Solutions to Problem Set 3: Production Economies

ECON 391, Equilibrium in Market Economies

Jean Guillaume Forand∗

1. Consider a one-consumer one-firm economy in which u(x) = min{xz , xq }, ω = 2 and f (z) =

3z.

(a) Derive a competitive equilibrium for this economy.

Solution. Letting m = 2pz +Π be the value of the consumer’s endowment and assum-
ing that pz , pq > 0, the consumer’s demand functions are x(p, m) = (m/pz +pq , m/pz +pq ).
The firm’s supply functions are

(0, 0) if pz/pq > 3,


φ(p) = any φ = (z, q) such that q = 3z if pz/pq = 3,


does not exist if pz/pq < 3,


and the firm’s corresponding profits are


0 if pz/pq ≥ 3,
Π=
does not exist if pz/pq < 3.

To find a competitive equilibrium, normalise p∗z = 1. In any equilibrium, the firm’s

optimal production plan must exist, so that from above we know that 1/p∗q ≥ 3. First,
can we have an equilibrium in which 1/p∗q > 3? In this case, the firm’s optimal output
level is q(p∗ ) = 0. However, the consumer’s demand for output is xq (p∗ ) = 2/1+p∗q > 0
(where I have used the fact that Π∗ = 0 at these prices). Therefore, the market
clearing condition for output is not satisfied. In words, at such prices the firm shuts
down but because the consumer still derives income from her endowment of the input
good, she has a positive demand for output. Second, can we have an equilibrium in
which 1/p∗q = 3? In this case, any level of output q ∗ = 3z ∗ is optimal for the firm, and
its profits are 0. The consumer’s demand for output is xq (p∗ ) = 3/2. Therefore, prices
(1, 1/3), allocations x∗ = (3/2, 3/2) and φ∗ = (1/2, 3/2), along with profits Π∗ = 0 form a
competitive equilibrium. This is illustrated in Figure 1.
∗
Room 131, Department of Economics, University of Waterloo, Hagey Hall of Humanities, Waterloo, Ontario,
Canada N2L 3G1. Email: [email protected].

1
Figure 1: Competitive equilibrium for Question 1

(b) Derive the Pareto set of this economy.

Solution. Pareto-efficient allocations are solutions to the problem
max min{xz , xq } subject to 3[2 − xz ] = xq .
0≤xz ≤2,xq ≥0

This utility-maximisation problem is illustrated in Figure 2. We know that the optimal

consumption bundle will be such that x∗z = x∗q , so that along with x∗q = 3[2 − x∗z ], we
have that x∗z = x∗q = 3/2. Therefore, allocations x∗ = (3/2, 3/2) and φ∗ = (1/2, 3/2) are
the unique Pareto-efficient allocations for this economy.
√ √
2. Consider a one-consumer one-firm economy in which u(x) = xz + xq and f (z) = z.

(a) Derive the Pareto set of this economy.

Solution. Pareto-efficient allocations are solutions to the problem
√ √
max xz + xq subject to ω − xz = xq ,
0≤xz ≤ω,xq ≥0

which is illustrated in Figure 3. Any solution to this problem in which x∗z , x∗q > 0 must
solve the first-order condition
∂ ∗
∂xz u(x ) 1
q
0 ∗
= f (ω − x z ), or 2 x∗q = √ .
∂ ∗
∂x u(x )
2 ω − x∗z
q

2
Figure 2: Pareto set for Question 1

√
By substituting x∗q = ω − x∗z , we have that x∗q = (1/2)4/3 and x∗z = ω − (1/2)8/3 .
Any solution to the Pareto problem with x∗z = 0 would have to be such that
∂ ∗
∂xz u(x )
∂
≤ f 0 (ω − x∗z ) ,
∗ x∗z =0
∂xq u(x ) x∗z =0
√
which, using that x∗q = w in this case, reduces to ω ≤ (1/2)8/3 . In words, if the
consumer’s endowment of input is small, then the Pareto-efficient allocations have
the firm use up all of the input to produce output, leaving the consumer with no
consumption of input. This is illustrated in Figure 4.
Any solution to the Pareto problem with x∗q = 0 would have to be such that
∂ ∗
∂xz u(x )
∂
≥ f 0 (ω − x∗z ) ,
∗ x∗q =0
∂xq u(x ) x∗q =0
√
which, using that x∗z = ω in this case, reduces to 0 ≥ 1/2 ω−x∗z |x∗z =0 ≈ ∞. This
condition never holds, so that x∗q = 0 is never optimal. In words, if the firm is
producing no output, then its marginal productivity is arbitrarily high, whereas if
the consumer is not consuming any output, then her marginal utility from output
is also arbitrarily high. Therefore, it would benefit the consumer and be feasible to
produce some output in this case, so that no-production is Pareto-dominated. This is
illustrated in Figure 5.

3
Figure 3: Pareto problem for Question 2

Figure 4: Corner solution with x∗z = 0

4
Figure 5: No corner solution with x∗q = 0

Therefore, the Pareto set of this economy contains only allocations x∗ = (ω−(1/2)8/3 , (1/2)4/3 )
√
and φ∗ = ((1/2)8/3 , (1/2)4/3 ) if ω > (1/2)8/3 , and it contains only allocations x∗ = (0, ω)
√
and φ∗ = (ω, ω) if ω ≤ (1/2)8/3 .

(b) Fix some Pareto-efficient allocations from (a) and derive a competitive equilibrium for
this economy that generates these allocations.
Solution. Suppose that ω > (1/2)8/3 , so that the Pareto-efficient allocations from
(a) have x∗z , x∗q > 0. The slope of the consumer’s indifference curve through the
Pareto-efficient allocations is 2 x∗q = 21/3 . Therefore, the prices for the equilibrium
p

that we are constructing must be p∗ = (1, (1/2)1/3 ). Given prices p∗ , we know that
the firm’s optimal production plan is φ∗ = (1/4(p∗q/p∗z )2 , 1/2(p∗q/p∗z )) = ((1/2)8/3 , (1/2)4/3 ),
with corresponding profits Π∗ = p∗q 2/4p∗z = (1/2)8/3 . Therefore, prices p∗ , allocations
x∗ = (ω − (1/2)8/3 , (1/2)4/3 ) and φ∗ = ((1/2)8/3 , (1/2)4/3 ), along with profits Π∗ = (1/2)8/3
form a competitive equilibrium. This is illustrated in Figure 6.

3. Consider a one-consumer one-firm economy in which u(x) = αxz +xq , ω = 1 and f (z) = z 2 .

(a) Derive a competitive equilibrium for this economy.

Solution. No competitive equilibrium exists for this economy. To see this, consider
any equilibrium prices p∗ . Given any level of input z > 0, the firm’s profit from
production plan φ = (z, z 2 ) is z[p∗q z − p∗z ]. This grows without bound as z → ∞, to
that no optimal production plan exists for the firm (referring back to ECON 290, notice
that the firm has increasing returns to scale). Intuitively, given any input and output
prices, the firm wants to “grow” as much as possible, so that no scale of production is
optimal because any larger scale would increase profits.

5
Figure 6: Competitive equilibrium generating Pareto-efficient allocations

(b) Derive the Pareto set of this economy.

Solution. Pareto-efficient allocations are solutions to the problem

max αxz + xq subject to (1 − xz )2 = xq ,

0≤xz ≤1,xq ≥0

which is illustrated in Figure 7. From the Figure, it follows that if α > 1, then
the unique Pareto-efficient allocations are x∗ = (1, 0) and φ∗ = (0, 0) (this case is
illustrated in Figure 7), if α < 1 then the unique Pareto-efficient allocations are x∗ =
(0, 1) and φ∗ = (1, 1), whereas if α = 1 then the Pareto set contains both these
allocations (and no others).
You may be wondering why this economy does not have competitive equilibrium allo-
cations and yet it has Pareto-efficient allocations. The key is that the Pareto problem
is a like a central planning problem in which feasibility is imposed on the problem,
whereas competitive equilibria are built up from the solutions to decentralised decision-
making by individual agents. It can be, as in this economy, that the various agent’s
problems are not compatible with each other and with aggregate constraints in the
economy. But how do we reconcile the results of this example with the Welfare The-
orems? The First Welfare Theorem says that any competitive equilibrium allocations
must be Pareto efficient; it has nothing to say if no competitive equilibrium allocations
exist. The Second Welfare Theorem says that any Pareto-efficient allocations can be
generated as competitive equilibrium allocations, and here we have Pareto-efficient

6
Figure 7: Pareto-efficient allocations for Question 3

allocations. However, this economy violates the convexity requirements of that result;
in particular, the firm’s production set is not convex.

4. Consider a two-consumer one-firm economy in which, as in class, there is a single input

good and a single output good.

(a) Suppose that consumer A owns a share 0 ≤ σ ≤ 1 of the firm (so that consumer B
owns share 1 − σ). Define a competitive equilibrium for this economy.
Solution. A competitive equilibrium for this economy consists of prices p∗ = (p∗z , p∗q ),
allocations (xA∗ , xB∗ , φ∗ ) and profits Π∗ satisfying
i. Given prices p∗ and profits Π∗ , the demands of consumer A, (xA∗ A∗
z , xq ) solve

max uA (xA ) subject to p∗z xA ∗ A ∗ A ∗

z + pq xq ≤ pz ω + σΠ .
xA A
z ,zq ≥0

ii. Given prices p∗ and profits Π∗ , the demands of consumer B, (xB∗ B∗

z , xq ) solve

max uB (xB ) subject to p∗z xB ∗ B ∗ B ∗

z + pq xq ≤ pz ω + (1 − σ)Π .
xB B
z ,zq ≥0

iii. Given prices p∗ , production plan φ∗ = (z ∗ , q ∗ ) solves

max p∗q q − p∗z z subject to f (z) ≥ q,

z,q≥0

and furthermore Π∗ = p∗q q ∗ − p∗z z ∗ .

7
 iv. The market for input clears: xA∗ B∗ ∗ A B
z + xz + z = ω + ω .
v. The market for output clears: xA∗ B∗ ∗
q + xq = q .

(b) Define Pareto-efficient allocations for this economy.

Solution. Feasible allocations (xA , xB , φ) Pareto-dominate feasible allocations (x0 A , x0 B , φ0 )
if uA (xA ) ≥ uA (x0 A ) and uB (xB ) ≥ uB (x0 B ), with at least one inequality strict. Fea-
sible allocations (xA , xB , φ) are Pareto-efficient if they are not Pareto-dominated by
any other feasible allocations.
P.S. You should convince yourself that you know how to define feasible allocations for
this economy.

5. Consider a two-consumer one-firm economy with a single input good and a single output
good. Consumers’ utility functions are uA (xA ) = xA A B B B
z + 3xq and u (x ) = 4xz + xq .
B
A B
Suppose that the input good is divisible and that ω = ω = 1/2. Suppose that the output
good is indivisible (i.e., that xq , q = 0, 1, 2, ...), and that the firm’s production technology
is such that it requires 1 unit of input to produce one unit of output.

(a) Derive the Pareto set of this economy.

Solution. In any Pareto-efficient allocations in this economy, either z = q = 0 and
xA B A B
z + xz = 1 or z = q = 1 and xz = xz = 0. In words, either no output is produce
and the consumers share the available input, or the firm uses all the available input
to produce one indivisible unit of output.
Are there Pareto-efficient allocations in which output is produced and given to con-
sumer A, i.e., such that z = q = xA A B B
q = 1 and xz = xq = xz = 0? In this case, we
have that uA (xA ) = 3 and uB (xB ) = 0. Can these allocations be Pareto-dominated
by allocations in which no output is produced but A obtains all of the input good, i.e.,
such that z 0 = q 0 = x0 A 0B 0B 0A A 0A
q = x q = x z = 0 and x z = 1? No, because u (x ) = 1 < 3.
Therefore, allocations (φ, xA , xB ) are in the Pareto set.
Are there Pareto-efficient allocations in which output is produced and given to con-
sumer B, i.e., such that z = q = xB A A B
q = 1 and xz = xq = xz = 0? In this case, we
have that uA (xA ) = 0 and uB (xB ) = 1. Can these allocations be Pareto-dominated
by allocations in which no output is produced but B obtains all of the input good,
i.e., such that z 0 = q 0 = x0 A 0B 0A 0B A 0A
q = x q = x z = 0 and x z = 1? Yes, because u (x ) = 0
and uB (x0 B ) = 4. Therefore, allocations (φ, xA , xB ) are not in the Pareto set.
Are there Pareto-efficient allocations in which no output is produced? First, we know
that the allocations with xAz = 1 are Pareto-dominated by the allocations in which the
output good is produced and given to consumer A. Second, from the argument from
the preceding paragraph we can also see that the allocations with xBz = 1 are Pareto-
efficient. Third, consider any allocations with z = q = xq = xq = 0 and xA
A B B
z + xz = 1

8
 with xA B
z , xz > 0. No such allocations are Pareto-dominated by allocations in which
q 0 = 1, because in that case only one consumer could obtain the output good, and
hence uJ (x0 J ) < uJ (xJ ) for some J = A, B. And no such allocations are Pareto-
dominated by allocations in which q 0 = 0 because no consumer can obtain more input
than under x without the other consumer receiving less. Therefore, all such allocations
are in the Pareto set.
(b) For all Pareto-efficient allocations from (a), derive a competitive equilibrium for this
economy that generates these allocations.
Solution. Consider the allocations with z = q = xA A B B
q = 1 and xz = xq = xz = 0. If
∗ ∗
pz = pq = 1, then the production plan (z, q) = (1, 1) is optimal for the firm (along with
all other feasible production plans) and it yields profits Π∗ = 0. If we redistribute the
input endowments so that ω 0 A = 1 and ω 0 B = 0, then demand xA = (1, 0) is optimal
for consumer A and demand xB = (0, 0) is optimal for consumer B, so that we have
a competitive equilibrium generating these Pareto-efficient allocations, as desired.
Now consider any allocations with z = q = xA B A B
q = xq = 0 and xz + xz = 1 with
xA ∗ ∗
z < 1. If pz = pq = 1, then the production plan (z, q) = (0, 0) is optimal for the firm
and yields profits Π∗ = 0. We must redistribute the endowments so that ω 0 A = xA z and
0 B B
ω = xz . In that case, even if consumer A would rather buy the output good than
the input good at prices p∗ , she cannot afford to buy any (integer) units of the output
good, so that demand xA = (xA z , 0) is optimal for her. Consumer B would rather
buy the input good than the output good at prices p∗ , so that demand xB = (xB z , 0)
is optimal for her. Therefore, we have a competitive equilibrium generating these
Pareto-efficient allocations, as desired.

6. Consider a two consumer one-firm economy with a single input good and a single output
good. Consumers’ utility functions are uA (xA ) = xA A B B B B A
z xq and u (x ) = xz xq , with ω = 2
and ω B = 1. The firm’s production function is f (z) = z /4 . Derive the Pareto set of this
1

economy.
Solution. Before tackling this problem, let me connect how to think about deriving the
Pareto set for a two-consumer production economy with what we know about two-consumer
exchange economies. To do this, I will write down the analogue of the “Pareto problem”
that we studied there. Fix a utility level u for consumer B. Then any allocations with
∗
φ∗ = (z ∗ , f (z ∗ )), xA∗ and xB = (ω A + ω B − z ∗ − xA∗ ∗ A∗
z , f (z ) − xq ) which are solutions to
the following problem must be Pareto-efficient:
max uA (xA ) s.t. uB (ω A + ω B − z − xA A
z , f (z) − xq ) = u.
0≤z≤ω A +ω B ,xA ∈R2+

That is, this problem maximises the utility of consumer A among all those allocations which
keep the utility of consumer B to u. For the output good, this problem looks just like the

9
related problem for exchange economies: what B gets is whatever A does not get. For the
input good, what B gets is whatever neither A nor the firm get. Pareto-efficient allocations
in production economies must also specify the scale of the firm’s production: this is why
the input level z is a choice variable in the Pareto problem above.
If you wrote down the Lagrangean corresponding to the problem above (as I have done in
class for related problems), then you would get a system of four first-order conditions (for
the three choice variables and the Lagrange multiplier). From these conditions, you can
recover the following tangency conditions (you should be able to do this on your own):
∂
∂xA
uA (xA∗ ) ∂
∂xB
uB (ω A + ω B − z ∗ − xA∗ ∗ A∗
z , f (z ) − xq )
z z
∂
= ∂
, and
∂xA
uA (xA∗ ) ∂xB
uB (ω A + ω B − z ∗ − xA∗ ∗ A∗
z , f (z ) − xq )
q q
∂
∂xA
uA (xA∗ )
∂
z
= f 0 (z ∗ ).
∂xA
uA (xA∗ )
q

The first condition is identical to the corresponding condition for exchange economies: at
Pareto-efficient allocations, there cannot be gains from trade between the consumers while
keeping the amount of goods produced fixed, which means that the consumer’s indifference
curves must be tangent. The second condition we know from one-consumer production
economies: at Pareto-efficient allocations there cannot be gains from increasing or decreas-
ing the scale of production, so that the consumers indifference curves must also be tangent
to the firm’s production function. The above give us two conditions in three unknowns,
but along with the utility constraint uB (ω A + ω B − z ∗ − xA∗ ∗ A∗
z , f (z ) − xq ) = u, we have a
system of three equations in three unknowns.
Now returning to our example economy, the two conditions from above reduce to
z ∗ /4 − xA∗
1
xA∗
q q
= ∗ − xA∗
, and
xA∗
z 3 − z z
xA∗
q 1
A∗
= ∗ 3/4 .
xz 4z
To derive the Pareto set for this economy, we would want to solve for all the combinations
of xzA∗ , xA∗
q and z ∗ that satisfy these two equations, and then obtain xB∗ z and xB∗ q by
∗ ∗
feasibility. In this economy, it turns out that there is a single production plan (z , q ) that
is included in every Pareto-efficient allocation. To see this, use the second equation above
to substitute xA∗ ∗ 3/4 into the first equation. This yields z ∗ = 3/5, so that q ∗ = 3/51/4 .
z = 4z
Therefore, finding the Pareto set of this economy reduces to finding the Pareto set of an
echange economy with consumers with Cobb-Douglas utility, an amount 12/5 of good 1 (this
is the amount of input good to distribute among the two consumers after the firm has used
3/5 of it in production) and an amount 3/51/4 of good 2 (the amount of output produced by

the firm in the unique Pareto-efficient production plan).

10
7. Consider a one-consumer two-firm economy in which, as in class, there is a single input
good and a single output good. Label the two firms a and b.

(a) Define a competitive equilibrium for this economy.

Solution. A competitive equilibrium for this economy consists of prices p∗ = (p∗z , p∗q ),
allocations (x∗ , φa∗ , φb∗ ) and profits (Πa∗ , Πb∗) satisfying
i. Given prices p∗ and profits (Πa∗ , Πb∗) , the consumer’s demands (x∗z , x∗q ) solve

max u(x) subject to p∗z xz + p∗q xq ≤ p∗z ω + Πa∗ + Πb∗ .

xz ,zq ≥0

ii. Given prices p∗ , the production plan φj ∗ = (z j∗ , q j∗ ) of firm j = a, b solves

max p∗q q j − p∗z z j subject to f (z j ) ≥ q j ,

z j ,q j ≥0

and furthermore Πj∗ = p∗q q j∗ − p∗z z j∗ .

iii. The market for input clears: x∗z + z a∗ + z b∗ = ω.
iv. The market for output clears: x∗q = q a∗ + q b∗ .

(b) Define Pareto-efficient allocations for this economy.

Solution. Feasible allocations (x, φa , φb ) Pareto-dominate feasible allocations (x0 , φ0 a , φ0 b )
if u(x) > u(x0 ). Feasible allocations (x, φa , φb ) are Pareto-efficient if they are not
Pareto-dominated by any other feasible allocations (again, make sure that you know
how to define feasible allocations for this economy).

8. Consider a one-consumer two-firm economy with a single input good and a single output
good. Suppose that the consumer’s utility function is u(x) = xz xq , and that the production
√
functions of firms a and b are f a (z a ) = z a and f b (z b ) = 2z b

(a) Derive the Pareto set of this economy.

Solution. To find Pareto-efficient allocations for this economy, we adapt the Pareto
problem from class. In particular, we must find solutions to the following utility-
maximisation problem:

max u(xz , xq ) subject to f a (z a ) + f b (ω − xz − z a ) = xq .

0≤xz ≤ω,xq ≥0,0≤z a ≤ω−xz

Notice that the amount of output good that the consumer obtains is the sum of the
outputs of the two firms. The difference with the one-firm one-consumer problem
from class is that with two firms Pareto-efficient allocations must determine how the
economy’s input is allocated to the two production technologies. This means that the
problem from above has an additional choice variable z a , which is the amount of input

11
that firm a receives (with firm b then getting whatever input is left over after the
consumer has also been served).
We could write down the Lagrangean corresponding to the problem above and take its
first-order conditions as in class. In Pareto-efficient allocations in which z a∗ , z b∗ , x∗z > 0
(i.e, in which both firms are operating and the consumer retains a positive amount of
the input good), we would obtain the tangency condition
∂ ∗ ∗
∂xz u(xz , xq ) 0
∂
= f a0 (z a∗ ) = f b (z b∗ ),
∗ ∗
∂xq u(xz , xq )

that is, the consumer’s indifference curve is tangent to the production functions of both
firms. In particular, all technologies that are operating must be equally productive at
the margin (otherwise, the consumer could be made better off by redistributing some
input from the less productive to the more productive firm). As we will see below,
it can be the case that some firm does not receive any input in some Pareto-efficient
allocations. In that case, it will have to be the case that the firm that is shut down is
less productive (again, at the margin).
To return to the example, consider Pareto-efficient allocations in which both firms are
using output. Using the condition above, we must have that
x∗q 1
= √ = 2.
x∗z 2 z a∗
The second of these equations says that z a∗ = 1/16. From the first of these equations,
we also know that x∗q = 2x∗z . We can combine this with the production constraint,
which is x∗q = 1/16 + 2[ω − x∗z − 1/16], to obtain that x∗z = ω/2 + 1/32 and x∗q = ω + 1/16.
p

There is an important constraint in the problem above that we need to worry about,
which requires that z a∗ < ω − x∗z . Remember that this ensures that firm b has some
input left over to produce. This requires that ω > 3/16. Therefore, if ω > 3/16, then only
allocations with x∗ = (ω/2 + 1/32, ω+ 1/16), φa∗ = (1/16, 1/4) and φb∗ = (w/2 − 3/32, w− 3/16)
are Pareto-efficient.
Suppose that ω ≤ 1/16, then what does the above tell us? That in Pareto-efficient
allocations it cannot be that both firms are producing. Notice that the marginal
productivity of firm b is always 2, whereas the marginal productivity of firm a is 2√1z a ,
which is declining in the input use of firm a but is arbitrarily high (and hence higher
than 2) when z a is small. Therefore, if the consumer’s input endowment is low, only
firm a will produce. In that case, we can solve for Pareto-efficient allocations by solving
the problem from class: notably, we have the tangency condition
∂ ∗ ∗
∂xz u(xz , xq )
∂
= f a0 (z a∗ ),
∗ ∗
∂xq u(xz , xq )

12
 1x∗
which in this problem reduces to x∗q = 2√ω−x ∗ . Along with the production constraint
z
√ z
∗
xq = ω − xz , we obtain that xz = /3 and x∗q = w/3. Therefore, if ω ≤ 3/16,
∗
p
∗ 2w

then only allocations with x∗ = (2w/3, w/3), φa∗ = (w/3, ω/3) and φb∗ = (0, 0) are
p p

Pareto-efficient.

(b) Fix some Pareto-efficient allocations from (a) and derive a competitive equilibrium for
this economy that generates these allocations.
Solution. Suppose that ω > 1/16, so that both firms produce in the unique Pareto-
efficient allocations. If p∗z = 1 and p∗q = 1/2, then the production plan φb∗ = (w/2 −
3/32, w− 3/16) is optimal for firm b (along with all other feasible production plans) and it
    
2
b∗ a 1 pq 1 pq
yields profits Π = 0. The supply functions of firm a are φ (p) = 4 pz , 2 pz ,
so that given prices p∗ the production plan φa∗ = (1/16, 1/4) is optimal for the firm,
and it generates profits Πa∗ = 1/16. We know that, given prices p, firm profits Πa
and Πb and endowment value m = pz w + Πa + Πb , the consumer’s demand functions
are x(p, m) = (m/2pz , m/2pq ). Therefore, given prices p∗ and profits Πa∗ and Πb∗ ,
demands x∗ = (ω/2 + 1/32, ω + 1/32) are optimal for the consumer. Therefore, we have
a competitive equilibrium generating these Pareto-efficient allocations, as desired.

9. Consider a production economy populated by a mass 1 of women and a mass 1 of men.

A fraction 0 < α < 1 of both women and men are high-skill, and a fraction 1 − α are
low-skill. There is a single labour market with a single firm. If the firm hires mass z of
high-skill workers and mass z of low-skill workers (irrespective of gender), then its output
is q = θz + θz, where θ > θ. An agent who works at wage w receives utility w. An agent
who does not work stays at home, say to take care of children or to engage in other forms
of domestic production. Suppose that women derive a higher utility from staying at home
than men (at this point if you’re getting angry with me let me ask you to suspend disbelief
for a moment, as the takeaway of this example, and with it the reason I am making this
assumption, will be come in (c) below). Specifically, women obtain utility u from staying
at home and men obtain utility u < u. Finally, assume that θ > u > θ > u.

(a) Define a competitive equilibrium for this economy.

Solution. A competitive equilibrium for this economy consists of skill-contingent wages
w∗ = (w, w) as well as gender-contingent allocations of workers into the firm z m∗ =
(z m∗ , z m∗ ) and z f ∗ = (z f ∗ , z f ∗ ), which satisfy
i. Given wages w∗ , allocations (z m∗ , z f ∗ ) solve the firm’s profit-maximisation prob-
lem

max [θ − w∗ ][z m + z f ] + [θ − w∗ ][z m + z f ].

z m ,z f ,z m ,z f ≥0

13
 ii. Given wages w∗ , allocations z m∗ and z f ∗ solve the utility-maximisation problem
of men and women of both high and low skill, and all labour markets clear. That
is,

α

 if w∗ > u
z m∗ = any 0 ≤ z m ≤ α if w∗ = u

if w∗ < u,

0



1 − α

 if w∗ > u
z m∗ = any 0 ≤ z m ≤ 1 − α if w∗ = u

if w∗ < u,

0


α

 if w∗ > u
zf ∗ = any 0 ≤ z f ≤ α if w∗ = u

if w∗ < u,

0


and

1 − α

 if w∗ > u
zf ∗ = any 0 ≤ z f ≤ 1 − α if w∗ = u

if w∗ < u

0


There are a few things to note about this definition. First, its general form closely
tracks the corresponding definition from the application to the sorting of worker skills
across sectors from class. Second, why do wages not depend on gender? Because
men and women of a given skill are equally productive, so there are markets for the
different skills in the non-home sector, but no separate markets for men and women.
Third, why do equilibrium allocations depend on gender? Because men and women
value staying at home differently, so that they will respond differently to the same
wage in the non-home sector.

(b) Derive a competitive equilibrium for this economy.

Solution. Let’s describe a competitive equilibrium in which the firm hires both high
and low-skill workers, that is, such that z m∗ + z f ∗ > 0 and z m∗ + z f ∗ > 0. From the
firm’s profit-maximisation problem, we know that this implies that the firm’s marginal
productivity from each type of workers must be equal to their equilibrium wages, or
w∗ = θ and w∗ = θ.
What is left to determine are equilibrium allocations, and for this we can use the
optimality and market clearing conditions for workers. Notice that because w∗ = θ >

14
 u > u, then by the market-clearing condition for high-skill men and women, we have
that z m∗ = z f ∗ = α. However, because u > θ = w∗ > u, then by the market-clearing
condition for low-skill men and women, we have that z m∗ = 1 − α and z f ∗ = 0.
Therefore, wages w∗ = (θ, θ) as well as allocations z m∗ = (α, 1 − α) and z f ∗ = (α, 0)
form a competitive equilibrium.

(c) Compute the gender wage gap for your equilibrium from (b), that is, the difference in
the average wages of men and women. How does that square with observed wage gaps
in real-world economies? Coming back to the model’s key assumption, do you buy the
idea that differences in men and women’s preferences for home production can explain
this wage gap?
Solution. Men of all skills are employed, and there is a fraction α of men who are high-
skill, so that mens’ average wage is αθ +(1−α)θ. Because only high-skill women work,
womens’ average wage is θ. Therefore, the man-woman wage gap is −(1 − α)[θ − θ],
which is negative. As you would know from many recent news articles and policy
discussions, the observed average wage gaps in the real world are positive, i.e., on
average, men earn higher wages than women.
Notice that attributing average wage gaps to gender discrimination is a bit of an odd
thing to do. Namely, in the model men and women with the same skills receive the
same wage. If we compute the wage gap within each skill level, it is zero. The average
wage gap arises because men and women sort into different sectors. In other words,
we cannot attribute the differences of wages between men and women to workplace
discrimination because if we condition on workers’ skills, then gender does not explain
wages. This is reflected in real-world data: when economists have access to better data
(e.g., detailed information about workers’ education levels, job titles, career paths, etc.,
which in practice is hard to obtain), they typically find that the wage differences that
can be explained by discrimination alone are small.
However, our model also points to why skill-specific wage equality may still mean
that there is gender discrimination in this economy. For example, if men and women
CEOs all earn the same salaries but only 20% of CEOs are women, is this consistent
with discrimination or not? The issue is that discrimination may influence the different
patterns of sorting across occupations between genders. In our model, sorting between
home and work sectors is determined by the utility attached to home production by
men and women. This might reflect a genuine preference difference; if this is the case
then we may respond that it is a free country and people can like what they want
(although some may then argue that preferences themselves are socially constructed
and may reflect a society’s biases). But differences in u may also capture factors like
direct coercion by spouses or social pressure more broadly. So whether or not these

15
 average wage differences reflect discrimination or not is still a very hard problem (and
getting reliable data on things like this is incredibly difficult), but nevertheless our
dinky model has given us a lot of information about where to look and why.
I have not yet fully answered the question that I asked you. Namely, in our model
women put a higher value on home production. In equilibrium, this means that the
only women who work are those women whose labour market wages compensate them
for their foregone utility from home production. Because market wages are tied to
productivity, then only high-skill women receive a wage high enough to bring them
into the labour market. Because men do not value home production, then even low-skill
men join the labour market. It follows that the average skill level of employed women
is higher than that of employed men, which explains the counterintuitive negative
men-women wage gap. Therefore, if someone ever claims that women are paid less
than men on average because they value home production more, then even our dinky
model gives you reason to be suspicious.

(d) Find a feature of real-world labour markets that the results of this model do fit.
Solution. I can think of a few things. First, in the model wages increase in workers’
skills. This is consistent with, for example, the existence of an education premium for
earnings (workers with more education earn more). This is not a feature that is specific
to this example: in all production economies with labour markets workers’ competitive
equilibrium wages will track their marginal productivities, so that if higher-skilled
workers are more productive they will always earn more. With respect to our particular
application, the model does predict that mens’ labour force participation rate will be
higher than womens’: 100% of men are employed in the model whereas α% of women
are employed. That working-age women are less likely to be in the labour force is a
well-know fact: for example, in Canada, men’s labour force participation rate is in the
73% range and women’s is in the 64% range.

16