Chapter 6: Optimization of Functions of One Variable
M. Hoy, J. Livernois, C. Mckenna, R. Rees and Th. Stengos
September 7, 2022
Hoy et al. Chapter 6 September 7, 2022 1 / 18
Introduction
Many economic models are based on the idea of optimization
Hoy et al. Chapter 6 September 7, 2022 2 / 18
Introduction
Many economic models are based on the idea of optimization
This is a key solution principle
Hoy et al. Chapter 6 September 7, 2022 2 / 18
Introduction
Many economic models are based on the idea of optimization
This is a key solution principle
For now, interpret this to mean maximizing or minimizing some
real-valued function y = f (x ), x ∈ R
Hoy et al. Chapter 6 September 7, 2022 2 / 18
Introduction
Many economic models are based on the idea of optimization
This is a key solution principle
For now, interpret this to mean maximizing or minimizing some
real-valued function y = f (x ), x ∈ R
Assume f (x ) is at least twice continuously differentiable and
possesses a max and/or min
Hoy et al. Chapter 6 September 7, 2022 2 / 18
Necessary Conditions
Some Definitions
Unconstrained max or min: all x ∈ R are feasible
Hoy et al. Chapter 6 September 7, 2022 3 / 18
Necessary Conditions
Some Definitions
Unconstrained max or min: all x ∈ R are feasible
Constrained max or min: only x ∈ X ⊂ R are feasible
Hoy et al. Chapter 6 September 7, 2022 3 / 18
Necessary Conditions
Some Definitions
Unconstrained max or min: all x ∈ R are feasible
Constrained max or min: only x ∈ X ⊂ R are feasible
Extreme values of y = f (x ) : a max or a min
Hoy et al. Chapter 6 September 7, 2022 3 / 18
Necessary Conditions
Some Definitions
Unconstrained max or min: all x ∈ R are feasible
Constrained max or min: only x ∈ X ⊂ R are feasible
Extreme values of y = f (x ) : a max or a min
Stationary values of y = f (x ) : points at which f ′ (x ) = 0
Hoy et al. Chapter 6 September 7, 2022 3 / 18
Necessary Conditions
Some Definitions
Unconstrained max or min: all x ∈ R are feasible
Constrained max or min: only x ∈ X ⊂ R are feasible
Extreme values of y = f (x ) : a max or a min
Stationary values of y = f (x ) : points at which f ′ (x ) = 0
Local max at x ∗ : f (x ∗ ) ≥ f (x ) for all x in a small neighbourhood of
x∗
Hoy et al. Chapter 6 September 7, 2022 3 / 18
Necessary Conditions
Some Definitions
Unconstrained max or min: all x ∈ R are feasible
Constrained max or min: only x ∈ X ⊂ R are feasible
Extreme values of y = f (x ) : a max or a min
Stationary values of y = f (x ) : points at which f ′ (x ) = 0
Local max at x ∗ : f (x ∗ ) ≥ f (x ) for all x in a small neighbourhood of
x∗
Global max at x ∗ : f (x ∗ ) ≥ f (x ) for all feasible x
Hoy et al. Chapter 6 September 7, 2022 3 / 18
Necessary Conditions
Some Definitions
Unconstrained max or min: all x ∈ R are feasible
Constrained max or min: only x ∈ X ⊂ R are feasible
Extreme values of y = f (x ) : a max or a min
Stationary values of y = f (x ) : points at which f ′ (x ) = 0
Local max at x ∗ : f (x ∗ ) ≥ f (x ) for all x in a small neighbourhood of
x∗
Global max at x ∗ : f (x ∗ ) ≥ f (x ) for all feasible x
Likewise for local and global min, with f (x ∗ ) ≤ f (x )
Hoy et al. Chapter 6 September 7, 2022 3 / 18
First Order Conditions
We are here concerned only with unconstrained max and min
solutions. Then:
Hoy et al. Chapter 6 September 7, 2022 4 / 18
First Order Conditions
We are here concerned only with unconstrained max and min
solutions. Then:
If y = f (x ) has an extreme value at x ∗ , then it has a stationary value
at x ∗ , that is
f (x ∗ ) ≥ f (x ) ⇒ f ′ (x ∗ ) = 0, or f (x ∗ ) ≤ f (x ) ⇒ f ′ (x ∗ ) = 0
Hoy et al. Chapter 6 September 7, 2022 4 / 18
First Order Conditions
We are here concerned only with unconstrained max and min
solutions. Then:
If y = f (x ) has an extreme value at x ∗ , then it has a stationary value
at x ∗ , that is
f (x ∗ ) ≥ f (x ) ⇒ f ′ (x ∗ ) = 0, or f (x ∗ ) ≤ f (x ) ⇒ f ′ (x ∗ ) = 0
To show this, consider dy = f ′ (x ∗ )dx.
Hoy et al. Chapter 6 September 7, 2022 4 / 18
First Order Conditions
We are here concerned only with unconstrained max and min
solutions. Then:
If y = f (x ) has an extreme value at x ∗ , then it has a stationary value
at x ∗ , that is
f (x ∗ ) ≥ f (x ) ⇒ f ′ (x ∗ ) = 0, or f (x ∗ ) ≤ f (x ) ⇒ f ′ (x ∗ ) = 0
To show this, consider dy = f ′ (x ∗ )dx.
If f ′ (x ∗ ) ̸= 0 then it is always possible to find feasible dx ̸= 0 such
that dy > 0 or dy < 0
Hoy et al. Chapter 6 September 7, 2022 4 / 18
First Order Conditions
We are here concerned only with unconstrained max and min
solutions. Then:
If y = f (x ) has an extreme value at x ∗ , then it has a stationary value
at x ∗ , that is
f (x ∗ ) ≥ f (x ) ⇒ f ′ (x ∗ ) = 0, or f (x ∗ ) ≤ f (x ) ⇒ f ′ (x ∗ ) = 0
To show this, consider dy = f ′ (x ∗ )dx.
If f ′ (x ∗ ) ̸= 0 then it is always possible to find feasible dx ̸= 0 such
that dy > 0 or dy < 0
So we say that f ′ (x ∗ ) = 0 is a necessary condition for a max or min
Hoy et al. Chapter 6 September 7, 2022 4 / 18
First Order Conditions
We are here concerned only with unconstrained max and min
solutions. Then:
If y = f (x ) has an extreme value at x ∗ , then it has a stationary value
at x ∗ , that is
f (x ∗ ) ≥ f (x ) ⇒ f ′ (x ∗ ) = 0, or f (x ∗ ) ≤ f (x ) ⇒ f ′ (x ∗ ) = 0
To show this, consider dy = f ′ (x ∗ )dx.
If f ′ (x ∗ ) ̸= 0 then it is always possible to find feasible dx ̸= 0 such
that dy > 0 or dy < 0
So we say that f ′ (x ∗ ) = 0 is a necessary condition for a max or min
Because this is a condition on the first order derivative, we also call it
a first order condition
Hoy et al. Chapter 6 September 7, 2022 4 / 18
Text Figure 6.2
Hoy et al. Chapter 6 September 7, 2022 5 / 18
Text Figure 6.3
Hoy et al. Chapter 6 September 7, 2022 6 / 18
Example 1
Monopoly with Linear Demand and Costs
Inverse demand function p (x ) = 100 − x, where p is price and x is
output
Hoy et al. Chapter 6 September 7, 2022 7 / 18
Example 1
Monopoly with Linear Demand and Costs
Inverse demand function p (x ) = 100 − x, where p is price and x is
output
Cost function: C (x ) = 25x
Hoy et al. Chapter 6 September 7, 2022 7 / 18
Example 1
Monopoly with Linear Demand and Costs
Inverse demand function p (x ) = 100 − x, where p is price and x is
output
Cost function: C (x ) = 25x
So profit function is π (x ) = p (x )x − C (x ) = 75x − x 2
Hoy et al. Chapter 6 September 7, 2022 7 / 18
Example 1
Monopoly with Linear Demand and Costs
Inverse demand function p (x ) = 100 − x, where p is price and x is
output
Cost function: C (x ) = 25x
So profit function is π (x ) = p (x )x − C (x ) = 75x − x 2
Then first order condition for profit maximization is
π ′ (x ) = 75 − 2x ∗ = 0
implying x ∗ = 37.5, p ∗ = $62.50, π ∗ = $1406.25
Hoy et al. Chapter 6 September 7, 2022 7 / 18
Example 1
Monopoly with Linear Demand and Costs
Inverse demand function p (x ) = 100 − x, where p is price and x is
output
Cost function: C (x ) = 25x
So profit function is π (x ) = p (x )x − C (x ) = 75x − x 2
Then first order condition for profit maximization is
π ′ (x ) = 75 − 2x ∗ = 0
implying x ∗ = 37.5, p ∗ = $62.50, π ∗ = $1406.25
Note that we could just as well have worked with demand and costs
as a function of price.
Hoy et al. Chapter 6 September 7, 2022 7 / 18
Example 2
Revenue Maximization
(Or why a publisher will always set a higher price and sell fewer copies of a
book than an author would like)
Suppose the firm in the above example is a book publisher, who sets
the book’s price
Hoy et al. Chapter 6 September 7, 2022 8 / 18
Example 2
Revenue Maximization
(Or why a publisher will always set a higher price and sell fewer copies of a
book than an author would like)
Suppose the firm in the above example is a book publisher, who sets
the book’s price
Why might the author be unhappy about this?
Hoy et al. Chapter 6 September 7, 2022 8 / 18
Example 2
Revenue Maximization
(Or why a publisher will always set a higher price and sell fewer copies of a
book than an author would like)
Suppose the firm in the above example is a book publisher, who sets
the book’s price
Why might the author be unhappy about this?
Suppose the author is paid a royalty of 10% of the book’s price
Hoy et al. Chapter 6 September 7, 2022 8 / 18
Example 2
Revenue Maximization
(Or why a publisher will always set a higher price and sell fewer copies of a
book than an author would like)
Suppose the firm in the above example is a book publisher, who sets
the book’s price
Why might the author be unhappy about this?
Suppose the author is paid a royalty of 10% of the book’s price
Then her income is Y (x ) = 0.1p (x )x = 0.1(100x − x 2 )
Hoy et al. Chapter 6 September 7, 2022 8 / 18
Example 2
Revenue Maximization
(Or why a publisher will always set a higher price and sell fewer copies of a
book than an author would like)
Suppose the firm in the above example is a book publisher, who sets
the book’s price
Why might the author be unhappy about this?
Suppose the author is paid a royalty of 10% of the book’s price
Then her income is Y (x ) = 0.1p (x )x = 0.1(100x − x 2 )
Then, assuming x is in units of 100, she would want to sell 5000
copies at a price of $50
Hoy et al. Chapter 6 September 7, 2022 8 / 18
The publisher’s profit function is now π (x ) = 0.9(100x − x 2 ) − 25x
Hoy et al. Chapter 6 September 7, 2022 9 / 18
The publisher’s profit function is now π (x ) = 0.9(100x − x 2 ) − 25x
So the publisher wants to sell 3610 copies at a price of $63.90 each
Hoy et al. Chapter 6 September 7, 2022 9 / 18
The publisher’s profit function is now π (x ) = 0.9(100x − x 2 ) − 25x
So the publisher wants to sell 3610 copies at a price of $63.90 each
The conflict of interest arises because the publisher wants to
maximize profit while the author wants to maximize sales revenue
Hoy et al. Chapter 6 September 7, 2022 9 / 18
The publisher’s profit function is now π (x ) = 0.9(100x − x 2 ) − 25x
So the publisher wants to sell 3610 copies at a price of $63.90 each
The conflict of interest arises because the publisher wants to
maximize profit while the author wants to maximize sales revenue
Under what condition would the conflict of interest not exist?
Hoy et al. Chapter 6 September 7, 2022 9 / 18
Example 3
A Monopolist Always Produces Where Demand Is Elastic
Let
R (x ) = p (x )x
be the firm’s revenue function
Hoy et al. Chapter 6 September 7, 2022 10 / 18
Example 3
A Monopolist Always Produces Where Demand Is Elastic
Let
R (x ) = p (x )x
be the firm’s revenue function
define
ϵ = −p/xp ′ (x )
as the price elasticity of demand (note p ′ (x ) < 0)
Hoy et al. Chapter 6 September 7, 2022 10 / 18
Example 3
A Monopolist Always Produces Where Demand Is Elastic
Let
R (x ) = p (x )x
be the firm’s revenue function
define
ϵ = −p/xp ′ (x )
as the price elasticity of demand (note p ′ (x ) < 0)
The firm’s marginal revenue is
R ′ (x ) = p (x ) + xp ′ (x )
= p (1 + xp ′ (x )/p )
= p (1 − 1/ϵ)
Hoy et al. Chapter 6 September 7, 2022 10 / 18
Example 3
A Monopolist Always Produces Where Demand Is Elastic
Let
R (x ) = p (x )x
be the firm’s revenue function
define
ϵ = −p/xp ′ (x )
as the price elasticity of demand (note p ′ (x ) < 0)
The firm’s marginal revenue is
R ′ (x ) = p (x ) + xp ′ (x )
= p (1 + xp ′ (x )/p )
= p (1 − 1/ϵ)
Since R ′ (x ∗ ) = C ′ (x ∗ ) ≥ 0 at the firm’s profit maximum, this implies
that ϵ ≥ 1 at that point.
Hoy et al. Chapter 6 September 7, 2022 10 / 18
Sufficient Conditions
Stationary points of a function may occur at a max, a min, or a point
of inflexion
Hoy et al. Chapter 6 September 7, 2022 11 / 18
Sufficient Conditions
Stationary points of a function may occur at a max, a min, or a point
of inflexion
Therefore it is not sufficient for a max or min that f ′ (x ∗ ) = 0
Hoy et al. Chapter 6 September 7, 2022 11 / 18
Sufficient Conditions
Stationary points of a function may occur at a max, a min, or a point
of inflexion
Therefore it is not sufficient for a max or min that f ′ (x ∗ ) = 0
For example, if we claim that x ∗ maximizes the function we might be
wrong because it really gives a min or point of inflexion
Hoy et al. Chapter 6 September 7, 2022 11 / 18
Sufficient Conditions
Stationary points of a function may occur at a max, a min, or a point
of inflexion
Therefore it is not sufficient for a max or min that f ′ (x ∗ ) = 0
For example, if we claim that x ∗ maximizes the function we might be
wrong because it really gives a min or point of inflexion
So we have to check that the claimed solution really is what we want
Hoy et al. Chapter 6 September 7, 2022 11 / 18
Sufficient Conditions
Stationary points of a function may occur at a max, a min, or a point
of inflexion
Therefore it is not sufficient for a max or min that f ′ (x ∗ ) = 0
For example, if we claim that x ∗ maximizes the function we might be
wrong because it really gives a min or point of inflexion
So we have to check that the claimed solution really is what we want
To do this we use the facts:
f ′ (x ∗ ) = 0 and f ′′ (x ∗ ) < 0 ⇒ x ∗ gives a max
f ′ (x ∗ ) = 0 and f ′′ (x ∗ ) > 0 ⇒ x ∗ gives a min
Hoy et al. Chapter 6 September 7, 2022 11 / 18
Sufficient Conditions
Stationary points of a function may occur at a max, a min, or a point
of inflexion
Therefore it is not sufficient for a max or min that f ′ (x ∗ ) = 0
For example, if we claim that x ∗ maximizes the function we might be
wrong because it really gives a min or point of inflexion
So we have to check that the claimed solution really is what we want
To do this we use the facts:
f ′ (x ∗ ) = 0 and f ′′ (x ∗ ) < 0 ⇒ x ∗ gives a max
f ′ (x ∗ ) = 0 and f ′′ (x ∗ ) > 0 ⇒ x ∗ gives a min
Since we apply this by checking whether f ′′ (x ∗ ) < 0 in case of a max,
or f ′′ (x ∗ ) > 0 in case of a min, these inequalities are called second
order conditions
Hoy et al. Chapter 6 September 7, 2022 11 / 18
To prove the above statements of sufficiency, and also to see what
happens when f ′′ (x ∗ ) = 0, we use the Taylor Series Expansion around
x∗
f ′ (x ∗ )(x − x ∗ ) f ′′ (ξ )(x − x ∗ )2
f (x ) = f (x ∗ ) + +
1! 2!
∗
for ξ ∈ (x , x ).
Hoy et al. Chapter 6 September 7, 2022 12 / 18
To prove the above statements of sufficiency, and also to see what
happens when f ′′ (x ∗ ) = 0, we use the Taylor Series Expansion around
x∗
f ′ (x ∗ )(x − x ∗ ) f ′′ (ξ )(x − x ∗ )2
f (x ) = f (x ∗ ) + +
1! 2!
∗
for ξ ∈ (x , x ).
Since f ′ (x ∗ ) = 0 and (x − x ∗ )2 > 0 we have
f ′′ (ξ ) < 0 ⇒ f (x ∗ ) > f (x )
f ′′ (ξ ) > 0 ⇒ f (x ∗ ) < f (x )
Hoy et al. Chapter 6 September 7, 2022 12 / 18
To prove the above statements of sufficiency, and also to see what
happens when f ′′ (x ∗ ) = 0, we use the Taylor Series Expansion around
x∗
f ′ (x ∗ )(x − x ∗ ) f ′′ (ξ )(x − x ∗ )2
f (x ) = f (x ∗ ) + +
1! 2!
∗
for ξ ∈ (x , x ).
Since f ′ (x ∗ ) = 0 and (x − x ∗ )2 > 0 we have
f ′′ (ξ ) < 0 ⇒ f (x ∗ ) > f (x )
f ′′ (ξ ) > 0 ⇒ f (x ∗ ) < f (x )
If f ′′ (ξ ) = 0 we continue the expansion until we find the first even
higher-order derivative that is non-zero and apply the same argument
- the ”n’th derivative test”
Hoy et al. Chapter 6 September 7, 2022 12 / 18
To prove the above statements of sufficiency, and also to see what
happens when f ′′ (x ∗ ) = 0, we use the Taylor Series Expansion around
x∗
f ′ (x ∗ )(x − x ∗ ) f ′′ (ξ )(x − x ∗ )2
f (x ) = f (x ∗ ) + +
1! 2!
∗
for ξ ∈ (x , x ).
Since f ′ (x ∗ ) = 0 and (x − x ∗ )2 > 0 we have
f ′′ (ξ ) < 0 ⇒ f (x ∗ ) > f (x )
f ′′ (ξ ) > 0 ⇒ f (x ∗ ) < f (x )
If f ′′ (ξ ) = 0 we continue the expansion until we find the first even
higher-order derivative that is non-zero and apply the same argument
- the ”n’th derivative test”
An odd higher derivative won’t do because (x − x ∗ )n ≶ 0 when n is
odd.
Hoy et al. Chapter 6 September 7, 2022 12 / 18
Text Figure 6.11
Hoy et al. Chapter 6 September 7, 2022 13 / 18
Example 4
Maximizing Tax Revenue
Competitive market supply and demand functions are:
S = bpS , D = a0 − a1 pB a0 , a1 , b > 0
Hoy et al. Chapter 6 September 7, 2022 14 / 18
Example 4
Maximizing Tax Revenue
Competitive market supply and demand functions are:
S = bpS , D = a0 − a1 pB a0 , a1 , b > 0
S is quantity supplied
Hoy et al. Chapter 6 September 7, 2022 14 / 18
Example 4
Maximizing Tax Revenue
Competitive market supply and demand functions are:
S = bpS , D = a0 − a1 pB a0 , a1 , b > 0
S is quantity supplied
D is quantity demanded
Hoy et al. Chapter 6 September 7, 2022 14 / 18
Example 4
Maximizing Tax Revenue
Competitive market supply and demand functions are:
S = bpS , D = a0 − a1 pB a0 , a1 , b > 0
S is quantity supplied
D is quantity demanded
pS is the price received by sellers
Hoy et al. Chapter 6 September 7, 2022 14 / 18
Example 4
Maximizing Tax Revenue
Competitive market supply and demand functions are:
S = bpS , D = a0 − a1 pB a0 , a1 , b > 0
S is quantity supplied
D is quantity demanded
pS is the price received by sellers
pB is the price paid by buyers, with pS = pB − t
Hoy et al. Chapter 6 September 7, 2022 14 / 18
Example 4
Maximizing Tax Revenue
Competitive market supply and demand functions are:
S = bpS , D = a0 − a1 pB a0 , a1 , b > 0
S is quantity supplied
D is quantity demanded
pS is the price received by sellers
pB is the price paid by buyers, with pS = pB − t
t ≥ 0 is a specific tax per unit bought and sold, i.e. $t per unit
Hoy et al. Chapter 6 September 7, 2022 14 / 18
The equilibrium condition D = S implies
b (pB − t ) = a0 − a1 pB
Hoy et al. Chapter 6 September 7, 2022 15 / 18
The equilibrium condition D = S implies
b (pB − t ) = a0 − a1 pB
with the equilibrium values
a0 b
pB∗ = + t
a1 + b a1 + b
S ∗ = D ∗ = a0 − a1 pB∗
Hoy et al. Chapter 6 September 7, 2022 15 / 18
The equilibrium condition D = S implies
b (pB − t ) = a0 − a1 pB
with the equilibrium values
a0 b
pB∗ = + t
a1 + b a1 + b
S ∗ = D ∗ = a0 − a1 pB∗
So equilibrium price increases with the tax, quantity bought and sold
decreases
Hoy et al. Chapter 6 September 7, 2022 15 / 18
After substitution, we have that tax revenue R (t ) = tD ∗ = tS ∗ is
given by the function
R (t ) = αt − βt 2
where
a1
α = a0 (1 − )>0
a1 + b
a1 b
β= >0
a1 + b
Hoy et al. Chapter 6 September 7, 2022 16 / 18
After substitution, we have that tax revenue R (t ) = tD ∗ = tS ∗ is
given by the function
R (t ) = αt − βt 2
where
a1
α = a0 (1 − )>0
a1 + b
a1 b
β= >0
a1 + b
So the tax that maximizes tax revenue is
α
t∗ =
2β
Hoy et al. Chapter 6 September 7, 2022 16 / 18
After substitution, we have that tax revenue R (t ) = tD ∗ = tS ∗ is
given by the function
R (t ) = αt − βt 2
where
a1
α = a0 (1 − )>0
a1 + b
a1 b
β= >0
a1 + b
So the tax that maximizes tax revenue is
α
t∗ =
2β
and we can confirm that this is a true maximum since
R ′′ (t ∗ ) = −2β < 0
Hoy et al. Chapter 6 September 7, 2022 16 / 18
Note that R ′′ (t ) < 0 for all values of t
Hoy et al. Chapter 6 September 7, 2022 17 / 18
Note that R ′′ (t ) < 0 for all values of t
In fact this is a strictly concave quadratic function with unique
stationary value at t ∗ = α/2β
Hoy et al. Chapter 6 September 7, 2022 17 / 18
Note that R ′′ (t ) < 0 for all values of t
In fact this is a strictly concave quadratic function with unique
stationary value at t ∗ = α/2β
The graph of this function has an inverted U-shape and is often called
”the Laffer Curve”
Hoy et al. Chapter 6 September 7, 2022 17 / 18
Note that R ′′ (t ) < 0 for all values of t
In fact this is a strictly concave quadratic function with unique
stationary value at t ∗ = α/2β
The graph of this function has an inverted U-shape and is often called
”the Laffer Curve”
An important point made by this curve is that if t > t ∗ , a reduction
in the tax increases tax revenue, gladdening the hearts of politicians
and taxpayers alike.
Hoy et al. Chapter 6 September 7, 2022 17 / 18
More generally, the example suggests that:
Hoy et al. Chapter 6 September 7, 2022 18 / 18
More generally, the example suggests that:
if a function is everywhere strictly concave (f ′′ (x ) < 0 at all points in
its domain) and has a stationary value, then this is a unique global
maximum
Hoy et al. Chapter 6 September 7, 2022 18 / 18
More generally, the example suggests that:
if a function is everywhere strictly concave (f ′′ (x ) < 0 at all points in
its domain) and has a stationary value, then this is a unique global
maximum
likewise if a function is everywhere strictly convex (f ′′ (x ) > 0 at all
points in its domain) and has a stationary value, then this is a unique
global minimum
Hoy et al. Chapter 6 September 7, 2022 18 / 18
More generally, the example suggests that:
if a function is everywhere strictly concave (f ′′ (x ) < 0 at all points in
its domain) and has a stationary value, then this is a unique global
maximum
likewise if a function is everywhere strictly convex (f ′′ (x ) > 0 at all
points in its domain) and has a stationary value, then this is a unique
global minimum
This is often used by economists to help in constructing
”well-behaved” models.
Hoy et al. Chapter 6 September 7, 2022 18 / 18
