Microeconomics

Introduction

GATE Lyon Saint-Etienne, Université de Saint-Etienne

Microeconomics 1 / 25
Plan of the course

This course contains three main parts plus an introduction

1 Social choice theory
2 Mechanism design
3 Matching
Textbooks
Danilov V., Stoskov A. 2002. Social Choice Mechanisms,
Springer.
Moulin H. 1983. The Strategy of Social Choice, North
Holland Publishing Cie, Amsterdam.
Roth A., Sotomayor M. 1990. Two-Sided Matching: A Study
in Game-Theoretic Modeling and Analysis, Econometric
Society Monographs. Cambridge University Press.
Evaluation: final exam.

Microeconomics 2 / 25
Plan of the introduction
Contents:
1 Social choice theory and its axiomatic method

2 Social states, preferences, social goals and mechanisms

3 The Mount-Reiter diagram

4 An example: King Solomon’s dilemma

5 Matching theory and Social choice theory

Recommended reading
Maskin E., 2007. Mechanism design: how to implement social
goals. Nobel Prize Lecture.
Roth A. 2012. The theory and practice of market design.
Nobel Prize Lecture.
Shapley L.S, 2012. Allocation games – the deferred
acceptance algorithm. Nobel Prize Lecture.
Thomson W., 2001. On the axiomatic method and its recent
applications to game theory and resource allocation. Social
Choice and Welfare, 18: 327-386.
Microeconomics 3 / 25
Social choice theory and its axiomatic method

Social choice theory is concerned with the relationships between

agents’ preferences or opinions over a set of social states and
collective choice. These relationships are represented by a social
choice function from an a priori specified domain of agents’
preferences to the set of social states.

Problem: Several agents have to decide on some issue or social

state of collective interest whereas their opinions or preferences
about the issue might differ. Despite their conflicting interests,
they must agree on one particular final decision. A social choice
function is any proposed solution of this problem.

Question: How to design a social choice function satisfying

desirables properties?

Microeconomics 4 / 25
Motivation

Example: Condorcet paradox

Each voter (agent) i of the set N = {1, 2, . . . , 5} reports his or her
preferences ≻i by providing a ranking of the elements belonging to
the set of candidates (social states) X = {x, y, z}.

Agent 1 x ≻1 y ≻1 z
Agent 2 x ≻2 y ≻2 z
Agent 3 y ≻3 z ≻3 x
Agent 4 z ≻4 y ≻4 x
Agent 5 z ≻5 x ≻5 y

How to select an element of X from (≻i )i∈N ?

Microeconomics 5 / 25
Social choice theory and its axiomatic method

A possibility is to use the majority rule: rank an element of X

above another one if a majority of agents do. Under the majority
rule, the social ranking ≻ is:

x ≻ y, y ≻ z and z ≻ x.

We end up with a cyclic sequence from which it is impossible to

make a rational social choice: for each candidate, there always
exists another one which is preferred by a majority of voters. This
constitutes the Condorcet paradox. □

Microeconomics 6 / 25
Social choice theory and its axiomatic method

If one wants to design a social welfare (or choice) function (with

|X| ≥ 3), the problem is thus to find a better method of
aggregation of agents’ preferences than the majority rule. Social
choice theorists approach this problem by using the so-called
axiomatic method (see Thomson, 2001):
the axiomatic study begins with the specification of the
domain of the problems (social choices, strategic games,
cooperative games. . .), and the formulation of a list of
desirable properties, called also axioms, for solutions (social
choice functions, equilibrium concepts, allocation rules . . .) on
the domain;
The axioms should express conceptually distinct
ideas/principles.

Microeconomics 7 / 25
Social choice theory and its axiomatic method
Example: the axiom of anonymity expresses an equity principle;
applied to social choice theory, it says that every agent’s
opinion/preference should influence equally the final social
decision. The axiom of Pareto optimality expresses an efficiency or
unanimity principle. Applied to social choice theory, it says that
whenever all agents rank a social state above another, so does the
social choice;
it ends with a description of the classes/sets of solutions
satisfying various combinations of the axioms. When such a
set is empty, we have an impossibility result.
The axioms characterizing a (possibly empty) set of solutions
must be logically independent. The axioms are independent if
by deleting any one of them, it is not true that the solution
set remains the only admissible one. A reason to establish
independence of axioms is to ensure that the results are
stated in the most general form.
Microeconomics 8 / 25
Social choice theory and its axiomatic method

It should also offer

a discussion of the implications of substituting for the axioms
natural variants of them;
a discussion of whether plausible alternative specifications of
the domain would affect the conclusions.

Microeconomics 9 / 25
Social states, preferences, social goals and mechanisms

The meaning of a social state depends on the context.

For an electorate seeking to fill a political office, a social state
is just a candidate for that office (see the previous example on
the Condorcet paradox).
For a government charged with delivering a public
infrastructure, the set of social states consists of the set of
possible infrastructures. For instance, assume there is a group
of islands that a government wishes to link with bridges so
that it is possible to travel from one island to any other in the
group. Then, the set of bridges which will enable to travel
from any island to any other forms an infrastructure.
...

Microeconomics 10 / 25
Social states, preferences, social goals and mechanisms

Each agent is endowed with a preference relation over the set of

social states. In most of social choice theory, preferences are
either a linear order, i.e. a binary relation which is transitive
antisymmetric and complete (and so reflexive),
or a weak order, i.e. a binary relation which is reflexive,
transitive and complete.

Microeconomics 11 / 25
Social states, preferences, social goals and mechanisms

Social goals/outcomes are given by the range/image of a

social choice function. Since a social choice function arises
from the combination of “desirable” axioms, it involves
prescriptive value judgements.
=⇒ Social choice theory is thus mostly a normative approach to
social decision-making.
The theory of mechanism design/Implementation theory
represents the positive approach of social decision-making. It
investigates the correspondence between the (normative)
social goals given by a social choice function and the
mechanisms/institutions designed to achieve those goals and
such that agents are assumed to act freely and rationally
inside these institutions.

Microeconomics 12 / 25
Social states, preferences, social goals and mechanisms

Much of theoretical or empirical work in economics tries to

explain or forecast the social outcomes of existing institutions.
The theory of mechanism design goes backward: it begins by
identifying our social goals/outcomes (by using the social
choice theory), and then ask whether institutions/mechanisms
could be designed to attain these social goals.
That’s the reason why the theory of mechanism design is
viewed as the engineering side of economic theory.

Microeconomics 13 / 25
Social states, preferences, social goals and mechanisms
What exactly do we mean by “designing mechanisms” and
“mechanisms attain these social goals”?
A mechanism is a set of incentives modeled through a
(strategic) game form. A game form contains an agent set, a
strategy/message set for each agent, and an outcome function
from the cartesian product of strategy sets to the social
states. A game form represents what we usually think of as
the rules of the game.
A game form augmented by a preference profile (i.e. for each
agent, a preference relation over the social states) is a game in
strategic form.
A mechanism attains a set of normative social goals when, for
each admissible preference profile, the equilibrium behavior
(e.g. Nash equilibrium) of the associated game in strategic
form is consistent with the social goals given by the social
choice function.
Microeconomics 14 / 25
Social states, preferences, social goals and mechanisms

By “consistent”, we mean that, for each preference profile, the

social states induced by the equilibrium behaviors coincide with
the image of the social choice function. Given a social choice
function and a mechanism, if the mechanism attains the set of
normative social goals given by the social choice function, one says
that the mechanism implements (at the equilibrium) the social
choice function.

Two important questions arise:

Can we identify some axioms/properties that a social choice
function must satisfy to be implementable (w.r.t. to some
equilibrium concept)?
What form might these mechanisms take when they exist?

Microeconomics 15 / 25
The Mount-Reiter diagram

To visualize in a more concise way the implementation problem,

the Mount-Reiter diagram is very useful. The notation is as
follows:
The fixed set of agents is N ;
The set of social states is X;
The set of admissible preference profiles ℓ := (≻i )i∈N is LN ;
The social choice function is F : LN −→ 2X \ {∅};
The mechanism M is given by a list of strategy sets (Ai )i∈N
Q
and an outcome function g : A −→ X, where A := i∈N Ai ;
A game in strategic form is a pair (M, ℓ), ℓ ∈ LN ;
A solution/equilibrium concept is a function ϕ such that, for
each ℓ ∈ LN , ϕ(M, ℓ) ⊆ A;

Microeconomics 16 / 25
The Mount-Reiter diagram

LN F (ℓ) X

ϕ(M, ℓ) g(ϕ(M, ℓ))

M ϕ-implements F if, for each ℓ ∈ LN , g(ϕ(M, ℓ)) = F (ℓ)

Microeconomics 17 / 25
King Solomon’s dilemma
King Solomon is called upon to resolve a dispute between two
women, Anna and Betta, who both claim to be the mother of
a baby. Therefore, the agent set N contains these two women
only.
The identity of the true mother is common knowledge
between Anna and Betta, but unknown to King Solomon,
whose objective is rightfully to resolve the dispute at no cost
to the true mother.
The set X of feasible King Solomon’s judgements (social
states) contains three elements:

x(A) := “give the baby to Anna”, x(B) := “give the baby to Betta”,

and

x(H) := “cut the baby in half; give each mother one half”.

Microeconomics 18 / 25
King Solomon’s dilemma
The set LN of preference profiles over X contained two
elements. If Anna is the real mother, then ℓA := (≻A A
A , ≻B ) is
given by:

x(A) ≻A
A x
(B)
≻A
A x
(H)
and x(B) ≻A
B x
(H)
≻A
B x
(A)
;

if Betta is the real mother, then ℓB := (≻B B

A , ≻B ) is given by:

x(A) ≻B
A x
(H)
≻B
A x
(B)
and x(B) ≻B
B x
(A)
≻B
B x
(H)
.

King Solomon’s (wise) judgement (social choice) is represented

by the social choice function F : LN −→ X defined as:

F (ℓA ) = x(A) := “give the baby to Anna”,

and
F (ℓB ) = x(B) := “give the baby to Betta”.
Microeconomics 19 / 25
King Solomon’s dilemma

Since King Solomon’s does know who is the real mother, he

wants to design a mechanism, or a game form, so that the
unique Nash equilibrium outcome of the mechanism is to give
the baby to the true mother, i.e. a mechanism that
implements F in Nash equilibrium.
The mechanism/game form M is a follows:
He asks to each mother to tell him in secret whether she is the
real mother. Thus, each mother can transmit one of the two
following secret messages to King Solomon:

m := “The baby is mine”, and ¬m := “The baby is not mine”.

The outcome fonction g is represented by the following table:

Microeconomics 20 / 25
King Solomon’s dilemma

Anna\Betta m ¬m
m x(H) x(B)
¬m x(A) x(H)
Recall that a Nash equilibrium of the strategic game (M, ℓ), where
ℓ := (≻A , ≻B ) ∈ LN is a message profile
a := (aA , aB ) ∈ AAnna × ABetta such that:

∀b ∈ AAnna \ {aA }, g(a) ≻A g(b, aB ).

∀b ∈ ABetta \ {aB }, g(a) ≻B g(aA , b).

Microeconomics 21 / 25
King Solomon’s dilemma

Anna\Betta m ¬m
Given g, m x(H) x(B)
¬m x(A) x(H)

if the preference profile is ℓ(A) ,

x(A) ≻A
A x
(B)
≻A
A x
(H)
and x(B) ≻A
B x
(H)
≻A
B x
(A)
,
then the unique Nash equilibrium is (m, ¬m), and
g(m, ¬m) = x(B) ̸= F (ℓ(A) ) = x(A) ;

if it is ℓB , x(A) ≻B
A x
(H) ≻B x(B) and x(B) ≻B x(A) ≻B x(H) ,
A B B
then the unique Nash equilibrium is (¬m, m), and
g(¬m, m) = x(A) ̸= F (ℓ(B) ) = x(B) .
Microeconomics 22 / 25
King Solomon’s dilemma

Conclusion: at the Nash equilibrium, the baby will not be cut in

half. Nevertheless, King Solomon’s mechanism always allocates
the baby to the wrong mother, i.e. the mechanism does not
implement F in Nash equilibrium.

King Solomon’s problem is a relevant parable for economists: it is

a special case of the economic problem of allocation of an
indivisible object: an indivisible “prize” (e.g. a master painting) is
to be allocated between two (or more) agents. An auctioneer
wants to design a mechanism to implement the social choice
function that assigns the prize to the agent who values it most.

Microeconomics 23 / 25
Matching theory

Matching theory focuses on the question of the allocation of

items or partners, based upon their preferences.

=⇒ Matching theory is thus well suited to study markets in

which agents can’t just choose what they want, they have to
be chosen, i.e. markets that can’t be solved by using the price
mechanism.
Matching theory provides methods to generate stable, efficient
and fair matches.
It has been applied successfully to concrete problems such as
student placement in schools (which students go to which
school?), labor markets (who works at which job?), and
transplantable organ exchange (who receives which
transplantable kidney?).

Microeconomics 24 / 25
Matching theory

Matching problems can be formulated into the framework of

Social choice theory.
Example: Two-sided matching.
- Agents in N belong to one of two disjoints sets, say U
(Universities), S (Students).
- Each agent in U can be matched to at most one agent in S, and
vice versa. This situation is called a one-to-one matching and
represents a social state.
- Each agent in N = U ∪ S has a preference relation over the set X
of one-to-one matchings.
- A social choice function F assigns to each feasible preference
profile a (set of) one-to-one matching(s). □

Microeconomics 25 / 25