ECON 403 Lecture Notes

Producer Theory.
Philip Gunby
Department of Economics
University of Canterbury
p.gunby@econ.canterbury.ac.nz
c _Philip P. Gunby 2006
DISCLAIMER: If you do not attend lectures, you are duly warned that these notes
do not contain all of the material presented in class. You do not attend lectures at
your own peril! Also note that not all material in these notes may be covered in class.
Finally, no matter how carefully I proof-read these lecture notes, there are still bound
to be mistakes in them!
1. Introduction
Economic activity not only involves consumption but also production and trade. Pro-
duction should be interpreted very broadly, however, to include production of both
physical goods such as rice or automobiles and services such as medical care or nan-
cial services.
A rm can be characterised by many factors and aspects such as sectors, production
scale, ownership, organisation structures, etc. But which are most important features
for us to study producers behaviour in making choices? To grasp the most important
features in studying producer behaviour and choices in modern producer theory, it
is assumed that the key characteristic of a rm is the production set. Producers
characteristics together with the behaviourial assumptions are building blocks in any
model of producer theory. The production set represents the set of all technologically
feasible production plans. The behaviourial assumption expresses the guiding principle
the producer uses to make choices. It is generally assumed that the producer seeks to
identify and select a production that is most protable.
We will rst present a general framework of production technology. By itself, the
framework does not describe how production choices are made. It only species basic
characteristic of a rm which denes what choices can be made; it does not specify
what choices should be made. Then we will discuss what choices should be made based
on the behaviourial assumptions placed on rms. A basic behaviourial assumption on
producers is prot maximisation. After that, we will describe production possibilities
in physical terms, which is recast into economic terms using cost functions.
2. Production Technology
Production is the process of transforming inputs to outputs. Typically, inputs consist
of labour, capital equipment, raw materials, and intermediate goods purchased from
other rms. Outputs consist of nished products or service, or intermediate goods to
1
be sold to other rms. Often alternative methods are available for producing the same
output, using dierent combinations of inputs. A rm produces outputs from various
combinations of inputs. In order to study rm choices we need a convenient way to
summarise the production possibilities of the rm, i.e., which combinations of inputs
and outputs are technologically feasible.
2.1 Measurement of Inputs and Outputs
It is usually most satisfactory to think of the inputs and outputs as being measured
in terms of ows: a certain amount of inputs per time period are used to produce a
certain amount of outputs per unit time period at some location. It is a good idea
to explicitly include the time and location dimensions in a specication of inputs and
outputs. The level of detail that we will use in specifying inputs and outputs will depend
on the problem at hand, but we should remain aware of the fact that a particular input
or output good can be specied in arbitrarily ne detail. However, when discussing
technological choices in the abstract, as we do in this chapter, it is common to omit the
time and location dimensions.
2.2 Specication of Technology
The fundamental reality rms must contend with in this process is technological feasi-
bility. The state of technology determines and restricts what is possible in combining
inputs to produce outputs, and there are several way we can represent this constraint.
The most general way is to think of the rm as having a production possibility set.
Suppose the rm has L possible goods to serve as inputs and/or outputs. If a rm uses
y
i
j
units of a good j as an input and produces y
o
j
of the good as an output, then the
net output of good j is given by y
j
= y
o
j
y
i
j
.
A production plan is simply a list of net outputs of various goods. We can represent a
production plan by a vector y R
L
where y
j
is negative if the jth good serves as a net
input and positive if the jth good serves as a net output. The set of all technologically
feasible production plans is called the rms production possibilities set and will be
denoted by Y of'
L
.
[see Typical Production Possibilities Set graph]
The set Y is supposed to describe all patterns of inputs and outputs that are technolog-
ically feasible. It gives us a complete description of the technological possibilities facing
the rm. When we study the behaviour of a rm in certain economic environments, we
2
may want to distinguish between production plans that are immediately feasible and
those that are eventually feasible. We will generally assume that such restrictions
can be described by some vector z '
L
. The restricted or short-run production
possibilities set will be denoted by Y (z); this consists of all feasible net output bun-
dles consistent with the constraint level z. The following are some examples of such
restrictions.
Example: (Input Requirement Set) Suppose a rm produces only one output. In this
case we write the net output bundle as (y, x) where x is a vector of inputs that
can produce y units of output. We can then dene a special case of a restricted
production possibilities set, the input requirement set:
V (y) = x '
L
+
: (y, x) Y .
The input requirement set is the set of all input bundles that produce at least y
units of output.
Note that the input requirement set, as dened here, measures inputs as positive num-
bers rather than negative numbers as used in the production possibilities set.
Example: (Isoquant) In the case above we can also dene an isoquant:
Q(y) = x '
n
+
: x V (y) and x / V (y

) y

> y.
The isoquant gives all input bundles that produce exactly y units of output.
[see Convex Input Requirement Sets graph]
[see Isoquant Curves graph]
Example: (Short-Run Production Possibilities Set) Suppose a rm produces some
output from labour and some kind of machine which we will refer to as capital.
Production plans then look like (y, l, k) where y is the level of output, l the
amount of labour input, and k the amount of capital input. We imagine that
labour can be varied immediately but that capital is xed at the level

k in the
short run. Then
Y (

k) = (y, l, k) Y : k =

k
is a short-run production possibilities set.
Example: (Production Function) If the rm has only one output, we can dene the pro-
duction function: f(x) = y ' : y is the maximum output associated with x
Y .
3
Example: (Transformation Function) There is an n dimensional analog of the produc-
tion function. A production plan y Y is (technologically) ecient if there is no
y

Y such that y

y and y

,= y; that is, a production plan is ecient if there

is no way to produce more output with the same inputs or to produce the same
output with less inputs. (Note carefully how the sign convention on inputs works
here.) We often assume that we can describe the set of technologically ecient
production plans by a transformation function T : '
n
' where T(y) = 0 i y is
ecient. Just as a production function picks out the maximum scalar output as a
function of the inputs, the transformation function picks out the maximal vectors
of net outputs.
Example: (Cobb-Douglas Technology) Let be a parameter such that 0 < < 1.
Then the Cobb-Douglas technology is dened in the following manner.
Y = (y, x
1
, x
2
) '
3
: y x

1
x
1
2

V (y) = (x
1
, x
2
) '
2
+
: y x

1
x
1
2

Q(y) = (x
1
, x
2
) '
2
+
: y = x

1
x
1
2

Y (z) = (y, x
1
, x
2
) '
3
: y x

1
x
1
2
, x
2
= z
T(y, x
1
, x
2
) = y x

1
x
1
2
f(x
1
, x
2
) = x

1
x
1
2
.
Example: (Leontief Technology) Let a > 0 and b > 0 be parameters. Then the
Leontief technology is dened in the following manner.
Y = (y, x
1
, x
2
) '
3
: y min(ax
1
, bx
2
)
V (y) = (x
1
, x
2
) '
2
+
: y min(ax
1
, bx
2
)
Q(y) = (x
1
, x
2
) '
2
+
: y = min(ax
1
, bx
2
)
T(y, x
1
, x
2
) = y min(ax
1
, bx
2
)
f(x
1
, x
2
) = min(ax
1
, bx
2
).
2.3 Common Properties of Production Sets
Although the production possibility sets of dierent processes can dier widely in struc-
ture, many technologies share certain general properties. If it can be assumed that these
properties are satised, special theoretical results can be derived. Some important prop-
erties are dened below:
4
Axiom (Possibility of Inaction): 0 Y .
Possibility of inaction means that no action on production is a possible production plan.
Axiom (Closeness): Y is closed. The possibility set Y is closed means that,
whenever a sequence of production plans y
i
, i = 1, 2, . . ., are in Y and y
i
y, then the
limit production plan y Y .
It guarantees that points on the boundary of Y are feasible. Note that Y is closed
implies that the input requirement set V (y) is a closed set y 0.
Axiom (Free Disposal): If y Y implies that y

Y for all y

y, then the set

Y is said to satisfy the free disposal or monotonicity property.
Free disposal implies that commodities (either inputs or outputs) can be thrown away.
This property means that if y Y , then Y includes all vectors in the negative orthant
translated to y, i.e. there are only inputs, but no outputs. A weaker requirement is
that we only assume that the input requirement is monotonic.
Axiom (Monotonicity): If x V (y) and x

x, then x

V (y).
Monotonicity of V (y) means that, if x is a feasible way to produce y units of output and
x

is an input vector with at least as much of each input, then x

should be a feasible
way to produce y.
Axiom (Irreversibility): Y Y = 0.
Irreversibility means a production plan is not reversible unless it is a non-action plan.
Axiom (Convexity): Y is convex if whenever y, y

Y , the weighted average

ty + (1 t)y

Y for any t with 0 t 1.

Convexity of Y means that, if all goods are divisible, it is often reasonable to assume
that two production plans y and y

can be scaled downward and combined. However,

it should be noted that the convexity of the production set is a strong hypothesis. For
example, convexity of the production set rules out start up costs and other sorts of
returns to scale.
5
Axiom (Strict Convexity): y is strictly convex if y Y and y

Y , then
ty + (1 t)y

intY 0 < t < 1, where intY denotes the interior points of Y .

As we will show, the strict convexity of Y can guarantee the prot maximising produc-
tion plan is unique provided it exists. A weak and more reasonable requirement is to
assume that V (y) is a convex set for all outputs y
o
:
Axiom (Convexity of Input Requirement Set): If x, x

V (y), then tx+(1

t)x

V (y) 0 t 1. That is, V (y) is a convex set.

Convexity of V (y) means that, if x and x

both can produce y units of output, then

any weighted average tx + (1 t)x

can also produce y units of output. We describe a

few of the relationships between the convexity of V (y), the curvature of the production
function, and the convexity of Y . We rst have,
Proposition 2.1 (Convex Production Set Implies Convex Input Requirement
Set): If the production set Y is a convex set, then the associated input requirement
set, V (y), is a convex set.
Proof: If Y is a convex set then it follows that for any x and x

such that (y, x) and

(y, x

) are in Y for 0 t 1, we must have (ty + (1 t)y, tx (1 t)x

) Y .
This is simply requiring that (y, (tx + (1 t)x

)) Y . It follows that if x, x

V (y),
tx + (1 t)x

V (y) which shows that V (y) is convex. Q.E.D.

Proposition 2.2: V (y) is a convex set i the production function f(x) is a quasi-
concave function.
Proof: V (y) = x : f(x) y, which is just the upper contour set of f(x). But a
function is quasi-concave i it has a convex upper contour set. Q.E.D.
2.4 Returns to Scale
Suppose that we are using some vector of inputs x to produce some output y and
we decide to scale all inputs up or down by some amount t 0. What will happen
to the level of output? The notions of returns to scale can be used to answer this
question. Returns to scale refer to how output responds when all inputs are varied in
the same proportion so that they consider long run production processes. There are
6
three possibilities: technology exhibits (1) constant returns to scale; (2) decreasing
returns to scale; and (3) increasing returns to scale. Formally, we have,
Denition ((Global) Returns to Scale): A production function f(x) is said to
exhibit:
(1) constant returns to scale if f(tx) = tf(x) t 0.
(2) decreasing returns to scale if f(tx) < tf(x) t > 1.
(3) increasing returns to scale if f(tx) > tf(x) t > 1.
Constant returns to scale (CRS) means that doubling inputs exactly double outputs,
which is often a reasonable assumption to make about technologies. Decreasing returns
to scale means that doubling inputs are less than doubling outputs. Increasing returns
to scale means that doubling inputs are more than doubling outputs.
Note that a technology has constant returns to scale if and only if its production
function is homogeneous of degree 1. Constant returns to scale is also equivalent to:
the statement y Y implies ty Y t 0; or equivalent to the statement x V (y)
implies tx V (ty) t > 1.
It may be remarked that the various kinds of returns to scale dened above are global
in nature. It may well happen that a technology exhibits increasing returns to scale
for some values of x and decreasing returns to scale for other values. Thus in many
circumstances a local measure of returns to scale is useful. To dene locally returns to
scale, we rst dene elasticity of scale.
Denition (Elasticity of Scale): The elasticity of scale measures the percent
increase in output due to a one percent increase in all inputs that is, due to an
increase in the scale of operations.
Let y = f(x) be the production function. Let t be a positive scalar, and consider the
function y(t) = f(tx). If t = 1, we have the current scale of operation; if t > 1, we
are scaling all inputs up by t; and if t < 1, we are scaling all inputs down by t. The
elasticity of scale is given by
e(x) =
dy(t)
y(t)
dt
t
,
evaluated at t = 1. Rearranging this expression, we have
e(x) =
dy(t)
dt
t
y
[t = 1 =
df(tx)
dt
t
f(tx)
[t = 1
7
Note that we must evaluate the expression at t = 1 to calculate the elasticity of scale
at the point x. Thus, we have the following the local returns to scale:
[see Returns to Scale and Varying Proportions graph]
Denition ((Local) Returns To Scale): A production function f(x) is said to
exhibit locally increasing, constant, or decreasing returns to scale as e(x) is greater,
equal, or less than 1.
2.5 The Marginal Rate of Technical Substitution
Suppose that technology is summarised by a smooth production function and that we
are producing at a particular point y

= f(x

1
, x

2
). Suppose that we want to increase
a small amount of input 1 and decrease some amount of input 2 so as to maintain a
constant level of output. How can we determine this marginal rate of technical
substitution (MRTS) between these two factors? The way is the same as for deriving
the marginal rate of substitution of an indierence curve. Dierentiating the production
function when output is constant, gives us the following total dierential,
0 =
f
x
1
dx
1
+
f
x
2
dx
2
which can be solved for
dx
2
dx
1
=
f
x
1
f
x
2

MP
x
1
MP
x
2
.
This gives us an explicit expression for the marginal rate technical substitution, which
is the rate of marginal production of x
1
and marginal production of x
2
.
[see MRTS graph]
Example: (MRTS for a Cobb-Douglas Technology) Given that f(x
1
, x
2
) = x

1
x
1
2
,
we can take the derivatives to nd,
f(bfx)
x
1
= x
1
1
x
1
2
=
_
x
2
x
1
_
1
f(x)
x
2
= (1 )x

1
x

2
= (1 )
_
x
1
x
2
_

.
It follows that,
x
2
(x
1
)
x
1
=
f
x
1
f
x
2
=

1
x
2
x
1
.
8
2.6 The Elasticity of Substitution
The marginal rate of technical substitution measures the slope of an isoquant. The
elasticity of substitution measures the curvature of an isoquant. More specically,
the elasticity of substitution measures the percentage change in the factor ratio divided
by the percentage change in the MRTS, with output being held xed. If we let (x
2
/x
1
)
be the change in the factor ratio and MRTS be the change in the technical rate of
substitution, we can express this as
=
(x
2
/x
1
)
x
2
/x
1
MRTS
MRTS
This is a relatively natural measure of curvature: it asks how the ratio of factor inputs
changes as the slope of the isoquant changes. If a small change in slope gives us a large
change in the factor input ratio, the isoquant is relatively at which means that the
elasticity of substitution is large.
In practice we think of the percent change as being very small and take the limit of
this expression as goes to zero. Hence, the expression for becomes
=
MRTS
(x
2
/x
1
)
d(x
2
/x
1
)
dMRTS
=
d ln(x
2
/x
1
)
d ln [MRTS[
.
(The absolute value sign in the denominator is to convert the MRTS to a positive
number so that the logarithm makes sense.) See pp.13-14 of Varian to see how to
derive this result.
Example: (The Cobb-Douglas Production Function) We have seen above that,
MRTS =

1
x
2
x
1
,
or,
x
2
x
1
=
1

MRTS
It follows that,
ln
x
2
x
1
= ln
1

+ ln [MRTS[.
This in turn implies,
=
d ln(x
2
/x
1
)
d ln [MRTS[
= 1.
Example: (The CES Production Function) The constant elasticity of substitution
(CES) production function has the form,
y = [a
1
x

1
+a
2
x

2
]
1

.
9
It is easy to verify that the CES function exhibits constant returns to scale. It
will probably not surprise you to discover that the CES production function has
a constant elasticity of substitution. To verify this, note that the marginal rate of
technical substitution is given by
MRTS =
_
x
1
x
2
_
1
,
so that
x
2
x
1
= [MRTS[
1
1
.
Taking logs, we see that
ln
x
2
x
1
=
1
1
ln [MRTS[.
Applying the denition of a using the logarithmic derivative,
=
d ln x
2
/x
1
d ln [MRTS[
=
1
1
.
[see Isoquants for CES Production Function graph]
3. Prot Maximisation
3.1 Producer Behaviour
A basic hypothesis on individual rm behaviour in the producer theory is that a rm
will always choose a most protable production plan from the production set. We will
derive input demand and output supply functions by considering a model of prot-
maximising behaviour coupled with a description of underlying production constraints.
Economic prot is dened to be the dierence between the revenue a rm receives and
the costs that it incurs. It is important to understand that all (explicit and implicit)
costs must be included in the calculation of prot. Both revenues and costs of a rm
depend on the actions taken by the rm. We can write revenue as a function of the level
of operations of some n actions, R(a
1
, . . . , a
n
), and costs as a function of these same
n activity levels, C(a
1
, . . . , a
n
), where actions can be in term of employment levels of
inputs or output levels of production or prices of outputs if the rm has market power
to set the prices.
10
A basic assumption of most economic analysis of rm behaviour is that a rm acts so
as to maximise its prots; that is, a rm chooses actions (a
1
, . . . , a
n
) so as to maximise
R(a
1
, . . . , a
n
) C(a
1
, . . . , a
n
). The prot maximisation problem facing the rm can
then be written as
max
a
1
,...,a
n
R(a
1
, . . . , a
n
) C(a
1
, . . . , a
n
)
The rst order conditions for interior optimal actions, a

= (a

1
, . . . , a

n
), is characterised
by the conditions
R(a

)
a
i
=
C(a

)
a
i
i = 1, . . . , n.
The intuition behind these conditions should be clear: if marginal revenue were greater
than marginal cost, it would pay to increase the level of the activity; if marginal revenue
were less than marginal cost, it would pay to decrease the level of the activity. In general,
revenue is composed of two parts: how much a rm sells of various outputs times the
price of each output. Costs are also composed of two parts: how much a rm uses of
each input times the price of each input.
The rms prot maximisation problem therefore reduces to the problem of determining
what prices it wishes to charge for its outputs or pay for its inputs, and what levels of
outputs and inputs it wishes to use. In determining its optimal policy, the rm faces
two kinds of constraints: technological constraints that are specied by production sets
and market constraints that concern the eect of actions of other agents on the rm.
[see Firms Prot Maximisation Problem graph]
The rms described in the remainder of this chapter are assumed to exhibit the sim-
plest kind of market behaviour, namely that of price-taking behaviour. Each rm
will be assumed to take prices as given. Thus, the rm will be concerned only with
determining the prot maximising levels of outputs and inputs. Such a price-taking
rm is often referred to as a competitive rm. For the more general case see the mono-
graphs mentioned in the references, usually under a Chapter title such as the theory
of markets.
3.2 Producers Optimal Choice
Let p be a vector of prices for inputs and outputs of the rm. The prot maximisation
problem of the rm can be stated as
(p) =max py (3.1)
s.t. y Y.
Note that since outputs are measured as positive numbers and inputs are measured as
negative numbers, the objective function for this problem is prots: revenues minus
11
costs. The function (p), which gives us the maximum prots as a function of the
prices, is called the prot function of the rm. There are several useful variants of
the prot function
Denition (Short-Run or Restricted Prot Function): In this case, we might
dene the short-run or restricted prot function as:
(p, z) =max py
s.t. y Y (z).
If the rm produces only one output, the prot function can be written as
(p, w) = max pf(x) wx
where p is now the (scalar) price of output, w is the vector of factor prices, and the
inputs are measured by the (nonnegative) vector x = (x
1
, . . . , x
n
).
The value of y that solves the prot problem (3.1) is in general not unique. When
there is such a unique production plan, the production plan is called the net out-
put function or net supply function, the corresponding input part is called the
producers input demand function and the corresponding output vector is called
the producers output supply function. We will see from the following proposition
that strict convexity of the production set will ensure the uniqueness of the optimal
production plan.
Proposition 3.1: Suppose Y strictly convex. Then, for each given p '
L
+
, the
prot maximising production is unique provided it exists.
Proof: Suppose not. Let y and y

be two prot maximising production plans for p

'
L
+
. Then, we must have py = py

. Thus, by the strict convexity of Y , we have

ty + (1 t)y

intY 0 < t < 1. Therefore, k > 1 s.t.

k[ty + (1 t)y

] intY. (3.2)
Then k[tpy + (1 t)py

] = kpy > py which contradicts the fact that y is a prot

maximising production plan. Q.E.D.
12
3.3 Producers First-Order Conditions
Prot-maximising behaviour can be characterised by calculus when the technology can
be described by a dierentiable production function. For example, the rst-order con-
ditions for the single output prot maximisation problem with interior solution are
p
f(x

)
x
i
= w
i
i = 1, . . . , n. (3.3)
Using vector notation, we can also write these conditions as
pDf(x

) = w.
The rst-order conditions state that the marginal value product of each factor must be
equal to its price, i.e., marginal revenue equals marginal cost at the prot maximising
production plan. This rst-order condition can also be exhibited graphically.
[see Prot Maximisation Condition graph]
In this two-dimensional case, prots are given by = py wx. The level sets of this
function for xed p and w are straight lines which can be represented as functions of the
form: y = /p+(w/p)x. Here the slope of the isoprot line gives the wage measured in
units of output, and the vertical intercept gives us prots measured in units of output.
At the point of maximal prots the production function must lie below its tangent line
at x

; i.e., it must be locally concave.

Similar to the arguments in the section on consumer theory, the calculus conditions
derived above make sense only when the choice variables can be varied in an open
neighbourhood of the optimal choice. The relevant rst-order conditions that also
include boundary solutions are given by the Kuhn-Tucker conditions:
p
f(x)
x
i
w
i
0 with equality if x
i
> 0. (3.4)
Remark: There may exist no prot maximising production plan when a production
technology exhibits constant returns to scale or increasing returns to scale . For exam-
ple, consider the case where the production function is f(x) = x. Then for p > w no
prot-maximising plan will exist. It is clear from this example that the only nontriv-
ial prot-maximising position for a constant-returns-to-scale rm is the case of p = w
and zero prots. In this case, all production plans are prot-maximising production
plans. If (y, x) yields maximal prots of zero for some constant returns technology,
then (ty, tx) will also yield zero prots and will therefore also be prot-maximising.
13
3.4 Suciency of Producers First-Order Condition
The second-order condition for prot maximisation is that the matrix of second deriva-
tives of the production function must be negative semi-denite at the optimal point;
that is, the second-order condition requires that the Hessian matrix
D
2
f(x

) =
_

2
f(x

)
x
i
x
j
_
must satisfy the condition that hD
2
f(x

)h

0 for all vectors h. (The prime indicates

the transpose operation.) Geometrically, the requirement that the Hessian matrix is
negative semi-denite means that the production function must be locally concave in
the neighbourhood of an optimal choice. Formerly, we have the following proposition.
Proposition 3.2: Suppose that f(x) is dierentiable and concave on '
+
and
(p, w) > 0. If x > 0 satises the rst-order conditions given in (3.4), then x is a
(globally) prot maximising production plan at prices (p, w).
Remark: The strict concavity of f(x) can be checked by verifying if the leading
principal minors of the Hessian must alternate in sign, i.e.,

f
11
f
12
f
21
f
22

> 0,

f
11
f
12
f
13
f
21
f
22
f
23
f
31
f
32
f
33

> 0,
and so on, where f
ij
=

2
f
x
i
x
j
. This algebraic condition is useful for checking second-
order conditions.
Example: (The Prot Function for Cobb-Douglas Technology) Consider the problem
of maximising prots for the production function of the form f(x) = x
a
where
a > 0. The rst-order condition is
pax
a1
= w,
and the second-order condition reduces to
pa(a 1)x
a2
0.
The second-order condition can only be satised when a 1, which means that
the production function must have constant or decreasing returns to scale for com-
petitive prot maximisation to be meaningful. If a = 1, the rst-order condition
14
reduces to p = w. Hence, when w = p any value of x is a prot-maximising choice.
When a < 1, we use the rst-order condition to solve for the factor demand func-
tion
x(p, w) =
_
w
ap
_ 1
a1
.
The supply function is given by
y(p, w) = f(x(p, w)) =
_
w
ap
_ 1
a1
,
and the prot function is given by
(p, w) = py(p, w) wx(p, w) = w
_
1 1
a
__
w
ap
_ 1
a1
.
3.5 Properties of Net Supply Functions
In this section, we show that the net supply functions that are the solutions to the
prot maximisation problem in fact impose certain restrictions on the behaviour of the
demand and supply functions.
Proposition 3.3: Net output functions y(p) are homogeneous of degree zero, i.e.,
y(tp) = y(p) t > 0.
Proof: It is easy to see that if we multiply all of the prices by some positive number
t, the production plan that maximises prots will not change. Hence, we must have
y(tp) = y(p) t > 0.
Proposition 3.4 (Negative Deniteness of Substitution Matrix): Let y =
f(x) be a twice dierentiable and strictly concave single output production function,
and let x(p, w) be the input demand function. Then, the substitution matrix,
Dx(p, w) =
_
x
i
(p, w)
w
j
_
is symmetric negative denite.
Proof: Without loss of generality, normalise p = 1. Then the rst-order conditions for
prot maximisation are
Df(x(w)) w 0
15
If we dierentiate with respect to w, we get
D
2
f(x(w))Dx(w)) I 0.
Solving this equation for the substitution matrix, we nd
Dx(w)) [D
2
f(x(w))]
1
Recall that the second-order condition for (strict) prot maximisation is that the Hes-
sian matrix is a symmetric negative denite matrix. It is a standard result of linear
algebra that the inverse of a symmetric negative denite matrix is a symmetric negative
denite matrix. Then, D
2
f(x(w)) is a symmetric negative denite matrix, and thus the
substitution matrix Dx(w)) is a symmetric negative denite matrix. This means that
the substitution matrix itself must be a symmetric, negative denite matrix. Q.E.D.
Remark: Note that since Dx(p, w) is symmetric, negative denite, we have in
particular:
(1) x
i
/w
i
< 0, for i = 1, 2, . . . , n since the diagonal entries of a negative denite
matrix must be negative.
(2) x
i
/w
j
= x
j
/w
i
by the symmetry of the matrix.
4. Prot Function
Given any production set Y , we have seen how to calculate the prot function, (p),
which gives us the maximum prot attainable at prices p. The prot function possesses
several important properties that follow directly from its denition. These properties
are very useful for analysing prot-maximising behaviour.
4.1 Properties of the Prot Function
The properties given below follow solely from the assumption of prot maximisation. No
assumptions about convexity, monotonicity, or other sorts of regularity are necessary.
Proposition 4.1 (Properties of the Prot Function): The following has the
following properties:
(1) Nondecreasing in output prices, nonincreasing in input prices. If p

i
p
i
for all
outputs and p

j
p
j
, for all inputs, then (p

) (p).
16
(2) Homogeneous of degree 1 in p. That is, (tp)) = t(p) t 0.
(3) Convex in p. That is, let p

= tp + (1 t)p

for 0 t 1 then (p

)
t(p) + (1 t)(p

).
(4) Continuous in p. That is, the function (p) is continuous, at least when (p) is
well-dened and p
i
> 0 for i = 1, . . . , n.
Proof:
(1) Let y be a prot-maximising net output vector at p, so that (p) = py and let y

be a prot-maximising net output vector at p

so that (p

) = p

. Then by denition
of prot maximisation we have p

y. Since p

i
p
i
for all i for which y
i
0 and
p

i
p
i
for all i for which y
i
0, we also have p

y py. Putting these two inequalities

together, we have (p

) = p

py = (p), as required.
(2) Let y be a prot-maximising net output vector at p, so that py py

for all y

Y .
It follows that for t 0, tpy = tpy

for all y

Y . Hence y also maximises prots at

prices tp. Thus (tp) = tpy = t(p).
(3) Let y maximise prots at p, y

maximise prots at p

, and y

maximise prots at
p

. Then we have,
(p

) = p

= (tp + (1 t)p

)y

= tpy

+ (1 t)p

()
By the denition of prot maximisation, we know that,
tpy tpy

= t(p)
(1 t)p

(1 t)p

= (1 t)(p

).
Adding these two inequalities and using (*), we have,
(p

) t(p) + (1 t)(p

),
as required.
4. The continuity of (p) follows from the Theorem of the Maximum. Q.E.D.
4.2 Deriving Net Supply Functions from the Prot Function
If we are given the net supply function y(p), it is easy to calculate the prot function.
We just substitute into the denition of prots to nd (p) = py(p). Suppose that
17
instead we are given the prot function and are asked to nd the net supply functions.
How can that be done? It turns out that there is a very simple way to solve this
problem: just dierentiate the prot function. The proof that this works is the content
of the next proposition.
Proposition 4.2 (Hotellings Lemma): Let y
i
(p) be the rms net supply func-
tion for good i. Then
y
i
(p) =
(p)
p
i
for i = 1, . . . , n
assuming that the derivative exists and that p
i
> 0.
Proof: Suppose y

is a prot-maximising net output vector at prices p

. Then dene
the function
g(p) = (p) py

.
Clearly, the prot-maximising production plan at prices p will always be at least as
protable as the production plan y

. However, the plan y

will be a prot-maximising
plan at prices p

, so the function g reaches a minimum value of 0 at p

. The assumptions
on prices imply this is an interior minimum. The rst-order conditions for a minimum
then imply that
g(p

)
p
i
=
(p

)
p
i
y

i
= 0 for i = 1, . . . , n
Since this is true for all choices of p

, the proof is done. Q.E.D.

Remark: Again, we can prove this derivative property of the prot function by
applying the Envelope Theorem:
d(p)
dp
i
=
py(p

)
p
i

x=x(a)
= y
i
.
This expression says that the derivative of with respect to a price is given by the
partial derivative of f with respect to p
i
, holding x xed at the optimal choice. This is
the meaning of the vertical bar to the right of the derivative.
5. Cost Minimisation
An important implication of the rm choosing a prot-maximising production plan is
that there is no way to produce the same amounts of outputs at a lower total input
cost. Thus, cost minimisation is a necessary condition for prot maximisation. This
18
observation motives an independent study of the rms cost minimisation. The problem
is of interest for several reasons. First, it leads to a number of results and constructions
that are technically very useful. Second, as long as the rm is a price taker in its input
market, the results owing from the cost minimisation continue to be valid whether or
not the output market is competitive and so whether or not the rm takes the output
price as given as. Third, when the production set exhibits nondecreasing returns to
scale, the cost function and optimising vectors of the cost minimisation problem, which
keep the levels of outputs xed, are better behaved than the prot function.
To be concrete, we will focus our analysis on the single-output case. We will assume
throughout that rms are perfectly competitive in their input markets and therefore
they face xed prices. Let w = (w
1
, w
2
, . . . , w
n
) 0 be a vector of prevailing market
prices at which the rm can buy inputs x = (x
1
, x
2
, . . . , x
n
).
5.1 First-Order Conditions of Cost Minimisation
Let us consider the problem of nding a cost-minimising way to produce a given level
of output:
min
x
wx
s.t. f(x) y.
We analyse this constrained minimisation problem using the Lagrangian function:
L(, x) = wx (f(x) y)
where the production function f is assumed to be dierentiable and is the Lagrange
multiplier. The rst-order conditions characterising an interior solution x

are:
w
i

f(x

)
x
i
= 0; i = 1, . . . , n (3.5)
f(x

) = y (5.2)
or in vector notation, the condition can be written as
w = Df(x

).
We can interpret these rst-order conditions by dividing the jth condition by the ith
condition to get
w
i
w
j
=
f(x

)
x
i
f(x

)
x
j
i, j = 1, . . . , n (5.3)
19
which means the marginal rate of technical substitution of factor i for factor j equals
the economic rate of substitution factor i for factor j at the cost minimising input
bundle.
This rst-order condition can also be represented graphically as shown:
[see Cost Minimisation graph]
where the curved lines represent iso-quants and the straight lines represent constant
cost curves. At a point that minimises costs, the isoquant must be tangent to the
constant cost line. When y is xed, the problem of the rm is to nd a cost-minimising
point on a given isoquant. It is clear that such a point will be characterised by the
tangency condition that the slope of the constant cost curve must be equal to the slope of
the isoquant. Again, the conditions are valid only for interior operating positions: they
must be modied if a cost minimisation point occurs on the boundary. The appropriate
conditions turn out to be

f(x

)
x
i
w
i
0 with equality if x
i
> 0, i = 1, 2, . . . , n (5.4)
Remark: It is known that a continuous function achieves a minimum and a max-
imum value on a closed and bounded set. The objective function wx is certainly a
continuous function and the set V (y) is a closed set by hypothesis. All that we need to
establish is that we can restrict our attention to a bounded subset of V (y). But this is
easy. Just pick an arbitrary value of x, say x

. Clearly the minimal cost factor bundle

must have a cost less than wx

. Hence, we can restrict our attention to the subset

x V (y) : wx wx

, which will certainly be a bounded subset, as long as w 0.

Thus the cost minimising input bundle always exists.
5.2 Suciency of First-Order Conditions for Cost Minimisation
Again, like consumers constrained optimisation problem, the above rst-order condi-
tions are merely necessary conditions for a local optimum. However, these necessary
rst order conditions are in fact sucient for a global optimum when a production
function is quasi-concave, which is formerly stated in the following proposition.
Proposition 5.1: Suppose that f(x) : '
n
+
' is dierentiable and quasi-concave
on '
n
+
and w > 0. If (x, ) > 0 satises the rst-order conditions given in (3.5) and
(5.2), then x solves the rms cost minimisation problem at prices w.
Proof: Since f(x) : '
n
+
' + is dierentiable and quasi-concave, the input requirement
set V (y) = x : f(x) y is a convex and closed set. Further, the object function wx
20
is convex and continuous, then by the Kuhn-Tucker theorem, the rst-order conditions
are sucient for the constrained minimisation problem. Similarly, the strict quasi-
concavity of f can be checked by verifying if the naturally ordered principal minors of
the bordered Hessian alternative in sign, i.e.,

0 f
1
f
2
f
1
f
11
f
12
f
2
f
21
f
22

> 0,

0 f
1
f
2
f
3
f
1
f
11
f
12
f
13
f
2
f
21
f
22
f
23
f
3
f
31
f
32
f
33

> 0,
and so on, where f
i
=
f
x
i
and f
ij
=

2
f
x
i
x
j
. For each choice of w and y there will
be some choice of x

that minimises the cost of producing y units of output. We will

call the function that gives us this optimal choice the conditional input demand
function and write it as x(w, y). Note that conditional factor demands depend on
the level of output produced as well as on the factor prices. The cost function is the
minimal cost at the factor prices w and output level y; that is, c(w, y) = wx(w, y).
Example: (Cost Function for the Cobb-Douglas Technology) Consider the cost min-
imisation problem
c(w, y) = min
x
1
,x
2
w
1
x
1
+w
2
x
2
s.t. Ax
a
1
x
b
2
= y.
Solving the constraint for x
2
, we see that this problem is equivalent to
min
x
1
w
1
x
1
+w
2
A
1/b
y
1/b
x

a
b
1
.
The rst-order condition is
w
1

a
b
w
2
A
1/b
y
1/b
x

a+b
b
1
= 0,
which gives us the conditional input demand function for factor 1:
x
1
(w
1
, w
2
, y) = A

1
a+b
_
aw
2
bw
1
_ b
a+b
y
1
a+b
.
The other conditional input demand function is
x
2
(w
1
, w
2
, y) = A

1
a+b
_
aw
2
bw
1
_

a
a+b
y
1
a+b
.
21
The cost function is thus
c(w
1
, w
2
, y) = w
1
x
1
(w
1
, w
2
, y) +w
2
x
2
(w
1
, w
2
, y)
= A

1
a+b
_
_
a
b
_ b
a+b
_
a
b
_

a
a+b
_
w
a
a+b
1
w
b
a+b
2
y
1
a+b
.
and when A = 1 and a + b = 1 (the latter being the constant returns to scale
assumption), this cost function reduces to,
c(w
1
, w
2
, y) = Kw
a
1
w
1a
2
y,
where K = a
a
(1 a)
a1
.
Example: (The Cost function for the CES Technology) Suppose that f(x
1
, x
2
) =
(x

1
+x

2
)
1/
. What is the associated cost function? The cost minimisation problem
is
min
x
1
,x
2
w
1
x
1
+w
2
x
2
s.t. x

1
+x

2
= y

The rst-order conditions are

w
1
x
1
1
= 0
w
2
x
1
2
= 0
x

1
+x

2
= y

.
Solving the rst two equations for x

1
and x

2
, we have
x

1
= w

1
1
()

1
(5.5)
x

2
= w

1
2
()

1
(5.6)
Substitute this into the production function to nd
()

1
_
w

1
1
+w

1
2
_
= y

Solve this for ()

1
and substitute into equations (5.5) and (5.6). This gives
us the conditional input demand functions
x
1
(w
1
, w
2
, y) = w
1
1
1
_
w

1
1
+w

1
2
_
1/
y
x
2
(w
1
, w
2
, y) = w
1
1
2
_
w

1
1
+w

1
2
_
1/
y.
22
Substituting these functions into the denition of the cost function yields
c(w
1
, w
2
, y) = w
1
x
1
(w
1
, w
2
, y) +w
2
x
2
(w
1
, w
2
, y)
= y
_
w

1
1
+w

1
2
_ _
w

1
1
+w

1
2
_
1/
= y
_
w

1
1
+w

1
2
_

.
This expression looks a bit nicer if we set r = /( 1) and write
c(w
1
, w
2
, y) = y[w
r
1
+w
r
2
]
1/r
.
Note that this cost function has the same form as the original CES production
function with r replacing . In the general case where
f(x
1
, x
2
) = [(a
1
x
1
)

+ (a
2
x
2
)

]
1/
,
similar computations can be done to show that
c(w
1
, w
2
, y) = [(w
1
/a
1
)
r
+ (w
2
/a
2
)
r
]
1/r
y.
Example: (The Cost function for the Leontief Technology) Suppose f(x
1
, x
2
) =
minax
1
, bx
2
. Since we know that the rm will not waste any input with a
positive price, the rm must operate at a point where y = ax
1
= bx
2
. Hence, if
the rm wants to produce y units of output, it must use y/a units of good 1 and
y/b units of good 2 no matter what the input prices are. Hence, the cost function
is given by,
c(w
1
, w
2
, y) =
w
1
y
a
+
w
2
y
b
= y
_
w
1
a
+
w
2
b
_
Example: (The Cost Function for the Linear Technology) Suppose that f(x
1
, x
2
) =
ax
1
+ bx
2
, so that factors 1 and 2 are perfect substitutes. What will the cost
function look like? Since the two goods are perfect substitutes, the rm will use
whichever is cheaper. Hence, the cost function will have the form
c(w
1
, w
2
, y) = minw
1
/a, w
2
/by.
In this case the answer to the cost-minimisation problem typically involves a boundary
solution: one of the two factors will be used in a zero amount. It is easy to see the
23
answer to this particular problem by comparing the relative steepness of the isocost line
and isoquant curve. If a
1
/a
2
< w
1
/w
2
then the rm only uses x
2
and the cost function
is given by
c(w
1
, w
2
, y) = w
2
x
2
= w
2
y
a
2
.
If a
1
/a
2
> w
1
/w
2
then the rm only uses x
1
and the cost function is given by
c(w
1
, w
2
, y) = w
1
x
1
= w
1
y
a
1
.
6. Cost Functions
The cost function measures the minimum cost of producing a given level of output for
some xed factor prices. As such it summarises information about the technological
choices available to the rms. It turns out that the behaviour of the cost function
can tell us a lot about the nature of the rms technology. In the following we will
rst investigate the behaviour of the cost function c(w, y) with respect to its price and
quantity arguments. We then dene a few related functions, namely the average and
the marginal cost functions.
6.1 Properties of Cost Functions
You may have noticed some similarities here with consumer theory. These similarities
are in fact exact when one compares the cost function with the expenditure function.
Indeed, consider their denitions.
(1) Expenditure Function: e(p, u) min
x
n
+
px s.t. u(x) u
(2) Cost Function: c(w, y) min
x
n
+
wx s.t. f(x) = y
Mathematically, the two optimisation problems are identical. Consequently, for every
theorem we proved about expenditure functions, there is an equivalent theorem for cost
functions. We shall state these results here, but we do not need to prove them. Their
proofs are identical to those given for the expenditure function.
Proposition 6.1 (Properties of the Cost Function): Suppose the production
function f is continuous and strictly increasing. Then the cost function has the following
properties:
24
(1) c(w, y) is nondecreasing in w.
(2) c(w, y) is homogeneous of degree 1 in w.
(3) c(w, y) is concave in w.
(4) c(w, y) is continuous in w, for w > 0.
(5) For all w > 0, c(w, y) is strictly increasing in y.
(6) Shephards lemma: If x(w, y) is the cost-minimising bundle necessary to produce
production level y at prices w, then x
i
(w, y) =
c(w,y)
w
i
for i = 1, . . . , n assuming the
derivative exists and that x
i
> 0.
6.2 Properties of Conditional Input Demand
As the solution to the rms cost-minimisation problem, the conditional input demand
functions possess certain general properties. These are analogous to the properties of
Hicksian compensation demands, so once again it is not necessary to repeat the proof.
Proposition 6.2 (Negative Semi-Denite Substitution Matrix): The ma-
trix of substitution terms, x
j
(w, y)/w
i
, is negative semi-denite. Again since the
substitution matrix is negative semi-denite, thus it is symmetric and has non-positive
diagonal terms. We then have in particular:
Proposition 6.3 (Symmetric Substitution Terms): The matrix of substitution
terms is symmetric, i.e.,
x
j
(w, y)
w
i
=

2
c(w, y)
w
j
w
i
=

2
c(w, y)
w
i
w
j
=
x
i
(w, y)
w
j
.
Proposition 6.4 (Negative Own-Substitution Terms): The compensated own-
price eect is non-positive; that is, the input demand curves slope downward:
x
i
(w, y)
w
i
=

2
c(w, y)
w
2
i
0.
Remark: Using the cost function, we can restate the rms prot maximisation
problem as
max
y0
py c(w, y). (6.1)
The necessary rst-order condition for y

to be prot-maximising is then
p
c(w, y

)
y
0, with equality if y

> 0. (6.2)
25
In other words, at an interior optimum (i.e., y

> 0), price equals marginal cost. If

c(w, y) is convex in y, then the rst-order condition (6.2) is also sucient for y

to be
the rms optimal output level.
6.3 Average and Marginal Costs
Let us consider the structure of the cost function. Note that the cost function can
always be expressed simply as the value of the conditional factor demands.
c(w, y) wx(w, y)
In the short run, some of the factors of production are xed at predetermined levels.
Let x
f
be the vector of xed factors, x
v
, the vector of variable factors, and break up
w into w = (w
v
, w
f
), the vectors of prices of the variable and the xed factors. The
short-run conditional factor demand functions will generally depend on x
f
, so we write
them as x
v
(w, y, x
f
). Then the short-run cost function can be written as
c(w, y, x
f
) = w
v
x
v
(w, y, x
f
) +w
f
x
f
.
The term w
v
x
v
(w, y, x
f
) is called short-run variable cost (SVC), and the term w
f
x
f
is the xed cost (FC). We can dene various derived cost concepts from these basic
units:
short-run total cost = STC = w
v
x
v
(w, y, x
f
) +w
f
x
f
short-run average cost = SAC =
c(w, y, x
f
)
y
short-run average variable cost = SAV C =
w
v
x
v
(w, y, x
f
)
y
short-run average xed cost = SAFC =
w
f
x
f
y
short-run marginal cost = SMC =
c(w, y, x
f
)
y
.
When all factors are variable, the rm will optimise in the choice of x
f
. Hence, the
long-run cost function only depends on the factor prices and the level of output as
indicated earlier. We can express this long-run function in terms of the short-run cost
function in the following way. Let x
f
(w, y) be the optimal choice of the xed factors,
and let x
v
(w, y) = x
v
(w, y, x
f
(w, y)) be the long-run optimal choice of the variable
factors. Then the long-run cost function can be written as
c(w, y) = w
v
x
v
(w, y) +w
f
x
f
(w, y) = c(w, y, x
f
(w, y)).
26
Similarly, we can dene the long-run average and marginal cost functions:
long-run average cost = LAC =
c(w, y)
y
long-run marginal cost = LMC =
c(w, y)
y
.
Notice that long-run average cost equals long-run average variable cost since all
costs are variable in the long-run; long-run xed costs are zero for the same reason.
The problem that is being looked at, that is the time period that is being looked at
and the factors involved, determines what factors are xed and what are variable and
can adjust over it.
Example: (The short-run Cobb-Douglas cost functions) Suppose the second factor in
a Cobb-Douglas technology is restricted to operate at a level k. Then the cost-
minimising problem is
min
x
1
,x
2
w
1
x
1
+w
2
k s.t. y = x
a
1
k
1a
.
Solving the constraint for x
1
as a function of y and k gives
x
1
= (yk
a1
)
1/a
.
Thus,
c(w
1
, w
2
, y, k) = w
1
(yk
a1
)
1/a
+w
2
k.
The following variations can also be calculated:
short-run average cost = w
1
_
y
k
_
1a
a
+
w
2
k
y
short-run average variable cost = w
1
_
y
k
_
1a
a
short-run average xed cost =
w
2
k
y
short-run marginal cost =
w
1
a
_
y
k
_
1a
a
Example: (Constant Returns to Scale and the Cost Function) If the production func-
tion exhibits constant returns to scale, then it is intuitively clear that the cost
function should exhibit costs that are linear in the level of output: if you want
to produce twice as much output it will cost you twice as much. This intuition is
veried in the following proposition:
Proposition 6.5 (Constant Returns to Scale): If the production function ex-
hibits constant returns to scale, the cost function may be written as c(w, y) = yc(w, 1).
27
Proof: Let x

be a cheapest way to produce one unit of output at prices w so that

c(w, 1) = wx

. We want to show that c(w, y) = wyx

= yc(w, 1). Notice rst that

yx

is feasible to produce y since the technology is constant returns to scale. Suppose

that it does not minimise cost; instead let x

be the cost-minimising bundle to produce

y at prices w so that wx

< wyx

. Then wx

/y < wx

and x

/y can produce 1 since

the technology is constant returns to scale. This contradicts the denition of x

. Thus,
if the technology exhibits constant returns to scale, then the average cost, the average
variable cost, and the marginal cost functions are all the same. Q.E.D.
6.4 The Geometry of Costs
Let us rst examine the short-run cost curves. In this case, we will write the cost
function simply as c(y), which has two components: xed costs and variable costs. We
can therefore write short-run average cost as
SAC =
c(w, y, x
f
)
y
=
w
f
x
f
y
+
w
v
x
v
(w, y, x
f
)
y
= SAFC + SAV C.
In most applications short-run xed factors include machinery, buildings, and other
types of capital, whereas variables factors are labour and raw materials.
As we increase output, average variable costs may initially decrease if there is some
initial region of economies of scale. However, it seems reasonable to suppose that the
variable factors required will eventually increase by the low of diminishing marginal
returns, as depicted in the following graph.
[see Average Variable, Average Fixed, and Average Cost Curves graph]
Average xed costs must of course decrease with output, as indicated in the graph.
Adding together the average variable cost curve and the average xed costs gives us
the U-shaped average cost curve in the graph. The initial decrease in average costs is
due to the decrease in average xed costs; the eventual increase in average costs is due
to the increase in average variable costs. The level of output at which the average cost
of production is minimised is sometimes known as the minimal ecient scale.
[see Relationship Between MC, AC, & TC graph]
In the long run all costs are variable costs and if the rms can replicate its production
processes then increasing average costs seems unreasonable. Instead the long-run av-
erage cost curve would be constant or even decreasing. On the other hand if there are
xed factors even in the long-run then the long-run average cost curve will be U-shaped
and thus eventually increasing.
28
Let us now consider the marginal cost curve. What is its relationship to the average
cost curve? Let y

denote the point of minimum average cost so that to the left of y

we know that average costs are declining or that for y y

,
d
dy
_
c(y)
y
_
0. (6.3)
Taking the derivative gives us,
yc

(y) c(y)
y
2
=
1
y
_
c

(y)
c(y)
y
_
. (6.4)
But, (6.3) and (6.4) together can only be true together if and only if
c

(y)
c(y)
y
0 ( 0)
This has an easy interpretation. The average variable cost curve is decreasing when
the marginal cost curve lies below the average variable cost curve, and it is increasing
when the marginal cost curve lies above the average variable cost curve. It follows that
average cost reach its minimum at y

when the marginal cost curve passes through the

average variable cost curve, i.e.,
c

(y

) =
c(y

)
y

.
Remark: All of the analysis just discussed holds in both the long and the short run.
However, if production exhibits constant returns to scale in the long run, so that the
cost function is linear in the level of output, then average cost, average variable cost,
and marginal cost are all equal to each other, which makes most of the relationships
just described rather trivial.
6.5 Long-Run and Short-Run Cost Curves
Consider now the relationship between the long-run cost curves and the short-run cost
curves. It is clear that the long-run cost curve must never lie above any short-run cost
curve, since the short-run cost minimisation problem is just a constrained version of
the long-run cost minimisation problem.
Let us write the long-run cost function as c(y) = c(y, z(y)). Here we have omitted
the factor prices since they are assumed xed, and we let z(y) be the cost-minimising
demand for a single xed factor. Let y

be some given level of output, and let z

= z(y

)
29
be the associated long-run demand for the xed factor. The short-run cost, c(y, z

),
must be at least as great as the long-run cost, c(y, z(y)), for all levels of output, and
the short-run cost will equal the long-run cost at output y

, so c(y

, z

) = c(y

, z(y

)).
Hence, the long and the short-run cost curves must be tangent at y

.
This is just a geometric restatement of the envelope theorem. The slope of the long-run
cost curve at y

is
dc(y

, z(y

))
dy
=
c(y

, z

)
y
+
c(y

, z

)
z
z(y

)
y
.
But since z

is the optimal choice of the xed factors at the output level y

, we must
have
c(y

, z

)
z
= 0.
Thus, long-run marginal costs at y

equal short-run marginal costs at (y

, z

).
Finally, note that if the long- and short-run cost curves are tangent, the long- and short-
run average cost curves must also be tangent. A typical conguration is illustrated in
the following gure:
[see Long-run and Short-run Average Cost Curves graph]
Note that the long-run and the short-run average cost curves must be tangent which
implies that the long-run and short-run marginal cost must equal.
7. Duality in Production
In the last section we investigated the properties of the cost function. Given any
technology, it is straightforward, at least in principle, to derive its cost function: we
simply solve the cost minimisation problem. In this section we show that this process
can be reversed. Given a cost function we can solve for a technology that could have
generated that cost function. This means that the cost function contains essentially
the same information that the production function contains. Any concept dened in
terms of the properties of the production function has a dual denition in terms of
the properties of the cost function and vice versa. This general observation is known
as the principle of duality.
7.1 Recovering a Production Set from a Cost Function
30
Given data (w
t
, x
t
, y
t
), dene V O(y) as an outer bound to the true input requirement
set V (y):
V O(y) = x : w
t
x w
t
x
t
t s.t. y
t
y.
It is straightforward to verify that V O(y) is a closed, monotonic, and convex technology.
Furthermore, it contains any technology that could have generated the data (w
t
, x
t
, y
t
)
for t = 1, . . . , T.
If we observe choices for many dierent factor prices, it seems that V O(y) should
approach the true input requirement set in some sense. To make this precise, let the
factor prices vary over all possible price vectors w 0. Then the natural generalisation
of V O becomes
V

(y) = x : wx wx(w, y) = c(w, y) w 0.

What is the relationship between V

(y) and the true input requirement set V (y)? Of

course, V

(y) clearly contain V (y). In general, V

(y) will strictly contain V (y). For

example, in the LHS graph in following gure
[see Relationship Between V (y) and V

(y) graph]
we see that the shaded area cannot be ruled out of V

(y) since the points in this area

satisfy the condition that wx c(w, y). The same is true as shown in the RHS of the
gure. The cost function can only contain information about the economically relevant
sections of V (y), namely, those factor bundles that could actually be the solution to a
cost minimisation problem, i.e., that could actually be conditional factor demands.
However, suppose that our original technology is convex and monotonic. In this case
V

(y) will equal V (y): This is because, in the convex, monotonic case, each point on
the boundary of V (y) is a cost-minimising factor demand for some price vector w 0.
Thus, the set of points where wx c(w, y) w 0 will precisely describe the input
requirement set. More formally:
Proposition 7.1 (Equality of V (y) and V

(y))): Suppose V (y) is a closed,

convex, monotonic technology. Then V

(y) = V (y):
Proof: (Sketch) We already know that V

(y) contains V (y), so we only have to show

that if x is in V

(y) then x must be in V (y). Suppose that x is not an element of V (y).

Then since V (y) is a closed convex set satisfying the monotonicity hypothesis, we can
apply a version of the separating hyperplane theorem to nd a vector w

0 such that
w

x < w

z z V (y). Let z

be a point in V (y) that minimises cost at the prices

w

. Then in particular we have w

x < w

= c(w

, y). But then x cannot be in

V

(y), according to the denition of V

(y). Q.E.D.
31
This proposition shows that if the original technology is convex and monotonic, then
the cost function associated with the technology can be used to completely reconstruct
the original technology. If we know the minimal cost for every possible price vector, w,
then we know the entire set of technological choices open to the rm.
This is a reasonably satisfactory result in the case of convex and monotonic technologies,
but what about less well-behaved cases? Suppose we start with some technology V (y),
possibly non-convex. We nd its cost function c(w, y) and then generate V

(y). We
know from the above results that V

(y) will not necessarily be equal to V (y), unless

V (y) happens to have the convexity and monotonicity properties. However, suppose
we dene
c(w, y) = min wx s.t. x V

(y)
What is the relationship between c

(w, y) and c(w, y)?

Proposition 7.2 (Equality of c(w, y) and c

(w, y))): It follows from the denition

of the functions that c

(w, y) = c(w, y).

Proof: It is easy to see that c

(w, y) c(w, y); since V

(y) always contains V (y), the

minimal cost bundle in V

(y) must be at least as small as the minimal cost bundle in

V (y). Suppose that for some prices w

the cost-minimising bundle x

(y) has the

property that w

= c

(w

, y) < c(w

, y). But this cant happen, since by denition

of V

(y), w

c(w

, y). Q.E.D.
This proposition shows that the cost function for the technology V (y) is the same as
the cost function for its convexication V

(y). In this sense, the assumption of convex

input requirement sets is not very restrictive from an economic point of view.
Let us summarise the discussion to date:
(1) Given a cost function we can dene an input requirement set V

(y).
(2) If the original technology is convex and monotonic, the constructed technology will
be identical with the original technology.
(3) If the original technology is non-convex or monotonic, the constructed input re-
quirement will be a convexied, monotonised version of the original set, and, most
importantly, the constructed technology will have the same cost function as the origi-
nal technology.
32
In conclusion, the cost function of a rm summarises all of the economically relevant
aspects of its technology. Since the cost function tells us all of the economically rele-
vant information about the underlying technology, we can then try to interpret various
restrictions on costs in the form of restrictions on technology. We will now consider a
specic example of this. We know that if a technology exhibits constant returns to scale,
then the cost function has the form c(w)y. Here we show that the reverse implication
is also true.
Proposition 7.3 (Constant Returns to Scale): Let V (y) be convex and mono-
tonic; then if c(w, y) can be written as yc(w), V (y) must exhibit constant returns to
scale.
Proof: Using convexity, monotonicity, and the assumed form of the cost function as-
sumptions, we know that
V (y) = V

(y) = x : wx yc(w) w 0.
We want to show that, if x V

(y), then tx V

(ty). If x V

(y), we know
that wx yc(w) w 0. Multiplying both sides of this equation by t we get:
wtx tyc(w) w 0. But this says tx V

(ty). Q.E.D.
7.2 Sucient Conditions for Cost Functions
We have seen in the last section that all cost functions are nondecreasing, homogeneous,
concave, continuous functions of prices. The question arises: suppose that you are given
a nondecreasing, homogeneous, concave, continuous function of prices is it necessarily
the cost function of some technology? The answer is yes, and the following proposition
shows how to construct such a technology.
Proposition 7.4 (When (w, y) is a cost function): Let (w, y) be a dieren-
tiable function satisfying
(1) (tw, y) = t(w, y) t 0;
(2) (w, y) 0 for w 0 and y 0;
(3) (w

, y) (w, y) for w

w;
(4) (w, y) is concave in w.
Then (w, y) is the cost function for the technology dened by
V

(y) = x 0 : wx (w, y), w 0.

33
Proof: Given a w 0 dene
x(w, y) =
_
(w, y)
w
1
, ,
(w, y)
w
n
_
and note that since (w, y) is homogeneous of degree 1 in w, Eulers law implies that
(w, y) can be written as
(w, y) =
n

i=1
w
i
(w, y)
w
i
= wx(w, y).
Note that the monotonicity of (w, y) implies x(w, y) 0. What we need to show is
that for any given w

0, x(w

, y) actually minimises w

x over all x V

(y):
(w

, y) = w

x(w

, y) w

x x V

(y).
First, we show that x(w

, y) is feasible; that is, x(w

, y) V

(y). By the concavity of

(w, y) in w we have
(w

, y) (w, y) +D(w, y)(w

w) w 0.
Using Eulers law as above, this reduces to
(w

, y) w

x(w, y) w 0.
It follows from the denition of V

(y), that x(w

, y) V

(y).
Next we show that x(w, y) actually minimises wx over all x V

(y). If x V

(y),
then by denition it must satisfy
wx (w, y).
But by Eulers law,
(w, y) = wx(w, y).
The above two expressions imply
wx wx(w, y) x V

(y)
as required. Q.E.D.
34
7.3 The Integrability of Conditional Factor Demand Functions
The proposition proved in the last subsection raises an interesting question. Suppose
you are given a set of functions, g
i
(w, y), that satisfy the properties of conditional factor
demand functions described in the previous sections, namely, that they are homogeneous
of degree 0 in prices and that
_
g
i
(w, y)
w
j
_
is a symmetric negative semi-denite matrix. Are these functions necessarily factor
demand functions for some technology?
Let us try to apply the above proposition. First, we construct a candidate for a cost
function:
(w, y) =
n

i=1
w
i
g
i
(w, y).
Next, we check whether it satises the properties required for the proposition just
proved.
1) Is (w, y) homogeneous of degree 1 in w? To check this we look at (tw, y) =

i
tw
i
g
i
(tw, y). Since the functions g
i
(w, y) are by assumption homogeneous of degree
0, g
i
(tw, y) = g
i
(w, y) so that
(tw, y) = t
n

i=1
wg
i
(w, y) = t(w, y).
2) Is (w, y) 0 for w 0? Since g
i
(w, y) = 0, the answer is clearly yes.
3) Is (w, y) nondecreasing in w
i
? Using the product rule, we compute
(w, y)
w
i
= g
i
(w, y) +
n

j=1
w
j
g
j
(w, y)
w
i
= g
i
(w, y) +
n

j=1
w
j
g
i
(w, y)
w
j
.
Since g
i
(w, y) is homogeneous of degree 0, the last term vanishes and g
i
(w, y) is clearly
greater than or equal to 0.
4) Finally is (w, y) concave in w? To check this we dierentiate (w, y) twice to get
_

2

w
i
w
j
_
=
_
g
i
(w, y)
w
j
_
.
For concavity we want these matrices to be symmetric and negative semi-denite, which
they are by hypothesis.
Hence, the proposition proved in the last subsection applies and there is a technology
V

(y) that yields g

i
(w, y) as its conditional factor demands. This means that the
35
properties of homogeneity and negative semi-deniteness form a complete list of the
restrictions on demand functions imposed by the model of cost-minimising behaviour.
Example: (Recovering Production from a Cost Function) Suppose we are given a
specic cost function c(w, y) = yw
a
1
w
1a
2
. How can we solve for its associated
technology? According to the derivative property
x
1
(w, y) = ayw
a1
1
w
a1
2
= ay
_
w
2
w
1
_
1a
x
2
(w, y) = (1 a)yw
a
1
w
a
2
= (1 a)y
_
w
2
w
1
_
a
.
We want to eliminate w
2
/w
1
from these two equations and get an equation for y
in terms of x
1
and x
2
. Rearranging each equation gives
w
2
w
1
=
_
x
1
ay
_ 1
1a
w
2
w
1
=
_
x
2
(1 a)y
_

1
a
.
Setting these equal to each other and raising both sides to the a(1 a) power,
x
a
1
a
a
y
a
=
x
1a
2
(1 a)
1a
y
1a
,
or,
[a
a
(1 a)
1a
]y = x
a
1
x
1a
2
.
This is just the Cobb-Douglas technology.
The same results hold for prot functions and unconditional demand supply functions.
If the prot function obeys the restrictions as listed before, or the demand and supply
functions obey the restrictions as listed before, then there must exist a technology that
generates the prot function or the demand and supply functions.
36
References
Jehle, G. and P. Reny. (1998). Advanced Microeconomic Theory. Reading, Mass.: Addison-Wesley.
[Chapter 5]
Kreps, D. (1990). A Course in Microeconomic Theory. Princeton: Princeton University Press. [Ap-
pendix 1]
Mas-Collel A., M. Whinston, and J. Green (1995). Microeconomic Theory. Oxford: Oxford University
Press. [Chapter 5]
Tian, G. (2005). Lecture Notes: Microeconomic Theory. Unpublished manuscript. Department of
Economics, Texas A&M University, College Station, Texas.
Varian H. (1992). Microeconomic Theory. 3rd ed. New York: Norton. [Chapters 1-6]
37
PRODUCER THEORY BASICS
y
i
j
= of good j as input.
y
o
j
= of good j as output.
y
j
= y
o
j
y
i
j
= net output of good j.
y R
L
= production plan.
Y '
L
= production possibilities set.
Y (z) = short-run production possibilities set (z '
L
).
V (y) = x '
L
+
: (y, x) Y .
Q(y) = x '
n
+
: x V (y) and x / V (y

) y

> y.
Y (

k) = (y, l, k) Y : k =

k.
f(x) = y ' : y is the maximum output associated with x Y
Example: (Cobb-Douglas Technology) Let 0 < < 1.
Y = (y, x
1
, x
2
) '
3
: y x

1
x
1
2

V (y) = (x
1
, x
2
) '
2
+
: y x

1
x
1
2

Q(y) = (x
1
, x
2
) '
2
+
: y = x

1
x
1
2

Y (z) = (y, x
1
, x
2
) '
3
: y x

1
x
1
2
, x
2
= z
T(y, x
1
, x
2
) = y x

1
x
1
2
f(x
1
, x
2
) = x

1
x
1
2
.
Example: (Leontief Technology) Let a, b > 0.
Y = (y, x
1
, x
2
) '
3
: y min(ax
1
, bx
2
)
V (y) = (x
1
, x
2
) '
2
+
: y min(ax
1
, bx
2
)
Q(y) = (x
1
, x
2
) '
2
+
: y = min(ax
1
, bx
2
)
T(y, x
1
, x
2
) = y min(ax
1
, bx
2
)
f(x
1
, x
2
) = min(ax
1
, bx
2
).
COMMON PROPERTIES OF PRODUCTION SETS
Axiom (Possibility of Inaction): 0 Y .
Axiom (Closeness): Y is closed.
Axiom (Free Disposal): If y Y implies that y

Y y

y, then Y satises
the free disposal or monotonicity property.
Axiom (Monotonicity): If x V (y) and x

x, then x

V (y).
Axiom (Irreversibility): Y Y = 0.
Axiom (Convexity): Y is convex if whenever y, y

Y ty+(1t)y

Y t s.t. 0
t 1.
Axiom (Strict Convexity): y is strictly convex if y Y and y

Y , then
ty + (1 t)y

intY 0 < t < 1.

Axiom (Convexity of Input Requirement Set): If x, x

V (y), then tx+(1

t)x

V (y) 0 t 1.
Proposition (2.1): Convex Production Set Implies Convex Input Requirement Set
If the production set Y is a convex set, then the associated input requirement set, V (y),
is a convex set.
Proof: If Y is a convex set then it follows that for any x and x

s.t. (y, x), (y, x

) Y
for 0 t 1, we must have (ty +(1 t)y, tx(1 t)x

) Y . This is simply requiring

that (y, (tx+(1t)x

)) Y . It follows that if x, x

V (y), tx+(1t)x

V (y) which
shows that V (y) is convex. Q.E.D.
Proposition (2.2): V (y) is a convex set i the production function f(x) is a
quasi-concave function.
PROPERTIES OF PRODUCTION PROCESSES
Denition ((Global) Returns to Scale): A production function f(x) is said to
exhibit:
(1) constant returns to scale if f(tx) = tf(x) t 0.
(2) decreasing returns to scale if f(tx) < tf(x) t > 1.
(3) increasing returns to scale if f(tx) > tf(x) t > 1.
Denition (Elasticity of Scale): The elasticity of scale is given by
e(x) =
dy(t)
y(t)
dt
t
,
evaluated at t = 1. Rearranging this expression, we have
e(x) =
dy(t)
dt
t
y
[t = 1 =
df(tx)
dt
t
f(tx)
[t = 1
Denition (Marginal Rate of Technical Substitution): The MRTS is given
by,
0 =
f
x
1
dx
1
+
f
x
2
dx
2
which can be solved for
dx
2
dx
1
=
f
x
1
f
x
2

MP
x
1
MP
x
2
.
Example: (MRTS for a Cobb-Douglas Technology) Given that f(x
1
, x
2
) = x

1
x
1
2
,
f(x)
x
1
= x
1
1
x
1
2
=
_
x
2
x
1
_
1
f(x)
x
2
= (1 )x

1
x

2
= (1 )
_
x
1
x
2
_

.
It follows that,
x
2
(x
1
)
x
1
=
f
x
1
f
x
2
=

1
x
2
x
1
.
Denition (The Elasticity of Substitution): The elasticity of substitution is
given by
=
(x
2
/x
1
)
x
2
/x
1
MRTS
MRTS
If the percent change is very small, taking the limit of this expression as goes to zero,
the expression for becomes
=
MRTS
(x
2
/x
1
)
d(x
2
/x
1
)
dMRTS
=
d ln(x
2
/x
1
)
d ln [MRTS[
.
Example: (The Cobb-Douglas Production Function) We know,
MRTS =

1
x
2
x
1
,
or,
x
2
x
1
=
1

MRTS
It follows that,
ln
x
2
x
1
= ln
1

+ ln [MRTS[.
This in turn implies,
=
d ln(x
2
/x
1
)
d ln [MRTS[
= 1.
Example: (The CES Production Function) The CES production function has the
form,
y = [a
1
x

1
+a
2
x

2
]
1

.
Note that the MRTS
MRTS =
_
x
1
x
2
_
1
,
so that
x
2
x
1
= [MRTS[
1
1
.
Taking logs, we see that
ln
x
2
x
1
=
1
1
ln [MRTS[.
Applying the denition of a using the logarithmic derivative,
=
d ln x
2
/x
1
d ln [MRTS[
=
1
1
.
PROFIT MAXIMISATION
The prot maximisation problem facing the rm:
max
a
1
,...,a
n
R(a
1
, . . . , a
n
) C(a
1
, . . . , a
n
)
The FOCs for interior optimal actions, a

= (a

1
, . . . , a

n
), are:
R(a

)
a
i
=
C(a

)
a
i
i = 1, . . . , n.
1. Producers Optimal Choice
(p) =max py (3.1)
s.t. y Y.
Denition (Short-Run or Restricted Prot Function):
(p, z) =max py
s.t. y Y (z).
If the rm produces only one output:
(p, w) = max pf(x) wx.
Proposition (3.1): Suppose Y is strictly convex. Then, for each given p '
L
+
, the
prot maximising production is unique provided it exists.
Proof: Suppose not. Let y and y

be two prot maximising production plans for p

'
L
+
. Then, we must have py = py

. Thus, by the strict convexity of Y , we have

ty + (1 t)y

intY 0 < t < 1. Therefore, k > 1 s.t.

k[ty + (1 t)y

] intY. (3.2)
Then k[tpy + (1 t)py

] = kpy > py which contradicts the fact that y is a prot

maximising production plan. Q.E.D.
2. Producers FOCs
The FOCs for the single output prot maximisation problem with interior solution are
p
f(x

)
x
i
= w
i
i = 1, . . . , n. (3.3)
or in vector notation,
pDf(x

) = w.
The FOCs also include:
p
f(x)
x
i
w
i
0 with equality if x
i
> 0. (3.4)
3. Suciency of Producers FOCs
The SOCs requires that the Hessian matrix
D
2
f(x

) =
_

2
f(x

)
x
i
x
j
_
must satisfy
hD
2
f(x

)h

0 h.
Proposition (3.2): Suppose that f(x) is dierentiable and concave on '
+
and
(p, w) > 0. If x > 0 satises the FOCs given in (3.4), then x is a (globally) prot
maximising production plan at prices (p, w).
Remark: The strict concavity of f(x) can be checked by verifying if the leading
principal minors of the Hessian must alternate in sign, i.e.,

f
11
f
12
f
21
f
22

> 0,

f
11
f
12
f
13
f
21
f
22
f
23
f
31
f
32
f
33

> 0,
and so on, where f
ij
=

2
f
x
i
x
j
.
Example: (The Prot Function for Cobb-Douglas Technology) The FOC is:
pax
a1
= w,
and the SOC reduces to
pa(a 1)x
a2
0.
which can only be satised when a 1.
When a < 1,
x(p, w) =
_
w
ap
_ 1
a1
.
The supply function is given by
y(p, w) = f(x(p, w)) =
_
w
ap
_ 1
a1
,
and the prot function is given by
(p, w) = py(p, w) wx(p, w) = w
_
1 1
a
__
w
ap
_ 1
a1
.
Properties of Net Supply Functions
Proposition 3.3: Net output functions y(p) are homogeneous of degree zero, i.e.,
y(tp) = y(p) t > 0.
Proposition 3.4 (Negative Deniteness of Substitution Matrix): Let y =
f(x) be a twice dierentiable and strictly concave single output production function,
and let x(p, w) be the input demand function. Then, the substitution matrix,
Dx(p, w) =
_
x
i
(p, w)
w
j
_
is symmetric negative denite.
Proof: Without loss of generality, we normalise p = 1. Then the rst-order conditions
for prot maximisation are
Df(x(w)) w 0
If we dierentiate with respect to w, we get
D
2
f(x(w))Dx(w)) I 0.
Solving this equation for the substitution matrix, we nd
Dx(w)) [D
2
f(x(w))]
1
Recall that the second-order condition for (strict) prot maximisation is that the Hes-
sian matrix is a symmetric negative denite matrix. It is a standard result of linear
algebra that the inverse of a symmetric negative denite matrix is a symmetric negative
denite matrix. Then, D
2
f(x(w)) is a symmetric negative denite matrix, and thus the
substitution matrix Dx(w)) is a symmetric negative denite matrix. This means that
the substitution matrix itself must be a symmetric, negative denite matrix. Q.E.D.
Remark: Note that since Dx(p, w) is symmetric, negative denite:
(1) x
i
/w
i
< 0, for i = 1, 2, . . . , n.
(2) x
i
/w
j
= x
j
/w
i
.
PROPERTIES OF THE PROFIT FUNCTION
Proposition 4.1 (Properties of the Prot Function):
(1) Nondecreasing in output prices, nonincreasing in input prices. If p

i
p
i
for all
outputs and p

j
p
j
, for all inputs, then (p

) = (p).
(2) Homogeneous of degree 1 in p. That is, (tp)) = t(p) t = 0.
(3) Convex in p. That is, let p

= tp + (1 t)p

for 0 t 1 then (p

)
t(p) + (1 t)(p

).
(4) Continuous in p. That is, the function (p) is continuous, at least when (p) is
well-dened and p
i
> 0 for i = 1, . . . , n.
Proof:
(1) Let y be a prot-maximising net output vector at p, so that (p) = py and let y

be a prot-maximising net output vector at p

so that (p

) = p

. Then by denition
of prot maximisation we have p

y. Since p

i
p
i
i for which y
i
0 and
p

i
p
i
i for which y
i
0, we also have p

y py. Putting these two inequalities

together, we have (p

) = p

py = (p), as required.
(2) Let y be a prot-maximising net output vector at p, so that py py

Y . It
follows that for t 0, tpy = tpy

Y . Hence y also maximises prots at prices

tp. Thus (tp) = tpy = t(p).
(3) Let y maximise prots at p, y

maximise prots at p

, and y

maximise prots at
p

. Then we have,
(p

) = (p

= t(p + (1 t)p

)y

= tpy

+ (1 t)p

()
By the denition of prot maximisation, we know that,
tpy tpy

= t(p)
(1 t)p

(1 t)p

= (1 t)(p

).
Adding these two inequalities and using (*), we have,
(p

) t(p) + (1 t)(p

),
as required.
4. The continuity of (p) follows from the Theorem of the Maximum. Q.E.D.
DERIVING NET SUPPLY FUNCTIONS
FROM THE PROFIT FUNCTION
Proposition 4.2 (Hotellings Lemma): Let y
i
(p) be the rms net supply func-
tion for good i. Then
y
i
(p) =
(p)
p
i
for i = 1, . . . , n
assuming that the derivative exists and that p
i
> 0.
Proof: Suppose y

is a prot-maximising net output vector at prices p

. Then dene
the function
g(p) = (p) py

.
Clearly, the prot-maximising production plan at prices p will always be at least as
protable as the production plan y

. However, the plan y

will be a prot-maximising
plan at prices p

, so the function g reaches a minimum value of 0 at p

. The assumptions
on prices imply this is an interior minimum. The rst-order conditions for a minimum
then imply that
g(p

)
p
i
=
(p

)
p
i
y

i
= 0 for i = 1, . . . , n
Since this is true for all choices of p

, the proof is done. Q.E.D.

Remark: Again, we can prove this derivative property of the prot function by
applying the Envelope Theorem:
d(p)
dp
i
=
py(p

)
p
i

x=x(a)
= y
i
.
COST MINIMISATION
w = (w
1
, w
2
, . . . , w
n
) 0.
x = (x
1
, x
2
, . . . , x
n
).
1. General Problem
min
x
wx
s.t. f(x) y.
or,
L(, x) = wx (f(x) y)
2. First-Order Conditions
w
i

f(x

)
x
i
= 0; i = 1, . . . , n (5.1)
f(x

) = y (5.2)
or
w = Df(x

).
Dividing the jth condition by the ith condition: to get
w
i
w
j
=
f(x

)
x
i
f(x

)
x
j
i, j = 1, . . . , n (5.3)
The appropriate conditions turn out to be

f(x

)
x
i
w
i
0 with equality if x
i
> 0, i = 1, 2, . . . , n (5.4)
3. Suciency of FOCs for Cost Minimisation
Proposition 5.1: Suppose that f(x) : '
n
+
' is dierentiable and quasi-concave
on '
n
+
and w > 0. If (x, ) > 0 satises the rst-order conditions given in (5.1) and
(5.2), then x solves the rms cost minimisation problem at prices w.
Proof: Since f(x) : '
n
+
' is dierentiable and quasi-concave, the input requirement
set V (y) = x : f(x) y is a convex and closed set. Further, the object function wx
is convex and continuous, then by the Kuhn-Tucker theorem, the rst-order conditions
are sucient for the constrained minimisation problem. Similarly, the strict quasi-
concavity of f can be checked by verifying if the naturally ordered principal minors of
the bordered Hessian alternative in sign, i.e.,

0 f
1
f
2
f
1
f
11
f
12
f
2
f
21
f
22

> 0,

0 f
1
f
2
f
3
f
1
f
11
f
12
f
13
f
2
f
21
f
22
f
23
f
3
f
31
f
32
f
33

> 0,
and so on, where f
i
=
f
x
i
and f
ij
=

2
f
x
i
x
j
. For each choice of w and y there will
be some choice of x

that minimises the cost of producing y units of output. We will

call the function that gives us this optimal choice the conditional input demand
function and write it as x(w, y). The cost function is the minimal cost at the factor
prices w and output level y; that is, c(w, y) = wx(w, y).
Example: (Cost Function for the Cobb-Douglas Technology)
c(w, y) = min
x
1
,x
2
w
1
x
1
+w
2
x
2
s.t. Ax
a
1
x
b
2
= y.
Solving for x
2
this is equivalent to
min
x
1
w
1
x
1
+w
2
A
1/b
y
1/b
x

a
b
1
.
The FOC is
w
1

a
b
w
2
A
1/b
y
1/b
x

a+b
b
1
= 0,
which gives us:
x
1
(w
1
, w
2
, y) = A

1
a+b
_
aw
2
bw
1
_ b
a+b
y
1
a+b
x
2
(w
1
, w
2
, y) = A

1
a+b
_
aw
2
bw
1
_

a
a+b
y
1
a+b
.
The cost function is thus
c(w
1
, w
2
, y) = w
1
x
1
(w
1
, w
2
, y) +w
2
x
2
(w
1
, w
2
, y)
= A

1
a+b
_
_
a
b
_ b
a+b
_
a
b
_

a
a+b
_
w
a
a+b
1
w
b
a+b
2
y
1
a+b
.
and when A = 1 and a +b = 1 this reduces to,
c(w
1
, w
2
, y) = Kw
a
1
w
1a
2
y,
where K = a
a
(1 a)
a1
.
Example: (The Cost Function for the Linear Technology)
c(w
1
, w
2
, y) = minw
1
/a, w
2
/by.
If a
1
/a
2
< w
1
/w
2
:
c(w
1
, w
2
, y) = w
2
x
2
= w
2
y
a
2
.
If a
1
/a
2
> w
1
/w
2
:
c(w
1
, w
2
, y) = w
1
x
1
= w
1
y
a
1
.
Example: (The Cost function for the CES Technology)
min
x
1
,x
2
w
1
x
1
+w
2
x
2
s.t. x

1
+x

2
= y

The FOCs are

w
1
x
1
1
= 0
w
2
x
1
2
= 0
x

1
+x

2
= y

.
Solving for x

1
and x

2
:
x

1
= w

1
1
()

1
(5.5)
x

2
= w

1
2
()

1
(5.6)
and substitute into the production function:
()

1
_
w

1
1
+w

1
2
_
= y

Solve for ()

1
and substitute into (5.5) and (5.6):
x
1
(w
1
, w
2
, y) = w
1
1
1
_
w

1
1
+w

1
2
_
1/
y
x
2
(w
1
, w
2
, y) = w
1
1
2
_
w

1
1
+w

1
2
_
1/
y.
Substituting these denition of the cost function:
c(w
1
, w
2
, y) = w
1
x
1
(w
1
, w
2
, y) +w
2
x
2
(w
1
, w
2
, y)
= y
_
w

1
1
+w

1
2
_ _
w

1
1
+w

1
2
_
1/
= y
_
w

1
1
+w

1
2
_

.
Set r = /( 1)
c(w
1
, w
2
, y) = y[w
r
1
+w
r
2
]
1/r
.
In the general case where
f(x
1
, x
2
) = [(a
1
x
1
)

+ (a
2
x
2
)

]
1/
,
c(w
1
, w
2
, y) = [(w
1
/a
1
)
r
+ (w
2
/a
2
)
r
]
1/r
y.
Example: (The Cost function for the Leontief Technology)
c(w
1
, w
2
, y) =
w
1
y
a
+
w
2
y
b
= y
_
w
1
a
+
w
2
b
_
PROPERTIES OF COST FUNCTIONS
(1) Expenditure Function: e(p, u) min
x
n
+
px s.t. u(x) u
(2) Cost Function: c(w, y) min
x
n
+
wx s.t. f(x) = y
Proposition 6.1 (Properties of the Cost Function): Suppose the production
function f is continuous and strictly increasing. Then the cost function has the following
properties:
(1) c(w, y) is nondecreasing in w.
(2) c(w, y) is homogeneous of degree 1 in w.
(3) c(w, y) is concave in w.
(4) c(w, y) is continuous in w, for w > 0.
(5) For all w > 0, c(w, y) is strictly increasing in y.
(6) Shephards lemma: If x(w, y) is the cost-minimising bundle necessary to produce
production level y at prices w, then x
i
(w, y) =
c(w,y)
w
i
for i = 1, . . . , n assuming the
derivative exists and that x
i
> 0.
Properties of Conditional Input Demand
Proposition 6.2 (Negative Semi-Denite Substitution Matrix): The ma-
trix of substitution terms, x
j
(w, y)/w
i
, is negative semi-denite. Again since the
substitution matrix is negative semi-denite, thus it is symmetric and has non-positive
diagonal terms.
Proposition 6.3 (Symmetric Substitution Terms): The matrix of substitution
terms is symmetric, i.e.,
x
j
(w, y)
w
i
=

2
c(w, y)
w
j
w
i
=

2
c(w, y)
w
i
w
j
=
x
i
(w, y)
w
j
.
Proposition 6.4 (Negative Own-Substitution Terms): The compensated
own-price eect is non-positive:
x
i
(w, y)
w
i
=

2
c(w, y)
w
2
i
0.
Remark: Using the cost function, we can restate the rms prot maximisation
problem as
max
y0
py c(w, y). (3.11)
The FOC is then
p
c(w, y

)
y
0, with equality if y

> 0. (3.12)
AVERAGE AND MARGINAL COSTS
c(w, y) wx(w, y)
x
f
= the vector of xed factors.
x
v
= the vector of variable factors.
w
f
= the vector of prices of the xed factors.
w
v
= the vector of prices of the variable factors.
x
v
(w, y, x
f
) = the short-run conditional factor demand functions.
c(w, y, x
f
) = w
v
x
v
(w, y, x
f
) +w
f
x
f
.
w
v
x
v
(w, y, x
f
) = short-run variable cost.
w
f
x
f
= xed cost.
short-run total cost = STC = w
v
x
v
(w, y, x
f
) +w
f
x
f
short-run average cost = SAC =
c(w, y, x
f
)
y
short-run average variable cost = SAV C =
w
v
x
v
(w, y, x
f
)
y
short-run average xed cost = SAFC =
w
f
x
f
y
short-run marginal cost = SMC =
c(w, y, x
f
)
y
.
x
f
(w, y) = the optimal choice of the xed factors.
x
v
(w, y) = x
v
(w, y, x
f
(w, y)) = the long-run optimal choice of the variable factors.
c(w, y) = w
v
x
v
(w, y) +w
f
x
f
(w, y) = c(w, y, x
f
(w, y)).
long-run average cost = LAC =
c(w, y)
y
long-run marginal cost = LMC =
c(w, y)
y
.
Example: (The short-run Cobb-Douglas cost functions) Suppose x
2
= k. Then
min
x
1
,x
2
w
1
x
1
+w
2
k s.t. y = x
a
1
k
1a
.
Solving for x
1
:
x
1
= (yk
a1
)
1/a
.
Thus,
c(w
1
, w
2
, y, k) = w
1
(yk
a1
)
1/a
+w
2
k.
and:
short-run average cost = w
1
_
y
k
_
1a
a
+
w
2
k
y
short-run average variable cost = w
1
_
y
k
_
1a
a
short-run average xed cost =
w
2
k
y
short-run marginal cost =
w
1
a
_
y
k
_
1a
a
Proposition 6.5 (Constant Returns to Scale): If the production function ex-
hibits constant returns to scale, the cost function may be written as c(w, y) = yc(w, 1).
Proof: Let x

be a cheapest way to produce one unit of output at prices w so that

c(w, 1) = wx

. We want to show that c(w, y) = wyx

= yc(w, 1). Notice rst that

yx

is feasible to produce y since the technology is constant returns to scale. Suppose

that it does not minimise cost; instead let x

be the cost-minimising bundle to produce

y at prices w so that wx

< wyx

. Then wx

/y < wx

and x

/y can produce 1 since

the technology is constant returns to scale. This contradicts the denition of x

. Thus,
if the technology exhibits constant returns to scale, then the average cost, the average
variable cost, and the marginal cost functions are all the same. Q.E.D.
THE GEOMETRY OF COSTS
SAC =
c(w, y, x
f
)
y
=
w
f
x
f
y
+
w
v
x
v
(w, y, x
f
)
y
= SAFC + SAV C.
We know for y y

,
d
dy
_
c(y)
y
_
0. (6.3)
Taking the derivative gives us,
yc

(y) c(y)
y
2
=
1
y
_
c

(y)
c(y)
y
_
. (6.4)
But, (6.3) and (6.4) together can only be true together i
c

(y)
c(y)
y
0 ( 0)
Long-Run and Short-Run Cost Curves
c(y) = c(y, z(y)) = long-run cost function.
z(y) = the cost-minimising demand for a single xed factor.
y

= some given level of output.

z

= z(y

) = the associated long-run demand for the xed factor.

c(y, z

) = the short-run cost.

We know that: c(y

, z

) = c(y

, z(y

)).
The slope of the long-run cost curve at y

is
dc(y

, z(y

))
dy
=
c(y

, z

)
y
+
c(y

, z

)
z
z(y

)
y
.
But since z

is the optimal choice of the xed factors at the output level y

, we must
have
c(y

, z

)
z
= 0.
RECOVERING A PRODUCTION SET
FROM A COST FUNCTION
V O(y) = x : w
t
x w
t
x
t
t s.t. y
t
y.
V

(y) = x : wx wx(w, y) = c(w, y) w 0.

Proposition 7.1 (Equality of V (y) and V

(y))): Suppose V (y) is a closed,

convex, monotonic technology. Then V

(y) = V (y):
Proof: (Sketch) We already know that V

(y) contains V (y), so we only have to show

that if x V

(y) then x V(y). Suppose that x / V(y). Then since V (y) is a

closed convex set satisfying the monotonicity hypothesis, we can apply a version of the
separating hyperplane theorem to nd a vector w

0 s.t. w

x < w

z z V (y).
Let z

V (y) that minimises cost at the prices w

. Then in particular we have

w

x < w

= c(w

, y). But then x / V

(y), according to the denition of V

(y).
Q.E.D.
c(w, y) = min wx s.t. x V

(y)
Proposition 7.2 (Equality of c(w, y) and c

(w, y))): It follows from the deni-

tion of the functions that c

(w, y) = c(w, y).

Proof: It is easy to see that c

(w, y) c(w, y); since V

(y) V (y), the minimal cost

bundle in V

(y) must be at least as small as the minimal cost bundle in V (y). Suppose
that for some prices w

the cost-minimising bundle x

(y) has the property that

w

= c

(w

, y) < c(w

, y). But this cant happen, since by denition of V

(y),
w

c(w

, y). Q.E.D.
Proposition 7.3 (Constant Returns to Scale): Let V (y) be convex and mono-
tonic; then if c(w, y) can be written as yc(w), V (y) must exhibit constant returns to
scale.
Proof: Using convexity, monotonicity, and the assumed form of the cost function as-
sumptions, we know that
V (y) = V

(y) = x : wx yc(w) w 0.
We want to show that, if x V

(y), then tx V

(ty). If x V

(y), we know
that wx yc(w) w 0. Multiplying both sides of this equation by t we get:
wtx tyc(w) w 0. But this says tx V

(ty). Q.E.D.
SUFFICIENT CONDITIONS
FOR COST FUNCTIONS
Proposition 7.4 (When (w, y) is a Cost Function): Let (w, y) be a dier-
entiable function satisfying
(1) (tw, y) = t(w, y) t 0;
(2) (w, y) 0 for w 0 and y 0;
(3) (w

, y) (w, y) for w

w;
(4) (w, y) is concave in w.
Then (w, y) is the cost function for the technology dened by
V

(y) = x 0 : wx (w, y), w 0.

Proof: Given a w 0 dene
x(w, y) =
_
(w, y)
w
1
, ,
(w, y)
w
n
_
and note that since (w, y) is homogeneous of degree 1 in w, Eulers law implies that
(w, y) can be written as
(w, y) =
n

i=1
w
i
(w, y)
w
i
= wx(w, y).
Note that the monotonicity of (w, y) implies x(w, y) 0. What we need to show is
that for any given w

0, x(w

, y) actually minimises w

x over all x V

(y):
(w

, y) = w

x(w

, y) w

x x V

(y).
First, we show that x(w

, y) is feasible; that is, x(w

, y) V

(y). By the concavity of

(w, y) in w we have
(w

, y) (w, y) +D(w, y)(w

w) w 0.
Using Eulers law as above, this reduces to
(w

, y) w

x(w, y) w 0.
It follows from the denition of V

(y), that x(w

, y) V

(y).
Next we show that x(w, y) actually minimises wx over all x V

(y). If x V

(y),
then by denition it must satisfy
wx (w, y).
But by Eulers law,
(w, y) = wx(w, y).
The above two expressions imply
wx wx(w, y) x V

(y)
as required. Q.E.D.
THE INTEGRABILITY OF CONDITIONAL
FACTOR DEMAND FUNCTIONS
Construct a candidate for a cost function:
(w, y) =
n

i=1
w
i
g
i
(w, y).
1) Is (w, y) homogeneous of degree 1 in w? To check this look at (tw, y) =

i
tw
i
g
i
(tw, y). Since the functions g
i
(w, y) are by assumption homogeneous of degree
0, g
i
(tw, y) = g
i
(w, y) so that
(tw, y) = t
n

i=1
wg
i
(w, y) = t(w, y).
2) Is (w, y) 0 for w 0? Since g
i
(w, y) 0, the answer is clearly yes.
3) Is (w, y) nondecreasing in w
i
? Using the product rule
(w, y)
w
i
= g
i
(w, y) +
n

j=1
w
j
g
j
(w, y)
w
i
= g
i
(w, y) +
n

j=1
w
j
g
i
(w, y)
w
j
.
Since g
i
(w, y) is homogeneous of degree 0, the last term vanishes and g
i
(w, y) is clearly
greater than or equal to 0.
4) Is (w, y) concave in w? Dierentiate (w, y) twice to get
_

2

w
i
w
j
_
=
_
g
i
(w, y)
w
j
_
.
Example: (Recovering Production from a Cost Function) Assume c(w, y) = yw
a
1
w
1a
2
.
According to the derivative property
x
1
(w, y) = ayw
a1
1
w
a1
2
= ay
_
w
2
w
1
_
1a
x
2
(w, y) = (1 a)yw
a
1
w
a
2
= (1 a)y
_
w
2
w
1
_
a
.
Rearranging each equation gives
w
2
w
1
=
_
x
1
ay
_ 1
1a
w
2
w
1
=
_
x
2
(1 a)y
_

1
a
.
Setting these equal to each other and raising both sides to the a(1 a) power,
x
a
1
a
a
y
a
=
x
1a
2
(1 a)
1a
y
1a
,
or,
[a
a
(1 a)
1a
]y = x
a
1
x
1a
2
.