At this stage another parallel with consumer theory can be seen: in that case the slope of the indifference curve was shown by the marginal rate of substitution (MRS). This was also decreasing in absolute magnitude from left to right, because of the law of diminishing marginal utility.
It was also seen that the MRS was given by the ratio of the marginal utilities of the two products. It should not be too difficult for the reader to draw another parallel at this point: the MRTS is given by the ratio of the marginal products of the two inputs. The mathematical proof of this is analogous to the one relating to the MRS.
When the firm moves from point B to point C it gains output from using more labour, given by L MPL, and it loses output from using less capital, given by K MPK. Since the points are on the same isoquant and therefore must involve the same total output, the gains must equal the losses, thus:
L MPL ¼ K MPK
Since the slope of the isoquant is given by K/ L, we can now express the absolute magnitude of the slope as:
K= L ¼ MPL=MPK (5:19)
There are two extreme cases of input substitutability. Zero substitutability occurs when the inputs are used in fixed proportions, for example when a machine requires two workers to operate it and cannot be operated with more or less than this number of workers. Isoquants in this case are L-shaped, meaning that the MRTS is either zero or infinity. Perfect substitutability is the opposite extreme, resulting in linear isoquants; this means that the MRTS is constant. It also implies that output can be produced using entirely one input or the other. These extremes are shown in Figure 5.6.
5.4.3 Returns to scale
We frequently want to analyses the effects on output of an increase in the scale of production. An increase in scale involves a proportionate increase in all the inputs of the firm. The resulting proportionate increase in output determines the physical returns to scale for the firm. Two points need to be explained before moving on to the description and measurement of returns to
scale:
Figure 5.6. Extreme cases of input substitutability.
1 Proportionate increase in all the inputs. It is always assumed in referring to returns to scale that all inputs increase by the same proportion. This is not necessarily optimal for the firm in terms of economic efficiency. If inputs increase by different proportions we have to talk about returns to outlay (measured in money terms).
2 Physical returns to scale. Returns to scale can be described in physical terms or in money terms, as will become clear in the next chapter. The two meanings do not necessarily coincide; for example, it is possible for a firm to experience constant physical returns to scale yet have increasing returns to scale in money terms (better known as economies of scale).
a. Types of returns to scale
Returns to scale, in physical or money terms, can be of three types. The following are the three types of physical return:
1 Constant returns to scale (CRTS). This refers to the situation where an increase in inputs results in an exactly proportional increase in output.
2 Increasing returns to scale (IRTS). This refers to the situation where an increase in inputs results in a more-than-proportional increase in output.
3 Decreasing returns to scale (DRTS). This refers to the situation where an increase in inputs results in a less-than-proportional increase in output.
The reasons for these different returns to scale will be considered in the next chapter, when they are compared with the monetary aspects of returns to scale. We can, however, use Table 5.1 to examine these different possibilities from the standpoint of quantitative measurement. The easiest way to do this is by examining the numbers in the leading diagonal. When inputs are increased from one worker/one machine to two workers/two machines this represents a doubling of inputs; however, output increases from 4 to 17 units, an increase of more than fourfold. Thus this situation involves, IRTS. If inputs increase from two of each factor to three of each factor this is an increase of 50 per cent; output increases from 17 to 39 units, over 100 per cent. Thus there are still IRTS. This situation continues until seven units of each input are used; when each input is increased to eight units this represents an increase of about 14 per cent, while output increases from 155 to 164 units, an increase of less than 6 per cent. Thus there are now DRTS.
Generalizing from this we can conclude that with a cubic production function the returns to scale are not the same at all levels of scale or output. The type or pattern of returns to scale will obviously depend on the nature of the mathematical form of the production function. In order to understand this more clearly we need to consider the concept of a homogeneous production function.
b. Homogeneous production functions*
These functions are useful for modelling production situations because of their mathematical properties. If the inputs in a function are multiplied by any constant l and if this constant can then be factored out of the function then the production function is said to be homogeneous. This can be explained more precisely in mathematical terms by stating that a production function is said to be homogeneous of degree n if:
F(λL, λk) = λnf(L,K)
If the degree of homogeneity is equal to 1 then the production function is said to be linearly homogeneous. The degree of homogeneity indicates the type of returns to scale:
if n = 1 there are CRTS
if n > 1 there are IRTS
if n< 1 there are DRTS.
These concepts now need to be applied to particular forms of production function. Let us take the simple linear form in (5.2) first:
Q = aL + bK
When each input is multiplied by l, output is given by:
a(λL)+b(λK) = λ(aL+bK)
Thus λ can be factored out of the function and the function is linearly homogeneous. This means that linear production functions like (5.2) feature constant returns to scale at all levels of output. This is not true for the linear function with a constant term in (5.3); this is not a homogeneous function. Nor is the linear function with an interaction term in (5.4).
Now let us consider the quadratic function in (5.5):
Q = aL2 + bK2 +cLK
When inputs are multiplied by λ, output is given by:
a(λL2)+b(λk2)+c(λL)( λK) = λ2 (aL2+bK2+cLK)