c

3.1 Concave and convex functions of a single variable

General definitions

The twin notions of concavity and convexity are used widely in economic theory, and are also
central to optimization theory. A function of a single variable is ?
? if every line segment
joining two points on its graph does not lie above the graph at any point. Symmetrically, a
function of a single variable is ?
 if every line segment joining two points on its graph does
not lie below the graph at any point. These concepts are illustrated in the following figure.

Here is a precise definition.

Definition
Let  be a function of a single variable defined on an interval. Then  is

{c ?
? if every line segment joining two points on its graph is never above the
graph
{c ?
 if every line segment joining two points on its graph is never below the
graph.

To make this definition useful we need to translate it into an algebraic condition that we can
check. Let  be a function defined on the interval [1, 2]. This function is concave according to
the definition if, for every pair of numbers  and  with 1 ”  ” 2 and 1 ”  ” 2, the line
segment from (,  ()) to (,  ()) lies on or below the function, as illustrated in the following
figure.
3.1 Concave and convex functions of a single variable
General definitions

The twin notions of concavity and convexity are used widely in economic theory, and are also
central to optimization theory. A function of a single variable is ?
? if every line segment
joining two points on its graph does not lie above the graph at any point. Symmetrically, a
function of a single variable is ?
 if every line segment joining two points on its graph does
not lie below the graph at any point. These concepts are illustrated in the following figure.

Here is a precise definition.

Definition
Let  be a function of a single variable defined on an interval. Then  is

{c ?
? if every line segment joining two points on its graph is never above the
graph
{c ?
 if every line segment joining two points on its graph is never below the
graph.
To make this definition useful we need to translate it into an algebraic condition that we can
check. Let  be a function defined on the interval [1, 2]. This function is concave according to
the definition if, for every pair of numbers  and  with 1 ”  ” 2 and 1 ”  ” 2, the line
segment from (,  ()) to (,  ()) lies on or below the function, as illustrated in the following
figure.

Denote the height of the line segment from (,  ()) to (,  ()) at the point  by ,(). Then for
the function  to be concave, we need

 () • ,() for all  with  ”  ”  (*)

for every pair of numbers  and  with 1 ”  ” 2 and 1 ”  ” 2.

Now, every point  with  ”  ”  may be written as  = (1 í Ȝ) + Ȝ, where Ȝ is a real number
from 0 to 1. (When Ȝ = 0, we have  = ; when Ȝ = 1 we have  = .) The fact that , is linear
means that

,((1 í Ȝ) + Ȝ) = (1 í Ȝ),() + Ȝ,()

for any value of Ȝ with 0 ” Ȝ ” 1. Further, we have ,() =  () and ,() =  () (the line
segment coincides with the function at its endpoints), so
,((1 í Ȝ) + Ȝ) = (1 í Ȝ)  () + Ȝ  ().

Thus the condition (*) is equivalent to

 ((1íȜ) + Ȝ) • (1 í Ȝ)  () + Ȝ  () for all Ȝ with 0 ” Ȝ ” 1.

We can make a symmetric argument for a convex function. Thus the definition of concave and
convex functions may be rewritten as follows.

Definition
Let  be a function of a single variable defined on the interval . Then  is

{c ?
? if for all  · , all  · , and all Ȝ · (0, 1) we have

  ((1íȜ) + Ȝ) • (1 í Ȝ)  () + Ȝ  ()

{c ?
 if for all  · , all  · , and all Ȝ · (0, 1) we have

 ((1íȜ) + Ȝ) ” (1 í Ȝ)  () + Ȝ  ().

In an exercise you are asked to show that  is convex if and only if í  is concave.

Note that a function may be 

 concave and convex. Let  be such a function. Then for all
values of  and  we have

 ((1íȜ) + Ȝ) • (1 í Ȝ)  () + Ȝ  () for all Ȝ · (0, 1)

and
 ((1íȜ) + Ȝ) ” (1 í Ȝ)  () + Ȝ  () for all Ȝ · (0, 1).
Equivalently, for all values of  and  we have
 ((1íȜ) + Ȝ) = (1 í Ȝ)  () + Ȝ  () for all Ȝ · (0, 1).
That is, a function is both concave and convex if and only if it is O
 (or, more precisely,

), taking the form  () = Į + ȕ for all , for some constants Į and ȕ.

Economists often assume that a firm's production function is increasing and concave. An
example of such a function for a firm that uses a single input is shown in the next figure. The fact
that such a production function is increasing means that more input generates more output. The
fact that it is concave means that the increase in output generated by a one-unit increase in the
input is smaller when output is large than when it is small. That is, there are "diminishing
returns" to the input, or, given that the firm uses a single input, "diminishing returns to scale".
For some (but not all) production processes, this property seems reasonable.

The notions of concavity and convexity are important in optimization theory because, as we shall
see, the first-order conditions are sufficient (as well as necessary) for a maximizer of a concave
function and for a minimizer of a convex function. (Precisely, every point at which the derivative
of a concave differentiable function is zero is a maximizer of the function, and every point at
which the derivative of a convex differentiable function is zero is a miniimizer of the function.)
The next example shows that 

?

?
?


 ?
? 
?



?
?.

Example
Let ] be a concave function and a nondecreasing and concave function. Define the
function  by  () = (]()) for all . Show that  is concave.

We need to show that  ((1íȜ) + Ȝ) • (1íȜ)  () + Ȝ  () for all values of  and  with
 ” .

By the definition of  we have

 ((1íȜ) + Ȝ) = (]((1íȜ) + Ȝ)).

Now, because ] is concave we have
]((1íȜ) + Ȝ) • (1 í Ȝ)]() + Ȝ]().
Further, because is nondecreasing,  •  implies () • (). Hence
(]((1íȜ) + Ȝ)) • ((1íȜ)]() + Ȝ]()).
But now by the concavity of we have
((1íȜ)]() + Ȝ]()) • (1íȜ) (]()) + Ȝ (]()) = (1íȜ)  () + Ȝ  ().
So  is concave.

6 ice-differentiable functions

We often assume that the functions in economic models (e.g. a firm's production function, a
consumer's utility function) are differentiable. We may determine the concavity or convexity of a
twice differentiable function (i.e. a function that is differentiable and that has a differentiable
derivative) by examining its second derivative: a function whose second derivative is nonpositive
everywhere is concave, and a function whose second derivative is nonnegative everywhere is
convex.

Here is a precise result.

Proposition
A twice-differentiable function  of a single variable defined on the interval is

{c concave if and only if  ''() ” 0 for all  in the interior of

{c convex if and only if  ''() • 0 for all  in the interior of .

The importance of concave and convex functions in optimization theory comes from the fact that
if the differentiable function  is concave then  point  at which  '() = 0 is a global
maximizer, and if it is convex then  such point is a global minimizer.

Example
 Is 2 í 2 + 2 concave or convex on any interval? Its second derivative is 2 • 0, so it is
convex for all values of .
Example
Is 3 í 2 concave or convex on any interval? Its second derivative is 6 í 2, so it is
convex on the interval [1/3, ’) and concave the interval (í’, 1/3].

The next example shows how the result in an earlier example may be established for twice-
differentiable functions. (The earlier result is true for OO functions, so the example proves a
result we already know to be true; it is included only to show how a version of the earlier result
for twice-differentiable functions may be established by using the characterization of concavity
in the previous Proposition.)

Example
Let ] be a concave function and a nondecreasing and concave function. Assume that ]
and are twice-differentiable. Define the function  by  () = (]()) for all . Show
that  is concave.

We have  '() = '(]())]'(), so that

 "() = "(]())·]'()·]'() + '(]())]"().

Since "() ” 0 ( is concave), '() • 0 ( is nondecreasing), and ]"() ” 0 (] is
concave), we have  "() ” 0. That is,  is concave.

A point at which a twice-differentiable function changes from being convex to concave, or vice
versa, is an inflection point.

Definition
? is an

?



 of a twice-differentiable function  of a single variable if for
some values of  and  with  < ? <  we have

{c either  "() • 0 if  <  < ? and  "() ” 0 if ? <  < 

{c or  "() ” 0 if  <  < ? and  "() • 0 if ? <  < .

An example of an inflection point is shown in the following figure.

Proposition

{c If ? is an inflection point of  then  "(?) = 0.

{c If  "(?) = 0 and  " changes sign at ? then ? is an inflection point of  .

Note, however, that  " does not have to change sign at ? for ? to be an inflection point of  . For
example, every point is an inflection point of a linear function.

3trict convexity and concavity

The inequalities in the definition of concave and convex functions are weak: such functions may
have linear parts, as in the following figure.

A concave function that has
linear parts is said to be 

 ?
O ?
?.

Definition
The function  of a single variable defined on the interval is

{c 
??
? if for all  · , all  · with   , and all Ȝ · (0,1) we have

 ((1íȜ) + Ȝ) > (1 í Ȝ)  () + Ȝ  ().

{c 
??
 if for all  · , all  · with   , and all Ȝ · (0,1) we have

 ((1íȜ) + Ȝ) < (1 í Ȝ)  () + Ȝ  ().

An earlier result states that if  is twice differentiable then

 is concave on [, ] if and only if  "() ” 0 for all  · (, ).

Does this result have an analogue for 
 ?
O concave functions? Not exactly. If  "() < 0 for all
 · (,) then  is strictly concave on [, ], 

?
 


: if  is strictly concave
then its second derivative is
 necessarily negative at all points. (Consider the function  () =
í4. It is concave, but its second derivative at 0 is zero,


 
.) That is,
 is strictly concave on [, ] if  "() < 0 for all  · (, ), ` if  is strictly concave on [, ]
then  "() is

necessarily negative for all  · (, ).
(Analogous observations apply to the case of convex and strictly convex functions, with the
conditions  "() • 0 and  "() > 0 replacing the conditions  "() ” 0 and  "() < 0.)