Bachelor X PR: Matías Núñez

October 2021 TA: Claire Leroy

Ecole Polytechnique

ECO201: Intermediate Microeconomics

Solutions Problem Set 2 - Exchange

Exercise 1 Pareto-optimal allocations

Consider an exchange economy consisting of two consumers, 1 and 2, and two goods, x and y.
The economy's initial endowment is one unit of each good. Determine (and draw in an Edgeworth

box) the Pareto-optimal allocations for the dierent cases listed below:

1. u1 (x1 , y1 ) = xα1 y11−α and u2 (x2 , y2 ) = xβ2 y21−β , where 0 < α, β < 1.
Answer: :

1 2 α y1 β y2 α y1 β 1 − y1
M RSx→y = M RSx→y ⇐⇒ = ⇐⇒ =
1 − α x1 1 − β x2 1 − α x1 1 − β 1 − x1
β(1 − α)x 1
⇐⇒ y1P O (x1 ) =
α(1 − β) + (β − α)x1

Moreover, since preferences are strictly monotonic, the allocations x1 = y1 = 0 and x1 =

y1 = 1 are also Pareto-optimal (we note that they satisfy the previous equation of equality of
MRS). We can also check (graphically) that the other corner allocations cannot be Pareto-

optimal (indeed, since the utility of one of the consumers is zero as soon as they consume 0

quantity of one of the two goods, we can always take all the positive quantity of the other

good from this consumer and redistribute it to the other consumer increasing his utility

without reducing the utility of the rst consumer which remains at zero). Finally, we note

that y1P O (x1 ) is a strictly increasing function in x1 , which is strictly convex (respectively

concave) when β<α (resp. β > α), and linear when α=β (i.e. y1P O (x1 ) = x1 ).

Figure 1: Set of Pareto-optimal allocations - Cobb-Douglas utility functions

1
2. u1 (x1 , y1 ) = x1 + y1 and u2 (x2 , y2 ) = x2 + y2
Answer: These utility functions represent perfect substitutes preferences. In this case, any

allocation is Pareto-optimal (i.e. for any given allocation, it is impossible to reallocate re-

sources in a way that would strictly improve one agent's utility without negatively aecting

the other agent). Indeed, since the MRS are constant and equal to 1, it is as if there was

a single good to consume (the utility is thus a strictly increasing function of the quantity

consumed). In this case, any reallocation (that strictly improves the utility of one agent)

would necessarily lead to a reduction in utility for the other consumer.

Figure 2: Set of Pareto-optimal allocations - Perfect substitutes & equal MRS

3. u1 (x1 , y1 ) = 2x1 + y1 and u2 (x2 , y2 ) = x2 + y2

Answer: Again, these utility functions represent perfect substitutes preferences. However,

the (constant) MRS are not equal to each other:

1
M RSx→y 2
= 2 > M RSx→y = 1. In this

case, any interior allocation (i.e. such that all quantities consumed are strictly positive) is

not Pareto-optimal because for such allocation, it is always possible to take a small quantity

ε>0 of good x from consumer 2 to redistribute it to consumer 1 in exchange for the same

quantity ε of good y. This exchange leaves consumer 2 indierent (∆u2 = −ε + ε = 0) but

strictly increases the utility of consumer 1 (∆u1 = 2ε − ε = ε > 0). Such an exchange is

only possible when consumer 2 has a strictly positive quantity of good x and consumer 1
a strictly positive quantity of good y. For any given allocation where consumer 1 has no

quantity of good y (y1 = 0), consumer 1 will not be willing to trade any quantity of good

x in exchange for good y (because good x always yields higher utility for this consumer),

any allocation with y1 = 0 is thus Pareto-optimal. Similarly, for any allocation where

consumer 2 has no quantity of good x (x2 = 0), no (utility improving) trade is possible

because consumer 1 will not be willing to trade any quantity of good x in exchange for

good y. The Pareto-optimal allocations are thus such that x1 = 1 (i.e x2 = 0) or y1 = 0.

2
 Figure 3: Set of Pareto-optimal allocations - Perfect substitutes & dierent MRS

4. u1 (x1 , y1 ) = x21 + y12 and u2 (x2 , y2 ) = x22 + y22

Answer: The consumers' preferences are strictly concave. We rst show that interior

allocations cannot be Pareto-optimal. Suppose that the allocation is such that x 1 ≥ y1 (if

x1 < y1 the roles played by good x and y below can be reversed) and consider the eect of a

reallocation which takes a quantity ε > 0 of good x from consumer 2 to give it to consumer 1
in exchange for a quantity ε of good y . Such an exchange increases the utility of consumer 1
(extremes are preferred to averages with concave preferences): ∆u1 = 2ε2 + 2(x 1 − y1 )ε >0
and also increases the utility of consumer 2: ∆u2 = 2ε2 +2(y 2 −x2 )ε. Indeed, since x1 ≥ y1
implies x2 ≤ y2 (x1 +x2 = 1 and y1 +y2 = 1), we also have ∆u2 > 0. Therefore, any interior
1
allocation cannot be Pareto-optimal . Conversely, from the strict concavity of preferences,

any corner allocation (i.e. such that at least one of the quantities of good x or y is zero) is

Pareto-optimal (see Figure 4).

Figure 4: Set of Pareto-optimal allocations - Strictly concave preferences

1
The intuition here is that both consumers have concave preferences and thus enjoy greater marginal utility
from an increase in quantity in the good that they possess in largest quantity. For any interior allocation and
given the symmetry of a given allocation (x1 ≥ y1 ⇒ y2 ≥ x2 ), both consumers can make utility improving
exchanges by increasing the quantity of the good they have in largest quantity and reducing the quantity of good
they have in smallest quantity.

3
 1

√
5. u1 (x1 , y1 ) = min 2 x1 , y1 and u2 (x2 , y2 ) = x2 y2
Answer: From the utility function of consumer 1 which represents perfect complements

preferences, we deduct that any interior allocation such that x1 6= 2y1 is not Pareto-

optimal. Indeed, we could either reduce x1 (if x1 > 2y1 ) or y1 (if 2y1 > x1 ) without

reducing consumer 1's utility and transfer all the quantity of good to consumer 2 which

would strictly increase his utility. However, this is not possible for allocations such that

x1 = 2y1 . In this case, in order not to decrease the utility of consumer 1, we would need

to transfer a positive quantity of both goods which would reduce the utility of consumer 2.
Such allocations are thus Pareto-optimal. We reason graphically for corner solutions (see

Figure 5).

Figure 5: Set of Pareto-optimal allocations - Perfect complements & Cobb-Douglas

1
u2 (x2 , y2 ) = min x2 , 21 y2



6. u1 (x1 , y1 ) = min 2 x1 , y1 and

Answer: Reason graphically (see Figure 6).

Figure 6: Set of Pareto-optimal allocations - Perfect complements

4
Exercise 2 Equilibrium in an exchange economy and decentralization

Consider an exchange economy consisting of two consumers, A and B, and two goods, x and

y. Consumers preferences can be represented by the utility functions uA (xA , yA ) = xA yA (for

√
A) and uB (xB , yB ) = x B yB (for B ). Consumer A's initial endowment is one unit of good x
(i.e., ωA = (1, 0)), while consumer B initial owns 3 units of good x and 4 units of good y (i.e.,

ωB = (3, 4)).

1. Draw the Edgeworth box, the initial endowment point as well as the initial indierence

curves for both consumers.

Answer:

Figure 7: Edgeworth box,initial endowment point and initial indierence curves

2. Characterize the Pareto-optimal allocations and draw the set of such allocations (also

known as the contract curve ) in the initial diagram. In the (uA , uB ) plane, draw the Pareto-
frontier (i.e., the set possible combinations of utility levels for Pareto-optimal allocations)?

Is the initial allocation Pareto-optimal?

Answer: Both consumers' preferences are represented by Cobb-Douglas utility functions.

We can thus use the results from Exercise 1 with α = β. Pareto-optimal allocations are

characterised by yA = xA (and yB = xB ). The initial allocation is not Pareto-optimal

because we can reallocate the unit of good x initially held by consumer A and transfer it to
consumer B (uB ↑) without reducing consumer A's utility (it remains at uA = 0). For any

Pareto-optimal allocation, let x be the quantity of good x allocated to consumer A. We

have: xA = yA = x. Moreover, given the initial endowments: xB = 4 − xA = 4 − x with

x ∈ [0, 4]. We also have: yB = xB = 4 − x. Therefore, for any Pareto-optimal allocation x:

√
uA (x) = x2 and uB (x) = 4 − x which can be rewritten: uB = 4 − uA with uA ∈ [0, 16].

5
 Figure 8: Pareto-frontier in the (uA , uB ) plane

3. Derive the equilibrium of this exchange economy and check that the equilibrium allocation

is Pareto-optimal.

Answer: To characterize the Walrasian equilibrium we rst maximize the utility of each

consumer to derive their demand functions (keeping in mind that their revenue is equal to

their initial endowment evaluated at market prices: Ri = p · ω i ) and then solve the system
P i
of market equilibrium conditions (i.e. for any good l, solve i xl (p) = ωl ). Here, the
1
consumers' utility functions are Cobb-Douglas with α= 2 . We can thus solve the system
given by the equality of the MRS with the price ratio and the binding budget constraint

to nd the consumers' demand functions

2 which are given by:

Ri Ri
xi (px , py , Ri ) = and y i (px , py , Ri ) =
2px 2py

Replacing Ri by the value of the initial endowments of each consumer (i.e. R A = px and

RB = 3px + 4py ) yields the following optimal choices:


xA (px , py ) = 1 px
2 and yA (px , py ) = 2py
x (p , p ) = 3px +4py 3px +4py
B x y 2px and yB (px , py ) = 2py

Solving xA (px , p, y) + xB (px , py ) = wx = 4 yields px = py . The aggregate excess demand

for good x is thus equal to zero if and only if px = py . Using Walras law, we do not need to

check that the market equilibrium condition is satised for good y (since it is necessarily the
∗
case). The equilibrium prices are thus such that px = p∗y and at the Walrasian equilibrium,
∗
the allocations are given by: xA = yA =
∗ 1
x∗B = yB ∗ = 7 . Since x∗ = y ∗ , the
2 and 2 A A
equilibrium allocation is Pareto-optimal.

2
We refer to Problem Set 1 for detailed derivations of the demand functions.

6
4. How would it be possible to reallocate the initial endowment in a way that the equilibrium

allocation (of this modied economy) would satisfy x∗A = x∗B ?

Answer: The equilibrium allocation that would satisfy x∗A = x∗B is symmetric. Indeed,

from the Cobb-Douglas preferences and the Pareto-optimality (the equilibrium allocation

is Pareto-optimal), we have: yA = xA and yB = xB . Moreover, from the total endowment

in good x and good y equal to 4, we have: xA = xB = yA = yB = 2. The marginal rates of

substitution are equal to 1 for both consumers. Therefore, at the equilibrium the prices of

both goods are equal: p∗x = p∗y (using the equality between the MRS and the price ratio).

Then, using the Second Welfare Theorem, we can deduce that any initial allocation such

that ωxA + ωyA = 4, the Walrasian equilibrium will be such that p∗x = p∗y and the equilibrium
∗
allocation will be xA = ∗
yA =2 ∗
(and thus xB = ∗
yB = 2).

Figure 9: Edgeworth box - Decentralization

5. Assuming that it is only possible to transfer (initial transfer prior to exchange) quantities

of good x, which allocations can be decentralized, that is, can be equilibrium allocations

(potentially after reallocation of the initial endowment in good x)?

Answer: For any Pareto-optimal allocation (such that xA = yA ), the marginal rates of

substitution are equal to 1 for both consumers. For such an allocation to be an equilibrium

allocation, the equality between the MRS and the price ratio must be satised which

is the case when the prices of both goods are equal (i.e. p∗x = p∗y ). Using the Second

Welfare Theorem, we deduce that it is possible to decentralize the Pareto-optimal allocation

x A = yA = x (and thus xB = yB = 4 − x) by reallocating the goods such that the initial

endowment is on the line with slope −1 passing through the point (xA , yA ) = (x, x) i.e. the

initial endowment is such that: xA + yA = 2x. When only the good x can be reallocated

(initial transfer prior to exchange), it will only be possible to decentralize the Pareto-

allocations (x, x, 4 − x, 4 − x) with x ≤ 2.

7
Exercise 3 Exchange and intertemporal choices

Consider an economy consisting of one good (say wealth or consumption), two types of agents

(young and old) and two periods (t = 1, 2). There are NY young people and NO old people in

this economy, and all have the same preferences that can be represented by the (intertemporal)

utility function u (c1 , c2 ) = cα1 c1−α

2 (where α ∈ ]0, 1[). Young people have no income in period 1

and a revenue equal to R in period 2, whereas old people have a revenue equal to R in period 1

but no income in period 2. Consumption in period 1 is used as the numeraire, that is, p1 = 1
1
and the interest rate is r, i.e., p2 = 1+r . Consumers can either borrow or save in period 1 at a
common interest rate r but they will only be able to do so eectively through inter-generational

exchanges.

1. Write down the budget constraints for young and old people.

Answer: Here, we can abstract away from the intertemporal problem and solve it as a

pure exchange economy problem (we have simply normalized the price of consumption

in period 1 to p1 = 1 and will nd the equilibrium interest rate the same way we nd an

equilibrium price ratio). For both the young and the old people, the budget constraint is

given by: p1 c1 + p2 c2 ≤ p1 ω1 + p2 ω2 . Plugging in prices and initial endowments for both

groups yields the intertemporal budget constraints (evaluated at t = 1):

ˆ cY1 + 1 Y
1+r c2 ≤ 1
1+r R for young people, and
ˆ cO
1 +
1 O
1+r c2 ≤R for old people.

2. Derive the demand and excess demand functions in period 1 and in period 2 for young and

old agents.

Answer: Preferences are represented by a Cobb-Douglas utility function. The demand

functions are thus given by (see Exercise 1, Problem Set 1):

 
cY (r, R) = αR cO (r, R) = αR
1 1+r 1
and
cY (r, R) = (1−α)R cO (r, R) = (1−α)R
2 (1+r)2 2 (1+r)

The excess demand functions are the dierence between the quantity the consumer/group

desires to consume (i.e. their demand) and their initial endowment. Therefore, we have:
 
z Y (r) = cY (r) − ω Y = αR z O (r) = cO (r) − ω O = (α − 1)R
1 1 1 1+r 1 1 1
and
z Y (r) = cY (r) − ω Y = (1−α)R z O (r) = cO (r) − ω O = (1−α)R
2 2 2 (1+r)2
−R 2 2 2 (1+r)

3. Write down the equilibrium condition in each period and derive the equilibrium interest

rate.

Answer: A Walrasian equilibrium is characterized by (i) individual allocations that are

equal to consumers' optimal choices at equilibrium prices (i.e. consumers maximize their

utility) and (ii) prices such that aggregate excess demands are all equal to zero (i.e. the

markets for each good clears).


z1 (r) = NY z Y (r) + N0 z O (r) = 0
1 1
z (r) = N z Y (r) + N z O (r) = 0
2 Y 2 0 2

8
 αR
The rst market clearing condition is equivalent to:
1+r +NO (α−1)R =NY 0 ⇐⇒ 1+r =
α NY ∗ α NY
1−α NO . The equilibrium interest rate is thus given by: r = 1−α NO − 1.

4. Under which condition is the equilibrium interest rate positive? Explain the intuition

behind this result.

Answer: r∗ > 0 ⇐⇒ αNY > (1 − α)NO . The interest rate will be positive if, adjusting

for preferences, there are more young people than old people. In other words, if the initial

endowment in good 1 is relatively scarce (ω1 = NO R), the interest rate will be positive

(and consumption in period 1 will be more expensive than consumption in period 2).

Exercise 4 Exchange economy: Pareto-optima and equilibrium

Consider an exchange economy consisting of two consumers, A and B, and two goods, X et Y.
1 2
Consumers preferences can be represented by the utility functions uA (xA , yA ) = xA − 4 yA (for A
√
who thus dislikes good y ) and uB (xB , yB ) = xB + yB (for B ). Consumer A's initial endowment
is eight units of good Y (i.e., ωA = (0, 8)), while consumer B initial owns 16 units of good X
and 8 units of good Y (i.e., ωB = (16, 8)).

1. Draw the Edgeworth box, the initial endowment point as well as the initial indierence

curves for both consumers.

Answer: The initial endowment point is represented in yellow in Figure 10, the initial

indierence curve of consumer A in red and the initial indierence curve of consumer B in

blue.

Figure 10: Edgeworth box with initial allocation and initial indierence curves

2. Is the initial allocation Pareto-optimal?

Answer: The initial allocation is not Pareto-optimal (it even yields negative utility for
√
consumer A: uA (0, 8) = −16; also note that uB (16, 8) = 16 + 8). Indeed, any reallocation

of good y from consumer A to consumer B will improve (strictly) both consumers' utility

since A's utility is strictly decreasing in y and B 's utility is strictly increasing in y.

9
3. Characterize the Pareto-optimal allocations and draw the set of such allocation (also known

as the contract curve ) in the initial diagram.

Answer: For any allocation such that yA > 0, there will always be a feasible utility

improving reallocation (since consumer A dislikes good y) in which a strictly positive

quantity of good y in (0, yA ] is transferred from consumer A to consumer B (thus improving

both consumers' utility). Therefore, any allocation with yA > 0 is not Pareto-optimal.

Conversely, if yA = 0 (and thus yB = 16), there is no possible Pareto-improvement (indeed,

the only way to increase consumer A's utility would be to increase his quantity of good

x which would decrease consumer B 's utility). Therefore, any allocation with yA = 0 is

Pareto-optimal (this corresponds to the x-axis for consumer A - see Figure 10 in green).

4. Derive the equilibrium of this exchange economy and check that the equilibrium allocation

is Pareto-optimal.

Answer: First, we maximize the consumers' utilities subject to their budget constraints

to derive their demand functions. For consumer A, the maximization is straightforward

because he/she only gets utility from consuming good x and disutility from consuming

good y. It is thus optimal for consumer A to allocate all of his revenue RA = 8py to the
RA 8py
consumption of good x: xA (px , py ) = px = px and yA (px , py ) = 0.
Consumer's B preferences are strictly convex and monotonic, we express x as a function
RB −py y B
of y (or viceversa) using the binding budget constraint (e.g. x= px for y ∈ [0, Rpy ]),
plug it into the utility function and maximize the new objective function (e.g. U (y) =
RB −py y √ p
px + y ). The derivative is given by: ∂U
∂y = − pxy + 1
√
2 y which is a decreasing function
in y. The interior solution is obtained by setting this derivative to zero . We have a corner
3
B p
solution if the derivative remains positive over the [0, Rpy ] interval, i.e. if − pxy + r1
B
≥0
2 Rp
y
RB px
(which is equivalent to M RSx→y (0, py ) ≤ py ). The demand functions are thus given by:
  2 q  q
 1 px if
px
<2 RB  RB − 1 px
if
px B
< 2 Rpy
4 py py py px 4 py py
yB (p, RB ) = q and xB (p, RB ) = q
 RB if
px
≥2 RB 0 if
px B
≥ 2 Rpy
py py py py

We can then replace the value of RB by the market value of B's initial endowment: RB =
16px + 8py .
The second condition for the Walrasian equilibrium is for markets to clear, i.e. for the

market equilibrium condition for each good to be satised. Walras' law implies that we

can simply solve the equilibrium condition for one good. Given the demand functions, and

assuming the solution is interior for consumer B , we

 solve
 the market clearing condition for
2
1 px px
good y : yA (px , py )+yB (px , py ) = wy = 16 ⇐⇒ 4 py = 16 ⇐⇒ py = 8. With this last
q
px RB
condition and RB = 16px + 8py , we indeed have <2
py (the condition for the optimal
py
∗ ∗
choice of consumer B to be interior holds). Therefore, equilibrium prices satisfy px = 8py ,
∗ ∗ ∗ ∗
and the equilibrium allocations are given by: (xA , yA ) = (1, 0) and (xB , yB ) = (15, 16).
∗
Since yA = 0, the equilibrium allocation is indeed Pareto-optimal.
3
It could also have been obtained with the equality between the MRS and the price ratio

10
Exercise 5 Exchange economy: Pareto-optima and equilibrium

Consider an exchange economy consisting of two consumers, A and B, and two goods, X and

Y. Consumers preferences can be represented by the utility functions uA (xA , yA ) = 2xA + yA

(for A) and uB (xB , yB ) = ln xB + ln yB (for B ). Consumer A's initial endowment is of one unit

of good X and two units of good Y (i.e., ωA = (1, 2)), while consumer B initial owns 2 units of

good X (i.e., ωB = (2, 0)).

1. Draw the Edgeworth box, the initial endowment point as well as the initial indierence

curves for both consumers.

Answer: See Figure 11: the initial endowment point is the green dot, the indierence

curves are represented in blue for A and in red for B.

yA
Ensemble des
xB optima de Pareto

xA
1 2 yB

Figure 11: Edgeworth box with initial allocation and initial indierence curves

2. Is the initial allocation Pareto-optimal?

Answer: The initial allocation is not Pareto-optimal: moving to xA = 3 and yA = 2 (i.e.,

giving everything to A) increases A's utility without aecting B 's utility.

3. Characterize the Pareto-optimal allocations and draw the set of such allocations (also

known as the contract curve ) in the initial diagram.

Answer: We rst characterize interior allocations Pareto-optimal allocations. Interior

allocations are Pareto-optimal if and only if:

B A yB
M RSx→y = M RSx→y ⇔ =2 ⇔ yB = 2xB .
xB
We then consider non-interior allocations:

ˆ if xB = 0 and yB = 0 , then the allocation is Pareto-optimal since consumer A holds

all the goods and no Pareto-improvement can be found.

ˆ if xB = 0 and yB > 0 , then a Pareto-improvement consists of allocating all the

quantity of good y held by consumer B to consumer A (uA ↑, ∆uB = 0).

11
 ˆ if xB > 0 and yB = 0 , then a Pareto-improvement consists of allocating all the

quantity of good x held by consumer B to consumer A (uA ↑, ∆uB = 0).

ˆ if xA = 0 and y A = 0, then the allocation is Pareto-optimal since consumer B holds

all the goods and no Pareto-improvement can be found.

yB
ˆ if xA = 0 (and thus xB = 3 ) and yA > 0, then
B
M RSx→y = 3 ≤ 2
3 < 2. Therefore,

there exist a Pareto-improvement which consists of giving some x to A in exchange

for some y (e.g. such that uB ↑, ∆uA = 0).

ˆ if xA > 0 and yA = 0 (and thus yB = 3), then
B
M RSx→y = 2
xB , and thus:

 if
B
xB < 1 (xA > 2), M RSx→y > 2, a Pareto-improvement consists of giving some

y (enough at least to keep ∆uA = 0) to consumer A in exchange for some x

(uB ↑).
 if
B
xB = 1 (xA = 2), M RSx→y = 2: Pareto-optimum.

 if
B
xB > 1, M RSx→y < 2; consumer B would oer some x to A in exchange for

good y but A has no y to give. Such allocations are thus Pareto-optimal.

In summary, the set of Pareto-optimal allocations is given by (also see the green line in

Figure 11):

PO
0, for 0 ≤ xA ≤ 2
yA =
2x − 4, for 2 < xA ≤ 3
A

4. Explain (in the simplest possible way  i.e., without fully deriving the equilibrium) why in

equilibrium: (1) equilibrium prices satisfy p∗X = 2p∗Y , and (2) the equilibrium allocation is
∗
such that xA =2 ∗
and yA = 0.
Answer: Option 1: To determine the equilibrium price ratio, we could simply compute

the demand for good x of consumer B (with Cobb-Douglas preferences):

RB 2pX
xB (pX , pY ) = = =1
2pX 2pX
From the First Welfare Theorem, we know that equilibrium allocation must be Pareto-
B pX
optimal. Therefore, yB = 2. Finally, from M RSx→y = pY (optimum is interior for B ), we

get the characterization of the price ratio.

Option 2: Given the perfect substitutes preferences of consumer A:

pX +2pY
ˆ if pX > 2pY , A only wants to consume good Y : yA (pX , pY ) = pY > 4;
pX +2pY
ˆ if pX < 2pY , A only wants to consume good X : xA (pX , pY ) = pX > 2.

In both cases, this leads to excessive aggregate demand (prices would have to adjust).

Therefore, p∗X = 2p∗Y . The equilibrium allocation then corresponds to the intersection be-

tween the contract curve (set of Pareto optima) and the budget line with slope −2 (passing
∗
through the initial endowment) which corresponds to xA =2 ∗
and yA = 0.

12