Master of Economics
Foundations of
Microeconomics
Chapter 4 - Utility Maximization and Choice
Utility Maximization and Choice p. 115
Complaints about the Economic Approach
C: Do individuals make the “lightning calculations” required for utility
maximization?
■ A: The utility-maximization model predicts many aspects of behavior
■ A: Economists assume that people behave as if they made such
calculations
C: Individuals are modeled as extremely selfish
■ A: Nothing in the model prevents individuals from getting satisfaction
from “doing good”
Slide 2
Utility Maximization, General Principle
■ Individual is endowed with some income, that she can spend to buy
the different goods
■ Budget constraint: which bundles can the individual afford?
■ Utility maximization is a constrained maximization problem: individual
chooses the feasible consumption bundle that yields the highest
utility
Slide 3
Budget Constraint
■ Budget to allocate between good and good y
■ p is the price of good and py is the price of good y
■ Budget constraint:
p + py y ≤ (4.1)
■ Individual cannot spend more than her income. When all the income
is spent
− p
y=
py
p
■ Defines a line, slope is − p . Bundles on the line are such that income
y
is entirely spent
■ Individual can afford all bundles below the line: feasible consumption
set
Slide 4
Utility Maximization, Graphically
(Figure 4.2) p. 118
py
= p + py y
0
p
Slide 5
Utility Maximization, Graphically
(Figure 4.2) p. 118
py
D
= p + py y
U3
0
p
Slide 5
Utility Maximization, Graphically
(Figure 4.2) p. 118
py
B
D
= p + py y
A
U3
U1
0
p
Slide 5
Utility Maximization, Graphically
(Figure 4.2) p. 118
py
B
D
C = p + py y
A
U3
U2
U1
0
p
Slide 5
Utility Maximization, Graphically
(Figure 4.2) p. 118
py
B
D
C = p + py y
y∗
A
U3
U2
U1
0 ∗
p
Slide 5
From Graphical to Analytical Solution (1)
■ Graphically, utility is maximized when the slope of the budget
constraint is equal to the slope of the indifference curve
■ Formally, the maximization problem is:
mx U (, y)
subject to the budget constraint
= p + py y
■ Set up the Lagrangian:
L = U (, y) + λ − p − py y
Slide 6
From Graphical to Analytical Solution (2)
■ First-order conditions (FOCs) for an interior maximum
∂L ∂U
∂
= ∂
− λp = 0
∂L ∂U
∂y
= ∂y
− λpy = 0
■ Rearranging the FOCs, we get:
∂U
p ∂
U
− =− ∂U
=−
py Uy
∂y
■ The slope of the budget constraint is indeed equal to the slope of the
indifference curve (or −MRS)
■ However, it relies on a well-behaved utility function to ensure a
unique solution to the maximization problem
Slide 7
Not Well-Behaved Utility Functions p. 119
■ Tangency is sufficient
with diminishing MRS
■ Non-diminishing MRS
requires checking
second-order conditions
for a maximum
p
■ At point C, MRS = p ,
y
but it does not maximize
utility
Example of indifference curves for which the
tangency condition does not ensure a maximum
Slide 8
Corner Solutions p. 120
■ Utility maximized if
only one of the goods is
consumed
■ At the optimal bundle,
the budget constraint
can be flatter or
steeper than the
indifference curve
■ Point E: Price ratio
smaller than MRS
■ Solution not
characterized by a FOC
Corner solution to utility maximization
Slide 9
The n-Good Case (1) p. 120f.
■ The individual’s objective is to maximize
U (1 , 2 , ..., n ) (4.4)
subject to the budget constraint
= p1 1 + p2 2 + ... + pn n (4.5)
■ Set up the Lagrangian:
L = U (1 , 2 , ..., n ) + λ ( − p1 1 − p2 2 − ... − pn n ) (4.7)
■ FOCs for an interior maximum
∂L ∂U
∂1
= ∂1
− λp1 = 0
∂L ∂U
∂2
= ∂2
− λp2 = 0
... (4.8)
∂L ∂U
∂n
= ∂n
− λpn = 0
∂L
∂λ
= − py 1 − p2 2 − ... − pn n = 0
Slide 10
The n-Good Case (2) p. 121f.
■ Implications of FOCs: For any two goods, and j
∂U
∂ p
∂U
= = MRS( ƒ or j ) (4.9/4.10)
pj
∂j
■ Lagrange multiplier (see also sections 2.5.4 and 2.5.5)
■ λ represents the utility increase associated with relaxing the budget
constraint (‘marginal utility of income’)
■ At utility-maximizing point, the marginal utility of income is the same
across goods ∂U ∂U ∂U
∂1 ∂2 ∂n
λ= = = ... = (4.11)
p1 p2 pn
■ The price represents the individual’s evaluation of the marginal utility, i.e.,
how much the individual is willing to pay for the last unit
∂U
∂
p = , for every (4.12)
λ Slide 11
Corner solutions in the n-Good Case p. 122f.
■ The FOCs must be modified:
∂L ∂U
= − λp ≤ 0 ( = 1, ..., n) (4.13)
∂ ∂
■ If
∂L ∂U
= − λp < 0 (4.14)
∂ ∂
∂U
∂
■ then, = 0 and p > λ
■ When the price exceeds the evaluation of the marginal utility for the
first unit of a good, this good not purchased
(see also section 2.7!)
Slide 12
Example 4.1: Demand Functions for
Cobb–Douglas Utility p. 123f.
Consider the Cobb-Douglas utility function:
U(, y) = α y β , where α + β = 1 (4.17)
Moreover, suppose the individual has a budget that she spends on the
two goods, with prices p and py .
■ What are the quantities of and y demanded?
■ What is the share of income allocated to each of the goods?
Comment.
■ Suppose p = 1, py = 4, = 8, and α = β = 0.5. What are the
quantities demanded and what is the resulting level of utility?
■ What is the value of the Lagrange multiplier λ associated with those
parameters? Interpret.
Slide 13
Example 4.2: CES Demand
1/ δ
Consider the CES utility function U(, y) = (δ + y δ ) . The budget is
spent on the two goods, with prices p and py .
■ What are the quantities of and y demanded for δ = 0.5 ?
■ How does the share of income spent on each good react to changes in
the price ratio?
■ Same questions for δ = −1 and for U(, y) = min(, 4y).
■ Do changes in income affect expenditure shares in any of the CES
functions discussed here? How is the behavior of expenditure shares
related to the homothetic nature of this function?
Slide 14
Indirect Utility Function (1) p. 126f.
■ In Examples 4.1 and 4.2 we have seen that the demand functions
depend on prices and income:
∗
1
= 1 (p1 , p2 , ..., pn , )
∗
2
= 2 (p1 , p2 , ..., pn , )
...
∗
n
= n (p1 , p2 , ..., pn , ) (4.41)
■ We can find the optimal level of utility by substituting these demand
functions into the original utility function to obtain the indirect
utility function
V (p1 , p2 , ..., pn , ) = (4.43)
U ∗
1
(p1 , p2 , ..., pn , ) , ∗
2
(p1 , p2 , ..., pn , ) , ..., ∗
n
(p1 , p2 , ..., pn , ) (4.42)
Slide 15
Indirect Utility Function (2) p. 126f.
■ The optimal level of utility depends indirectly on prices and income
■ Indirect utility function is an example of a value function:
■ „solves out“ all endogenous variables in an optimization problem, such
that the optimal value only depends on exogenous variables (e.g.,
prices)
■ shortcut to exploring how changes in exogenous variables affect
outcome
Slide 16
The Lump Sum Principle p. 127f.
Income taxes are better than taxes on a specific good
■ Both taxes reduce the individual’s purchasing power
■ An income tax does not distort relative prices and allows to freely
allocate the remaining income
■ A tax on a specific good distorts relative prices and thus also the
choice
■ Thus, the utility burden of an income tax is smaller
■ Same for income grants compared to subsidies for specific goods
Slide 17
Lump-sum Principle of Taxation
(Figure 4.5) p. 128
y∗
U3
∗
Slide 18
Lump-sum Principle of Taxation
(Figure 4.5) p. 128
y1
y∗
U3
U1
1 ∗
Slide 18
Lump-sum Principle of Taxation
(Figure 4.5) p. 128
′
y1
y∗
U3
U1
1 ∗
Slide 18
Lump-sum Principle of Taxation
(Figure 4.5) p. 128
′
y1
y∗
y2
U3
U1 U2
1 2 ∗
Slide 18
Lump Sum Principle, Interpretation of
the Graph p. 127f.
■ A tax on good changes the utility-maximizing choice from (∗ , y ∗ )
to (1 , y1 )
■ Changes the slope of the budget line and collects t1 :
= (p + t) + py y
■ An income tax leaves the individual with ′ and collects − ′
■ Slope of the budget line is unchanged
′ = p + py y
■ When both the budget line with a tax on and an income tax go
through (1 , y1 ), they collect the same taxes t1 :
− ′ = ((p + t)1 + py y1 ) − (p 1 + py y1 ) = t1
■ However, higher indifference curve (U2 ) can be achieved with the
income tax than with the tax on alone (U1 )
Slide 19
Example 4.3 (I): Indirect Utility and the
Lump Sum Principle
Suppose p = 1, py = 4, and = 8.
■ What is the indirect utility with a Cobb-Douglas utility function
U(, y) = 0.5 y 0.5 ?
From Example 4.1: ∗ = 2p and y ∗ = 2p .
y
■ Suppose we impose a tax of 1 per unit of . What is the resulting level
of utility? What would have been the resulting utility with an income
tax that collects the same amount?
■ Same questions with the fixed-proportion utility function
U(, y) = min(, 4y).
From Example 4.2: ∗ = p +0.25p and y ∗ = 4p +p .
y y
Slide 20
Example 4.3 (II): Property of Indirect Uti-
lity
■ Consider the indirect utility functions derived in Example 4.3 (I). What
would happen if we double income and all prices?
■ What does it suggest for the homogeneity of indirect utility functions?
■ Explain why you would expect this to be a property of all indirect
utility functions.
Slide 21
Primal and Dual Problems p. 129f.
■ So far, we maximized utility subject to budget constraint
■ This is called the primal problem
■ Alternative approach: reverse the problem and minimize expenditure
subject to the constraint that utility must have a given value
■ This is called the dual problem
Slide 22
Expenditure Minimization (1) p. 129f.
Dual minimization problem
to the case of utility maximiza-
tion
■ Allocate income to achieve a
given level of utility with the
minimal expenditure
■ The goal and the constraint
are reversed
The dual expenditure-minimization problem
Slide 23
Expenditure Minimization (2) p. 130f.
■ The individual’s problem is to choose 1 , 2 , . . . , n to minimize total
expenditures E:
E = p1 1 + p2 2 + ... + pn n (4.48)
subject to the constraint
U (1 , 2 , ..., n ) = U (4.49)
■ Set up the Lagrangian:
L = p1 1 + p2 2 + ... + pn n + μ[U − U (1 , 2 , ..., n )]
Slide 24
Expenditure Minimization (3) p. 130f.
■ The first order conditions are
∂L ∂U
∂1
= p1 − μ ∂ =0
1
∂L ∂U
∂2
= p2 − μ ∂ =0
2
...
∂L ∂U
∂n
= pn − μ ∂ =0
n
∂L
∂μ
= U − U(1 , 2 , ..., n ) = 0
■ Note that the relation between the Lagrange multipliers of the primal
and dual problem is λ = μ1
Slide 25
Expenditure Minimization (4) p. 131
■ The optimal quantities are c1 (p1 , ..., pn , U), ..., cn (p1 , ..., pn , U). They
are called the Hicksian or compensated demands. They depend
on the prices of the goods and the required utility level.
■ The result of the minimization problem is the expenditure function:
E p1 , p2 , ..., pn , U = p1 c1 (p1 , ..., pn , U) + ... + pn cn (p1 , ..., pn , U)
■ The expenditure function shows the minimal expenditures necessary
to achieve a given utility level for particular prices. It is also a value
function.
■ The expenditure and indirect utility functions are inversely related
and depend on prices. However, they involve different constraints
(income or utility).
Slide 26
Primal and Dual, Summary
■ Primal problem ■ Dual problem
■ Maximize utility subject to ■ Minimize expenses subject to
budget constraint utility constraint
■ Arguments solving the ■ Arguments solving the
maximization problem are the minimization problem are the
(uncompensated) demands, also compensated or Hicksian
called Marshallian demands demands
■ Value of the maximum is the ■ Value of the minimum is the
indirect utility function expenditure function
Slide 27
Example 4.4: Two Expenditure Functions
(1/2)
■ Solve the expenditure-minimization problem to compute the
expenditure function with a Cobb-Douglas utility function
U(, y) = 0.5 y 0.5 .
■ In Example 4.3 (I), we have established that V(p , py , ) = p .
2 p py
Show that this result can be used to compute the same expenditure
function as in the previous question.
■ Suppose p = 1, py = 4. What is the minimal expenditure needed to
achieve U = 2?
■ Suppose the price of y increases to py = 5. What is the additional
expenditure needed to maintain U = 2?
Slide 28
Example 4.4: Two Expenditure Functions
(2/2)
Consider the fixed-proportion utility function U(, y) = min(, 4y). In
Example 4.3 (I), we have established that V(p , py , ) = p +0.25p .
y
■ Use this indirect utility to recover the expenditure function.
■ Suppose p = 1, py = 4. What is the expenditure needed to achieve
U = 4? What is the additional expenditure needed to maintain U = 4 if
py = 5?
■ Compare the results with the Cobb-Douglas and the fixed-proportion
utility functions.
■ How would a person be compensated for a price decrease?
Slide 29
Properties of Expenditure Functions (1) p. 132ff.
■ Homogeneity
■ A doubling of all prices precisely doubles the value of the required
expenditures
■ Homogeneous of degree one in all prices:
E(tp1 , tp2 , ..., tpn , U) = tE(p1 , p2 , ..., pn , U)
■ Nondecreasing in prices
∂E
≥ 0 for every good (4.55)
∂p
■ Concave in each price
■ The composition of the bundle changes with the prices
Slide 30
Properties of Expenditure Functions (2) p. 133
■ At p∗ , expenditures are
1
∗
E p1 , ... .
■ If the same bundle is
purchased as p1
changes, expenditures
would be given by
Epsedo .
■ As bundle likely
changes with varying
p1 , actual expenditures
will be below Epsedo
(concavity in prices).
Concavity in Prices
see Example 4.4 "Compensating for a Price Change"
Slide 31
