PK and PL into (4.2)
(4.8K –0.6L0.4)/40 = (4.8K 0.4L-0.6)/5
Dividing both sides by 4.8 and multiplying by 40 gives
K-0.6L0.4 = 8K0.4L”0.6
Multiplying both sides by K0.6L0.6 gives
L = 8K (4.5)
Substituting (4.5) for L into the budget constraint (4.1) gives
40K + 5(8K) = 800
40K + 40K = 800
80K = 800
Thus, the optimal value of K is
K = 10
and, from (4.5), the optimal value of L is
Self-Instructional
L = 80 Material 107

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES Note that although this method allows us to derive optimum values of K and L
that satisfy condition (4.2) above, it does not provide a check on whether this is a
unique solution, i.e., there is no second-order condition check. However, it may be
assumed that in all the problems, in this section, the objective function is maximised (or
minimised depending on the question) when the basic economic rules for an optimum
are satisfied.

4.5 THE LAGRANGE MULTIPLIER: CONSTRAINED

MAXIMISATION WITH TWO VARIABLES

The Lagrange multiplier method is a strategy for finding the local maxima and minima
of a function subject to equation constraints. In the case of constrained maximisation
with two variables, the method works by introducing a new variable, called the Lagrange
multiplier, and forming a new function called the Lagrangian. The Lagrangian is equal
to the original function plus the product of the Lagrange multiplier and the constraint
function. The Lagrange multiplier is then found by setting the partial derivatives of the
Lagrangian with respect to each of the variables equal to zero. The solutions to these
equations are the points that maximise the original function subject to the constraint.

Utility Maximisation

Let the utility of the consumer is the function of two good called x1, x2, i.e.,
u = f (x1 x2) (4.6)
Equation (4.6) is known as objective function.
Let the consumer have ‘M’ money income which has to be spent on two goods,
x1 and x2, i.e.,
M = P1 x1 + P2 x2 (4.7)
Where P1 and P2 are unit prices of good x1 and x2, respectively.
L = u(x1 x2 ) + (M - P1 x1 + P2 x2) (4.8)
Where L is Lagrange function.
Self-Instructional
108 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Constrained Optimisation

x1 , x2 are two goods. NOTES

 is Lagrangian multiplier
P1 and P2 are price of goods x1 and x2, respectively.
‘M’ is Money income
From Equation (4.8) through Lagrangian multiplier, we have to find out the
values of x1 x2, within the given constraint.
M = P1 x1 + P2 x2, maximise the utility, ‘u’
For this, we differentiate Equation (4.8) w.r.t. x1, x2 and ; and equate to zero
for the 1st order condition.
L = u(x1 x2) + [M – P2x2]

L
 fx1  P1  0 (4.9)
x1

L
 fx2  P2  0 (4.10)
x2

L
 M  P1x1  P2 x2  0 (4.11)

For critical value of x1 and x2 dividing equations (4.9) by (4.10)

fx1 P1
 (4.12)
fx2 P2

we know that is also known as i.e., for equilibrium fx1/fx2 = MRS

MRS = p1/p2 (4.13)
where MRS is the marginal rate of substitution
To find out whether ratio of Equation (4.12) or (4.13) maximises the utility or
not may be checked by border Hessian condition in which we use constraint as a
border in the matrix as follows:

Self-Instructional
Material 109

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES In the border Hessian, the constraint is used as a border.

If

Then it represents minima

And if

Then it represents maxima.

To optimise above objective function, we may use the Lagrangian multiplier
method as follows:
Example 4.3:
The firm is trying to maximise the output.
Q = 12K0.4L0.4
subject to the budget constraint
40K +5L = 800.
The first step is to rearrange the budget constraint so that zero appears on one
side of the equality sign. Therefore,
0 = 800 – 40K – 5L (4.14)
We then write the ‘Lagrange equation’ or ‘Lagrangian’ in the form
Self-Instructional F(L,K,) = (function to be optimised) + (constraint)
110 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Constrained Optimisation

where F(L,K,) is just the value of the Lagrangian function and NOTES
 is known as the ‘Lagrange multiplier’.
In this problem, the Lagrange function is thus
F(L,K,) = 12K0.4L0.4 + (800 – 40K – 5L) (4.15)
Then derive the partial derivatives of F(L,K,) with respect to K, L and  and
set them equal to zero, i.e., find the stationary points of L that satisfy the first-order
conditions for a maximum.

f  L, K, λ 
= 4.8K–0.6L0.4 – 40 = 0 (4.16)
L

f  L, K, λ 
= 4.8K0.4L-0.6 – 5 = 0 (4.17)
L

f  L, K, λ 
= 800 – 40K – 5L = 0 (4.18)
L
You will note that (4.18) is the same as the budget constraint (4.14). We now
have a set of three linear simultaneous equations in three unknowns to solve for K and
L. The Lagrange multiplier,  can be eliminated as, from (4.16),
0.12K–0.6L0.4 =  and from (4.17)
0.96K0.4L–0.6 = 
Therefore,
0.12K–0.6L0.4 = 0.96K0.4L”0.6
Multiplying both sides by K0.6L0.6,
0.12L = 0.96K
L = 8K (4.19)
Substituting (4.19) into (4.18),
800 – 40K – 5(8K) = 0
800 = 80K
10 = K Self-Instructional
Material 111

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES Substituting back into (4.18),

800 – 40(10) – 5L = 0
400 = 5L
80 = L
These are the same values of K and L as those obtained by the substitution
method. Thus, the values of K and L that satisfy the first-order conditions for a maximum
value of the Lagrangian function are the values that will maximise the output subject to
the given budget constraint.
Strictly speaking we should now check the second-order conditions in the above
problem to ensure that we actually have a maximum rather than a minimum. These
conditions, however, are rather complex, involving an examination of the function at
and near the stationary points found, and are discussed in the next section. For the
time being, you can assume that once the stationary points of the Lagrangian function
have been found, the second-order conditions for a maximum will automatically be
met.
Example 4.4:
The prices of inputs, K and L are given as £12 per unit and £3 per unit, respectively,
and a firm operates with the production function Q = 25K0.5L0.5.
(i) What is the minimum cost of producing 1,250 units of output?
(ii) Demonstrate that the maximum output that can be produced for this budget will
be the 1,250 units specified in (i) above.
Solution:
This question essentially asks us to demonstrate that the constrained maximisation and
minimisation methods give consistent answers.
(i) The output constraint is that
1,250 = 25K0.5L0.5
The objective function to be minimised is the cost function.
TC = 12K + 3L
The corresponding Lagrange function is therefore,
Self-Instructional
112 Material
G = 12K + 3L + (1,250 – 25K0.5L0.5)

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Constrained Optimisation

First-order conditions require: NOTES

G 12 K 0.5
 12  12.5 K L  0 giving 12.5L0.5  
0.5 0.5 (4.20)
K
G 3L0.5
0.5 0.5
 3  12.5 K L  0 giving  (4.21)
L 12.5K 0.5
G
 1, 250  25 K 05 L0.5  0 (4.22)

Setting (4.20) equal to (4.21)

12 K 0.5 3L0.5

12.5 L0.5 12.5K 0.5
4K=L (4.23)
Substituting (4.23) into (4.22)
1,250 – 25K0.5(4K)0.5 = 0
1,250 = 25K0.5(4)0.5K0.5
1,250 = 50K
25 = K
Substituting this value into (4.23)
4(25) = L
100 = L
When these optimum values of K and L are used, the actual minimum cost will
be
TC = 12K + 3L = 12(25) + 3(100) = 300 + 300 = £600
(ii) This part of the question requires us to find the values of K and L that will
maximise the output subject to a budget of £600, i.e., the answer to (i) above.
The objective function to be maximised is therefore Q = 25K0.5L0.5 and the
constraint is 12K + 3L = 600. The corresponding Lagrange equation is thus
G = 25K0.5L0.5 + (600 –12K–3L)
Self-Instructional
Material 113

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES First-order conditions require:

G 12.5L0.5
 12.5 K 0.5 L0.5  12  0 giving  (4.24)
K 12 K 0.5
G 12.5K 0.5
 12.5 K 0.5 L0.5  3  0 giving  (4.25)
L 3L0.5
G
 600  12 K  3 L  0 (4.26)

Setting (4.24) equal to (4.25)

12.5 L0.5 12.5K 0.5


12 K 0.5 3L0.5
3L = 12K
L = 4K (4.27)
Substituting (4.27) into (4.26)
600 – 12K – 3(4K) = 0
600 – 12K – 12K = 0
600 = 24K
25 = K
Substituting this value into (8) L = 4(25) = 100
These are the same optimum values of K and L that were found in part (i)
above. The actual output produced by 25 of K plus 100 of L will be
Q = 25K0.5L0.5 = 25(25)0.5(100)0.5 = 1,250

In-Text Questions
5. What is a Lagrange multiplier used for?
a) Maximising the objective function
b) Minimising the objective function
c) Finding local maxima and minima subject to constraints
Self-Instructional
d) Evaluating the feasibility of a solution
114 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Constrained Optimisation

6. What is the purpose of the Lagrangian function in constrained optimisation? NOTES

a) Maximising the objective function
b) Minimising the objective function
c) Evaluating the feasibility of a solution
d) Combining the objective function and constraints
7. In the Lagrange multiplier method, a variable introduced to find constrained
maxima and minima is called the:
a) Feasible region
b) Objective function
c) Lagrange multiplier
d) Substitution method

4.6 SECOND-ORDER CONDITION

Optimisation is the process of finding the best possible outcome, whether that means
maximising or minimising something. For example, we might want to maximise bus
service in a city while minimising pollution, or we might want to maximise our grades
while minimising our efforts.
In order to find the optimal outcome, we need to consider all the possible
constraints. For example, we may have a limited amount of money or time.
The first-order conditions for optimisation are the same for both maximisation
and minimisation. These conditions state that the partial derivatives of the objective
function with respect to all of the variables must be equal to zero.
The second-order conditions can help us determine whether a solution is a
maximum or a minimum. These conditions state that the Hessian matrix of the objective
function must be a positive definite for a maximum, and a negative definite for a minimum.
The Hessian matrix is a square matrix that contains all of the second-order
partial derivatives of the objective function. If the Hessian matrix is a positive definite,
then the objective function is concave up at the solution, which means that it is a
maximum. If the Hessian matrix is a negative definite, then the objective function is Self-Instructional
concave down at the solution, which means that it is a minimum. Material 115

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES The total differential is a tool that can be used to calculate the change in an
objective function due to a change in the variables.
Example 4.5:
Find the extremum of z = xy subject to x + y = 6.
Solution:
The Lagrange function is Z = xy + (6 – x – y).
The first-order condition is Z = 6 – x – y = 0
Zx = y –  = 0;
Zy = x –  = 0.
Thus, optimum values are:  = 3, x = 3, y = 3.
An Interpretation of the Lagrange Multiplier
The Lagrange multiplier,  measures the sensitivity of Z to change in the constraint.
If we can express the solution  , x , and y (optimum values) all as implicit functions
of the parameter, c:

   (c), x  x  c  , and y  y  c 
All of which will have continuous derivative, we have the identities:

Thus, we can consider Z as a function of c:

Therefore, we have

Self-Instructional
116 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Constrained Optimisation

Total Differential NOTES

The concept of total differential is very useful in constrained optimisation. In discussing

the concept of total differential, remember that if we have a function, f (x,y) then the
f
symbol fx stands for .
x
If z = f (x,y) is a homogeneous function of the first degree, then the total differential
dz can be expressed as:

It may be noted that this formula holds well whether x and y are dependent or
independent variables. The expression, dz shows the increment in the function z f xy,
when there is an infinitesimal increments in x and well as y. For example, If z = x3 + y3
then total differential can be expressed as:
dz = fxdx + fydy = 3x2dx + 3y2dy
We can obtain the second-order total differential, d2z:

We know that if f(x, y) satisfies the conditions of Schwarz’s theorem, we have

fxy = fyx.
Example 4.6:
Find du when u =3x3+2y2+y3.
Solution: Total differential du is given by:
du = fxdx + fydy = 9x2dx + (uy + 3y2)dy
= 9x2dx + y(u + 3y)dy Self-Instructional
Material 117

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES Second-Order Conditions

For a constrained extremum of z = f(x, y), subject to g(x, y) = c, the second-order

necessary-and-sufficient conditions still revolve around the algebraic sign of the second-
order total differential d2z, evaluated at a s-stationary point. However, there is one
important change. In the present context, we are concerned with the sign definiteness
or semi-definiteness of d2z, not for all possible values of dx and dy (not both zero), but
only for those dx and dy values (not both zero) satisfying the linear constraint gxdx +
gydy = 0.
The set of (dx, dy) pairs that satisfy the constraint gxdx + gydy = 0 forms only a
subset of all possible dx and dy pairs. As a result, the constrained sign definiteness is
not as strict as it is in the case of unconstrained problems. In simpler terms, the second-
order sufficient condition for a constrained extremum problem is less stringent than
that for an unconstrained extremum problem.
The second order sufûcient conditions are:
 For maximum of z: d2z negative deûnite, subject to dg = 0
 For minimum of z: d2z positive deûnite, subject to dg = 0
It can be summarised as follows:
Condition Maximum Minimum
First-order necessary condition fx = fy = 0 fx = fy = 0
Second-order sufficient condition fxx, fyy < 0 fxx, fyy > 0
and fxx fyy > fxy2 and fxx fyy > fxy2

4.7 THE BORDERED HESSIAN

The bordered Hessian, also known as the augmented Hessian or the bordered matrix,
is a concept used in the optimisation theory to determine the nature of critical points
(e.g., minima, maxima, or saddle points) for a function subject to equality constraints.
It is an extension of the Hessian matrix, which is used in unconstrained optimisation
Self-Instructional problems.
118 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Constrained Optimisation

The bordered Hessian matrix includes an additional row and column compared NOTES
to the original Hessian matrix to incorporate the derivatives of the constraint function
g(x, y). The last row and column are filled with the partial derivatives of g(x, y) with
respect to x and y, respectively, while the last entry is always zero.
This determinant used in this case is nothing but the original determinant, with a
border (row) placed on the top and a similar border (column) on the left. Moreover,
this border is merely composed of the coefficients from the constraint, with a zero in
the principal diagonal. We had expressed the second-order differential as given in the
equation z = f (x, y), its second-order total differential is:
d2z = fxxdx2 + 2fxydxdy + fyydy2 + fyd2y.
The first-order conditions is given by:
Lx = fx – gx = 0
Ly = fy – gy = 0
L = c – g (x, y) = 0
We can partially differentiate the derivatives again to get:
Lxx = fxx – gxx = 0
Lyy = fyy – gyy = 0
Lxy = fxy – gxy = Lyx
Making use of the Lagrangian, we can express d2z as:
d2z = fxxdx2 + Lxydxdy + Lyxdydx + Lyydy2.
When we apply the bordered Hessian to the above case, we get the conditions
d2z is positive definite subject to dg = 0 if

 0 gx gy 
 
 g x Lxx Lxy 
(determinant) < 0
g Lyy 
 y Lyx

The condition for negative definiteness is similar with the sign of bordered Hessian
reversed (> 0).
Self-Instructional
Material 119

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES To analyse the critical points using the bordered Hessian, we consider the
eigenvalues of the bordered Hessian matrix. The nature of the critical point depends
on the eigenvalues:
1) If all eigenvalues are positive (Lxx > 0, Lyy > 0 ), the critical point is a
local minimum.
2) If all eigenvalues are negative (Lxx < 0, Lyy < 0), the critical point is a
local maximum.
3) If the eigenvalues have both positive and negative values, the critical point
is a saddle point.
4) If there are zero eigenvalues, further analysis is required.
The bordered Hessian provides a more comprehensive analysis of critical points
in constrained optimisation problems by considering both the objective function and
the constraints. It helps in determining the nature of the critical points and optimising
the objective function subject to the given constraints.
For Practice
(i) What is a bordered Hessian?

.........................................................................................................................................

.............................................................................................................................

.........................................................................................................................................

.........................................................................................................................................

(ii) State the second-order sufficient condition for a minimum in the case of
constrained minimum.

.........................................................................................................................................

.............................................................................................................................

.........................................................................................................................................

.........................................................................................................................................
Self-Instructional
120 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Constrained Optimisation

General Case (n-variable case) NOTES

The generalisation of the Lagrange-multiplier method to n variable can be easily carried
out. The objective function is

z = f(x1, x2, · · · , xn)

subject to
g(x1, x2, · · · , xn) = c.
It follows that the Lagrange function will be
Z = f(x1, x2, · · · , xn)+ [c – g(x1, x2, · · · , xn)],
for which the ûrst-order condition will be given by
Z = c – g(x1, x2, · · · , xn) = 0;
Zi = fi(x1, x2, · · · , xn) –  gi(x1, x2, · · · , xn)= 0 [i = 1, 2, · · · , n].
Define the bordered Hessian determinant |H | by
0 g1 g2  gn
g1 Z11 Z12  Z1n
H  g 2 Z 21 Z 22  Z 2n
    
g n Z n1 Z n 2  Z nn
where in the newly introduced symbols, the horizontal bar above H mean-s
bordered, and
Zij = fij – gij.
Note that by the first-order condition,
f1 f 2 f
     n.
g1 g 2 gn
The bordered principal minors can be defined as:
0 g1 g 2 g3
0 g1 g 2
g Z11 Z12 Z13
H 2  g1 Z11 Z12 , H 3  1
g 2 Z 21 Z 22 Z 23
g 2 Z 21 Z 22 Self-Instructional
g3 Z 31 Z 32 Z33 Material 121

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES with the last one being |Hn| = |H |, where the subscript indicates the order of the
leading principal minor being bordered. For instance, |H2| involves the second leading
principal minor of the (plain) Hessian, bordered with 0, g1, and g2; and similarly for the
others. The conditions for positive and negative deûniteness of d2z are then:
d2z is negative deûnite subject to dg = 0 if

H 2  0, H 3  0, H 4  0,  , (1)n H n  0,

and d2z is positive definite subject to dg = 0 if

H 2  0, H 3  0, H 4  0,  , H n  0.

In the former, all the bordered leading principal minors, starting with |H2|, must
be negative; in the latter, they must alternate in the sign. As previously, a positive
deûnite d2z is sufûcient to establish a stationary value of z as its minimum, whereas a
negative deûnite d2z is sufûcient to establish it as a maximum.
In short,

Conditions for maxima:

1. Z = Z1 = Z2 = .... = Zn = 0 (neccesary condition);

2. H 2  0, H 3  0, H 4  0,  , (1)n H n  0.

Conditions for minima:

1. Z = Z1 = Z2 = .... = Zn = 0 (neccesary condition);

2. H 2  0, H 3  0, H 4  0,  , H n  0.

Example 4.7:
Find the extremum of z = xy subject to x + y = 6.
Solution:
The Lagrange function is given by
Z = xy + (6 – x – y).
Self-Instructional
122 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Constrained Optimisation

The ûrst-order condition is NOTES

Zx = y –  = 0;
Zy = x –  = 0;
Z = 6 – x – y = 0
Thus, optimal values are  = 3, x = 3, y = 3.
(x, y, z) = (3, 3, 9) is a possible extremum solution. Since
Zx = y– and
Zy = x–, then
Zxx = 0,
Zxy = 1, and
Zyy = 0,
gx = gy = 1.
Thus, we find that

This establishes the stationary value of z = 9 as a maximum.

In-Text Questions
8. The second-order condition used to determine if a critical point is a maximum
or minimum is called the:
a) Border Hessian condition b) Budget constraint
c) Isoquant d) Lagrangian multiplier
9. The condition where the last unit of currency spent on each input adds the
same amount to output is known as:
a) Optimisation condition b) Marginal product
c) Substitution method d) Second-order condition
Self-Instructional
Material 123

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES Multiple Constraints

If the objective function has more than one constraint, say m constraint in n variables
i.e.

Optimise Z = f(x1, x2,    , xn)

Sub. to,
gj(x1, x2,    , xn)= bj
where j= 1,2…..m and m<n
The Lagrange function is then deûned by
Z = f(x1, x2,    , xn)+” j[bj – gj(x1, x2,    , xn)]
and the bordered Hessian will appear as

Let f and g1, . . ., gm are twice continuously differentiable functions and let x”
satisfy the necessary condition for the problem.
Let |Hr(x*)| be the bordered Hessian determinant evaluated at x*. Then
r  m1
1. if  1 H r ( x*)  0, r  m  1, , n, then x* is a local maximum point
for the problem
m
2. if  1 H r ( x*)  0, r  m  1, , n, the x* is a local minimum point for the
problem
Note that when m = 1, the above second-order sufficient conditions reduce to
the second-order sufficient conditions in the previous section.

Self-Instructional
124 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Constrained Optimisation

NOTES
4.8 SUMMARY

Constrained optimisation is a mathematical optimisation problem where the objective

function is optimised while satisfying a set of constraints. The constraints can be
inequalities or equalities that restrict feasible solutions. The goal is to find the optimal
solution that maximises or minimises the objective function while satisfying all the given
constraints.
The free extremum represents the maximum or minimum value of the objective
function without any constraints, while the constrained extremum represents the maximum
or minimum value subject to the given constraints. Generally, a constrained maximum
will have a lower value than a free maximum because it is limited by constraints.
To find stationary values in constrained optimisation, a common approach is to
use the Lagrange multiplier method or the substitution method. The Lagrange multiplier
method involves introducing a Lagrange multiplier and forming a new function called
the Lagrangian. The Lagrange multiplier is found by setting the partial derivatives of
the Lagrangian equal to zero, and the solutions to these equations give the points that
maximise or minimise the objective function subject to the constraints.
The substitution method is suitable for problems with two variables and one
constraint. It involves reformulating the problem as a mathematical optimisation problem
and solving it by substituting one variable from the constraint equation into the objective
function. The resulting function is then differentiated to find the critical points, which
are the solutions that satisfy the first-order conditions for a maximum or minimum.
Both methods allow for finding the optimal solution in constrained optimisation
problems and can be applied to the various types of objective functions and constraints.
It is important to check the second-order conditions to ensure that the found solution
corresponds to a maximum or minimum.

4.9 GLOSSARY

 Constrained optimisation: It is a mathematical optimisation problem with Self-Instructional

objective functions and constraints. Material 125

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES  Objective function: It is a function to be maximised or minimised in the

optimisation problem.
 Constrained extremum: It is a maximum or minimum value of the objective
function subject to constraints.
 Free extremum: It is a maximum or minimum value of the objective function
without any constraints.
 Stationary values: These are critical points of the objective function satisfying
the constraints.
 Lagrange multiplier method: It is the method for finding local maxima and
minima of a function subject to constraints.
 Lagrangian/Lagrange function: It is the function obtained by adding the
product of the Lagrange multiplier and the constraint function to the objective
function.
 Optimisation condition: It is a condition where the last unit of currency spent
on each input adds the same amount to the output.
 Lagrangian multiplier: It is a variable introduced in the Lagrange multiplier
method to find constrained maxima and minima.
 Substitution method: It is a method used to solve optimisation problems with
a single constraint by substituting one variable from the constraint into the objective
function.
 Second-order conditions: These are criteria are used to determine if a critical
points is a maximum, minimum, or neither.
 Optimisation problem: It is a problem of finding the optimal solution that
optimises the objective function while satisfying the given constraints.

4.10 ANSWERS TO IN-TEXT QUESTIONS

1. (c) Maximising or minimising a function subject to constraints

2. (b) Objective function
Self-Instructional
126 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Constrained Optimisation

3. (a) Solutions that satisfy all the given constraints NOTES

4. (c) Substitution method
5. (c) Finding local maxima and minima subject to constraints
6. (d) Combining the objective function and constraints
7. (c) Lagrange multiplier
8. (a) Border Hessian condition
9. (a) Optimisation condition

4.11 SELF-ASSESSMENT QUESTIONS

1. Outline the method of substitution in obtaining a solution to a constrained

optimisation problem.
2. Describe the method of a Lagrange multiplier to obtain a solution to a constrained
optimisation problem, outlining the various steps.
3. What is constrained optimisation? What is the role of constraints in constrained
optimisation? What is the objective function in a constrained optimisation
problem?
4. What is the border Hessian condition used in constrained optimisation?
5. What is the optimisation condition in resource allocation problems?
6. Consider a constrained optimisation problem with the objective function f(x) =
3x + 2y and the constraint 2x + y = 10. Find the maximum value of f(x) within
the feasible region.
7. A utility maximisation problem involves two goods, x and y, with utility function
U(x, y) = x2 + y2. The consumer’s budget constraint is given by 2x + 3y = 12.
Find the combination of x and y that maximises the utility.
8. A production process has two inputs, labor (L) and capital (K), with the
production function, Q(L, K) = 3L2 + 2K2. The firm’s budget constraint is
given by 2L + 3K = 12. Find the combination of L and K that maximises the
production output. Self-Instructional
Material 127

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES 9. A firm wants to minimise the cost of production for a given level of output. The
production function is Q = 2L2 + 3K2, and the firm’s budget constraint is 4L +
5K = 30. Find the combination of L and K that minimises the cost of production.
10. A firm wants to maximise profit, given the production function P = 4L2 + 3K2
and the budget constraint 2L + 3K = 12. Find the combination of L and K that
maximises the profit.
11. A consumer wants to maximise utility U = x2 + y2, given a budget of $50 to
spend on two goods, x and y, with prices px = $2 and py = $3. Find the
combination of x and y that maximises the consumer’s utility while staying within
the budget.

Self-Instructional
128 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Convex and Concave Functions

LESSON 5 NOTES

CONVEX AND CONCAVE FUNCTIONS

Sarabjeet Kaur
Assistant Professor,
Department of Economics,
Zakir Husain College (E), University of Delhi

Structure
5.1 Introduction
5.2 Convex Set
5.3 Convex Functions
5.4 Concave Functions
5.5 Test of Concavity and Convexity
5.6 Quasi-Concavity and Quasi-Convexity
5.7 Envelope Theorem
5.8 Summary
5.9 Glossary
5.10 Answers to In-Text Questions
5.11 Self-Assessment Questions
5.12 References

5.1 INTRODUCTION

Convex and concave functions play a crucial role in economic analysis and modelling.
They are used to represent various economic concepts and relationships, including
utility functions, production functions, cost functions, and demand curves. They are
discussed as follows:
1. Utility functions: In economics, individuals’ preferences and satisfaction are
often represented by utility functions. A utility function is typically assumed to be
either concave or quasi-concave to reflect diminishing marginal utility. Concavity
implies that as individuals consume more of a good, the additional satisfaction Self-Instructional
Material 129

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES they gain from each additional unit diminishes. This concept is essential in
understanding consumer behavior, demand analysis, and welfare economics.
2. Production functions: Production functions describe how inputs such as labour
and capital are combined to produce outputs in an economy. Convex production
functions are commonly used to represent increasing returns to scale, where a
proportional increase in inputs leads to a more than proportionate increase in
output. Convexity implies that as inputs are increased, the marginal product of
each input diminishes slowly. This has implications for technological progress,
efficiency analysis, and optimal production decisions.
3. Cost functions: Cost functions measure the costs associated with producing a
given level of output. Convex cost functions are often used to represent
economies of scale, where the average cost declines as output increases. This is
due to spreading fixed costs over a larger output. The convexity of cost functions
is vital in understanding firms’ cost behaviour, pricing decisions, and industry
structure analysis.
4. Demand curves: Demand curves illustrate the relationship between the price
of a product and the quantity demanded by consumers. Typically, demand curves
are downward sloping, reflecting the law of demand. Under standard
assumptions, demand curves are assumed to be convex at the individual consumer
level and global aggregate level. This convexity implies that as the price decreases,
the quantity demanded increases at a decreasing rate.
Convex and concave functions are not only used to represent economic
relationships but are also crucial in optimising economic decision-making. Optimisation
problems, such as profit maximisation or welfare maximisation, often involve maximising
or minimising convex or concave objective functions subject to certain constraints.
The convexity or concavity of these objective functions plays a key role in designing
efficient optimisation algorithms and deriving optimal solutions.
Overall, convex and concave functions provide a powerful framework for
analysing and modelling economic phenomena, enabling economists to make predictions,
and policy recommendations, and evaluate efficiency and welfare implications in various
economic contexts.

Self-Instructional
130 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Convex and Concave Functions

NOTES
5.2 CONVEX SET

A convex set is a set where any line segment connecting two points in the set lies
entirely within the set. In other words, if you take any two points in a convex set and
draw a line segment between them, the entire line segment will also be in the set. In
other words, a set U is a convex set if whenever X and Y are points in U then every
point on the line segment joining X to Y is in U.
a(X,Y) =[(tX + ( 1 – t)Y : t lies between 0 and 1]
Choose any two points, b and d on the line and connect them by a line. The line
connecting the points, b and d also lies on the straight line which we drew in the
beginning. All the points on the line connecting b and d lie on the original line.

Fig. 5.1(a) Fig. 5.1(b)

Look at Fig. 5.1(a), choose any two points on or in the circle and connect them
by a straight line, x1x2. This straight line also lies within the circle. A set is convex if the
line joining any two points of the set lies entirely within the set. The straight line and the
circle which includes the area within the circle, hence, is an example of a convex set.
Let us see a few more examples of a convex set.

Self-Instructional
Fig. 5.2 Material 131

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES Now consider Fig. 5.1(b). In this case, if we connect the two points, x1 and x2 by a
straight line, we find that there are many points on x1x2 which do not lie within the
figure, as shown above. Therefore, it is not a convex set.

Fig. 5.3

In these figures, there is a feature of reentrance (and also a hole). This is a cause
for non-convexity. To qualify as a convex set, the set of points in the figure must
contain no holes, and its boundary points must not be reentered anywhere.
The geometric definition of convexity also applies to points in 3-space as well
as in n-space. A solid cube is a convex set. A hollow cylinder is non-convex.
A concave set is a set where there are at least two points such that the line
segment connecting them does not lie entirely within the set. In other words, there are
at least two points in a concave set such that the line segment between them intersects
the boundary of the set.
Convex and concave functions are used to describe the curvature or shape of a
function’s graph. The distinction between convex and concave functions depends on
the behaviour of the function when two points on its graph are connected by a line
segment.

5.3 CONVEX FUNCTIONS

A function is considered convex if the line segment between any two points on the
Self-Instructional
132 Material
graph of the function lies above or on the graph itself. In other words, if you pick any

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Convex and Concave Functions

two points (x1, y1) and (x2, y2) on the graph of a convex function, the line connecting NOTES
these points should not dip below the graph at any point in between.
Graphically, a convex function has a ‘smiling’ shape, where it curves upward
and does not have any ‘dips’. For a function to be convex, it must satisfy the following
property for all x within its domain and for any t in the interval [0, 1];
f(tx1 + (1 - t)x2)  tf(x1) + (1 - t)f(x2)
In this property, the left side represents the value of the function at the weighted
average of x1 and x2, and the right side represents the weighted average of the function
values at x1 and x2.
Examples of convex functions include linear functions, quadratic functions with
non-negative leading coefficient, and exponential functions with a positive base.

Fig. 5.4

5.4 CONCAVE FUNCTIONS

A function is considered concave if the line segment between any two points on the
graph of the function lies below or on the graph itself. In other words, if you pick any
two points (x1, y1) and (x2, y2) on the graph of a concave function, the line connecting
these points should not rise above the graph at any point in between. Self-Instructional
Material 133

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES Graphically, a concave function has a ‘frowning’ shape, where it curves

downward and does not have any ‘peaks’. For a function to be concave, it must
satisfy the following property for all x within its domain and for any t in the interval [0,
1]:
f(tx1 + (1 - t)x2)  tf(x1) + (1 - t)f(x2)
Similar to the convexity property, the left side represents the value of the function
at the weighted average of x1 and x2, and the right side represents the weighted average
of the function values at x1 and x2.
Examples of concave functions include linear functions, quadratic functions with
a non-positive leading coefficient, and logarithmic functions with a base greater than 1.

Fig. 5.5

The conditions for concavity and convexity, strict and non-strict can be defined for
functions of many variables. We shall discuss the concept of concavity and convexity
for a two-variable function. z = f(x1, x2)

The function f(x,y) is concave (convex) if and only if for any pair of distinct
points, A and B on its graph (a-surface) the line segment lies either on or below (above)
the surface except at points, A and B. Strict concavity requires the line segment AB to
lie below the arc AB. Imagine a dome-shaped surface. The surface of a convex function

Self-Instructional
134 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Convex and Concave Functions

typically is bowl-shaped. For non-strictly concave and convex functions, the line NOTES
segment AB is allowed to lie on the surface itself, some portion of the surface, or even
the entire surface may be flat rather than curved.

Fig. 5.6

In-Text Questions
1. Which of the following sets is considered convex?
a) A set where the line segment between any two points lies above the set
b) A set where the line segment between any two points lies below the set
c) A set with non-linear elements
d) A set with undefined elements
2. A convex function satisfies the property of:
a) Concavity b) Convexity
c) Linearity d) Non-differentiability
3. Concave functions have the property that the line segment between any two
points on the graph lies:
a) Above the graph b) Below the graph
c) On the graph d) Parallel to the graph

Self-Instructional
Material 135

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES
5.5 TEST OF CONCAVITY AND CONVEXITY

Single Variable Case

To test the concavity or convexity of a function in the case of a single variable, you can
follow the following steps:
1. Find the second derivative of the function.
2. Calculate the second derivative of the function. If the function is given as f(x),
the second derivative is denoted as f(x) or d²f/dx².
3. Set up the inequality based on the type of function you are testing:
 For concave functions: f(x)  0
 For convex functions: f(x)  0
4. If the second derivative is zero for some values of x, then the function has an
inflection point at those values of x.
An inflection point is a point on the graph of a function where the concavity
changes. In other words, it is a point where the graph of the function changes from
being concave up to concave down, or vice versa.
Example 5.1:
Let the function be
f(x) = -1/3X3 + 3X2 -5X
The first derivative of the function is
f’(X) = - X2 + 6X -5
and the second derivative of function is
f” (X) = - 2X + 6
i.e., f(x) will be greater than 0 if X is less than 3 and function is convex in this
range and
f”(X) will be less than 0 if X is more than 3 hence function is concave.

Self-Instructional
136 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Convex and Concave Functions

Example 5.2: NOTES

Suppose a single input ‘l’ is used to produce ‘q’ prove that if production q = l1/3, l > 0
then the cost function c(q) is convex, whereas the production function is concave.
Solution:
Production function f (l) = l1/3

1 -2/3
The first derivative: f’(l) = l
3

2 -5/3
The second derivative f(l) = l < 0 if l > 0
9
Therefore, the production function is concave.
The cost function is C(l) =  +l where  is fixed cost and  is marginal cost.
Since q = l1/3, l = q3 substituting in the cost function in term of output
C (q) = + q3
The first derivative of cost function C(q) = 3q2
The second derivative of cost function C(q) = 6q > 0
Therefore, the cost function is convex.

Multi Variable Case

When dealing with functions of two or more variables, it can be challenging to determine
the concavity or convexity of the function using diagrammatic or algebraic methods.
These functions often involve complex algebraic manipulations that make it cumbersome
to apply algebraic formulas. However, a viable approach relies on derivatives if the
function is differentiable. The test of concavity and convexity is a way to determine the
concavity or convexity of a function using its second derivative.
The second derivative of a function is the derivative of its first derivative. In
other words, it is the rate of change of the rate of change of the function. If the second
derivative of a function is positive, then the function is convex. If the second derivative
of a function is negative, then the function is concave.
Self-Instructional
Material 137

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES A differentiable function f(x) = f(x1, x2,..., xn) is concave if and only if for any
given point p = ( x10 , x20, …..xn0 ) and any other point, R = ( x11 , x21, …..xn1 ) in a
convex domain.
n
f  R   f ( p )   fi ( f ) ( R  p )
L 1
In the case of a function of two variables, this can be written as:

given P   x10 , x20  and R   x11 , x12 

f ( R)  f ( p)  f1 ( xi )  x20  x1   f 2  x20  x12  x11 

In the case of a convex function, the inequality will be reversed.
In geometric terms, concavity of a function implies that the tangent plane at the
point, p on the function’s graph is positioned above the graph initially. On the other
hand, in the case of a convex function, the graph of the function lies strictly above all
tangent planes or hyperplanes, except at the point of tangency.
Example 5.3:

Z = x12 + x22

The function is convex if for all X = (x1, x2) and Y = (y1, y2)

f (Y )  f ( X )   fi ( X )( yi  xi )   y12  y22    x12  x22 

 2 x1 ( y1  x1 )  2 x2 ( y2  x2 )
f f
(where x  2 x1 and x  2 x2 ) = 2 x1 y1  2 x12  2 x2 y2  2 x22
1 2

Shifting the right-hand side to left-hand side, we get,

y12  y 2  x12  x22  2 x1 y1  2 x12  2 x2 y2  2 x22  0

 y12  x12  2 x1 y1  y22  x22  2 x2 y2  0

  y1  x1    y2  x2   0
2 2

Self-Instructional
138 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Convex and Concave Functions

The expression in the brackets will remain positive irrespective of the values of NOTES
(x1, x2) and (y1, y2). This proves that the function is convex.

Properties

Let f and g be functions defined over a convex set S in Rn. Then:

a) f and g concave and a  0, b  0  af + bg concave
b) f and g convex and a  0, b  0  af + bg convex
c) f(x) concave and F(u) concave and increasing  U(x) = F (f(x)) concave
d) f(x) convex and F(u) convex and increasing  U(x) = F (f(x)) convex
e) f and g concave  h(x) = min{f(x), g(x)}is concave.
f) f and g convex  h(x) = max{f(x), g(x)}is convex.
Let z = f(x1, x2) be a function defined over an open set P in the plane with
continuous partial derivatives of the first and second orders. Then,

fxx fxy 
a. f is concave  fxx  0, fyy  0, and 0
fyx fyy 

fxx fxy 
b. f is concave  fxx  0, fyy  0, and 0
fyx fyy 

where all inequalities should hold throughout z.

Sufficient Conditions for Global Extreme Points

Suppose we have a function f(x, y) defined on a convex domain P, with continuous

partial derivatives up to the first and second orders. Let (x0, y0) be an interior point
within P where the function becomes stationary (i.e. the partial derivatives are zero at
that point). If, for all other points (x, y) in the domain P, the second-order mixed partial
derivatives of f satisfy a specific condition, then we can make a conclusion about the
nature of the function at the point (x0, y0).
a) If the second-order partial derivatives fxx(x, y) and fyy(x, y) are both less
than or equal to zero, and the determinant fxx(x, y) fyy(x, y) - [fxy(x, y)]2 is
Self-Instructional
greater than or equal to zero, then we can conclude that the point (x0, y0) is a Material 139

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES maximum point for the function f(x, y) within the domain P.
b) On the other hand, if the second-order partial derivatives fxx(x, y) and fyy(x,
y) are both greater than or equal to zero, and the determinant fxx(x, y) fyy(x,
y) - [fxy(x, y)]2 is greater than or equal to zero, then we can infer that the point
(x0, y0) is a minimum point for the function f(x, y) within the domain P.
For any function y = f(x1, x2, x3, ...., xn) = f(X) where X............. which is twice
differentiable with Hession H, the function f is strictly concave on Rn if H is a negative
definite for all X in Rn, that is
Let dx = [dx1 dx2] Then
d2y = dxT H dX d2y = dxT H dx < 0
The function f is strictly convex on Rn if and only if H is positive definite for all x
€ Rn, that is
d2y = dXT H dx > 0
The Hession H is a positive definite if and only if all the leading principal minors
of Matrix H are positive.
 f11 f12 f13 
 
H   f 21 f 22 f 23 
f f32 f33 
 31
For example, if y = f(x1, x2, x3) the leading principal minors are
f11 f12
|H1| = | f1|, H2  |H3| = | H| = | f11 f12 f13 |
f 21 f 22
If |H1|<0, |H2|>0 ; |H3|>0 ; the f(x1, x2, x3) is strictly convex (the Hession is
positive definite.
If |H1|<0, |H2|>0, |H3|<0
Then H is a negative definite, and the f(x1, x2, x3) is strictly concave. We now
revise the condition for concavity and convexity.
Suppose H is the Hession matrix associated with twice differentiable function y
= f(X) = f(x1, x2, x3..., xn) x  Rn then
Self-Instructional
140 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Convex and Concave Functions

H is a positive definite on Rn if and only if its leading principal minors are positive NOTES
|H1|>0,
|H2|>0, .... |Hn| = |H|>0 for X  Rn
This means that d2y>0 and so f is strictly concave.
H is a negative definite on Rn if and only if its leading principal minors alternate
in sign beginning with a negative value for |H1|.
|H1|<0 |H2|>0, ... |Hn| = |H|
This means that d2y<0 and f is strictly concave.
Note: A leading principle minor of order r of Hession matrix is found by suppressing
the last n–r rows and columns.

Rules for concavity and convexity

(–1)k Dk(x) > for k = 1, ..., n and for all x  P.  f is strictly concave in P.
Dk(x) > for k = 1, ..., n and for all x  P.  f is strictly concave in P.

Example 5.4:

For n = 2, strictly concavity is implied by

For n = 2, strictly convexity is implied by

Note: If f (x) = f (x1,…,xn) is defined in an open and convex set then,

Self-Instructional
Material 141

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES Example 5.5:

Suppose that a competitive ûrm produces two products. Let Qi represents the output
level of the i-th product and let the prices of the products are denoted by P1 and P2.
Since the firm is a competitive firm, it takes the prices as given. Accordingly, the firm’s
revenue function will be
T R = P1Q1 + P2Q2
The firm’s cost function is assumed to be
C = 2Q21 + Q1Q2 + 2Q22.
Then, the profit function of this hypothetical firm is given by
 = T R – C = P1Q1 + P2Q2 – 2Q21 – Q1Q2 – 2Q22.
The firm wants to maximise the profit by choosing the levels of Q1 and Q2. For
this purpose, setting

we have
4Q1 + Q2 = P1;
Q1 + 4Q2 = P2,
and thus

Also, the Hessian matrix is

Since |H1| = –4 < 0 and |H2| = 16 – 1 > 0, the Hessian matrix (or d2z) is a
negative definite, and the solution does maximise / concave. In fact, since H is a negative
definite everywhere, the maximum profit found above is actually a unique absolute
Self-Instructional maximum.
142 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Convex and Concave Functions

NOTES
In-Text Questions
4. The second derivative test is used to test the:
a) Continuity of a function
b) Differentiability of a function
c) Convexity or concavity of a function
d) Existence of global extreme points
5. Global extreme points represent the:
a) Maximum values of a function
b) Minimum values of a function
c) Maximum or minimum values of a function over the entire domain
d) Inflection points of a function
6. Sufficient conditions guarantee the existence of:
a) Convex functions
b) Concave functions
c) Inflection points
d) Maximum or minimum points
7. The Hessian matrix is used to determine the properties of strict:
a) Convexity
b) Concavity
c) Differentiability
d) Linearity

5.6 QUASI-CONCAVITY AND QUASI-CONVEXITY

In our previous discussion on concave and convex functions, we observed that if a

function is concave or convex, there is no requirement to examine the second-order
condition in order to determine whether the function attains a maximum or minimum
Self-Instructional
Material 143

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES value. Similarly, when dealing with constrained optimisation problems, it becomes
possible to bypass the second-order condition. In the case of constrained optimisation,
the property of quasi-concavity of a function eliminates the need for the second-order
condition in determining the maximum value. Similarly, for finding the minimum value of
a function, the property of quasi-convexity eliminates the necessity of the second-
order condition.
To begin with, we will delve into the concept of quasi-concavity and quasi-
convexity in the context of functions of a single variable.
1) Let f be a function of x [y = f(x)].
2) x and y are non-negative [x, y  0].
This means that we are in the first quadrant.
3) The domain of the function is a convex set.
4) Choose two distinct points, xi and xj (or x1 and x2) such that xi < xj in the
convex domain of the function.
5) The function f(x) forms an arc between xi and xj such that f (xi) = A and f
(xj) =B

Fig. 5.7 (a) Fig. 5.7 (b)

In Fig. 5.7 (a), the point, B is higher in height than A. In other words, f(xj) >
f(xi). The function is strictly quasi-concave if all other points on arc AB are higher in
height than the point, A.
In Fig. 5.7 (b),
1) The domain of the Ist quadrant (xi, xj  0)
Self-Instructional The domain is a convex set.
144 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Convex and Concave Functions

2) Two distinct points are xi and xj such that xi < xj. NOTES
3) The function f(x) forms an arc CD between xi and xj such that f(xi) = c and
f(xj) =
In Fig. 5.7 (b) f(xj) < f(xi). The function is strictly quasi-concave if all other
points on the arc are lower in height than f(xi).
A function f(x) defined on a convex domain is said to be quasi-convex if, for any
two points, a and b in the domain, the function values at the points lying between a and
b on the line segment connecting them are less than or equal to the maximum of the
function values at a and b. In other words, if f(a) and f(b) are two function values and
 is a scalar between 0 and 1, then f(a + (1-)b)  max(f(a), f(b)). Furthermore, if
the weak inequality “” is replaced by the strict inequality “<”, f is said to be strictly
quasi-convex. Geometrically, this means that the function lies entirely or above any line
segment connecting two points in its domain.
On the other hand, a function f(x) defined on a convex domain is said to be
quasi-concave if, for any two points, a and b in the domain, the function values at the
points lying between a and b on the line segment connecting them are greater than or
equal to the minimum of the function values at a and b. In other words, if f(a) and f(b)
are two function values and  is a scalar between 0 and 1, then f(a + (1-)b) 
min(f(a), f(b)). Furthermore, if the weak inequality “” is replaced by the strict inequality
“>”, f is said to be strictly quasi-concave. Geometrically, this means that the function
lies entirely or below any line segment connecting two points in its domain.

Fig. 5.8 Graphical Representations of Quasi-concavity and Quasi-convexity

Self-Instructional
Material 145

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES (a) The function exhibits strict quasi-concavity.

(b) The function demonstrates strict quasi-convexity.
(c) The function displays quasi-concavity, but it is not strictly quasi-concave.

Example of quasi-convex functions:

They are given as follows:

 The function f(x) = x2 is quasi-convex because its lower contour sets are all
lines.
 The function f(x) = x3 is quasi-concave because its upper contour sets are all
lines.
Some examples of functions that are not quasi-convex or quasi-concave are as follows:
 The function f(x) = |x| is not quasi-convex because its lower contour sets are
not all lines.
 The function f(x) = x2 + x is not quasi-concave because its upper contour sets
are not all lines.

Properties

They are listed as follows:

1. If f(x) is quasi-concave (strictly quasi-concave), then “f(x) is quasi-convex
(strictly quasi-convex).
2. Any (strictly) concave (convex) function is (strictly) quasi-concave (quasi-
convex), but the converse may not be true.
3. If f(x) is linear, then it is quasi-concave as well as quasi-convex.
4. If f is a function of one variable (linear function), then it is quasi-concave as well
as quasi-convex.

Functions of multiple variables

Let a function z = f(x, y) of two variables

The function is defined on R2++ (x, y  0)
Self-Instructional Let u = (x1, y1) and v = (x2, y2) and v > u then f(x, y) is quasi-concave if
146 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Convex and Concave Functions

f(v) > f(x)  f [x + (1 – )v] > f(x) strictly quasi-concave NOTES
f(v)  f(x)  f [x + (1 – )v]  f(x) quasi-concave
f(v) < f(x) f [x + (1 – )v] < f(x) strictly quasi-convex
f(v)  f(x)  f [x + (1 – )v]  f(x) quasi-convex
Example 5.6: For what values of the parameters a, b, c and d is the function ax3 + bx2
+ cx + d quasi-concave? (Use the characteristics of quasi-concave functions of a
single variable. There are several cases to work through.)
A function f of a single variable is quasi-concave if and only if either it is non-
decreasing or it is non-increasing or there is some x* such that f is non-decreasing for
x < x* and non-increasing for x > x*. Consider each case in turn.

f is non-decreasing

The derivative of the function is 3ax2 + 2bx + c. In order for this to be non-negative for
all x, we certainly need c  0 (take x = 0). Now, we can consider three cases separately.
 If a > 0 then the derivative is a convex quadratic, with a minimum at x = –
b/3a. (Take the derivative of the derivative and set it equal to zero.) The minimal
value of the derivative is thus 3a(b/3a)2–2b2/3a + c = –b2/3a + c. For the
derivative to be non-negative for all x, it is necessary and sufficient that this
minimum be  0, which is equivalent to c  b2/3a.
 If a = 0 then the derivative is non-negative for all x if and only if b = 0 and c  0.
 If a < 0 then the derivative is a concave quadratic, and hence is negative for
some values of x, no matter what the values of b and c are.

f is non-increasing

The derivative of the function is 3ax2 + 2bx + c. In order for this to be non-positive for
all x, we certainly need c  0 (take x = 0). Now, as in the previous case, we can
consider three cases separately.
 If a > 0 then the derivative is a convex quadratic, and hence is not non-positive
for all values of x no matter what the values of b and c are.
 If a = 0 then the derivative is non-positive for all x if and only if b = 0 and c  0.
Self-Instructional
Material 147

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES  If a < 0 then the derivative is a concave quadratic, with a maximum at

x = – b/3a. (Take the derivative of the derivative, and set it equal to zero.) The
maximal value of the derivative is thus 3a.(b/3a)2–2b2/3a + c = – b2/3a + c. For
the derivative to be non-positive for all x, it is necessary and sufficient that this
maximum be  0, which is equivalent to c b2/3a.

f increases then decreases

If a > 0 then the derivative of f is a quadratic; since no quadratic has the property that
it is positive up to some point and negative thereafter, the only possibility is that a = 0.
In this case, f itself is a quadratic; it increases and then decreases if and only if b < 0.
In summary, f is quasi-concave if and only if either a > 0 and c  b2/3a, or a <
0 and c  b2/3a, or a = 0 and b  0.

In-Text Questions
8. A function that is quasi-convex is also convex. a. True b. False
9. If a function is quasi-convex, then the upper-level set of any point is a convex
set. a. True b. False
10. A function that is quasi-concave is also concave. a. True b. False
11. If a function is quasi-concave, then the lower-level set of any point is a convex
set. a. True b. False
12. Quasi-convexity and quasi-concavity are properties that apply only to
continuous functions. a. True b. False
13. A function that is both quasi-convex and quasi-concave must be constant.
a. True b. False
14. The sum of two quasi-convex functions is always quasi-convex.
a. True b. False
Example 5.7:
Use the bordered Hessian condition to determine whether the function f(x,y) = ye-x is
quasi-concave for the region in which x  0 and y  0.

Self-Instructional
148 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Convex and Concave Functions

Solution: NOTES
The bordered Hessian of f is
0  ye x e x
 ye x ye x e  x
e x e x 0

We have D1(x,y) = -y2e-2x  0 and D2(x,y) = ye-3x + e-x (ye-2x - ye-2x ) = ye-3x 
0. Both determinants are zero if y = 0, so while the bordered Hessian is not inconsistent
with the functions being quasi-concave, it does not establish that it is in fact quasi-
concave either. However, the test shows that the function is quasi-concave on the
domain in which y is restricted to be positive (rather than only non-negative).
Example 5.8:
Show that a concave function is quasi-concave by using the fact that a function f is
quasi-concave if and only if for all x  S, all y  S and all   [0.1], we have
If f(x)  f(y), then f((1-)x + y)  f(y).
Solution:
Take x and y such that f(x)  f(y). By concavity we have
f((1–)x + y)  (1–)f(x) + f(y) for all  with 0    1, so that
f((1–)x + y)  (1–)f(y) + f(y) = f(y) for all  with 0    1.
Thus f is quasi-concave.

Differentiable function

If we consider a function z = f(x1, x2, ..., xn) that is twice continuously differentiable,
we can assess the quasi-concavity and quasi-convexity of the function by examining
the first and second partial derivatives of the function arranged in a bordered determinant.

0 f1 f2    fn
f1 f11 f12    f1n
B  f2 f11 f 22     f 2n
: : :
Self-Instructional
fn f n1 f n 2  f nn Material 149

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES This bordered determinant bears resemblance to the bordered Hessian

determinant |H| used in the context of constrained maximisation or minimisation. In the
case of the bordered Hessian |H|, the first row (or column) of |H| comprises zeros and
the first derivatives of the constraints. However, in this scenario, the first row (or
column) of the bordered determinant |B| consists of zeros and the first derivatives of
the function f(x1, x2, ..., xn) itself. This is because the notion of quasi-concavity or
quasi-convexity solely relies on the partial derivatives of the function itself. We use B
along with its leading principal minor.
0 f1 f2    fn

0 f1 f2 f1 f11 f12    f1n

0 f1 B2  f1 f11 f12 Bn  B  f 2 f11 f 22    f 2n
B1 
0 f2 : : :
f2 f 21 f 22
fn f n1 fn2    f nn
The domain of the function is non-negative or thant (that is x1, x, x3 .......... xn
 0)
z = f(x1, x2,.......xn) is quasi-concave on the Rn+.
|B1|  0, |B2|  0, |Bn| = |B|  0 if n is odd
= |B|  0 if n is even
The function is strictly quasi-concave on the Rn+
If
|B1| < 0, |B2| > 0,..... |Bn| = |B| < if n is odd
= |B| > if n is even
Quasi-convex.
|B1|  0, |B2|  0, |B3|  0,...., |Bn| = |B|  0
Strictly quasi-convex if
|B1| < 0, |B2| < 0,...... |Bn| = |B| < 0.

Self-Instructional
150 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Convex and Concave Functions

NOTES
5.7 ENVELOPE THEOREM

The envelope theorem is a mathematical result in economics that describes how the
value of an optimal function changes when a parameter is varied. The theorem states
that the change in the value of the function can be estimated by the change in the
function evaluated at the optimal value of the parameter.
To understand this better, let’s consider a simple example. Suppose you have
an optimisation problem of the form:
Maximise f(x;)
subject to g(x;) = 0,
where x is the decision variable,  is the parameter, f(·) is the objective function,
and g(·) is the constraint function.
The envelope theorem states that the derivative of the optimal value of f(x;)
with respect to  can be calculated as follows:
d/d f(x*;) = *(),
where x* is the optimal solution to the problem, and *() is the Lagrange
multiplier associated with the constraint evaluated at .
In other words, to calculate the derivative of the objective function with respect
to the parameter, you only need to find the Lagrange multiplier associated with the
constraint and evaluate it at the given parameter value.
The envelope theorem can be used to analyse the effects of changes in parameters
on optimal decisions. For example, suppose we are interested in the effect of a change
in the price of a good on the consumer’s optimal consumption of that good. We can
use the envelope theorem to approximate the change in the consumer’s optimal
consumption by evaluating the consumer’s indirect utility function at the optimal price.
The envelope theorem is a powerful tool for comparative statics, which studies
how changes in parameters affect economic equilibria. It is used in a wide variety of
areas of economics, including consumer theory, production theory, and game theory.

Self-Instructional
Material 151

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES Keypoints

They are discussed as follows:

 The envelope theorem only considers the direct effects of a change in a parameter.
This means that it ignores any indirect effects that may occur through the optimal
value of decision variables.
 The envelope theorem is only valid under certain conditions. These conditions
typically involve the assumption that the decision variables are continuously
differentiable and that the objective function is well-behaved.
 The envelope theorem can be used to estimate the change in an optimal function
when a parameter is varied. This approximation can be used to analyse the
effects of changes in parameters on economic equilibria.
The envelope theorem is a powerful tool in economics and optimisation because
it simplifies the calculation of derivatives in optimisation problems. It allows researchers
to focus on the relevant parts of the problem and ignore certain complexities, leading
to more efficient analysis and computations.
A few examples of how the envelope theorem can be used are as follows:
 A firm is considering changing the price of its product. The envelope theorem
can be used to estimate how the firm’s profits will change as a result of the
price change.
 A government is considering imposing a tax on a certain good. The envelope
theorem can be used to estimate how the tax will affect the quantity of the
good that is produced and consumed.
 A consumer is considering buying a new car. The envelope theorem can be
used to estimate how the consumer’s utility will change as a result of the
purchase.

5.8 SUMMARY

A function f(x) is considered convex if the line connecting any two points on its graph
Self-Instructional never falls below the curve, while it is concave if the line connecting any two points on
152 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Convex and Concave Functions

the graph never rises above the curve. If a function f(x) is defined over a convex subset NOTES
S in Rn, it is quasi-concave if for any real number a, the set U  {x  S: f(x)  a} is
convex. Similarly, if U  {x  S: f(x)  a} is a convex set for every real number a, then
the function f(x) is quasi-convex.
Assuming a function f(x) is continuous over the interval S and twice differentiable
within its interior, denoted as S0, the following conditions hold:
- f is convex on S0 if and only if f(x)  0 for all x in S0.
- f is concave on S0 if and only if f(x)  0 for all x in S0.
A set is considered convex if it is possible to connect any pair of points within
the set using a line segment that lies entirely within the set. For a function f(x) = f(x1, ...,
xn) defined over a convex set S, the function is concave in S if the inequality f((1-)x0
+ x)  (1-)f(x0) + f(x) holds for all x0, x  S, and   [0, 1]. For a function f(x)
= f(x1, ..., xn) defined over a convex set S, the function is convex in S if the inequality
f((1-)x0 + x)  (1-)f(x0) + f(x) holds for all x0, x  S, and   [0, 1]. A function
f defined over a convex set S in Rn is considered quasi-concave if its upper level sets
Pa = {x  S: f(x)  a} are convex for every value of a, and it is quasi-convex if its
lower level sets Pa = {x  S: f(x)  a} are convex for every value of a.
The envelope theorem describes how the value of an optimal function changes
when a parameter is varied. The theorem states that the change in the value of the
function can be determined by the change in the function evaluated at the optimal value
of the parameter.

5.9 GLOSSARY

 Convex sets: These are sets in which the line segment between any two points
lies above or on the set.
 Convex functions: These functions satisfy the property of convexity, where the
line segment between any two points on the graph lies above or on the graph.
 Concave functions: These functions satisfy the property of concavity, where
the line segment between any two points on the graph lies below or on the
Self-Instructional
graph. Material 153

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES  Inequality properties: These are mathematical properties that describe the
relationships between the values of a function and its derivatives or second
derivatives.
 Second derivative test: It is a method for testing the convexity or concavity
of functions using the sign of the second derivative.
 Global extreme points: These points on a function represent the maximum or
minimum values over the entire domain.
 Sufficient conditions: These conditions, if met, guarantee the existence of
maximum or minimum points.
 Strict concavity and convexity: These are properties of functions where the
line segments between points lie strictly below or strictly above the graph,
respectively.
 Hessian matrix: It is a matrix used to determine the properties of strict concavity
or convexity of functions.
 Curvature: It is the measure of how a function deviates from being linear,
indicating its convex or concave nature.

5.10 ANSWERS TO IN-TEXT QUESTIONS

1. (a) A set where the line segment between any two points lies above the set
2. (b) Convexity
3. (b) Below the graph
4. (c) Convexity or concavity of a function
5. (c) Maximum or minimum values of a function over the entire domain
6. (d) Maximum or minimum points
7. (b) Concavity
8. (b) False
9. (a) True
Self-Instructional
154 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
Convex and Concave Functions

10. (b) False NOTES

11. (a) True
12. (b) False
13. (a) True
14. (b) False

5.11 SELF-ASSESSMENT QUESTIONS

1. How is convexity defined for a function, and what is the criterion for a function
to be considered convex or concave?
2. Explain the concept of quasi-concave and quasi-convex functions. Also, provide
examples to illustrate their definitions.
3. What are the conditions for a function to be convex or concave over a given
interval, assuming it is continuous and twice differentiable?
4. Define a convex set and explain how it is related to the concept of convex
functions.
5. In the context of convexity, what is the significance of the upper-level set and
lower-level set of a function? How do they determine whether a function is
quasi-concave or quasi-convex?
6. Consider a function f(x) = 2x2 - 3x + 1. Determine whether this function is
convex or concave over its entire domain.
7. Let f(x) = 4x - 2 be a function defined over the convex set S = [0, 5]. Find the
upper-level set P2 = {x  S: f(x)  2}.
8. For a function f(x) = x3 - 2x2 + x defined over the interval S = [-1, 2], calculate
the second derivative f(x) and determine if the function is convex or concave
on S.
9. Given a convex set S with two points A(2, 5) and B(4, 1), find the equation of
the line segment that joins these two points.
Self-Instructional
Material 155

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
 Optimisation Methods for Economic Analysis

NOTES 10. Consider a function f(x) = 3x2 - 4x + 1 defined over a convex set S. Determine
if this function is concave or convex based on the inequality f((1-)x0 + x) 
(1-)f(x0) + f(x), where x0 = 2, x = 5, and  = 0.5.

5.12 REFERENCES

 Allen, R.G.D. Mathematical Analysis for Economists, London: Macmillan

and Co. Ltd.
 Sydsaeter, Knut and Peter J. Hammond. Mathematics for Economic Analysis,
Prentice Hall.
 Simon P. Carl and Lawrence Blume. Mathematics for Economists. London:
W .W. Norton & Co.

Self-Instructional
156 Material

© Department of Distance & Continuing Education, Campus of Open Learning,

School of Open Learning, University of Delhi
DEPARTMENT OF DISTANCE AND CONTINUING EDUCATION
UNIVERSITY OF DELHI