Upload
allison-snipes
View
220
Download
4
Tags:
Embed Size (px)
DESCRIPTION
nmnnlnln
Citation preview
151
Game Theory
Definition: The players are the decision-makers whose behavior we are trying to predict.
Definition: The strategies are the potential choices that can be made by the players.
Definition: The payoffs are the outcomes, or consequences, of the strategies chosen.
Definition: A strategic form game is a tuple 1 1 1 1[ , ... , , ( ,..., ), ... , ( ,..., )]n n n nG S S u S S u S S , where
for each player 1,...,i n , iS is the set of strategies available to player i , and
1( , ..., )i nu S S describes player 'i s payoff as a function of the strategies chosen by all
players.
A strategic form game is finite if each players strategy set contains finitely many elements.
Example: The Prisoners Dilemma Game
Penalty B
Confess Not confess
A Confess $7m, $7m $1m, $10m
Not confess $10m, $1m $2m, $2m
{confess, not confess}
(confess, confess) (confess, confess) $7
(confess, not confess) (not confess, confess) $1
(not confess, confess) (confess, not confess) $10
(not confess, n
A B
A B
A B
A B
A
S S
u u m
u u m
u u m
u
ot confess) (not confess, not confess) $2 Bu m
Note: A 2-player strategic form games with finite strategy sets can always be represented in
matrix form.
Definition: is , for player i , is a strictly dominant strategy if ( , ) ( , ) ( , )i i i i i i i iu s s u s s s s .
Note: A strictly dominant strategy is a strategy a player will adopt no matter what strategy the
other player adopts.
Definition: A strictly dominant strategy equilibrium is the resulting equilibrium of all players
adopt their dominant strategy.
152
Note that in this game, the best outcomes for both 2 players together is when neither one of them
confesses. The worst outcomes for both 2 players together is when both of them confess.
The Prisoners Dilemma game predicts that both player will confess.
In this game,
i) the dominant strategy for A is confess.
ii) the dominant strategy for B is confess.
iii) the dominant strategy equilibrium is (confess, confess).
Cheating in the cartel
Profit B
Cheat Comply
A Cheat $10, $10 $25, $5
Comply $5, $25 $20, $20
Note that in this game, the best outcomes for both 2 players together is when neither one of them cheats.
The worst outcomes for both 2 players together is when both of them cheat.
The cheating in the cartel game predicts that both player will cheat.
In this game,
i) the dominant strategy for A is cheat.
ii) the dominant strategy for B is cheat.
iii) the dominant strategy equilibrium is (cheat, cheat).
Example: High output or low output (a game with no DS equilibrium)
Profit B
Low output High output
A Low output $22, $20 $9, $30
High output $20, $17 $18, $25
In this game,
i) there is no strictly dominant strategy for A.
ii) the strictly dominant strategy for B is high output.
iii) there is no strictly dominant strategy equilibrium.
Definition: Player 'i s strategy is , strictly dominates another of his strategies is , if
( , ) ( , ) i i i i i i iu s s u s s s . In this case, we also say that is is strictly dominated.
153
Solution by iterated elimination of strictly dominated strategies
Example:
2
L C R
T 3, 3 2, 6 3, 1
1 M 2, 4 1, 4 0, 4
B 1, 5 0, 2 6, 0
2
L C R
1 T 3, 3 2, 6 3, 1
B 1, 5 0, 2 6, 0
2
L C
1 T 3, 3 2, 6
B 1, 5 0, 7
2
L C
1 T 3, 3 2, 6
2
C
1 T 2, 6
Definition: Let 0 for each player i iS S i , and for 1N , let N
iS denotes those strategies of player i
surviving after the thN round of elimination. That is 1, if N Ni i i is S s S is not strictly
dominated in 1NS .
Definition: A strategy is for player i is iteratively strictly undominated in S (or survives iterative
elimination of strictly dominated strategies) if 1Ni is S N .
Example:
2
L C R
T 3, 3 2, 1 3, 1
1 M 2, 4 2, 4 0, 4
B 1, 5 0, 2 6, 5
both players do not have
dominate strategies.
For 1: M is strictly
dominated by T.
For 2: R is strictly
dominated by either L or C.
For 1: B is strictly
dominated by T.
solution
neither player has strictly
dominated strategy.
154
Definition: Player 'i s strategy is , weakly dominates another of his strategies is , if
( , ) ( , ) with at least 1 strict inequalityi i i i i i i iu s s u s s s S . In this case, we also
say that is is weakly dominated.
Definition: Let 0 for each player i iW S i , and for 1N , let N
iW denotes those strategies of player i
surviving after the thN round of elimination. That is 1, if N Ni i i is W s W is not strictly
dominated in 1NW .
Definition: A strategy is for player i is iteratively weakly undominated in S (or survives iterative
elimination of weakly dominated strategies) if 1Ni is W N .
Solution by iterated elimination of weakly dominated strategies
2
L C R
T 3, 3 2, 1 3, 1
1 M 2, 4 2, 4 0, 4
B 1, 5 0, 2 6, 5
2
L C R
1 T 3, 3 2, 1 3, 1
B 1, 5 0, 2 6, 5
2
L
1 T 3, 3
B 1, 5
2
L
1 T 3, 3
neither player has strictly
dominated strategy.
For 1: M is weakly
dominated by T.
For 2: L and R are weakly
dominated by L.
solution
155
Example
2
L C R
T 3, 3 2, 1 3, 6
1 M 4, 5 2, 7 0, 4
B 1, 5 0, 2 6, 5
Definition: A Nash equilibrium is a set of strategies (or choices) such that each players choice is the best one possible given the strategy chosen by the other players.
Formally, given a strategic form game 1 1 1 1[ , ... , , ( ,..., ), ... , ( ,..., )]n n n nG S S u S S u S S ,
( 1 , ..., ns s ) is a Nash equilibrium if for each player i , 1 ( , ..., ) ( , ) i n i i i i iu s s u s s s S .
Example
2
L C R
T 3, 3 2, 1 3, 6
1 M 4, 5 2, 7 0, 4
B 1, 5 0, 2 6, 5
(1, 2) 12 12
(T, L ) TR X
(T, C) TR X
(T, R) TR BR X
(M, L) MC X
(M, C) MC T,MC Nash equilibrium
(M, R) MC X
(B, L) BL,R ML X
(B, C) BL,R X
(B, R) BL,R BR Nash equilibrium
neither player has weakly
dominated strategy.
neither player has weakly
dominated strategy.
156
Example: (A game with no Nash equilibrium)
B
I II
A I $0, $0 $0, $1
II $1, $0 $1, $3
(A, B) AB AB
(I, I ) II III X
(I, II) II X
(II, I) IIII X
(II, II) IIII III X
157
Mixed Strategies Nash Equilibrium
Definition: Fix a strategic form game 1 1 1 1[ , ... , , ( ,..., ), ... , ( ,..., )]n n n nG S S u S S u S S . A mixed
strategy for player i is a probability distribution over iS .
iM : set of mixed strategies for player i
iS : set of pure strategies for player i
Example:
2
L R
1 T 3, 1 2, 4
B 2, 2 3, 1
The game has no pure-strategy Nash equilibrium.
Lets assume the players have mixed strategies, in which a player randomizes by choosing the probabilities for playing the possible pure strategies.
Let p be the probability for 1 to choose T ,
1 p be the probability for 1 to choose B .
Let q be the probability for 2 to choose L ,
1 q be the probability for 2 to choose R .
Expected payoff for 1: 1( ) (3) (1 ) (2) (1 )(2) (1 )(1 )(3) 2 3E pq p q p q p q pq p q
Expected payoff for 2: 2( ) (1) (1 ) (2) (1 )(4) (1 )(1 )(1) 4 3 1E pq p q p q p q pq p q
Problem for 1:
1max ( ) 2 3p
E pq p q
FOC: 1( ) 1
2 1 0 *2
dEq q
dp
Problem for 2:
2max ( ) 4 3 1q
E pq p q
FOC: 2( ) 1
4 1 0 *4
dEp p
dq
1
1 1 1 1( ) 2( )( ) ( ) ( ) 3 2.5
4 2 4 2E
2
1 1 1 1( ) 4( )( ) 3( ) ( ) 1 1.75
4 2 4 2E
158
Remark: There is a problematic feature of a mixed-strategy equilibrium: neither player has any
positive incentive to play the equilibrium strategy.
Player 1: 11 1
( ) 2( )( ) ( ) ( ) 3 2.5 (no matter what is!!!)2 2
E p p p
Player 2: 21 1
( ) 4( )( ) 3( ) ( ) 1 1.75 (no matter what is!!!)4 4
E q q q
Theorem: Every n-player game with finite pure-strategy sets has at least 1 pure or mixed strategy
Nash equilibrium.
159
Oligopoly
Properties:
1. a few sellers (firms)
2. many buyers (consumers)
3. homogeneous products
4. difficult to enter into the market
Cournot model with 2 identical firms
1Firm i's share (1 j's share)
2
: competitive outputCQ
P P
MC MC
Q Q 0
MR D MR D
Residual D for firm 2 Firm 1 Firm 2
1 1
2cQ
0
2 1
2cQ
1 1 1( )
2 2 4c c cQ Q Q
3 1 1 3( )
2 4 8c c cQ Q Q
1
4cQ
4 3
8cQ
1 3 5( )
2 8 16c c cQ Q Q
5 1 5 11( )
2 16 32c c cQ Q Q
5
16cQ
6 11
32cQ
1 11 21( )
2 32 64c c cQ Q Q
7
8 1
3cQ
1
3cQ
Nash equilibrium:
1 1 1 1 1 1Firm 1: ( ) Firm 2: ( )
2 3 3 2 3 3c c c c c cQ Q Q Q Q Q
Nash
Equilibrium
160
Cournot model of 3 identical firms
Firm 1 Firm 2 Firm 3
1 1
2cQ
0 0
2 1
2cQ
1 1 1( 0)
2 2 4c c cQ Q Q
0
3 1
2cQ
1
4cQ
1 1 1 1( )
2 2 4 8c c c cQ Q Q Q
4 1 1 1 5( )
2 4 8 16c c c cQ Q Q Q
1
4cQ
1
8cQ
5 5
16cQ
1 5 1 9( )
2 16 8 32c c c cQ Q Q Q
1
8cQ
6 5
16cQ
9
32cQ
1 5 9 13( )
2 16 32 64c c c cQ Q Q Q
7
8 1 Nash Equilibrium
4cQ
1 Nash Equilibrium
4cQ
1 Nash Equilibrium
4cQ
Nash Equilibrium
1 1 1 1Firm 1: ( )
2 4 4 4c c c cQ Q Q Q
1 1 1 1Firm 2: ( )
2 4 4 4c c c cQ Q Q Q
1 1 1 1Firm 3: ( )
2 4 4 4c c c cQ Q Q Q
Cournot model with n identical firms
Since the firms are identical, hence the shares are the same.
Let x be firm is share.
Firm is share: 1
( ... )2
cx Q x x
1 'n x s
1 ( 1) 1[ ( 1) ] [1 ]
2 2 2 2 2
* Nash Equilibrium1
c c
c
c
c
Q Qn nx Q n x x x
n
Total output1
c
nQ
n
As , total output cn Q Competitive market is the limiting case of oligopoly.
161
Example
Consider a duopoly market with a demand function : 100D P Q . Let the cost function of firm 1 be 2
1 1 110 4TC Q Q and the cost function of firm 2 be 2
2 2 210 2TC Q Q .
Firm 1: 2 2# #
1 1 2 1 1 1 1 1 1 2max [100 ( )] [10 4 ] 10 96 2Q Q Q Q Q Q Q QQ
FOC: #1 1 21
96 4 0d
Q QdQ
Reaction function: #
21
96*
4
Firm 2: 2 2# #
2 1 2 2 2 2 2 2 1 2max [100 ( )] [10 2 ] 10 99 3Q Q Q Q Q Q Q Q Q
FOC: #2 2 12
99 6 0d
Q QdQ
Reaction function: #
12
99*
6
Nash equilibrium: # #
1 1 2 2* , *Q Q Q Q
1 2
1 2
96 4 * * 0
99 * 6 * 0
Q Q
Q Q
1 2
477 300* 20.74, * 13.04
23 23Q Q
Collusion equilibrium
1 2
2 2
1 2 1 2 1 1 2 2,
2 2
1 2 1 2 1 2
max [100 ( )]( ) [10 4 ] [10 2 ]
20 96 99 2 3 2
Q QQ Q Q Q Q Q Q Q
Q Q Q Q Q Q
FOC:
1 2
1
1 2
2
96 4 2 0
99 2 6 0
Q QQ
Q QQ
1 2 1 2
189 102 291* , * * *
10 10 10Q Q Q Q Q
Stackelberg Equilibrium (Firm 1 is the leader)
211 1 1 1
99max [100 ] [10 4 ]
6
QQ Q Q Q
1 2
477 567* 21.68, * 12.89
22 44Q Q
162
The dominant Firm Model (price leadership)
1 dominant firm: behaves as if it is a monopoly price-setter a few small firms: behave as if they are competitive firms price-takers P
Ssmall
MCbig
P*
Dbig = Residual demand facing the dominant
firm
D
MRbig
Q
Qbig* Qsmall*
Example:
10 small firms: 22 2TC Q Q P S(one firm)
1 dominant firm: 210TC Q Q
Demand function: 100Q P S(10 firms)
small firm: 22 2 2 2TC Q Q MC Q
For a competitive firm, MC curve is the supply curve.
2: 2 2 1
2 2small
P PS P Q Q
Q
10 small firms: 10( 1) 5 102
PQ P
Residual demand for the big firm:110
(100 ) (5 10) 110 66
QQ P P P P
Problem facing the big firm: 2 2
2 2110 110 104 7max ( ) (10 ) 10 106 6 6 6 6Q
Q Q QQ Q Q Q Q Q Q
FOC:
52110
7 104 52 110 718 35970 * *3 6 7 6 6 42 21
big
d QQ Q P
dQ
359
359 317211 1 12 2 42 42
small
PQ
163
Kinked demand curve
Assumptions:
1. When a firm lowers its price, every competitor will follow (lower its price).
2. When a firm raises its price, no competitor will follow (raise its price).
P
MC
D
MR
Q
Q*
Prediction:
When MC changes slightly, * and * will not be affectedP Q .
164
Monopolistic Competition
Properties of a monopolistic competitive market
1. Many sellers (firms)
2. Many buyers (consumers)
3. Heterogeneous product
4. Free entry/exit
Short-run equilibrium ( 0 ) P MC
P* AC D
MR
Q Q*
Long run equilibrium ( 0 )
P
MC
AC
P*
D
MR
Q
Q*
165
A Model of Monopolistic Competition
Consumers
Consider a single (representative) consumer whose preferences exhibit the love-for variety property.
The utility function of the representative consumer is given by a constant-elasticity-of substitution
(CES) utility function:
1 2
1
( , ,..., )M
M i
i
U x x x x
where M is a large number.
This type of utility exhibits love for variety since the marginal utility of each brand at a zero
consumption level is infinite.
The indifference curves (for the 2 goods case) are convex to the origin, indicating the consumers will
usually buy more than 1 good.
The indifference curve (for the 2 goods case) will cut the horizontal axis and vertical axis, therefore it is
possible that the consumers do not purchase the entire variety of products.
The consumers income (I) is composed of the total wages paid by the producing firms plus the sum of profits (if any).
( ) : i ix the profit of the firm producing brand i
We also normalize the wage rate to $1.
Problem:
1 2
1
1 1
max ( , ,..., )
. . ( )
i
M
M ix
i
M M
i i i i
i i
U x x x x
s t p x I L x
Lagrangian: 1 1
[ ]M M
i i i
i i
L x I p x
FOC for each brand i :
10, 1,2,...,
2i
i i
Lp i N
x x
2 2
1
4i
i
xp
166
22 2 11 1 12
1 1 2 2
2 2
1 11 1 2 1 12 2
1 2
2 2 2 2
1 1 1 11 1 1 1 1 1
2 2
2
1 1 1211
1
1 1 1... ...
2 2 2
( ) ... ( )
... ( ... )
1( ) or
1( )
M M i
iM M
M
i i n
i M
M M
M
Mi i
i i
pp x p x x x
pp x p x p x
p pI p x p x p x p x I
p p
p p p pp x x x I x p I
p p p p
Ix p I x
Pp
P
2
1
1
( )i M
i
i i
Ix
pP
From the FOC, we have 1
2 i ip x
1
2
2 1 12
1
1
1 1 1
1 122
114 ( )( )
4
M
i ii M M
ii Mi ii i
ii
i i
p Ix
IIpp
p ppp
Ip
When M is large, 0ip
3
2 2 3
2 3
2 2
1 1( 2)
4 2
1( ) 2
12
4
ii
i i
i i i
i i i
i
xp
p p
x p p
p x p
p
Brand-producing firms
Each firm is produced by a single firm. All (potential) firms have identical technologies (identical cost
structure) .
The total cost of a firm producing ix units of brand i : if 0
( )0 if 0
i i
i i
i
F cx xTC x
x
Definition: The triplet , , , 1,..., }mc mc mc mci iM p x i M is called a Chamberlinian monopolistic-competition equilibrium if
1) Firms: Each firm behaves as a monopoly over its brand; that is, given the demand for brand i ,
2 2
1
4i
i
xp
, each firm i chooses mcix to solve the problem max ( ) ( )i
i i i i ix
p x x F cx
more firms iM x
167
2) Consumers: Each consumer takes his income and prices as given and solve the utility
optimization problem.
3) Free entry: Free entry of firms (brands) will result in each firm making zero profits;
( ) 0 1,2,...,mci ix i M .
4) Resource constraint: Labor demanded for production equals the total labor supply,
( ) total wage bill $1N
i
i
F cx L L
,
2mcip c
( )i ii
FAC x c
x
( )i iMC x c
( , )i iD p M ix
mc
i
Fx
c
The demand facing each (existing) brand-producing firm depends on the total number of brands in the
industry, M . When M , the demand facing each brand-producing firm shifts downward, reflecting the fact that consumers partially substitute higher consumption levels of each brand with a lower
consumption spread over a large number of brands.
Therefore, free entry increases the number of brands until the demand facing each firm becomes tangent
to the firm's average cost function. At this point, each existing brand-producing firm makes zero profit,
and entry stops.
The equilibrium condition in which demand becomes tangent to the average cost of each firm is known
as Chamberlin's tangency condition.
Note:
i) In equilibrium the price of each brand equals average cost.
ii) In equilibrium all brand-producing firms produce on the downward sloping part of the average
cost curve. Thus, firms do not minimize average cost under a monopolistic -competition market
structure.
168
Solving for a monopolistic-competition equilibrium
A firm's profit-maximization problem is MR MC .
1 1( ) (1 ) (1 ) ( )
2 2
ii i i i i
pMR x MC p p c MC x
Hence, the equilibrium price of each brand is given by 2 (twice the MC)mcip c .
The zero-profit condition 0 ( ) ( ) (2 )mc mc mc mc mc mci i i i i i iF
x p c x F c c x F cx F xc
The resource-constraint condition ( ) ( )2
mc
i
F LM F cx M F c L M
c F
Proposition
1) In a monopolistic competition equilibrium with strictly positive fixed and marginal cost, only a
finite number of brands will be produced. The equilibrium is given by
2 ; ; 2
mc mc mc
i i
F Lp c x M
c F
2) When the fixed cost is large, there will be a low variety of brands, but each brand will be
produced/consumed in a large quantity.
When the fixed cost is small. there will be a large variety of brands, and each will be
produced/consumed in a small quantity.
169
Monopsony
Properties of a monopsony:
1. many sellers (wokers)
2. 1 buyer (firm)
P MCL S 1
2
8
P*=6
D
2 3 Q
Example:
Lau's company is the only company hiring economists in Lau's Island. Suppose its demand function is
D: 10P Q . Suppose the supply function for economists is S: 4P Q
Equilibrium
: 4S P Q Total cost of hiring2(4 ) 4Q Q Q Q
Marginal cost of labor 4 2Q
maximizing condition: Marginal cost of labor
4 2 10 3 6 * 2
*(wage*) 6
D
Q Q Q Q
P
A minimum wage of $7 will eliminate the DWL.
P MCL S
P'=$7
D
3 Q
$2 1$1
2DWL
If minimum wage $7 , # no DWLCQ Q
170
Example: Rent-seeking behavior
In many situations economic agents are willing to spend resources in pursuit of some economic gains.
For example: lobbying for tax breaks or trade protection.
2-player game
Consider 2 players who split an economic rent of $R . Each player spends money in pursuit of the rent.
Let 1 2 and x x represent the 2 players respective expenditures.
Assume the share of 1: 111 2
xs
x x
Assume the share of 2: 221 2
xs
x x
Problem of 1:
1
11 1 1 1
1 2
max ( )x
xs R x R x
x x
FOC: 1 1 2 1 22 2
1 1 2 1 2
( )[ ] 1 0 1 0
( ) ( )
x x x x RR
x x x x x
2
2 1 2
1 2 2
1 2 2
( )
* reaction function (nonlinear)
x R x x
x x x R
x x R x
Similarly, 2 1 1* reaction function (nonlinear)x x R x
Nash equilibrium: 1 1 2 2* and *x x x x
1 2 2 1 1 1 1
1 1 1
1 1 1
1 1 1
1 1
2
1 1
1 2
( ) ( )
( )
( )
2
4
* Similarly, *4 4
x x R x x R x R x R x
x R x R x R
x R x R x R
x R x x
x R x
x R x
R Rx x
4R
4
R
171
Production with Teams
The inability to monitor a worker's effort also generates an inefficiency when the output of the firm
depends on the effort levels of all workers assigned to work on a certain project(which we call the joint
effort of a team). This type of externality is called the free-rider effect, in which a worker, knowing
that all other workers in a team are putting a lot of effort into the project, will have an incentive not to
work hard, given that the group as a whole is rewarded on the value of the project, that is, when the
individual workers are not rewarded according to their individual effort levels.
Consider a lab developing the future product whose value is V.
In the lab there are N scientists (workers) who work on the project.
: the effort put in by scientist , 1,2,..,ie i i N
The value of the jointly developed product depends on the effort levels of all the N scientists and is
given by 1
(A)N
i
i
V e
The equation (A) can be viewed as a production function where the inputs are the efforts put out by the
scientists.
: the compensation given to scientist after the project is completed.iw i
We assume that the value of the product is distributed to the workers so 1
n
i
i
w V
.
All scientists have identical preferences: , 1,2,..., (B)i i iU w e i N
Optimal effort levels
Assume each scientist can observe the efforts of his other colleagues, and we suppose that they collude
to maximize their utility levels.
We now wish to calculate what the optimal symmetric allocations of effort and output shares (wages)
are, and therefore, we set and for every 1,2,...,i iV
e e w w i NN
.
If we substitute into (B), the representative effort level *e that maximizes a representative worker's utility.
max V N e
w e e e e eN N
172
FOC:( ) 1 1 1
1 0 2 1 *2 42
d w ee e e
de e
Note:
If the workers can (theoretically) collude, observe each other's effort, and adjust their efforts to
maximize their utility, each should put out 1
*4
e level of effort, and the resulting total value would be
1*
4 2
NV N
173
The equal-division economic mechanism
Suppose that the manager of this firm rewards the scientists according to their equal share of the total
value of output.
Let us suppose that the manager sets iV
wN
.
Assume that each scientist takes the effort levels of his colleagues as given and chooses his effort level
to maximize his utility (B).
Problem:
max j ij i
i i
e eU e
N
FOC:2
1 1 11 0 2 1 * (C)
2 42
nii i i
i i
UN e e e e e
e N NN e
Proposition:
Under the equal-division rule:
1. If the team consists of a single worker, the worker will provide the optimal level of effort. That
is, if 1
1, then *4
nN e e .
2. If the team consists of more than 1 worker, each worker would devote less than the optimal level
of efforts. That is, if 1
1, then *4
nN e e .
3. The larger the team is, the lower will be the effort put out by each worker (each would have a
greater incentive not to work hard). That is, as N increases, ne decreases.
Note:
1. This proposition shows that offering the workers equal shares of the value of the output is
insufficient to induce them to devote the optimal level of effort to their work. So, why not offer them a
higher share of the output? Although it may be possible to induce them to work harder, if all workers
are offered a higher share of the output (i.e. iV
wN
), the total wage bill will exceed the value of output.
2. The above equilibrium is called a Nash equilibrium.
174
Effect of the size of the workforce on the total output as well as the worker's welfare level
1. 2
1 1( ) ( ) (D)
4 2
n nC A V N e NN
2
1 1 1 1* *
4 4 2 2 2
n n N NV V N e N e N NN
2. 2 2
2 3 3
11 1 12( ) ( )
4 2 4
1 1 10 if 1
2 2 2
nn
i
i
VC B U e
N N N N N
U NN
N N N N
Proposition:
1. An increase in the number of workers on the team will increase the difference between the
optimal output level and the Nash equilibrium output level.
2. An increase in the number of workers will reduce the welfare levels of each worker.
Note:
1. Part 2 of the proposition shows that the free-rider effect intensifies when the number of workers
increases, causing a further deviation from the optimal output level.
2. The optimal output level ( *2
NV ) increases with the team size, but the equilibrium level
(1
2
nV ) does not with the size of the team.
An economic mechanism that works
Following Holmstrom (1982)1, we now discuss a (rather tough) incentive mechanism that would induce
all the N workers to put forth the optimal effort level.
Suppose the team sets the following rule: If the team as a group achieves the optimal output level *V ,
then each team member receives *V
N. If the teams output is different from *V , then all team members
receive 0.
1
* if *
0 otherwise
N
i
ii
Ve V
Nw
1 Holmstrom, B. (1982): "Moral Hazard in Teams", Bell Journal of Economics, 13, pp. 324-340.
175
This mechanism makes each team member responsible for the entire output level of the team. Under the
equal-division rule, the marginal effect of each team member is lower than the marginal social value. In
this mechanism, however, the marginal value of each worker effort is the entire enterprise.
Does this work?
1. This allocation mechanism may suffer from a problem known as time inconsistency. That is,
this mechanism can work only if after each time the output is produced, the manager fires and
replaces all the workers.
2. If workers continue to work on a new project, it seems unlikely that workers would agree to let
the manager confiscate all the output just because somebody has intentionally or unintentionally
deviated from the optimal effort level.
3. Even if some deviation has occurred, it looks as if the workers would be able to negotiate with
manager or among themselves a re-division of the output, given that some output has already
been produced. Since workers anticipate that the manager will re-negotiate the contract, the
workers may not take this contract too seriously.
176
General Equilibrium II
Example: Suppose there are 2 price-taking consumers, consumer A and consumer B. Consumer A
has an initial endowment of ( , ) (1,0)X Y and a utility function of 1( , ) a aU X Y X Y . Consumer B has
an initial endowment of ( , ) (0,1)X Y and a utility function of 1( , ) b bU X Y X Y .
Calculate the competitive equilibrium price.
Consumer A:
,max ln (1 ) ln
s.t. (1) (0)
X Y
X Y X Y X
a X a Y
P X P Y P P P
Cobb-Douglas Utility function
(1 )
1and (1 ) (1 )
(1 )
X X A
XY X A
Y
aP X M aP X a
a a
a PP Y M a P Y a
a a P
Excess demand:
(1) 1 0
(1 ) (0) (1 ) 0
A
X
A X XY
Y Y
E a a
P PE a a
P P
Consumer B:
,max ln (1 ) ln
s.t. (0) (1)
X Y
X Y X Y Y
b X b Y
P X P Y P P P
Cobb-Douglas Utility function
(1 )
1and (1 ) 1
(1 )
YX Y B
X
Y Y B
PbP X M bP X b
b b P
bP Y M b P Y b
b b
Excess demand:
(0) 0
1 (1) 0
B Y YX
X X
B
Y
P PE b b
P P
E b b
177
Market:
( 1) ( )
(1 ) ( )
A B YX X X
X
A B XY Y Y
Y
PE E E a b
P
PE E E a b
P
At equilibrium: 0X YE E
10 1 0 (1)
0 (1 ) 0 (2)1
Y YX
X X
X XY
Y Y
P P aE a b
P P b
P P bE a b
P P a
Note that (1) and (2) are exactly the same, hence it is impossible to find XP and YP separately.
178
Example: Suppose there are 2 price-taking consumers, consumer A and consumer B. Consumer A
has an initial endowment of ( , ) (2,2)X Y and a utility function of ( , ) min( , )U X Y X Y . Consumer B
has an initial endowment of ( , ) (1,2)X Y and a utility function of 1/ 2 1/ 2( , )U X Y X Y .
Calculate the competitive equilibrium price.
A: A( , ) min( , ) (2,2) 2 2A X YU X Y X Y M P P
2 2( , ) min( , ) * * =2 X Y
X Y
P PU X Y X Y X Y X Y
P P
The individual will buy his endowment!!
(2) 2 0
(2) 2 0
A
X
A
Y
E
E
B: 1 1
2 2( , ) (1,2) 2B B X YU X Y X Y M P P
1
21 12 *1 1 2 2 2 2
2 2
1
212 * 11 1 2 2 2 2
2 2
X Y YX
X X X
X Y XY
Y Y Y
P P PMP X M M X
P P P
P P PMP Y M M Y
P P P
1 1( ) 12 2
( 1) 2 12 2
B Y YX
X X
B X XY
Y Y
P PE
P P
P PE
P P
Market:
1 1(0) ( )
2 2
(0) ( 1) 12 2
A B Y YX X X
X X
A B X XY Y Y
Y Y
P PE E E
P P
P PE E E
P P
At equilibrium:
1 10
2 2
1 0 1 22 2
Y YX
X X
X X XY
Y Y Y
P PE
P P
P P PE
P P P
179
Walras Law: 1
0n
j j
j
P E
Proof:
From the budget constraint of individual i , we have 1
0 1,...,n
i
j j
j
P E i m
1 1 1 1 1
( ) 0 ( )m n n m n
i i
j j j j j j
i j j i j
P E P E P E
Note: i) By the Walras Law, 0 0X YE E
ii) If 1 markets are in equilibrium the remaining market is also in equilibriumn .
iii) There is no way to solve out the n individual prices, we can only solve out 1n relative prices.
Pareto optimality
Example: Consider an economy composed of 2 1N individuals, N own one right shoe and 1N own one left shoe. Shoes are indivisible. Every one has the same utility function which is
min[ , ]U R L , where and R L are the quantities of right and left shoes consumed respectively. Find
the set of Pareto Optimal allocations.
Solution: A Pareto Optimal allocation is an allocation where all the shoes are in pairs. (It does not
matter who own the shoes. It is possible that a single individual own all N pairs.)
There will be an individual which owns the remaining left shoe.
180
Pareto Optimality in a Consumption Economy
,max ( , )
. . ( , )
A A
A A AX Y
B B B B
A B
A B
U X Y
s t U X Y U
X X X
Y Y Y
Method 1
,max ( , )
. . ( , )
A A
A A AX Y
B A A B
U X Y
s t U X X Y Y U
( , ) [ ( , )]A A A B B A AL U X Y U U X X Y Y
FOC:
0
0
A B
A A B
A B
A A B
U UL
X X X
U UL
Y Y Y
A A
A A
B B
B B
U U
X Y
U U
X Y
B A
B A
B A
B A
U U
Y Y
U U
X X
A B
XY XYMRS MRS
Method 2
( , ) [ ( , )] [ ] [ ]A A A B B B B A B A BL U X Y U U X Y X X X Y Y Y
FOC:
0
0
A
A A
A
A A
UL
X X
UL
Y Y
A
A
A
A
U
X
U
Y
0
0
B
B B
B
B B
UL
X X
UL
Y Y
B
B
B
B
U
X
U
Y
A B
A B
A B
A B
U U
X X
U U
Y Y
B A
B A
B A
B A
U U
Y Y
U U
X X
A B
XY XYMRS MRS
181
Pareto Optimality in a Production Economy
,max ( , )
. . ( , )
X X
X X XK L
Y Y Y
X Y
X Y
F K L
s t F K Y Y
K K K
L L L
or
,max ( , )
. . ( , )
X X
X X XK L
Y X X
F K L
s t F K K L L Y
( , ) [ ( , )]X X X Y X XF K L Y F K K L L
FOC:
0
0
X Y
X X Y
X Y
X X B
F F
K K K
F F
L L L
X X Y X
X X Y X
Y Y Y X
Y Y Y X
F F F F
K L L L
F F F F
K L K K
X Y
KL KLMRTS MRTS
182
Pareto Optimality in General
Consider an economy with m consumers, n producers, r primary factors and s produced goods.
Assume each consumer consumes all produced goods, and each producer user all primary factors and
produces all goods.
The consumers utility functions are 1* * 1 1* *( , ... , , , ... , ) 1,...,s r ri i i i i i i iU U X X Y Y Y Y i m
The firms implicit production functions are 1 1( ,..., , ,..., ) 0 1,...,s rh h h h hF X X Y Y h n
At equilibrium,
a) the aggregate amount of primary factors supplied by the consumers is equal to the aggregate
amounts used by producers: *
1 1
1,...,m n
j j
i h
i h
Y Y j r
b) the aggregate consumption levels of produced goods is equal to their aggregate output levels:
*
1 1
1,...,m n
j j
i h
i h
X X k s
quantity of
produced goods kX consumed
by consumer i
j
iY : endowment of the jth
primary factor of
consumer i *j
iY : quantity of primary input j that
consumer i supplied to the producers *j j
i iY Y : quantity of primary input j that
consumer i consumes
quantity of
produced goods kX produced by
firm h
quantity of input k that firm h uses
183
To attain Pareto Optimality, consumer i is to solve the following problem:
1* * 1* *1 1 1 1
1* * 1 1* *
1 1 1 1 1 1 1,..., , ,...,
1* * 1 1* *
1
max ( , ... , , , ... , )
. . ( , ... , , , ... , ) 2,...,
( ,...,
s r
s r r
X X Y Y
s r r
i i i i i i i i
h h h
U X X Y Y Y Y
s t U X X Y Y Y Y U i m
F X X
1, ,..., ) 0 1,...,s rh hY Y h n
*
1 1
1,...,m n
j j
i h
i h
Y Y j r
*
1 1
1,...,m n
k k
i h
i h
X X k s
1* * 1 1* * 1* * 1 1* *
1 1 1 1 1 1 1
2
1 1 * *
1 1 1 1 1
( , ... , , , ... , ) ( , ... , , , ... , )
+ ( ,..., , ,..., ) +
ms r r s r r
i i i i i i i i i
i
r m n ns r j j k k
h h h h h h j i h k h i
j i h h i
L U X X Y Y Y Y U U X X Y Y Y Y
F X X Y Y Y Y X X
1 1
n s m
h k
FOC:
1
* *
1 1
0 1,...,kk kUL
k sX X
* *0 2,..., ; 1,...,ii kk k
i i
ULi m k s
X X
0 1,..., ; 1,...,hh kk kh h
FLh n k s
X X
1
* *
1 1 1
0 1,...,( )
jj j j
ULj s
Y Y Y
* *0 2,..., ; 1,...,
( )
ii jj j j
i i i
ULi m j s
Y Y Y
0 1,..., ; 1,...,hh jj jh h
FLh n j s
Y X
The constraints also have to be satisfied.
From the FOCs, we have the following conditions for Pareto Optimality:
a) 1 1... ...m njk jk jk jkMRS MRS MRTS MRTS for each pair of produced goods
1 1
**
1 1
1 1
* *
1 1
... ... , 1,...,
m n
j jj jj m n
m nkk kk k
m n
U FU F
X XX Xj k s
U U F F
X XX X
184
b) 1 1... ...m njk jk jk jkMRS MRS MRTS MRTS for each pair of primary goods
1 1
* ** *
1 1 1
1 1 1
* * * *
1 1 1
( )( )... ... , 1,...,
( ) ( )
m n
j j jj j jj m m n
nkk k k k k k
m m n
U FU F
Y Y YY Y Yj k r
U U F F
Y Y Y Y Y Y
c) between factors and commodities (consumers)
between factors and commodities (producers)
MRS
MRTS
1 1
* ** *
1 1 1
1 1
* *
1 1
( )( )... ... 1,..., ; 1,...,
m n
j j jj j jj m m n
m nkk kk k
m n
U FU F
Y Y YY Y Yj r k s
U U F F
X XX X
185
External Effects in Consumption
Interdependent utility function
Assume that the utility level of one consumer depends upon the consumption of another. Extreme
altruism may increase the satisfaction of the ith
consumer if the consumption level of the jth
consumer is
raised. The desire to keep up with the Joneses may have the opposite effect.
,max ( , , , )
. . ( , , , )
A A
A A A B BX Y
B A A B B B
A B
A B
U X Y X Y
s t U X Y X Y U
X X X
Y Y Y
or
,max ( , , , )
. . ( , , , )
A A
A A A A AX Y
B A A A A B
U X Y X X Y Y
s t U X Y X X Y Y U
( , , , ) [ ( , , , )]A A A A A B B A A A AU X Y X X Y Y U U X Y X X Y Y
FOC:
0
0
A A B B
A A B A B
A A B B
A A B A B
U U U U
X X X X X
U U U U
Y Y Y Y Y
A A A A A A A A
A B A B A B A B
B B B B B B B B
A B A B B A B A
U U U U U U U U
X X Y Y Y Y X X
U U U U U U U U
X X Y Y Y Y X X
The Pareto optimal allocations are no longer points where the indifference curves are tangent to
each other!!
186
Public Goods
2 consumers: and A B who produce and consume a public good X and a private good Y , using a given amount of labor
,max ( , )
. . ( , )
A
A A AX Y
B B B
U U X Y
s t U U X Y
( )
( )
X
A B Y
X X L
Y Y Y L
production function
X YL L L
1 2 3 4( , ) [ ( , )] [ ( )] [ ( )] [ ]A A B B B X A B Y X YU X Y U U X Y X X L Y Y Y L L L L
FOC:
1 2 0A BU U
X X X
(1)
3 0A
A A
U
Y Y
(2)
1 3 0B
B B
U
Y Y
(3)
2 4 0X X
X
L L
(4)
3 4 0Y Y
Y
L L
(5)
(4) and (5) 24 2 3
3
Y
X Y
X
Y
LX YMRT
XL L
L
(6)
(2) and (3) 31
A
A
B B
B B
U
Y
U U
Y Y
(7)
2 1(1)
A
A B A A B
B
B
U
U U U Y U
UX X X X
Y
marginal rate of transformation
187
(2), (6) and (8) 2
3
A
A A B
BA B
B
A A B
A A B
U
U Y U
UX X U UY X X MRT
U U U
Y Y Y
A B
YX YXMRS MRS MRT
188
Theory of Second Best
Consider a simplified system with 1 consumer, 1 implicit production function, n commodities, and a fixed supply of 1 primary factor not desired by the consumer.
The necessary conditions for Pareto optimality are obtained by maximizing the consumers utility subject to the production function.
0
1 1( ,..., ) ( ,..., , )n nL U x x F x x x
FOC:
0 1,...,i ii
LU F i n
x
or i i
j j
U FMRS MRTS
U F
Assume that institutional conditions prevent the attainment of one of the FOC.
Lets say, 1 1 0 where U kF k (**)
The conditions for second-best welfare optimum are obtained by maximizing utility subject to the
aggregate production function and (**).
0
1 1 1 1( ,..., ) ( ,..., , ) ( )n nL U x x F x x x U kF
FOC:
1 1
0
1
1 1
( ) 0 1,...,
( ,..., , ) 0
0
i i i i
i
n
LU F U kF i n
x
LF x x x
LU F
1 1
1 1
( ) , 1,...,
( )
i i i i
j j j j
U F U kFi j n
U F U kF
In general, nothing is known a prior about the signs of the cross partial derivatives. Therefore,
the usual conditions for Pareto optimality no longer hold.
The theory of second best has been used to questions the desirability of partial-equilibrium
policies that might be used to attain the Pareto conditions on a piecemeal basis for markets
considered in isolation.