A Simple Axiomatization of The

M PRA
Munich Personal RePEc Archive
A simple axiomatization of the

egalitarian solution
Ismail Saglam
TOBB University of Economics and Technology
19. February 2012
Online at http://mpra.ub.uni-muenchen.de/36773/
MPRA Paper No. 36773, posted 20. February 2012 01:39 UTC
A simple axiomatization of the egalitarian solution

Ismail Saglam
Department of Economics, TOBB University of Economics and Technology,
Sogutozu Cad. No: 43, Sogutozu 06560 Ankara, Turkey.
Abstract. In this paper, we present a simple axiomatization of the n-person egalitarian

solution. The single axiom sufficient for characterization is a new condition which we call
symmetric decomposition.
Keywords: Cooperative bargaining; egalitarian solution.
JEL Classification Numbers: C71, C78
Introduction
Since the pioneering work of Nash (1950) on cooperative bargaining theory, a voluminous number of bargaining solutions have been proposed and axiomatized in the
literature. Among the solutions that have been most studied is the egalitarian solution which was recommended by Rawls (1971). Given a bargaining problem faced by
individuals in the society, this solution implies maximization of the utility of the worstoff individual over the bargaining set. A characterization of the egalitarian solution
when the number of individuals is fixed and the bargaining set is convex, compact, and
comprehensive was proposed by Kalai (1977) using symmetry, weak Pareto optimality,
Fax: +(90) 312 292 4213.
E-mail addresses: [email protected] (I. Saglam).
and strong monotonicity conditions. Symmetry axiom says that if the bargaining set is
invariant under all exchanges of agents, then the solution must be invariant too. Weak
Pareto optimality requires that all gains from cooperation should be weakly exhausted,
whereas strong monotonicity demands that all agents should benefit from any expansion of the bargaining set. Kalai (1977) also shows that strong monotonicity condition
can be replaced by step-by-step negotiation, a decomposability condition which says
that if the bargaining set expands from U to S, the solution on S can be calculated by
first finding the solution on U (step 1) and then adding it to the solution on the set of
individually rational options in S with respect to the solution in step 1.
In this paper, we propose a simple axiomatization of the egalitarian solution in
the same bargaining domain as studied by Kalai (1977). We show that the n-person
egalitarian solution is the only bargaining solution that satisfies a new condition which
we call symmetric decomposition.
A number of studies in the literature have axiomatized the egalitarian solution in
alternative domains of bargaining problems. As such, Thomson (1983a, 1984) consider
bargaining problems where the number of bargaining individuals may vary, Conley and
Wilkie (2000) relax the restriction that the bargaining set is convex, Rachmilevitch
(2011) considers a restricted domain where the bargaining set is strictly comprehensive,
and Conley and Wilkie (2012) study domains where the bargaining set is finite. There
are also other studies that consider the characterization of related solutions, such as the
lexicographic egalitarian solution and proportional solutions. See, for example, Myerson
(1977), Roth (1979), Myerson and Thomson (1980), Myerson (1981), Thomson (1983b),
and Chun and Thomson (1990).
The paper is organized as follows: In Section 2 we introduce the basic structures
and in Section 3 we present our characterization result. Finally, Section 4 contains
some concluding remarks.
Basic Structures
We consider a society with the set of individuals N = {1, 2, ..., n}. A bargaining
n
problem for this society consists of a pair (S, d) where S is a non-empty subset of R+
and d S. Here, S represents von Neumann-Morgenstern utilities attainable through

the cooperative actions of n individuals. If the individuals fail to agree on an outcome
in S, then the bargaining is settled at the point d, which is called the disagreement
point. In this paper, we consider the domain n of bargaining problems where
(a) S is convex and compact, and there exists x S such that x > d.1
n
, x y d and x 6= y then y S
(b) S is d-comprehensive; i.e., if x S, y R+
(the possibility of free disposal of utility).

We define the weak Pareto set of S as
W P (S) = {x S | y > x implies y
/ S}
and the strong Pareto set of S as
P (S) = {x S | y x and y 6= x implies y

/ S}.
n
A solution F is a mapping from n to R+
such that for each (S, d) n , F (S, d) S.
The egalitarian solution maps each bargaining problem (S, d) n to the point E(S, d)
of W P (S) such that Ei (S, d) di = Ej (S, d) dj for all i, j N .
Given a bargaining problem (S, d), we denote by aj (S, d) the maximal net utility
attainable by agent j N ; i.e., aj (S, d) = maxxS (xj dj ). For any real (0, 1], we
define the reference point c(S, d, ) S such that ci (S, d, ) = di + minjN aj (S, d)/2
for all i N . Clearly, c(S, d, 1) d is the symmetric point in the Pareto frontier
of the convex hull of the set of vectors v 1 , v 2 , . . . , v n where for each k N , vkk =
minjN aj (S, d) and vlk = 0 for each l N \{k}.
1
n
Given two vectors x and y in R+
, x y means xi yi for all i N and x > y means xi > yi for
all i N .
Given a bargaining problem (S, d), we denote by IR(S, d) the individually rational
n
allocations; i.e., IR(S, d) = {x S | d x}. For any bargaining set S R+
and
n
any z R+
we define S z = {x Rn | y S such that x = y z}. Apparently,
IR(S, c(S, d, )) c(S, d, ) = IR(S c(S, d, ), 0) n for any (S, d) n and for
any (0, 1]. In Fig. 1, we plot a 2-person bargaining problem with = 1.
Agent 2
S
IR(S,d)
IR(S,c(S,d,1))
E(S,d)
a2(S,d)
90
c(S,d,1)
45
d
a1(S,d)
(0,0)
Agent 1
Fig. 1
We investigate the existence of solutions satisfying the following new axiom:
Symmetric decomposition. There exists (0, 1] such that F (S, d) = c(S, d, ) +
F (IR(S c(S, d, ), 0), 0) for all (S, d) n .
The above axiom says that the solution F (S, d) can be calculated in two steps, by
first obtaining, for some (0, 1], the symmetric reference point c(S, d, ) in S and
then taking it to be the starting point for the distribution of the utilities in S.
Characterization Result
Theorem 1. A solution satisfies symmetric decomposition if and only if it is the

egalitarian solution.
Proof. It is clear that the egalitarian solution satisfies symmetric decomposition since
for all (0, 1] we have E(S, d) = c(S, d, ) + E(IR(S c(S, d, ), 0), 0) for all
(S, d) n . Conversely, let F be a solution on n satisfying symmetric decomposition.
First, pick any (0, 1] such that F (S, d) = c(S, d, ) + F (IR(S c(S, d, ), 0), 0)
for all (S, d) n , and then pick any (S, d) n .
Consider the sequence of
0
= S, d0 = d, and dt = 0 and S t =
bargaining problems (S t , dt )
t=0 where S
IR(S t1 c(S t1 , dt1 , ), 0) for each integer t 1. We say that agent k deter ) in a given bargaining problem (S,
n if
d,
d)
mines the reference point c(S,
d)}.
k = min{i N | i = argminjN aj (S,
Then, for any integer m 0, there must
exist an agent, say k(m), determining the reference point in at least m + 1 of the
first nm + 1 problems in the sequence (S t , dt )
t=0 . For each integer m 0, we denote by (S t , dt )t{t1 ,t2 ,...,tm+1 } the first m + 1 problems in which agent k(m) determines
the reference point. Clearly, we have ak(m) (S ti+1 , dti+1 ) (1 2 )ak(m) (S ti , dti ) for
all i {1, 2, . . . , m} for each integer m 0. Using the fact that ak(m) (S t1 , dt1 )
ak(m) (S 0 , d0 ), we have ak(m) (S tm+1 , dtm+1 ) (1 2 )m ak(m) (S 0 , d0 ) for each integer
m 0. Now suppose that F (S 0 , d0 ) 6= E(S 0 , d0 ). Pick any integer m
such that
0 0
0 0
0 0
tm+1
(1 2 )m ak(m)
, dtm+1
)<
(S , d ) < |Fk(m)
(S , d ) minjN Fj (S , d )|. Then ak(m)
(S
P
tm
0 0
0 0
0 0
|Fk(m)
/ S tm+1
.
(S , d ) minjN Fj (S , d )| and therefore F (S , d )
=1 c(S , d , )
P
However, we have F (S t+1 , dt+1 ) = F (S 0 , d0 ) t =1 c(S , d , ) for all integer t 0 by
symmetric decomposition. This implies that F (S tm+1

, dtm+1
)
/ S tm+1
, a contradiction.
Therefore, F (S 0 , d0 ) = E(S 0 , d0 ). Since (S 0 , d0 ) = (S, d) n was arbitrarily picked,

F must be equal to the the egalitarian solution on n .
Concluding Remarks
In this paper, we have considered an alternative characterization of the egalitarian

solution in a class of bargaining problems with convex, compact and comprehensive
bargaining sets. We show that the n-person egalitarian solution is the only bargaining
solution that satisfies a new condition which we call symmetric decomposition. This single condition replaces the three conditions in a characterization result of Kalai (1977),
step-by-step negotiation, symmetry and weak Pareto optimality. Dropping symmetry in
Kalai (1977), any n-person weighted egalitarian solution (n-person proportional solution) with weights in the n 1 dimensional simplex, selecting the maximal point of
the bargaining set in the direction of also become admissible. Alternatively, dropping
weak Pareto optimality in the characterization result of Kalai (1977), any n-person contracted egalitarian solution with the contraction factor in [0, 1] also become admissible.
The reason why the decomposition condition of ours successfully strengthens that
of Kalai (1977) is that given any bargaining problem (S, 0), step-by-step negotiation
requires any bargaining solution F to be decomposable with regard to the reference
point F (T, 0) for each T S, whereas symmetric decomposition requires the decomposability, for some (0, 1], only with regard to the symmetric point c(T, 0, ) of the
greatest symmetric and strictly comprehensive set T contained by S such that T has
a linear Pareto frontier. Thus, the reference point in our case is proportional to the
egalitarian solution on T ; i.e., c(T, 0, ) = E(T, 0), where (0, 1]. Since it is true
that E(T, 0) = E(T, 0) for any (0, 1], the reference point we use in our characterization is the symmetric and weakly Pareto optimal point in T for our particular
choice of T . Indeed, what eliminate the admissibility of any n-person contracted, nonegalitarian proportional solution in the sole presence of decomposability are entirely
the notions of symmetry and weak Pareto optimality embedded in our decomposition
condition.
References
Chun, Y. & Thomson, W. (1990). Egalitarian solutions and uncertain disagreement
points. Economics Letters, 33, 29-33.
Conley, J. P. & Wilkie, S. (2000). The bargaining problem without convexity: extending the egalitarian and Kalai-Smorodinsky solutions. Mimeo.
Conley, J. P. & Wilkie, S. (2012). The ordinal egalitarian bargaining solution for finite
choice sets. Social Choice and Welfare, 38, 23-42.
Kalai, E. (1977). Proportional solutions to bargaining situations: Interpersonal utility
comparisons. Econometrica, 45, 1623-1630.
Myerson, R. B. (1977).
Two-person bargaining problems and comparable utility.
Econometrica, 45, 1631-1637.

Myerson, R. B. (1981). Utilitarianism, egalitarianism and the timing effect in social
choice problems. Econometrica, 49, 883-897.
Nash, J. F. (1950). The bargaining problem. Econometrica, 28, 155-162.
Rachmilevitch, S. (2011). Disagreement point axioms and the egalitarian bargaining
solution. International Journal of Game Theory, 40, 63-85.
Rawls, J. (1971). A theory of justice. Cambridge, Massachusetts: Harvard University
Press.
Roth, A. E. (1979). Proportional solutions to the bargaining problem. Econometrica,
47, 775-778.
Thomson, W. (1983a). Problems of fair division and the egalitarian solution. Journal
of Economic Theory, 31, 211-226.
Thomson, W. (1983b). Truncated egalitarian and monotone path solutions. University

of Minnesota Discussion Paper, No. 174.
Thomson, W. (1984). Monotonicity, stability and egalitarianism. Mathematical Social
Sciences, 8, 15-28.
Thomson, W. & Myerson, R. B. (1980). Monotonicity and independence axioms. International Journal of Game Theory, 9, 37-49.

A Simple Axiomatization of The

Uploaded by

Copyright:

Available Formats

A Simple Axiomatization of The

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

A Simple Axiomatization of The

Uploaded by

Copyright:

Available Formats

M PRA

Munich Personal RePEc Archive

A simple axiomatization of the

19. February 2012

A simple axiomatization of the egalitarian solution

Abstract. In this paper, we present a simple axiomatization of the n-person egalitarian

Fax: +(90) 312 292 4213.

E-mail addresses: [email protected] (I. Saglam).

and d S. Here, S represents von Neumann-Morgenstern utilities attainable through

(the possibility of free disposal of utility).

P (S) = {x S | y x and y 6= x implies y

Theorem 1. A solution satisfies symmetric decomposition if and only if it is the

Consider the sequence of

symmetric decomposition. This implies that F (S tm+1

Therefore, F (S 0 , d0 ) = E(S 0 , d0 ). Since (S 0 , d0 ) = (S, d) n was arbitrarily picked,

In this paper, we have considered an alternative characterization of the egalitarian

Two-person bargaining problems and comparable utility.

Econometrica, 45, 1631-1637.

Thomson, W. (1983b). Truncated egalitarian and monotone path solutions. University

You might also like