You are on page 1of 25

Microeconomic Analysis (806L1)

Topic 1: Preference and Utility

Matthew Embrey

University of Sussex

Autumn 2023

Matthew Embrey (University of Sussex) Microeconomic Analysis (806L1) Autumn 2023 1 / 13


Rational Choice
Choice problem

Economic agent (consumer) faces a choice problem; she has to


choose among alternatives.
Rational choice:
What is desirable? (a ranking of alternatives)
What is feasible? (a budget set)
Pick the most desirable from the feasible alternatives.

Matthew Embrey (University of Sussex) Microeconomic Analysis (806L1) Autumn 2023 2 / 13


Rational Choice
Choice problem

Economic agent (consumer) faces a choice problem; she has to


choose among alternatives.
Rational choice:
What is desirable? (a ranking of alternatives)
What is feasible? (a budget set)
Pick the most desirable from the feasible alternatives.
Note: an agent’s desires are not shaped by the set of feasible
alternatives.

Matthew Embrey (University of Sussex) Microeconomic Analysis (806L1) Autumn 2023 2 / 13


Rational Choice
Preferences

Let X be a set of alternatives; e.g.


Finite: X = {x1 , ...., xn } such as {Cancun, Hawaii, Madrid, ...}
Uncountable: X ⊂ Rn such as consumption bundles (apples/pears)

Matthew Embrey (University of Sussex) Microeconomic Analysis (806L1) Autumn 2023 3 / 13


Rational Choice
Preferences

Let X be a set of alternatives; e.g.


Finite: X = {x1 , ...., xn } such as {Cancun, Hawaii, Madrid, ...}
Uncountable: X ⊂ Rn such as consumption bundles (apples/pears)
Preferences over a set of alternatives X is a binary relation ≿ (“at
least as preferred as”) on X satisfying:
Complete: For any x, y ∈ X , x ≿ y , or y ≿ x.
Reflexive: For any x ∈ X , x ≿ x.
Transitivity: For any x, y , z ∈ X , if x ≿ y and y ≿ z, then x ≿ z.
Often referred to collectively as Ordering
Note: if x ≿ y and not y ≿ x, then x ≻ y (strict preference).
Note: if x ≿ y and y ≿ x, then x ∼ y (indifference).

Matthew Embrey (University of Sussex) Microeconomic Analysis (806L1) Autumn 2023 3 / 13


Rational Choice
Preferences

Let X be a set of alternatives; e.g.


Finite: X = {x1 , ...., xn } such as {Cancun, Hawaii, Madrid, ...}
Uncountable: X ⊂ Rn such as consumption bundles (apples/pears)
Preferences over a set of alternatives X is a binary relation ≿ (“at
least as preferred as”) on X satisfying:
Complete: For any x, y ∈ X , x ≿ y , or y ≿ x.
Reflexive: For any x ∈ X , x ≿ x.
Transitivity: For any x, y , z ∈ X , if x ≿ y and y ≿ z, then x ≿ z.
Often referred to collectively as Ordering
Note: if x ≿ y and not y ≿ x, then x ≻ y (strict preference).
Note: if x ≿ y and y ≿ x, then x ∼ y (indifference).
We exclude:
An inability to compare e.g. “I don’t know”
A dependance on other factors e.g. “Depends on what my friend says”
Intensity of preferences e.g. “I slightly prefer x”.
Key idea: Can always compare and rank two alternatives from a
choice set, and this ranking is transitive.
Matthew Embrey (University of Sussex) Microeconomic Analysis (806L1) Autumn 2023 3 / 13
Rational Choice
Examples

Spring break locations


Finite: X = {Cancun, Hawaii, Madrid}
≿ is effectively a matrix of zeros and ones
For example ... M ≻ H and H ≻ C

Matthew Embrey (University of Sussex) Microeconomic Analysis (806L1) Autumn 2023 4 / 13


Rational Choice
Examples

Spring break locations


Finite: X = {Cancun, Hawaii, Madrid}
≿ is effectively a matrix of zeros and ones
For example ... M ≻ H and H ≻ C
Consumption bundles
Uncountable: X ⊂ Rn
n is the number of commodities in a consumption bundle
An alternative x ∈ X is a list of quantities (i.e. a vector) of each
commodity: x = (x1 , x2 , ..., xn )
Typically agent only buys/consumes commodities rather than possibly
selling them, so xi ≥ 0 (although for more general analysis this could
be generalised)

Matthew Embrey (University of Sussex) Microeconomic Analysis (806L1) Autumn 2023 4 / 13


Rational Choice
Examples

Spring break locations


Finite: X = {Cancun, Hawaii, Madrid}
≿ is effectively a matrix of zeros and ones
For example ... M ≻ H and H ≻ C
Consumption bundles
Uncountable: X ⊂ Rn
n is the number of commodities in a consumption bundle
An alternative x ∈ X is a list of quantities (i.e. a vector) of each
commodity: x = (x1 , x2 , ..., xn )
Typically agent only buys/consumes commodities rather than possibly
selling them, so xi ≥ 0 (although for more general analysis this could
be generalised)
For illustrative examples, simplify to n = 2 and X is the positive
orthant of a 2-D graph

Matthew Embrey (University of Sussex) Microeconomic Analysis (806L1) Autumn 2023 4 / 13


Rational Choice
Examples

Spring break locations


Finite: X = {Cancun, Hawaii, Madrid}
≿ is effectively a matrix of zeros and ones
For example ... M ≻ H and H ≻ C
Consumption bundles
Uncountable: X ⊂ Rn
n is the number of commodities in a consumption bundle
An alternative x ∈ X is a list of quantities (i.e. a vector) of each
commodity: x = (x1 , x2 , ..., xn )
Typically agent only buys/consumes commodities rather than possibly
selling them, so xi ≥ 0 (although for more general analysis this could
be generalised)
For illustrative examples, simplify to n = 2 and X is the positive
orthant of a 2-D graph
In addition to the ordering axioms will typically assume that
consumers’ preferences over consumption bundles also satisfy
continuity, monotonicity and convexity
Matthew Embrey (University of Sussex) Microeconomic Analysis (806L1) Autumn 2023 4 / 13
Consumption Bundles
Indifference curves

≿ is now a much “richer” object


than in the simple finite case
Indifference curves provide a
graphical depiction of preferences.
Given a bundle x, the indifference
set (curve) is the set of all
consumption bundles y ∈ X such
that y ∼ x.
Upper contour (preferred) set of
x is the set of all bundles y ∈ X
such that y ≿ x
Lower contour set of x is the set
of all bundles y ∈ X such that
x ≿y
Matthew Embrey (University of Sussex) Microeconomic Analysis (806L1) Autumn 2023 5 / 13
Consumption Bundles
Continuity of Preferences

Continuous preferences
Preferences do not exhibit
sudden “jumps”
Formally, the preference
relation ≿ on X is continuous
if for any sequence of pairs
(x k , y k )∞ k k
n=1 with x ≿ y for
all k, x ≿ y where
limk→∞ x k = x and
limk→∞ y k = y
Equivalently, upper contour
sets are closed sets.
Technical condition
Continuity rules out ”holes” in
the indifference curve
Ensures a continuous utility
function (later)
Matthew Embrey (University of Sussex) Microeconomic Analysis (806L1) Autumn 2023 6 / 13
Consumption Bundles
Monotonicity of Preferences

Monotonicity
A preference relation ≿ on X
is strongly monotone if for
every x, y ∈ X such that
x ≥ y (i.e. xi ≥ yi for all
i = 1, 2..., N goods in the
bundle) and x ̸= y (there is at
least one good j such that
xj > yj ), then x ≻ y .
Key idea: more is better
Rules out ‘thick’ indifference
curves
Implies negatively sloped
indifference curves
(Technical aside: weaker local
non-satiation often sufficient)
Matthew Embrey (University of Sussex) Microeconomic Analysis (806L1) Autumn 2023 7 / 13
Consumption Bundles
Convexity of preferences

Convexity
Consumers prefer to diversify -
mixture over extremes.
Formally, the preference
relation ≿ on X is convex if
for every x ∈ X , the upper
contour set is convex i.e. if
y , z ≿ x, then
αy + (1 − α)z ≿ x for any
α ∈ [0, 1].
In Fig (1), the dashed line which
is the convex combination of y
and z where y , z ≿ x does not
lie in the upper contour set -
preferences are not convex.
Matthew Embrey (University of Sussex) Microeconomic Analysis (806L1) Autumn 2023 8 / 13
Utility Maximisation
Preference Maximisation verus Utility Maximisation

Rational choice by our agent is associated with maximising their


preferences given what is feasible
Choose an x ∈ X such that x ≿ y for all y ∈ X

Matthew Embrey (University of Sussex) Microeconomic Analysis (806L1) Autumn 2023 9 / 13


Utility Maximisation
Preference Maximisation verus Utility Maximisation

Rational choice by our agent is associated with maximising their


preferences given what is feasible
Choose an x ∈ X such that x ≿ y for all y ∈ X
This is preference maximisation
While an appealing definition of rational behaviour, often not
convenient for analysis

Matthew Embrey (University of Sussex) Microeconomic Analysis (806L1) Autumn 2023 9 / 13


Utility Maximisation
Preference Maximisation verus Utility Maximisation

Rational choice by our agent is associated with maximising their


preferences given what is feasible
Choose an x ∈ X such that x ≿ y for all y ∈ X
This is preference maximisation
While an appealing definition of rational behaviour, often not
convenient for analysis
Suppose there exists a utility function U : X → R that represents the
agents preferences, then
Choose x ∈ X to maximise U (x) subject to x ∈ X
This is utility maximisation
When is this equivalent to preference maximisation?

Matthew Embrey (University of Sussex) Microeconomic Analysis (806L1) Autumn 2023 9 / 13


Utility Maximisation
Utility representation of preferences

A function U : X → R represents the preference ≿ if for all x, y ∈ X ,


x ≿ y if and only if U(x) ≥ U(y ).
If U represents the preference relation ≿, we call U a utility function,
and we say that ≿ has a utility representation.
Note: utility functions are numerical and make optimisation more
convenient.

Matthew Embrey (University of Sussex) Microeconomic Analysis (806L1) Autumn 2023 10 / 13


Utility Maximisation
Utility representation of preferences

A function U : X → R represents the preference ≿ if for all x, y ∈ X ,


x ≿ y if and only if U(x) ≥ U(y ).
If U represents the preference relation ≿, we call U a utility function,
and we say that ≿ has a utility representation.
Note: utility functions are numerical and make optimisation more
convenient.
If a preference relation has a utility representation, it has infinite other
such representations:
If U represents ≿, then for any strictly increasing function f : R → R,
the function V (x) = f (U(x)) represents ≿ as well.
Key idea: utility representations are ordinal (as underlying preferences
are rankings), and therefore any monotonically increasing
transformations also represent the same preferences.

Matthew Embrey (University of Sussex) Microeconomic Analysis (806L1) Autumn 2023 10 / 13


Utility Maximisation
When can we represent preferences by a utility function?

In the finite case,


If preferences are complete, reflexive and transitive, there exists a utility
function representation

Matthew Embrey (University of Sussex) Microeconomic Analysis (806L1) Autumn 2023 11 / 13


Utility Maximisation
When can we represent preferences by a utility function?

In the finite case,


If preferences are complete, reflexive and transitive, there exists a utility
function representation
In the consumption bundles case,
If preferences are complete, reflexive and transitive, and continuous,
there exists a continuous utility function representation

Matthew Embrey (University of Sussex) Microeconomic Analysis (806L1) Autumn 2023 11 / 13


Revealed Preference
Preferences and Utility

So far, two concepts for our consumer’s behaviour


Let A ⊂ X
Preference maximisation: choose x ∈ A such that

x ≿y for all y ∈ A

Utility maximisation: choose x ∈ A such that

u (x) ≥ u (y ) for all y ∈ A

... (simplifying to X finite) our utility representation theorem


essentially tells us that
preference maximisation ⇐⇒ utility maximisation
as long as ≿ satisfies completeness, reflexivity and transitivity.

Matthew Embrey (University of Sussex) Microeconomic Analysis (806L1) Autumn 2023 12 / 13


Revealed Preference
Preferences and Utility

So far, two concepts for our consumer’s behaviour


Let A ⊂ X
Preference maximisation: choose x ∈ A such that

x ≿y for all y ∈ A

Utility maximisation: choose x ∈ A such that

u (x) ≥ u (y ) for all y ∈ A

... (simplifying to X finite) our utility representation theorem


essentially tells us that
preference maximisation ⇐⇒ utility maximisation
as long as ≿ satisfies completeness, reflexivity and transitivity.
However, we do not directly observe preference relations.
And it seems unlikely consumers actively make utility calculations.
Matthew Embrey (University of Sussex) Microeconomic Analysis (806L1) Autumn 2023 12 / 13
Revealed Preference
Choice Functions

Instead we observe choice.


Consumer faces a menu A ⊂ X and chooses C (A) ⊂ A.
C is called a choice rule.
Weak Axiom of Revealed Preference (WARP)
Let x, y ∈ X and A, B ⊂ X that both contain x and y . Then, a choice
rule satisfies WARP if

x ∈ C (A) and y ∈ C (B) =⇒ x ∈ C (B)

Revealed Preference
Define the preference relations ≿R by x ≿R y if x ∈ C ({x, y })
Revealed preference satisfies completeness, reflexivity and transitivity
(“rational” preferences) if and only if it satisfies WARP.

Matthew Embrey (University of Sussex) Microeconomic Analysis (806L1) Autumn 2023 13 / 13


Revealed Preference
Choice Functions

Instead we observe choice.


Consumer faces a menu A ⊂ X and chooses C (A) ⊂ A.
C is called a choice rule.
Weak Axiom of Revealed Preference (WARP)
Let x, y ∈ X and A, B ⊂ X that both contain x and y . Then, a choice
rule satisfies WARP if

x ∈ C (A) and y ∈ C (B) =⇒ x ∈ C (B)

Revealed Preference
Define the preference relations ≿R by x ≿R y if x ∈ C ({x, y })
Revealed preference satisfies completeness, reflexivity and transitivity
(“rational” preferences) if and only if it satisfies WARP.
Consistent choice ⇐⇒ rational preference maximisation ⇐⇒ utility
maximisation

Matthew Embrey (University of Sussex) Microeconomic Analysis (806L1) Autumn 2023 13 / 13

You might also like