Cobb Douglas Production

Estimating a Cobb Douglas Production Function
1.
The Cobb Douglas Production Function

Suppose the Cobb-Douglas production function describes how a particular economy's output
level
is determined from the inputs L and K:
Y ( L, K ) ALa1 K a2
(1)
A>0, 0 < a1 < 1, 0 < a2 < 1
The variables A, a1, and a2 describe the economy's technology. The variable A can be thought of as the
general level of technology. The production function indicates that an increase in the parameter A---a
technological improvement---will increase output. The technological parameters a1 and a2 measure the
respective contributions of L and K to the production process, as will now be shown more carefully.
For this Cobb-Douglas production function, the marginal product of labor can be calculated as
YL
(2)
Y ( L, K ) a1 ALa1 1 K a2
L L
L
a1 ALa1 1K a 2
L
a1
ALLa1 1 K a 2
L
,
ALa1 K a 2
a1
L
which means
Y
Y
a1
L
L
The quantity Y/L is the average product of labor. Thus, we see that the marginal product of labor Y/L
and the average product of labor Y/L are each measures of labor productivity, and the last equation
indicates the two are related.
In fact, rearranging the last equation, we find that the parameter a1 is given by the ratio of the
marginal product to the average product:
(4)
Y
a1 L
Y
L
The ratio of the marginal product to the average product defines the elasticity of output with respect to
labor input. An elasticity always gives the percentage change in one variable divided by the percentage
change in another variable. By rearranging the right side of the last equation, we see this more clearly:
Y s Y s
L Y s % OUTPUT
L
Ys
% LABOR
L
L
(5)
If a1 > 1, then a given percentage change in labor would generate a larger percentage change in
output. The term elastic is used to describe the responsiveness of output to labor input in this
case. If a1 < 1, then a given percentage change in labor would generate a smaller percentage
change in output. The term inelastic is used to describe the lack responsiveness of output to
labor input in this case. If the production process exhibits diminishing returns relative to labor,
then a1 < 1 must hold.
The marginal product of labor decreases and the employment level increases whenever
diminishing returns is present. Mathematically, the rate of change in the marginal product is
found by taking the derivative of the marginal product function with respect to the employment
level L. For the Cobb-Douglas production function the rate of change in the marginal product is:
(6)
[YL ]
a1 ALa1 1 K a2
L
L
a1[a1 1] ALa1 2 K a 2
.
As long as a1 < 1 in the Cobb-Douglas production function, the derivative (6) is negative. This is
equivalent to assuming diminishing returns. That is, a1 < 1 indicates that the marginal product of
labor decreases as the employment level increases.
2.
Estimating the Cobb Douglas Production Function
By applying econometric tools to our Cobb-Douglas production function, we can try to obtain
estimates of the parameters A, a1, and a2. In fact, this is a classic econometric problem.
Our production function indicates the output level Y depends upon employment L and
capital K. By gathering data on output, employment, and capital stock, we could regress Y on
L and K. However, such a regression would not be consistent with the Cobb Douglass
production function relationship, but rather would be consistent with the linear production
function relationship
Y a1 L a 2 K
(7)
a1
a2
a1
The regression would provide us with estimates of

and
. The estimate of
would be
the estimate of the marginal product of labor for the linear production function (7). Indeed, the
linear production function (7) indicates marginal labor productivity cannot decrease as
employment increase.
Moreover, the level of capital does not impact labor productivity in the linear production function
(7) as it does in the Cobb Douglas production function (7). Economists have long favored
using the Cobb-Douglas production function over the linear production function because the
Cobb Douglas allows for diminishing returns and because it allows the level of one input (e.g.,
capital) to affect the productivity of another input (e.g., labor).
Y ALa1 K a2
In the Cobb-Douglas production function

, there is a nonlinear relationship between
the inputs L and K and the output Y, and the two input interact. This is possible, which is
another reason why the Cobb-Douglas is so attractive. Taking the natural log of both sides of
Y ALa1 K a2
, we obtain
ln( Y ) ln( A) a1 ln( L) a2 ln( K )

(8)
.
Using available data, we can take the natural log of each data series to create variables
ln( Y )
that are in the log levels rather than the levels. Regression the
ln( L)
on the
and the
ln( K )
, we obtain
ln( Y ) 7.08
0.94 ln( L)
0.51 ln( K )
(9)
,
(0.69)***
(0.012)***
R2=.9975
(0.07)***
Increasing returns to scale means a proportionate increase in all inputs leads to a more than
proportional increase the output. For example, doubling all inputs would lead to more than a
a1 a 2 1.45
doubling of output. In this case,

indicates a one hundred percent increase in (or
doubling of) the inputs leads to a 145 percent increase in the output level. With constant
technology, it is difficult to conceive how output could more than double from a doubling of the
inputs.
1 dY 1 dA
1 dL
1 dK
a1
a2
Y dt
A dt
L dt
K dt
(10)
gY
[1 / Y ][ dY / dt ]
Recognizing
is the growth rate of Y, we can define a new variable
gA gL
growth rate of Y. Similarly, we can define the variables
technology, labor, and capital, and we can rewrite (10) as
for the
gK
, and
for the growth rates of
g Y g A a1 g L a 2 g K
(11)
Whereas the level of technology was assumed constant in the model (8), it is the growth
rate of technology that is assumed constant in the model (11), if we regress the growth rate of
output on the growth rates of employment and capital.

Cobb Douglas Production

Uploaded by

Copyright:

Available Formats

Cobb Douglas Production

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Cobb Douglas Production

Uploaded by

Copyright:

Available Formats

Estimating a Cobb Douglas Production Function

The Cobb Douglas Production Function

is determined from the inputs L and K:

A>0, 0 < a1 < 1, 0 < a2 < 1

Estimating the Cobb Douglas Production Function

The regression would provide us with estimates of

In the Cobb-Douglas production function

ln( Y ) ln( A) a1 ln( L) a2 ln( K )

doubling of output. In this case,

is the growth rate of Y, we can define a new variable

for the growth rates of

You might also like