Summary of

Production Theory

The Production Function

The production function is a statement of the relationship between a firm’s scarce resources (i.e.,
its inputs) and the output that results from the use of these resources. Economic cost analysis can then
be seen as the application of a monetary unit such as dollars to measure the value of this input usage in
the production process. There are two other key assumptions that you should be aware of. First, we are
assuming some given “state of the art” in the production technology. Any innovation in production (e.g.,
the use of robotics in manufacturing or a more efficient software package for financial analysis) would
cause the relationship between given inputs and their output to change. Second, we are assuming
whatever input or input combinations are included in a particular function, the output resulting from
their utilization is at the maximum level. With this in mind, we can offer a more complete definition of a
production function: “A production function defines the relationship between inputs and the maximum
amount that can be produced within a given period of time with a given level of technology”.

The Law of Diminishing Returns

The key to understanding the pattern of change in Q AP, and MP is the phenomenon known as the
law of diminishing returns. This law states: “As additional units of a variable input are combined with a
fixed input, at some point the additional output (i.e., marginal product) starts to diminish”. In situations
in which it would not be possible to consider intervals of less than one unit, it is necessary to
approximate the point at which diminishing returns occurs. Suppose you were a manager who could
actually measure and track your employees’ marginal product.
There are two key concerns of a practical nature that we advise readers to keep in mind when
considering the impact of the law of diminishing returns in actual business situations. First, there is
nothing in the law that states when diminishing returns will start to take effect. The law merely says that
if additional units of a variable input are combined with a fixed input, at some point, the marginal
product of the input will start to diminish. Therefore, it is reasonable to assume a manager will only
discover the point of diminishing returns by experience and trial and error. Hindsight will be more
valuable than foresight. Second, when economists first stated this law, they made some very restrictive
assumptions about the nature of the variable inputs being used. Essentially, they assumed all inputs
added to the production process were exactly the same in individual productivity.

The Long-Run Production Function

In the long run, a firm has time enough to change the amount of all its inputs. Thus, there is really
no difference between fixed and variable inputs. The resulting increase in the total output as the two
inputs increase is called returns to scale. According to economic theory, if an increase in a firm’s inputs
by some proportion in an increase in output by a greater proportion, the firm experiences increasing
returns to scale. If output increases by the same proportion as the inputs increase, the firm experiences
constant returns to scale. A less than proportional increase in output is called decreasing returns to
scale.
Using this notation, we can summarize returns to scale in the following way:
 If h > k, the firm experiences increasing returns to scale (E Q > 1).
 If h = k, the firm experiences constant returns to scale (E Q = 1).
 If h < k, the firm experiences decreasing returns to scale (E Q < 1).
The Estimation of Production Functions
First, the possible shapes of production functions. Second, we discuss the Cobb-Douglas production
function, a form that has been commonly used by economists since its introduction in the 1920s. Last,
the data needed for estimation and some of the production function studies published by economists.

 The Various Forms of a Production Function

The short run is characterized by the existence of a fixed factor to which we add a variable factor.
Thus, the simple function, containing just one variable factor and one fixed factor, can be written as
follows: Q = f(L)k.
Using more than one independent variable in a production function is certainly more realistic than
limiting the analysis to only one, and when it is assumed all inputs are variable, we have then moved
from short-run analysis to the long run. Indeed, this function can be employed in both analyses. In a
simple two-variable model (e.g., labor and capital), the power function permits the estimation of
marginal product (e.g., when labor changes and capital remains the same) and of returns to scale (when
both variables change).

 The Cobb-Douglas Production Function

The Cobb-Douglas production function was introduced in 1928,10 and it is still a common functional
form in economic studies today. It has been used extensively to estimate both individual firm and
aggregate production functions. It has undergone significant criticism but has endured. “It is now
customary practice in economics to deny its validity and then to use it as an excellent approximation.”11
It was originally constructed for all the manufacturing output (Q) in the United States for the years 1899
to 1922. The two inputs used by the authors were number of manual workers (L) and fixed capital (K).
The formula for the production function, which was suggested by Cobb, was of the following form:
Q = aLbK 1-b.
The Cobb-Douglas function has the following shortcomings:
1. The Cobb-Douglas cannot show the marginal product going through all three stages of
production in one specification. (A cubic function would be necessary to achieve this.)
2. Similarly, it cannot show a firm or industry passing through increasing, constant, and
decreasing returns to scale.
3. There are also important problems with specification of data to be used in empirical
estimates. These problems are discussed next.

 Statistical Estimation of Production Functions

If only one product is produced in a plant, Q is specified in physical units (e.g., number, tons,
gallons). However, if a plant produces a number of different products, and it is not possible to segregate
properly the inputs and outputs of the products, estimation becomes considerably more difficult. In such
a case, the investigator probably must settle for some measure of value, assigning weights to products
depending on the value (in terms of cost or selling price) produced. There are some obvious problems
with this procedure. First, over time the data will have to be deflated to account for price or cost
changes. Second, the price or cost of a product may not be an exact reflection of the inputs combined in
the total value. However, until better measurement methods are found, such valuing methods will have
to suffice.
Cross-sectional analysis is favored when the data collected cover a number of plants in a given time
period. But here again, problems may arise. The various plants may not employ the same level of
technology. If the data are in monetary terms, an adjustment for differential price or wage levels at
different geographic locations would be necessary. Although a theoretical production function assumes
 output is produced at the most efficient input combinations, in reality such an ideal situation is certainly
not assured whichever estimating method is used. Ultimately, there is no perfect way to measure and
analyze the data. The researcher must choose the most appropriate method.

 A Numerical Example of a Cobb-Douglas Production Function

A cross-sectional sample of twenty soft drink bottling plants has been selected. The data are given
for a specific month in 1998. Only two independent variables are used: (1) number of direct workers and
(2) plant size. Production, the dependent variable, is stated in terms of gallons of product shipped during
the period.

 Aggregate Production Functions

A large proportion of the studies performed using the Cobb-Douglas function did not deal with data
for individual firms, but rather with aggregations of industries or even the economy as a whole.
Although much of this work has proved to be quite fruitful in describing production functions, the
interpretation of the results may not be quite as meaningful as for individual production functions.
When data for the economy as a whole are used, the model must accommodate different technologies
and different processes, and thus does not represent a specific technological process of a given firm.
When the aggregation is done at the level of an industry rather than the overall economy, the
assumption of similar technology is more appropriate, but even in such a case many dissimilarities may
occur.
Gathering data for such aggregate functions can be difficult. For the economy as a whole, gross
national or domestic product—in real terms—could be used to measure output. For specific industries,
data from the Census of Manufactures or the production index published by the Federal Reserve Board
can be employed. Data for investment and depreciation by industry are also available for the
construction of appropriate indexes for the capital variable. The Bureau of Labor Statistics publishes a
great deal of data on employment and work hours.
In another study, the author performed a cross-sectional study of eighteen industries using data
from the Census of Manufactures. The observation units were individual states. Three independent
variables were used:
1. Production worker hours
2. Nonproduction worker years
3. Gross book value of depreciable and depletable assets

A rather interesting application of the cross-sectional Cobb-Douglas function is represented by a

study of the importance of various baseball statistics to the success of a baseball team. The dependent
variable was a team’s win/loss percentage relative to the league average. The independent variables
were slugging percentage, on-base percentage, number of stolen bases, unearned runs allowed, and
earned run average. Surprisingly, the offensive statistics proved to be more important than the
defensive statistics. All coefficients were statistically significant, and the equation exhibited increasing
returns to scale.

The Importance of Production Functions in Managerial Decision Making

The production function is an important part of the economic analysis of the firm because it serves
as the foundation for the analysis of cost examined. But for managers, an understanding of the basic
concepts discussed in this chapter also provides a solid conceptual framework for decisions involving the
allocation of a firm’s resources both in the short run and in the long run. Another key management
principle illustrated in the economic theory of production is planning, discussed in the following
subsection.
 Careful Planning Can Help a Firm to Use Its Resources in a Rational Manner
Good capacity planning requires two basic elements: (1) accurate forecasts of demand, and (2)
effective communication between the production and marketing functions (particularly in large
organizations where these functions are often handled by separate work groups).
The first element is rather obvious but, as you have seen in chapter 5, not easy to achieve. The
second element may not be so obvious, especially for those who have not had work experience in large
organizations. It is not uncommon for manufacturing people to proceed merrily with their production
plans on a purely technical basis (i.e., from a strictly engineering point of view) without fully
incorporating the marketing plans of those whose main responsibility is to sell the products.
It is also quite possible for marketing people to try to sell as many units of the product as possible
(as marketing people are supposed to do) without consulting the production people as to whether the
firm has the capacity to meet the increase in demand. A full discussion of these problems of
management and organization is beyond the scope of this chapter. However, they are mentioned to
underscore the importance for managers of understanding production theory.

The Production Process

The relationship between inputs and outputs (the production technology) expressed numerically or
mathematically is called a production function or total product function. The marginal product of a
variable input is the additional output that an added unit of that input will produce if all other inputs are
held constant. According to the law of diminishing returns, when additional units of a variable input are
added to fixed inputs, after a certain point, the marginal product of the variable input will decline.
Average product is the average amount of product produced by each unit of a variable factor of
production. If marginal product is above average product, the average product rises; if marginal product is
below average product, the average product falls. Capital and labor are at the same time complementary
and substitutable inputs. Capital enhances the productivity of labor, but it can also be substituted for
labor.

Going “Beyond the Curves”: Current Production Issues and Challenges for Today’s Managers
The heart of the economic theory of production can be expressed as production curves showing the
relationship between inputs and output. But the intensity of current global competition often requires
managers to go beyond these curves. Being competitive in production today mandates that today’s
managers also understand the importance of speed, flexibility, and what is commonly called “lean
manufacturing.”
Speed of production is required because rapidly changing technology renders products on the
market obsolete very quickly and because today’s consumers often change their tastes and preferences at
the same fast pace. These “supply side” and “demand side” changes are quite evident in the world of
consumer electronics, in particular the market for mobile devices such as smartphones and mobile
phones.
Some important aspects of the lean approach are (1) kaizan, or continuous Improvement events;
(2) the elimination of muda (Japanese for “waste”); and (3) the 5 S’s. Examples of muda are:
1. Overproduction
2. Waiting (underutilization of assets affects downstream production)
3. Unnecessary movement of materials (wastes time and increases chances of handling damage)
4. Extra processing (extra operations such as rework, preprocessing, handling, or storage)
5. Inventory
6. Motion (unnecessary motion)
7. Defects

Call Centers: Applying the Production Function to a Service

Most production function examples involve manufacturing or agriculture. To illustrate how the
concepts in this chapter can be used for a service activity, let us consider the example of a call center
represented by the following production Q = f (X, Y ) function:

Where:
Q = number of calls.
X = variable input (this includes call center representatives and the complementary hardware such as PCs,
desks, and software if site licenses are sold on a per-user basis).
Y = fixed input (this includes the call center building, hardware such as servers and telecommunications
equipment, and software site licenses up to the designated maximum number of users).

Using this basic production function, it is possible to consider a number of applications of concepts.
The following are offered as examples.
 Three Stages of Production: Stage I could be a situation in which there is so much fixed capacity
relative to number of variable inputs that many representatives sit around idle, waiting for calls to come
in. Stage II could be a situation in which representatives are constantly occupied and callers are
connected to representatives immediately after the call is answered or are kept waiting no more than a
certain amount of time (e.g., 3 minutes). If callers are kept waiting for longer than 3 minutes, the call
center manager might consider adding more call center representatives. Stage III could be a situation in
which callers begin to experience a busy signal on a more frequent basis or call representatives may
begin to experience a slower computer response or more frequent computer “down times.” These are
all technical manifestations of overloaded computer processing power or transmission capacity (i.e.,
insufficient bandwidth).
 Input Combinations: A number of input combinations could be considered in the operation of a
call center. To begin with, there is the location trade-off in which the productivity of call center
representatives in two or more locations might be considered relative to the cost of each representative
(including wages, training costs, costs associated with turnover rates, etc.).
 Returns to Scale: One can well imagine that a call center would be amenable to Increasing
returns to scale. This is evidenced by the number of smaller companies who outsource their operations
to companies that specialize in call center operations. These “third-party” vendors are able to provide
call center services at a lower cost to the company mainly because their size enables them to take
advantage of increasing returns to scale.

Global Application: Shifting Trends in Global Outsourcing

China is sometimes called the world’s factory and India is sometimes called the world’s back office.
Economic data indeed supports these generalizations. China began its climb to global prominence by
manufacturing products that required a high labor content and relatively less capital and technology.
The apparel industry is a good example. By the same token, in the service sector, India began by gaining
prominence in call center operations. In recent years, companies have been starting to shift their
production of apparel and their call center operations to countries other than China and India,
respectively. One would assume the wage rate is the primary factor driving their decisions to do so. But
this is not quite the case. True, wage rates in China have been rising significantly. However, there are
other factors such as speed to market.
 For years, India has led the world in global outsourcing revenue. But in recent years, call center
operations have grown considerably in the Philippines, to the point where it appears to have passed
India in the number of call center workers. Going back to the point made about the apparel industry in
China, we see that, in fact, the Chinese government is carrying out policies to focus the country more on
higher value-added manufacturing in the high-tech sector and less on lower-value, labor-intensive
sectors such as apparel and toys.