Chapter 8: Optimal Input Combination and Cost Functions

I. Optimal combination of inputs for a given level of output. We may rearrange the formula for the
budget line into a form where

R P
K = − L L
PK PK

The budget R divided by PK gives the Y intercept, the number of units of capital that could be
purchased if the entire budget were spent on capital. PL/PK is the slope of the budget line. Notice
it is the price of what is on the X-axis, labor, divided by the price of what is on the Y-axis, capital.
We have suggested previously that a budget line may be called a price line, a line of attainable
combinations, or an isocost line.

Capital

R/PK
Slope = -PL/PK

Labor

Equilibrium occurs where the slope of the budget line and the slope of the isoquant are tangent.

MPP L P
=− L 8.2
MPP K PK

Capital

Tangency where
equilibrium occurs. The
slope of the budget line
. -PL/PK equals the slope
of the isoquant
MPPL/MPPK

Labor
 Capital

II. Cost definitions

A. Opportunity costs--return of resources in next best use
B. See figure 8.5 for application to pollution control
C. Social costs include costs that the consumer may not pay--external costs such as pollution
D. Explicit costs--costs paid by the firm for resources
E. Implicit costs--costs of owner inputs that may not be paid by the firm

III. Costs in the short run.

A. Fixed costs--costs associated with fixed inputs such as plant and capital
B. Variable costs--costs of inputs that can be changed in the short run.

1.
$/Q TC

N TVC or VC

Average total cost at point N= segment A/segment B on the graph

A
TFC or FC

Output or Q
B

C. Total costs--sum of fixed and variable costs.

D. Average costs--average total costs = TC/Q, slope of a line drawn from origin to point on TC
E. Average variable costs are VC/Q; average fixed costs are FC/Q.
F. Average cost diagrams

Average total cost = average fixed cost + average variable cost

$/Q

D1=AVC

Average variable cost or AVC

Average fixed cost or AFC

D2=AVC Quantity of Output
A. Notice the graphical relationship. The distance from X-axis to ATC measures average total
cost. It is equal to the sum of the distances from X-axis to AVC plus the distance from the X-
axis to AFC line. This gives a way to see if the graph is drawn properly. The distance from
ATC to AFC (D1) must equal the distance from the X-axis to AVC (D2).
B. Notice that average fixed cost continues to fall because FC is constant and since it is divided
by quantity of output to get average fixed cost, it continues to fall and is asymptotic to the X-
axis.
C. Notice also that average variable cost reaches a minimum and rises (it is U shaped).
D. Average total cost is the sum of AVC and AFC and reaches a minimum either at the same
output where the average variable cost reaches its minimum or a few units beyond that output
(as long as AFC falls faster than AVC rises, ATC will continue to fall).
E. Marginal cost is the change in total cost from a one-unit change in output.

∆TC dTC
MC = or MC =
∆Q dQ

F. The second definition is a derivative in calculus.

G. In earlier chapters, a relationship between the average product and marginal product and
average revenue and marginal revenue was described. A similar relationship applies to
average cost and marginal cost. When the average (cost) is falling, the marginal (cost) must
be less than the average; when the average (cost) is rising, the marginal (cost) must be greater
than average (cost). And when the average is neither rising nor falling, the marginal (cost)
must equal the average (cost).

ATC is falling ATC is rising

$/Q
ATC

MC
MC is greater than ATC

Output or Q

MC is less than ATC

 H. When average total cost is neither rising nor falling, the marginal must equal the average.1
I. Notice that since FC portion of TC does not change, marginal cost is both the change in total
cost from a one unit change in output and the change in total variable cost from a one unit
change in output because they are equal.

II. All costs in the long run are variable cost since all inputs can be changed, even plant size. The
long run costs are planning costs since once the plant is built the firm faces fixed costs. The long
run average cost shows the minimum average cost for producing each output. There is usually a
different scale of plant the gives minimum average cost for each output.

$/Q Short run ATCs

LRAC

Output

1
Taking the derivative of total cost with respect to output and setting it equal to zero to find the minimum:

 TFC TVC  TVC dTVC dQ

d Q + Q 
 d Q − TVC
dATC Q dQ dQ
=   =0+ = 2
=0
dQ dQ dQ Q
MC (Q) − TVC = 0
MC = AVC
 V. Expansion path and long run total costs:

Capital

Expansion path
Budgets $500, $600, and $700
(PL = $30 and PK = $25)

K
Outputs 300, 400, and 500

Labor
L

D. There are three budget lines representing different costs. They are tangent to three isoquants
representing different outputs. Combining the output and cost information together on a
single diagram will give a total cost function. On either axis the distance to each of the
budget lines times the price of that input gives the budget or cost of that output. For example

LPL = Budget 1

or
KP K = Budget 1

Costs

Total Cost
$700

$600

$500

Output or Q

300 400 500

I. Theoretically, if all inputs are variable, does the average have a rising portion? Probably yes,
because although management is a variable input, multiplying the number of managers leads
 eventually to loss of information up through the hierarchy, which increases average cost.
Empirically, over a considerable range of output the long run average cost curve is constant.
A. When firms produce many products, after some point there begins diseconomies of scope:
adding more products can only be done at increasingly higher costs per product. Perhaps
adding the first few products may yield economies of scope.

VII. Empirical studies are both time series and cross sectional analysis. Engineering approaches are
used as well. All approaches have unique shortcomings. Generally notions of short and long run
complicate data problems. Depreciation in accounting data is set by assumption. Engineering
studies may err because real world costs often differ from initial projections. Finally, theoretical
minimum costs and actual minimum costs may differ considerably. See Table 8.5