Economics of Production
Economics of Production
Economics of Production
produced by one unit of x1, when all other input quantities are held constant and only the
1
Concept of iso-quant and rate of technical substitution (RTS)
In case of two input case, production function is Q = f (x1, x2), and production function for a
specific iso-quant is Q0 = f (x1, x2). Along a definite iso-quant curve, dQ = 0. Hence, f1dx1 +
𝑑𝑥2 𝑓
f2dx2 = 0 and
𝑑𝑥1
= – 𝑓1 = RTS
2
In two-input case, cost equation is, C = r1x1 + r2x2 + b and equation form of isocost line is,
C0 = r1x1 + r2x2 + b.
𝐶0 − 𝑏 𝑟1 𝑥1 𝑑𝑥2 𝑟
Or, x2 = − . Hence, slope of iso-cost line = = − 𝑟1 = ratio of input prices.
𝑟2 𝑟2 𝑑𝑥1 2
2
𝑑 𝑑𝐾 𝑑2𝐾 𝑑 𝛽𝐾 𝛽 𝑑 𝐾
Now, ( )= = (− )=− ( )
𝑑𝐿 𝑑𝐿 𝑑𝐿2 𝑑𝐿 𝛼𝐿 𝛼 𝑑𝐿 𝐿
𝑑𝐾 𝑑𝐿 𝛽𝐾
𝛽 (𝐿 𝑑𝐿 ‒ 𝐾 𝑑𝐿) 𝛽 {𝐿 (‒ 𝛼𝐿 ) ‒ 𝐾}
=− =−
𝛼 𝐿2 𝛼 𝐿2
𝛽𝐾 𝛽 𝛽𝐾
=
𝛼𝐿2
( 𝛼 + 1) =
𝛼 2 𝐿2
(α + β) ˃ 0 (K ˃ 0; L ˃ 0; α ˃ 0; β ˃ 0)
3
Characteristic features of homogeneous production functions
a) These production functions give rise to downward sloping and convex to the origin
indifference curves.
b) In case of these functions, expansion paths are straight lines passing through the origin.
c) In case of these production functions, ridge lines are also straight lines passing through the
origin.
Marginal and average products of capital and labour are homogeneous of degree α+β–1
for unspecified values of α and β and are homogeneous of degree zero when α + β = 1
Given Q = A Kα Lβ ;
𝜕Q
MPK = = α A Kα – 1Lβ ,
∂K
Exponents of the variables express output elasticity with respect to those variables
Given Q = A K αL β
𝜕Q 𝐾𝛼 𝐿𝛽 𝛼𝑄 𝜕𝑄 𝐾𝛼 𝐿 𝛽 𝛽𝑄
∂K
= α A K α – 1L β = α A = and,
𝜕𝐿
= β A K αL β–1 = β A =
𝐿
𝐾 𝐾 𝐿
𝜕𝑄 𝐾 𝛼𝑄 𝐾
Now, elasticity of output with respect to capital (K) = EQK = = =α
𝜕𝐾 𝑄 𝐾 𝑄
𝜕𝑄 𝐿 𝛽𝑄 𝐾
And, elasticity of output with respect to labour (L) = EQL = = =β
𝜕𝐿 𝑄 𝐾 𝑄
4
Exponents of each variable show relative share of that input in total output
Given Q = A K α L β
𝑀𝑃𝐾 𝑥 𝐾 𝛼𝑄 𝐾
Capital’s share in total output = = = α (constant)
𝑄 𝐾 𝑄
𝑀𝑃𝐿 𝐿 𝛽𝑄 𝐾
Similarly, labour’s share in total output = = = β (constant)
𝑄 𝐾 𝑄
Euler’s Theorem
Leonard Euler, a Swiss mathematician developed this theorem at the end of 1930s. The
theorem played a major role in the development of ‘Marginal Productivity Theory of
Distribution’. The theorem is a mathematical proposition which states that if a production
function is homogeneous of degree one (i.e. it exhibits constant returns to scale) and the
participating inputs are paid in accordance to their marginal productivity; then, while making
payments to inputs, total product will be fully exhausted with no surplus and deficit.
b) If production function is homogeneous of degree one and inputs are paid at the rate of their
MPs, then monetary value of output will be equal to total cost of production.
As we have seen, Q = MPK.K + MPL.L
Or, PQ = P. MPK.K + P.MPL.L (dividing both sides by ‘P’)
= VMPK.K + VMPL.L (VMPK= Value of MPK ; VMPL = Value of MPL)
= rK + wL = TC
5
According to (b) PQ = wL + rK and, TC = wL + rK, hence, TR = TC ; that is, ∏ = 0
d) Under perfect competition, factor shares in total output equal total output.
𝑤𝐿 𝑟𝐾
Labour’s share in total output = , and capital’s share in total output =
𝑃𝑄 𝑃𝑄
𝑤𝐿 𝑟𝐾 𝑤𝐿 + 𝑟𝐾 𝑃𝑄
Hence, combined share of labour and capital = + = = =1
𝑃𝑄 𝑃𝑄 𝑃𝑄 𝑃𝑄
a) In case of constant returns to scale, α + β = 1; hence, total output will be fully exhausted in
making payments to inputs used.
b) In case of increasing returns to scale, α + β ˃ 1; hence, total payments would exceed
output. That is total output will fall short while making payments to inputs used.
c) In case of decreasing returns to scale, α + β < 1, hence, total payments would be less than
output. That is, while making payments to inputs used, there will be surplus retained.
𝛼𝑄 𝛽𝑄
In case of Q = A K αL β ; fk = and fL = 𝐿
.
𝐾
6
𝛽𝑄
𝑓𝐿 𝐿 𝛽𝑄 𝐾 𝛽𝐾
Therefore, = 𝛼𝑄 = =
𝑓𝑘 𝐿 𝛼𝑄 𝛼𝐿
𝐾
𝐾
𝐾 𝛼 𝑓𝐿 𝑑 𝛼
𝐿
Or,
𝐿
=
𝛽 𝑓𝑘
Therefore, 𝑓𝐿 =
𝑑𝑓 𝛽
𝐾
𝐾 𝑓𝐿 𝛽𝐾
𝑑𝐿 𝑓𝐾 𝛼 𝛼𝐿 𝛼 𝛽𝐾 𝐿
Hence, σ = 𝑓 x 𝐾 = x 𝐾 = =1
𝑑 𝐿 𝛽 𝛽 𝛼𝐿 𝐾
𝑓𝐾 𝐿 𝐿
Expansion path is an upward sloping straight line passing through the origin
𝑀𝑃𝐾 𝑃𝐾
According to producers’ equilibrium condition, =
𝑀𝑃𝐿 𝑃𝐿
𝛼𝑄 𝛽𝑄
Given Q = A K α L β ; MPK = and MPL =
𝐾 𝐿
Let, PK = r and, PL = w
𝛼𝑄
𝑀𝑃𝐾 𝐾 𝛼𝑄 𝐿 𝑃𝐾 𝑟
Hence, = 𝛽𝑄 = x = =
𝑀𝑃𝐿 𝐾 𝛽𝑄 𝑃𝐿 𝑤
𝐿
𝛼𝑄𝐿 𝑟 𝛽𝑟 𝛼𝑤
Therefore, = Or, L = K and K = L (β ˃ 0; α ˃ 0; r ˃ 0; w ˃ 0)
𝛽𝑄𝐾 𝑤 𝛼𝑤 𝛽𝑟
Hence, ‘L’ is a positive and linear function of ‘K’ and vice versa. Moreover, both equations
have no intercept term. That is, in case of CDPF, the expansion path is an upward sloping
straight line passing through the origin for any degree of homogeneity.
𝛽𝑟 𝛼𝑤
Therefore, L = K , and K= L
𝛼𝑤 𝛽𝑟
𝛽𝑟
Now, cost function is C = wL + rK = w K + rK = rK ( 𝛼+𝛽
𝛼 )
𝛼𝑤
7
𝛼𝐶 𝛽𝐶
Therefore, K = , and L =
𝑟(𝛼+ 𝛽) 𝑤(𝛼+ 𝛽)
𝛼𝐶 𝛼
𝛽𝐶 𝛽 𝐴𝛼 𝛼 𝛽𝛽 𝐶 𝛼+𝛽
Q = A(𝑟(𝛼+ ) ( ) = 𝛼 𝛽
𝛽) 𝑤(𝛼+ 𝛽) 𝑟 𝑤 (𝛼+𝛽)𝛼+𝛽
𝑟 𝛼 𝑤 𝛽 (𝛼+𝛽)𝛼+𝛽
Or, Cα + β = Q
𝐴𝛼𝛼 𝛽 𝛽
1 1
1 1
𝑟𝛼𝑤𝛽 𝛼+𝛽 𝑟𝛼𝑤𝛽 𝛼+𝛽
Or, C = (𝛼 + 𝛽) ( ) (𝑄 ) 𝛼+𝛽 Or, C = 𝑚𝑄 𝛼+𝛽 (𝛼 + 𝛽) ( 𝛽) =m
𝐴𝛼𝛼 𝛽𝛽 𝐴𝛼𝛼𝛽
1 1
1 − 𝛼−𝛽 1 − 𝛼−𝛽
𝐶 𝑚𝑄𝛼+𝛽 𝑚𝑄𝛼+𝛽 𝑑𝐶
AC =
𝑄
=
𝑄
=
𝑄
= 𝑚𝑄 𝛼+𝛽 ; MC =
𝑑𝑄
= (𝛼+1 𝛽) 𝑚𝑄 𝛼+𝛽
1
= (𝛼+ 𝛽) AC
1
𝑇𝐶 𝑚𝑄 𝑑𝐶
i) Now, if, α + β = 1; TC = 𝑚𝑄 𝛼+𝛽 = mQ ; AC = = = m ; MC = =m
𝑄 𝑄 𝑑𝑄
Here, α, β, r, w, all being positive, m is also positive. That is, TC is a positive and linear
function of Q. Moreover, it has no intercept term. That is, in case of α + β = 1, the total cost
curve is an upward sloping straight line passing through the origin. Furthermore, in this case,
average cost, AC = marginal cost, MC = m. That is, AC and MC are not only constants, they
are also equal; they, therefore, coincide with each other and are parallel to quantity axis.
C,
MC, AC TC
AC = MC = m
O Q
1
ii) When, α + β > 1, following the relation, MC = AC, we get, MC < AC. In this case,
(𝛼+ 𝛽)
TC, AC and MC all are non-linear. TC is upward sloping and bending towards x-axis. MC
and AC are downward sloping and MC lies below AC.
TC AC, MC
TC
AC
MC
O Q O Q
8
1
iii) When, α + β < 1, following the relation, MC = AC, we get, MC < AC. In this case
(𝛼+ 𝛽)
also TC, AC and MC all are non-linear. TC is upward sloping and bending towards y-axis.
MC and AC are also upward sloping and MC lies above AC.
TC AC, MC MC AC
O Q O Q
In actual form, the three curves (TC, AC and MC) are as follows:
TC TC AC, MC MC AC
O Q O Q
Using above result, supply function corresponding to CDPF can be derived as follows:
𝛼𝑃 𝛼 𝛽𝑃 𝛽 𝛼 𝛼 𝛽 𝛽 𝛼 𝛼 𝛽 𝛽
𝑄 = 𝐴𝐾 𝛼 𝐿𝛽 = 𝐴 ( ) ( ) = 𝐴 ( ) ( ) 𝑃𝛼+𝛽 = 𝐴 ( ) ( ) 𝑃 (when α +β = 1).
𝑟 𝑤 𝑟 𝑤 𝑟 𝑤
That is, under constant returns to scale, supply curve is an upward sloping straight line
passing through the origin.
9
Criticisms of C-D Production Function:
The C-D production function has been criticizd by Arrow, Chenery, Minhas and Solow as
discussed below:
a) The C-D production function considers only two inputs, labour and capital, and neglects
some important inputs, like raw materials, which are used in production. It is, therefore, not
possible to generalize this function to more than two inputs.
b) In the C-D production function, the problem of measurement of capital arises because it
takes only the quantity of capital available for production. But the full use of the available
capital can be made only in periods of full employment. This is unrealistic because no
economy is always fully employed.
c) The C-D production function is criticized because it shows constant returns to scale. But
constant returns to scale are not an actuality, for either increasing or decreasing returns to
scale are applicable to production. Moreover, it is not possible to change all inputs to bring a
proportionate change in the outputs of all the industries. Some inputs are scarce and cannot be
increased in the same proportion as abundant inputs. On the other hand, inputs like machines,
entrepreneurship, etc. are indivisible. Thus when the supply of inputs is scarce and
indivisibilities are present, constant returns to scale are not possible. Whenever the units of
different inputs are increased in the production process, economies of scale and specialization
lead to increasing returns to scale.
d) The C-D production function is based on the assumption of substitutability of factors and
neglects the complementarity of factors.
e) This function is based on the assumption of perfect competition in the factor market which
is unrealistic. If, this assumption is dropped, then, the coefficients α and β do not represent
factor shares.
f) One of the weaknesses of C-D function is the aggregation problem. This problem arises
when this function is applied to every firm in an industry and to the entire industry. In this
situation, there will be many production functions of low or high aggregation. Thus the C-D
function does not measure what it aims at measuring.
Thus the practicability of the C-D production function in the manufacturing industry is a
doubtful proposition. This is not applicable to agriculture where for intensive cultivation,
increasing the quantities of inputs will not raise output proportionately. Even then, it cannot
be denied that constant returns to scale are a stage in the life of a firm, industry or economy.
It is another thing that this stage may come after some time and for a short while.
10
Significance
Despite these criticisms, the C-D function is of much importance.
a) It has been used widely in empirical studies of manufacturing industries and in inter-
industry comparisons. (b) It is used to determine the relative shares of labour and capital in
total output. (c) It is used to prove Euler’s Theorem. (d) Its parameters α and β represent
elasticity coefficients that are used for inter-sectoral comparisons. (e) This production
function is homogeneous of degree one; which shows constant returns to scale. If α + β > 1,
there is increasing returns to scale and if α + β < 1, there is diminishing returns to scale.
1
−
= A [λ−𝜌 {𝛿𝐾 −𝜌 + (1 − 𝛿)𝐿−𝜌 ] 𝜌
1
−
= λ(−𝜌)( ̶ 1/𝜌)
A [𝛿𝐾 −𝜌 + (1 − 𝛿)𝐿−𝜌 ] 𝜌 = 𝜆Q (proved)
𝜕𝑄 1 ̶ 1/𝜌 −1
MPK = = (− )A 𝑍 .δ. ( ̶ ρ) 𝐾 −𝜌 −1 = δ A 𝑍 ̶ 1/𝜌 −1
𝐾 −(𝜌+1)
𝜕𝐾 𝜌
11
̶ 1/𝜌 −1
= δA𝑍 (𝐾)−(𝜌+1) (λ)−(𝜌+1) +(𝜌+1) = λ0 δ A 𝑍 ̶ 1/𝜌 −1
(𝐾)−(𝜌+1) = λ0 MPK = MPK
𝜕𝑄 1 ̶ 1/𝜌 −1
MPL = = (− )A 𝑍 .(1 ̶ δ). ( ̶ ρ) 𝐿−𝜌 −1 = (1 ̶ δ) A 𝑍 ̶ 1/𝜌 −1
𝐿−(𝜌+1)
𝜕𝐿 𝜌
Now, to estimate the degree of homogeneity; K and L are changed by λ times. Thus, the new
function is:
𝑀𝑃𝐿∗ = (1 ̶ δ)A [𝛿(λ𝐾)−𝜌 + (1 − 𝛿)(λ𝐿)−𝜌 ] ̶ 1/𝜌 −1
(λ𝐿)−(𝜌+1)
= (1 ̶ δ)A [λ−𝜌 {𝛿𝐾 −𝜌 + (1 − 𝛿)𝐿−𝜌 }] ̶ 1/𝜌 −1
(λ)−(𝜌+1) (𝐿)−(𝜌+1)
̶ 1/𝜌 −1
= (1 ̶ δ) A 𝑍 (λ)−(𝜌+1) (𝐿)−(𝜌+1) (λ)−𝜌 ( ̶ 1/𝜌 −1)
̶ 1/𝜌 −1
= (1 ̶ δ) A 𝑍 (𝐾)−(𝜌+1) (λ)−(𝜌+1) +(𝜌+1)
= λ0 (1 ̶ δ) A 𝑍 ̶ 1/𝜌 −1
(𝐿)−(𝜌+1) = λ0 MPL = MPL
̶ 1/𝜌 −1
MPK = δ A 𝑍 𝐾 −(𝜌+1) , and, MPL = (1 ̶ δ) A 𝑍 ̶ 1/𝜌 −1
𝐿−(𝜌+1)
̶ 1/𝜌 −1
MPK.K= δ A 𝑍 𝐾 −(𝜌+1) . K = δ A 𝑍 ̶ 1/𝜌 −1
𝐾 −𝜌 , and,
̶ 1/𝜌 −1 −(𝜌+1) ̶ 1/𝜌 −1
MPL.L= (1 ̶ δ) A 𝑍 𝐿 . L = (1 ̶ δ) A 𝑍 𝐿−𝜌
Therefore, MPK. K + MPL. L
̶ 1/𝜌 −1
=δA𝑍 𝐾 −𝜌 + (1 ̶ δ) A 𝑍 ̶ 1/𝜌 −1
𝐿−𝜌
̶ 1/𝜌 −1
=A𝑍 [δ𝐾 −𝜌 + (1 ̶ δ)𝐿−𝜌 ] = A 𝑍 ̶ 1/𝜌 −1
𝑍=A𝑍 ̶ 1/𝜌
= Q (proved)
⁄𝑓 𝐿
𝑘
𝑓𝐿 (1 ̶ δ) A 𝑍 ̶ 1/𝜌 −1 𝐿−(𝜌+1) (1 ̶ δ) 𝐾 𝜌 +1
Here,
𝑓𝐾
=
δ A 𝑍 ̶ 1/𝜌 −1 𝐾 −(𝜌+1)
=
δ
(𝐿 )
1 1
𝐾 𝜌 +1 δ 𝑓𝐿 𝐾 δ 𝜌 +1 𝑓 𝜌 +1
Or, ( )
𝐿
=
1 ̶ δ 𝑓𝐾
, therefore,
𝐿
=(
1 ̶δ
) (𝑓𝐿 )
𝐾
1 1
𝑑𝐾⁄ 1 δ 𝑓𝐿 𝜌 +1 ̶ 1
𝐿 𝜌 +1
Hence, 𝑑𝑓𝐿 =
ρ+ 1
(1 ̶δ
) (𝑓 )
⁄𝑓 𝐾
𝑘
𝑑𝐾⁄ 𝑓𝐿 1 1 𝑓𝐿
𝐿
⁄𝑓 1 δ 𝜌 +1 𝑓 𝜌 +1 ̶ 1 ⁄𝑓 1
Or, σ = 𝑑𝑓𝐿 . 𝐾⁄
𝑘
= (1 ) (𝑓𝐿 ) . 1
𝑘
1 =
⁄𝑓 𝐿 ρ+ 1 ̶δ 𝐾 δ 𝜌 +1 𝑓 𝜌 +1
ρ+ 1
𝑘 ( ) ( 𝐿)
1 ̶δ 𝑓𝐾
12
In case of CES production function, shape of the isoquant depends on the value of σ
1 1
We have seen that, σ = , therefore, ρ = ̶ 1
ρ+1 𝜎
a) When, σ = 0; ρ = ∞ K
This is the case when in the production process; substitution
between inputs is not possible. Consequently, the production b
function is characterized by the existence of fixed input a IQ2
proportion. That is, the corresponding isoquant curves are IQ1
right-angled shaped. Points ‘a’ and ‘b’ indicate the input ratio. O L
b) When, σ = ∞; ρ = ̶ 1 K
This is the case when in the production process perfect
substitution between inputs is possible. Consequently, the
production function is characterized by the existence of
constant MRTS. That is, the corresponding isoquant curves are
downward sloping straight lines. O IQ1 IQ2 ` L
c) When, σ = 1; ρ = 0 K
In this case CES production function becomes Cobb-Douglas
production function. Here, in the production process,
substitution between the inputs is possible but the rate of IQ2
MRTS decreases. Consequently, the corresponding isoquant IQ1
curves are downward sloping and convex to the origin. O `L
d) When, 0 < σ < 1; ρ > 0 K
In this case in the production process, substitution between the
inputs is possible only over a definite range of the isoquant (ab a c
for IQ1 and cd for IQ2). Consequently, they are downward
sloping and convex to the origin in this range. The range where d IQ2
substitution between the inputs is not possible, segments of the b IQ1
corresponding isoquant curves are either parallel to ‘x’ axis or O `L
are parallel to ‘y’ axis.
e) When, σ > 1; ̶1<ρ<0 K
In this case in the production process, substitution between the
inputs is possible over the entire range of the isoquants.
Moreover, consequent isoquant curves touch both the axes.
O IQ1 IQ2 `L
13
Examples of Production Function
a) Production function modeling case of fixed proportions
An important family of production functions models technologies involving a single technique of
production. The only way to produce a unit of output, for example, may be to use 1 machine and 2
workers; if the firm has available 2 machines and 2 workers then the extra machine simply sits idle,
and if it wants to produce two units of output then it has to use 2 machines and 4 workers. Such a
production function has fixed proportions. If the only way to produce one unit of output is to one
machine and two workers then the output from ‘M’ number of machines and L workers is the
minimum of M and L/2. This case is portrayed above as case (a), when, σ = 0; ρ = ∞.
c) Production function modeling smooth but not perfect substitution between inputs
Many technologies allow inputs to be substituted for each other, but not at a constant rate. A class of
production functions that models situations in which inputs can be substituted for each other
to produce the same output, but cannot be substituted at a constant rate, contains functions of
the form f (K, L) = AKα Lβ , for some constants A, α, and β. Such a production functions are known as
Cobb-Douglas production function. This case is portrayed above as case (c), when, σ = 1; ρ = 0.
14
Transcendental Logarithmic (Translog) Production Function
Translog production function is a generalization of the Cobb-Douglas production function.
The name stands for 'transcendental logarithmic'. The translog production functions occurred
in the context of researches related to the discovery and definition of new flexible forms of
production functions and to the approximation of CES production function; when the
elasticity of substitution is very close to the unitary value, which is the case of Cobb-Douglas
production function.
The term ‘translog production function’ coming from the term ‘transcendental logarithmic
production function’ was proposed by Christiansen, Jorgensn and Lau in two papers
published in 1971 and 1973, which dealt with the problems of strong separability (additivity)
and homogeneity of Cobb-Douglas and CES production functions and their implications for
the production frontier. The translog production functions represent in fact a class of flexible
functional forms for the production functions. One of the main advantages of the respective
production function is that, unlike in case of Cobb-Douglas production function, it does not
assume rigid premises such as: perfect or smooth substitution between production factors or
perfect competition in the product and factor markets.
Also, the concept of the translog production function permits to pass from a linear
relationship between the output and the production factors; thus, to take into account of
nonlinear ones. Due to its properties, the translog production function can be used for second
order approximation of a linear-homogenous production, estimation of the ‘Allen elasticities
of substitution’, and estimation of the production frontier or the measurement of the total
factor productivity dynamics.
The translog production function is much more general (it has a flexible functional form
permitting the partial elasticities of substitution between inputs to vary). Cobb-Douglas
production function and constant elasticity of substitution (CES) production function place
restriction on elasticity of substitution. Alternatively, the generalized Leontief, generalized
Cobb-Douglas and Translog function all are sufficiently flexible. The Translog function
allows for variability of Allen partial elasticities of substitution and for using any number of
inputs. Moreover, this function has both linear and quadratic terms with the ability of using
more than two factor inputs.
15
Two factor translog production function
lnY = ln(A) + a. ln K + b. ln L + c. ln2 (K/L)
Where, Y= Output, K= Fixed capital, L= Employed population, and, A, a, b, c are parametres
to be estimated.
These four variants of isoelastic elasticity of substitution (IEES) production functions have a
range of intuitively desirable properties and yield empirically testable predictions for the
functional relationship between relative factor shares and capital-labor ratios. IEES functions
generalize the CES function in the same way as the CES function generalizes the Cobb–
Douglas. The Cobb–Douglas production function is isoelastic and implies constant factor
shares; the CES production function implies isoelastic factor shares and has a constant
elasticity of substitution; whereas IEES functions have an isoelastic elasticity of substitution
and a constant elasticity of elasticity of substitution.
16