IE54500 - Problem Set 2: 1. CES Production
IE54500 - Problem Set 2: 1. CES Production
The elasticity of substitution measures the curvature of the isoquant (i.e., the percentage change in the ratio of
inputs divided by the percentage change in the MRTS, with output held constant). For this production function,
1
𝑓𝐿 (1 − 𝛼)𝐿−𝜌−1 [𝛼𝐾 −𝜌 + (1 − 𝛼)𝐿−𝜌 ]− ⁄𝜌−1 𝛼 𝐾 1+𝜌
𝑀𝑅𝑇𝑆 = = 1 = ( )
𝑓𝐾 𝛼𝐾 −𝜌−1 [𝛼𝐾 −𝜌 + (1 − 𝛼)𝐿−𝜌 ]− ⁄𝜌−1 1−𝛼 𝐿
2. Cobb-Douglas Production
Consider the following Cobb-Douglas production function:
𝑦 = 𝐾 𝛼 𝐿𝛽
a) Show that this production function exhibits constant returns to scale when 𝛼 + 𝛽 = 1, increasing
returns to scale when 𝛼 + 𝛽 > 1, and decreasing returns to scale when 𝛼 + 𝛽 < 1.
The return to scale of the Cobb-Douglas production function depends on the values of 𝛼 and 𝛽. Suppose
all inputs were increased by a factor of 𝑡 > 1, then
𝑊ℎ𝑒𝑛 𝛼 + 𝛽 = 1, then 𝑓(𝑡𝐾, 𝑡𝐿) = 𝑡𝑓(𝐾, 𝐿), the definition of a homogeneous function of degree 1. This
is a constant returns to scale production function, where doubling all inputs doubles output.
When 𝛼 + 𝛽 < 1, then 𝑓(𝑡𝐾, 𝑡𝐿) < 𝑡𝑓(𝐾, 𝐿) because 𝑡 𝛼+𝛽 < 𝑡 when 𝑡 > 1. This implies decreasing
returns to scale. By a similar argument, the function implies increasing returns to scale when 𝛼 + 𝛽 < 1.
b) Show that all Cobb-Douglas production functions exhibit constant elasticity of substitution.
3. Comparative Statics
Assume there exists some firm producing a good 𝑦 = 𝑓(𝐾, 𝐿), with capital 𝐾 and labor 𝐿. The firm can sell 𝑦 at a
price 𝑝 per unit, and the costs of capital and labor are 𝑟 and 𝑤, respectively.
𝜕𝐾
a) Set up the firm’s cost minimization problem, and use comparative statics to show that the sign of 𝜕𝑦 is
indeterminant.
The cost-minimizing firm will choose the level of K and L that solves the following problem:
min 𝑟𝐾 + 𝑤𝐿 + 𝜆(𝑦 − 𝑓(𝐾, 𝐿))
𝐾,𝐿,𝜆
This has first-order conditions
𝑟 − 𝜆𝑓𝐾 (𝐾, 𝐿) = 0
𝑤 − 𝜆𝑓𝐿 (𝐾, 𝐿) = 0
𝑦 − 𝑓(𝐾, 𝐿) = 0
𝑇ℎ𝑒𝑠𝑒 𝐹𝑂𝐶𝑠 𝑑𝑒𝑠𝑐𝑟𝑖𝑏𝑒 𝑡ℎ𝑒 𝑜𝑝𝑡𝑖𝑚𝑎𝑙 𝑐ℎ𝑜𝑖𝑐𝑒𝑠 𝐾 ∗ ≔ 𝐾 ∗ (𝑟, 𝑤, 𝑦), 𝐿∗ ≔ 𝐿∗ (𝑟, 𝑤, 𝑦), and 𝜆∗ ≔ 𝜆∗ (𝑟, 𝑤, 𝑦).
In this problem, we want to know how the optimal choice of capital 𝐾 ∗ changes when the level of output
𝑦 changes. We can derive this by first rewriting the FOCs in terms of the exogenous parameters:
Next, because we are interested in analyzing the impact of a change in y, differentiate with respect to y:
𝜕𝜆 𝜕𝐾 𝜕𝐿
− 𝑓𝐾 − 𝜆𝑓𝐾𝐾 − 𝜆𝑓𝐾𝐿 =0
𝜕𝑦 𝜕𝑦 𝜕𝑦
𝜕𝜆 𝜕𝐾 𝜕𝐿
− 𝑓𝐿 − 𝜆𝑓𝐿𝐾 − 𝜆𝑓𝐿𝐿 =0
𝜕𝑦 𝜕𝑦 𝜕𝑦
𝜕𝐾 𝜕𝐿
1 − 𝑓𝐾 − 𝑓𝐿 =0
𝜕𝑦 𝜕𝑦
These equations capture how the optimal choices of K, L, and 𝜆 change when 𝑦 changes. We can rewrite
this in matrix form:
𝜕𝐾
𝜕𝑦
−𝜆𝑓𝐾𝐾 −𝜆𝑓𝐾𝐿 −𝑓𝐾 0
𝜕𝐿
[ −𝜆𝑓𝐿𝐾 −𝜆𝑓𝐿𝐿 −𝑓𝐿 ] =[ 0 ]
𝜕𝑦
−𝑓𝐾 −𝑓𝐿 0 −1
𝜕𝜆
[ 𝜕𝑦]
𝜕𝐾
Now, use Cramer’s Rule to solve for 𝜕𝑦 , where the 3x3 matrix above is referred to as 𝐁:
0 −𝜆𝑓𝐾𝐿 −𝑓𝐾
| 0 −𝜆𝑓𝐿𝐿 −𝑓𝐿 |
𝜕𝐾 −1 −𝑓𝐿 0 −𝜆(𝑓𝐾𝐿 𝑓𝐿 − 𝑓𝐿𝐿 𝑓𝐾 )
= =
𝜕𝑦 |𝐁| |𝐁|
𝑇ℎ𝑒 𝑠𝑖𝑔𝑛 𝑜𝑓 𝑡ℎ𝑖𝑠 𝑑𝑒𝑟𝑖𝑣𝑎𝑡𝑖𝑣𝑒 𝑖𝑠 𝑡ℎ𝑒𝑟𝑒𝑓𝑜𝑟𝑒 𝑖𝑛𝑑𝑒𝑡𝑒𝑟𝑚𝑖𝑛𝑎𝑛𝑡 𝑏𝑒𝑐𝑎𝑢𝑠𝑒 𝑤𝑒 𝑑𝑜 𝑛𝑜𝑡 𝑘𝑛𝑜𝑤 𝑤ℎ𝑒𝑡ℎ𝑒𝑟 𝑓𝐾𝐿 𝑓𝐿 −
𝑓𝐿𝐿 𝑓𝐾 is positive, negative, or equal to zero.
b) Set up the firm’s profit maximization problem, and use comparative statics to show that the sign of
𝜕𝐾/𝜕𝑝 is indeterminant.
max 𝑝𝑓(𝐾, 𝐿) − 𝑟𝐾 − 𝑤𝐿
𝐾,𝐿
This produces FOCs
𝑝𝑓𝐾 (𝐾, 𝐿) − 𝑟 = 0
𝑝𝑓𝐿 (𝐾, 𝐿) − 𝑤 = 0
These FOCs provide a solution for the optimal level of capital and labor as functions of the exogenous
parameters,𝐾 ∗ (𝑟, 𝑤, 𝑝) and 𝐿∗ (𝑟, 𝑤, 𝑝). Next, we want to find how the demand for capital changes when
𝜕𝐾
the price of output increases; in other words, we are interested in . First, we totally differentiate the
𝜕𝑝
FOCs with respect to all of the exogenous and endogenous variables:
𝑝𝑓𝐾𝐾 𝑑𝐾 + 𝑝𝑓𝐾𝐿 𝑑𝐿 + 𝑓𝐾 𝑑𝑝 − 𝑑𝑟 = 0
𝑝𝑓𝐿𝐾 𝑑𝐾 + 𝑝𝑓𝐿𝐿 𝑑𝐿 + 𝑓𝐿 𝑑𝑝 − 𝑑𝑤 = 0
These equations describe how the optimal choice of K and L change when the exogenous variables in the
model change. We want to hold the costs of capital and labor constant, so set dr=dw=0, then put these
relationships into matrix form:
𝑝𝑓𝐾𝐾 𝑝𝑓𝐾𝐿 𝑑𝐾 −𝑓 𝑑𝑝
[ ][ ] = [ 𝐾 ]
𝑝𝑓𝐿𝐾 𝑝𝑓𝐿𝐿 𝑑𝐿 −𝑓𝐿 𝑑𝑝
𝐴𝑔𝑎𝑖𝑛, 𝑤𝑒 𝑐𝑎𝑛𝑛𝑜𝑡 𝑑𝑒𝑡𝑒𝑟𝑚𝑖𝑛𝑒 𝑡ℎ𝑒 𝑠𝑖𝑔𝑛 𝑜𝑓 𝑡ℎ𝑖𝑠 𝑒𝑥𝑝𝑟𝑒𝑠𝑠𝑖𝑜𝑛 𝑏𝑒𝑐𝑎𝑢𝑠𝑒 𝑡ℎ𝑒 𝑠𝑖𝑔𝑛 𝑜𝑓 𝑓𝐾𝐿 𝑓𝐿 − 𝑓𝐾 𝑓𝐿𝐿 is
unknown.
c) If 𝜕𝐾/𝜕𝑝 is positive, then what is the sign of 𝜕𝐾/𝜕𝑦? Explain your answer.
𝜕𝐾 𝜕𝐾
If 𝜕𝑝 > 0, then 𝑓𝐾𝐿 𝑓𝐿 − 𝑓𝐾 𝑓𝐿𝐿 > 0 since we know |𝐇| > 0. If 𝑓𝐾𝐿 𝑓𝐿 − 𝑓𝐾 𝑓𝐿𝐿 > 0, then 𝜕𝑦 > 0 because
|𝐁| < 0.
To make sense of this, suppose output increases the demand for capital. If that is the case, then because
output increases with price (this property is due to the convexity of the profit function), the demand for
capital should also increase with price.
4. Duality
Find the production function associated with the following cost function: 𝑐 = (𝑤1 𝑤2 )1⁄2 𝑒 𝑦⁄2.
We want to find the production function 𝑦(𝑥1 , 𝑥2 ). With the given cost function, we can use Shepard’s Lemma to
derive 𝑥𝑖 (𝑤1 , 𝑤2 , 𝑦):
1
𝜕𝑐 1 𝑤2 2 𝑦
𝑥1 (𝑤1 , 𝑤2 , 𝑦) = = ( ) 𝑒2
𝜕𝑤1 2 𝑤1
1
𝜕𝑐 1 𝑤2 −2 𝑦
𝑥2 (𝑤1 , 𝑤2 , 𝑦) = = ( ) 𝑒2
𝜕𝑤2 2 𝑤1
𝑤2
𝑊𝑒 𝑤𝑎𝑛𝑡 𝑡𝑜 𝑒𝑙𝑖𝑚𝑖𝑛𝑎𝑡𝑒 from these equations to obtain an equation for y in terms of 𝑥1 and 𝑥2 :
𝑤1
2 −2
𝑤2 2𝑥1 2𝑥2
=( 𝑦) =( 𝑦)
𝑤1 𝑒2 𝑒2
⇒ 4𝑥1 𝑥2 = 𝑒 𝑦 ⇒ 𝑦 = ln 4𝑥1 𝑥2
5. Factory Location
A firm produces output through the use of capital, 𝐾, and land, 𝐿. The price of capital is constant, 𝑟, but the
price of land, 𝑤, is a decreasing function of its distance, 𝑑, from a distribution center. The firm must ship its
output from its factory to the distribution center; transportation costs are 𝑡 dollars per mile per unit of output
shipped.
b) Derive the system of equations that describes the firm’s optimal choices of 𝐾, 𝐿, and 𝑑.
Take the derivatives of the objective function with respect to the choice variables:
ℒ𝐾 = 𝑟 − 𝜆𝑓𝐾 = 0
ℒ𝐿 = 𝑤(𝑑) − 𝜆𝑓𝐿 = 0
ℒ𝑑 = 𝑤 ′ (𝑑)𝐿 + 𝑡𝑦 = 0
ℒ𝜆 = 𝑦 − 𝑓(𝐾, 𝐿) = 0
From the first two equations, we can derive the usual FOC that the ratio of costs equals the ratio of
marginal production (or equivalently, the marginal costs of K and L are equalized):
𝑟 𝑓𝐾
=
𝑤(𝑑) 𝑓𝐿
When deciding where to locate, the firm chooses the distance from downtown such that the marginal
transportation costs are equal to the marginal cost of land. In other words, in equilibrium, moving one
mile closer to downtown would increase the cost of land by the same amount as the decrease in
transportation costs.
This suggests that in equilibrium, the firm wants to equate the ratio of marginal costs to marginal product across
different factories. The firm should be indifferent between investing one more unit of input in the first or second
factory. This can be achieved when the costs per additional output is the same for both plants. If the plants have
the same technology with constant returns to scale, then 𝑓𝑥1 = 𝑔𝑥2 . In this case, the firm would choose to
produce all output at the cheaper plant only. If the technology is characterized by decreasing returns to scale, the
firm would use more inputs in the less expensive factory, but it may not be able to shift all production there.