Chapter 11

Utility
 Structure
Utility function (效用函数）
– Definition
– Monotonic transformation （单调转换）
– Examples of utility functions and their
indifference curves
Marginal utility （边际效用）
Marginal rate of substitution 边际替代率
– MRS after monotonic transformation
 Utility Functions
A utility function U(x) represents a
preference relation f ~ if and only if:
p
x’ x” U(x’) > U(x”)

x’ p x” U(x’) < U(x”)

x’ ~ x” U(x’) = U(x”).
 Utility Functions
Utility is an ordinal (i.e. ordering)
concept. [序数效用]
E.g. if U(x) = 6 and U(y) = 2 then
bundle x is strictly preferred to
bundle y. But x is not preferred three
times as much as is y.
Utility is always non-minus.
Utility Functions & Indiff. Curves
Consider the bundles (4,1), (2,3) and
(2,2).
(4,1) ~ (2,2).
p
Suppose (2,3)
Assign to these bundles any
numbers that preserve the
preference ordering;
e.g. U(2,3) = 6 > U(4,1) = U(2,2) = 4.
Call these numbers utility levels.
Utility Functions & Indiff. Curves
An indifference curve contains
equally preferred bundles.
Equal preference  same utility level.
Therefore, all bundles in an
indifference curve have the same
utility level.
Utility Functions & Indiff. Curves
So the bundles (4,1) and (2,2) are in
the indiff. curve with utility level U  
But the bundle (2,3) is in the indiff.
curve with utility level U  6.
On an indifference curve diagram,
this preference information looks as
follows:
Utility Functions & Indiff. Curves

x2 (2,3)
p (2,2) ~ (4,1)

U6
U4

x1
Utility Functions & Indiff. Curves
Comparing more bundles will create
a larger collection of all indifference
curves and a better description of
the consumer’s preferences.
Utility Functions & Indiff. Curves

x2

U6
U4
U2
x1
Utility Functions & Indiff. Curves
The collection of all indifference
curves for a given preference relation
is an indifference map.
An indifference map is equivalent to
a utility function; each is the other.
 Utility Functions
There is no unique utility function
representation of a preference
relation.
Suppose U(x1,x2) = x1x2 represents a
preference relation.
Again consider the bundles (4,1),
(2,3) and (2,2).
 Utility Functions
U(x1,x2) = x1x2, so

U(2,3) = 6 > U(4,1) = U(2,2) = 4;

(4,1) ~ (2,2).
p
that is, (2,3)
 Utility Functions
(4,1) ~ (2,2).
p
U(x1,x2) = x1x2 (2,3)
Define V = U2.
 Utility Functions
(2,3) (4,1) ~ (2,2).
p
U(x1,x2) = x1x2
Define V = U2.
Then V(x1,x2) = x12x22 and
V(2,3) = 36 > V(4,1) = V(2,2) = 16
so again
(2,3) (4,1) ~ (2,2).
p
V preserves the same order as U and
so represents the same preferences.
 Utility Functions
(4,1) ~ (2,2).
p
U(x1,x2) = x1x2 (2,3)
Define W = 2U + 10.
 Utility Functions
(2,3) (4,1) ~ (2,2).
p
U(x1,x2) = x1x2
Define W = 2U + 10.
Then W(x1,x2) = 2x1x2+10 so
W(2,3) = 22 > W(4,1) = W(2,2) = 18.
Again,
(2,3) (4,1) ~ (2,2).
p
W preserves the same order as U and V
and so represents the same preferences.
Utility Functions: Monotonic
Transformation
If
– U is a utility function that
represents a preference relation f ~
and
– f is a strictly increasing function,
then V = f(U) is also a utility function
representing f .
~
 Goods, Bads and Neutrals
A good is a commodity unit which
increases utility (gives a more
preferred bundle).
A bad is a commodity unit which
decreases utility (gives a less
preferred bundle).
A neutral is a commodity unit which
does not change utility (gives an
equally preferred bundle).
 Goods, Bads and Neutrals
Utility
Utility
function
Units of Units of
water are water are
goods bads

x’ Water
Around x’ units, a little extra water is a neutral.
Some Other Utility Functions and
Their Indifference Curves
Instead of U(x1,x2) = x1x2 consider

V(x1,x2) = x1 + x2.

What do the indifference curves for

this “perfect substitution” utility
function look like?
Perfect Substitution Indifference
Curves
x2
x1 + x2 = 5
13
x1 + x2 = 9
9
x1 + x2 = 13
5
V(x1,x2) = x1 + x2.

5 9 13 x1
Perfect Substitution Indifference
Curves
x2
x1 + x2 = 5
13
x1 + x2 = 9
9
x1 + x2 = 13
5
V(x1,x2) = x1 + x2.

5 9 13 x1
All are linear and parallel.
Some Other Utility Functions and
Their Indifference Curves
Instead of U(x1,x2) = x1x2 or
V(x1,x2) = x1 + x2, consider

W(x1,x2) = min{x1,x2}.

What do the indifference curves for

this “perfect complementarity” utility
function look like?
 Perfect Complementarity
Indifference Curves
x2
45o
W(x1,x2) = min{x1,x2}

8 min{x1,x2} = 8
5 min{x1,x2} = 5
3 min{x1,x2} = 3

3 5 8 x1
 Perfect Complementarity
Indifference Curves
x2
45o
W(x1,x2) = min{x1,x2}

8 min{x1,x2} = 8
5 min{x1,x2} = 5
3 min{x1,x2} = 3

3 5 8 x1
All are right-angled with vertices on a ray
from the origin.
Some Other Utility Functions and
Their Indifference Curves
A utility function of the form

U(x1,x2) = f(x1) + x2

is linear in just x2 and is called quasi-

linear （准线性）.
E.g. U(x1,x2) = 2x11/2 + x2.
 Quasi-linear Indifference Curves
x2 Each curve is a vertically shifted
copy of the others.

x1
Some Other Utility Functions and
Their Indifference Curves
Any utility function of the form

U(x1,x2) = x1a x2b

with a > 0 and b > 0 is called a Cobb-

Douglas utility function.
E.g. U(x1,x2) = x11/2 x21/2 (a = b = 1/2)
V(x1,x2) = x1 x23 (a = 1, b = 3)
 Cobb-Douglas Indifference
x2 Curves

x1
 Marginal Utilities
Marginal means “incremental”.
The marginal utility of commodity i is
the rate-of-change of total utility as
the quantity of commodity i
consumed changes; i.e.
U
MU i =
 xi
 Marginal Utilities
E.g. if U(x1,x2) = x11/2 x22 then

 U 1 − 1/ 2 2
MU1 = = x1 x2
 x1 2
 Marginal Utilities
E.g. if U(x1,x2) = x11/2 x22 then

 U 1 − 1/ 2 2
MU1 = = x1 x2
 x1 2
 Marginal Utilities
E.g. if U(x1,x2) = x11/2 x22 then

U 1/ 2
MU 2 = = 2 x1 x2
 x2
 Marginal Utilities
E.g. if U(x1,x2) = x11/2 x22 then

U 1/ 2
MU 2 = = 2 x1 x2
 x2
 Marginal Utilities
So, if U(x1,x2) = x11/2 x22 then
 U 1 − 1/ 2 2
MU1 = = x1 x2
 x1 2
U 1/ 2
MU 2 = = 2 x1 x2
 x2
Marginal Utilities and Marginal
Rates-of-Substitution
The general equation for an
indifference curve is
U(x1,x2)  k, a constant.
Totally differentiating this identity gives
U U
dx1 + dx2 = 0
 x1  x2
Marginal Utilities and Marginal
Rates-of-Substitution
U U
dx1 + dx2 = 0
 x1  x2
rearranged is
U U
dx2 = − dx1
 x2  x1
Marginal Utilities and Marginal
Rates-of-Substitution
And U U
dx2 = − dx1
 x2  x1
rearranged is
d x2  U /  x1
=− .
d x1  U /  x2
This is the MRS.
Marg. Utilities & Marg. Rates-of-
Substitution; An example
Suppose U(x1,x2) = x1x2. Then
U
= (1)( x2 ) = x2
 x1
U
= ( x1 )(1) = x1
 x2
d x2  U /  x1 x2
so MRS = =− =− .
d x1  U /  x2 x1
Marg. Utilities & Marg. Rates-of-
Substitution; An example
x2
x2 U(x1,x2) = x1x2; MRS = −
x1
8 MRS(1,8) = - 8/1 = -8
6 MRS(6,6) = - 6/6 = -1.

U = 36
U=8
1 6 x1
 Marg. Rates-of-Substitution for
Quasi-linear Utility Functions
A quasi-linear utility function is of
the form U(x1,x2) = f(x1) + x2.
U U
= f ( x1 ) =1
 x1  x2
d x2  U /  x1
so MRS = =− = − f  ( x1 ).
d x1  U /  x2
Marg. Rates-of-Substitution for
Quasi-linear Utility Functions
MRS = - f  (x1) does not depend upon
x2 so the slope of indifference curves
for a quasi-linear utility function is
constant along any line for which x1
is constant. What does that make
the indifference map for a quasi-
linear utility function look like?
 Marg. Rates-of-Substitution for
x2
Quasi-linear Utility Functions
MRS = Each curve is a vertically
- f(x1’) shifted copy of the others.
MRS = -f(x1”) MRS is a
constant
along any line
for which x1 is
constant.

x1’ x1” x1
Monotonic Transformations &
Marginal Rates-of-Substitution
Applying a monotonic transformation
to a utility function representing a
preference relation simply creates
another utility function representing
the same preference relation.
What happens to marginal rates-of-
substitution when a monotonic
transformation is applied?
Monotonic Transformations &
Marginal Rates-of-Substitution
For U(x1,x2) = x1x2 the MRS = - x2/x1.
Create V = U2; i.e. V(x1,x2) = x12x22.
What is the MRS for V?
 V /  x1 2
2 x1 x2 x2
MRS = − =− =−
 V /  x2 2
2 x1 x2 x1
which is the same as the MRS for U.
 Monotonic Transformations &
Marginal Rates-of-Substitution
More generally, if V = f(U) where f is a
strictly increasing function, then
 V /  x1 f  (U )   U / x1
MRS = − =−
 V /  x2 f '(U )   U / x2
 U /  x1
=− .
 U /  x2
So MRS is unchanged by a positive
monotonic transformation.