Higher Order Differentiation and Its Applications

DC-1
Semester-II
Paper-IV: Mathematical methods for Economics-II
Lesson: Higher Order Differentiation and Its Applications
Lesson Developer: Sarabjeet Kaur
College/Department: Department of Economic, P.G.D.A.V College,
University of Delhi

1
 Higher Order Differentiation and Its Applications

CONTENTS

1. Learning Outcomes
2. Higher Order Differentiation
3. Partial Derivatives
3.1Higher order partial derivative
3.2Partial derivative with many variables
4. Quadratic Forms
5. Exercise
6. References
1. Learning Outcomes

After completing of the present chapter, you should able to:-

1. Higher Order Differentiation

2. Examples of Higher Order Differentiation
3. Partial Differentiation
4. Examples of Partial Differentation
5. Clairant Theorm/Young’s theorm

2. Higher Order Differentiation

If f (x) be differentiable function of x, then f'(x) or is the first derivative or first order

derivative of y = f (x) w.r.t ‘x’. Since the derivative of function is also a function, therefore
another derivative can also be find. The second order derivative, or second derivative, is the
derivative of the first derivative of the function f(x). Other notations are:

or or or f '' (x)

Since f '' (x) is also a function, therefore, its derivative can also be find which is denoted as f '''
(x). For higher order derivatives superscripts can be used i.e. f4 = fourth derivative etc.
Example: -f(x) = 4x5 + 6x3+2x+1

f''(x) = 20x4 + 18x2 + 2

2
 Higher Order Differentiation and Its Applications

f''(x) = 80x3 + 36x

f'''(x) = 240x2 + 36

3. Partial Derivatives

Given a function y = f(x), the derivative f '(x), represents the rate of change of the function as x
changes. For a function of two variables, such as z = f (x,y), one variable could be changing
faster than the other variables. It will be completely possible for the function to be changing
differently.

For a function of two independent variables, z = f(x,y), the partial derivative ‘z’ with respect to x
may be found as normal rule of differentiation. The only difference is that, whenever or
wherever the second independent variable ‘y’ appears, it will be treated as constant in every
respect. Also the partial differentiation of y can be found by treating x variable as constant.
Notations of partial differentiation are given below:

Notations of Partial Differentiation

Partial derivative of z w.r.t x fx

Partial derivative of z w.r.t y Fy

Example: Z = x4 y2 – x2 y6

= 4x3y2 – 2xy6

= 2x4y – 6x2 y5

Partial derivative can be defined as:-

If z = f(x,y) is a function of two variables, then and

, called partial derivatives of z with respect to x and y respectively, be the derivative z w.r.t.x

3
 Higher Order Differentiation and Its Applications

by keeping y as constant and the derivative z w.r.t.y by keeping x as constant. All the rules of
differentiation can be applied when partial differentiation can be calculated.

Symbolically if f =f (x,y) then

fx = =LtΔx →0

fy = = LtΔy →0

provided these limits exist.

Example 1.z = f (x, y) = x3y+x2y2 +xy +x+y2

= 3x2y + 2xy2 + y + 1

= x3 + 2x2y + x +2y

3.1 Higher order partial derivative

For a function z = f (x, y); fˈ (x) & fˈ (y) are the two first order partial derivatives with respect to
x and y respectively. Since ‘z’ is a function hence fˈ (x) and fˈ (y) are also a function, hence,
second order partial differentiation can also be found.

The second order partial derivatives are called mixed partial derivative because derivatives of
more than one variable are to be observed. e.g differentiating a function with respect to ‘x’ first
and then ‘y’ is called as mixed partial derivative. The various notation of partial derivative are
given in table:

Notations of second order partial derivatives

Partial derivatives of z Notation 1 Notation 2 Notation 3

w.r.t. x twice ( ) fxx

4
 Higher Order Differentiation and Its Applications

w.r.t. y twice ( ) fyy

w.r.t. x first than y ( ) fxy

w.r.t. y y first than x ( ) fyx

A function has four possible second partial derivatives ones that are obtained by differentiating
function w.r.t ‘x’ twice, w.r.t. y twice, w.r.t. x first than y and w.r.t. y first then x. All derivatives
have sign (+ or -) interpretation of these signs are as follows.

Partial Sign Interpretation

derivative
+ Slopes in x direction is positive
- Slopes in x direction is negative
+ Slopes in x direction increases as x increases(y constant)
- Slopes in x direction decreases as x decreases(y constant)
+ Slopes in y direction is positive
- Slopes in y direction is negative
+ Slopes in y direction increases as y increases(x constant)
- Slopes in y direction decreases as y decreases(x constant)
+ Slopes in x direction increases as y increases(x constant)
- Slopes in x direction decreases as y decreases(x constant)
+ Slopes in y direction increases as x increases(y constant)
- Slopes in y direction decreases as x decreases(y constant)

Example z = x0.5 y0.5 – 60 find interprate the result

sol. = 0.5 x0.5 y0.5

5
 Higher Order Differentiation and Its Applications

= (0.5 x0.5 y0.5)

= (0.5) (-0.5 x-0.5 y0.5)

= - 0.25 x-1.5 y0.5

Since x and y are positive, positive number raised to any power is positive; hence y0.5 and x-1.5
are positive , the term -0.25 in equation show that second order differentiation of z with respect
to x twice is negative meaning that the slope in the x direction decreases as x increases when y is
constant.

Example : z = f(x, y) = x3y +x2 y +2x + xy + x + y2

= x3 + 2x2y + 2xy2 +y +1

fy= x3 + 2x2y + x + 2y

fxx= 6xy +2y2

fyy = 2x2 + 2

fxy = 3x2 + 4xy +1

fyx = 3x2 + 4xy +1

.: fxy = fyx .

The two mixed second order partial derivatives (also called as cross partial derivatives) are
always equal when fxy and fyx are continuous. It is explained by the following theorem given by
Alexis Clairant also know as Young’s theorem.

Theorem: Suppose f is defined on a disk D, which contains the point (a, b). If the partial
derivatives fxy and fyx are both continuous on disk D, then
fxy (a, b) = fyx (a, b).

Example:- Verify Young’s theorem f(x,y) = x

Solution:-

6
 Higher Order Differentiation and Its Applications

fx(x,y) = -2 x2 y2

fy(x,y) = -2 x3 y

Now, compute the two mixed partial derivatives.

fxy(x,y) = -2 x2 y -4 x2 y +4 x4y3

= -6 x2 y +4 x4y3

fyx(x,y)= -6 x2 y +4 x4y3

.: fxy = fyx

Hence proved.

3.2 Partial derivative with many variables

If z = f (x1, x2 ……. xn) then

is the differentiation of the function w.r.t. xi when all the other variables xj (j≠i ) are held

constant.

i.e =

and = and so on

Suppose, there is a function which consists of three variables v = f (x, y, z). For such a function,
there are partial derivatives of w.r.t x, y and z. When partial derivative has to take with respect to
one of x, y and z assuming other two independent variables are constant.

In general, function consists of n variables. If Z = f (x1 x2,….,xn) then partial derivative of z

w.r.t. xi is when all the other variables xj (j≠i) are held constant.

fxi = = Lth →0

7
 Higher Order Differentiation and Its Applications

provided limit exists.

Example:- f (x, y, z) = x2 +y3 +z4

fx = = 2x

fy=3y2

fz = 4z3

Example: Find Zxxx, Zxyx, Zyyy,Zyxy of the function

Z= 3x2(5x+7y)

Zx= 3x2(5) + (5x+7y)(6x)

= 45x2 + 42xy

Zxx= 90x+42y

Zxxx= 90.

Zxy = 0 + 42x

= 42x

Zxyx = 42

Zy= 3x2(7) + (5x+7y)(0)

=21 x2

Zyy = 0

Zyyy = 0.

Zyx = 42x

Zyxy = 0.

Example: Find Zxxx, Zxyy, Zyyy,Zxxy of the function Z= ( 9x – 4y)(12x + 2y)

8
 Higher Order Differentiation and Its Applications

Zx = (9x – 4y)(12) + (12x + 2y) (9)

= 108x – 48y + 108 x + 18 y

= 216 x – 30 y

Zxx = 216

Zxyy = 0

Zxxx = 0

Zxy = -30

Zxyy = 0

Zy = (9x-4y) (2) + (12x +2y) (-4)

= 18x – 8y -48x -8y

= -30x -16y

Zyy = -16

Zyyy =0

Example: Find Zxx, Zyy of the function Z =

Zx =

Zxx= 0

Zy =

= =

Zyy = - (-2xy-3) = xy-3

9
 Higher Order Differentiation and Its Applications

Clairaut theorem (Young’s theorem) can be extended to any function of ‘n’ number of variables
and their mixed partial derivations. The only thing has to remember that in each derivative, we
differentiate with respect to each variable the same number of times.

For three variables, according to clairaut theorem,

fxz (x, y, z) = fzx (x,y,z)

provided with the derivatives are continuous.

The partial derivative is approximate equal to the change in function.i.e,

fi(x1,….xn)≈ fi(x1,…,xi-1, xi+h, xi+1…..xn)- fi(x1,…,xi-1, xi, xi+1…..xn)

There are n partial derivatives of first order. For each of the first order partial order derivative of
the function, there are n second order derivatives. i.e.,

( )= =fxixj (i=1..n;j=1..n)

So, total n2 elements are there. Therefore, n*n matrix of second order partial derivative is the
Hessian matrix which is symmetric and all f11=f22=….=fn(Clairant theorem)

Example:

If the two demand functions for the two commodities are given by

x= y=

then the marginal demand functions are

= =

10
 Higher Order Differentiation and Its Applications

= =

Since ≥0 and ≥0, therefore, two commodities are competitive.

Example:

If the demand functions for two related commodities are given by

x=ae-pq and y=bep-q where a ≥ 0 b ≥ 0

Solution: Since two demand functions are given as

x=ae-pq

y=bep-q

their marginal demand functions can be calculated as:

= -aqe-pq = bep-q

= -ape-pq = -bep-q

Because ≤ 0 and ≥0, therefore the given commodities are neither competitive nor

complementary.

Example

Consider two products, A and B. the demand for good A and B, & described by following two
equations

qa =

qb = find q/ q given the result explain A and Bare complementary or substitutes.

Solution

11
 Higher Order Differentiation and Its Applications

qa = =

= 200( ( pb-1/2) /pa)

= -100

=-100 pa-1 pb -3/2

qb =

= =( )

= ( (pa-1/3)

= ( )

= ( pa-1/3-1)

=
pa-4/3 . pb-1

We know that pa and pb are positive because prices can never be negative therefore:

=-( ) = -( )<0

=- =-( )<0

Because both cross elasticities are – ve

Example: Two goods are complement goods.

The Stone-Geary Utility function is written as u= log U = β1 log (q1 – γ1) + β2( q2 – γ2), where u is
the utility index, qi is the quantity of commodity i, 0 < i < 1, i > 0, qi - i > 0 and i = 1,2.
a. Find the marginal utility of this function with respect to q1 and determine its sign.

12
 Higher Order Differentiation and Its Applications

b. What is the significance of a positive marginal utility?

c. Find the second derivative of this function with respect to q1. Does the utility function
exhibit diminishing marginal utility?
Solution: Utility function is: u= log U = β1 log (q1 – γ1) + β2( q2 – γ2)

a. Marginal utility is given by : = ; which is greater than zero; because both the

numerator and the denominator are positive.

b. Since marginal utility is positive; this implies that as utility increases monotonically with
increase in q1.

c. ; which is less than zero. Since the second derivative is negative, the

utility function exhibits diminishing marginal utility.

Example: Given the production function:

P(L,K)=5L1/2 K1/2 +L.

Find out the partial elasticity with respect to labor at (L,K)= (1024,27).

Solution: P(L,K)=5L1/2 K1/2 +L

ϵL= PˈL (L,K).L

= (L-4/5 K-1/3 +1)

Therefore, at (L,K)=(1024,27) we have ϵL= =

This explains that if capital remains constant at K=27 and at L=24 labour increases with 1
percent, then output will increase by percent.

Example: Utility function is given:

13
 Higher Order Differentiation and Its Applications

U = U= X0.5Y0.5.

Calculate the marginal rate of substitution between X,Y.

Function is U= X0.5Y0.5.

First, take the partial derivative of U with respect to X to get MUx.

MUx= = 0.5X-0.5Y0.5.

Next, take the partial derivative with respect to Y to get MUy.

MUy= = 0.5X0.5Y-0.5.

Dividing MUx by MUy we get

MRS= = = y/x

Example: Given an isoquant

Q = K1/6L1/2
Find out slope of isoquant.

Solution: Slope of Isoquant =

=- /

= K1/6L-1/2

= K-5/6L1/2

= −( K1/6L-1/2)/ ( K-5/6L1/2)

=- K1/6 K5/6L-1/2L-1/2

=-3(K/L)
Therefore, the slope of isoquant is 3(K,L).

Example: Given demand function Q- 90+ 2P=0; and average cost function
AC= Q2- 39.5Q+ 120+ 125/Q

14
 Higher Order Differentiation and Its Applications

Calculate the level of output where:

(a) total revenue is maximum,
(b) marginal cost is minimum,
(c) profits is maximum.
Solution: (a) The demand function is Q- 90+ 2P=0.
Written as 2P=90 - Q
P= 45 – 0.5Q
TR= PQ= (45 – 0.5Q)Q
=45Q – 0.5 Q2
For maximizing TR, first-order condition is:

=45 – Q=0

Q=45

and second-order condition is, = -1<0.

Therefore, at Q=45, TR is maximized.

(b) AC= Q2- 39.5Q+ 120+ 125/Q
TC= AC.Q= (Q2- 39.5Q+ 120+ 125/Q)Q
= Q3- 39.5Q2+ 120Q+ 125

MC= = 3Q2 -79Q+ 120

MC is minimum when, =0 and

= 6Q -79=0

Q= 13.167

And, =6>0.

Hence, at Q= 13.167, MC is minimum.

(c) Profit (π) = TR- TC
=45Q – 0.5 Q2 – (Q3- 39.5Q2+ 120Q+ 125)
= -Q3 + 39 Q2 - 75Q- 125
For maximization of profit, first order condition

= - 3Q2 + 78Q- 75=0

15
 Higher Order Differentiation and Its Applications

(-3Q + 3)(Q – 25) =0

Q=1 and Q= 25
and for second order condition,

= - 6Q + 78

When Q = 1 then,

= 72>0.

When Q=25 then,

= -72<0.

Therefore, profit is maximum when Q=25

Maximum = -(25)3+ 39 (25)2-75(25) -125= 6750.

Example: Two different demand functions are given:

Q1= 21 – o.1P1 and Q2 = 50 – 0.4 P2
TC = 2000 + 10 Q where Q= Q1+ Q2, what price will the firm charge (a) with discrimination and
(b) without discrimination between markets?
Solution: since demand function in first market is, Q1= 21 – o.1P1
Therefore, P1= 210- 10 Q1
And, TR1= P1Q! = (210-10Q1)Q1 = 210Q1 – 10 Q12

MR1= = 210- 20Q1

Profit is maximum when MR= MC,

MC = = 10

MR1= MC
210- 20Q1=10
Q1= 10
When Q1= 10, P1 = 210- 10(10)= 110
demand function in second market is, Q2= 50 – o.4P2
hence, P2= 125 – 2.5 Q2
TR2= (125 – 2.5 Q2)Q = 125Q – 2.5 Q22

16
 Higher Order Differentiation and Its Applications

MR2= = 125- 5Q2

When MR2=MC
125- 5Q2=10
Q2= 23
When Q2= 23, then P2= 125 – 2.5(23)= 67.5
The discriminating monopoly charges a lower price in the second market where the demand is
relatively more elastic, and a higher price in the first market where the demand is relatively less
elastic.
Example: A producer is a price-taker on both the market for input factors labor and capital, and
the market for end products. The cost of one unit of labor equals w = 2, the cost of one unit of
capital equals r = 32 , while the selling price of the end products equals p =32. The production
function of this producer is given by Y(L,K) = L1/8 K1/2. Determine the maximum profit.
Solution:
The revenue function is R(L,K) = p.Y(L,K) = 32. L1/8 K1/2
Cost function
C(L,K) = wL + rK= 2L+32K, and
Hence, profit function becomes
Π(L,K) = 32 L1/8 K1/2 – 2L – 32K
Partial derivative of π(L,K) is given by:
= 4L-7/8K1/2 -2 and
= 16 L1/8K-1/2 -32
the stationary points of profit function are solutions of the following system
4L-7/8K1/2 -2 = 0
16 L1/8K-1/2 -32 = 0
Hence, K1/2=1/2L7/8 and
therefore, K = ¼ L14/8
Consequently, L1/8(1/4 L14/8)-1/2= 1
which gives L=1 and therefore, K=1/4
Hence, (L,K) = (1,1/4) is the only stationary point. By the use of the criterion function we
investigate whether or not this point is a maximum location.
= -3.1/2L-15/8K1/2;

17
 Higher Order Differentiation and Its Applications

= -8L1/8K-3/2 and
= 2L-7/8K-1/2, which implies that the criterion function is given by
C(L,K) = . -( )2
= (-3.1/2L-15/8K1/2)(-8L1/8K-3/2)- (2L-7/8K-1/2)2
= 28 L-14/3K-1 – 4L-14/8K-1
= 24 L-14/8K-1>0
Hence, as C(1,1/4) >0 and (1,1/4) <0 it follows that has a maximum profit
at (L,K)= (1,1/4), with value π=6.

4. Quadratic Forms
A quadratic form of two variables is
f(x,y) = ax2 + 2bxy +cy2;
a,b,and c are constants. Now, using matrix notation:

f(x,y) = (x,y)

= 2a, = = 2b and = 2c are the second order partial derivatives of the function f(x,y)
Therefore, the Hessian of f is given by

The given quadratic form is said to positive definite if f(x,y) >0; for all values of x and y i.e,
(x,y) ≠ (0,0), and positive semidefinite if f(x,y)≥ 0 for all values of (x,y). The given function is
negative definite if f(x,y)<0; for all values of x and y; and it is negative semidefinite if f(x,y)≤0.
And it is indefinite we have two different pairs of x and y; (x-,y-) and (x+,y+); and also f(x+,y+)
>0.
Example: Express the quadratic form below as a matrix form. Determine the definiteness of the
equations:
a) f(x1,x2) = 4x2 +8xy +5y2
b) f(x1,x2) = -x2 +xy - 3y2

Solution: a) f(x,y) = (x,y)

Therefore, symmetric matrix is , whose determinant is positive. Hence,f(x,y) > 0 for all

values of x and y. Therefore, the quadratic form is positive definite.

18
 Higher Order Differentiation and Its Applications

f(x,y) = (x,y)

Therefore, symmetric matrix is , whose determinant is negative. Hence,f(x,y)<

0 for all values of x and y. Therefore, the quadratic form is negative definite.

5. Exercise:
1. Find the second – order partial derivatives fxx, fyy and fxy for each of the following
functions:
(a) Z=

(b) Z = (7x +3y)3

(c) Z= (x3+ 5y)5
(d) Z = (2x+5y)ey
(e) Z=log (1+ x2) +y2
(f) Z = 3x2e2y
2. Consider the function: f ( x1 , x2 )  (3x12  5x1  1)  ( x2  4) .
a. Find f1 and f2.
b. Find f11, f22, f12 and f21.
3. Assume the demand for sugar is a function of income (Y), the price of sugar (Ps) and the
price of saccharine (Pc), a sugar substitute, as follows:
Qd  f (Y , Pc , Ps )  0.05Y  10 Pc  5Ps2 .

a. Find the partial derivatives of this demand function.

  Qd Y 
b. Find the elasticity of demand with respect to income    when Y = 10,000, Ps =
  Y Qd 
5 and Pc = 7.
  Qd Ps 
c. Find the own-price elasticity of demand    when Y = 10,000, Ps = 5 and Pc = 7.
  Ps Qd 

  Qd Pc 
d. Find the cross-price elasticity of demand    when Y = 10,000, Ps = 5 and Pc = 7.
  Pc Qd 
4. Show that fxz = fzx and fxzz = fzxz = fzzx from the following function:

19
 Higher Order Differentiation and Its Applications

F(x,y,z) = y

5. The demand function of two related commodities are given by

X1= p1-1.7p20.8

X2= p10.5p2-0.2

What can you say about the two commodities X1 and X2 and also find all partial elasticties.

6. A firm produces two commodities: commodity X and commodity Y. the demand

functions are:
p1 =8 – 2x
p2 =14 –y2
The combined cost of production of these unit is given by C = 10+4x+2y. What will be
the prices of two products so that joint profit will be the maximum.
1 1
7. Consider a production function that takes the form y  10 L2 K 2 , and assume that capital
(K) is constant at K0 = 64.
y
a. Find the marginal product of labor, .
L
b. If the labor were paid real wage equivalent to the marginal product of labor, how many
labors would be employed when the going wage rate is 10?
c. What happens to the number of labor demanded when the wage declines to 8?
d. How many labor would be demanded if wage remains 8, but the capital is increased to
100?
 2y
e. Find the cross partial derivative, .
 K L
8. Example: Two different demand functions are given:
Q1= 11 – 2p1-2p2 and Q2 = 16 – 2p1- 3p2
TC = 10 +4x+2y. Determine the quantities that maximize the profit of monopolist and also find
maximum profit.
9. Two different demand functions of discriminating monopoly are given:
p1= 140 – 7q1 and p2 = 90 – 0.4 q2/2

20
 Higher Order Differentiation and Its Applications

TC = 20 + 2q + 3q2 where q= q1+ q2, what price will the firm charge in two markets to maximize
profit?

Solution:

1. a. fxx= ; fyy = and fxy = fyx =

b.fxx = 294(7x + 3y); fyy = 54(7x + 3y) and fxy=fyx= 126 (7x +3y)

c. fxx= 30x((x3+5y)4+6x2(x3 +5y)3);fyy=50(x3 +5y)3 and fxy= 300 (x3 +5y)3x2

d. fxy= 2ey fxx=0.

e. fxx= - ; fyy= 2 and fxy= = fyx

f. fxx= 6x e2y; fyy= 4x3e2y and fxy= 6x2e2y

2. Derivatives
a. f1 = 6x1x2 + 24x1 + 5x2 + 20; f2 = 3x12 + 5x1 + 1
b. f11 = 6x2 + 24; f22 = 0; f12 = 6x1+ 5; and f21 = 6x1 + 5. Note both cross partial derivatives
are equal, as they should be, according to the Young's Theorem.
3. Answers:
 Qd  Qd  Qd
a.  0.05;  10;   10 Ps
Y  Pc  Ps
b. 1.12
c. - 0.56
d. 0.16
4. Apply Young’s theorm

5. Since and are both greater than zero. Hence, the commodities X1 and X2 are

competitive.
6. p1= 3.2 and p2= 3.9; e11= -1.7; e21= 0.8; e22= -.2 and e12 = 0.5.
7. Answers:

21
 Higher Order Differentiation and Its Applications

1
y  K  2 40
a.  5   1
L  L  2
L
b. L = 16
c. L = 25; labor demand increases with wage decline.
1
y  K  2 50
d.  5   1 = 8; i.e., L = 39.
L  L  2
L
1 1
 
e. 2.5L 2 K 2

8. x=1 and y=2; π = 8

9. p1= 110.52 and p2 = 85.52 and q= 13.17

6.References:
1. K. Sydaster and P. Hammond, Mathematics for Economic Analysis, Person Educational
Asia, Delhi, 2002.
2. M. Hoy et.al, Mathematics for Economics, PHI Learning Private Limited, Delhi, Second
Edition, 2001.
3. J.E. Draper and J.S. Klingman, Mathematical Analysis Bussiness and Economic
Applications, Harper & Row Publishers, New York, 1967.
4. Rosser, Mike, Basic Mathematics for Economists Second Edition, London, 2003.

22