Nature of Comparative Statics

 A comparative-static analysis can be either qualitative or quantitative in nature.

 If we are interested only in the question, whether an increase in investment I 0 will

increase or decrease the equilibrium income Y, the analysis will be qualitative

because the direction of change is the only matter considered.
 But if we are considered the magnitude of the change in income Y resulting from a
given change in I 0 (that is, the size of the investment multiplier), the analysis will

obviously be quantitative.
 By obtaining a quantitative answer, we may easy to identify the direction of change
from its algebraic sign.
 Hence the quantitative analysis always embraces the qualitative.
 It indicates the rate of change of the equilibrium value of an endogenous variable with
respect to the change in a particular parameter or exogenous variable.
 For this reason, the mathematical concept of derivative takes on preponderant
significance in comparative statics, because that concept is the branch of mathematics
known as differential calculus which directly connected with the notion of rate of
change.
 Therefore, to measure the rates of change of the equilibrium values of the variables in
a model, we may explain the rate of change of any variable y in response to a change
in another variable x.
 For considering two variables which are related to each other then the mathematical
function is defined as
y  f (x).
 Here, to explain comparative-static model, the variable y represents the equilibrium
value of an endogenous variable and x represents the some parameters.

1
 Difference Quotient
When the variable x changes from the value x 0 to a new value x 1 , the change is

measured by the difference x1  x 0 i.e., x  x1  x0 , where  is the Greek capital

delta, here it indicates difference. If we denote the function f(x) for various values of x ,

then it is denoted by f(xi ) when x  x i . Therefore, if we define f(x) as the function

f(x)  x 2  5 , then we have f(0)  0 2  5  5 ; and similarly, f(2)  2 2  5  9 , etc.

When x changes from an initial value x 0 to a new value ( x0  x) , the value of the

function y  f(x) changes from f(x0 ) to f(x0  Δx) . The change in y per unit of
change in x can be represented by the difference quotient.
y f ( x0  x)  f ( x0 )

x x
This quotient, which measures the average rate of change of y , can be calculated if we
know the initial value of x , or x 0 , and the magnitude of change in x , or x . That is,
Δy/Δx is a function of x 0 and x .

Example : f(x)  3x  4 , we can write f(x0 )  3( x 0 )  4 and

2 2

f(x0  x)  3(x 0  x) 2  4 .

Therefore, the difference quotient is

y 3( x0  x) 2  4  (3x02  4) 6 x0 x  3(x) 2

   6 x0  3x , (1)
x x x
which can be evaluated if we are given x 0 and x . Let x0  3 and x  4 ; then the

average rate of change of y is 6(3) + 3(4) = 30. This means that, on the average, as x

changes from 3 to 7, the change in y is 30 units per unit change in x.

2
 Derivative
If we interest to know the change of y when x is very small, in that situation, it is
possible to obtain an approximation of Δy/Δx by dropping all the terms in the
difference quotient involving the expression x . In equation (1), for instance, if x is
very small, we may simply take the term 6x0 on the right as an approximation of Δy/Δx

. The smaller value of x , it indicates that the closer is the approximation to the true
value of Δy/Δx . As x approaches zero, the term (6 x0  3x) indicates the value

6x0 which is equivalent to Δy/Δx of equation (1). That is, Δy/Δx  6x 0 as

Δx  0 then the equation is expressed as
y
lim  lim
x  0 x x0
(6 x0  3x)  6 x0 .

If Δx  0 is exists for the difference quotient Δy/Δx then the limit is called the

derivative of the function y  f (x). This notation may be expressed as dy / dx to

explain the rate of change of y with respect to x . Therefore, the derivative of a given

function y  f (x) is defined as

dy y
 f ( x)  lim .
dx x0 x

dy
Example : Let y  3x 2  4 then  6 x or f ( x)  6 x . Note that different
dx
values of x will give the derivative correspondingly different values. For instance, when
x  3 , we find, by substituting x  3 in the f (x) expression, that f (3)  6(3)  18
; similarly, when x  4 , we have f (4)  6(4)  24 . Thus, whereas f (x) denotes a

derivative function, the expressions f (3) and f (4) each represents a specific
derivative value.

3
  Derivative and Slope
Let us consider the total cost function C  f (Q) , where C represents the total cost and

Q represents the output. Then the marginal cost (MC) is defined as the change in total
cost resulting from a unit increase in output i.e., MC  C / Q . Here, Q indicates

an extremely small change, which is applicable for discrete cases. For continuous
variable, Q refer to an infinitesimal change. In this case, the marginal cost is measured

by the slope of the total-cost curve. The slope of the total-cost is nothing but the limit of
the ratio C / Q , when Q approaches zero.

In the above figure, we draw a total-cost curve. Here, we consider Q0 as the initial
output level from which an increase in output is measured; then the relevant point on the

4
cost curve is the point A. If output is to be raised to Q0  Q  Q2 , the total cost will
be increased from C0 to C0  C  C2 ; thus ΔC/ΔQ  (C 2  C0 )/(Q 2  Q0 ) . This

the ratio of two line segments, EB/ AE , or the slope of the line AB . It is the particular
ratio which measures an average rate of change or the average marginal cost for the
particular value of Q .

For smaller output increment (from Q0 to Q1 ), the average marginal cost may be
measured by the slope of the line AD. If we reduce the output increment further and
further, flatter and flatter lines (as Q  0 ) then we obtain the line KG as the relevant
line. The slope of KG (  HG / KH ) measures the slope of the total-cost curve at point
A and represents the limit of C / Q , as Q  0 , when initial output is at Q  Q0 .
If the initial output level is changed from Q0 to Q2 , then the point A shifts to the point
B and the slope of the curve indicates the point B i.e., f (Q2 ).

Rules of Differentiation

 Constant-Function Rule
Let us consider a constant function y  k , or f ( x)  k , is identically zero, i.e., is zero
for all values of x . Symbolically, for given y  f ( x)  k this rule may be explained by
the following derivative
dy dk
  0 or f ( x)  0 .
dx dx
This notation may be stated as
d d d
y f ( x)  k  0 .
dx dx dx

 Power-Function Rule
n1
The derivative of a power function y  f ( x)  x
n
is nx . Symbolically, this is
expressed as
d n
x  nx n1 or f ( x)  nx n1
dx

5
 dy d 3
Examples : Let us consider y  x , then
3
 x  3x 2 . Further, let us consider
dx dx
d
f ( x)  x , then f ( x)  x  1( x 0 )  1 and consider y  x ,
0
then
dx
d 0 d 3
x  0( x 1 )  0 or consider y  x 3 , then x   3x 4 or consider
dx dx
d 1/ 2 1 1/ 2 dy
y  x  x1/ 2 , then x  x consider y  2 x , then  2x0  2 .
dx 2 dx

 Sum-Difference Rule
The derivative of a sum (difference) of two functions is the sum (difference) of the
derivatives of the two functions :
d
 f ( x)  g ( x)  d f ( x)  d g ( x)  f ( x)  g ( x)
dx dx dx

dy
Examples : Let us consider y  14x , then
3
 42x 2 . Further, we consider
dx
14 x 3  5x 3  9 x 3 , then there we also consider f ( x)  5 x 3 and g ( x)  9 x 3 .
Therefore, according to the sum rule, we then have
dy d d d
 (5 x 3  9 x 3 )  5 x 3  9 x 3  15 x 2  27 x 2  42 x 2
dx dx dx dx
which is identical with earlier result.
Further, the function, y  14x can be written as y  2 x  13x  x . Then the
3 3 3 3

derivative of this function is explained as

dy d
 (2 x 3  13x 3  x 3 )  6 x 2  39 x 2  3x 2  42 x 2
dx dx
which is also identical with previous result. Again, this derivative rule may be explained
by
d
(ax 2  bx  c)  2ax  b .
dx
In this derivative the constant term `c’ does not produce any effect, because the derivative
of a constant term is zero. Here it is also clear that the multiplicative constants `a’ and `b’
are retained during differentiation, and the additive constant drops out.

6
  Application of derivative in the Field of Economics
To justify the fixed cost of a firm does not affect its marginal cost, we consider following
short-run total-cost function
C  Q 3  4Q 2  10Q  75
where C indicates total-cost of production function and Q indicates the production of
the firm. In this function additive constant 75 indicates fixed cost of the firm. Here the
marginal cost function (for infinitesimal output change) is the limit of the quotient
C / Q , or the derivative of the C function is
dC
 3Q 2  8Q  10
dQ
In this derivative, the additive constant 75 is dropped out. So, the magnitude of the fixed
cost obviously cannot affect the marginal cost. In general, if a primitive function
dy
y  f (x) represents a total function, then the derivative function is its marginal
dx
function.

 Product Rule
The derivative of the product of two (differentiable) functions is equal to the first
function times the derivative of the second function plus the second function times the
derivative of the first function.
d
 f ( x) g ( x)  f ( x) d g ( x)  g ( x) d f ( x)
dx dx dx
 f ( x) g ( x)  g ( x) f ( x) .
Example : Let us consider a function y  (2 x  3)(3x ) ,
2
here we also consider

f ( x)  2 x  3 and g ( x)  3x 2 . Then we have f ( x)  2 and g ( x)  6 x .

Therefore,
d
dx
 
(2 x  3)(3x 2 )  (2 x  3)(6 x)  (3x 2 )(2)  18 x 2  18 x .

This result can be checked by first multiplying out f ( x) g ( x) and then taking the
derivative of the product polynomial. The product polynomial is in this case
f ( x) g ( x)  (2 x  3)(3x 2 )  6 x 3  9 x 2 , and direct differentiation does yield the
same derivative, 18 x  18 x .
2

7
  Application of derivative in the Field of Economics
To find the marginal-revenue function from average-revenue function, we consider
following average-revenue (AR) function
AR 15  Q ,
the marginal-revenue (MR) function can be found by first multiplying AR by Q to get the
total-revenue (R) function which is explained as
R  AR . Q  (15  Q)Q  15Q  Q 2 .
Now the differentiation of R with respect to Q then we have
dR
MR   15  2Q .
dQ
Here we consider a general function, AR  f (Q) then the total-revenue function is
defined as
R  AR . Q  f (Q) . Q .
Therefore, R is a product of two functions of Q, namely, f (Q) and Q itself. Thus we
can differentiate R to get the MR function as follows :
dR
MR   f (Q) . 1  Q . f (Q)  f (Q)  Q f (Q)
dQ
dR
or MR   f (Q) . 1  Q . f (Q)  AR  Q f (Q)
dQ
i.e., MR  AR  MR  f (Q)  Q f (Q) .
In this expression Q f (Q) , Q denotes the output and f (Q) represents the slope of
the AR curve plotted against Q, then we have
R PQ
AR    P,
Q Q
where P  f (Q) . That is, “average revenue” and “price” are different names but they
are same thing.
Under pure competition, MR-AR=0 because 𝑓 ′ 𝜃 = 0, i.e. the MR curve and AR curve
must coincide. Under imperfect competition the AR curve is normally downward-
sloping, 𝑓 ′ 𝜃 < 0, MR-AR<0 for all positive levels of output. In this cause, the MR
curve must lie below the AR curve.

8
  Quotient Rule

The derivative of the quotient of two functions, f ( x) / g ( x) , is

d f ( x) f ( x) g ( x)  f ( x) g ( x)
 .
dx g ( x) g 2 ( x)
Example :
d  2 x  3  2( x  1)  (2 x  3)(1) 5
1.   
dx  x  1  ( x  1) 2
( x  1) 2

d  ax 2  b  2ax(cx)  (ax 2  b)(c) c(ax 2  b) ax 2  b

2.    
dx  cx  (cx) 2 (cx) 2 cx 2

 Economic Application of Quotient Rule (Marginal Cost and

AverageCost)

To identify the relationship between Marginal-Cost and Average-Cost, let us consider the
rate of change of average cost when output varies. Suppose a total-cost function C=C(Q),
the average-cost (AC) function is a quotient of two functions of Q, since
AC  C (Q) / Q , defined as long as Q  0 . Therefore, the rate of change of AC with
respect to Q can be found by differentiating AC:

d C (Q) C (Q) . Q  C (Q) . 1 1  C (Q) 

   
Q  Q 
C (Q ) . (1)
dQ Q Q2
d C (Q) C (Q)
Then 0 if C (Q) , (2)
dQ Q Q
where C (Q) represents the marginal-cost (MC) function, and C (Q) / Q represents the
AC function. The economic meaning of (2) is that the slope of the AC curve will be
positive, zero, or negative if and only if the marginal-cost curve lies above, intersects, or
lies below the AC curve. In the following figure, the MC and AC functions plotted are
based on the specific total-cost function which is shown below:
C  Q 3  12Q 2  60Q (3)

9
C
MC=3Q2-24Q+60

AC=Q2-12Q+60

0 6
Q

To the left of this figure, Q = 6 indicates AC is declining and MC lies below it; to the
right these curves indicate opposite direction. Further, at Q = 6, AC has a slope of zero,
and MC and AC have the same value. Again, we interpret C (Q) as any other
differentiable total function with C (Q) / Q and C (Q) as its corresponding average and
marginal functions. Thus this result gives us a general marginal-average relationship.

 Chain Rule
If we have a differentiable function z  f ( y) , where y is in turn a differentiable function
of another variable x , say, y  g (x) , then the derivative of z with respect to x is
equal to the derivative of z with respect to y with respect to x . Expressed symbolically,
dz dz dy
  f ( y ) g ( x) .
dx dy dx
This rule is known as the chain rule. This rule may extent for three or more functions. If
we have z  f ( y), y  g ( x) , and x  h(w) , then
dz dz dy dx
  f ( y ) g ( x)h( w) .
dw dy dx dw

10
 dz
Example : Let us consider a function, z  ( x  3x  2) . To identify
2 17
the chain
dx
rule is considered by defining a new variable y  x  3x  2 , and then we have
2

z  y17 . Therefore, this differentiation can be defined as

dz dz dy
  17 y16 (2 x  3)  17( x 2  3x  2)16 (2 x  3) .
dx dy dx

Example : Let us consider a total-revenue function of a firm, R  f (Q) , where

output Q is a function of labour input L , or Q  g (L) . To find the marginal-revenue-
product-of-labour ( MRPL ) the chain rule is defined as
dR dR dQ
  f (Q) g ( L) ,
dL dQ dL
dR dQ
where is the marginal revenue (MR) function and is the marginal-physical-
dQ dL
product-of-labour ( MPPL ) function. Therefore, MRPL  MR . MPPL .

 Partial Differentiation
Hitherto, we have considered only the derivatives of functions of a single independent
variable. However, different comparative-static models analysis appear several types of
parameters, so that the equilibrium value of each endogenous variable may be a function
of more than one parameter. Therefore, as a final preparation for the application of the
concept of derivative to comparative statics, we must learn how to find the derivative of a
function of more than one variable.

 Partial Derivatives
Let us consider a function
y  f ( x1 , x2 , ...., xn ) , (4)
where the variables xi (i  1, 2, 3, ......,n) are all independent of one another, so that
each can vary by itself without affecting the others. If the variable x1 undergoes a change
xi while x2 , ...., xn all remain fixed, there will be a corresponding change in y ,
namely, y . The difference quotient in this case can be expressed as

11
 y f ( x1  x1 , x2 , ...., xn )  f ( x1 , x2 , ...., xn )
 .
xi x1
y
If we take the limit of as x1  0 , that limit will constitute a derivative. We call it
x1
the partial derivative of y with respect to x1 , to indicate that all the other independent
variables in the function are held constant when taking this particular derivative. Similar
partial derivatives can be defined for infinitesimal changes in the other independent
variables. The process of taking partial derivatives is called partial differentiation.

Partial derivatives are assigned distinctive symbols. In lieu of the letter d (as in dy / dx
), we employ the symbol  . Thus we shall now write y / x , which is read: “the partial
derivative of y with respect to xi ”.

 Economic Application of Partial Derivatives

Let us consider a production function Q  Q( K , L) , where Q, K , and L denote
output, capital input, and labour input, respectively. This function is a particular two-
variable version of (4), with n  2 . We can therefore define two partial derivatives
Q / k (or Qk ) and Q / L (or QL ). The partial derivative Qk relates to the rates of
change of output with respect to infinitesimal changes in capital, while labour input is
held constant. Thus Qk symbolizes the marginal-physical-product-of-capital ( MPPk )
function. Similarly, the partial derivative QL is the mathematical representation of the
MPPL function.

 Applications to Comparative-Static Analysis

Equipped with the knowledge of the various rules of differentiation, we can at last tackle
the problem posed in comparative-static analysis: namely, how the equilibrium value of
an endogenous variable will change when there is a change in any of the exogenous
variables or parameters.

12
 National-Income Model

Let us consider a national-income model with three endogenous variables, Y (national

income), C (consumption), and T (taxes).
Y  C  I 0  G0
C     (Y  T ) (  0, 0    1) (5)
T    Y (  0, 0    1)
where I 0 represents the investment and G0 represents the government expenditure.
 The first equation in this system gives the equilibrium condition for national income.
 The second and third equations show, respectively, how C and T are determined in
the model.
 The restrictions on the values of the parameters  ,  ,  and  are explained as
  is positive because consumption is positive even if disposable income
(Y  T ) is zero;
  is a positive fraction because it represents the marginal propensity to
consume;
  is positive because even if Y is zero the government will still have a positive
tax revenue (from tax bases other than income);
  is a positive fraction because it represents an income tax rate, and as such it
cannot exceed 100 percent.
 The exogenous variables I 0 (investment) and G0 (Government expenditure) are
nonnegative.
 All the parameters and exogenous variables are assumed to be independent of one
another.

*
For solving this model, we consider Y by substituting the third equation of (5) into the
second and then substituting the resulting equation into the first. The equilibrium income
(in reduced form) is
    I 0  G0
Y*  . (6)
I    

13
Similar equilibrium values can also be found for the endogenous variables C and T, but
we shall concentrate on the equilibrium income. From the equation (6) following three
equations are obtained :
Y * 1
 0 (7)
G0 1    
Y * 
 0 (8)
 1    
Y *   (    I 0  G0 )  Y *
  0 (9)
 (1     ) 2 1    

 The equation (7) gives us the government-expenditure multiplier. It is positive term

because  is less than 1, and  is greater than zero. If we have numerical values
for the parameters  and  , we can easy to estimate the numerical value of this
multiplier. Further, the equation (8) indicates nonincome-tax multiplier, because it
shows how a change in  , the government revenue from nonincome-tax sources,
will affect the equilibrium income. Again, the equation (9) which is not in the nature
a multiplier, because it does not relate a dollar change to another dollar change as the
equation (7) and (8).

 Total Differentials
The concept of differentials can easily be extended to a function of two or more
independent variables. Consider a saving function
S  S (Y , i) (10)
where S is savings, Y is national income, and i is the interest rate. Here, the partial
S
derivative measures the marginal propensity to save. Thus, for any change in Y ,
Y
dY , the resulting change in S can be approximated by the quantity (S / Y )dY .
Similarly, given a change in i , di , we may take ( S / i)di as the approximation to the
resulting change in S . The total change in S is then approximated by the differential
S S
dS  dY  di .
Y i

14
Then it may be written as
dS  S y dY  Si di .
Note that the two partial derivatives S y and S i again play the role of “converters” that

serve to convert the changes dY and di , respectively, into a corresponding change dS .

The expression dS , being the sum of the approximate changes from both sources, is
called the total differential of the saving function. Therefore, total differential is called
total differentiation. If it is happened that Y may change while i remains constant then
we may write di  0 , and the total differential will reduce to dS  ( S / Y ) dY which
may be written as
S  dS 
  .
Y  dY  t cos ns tan t

The more general case of a function of n independent variables can be exemplified by,
say, a utility function in the general form
U  U ( x1 , xs , ...., xn ) (11)
The total differential of this function can be written as
U U U
dU  dx1  dx2  ........ dxn
x1 x2 xn
(12)
n
or dU  U1dx1  U 2 dx2  ........ U n dxn  U i dxi
i 1

in which each term on the right side indicates the approximate change in U resulting from
a change in one the independent variables.
 Economically, the first term, U1dx1 , means the marginal utility of the first
commodity times the increment in consumption of that commodity, and similarly
for the other terms.
 The sum of these, dU , thus represents the total approximate change in utility
originating from all possible sources of change.

15
Like any other function, the saving function (10) and the utility function (11) can both be
expected to give rise to point-elasticity measurement. Each elasticity measurement
indicates the change in one of the independent variables only which is measured for
saving function and utility function. These are accordingly called partial elasticities. For
saving function, the partial elasticities may be defined as
S / Y S Y S / i S i
 SY   and  Si   .
S /Y Y S S /i i S
For the utility function, the n partial elasticities can be concisely denoted as follows :
U xi
Ux  (i  1, 2, .....,n) .
i xi U

Example : Find the total differential for the following utility functions, where a, b  0 :
(a) U ( x1 , x2 )  ax1  bx2

(b) U ( x1 , x2 )  x1  bx2  x1 x2
2 3

a b
(c) U ( x1 , x2 )  x1 x2

The total differentials are as follows :

U U
(a)  U1  a and U2  b
x1 x2
then dU  U1dx1  U 2 dx2  a dx1  b dx2 .

U U
(b)  U1  2 x1  x2 and  U 2  3x22  x1
x1 x2
then dU  U1dx1  U 2 dx2  (2 x1  x2 ) dx1  (3x2  x1 ) dx2 .
2

U n1 b ax1a x2b U bx1a x2b

(c)  U1  ax1 x2  and
a b1
 U 2  bx1 x2 
x1 x1 x2 x2
ax1a x2b bx a x b
then dU  U1dx1  U 2 dx2  dx1  1 2 dx2 .
x1 x2
You have done the Exercise 8.2 of the book of Alpha C. Chiang
except the question No. 1 and 5.

16
Rules of Differentials

A straightforward way of finding the total differential dy , given a function

y  f ( x1 , x2 )
is to find the partial derivatives f1 and f 2 , and substitute these into the equation
dy  f1dx1  f 2 dx2 .
But sometimes it may be more convenient to apply certain rules of differentials which are
shown below:
Let k be a constant and u and v be two functions of the variables x1 and x 2 , here we
assume that u and v are independent . Then the following rules are considered
Rule I : dk  0 (constant-function rule).

Rule II : d (cu n )  cnu n1du (power-function rule).

Rule III : d (u  v)  du  dv (sum-difference rule).
Rule IV : d (uv)  v du  u dv (product rule).
u 1
Rule V : d ( )  2 (v du  u dv) (quotient rule).
v v
Rule VI : d (u  v  w)  du  dv  dw
Rule VII : d (uvw)  vw du  uw dv  uv dw , if we consider z  vw , then
we may write d (uvw)  d (uz )  z du  u dz . Here, dz  d (vw)  w dv  v dw ,
it is substituted into the preceding equation, we may have
d (uvw)  vw du  u(w dv  v dw)  vw du  uw dv  uv dw .

Total Derivatives

Let us consider any function y  f ( x, w) , where x  g (w)

y x w

17
The two functions f and g can also be combined into a composite function
y  f g (w), w . (13)
The relationships among three variables y, x, and w are shown in this figure. Form
this figure it is clear that w affect the y through separate channels: (i) indirectly, via the
function g and then f , and (ii) directly, via the function f . The direct effect can simply
be represented by the partial derivative f w . But the indirect effect can only be expressed
dx y dx
by a product of two derivatives, f x , or , by the chain rule for a composite
dw x dw
function. Adding up the two effects gives us the desired total derivative of y with respect
to w :
dy dx
 fx  fw
dw dw (14)
y dx y
 
x dw w
dy
Example : Find the total derivative , given the function
dw
y  f ( x, w)  3x  w2 where x  g (w)  2w2  w  4
By virtue of (14), the total derivative should be
dy
 3(4w  1)  (2w)  10w  3
dw
Example : If we have a utility function U  U (c, s) , where c is the amount of coffee
consumed and s is the amount of sugar consumed, and another function s  g (c)
indicating the complementarity between these two goods, then we can simply write the
composite function
U  U (c, g (c))
from which it follows that
dU U U
  g (c) .
dc c g (c)

18
The situation is only slightly more complicated when we have
 x1  g ( w)
y  f ( x1 , x2 , w) where  (15)
 x2  h( w)
This relationship for four variables y, x1 , x2 and w can be shown by the following
figure

x1

y w

x2

From this figure it is clear that the variable w can affect y through three channels: (i)
indirectly, via the function g and then f , (ii) again indirectly, via the function h and
then f , and (iii) directly via f . These three effects are expected to the expressible,

y dx1 y dx2 y
respectively as , , and . By adding these together, we get
x1 dw x2 dw w
the total derivative
dy y dx1 y dx2 y
  
dw x1 dw x2 dw w
dx1 dx
 f1  f2 2  fw
dw dw

19
Example : Let the production function be Q  Q(k , L, t ) , where aside from the two
inputs K and L , there is a third argument t , denoting time. The presence of the t
argument indicates that the production function can shift over time in reflection of
technological changes. Thus this is a dynamic rather than a static production function.
Since capital and labour, too, can change over time, we may write
K  K (t ) and L  L(t ) .
Then the rate of change of output with respect to time can be expressed by
dQ Q dK Q dL Q
  
dt K dt L dt  t
 Qk K (t )  QL L(t )  Qt

Jacobian Determinants
The partial derivatives also provide a means of testing whether the given functional
relationships are dependent or not in the set of n functions in n variables. This is related
to the notion of Jacobian determinants. Consider the two functions
y1  2 x1  3x2
y2  4 x12  12 x1 x2  9 x22
Then we have following four partial derivatives

y1 y1 y2 y2

 2,  3,  8 x1  12 x2 and  12 x1  18 x2 .
x1 x2 x1 x2
These partial derivatives we arrange into a square matrix in a prescribed order and it
makes a matrix. This matrix is called a Jacobian matrix which denoted by J , and then
take its determinant is known as a Jacobian determinant, denoted by J :

y1 y1
x1 x2 2 3
J  
y 2 y 2 (8 x1  12 x2 ) (12 x1  18 x2 )
x1 x2
 (24 x1  36 x2 )  (24 x1  36 x2 )  0
That is, the Jacobian vanishes for all values of x1 and x 2 . This result indicates given two
functions are dependent.

20
Differentials and Point Elasticity

For economic application of differentials, let us consider the notion of the elasticity of
demand function, Q  f (P) . For this function the elasticity is defined as
(Q / Q) /(P / P) . Using this idea, we can replace the independent change P and
dependent change Q with the differentials dP and dQ , respectively, to get an
approximation elasticity measure known as the point elasticity of demand and denoted by
d :
dQ / Q dQ / dP
d  
dP / P Q/ P
Here, dQ / dP indicates the marginal function and Q / P indicates the average function
of the demand function. Therefore, the point elasticity of demand  d is explained by the
ratio of the marginal function to the average function of the demand function.

Indeed, this last-described relationship is valid not only for the demand function but also
for any other function, because for any given total function y  f (x) we can write the
formula for the point elasticity of y with respect to x as
dy / dx m arg inal function
 yx  
y/ x average function

The absolute value of elasticity is measured in deciding whether the function is elastic at
a particular point. In the case of demand function, we stipulate :
elastic 
 
The demand is of unit elasticity  at a po int when  d  1 .
inelastic 
 

Example : Find  d if the demand function is Q  100  2P .

The marginal function and the average function of the given demand function are
dQ Q 100  2 P
2 and  ,
dP P P

21
so the ratio will give us
P
d  .
50  P
This elasticity indicates the function of P. A specific price is chosen, however, the point
elasticity will be determinate in magnitude. For P  25 , we have  d   1 , or  d  1,
so that the demand elasticity is unitary at that point. When P  30 , we have  d  1.5 ;
which indicates demand is elastic at that price.

Example : Find the point elasticity of supply  s from the supply function

Q  P 2  7 P , and determine whether the supply is elastic at P  2 .

This supply function explains the following marginal and average functions respectively,
dQ Q
 2P  7 and  P 7,
dP P
their ratio gives us the elasticity of supply
2P  7
S  .
P7
When P  2 , this elasticity has the value 11 9  1; thus the supply is elastic at P  2 .

Example : Given the demand equation

Q  150  15P ;
compute elasticity when P  4 and P  8 .
When P  4 ,
Q  150  15(4)  90
dQ
  15
dP

22
 dQ P
D .
dP Q
4 2
 (15) . 
90 3

When P  8 ,
Q  150  15(8)  30
dQ
  15
dP
dQ P
D .
dP Q
8
 (15) . 4
30
We are normally interested in elasticity of demand because we wish to know how total
expenditure will change when the price of a commodity rises or falls. The relationship
among expenditure, price changes and elasticity can be shown as follows. If E denotes
expenditure, P indicates price and Q explains quantity, then we have E  P . Q . To
find the effect upon expenditure when there is a change in price, we differentiate the
function with respect to P , using the product rule :
dE dQ
Q .P
dP dP
 dQ P 
 Q 1  . 
 dP Q 
dQ P dE
But since D  . , we have  Q (1  D) .
dP Q dP
dE
If D   1, is negative, then it indicates expenditure will fall when price rises and
dP
rise when price falls. If D   1, there is no change in expenditure when the price
dE
changes. If D   1, is positive, then it explains expenditure will rise when price
dP
rises and fall when price falls.

23
24