Preference and Utility

Outline
1. Utility Functions and Preference Representation
2. Continuous Preferences
3. Existence of Utility Functions
 Preference and
Preferences and Utility
UtilityFunctions
Functions
Last week, we discussed what rationality implies for choices: given a preference
relation %, the induced choice rule C% is defined by
C% (A) = {x 2 A : x % y for all y 2 A}.
On the other hand, from initermediate micro, we are used to modeling
consumers’ behavior using a utility function: the consumer chooses, among all
a§ordable bundles, one that maximizes utility:
Cutility (A) = {x 2 A : u(x) ≥ u(y ) for all y 2 BudgetSet}.
Next we study when these two ways of looking at choices are equivalent.
When is it possible for a preference relation to be described by a function?
What does this mean?
A utility function maps consumption bundles to real numbers.
It represents preferences if the function’s values rank any pair of consumption
bundles in the same way the preferences would rank them.
If it exists, a utility function provides a tremendous simplification: instead of
comparing two highly dimensional objects using preferences, just look at which
of two real numbers is largest.

Question
Under what assumptions can a preference relation be represented by a utility
function? 2
 Preference Representation
Preference Representation
Definition
A utility function on X is a function u : X ! R.
Preference Representation
Definition
The utility function u : X ! R represents the binary relation % on X if
Definition
x % y u,: u(x)
A utility function on X is a function X !≥R.u(y).

The utility
•Question
Definition values do
1, Problem Setnot have Monday
2 (due any importance
September in the
14).sense that they just
Letrepresent
The utility the order
% be afunction
preference X of
!preferences.
u : relation. Prove thatthe
R represents u : binary
X ! Rrelation
represents %X
% on if and
if only if:
• The utilityx function
% y ) u(x) just≥attaches
x % y ,au(x)
u(y); number
and x to
≻ yall)elements.
≥ u(y). u(x) > u(y).

Exercise
This 11,
Question (Part
result of HWSet
provides
Problem 3):
an 2alternative way toSeptember
(due Monday show that that
14).a preference is
represented by some function (very useful).
Let % be a preference relation. Prove that u : X ! R represents % if and only if:
x % y ) u(x) ≥ u(y); and x ≻ y ) u(x) > u(y).
• This result provides an alternative way to show that a preference is
represented by some
This result provides anfunction (very
alternative wayuseful).
to show that that a preference is 3
 ExistenceofofaaUtility
Existence Utility Function
Functionand
andAxioms
Axioms
Thefollowing
• The followingresult
result says
says that
that completeness
completeness and transitivity
and are are
transitivity necessary for
necessary
existence of a utility function representing some preference relation.
for existence of a utility function representing some preference relation.
Proposition
If there exists a utility function that represents %, then % is complete and transitive.

Proof.
Let u (·) be a utility function that represents %.
Transitivity.
Suppose x % y and y % z.
Since u (·) represents %, we have u(x) ≥ u(y) and u(y) ≥ u(z).
Therefore, u(x) ≥ u(z).
Since u (·) represents %, we conclude x % z. Hence, % is transitive.
Completeness.
Since u : X ! R, for any x, y 2 X , either u(x) ≥ u(y) or u(y) ≥ u(x).
Since u (·) represents %, we have x % y or y % x.
Therefore, % is complete. 4
 Preferences, Utility Functions, and Ordinality
osition Preferences,
Preferences,
Preferences,
oposition
UtilityFunctions,
Utility
Utility Functions, and
Functions, andOrdinality
and Ordinality
Ordinality
ose
A unatural
: X !question
oposition R representsto %. isThen
ask v : X the
whether !R represents
utility % if and only
representation of a ifgiven
Proposition
ppose
exists ua: strictly
X ! R increasing %. Then hv ::u(X
represents function X ! ) R represents
! R such % vif =
that andh only
◦ u. if
preference
ppose relation
u :aXustrictly is unique.
%. Thehvfollowing proposition %vasserts that ifthe
ere exists
Suppose :!XR !represents
increasing
R represents %.Then
function
Then :v:u(X
:XX!)!!RRR represents
such that
represents % ifif=and
and only
h ◦only
u. if
answer
erethere
exists to
exists this
a strictlyquestion
a strictly is
increasing “no.”
increasingfunction
functionh h: :u(X
u(X))!!R R such thatvv==hh◦ ◦u.u.
such that
Preferences,
f. Exercise
oof.
2 (Part of HW3):Utility
Prove theFunctions,
proposition below. and Ordinality
oof.
tion 2, Problem Set 2; due Monday September 14.
Proof.
Proposition
estion 2, Problem Set 2; due Monday September 14.
One2,direction
e: Question
uestion
Suppose
Problem isSet
u2,: Problem
X !R
easy2; while
Set due
2; due
represents
the
Monday
%.
other
Monday
Then
is: Xnot.
September
vSeptember
14.
14.
! R represents % if and only if
ote: One direction is easy while the other
ote: One direction is easy while the other is not. is not.
Note: One adirection
there exists is easy while
strictly increasing the other
function is not.
h : u(X ) ! R such that v = h ◦ u.
cise
ercise
ercise
ne Exercise
Proof.
the binary relation ≍ on RX XXby u ≍ v if there exists a strictly increasing
fine
fine the
the binary
Define binary
the
relation
relation
binary
≍≍ on
relation
RR Rby
onon
≍ Xbyuu≍
by ≍≍
u
vvvifififthere
there
there
exists
exists
exists
aaastrictly
strictlyincreasing
strictly increasingX
increasing
(X Question
) ! R 2,
suchProblem
that v Set
= h2;
◦ due
u. Monday
Show September
that ≍ is an 14.
equivalence relation ononRRX X.
u(X
u(X )
)
hNote:!
! RR such
: u(X )Onesuch that
that v
such that
! Rdirection v == hh◦◦u.
u.
v = hwhile
is easy Show
Show
◦ u. the
Showthat
that ≍ ≍
thatis ≍
other isisan
an equivalence
equivalence relation
relationon
is an equivalence relation on R .
not. RX . .

These
These
results
These are
results
results
Exercise
uniqueness
are
are uniqueness
uniqueness
statements
statementsof
statements
sorts.
ofofofsorts.
sorts.
These results are uniqueness statements sorts.
notnot
literally
literallybecause
because there
thereareare
many
X many utility
utilityfunctions,
functions,rather
ratherthan
thana asingle
singleutility
utility
Define
notthe notbinary
literally relation
because
literally because ≍ on
there R
are
there by
many
are u
many≍ v
utilityif
utility there exists
functions,
functions, a strictly
rather
rather than
than increasing
a single
single utility
utility
h ◦ u.%.
function,
function, in in
R X X
RXthatX represent
that represent %.
h : u(X )!
function, R insuch
function, Rin that
that
R vrepresent
that= represent%. %. that ≍ is an equivalence relation on RX .
Show
There
There is is single
a a single equivalence
equivalence class,
class, under≍,≍,ofofutility
under utilityfunctions
functionsthat thatrepresents
represents
ThereThere is a single
is representation equivalence
a single equivalence class, under
class, under ≍,
≍,inofR of utility functions that represents
utility functions that represents
%.
%.%.The
The is therefore unique X XX / ≍.
%.
The representation
These The representation
representation
results isistherefore
therefore
are uniqueness unique
is thereforeunique
statements inin
unique ofinRR X/ ≍.
R / /≍.
sorts. ≍.
5
 Existence
Existence
Existence
Existence ofof
of
of aUtility
UtilityFunction
aaaUtility
Utility FunctionWhen
Function
Function When
When
WhenXXXXisisisFinite
Finite
isFinite
Finite
Theorem
Theorem
Theorem
SupposeXXisisfinite. Then%%isisa apreference
finite.Then preferencerelation
relationif ifand
Suppose
Suppose X Then % is a preference relation if and
and onlyif ifthere
only ifthere
only existssome
existssome
exists
there some
utility function u : X ! R that represents %.
utility function
utility function u : XX! !RR that represents%.%.
thatrepresents

• When
When the
the consumption
consumption setisisfinite
set finitethe theproblem
problem of of finding
finding a a utility
utility function
When the consumption
When consumption set set isisfinite
finite the
the problem
problem ofoffinding
findinga utility
a function
function
that that represents
represents some given some given preferences
preferences is easy, almost is easy,
trivial. almost function
utility trivial.
that represents some given preferences is easy,
that represents some given preferences is easy, almost trivial. almost trivial.
• TheTheproof
The proofisisisconstructive:
proof
constructive:
constructive: a a function
function that thatworks
worksis isthe the one
one that
that counts
counts thethe
The proof is constructive: aafunction functionthat thatworks
worksis isthetheone that
one counts
that countsthe the
number
number
number of of
of elements
elements
of elements that
that are
are not
not asas good
good asasthethe bundle
bundle in in question.
question.
number elements that that are
arenot notasasgoodgoodasasthe thebundle
bundle in in
question.
question.
• This This
function
Thisfunction
This is well
function defined
is well
function isis well
defined
welldefined
since
definedsince
sincethere
sincethe theare
are
the areonly
only
are only
only finitely
finitely
finitely
finitely
many
manybundles
many bundles
many bundles
bundles
thatcan
that canbebe
that can be
that worse
can be
worse
worse than
thanworse
than than something.
something.
something.
something.
In
• In other
InInother
other words,
otherwords,
words,the
words, the utility
theutility
the
utility
utility function
function
function
function
isis
isis
u(x)==|-|-(x)|
u(x) (x)|
u(x) = |- (x)|
where--(x)
where (x) ={y{y22XX: x: x%%y}y}isisthe
the lowercontour
contoursetsetofofx.x.
where - (x) = = {y 2 X : x % y} is thelower lower contour set of x.
Noticethat
thatweweonly
onlyneed
needtotoprove
provethat
thatthisthisfunction represents--(x)
functionrepresents (x)because
because
• Notice
Notice that we only need to prove that this isfunction represents - (x) because
from a previous
from aa previous result
previous result wewe know
know that
that if that
ififthat the
is isthe case
case then
then --- is a preference
is aispreference
from
relation. result we know that that the case then a preference
relation.
relation. 6
 Existence
Existence of of a Utility Function
a Utility Function When
WhenX isXFinite
is Finite
If X is finite and % is a preference relation ) 9 u : X ! R that represents %.
Proof.
Let u(x) = |- (x)|. Since X is finite, u(x) is finite and therefore well defined.
Suppose x % y. I claim this implies u(x) ≥ u(y).
Let z 2- (y), i.e. y % z.
By transitivity, x % z, i.e. z 2- (x). Thus - (y) ⊂- (x).
Therefore |- (y)| ≤ |- (x)|. By definition, this means u(y) ≤ u(x).
Now suppose x ≻ y. I claim this implies u(x) > u(y).
x ≻ y implies x % y, so the argument above implies - (y) ⊂- (x).
x % x by completeness, so x 2- (x). Also, x ≻ y implies x 2-
/ (y).
Hence - (y) and {x} are disjoint, and both subsets of - (x). Then
- (y) [ {x} ⊂ - (x)
|- (y) [ {x}| ≤ |- (x)|
|- (y)| + |{x}| ≤ |- (x)|
u (y) + 1 ≤ u (x)
u (y) < u (x)

x % y ) u(x) ≥ u(y)
This proves and thus we are done.
x ≻ y ) u(x) > u(y) 7
 ean
xprecisely
2X
be k . This
done function
in a finite u(x)
number will represent
or steps. % u(x)
Then define in X.
=k
bj is the j th element in B
Utility% Function
n u(x) will represent in X. for a Countable Space
⇤
2
The intuition iscthat we can rank x x > kDx
... max.i;
andj /so is max
Citi C eas
etilities
can rank xn) >x0 ; c ; : : : ; c ; : : : : n n
1 > ... nx and so it is easy to assign
1 1
Intuition:
u(x n we can(x
n 1
n rank
1 ) >x...
> ... > u(x1 ). This can be done
≻ x
>
n n−1 ...1≻task
xrational
1 and sopre↵erences
it isaeasy towhen
assign
1 u(x ). This can be done for any rational p
≻any
forThe of finding listing one or both o
f the
ure, weutilities
set
but also u(x
defined if an)function
is finite
the >set
(xn−1
for is )countable.
>f...but
sure, > u(x
on the
also).nonnegative
if theatset
1problems the isintegers
endcountable. by the rule
of the chapter.
then
f X
f would
is
be a(e.g.
countable
bijection
N or
from
Q),
N to Cany
then
. More formally,
preference relationspace.
on X h
N• orExistence
Q), then of a utility
any function
preference can also
relation on be
X proven
has a for a countable
utility
epresentation.
1. AExercise
set C is4. countably infinite(e.g.
If X is countable iff bijQCIs
N or
13.2.4 ),. then
A setany
is countable
Countable preferenceiff
Microeconomic Theory
ntably infinite.
relation on X has
Econ a utility representation.
711-UW-Madison
1
Theorem 13.2.3 also has By:a Lones SmithCorollary
surprising &Tono C
elements
Proof:
Proof: Takein X
Take such
any
CHAPTER
any that
listlist3 of
of {x n }n=1 =
elements
-elements
Consumer inX.
inXXFor
Theory is the
such element
that
that {xnx}11 Updated: October 19
numbers countable. n=1 = X. For the
Then for x2 assign
nd For x1, assign
assign u(x1 ) u(x
= 0. 1) =Then
0. Then,
for for x2 assign
x2 assign
8 Corollary 13.2.4. The set of rational numbers Q
2 ⇠assign
< 0Takeif x3xand x1 8
u(x2 ) = 1 if x2 x1 < 0 if x2 ⇠ x1
: 8
1 if x1 x u(x
2 2) =
>
> 01 ifif x2 xx11 ⇠ x3
>
> :
< u(x2 )1 ifif x1 xx22 ⇠ x3
Then, for x3 assign: u(x3 ) = u(x2 )/2 if x1 x3 x2 or x2 x3 x1
>
>
3 >
> 2 if x1 % x2 x3
:
So on and so forth... 2 if x3 x2 % x1
5
QED. 4 so on and so forth... . 8
 Lexicographic Preferences
Lexicographic
Lexicographic Preferences
Preferences
• We want conditions on a preference relation which guarantee the
We existence
want
We want conditions
of a utilityon
conditions aa preference
preference
function
on representingrelation
relation which
those
which guarantee
preferences
guarantee the even the
if X exi
existence of
isanot
a utility countable.
function
utility functionrepresenting those
representing those preferences
preferences even even
if X is ifnot
X countable.
is not coun
n
Typically,
Typically,
•Typically, wewe assumethat
weassume
assume that
that XXXlives
lives
in R
lives in ,, but
in ,Rbut but ideally
ideallywewewould
wouldwant
n ideally we would want as few
want as
restrictions on X as we can get away with.
as few restrictions
restrictions on X as on we Xcanas wegetcanawaygetwith.
away with.
We know that without completeness and transitivity a utility function does not
• We
We know know
exist, sothat that without
without
we cannot
completeness
completeness
dispense
and transitivity a utility function
of those. and transitivity a utility function
does
exist, so not
we exist, so we cannot dispense
of those.of those.
What elsecannot dispense
is needed?
• What
What elseelse
Next,
is needed? of the existence of a representation.
a iscounterexample
needed?
• Next, a counterexample of the existence of a representation.
Next, a counterexample of the existence of a representation.
Exercise 53,(Part
Question of HW3)
Problem Set 2 (due Monday September 14).
Show that the lexicographic ordering on R2 defined by
uestion 3, Problem Set 2 (due Monday 8 September 14).
< x1 > y1
(x1 , x2 ) % (yordering
ow that the lexicographic 2
1 , y2 ) ,on R defined or by
:8
x1 = y1 and x2 ≥ y2
< x1 > y1
is complete, transitive, and antisymmetric (i.e. if x % y and y % x, then x = y ).
(x1 , x2 ) % (y1 , y2 ) , or
: 9
x1 = y1 and x2 ≥ y2
m0)
n!1and
xny==(0,
lim0)n!1
and
13.2.3
countable,
yy =tocountable,
nCross
we need
(0,
lim 1).
Products
find
n!1
a listing
yn =
we need
(0, 1).
of thetoelements
find a listing of the elements
d. Lexicographic
Somewhat Preferences
surprisingly, cross products 13.2.3 Cross
of countable setsProducts
are also countable. At first,
f .a; b/ j a 2 A; b 2 fB.a; g: b/ j a 2 A; b 2 B g:
you might be tempted to think that “infinity times infinity” (whatever that means)
Exercise 6: Show that somehow
lexicographic
results in a
preferences
larger infinity,
Somewhat
but this
on is
[0,
not
1]case.
the
× [0,cross
surprisingly, 1] do not of countab
products
do 1] ⇥admit
[0, not [0, 1] do
a utility
not admit
admit a utility There
functionfunction
a utility function
are many suchThere listings.
representation. are many
One issuch shownlistings.
you in might
Figure
One be13.1
istempted
shown
for the in case
to Figure
think when 13.1
that Afor thetime
“infinity cas
and B Theorem
are both infinite andsets.
13.2.3. B are
The Inboth
cross this infinite
listing,of
product sets.
somehow
.a jIn/ this
isresults
i ; bcountable
two thelisting,
in aelement
kth
sets islarger /inisthe
thelist
i ; bjinfinity,
.acountable. but element
kththis is no
Proof: Let’s assume a utility function
for C where for C whereu exists for contradiction.
Theorem 13.2.3. Thethat cross
Proof. Let A and B be any pair of countable sets. To show C product
D A ⇥ Bofistwo count
also
For any x ∈ [0,1], we
tion
thatsoa assume
utility function, haveweu(x,1)
that a utility
countable, u, exists. >find
ai istothe
need u(x,0).
function, u, exists.
i tha element
listing ofainithe
is
A,the i th element in A,
elements
Proof. Let A and B be any pair of countable sets
b is the j th elementbcountable,in
j isB,theand j thwe element toinx. B, and
Define I(x) = [u(x,0),u(x,1)] ⊆ j , an interval
u(x,
I(x)0).= Define
[u(x, 0),I(x)
u(x,=1)] [u(x,✓ R, 0),an u(x, 1)] ✓ R, an
f .a;
2
b/ j in
a 2 A; ,b for
2 B 2
each
need
g: find a listing of the elemen
k D max.i; j / C i C kD max.i
max.i;j; j /0/:C i C max.i j; 0/:
for
I(x) any\ I(y)
x 6= = ;, since
y I(x)
Note that for any xThere \ I(y)
≠ y I(x) u(x,=0);,>since u(x, 0) >
∩ I(y)
are many such= listings.
∅, since Oneu(x,0)
is shown > u(y,1)
in Figurewhenever 13.1 for the f .a;xb/> jwhen
case y.a 2 A;A b 2
The task
andofBfinding
are bothaThe
listing
taskwhen
infinite of finding
sets. one ora both
In this listing ofwhen
listing, A.aand oneB or areboth
finiteofisAleft andtoBtheare finite is
i ; bj / is the kth element in the list
function
is an injection
I(x) isfrom
Then, the function an injection
problems
for
the realsfrom
I(x)atC the
is an
endinjection
where of
problemsin the reals in
at from
the chapter. the end the reals
There
of the are in
chapter. many[0,1] suchto the space
listings. One is shown in
and B are both infinite sets. In this ⌅ listing, .ai ;
formed by these intervals.
tervals.
se, define Likewise,
the functiondefineQ(I(x))the function : Q(I(x)) : for C in where
ai is the i th element A,
x).Define
the thetheir
interval
From function
13.2.4
I(x). Q (I(x))
disjointness,
From :13.2.4
I(·) →
Is Countable
their this that
bjQisIs
disjointness,
the takes
Countable
j th element any rational
this
in B, and ini isthe
a the iinterval
th element in A,
I(x).
sthese
to theintervalsTheorem
rationals.to 13.2.3
the alsoTheorem
Therefore has a we
rationals. surprising
k 13.2.3 also
Corollary;
j /hasCai surprising
Therefore
D max.i; 2
Cnamely
we
max.i that
Corollary;
theb set
j; 0/: namely
of rationalthat the set o
j is the j th element in B,
numbers is countable.numbers is countable.
From
0,onals their
1] towhich disjointness,
the rationals
implies |[0,this
which
The task of 1]| is
 aalso
listingan
|Q|
implies
finding wheninjection
|[0, 1]|
one or  from
both|Q|
of A and the
B karespace
max.i;
D finite isof/2 these
jleft C C max
to ithe
Corollary 13.2.4. TheCorollary
set of the13.2.4.
of rational numbers
The set Qof is
rational
countable.
numbers Q is countable.
ble intervals
1] iswhile to the rationals.
problems
Q is justwhile
uncountable at the end
Q is just⇤countable.
countable. chapter.
The task of⇤ finding a listing when one or both o
⌅
Therefore, we have constructed an injection problems from [0,at 1]
the to
end the
of therationals
chapter.
which implies |[0, 1]| ≤ | Q|.
13.2.4 Is Countable
6 A contradiction because
Theorem[0,1]
13.2.3isalso
uncountable while
has a surprising13.2.4 Q Isis
Corollary; just that
countable.
Countable
namely QED.
the set of rational
numbers is countable.
Theorem 13.2.3 also has a surprising Corollary
10
Corollary 13.2.4. The set5of rational numbers
numbers 5isQcountable.
is countable.
 Continuous Preferences
Microeconomic Theory
Continuous Preferences
Econ 711-UW-Madison By: Lones Smith &To
y
Continuity will get rid
CHAPTER 3 - of this example.
Consumer Theory Updated: Octo
• Continuity will get rid of this example.
Definition
conomic Theory
• The upper contour sets for lexicographic preferences is depicted in Fi
A binary relation % on the metric space XBy:
11-UW-Madison Lones Smith
is continuous if, &Tono
for all xCarrasco
2 X , the upper
isoquant is just a point; the upper contour set 19,is not closed, and
TER
and3 lower
- Consumer sets, {y 2 X : y % x} and {y 2Updated:
contourTheory X : x % October 2015
y}, are closed.
continuous.
A set is closed if
Examples 1 y
Consumer Theory -I
• its complement is open
U (u0 )
8
< x1 > y1
• it producer
Unlike contains
Define % all of
theory,
by (x its
consumer
, x limit
) % (ypoints
theory
, y ) is ,
an ordinal theory.
or Everything is about
omparingy0alternatives. 1 2 1 2
:
x1 = y1 and x2 ≥ y2
Lexicographic preferences:
When comparingDrawalternatives, constraints
the upper contour set of will matter
(1, 1); and thus on
this preference R2 is not
trade-o↵s arise.
continuous.
Each utility “curve” (set) is 8 U(u0 )
< xmain
1 ≥ yapproaches:
a point.
To model individual decision making there are two
Define % by (x1 , xx
1
2 ) % (y1 , y2 ) , and
-Upper contour
The first one, a set is not
0
preference-based y0
approach, :assumes thatxconsumer has some
x2 ≥ y2
closed and,
preference
Figure
Draw2:thus,
relation
upperiscontour set
% (i.e the primitive)
Lexicographic
the Preferences over a given set of choices X. This
of (1, 1); this preference on R2 is continuous.
preference relation % has to satisfy some axioms of rational choice.
not continuous.
), -UThe are
are bundles
(u0 )secondallone, (x,y):
a choice-based
bundles approach,
(x, y) such that focuses
u(x, y)xon
0
the
u(xconsumer’s
0 , y 0 ) = u choice
0 . x
behavior and takes observables choices as the primitives. When this choices
u(x, y) ≥some
satisfy u(x0consistency
, y0) = u0. restrictions, they are equivalent to the axioms of ra-
11
Figure providing
Curves: A set of consumption bundles 2: Lexicographic Preferences
the same level of utility
 Continuous Preferences
Continuous Preferences
Definition
A binary relation % on the metric space X is continuous if, for all x 2 X , the upper
and lower contour sets, {y 2 X : y % x} and {y 2 X : x % y}, are closed.

Proposition
A binary relation % is continuous if and only if:
1 If xn % y for all n and xn ! x, then x % y; and
2 If x % yn for all n and yn ! y, then x % y.

Proof.
This follows from the fact that a set is closed if and only if it contains all of its
limit points.

12
 Continuous Preferences
Continuous Preferences

◦ Intuitively, there can be no sudden jumps (i.e.,

preference reversals) in an individual preference
over a sequence of bundles.

Advanced Microeconomic Theory 87

13
 Continuous Preferences
Example: Lexicographic preferences are not continuous
⎛1 ⎞
Consider the sequence n
x = ⎜ ,0 ⎟ and yn = (0,1), where n={1, 2,…}
⎝n ⎠
The sequence yn = (0,1) is constant in n.
⎛1 ⎞
Continuousis not:
The sequence Preferences
n
x = ⎜ ,0 ⎟
⎝n ⎠ ⎛1 ⎞ ⎛1 ⎞
Starts at = (1,0), and moves leftwards to
x1 x 2 = ⎜ ,0 ⎟ , x 3 = ⎜ ,0 ⎟ , …
Thus, the individual ⎝2 ⎠ ⎝3 ⎠
Thus, the DM prefers:
prefers: x
1 1
2 y n, n , y 1 = y 2 = = yn
x 𝑥 1≻=(0,1)
= (1,0) 1,0= y≻ 0,1 = 𝑦1 1

2 1 𝑥 2 = 1 , 0 ≻2 0,1 = 𝑦 2
x = ( ,0) ≻ (0,1) 2 = y
2 3 1
𝑥 = , 0 ≻ 0,1 = 𝑦 3
1 3
x = ( ,0) ≻ (0,1) = y⋮3
3

3 lim x n = (0,0)
!
But, n

𝑛 = 0,0 ≺ 0,1
lim 𝑥
𝑛→∞ ≺ (0,1) = lim y n
lim x n = (0,0)
n→∞ = lim 𝑦 𝑛 n→∞
𝑛→∞ x4 x3 x2 x1
14
Preference reversal!
Preference reversal! 0 ¼ ⅓ ½ 1 x1
 Debreu’s Representation
Debreu’s Representation Theorem
Theorem
Debreu’s Representation Theorem
TheGerard Debreu theorem provides necessary
Debreu, and sufficient
•The main result of today is due to Gerard Debreu, and it provides conditions for
necessary and
main result of today is due to Gerard and it provides necessary and
the existence
su¢cient of
conditions a continuous
for utility
the existence of a function.
continuous utility function.
su¢cient conditions for the existence of a continuous utility function.
Theorem (Debreu)
Theorem (Debreu)
Suppose X ⊂ Rnn . The binary relation % on X is complete, transitive, and
Suppose X ⊂
continuous R . only
if and The ifbinary exists a %
thererelation on X is complete,
continuous transitive, and
utility representation u : X ! R.
continuous if and only if there exists a continuous utility representation u : X ! R.

We will prove su¢ciency next (under a couple of simplifying assumptions, but

• We
youwill
We willprove
prove
should sufficiency
su¢ciency
ckeck next(under
next
what happens under aa couple
without couple
those).ofofsimplifying
simplifying assumptions
assumptions, but
you should ckeck what happens without those).
You will do necessity as homework.
You will do necessity as homework.
• You will do necessity as homework (the easier direction.)
Question 4, Problem Set 2.
Exercise
Question7:4, (Part
n of HW3)
Problem Set 2.
Suppose X ⊂ R . Prove that if u : X ! R is a continuous utility function
n
representing %, then
Suppose X ⊂ R % isthat
. Prove if u : X !
a complete, R is a continuous
transitive, utilitypreference
and continuous function relation.
representing %, then % is a complete, transitive, and continuous preference relation.

15
 Debreu’s
Debreu’s Representation Theorem
Representation Theorem
Theorem (Debreu)
Suppose X ⊂ Rn . The binary relation % on X is complete, transitive, and
continuous if and only if there exists a continuous utility representation u : X ! R.

Two simplifying assumptions:

X = Rn ; and
% is strictly monotone, (i.e. if xi ≥ yi for all i and x 6= y, then x ≻ y).
When strict monotonicity holds we have
α ≥ β , αe % βe. (mon)
where e = (1, 1, . . . , 1) (make sure you check this).

To prove:
if % is a continuous and strictly monotone preference relation on Rn , then
there exists a continuous utility representation of %.
How do we find a utility function? Look at the point on the 45o line that is
indi§erent:
u(x) = α∗ (x) where α∗ (x) is the real number α∗ such that α∗ e ∼ x
The proof is in 3 steps. 16
 Step 1: There exists a unique α∗ (x) such that α∗ e ∼ x.

Proof.
Let B = {β 2 R : βe % x} ⊂ R and define
α∗ = inf {β 2 R : βe % x} .
| {z }
B

Oviously, (maxi xi )e ≥ x, so by strict monotonicity (maxi xi )e % x which

implies B is nonempty and α∗ is well-defined.
Now show that α∗ e % x and α∗ e - x, so that α∗ e ∼ x.
Since α∗ is the infimum of B , there exists a sequence β n 2 B s.t. β n ! α∗ .
Then β n e ! α∗ e (in Rn ) and β n e % x because β 2 B .
By continuity, one gets: α∗ e % x as desired.
Since α∗ is a lower bound of B , if α 2 B ) α ≥ α∗ .
Using the contrapositive:
α < α∗ ) αe ≺ x. (A)
Let αn = α∗ − n1 . By (A), αn e ≺ x, so αn e - x.
Also, αn e ! α∗ e (in Rn ). Hence, by continuity, α∗ e - x as desired.
To prove uniqueness, suppose α̂e ∼ x.
By transitivity, α̂e ∼ α∗ e. Then, by (mon), α̂ = α∗ .
17
Step 2: The u(x) that represents % is defined as follows:
u(x) = α∗ (x) where α∗ (x) is the unique α∗ such that α∗ e ∼ x.

Proof.
Show that u(x) represents %.
Suppose x % y.
By construction of α∗ , we have:
x ∼ α∗ (x)e and α∗ (y )e ∼ y
By transitivity, we have:
x ∼ α∗ (x)e % α∗ (y)e ∼ y
By (mon) this is equivalent to
α∗ (x) ≥ α∗ (y)

Repeat the same argument to show that x ≻ y implies α∗ (x) > α∗ (y).
x % y ) u(x) ≥ u(y)
Since we have shown that we are done.
x ≻ y ) u(x) > u(y)

18
 Step 3: The defined u(x) is continuous.
u(x) = α∗ (x) where α∗ (x) is the unique α∗ such that α∗ e ∼ x

Proof.
To prove that a function f : Rn ! R is continuous, it su¢ces to show that
f −1 ((a, b)) is open for all a, b 2 R.
Since ae ∼ ae, we have u(ae) = α∗ (ae) = a for any a 2 R.
Notice that
u −1 ((a, b)) = u −1 ((a, 1) \ (−1, b)) = u −1 ((a, 1)) \ u −1 ((−1, b)).

But u(ae) = a, so
u −1 ((a, 1)) = u −1 ((u(ae), 1)) = {x 2 Rn : x ≻ ae},
this is open because the strict upper contour set of ae is open whenever % is
complete and continuous (why?).
An entirely symmetric argument proves that u −1 ((−1, b)) is the strict lower
contour set of be, hence it is also an open set.
Since u −1 ((a, b)) is the intersection of two open sets, it is open.
19
 Debreu’s Representation
Representation Theorem Theorem
Debreu’s
Debreu’s
The main result Representation
of today is due
Back to the theorem
to Gerard Debreu, andTheorem
it provides necessary and
Back to the theorem
•su¢cient
Back toconditions
theoremfor the existence of a continuous utility function.
Theorem (Debreu)
Theorem
Theorem (Debreu)
(Debreu)
Suppose X ⊂ Rnnn. The binary relation % on X is complete, transitive, and
Suppose
Suppose XX⊂ ⊂R R .. The
The binary relation %
binary relation % on
on XX isis complete,
complete, transitive,
transitive, and
and
continuous if and only if there exists a continuous utility representation u : X ! R.
continuous
continuous ifif and
and only
only ifif there
there exists
exists aa continuous
continuous utility
utility representation
representation uu ::XX !
!R.
R.
Conclusion
Conclusion
We willher
prove su¢ciency nextis(under a couple of simplifying assumptions, but
When preference relation complete, transitive, and continuous a
When her preference
you should ckeck whatrelation
happensis complete, transitive, and continuous a
without those).
consumer’s
consumer’s
You will
that
behavior
do behavior
necessity
represents
is entirely
isasentirely
described Existence of
by some continuous utility function
described by some continuous utility function
homework.
that preference relation.
that represents that preference relation.
◦ And by monoto
•Question
The 4, Problem
Thetheorem
theorem asserts
assertsSet 2. one of the utility representations for % must be
that
that one of the utility representations for 𝑥≿𝑦 must
⟺ be 𝑡 𝑥
The theorem n asserts that one of the utility representations for % must be
Suppose X ⊂ R not
continuous, . Prove
that that ofif the
u:X is a continuousmust
! Rrepresentations utility
befunction
continuous, notnot that all
continuous, all of the utility
utility representations must be continuous.
continuous.
that all of the utility representations must be continuous.
%, is % is a complete,
•representing
Continuity
Continuity then
is aa cardinal
cardinal feature
feature oftransitive,
of
the the
utility and continuous
utility not anpreference
function,
function,
Continuity is a cardinal feature of the utility function, not an ordinal not an
ordinal
◦ Note:
relation.
Aordinal
feature,
utility fu
feature,
since
feature,it is not robust to strictly increasing transformations.
since it since
is not itrobust
is nottorobust
strictlytoincreasing
strictly increasing transformations.
transformations. but still be non-
Exercise For instance, th
Exercise
Construct a preference relation on R that is not continuous, but admitsmin{𝑎𝑥
a utility1 ,𝑏𝑥2 },
Construct a preference relation on R that is not continuous, but admits differentiated
a utility a
representation.
representation.
20
 Choice Rules
Choice Rules and
and Utility
Utility Functions
Functions
The induced choice rule for % is C% (A) = {x 2 A : x % y for all y 2 A}

Proposition
If % is a continuous preference relation and A ⊂ Rn is nonempty and compact,
then C% (A) is nonempty and compact.

Proof.
Suppose % is continuous.
By Debreu’s Theorem, there exists some continuous function u representing
the preferences.
One can show (do it as exercise) that
C% (A) = arg max u(x).
x2A
Nonemptiness then follows from continuity of u and the Extreme Value
Theorem.
Compactness follows from the fact u −1 (·) is bounded (it is a subset of the
bounded set A), and closed (since the inverse image of a closed set under a
continuous function is closed). 21