Vous êtes sur la page 1sur 6

Lecture 15: Competitive equilibrium in a pure

exchange economy
1 A pure exchange economy
Suppose there are l goods (commodities) in the economy. A pure exchange economy is a
tuple u
i
; !
i

i2N
, where N = 1; : : : ; n denotes a set of consumers, and u
i
: R
l
+
R; !
i
=
(!
i1
; : : : ; !
il
) R
l
++
are their utility functions and endowments, respectively. Further, we
assume that the u
i
s are continuous, strictly increasing and strictly quasi-concave.
Given their endowments, consumers trade with each other in the market to get better
consumption bundles and maximize their utilities. If n is large enough, then each consumer
would have no impact on the market prices of the goods. That is, each consumer takes the
market price vector p R
l
+
as given and chooses the best consumption bundle by solving
max
x
i
2R
l
+
u
i
(x
i
) = u
i
(x
i1
; x
i2
; : : : ; x
il
) (1.1)
subject to p x
i
_ p !
i
; and x
i
_ 0:
It is easy to see that only relative prices matter; if you change the price vector p to p,
1
the constraint set will not change. Hence, we can normalize the prices to be
p
l1
= p R
l
+
:
l
X
j=1
p
j
= 1; p
j
_ 0;
where
l1
R
l
+
is the l-dimensional simplex.
Suppose that given p; the unique solution to Equation 1.1 is x
i
(p); which is consumer is
demand. Then the market demand is
D(p) =
n
X
i=1
x
i
(p):
The market supply, on the other hand, is the sum of individual endowments:
S(p) =
n
X
i=1
!
i
:
The market supply is independent of p because there is no production here. The question
is if there is some price vector p such that markets of all goods clear, that is, D( p) = S( p):
Example 1.1. A pure exchange economy of two goods and two consumers: N = 1; 2; !
1
=
(1; 2); !
2
= (3; 1): Therefore, the total supply is S = (4; 3).
We can use the Edgeworth Box to illustrate all possible trading points x
1
= (x
11
; x
12
)
and x
2
= (x
21
; x
22
). Market clearing requires that x
1
+ x
2
= !
1
+ !
2
: That is,
8
>
<
>
:
x
11
+ x
21
= 4
x
12
+ x
22
= 3
:
2
2 Competitive equilibrium
Denition 2.1. Acompetitive equilibrium(or Walrasian equilibrium) is a tuple ( x
1
; : : : ; x
n
; p)
R
nl

l1
consists of n consumption bundles xs, one for each consumer, and a price vector
p
l1
such that
1. (Individual rationality) For each consumer i; x
i
= x
i
( p); that is, x
i
solves
max u
i
(x
i
) = u
i
(x
i1
; : : : ; x
il
)
s.t p x
i
_ p !
i
; and x
i
_ 0:
2. (Market clearing condition) The market demand equals the market supply,
n
X
i=1
x
i
=
n
X
i=1
!
i
:
Let z(p) = (z
1
(p); : : : ; z
l
(p)) dened by
z(p) = D(p) S(p) =
n
X
i=1
x
i
(p)
n
X
i=1
!
i
be the vector of excess demands under price vector p. Then the market clearing condition
is equivalent to z( p) = 0. If p
i
= 0 for some good i, that is, if good i is free, then since u
i
is
strictly increasing, everybody would demand the maximal amount of good i. Then z
i
(p) > 0;
the market for good i can never clear. Hence at any equilibrium, p
i
> 0; \i N.
Also, since \i; u
i
is strictly increasing, given any price vector p, at the optimal solution,
p z(p) = 0:
This is called the Walras Law. It holds because for each consumer i, p !
i
p x
i
(p) = 0;
3
that is, everybody uses up the budget.
Theorem 2.1. Any pure exchange economy satisfying all assumptions above has at least one
competitive equilibrium.
Proof. Since p > 0; given the assumptions on u
i
; Weierstrass theorem and Theorem of
Maximum imply that the optimal solution to the utility maximization problem, x
i
(p); is a
continuous function. Hence z(p) is continuous.
We are left to show that there exists a price vector p
l1
such that the excess demand
vector z( p) = 0: The idea is to dene a function f :
l1

l1
which illustrates the
dynamic adjustment of prices. More concretely, in the adjustment process,
1. at any price vector p, if the market is not at equilibrium, that is, if z(p) ,= 0, we
increase the prices of the goods that are in excess demand. And since only relative
prices matter, we dont need to decrease the prices of the goods in excess supply.
2. the size of the increase in the price of each good depends on the excess demand.
3. the new prices are still normalized to be in
l1
.
Let the price adjustment function f :
l1

l1
be dened by f(p) =
l
j=1
f
j
(p);
where
f
j
(p) =
p
j
+ max0; z
j
(p)
P
l
k=1
(p
k
+ max0; z
k
(p))
=
p
j
+ max0; z
j
(p)
1 +
P
l
k=1
max0; z
k
(p)
denes the new price for good j after adjustment, given p and z(p). You should also see
that if z(p) = 0; then f
j
(p) = p
j
; the adjustment process stops. We thus expect any xed
point of f to be an equilibrium price.
Since f is continuous and
l1
is nonempty, convex and compact, by Brouwers xed
point theorem, there exists p
l1
such that f( p) = p, that is,
p
j
=
p
j
+ max0; z
j
( p)
1 +
P
l
k=1
max0; z
k
( p)
; \1 _ j _ l:
4
Hence
p
j
l
X
k=1
max0; z
k
( p) = max0; z
j
( p)
l
X
j=1
z
j
( p) p
j
l
X
k=1
max0; z
k
( p) =
l
X
j=1
z
j
( p) max0; z
j
( p):
By Walras Law,
P
l
j=1
z
j
( p) p
j
= 0. Therefore,
l
X
j=1
z
j
( p) max0; z
j
( p) = 0:
Consequently,
z
j
( p) _ 0; \1 _ j _ l:
Walras Law also implies that p
j
z
j
( p) =
P
kj
p
k
z
k
( p) _ 0; \j. Hence z
j
( p) = 0; \j.
In general, to solve for a competitive equilibrium, we need to solve the utility maximiza-
tion problem for each consumer, and then plug in the demand functions x
i
(p) in l market
clearing conditions to get the equilibrium price. Thanks to Walras Law, we only need to
solve (l 1) prices from (l 1) market clearing conditions.
Example 2.2. Consider a two-good two-consumer pure exchange economy. Suppose
u
1
(x
11
; x
12
) = (x
11
)
1
4
(x
12
)
3
4
; (!
11
; !
12
) = (2; 1):
u
1
(x
21
; x
22
) = (x
21
)
3
4
(x
22
)
1
4
; (!
21
; !
22
) = (2; 2):
Since we care only about relative prices, its usually convenient to normalize p
1
= 1.
Consumer 1s utility maximization problem is
max(x
11
)
1
4
(x
12
)
3
4
s.t. x
11
+ p
2
x
12
= !
11
+ p
2
!
12
:
5
Solving it gives
x
11
(1; p
2
) =
2 + p
2
4
; x
12
(1; p
2
) =
3(2 + p
2
)
4p
2
:
Similarly,
x
21
(1; p
2
) =
3(2 + 2p
2
)
4
; x
22
(1; p
2
) =
2 + 2p
2
p
2
:
The clear of the market of good 1 requires
2 + p
2
4
+
3(2 + 2p
2
)
4
= 2 + 2:
p
2
=
8
7
:
Hence x
11
=
11
14
; x
12
=
33
16
; x
21
=
45
14
; x
22
=
15
16
: And the competitive equilibrium is
((x
11
; x
12
); (x
21
; x
22
); p) = ((
11
14
;
33
16
); (
45
14
;
15
16
); (1;
8
7
)):
2.1 Welfare implications of competitive equilibrium
An allocation is (Pareto) ecient if no alternative can make everyone better o. The rst
welfare theorem states that any competitive equilibrium is ecient. This is viewed as one
of the most important results in economics since it best explains merit of market economy.
However, market economy may not reduce inequality. Given every unequal endowments, the
equilibrium allocation may also be very unequal.
The second welfare theorem states that if income transfers among consumers are per-
mitted, any ecient allocation can be achieved through market economy as an competitive
equilibrium allocation.
6

Vous aimerez peut-être aussi