ENSAE Année académique 2022-2023

ISE2

Théorie des jeux

TD 2

Exercice 1 :

FOFANA et BALDE veulent partager une bûche glacée. FOFANA propose un partage initial. Si
BALDE refuse, c’est à lui de faire une offre, mais une partie de la glace aura fondu ( il fait chaud
et ils n’ont pas la possibilité de la laisser dans un congélateur le temps de la négociation ). De
même si c’est FOFANA qui refuse il doit faire une contre-offre. Quelle va être le partage décidé
par ces deux ? En considérant les différentes possibilités de négociations suivantes. Ces partages
constituent ils des équilibres de Nash ?

1. La glace fond une seule fois au cas où BALDE refuse la première offre de FOFANA (On
suppose que BALDE préfère prendre une miette que de ne rien avoir)
2. La glace prend deux périodes pour fondre totalement ?
3. Un tiers de la glace fond à chaque période ?
4. Seulement ¼ de la glace fond en cas de désaccord ?
5. Seulement 1/5 de la glace fond en cas de désaccord ?
6. Donner une généralisation de pour ( t périodes :on distinguera≤cas t pair et t impair )
7. Donner la valeur du partage pour (t=101). Quelle conclusion pouvez-vous tirer de cet
exemple ?

Exercice 2 : Un simple jeu de pollution transfrontalière à temps discret

To illustrate the difference between OLNE and MPNE, let us consider a simple example: a two-
period model of transboundary pollution. There are two players (say two countries), called
SMALL and CAP. Player SMALL chooses its level of CO2 emissions for periods 1 and 2,
denoted by the lowercase symbols x and y. Player CAP’s levels of emissions for period 1 and 2
are denoted by the capital letters X and Y . Let St be the global stock of CO2 concentration in
the atmosphere at the beginning of period t. The initial stock (at the beginning of period 1) is
denoted by S1. Assume that the stock at the beginning of period 2 is determined by

S2=S 1+ x + X

and that the stock at the end of period 2 (which is the beginning of period 3) is S3=S 2+ y+Y
The variable St is called the state variable of the system. The control variables in this model are
the emission levels.

The period 2 payoff to SMALL is the benefit from consuming the output, ay − ( 12 ) y , minus the
2

cost of period 2 environmental damages, ( 12 ) S 2

2

u2=ay− ( 12 ) y −( 12 ) S
2 2
2

where a is a positive constant. Similarly, CAP’s payoff in period 2 is

U 2= AY − ( 12 ) Y −( 12 ) S
2 2
2

where A is a positive constant. We assume SMALL’s payoff in period 1 is

u1=ax− ( 12 ) x −( 12 ) S
2 2
1

and that of CAP is

U 1=aX− ( 12 ) X −( 12 ) S
2 2
1

The objective of SMALL is to maximize its overall payoff w, defined as the sum of its period 1
and period 2 payoffs, minus a term that reflects its guilt of passing on to the next generation the
stock of pollution S3. We denote this term by g3 (S3 ) and assume it is increasing in S3. For

()
simplicity, let g3 (S3 ) take the simple form
1 2
2 3
S . Thus, w ≡ u1 +u2−g3 =u1+ u2−
1 2
2 3
S ()
Similarly, CAP’s guilt function is G ( S )=( ) S
1 2
3 3 3 and it wants to maximize
2

W ≡U +U −G =U +U −( ) S
1 2
1 2 3 1 2 3
2

1. Find the open-loop Nash equilibrium (OLNE ) : open-loop strategy of a player is a

planned time path of actions over the time horizon. Each player assumes that the other
player is going to carry out its planned course of actions.
2. Find the Markov-perfect strategies, it’s found by solving the game backward. This
method gives us “feedback decision rules”
3. Comparing OLNE and MPNE of the transboundary pollution game
4. Consider a numerical example. L et A = a = 20, and S1 = 2.