BASIC MATHEMATICS FOR ECONOMISTS

Derivatives in Use

6.1 Implicit Differentiation

6.2 Differentiating the Inverse
6.3 Linear Approximations
6.4 Polynomial Approximations
6.5 Elasticities
6.6 Continuity
6.7 More on Limits
6.8 L’Hopital’s Rule

1
6.1 Implicit Differentiation
In many economic models, the variables involved are not linked to each other
in an explicit (and obvious) way. Many of them are defined implicitly by one
or more equations. We will learn how to compute derivative of such
functions and the inverse.

Example: Consider the simple equation = 5. For this simple equation, we

can represent y in terms of directly, i.e. = . What is ( )= ( )=?
o Suppose we interpret as an implicit function of . Then we can re-express
the equation as × ( ) = 5.
o The implicit function, × ( ) = 5, must hold for all values of . As a
result, the derivatives (with respect to ) on both sides of the equation must
also be the same.
o LHS: 1 ( ) + ( ) = 0 :RHS
( )
o ( )= .
o Substitute ( ) away with . Therefore, ( )=
2
 However not every example is as simple as the previous one. In many cases,
the variable y cannot be expressed explicitly in terms of another variable, x.

Example: Consider the equation +3 = 13. It is difficult to express y

explicitly as the function of . But we can still find the slope of the graph, .

Question: What is the slope at ( , ) = (2,1)?

3[ ( )] ( ) + 6 ( ) + 3 ( )=0
( )
( )= (this is the slope at ( , ))
[ ( )]
6×2×1 4
(2) = =
3×1+3×2 5
3
 In short, the method of implicit differentiation involves finding the derivative
of y with respect to x ( i.e., ) when two variables, x and y, are related by an
equation.
Differentiate each side of the equation with respect to x, treating y as a
function of x (usually requiring the chain rule).
Solve the resulting equation for .

The second derivative of functions defined implicitly is found by repeating

the procedure we used to find earlier.

Example: Suppose = 5. Find

Recall: When we take first derivative, 1 ( ) + ( )= 0
o Take the second derivative, ( ) + × ( )+ × ( )=0.
o Simplify the derivative, ( )+ ( )+ ( ) = 0.
( )
o Therefore, ( )= , where ( )=

4
Example: In a standard macroeconomic model for determining national
income in a closed economy, the following assumptions are made.
(a) = +
(b) = 200 + 0.8

(i) Find the equilibrium of in terms of , where is expressed as a

function of .
= + = 200 + 0.5 + . Therefore, = 400 + 2
(ii) Find

=0+2× =2

You can observe that the larger is MPC, the larger becomes.
If MPC =0.5, = 2.
If MPC =0.8, = 5.
If MPC =0.9, = 10.
5
6.2 Differentiating the Inverse
Suppose ( ) is a one-to-one function. It has an inverse function ( ).
In the lecture note 4, (Section 4.3), we use ( ) to represent the inverse
function of ( ). But in this context, we will use a more general and
friendly notation ( ) where ( ) = ( ).

Through implicit differentiation, we can easily establish the relationship

between the derivative of and . If both are differentiable, then =
.
( )

= is the slope, indicating that a one-unit change in the input

variable results in a change of the output variable . For example, if
= = 2, it means that when changes by one unit, increases by 2
units.
For the inverse function of , denoted as , the input now is and
output is . Therefore, the slope of the inverse function is ( ) = .

6
 = = or ( ) =
( ) .

Since ( ) = , the slope of the inverse function is the reciprocal of the

slope of the original function.

Example: Consider the function = 2 .

(a) What is the slope of the function?
= =2

(b) What is the slope of the inverse function?

( ) = .
Alternatively, we can find the inverse function directly, which is = .
Then the slope is =

7
Example: Consider the function = . Find the slope of the inverse function.

[Answer]
Find the slope of function directly, and =5 .

Now, to find the slope of the inverse function, take the reciprocal of the original
slope: =

Using the inverse directly, = ln( ), we find = 0.2 .

8
6.3 Linear Approximations
A linear approximation around point is given by the formula:
( ) ( ) + ( )( ) where is close to .

Example: Find the linear approximation of around = 1.

Use the formula, ( ) ( )+ ( )( ) and = .
Therefore, ( ) (1) + (1)( 1)
1+ ( 1) (How accurate is it?)

9
Example: Find 1.03 using the linear approximation, 1+ ( 1)
[Answer] We can use the derived formula as = 1.03 is very close to = 1.
1.03 = 1 + (1.03 1) = 1.01.
The correct value to four decimals is 1.0099

Example: Find 8 using the linear approximation 1+ ( 1).

[Answer] 8 = 1 + (8 1) =3.33. The correct value is 2.
The error is big because = 8 is far from = 1.

Another very popular linear approximation is ln (1 + ) around = 0.

ln(1 + ) ln(1 + 0) + × =0+1× =

Example: ln(1.01) = 0.01

10
6.4 Polynomial Approximations
Sometimes the preceding linear approximation is not very precise. However,
most functions can be approximated using higher order linear approximations.

Quadratic approximation: Consider a function ( ) around point = .

( ) ( )+ ( )( )+ ( ) where is close to .
!

Example: Find the quadratic approximation of around = 1.

Use the formula, ( ) ( )+ ( )( )+ ( )( ) and
! !

( )= , ( )=
Therefore, 1+ ( 1) + × ( 1)
!
If we use the quadratic approximation to find 1.03, then it will be 1.0099 which
is more precise than using the linear approximation because the linear
approximation gives the answer 1.01.
11
 So far, we have considered linear and quadratic approximations. We can
extend this logic and consider the higher order approximations so long as the
higher order derivatives exist. The general expression is as follows.

Approximation to ( ) about =
( ) ( )+ ( )( )+ ( )( ) + ( )( ) + + ( )( )
! ! !

This is called the - order Taylor polynomial (or approximation) for

about = .

The difference between ( ) and the right-hand side is called “remainder”.

That is,
1 1 1
( )= ( )+ ( )( )+ ( )( ) + ( )( ) + + ( )( )
2! 3! !
+

Linear and quadratic approximations are the most popular approximations in

many economic models,
12
6.5 Elasticities
In economics, the derivative is that it captures the effect of a variation of the
independent variable (e.g., ) by 1 unit on the dependent variable (e.g., ).

Example:
Let ( ) denote the demand function. Then ( ) measures the reduction in
demand if price increase by S$1. The concept has a number of problems:

(1) The concept ignores size effects. Raising price of coffee by $1 is not the
same as raising the price of cars.

(2) The concept ignores measurement scale effects. For example, whether
quantities are measured in kg or pounds, or prices in S$1 or US$ would
significantly affect the result.

As a result, economists have devised a way to measure in terms of % change.

Consider the demand function ( ). The following term measures the %
change in demand if price varies by / :

13
Elasticity measures the percentage change in the dependent variable
concerning a percentage change in the independent variable. In the example
of the demand function. we obtain:
= = × (**)

If converges to zero, then = lim × = × .

Economists consider income elasticities of demand, supply elasticities,

substitution elasticities, and various other types of elasticity. The general
definition of a differentiable function at the point with ( ) 0 is
( )
= × = ( )×
( ) ( )

( ) ( )
( ) ( ) ( )
: = × = = lim (Conceptually, it is same as (**).)
( )

14
Example: Consider the function ( ) = . Find the elasticity.

[Answer]
= ( )× = × =
( )

Remarks:
(1) When considering the function f(x), economists sometimes use the following
terminology

15
(2) Who bears the burden of a tax?
Recall the tobacco tax in Assignment 4. Consumers and producers both
share the burden of the tax equally.

However, this is not always the case

o What if the demand curve is perfectly inelastic?
o What if the demand curve is perfectly elastic?

16
6.6 Continuity
Intuitively, a function is continuous if small changes in the independent
variable induce small changes in the function value.

Geometrically, a function is continuous on an interval if its graph is

connected, i.e., it has no breaks.
If you draw a curve, and never lift up your pencil, then the curve is
continuous.

17
 Continuity in terms of Limits: Analytically, a function is said to be
continuous at point = if ( ) = lim ( ).
(i) The function must be defined at = .
(ii) The limit ( )as tends to must exist.
(iii) This limit must be exactly equal to ( )

Graphs (violations of the conditions)

18
Any function that can be constructed from continuous functions by
combining one or more operations of addition, subtraction, multiplication,
division (except by zero) and composition, is continuous at all points where
it is defined.

19
6.7 More on Limits (continuity versus differentiability)
If a function is continuous, it doesn’t guarantee the function is differentiable.
Continuity does not imply differentiability.

At the point ( , ( )) the function does not (have a unique) tangent. Thus,
has no derivatives at = , but is continuous at = .

But differentiability implies continuity. (Can you prove it?)

( ) ( )
o is differentiable at = . Then ( ) = li .
o If is continuous at = , it means lim ( ) = ( )

20
6.8 L’Hopital’s Rule
We often need to examine the limit as tends to of a quotient in which
both numerator and denominator tend to zero (or infinity). For example,
( )
lim , where lim ( ) = 0 and lim ( ) = 0. (or lim ( ) = and
( )
lim ( ) = )

The French mathematician Guillaume de L'Hôpital published the following

rule, which was proved by the Swiss mathematician Johann Bernoulli, and it
is now known as L'Hôpital's rule.

( ) ( )
lim = lim , where lim ( ) and lim ( ) exist.
( ) ( )

21
Exercise:
2 4 2 1
( ) lim = lim =
8 3 6

2 4 1 4 2
( ) lim = lim = lim =
1 5 10 10 5

( ) lim = lim
×( 1) ×( 1) × … × 1
= lim = = lim =0

(d) lim ( )=?

22