ACTA UNIVERSITATIS DANUBIUS Vol 11, no 1, 2015

Microeconomics

The Complete Theory of Cobb-Douglas Production Function

Cătălin Angelo Ioan1, Gina Ioan2

Abstract. The paper treats various aspects concerning the Cobb-Douglas production function. On the
one hand were highlighted conditions for the existence of the Cobb-Douglas function. Also were
calculated the main indicators of it and short and long-term costs. It has also been studied the
dependence of long-term cost of the parameters of the production function. The determination of
profit was made both for perfect competition market and maximizes its conditions. Also we have
studied the effects of Hicks and Slutsky and the production efficiency problem.
Keywords: production function; Cobb-Douglas; Hicks; Slutsky

1. Introduction
To conduct any economic activity is absolutely indispensable the existence of
inputs, in other words of any number of resources required for a good deployment
of the production process. We will assume that all resources are indefinitely
divisible.
We define on Rn the production space for n fixed resources as SP=(x1,...,xn)xi0,
i= 1, n  where xSP, x=(x1,...,xn) is an ordered set of resources and, because inside
a production process, depending on the nature of applied technology, not any
amount of resources is possible, we will restrict production space to a convex
subset DpSP – called the domain of production.
We will call a production function an application:
Q:DpR+, (x1,...,xn)Q(x1,...,xn)R+ (x1,...,xn)Dp
which satisfies the following axioms:

1
Associate Professor, PhD, Danubius University of Galati, Faculty of Economic Sciences, Romania,
Address: 3 Galati Blvd, Galati, Romania, Tel.: +40372 361 102, Fax: +40372 361 290,
Corresponding author: catalin_angelo_ioan@univ-danubius.ro.
2
Assistant Professor, PhD in progress, Danubius University of Galati, Faculty of Economic Sciences,
Romania, Address: 3 Galati Blvd, Galati, Romania, Tel.: +40372 361 102, Fax: +40372 361 290, E-
mail: gina_ioan@univ-danubius.ro.
AUDŒ, Vol. 11, no. 1, pp. 74-114
74
 ŒCONOMICA

A1. Q(0,...,0)=0;
A2. The production function is of class C2 on Dp that is it admits partial derivatives
of order 2 and they are continuous on Dp;
A3. The production function is monotonically increasing in each variable, that is:
Q
0, i= 1, n ;
x i
A4. The production function is quasi-concave (see Appendix).
Considering a production function Q:DpR+ and Q0 R+ - fixed, the set of inputs
which generate the production Q0 called isoquant. An isoquant is therefore
characterized by: {(x1,...,xn)DpQ(x1,...,xn)= Q0 } or, in other words, it is the
inverse image Q 1 Q 0  .
We will say that a production function Q:DpR+ is constant return to scale if
Q(x1,...,xn)=Q(x1,...,xn), with increasing return to scale if
Q(x1,...,xn)>Q(x1,...,xn) and decreasing return to scale if
Q(x1,...,xn)Q(x1,...,xn) (1,) (x1,...,xn)Dp.

2. The Cobb-Douglas Production Function

The Cobb-Douglas function has the following expression:
Q:D R n -{0}R+, (x1,...,xn)Q(x1,...,xn)= Ax11 ...x n n R+ (x1,...,xn)D,
A R * , 1,...,nR*
Computing the partial derivatives of first and second order, we get:
iQ
Q'x i  i Ax11 ...x i i 1...x n n  i= 1, n
xi

 1 i  jQ
Q"x i x j  i  jAx11 ...x i i 1...x j j ...x n n  ij= 1, n
xix j

 i  i  1Q
Q"xi xi   i  i  1Ax11 ...x ii 2 ...x n n  i= 1, n
x i2
Let the bordered Hessian matrix:

75
ACTA UNIVERSITATIS DANUBIUS Vol 11, no 1, 2015

 1Q  2Q nQ 
 0 ... 
 x1 x2 xn 
 1Q 1 1  1Q 1 2Q 1 n Q 
 x ... 
 1 x12 x1 x 2 x 1x n 
H (Q)    2 Q 1 2Q  2  2  1Q  2 n Q 
B
...
 x x1 x 2 x 22 x 2x n 
 ...2 ... ... ... ... 
 Q 1 n Q  2 n Q  n  n  1Q 
 n ...
 x x 2n 
 n x1x n x 2x n 
k k
 i  i
We find (not so easy): Bk =  1 Q k 1 i 1 i 1
k
2
, k= 1, n .
 k 
  xi 
 i1 
k k
 i  i k k
Because (-1)k Bk = Q k 1 i 1 i 1
2
, if   i   i 0, k= 1, n it follows that the
 k  i 1 i 1
  xi 
 i1 
function is strictly quasi-concave. Also, if the function is quasi-concave we have
k k
that   i   i 0.
i 1 i 1

iQ
But from the axiom A3 we must have that Q'x i  0 that is i0. After these
xi
considerations we have that if i0, i= 1, n the Cobb-Douglas function is strictly
quasi-concave.
n
We have now: q1 ,..., n1  = Q1 ,..., n1 ,1 = A11 ... n n11 and r=   k .
k 1

The main indicators are:

iQ
  x i = A i x11 ...x ii 1...x n n = , i= 1, n
xi
Q
 w x i = Ax11 ...x ii 1...x n n = , i= 1, n
xi

76
 ŒCONOMICA

i x j
 RMS(i,j)= , i,j= 1, n
 jx i
i
 RMS(i)= , i= 1, n
n 1  2

j
xi 2
j1 x j
ji

  x i = i , i= 1, n
 ij=-1, i,j= 1, n
Reciprocally, if for a homogenous production function of degree r:  x i = i , i= 1, n
n 1 q
q
rq   i
 i i 1  i
we have that:   i , i= 1, n  1 and  n .
q q
i
But now, we have:
q q
 ln 
q q  ln i i  i  ln  ii  ln q  ln ii
i i
 i  q      0
i i i q  i i i i

= Fi 1,..., ˆ i ,..., n1  (where ^ means that the variable is missing).

q
ln
 i i
Fi
We have now: q   ii e Fi 1 ,...,i ,...,n 1  . For ji we obtain now:  j   j
ˆ
therefore:
 j
Fi  j
 . Integrating with respect to  j :
 j  j
 

Fi   j ln  j  g i 1 ,..., ˆ i ,...ˆ j ,...,  n 1  therefore:

q  ii  j j e
gi 1 ,...,ˆ i ,...ˆ j ,...,n 1
.
Analogously, by recurrence: q  A11 ... n n11 with A=constant with respect to
n 1 q
rq   i
i 1 i
n 1 n
1 ,..., n1 . But: n   r   i  r    i . After these
q i 1 i 1
n
considerations it follows that if it is homogenous of degree r, r must be  i .
i 1

77
ACTA UNIVERSITATIS DANUBIUS Vol 11, no 1, 2015

Finally: q  A11 ... n n11 implies that:

n 1
r   k
x11 x n n11
Q(x1,...,xn)= x rn q 1 ,...,  n 1  = Ax rn ... = Ax k 1
x11 ...x n n11 =
x n 1 x n n 1
n

Ax11 ...x n n11 x n n - the Cobb-Douglas production function.

Considering now again the Cobb-Douglas production:
1 n
Q(x1,...,xn)= Ax1 ...x n let search the dependence of the parameters 1,...,n.

Q
We have: = Ax11 ...x n n ln x i = Q ln x i 0 xi1, i= 1, n . From this relation we
 i
have that at an increasing of a parameter i the production Q will increase also.
In particular, for the Cobb-Douglas function related to capital K and labor L:
Q=AKL we have that the main indicators are:
 K = AK 1L
 L = A K  L 1
 w K = AK  1L
 w L = AK  L 1
L
 RMS(K,L)=RMS(K)=
K
K
 RMS(L,K)= RMS(L)=
L
 K =
  L =
 =KL=-1

3. The Costs of the Cobb-Douglas Production Function

Considering now the problem of minimizing costs for a given production Q0,
where the prices of inputs are pi, i= 1, n , we have:

78
 ŒCONOMICA

 n

 min  pk x k
k 1
 1  n
Ax1 ...x n  Q0
x ,..., x  0
 1 n

 1 
  ...  n
From the obvious relations:  p1x1 p n x n we obtain:
Ax 1 ...x  n  Q
 1 n 0

 k pn
x k  x n , k  1, n - 1
 n pk and from the second equation:
Ax 1 ...x  n  Q
 1 n 0
n 1
  k n 1
p nk 1   k k n
 k n
A n 1
k 1 k 1
xn  Q0 . Noting r=   k we finally obtain:
  k n 1 k 1
 nk 1  pk k
k 1

1/ r
 n k 
  pk 
 k 1   k Q10/ r
xk = 1/ r 1/ r
, k= 1, n
 n  k  pk A
  k 
 k 1 
The total cost is:
1/ r
 n i 
  pi  1/ r
TC=  pk x k =  i 1  rQ 0 .
n

1/ r 1/ r
k 1  n i  A
  i 
 i 1 

79
ACTA UNIVERSITATIS DANUBIUS Vol 11, no 1, 2015

At a price change of one factor, i.e. xk, from the value pk to pk we have:
1/ r
 n 
 i 
  pi  pk
k / r

 ik
i 1  rQ10/ r
TC =  
1/ r
 n i  A1 / r
  i 
 i 1 
k / r
TC TC  TC  pk 
where the relative variation of the total cost is:     1.
TC TC  pk 
Let us now consider the behavior of the total cost of production function at a
parameters variation. We have:
1
 n p i r
 i Q   n
 j 
 in1 0   j1 n
i

 Ap k   i 
j k
  i  A 
TC  i1 i   
i 1
   ln
i k
 r
 k 2 n
 n    j 
 j  j1 n
 Q 0  kj k  p ii 
 j1   i 1 
 ik 
n
j
j1 n

n
j k
Apk  i i
n j i 1
TC   j j1
ik
n
Therefore:
 k
0 e j1  kjk  n
. If we note:     j 0 and
Q0  pi i j1
j k
i 1
ik
n
 j
j1 n

Ap kj k
 i i
i 1
ik TC M
M= 0 we obtain that: 0 e  k  k   .
n
Q 0  p i i  k e
i 1
ik

Considering now the function f:(0,)R, f(x)= x e x we have

f ' ( x )  x e x    0 therefore f is strictly increasing. Because lim x  e x  0
 1 x
x 0

80
 ŒCONOMICA

M
and lim x  e x   the equation: x e x  has a unique solution  0k called cost
x  e
threshold with respect to the k-th parameter. After these considerations we have
that for k  0k the total cost will increase at an increasing of k and after it will
decrease.
The situation may seem paradoxical that at the growth of the elasticity of one input,
total cost increases. Fortunately, due to the sharp rise of f, the values of  0k are
very small so it does not significantly affect processes.
Like an example, considering the production function Q(K,L)=K L, ,0 we
have that the behavior of 0 related to  is (for Q=5):

0.06

0.05

0.04

0.03

0.02

0.01

0.2 0.4 0.6 0.8 1.0

Figure 1
The long-term total cost for =0.7 and variable  is shown in figure 2:

81
ACTA UNIVERSITATIS DANUBIUS Vol 11, no 1, 2015

TotalCost

22

20

18

16

14
0.0 0.2 0.4 0.6 0.8

Figure 2
where the maximum value is reached for =0.025.
If we consider for a given output Q0, the inputs x1,...,xn such that: Ax11 ...x n n  Q0
1
k
Q0
let x k  1 1 n
where ^ means that the term is missing.
A x1 ...x̂ k ...x n k
k k

1
n n k
Q
We have STCk=  p i x i =  p i x i  pk 1  0
1 n
representing the short-
i 1 i 1
ik A x1 ...x̂ k ...x n k
k k

term total cost when factors x1,..., x̂ k ,..., x n remain constant.

We put now the question of determining the envelope of the family of
hypersurfaces:
1
n k
f Q0 , x1,..., x̂ k ,..., x n  =  p i x i 
Q
pk 1  0 n
1
i 1
ik A x1 ...x̂ k ...x n k
k k

Conditions to be met are:

82
 ŒCONOMICA

TC  f Q0 , x1 ,..., x̂ k ,..., x n 

 f
  0, i  1, n, i  k
 x i

After the elimination of parameters x1 ,..., x̂ k ,..., x n we have either the locus of
singular points of hypersurfaces (which is not the case for the present issue) or
envelope sought.
We have therefore:
 1
k
 n Q0
TC   pi x i  p k 1 1 n
 i 1
 ik A x1 ...x̂ k ...x n k
 k  k

 1
 k

p i  i p Q
1
k 0
i n
 0, i  1, n , i  k
 k 1
1
 A  k x1 k ...x i k ...x̂ k ...x n k
1
k
Q0
Noting:  = 1 1 n
it follows:
A  k x1 k ...x̂ k ...x n k

 n

TC   pi x i  p k 
 i 1
 ik
p  i p k   0, i  1, n, i  k

 i  x
 k i

 p   p 1 ...p n n Q10/ r
from where: x i  i k , i  1, n, i  k . Finally:  = k 1
 
1/ r

 
and
 k pi A1 / r p  1 ...  n
1/ r
k 1 n
replacing:
1/ r
 n 
 i   p ii  Q10/ r
x i   i1 
1/ r
, i  1, n , i  k
1/ r  i 
n
A pi   i 
 i1 

83
ACTA UNIVERSITATIS DANUBIUS Vol 11, no 1, 2015

1/ r
 n i 
  pi  1/ r
TC=  i 1  rQ 0
1/ r 1/ r
 n i  A
  i 
 i 1 
We obtained so that the envelope of the family of hypersurfaces of the short-term
total cost when all inputs are constant except one is just the long-term cost obtained
from nonlinear optimization problem with respect to the minimizing of the cost for
a given production.
Calculating the costs derived from the (long-term or short-term) total cost now, we
have:
1/ r
 n i 
  pi  1 / r 1
TC  i 1  rQ 0
ATC= = (average long-term total cost)
Q 0  n  1 / r A1 / r
  i i 
 i 1 
1/ r
 n i 
  pi  1 / r 1
TC  i1  Q 0 = ATC (marginal long-term total cost)
MTC= =
Q 0  n  1/ r A1/ r r
  i i 
 i1 
n
 pi x i 1
1
i 1 k
STC k i k Q0
ASTCk= =  pk 1  n
(average short-term total cost)
Q0 Q0 1

A x1 ...x̂ k ...x nk

k k

1
1
k
STC k Q0
MCk= = pk 1 n
(marginal short-term total cost)
Q 0 1

 k A x1 ...x̂ k ...x n k
k k

1
k
Q
VTCk= p k 1  0 n
(variable short-term total cost)
1

A x1 ...x̂ k ...x n k
k k

84
 ŒCONOMICA

1
1
k
VTC k Q0
AVTCk= = pk 1  n
(average variable short-term total cost)
Q0 1

A x1 ...x̂ k ...x n k
k k

n
FTCk=  p i x i (fixed short-term total cost)
i 1
ik

n
 pi x i
i 1
FTC k i  k
AFTCk= = (average fixed short-term total cost)
Q0 Q0
n
 pi x i 1
1
i 1
i k Q 0k
The extreme of the function ASTCk(Q0)=  pk 1 1 n
are given
Q0 k k
A x1 ...x̂ k ...x nk
by:
1 1 1 n
 1  n
p k   1Q 0k  A  k x1 k ...x̂ k ...x nk  p i x i
 k  i 1
ASTC k ' Q0  =
ik
1 1 n
=0 from where:
k k
Q 02 A x1 ...x̂ k ...x n k
k
 n 
n  
Ax1 ...x̂ k ...x n   p i x i 
1

 ii1k 
Q 0,droot    when k1 and the minimum value is:
k
 1 
p k k   1
 k 
 k 1
1 
p k k 
 1
ASTCk Q0droot  =  k  .
 k 1
n 
 
A k x11 ...x̂ k ...x n n   p i x i 
 ii1k 
 

85
ACTA UNIVERSITATIS DANUBIUS Vol 11, no 1, 2015

If k1 it follows that ASTC k ' Q0  0 therefore the average short-term total cost
will decrease.
Finally we have:
CT
p k 
 pk   k - the coefficient of elasticity of long-term total cost with respect
CT r
pk
to the price factor i
CT
Q 0 1
Q   - the coefficient of elasticity of long-term total cost with respect to
CT r
Q0
the production Q0
ATC
p k 
 av ,pk   k - the coefficient of elasticity of average long-term total cost
ATC r
pk
with respect to the price factor i
MTC
p k 
 m arg, pk   k - the coefficient of elasticity of marginal long-term total
MTC r
pk
cost with respect to the price factor i
In particular, for the Cobb-Douglas function related to capital K and labor
L: Q=AKL we have:

p p 
  1 /(  )
K L  Q10/( )
  
K=
  1 /(  ) pK A1 /( )

p p 
  1 /(  )
K L  Q10/( )
  
L=
  1 /(  ) p L A1 /( )

86
 ŒCONOMICA

p p 
  1 /(  )
K L (  )Q10/(  )
  
TC=
  1 /(  ) A1 /( )
1

Q0
On the short-term, we have for constancy of K: STCL= p K K  p L 1 
and
 
A K

p p 
  1 /(  )
K L (  )Q10/( ) 1
  
ATC=
  1 /(  ) A1 /( )

p p 
  1 /(   )
K L Q10/( ) 1
  
MTC=
  1 /(   ) A1 /( )
1
1

pK K Q0
ASTCL=  pL 1 
Q0  
A K
1
1

pLQ0
MCL= 1 
 
A K
1
Q 
VTCL= p K 1 0 
A  L
1
1
Q0
AVTCL= p K 1 
A  L
FTCL= p L L

pLL
AFTCL=
Q0

87
ACTA UNIVERSITATIS DANUBIUS Vol 11, no 1, 2015

ApK K 
The extreme of the function: ASTCL(Q0) is given by: Q 0,d root  
when
1 
p L   1
 
pL 1  
1
1 and the minimum value is: ASTC L Q0droot  .
ApK1 K 1

  1  
 pK  ,  pL  , Q  ,  av ,pK  ,  av ,pL  ,
    
 
 m arg, pK  ,  m arg, pL  .
 

4. The Profit
Now consider a sale price of output Q0: p(Q0). The profit is therefore:
(Q0)= p(Q0)Q0-TC(Q0)
It is known that in a market with perfect competition, the price is given and equals
marginal cost. The profit on long-term becomes:
(Q0)=p(Q0)Q0-TC(Q0)=MTC(Q0)Q0-
1/ r
 n i 
  pi  1 / r 1
TC(Q0)= ATC ' (Q 0 )Q 0 =  i1
2  rQ 0
1/ r
 n i  A1 / r
  i 
 i1 
In particular, for the Cobb-Douglas function related to capital K and labor L:

Q=AKL we have: (Q0)=

p p  
  1 /  
K L   Q10/ r1 .
   
  1/   A1/ r

On short-term, when factors x1,..., x̂ k ,..., x n remain constant, we have:

(Q0)=p(Q0)Q0-STCk(Q0)=MTC(Q0)Q0-STCk(Q0)= AVTC k ' Q 0 Q 02  FTC k

therefore:

88
 ŒCONOMICA

1
k
 1  Q n  1  n
(Q0)=   1p k 1  0 n
  p i x i =   1p k x k   p i x i
 k   k 
1
i 1 i 1
A x1 ...x̂ k ...x n k
k k ik ik

Like a conclusion, the company will make a profit in the short-term if, under
constancy factors x1,..., x̂ k ,..., x n , the amount of factor xk will be higher than
n n
 pi x i  pi x i
i 1 i 1
ik ik
if k1 and less than if k1. If k=1 then the firm will
 1   1 
  1p k   1p k
 k   k 
incur losses.
For Q=AKL we have that if K=constant, the company will make a profit in the
p K K p K K
short-term in the case 1 if L and in the case 1 if L . If
1  p L 1  p L
=1 the firm will incur losses.
The condition of profit maximization for an arbitrarily price p, depending on the
 n

factors of production, is: max x1 ,..., x n  = max  pQx1 ,..., x n    p i x i  from
 i 1 
Q p i Q pi  pQ
where  , i= 1, n or otherwise:  i  and finally: x i  i . Because
x i p xi p pi
Q is quasi-concave the solution of the characteristic system is the unique point of
maximum. How Q= Ax11 ...x n n we obtain that the appropriate production is:
1
n
 n k  r 1
Ap r
k k
  pk 
Ax 11 ...x n n therefore, if r1: Q   k 1 n 
r
k 1
Q = Q and the
n  r 
 pk  Ap   k
k k

k 1  k 1 
1
 n
 r 1
  ir 1i  p k k 
 k 1 
k i
factors: x i    , i= 1, n .
 n

 App r 1  i
i   k k 
 k 1 
 k i 

89
ACTA UNIVERSITATIS DANUBIUS Vol 11, no 1, 2015

1
 n k  r 1
  pk 
The maximum profit is: x1 ,..., x n  = p  r  k 1n  .
 r 
 Ap   k
k

 k 1 
n n
If r=1 the necessary condition for profit maximization is:  pk k
 Ap  kk or p
k 1 k 1
n
 pk k

k 1
must be: p= n
therefore the amount of factors are not independent, that is,
A  k k
k 1
n
Aps   k k
i ps
for a fixed factor, let say xs: Q  k 1
x s and: x i  x s , i= 1, n , is. The
n
s pi
 s  p k k
k 1
profit is: x1 ,..., x n  =0 for any amount of xs.
For Q=AKL we have that, if +1:
1 1 1
 p  p  1   1pL  1  1p K  1
Q   rK L   , K   
1  
, L    ,
 App 1   
 Ap     App K    L 
1
 p  p    1
K, L   p      K L  
 Ap   
and if +=1 the necessary condition for profit maximization is:
p1K pL p K A pK
p , L  K , Q  K , K, L   0 .
A1   
1 
1  p L 1   pL
At a variable price p(Q) we have now: (Q)=p(Q)Q-CT(Q) therefore the
necessary condition for profit maximization is  ' (Q) =0 therefore:
p' (Q)Q  p(Q)  MTC (Q)  0 .

90
 ŒCONOMICA

Substituting the expression of MTC we obtain: p' (Q)Q  p(Q)  Q1 / r 1  0 where
1/ r
 n i 
  pi 
we noted:  =  i 1 
1/ r
. But this differential equation gives us:
1/ r  i 
n
A   i 
 i 1 
C
p(Q)  rQ1 / r 1  , CR+ and the profit (Q)=p(Q)Q-CT(Q)=C. Therefore, for
Q
C
the maximum of the profit (Q)=C we must have the price p(Q)  rQ1 / r 1  ,
Q

CR+. Because: p' (Q) 

1  r Q1 / r  C we have that if r1, the price will
Q2
r
 C 
decrease with production and if r1 for Q    the price will decrease and
 1  r  
r
 C  C
for Q    the price will increase. If r=1 we have that p(Q)    and
 1  r   Q
the price will decrease with production.
1 /    
 p  p 
For Q=AKL we have    K  L  and for the profit (Q)=C we must
 A  
have the price p(Q)    Q1 /   1 
C
. If +1, the price will decrease
Q

 C 
with production and if +1 for Q    the price will decrease and
 1     
 
 C 
for Q    the price will increase. If +=1 we have that
 1     
C
p(Q)    and the price will decrease with production.
Q

91
ACTA UNIVERSITATIS DANUBIUS Vol 11, no 1, 2015

5. The Hicks and Slutsky Effects for the Cobb-Douglas Production

Function
Now consider the production function Q(x1,...,xn)= Ax11 ...x n n and factor prices
pi i 1, n . The non-linear programming problem relative to maximize production at
a given total cost (CT0) is:
max Ax11 ...x n n
 n
  p k x k  CT0
k 1
x1 ,..., x n  0

Because the objective function is quasi-concave and also the restriction (being
affine) and the partial derivatives are all positive we find that the Karush-Kuhn-
Tucker conditions are also sufficient. Therefore, we have:
 Q Q
 x x1 ,..., x n  xn
x1 ,..., x n 
 1
 ... 
 p1 pn
n
  p k x k  CT0
k 1
From the first equations we obtain:
 1 n
 p x  ...  p x
 n1 1 n n

  p k x k  CT0
k 1

therefore:
 k pn
x k   p x n , k  1, n - 1
n n k

  p k x k  CT0
k 1

92
 ŒCONOMICA

Substituting the first n-1 relations into the last we finally find that:
 CT
x 0,k  k 0 , k  1, n and the appropriate production:
rp k
n
  i i

i 1
Q0(x1,...,xn)= A n
CT0r .
r r
 pi i

i 1

Suppose now that some of the prices of factors of production (possibly after
renumbering, we may assume that they are: x1,...,xs) is modified to values p1, , ,.ps ,
the rest remain constant.
From the above, it results:


x f ,k   k CT0 , k  1, s
 rp k



  k CT0
 x f ,k  , k  s  1, n
 rp k

 n
   ii
Q  A i 1
CT0r
 f s n
j
 r  pi  p j
r i

 i 1 js 1

We will apply in the following, the method of Hicks. To an input price change, let
consider that it remains unchanged, leading thus to a change of the total cost. We
therefore have:
n n
 i i
r
 i i

A s
i 1
n
CT0 A i 1
n
CT0r
 
r r
 pi i  pj j r r
 pj j
i 1 j  s 1 j 1

from where:

93
ACTA UNIVERSITATIS DANUBIUS Vol 11, no 1, 2015

r
 pi i

CT0  i 1
s
CT0r

 pj j
j 1

With the new total cost, the optimal amounts of inputs become:
  s 
1/ r

  k   pi i  CT0
x   i 1  , k  1, s
 int,k  s 
1/ r
 
rpk   p j j 
  j1 
 1/ r
  s

  d   pi i  CT0
 i 1 
x int,d  1/ r
, d  s  1, n
  s
j 
 rp d   p j 
  j1 
The Hicks substitution effect which preserves the production is therefore:
  p s  p  i / r   CT
1H x k  x int,k  x 0, k   k   i   1 k 0 , k  1, s
  pk i 1  pi   rp k

  s  p  i / r   CT
1H x d  x int,d  x 0, d    p   1 d 0 , d  s  1, n
i

 
  i
i 1  rp d

The difference caused by the old cost instead the new total cost one is therefore:
  s 
p    k CT0
i / r
 2 H x k  x f , k  x int,k  1      i
, k  1, s
  i 1  pi   rpk

  s 
pi    d CT0
i / r

 2 H x d  x f , d  x int,d  1    p   rp , d  s  1, n
  
  i 1  i   d

Let now calculate the new prices influence to the effects of substitution and of new
cost in the Hicks effect.
We have:

94
 ŒCONOMICA

i / r
1H x k p s p   k CT0 n

p k
  k2   i   i , k  1, s
pk i 1  pi  r 2pk i 1
ik

 i / r 
 2 H x k  k CT0  s  pi 
  2 2      j  r , k  1, s
n

p k r pk  i 1  pi  j1 
 j k 
i / r
s 
1H x k p   k  t CT0
   i  , k  1, s, t  1, s, t  k
pt i 1  pi  r 2 pk p t
i / r
s 
 2H x k p   k  t CT0
   i  , k  1, s, t  1, s, t  k
p t i 1  pi  r 2 pk p t
i / r
s 
1H x d p   k d CT0
   i  , d  s  1, n, k  1, s
pk i 1  pi  r 2 p d pk
i / r
s 
 2 H x d p  d  k CT0
   i  , d  s  1, n, k  1, s
pk i 1  pi  r 2 p d pk
After these relations, it follows that the effect of substitution at the increase of the
price xk, k= 1, s is reduced, while the effect of new cost is reduced if
s i / r  s i / r 
  pi     pi  
n n

  p    j  r   0 or it increase if   p    j  r  0 .
  i  j k   i  j k
i 1 j 1 i 1 j 1
 
We shall apply now the Slutsky method for our analysis.
At the modify of the price of the factors x1,...,xs, the total cost for the same optimal
combination of factors is:

CT0  n s p 
CTint     j   i i 

r  js1 i 1 p i 

therefore:

95
ACTA UNIVERSITATIS DANUBIUS Vol 11, no 1, 2015

  k CT0  n s p 
x int,k  2    j    i i , k  1, s
 
r p k  js1 i 1 p i 

x  d CT0  n s p 
 int,d  2

   j    i i , d  s  1, n
 r p d  js1 i 1 pi 

The appropriate production is:

n
 i CT0r
i
 n p 
r

Qint x int,1 ,..., x int,n  = A

s
i 1    j   i i 
s n  j s 1 pi 
r 2 r  pk k  p d d  i 1

k 1 d  s 1

The Slutsky substitution effect which not preserves the production is therefore:
  k CT0  n s p p 
1S x k  x int,k  x 0,k  2    j    i i  r k , k  1, s
 r p k  js1 i 1 pi p k 

 x  x  d CT0  n s pi 
 1S d int,d  x 0,d  2
  
 js1 j    i  r , d  s  1, n

 r p d  i  1 p i 
and the difference caused by the old production instead the new production one is
therefore:
  k CT0 s  p 
 2S x k  x f ,k  x int,k  2   i 1  p i , k  1, s
 r p k i1  i 

 x  x  x  CT s  p 
 2S d f ,d int,d 
d
2
0
  i 1  i , d  s  1, n
 r p d i1  pi 

Let us now calculate the influence of the new prices on the effects of substitution
and new cost in Slutsky effect.
We have:
 
1S x k  k CT0  n s pi 
  2 2    j   i , k  1, s
pk r pk  j s 1 i 1 pi 
 ik 
 
 2S x k  k CT0  s  pi  
  2 2   k   i 1   , k  1, s
p k r pk  i 1  pi  
 ik 
96
 ŒCONOMICA

1S x k  k  t CT0
 2 , k  1, s, t  1, s, t  k
p t r pk p t

 2S x k   CT
  k2 t 0 , k  1, s, t  1, s, t  k
p t r pk p t

1S x d  d  k CT0
 2 , d  s  1, n, k  1, s
p k r pd p k

 2S x d   CT
  d2 k 0 , d  s  1, n, k  1, s
p k r pd p k

Therefore the effect of substitution at a price increase of the factor xk, k= 1, s is

s  p 
reduced, while the effect due to production decrease in case  k    i 1  i  0
i 1  pi 
ik

 p 
s
or it increase if  k    i 1  i  0.
i 1  pi 
ik

6. Production Efficiency of Cobb-Douglas Production Function

Let now two Cobb-Douglas production functions for two goods  ,  and a
number of n inputs F1,...,Fn available in quantities x1, , ,.x n . The production
functions of  or  are:
Q  x1 ,..., x n   Ax11 ...x n n , Q  x1 ,..., x n   Bx 11 ...x nn

appropriate to the consumption of xk units of factor Fk, k= 1, n .

We have seen that: , x i  A i x11 ...x i i 1...x n n ,  , x i  Bi x11 ...x i i 1...x n n , i= 1, n .

The production contract curve satisfies:

 , x i A i x11 ...x i i 1...x n n
   , i= 1, n
Bi x1  x1  1 ...x i  x i  i ...x n  x n  n
  1 
 , x i

97
ACTA UNIVERSITATIS DANUBIUS Vol 11, no 1, 2015

 ji x jx i
xj 
Dividing for ij:
     x
j i i j i   i j x i
and for i=1:

 j1x jx1
xj 
  
 1 j x1  1 jx1
j 1
, j= 2, n . Finally, for xi= we have the equation of

production contract curve:

 x1  
  j1x j
x  ,R


j
 
j 1  
1 j     x
1 j 1 
If we consider now the input prices: p1,...,pn we have that for the production
contract curve: x1=g1(),...,xn=gn(), R:
x1  g1    
  j1x j
x  g    ,R


j j
 
j 1  
 
1 j     x
1 j 1 
and:

pj 
,x j (g1 (),..., g n ())

 1
A jg11 ()...g j j ()...g n n ()

 jg1 ()

         x
j 1 1 j 1 j 1

,x1 (g1 (),..., g n ()) A1g111 ()g 2 2 ()...g n n () 1g j () 11x j
, j= 1, n .

For =1 we then obtain: p1=1, p j 

 
j 1 
 1 j   1 jx1
, j= 2, n .
11x j

If the initial allocation of factors of production was x   a1,..., a n  we have that

n
 p j (a j  x j )  0 therefore:
j 1
n a j j n a j j
a111  1x1  a111  1x1 
xj
j 2 xj
j2
n
*  where r1=   k .
11  
n a j j1  a j1 j   j1x j
=
r11  
n 
a j  j1  1 j  k 1

j 2 xj j2 xj

x1  *
  j1x j*
For this value we find now the final allocation: 


x
 j      *    x
j 1 1 j 1 j 1  
98
 ŒCONOMICA

If now the two production functions are: Q  K, L   AK 1 L 2 ,

Q  K, L   AK 1 L 2 we have that for limited quantities of capital K  and labor
L  the equation of production contract curve is:
K  
  21L
L  , R
  21  12   12 K
and the final allocation for an initial one: x   K1, L1  :

K  *
 K1L11  1KL12
L   21L* where *  .
  21  12   12 K 1   2 1L  L1  21  12 

7. The Concrete Determination of the Cobb-Douglas Production

Function
Considering an affine function: f:RnR, f(x1,...,xn)=1x1+...+ nxn+ n+1 and a set of
 
m>n+1 data: x1k ,..., x kn , f k , k= 1, m the problem of determining  i, i= 1, n  1 using
the least square method is to minimize the expression:

 1x1k  ...  n x kn  n1  f k 

m 2
that is to solve the system:
k 1

 m k k m m m

  x
 1 k 1 1 i
x  ...   n  x k k
x
n i   n 1  x k
i   f k x ik , i  1, n
k 1 k 1 k 1
 m m m
1  x1  ...   n  x n  m n 1   f
k k k
 k 1
 k 1 k 1

Considering the matrix:

m k 2
  
m m
  x1  x1k x k2 ...  x1k 
 k 1 k 1 k 1 
 
m k k m m

   k ...  
2
x1 x 2 x k2 x k2
1 k 1 k 1
 ... ... ... ... 
 m m
  x1k  x k2 ... m 
 k 1 k 1 

99
ACTA UNIVERSITATIS DANUBIUS Vol 11, no 1, 2015

and  ij the cofactor of the (i,j)-element in  we will obtain:

m m m
1i  f k x1k  ...   ni  f k x kn   n 1,i  f k
k 1 k 1 k 1
i  , i  1, n  1
det 
Considering now a production function Q(x1,...,xn)= Ax11 ...x n n we put the problem
of concrete determination of the parameters A, i, i= 1, n .


Let therefore a set of m>n+1 data: x 1k ,..., x kn , Q k , k= 1, m . 
Considering the logarithm of Q, we have: ln Q  1 ln x1  ...  n ln x n  ln A
therefore we will modify the data set to the new one: ln x1k ,..., ln x kn , ln Q k ,  
k= 1, m .
From above:
m m m
1i  ln Q k ln x1k  ...   ni  ln Q k ln x kn   n 1,i  ln Q k
k 1 k 1 k 1
i  , i  1, n
det 
m m m
1,n 1  ln Q k ln x1k  ...   n ,n 1  ln Q k ln x kn   n 1,n 1  ln Q k
k 1 k 1 k 1
ln A 
det 
 m
  
m m
  ln x1k  ln x1k ln x k2  ln x1k 
2
...
 k 1 k 1 k 1 
 ln 
m m m

where    k  ln
2
ln x1k ln x k2 x k2 ... x k2 .
 1 ... k 1
... ...
k 1
... 
 m m 
  ln x1k  ln x k2 ... m 
 k 1 k 1 
For the Cobb-Douglas Q=Q(K,L)=AKL we have therefore, for the set:
K i , L i , Q i i1,m :

100
 ŒCONOMICA

m m m
 ln Q i ln K i  ln K i ln L i  ln K i
i 1 i 1 i 1
m m m
 ln Q i ln L i  ln 2
Li  ln L i
i 1 i 1 i 1
m m
 ln Q i  ln L i m
i 1 i 1
 m m m
,
 ln 2 K i  ln K i ln L i  ln K i
i 1 i 1 i 1
m m m
 ln K i ln L i  ln 2
Li  ln L i
i 1 i 1 i 1
m m
 ln K i  ln L i m
i 1 i 1
m m m
 ln 2 K i  ln Q i ln K i  ln K i
i 1 i 1 i 1
m m m
 ln K i ln L i  ln Q i ln L i  ln L i
i 1 i 1 i 1
m m
 ln K i  ln Q i m
i 1 i 1
 m m m
,
 ln 2
Ki  ln K i ln L i  ln K i
i 1 i 1 i 1
m m m
 ln K i ln L i  ln 2
Li  ln L i
i 1 i 1 i 1
m m
 ln K i  ln L i m
i 1 i 1

101
ACTA UNIVERSITATIS DANUBIUS Vol 11, no 1, 2015

m m m
 ln 2 K i  ln K i ln L i  ln Q i ln K i
i 1 i 1 i 1
m m m
 ln K i ln L i  ln 2
Li  ln Q i ln L i
i 1 i 1 i 1
m m m
 ln K i  ln L i  ln Q i
i 1 i 1 i 1
ln A  m m m
 ln 2 K i  ln K i ln L i  ln K i
i 1 i 1 i 1
m m m
 ln K i ln L i  ln 2
Li  ln L i
i 1 i 1 i 1
m m
 ln K i  ln L i m
i 1 i 1

8. Conclusions
The above analysis reveals several aspects. On the one hand were highlighted
conditions for the existence of the Cobb-Douglas function. Also were calculated
the main indicators of its and short and long-term costs. It has also been studied the
dependence of long-term cost of the parameters of the production function. The
determination of profit was made both for perfect competition market and
maximize its conditions. Also we have studied the effects of Hicks and Slutsky and
the production efficiency problem.

9. Appendix
A.1. Mathematical concepts
A function Q:DRnR, D – convex set, is quasi-concave if:
Q(x+(1-)y)min(Q(x),Q(y)) [0,1] x,yD
and is strictly quasi-concave if:
Q(x+(1-)y)min(Q(x),Q(y)) (0,1) x,yD
A function Q:DRnR, D – convex set, is quasi-convex if:
Q(x+(1-)y)max(Q(x),Q(y)) [0,1] x,yD
and is strictly quasi-convex if:
Q(x+(1-)y)max(Q(x),Q(y)) (0,1) x,yD
102
 ŒCONOMICA

Geometrically speaking, a quasi-concave function has the property to be above the

lowest values recorded at the ends of some segment. This property is equivalent
with the convexity of the set Q-1[a,)={xDQ(x)a} aR.
Note also that if f and g are arbitrary functions:
 f – quasi-concave (quasi-convex) implies that -f is quasi-convex (quasi-
concave);
 f – strictly quasi-concave (quasi-convex) implies that f is quasi-concave (quasi-
convex);
 f - quasi-concave (quasi-convex) implies that f is quasi-concave (quasi-
convex) for any 0;
 f,g - quasi-concave (quasi-convex) imply that min(f,g) (max(f,g)) is
quasi-concave (quasi-convex) for any ,0;
 f - quasi-concave (quasi-convex) and g:RR is increasing imply that
gf:DR is quasi-concave (quasi-convex);
 fC1(D) is (strictly) quasi-concave if and only if: f(x)f(y)
n f
 ( y)x i  y i   ()0 x,yD;
i 1 x i

 fC1(D) is (strictly) quasi-convex if and only if: f(x)f(y)

n f
 ( x )x i  y i   ()0 x,yD;
i 1 x i

 A monotonically function f:DRR is quasi-concave and quasi-convex;

 Any affine function is quasi-concave and quasi-convex.
Considering now the bordered hessian matrix:
 0 f 'x 1 f 'x 2 ... f 'x n 
 
 f 'x 1 f "x 1x 1 f "x1x 2 ... f "x 1x n 
H (f ) =  f 'x 2
B
f "x 1x 2 f "x 2 x 2 ... f "x 2 x n 

 ... ... ... ... ... 
 f 'x f "x 1x n f "x 2 x n ...

f "x n x n 
 n
and the bordered principal diagonal determinants:

103
ACTA UNIVERSITATIS DANUBIUS Vol 11, no 1, 2015

0 f 'x 1 f 'x 2 ... f 'x k

f 'x 1 f "x 1 x 1 f "x 1 x 2 ... f "x 1 x k
Bk = f 'x 2 f "x 1 x 2 f "x 2 x 2 ... f "x 2 x k , k= 1, n
... ... ... ... ...
f 'x k f "x 1 x k f "x 2 x k ... f "x k x k

we have the following theorems:

Theorem If the function f:D R n R, D – convex, fC2(D) is quasi-concave then
 1k Bk 0, k= 1, n .
Theorem In order that the function f:D R n R, D – convex, fC2(D) be quasi-
concave is sufficient that  1 Bk 0, k= 1, n .
k

A.2. The main indicators of production functions

Let a production function:
Q:DpR+, (x1,...,xn)Q(x1,...,xn)R+ (x1,...,xn)Dp
Q
We will call the marginal productivity relative to a production factor x i:  x i =
x i
representing the trend of variation of production at the variation of the factor x i. In
Q
particular, for a production function of the form: Q=Q(K,L) we have K= -
K
Q
called the marginal productivity of capital and L= - called the marginal
L
productivity of labor.
Q
We call the average productivity relative to a production factor x i: w x i =
xi
representing the value of production at the consumption of a unit of factor xi. In
Q
particular, for a production function of the form: Q=Q(K,L) we have: wK= -
K
Q
called the productivity of capital, and wL= - the productivity of labor.
L
Considering the factors i and j with ij, we define the restriction of production
area: Pij=(x1,...,xn)xk=ak=const, k= 1, n , ki,j, xi,xjDp relative to the two factors

104
 ŒCONOMICA

when the others have fixed values. Also, let: Dij=(xi,xj)(x1,...,xn)Pij - the
domain of production relative to factors i and j.
We define: Qij:DijR+ - the restriction of the production function to the factors i
and j, i.e.: Qij(xi,xj)=Q(a1,...,ai-1,xi,ai+1,...,aj-1,xj,aj+1,...,an). The functions Qij define a
surface in R3 for every pair of factors (i,j).
We will call partial marginal rate of technical substitution of the factors i and j,
relative to Dij (caeteris paribus), the opposite change in the amount of factor j to
substitute a variation of the quantity of factor i in the situation of conservation
production level.
dx j
We will note: RMS(i,j)=  and we have, since Qij(xi,xj)=Q0=constant:
dx i
xi
. Obviously RMS i, j 
D ij 1
RMS(i,j)= . We also define the global
x j D ij
RMS  j, i 
marginal rate of substitution between the i-th factor and the others:
x i
RMS(i)= . The global marginal rate of technical substitution is the
n
 2x j
j1
j i

minimum (in the meaning of norm) of changes in consumption of factors so that

the total production remain unchanged.
In particular, for a production function of the form: Q=Q(K,L) we have:
 
RMS(K,L)= RMS(K)= K , RMS(L,K)=RMS(L)= L .
L K
It is called elasticity of production in relation to a production factor x i:
Q
x x
 x i = i = i - the relative variation of production at the relative variation of
Q w xi
xi
factor xi.

105
ACTA UNIVERSITATIS DANUBIUS Vol 11, no 1, 2015

K
In particular, for a production function of the form: Q=Q(K,L) we have K= -
wK
L
called the elasticity of production in relation to the capital and L= - the
wL
elasticity factor of production in relation to the labor.
xi
Let note now for arbitrary factors xi, xj: ij= , i,j= 1, n , ij and we call the factor
xj
endowment ratio with the factor i relative to factor j.
It is called the elasticity of marginal rate of technical substitution for a production
RMS (i, j)
ij
function relative to inputs i and j: ij= , i,j= 1, n , ij and represents the
RMS (i, j)
ij
relative variation of marginal rate of technical substitution relative to factors i and j
at the relative variation of the factor endowment ratio with factor i relative to factor
j.
RMS (i, j)
xi
x i  ln RMS (i, j)
We have therefore: ij= = xi .
RMS (i, j) x i

Considering now a production function Q:DpR+, (x1,...,xn)Q(x1,...,xn)R+

(x1,...,xn)Dp, homogenous of degree r, let note for an arbitrary factor (for
x 
example xn): i= i , i= 1, n  1 . Of course: ij= i .
xn j
We obviously have:
x 
 = x rn Q1 ,...,  n 1 ,1
x x
Q(x1,...,xn)= x rn Q 1 ,..., n 1 , n
 n
x x n xn 

Considering the restriction of the production function at Dp R n1  1 :

q1 ,..., n 1  = Q1 ,..., n 1 ,1 we can write:

Q(x1,...,xn)= x rn q 1 ,...,  n 1 

With the new function introduced, the above indicators are:

106
 ŒCONOMICA

q
  x i = x rn1 , i= 1, n  1
 i
 n 1 q 
  x n = x rn1  rq    i 
 i 1  i 
q
 w x i = x rn1 , i= 1, n  1
i
 w x n = x rn1q
q

 RMS(i,j)= i , i,j= 1, n  1
q
 j
q
i
 RMS(i,n)= n 1 q
, i= 1, n  1
rq   i
i 1  i

q
 i
 RMS(i)= , i= 1, n  1
2 2
   
 rq   q  j     q 
n 1 n 1

 j1  j
  
j1   j 
  ji
n 1 q
rq   j
j1  j
 RMS(n)=
2
n 1 q 
   
j1  j 

q

  x i = i , i= 1, n  1
q
i
n 1 q
rq   i
i 1  i
 xn =
q

107
ACTA UNIVERSITATIS DANUBIUS Vol 11, no 1, 2015

 2q q q  2q

i2  j i i  j
 ij= i
q q
i  j
2
 2q  q  q n 1  2q  2q n 1 q
rq 2  1  r        k
i  i  i k 1  k  i i2 k 1  k
k

k i k i
 in= i
q  n 1 q 
 rq    k 
i  k 1  k 
, q  Q,1 :
K
For a production function of the form: Q=Q(K,L),  
L
q
 K = Lr 1

 q 
 L = Lr 1  rq   
  
q
 w K = Lr 1

 w L = Lr 1q
q

 RMS(K,L)=RMS(K)=
q
rq  

q

 K =
q

q
rq  

 L =
q

108
 ŒCONOMICA

2
 2q  q 
rq 2  1  r  
   
 =KL= 
q  q 
 rq   
   
A.3. Necessary and sufficient conditions for nonlinear optimization
Considering now the non-linear programming problem:
max(min) f ( x1 ,..., x n )

g i ( x1 ,..., x n )  0, i  1, p
x ,..., x  0
 1 n

where f,gC2(Dp) and a solution x = x1 ,..., x n  the Karush-Kuhn-Tucker

conditions occur: iR+, i= 1, p so that:

 p

  f x 1 ,..., x n     ig i x1 ,..., x n   0

 i 1

g i x1 ,..., x n   0, i  1, p
 g x ,..., x   0, i  1, p
 i i 1 n


 F F 
where F is the gradient of F defined by: F=  ,...,  and =1 for the case

 1x xn 
of maximizing and =-1 in the case of minimizing.
If f,gi, i= 1, p are of class C2, from [1] follows, for the maximizing case, the
sufficiency of Karush-Kuhn-Tucker conditions takes place in the broader
framework of quasi-concavity of functions f and g and, moreover, if for a solution
x = x1 ,..., x n  one of the conditions occurs:
f
 k= 1, n such that x  0;
x k
f
 k= 1, n such that x  0 and x k 0;
x k
 f 0;
 f is concave.

109
ACTA UNIVERSITATIS DANUBIUS Vol 11, no 1, 2015

min f ( x1 ,..., x n )

For the problem: g i ( x1 ,..., x n )  0, i  1, p
x ,..., x  0
 1 n

replacing f with –f and taking into account that

min f (x1 ,..., x n )  max  f (x1 ,..., x n )
= follows that Karush-Kuhn-Tucker
conditions becomes:
 p

   f x 1 ,..., x n     ig i x1 ,..., x n   0


i 1

g i x1 ,..., x n   0, i  1, p , iR+, i= 1, p

 g x ,..., x   0, i  1, p
 i i 1 n


and sufficiency reduces to one of the cases:

f
 k= 1, n such that x  0;
x k
f
 k= 1, n such that x  0 and x k 0;
x k
 f 0;
 f is convex.
In the particular case of the problem of minimizing the total cost (TC) relative to a
production function Q=Q(x1,...,xn) and pi, i= 1, n - the prices of inputs:
 n

min  p k x k
 k 1
Qx1 ,..., x n   Q 0  0 the Karush-Kuhn-Tucker conditions are:
x ,..., x  0
 1 n

 Q
 p k   x x1 ,..., x n   0, k  1, n
 k
Qx1 ,..., x n   Q 0 . Because =0 implies pk=0 – which is absurd
 Qx ,..., x   Q   0
 1 n 0


110
 ŒCONOMICA

 Q
 x ,..., x n   p k , k  1, n
in economic terms, results: 0 then:  x k 1 or, with
Qx1 ,..., x n   Q 0
 Q Q
 x x1 ,..., x n  x n
x1 ,..., x n 
another expression:  1
 ...  . Because the objective
 p1 pn
Qx1 ,..., x n   Q 0
n
function f x1 ,..., x n    p k x k is affine, Q is quasi-concave and, in addition
k 1
f
x  =pk0 follows, from the foregoing, that these conditions are sufficient.
x k
A.4. Production efficiency
Let us consider in the following two goods  ,  and a number of n inputs
F1,...,Fn available in quantities x1, , ,.x n , and the production functions of  or 
as follows:
Q  Q x1,..., x n  , Q  Q x1,..., x n 

appropriate to the consumption of xk units of factor Fk, k= 1, n . We will assume that

the production functions are of class C2 inside space production SP.
We will build the Edgeworth’s box consisting in a n-dimensional parallelepiped:
0, x1  ... 0, x n  the quantities of  being relative to O0,...,0 and those of 
relative to Fx1, , ,.x n  on the parallelepiped sides. Let consider an initial allocation
of inputs for  and  :
x   a1,..., a n  , x   b1,..., bn 

where ai+bi= x i , i= 1, n . The productions appropriate to the initial allocation are:

Q,0 a1 ,..., a n  , Q,0 b1,..., bn  relative to O and F, respectively. Because bi= x i -ai,
i= 1, n we have: Q,0 b1,..., bn  = Q x1  a1,..., x n  a n  . The production function
of  is therefore: Q̂   Q  x1  x1 ,..., x n  x n  and means the production of 
Q̂  Q  2Q̂   2Q 
relative to the origin of axes. We have now:   ,  ,
x i x i x i x j x i x j
i,j= 1, n therefore Q̂  is also quasi-concave but with negative partial derivatives of

111
ACTA UNIVERSITATIS DANUBIUS Vol 11, no 1, 2015

order 1. Considering the isoproduction hypersurfaces, it follows that (relative to O)

those of  is convex, while that of ̂ is concave.

Let PZ ,0  x1 ,..., x n   SP Q  x1 ,..., x n   Q ,0 - the production zone of 
superior to Q  ,0 and 
PZ ,0  x1 ,..., x n   SP Q  x1 ,..., x n   Q  ,0  - the
production zone of  superior to Q  ,0 .

Suppose now that int PZ,0  PZ ,0   (int means the interior of the set, i.e.
those points for which there is a n-dimensional cube centered in them sufficiently
small side and included in the given set).
Let now a point Cc1,..., cn   int PZ,0  PZ ,0  and also let the straight line that
passes through the origin and C. Let note Dd1,..., dn  the intersection with the
isoproduction hypersurface Q x1 ,..., x n   Q,0 and Ee1,..., en  the intersection
with the isoproduction hypersurface Q̂ x1 ,..., x n   Q̂ ,0 . We have now
Q d1 ,..., d n   Q,0 and Q̂ e1 ,..., en   Q̂,0 . Because  is convex we obtain
that: Q d1,..., dn   Q c1,..., cn  and Q̂  - concave implies that
Q̂  e1 ,..., e n   Q̂  c1 ,..., c n  or Q x1  e1,..., x n  en   Q x1  c1,..., x n  cn  .
After these inequalities follows that the production of each good can increase, so
the initial allocation is not optimal.
We call Pareto’s efficiency the situation where new production can not improve
without affecting the other’s production. From the foregoing, it follows that the
Pareto’s efficiency is obtained if the isoproduction hypersurfaces are tangent.
The condition of tangency for Q= Q x1,..., x n  and Q= Q̂  x1 ,..., x n  =
Q x1  x1,..., x n  x n  is reduced to the determination of those points x1 ,..., x n 
Q Q̂
where    , i= 1, n , R i.e. those points where hypersurfaces intersect
x i x i
and have the same tangent hyperplane (directors parameters are proportional).
Taking into account that Q̂  x1 ,..., x n  = Q x1  x1,..., x n  x n  we have that:
Q Q
   , i= 1, n , =-R.
x i x i

In marginal notation, we have: ,i x1 ,..., x n   ,i x1  x1 ,..., x n  x n  , i= 1, n ,

R.

112
 ŒCONOMICA

 , K  , K
For two inputs (K and L) the above relations are equivalent with:  .
 , L  , L
 , K dL
On the other hand:   = RMS K, L - marginal rate of technical
 , L dK
 , K dL
substitution of capital for  and   = RMS K, L - marginal rate of
 , L dK
technical substitution of capital for  . The upper equality becomes:
RMS K, L = RMS K, L .
All of the points where the allocation is Pareto’s efficient generates the production
contract curve.
Contract curve represents all combinations of goods for which no party can
maximize its production without diminishing the other’s production. On the other
hand, any point on the curve represents a possible allocation contracts. The
problem is this: if one good will be produced in order to reach the maximum level,
what will do the other?
n
Considering now the prices of n inputs as p1,...,pn the total cost is: TC=  p i x i and
i 1
it maximize the production if it is tangent to the isoproduction hypersurface. But
each good want to be produced in maximum quantity therefore:
,1 ( x1 ,..., x n )  ( x ,..., x n )
 ...  , n 1 ,
p1 pn
 ,1 (x1  x1 ,..., x n  x n )  (x  x ,..., x n  x n )
 ...   ,n 1 1
p1 pn
or, in other words, the cost hyperplane will be tangent to both isoproduction
hypersurfaces, that is it will coincide with the common tangent hyperplane.
Considering the production contract curve of the form:
x1=g1(),...,xn=gn(), R
follows:
,1 (g1 (),..., g n ()) ,n (g1 (),..., g n ())
 ... 
p1 pn
from where:

113
ACTA UNIVERSITATIS DANUBIUS Vol 11, no 1, 2015

 ,k (g1 (),..., g n ())

pk=  , 0, k= 1, n
 ,1 (g1 (),..., g n ())

We note that prices are determined up to a multiplicative factor, which does not
affect the result of the problem and can therefore consider =1. If the initial
allocation of factors of production was x   a1,..., a n  , x   b1,..., bn  the total
n
cost of production of  is TC  =  p k a k . The new amounts of factors (which
k 1
n
also satisfy the same total cost) involves:  pk (a k  x k )  0 . Replacing the values
k 1
of pk into this equation:
n  (g (),..., g ())
 , k (g 1(),..., g n ()) a k  g k ()  0
k 1  ,1 1 n

hence we will find R. Substituting in the appropriate expressions, will result pk
and xk, k= 1, n .

10. References
Arrow, K.J. & Enthoven, A.C. (1961). Quasi-Concave Programming. Econometrica, vol.29, no.4, pp.
779-800.
Boyd S. & Vandenberghe, L. (2009). Convex Optimization. Cambridge University Press.
Chiang ,A.C. (1984). Fundamental Methods of Mathematical Economics. McGraw-Hill Inc.
Harrison, M. & Waldron, P. (2011). Mathematics for Economics and Finance. Routledge.
Ioan, C.A. & Ioan, G. (2011). The Extreme of a Function Subject to Restraint Conditions. Acta
Universitatis Danubius. Oeconomica, vol.7, no. 3, pp. 203-207.
Ioan, C.A. & Ioan, G. (2012). On the General Theory of Production Functions. Acta Universitatis
Danubius, Oeconomica, vol. 8, no.5, pp. 223-236.
Luenberger, D.G. (1968). Quasi-Convex Programming. Siam J. Appl. Math., Vol.16, No.5, pp. 1090-
1095.
Pogany, P. (1999). An Overview of Quasiconcavity and its Applications in Economics. Office of
Economics, U.S. International Trade Commission.
Simon, C.P. & Blume, L.E. (2010). Mathematics for Economists. W.W.Norton & Company.
Stancu, S. (2006). Microeconomics. Bucharest: Economica.

114