Professional Documents
Culture Documents
CHAPTER 4
lEARN1NG OBJECTIVeS
The Production Function with
One Vanahle
The Law 01 Dlfmnishlng
Returns
I~argifla!
PRODUCTION THEOR
Corner Solutions
Relurns 10 Scale
The Output Elasticity
Eslimattons 01 Produclion
Funclions
Appendix. Lagrangian
Moltipliers and Oplimallnput
Combinations
Once managers determine the demand for the firm's product or service, their'
is far from over. Now they must choose the optimal method to
need to be as efficient
hallmark of good managers. Efficiency
tion process. Simply staled, a production process explains
or service (output). The production
precisely specifies the relationship between inputs and outputs.
Production issues are not confined to the physical transformation of'
into outputs. In business,
goods and services, such as employment
distribution.
managers are concerned with rtflnenttv
intellectual resources.
into
demand, is required
for managers to optimize
process.
cannot understand their firm's cost structure unless they understand
the production process.
CHAPitR
I.:
PRODliCTION [HEORY
TABlE4.1
~"' '~~
otparts
'
1,~OO
-Tolatoutpul
1.400 .
Amount 01
Amount of
Cdpital [Number of
Labor
Machines]
Output 01 Parts
10. hundreds ~er year]
o
?
3
1.300
1.200
1,9
1,1DO
132
1,000
243
376
5
6
6.67
7
52'J
684
792.'39
gOO
800
700
847
600
1,008
500
1.161
10
1,300
11
1.419
12
13
5
5
1,512
1.573
1,596
1,:175
14
15
II~PUj
400 "
300"
200"
:l!IIf,;rL
numbers in Table 4.1 (and Table 4.2) are derived from the
to Q 30L + 20U P. Lis
We can think of the production lunctlOn as
technology use. Thomas is clearly interested in
the number of machinists varies. One common measure used by m,lIlY manager~iII
is 01ltput Jl9T wOIker. This measure is Wh~li tconOinim call aVefdge product
Because we are varvin~ machinists, this is output per worker or
per mpul
-X~
AP
Metric lor
eSllmallog the efficiellcy 01 each
rnpullo which the "'pufs MP is
(qual 10 Ihe Incremental change
rn output created by a small
change in the input
. holding X2 conslanl
3
4
5
6
6.67
MP
IlQ
f,X I '
x? constanl
10
11
12
13
Output of
Parts !Q,
Hundreds
of Parts!
tA..verage
Marginal
Product
Product
!D/L1
IllOIM I'
Ida/dUo
Product
5,
5
5
49
49
49
1:l2
83
21,3
66
81
111
5
5
5
5
5
5
5
5
5
5
5
376
9t,
133
142
525
105
149
1~5
67
98
123
684
114
159
162
792 59
847
118.89
162.89
16133
121
163
163
1,008
126
161
158
1,161
129
153
1&7
1,300
1:jQ
139
119
130
107
78
43
2
-45
1,419
129
1,512
126
93
'1,573
121
61
23
-19
14
1,596
114
15
IS/5
105
ligures In Ihe I'lO/M column pertain to the ,nterval belween Ihe indicated amounl 01 tabor and
one Ilnitlesslhan Ihe mdicaled amount 01 tabor The ligures inlhe dO/dt column <lre the conlinuous
marginat product-Ihalls, dO/dL = MPL 30+ 'Ot - 3L 2
average
parts per machinist between the fllst and second hires. Results for
hires are shown in Table 1.2.
QUANT OPTION
More precisely. the
product 01 an Input
to the quantity 01 the Input ThaI IS, rf Q is the output and x Ihe
quantily ollhe input, the marginal product 01 the input equals dO/dx If the
With
CHAPTER
I, pr'ODUcrIOI~
THEORY
FIGURE 4.2
l} .il
Oilipul
dx
0'
( clx
dx
IS
C';('d~"~1
lcro.
And hence
flAP
dx
AP
I. I
3 4
I
B
I I I
10 11 12 13 14 1:)
It is not hard
Amounlo1
rullor
The second (dQI elL) assumes that Thomas can employ labor continuously, as
J.25 workers or J,33 workers. This could be achieved by
or workers who work more or Jess time than in a standard day's work.
MP equals AP when AP is maximized. A
llltuitive frame may help. Assume your IJrofessOT is
average of test scores by
II.
score is
must decrease. This is a natural law of mathern<ltics. So as
than Ap, AP mllst be increasl11g. \lJhcn MP is less than AP, AP must be
MPinlersects wilh ilPwhen AP is nt a maximum.
jml"l('\rt~l'\t
prod
five machines. As more machin
will have to ration machines, or new hires will be assigned to
tasks.
ISOQUANTS
TABLE 4,3
Production
Two Variable
Amount of Labor
!Unitsl
11
18
11.
30
22
30
35
60
50
80
Bl
84
1/,0
2
3
I.
115
f(X~,
X2)
Tolal I
oulpul'
B,
B,
B
Amellol 01 labor
QUANT OPIION
the iaslldio1l5, we have
MP, ~ ~X,
10?
MP
iJO
ilX2
bundles capable of
HI'\
ISOQUANl S
G'G
past
recyder, repro
';'
'\.
It maintains a
\.
"
K1~--""
..,.
slre.amlined
'\
cenler and
300
200
Ko'-
100
L2
L,
1.0
105
angles,
the output
MRTS ~
300
ilX1
200
QUANlOPllON
100
Labor
are
dO
tines IhJI
profil' maxlmillng fir ms
aXJ ) dX1
+ ( aX2 1dX2
Therefore,
MP2
(4,3)
negative,
FIGUHE 4.6
out'.lde the
OUafld OV
the combinations of
aiM.
Amounl01
capilalused
100
P.
Labor
K = PI(
PI(
MIP,
combination
to choose an mput bundle where the margi.nal prod
per dollar spent of Jabor and capital are identical. If they are ]]ot, the manager
increase the use of the immt with the higher marginal per dollar value.
the manager maximizes output by distrib
so the maminal oroduct of a dollar's worth
RETURNS TO SCALE
FIGURE 4.8
---- --- - - - - - -
C"""'Mn~;nfl
tlOn at pOint R
Amount j
to Ihis 150
An10tJni I
of capital.
of ("pitat
fsoquent
WMlS
Arnollnt of labor
of one
used. In
of a dollar's worth
bundle such that
MPa
p.
where
!:'!PIJ
Pb
l1n
Amount of labO!
CORNER SOLUTIONS
P"
products
a, b, .. , 11.
bundle that minimizes production costs, we
along the isoquant of the stipulated
that lies on the lowest isocost curve-for example, S
Input bundles on isocost curves like Co that iie below S are cheaper
the desired output.
like Cz that lie above S
is obvious that the optimal bundle S is a point where the isocost curve is tafi!\cnJijl,
to the isoquant. Therefore, to minimize the cost of producing a
to maximize the output from a given cost outlay, the finn must equa te
and P/Pi> this means that
needed, the manager must
MP'F'"
an isocost curve. In the two- input case, this means that just one input is used
in the least expensive way
to produce the most output
4.6 will now be an inequality reading MPKIPK >
cases where just capilal is used and MPKIPK MP,lPL for cases where
labor is used. The former case is shown in Figure 4.1 O.
111
flETurlNSTO SCALE
FIGURE 4.10
---------
-----
At first glance, some managers may believe thai production functions neces
exhibit constant returns to scale. After all, if a manager can build two fa(
size and Iypes of workers, can't she achieve the same
twice the size' Blit things arc not this simple. jf man
Amount
olr.apltal
M'I!'K
!!.Q
!!.T
MP,
~~;'O!Iffi;.lU>~I:.~
03
f2 + 12T 0.5P
(4.7)
12 -'. T
2,000
20 - 2E
4,000
2,000(20 -2E)
4,000
0,
12
10
12
E 12 ~ T
T E-l 2
M'IP,
Amount of tabor
Pc
(4.8)
P7
have
4,000E + 2,ooOT
Suppose we consider a
managers increase the level
to
112
28,000
IE + 21 lor Tgives us
, and PI is the wage of a lech
!!.Q
!!.E = 20
2E
(4.9a)
4,000E + 2,000( + 2)
28,000
MPL ~. ~C!
ilL
5( KL
MPK ~ ilK
._\ 1<
~.
5K
80L
which means that K .cc 4L. Because 0 ~ 800,
OUTPUT ELASTICITY
increasing, decreasing, or constant returns to
The output elasticity is dcfilled as the percent
in omput resulting from a 1 percent increase in all inDuts. If the
increasing returns to
if it
are constant returns to scale; and if it is Jess than 1, there are decreasing returns
a maker of aircraft parts,
function:
nrArll1rti()IJ
aLbKc
(4.10)
CWAiH OPTION,
Year
=c
baLbWIL = b(OIL)
to trayel at 15 knots.
requires 42.500 horseoower to do the same. So 2.7
times the
of labor
be
S()4tCS;
baL'J-iKC =
blAPd
www,oceanatiils.orglunattas/uses/transportalion;
logO
ioga+blogL+clogK
(4,11)
Note that if managers use the Cobb-Douglas form, they can easily estimate
returns
to scale. If the Sllm of the exponents (that is, b + c) exceeds 1, increas
To calculate the output elasticity at the Lone Star
returns
to scale are indicated; if the sum of the exponents equals l, constant
to Qif we multiply ooth inputs (Land K) oy 1.01.
to
scale
prevail; and jf the sum of the exponents is less than 1, decreasing
Q(that is,
to scale are indicated. This is true because if the Cobb-Douglas produc
0' =
the output elasticity equals the sum of the exponents, For
section the output elasticity of the Lone Star Company
the sum of the exponents (OJ and 0.8).
There is no cut-and-dried way to determine which mathematical form is best
= 1.0 11 0054840
:;.-..v"aUse the answer depends on the particular situation. Frequently a good proce
dure is to try more than one mathematical form and see which fits the data best
Therefore, If a manager illcreases the use of both
The
important thing is that the chosen form provide a faithful representation of
more than Ll percent; this means the output
thc actual situation. '1'0 dctermine whether this is the case, it often is useful to see
nnwwim,tpiv 1.J. It is exactly U for all infinitesimal
is larQcr than infinitesimal, the inrrp1,P~1 how weJl a oarticular estimated oroduction function can forecast the Quantity of
output
med.
in output is
a One-Pound Weight Gain for a Broiler and Isocost Curve If Corn Price Is %
Oilmeal Price
input combination i5 1.35 pounds of corn and 0.61 pounds of soybean OIlrnea!.
lor hOI,omle Cooperalion and Devclopment.lnterdisciplin;;ry Reseqrch
PouIK1s 01
pmilrtiler
0.61
1.35
110
PROBLEMS
SUMMARY
1. The production function. defines the relationship among variolls .
and the maximum quantity of a good that can be procluced.l\1anagers study
duction hmctions to gain insights into the firm's cost structure.
2. An isoquant is a curve showing all possible (effJCient) com' .
inpuls capable .of producing a panicuiar quantity of oUlpl1L Th~
.of technical substitution shows the rate at which one input Gill be
for allother
to
isocost curve.
4. lf a manager increases all
increases by more
than this
retUrllS to scale,
of inputs, various geometrical relations, or
returns to scale can also occur; the most frequently cited reaSO]1 is the
CUllY of managing a huge enterprise. Whether there are constant, i
or decreasing returns to scale is an empirical question that mnst be settled
by case.
5. Managers have estimated production functions in many firms
industries. Many studies show that a Cobb-Douglas function is the best fit
the data.
wwnorton.com/studyspace ~
PROBLEMS
t In the
the
between output
and unskilled labor ( U) is
Q 3m') + tOOU
0.25 2
0.3U2
and
is $5. The firm can hire as much labor as it wants at these wage rates.
a.
chief
recommends that the firm hire 400 hours
of skilled labor and 100 hours of unskilled labor. Evaluate this
recommendation.
b. If the Elwyn Company decides to spend a total of $5,000 on skilled and
unskilled labor, how many hOllrs of each type of labor should it hire?
c. If the price of a unit of output is $10 (and does not vary with output
level), how many hours of un;;killed labor ,hould the company hire?
120
Pounds of Grain
130.9
125.1
120.1
15.7
1.8
108.3
102.3
97.4
93.8
esti~
Q = 0.9P + 0.06L
where Qis the llumber of pounds of stationery produced by Ascot per year,
l. is the number of hours of labor per year, and Pis the number of pounds of
paper used per year.
a. Does this production function seem to include all the relevant in puts).
Explain.
1'}1
CIlAPTEI~ I"
PRODUCTION THEORY'
tOMI3INAIIUN~
is fixed at 6,500 _
wheD each amount is used.
returns?
4. A Cobb-Douglas production function was estimated for six types of
There were five inpnts in the production function: (lJ land, (2) labor,
livestock and feed, and (5) other resolllc(' services. The
was as follows:
Exponent
Livestock
Farm Type
Land
Labor
Equipment
and Feed
Crop farms
farms
farms
General farms'
Large farms
Small farms
0,24
0.07
0,02
0.01
0.08
0.10
0.06
0.16
0.11
0.08
DC"
.J..)
0.02
0.74
0.03
0.63
0.02
0,46
0.03
003
0.03
0.07
0.10
0.17
0.28
0.21
0,12
0,01
0.05
0.53
0.43
-"-"-"-~-
the price of labor increases to $2 per uniL What effect will this
have on output per unit of labor?
c. Is this plant subject to decreasing returns to scale? ','lhy or why not?
Volvo A,B., the Swedish auto firm, operated a car
in 1988, The idea was to have a small team of highly skilled workers build an
entire car. According to the proponents, this would reduce the tedium associated
with the conventional assembly Iim~ and cut absenteeism and turnover among
workcIs.ln J99 J there were reports that it took 50 hours of labor to assemble a
car at Uddevalla, in contrast to 25 hours at Volvo's conventional
at Ghent, Belgium. If you were Volvo's chief executive officer, what
take?
would you ask Uddevalla's managers, and what
is as follows:
Amount of Milk
(Pounds)
1,200
1,800
2,400
5,917
'1,250
8,379
3,000
9,371
0=
+ X.Po
.~
OJ)
E-'
1')1
CHAPTER
I,;
PI(O[)UCTION TH[ORY
by the rf)l"rp<:ntinnil1o
where PI is the
manager seeks IU)''':t,u'"11C
wants to maximize Q, where
viP,
Q~
MP2
P2
P1
P2
f'
L,
+ )..(f> . Xl,
f\X"
X,P2)
respect to
decision rule in
managers want to equate the
of Lagrangian
constraint.
aX
'AP,
X,P1
f'
iJ'A
Q'
[(X"
)..P?
X2P2
ilX2
iJX2
Thus
of output, Q'
problem is to minimize
(4.22)
0'
(4.23)
we first construct
DQ ~ MP 1
ax,
iJX,
al\X" X2 ) =
iJX2
ilQ ~ MP?
iJX2
)..P1
=0
? -
iJX; -
iiA
MP2
'hP2
A ill(X 1,X2)
= Q'
(4,24)
=0
(4,25)
aX 1
12<1
(4.21)
iJX2
l\Xl' X2)
14.26}
for
and
for
CHAPTER 5
~-
P,
P2
LEi'.RNING 08Jf:ClIVES
P, = 'AMP,
P2
side
(1127) by
we find that
P2
r(onomies
or
EANALYSIS OF COSTS
P,
P"
all business
comparisons of costs and benefits. A manager wants to under
an action if the additional (marRinall revenue attributable to that action
wishes to produce at an output level where the
cost. Obviously this calculation is not
many four-lettered words) jnvokes multiple interpretations. Man
find that what seems like a simple concept often provokes controversy over
accounting in virtually every MBA program),
of cost is necessary for a variety of basic
cost control, and
for future
consideration of costs must include both short-run and
consequences for an
Inntr_
managerial decisions
170
tJ.t'Tn
vision.
127
Appendix 8: Measurement of
Short-Run Cosl Functions; H,c_
Choice of a Mathematical Fon-_;