Master of Economics

Foundations of
Microeconomics

Chapter 9 - Production Functions

Introduction

■ In previous chapters, we studied consumer theory

■ In chapters 9, 10, and 11, we will consider producers/firms

■ The (representative) firm uses two inputs:

■ Homogeneous labor (), measured in labor-hours
■ Homogeneous capital (k), measured in machine-hours

■ Output is determined by the production function

q = ƒ (k, ) (9.2)

Slide 2
Marginal Productivity p. 297f.

Marginal productivity:
The additional output that can be produced by employing one more unit
of the input holding other inputs constant

■ Marginal productivities of capital and labor are

∂ƒ (k, ) ∂ƒ (k, )
MPk = = ƒk and MP = = ƒ
∂k ∂
■ We assume ƒ , ƒk ≥ 0. Negative marginal productivity does not make
sense economically: increase output by decreasing input.

Slide 3
Diminishing Marginal Productivity p. 298

■ We expect the marginal productivity of an input to be lower if the

quantity of this input is large

■ We assume diminishing marginal productivity:

∂MPk ∂2 ƒ
= = ƒkk ≤ 0,
∂k ∂k 2
∂MP ∂2 ƒ
= = ƒ ≤ 0. (9.4)
∂ ∂2

Slide 4
Cross-Productivity p. 298

■ The marginal productivity of an input can also depend on the other

input
■ For instance, ƒk is often positive, as more capital increases the
productivity of an additional worker:

∂MP
ƒk = >0
∂k

Slide 5
Average Productivity p. 299

■ Average productivities of capital and labor

ƒ (k, ) ƒ (k, )
APk = and AP = (9.5)
k 
■ Depend on the amount of both inputs

Slide 6
Example 9.1:
A Two-Input Production Function (1) p. 299

■ Suppose the production function is

ƒ (k, ) = 600k 2 2 − k 3 3 (9.6)

■ We often assume that k is fixed in the short-run. With k = 10, the

production function becomes

ƒ () = 60, 0002 − 1, 0003

■ Marginal productivity of labor:

ƒ = 120, 000 − 3, 0002 (9.9)

■ Production reaches its maximum when ƒ = 0, i.e. for  = 40 (check

first and second-order conditions)
■ For  > 40, we have negative marginal productivity of labor (!)
Slide 7
Example 9.1:
A Two-Input Production Function (2) p. 299f.

■ Considering average productivity,

ƒ (k, )
AP = = 60, 000 − 1, 0002 (9.10)


■ AP reaches its maximum for  = 30

■ When  = 30, AP = ƒ (k, ) = 900, 000
■ Property: when AP is at its maximum, AP = ƒ (k, )

ƒ (k, ) ∂AP ƒ (k, ) ·  − ƒ (k, )

AP = ⇒ =
 ∂ 2
∂AP ƒ (k,)
■ Thus, when ∂
= 0, ƒ (k, ) ·  = ƒ (k, ) ⇔ ƒ (k, ) = 
⇔ AP = ƒ (k, ).

Slide 8
Marginal Rate of Technical
Substitution (1) p. 301

■ Isoquants show all

combinations of inputs
that yield a given level
of output.
■ Slope of the curves
shows the rate at which
 can be substituted for
k while keeping output
constant.
■ The negative of this
slope is called the
marginal rate of
technical substitution
Figure 9.1: Isoquant map
(MRTS).
Slide 9
Marginal Rate of Technical
Substitution (2) p. 301

■ Marginal rate of technical substitution (MRTS): Rate at which

labor can be substituted for capital holding output constant

dk
MRTS ( for k) = − (9.13)
d q=q0

Slide 10
MRTS and Marginal Productivities p. 302

■ Holding production constant, k becomes an implicit function of :

ƒ (k(), ) = q0 , (9.14)
■ Taking the total derivative:

dk
ƒk + ƒ = 0 (9.15)
d
■ Which implies

dk ƒ MP
MRTS ( for k) = − = = (9.16)
d q=q0 ƒk MPk

■ MRTS is the ratio of the inputs’ marginal productivity

■ MRTS is positive as we assumed both ƒ and ƒk are positive

Slide 11
Isoquants are Convex (1) p. 302f.

€ Š
ƒ
€ Š € Š
dMRTS d ƒk ƒk ƒ + ƒk dk
d
− ƒ ƒ
 k + ƒ dk
kk d
= = (9.17/18)
d d ƒk2
dk ƒ
■ As d
= − ƒ  along an isoquant and ƒk = ƒk (Young’s theorem)
k

dMRTS ƒk2 ƒ − 2ƒk ƒ ƒk + ƒ2 ƒkk

= (9.19)
d ƒk3
■ Denominator is positive because we have assumed ƒk > 0
■ MRTS is decreasing if production is quasi-concave

Slide 12
Isoquants are Convex (2) p. 302f.

dMRTS ƒk2 ƒ − 2ƒk ƒ ƒk + ƒ2 ƒkk

= (9.19)
d ƒk3

■ Recall we assumed ƒ and ƒkk are negative

■ A sufficient condition for the ratio to be negative is ƒk = ƒk ≥ 0.
■ E.g., if workers have more capital, they will be more productive
■ However, it could be that ƒk < 0
■ We assume the decrease of marginal productivities compensates for
any possible negative cross-productivity effects, i.e., we assume
diminishing MRTS

Slide 13
Example 9.2: Diminishing MRTS p. 302f.

Consider the production function ƒ (k, ) = 600k 2 2 − k 3 3 . For

which values of (k, ) does this function satisfy our assumptions
on production functions? To answer, express and comment on
each element of

dMRTS ƒk2 ƒ − 2ƒk ƒ ƒk + ƒ2 ƒkk

= (9.19)
d ƒk3

Slide 14
Returns to Scale (1) p. 304

■ How does output respond to a proportional increase in all inputs?

■ Suppose we multiply k and  by t > 1
■ Greater division of labor and specialization of tasks: production could be
multiplied by more than t
■ Loss in efficiency: production could be multiplied by less than t
■ These effects are captured by returns to scale
ƒ (tk, t) = t γ ƒ (k, )

■ If γ = 1, constant returns to scale

■ If γ < 1, decreasing returns to scale
■ If γ > 1, increasing returns to scale

Slide 15
Returns to Scale (2) p. 305

■ Returns to scale are generally defined within a narrow range of

variation in input
■ A local measure of returns to scale is the scale elasticity:

∂ƒ (tk, t) t
eq,t = ·
∂t ƒ (tk, t)

Slide 16
Brief Excursion on
Homogeneous Functions, Take 3

■ Constant returns to scale production functions are homogeneous of

degree one in inputs
ƒ (tk, t) = tƒ (k, ) (9.24)
■ This implies that the marginal productivities are homogeneous of
degree zero. To see this, take the partial derivative of both sides with
respect to k:

∂ƒ (tk, t) ∂ƒ (k, )

=t ⇒ tƒk (tk, t) = tƒk (k, )
∂k ∂k
⇒ ƒk (tk, t) = ƒk (k, )
■ Indeed, ƒk is homogeneous of degree zero. A similar argument can be
made for ƒ .
■ General property: if a function is homogeneous of degree k, its partial
derivatives are homogeneous of degree k − 1.
Slide 17
Constant Returns to Scale (1) p. 305

■ For constant returns to scale functions, the marginal productivity

functions are homogeneous of degree zero:

∂ƒ (k, ) ∂ƒ (tk, t)

MPk = =
∂k ∂k
∂ƒ (k, ) ∂ƒ (tk, t)
MP = =
∂ ∂
1
■ For t =  € Š € Š
k k
∂ƒ 
,1 ∂ƒ 
,1
MPk = , MP = (9.26)
∂k ∂
■ The marginal productivity of any input depends on the ratio of capital
and labor, not their absolute levels.

Slide 18
Constant Returns to Scale (2) p. 306

Homothetic production function:

all isoquants are radial expansions
of one another.
Constant returns to scale produc-
tion functions are homothetic:
■ MRTS depends only on the ratio
of k to , not on the scale of
production.
■ Thus, along any ray through
the origin (a ray of constant k ),
the MRTS is the same on all
isoquants.
■ Production increases
proportionally with inputs. Figure 9.2: Isoquants for constant returns to scale
Slide 19
Homothetic Production Function p. 306f.

■ Suppose ƒ (k, ) is a constant returns to scale production function. We

take the monotonic transformation:

F(k, ) = [ƒ (k, )] γ (9.27)

where γ > 1 captures the degree of the increasing returns to scale. To

see this:

F(tk, t) = [ƒ (tk, t)] γ = [tƒ (k, )] γ = t γ [ƒ (k, )] γ = t γ F(k, ) (9.28)

■ F(k, ) is homogeneous of degree γ (and homothetic)

■ A function with increasing (or decreasing) returns to scale can also be
homothetic

Slide 20
The n-Input Case p. 307

■ Returns to scale can be generalized to a production function with n

inputs
q = ƒ (1 , 2 , ..., n ) (9.29)

■ If all inputs are multiplied by t > 0:

ƒ (t1 , t2 , ..., tn ) = t γ ƒ (1 , 2 , ..., n ) (9.30)

■ If γ = 1, constant returns to scale

■ If γ < 1, decreasing returns to scale
■ If γ > 1, increasing returns to scale
However, a proportional increase of all inputs may make little
economic sense in practice.

Slide 21
The Elasticity of Substitution (1) p. 307f.

■ Recall that the elasticity of substitution σ is a scale-free measurement

of how the ratio of inputs changes, as the slope of the isoquant
changes.
■ It measures the curvature of an isoquant.
Δ(k/ )/ k/  d (k/ ) MRTS d n (k/ ) d n (k/ )
σ= = = = (9.33)
ΔMRTS/ MRTS dMRTS k/  d n MRTS d n (ƒ/ ƒk )
■ σ is positive because k/  and MRTS move in the same direction
k
■ If σ is low, the MRTS changes by a substantial amount as 
changes.
The isoquant is sharply curved.

Slide 22
The Elasticity of Substitution (2) p. 309

■ σ can change along an

isoquant (e.g., at A and
B) or as the scale of
production changes
■ However, σ is the same
for all isoquants of
homothetic production
functions along a ray
going through the
origin
■ MRTS and k are the
same at points A and
C and at B and D
■ Consequently, σ is
the same along the Figure: Elasticity of substitution for homothetic production
two isoquants functions
Slide 23
Linear Production Function (1) p. 310

■ Linear production function:

ƒ (k, ) = αk + β (9.34)
■ Constant returns to scale:

ƒ (tk, t) = αtk + βt = t(αk + β) = tƒ (k, ) (9.35)

β
■ All isoquants are straight lines with slope − α
β
■ MRTS = α
is constant
■ As the denominator of σ in (9.33) is ΔMRTS = 0, σ = ∞

Slide 24
Linear Production Function (2) p. 311

■ Capital and labor are perfect

substitutes.

■ The MRTS does not change

as the capital–labor ratio
changes.

Figure 9.4 (a): Isoquants for perfect substitutes

Slide 25
Fixed Proportions
Production Function (1) p. 310f.

■ Fixed proportions production function:

ƒ (k, ) = mn(αk, β), α, β > 0 (9.36)

■ Capital and labor must always be used in a fixed ratio

k β
■ The firm will always operate along a ray where 
= α
, i.e., at a
constant ratio of capital and labor

■ Moving from the vertical part of the L-shaped isoquant to the

horizontal part, the MRTS jumps from from ∞ to 0. As the
denominator of σ in (9.31) becomes ΔMRTS = ∞ − 0, σ = 0.

Slide 26
Fixed Proportions
Production Function (2) p. 311

■ No substitution is possible.

■ The capital–labor ratio is

β
fixed at α .

Figure 9.4 (b): Isoquants for fixed proportions

Slide 27
Cobb-Douglas
Production Function (1) p. 312f.

■ Cobb-Douglas production function:

ƒ (k, ) = Ak α  β
A, α, β > 0 (9.37)

■ Can exhibit any returns to scale:

ƒ (tk, t) = A(tk)α (t) β

= At α+β k α  β
= t α+β ƒ (k, ) (9.38)

■ Returns to scale are constant, increasing and decreasing for α + β = 1,

α + β > 1 and α + β < 1, respectively.
■ Cobb-Douglas has σ = 1 (see slides of Chapter 3 or Chapter 9 in the
book for a proof).

Slide 28
Cobb-Douglas
Production Function (2) p. 312f.

The Cobb-Douglas production func-

tion is linear in logarithms:

n(q) = n(A) + α n(k) + β n() (9.39)

■ α is the elasticity of output with

respect to k
■ β is the elasticity of output with
respect to 

Figure 9.4 (c): Isoquants for Cobb-Douglas

Slide 29
Constant Elasticity of Substitution (CES)
Production Function p. 313

γ
ƒ (k, ) = [k ρ +  ρ ] ρ , ρ ≤ 1, ρ ̸= 0, γ > 0 (9.40)

■ Increasing returns to scale if γ > 1; decreasing returns to scale if

γ<1⇒

1
■ σ= 1−ρ

ρ = 1 ⇒ linear production function

ρ = −∞ ⇒ fixed proportions production function
ρ = 0 ⇒ Cobb-Douglas production function

Slide 30
Example 9.3: Generalized Leontief
Production Function p. 313f

Consider the production function

p
ƒ (k, ) = k +  + 2 k · .

■ Show that the function exhibits constant returns to scale.

■ How does the MRTS depend on the ratio of the two inputs?
■ Calculate the elasticity of substitution.

Slide 31