Engineering Economic Analysis

2019 SPRING
Prof. D. J. LEE, SNU

Chap. 22
COST CURVES
 Average cost & Marginal cost
§ Total cost function can be derived from the cost
min. problem: c(w1,…,wn, y)
§ Various costs can be defined based on the total
cost function
• variable cost, fixed cost,
• average cost, marginal cost,
• long-run cost, short-run cost

1
 Average cost & Marginal cost
§ Let x f : vector of fixed inputs, x v : vector of variable factors
(
w = wv , w f )
§ Short-run cost function
( ) ( )
c w, y, x f = wv × x v w, y, x f + w f × x f

• Short-run average cost (SAC):

(
c w, y, x f )
y

• Short-run average variable cost (SAVC): (

wv × x v w, y, x f )
y

wf × xf
• Short-run average fixed cost (SAFC): y

• Short-run marginal cost (SMC): (

¶c w, y, w f )
¶y

2
 Average cost & Marginal cost

§ Long-run cost function There is no fixed input

( ) ( ) (
c w, y = wv × x v w, y + w f × x f w, y )
• Long-run average cost (LAC):
(
c w, y )
y

• Long-run marginal cost (LMC): (

¶c w, y )
¶y

• Note that ‘long-run average cost’ equals ‘long-run

average variable cost’ and ‘long-run fixed costs’ are
zero’

3
 Average cost & Marginal cost

§ Example: Short-run Cobb-Douglas cost function

• In a short-run, x2 = k
• Cost-min. problem: min w1 x1 + w2 k a -1 1
x1 = k a
×y a
a 1- a
s.t. y = x k
1
a -1 1

• SR Cost function: ( )
c w, y, k = w k 1
a
× y + w2 k
a

1- a
æ yö a w2 k
SAC = w1 ç ÷ +
èkø y
1- a
æ yö a
SAVC = w1 ç ÷
èkø
wk
SAFC = 2
y
1- a
1 æ yö a
SMC = w1 ç ÷
a èkø
4
 Cost curves (Geometry of costs)
§ Total cost function can be derived from the cost
min. problem: c(w1,…,wn, y)
§ Assume that the factor prices to be fixed, then
c(y).
§ Total cost, c(y), is assumed to be monotonic in y.
§ Various cost curves: Average cost curve, Marginal
cost curve, LR cost curve, SR cost curve etc.
§ How are these cost curves related to each other?

5
 Cost curves (Geometry of costs)

§ SR Average cost
• SR total cost = VC + FC: c ( y ) = cv ( y ) + F
• SAC = SAVC + SAFC
c ( y ) cv ( y ) F wv × x v ( w, y, x f ) w f × x f
= + = +
y y y y y

U-shaped
SAC

6
 Cost curves (Geometry of costs)

§ Marginal cost vs. Average cost

dAC( y ) d æ c ( y ) ö c¢ ( y ) y - c ( y )
= ç ÷=
dy dy è y ø y2
1æ c( y) ö 1
= ç c¢ ( y ) - ÷ = ( MC( y ) - AC( y ) )
yè y ø y

• Thus, AC is increasing (AC¢( y ) > 0) Û MC ( y ) > AC ( y )

AC is decreasing (AC¢( y ) < 0) Û MC ( y ) < AC ( y )
AC is minimum (AC¢( y ) = 0) Û MC ( y ) = AC ( y )

• MC for the first small unit of amount equals AVC for a

single unit of output
TC (1) - TC (0) cv (1) + F - cv ( 0 ) - F cv (1)
MC (1) = = = = AVC (1)
1 1 1

7
 Cost curves (Geometry of costs)

§ Marginal cost vs. Average cost

AC decreasing AC increasing

8
 Cost curves (Geometry of costs)

§ Marginal cost vs. Variable cost

• Since
dcv ( y ) y
MC ( y ) = cv ( y ) = ò MC ( z )dz
dy 0

• The area beneath the MC curve up to y gives us the

variable cost of producing y units of output
Cost

MC(y)

Area is the variable

cost of making y units
0 y¢ y
9
 Cost curves (Geometry of costs)

§ Example
• SR total cost function: c ( y ) = y2 +1

• Variable cost: cv ( y ) = y 2
• Fixed cost: c f ( y ) = 1
AVC ( y ) = y 2 / y = y
AFC ( y ) = 1/ y

AC ( y ) = y + 1/ y
MC ( y ) = 2 y

10
 Cost curves (Geometry of costs)

§ Example: C-D Technology

• Recall that -1 é
æ a ö
b
a +b æ a ö
-a
a +b
ù a b 1
c ( w1 , w2 , y ) = A a +b êç ÷ + ç ÷ ú w1a +b w2a +b y a +b
êè b ø èbø ú
ë û
1
• For a fixed w1, w2, c ( y ) = Ky a +b
, a +b £1

1- a -b
AC ( y ) = Ky a +b

K 1-aa+-bb
MC ( y ) = y
a+b
a -1 1

• In the short-run, recall that ( )

c w, y, k = w1k a
× y + w2 k
a

1
c ( y ) = Ky + F
a

1- a
F
AC ( y ) = Ky a
+
y

11
 Cost curves (Geometry of costs)

§ Example: MC curves for two plant

Plant 1 Plant 2
c1(y) y1 c2(y) y2

• How much should you produce in each plant?

min c1 ( y1 ) + c2 ( y2 ) y2 = y - y1
{ y1 , y2 }
s.t. y1 + y2 = y y1
(
min c1 ( y1 ) + c2 y - y1 )

¶c dc1 ( y1 ) dc2 ( y2 ) dy2 dy2

= + = 0 and = -1
¶y1 dy1 dy2 dy1 dy1

• Therefore, the optimality condition is

( )
MC1 y1* = MC2 y2* ( )
12
 Long-run vs. Short-run Cost Curves

§ In the long run, all inputs are variable

• LR problem: Planning the type and scale investment
• SR problem: Optimal operation

§ Given a fixed factor: Plant size k

• SR cost function: cs ( y, k )
• In LR, let the optimal plant size to produce y be k(y)
Øfirm’s conditional factor demand for plant size!

§ Then LR cost function is

c ( y ) = cs ( y , k ( y ) ) How this looks graphically?

13
 Long-run vs. Short-run Cost Curves

• For some given level of output y*

( )
Þ k * = k y* : optimal plant size for y*
Þ c ( y, k ) : SR cost function for a given k
s
* *

Þ cs ( y, k ( y ) ) : LR cost function

• Since SR cost min problem is just a constrained version of the

LR cost min problem, SR cost curve must be at least as large
as the LR cost curve for all y

( )
c ( y ) £ cs y, k * for all level of y

( ) ( )
c y* = cs y* , k * when k * = k y* ( )

14
 Long-run vs. Short-run Cost Curves

• Also LAC ( y ) £ SAC ( y, k * ) for all level of y

( ) ( )
LAC y* = SAC y* , k * when k * = k y* ( )
• Hence SR and LR cost curves must be tangent at y*

( )
argmin SAC y, k * = ymin

(
Note that LAC ( ymin ) ¹ SAC ymin , k * , )
SAC(ymin)
LAC(ymin)
(
that is, LAC ( ymin ) £ SAC ymin , k * )

ymin

15
Long-run vs. Short-run Cost Curves

16
 Long-run vs. Short-run Cost Curves

• Discrete levels of plant size: k1 , k2 , k3 , k4

• LR average cost curve is the lower-envelop of SR average

cost curves
17
 Long-run vs. Short-run Cost Curves

§ LR marginal cost
• When there are discrete levels of the fixed factor, the firm will
choose the amount of the fixed factor to minimize costs.
• Thus the LRMC curve will consist of the various segments of
the SRMC curves associated with each different level of the
fixed factor.

18
 Long-run vs. Short-run Cost Curves

§ LR marginal cost
• This has to hold no matter how many different plant sizes
there are !
Cost
SACs

LAC(y)

y
19
Short-Run & Long-Run Marginal Cost Curves

Cost
SRMCs LMC(y)

LAC(y)

y
20
 Long-run vs. Short-run Cost Curves

§ LR marginal cost
• LR cost function

• Differentiating LR cost function w.r.t. y

• Since k* is the optimal at y=y*,

• Thus LRMC at y* equals to SR marginal cost at (k*, y*)

21