Study Guide CH 16

C H A P T E R
16
General Equilibrium
Chapter 14 introduced the positive idea of equilibrium in the context of a competi-

tive environment — and Chapter 15 moved onto the more normative assessment of
a competitive equilibrium within the context of the first welfare theorem. We now
extend the basic insights from these chapters to economies that are more “gen-
eral” in the sense that we don’t look just at one isolated market but rather treat
the economy as an interconnected web of markets. It admittedly may at times not
seem like we are extending the analysis to something more “general” — because
the economies we discuss in this chapter seem very “simple” in that they only in-
volve a few individuals and often abstract away from issues like production. But
these “simple” economies can be extended to much more complex models with
many goods, many individuals, many production processes, etc. And the basic in-
sights that emerge from the “simple” models continue to hold in these more com-
plicated models that indeed are more “general” than the partial equilibrium models
of Chapters 14 and 15.
Chapter Highlights
The main points of the chapter are:
1. An exchange economy is an economy in which individuals trade what they

own but no production takes place. Typically, individuals can improve their
welfare in such economies by engaging in mutually beneficial trades. The set
of efficient allocations of goods in such an economy is known as the contract
curve.
2. A competitive equilibrium in an exchange economy consists of a set of prices

and an allocation of the goods in the economy such that all individuals would
agree to trade to that allocation at these prices. (The equilibrium is “compet-
itive” because all individuals are assumed to be price takers.)
3. The equilibrium allocation of goods always lies on the contract curve (and
is thus efficient). This is our generalization of the first welfare theorem. It
1
General Equilibrium 2
furthermore lies in the core of the economy — where the core is defined as
the set of allocations under which no coalition of individuals could leave the
economy with their endowment and do better on their own. (In the 2-person
exchange economy, the core allocations are equal to the contract curve allo-
cations that lie in the region of mutually beneficial trades.)
4. If governments can use lump sum transfers, then any allocation on the con-
tract curve can become a competitive equilibrium allocation after appropri-
ate lump sum transfers have been made. This is what is known as the second
welfare theorem. If, however, the government can only use distortionary
taxes, then a tradeoff emerges between notions of “equity” and “efficiency”
because it is no longer possible for the government to redistribute and ex-
pect an efficient outcome.
5. Production can be introduced into general equilibrium economies, and the

same basic welfare results hold. The simplest example of such an economy
is the Robinson Crusoe economy where a single individual acts as producer,
worker and consumer.
16A Solutions to Within-Chapter-Exercises for

Part A
Exercise 16A.1
What would the Edgeworth Box for this example look like if oranges appeared
on the vertical and bananas on the horizontal axis?
Answer: The Edgeworth Box would then be 10 units (bananas) long and 13 units
(oranges) high, with the endowment point as pictured in panel (a) of Exercise Graph
16A.1.
Exercise 16A.2
What would the Edgeworth Box for this example look like if my wife’s axes had
the origin in the lower left corner and my axes had the origin in the upper right
corner?
Answer: The box would have the same dimensions, but the endowment point
would be located as pictured in panel (b) of Exercise Graph 16A.1 in the previous
within-chapter exercise solution.
3 16A. Solutions to Within-Chapter-Exercises for Part A
Exercise Graph 16A.1 : Edgeworth Boxes
Exercise 16A.3
True or False: Starting at point A, any mutually beneficial trade will involve me
trading bananas for oranges, and any trade of bananas for oranges will be mutually
beneficial. (Hint: Part of the statement is true and part is false.)
Answer: The first part of the statement is true — any mutually beneficial trade
from A will involve me giving up bananas and getting oranges. This is because
all the allocations of bananas and oranges that lie within the lens shape (and thus
above both of our indifference curves through A) involve more oranges and fewer
bananas for me. But it is not true that any trade of bananas for oranges will be
mutually beneficial — only those lying above both our indifference curves. Put dif-
ferently, the allocations that involve mutually beneficial trades from A do lie to the
southeast of A (with fewer bananas and more oranges for me), but there are allo-
cations that lie to the southeast of A that do not lie within the lens that represents
allocations which are better for both me and my wife.
Exercise 16A.4
In Chapter 6, we argued that consumers leave Wal-Mart with the same tastes
“at the margin” — i.e. with the same marginal rates of substitution between goods
that they have purchased, and that this fact implies that all gains from trade have
been exhausted. How is this similar to the condition for an efficient distribution of
an economy’s endowment in the exchange economy?
Answer: Once gains from trade have been exhausted in the Edgeworth Box, it is
similarly true that the marginal rates of substitution of the two individuals will be
equal to one another — just as when they come out of Wal-Mart after maximizing
subject to facing the same prices. Thus, in both cases, the tastes are the same at the
margin once all gains from trade have been exhausted. (This presumes that both
individuals will optimize at an interior solution — but the efficiency logic extends
to the case where individuals end up at corner solutions).
Exercise 16A.5
Starting at the point where my wife gets the entire endowment of the economy,
are there points in the Edgeworth Box that make my wife worse off without making
me better off (assuming that bananas and oranges are both essential goods for me)?
Answer: If both goods are essential, this means that the only way I can become
better off than I am at the bundle (0,0) is for me to get at least some of each of the
two goods. Thus, I do not become better off if you just take bananas away from my
wife and give them to me, nor am I better off if you take oranges away from her and
give them to me. Thus, any movement from (0,0) along either of the axes that refer
to me would make my wife worse off without making me better off.
Exercise 16A.6
Is the point on the upper right hand corner of the Edgeworth Box Pareto effi-
cient?
Answer: Yes — it is the point at which I get all bananas and oranges and my
wife gets none. If that is the allocation, then any move away from this point within
the box will make me worse off — and there is therefore no way to make someone
better off without making anyone else worse off.
Exercise 16A.7
If bananas and oranges are essential goods for both me and my wife, can any
points on the axes (other than those at the upper right and lower left corners of the
Edgeworth Box) be Pareto efficient?
Answer: No. Consider a bundle that lies on one of the axes. If both goods are
essential, it means that one of us is no better off than he/she would be if he/she
had nothing. That means that we could simply give everything to the other person
— make him/her better off without making the other worse off.
Exercise 16A.8
What would the contract curve look like if bananas and oranges were perfect
complements for both me and my wife? (Hint: It is an area rather than a “curve”.)
What if they were perfect complements for me and perfect substitutes for my wife?
Answer: The case of perfect complements for both of us is depicted in panel (a)
of Exercise Graph 16A.8 where u1 and u 1 are two indifference curves for me and
u2 and u 2 are two indifference curves for my wife. These are “tangent” along the
darkened parts which lie in the shaded region. And the shaded region is the region
in which all such “tangencies” between two of our indifference curves lie — i.e. it is
the contract curve. (Note that the region is bounded by the two 45 degree lines that
emanate from our two origins of the two sets of axes — because it is along these
45 degree lines that the corners of our indifference curves lie.) Panel (b) shows
the case where bananas and oranges are perfect complements for me and perfect
substitutes for my wife. Again, u1 and u 1 are two indifference curves for me, and u2
and u 2 are two indifference curves for my wife. These are now “tangent” just at the
corner of my indifference curves — making the 45 degree line emanating from my
origin the contract curve.
Exercise Graph 16A.8 : Contract Curves
Exercise 16A.9
What does the contract curve look like if bananas and oranges are perfects sub-
stitutes (one for one) for both me and my wife? (Hint: You should get a large area
within the Edgeworth Box as a result.)
Answer: The entire Edgeworth Box is then the contract curve. Because our in-
difference curves have the same slopes, there will be — for any arbitrary indiffer-
ence curve for me — an indifference curve for my wife that lies right on top of mine
and is thus “tangent” everywhere. Every allocation in the Edgeworth Box is there-
fore efficient. This should make intuitive sense: Suppose the two goods are Coke
and Pepsi and neither of us can tell Pepsi apart from Coke. In that case, Coke and
Pepsi are perfect substitutes — and the only thing we care about is how much Coke
and Pepsi we have in total. For any allocation in the Edgeworth Box, we will then
not be able to make one of us better off without making the other worse off — be-
cause we either keep the overall quantity of soft drinks the same by changing only
the mix of Coke and Pepsi in each of our bundles (thus making neither of us better
or worse off), or we will increase the total quantity of soft drinks for one of us (thus
making the other worse off). There is no way to make one of us better off without
making the other worse off.
Exercise 16A.10
What are the intercepts of this budget on my wife’s axes for oranges and ba-
nanas?
Answer: She begins at her endowment point E 2 = (10, 4), with 10 oranges and 4
bananas. If she sells the 4 bananas at a price of 1.5, she collects $6 — for which she
can buy 6 oranges at a price of 1. Thus, the intercept of her budget on the oranges
axis is 16. Alternatively, she could sell her 10 oranges for $10 at a price of 1, allowing
her to buy 6.67 bananas at a price of 1.5. Her bananas intercept is therefore 10.67.
Exercise 16A.11
Suppose both oranges and bananas are normal goods for both me and my wife.
Draw separate graphs for me and my wife — with the initial budget constraint when
the prices were both equal to 1 and the new budget constraint when the price of
bananas is raised to 1.5. Illustrate — using substitution and wealth effects — why
my demand for oranges will unambiguously increase and my wife’s demand for ba-
nanas will unambiguously decrease. Can you say unambiguously what will happen
to my demand for bananas and my wife’s demand for oranges?
Answer: This is illustrated in Exercise Graph 16A.11.
In panel (a), the graph for “me” is drawn — where the endowment point E 1
is the initial 3 oranges and 6 bananas and the solid original budget has slope −1
and passes through E 1 . The new slope when bananas are priced at 1.5 (relative to
oranges at 1) becomes −2/3 and must still pass through E 1 . The bundle A is opti-
mal on the original budget, and the move from A to A ′ is a substitution effect that
implies more consumption of oranges and less of bananas. Since both goods are
normal, the movement from the compensated budget to the new budget leads to
an increase in consumption of both goods relative to A ′ — i.e. a wealth effect in the
same direction as the substitution effect for oranges but not for bananas. We can
therefore unambiguously say that I will consume more oranges when the price of
bananas increases, but we cannot say whether I will consume more bananas (since
we have offsetting substitution and wealth effects on the bananas axis). In panel
(b), the graph for “my wife” is drawn, with the solid budget through E 2 again rep-
resenting the initial budget with slope −1. The new budget again passes through
E 2 but has shallower slope −2/3 — giving us a substitution effect that implies more
oranges and fewer bananas (just as in panel (a)). For my wife, however, the move-
Exercise Graph 16A.11 : Price Changes in the Edgeworth Box
ment from the compensated budget to the new budget is a decrease in income —
so the wealth effect points in the opposite direction (relative to the substitution ef-
fect) for oranges and in the same direction for bananas. Thus, for my wife we can
unambiguously say that her consumption of bananas will fall as the price of ba-
nanas increases, but we cannot be sure whether her consumption of oranges will
increase or decrease.
Exercise 16A.12
Suppose you had decided to leave the price of bananas at 1 and to rather change
the price of oranges. What price (for oranges) would you have to set in order to
achieve the same equilibrium outcome?
Answer: The equilibrium outcome will be achieved for any set of prices that
result in a slope of the budget equal to −2/3. With oranges denoting good 1 and
bananas denoting good 2, the slope of the budget constraint is −p 1 /p 2 . If we keep
the price of bananas at 1, we therefore have to set the price of oranges (p 1 ) to 2/3 in
order to get the right slope for the budget constraint.
Exercise 16A.13
Suppose you set the price of oranges equal to 2 instead of 1. What price for
bananas will result in the same equilibrium outcome?
Answer: The equilibrium outcome will be achieved for any set of prices that
result in a slope of the budget equal to −2/3. With oranges denoting good 1 and
bananas denoting good 2, the slope of the budget constraint is −p 1 /p 2 . If we keep
the price of oranges at p 1 = 2, we therefore have to set the price of bananas (p 2 ) to

3 in order to get the right slope for the budget constraint.
Exercise 16A.14
True or False: When the First Welfare Theorem holds, competitive equilibria in
an exchange economy result in allocations that lie on the contract curve but not
necessarily in the core.
Answer: This is false — the competitive equilibria in an exchange economy re-
sult in allocations that lie in the core but not necessarily on the contract curve. This
is because no one would agree to trade at equilibrium prices if she became worse
off — and lots of allocations on the contract curve have one person worse off than
she is at the endowment point. Competitive equilibrium allocations are, however,
efficient in addition to making no one worse of than she is at the endowment point
— which implies that such allocations are on the portion of the contract curve that
lies in the lens between the indifference curves that pass through the endowment
point — i.e. in the core.
Exercise 16A.15
Can you think of other redistributions of oranges and bananas that would be
“appropriate” for insuring that D is the competitive equilibrium outcome?
Answer: Any redistribution that results in a new endowment point that lies on
the green budget (that is tangent to both indifference curves at D) would work. For
instance, transferring 4 bananas and 3 oranges to me would leave my wife with
7 oranges and no bananas — which is the intercept of the green budget with her
horizontal axis (oranges). Alternatively, you could give me 10 oranges and take
3 bananas away — leaving my wife with 7 bananas and no oranges. This would
correspond to the allocation on the bananas axis. Of course there are many other
possibilities in between these that would also work.
Exercise 16A.16
Does this production frontier exhibit increasing, decreasing or constant returns

to scale? Is the marginal product of labor increasing, constant or decreasing?
Answer: This production frontier has decreasing returns to scale and diminish-
ing marginal product of labor throughout.
Exercise 16A.17
As drawn, which of our usual assumptions about tastes — rationality, convexity,

monotonicity, continuity — are violated?
Answer: The monotonicity assumption is violated — more is not better since I
prefer less labor over more, all else being equal.
Exercise 16A.18
In votes on referenda on school vouchers, researchers have found that renters

vote differently than homeowners. Consider a renter and homeowner in a bad pub-
lic school district. Who do you think will be more likely to favor the introduction of
school vouchers and who do you think will be more likely to be opposed?
Answer: The homeowner should be more likely to favor vouchers as demand
for housing in bad public school districts will increase and thus drive up housing
prices. Put differently, the homeowner in the bad public school district may favor
private school vouchers because he expects it to increase the value of his home.
The renter, on the other hand, would experience higher rents without the capital
gain that comes from homeownership. Thus, the renter in the bad public school
district is less likely to favor private school vouchers.
Exercise 16A.19
How do you think the elderly (who do not have children in school but who
typically do own a home) will vote differently on school vouchers depending on
whether they currently live in a good or bad public school district?
Answer: Demand for housing in bad public school districts will increase while
demand for housing in good school districts will decrease — implying that housing
prices will rise in bad school districts and fall in good school districts. The elderly
would therefore be more likely to favor vouchers if they own a house in bad public
school districts than if they owned a house in a good public school district.
Exercise 16A.20
If you were considering opening up a private school following the introduction

of private school vouchers, would you be more likely to open your school in poor
or in rich districts?
Answer: You would be more likely to open it in a poor district where public
schools are likely to be relatively poor. This is because you will be able to attract
no only parents who currently send their children to the local public school, but
you will also be able to attract parents who currently live in better school districts
but are willing to move to get better housing deals if they can send their children to
private schools. (Some research suggests that the latter demand for private schools
may be twice as large as the former.)
Exercise 16A.21
Suppose two different voucher proposals were on the table: The first proposal
limits eligibility for vouchers only to families below the poverty line, while the sec-
ond limits eligibility to those families who live in bad public school districts. Which
policy is more likely to lead to general equilibrium effects in housing markets?
Answer: The second policy that targets geographically to bad (typically poor)
public school districts will cause general equilibrium effects in housing markets
because families — regardless of income — can qualify for the vouchers by moving
to the targeted areas. The first policy targets primarily people who already tend to
live in bad public school districts — and is therefore less likely to induce families to
move (since middle income families could not qualify for the vouchers even if they
moved.)
11 16B. Solutions to Within-Chapter-Exercises for Part B
16B Solutions to Within-Chapter-Exercises for

Part B
Exercise 16B.1
Can you see how the Edgeworth Box we drew in Section A contains all the allo-
cations in this set?
Answer: The Edgeworth Box has length equal to the sum of the endowments of
good 1 and height equal to the sum of the endowments of good 2 — i.e. length equal
to x11 + x12 = 13 and height equal to x21 + x22 = 10 — just as is stated in the definition
for F A. The set F A is then simply the set of all the possible ways of dividing 13 units
of x1 and 10 units of x2 between two people — which is exactly what the Edgeworth
Box is.
Exercise 16B.2
True or False: The Edgeworth Box represents a graphical technique that allows
us to graph in a 2-dimensional picture points that lie in four dimensions.
Answer: This is true as is easily seen in the definition of the set F A which is in
fact the definition of the Edgeworth Box. This definition has points that have four
components — i.e. that lie in four dimensions. The clever trick used in graphing
the Edgeworth Box is that we use two different two-dimensional axes to assign four
values to each point — thus graphing a four dimensional space in 2 dimensions.
Exercise 16B.3
What are the reservation utilities for me and my wife in our example (given the
utility functions specified above)?
Answer: The utility functions are u 1 (x1 , x2 ) = x13/4 x21/4 and u 2 (x1 , x2 ) = x11/4 x23/4 ,
and our endowments are (e 11 , e 21 ) = (3, 6) and (e 12 , e 22 ) = (10, 4). Thus,
U 1 = 33/4 61/4 ≈ 3.5676 and U 2 = 101/4 43/4 = 5.0297. ( 16B.3)
Exercise 16B.4
For the example of me and my wife, write the set of mutually beneficial alloca-
tions in the form of equation (16.6). Can you see that the lens-shaped area identi-
fied in the Edgeworth Box in Graph 16.2 is equivalent to this set?
Answer: The set would be
¡ ¢3/4 ¡ 1 ¢1/4
MB = {(x11 , x21 , x12 , x22 ) ∈ F A | x11 x2 ≥ 3.5676 and
¡ 2 ¢1/4 ¡ 2 ¢3/4 ( 16B.4.i)
x1 x2 ≥ 5.0297},
where
F A = (x11 , x21 , x12 , x22 ) ∈ R4+ | x11 + x12 = 13 and x21 + x22 = 10 .
© ª
( 16B.4.ii)
Exercise 16B.5
Can you see that no allocation in the set PE of an Edgeworth Box could have
indifference curves pass through it in a way that creates a lens-shaped area between
them?
Answer: Such a lens shape between indifference curves would imply a region
that is feasible and contains bundles that lie above both indifference curve — i.e.
there exist ways of making both people better off. If an allocation is such that we
can make both people better off, it cannot lie in the PE set.
Exercise 16B.6
Verify that the contract curve we derived goes from one corner of the Edgeworth
Box to the other.
Answer: Substituting x11 = 0 into the equation for the contract curve, we get
1
x2 = 0 — i.e. the contract curve passes through the lower left hand corner of the
Edgeworth Box at (0,0). Substituting x11 = 13 into the equation for the contract
curve, we get x21 = 10 — i.e. the contract curve passes through (13,10), the upper
left corner of the Edgeworth Box. The rest of the function is pictured in Exercise
Graph 16B.6.
10
2 4 6 8 10 12
Exercise Graph 16B.6 : Contract Curve
Exercise 16B.7
A different way to find the contract curve would be to maximize my utility sub-
ject to the constraint that my wife’s utility is held constant at utility level u ∗ and
that her consumption bundle is whatever is left over after I have been given my con-
sumption bundle. Put mathematically, this problem can be written as max x1 ,x2 x13/4 x21/4
s.t. u ∗ = (13 − x1 )1/4 (10 − x2 )3/4 (where we drop the superscripts given that all vari-
ables refer to my consumption.) Demonstrate that this leads to the same solution
as that derived in equation (16.9).
Answer: Setting up the Lagrange function for this problem, we get
L = x13/4 x21/4 + λ u ∗ − (13 − x1 )1/4 (10 − x2 )3/4 .

¡ ¢
( 16B.7.i)
Taking first order conditions, we get
∂L 3x 1/4 (10 − x2 )3/4 ∂L x 3/4 3(13 − x1 )1/4

= 21/4 +λ 3/4
= 0 and = 13/4 +λ = 0. ( 16B.7.ii)
∂x1 4x1 4(13 − x1 ) ∂x1 4x2 4(10 − x2 )1/4
Collecting the λ terms on one side and dividing the two first order conditions
by one another, we then get
3x2 (10 − x2 )
= ( 16B.7.iii)
x1 3(13 − x1 )
which is identical to the condition in the text from which we derived the con-
tract curve
10x1
x2 = . ( 16B.7.iv)
(117 − 8x1 )
Exercise 16B.8
Verify that these are the correct demands for this problem.
Answer: Consider the utility function u(x1 , x2 ) = x1α x2(1−α) for an individual with
endowments e 1 and e 2 . At prices p 1 and p 2 , the value of the endowment for this
individual is p 1 e 1 + p 2 e 2 , which allows us to write the utility maximization problem
as
max x1α x2(1−α) subject to p 1 x1 + p 2 x2 = p 1 e 1 + p 2 e 2 . ( 16B.8.i)
x 1 ,x 2
We can then write the Lagrange function as
L = x1α x2(1−α) + λ p 1 e 1 + p 2 e 2 − p 1 x1 − p 2 x2 .
¡ ¢
( 16B.8.ii)
The first two first order conditions are then
∂L ∂L
= αx1(α−1) x2(1−α) − λp 1 = 0 and = (1 − α)x1α x2−α − λp 2 = 0, ( 16B.8.iii)
∂x1 ∂x2
which solve to give us
(1 − α)p 1 x1
x2 = . ( 16B.8.iv)
αp 2
Substituting this into the budget constraint and solving for x1 , we then get
α(p 1 e 1 + p 2 e 2 )
x1 = , ( 16B.8.v)
p1
and substituting this back into ( 16B.8.iv), we get
(1 − α)(p 1 e 1 + p 2 e 2 )
x2 = . ( 16B.8.vi)
p2
For me, α = 3/4 and (e 1 , e 2 ) = (3, 6) whereas for my wife α = 1/4 and (e 1 , e 2 ) =
(10, 4). Substituting these into equations ( 16B.8.v) and ( 16B.8.vi), we can verify
that these become equivalent to the equations in the text.
Exercise 16B.9
Write down the equilibrium condition in the x2 market (from the second equa-
tion in expression (16.14)) using the appropriate expressions from (16.15) and solve
for the equilibrium price ratio. You should get the same answer.
Answer: Equilibrium in the x2 market is reached when
(3p 1 + 6p 2 ) 3(10p 1 + 4p 2 )
+ = 6 + 4. ( 16B.9)
4p 2 4p 2
Solving this for p 2 , we again get p 2 = 3p 1 /2.
Exercise 16B.10
Demonstrate that the same equilibrium allocation of goods will arise if p 1 = 2

and p 2 is one and a half times p 1 — i.e. p 2 = 3.
Answer: Plugging the prices into the demand equations for the two individuals,
we get
3(3(2) + 6(3)) (3(2) + 6(3))

x11 (2, 3)) = = 9 and x21 (2, 3) = =2
4(2) 4(3)
(10(2) + 4(3)) 3(10(2) + 4(3))
x12 (2, 3) = = 4 and x22 (2, 3) = = 8. ( 16B.10)
4(2) 4(3)
Exercise 16B.11
Can you demonstrate that the equilibrium allocation we derived for me and my
wife lies in the core that we defined in equation (16.12)?
Answer: To do this, we have to demonstrate that the allocation (x11 , x21 , x12 , x22 ) =
(9, 2, 4, 8) lies in the PE set as well as the MB set. To demonstrate that it lies in the
PE set, we have to show that
10x11
x21 = , x12 = 13 − x11 and x22 = 10 − x21 . ( 16B.11.i)
(117 − 8x11 )
Plugging x11 = 9 into the first two of these equations, we get
10(9)
x21 = = 2 and x12 = 13 − 9 = 4, ( 16B.11.ii)
(117 − 8(9))
and plugging x21 = 2 into the last of the three equations, we get x22 = 10 − 2 = 8.
Thus, the allocation (x11 , x21 , x12 , x22 ) = (9, 2, 4, 8) indeed lies in the PE set. To verify
that it also lies in the MB set, we just need to verify that the utility for each person
lies above their reservation utilities of U 1 = 3.57 and U 2 = 5.03. Evaluating utility
for each of the individuals at the allocation (x11 , x21 , x12 , x22 ) = (9, 2, 4, 8), we get
u 1 (9, 2) = 93/4 21/4 ≈ 6.18 and u 2 (4, 8) = 41/4 83/4 ≈ 6.73. ( 16B.11.iii)
Exercise 16B.12
In our proof, we began by assuming that there exists a (y 11 , y 21 , y 12 , y 22 ) that is

strictly preferred by everyone to (x11 , x21 , x12 , x22 ) and showed that there cannot be
such an allocation within this economy. Can you see how the same logic also goes
through if we assume that there exists an allocation (z11 , z21 , z12 , z22 ) that is strictly pre-
ferred by one of the individuals while leaving the other indifferent to (x11 , x21 , x12 , x22 )?
Answer: Suppose the allocation (z11 , z21 , z12 , z22 ) is strictly preferred by individual
1 to the allocation (x11 , x21 , x12 , x22 ) and that individual 2 is indifferent between the two
allocations. Then (z11 , z21 ) could not have been affordable for individual 1 under the
equilibrium prices (p 1 , p 2 ) (or else he would have chosen it) — i.e.
p 1 z11 + p 2 z21 > p 1 x11 + p 2 x21 . ( 16B.12.i)
If individual 2 is indifferent between (z11 , z21 , z12 , z22 ) and (x11 , x21 , x12 , x22 ), then we
can conclude that the bundle (z12 , z22 ) could not have been cheaper at the equilib-
rium prices than (x12 , x22 ). This is because if (z12 , z22 ) actually were to lie strictly inside
the budget set (and not on the budget line), the indifference curve that contains
(z12 , z22 ) would cut the budget line — leaving some bundles on higher indifference
curves that still lie on the budget constraint — which in turn implies that the opti-
mal choice (x12 , x22 ) could not have been one that yields the same utility as (z12 , z22 ).
Concluding that (z12 , z22 ) could not have been cheaper at the equilibrium prices than
(x12 , x22 ) is then the same as concluding that
p 1 z12 + p 2 z22 ≥ p 1 x12 + p 2 x22 ( 16B.12.ii)
Adding equations ( 16B.12.i) and ( 16B.12.ii), we get
p 1 (z11 + z12 ) + p 2 (z21 + z22 ) > p 1 (x11 + x12 ) + p 2 (x21 + x22 ) ( 16B.12.iii)
just as in the proof in the text (where we assumed that there exists a bundle
which is strictly preferred by both individuals). The rest of the proof is then iden-
tical to the one in the text — with the ultimate contradiction that the allocation
(z11 , z21 , z12 , z22 ) is not feasible.
Exercise 16B.13
Can you demonstrate that this is in fact the case for an N -person, M-good econ-
omy?
Answer: The proof is virtually identical to what we have done for the 2-person
case. We begin with the equilibrium defined by prices (p 1 , p 2 , ..., p M ) and the allo-
cation
1
(x11 , x21 , ..., x M 2
, x12 , x22 , ..., x M N
, ..., x1N , x2N , ...x M NM
) ∈ R+ . ( 16B.13.i)
To say that this allocation is an equilibrium allocation under prices (p 1 , p 2 , ..., p M )

is the same as saying that each individual n, taken his/her endowment and prices
as given, in fact maximizes utility by choosing his/her portion of the allocation —
i.e. (x1n , x2n , ..., x M
n
). We are trying to prove that this equilibrium allocation must be
efficient — i.e. there does not exist an alternative allocation
(y 11 , y 21 , ..., y M
1
, y 12 , y 22 , ..., y M
2
, ..., y 1N , y 2N , ...y M
N NM
) ∈ R+ ( 16B.13.ii)
that is strictly preferred by at least one person and viewed as at least as good
by all others. So suppose that such an allocation did exist, and suppose we identify
the person for whom this allocation is strictly better as person 1. We then know that
(y 11 , y 21 , ..., y M
1
) could not have been affordable for person 1 at prices (p 1 , p 2 ) (or else
it would have been chosen) — i.e. we know
1
p 1 y 11 + p 2 y 21 + ... + p M y M 1
> p 1 x11 + p 2 x21 + ... + p M x M . ( 16B.13.iii)
Similarly — and for reasons analogous to those explained in the previous exer-
cise — it must be the case that
p 1 y 1n + p 2 y 2n + ... + p M y M
n
≥ p 1 x1n + p 2 x2n + ... + p M x M
n
for all n = 2, 3, ..., N .
( 16B.13.iv)
Adding equation ( 16B.13.iii) to the (N − 1) equations in expression ( 16B.13.iv),
we get
N N N N N N
y 1n + p 2 y 2n + ... + p M n
x1n + p 2 x2n + ... + p M n
X X X X X X
p1 yM > p1 xM .
n=1 n=1 n=1 n=1 n=1 n=1
( 16B.13.v)
Walras’ Law tells us that the right hand side is equal to the value of the endow-
ments — which implies we can re-write this as
N N N N N N
y 1n + p 2 y 2n + ... + p M n
e 1n + p 2 e 2n + ... + p M e nM
X X X X X X
p1 yM > p1
n=1 n=1 n=1 n=1 n=1 n=1
( 16B.13.vi)
or as
Ã ! Ã ! Ã !
N N N N N N
y 1n − e 1n y 2n e 2n n
e nM
X X X X X X
p1 + p2 − + ... + p M yM − > 0.
n=1 n=1 n=1 n=1 n=1 n=1
( 16B.13.vii)
But, since all prices are positive, this implies that at least one of the bracketed
terms in equation ( 16B.13.vii) must be greater than zero — i.e.
N N
n n
X X
ym − em > 0 for some good m. ( 16B.13.viii)
n=1 n=1
But that implies that the y allocation is not feasible — i.e. the assumption of in-
efficiency of the equilibrium allocation has led to a contradiction. Thus, the equi-
librium allocation is efficient.
Exercise 16B.14
Verify the results in equation (16.27).

Answer: We can solve this either by setting up the Lagrange function or sim-
ply by substituting the constraint into the objective. Doing the former, we get the
Lagrange function ³ ´
L = x α (L − ℓ)(1−α) + λ x − Aℓβ . ( 16B.14.i)
The first two first order conditions are then
∂L
= αx (α−1) (L − ℓ)(1−α) + λ = 0 and
∂x
( 16B.14.ii)
∂L
= −(1 − α)x α (L − ℓ)−α − λβAℓ(β−1) = 0.
∂ℓ
which can also be written as
αx (α−1) (L − ℓ)(1−α) = −λ and (1 − α)x α (L − ℓ)−α = −λβAℓ(β−1) . ( 16B.14.iii)
Dividing the two equations in expression ( 16B.14.iii) by one another and solv-
ing for x, we get
αβA(L − ℓ)ℓ(β−1)
x= , ( 16B.14.iv)
(1 − α)
and substituting this into the constraint from the optimization problem, we get
αβA(L − ℓ)ℓ(β−1)
Aℓβ = ( 16B.14.v)
(1 − α)
which we can in turn solve for ℓ to verify
αβL
ℓ= . ( 16B.14.vi)
1 − α(1 − β)
Plugging this into ( 16B.14.iv) and solving for x, we furthermore verify the text’s
expression for x. This involves a bit of algebra — the quickest way to get there is as
follows: Begin by substituting ( 16B.14.vi) into (L − ℓ) to get
αβL [1 − α(1 − β)]L − αβL (1 − α)L

(L − ℓ) = L − = = . ( 16B.14.vii)
1 − α(1 − β) 1 − α(1 − β) 1 − α(1 − β)
Substituting this for (L − ℓ) into equation ( 16B.14.iv), we get
αβAℓ(β−1) αβALℓ(β−1)
µ ¶
(1 − α)L
x= = . ( 16B.14.viii)
(1 − α) 1 − α(1 − β) 1 − α(1 − β)
Substituting ( 16B.14.vi) into this for ℓ, we then get
µ ¶(β−1) µ ¶β
αβAL αβL αβL
x= =A . ( 16B.14.ix)
1 − α(1 − β) 1 − α(1 − β) 1 − α(1 − β)
Exercise 16B.15
Verify that this is the correct solution.

Answer: Taking the first derivative of (p Aℓβ −wℓ) (with respect to ℓ) and setting
it to zero, we get
βp Aℓ(β−1) − w = 0, ( 16B.15.i)
and solving for ℓ, we get

µ ¶1/(β−1) µ ¶1/(1−β)
w βp A
ℓ= = . ( 16B.15.ii)
βp A w
Substituting this back into the production function, we get
Ãµ ¶1/(1−β) !β µ ¶β/(1−β)
βp A βp A
x=A =A ( 16B.15.iii)
w w
Exercise 16B.16
Verify that these solutions for labor supply and banana demand are correct.
Answer: The Lagrange function for the maximization problem is
L = x α (L − ℓ)(1−α) + λ(wℓ + π(w, p) − p x) ( 16B.16.i)
giving rise to first order conditions

∂L
= αx (α−1) (L − ℓ)(1−α) − λp = 0 and
∂x
( 16B.16.ii)
∂L
= −(1 − α)x α (L − ℓ)−α + λw = 0.
∂ℓ
Rearranging these with the λ terms on one side, dividing the equations by each
other and solving for x, we get
α(L − ℓ)w
x= . ( 16B.16.iii)
(1 − α)p
Substituting this into the constraint wℓ+π(w, p) = p x and solving for ℓ, we then
get
(1 − α)π(w, p)
ℓ = αL − , ( 16B.16.iv)
w
and substituting this back into equation ( 16B.16.iii), we get
α¡ ¢
x= wL + π(w, p) . ( 16B.16.v)
p
These labor supply and banana demand equations can then be expanded by
simply replacing π(w, p) with the profit function
µ ¶β/(1−β)
β
π(w, p) = (1 − β)(Ap)1/(1−β) ( 16B.16.vi)
w
to give us
(1 − α)(1 − β) βp A 1/(1−β)
µ ¶
ℓ = αL −
β w
Ã ! ( 16B.16.vii)
(1 − β) βp A 1/(1−β)
µ ¶
αw
x= L+ .
p β w
Exercise 16B.17
Verify that the same equilibrium relationship between prices and wages arises
by solving x S = x D .
Answer: We need to solve
¶β/(1−β) Ã !
(1 − β) βp A 1/(1−β)
µ µ ¶
S βp A αw
x (w, p) = A = L+ = x D (w, p).
w p β w
( 16B.17.i)
The problem is simplified if we re-write the right hand side as
Ã !
(1 − β) βp A 1/(1−β)
µ ¶
D αw
x (w, p) = L+ =
p β w
( 16B.17.ii)
βp A β/(1−β)
µ ¶
αw
= L + α(1 − β)A .
p w
Equation ( 16B.17.i) then becomes

¶β/(1−β)
βp A β/(1−β)
µ µ ¶
βp A αw
A = L + α(1 − β)A ( 16B.17.iii)
w p w
which can be re-written as
A(βp A)β/(1−β) (1 − α(1 − β))p

w 1/(1−β) = =
αL
( 16B.17.iv)
(Aβ)1/(1−β) (1 − α(1 − β))p 1/(1−β)
= .
αβL
Taking both the left and right hand sides to the power (1 − β) and re-arranging
terms slightly, we then arrive at
µ ¶(1−β)
1 − α(1 − β)
w = Aβ p. ( 16B.17.v)
αβL
Exercise 16B.18
Can you tell from the graph of an equilibrium in Graph 16.10 that any combina-
tion of w and p that satisfies a particular ratio will generate the same equilibrium
in the labor and banana markets?
Answer: In panel (b) of the graph we see that the equilibrium point C occurs at
the tangency of the indifference curve u ∗ and the production frontier. The input
and output prices that support this as an equilibrium are those that result in the
isoprofit/budget line with slope w ∗ /p ∗ such that the line is tangent to both the
indifference curve and the production frontier at C — and what matters for this is
the ratio of the prices. (Economic theorists will often call the grey line in the graph
a “separating hyperplane” — i.e. the “hyperplane” that separates the production
frontier and indifference curve at precisely one point.)
Exercise 16B.19
Can you tell from the graph of an equilibrium in Graph 16.10 whether profit will
be affected by different choices of w and p that satisfy the equilibrium ratio? Verify
whether your intuition holds mathematically.
Answer: Suppose w ∗ and p ∗ satisfy the equilibrium ratio — resulting in profit
π . The intercept of the grey isoprofit/budget line in panel (b) of the graph is π∗ /p ∗ .
∗
Now suppose that we increase both the wage and the output price by a factor t ,
resulting in profit π′ . The intercept of the grey line that separates the indiffer-
ence curve and production frontier in the graph does not change, which implies
π′ /(t p ∗ ) = π∗ /p ∗ — which can only hold if π′ = t π∗ . Thus, as the input and out-
put prices are scaled up and down, profit is scaled with it at the same rate. We can
see this from the profit function by illustrating that π(t w, t p) = t π(w, p) — i.e. by
showing that the profit function is homogeneous of degree 1:
β β/(1−β)
¶ µ
π(t w, t p) = (1 − β)(At p)1/(1−β) =
tw
µ ¶β/(1−β) ( 16B.19)
β
= t (1 − β)(Ap)1/(1−β) = t π(w, p).
w
Exercise 16B.20
Demonstrate that the equilibrium banana consumption (and production) is

equal to the optimal level of banana consumption.
Answer: When first considering the optimal decision for Robinson Crusoe in
the text, we derived the pareto optimal consumption level for bananas as
µ ¶β
αβL
x=A . ( 16B.20.i)
1 − α(1 − β)
The equilibrium level can be determined by plugging our expression for the
equilibrium wage w ∗ (in terms of p) into either the output supply or demand equa-
tion in the text. Using the output supply equation, we get
µ ¶β/(1−β) µ ¶β/(1−β) µ ¶β
βp A ¢β/(1−β) 1 αβL
xS = A
¡
= A βp A =
w∗ βAp 1 − α(1 − β)
µ ¶β ( 16B.20.ii)
αβL
=A .
1 − α(1 − β)
Substituting our expression for w ∗ into x D gives the same answer.
Exercise 16B.21
Why is there no coalition to block A in the 2-person version of this economy?

Answer: In the 2-person economy, there are only three coalitions: (1) a coalition
of both (and thus all) the individuals in the economy; (2) a coalition composed of
a single individual of type 1; and (3) a coalition of a single individual of type 2.
At A, “coalition” (2) is just as well off as at E while “coalition” (3) is strictly better
off. Thus, neither “coalition” (2) nor “coalition” (3) would block the allocation A
because neither can do better by splitting away from the economy. That leaves only
coalition (1) — composed of both type 1 and type 2. At A, their indifference curves
are tangent to one another — which implies there is no way we can make either
of them better off without making the other worse off. Thus, coalition (1) has no
reason to block A. As a result, none of the coalitions in the economy can do better
by splitting off than the individuals in the coalitions do at A. The difference in the
4-person economy is that there are now more coalitions to consider — including
3-person coalitions composed of two of one type and one of the other.
Exercise 16B.22
Can you demonstrate that a coalition of two of the “type 2” individuals with on
“type 1” individual will block the allocation B?
Answer: Such a coalition will block B for exactly the same reasons that the coali-
tion of two type 1 individuals with 1 type 2 individual blocks A. Suppose B is pro-
posed as the allocation for the economy — with each type getting what the axes for
that type indicate at B. Now suppose the coalition of two type 2 and one type 1 indi-
vidual splits off with their endowments. In the Edgworth Box, draw a line that goes
through E and B. With both types starting at E , a trade in which the two type 2 con-
sumers give up 1 unit of x2 results in the one type 1 individual getting two units of
x2 . Thus, if the individuals in the coalition trade along the line that passes through
E and B, the one type 1 individual moves down the line twice as quickly as the two
type 2 individuals do. Thus, the coalition can trade among itself and end up with
an allocation for the type 1 consumer that lies to the southeast of B and an alloca-
tion for the type 2 consumers that lies to the northwest of B — with both consumer
types ending up on an indifference curve that is higher than the one that passes
through B. Thus, the coalition can do better by splitting off the economy with its
endowment than their members do under the allocation B — which implies that
the coalition will “block” B from being implemented.
Exercise 16B.23
Why must the distance between E and D ′ be twice the distance from E to C ′ ?
Answer: This is because there are two type 1 individuals and one type 2 individ-
ual — which means that everything a type 1 individual gives up is received twice by
the single type 2 — and anything that the type 2 individual gives up in exchange is
split between the two type 1 people.
Exercise 16B.24
Demonstrate that the competitive equilibrium allocation must lie in the core of
the replicated exchange economy.
Answer: An equilibrium allocation X has the feature that the indifference curves
for type 1 and type 2 consumers are tangent at X in the Edgeworth Box — with the
line that passes through E and X “separating” the two indifference curves at X . If
any coalition splits off, it will trade along a line that passes through E — and this
line is either shallower or steeper than the one that passes through E and X . Sup-
pose it is shallower. In that case all points on that line lie “below” the equilibrium
indifference curve for type 2 consumers — and thus there is no coalition that in-
cludes a type 2 individual which could in fact do better for type 2. Suppose instead
that the line is steeper. Then it lies entirely “below” the equilibrium indifference
curve for type 1 consumers — implying that no type 1 consumer could be part of a
blocking coalition. Thus, neither consumer could be part of a blocking coalition —
which implies that the equilibrium allocation X lies in the core.
16C Solutions to Odd Numbered

End-of-Chapter Exercises
Exercise 16.1
Consider a 2-person/2-good exchange economy in which person 1 is endowed

with (e 11 , e 21 ) and person 2 is endowed with (e 12 , e 22 ) of the goods x1 and x2 .
A: Suppose that tastes are homothetic for both individuals.
(a) Draw the Edgeworth Box for this economy, indicating on each axis the di-
mensions of the box.
Answer: This is illustrated in panel (a) of Exercise Graph 16.1(1) where
the width of the box is (e 11 + e 12 ) and the height is (e 21 + e 22 ).
Exercise Graph 16.1(1) : Contract Curve with Homothetic Tastes
(b) Suppose that the two individuals have identical tastes. Illustrate the con-
tract curve — i.e. the set of all efficient allocations of the two goods.
Answer: Homothetic tastes have the characteristic that the MRS is the
same along any ray from the origin. Consider the ray that passes from
the lower left to the upper right corners of the box — i.e. from the origin
for person 1 to the origin for person 2. If tastes are homothetic for each
of the two individuals, this means that, for each individual, it is the case
that the MRS is constant along this ray. And if their tastes are identical,
then their MRS’s are the same along that ray — i.e. on each point of the
25 16C. Solutions to Odd Numbered End-of-Chapter Exercises
ray, the indifference curves that pass through that point are tangent to
one another. Since the contract curve is the set of allocations where the
indifference curves are tangent, this ray is then the contract curve. It is
depicted in panel (b) of Exercise Graph 16.1(1).
(c) True or False: Identical tastes in the Edgeworth Box imply that there are no
mutually beneficial trades.
Answer: This is false. In panel (b) of Exercise Graph 16.1(1), for instance,
the indifference curves u 1 and u 2 contain the endowment bundle E —
with allocations inside the lens created by these indifference curves rep-
resenting mutually beneficial trades. The only way that there are no mu-
tually beneficial trades when both individuals have identical homothetic
tastes is if the endowment bundle falls on the contract curve — i.e. on the
line connecting the origins for the two individuals.
(d) Now suppose that the two individuals have different (but still homothetic)
tastes. True or False: The contract curve will lie to one side of the line that
connects the lower left and upper right corners of the Edgeworth Box — i.e.
it will never cross this line inside the Edgeworth Box.
Answer: This is true (almost). If the two individuals’ tastes are not identi-
cal, then their indifference curves are not likely to be tangent on the line
connecting the lower left and upper right corners of the box. Take one
point on that line — it is likely the case that the indifference curve for
person 1 is steeper or shallower than that for individual 2 at this point.
Suppose first that individual 1’s indifference curve is shallower. Then the
two indifference curves form a lens shape — with the entire area of the
lens lying above the line connecting the corners of the box. Since the
slopes of the indifference curves are constant along this line, this same
lens shape will appear above the line for any allocation on the line. This
implies that the tangencies of indifference curves (which form the con-
tract curve) must also lie above the line (because these tangencies will be
found within the lens shapes formed from indifference curves that cross
on the line.) The reverse will be true if individual 1’s indifference curve
is steeper along the line than indifference curves for individual 2 — with
the entire contract curve now lying below the line. The reason the answer
is true (almost) is that it is still possible that the marginal rates of substi-
tution for the two individuals are in fact equal along the line connecting
the lower left and upper right corners of the box. For instance, it might be
that the tastes have different degrees of substitutability (and are therefore
different) but still have the same marginal rates of substitution on that
line. In that case, the contract curve lies on the line connecting the lower
left and upper right corners. Thus, homothetic tastes imply that the con-
tract curve lies either on the line connecting the corners or all to one side
of that line.
B: Suppose that the tastes for individuals 1 and 2 can be described by the utility
β (1−β)
functions u 1 = x1α x2(1−α) and u 2 = x1 x2 (where α and β both lie between
0 and 1). Some of the questions below are notationally a little easier to keep
track off if you also denote E 1 = e 11 + e 12 as the economy’s endowment of x1 and
E 2 = e 21 + e 22 as the economy’s endowment of x2 .
(a) Let x 1 denote the allocation of x1 to individual 1, and let x 2 denote the
allocation of x2 to individual 1. Then use the fact that the remainder of the
economy’s endowment is allocated to individual 2 to denote individual 2’s
allocation as (E 1 − x 1 ) and (E 2 − x 2 ) for x1 and x2 respectively. Derive the
contract curve in the form x 2 = x2 (x 1 ) — i.e. with the allocation of x2 to
person 1 as a function of the allocation of x1 to person 1.
Answer: You can derive this either by setting the MRS for individual 1
equal to the MRS for individual 2 — or you can solve the problem
(1−α)
max x α
1 x2 subject to u 2 = (E 1 − x 1 )β (E 2 − x 2 )(1−β) (16.1.i)
x 1 ,x 2
where person 1’s utility is maximized subject to getting person 2 to a par-

ticular indifference curve u 2 . Either way, you will get to the point where
you have an expression that sets the marginal rates of substitution equal
to one another — i.e.
∂u 1 (x 1 , x 2 )/∂x1 αx 2 β(E 2 − x 2 )
1
= = =
∂u (x 1 , x 2 )/∂x2 (1 − α)x 1 (1 − β)(E 1 − x 1 )
(16.1.ii)
∂u 2 ((E 1 − x 1 ), (E 2 − x 2 ))/∂x1
= .
∂u 2 ((E 1 − x 1 ), (E 2 − x 2 ))/∂x2
Solving the middle of this expression for x 2 , we then get the contract
curve
(1 − α)βE 2 x 1
x2 (x 1 ) = . (16.1.iii)
α(1 − β)E 1 + (β − α)x 1
(b) Simplify your expression under the assumption that tastes are identical —
i.e. α = β. What shape and location of the contract curve in the Edgeworth
Box does this imply?
Answer: Replacing β with α, the expression then simplifies to
(1 − α)αE 2 x 1 E2
x2 (x 1 ) = = x 1. (16.1.iv)
α(1 − α)E 1 + (α − α)x 1 E 1
This is simply the equation of a line with zero vertical intercept and slope
E 2 /E 1 — which is the slope of the ray that passes from the lower left to the
upper right corner of the Edgeworth Box. Thus, when tastes are identical,
we get that the contract curve is the line that connects the origins for the
two individuals in the Edgeworth Box — exactly as we did for homothetic
tastes in part A of the question (and as depicted in panel (b) of Exercise
Graph 16.1(1).)
(c) Next, suppose that α 6= β. Verify that the contract curve extends from the
lower left to the upper right corner of the Edgeworth Box.
Answer: Evaluating the contract curve equation (16.1.iii) at x 1 = 0, we get
x2 (0) = 0 — i.e. the contract curve passes through the lower left hand
corner of the Edgeworth Box. Evaluating the contract curve at x 1 = E 1 , we
get
(1 − α)βE 2 E 1 (1 − α)βE 2 E 1
x2 (E 1 ) = = = E 2; (16.1.v)
α(1 − β)E 1 + (β − α)E 1 (1 − α)βE 1
i.e. the contract curve passes through the upper right corner of the box
where individual 1 gets all the goods in the economy.
(d) Consider the slopes of the contract curve when x 1 = 0 and when x 1 = E 1 .
How do they compare to the slope of the line connecting the lower left and
upper right corners of the Edgeworth Box if α > β? What if α < β?
Answer: The slope of the contract curve is the derivative of equation (16.1.iii)
with respect to x1 —
∂x2 (x 1 ) (1 − α)βE 2 (β − α)(1 − α)βE 2 x 1

= −£ ¤2 =
∂x1 α(1 − β)E 1 + (β − α)x 1 α(1 − β)E 1 + (β − α)x 1
(16.1.vi)
αβ(1 − α)(1 − β)E 1 E 2
=£ ¤2 .
α(1 − β)E 1 + (β − α)x 1
Evaluated at x 1 = 0 and at x 1 = E 1 , we get
∂x2 (0) β(1 − α)E 2 ∂x2 (E 1 ) α(1 − β)E 2

= and = . (16.1.vii)
∂x1 α(1 − β)E 1 ∂x1 β(1 − α)E 1
The slope of the line connecting the two corners of the Edgeworth box
is E 2 /E 1 . If α = β, both derivatives in expression (16.1.vii) are equal to
E 2 /E 1 — i.e. the slope of the contract curve is exactly the slope of the line
connecting the corners (as we already concluded previously). If α > β,
then
β(1 − α) α(1 − β)
< 1 and >1 (16.1.viii)
α(1 − β) β(1 − α)
implying that
∂x2 (0) E 2 ∂x2 (E 1 ) E 2

< and > . (16.1.ix)
∂x1 E1 ∂x1 E1
The reverse relationship holds when α < β.
(e) Using what you have concluded, graph the shape of the contract curve for
the case α > β and for the case when α < β?
The contract curves consistent with these relationships are graphed in
Exercise Graph 16.1(2).
Exercise Graph 16.1(2) : Contract Curves when (a) α > β and when (b) α < β
(f) Suppose that the utility function for the two individuals instead took the
more general constant elasticity of substitution form u = (αx1 + (1 − α)x2 )−1/ρ .
If the tastes for the two individuals are identical, does your answer to part
(b) change?
Answer: No, the answer does not change. The MRS for this utility func-
tion (derived in Chapter 5) is
µ ¶µ ¶ρ+1
α x2
MRS = − . (16.1.x)
(1 − α) x1
Using our notation and setting the MRS’s equal to each other for the two
individuals, we then get
µ ¶µ ¶ρ+1 µ ¶µ ¶ρ+1
α x2 α (E 2 − x 2 )
= (16.1.xi)
(1 − α) x1 (1 − α) (E 1 − x 1 )
which we can solve for x 2 to get the contract curve

µ ¶
E2
x2 (x 1 ) = x1; (16.1.xii)
E1
i.e. the contract curve again has zero vertical intercept and slope E 2 /E 1 ,
the slope of the ray that connects the two corners of the Edgeworth Box.
Exercise 16.3
Suppose you and I have the same homothetic tastes over x1 and x2 , and our en-
dowments of the two goods are E M = (e 1M , e 2M ) for me and E Y = (e 1Y , e 2Y ) for you.
A: Suppose throughout that, when x1 = x2 , our MRS is equal to −1.
(a) Assume that e 1M = e 2M = e 1Y = e 2Y . Draw the Edgeworth box for this case and
indicate where the endowment point E = (E M , E Y ) lies.
Answer: This is done in panel (a) of Exercise Graph 16.3 where the Edge-
worth Box is drawn as a square (because the overall endowments of x1 are
equal to those of x2 ), with the endowment bundle E located in the center
(since we are endowed with equal amounts of everything.)
Exercise Graph 16.3 : Equal Endowments and Same Tastes
(b) Draw the indifference curves for both of us through E . Is the endowment
allocation efficient?
Answer: This is also done in panel (a). Since our endowment lies on the
45-degree line and our MRS along the 45-degree line is always −1, the in-
difference curves through E are tangent to one another. This implies that
the endowment allocation is efficient — because there is no lens shape
between our indifference curves that would give us room to make both
of us better off (or at least one better off without making the other worse
off).
(c) Normalize the price of x2 to 1 and let p be the price of x1 . What is the
equilibrium price p ∗ ?
Answer: The equilibrium price must pass through E and induce budget
constraints for me and you such that both of us optimize at the same
point within the Edgeworth Box. In this case, the only way to do this is
to let p ∗ = 1 — resulting in a budget with slope −1 through E . Since the
MRS at E is −1 for both of us, we will both choose to remain at our endow-
ment bundle at this price. This is also illustrated in panel (a) of Exercise
Graph 16.3.
(d) Where in the Edgeworth Box is the set of all efficient allocations?
Answer: The set of all efficient allocations lies on the 45-degree line —
because along the 45 degree line, our MRS’s are equal to 1 and thus equal
to one another, implying indifference curves that are tangent to one an-
other. This is the contract curve.
(e) Pick another efficient allocation and demonstrate a possible way to re-
allocate the endowment among us such that the new efficient allocation
becomes an equilibrium allocation supported by an equilibrium price. Is
this equilibrium price the same as p ∗ calculated in (c)?
Answer: This is illustrated in panel (b) of Exercise Graph 16.3 where we
consider the equilibrium if the endowment is redistributed so that it moves
from E to A. The only place where the indifference curves are tangent to
one another is on the 45-degree line where their slope is −1. Thus, the
new equilibrium price must again be 1 — and the new budget must pass
through the new endowment A as drawn. This will cause us to trade from
A to B along the budget line with slope p = 1 — with me giving up x1 to
get more x2 and you giving up an equal amount of x2 to get more x1 (as
indicated by the arrows on the axes.)
B: Suppose our tastes can be represented by the CES utility function u(x1 , x2 ) =
¡ −ρ −ρ ¢−1/ρ
0.5x1 + 0.5x2 .
(a) Let p be defined as in A(c). Write down my and your budget constraint
(assuming again endowments E M = (e 1M , e 2M ) for me and E Y = (e 1Y , e 2Y ).)
Answer: The value of or endowments has to be equal to the value of what
we consume. For me, this implies
pe 1M + e 2M = p x1M + x2M , (16.3.i)
and for you it means
pe 1Y + e 2Y = p x1Y + x2Y . (16.3.ii)

(b) Write down my optimization problem and derive my demand for x1 and
x2 .
Answer: My optimization problem is then
−ρ −ρ ¢−1/ρ
subject to pe 1M + e 2M = p x1 + x2
¡
max 0.5x1 + 0.5x2 (16.3.iii)
x 1 ,x 2
where, for now, we suppress the M superscripts on the x variables. Setting

up the Lagrangian and solving in the usual way, we get
pe 1M + e 2M p 1/(ρ+1) e 1M + e 2M
x1M = and x2M = . (16.3.iv)
p + p 1/(ρ+1) p + p 1/(ρ+1)
(c) Similarly, derive your demand for x1 and x2 .

Answer: Repeating the steps in the previous part for you, we get
pe 1Y + e 2Y p 1/(ρ+1) e 1Y + e 2Y
x1Y = and x2Y = . (16.3.v)
p + p 1/(ρ+1) p + p 1/(ρ+1)
(d) Derive the equilibrium price. What is that price if, as in part A, e 1M = e 2M =
e 1Y = e 2Y ?
Answer: In equilibrium, the price has to be such that demand is equal to
supply in both markets. Because of Walras’ Law, we only have to solve for
p in one of the markets though — and either one will work. Choosing the
market for x1 , it must therefore be the case that x1M + x1Y = e 1M + e 1Y or,
plugging in our demands from the previous parts,
pe 1M + e 2M pe 1Y + e 2Y
+ = e 1M + e 1Y . (16.3.vi)
p + p 1/(ρ+1) p + p 1/(ρ+1)
Multiplying both sides by the denominators on the left hand side, we get
pe 1M + e 2M + pe 1Y + e 2Y = e 1M + e 1Y p + p 1/(ρ+1)
¡ ¢¡ ¢
(16.3.vii)
and, rearranging terms,
p e 1M + e 1Y + e 2M + e 2Y = p e 1M + e 1Y + p 1/(ρ+1) e 1M + e 1Y .
¡ ¢ ¡ ¢ ¡ ¢ ¡ ¢
(16.3.viii)
Subtracting out the first term on each side and then solving for p, we get
Ã !(ρ+1)
∗ e 2M + e 2Y
p = . (16.3.ix)
e 1M + e 1Y
When e 1M = e 2M = e 1Y = e 2Y , this simplifies to p ∗ = 1 — consistent with

what we did in part A.
(e) Derive the set of pareto efficient allocations assuming e 1M = e 2M = e 1Y = e 2Y .

Can you see why, regardless of how we might redistribute endowments, the
equilibrium price will always be p = 1?
Answer: Let e = e 1M = e 2M = e 1Y = e 2Y . Then the economy is endowed with
2e of each good, which implies that, for any allocation (x1M , x2M ) that I get,
what’s left over for you is (2e − x1M ), (2e − x2M ). The pareto efficient set of
(x1M , x2M ) (with its implied consumption levels for you) is then defined as
the set where our MRS’s are equal to one another. The MRS for me at a
bundle (x1M , x2M ) is
Ã !(ρ+1)
M
∂u(x1M , x2M )/∂x1 x2M
MRS =− =− (16.3.x)
∂u(x1M , x2M )/∂x2 x1M
and the MRS for you at the left-over bundle ((2e − x1M ), (2e − x2M )) is
Ã !(ρ+1)
Y
∂u((2e − x1M ), (2e − x2M ))/∂x1 2e − x2M
MRS = − =− . (16.3.xi)
∂u((2e − x1M ), (2e − x2M ))/∂x2 2e − x1M
Setting MRS M equal to MRS Y and solving for x2M , we get
x2M = x1M ; (16.3.xii)
i.e. the contract curve is a straight line with slope 1 and intercept 0 —
the 45-degree line in the Edgeworth Box. Since all efficient allocations
happen on this line, and since equilibria are efficient, we know that any
competitive equilibrium lies on the 45-degree line. This further implies
that, when we plug x1M = x2M and 2e − x1M = 2e − x2M into the equations
for marginal rates of substitution, we get MRS M = −1 = MRS Y in any
equilibrium, which can only hold if the slope of the budget is −1. And
that can only be true if p = 1.
Exercise 16.5
In this exercise we explore some technical aspects of general equilibrium theory

in exchange economies and Robinson Crusoe economies. Unlike in other problems,
parts A and B are applicable to both those focused on A-Section material and those
focused on B-Section material. Although the insights are developed in simple exam-
ples, they apply more generally in much more complex models.
A: The role of convexity in Exchange Economies: In each part below, suppose

you and I are the only individuals in the economy, and pick some arbitrary al-
location E in the Edgeworth Box as our initial endowment. Assume throughout
that your tastes are convex and that the contract curve is equal to the line con-
necting the lower left and upper right corners of the box.
(a) Begin with a depiction of an equilibrium. Can you introduce a non-convexity

into my tastes such that the equilibrium disappears (despite the fact that
the contract curve remains unchanged?)
Answer: This is done in panel (a) of Exercise Graph 16.5 where the equi-
librium budget passes through E and is tangent to both solid (and con-
vex) indifference curves at A. Thus, A is an equilibrium allocation. How-
ever, if I permit my indifference curves to have non-convexities, I can
maintain the tangency at A but lose the equilibrium at A by having my
indifference curve continue along the dashed curve beginning at B and
moving right. Notice that A is still efficient — but, when faced with the
budget line that previously supported A as an equilibrium, I now no longer
optimize at A but rather at C which lies on a higher dashed (and non-
convex) indifference curve.
Exercise Graph 16.5 : Convexity Assumptions in General Equilibrium
(b) True or False: Existence of a competitive equilibrium in an exchange econ-

omy cannot be guaranteed if tastes are allowed to be non-convex.
Answer: This is true, as we have just shown.
(c) Suppose an equilibrium does exist even though my tastes exhibit some non-
convexity. True or False: The first welfare theorem holds even when tastes
have non-convexities.
Answer: The allocation A in panel (a) of Exercise Graph 16.5 would con-
tinue to be an equilibrium so long as the non-convexity that is introduced
is not sufficiently pronounced so as to cause the indifference curve that
is tangent at A to cross the budget line. Thus, had we drawn the non-
convexity in a less pronounced manner, the budget line through A and
E would still have been such that I optimize at A — and thus A would
have continued to be an equilibrium. We can conclude that, if an equi-
librium exists in the presence of non-convex tastes, then it will indeed still
be efficient. The first welfare theorem therefore holds in the presence of
non-convexities.
(d) True or False: The second welfare theorem holds even when tastes have
non-convexities.
Answer: The second welfare theorem says that any efficient allocation
can be an equilibrium allocation so long as endowments can be appro-
priately redistributed. We have just shown in panel (a) of Exercise Graph
16.5 an example of an efficient allocation A that cannot be supported as
an equilibrium no matter where we move the endowment. This is be-
cause, in order to support A as an equilibrium, the budget line has to be
the line that is draw in the graph — because that is the only budget that
will cause you to optimize at A. But that line crosses the dashed exten-
sion of my indifference curve that is tangent at A — implying that I will
not optimize at A if my tastes are the non-convex kind in the graph. Thus,
we have identified a case where an efficient allocation cannot become an
equilibrium allocation regardless of where we put the endowment. The
statement is therefore false — the second welfare theorem may not hold
when tastes have non-convexities.
B: The role of convexity in Robinson Crusoe Economies: Consider a Robinson
Crusoe economy. Suppose throughout that there is a tangency between the worker’s
indifference curve and the production technology at some bundle A.
(a) Suppose first that the production technology gives rise to a convex produc-
tion choice set. Illustrate an equilibrium when tastes are convex. Then
show that A may no longer be an equilibrium if you allow tastes to have
non-convexities even if the indifference curve is still tangent to the produc-
tion choice set at A.
Answer: This is illustrated in panel (b) of Exercise Graph 16.5. The solid
indifference curve is tangent to the convex production choice set at A,
with both tangent to the isoprofit/budget line (that has slope w/p). When
viewed as a budget line, the worker is doing the best he can by choosing
A, and when viewed as an isoprofit line, the firm is doing the best it can at
A, with the wage/price ration w/p supporting A as an equilibrium. But
we can take the same indifference curve, keep it tangent to the budget at
A, but then change its shape from B on to take the shape of the dashed
curve. When we do this, we introduce a non-convexity — and, as a result,
the worker is no longer doing the best he can by choosing A when con-
fronting the budget formed by the former equilibrium wage/price ratio.
In particular, the worker would now be better off optimizing at C — but
that lies outside the production frontier and is therefore not an equilib-
rium. Thus, by introducing the non-convexity, A ceases to be a competi-
tive equilibrium in this economy.
(b) Next, suppose again that tastes are convex but now let the production choice
set have non-convexities. Show again that A might no longer be an equi-
librium (even though the indifference curve and production choice set are
tangent at A).
Answer: This is shown in panel (c) of Exercise Graph 16.5 where the pro-
duction frontier f is tangent to the indifference curve u — thus making
A an efficient production plan. The budget that is tangent to both the

production frontier and the indifference curve at A — with slope w/p —
causes the worker to optimize at A where his indifference curve is tan-
gent. However, the firm would not be optimizing at A — because it can
reach a higher isoprofit curve and would maximize profit at C instead.
The production plan A would be optimal for the firm (and would thus be
an equilibrium) if the production frontier took on the dashed shape fol-
lowing A — i.e. if the production choice set were convex. But A is lost
as an equilibrium because of the non-convexity of the solid production
choice set.
(c) True or False: A competitive equilibrium may not exist in a Robinson Cru-
soe economy that has non-convexities in either tastes or production.
Answer: This is true as shown in the previous two parts.
(d) True or False: The first welfare theorem holds even if there are non-convexities
in tastes and/or production technologies.
Answer: This is true. The non-convexities may cause there to be no equi-
librium, but if there is an equilibrium, it will again have the feature that
the indifference curve is tangent to the production frontier at that point
— which will make it efficient. You can see this in panels (b) and (c) if
you imagine the non-convexity that was introduced as being less pro-
nounced. In panel (b), A would remain an equilibrium so long as the
dashed portion of the indifference curve does not cross the budget line to
the right of B — which is certainly possible even if there were a less pro-
nounced non-convexity. And that equilibrium would be efficient. Sim-
ilarly, in panel (c) you can imagine a non-convexity in the production
choice set either to the left of A or some distance to the right of A — and
you can imagine such a non-convexity to not be sufficiently pronounced
so as to cross the isoprofit line that is tangent at A. In that case, A would
remain as an equilibrium — and it would be efficient. Thus, the first wel-
fare theorem holds — every equilibrium (that exists) is indeed efficient.
(e) True or False: The second welfare theorem holds regardless of whether there
are non-convexities in tastes or production.
Answer: This is false. In panel (b) of the graph, we have shown an efficient
point A that cannot be an equilibrium because the budget line that must
support it crosses the indifference curve that is tangent at A. In panel (c)
we have shown another efficient point A that cannot be supported as an
equilibrium because the isoprofit line that is needed to support it as an
equilibrium crosses the production frontier because of a non-convexity.
We have therefore shown that, when there are non-convexities, there may
be efficient outcomes that cannot be supported as equilibria.
(f) Based on what you have done in parts A and B, evaluate the following:
“Non-convexities may cause a non-existence of competitive equilibria in
general equilibrium economies, but if an equilibrium exists, it results in
an efficient allocation of resources. However, only in the absence of non-
convexities can we conclude that there always exists some lump-sum re-
distribution such that any efficient allocation can also be an equilibrium

allocation.” (Note: Your conclusion on this holds well beyond the exam-
ples in this problem — for reasons that are quite similar to the intuition
developed here.)
Answer: The statement is fully consistent with everything we have done
in this exercise. We have shown — in both exchange and Robinson Cru-
soe economies — that non-convexities may lead to a non-existence of
equilibria; that if equilibria exist, they will be efficient (i.e. the first wel-
fare theorem holds); but not all efficient outcomes can be supported as
equilibria (i.e. the second welfare theorem fails in the presence of non-
convexities).
Exercise 16.7
Everyday Application: Parents, Children and the Degree of Substitutability across

Time: Consider again exactly the same scenario as in exercise 16.6.
A: This time, however, suppose that parent and child tastes treat consumption
now and consumption in the future as perfect complements.
(a) Illustrate in an Edgeworth Box an equilibrium with a single parent and a
single child.
Answer: Perhaps the most obvious equilibrium is the one with price equal
to 1 and thus a budget line that runs from E in the lower right corner to
the upper left corner of the box — with the equilibrium allocation at A,
pictured in panel (a) of Exercise Graph 16.7.
Exercise Graph 16.7 : Parents and Children: Part 1
(b) Is the equilibrium you pictured in (a) the only equilibrium? If not, can you
identify the set of all equilibrium allocations?
Answer: It is not the only equilibrium — in fact, panel (a) of Exercise
Graph 16.7 picture two others, with allocations at B and at C . Because
of the sharp corners on indifference curves for perfect complements, any
budget line with negative slope can be fit to any “tangency” of the two
indifference curves on the 45 degree line. Thus, all allocations on the 45
degree line have some budget line that passes through the endowment
allocation E and is “tangent” to both indifference curves on that point of
the 45 degree line. The entire 45-degree line in the box is therefore the set
of possible equilibrium allocations.
(c) Now suppose that there were two children and one parent. Keep the Edge-
worth Box with the same dimensions but model this by recognizing that,
on any equilibrium budget line, it must now be the case that the parent
moves twice as far from the endowment E as the child (since there are two
children and thus any equilibrium action by a child must be half the equi-
librium action by the parent). Are any of the equilibrium allocations for
parent and child that you identified in (b) still equilibrium allocations?
(Hint: Consider the corners of the box.)
Answer: For any budget line that intersects the 45-degree line inside the
box, both parent and child will optimize on the 45 degree line. But with
two children and one parent, that cannot be an equilibrium — because
the parent’s action must be twice the children’s in the opposite direction
in order for demand to equal supply. Thus, none of the efficient alloca-
tions on the 45 degree line inside the box can be an equilibrium alloca-
tion. However, suppose that p = ∞. Then the budget line becomes ver-
tical and passes through E . The parent will optimize at the top corner
(point D in panel (b) of Exercise Graph 16.7), and the children don’t care
where on the budget they optimize because all the bundles on that bud-
get lie on the same indifference curve. Thus, it is not inconsistent with
optimization to assume that the children will choose F — halfway up the
budget and halfway to D where the parent optimizes. Thus, children con-
sume nothing now and give half of what they earn in the future to the par-
ent, and parents consume everything now and half of everything (i.e. half
of what each of the two children earns) in the future.
(d) Suppose instead that there are two parents and one child. How does your
answer change?
Answer: No equilibrium allocation can lie on the 45 degree line for the
same reason as in the previous case — and now we end up with the child
optimizing at G in panel (b) of Exercise Graph 16.7 and the two parents
optimizing at H , with p = 0. Thus, parents consume half their income
now and nothing in the future, while children consume half of each par-
ents’ income now and everything in the future.
(e) Repeat (a) through (d) for the case where consumption now and consump-
tion in the future are perfect substitutes for both parent and child.
Answer: When consumption across time is perfectly substitutable, the in-
difference curves have slope −1 at every allocation in the Edgeworth Box.
Thus any equilibrium allocation inside the box must lie on the line con-
necting the upper left to the lower right corners of the box — the line pic-
tured in panel (c) of Exercise Graph 16.7. Neither parent nor child cares
where on that line they consume — and thus any split of the economy’s
endowment that falls on this line will be an equilibrium allocation with
p = 1. For instance, when there is one child and one parent, A is a pos-
sible equilibrium allocation, as is C and B. When there are two children
and 1 parent, any allocation that has the parent’s bundle twice as far from
E as the children’s works — for instance A for the parent and C for the
children. When there are two parents and one child, then any allocation
that has the child twice as far as the parents from E works. In all cases, the
equilibrium price continues to be p = 1 — because it makes no sense for
individuals to trade on other terms when consumption now is the same
as consumption in the future.
(f) Repeat for the case where consumption now and consumption in the future
are perfect complements for parents and perfect substitutes for children.
Answer: Consider first the case of one parent and one child. For any bud-
get with positive slope (not equal to infinity), the parent will optimize on
the 45-degree line. For any price not equal to 1, the child will choose a
corner solution (since consumption now and in the future are the same
for her). Thus, the only way the child will trade to permit the parent to
get to the 45 degree line is if p = 1 and the budget line takes the shape
graphed in panel (c) of Exercise Graph 16.7. The equilibrium allocation
is then A — where the parent’s indifference curve is drawn as a dotted
L-shape. Next, suppose there are two children. Nothing has changed in
terms of the children’s willingness to trade to an interior solution only at
p = 1 and in terms of the parent’s optimal bundle falling on the 45 degree
line for any positive price. Thus, p will remain 1, the parent will optimize
at A and the children will each optimize at C — halfway between A and E .
Finally, suppose there are two parents and one child. Again, for the same
reasons as before, price has to remain 1, and the parents’ optimization
has to lead to A. Thus, parents end up at A and the child ends up at the
top left corner D — twice as far from E as the two parents.
(g) True or False: The more consumption is complementary for the parent rel-
ative to the child, and the more children there are per parent, the more
gains from trade will accrue to the parent.
Answer: This is roughly true, as illustrated in the previous parts of the
question. For instance, when parent viewed consumption as perfectly
complementary across time while children viewed it as substitutable (in
panel (c) of Exercise Graph 16.7), the children gain no utility from trading
while the parent(s) get all gains from trade. Similarly, we saw in this and
the previous exercise that more gains typically accrue to the party that
is in control of the goods that are scarcer. Parents are in control of con-
sumption now — which is relatively more scarce the more children there
are per parent.
B: Suppose that parent and child tastes can be represented by the CES utility
¡ −ρ −ρ ¢−1/ρ
function u(c 1 , c 2 ) = 0.5c 1 + 0.5c 2 . Assume that the income earned by
parents in period 1 and by children in period 2 is 100.
(a) Letting p denote the price of consumption now with price of future con-
sumption normalized to 1, derive parent and child demands for current
and future consumption as a function of ρ and p.
Answer: We want to maximize utility (which is the same for parents and
children) subject to the budget constraint — which is 100p = pc 1 + c 2 for
parents (who are endowed with 100 now) and 100 = pc 1 + c 2 for children
(who are endowed with 100 in the future). Solving this in the usual way,
we get
100p 1/(ρ+1) 100p

c 1P = and c 2P = for parents, and (16.7.i)
p ρ/(ρ+1) + 1 p ρ/(ρ+1) + 1
100 100
c 1C = and c 2C = for children. (16.7.ii)
p + p 1/(ρ+1) p ρ/(ρ+1) + 1
(b) What is the equilibrium price — and what does this imply for equilibrium
allocations of consumption between parent and child across time. Does
any of your answer depend on the elasticity of substitution?
Answer: This solves slightly more easily if we set demand and supply in
the c 2 market equal to one another (rather than setting it equal to one
another in the c 1 market. Of course the latter would give the same answer
even if it is slightly more burdensome to get there.) Thus, we need to solve
100p 100
+ = 100. (16.7.iii)
p ρ/(ρ+1) + 1 p ρ/(ρ+1) + 1
Dividing by 100, multiplying by the denominator on the left hand side,

and simplifying, we get
p = p ρ/(ρ+1) or 1 = ρ −1/(ρ+1) (16.7.iv)
which solves to p = 1. The answer therefore does not depend on ρ and

thus is independent of the elasticity of substitution. (This is because the
indifference curves for the utility function always have MRS = −1 along
the 45 degree line no matter what elasticity of substitution is assumed.)
(c) Next, suppose there are 2 children and only 1 parent. How does your an-
swer change?
Answer: We now have to sum twice the child demands with the parent
demand for c 2 and set it equal to overall consumption in the future —
which is 200 when there are two children. This implies we need to solve
µ ¶
100p 100
+2 = 200 (16.7.v)
p ρ/(ρ+1) + 1 p ρ/(ρ+1) + 1
which solves to
p = 2ρ+1 . (16.7.vi)
The equilibrium price now depends on ρ and thus on the elasticity of

substitution. As ρ increases — which implies the elasticity of substitution
falls — price increases. In the limit, as ρ approaches infinity — and con-
sumption becomes perfectly complementary across time — price rises to
infinity. This is exactly what we concluded in part A for perfect comple-
ments. As ρ falls to −1 — and consumption becomes perfectly substi-
tutable across time, on the other hand, price becomes 1 — again exactly
as we concluded in part A.
(d) Next, suppose there are 2 parents and only 1 child. How does your answer
change?
Answer: We now have to sum twice the parent demands with the child
demand for c 2 and set it equal to overall consumption in the future —
which is 100 when there is only one child. This implies we need to solve
µ ¶
100p 100
2 + = 100 (16.7.vii)
p ρ/(ρ+1) + 1 p ρ/(ρ+1) + 1
which solves to
µ ¶ρ+1
1
p= . (16.7.viii)
2
The equilibrium price again depends on ρ and thus on the elasticity of

substitution. As ρ increases — which implies the elasticity of substitution
falls — price falls. In the limit, as ρ approaches infinity — and consump-
tion becomes perfectly complementary across time — price falls to zero.
This is exactly what we concluded in part A for perfect complements. As
ρ falls to −1 — and consumption becomes perfectly substitutable across
time, on the other hand, price becomes 1 — again exactly as we con-
cluded in part A.
(e) Explain how your answers relate to the graphs you drew for the extreme
cases of both parent and child preferences treating consumption as perfect
complements over time.
Answer: We already did this. We showed that, as tastes become perfectly
complementary, then p approaches infinity if there are two children and
one parent and to zero if there are two parents and one child. We illus-
trated precisely this extreme case in panel (b) of Exercise Graph 16.7.
(f) Explain how your answers relate to your graphs for the case where con-
sumption was perfectly substitutable across time for both parents and chil-
dren.
Answer: Again, we already did this. We showed that, when consumption
is perfectly substitutable across time, then price will be 1 regardless of the
number of children relative to parents.
Exercise 16.9
Business Application: Hiring an Assistant: Suppose you are a busy CEO — with
lots of consumption but relatively little leisure. I, on the other hand, have only a
part-time job and therefore lots of leisure with relatively little consumption.
A: You decide that the time has come to hire a personal assistant — someone
who can do some of the basics in your life so that you can have a bit more leisure
time.
(a) Illustrate our current situation in an Edgeworth Box with leisure on the
horizontal and consumption on the vertical axis. Indicate an endowment
bundle that fits the description of the problem and use indifference curves
to illustrate a region in the graph where both of us would benefit from me
working for you as an assistant.
Answer: This is illustrated in panel (a) of Exercise Graph 16.9 where the
endowment allocation E has me (on the lower axes) with lots of leisure
but little consumption and you (on the upper axes) with the reverse. The
mutually beneficial region is formed by the lens made from our indiffer-
ence curves that pass through E . Both of us would prefer any allocation
in that lens shape to the endowment bundle E .
Exercise Graph 16.9 : Cheerfulness in Office Assistants
(b) Next, illustrate what an equilibrium would look like. Where in the graph
would you see the wage that I am being paid?
Answer: This is also illustrated in panel (a) of Exercise Graph 16.9 where
the budget line that passes through A and E has slope −w (where w is the
wage when the price of consumption is normalized to 1).
(c) Suppose that anyone can do the tasks you are asking of your assistant —
but some will do it cheerfully and others will do it with attitude. You hate
attitude — and therefore would prefer someone who is cheerful. Assuming
you can read the level of cheerfulness in me, what changes in the Edge-
worth box as your impression of me changes?
Answer: As you think I am more cheerful, you will be willing to trade more
of your consumption for an increase in your leisure. Thus, your indiffer-
ence curves become steeper.
(d) How do your impressions of me — i.e. how cheerful I am — affect the region
of mutually beneficial trades?
Answer: This is illustrated in panel (b) of Exercise Graph 16.9 where your
original indifference curve through E is illustrated as a dashed indiffer-
ence curve and your new indifference curve (that contains E ) as my cheer-
fulness increases is illustrated as a bold curve. This increases the lens
formed by our indifference curves through E — and thus the mutually
beneficial region.
(e) How does increased cheerfulness on my part change the equilibrium wage?
Answer: This is illustrated in panel (c) of Exercise Graph 16.9 where A
is the original equilibrium at low levels of cheerfulness and B is the new
equilibrium at higher levels of cheerfulness. As my cheerfulness increases,
your indifference curve through A becomes steeper — rotating from the
dashed curve to the solid one. Thus, A can’t be an equilibrium anymore
because you now want more of me but I am not willing to offer any more
at the original wage. Thus, the wage must increase in order to get me to
offer more of myself and you to reduce your demand for me. This leads
us to the steeper budget through B — with a higher wage. Cheerfulness is
rewarded in the competitive market.
(f) Your graph probably has the new equilibrium (with increased cheerful-
ness) occurring at an indifference curve for you that lies below (relative to
your axes) the previous equilibrium (where I was less cheerful). Does this
mean that you are worse off as a result of me becoming more cheerful?
Answer: No, it does not. It is indeed true that your indifference curve
through B in panel (c) of Exercise Graph 16.9 lies below A (relative to
your axes). But this does not mean you are less happy — because my
cheerfulness is what made your indifference curves get steeper. In terms
of some of the earlier problems in our development of consumer theory,
cheerfulness is a third good you care about — and as it changes in the
problem, you switch to a different “slice” of your 3-dimensional indiffer-
ence surfaces. Increased cheerfulness switches you to a slice where you
are happier for any level of consumption and leisure than you were be-
fore — and so an indifference curve with more cheerfulness can lie below
one with less cheerfulness and still be preferred.
B: Suppose that my tastes can be represented by u(c, ℓ) = 200ln ℓ + c while yours

can be represented by u(c, ℓ, x) = 100x ln ℓ+c where ℓ stands for leisure, c stands
for consumption and x stands for cheerfulness of your assistant. Suppose that,
in the absence of working for you, I have 50 leisure hours and 10 units of con-
sumption while you have 10 leisure hours and 100 units of consumption.
(a) Normalize the price of c as 1. Derive our leisure demands as a function of

the wage w.
Answer: My budget constraint is wℓ+c = 50w +10 while yours is wℓ+c =
10w + 100. Maximizing our utilities subject to these constraints, we get
(by solving this in the usual way)
200 100x
ℓM = for me and ℓY = for you. (16.9.i)
w w
(b) Calculate the equilibrium wage as a function of x.

Answer: The sum of our leisure demands has to be equal to the leisure
supply of 60 in equilibrium — i.e.
200 100x
+ = 60 (16.9.ii)
w w
which implies that the equilibrium wage is
10 + 5x
w∗ = . (16.9.iii)
3
(c) Suppose x = 1. What is the equilibrium wage, and how much will I be
working for you?
Answer: Substituting x = 1 into our equation for w ∗ , we get an equilib-
rium wage of 5. Plugging this wage into our leisure demand equations,
we get that you will have 20 hours of leisure and I will have 40 — which
is 10 less for me and 10 more for you than what we were endowed with.
Thus, I’ll be working for you for 10 hours.
(d) How does your MRS change as my cheerfulness x increases?
Answer: Your MRS is
∂u(c, ℓ, x)/∂ℓ 100x

MRS Y = − =− . (16.9.iv)
∂u(c, ℓ, x)/∂x ℓ
Thus, for any bundle (ℓ, c), the MRS gets larger in absolute value as x in-
creases — i.e your indifference curves become steeper as my cheerfulness
increases.
(e) What happens to the equilibrium wage as x increases to 1.2? What hap-
pens to the equilibrium number of hours I work for you? What if I get
grumpy and x falls to 0.4?
Answer: When x goes to 1.2, the equilibrium wage rises to 5.33 and the
number of hours I work for you increases to 12.5. When x falls to 0.4,
the equilibrium wage falls to 4 but you no longer hire me and we simply
consume at our endowment bundles.
Exercise 16.11
Policy Application: Distortionary Taxes in General Equilibrium: Consider, as in

exercise 16.10, a 2-person exchange economy in which I own 200 units of x1 and 100
units of x2 while you own 100 units of x1 and 200 units of x2 .
A: Suppose you and I have identical homothetic tastes.

(a) Draw the Edgeworth Box for this economy and indicate the endowment
allocation E .
Answer: This is illustrated in panel (a) of Exercise Graph 16.11 where the
box takes on the shape of a square since the economy’s endowment of
both goods is 300.
Exercise Graph 16.11 : Distortionary Taxes
(b) Normalize the price of good x2 to 1. Illustrate the equilibrium price p ∗

for x1 and the equilibrium allocation of goods in the absence of any taxes.
Who buys and who sells x1 ?
Answer: This is also done in panel (a) of Exercise Graph 16.11 where A
is the equilibrium allocation (which appears in the center of the box on
the 45 degree line because of our identical homothetic tastes and endow-
ments.) Thus, I sell 50 units of x1 to you for the price of p ∗ = 1.
(c) Suppose the government introduces a tax t levied on all transactions of x1
(and paid in terms of x2 ). For instance, if one unit of x1 is sold from me to
you at price p, I will only get to keep (p −t ). Explain how this creates a kink
in our budget constraints.
Answer: This implies that the price p paid by the buyer is greater than the
price (p − t ) received by the seller. On my budget constraint, I am a seller
to the left of E and a buyer to the right of E — implying that my budget
has shallower slope −(p − t ) to the left of E and steeper slope −p to the
right of E , with a kink at E . The same is true for you — except that “right”
and “left” are reversed when we flip your axes to create the Edgeworth
Box. The portions along which I am a seller and you are a buyer of x2 are
illustrated as the solid lines in panel (b) of Exercise Graph 16.11, with the
remaining portion of the constraints dashed to the right of E .
(d) Suppose a post-tax equilibrium exists and that price increases for buyers
and falls for sellers. In such an equilibrium, I will still be selling some
quantity of x1 to you. (Can you explain why?) How do the relevant por-
tions of the budget constraints you and I face look in this new equilibrium,
and where will we optimize?
Answer: This is illustrated in panel (b) of Exercise Graph 16.11 where the
steeper (solid) constraint is yours (with the higher post-tax price) and the
shallower one is mine (with the lower pre-tax price). In equilibrium, it
will still have to be the case that the amount of x1 I sell to you is equal to
the amount of x1 you buy. Thus, in equilibrium, our two budgets have
to be such that your optimum B lies right above my optimum C in the
Edgeworth Box. We know that this will be to the right of the original equi-
librium A — because your budget is steeper than before and mine is shal-
lower than before. The fact that it is shallower for me means that I will be
optimizing on a shallower ray from the origin (given that my tastes are
homothetic), and the fact that it is steeper for you implies you will be op-
timizing on a steeper ray from your origin. Thus, the amount we trade will
fall by the amount of the arrows in the graph. (The reason we know that I
will still be selling (or at least not buying) x1 under the tax is as follows: My
budget under the tax has a kink at E — and becomes steeper to the right
of E . Given that my tastes are homothetic, it cannot be that I optimize
on that steeper portion — because the steeper parts of my indifference
curves lie to the left of E ).
(e) When we discussed price changes with homothetic tastes in our develop-
ment of consumer theory, we noted that there are often competing income
(or wealth) and substitution effects. Are there such competing effects here
relative to our consumption of x1 ? If so, can we be sure that the quantity
we trade in equilibrium will be less when t is introduced?
Answer: Both of us experience a negative wealth effect — me because
what I am selling has fallen in price, you because what you are buying
has increased in price. Thus, the wealth effect says “consume less of x1 ”
for both of us. But the substitution effects operate in opposite directions
for the two of us. For me, the price of x1 falls as a result of the tax — which
means the substitution effect will tell me to consume more of x1 . For you,
on the other hand, the price of x1 has increased — with the substitution
effect therefore telling you to consume less of x1 . The wealth and sub-
stitution effects therefore point in opposite directions for me but not for
you. This implies you will consume less x1 under the tax, which means
in equilibrium the prices have to adjust such that I will sell you less (and
therefore consume more) even though the wealth effect tells me to con-
sume less. (This implies that the equilibrium that we assume exists (with
price increasing for buyers and falling for sellers) requires that the goods
are sufficiently substitutable to create the necessary substitution effect.)
(f) You should see that, in the new equilibrium, a portion of x2 remains not
allocated to anyone. This is the amount that is paid in taxes to the govern-
ment. Draw a new Edgeworth box that is adjusted on the x2 axes to reflect
the fact that some portion of x2 is no longer allocated between the two of
us. Then locate the equilibrium allocation point that you derived in your
previous graph. Why is this point not efficient?
Answer: The portion of x2 that remains not allocated in our tax-equilibrium
in panel (b) of the graph is the vertical difference between B and C — la-
beled T R in the graph. Thus, the amount that gets allocated is T R less of
x2 than what is available — because the difference is collected by the gov-
ernment. If we shrink the Edgeworth Box by that vertical amount, we get
the box illustrated in panel (c) of Exercise Graph 16.11. By shrinking the
height of the box, we move B on top of C and now see even more clearly
than in panel (b) that this allocation is not efficient. The reason it is in-
efficient is that both you and I would prefer to divide everything that was
not taken by the government differently — with all the allocations in the
lens shape between our indifference curves through B = C all preferred
by both of us. We could thus make everyone better off by moving the al-
location into that lens shape without taking any of the tax revenue the
government has raised back.
(g) True or False: The deadweight loss from the distortionary tax on trades
in x1 results from the fact that our marginal rates of substitution are no
longer equal to one another after the tax is imposed and not because the
government raised revenues and thus lowered the amounts of x2 consumed
by us.
Answer: This is true. The inefficiency we show in panel (c) arises from
the fact that there is a lens shape between our indifference curves — and
that lens shape arises from the fact that our marginal rates of substitution
are not equal to one another (which is due to the fact that the prices we
face as buyers and sellers is different when the government uses price-
distorting taxes). The fact that the box has shrunk is not evidence of an
inefficiency — because the government now has the difference and may
well be doing some very useful things with the money. The problem is
that what remains is not allocated efficiently due to distorted prices.
(h) True or False: While the post-tax equilibrium is not efficient, it does lie in
the region of mutually beneficial trades.
Answer: This is true. In panel (b), the indifference curves through B and
C still lie above E for both of us — i.e. trade is still making us better off
than we would be without trade, just worse off than we would be if we
could trade without price distortions. (Even if it is not obvious from the
graph that our indifference curves through B and C lie above E , it should
intuitively make sense that this has to be the case: After all, even in the
presence of the distortionary tax, no one is forcing us to trade with one

another — and we would not do so if trade made us worse off than we
would be if we simply consumed our endowments.)
(i) How would taxes that redistribute endowments (as envisioned by the Sec-
ond Welfare Theorem) be different than the price distorting tax analyzed
in this problem?
Answer: Redistributions of endowments would involve lump sum taxes
and subsidies that do not distort prices — because they would simply
shift E around in the box. From the new E , markets could act as before —
finding the competitive equilibrium price and causing the individuals to
optimize where their indifference curves are tangent to one another and
the resulting allocation is therefore efficient.
B: Suppose our tastes can be represented by the utility function u(x1 , x2 ) = x1 x2 .

Let our endowments be specified as at the beginning of the problem.
(a) Derive our demand functions for x1 and x2 (as functions of p — the price
of x1 when the price of x2 is normalized to 1).
Answer: My budget constraint is p x1 +x2 = 200p +100 while yours is p x1 +
x2 = 100p + 200. Solving our utility maximization problems subject to
these constraints in the usual way, we get
100p + 50
x1M = and x2M = 100p + 50 for me, and (16.11.i)
p
50p + 100
x1Y = and x2Y = 50p + 100 for you. (16.11.ii)
p
(b) Derive the equilibrium price p ∗ and the equilibrium allocation of goods.
Answer: To derive the equilibrium price, we can sum the demands for
x1 and set them equal to 300 — the amount of x1 that the economy is
endowed with. Solving for p, we get p ∗ = 1. Substituting back into the
demand equations, we get x1M = x2M = x1Y = x2Y = 150.
(c) Now suppose the government introduces a tax t as specified in A(c). Given
that I am the one that sells and you are the one that buys x1 , how can you
now re-write our demand functions to account for t ? (Hint: There are two
ways of doing this — either define p as the pre-tax price and let the relevant
price for the buyer be (p + t ) or let p be defined as the post-tax price and let
the relevant price for the seller be (p − t ).)
Answer: Letting p indicate the price paid by you and (p − t ) be equal to
the price received by me (as the seller), we can substitute (p − t ) into my
demand equations to get
100(p − t ) + 50
x1M (t ) = and x2M (t ) = 100(p − t ) + 50 (16.11.iii)
(p − t )
Your demand functions would remain the same as before.

(d) Derive the new equilibrium pre- and post-tax prices in terms of t . (Hint:
You should get to a point where you need to solve a quadratic equation us-
ing the quadratic formula that gives two answers. Of these two, the larger
one is the correct answer for this problem.)
Answer: We again set demand for x1 equal to supply to get the equation
100(p − t ) + 50 50p + 100

x1M (t ) + x1Y = + = 300. (16.11.iv)
(p − t ) p
Multiplying both sides by (p − t )p, taking all terms to one side, summing
like terms and dividing by 50, we get
3p 2 − 3(t + 1)p + 2t = 0. (16.11.v)
Applying the quadratic formula (and accepting the higher of the two so-
lutions), we get
q
2 2
9(t + 1)2 − 4(3)(2t ) (t + 1) + t − 3 t + 1
p
3(t + 1) +
p= = (16.11.vi)
6 2
which is the post-tax equilibrium price that buyers pay. The pre-tax price
that sellers receive is then simply t less; i.e.
q
(1 − t ) + t 2 − 32 t + 1
(p − t ) = . (16.11.vii)
2
(e) How much of each good do you and I consume if t = 1?
Answer: Plugging t = 1 into our equations for p and (p − t ), we get p ≈
1.5774 and (p − t ) ≈ 0.5774. Plugging these into our demand equations,
we get
x1M ≈ 186.60, x2M ≈ 107.74, x1Y ≈ 113.40 and x2Y ≈ 178.87. (16.11.viii)
(f) How much revenue does the government raise if t = 1?

Answer: The tax revenue must be the difference between the 300 units of
x2 that were available in the economy and the sum of our consumption
levels of x2 ; i.e. tax revenue must be 300 − (107.74 + 178.87) = 13.39. We
can verify that this is the case by multiplying t = 1 times the quantity of
x1 that is sold by me to you in equilibrium — i.e. (1)(200−186.60) = 13.40.
(The difference between the two values for tax revenue is rounding error.)
(g) Show that the equilibrium allocation under the tax is inefficient.
Answer: To show that the equilibrium allocation is inefficient, all we have
to show is that our marginal rates of substitution at the equilibrium con-
sumption bundles are not the same. For the utility function we are using,
the MRS is given by
∂u/∂x1 x2
MRS = − =− . (16.11.ix)
∂u/∂x2 x1
Plugging in our consumption levels from equation (16.11.viii), we get MRS M ≈

0.5774 and MRS Y ≈ 1.5774 for you — which are, of course, equal to the
negative (p−t ) and p values we calculated earlier and that form the slopes
of our two equilibrium budget constraints.
Conclusion: Potentially Helpful Reminders

1. Keep in mind that the dimensions of the Edgeworth Box are determined by
the overall endowment of each of the goods in the economy. A point in the
Edgeworth Box has four components — two measured on each of the axes
that correspond to the two individuals in the economy.
2. The set of mutually beneficial trades can easily be found by drawing the in-
difference curves (for the two individuals) that pass through the endowment
point in the Edgeworth Box. (This usually gives us a lens-shaped set of mu-
tually beneficial trades.) Within this set, only some of the allocations are effi-
cient — because only some of them have the characteristic that the marginal
rates of substitution for the two individuals are equal to one another.
3. A competitive equilibrium in the Edgeworth Box always has the following fea-
tures: It consists of prices that form a budget line passing through the endow-
ment, with indifference curves for both individuals tangent to this budget at
the equilibrium allocation. The allocation is efficient because this tangency
implies that the marginal rates of substitution for the two individuals are the
same at that allocation — with no further gains from trade possible.
4. A competitive equilibrium in the Robinson Crusoe economy has similar fea-

tures: It consists of an isoprofit line that also doubles as a worker budget con-
straint, with this line tangent to both the production frontier and the worker’s
indifference curve.
5. Keep in mind that we are still assuming that individuals are all price takers
— and so we do not have to think about relative bargaining power when we
investigate competitive equilibria. This is sometimes hard to keep in mind
because the simple economies we are dealing with in this chapter only have
two individuals in them — and it is therefore artificial for us to treat them as if
they were price takers. (It seems even sillier in the Robinson Crusoe economy
where we treat a single individual as if he had a split personality!) But the
point here is to illustrate the basic intuitions that continue to hold when the
economies get much larger and the assumption becomes natural.
6. The mathematical steps in calculating an equilibrium in a general equilib-

rium economy follow straightforwardly from the Edgeworth Box (or Robin-
son Crusoe) pictures — so keep going back to the underlying pictures if you
get lost in the math steps.

Study Guide CH 16

Uploaded by

Copyright:

Available Formats

Study Guide CH 16

Uploaded by

Document Information

Original Title

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

Study Guide CH 16

Uploaded by

Copyright:

Available Formats

C H A P T E R

Chapter 14 introduced the positive idea of equilibrium in the context of a competi-

1. An exchange economy is an economy in which individuals trade what they

2. A competitive equilibrium in an exchange economy consists of a set of prices

5. Production can be introduced into general equilibrium economies, and the

16A Solutions to Within-Chapter-Exercises for

Exercise Graph 16A.1 : Edgeworth Boxes

Exercise Graph 16A.8 : Contract Curves

Exercise Graph 16A.11 : Price Changes in the Edgeworth Box

the price of oranges at p 1 = 2, we therefore have to set the price of bananas (p 2 ) to

Does this production frontier exhibit increasing, decreasing or constant returns

As drawn, which of our usual assumptions about tastes — rationality, convexity,

In votes on referenda on school vouchers, researchers have found that renters

If you were considering opening up a private school following the introduction

16B Solutions to Within-Chapter-Exercises for

U 1 = 33/4 61/4 ≈ 3.5676 and U 2 = 101/4 43/4 = 5.0297. ( 16B.3)

Exercise Graph 16B.6 : Contract Curve

L = x13/4 x21/4 + λ u ∗ − (13 − x1 )1/4 (10 − x2 )3/4 .

Taking first order conditions, we get

∂L 3x 1/4 (10 − x2 )3/4 ∂L x 3/4 3(13 − x1 )1/4

We can then write the Lagrange function as

which solve to give us

Demonstrate that the same equilibrium allocation of goods will arise if p 1 = 2

3(3(2) + 6(3)) (3(2) + 6(3))

Plugging x11 = 9 into the first two of these equations, we get

In our proof, we began by assuming that there exists a (y 11 , y 21 , y 12 , y 22 ) that is

p 1 z11 + p 2 z21 > p 1 x11 + p 2 x21 . ( 16B.12.i)

p 1 z12 + p 2 z22 ≥ p 1 x12 + p 2 x22 ( 16B.12.ii)

Adding equations ( 16B.12.i) and ( 16B.12.ii), we get

To say that this allocation is an equilibrium allocation under prices (p 1 , p 2 , ..., p M )

Verify the results in equation (16.27).

The first two first order conditions are then

which can also be written as

αx (α−1) (L − ℓ)(1−α) = −λ and (1 − α)x α (L − ℓ)−α = −λβAℓ(β−1) . ( 16B.14.iii)

αβL [1 − α(1 − β)]L − αβL (1 − α)L

Substituting this for (L − ℓ) into equation ( 16B.14.iv), we get

Verify that this is the correct solution.

and solving for ℓ, we get

L = x α (L − ℓ)(1−α) + λ(wℓ + π(w, p) − p x) ( 16B.16.i)

giving rise to first order conditions

Equation ( 16B.17.i) then becomes

A(βp A)β/(1−β) (1 − α(1 − β))p

Demonstrate that the equilibrium banana consumption (and production) is

Substituting our expression for w ∗ into x D gives the same answer.

Why is there no coalition to block A in the 2-person version of this economy?

16C Solutions to Odd Numbered

Consider a 2-person/2-good exchange economy in which person 1 is endowed

A: Suppose that tastes are homothetic for both individuals.

Exercise Graph 16.1(1) : Contract Curve with Homothetic Tastes

where person 1’s utility is maximized subject to getting person 2 to a par-

∂x2 (x 1 ) (1 − α)βE 2 (β − α)(1 − α)βE 2 x 1

Evaluated at x 1 = 0 and at x 1 = E 1 , we get

∂x2 (0) β(1 − α)E 2 ∂x2 (E 1 ) α(1 − β)E 2

∂x2 (0) E 2 ∂x2 (E 1 ) E 2

which we can solve for x 2 to get the contract curve

Exercise Graph 16.3 : Equal Endowments and Same Tastes

pe 1M + e 2M = p x1M + x2M , (16.3.i)

and for you it means

pe 1Y + e 2Y = p x1Y + x2Y . (16.3.ii)