ECN 2311

Mathematics for Economics

Mathematics for Economics Department of Economics

 UNIT 4

OPTIMISATION

Mathematics for Economics Department of Economics

 Introduction
• Optimisation means the quest for the best.
• The optimization framework is one of the most important
concepts in economics because of its widespread applicability in
both microeconomics and macroeconomics.
• Economics is the science which studies human behaviour as a
relationship between ends and scarce means which have
alternative uses.
• To an economist, optimisation is the process of finding the best
choice available.
• Optimization is the process of finding the relative maximum or
minimum of a function.

Mathematics for Economics Department of Economics

• For instance, in economics we deal with maximizing utility, profit,
growth rate of Zambia or minimizing pollution and the cost of
producing a given output and so forth.
• For any optimization problem, one needs to know the objective
function which essentially illustrates the relationship between a
dependent variable and independent variables.

Mathematics for Economics Department of Economics

Optimization (2)

• In mathematical terms,
optimization is the process of
finding the relative maximum or
minimum of a function
• To be at the maximum or
minimum, the function must not
be increasing or decreasing
• At this point
o 𝑓′ 𝑥 = 0
o Known as critical point

Mathematics for Economics Department of Economics 5

Optimization (2)
y y

x x

Case 1. Case 2.
y
E

F
x
Case 3.
Mathematics for Economics Department of Economics 6
First Derivative Test
• If the first derivative of a function 𝑓(𝑥) at a critical value 𝑥 = 𝑥𝑜 is
𝑓 ′ 𝑥𝑜 = 0, then the value of the function at 𝑥𝑜 , 𝑓 ′ 𝑥𝑜 , will be
i. A relative maximum if the derivative 𝑓 ′ 𝑥 changes its sign from
positive to negative from the immediate left of the point x0 to its
immediate right.
ii. A relative minimum if 𝑓 ′ 𝑥 changes its sign from negative to
positive from the immediate left of 𝑥𝑜 to its immediate right.
iii. Neither a relative maximum nor a relative minimum if 𝑓 ′ 𝑥 has
the same sign on both the immediate left and the immediate right
of point 𝑥𝑜 .

Mathematics for Economics Department of Economics

In-class exercise
• Find the relative extremum of the average-cost function
𝐴𝐶 = 𝑓 (𝑄) = 𝑄 2 − 5𝑄 + 8

Mathematics for Economics Department of Economics

Interpretation of the Second Derivative
• Since 𝑓 ′ (𝑥) measures the rate of change of the function 𝑓 𝑥 then 𝑓 ′′ (𝑥)
measures the rate of change of 𝑓 ′ (𝑥)
• That is, the second derivative measures the rate of change of the rate of change
of the original function 𝑓(𝑥). That is

𝑓 ′ (𝑥𝑜 ) > 0 increase

ቋ means that the 𝑣𝑎𝑙𝑢𝑒 𝑜𝑓 𝑡ℎ𝑒 𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛 tends to ቊ
𝑓 ′ (𝑥𝑜 ) < 0 Decrease

• whereas, with regard to the second derivative

𝑓 ′′ (𝑥𝑜 ) > 0 increase

ቋ means that the 𝑠𝑙𝑜𝑝𝑒 𝑜𝑓 𝑡ℎ𝑒 𝑐𝑢𝑟𝑣𝑒 tends to ቊ
𝑓 ′′ (𝑥𝑜 ) < 0 Decrease

Mathematics for Economics Department of Economics

Second-derivative test for relative extremum
• If the value of the first derivative of a function 𝑓 𝑥 at 𝑥 = 𝑥𝑜 is
𝑓 ′ 𝑥 = 0, then the value of the function at 𝑥𝑜 ,𝑓(𝑥𝑜 ), will be
i. A relative maximum if 𝑓 ′′ (𝑥𝑜 ) < 0.
ii. A relative minimum if 𝑓 ′′ (𝑥𝑜 ) > 0.

In-class exercise
• Find the relative extrema of the function
𝑦 = 𝑥 3 − 3𝑥 2 + 2

Mathematics for Economics Department of Economics

 Unconstrained optimisation
• Simplest form of optimization without constraints
• Objective function without a single constraint

Mathematics for Economics Department of Economics

 Conditions for unconstrained optimization
single variable
• Both the first and the second derivative tests given so far represent
the necessary and sufficient conditions for optimization of
unconstrained single choice variables.
• Therefore, if 𝑦 = 𝑓(𝑥) then

Condition Type of condition Maximum Minimum

First Order Necessary 𝑓′ 𝑥 = 0 𝑓′ 𝑥 = 0

Second order Necessary 𝑓 ′′ 𝑥 ≤ 0 𝑓 ′′ 𝑥 ≥ 0
Second order Sufficient 𝑓 ′′ 𝑥 < 0 𝑓 ′′ 𝑥 > 0

Mathematics for Economics Department of Economics

Economic Application: Profit Maximization Problem

• The profit maximization problem is:

max 𝜋 = 𝑃 𝑞 𝑞 − 𝑇𝐶 𝑞
𝑞

• where 𝜋 is the optimal level of profit, 𝑃(𝑞) denotes the output price
as a function 𝑞, 𝑞 is output, and 𝑇𝐶(𝑞) is the total cost of production
as a function of 𝑞
• The product 𝑃(𝑞)𝑞 is basically the total revenue [𝑇𝑅 𝑞 ] which is
also a function of 𝑞.
Mathematics for Economics Department of Economics 13
 Necessary condition for profit maximization
• The first order condition for this single variable case solves for the
value of output (q) that maximizes profits:

𝑑𝜋 𝑑𝑇𝑅 𝑑𝑇𝐶
= − =0
𝑑𝑞 𝑑𝑞 𝑑𝑞

• Thus, the necessary condition for profit maximization is that the firm
should choose the output for which 𝑀𝑅 = 𝑀𝐶

𝑑𝑇𝑅 𝑑𝑇𝐶
𝑀𝑅 = = = 𝑀𝐶
𝑑𝑞 𝑑𝑞

Mathematics for Economics Department of Economics 14

 Sufficient condition for profit maximization

• The following second order condition is sufficient for profit

maximization:

𝑑2 𝜋
2
<0
𝑑𝑞

• For any q greater than the optimal level (q*), profit must be
decreasing

Mathematics for Economics Department of Economics 15

In-Class Exercise
1. Suppose a firm faces a constant elasticity demand curve of the form 𝑞 =
256𝑃−2 and has a total cost function of the form 𝑇𝐶 𝑞 = 0.0005𝑞 2 .
a. Set up the profit maximization problem and derive the first order conditions.
b. Is the sufficiency condition satisfied? Explain.
c. What is the firm’s profit maximizing output level?

2. The total cost function for a commodity is given by

𝐶 = 3𝑄3 − 6𝑄2 + 4𝑄
a. Find the value of 𝑄 that optimises the function.
b. Determine if this value of 𝑄 minimises or maximises the profit function.

Mathematics for Economics Department of Economics 16

 Optimization procedure in more than one
choice variable
• Consider the following relationship 𝑧 = 𝑓(𝑥, 𝑦)

• Suppose my interest is to find the choices of 𝑥 and 𝑦 at the maximum

max 𝑧 = 𝑓(𝑥, 𝑦)
𝑥,𝑦

• Suppose my interest is to find the choices of 𝑥 and 𝑦 at the minimum

min 𝑧 = 𝑓(𝑥, 𝑦)
𝑥,𝑦

Mathematics for Economics Department of Economics 17

• In order for the function to be optimal, three conditions need to exist
at the critical point:
a. First order partial derivatives must equal zero
b. Second order direct partial derivatives must be negative for a relative
maximum and positive for a relative minimum
c. Product of the second-order direct partial derivatives must exceed the
product of the cross partial derivatives

Cross partial derivative

simply means that after Second-order direct
taking the first derivative in partial derivative simply
terms of one choice variable, means taking derivatives
the second-order partial with respect to the same
derivative would be in terms choice variable twice
of the other choice variable

Mathematics for Economics Department of Economics 18

• Thus, for a relative maximum, • For a relative minimum,

o𝑓𝑥 = 0 and 𝑓𝑦 = 0 o𝑓𝑥 = 0 and 𝑓𝑦 = 0

o𝑓𝑥𝑥 < 0 and 𝑓𝑦𝑦 < 0 o𝑓𝑥𝑥 > 0 and 𝑓𝑦𝑦 > 0

o𝑓𝑥𝑥 ∙ 𝑓𝑦𝑦 > (𝑓𝑥𝑦 )2 right side o𝑓𝑥𝑥 ∙ 𝑓𝑦𝑦 > (𝑓𝑥𝑦 )2 right side
of this inequality is true of this inequality is true
because, by Young’s theorem, because, by Young’s theorem,
𝑓𝑥𝑦 = 𝑓𝑦𝑥 𝑓𝑥𝑦 = 𝑓𝑦𝑥

Mathematics for Economics Department of Economics 19

 Inflection point cannot be a optimal
• Inflection point is
when:
o 𝑓𝑥 = 0, 𝑓𝑦 = 0
o 𝑓𝑥𝑥 ∙ 𝑓𝑦𝑦 <
(𝑓𝑥𝑦 )2
o 𝑓𝑥𝑥 and 𝑓𝑦𝑦
have the same
signs

Mathematics for Economics Department of Economics 20

 Saddle point cannot be optimal

Saddle point is when:

• 𝑓𝑥 = 0, 𝑓𝑦 = 0
• 𝑓𝑥𝑥 ∙ 𝑓𝑦𝑦 <
(𝑓𝑥𝑦 )2
• 𝑓𝑥𝑥 and 𝑓𝑦𝑦
have different
signs

Mathematics for Economics Department of Economics 21

 In-class Exercise

For the following function, find the critical points and determine
whether the function at each critical point is maximized, minimized, is
at an inflection point, or is at a saddle point.

𝑧 = 2𝑦 3 − 𝑥 3 + 147𝑥 − 54𝑦 + 12

Mathematics for Economics Department of Economics 22

 Constrained Optimization
• Constrained optimization includes optimization problems that
maximize or minimize a function subject to constraints

Mathematics for Economics Department of Economics

 Constrained Maximization
• Step 1: Set the constraint equal to zero
• For a constrained maximization
𝑘 − 𝑔 𝑥, 𝑦 = 0
problem, the general set up is

• Step 2: Multiply the rearranged constraint

max 𝑓(𝑥, 𝑦) by 𝜆 (the Lagrange multiplier)
𝑥,𝑦
𝜆 𝑘 − 𝑔 𝑥, 𝑦 = 0
Subject to: 𝑘 = 𝑔(𝑥, 𝑦)

• Step 3: Add the product from step 2 to the

• New function can be formed by objective function to get Lagrangian
combining the objective function function
and the constraint ℒ 𝑥, 𝑦, 𝜆 = 𝑓(𝑥, 𝑦) + 𝜆 𝑘 − 𝑔 𝑥, 𝑦

Mathematics for Economics Department of Economics 24

 Constrained Minimization
• Step 1: Set the constraint equal to zero
• For a constrained minimization
𝑘 − 𝑔 𝑥, 𝑦 = 0
problem, the general set up is

• Step 2: Multiply the rearranged constraint

min 𝑓(𝑥, 𝑦) by 𝜆 (the Lagrange multiplier)
𝑥,𝑦
𝜆 𝑘 − 𝑔 𝑥, 𝑦 = 0
Subject to: 𝑘 = 𝑔(𝑥, 𝑦)

• Step 3: Add the product from step 2 to the

• New function can be formed by objective function to get Lagrangian
combining the objective function function
and the constraint ℒ 𝑥, 𝑦, 𝜆 = 𝑓(𝑥, 𝑦) + 𝜆 𝑘 − 𝑔 𝑥, 𝑦

Mathematics for Economics Department of Economics 25

 Economic Applications of Constrained
Optimization
• We shall look at:

1. Utility Maximization

2. Cost Minimization

Mathematics for Economics Department of Economics

 Utility Maximization Problem: Definition

• The Choice Problem: Choose 𝑥 and 𝑦 so as to maximize utility subject

to the budget constraint:
max 𝑢(𝑥, 𝑦)
𝑥,𝑦
Subject to: 𝐼 = 𝑃𝑥 𝑥 + 𝑃𝑦 𝑦

• Utility maximization is the way we model a consumer’s choice of

goods.

• Utility maximization leads to the consumer’s demands for goods –

Marshallian demands

Mathematics for Economics Department of Economics 27

 Lagrangian Equation and FOCs

• Set up the maximization in terms • The first order conditions are:

of the Lagrangian equation:
𝛿𝐿 𝛿𝑢
= − 𝜆𝑃𝑥 = 0 (1)
𝛿𝑥 𝛿𝑥
ℒ 𝑥, 𝑦, 𝜆 = 𝑢(𝑥, 𝑦) + 𝜆 𝐼 − 𝑃𝑥 𝑥 − 𝑃𝑦 𝑦
𝛿𝐿 𝛿𝑢
= − 𝜆𝑃𝑦 = 0 (2)
𝛿𝑦 𝛿𝑦

𝛿𝐿
= 𝐼 − 𝑃𝑥 𝑥 − 𝑃𝑦 𝑦 = 0 (3)
𝛿𝜆

Mathematics for Economics Department of Economics 28

 𝑷𝒙
Solve for Choice Condition: 𝑴𝑹𝑺 =
𝑷𝒚

ℒ 𝑥, 𝑦, 𝜆 = 𝑢(𝑥, 𝑦) + 𝜆 𝐼 − 𝑃𝑥 𝑥 − 𝑃𝑦 𝑦

• The first order conditions are: 𝛿𝑢

𝛿𝐿 𝛿𝑢 𝛿𝑢 𝛿𝑥
= − 𝜆𝑃𝑥 = 0 → = 𝜆𝑃𝑥 → 𝜆 =
𝛿𝑥 𝛿𝑥 𝛿𝑥 𝑃𝑥

𝛿𝐿 𝛿𝑢
= − 𝜆𝑃𝑦 = 0
𝛿𝑦 𝛿𝑦
𝛿𝐿
= 𝐼 − 𝑃𝑥 𝑥 − 𝑃𝑦 𝑦 = 0
𝛿𝜆

Mathematics for Economics Department of Economics 29

 𝑷𝒙
Solve for Choice Condition: 𝑴𝑹𝑺 =
𝑷𝒚
ℒ 𝑥, 𝑦, 𝜆 = 𝑢(𝑥, 𝑦) + 𝜆 𝐼 − 𝑃𝑥 𝑥 − 𝑃𝑦 𝑦

• The first order conditions are:

𝛿𝑢
𝛿𝐿 𝛿𝑢 𝛿𝑢 𝛿𝑥
= − 𝜆𝑃𝑥 = 0 → = 𝜆𝑃𝑥 → 𝜆 =
𝛿𝑥 𝛿𝑥 𝛿𝑥 𝑃𝑥
𝛿𝑢
𝛿𝐿 𝛿𝑢 𝛿𝑢 𝛿𝑦
= − 𝜆𝑃𝑦 = 0 → = 𝜆𝑃𝑦 → 𝜆 =
𝛿𝑦 𝛿𝑦 𝛿𝑦 𝑃𝑦

𝛿𝐿
= 𝐼 − 𝑃𝑥 𝑥 − 𝑃𝑦 𝑦 = 0
𝛿𝜆

Mathematics for Economics Department of Economics 30

 𝑷𝒙
Solve for Choice Condition: 𝑴𝑹𝑺 =
𝑷𝒚
ℒ 𝑥, 𝑦, 𝜆 = 𝑢(𝑥, 𝑦) + 𝜆 𝐼 − 𝑃𝑥 𝑥 − 𝑃𝑦 𝑦

• The first order conditions are:

𝛿𝑢
𝛿𝐿 𝛿𝑢 𝛿𝑢 𝛿𝑥 𝛿𝑢 𝛿𝑢
= − 𝜆𝑃𝑥 = 0 → = 𝜆𝑃𝑥 → 𝜆 = 𝛿𝑥 = 𝛿𝑦
𝛿𝑥 𝛿𝑥 𝛿𝑥 𝑃𝑥
𝑃𝑥 𝑃𝑦
𝛿𝑢
𝛿𝐿 𝛿𝑢 𝛿𝑢 𝛿𝑦
= − 𝜆𝑃𝑦 = 0 → = 𝜆𝑃𝑦 → 𝜆 =
𝛿𝑦 𝛿𝑦 𝛿𝑦 𝑃𝑦
𝛿𝐿
= 𝐼 − 𝑃𝑥 𝑥 − 𝑃𝑦 𝑦 = 0
𝛿𝜆
Mathematics for Economics Department of Economics 31
 𝑷𝒙
Solve for Choice Condition: 𝑴𝑹𝑺 =
𝑷𝒚

ℒ 𝑥, 𝑦, 𝜆 = 𝑢(𝑥, 𝑦) + 𝜆 𝐼 − 𝑃𝑥 𝑥 − 𝑃𝑦 𝑦

• The first order conditions are:

𝛿𝑢
𝛿𝐿 𝛿𝑢 𝛿𝑢 𝛿𝑥 𝛿𝑢 𝛿𝑢
= − 𝜆𝑃𝑥 = 0 → = 𝜆𝑃𝑥 → 𝜆 = 𝛿𝑥 = 𝛿𝑦
𝛿𝑥 𝛿𝑥 𝛿𝑥 𝑃𝑥
𝑃𝑥 𝑃𝑦
𝛿𝑢
𝛿𝐿 𝛿𝑢 𝛿𝑢 𝛿𝑦
= − 𝜆𝑃𝑦 = 0 → = 𝜆𝑃𝑦 → 𝜆 =
𝛿𝑦 𝛿𝑦 𝛿𝑦 𝑃𝑦
𝛿𝐿 𝛿𝑢
𝑃𝑥
= 𝐼 − 𝑃𝑥 𝑥 − 𝑃𝑦 𝑦 = 0 𝑀𝑅𝑆 = 𝛿𝑥 =
𝛿𝜆 𝛿𝑢 𝑃𝑦
𝛿𝑦
Mathematics for Economics Department of Economics 32
A Consumer’s Utility Maximizing
Choice of Goods x and y
The subjective relative valuation of x and y equals the market valuation

y
u
MRS = x = x
p
u p y
I/py y

y* . y=
I p
− xx
py py

u = u(x,y)

Mathematicsx* I/px
for Economics x Department of Economics 33
Solution of utility maximization

• Marshallian demand for 𝑥 • Lagrange multiplier (𝜆), which is

(quantity of good 𝑥 which is a the amount by which a
function of prices and income) consumer’s utility could be
𝑥 ∗ = 𝑑𝑥 (𝑃𝑥 , 𝑃𝑦 , 𝐼) increased if income is increased
by one unit
• Marshallian demand for y
(quantity of good y which is • Indirect utility function
function of prices and income) 𝑣(𝑃𝑥 , 𝑃𝑦 , 𝐼)
𝑦 ∗ = 𝑑𝑦 (𝑃𝑥 , 𝑃𝑦 , 𝐼)

Mathematics for Economics Department of Economics 34

In-Class Exercise

Consider a consumer with the following Cobb-Douglas utility function

𝑥𝑦, facing the following prices, 𝑃𝑥 and 𝑃𝑦 and has income 𝑚.
a) Set up the consumer’s utility maximization problem and derive the FOCs.
b) Derive the Marshallian demand functions for 𝑥 and 𝑦

Mathematics for Economics Department of Economics 35

 Cost Minimization Problem: Definition

• The Choice Problem: Choose 𝐾 and 𝐿 so as to minimize firm costs

subject to the production function constraint:

min 𝑇𝐶 = 𝑤𝐿 + 𝑟𝐾
𝐾,𝐿
𝑠. 𝑡. 𝑞0 = 𝑓(𝐾, 𝐿)

• Cost minimization is the way we model a firm’s choice of input levels

required in the production process

Mathematics for Economics Department of Economics 36

 Lagrangian Equation and FOCs

• Set up the maximization in terms • The first order conditions are:

of the Lagrangian equation:
𝛿ℒ 𝛿𝑓
ℒ = 𝑤𝐿 + 𝑟𝐾 + 𝜆[𝑞0 − 𝑓 𝐾, 𝐿 ] =𝑤−𝜆 =0 (1)
𝛿𝐿 𝛿𝐿
𝛿ℒ 𝛿𝑓
=𝑟−𝜆 =0 (2)
𝛿𝐾 𝛿𝐾
𝛿ℒ
= 𝑞0 − 𝑓 𝐾, 𝐿 = 0 (3)
𝛿𝜆

Mathematics for Economics Department of Economics 37

 𝒘
Solve for Choice Condition: 𝑴𝑹𝑻𝑺 =
𝒓

• Simplify FOC 1 to get the value of • Equate the two fractions since
lambda they are both equal to 𝜆
𝛿𝑓 𝑤
𝑤=𝜆 → 𝜆=
𝛿𝐿 𝛿𝑓
𝛿𝐿 𝛿𝑓
𝑤 𝑟 𝛿𝐿 𝑤 𝑀𝑃𝐿 𝑤
= → = → =
𝛿𝑓 𝛿𝑓 𝛿𝑓 𝑟 𝑀𝑃𝐾 𝑟
• Simplify FOC 2 to get the value of 𝛿𝐿 𝛿𝐾 𝛿𝐾
lambda
𝛿𝑓 𝑟
𝑟=𝜆 → 𝜆=
𝛿𝐾 𝛿𝑓
𝛿𝐾

Mathematics for Economics Department of Economics 38

Cost minimizing input choices

The Choice Condition:

¶f Market
w valuation
K RTS = ¶L =
¶f r
¶K

TC/r firm’s
valuation

K* .
q

MathematicsL* TC/w
for Economics L Department of Economics 39
Solution of cost minimization

• Input demand for 𝐿 (quantity of the • Lagrange multiplier (𝜆), which is

labor input 𝐿 which is a function of interpreted as the amount by which a
input prices and the level of output) firm’s total cost could be changed if
denoted as level of output is changed by one unit
𝐿∗ = 𝐿 𝑤, 𝑟, 𝑞
• Total cost function denoted as 𝑇𝐶 ∗ =
• Input demand for 𝐾 (quantity of the 𝑇𝐶(𝑤, 𝑟, 𝑞)
capital input 𝐾 which is a function of • From the total cost function, the
input prices and the level of output) marginal and average cost
denoted as functions can easily be derived
𝐾 ∗ = 𝐾(𝑤, 𝑟, 𝑞)

Mathematics for Economics Department of Economics 40

In-Class Exercise

Consider the following Cobb-Douglas production function 𝑞 = 𝐾𝐿 for

a firm. The firm is faced with market wage rates of w and rental rates of
r. If the firm behaves optimally (minimizes costs):
a) What are the optimal levels of 𝐾 and 𝐿 required by the firm?
b) Derive the total cost function, average and marginal cost function

Mathematics for Economics Department of Economics 41

 End of Lecture 4

Mathematics for Economics Department of Economics