2

UTILITY MAXIMIZATION
AND CHOICE
Topics
• Lagrange multipliers method. First-
order conditions. Second-order
conditions. Interior and corner
solutions. Demand functions.
• Duality. Indirect utility function.
Expenditure minimization.
Properties of expenditure functions.
Reading
[В, гл 5],
[ЧФ, гл 1, 1.5; гл 7, 7.1-7.3],
[К, гл 4, 4.5-4.6],
[ДР, гл 1, 1.3-1.5]
Consumer Choice
• In first lecture we described what a
consumer can afford (budget set)
• Then you had to focus on the concept of
how the consumer determines what is best
(preferences and utility function)
• Now we are able to undertake a detailed
study of the implications of this simple
model of consumer behavior, or consumer
choice
Optimization Principle
• To maximize utility, given a fixed amount of
income to spend, an individual will buy the
goods (two goods):
 that exhaust his total income
 for which the MRS (the rate at which an
individual is willing to trade one good for
another while remaining equally well off) is
equal to the ratio of the prices of the two goods
(the rate at which the two goods can be
substituted for each other without changing
the total amount of money spent)
– in short: the MRS is equal to the ratio of the prices
The Budget Constraint
Assume that an individual has m dollars to
allocate between good 1 and good 2
𝑝1𝑥1 + 𝑝2𝑥2 ≤ 𝑚

The individual can afford

𝑥2 If all income is spent to choose only combinations
on good 2, this is the amount of two goods in the shaded
of good 2 that can be purchased triangle
𝑚/𝑝2

If all income is spent

on good 1, this is the amount
of good 1 that can be purchased

𝑚/𝑝1 𝑥1
First-Order Conditions for a Maximum
We can add the individual’s utility map to show the utility-
maximization process
As we move along the budget line (simply start at the left-
hand corner and move to the right) we note that we are
moving to higher and higher indifference curves.

𝑥2 The individual can do better than point A

by reallocating his budget
A
C The individual cannot have point C
B because income is not large enough
𝑈3
Point B is the point of utility
𝑈2
𝑈1 maximization

𝑥1
First-Order Conditions for a Maximum

Utility is maximized where the indifference curve is

tangent to the budget constraint (where we get to the
highest indifference curve)
𝑝1
slope of budget constraint = −
𝑥2 𝑝2
𝑑𝑥2
slope of indifference curve = −
𝑑𝑥1 𝑈=𝑐𝑜𝑛𝑠𝑡
B
𝑝1 𝑑𝑥2
= = 𝑀𝑅𝑆
𝑈2 𝑝2 𝑑𝑥1 𝑈=𝑐𝑜𝑛𝑠𝑡

𝑥1
Second-Order Conditions for a Maximum

The tangency rule is only necessary but not sufficient

unless we assume that MRS is diminishing
• if MRS is diminishing, then indifference curves are strictly
convex
If MRS is not diminishing, then we must check second-
order conditions to ensure that we are at a maximum
Second-Order Conditions for a Maximum

The tangency rule is only a necessary condition

• we need MRS to be diminishing

𝑥2
There is a tangency at point A,
but the individual can reach a higher
level of utility at point B
B

A
𝑈2
𝑈1

𝑥1
Corner Solutions
In some situations, individuals’ preferences may be
such that they can maximize utility by choosing to
consume only one of the goods

𝑥2
At point A, the indifference curve
𝑈1 𝑈 𝑈3
2 is not tangent to the budget constraint

Utility is maximized at point A

A 𝑥1
 Methods for Solving the Optimization
Problem
The individual’s objective is to maximize his
utility subject to (s.t.) his budget constraint:
𝑈 𝑥1 , 𝑥2 = 𝑥1 𝑥2 → max
𝑥1 ,𝑥2 ≥0
s.t. 𝑝1 𝑥1 + 𝑝2 𝑥2 = 𝑚
 Methods for Solving the Optimization
Problem
Method 1: Converting the Constrained Optimization
Problem into an Unconstrained Optimization Problem
• first, we can substitute the budget constraint into the
utility function. Using algebra, we can rewrite the budget
constraint as
𝑝1 𝑚 𝑚 − 𝑝1𝑥1
𝑥2 = − 𝑥1 + =
𝑝2 𝑝2 𝑝2
• if we substitute this expression for x2 in the utility function
we can rewrite individual’s problem as
𝑚 − 𝑝1𝑥1
max 𝑥1 ( )
𝑥1 ≥0 𝑝2
Unconstrained Optimization
Converting the Constrained Optimization
Problem into an Unconstrained Optimization
Problem
𝑈 𝑥1 , 𝑥2 = 𝑥1 𝑥2 → max
𝑥1 ,𝑥2 ≥0 → max 𝑓(𝑥1 )
𝑝1 𝑥1 + 𝑝2 𝑥2 = 𝑚 𝑥1 ≥0
𝑚 − 𝑝1𝑥1
→ max 𝑥1 ( )
𝑥1 ≥0 𝑝2
Unconstrained Optimization
This problem is an unconstrained problem, so we
can use standard maximization techniques to solve
it.
The first-order condition (FOC) is obtained by
setting the derivative of the utility function with
respect to the only remaining control variable x1
equal to zero:
𝑑𝑈 𝑚 − 2𝑝1𝑥1
= =0
𝑑𝑥1 𝑝2
𝑚 𝑚
𝑥1∗ = , 𝑥2∗ =
2𝑝1 2𝑥2
Unconstrained Optimization
To be sure that we have a maximum, we
need to check that the second-order
condition (SOC) holds.
This condition holds if the utility function is
quasiconcave, which implies that the
indifference curves are convex to the origin:
𝑓(𝑥1 )
the MRS is diminishing as we move down
and to the right along the curve

𝑚 𝑚
𝑥1∗ = 𝑥1 = 𝑥1
2𝑝1 𝑝1
Unconstrained Optimization
In general form:
max 𝑈 𝑥1 , 𝑥2
𝑥1 ,𝑥2 ≥0

s.t. 𝑝1 𝑥1 + 𝑝2 𝑥2 = 𝑚
𝑝1 𝑚 𝑚 − 𝑝1𝑥1
𝑥2 = − 𝑥1 + =
𝑝2 𝑝2 𝑝2
𝑚 − 𝑝1𝑥1
max 𝑈 𝑥1 ,
𝑥1 ≥0 𝑝2
Unconstrained Optimization
In general form:

𝑑𝑈 𝜕𝑈 𝜕𝑈 𝑑𝑥2 𝜕𝑈 𝜕𝑈 𝑝1
= + = − =
𝑑𝑥1 𝜕𝑥1 𝜕𝑥2 𝑑𝑥1 𝜕𝑥1 𝜕𝑥2 𝑝2
𝑝1
= 𝑀𝑈1 − 𝑀𝑈2 = 0
𝑝2
To be sure that we have a maximum, we need to check that
the SOC holds. This condition holds if the utility function is
quasiconcave, which implies that the indifference curves
are convex to the origin: the MRS is diminishing as we move
down and to the right along the curve.
Unconstrained Optimization
By rearranging, we get the same condition for an
optimum
𝑀𝑈1 𝑝1
=
𝑀𝑈2 𝑝2
or
𝑝1
𝑀𝑅𝑆 =
𝑝2

If we combine the 𝑀𝑅𝑆 = 𝑝1 /𝑝2 condition with the

budget constraint, we have two equations in two
unknowns, 𝑥1 and 𝑥2 , so we can solve…
Methods for Solving the Optimization
Problem
Method 2: Using 𝑀𝑅𝑆 = 𝑝1 /𝑝2 to solve the
constrained optimization problem
Tangency rule: the optimal bundle is the
point (𝑥1∗ , 𝑥2∗ ) where an indifference curve is
tangent to the budget line, i.e. the slope of
the indifference curve (MRS) is equal to the
slope of the budget line (𝑝1 /𝑝2 )
Tangency Rule
Utility is maximized where the indifference curve is
tangent to the budget constraint (where we get to the
highest indifference curve)

𝑥2

𝑝1
𝑀𝑅𝑆 =
𝑝2
𝑥2∗

𝑈2

𝑥1∗ 𝑥1
 Methods for Solving the Optimization
Problem
Method 3: The Lagrange Method for Solving the
Constrained Optimization Problem
• The individual’s objective is to maximize utility
subject to the budget constraint
max 𝑈 𝑥1 , 𝑥2
𝑥1 ,𝑥2 ≥0

s.t. 𝑝1 𝑥1 + 𝑝2 𝑥2 = 𝑚
• Set up the Lagrangian (Lagrangian function):
𝓛 = 𝑈 𝑥1 , 𝑥2 + (𝑚 − 𝑝1 𝑥1 − 𝑝2 𝑥2 )
 Lagrange Method

First-order conditions for an interior

maximum:
𝜕𝓛 𝜕𝑈
= − 𝜆𝑝1 = 0
𝜕𝑥1 𝜕𝑥1
𝜕𝓛 𝜕𝑈
= − 𝜆𝑝2 = 0
𝜕𝑥2 𝜕𝑥2
𝜕𝓛
= 𝑚 − 𝑝1 𝑥1 − 𝑝2 𝑥2 = 0
𝜕
Lagrange Method
For any two goods,
𝜕𝑈/𝜕𝑥1 𝜕𝑈/𝜕𝑥2
= =
𝑝1 𝑝2

This implies that at the optimal allocation

of income
𝑝1
𝑀𝑅𝑆 =
𝑝2
Interpreting the Lagrangian
Multiplier
𝜕𝑈/𝜕𝑥1 𝜕𝑈/𝜕𝑥2
= =
𝑝1 𝑝2
𝑀𝑈1 𝑀𝑈2
= =
𝑝1 𝑝2
 is the marginal utility of an extra dollar of
consumption expenditure
– the marginal utility of income
Interpreting the Lagrangian
Multiplier
At the margin, the price of a good
represents the consumer’s evaluation of the
utility of the last unit consumed
– how much the consumer is willing to pay for
the last unit
𝑀𝑈𝑖
𝑝𝑖 =

Corner Solutions
When corner solutions are involved, the
first-order conditions must be modified:
𝓛/𝑥𝑖 = 𝑈/𝑥𝑖 − 𝑝𝑖  0 (𝑖 = 1,2)
If 𝓛/𝑥𝑖 = 𝑈/𝑥𝑖 − 𝑝𝑖 < 0, then 𝑥𝑖 = 0
This means that
𝑈/𝑥𝑖 𝑀𝑈𝑖
𝑝𝑖 > =
 
– any good whose price exceeds its marginal
value to the consumer will not be purchased
 Corner Solutions: Perfect Substitutes
Here the tangency rule is not necessary condition
for a point to be optimal!
𝑥2 𝑥2

𝑚
𝑥2∗ =
𝑝2

𝑥2∗ = 0
𝑥1∗ = 0 𝑥1 𝑚 𝑥1
𝑥1∗ =
𝑝1

Here the tangency rule doesn’t work because the consumer’s

preferences are not strictly convex.
Corner Solutions: Perfect Substitutes
Problem:
max 𝑈 𝑥1 , 𝑥2 = 𝛼𝑥1 + 𝛽𝑥2
𝑥1 ,𝑥2 ≥0

s.t. 𝑝1 𝑥1 + 𝑝2 𝑥2 = 𝑚
Then we convert it into an unconstrained
optimization problem:
𝑚 𝑝1
max 𝛼𝑥1 + 𝛽 − 𝑥1 =
𝑥1 ≥0 𝑝2 𝑝2
𝑝1 𝑚
= (𝛼 − 𝛽 𝑥1) + 𝛽
𝑝2 𝑝2
Corner Solutions: Perfect Substitutes
𝑝1 𝑚 𝛼 𝑝1 𝑚
𝑈 𝑥1 = (𝛼 − 𝛽 𝑥1) + 𝛽 = ( − )𝛽𝑥1 + 𝛽
𝑝2 𝑝2 𝛽 𝑝2 𝑝2

𝛼 𝑝1 𝑚
𝑈 𝑥1 = ( − )𝛽𝑥1 + 𝛽
𝛽 𝑝2 𝑝2
𝑥1 ∈ 0; 𝑚/𝑝1
Corner Solutions: Perfect Substitutes

𝛼 𝑝1 𝑚
𝑈 𝑥1 = ( − )𝛽𝑥1 + 𝛽
𝛽 𝑝2 𝑝2
𝑥1 ∈ 0; 𝑚/𝑝1

This utility function is linear and depends on

𝑥1 (only!).
Corner Solutions: Perfect Substitutes

So:
𝛼 𝑝1
a) if > then 𝑈 𝑥1 is increasing function and
𝛽 𝑝2
𝑚
maximized when 𝑥1∗ = 0; 𝑥2∗ =
𝑝2
𝛼 𝑝1
b) if < then 𝑈 𝑥1 is decreasing function and
𝛽 𝑝2
𝑚
maximized when 𝑥1∗ = ; 𝑥2∗ = 0
𝑝1
𝛼 𝑝1
c) if = then 𝑈 doesn’t depend on 𝑥1
𝛽 𝑝2
𝑥1∗ ∈ 0; 𝑚/𝑝1 ; 𝑥2∗ ∈ 0; 𝑚/𝑝2
Corner Solutions: Perfect Substitutes

Finally:
𝛼 𝑝1
𝑚 𝑝1 , if >
𝛽 𝑝2
𝛼 𝑝1
𝑥1∗ = ∈ 0; 𝑚 𝑝1 , if 𝛽
=
𝑝2
𝛼 𝑝1
0, if <
𝛽 𝑝2
 𝑥2 𝑥2

𝑚 𝛼 𝑝1
𝑥2∗ = 𝑥1∗ = 0 if <
𝑝2 𝛽 𝑝2

𝑚 𝛼 𝑝1
𝑥1∗ = if >
𝑝1 𝛽 𝑝2

𝑥2∗ = 0
𝑥1∗ = 0 𝑥1 𝑚 𝑥1
𝑥2 𝑥1∗ =
𝑝1

𝑚
𝑥2 = 𝛼 𝑝1
𝑝2 𝑥1∗ ∈ 0; 𝑚/𝑝1 if =
𝛽 𝑝2

𝑚 𝑥1
𝑥1 =
𝑝1
Corner Solutions: Perfect Complements

Here the tangency rule is not necessary condition

for a point to be optimal!
𝑥2

𝑥2∗

𝑥1∗ 𝑥1

Here the tangency rule doesn’t work because the consumer’s preferences
are not strictly convex.
 Corner Solutions: Perfect Complements
Problem:
max 𝑈 𝑥1 , 𝑥2 = min{𝛼𝑥1 , 𝛽𝑥2 }
𝑥1 ,𝑥2 ≥0

s.t. 𝑝1 𝑥1 + 𝑝2 𝑥2 = 𝑚
Then we convert it into an unconstrained
optimization problem:
𝑚 𝑝1
max 𝑈 𝑥1 = min{𝛼𝑥1 , 𝛽( − 𝑥1)}
𝑥1 ≥0 𝑝2 𝑝2
Corner Solutions: Perfect Complements

In curly brackets we have linear

relationships:
𝑈 𝑥1 = 𝛼𝑥1
and
𝑚 𝑝1
𝑈 𝑥1 = 𝛽 − 𝛽 𝑥1
𝑝2 𝑝2
Corner Solutions: Perfect Complements
𝑈 𝑥1
𝑚 𝑝1
𝑈 𝑥1 = 𝛽 − 𝛽 𝑥1
𝑝2 𝑝2

𝑈 𝑥1 = 𝛼𝑥1

𝑥1
𝑥1∗ 𝑥1
Corner Solutions: Perfect Complements
In order to determine the minimum value of
the function U, we compare the two values for
each of the lines.
In the interval 𝑥1 ∈ [0, 𝑥1∗ )
𝑚 𝑝1
𝛼𝑥1 < 𝛽 − 𝛽 𝑥1 ⇒ 𝑈 𝑥1 = 𝛼𝑥1
𝑝2 𝑝2

In the interval 𝑥1 ∈ (𝑥1∗ , 𝑥1 )

𝑚 𝑝1 𝑚 𝑝1
𝛼𝑥1 > 𝛽 − 𝛽 𝑥1 ⇒ 𝑈 𝑥1 = 𝛽 − 𝛽 𝑥1
𝑝2 𝑝2 𝑝2 𝑝2
Corner Solutions: Perfect Complements

In point 𝑥1∗ :
𝑚 𝑝1
𝑈 𝑥1 = 𝛼𝑥1 = 𝛽 − 𝛽 𝑥1
𝑝2 𝑝2
I.e., we have defined 𝑈 𝑥1 as a minimum of
two observed values of U (highlighted in
red).
To maximize the utility function, we need to
find the maximum value from the minimum
ones
Corner Solutions: Perfect Complements
We can see on the diagram that a maximum of
𝑈 𝑥1 is reached with 𝑥1∗ where:
∗
𝑚 𝑝1 ∗
𝛼𝑥1 = 𝛽 − 𝛽 𝑥1
𝑝2 𝑝2
We solve the equation for 𝑥1∗ :
𝛽𝑚 𝑚
𝑥1∗ = or 𝑥1∗ = 𝛼
𝛽𝑝1+𝛼𝑝2 𝑝1+ 𝑝2
𝛽

Then:
𝛼𝑚 𝑚
𝑥2∗ = or 𝑥2∗ =𝛽
𝛽𝑝1+𝛼𝑝2 𝑝 +𝑝2
𝛼 1
Corner Solutions: Perfect Complements
It is possible to use another way for the problem with
perfect complements

𝑥2

𝛼𝑥1 = 𝛽𝑥2

𝑥2∗

𝑥1∗ 𝑥1

The optimal point lies on the ray from the origin with
𝛼
the slope
𝛽
 Corner Solutions: Perfect Complements
Problem:
max 𝑈 𝑥1 , 𝑥2 = min{𝛼𝑥1 , 𝛽𝑥2 }
𝑥1 ,𝑥2 ≥0

s.t. 𝑝1 𝑥1 + 𝑝2 𝑥2 = 𝑚
can be converted into:
𝛼𝑥1 = 𝛽𝑥2
𝑝1 𝑥1 + 𝑝2 𝑥2 = 𝑚
and so on…
Methods for Solving the Optimization
Problem: Quasilinear Utility Function

Quasilinear preferences are monotonic and strictly

convex. But the indifference curves hit the axes, so
either an interior or corner solution is possible.
𝑥2 𝑥2

𝑥1 𝑥1
Methods for Solving the Optimization
Problem: Quasilinear Utility Function

Suppose a consumer has a quasilinear utility

function 𝑢 𝑥1 , 𝑥2 = 4 𝑥1 + 𝑥2 .
We use a two-step procedure to determine the
optimal bundle.
‒ We first check to see if there is an interior solution
using the tangency condition and the budget
constraint. If we find that these conditions imply that
he wants to buy positive quantities of both goods, we
have found an interior solution.
‒ Otherwise, we have to determine a corner solution
as a second step.
Methods for Solving the Optimization
Problem: Quasilinear Utility Function

At an interior solution his indifference curve is tangent to

his budget line, so that the 𝑀𝑅𝑆 = 𝑝1/𝑝2.
His marginal utility of good 1 is 𝑀𝑈1 = 2𝑥1−0,5 and his
marginal utility of good 2 = 1, so his 𝑀𝑅𝑆 = 𝑀𝑈1 /𝑀𝑈2 =
2𝑥1−0,5 .
We can use this tangency condition 2𝑥1−0,5 = 𝑝1/𝑝2,
to show that 𝑥1∗ = 4(𝑝2 𝑝1 )2 (prove it).
Because 𝑥1 does not depend on m, consumer buys the same
quantity of 𝑥1 regardless of his income, given that
we have an interior solution where the tangency condition
holds.
By substituting this value of 𝑥1 into the budget constraint,
we can solve for 𝑥2 : 𝑥2∗ = 𝑚/𝑝2 − 4(𝑝2 𝑝1 ) (prove it).
Methods for Solving the Optimization
Problem: Quasilinear Utility Function

Thus, if 𝑚 > 4, we have an interior solution

where both quantities are positive.
Suppose the prices are 𝑝1 = 𝑝2 = 1,
and 𝑚 = 6.
Then 𝑥1∗ = 4 and 𝑥2∗ = 𝑚 − 4 = 2
Methods for Solving the Optimization
Problem: Quasilinear Utility Function
Methods for Solving the Optimization
Problem: Quasilinear Utility Function

At incomes 𝑚 ≤ 4, we do not have an interior

solution - a tangency at positive levels of 𝑥1 and 𝑥2
We first show that the only possibly point of
tangency involves a negative quantity of one good.
Then we determine the corner solution.
For example, if 𝑚 = 2, we were at a point of
tangency where 𝑥2 = 2 − 4 = −2, which is not
plausible.
Methods for Solving the Optimization
Problem: Quasilinear Utility Function
Methods for Solving the Optimization
Problem: Quasilinear Utility Function
If 𝑚 = 2 the only solution with nonnegative
quantities of the two goods is corner one where
𝑥1 = 2 and 𝑥2 = 0.
So, if a consumer’s income is low (𝑚 ≤ 4), he
spends all his money on 𝑥1 , buying 𝑥1 = 𝑚 𝑝1 =
𝑚, which is a corner solution.
If he has enough income (𝑚 > 4) for an interior
solution, he buys a fixed amount of 𝑥1∗ =
4(𝑝2 𝑝1 )2 = 4 and spends all his extra money on
𝑥2 as his income rises.
Loosely speaking, a consumer views 𝑥1 as a
necessity and 𝑥2 as a luxury.
Methods for Solving the Optimization
Problem: Quasilinear Utility Function
Methods for Solving the Optimization
Problem: Quasilinear Utility Function

Now we use unconstrained optimization method

(converting the constrained optimization problem
into an unconstrained optimization problem),
i.e.*
max 𝑢 𝑥1 , 𝑥2 = 4 𝑥1 + 𝑥2
𝑥1 ,𝑥2 ≥0

s.t. 𝑥1 + 𝑥2 = 6
⇒ max 4 𝑥1 + (6 − 𝑥1 )
𝑥1 ≥0

* 𝑝1 = 𝑝2 = 1, 𝑚 = 6.
Methods for Solving the Optimization
Problem: Quasilinear Utility Function
The first-order condition (FOC) for interior
solution:
𝑑𝑢 2
= −1=0
𝑑𝑥1 𝑥1
FOCs are not only necessary but also sufficient
because the objective function for the problem is
strictly concave (its second derivative is strictly
less than zero)
So, from FOC we get 𝑥1 = 4 and 𝑥2 = 2
Methods for Solving the Optimization
Problem: Quasilinear Utility Function
Suppose 𝑚 = 2.
The first-order condition (FOC) for interior
solution:
𝑑𝑢 2
= −1=0
𝑑𝑥1 𝑥1
From FOC we get 𝑥1 = 4 and 𝑥2 = −2
But it is nonsense!
We have to employ a more complicated method
that explicitly introduces nonnegativity constraints
for all consumption goods.
Methods for Solving the Optimization
Problem: Quasilinear Utility Function
We have to employ a more complicated method
that explicitly introduces nonnegativity constraints
for all consumption goods.
This more complicated method is a generalization
of the Lagrange Method known as the “Kuhn-
Tucker method”
Kuhn-Tucker Conditions
Lagrangian for the consumer problem:*
𝓛 = 4 𝑥1 + 𝑥2 + (𝑚 − 𝑥1 − 𝑥2 )
Kuhn-Tucker conditions:
𝜕𝓛 𝜕𝑥1 = 2/ 𝑥1 −  ≤ 0 and = 0 if 𝑥1 > 0 (1)
𝜕𝓛 𝜕𝑥2 = 1 −  ≤ 0 and = 0 if 𝑥2 > 0 (2)
𝜕𝓛 𝜕 = 𝑚 − 𝑥1 − 𝑥2 ≥ 0 and = 0 if  > 0 (3)

* 𝑝1 = 𝑝2 = 1, 𝑚 = 6. The problem would be a little bit

harder if 𝑝1 ≠ 𝑝2 > 1 or in case of general form for a
budget constraint 𝑝1𝑥1 + 𝑝2𝑥2 ≤ 𝑚
Kuhn-Tucker Conditions
Condition (2) implies that 1 ≤ . If  = 0 this
condition (2) is not met, and hence  > 0, and thus
condition (3) is the budget equation.
So we need a look for 4 cases:
𝑥1 = 0, 𝑥2 =0
𝑥1 = 0, 𝑥2 >0
𝑥1 > 0, 𝑥2 =0
𝑥1 > 0, 𝑥2 >0
Kuhn-Tucker Conditions
𝑥1 = 0, 𝑥2 = 0 do not satisfy the budget equation
which must hold. The economics of the matter is
that the goods have positive marginal utilities.
We can see that 𝑥1 > 0, 𝑥1 ≠ 0
Suppose ε > 0.
Compare two bundles: (𝑥1 = 𝜀, 𝑥2 = 𝑚 − 𝜀) and
(𝑥1 = 0, 𝑥2 = 𝑚)
We can show that there is always such a value 0 <
𝜀 < 16 → 𝑢(𝑥1 , 𝑥2 ) > 𝑢(𝑥1 , 𝑥2 ):
4 𝜀 + 𝑚 − 𝜀 > 𝑚 ⇒ 16 > 𝜀
(𝑥1 = 0, 𝑥2 = 𝑚) can not be a solution!
Kuhn-Tucker Conditions
Because of 𝑥1 ≠ 0 condition (1) is satisfied as an
equality : 2/ 𝑥1 = 
From (2) we have 1 ≤ 
Hence 2/ 𝑥1 ≥ 1 and 𝑥1 ≤ 4
𝑥1 = 4 if 𝑥2 = 0
At the same time 𝑥1 ≤ min{4, 𝑚}
We get 𝑥1 ≤ 𝑚; 𝑥1 = 𝑚 if 𝑥2 = 0
If 𝑥2 = 0 ⇒ 𝑥1 = 𝑚. If 𝑥2 > 0 ⇒ 𝑥1 = 4
If 𝑥1 = min{4, 𝑚}
Then 𝑥1 = 𝑚 ⇔ 𝑥2 = 0 ⇒ 𝑚 ≤ 4
So 𝑥1∗ = 𝑚, 𝑥2∗ = 0, if 𝑚 ≤ 4;
𝑥1∗ = 4, 𝑥2∗ = 𝑚 − 4, if 𝑚 > 4
Kuhn-Tucker Conditions
If budget constraint is given in general form, 𝑝1𝑥1 +
𝑝2𝑥2 ≤ 𝑚, we have solutions:
𝑚 4𝑝22 4𝑝22
, if 𝑚≤ 0, if 𝑚 ≤
𝑝1 𝑝1 𝑝1
𝑥1∗ = 𝑥2∗ =
4𝑝22 4𝑝22 𝑚 4𝑝2 4𝑝22
2 , if 𝑚> − , if 𝑚>
𝑝1 𝑝1 𝑝2 𝑝1 𝑝1
Demand Functions
In all examples above the solution to
problem can be written as the demand
functions
𝑥1∗ = 𝑥1 (𝑝1 , 𝑝2 , 𝑚)
𝑥2∗ = 𝑥1 (𝑝1 , 𝑝2 , 𝑚)
Demand Functions: Cobb-Douglas

Cobb-Douglas utility function:

𝛼 𝛽
𝑢 𝑥1 , 𝑥2 = 𝑥1 𝑥2
Solution (demand functions):
𝛼 𝑚
𝑥1∗= 𝑥1 𝑝1 , 𝑝2 , 𝑚 =
𝛼 + 𝛽 𝑝1
∗
𝛽 𝑚
𝑥2 = 𝑥1 𝑝1 , 𝑝2 , 𝑚 =
𝛼 + 𝛽 𝑝2
Demand Functions: Perfect Substitutes

Perfect substitutes utility function:

𝑢 𝑥1 , 𝑥2 = 𝛼𝑥1 + 𝛽𝑥2
Solution (demand functions):
𝑚 𝛼 𝑝1 𝑚 𝛼 𝑝1
, if > , if <
𝑝1 𝛽 𝑝2 𝑝2 𝛽 𝑝2
𝑚 𝛼 𝑝1 𝑚 𝛼 𝑝1
𝑥1∗ = ∈ 0;
𝑝1
, if =
𝛽 𝑝2
𝑥2∗ = ∈ 0;
𝑝2
, if =
𝛽 𝑝2
𝛼 𝑝1 𝛼 𝑝1
0, if < 0, if >
𝛽 𝑝2 𝛽 𝑝2
Demand Functions: Perfect Complements

Perfect complements utility function:

𝑢 𝑥1 , 𝑥2 = min{𝛼𝑥1 , 𝛽𝑥2 }
Solution (demand functions):
𝛽𝑚 𝛼𝑚
𝑥1∗ = 𝑥2∗ =
𝛽𝑝1 +𝛼𝑝2 𝛽𝑝1 +𝛼𝑝2
Demand Functions: Quasi-linear
Quasi-linear utility function:
𝑢 𝑥1 , 𝑥2 = 𝛼 𝑥1 + 𝛽𝑥2
Solution (demand functions):
𝑚 𝛼2 𝑝22
, if 𝑚≤
𝑝1 4𝛽2 𝑝1
𝑥1∗ =
𝛼2 𝑝22 𝛼2 𝑝22
, if 𝑚 >
4𝛽2 𝑝12 4𝛽2 𝑝1

𝛼 2 𝑝22
0, if 𝑚 ≤ 2
∗ 4𝛽 𝑝1
𝑥2 =
𝑚 𝛼 2 𝑝2 𝛼 2 𝑝22
− , if 𝑚 > 2
𝑝2 4𝛽2 𝑝1 4𝛽 𝑝1
Demand Functions: CES
CES utility function:
𝜌 𝜌 1/𝜌
𝑢 𝑥1 , 𝑥2 = (𝑥1 +𝑥2 )
Solution (demand functions):
𝑚𝑝1𝜎 𝑚𝑝2𝜎
𝑥1∗ = 𝑥2∗ =
𝑝1𝜎+1 +𝑝2𝜎+1 𝑝1𝜎+1 +𝑝2𝜎+1
where
1
𝜎=
𝜌−1
Demand Functions
Again: the solution to consumer’s utility
maximization problem can be written as the
demand functions and the optimal values 𝑥 ∗
depend on the prices of all goods and income:
𝑥1∗ = 𝑑1 (𝑝1 , 𝑝2 , 𝑚)
𝑥2∗ = 𝑑2 (𝑝1 , 𝑝2 , 𝑚)
These functions are known as
uncompensated (or Marshallian) demand
functions
Demand Functions and Homogeneity
Marshallian demand functions are homogeneous
of degree zero in prices and income. That is, for
any positive constant t,
𝑑1 𝑡𝑝1 , 𝑡𝑝2 , 𝑡𝑚 = 𝑡 0 𝑑1 𝑝1 , 𝑝2 , 𝑚 = 𝑑1 (𝑝1 , 𝑝2 , 𝑚)
𝑑2 𝑡𝑝1 , 𝑡𝑝2 , 𝑡𝑚 = 𝑡 0 𝑑2 𝑝1 , 𝑝2 , 𝑚 = 𝑑2 (𝑝1 , 𝑝2 , 𝑚)
The reason Marshallian demands are
homogeneous of degree zero is that consumption
opportunities do not change if prices and income
change by the same proportion.