R.N. Batra
R.N. Batra
R.N. Batra
INTERNATIONAL TRADE
STUDIES IN THE
PURE THEORY OF
INTERNATIONAL
TRADE
Raveendra N. Batra
Palgrave Macmillan
©Raveendra N. Batra 1973
Softcover reprint of the hardcover I st edition 1973 978-0-333-13367-5
All rights reserved. No part of this publication may be
reproduced or transmitted, in any form or by any means,
without permission.
PREFACE xi
I INTRODUCTION I
l.l The Production Function I
1.2 Technical Progress 6
1.3 The Transformation Curve II
t Oniki, H., and Uzawa, H., 'Patterns of Trade and Development in a Dynamic
Model of International Trade', Review of Economic Studies, xxxn (Jan 1965) 15-38.
Preface xiii
that is, if both inputs are changed in some proportion A., the
output also changes in the same proportion.
(iii) The first partial derivatives of F(K,L), denoted by FK and FL,
are positive, but the second derivatives, FKK and FLL• are
negative. This implies that all marginal productivities are
positive but diminishing:
F. _ oF(K,L) O F. _ oF(K,L) O
K- oK > ' L- oL > '
82 F(K,L) 82 F(K,L)
FKK = oK2 < 0, and FLL = oL2 < 0.
and
au- kh.) ok ok ok k2
FLL = oL = fk oL -kfkk oL -h. oL = L fkk·
The negative signs of FKK and FLL imply that Ak =
o.t,;/ok < 0,
which, in turn, implies that the marginal product of capital declines
Introduction 3
where IX 1 is the Lagrangian multiplier. The first-order conditions for the minimisation
of T for given w and r are then given by
aH aH
oK = r-IX,FK = oL = W-IX 1 FL = 0.
Note that this condition, associated with the minimisation of the unit cost, must be
valid so long as there is perfect competition in factor markets, irrespective of con-
ditions in the product markets.
4 Studies in the Pure Theory of International Trade
FL w
w = - = -. (1.5)
Fx r
Thus, so long as there is perfect competition in factor markets, the
ratio of the two marginal products equals the wage/rental ratio.
From (1.3) follows another well-known result: namely, the total
product is fully exhausted by the contribution by the two factors:
LFL + KFx = L(f- kfk) + Kfk = X.
There is yet another way in which the production function can be
expressed. With the marginal products of both factors determined
exclusively by the factor proportions, the entire information con-
cerning the production surface in the two-dimensional capital and
labour space can be summed up in the unit isoquant, representing
oneunitoftheoutpu t. Supposd = 1/X. Then(l.l)can be written as
1 = F(Cx,Cd
where Cx = K/X and CL = L/X are the input-output coefficients.
From this and (1.5) the definition of the elasticity of substitution
given in (1.4) becomes
(J
K.*-L*
= ------,--
(K.*-X*)-(L*-X* ) q-q
w*-r* w*-r* w*-r*
given prices:
and (1.7)
L
Figure 1.1
Substituting, we get
1
FL(OD+DB) = 1, or FL = OB'
Similarly, it can be shown that FK = 1/0A.
Now suppose that AB also represents the isocost line, so that the
slope of AB equals the wage/rental ratio. PointE then furnishes the
capital/labour ratio (k = DE/OD), the capital/output ratio (CK =
DE) and the labour/output ratio (CL = OD), which are uniquely
determined, provided the entrepreneur seeks to minimise his unit
cost. Thus for any given wjr ratio, and a given technology repre-
sented by the neo-classical production function, the equilibrium
values of k, CK and CLare uniquely determined. Furthermore, with
production taking place under conditions of perfect competition,
the commodity price from (1.7) can be expressed in terms of the
cost of capital or labour alone. That is, p = wjFL = r/FK, or, in
termsofFig.l.l,p = w.OB = r.OA.
Figure 1.2
Figure 1.3
10 Studies in the Pure Theory of International Trade
Figure 1.4
t With non-neutral improvements, it is, of course, possible that one of the input-
output coefficients may rise at the original wage/rental ratio. Figs. 1.3 and 1.4 can be
modified to portray this case.
Introduction 11
•'
t
>(
:! 0
·a .;a.
"
<.J ;
0
- - Output of X 2
Figure 1.5
commodity space. The locus of points such asP and Q is called the
transformation curve, which in Fig. 1.5 is given by 0 1 PQ02 • The
transformation curve is thus a mirror-image of the contract curve.
One may observe that the curve 0 1 PQ0 2 is wholly concave towards
the origin, 0.
What is the slope of the transformation curve in equilibrium?
To answer this question, we need information about the type of
markets in which production takes place. The transformation curve
is the locus of efficient points and its curvature depends on market
conditions in which the firms operate. The answer turns out to be
very simple and elegant, if we assume that all markets are perfect.
In the presence of market imperfections, of course, it is possible
that the transformation curve, as a locus of efficiency points, may
cease to exist, simply because production efficiency cannot be
attained under such imperfections. For the time being, however, we
gloss over this question, but we shall examine it in detail in the
chapters concerning market imperfections.
Under the assumption of perfect competition in all markets, the
reward of each factor in equilibrium equals the value of its marginal
product and is the same in both commodities. Let M Pu be the
marginal productivity of the ith factor in the jth commodity and Pi
be the commodity price (i = K,L;j = 1,2). Then
w = P1MPL1 = P2MPL2
and
x,
Figure 1.6
origin. Lines such as RS and M N are the budget lines which measure
the level of national income in the economy at different production
points. The slope of each budget line reflects the relative price of X 1 ,
ptfPJ. The equilibrium production point is obtained by drawing the
given budget line tangential to the transformation curve. Thus, with
RS, the production point is given by P, where the slope of TT' equals
the relative price of X 1 • The concavity of TT' towards the origin
implies that a rise in the output of a commodity calls forth a rise in
its relative price. This point is illustrated by the fact that the pro-
duction point associated with MN, showing a higher relative price
of X1 , is Q, which shows that the output of X 1 rises from Oa to Oa'
and that of X 2 declines from Ob' to Ob.
What is the reasoning behind the concavity of the transformation
curve, or, stated differently, what is the economic explanation behind
the phenomenon of greater output requiring a greater relative price,
16 Studies in the Pure Theory of International Trade
Figure 1.7
Introduction 17
REFERENCES
[I] Hicks, J. R., The Theory of Wages (London: Macmillan, 1932).
[2] Savosnick, K. M., 'The Box Diagram and the Production-Possibility Curve',
Ekonomisk Tidskrift, LX (Sep 1958) 183-97.
2 Structure of Production
in Autarky
2.1 Introduction
One of the elemental, though most important, questions in inter-
national trade concerns the determination of a country's pattern of
trade. Stated differently, why are countries induced to import goods
if they can produce them at home? The answer to these questions,
however, requires a prior comprehension of the structure of produc-
tion in a closed economy or autarky. It is this latter question with
which the present chapter is concerned. Specifically, we wish to deter-
mine the factors which influence commodity prices in the absence
of trade.
and
K 1 +K2 = K.
These equations can be transformed in a manner which, by means of
the concept of input-output coefficients, will determine the output
level of each commodity. Let Cii be the quantity of the ith factor used
in the production of one unit of the jth commodity. For example,
eLl = LtfXl, and CKl = KtfXl, so that Ll = CLlxl and Kl =
CK 1 X 1 , and so on. The full-employment equations may then be
written as
(2.3)
(2.4)
(2.6)
and
(2.7)
Equations (2.6) and (2. 7) state that the demand for each com-
modity depends on the commodity-price ratio and national income,
which in turn may be written as
(2.8)
that is, national income simply equals the total value of the two
outputs or the payments made to the factors of production. With
this equation, the specification of our two-commodity, two-factor
autarky model is complete. We have now added five equations in
four more variables, D 1 , D 2 , p and Y, but from Walras's Law only
one of the two equations, Di = Xi, is independent, so that the system
continues to be determinate. One may have noticed the differential
treatment accorded to p on the demand and the production sides
of the model. On the demand side,p is treated as a variable. However,
on the production side p is treated as a parameter, because the
competitive firm takes the commodity price as given.
The significance of the assumption of differing capital/labour
ratios in the two commodities for the unique determination of out-
puts may now be shown. For a given t = t, the economy's state of
technology is completely defined by the input-output coefficients,
which can be collectively expressed by a technology matrix [CJ,
that is,
Structure of Production in Autarky 21
(2.11)
where k = K/ L.
As far as the price equations (2.3) and (2.4) are concerned, the
matrix of the input-output coefficients associated with w and r is
defined by the rows of the C matrix, or simply by its transpose.
However, the sign of the matrix is still given by the sign of ICJ. The
solution of (2.3) and (2.4) gives us
(2.12)
where A.ii is the proportion of the ith factor employed in the jth
commodity and eii is the share of the ith factor in the total earnings
in the jth commodity (i = L,K; j = 1,2). For example, A.Ll =
(CL 1XdL) = LdL equals the proportion of labour employed in the
first commodity, whereas eLl = (wCLl/p 1) = wLifp 1X 1 equals the
Structure of Production in Autarky 23
relative share of labour in the total value of output of the first com-
modity. By very definition, A.il +A.i 2 = I and (}Li+(}Ki = I. Let [A.]
and [ (}] respectively denote the matrices of A. coefficients and (} co-
efficients in equations (2.13)-(2.16); that is,
[A_] =[ALl
AK!
AL2],
AK2
[(}] =[(}Ll
(}L2
(}K!J.
(}K2
The signs of the determinants lA. I and 1e1 are the same as the sign of
the determinant Iq, because
,
IA. I = A.LIAK2-A.K!A.L2 L 1L 2
= LK (k2-kl)
wrL 1 L 2
1e1 = eLleK2-eL2eK! = x x (k2-kl). (2.17)
P1 1P2 2
Furthermore, since each row in IA.I and 1e1 adds to unity, the deter-
minants lA. I and 181 are also given by
IA.I = A.Ll-A.Kl = A.K2-A.L2
I(}I = (}Ll- (}L2 = (}K2- (}Kl (2.17*)
Clearly, the next step in the solution of the system is to obtain
expressions for Cij. Equation (2.5) can be totally differentiated to
obtain
(2.18)
where Aij = (l/Cii)(oCufow) dw is the change in the input-output
coefficient that occurs as a result of a change in the wage/rental
ratio (the partial derivative notation in writing Aij signifies that
technology remains unchanged) and Bij = (- 1/Cii)(oCii/ot) dt is
a measure of technical progress that results in a change in Cii; the
partial derivative notation in Bij shows that the wage/rental ratio is
kept constant. Since technical improvements ordinarily involve a
reduction in input-output coefficients, Bij is defined to be non-
negative. Simple expressions for Aij can be derived by introducing
the definition of the elasticity of factor substitution and taking into
account the implications of the unit cost minimisation. Let ai be the
elasticity of factor substitution in the jth sector. As defined in the
previous chapter,
(2.19)
24 Studies in the Pure Theory of International Trade
The unit cost in the first commodity equals wCLl + rCKl· For
given factor prices in competitive factor markets, the unit cost in
the first commodity is minimised by setting the first derivative
(wCL 1A! 1 + rCK 1At 1) equal to zero.t Dividing through by Pt> we get
(2.21)
t Since at present we are trying to assess the effect of a change in won the input-
output coefficients as of a given technology, the notation used for dC,;/C,; is A;j
instead of C;j. which actually denotes the rate of total change in c,i.
Structure of Production in Autarky 25
and
llj = (}LjBfj + (}KjB:j (J = l ,2).
Before we proceed further with the analysis, a few explanatory
remarks concerning /3;, ll; and nj are in order. If production co-
efficients are fixed, crj = 0, in which case /3; also equals zero. In the
general case of variable coefficients, however, /3; reflects the percent-
age change in the use of the ith factor per unit of output that occurs
in both commodities as a result of a change in the wage/rental ratio
alone. By contrast, ll; represents the percentage reduction in the ith
factor that occurs owing to the occurrence of technical progress in
both commodities when the wage/rental ratio is kept constant; for
example,
av = P
(DdD2)
• o(DdD2) =
op
p[-1 oD 1_. oD 2]
1_
D1 op D2 op
= (XI +tx2 > 0.
(2.31)
t This elasticity was first introduced by Jones [2]. Along any transformation
curve, factor supplies and the level of technology are given, so that K* = L * = n L =
ll 1 = ll 2 = 0. Then, from equation (2.32),
(X!-X!)
= (15.
P!-p!
In the absence of market distortions, u,, which also equals [(PL + PK)!(jA.JJOJ)], is
positive, because PL > 0, PK > 0 and JA.J and JOJ always possess the same sign.
28 Studies in the Pure Theory of International Trade
Figure 2.1
n
R
\ '•-----------+--:.......o<------,r
Figure 2.2
D
R
p 0 k
Figure 2.3
attributable to the fact that the unit isoquants intersect only once.
A similar diagram can be drawn for the case where k 2 < k 1 .
The 'one-to-one' relationship between w and p on the one hand,
and between w and ki on the other, is depicted in Figs. 2.2 and 2.3.
30 Studies in the Pure Theory of International Trade
Factor-Intensity Reversals
Until now we have assumed that factor intensities are non-reversible.
The analysis becomes more complicated and vexing if factor inten-
sities are reversible. Symbolically, the problem is one of confronting
the multiple signs of I01 at different levels of w. For if k 2 ~ k 1 at
different w's, 101 ~ 0. Evidently, then, the relationship betweenp and
w will no longer be unique. Geometrically, the problem is illustrated
in Fig. 2.4. The only difference between Figs. 2.4 and 2.1 lies in the
presence of two points of intersection between the two unit isoquants
in Fig. 2.4, as against the single intersection point in Fig. 2.1. As
before, we commence with the isocost line AB and a unit commodity-
price ratio (p). As the wage/rental ratio declines, we find that X 2
becomes labour-intensive relative to X1 because Ob 2 is less steep
than Oa 2 , although in the initial situation X 2 was capital-intensive
Structure of Production in Autarky 31
K
L
Figure 2.4
p
Figure 2.5
0 p
Figure 2.6
Figure 2.7
be even greater if the demand change due to the income change was
biased in favour of the second commodity, that is if '7 2 > 'It. The
same reasoning applies to the case of a decline in k.
The relationship between k and p continues to be unique if lA. I > 0,
that is, k 2 > k t, and '7 2 < 'It· Here a rise in k will result in a larger
decline in p than before, and conversely.
The picture becomes blurred if the terms on the right-hand side
of (2.33*) possess opposite signs. This will be the case if (i) lA. I < 0
but '7 2 < 'It• or (ii) IA.I > 0 but '7 2 > 'It· To say the least, there-
lationship between k and p is now weakened. Furthermore, the
previous relationship between k and p may be reversed; that is to
say, if k and p were positively related when '7t = 17 2 , they may be
negatively related when the influence of differing income elasticities
of demand is taken into account, and conversely.
(2.34)
r*-pt =
-fh·<Pt-Pn
J IOI u = 1,2). (2.36)
n - AuK*-A.KlL*
(2.38)
2 - 121
With a rise in the supply of capital alone, K* > 0 but L* = 0. The
proof of the Rybczynski theorem then follows directly by observing
that
Xi§ 0,
A simple diagrammetic derivation of the Rybczynski theorem can
be accomplished by utilising the box diagram presented in the previ-
ous chapter. Consider Fig. 2.8, where the factor supplies are
measured along the axes merging in the origins 0 1 and 0 2 and the
outputs by the distance of the production point from the respective
origins. Suppose A is the initial production point lying on the con-
tract curve 0 1 A0 2 , so that the capital/labour ratio in each industry
is given by the slope of 0 1 A and 0 2 A, reflecting that the first com-
modity is labour-intensive relative to the second commodity, and
38 Studies in the Pure Theory of International Trade
Labour
r----------------------------,r--------,0~
I
I
I
AI/
,-;~
I
/1 ><"
0
'SQ.
...
:I
0
0 o"
------Output of x2
Figure 2.8
r*
PT
,_w
eu (2.36*)
X! ;.L2
L* l+g (2.37*)
X! Au
L* '-m· (2.38*)
Equations (2.35*) and (2.36*) are derived under the condition that
only the price of the first commodity changes, e.g. P! = 0, whereas
the condition behind (2.37*) and (2.38*) is that K* = 0. It may
now be evident that the link provided by I01 between commodity
prices and factor rewards is symmetrical to the link provided by IJ.I
between factor endowments and the individual outputs. Further-
more, each equation serves to show the magnification effect. For
example, suppose that k 2 > k 1 so that both IA.I and 101 are positive.
Then a rise in the price of the labour-intensive commodity (implying
that PT > 0) stimulates, at constant k, a more than proportionate
rise in the reward of labour, whereas a rise in the supply of labour
alone at constant p results in a more than proportionate increase in
the output of the labour-intensive commodity (see equations (2.35*)
and (2.37*)). On the other hand, ifk 2 < k 1 , i.e. ifiJ.I and 101 are
negative, the magnification and the dual effects are reflected in
equations (2.36*) and (2.38*). t
t Any remaining doubts concerning the dual relationship may be dispelled by
observing that, if P! = K* = 0, then the substitution of (2.17) in (2.35) and (2.37)
yields
dw dX 1
dp 1 dL
On the other hand, if p! = L* = 0, then
dr dX 2
dp, dK CL,(k,-k!)
40 Studies in the Pure Theory of International Trade
x,
Figure 2.9
u,
0
Figure 2.10
output bundles along any one of these curves as the latter are not
uniquely related to a single distribution of utilities among indivi-
duals. However, if individual utility functions are homothetic and
if conditions (I) and (2) are simultaneously satisfied, then the com-
munity indifference curves obtained will constitute a genuine
counterpart of each individual's indifference curves both in the
behavioural and in the welfare sense.
2.10 Summary
The purpose of this chapter was to analyse factors that affect the
commodity-price ratio in a closed economy. We studied how the
commodity-price ratio changes under the thrust of changes in factor
prices, factor endowments and levels of technology in the two
industries. Using a two-commodity, two-factor model with perfect
competition, neo-classical production functions, full employment
and non-reversible factor intensities, etc., the following results were
derived:
I. At constant factor endowment and technology, there exists a
unique relationship between the commodity-price ratio and the
factor-price ratio. This relationship is, however, no longer
unique if factor intensities are reversible. Furthermore, a rise in
the price of a commodity leads to a more than proportionate
rise in the reward of its intensive factor and a decline in the
reward of its unintensive factor, and conversely. The latter
result is usually called the Stolper-Samuelson theorem.
2. Under unchanged technology and homothetic preferences,
there exists a unique relationship between the overall capital-
labour ratio (k) and the commodity-price ratio (p). However,
when income effects are introduced, that is, when community
preferences are non-homothetic, or when factor intensities are
reversible, the relationship between k and p may no longer be
monotonic, in which case there arises a strong possibility of
multiple equilibria.
3. At constant technology and commodity prices, a rise in the
supply of a factor results in a more than proportionate rise in
the output of the commodity utilising the expanded factor in-
tensively at the expense of the output of the other commodity,
and vice-versa. This is the so-called Rybczynski theorem.
4. For any k and homothetic preferences, there exists a monotonic
Structure of Production in Autarcky 45
REFERENCES
[1] Chipman, J. S., 'A Survey of the Theory of International Trade: Part 2, the Neo-
Classical Economy', Econometrica, XXXIII (Oct 1965) 685-760.
[2] Jones, R. W., 'The Structure of Simple General Equilibrium Models', Journal
of Political Economy, LXXIII (Dec 1965) 557-72.
[3JRybczynski, T. M. 'Factor Endowment and Relative Commodity Prices',
Economica, XXII (Nov 1955) 336-41.
[ 4] Scitovsky, T., 'A Reconsideration of the Theory of Tariffs', Review of Economic
Studies, IX (summer 1942) 89-110.
[5] Stolper, W. F., and Samuelson, P. A., 'Protection and Real Wages', Review of
Economic Studies, IX (Nov 1941) 58-73.
3 The Basis of International
Trade
3.1 Introduction
In the last chapter we studied how changes in the commodity-price
ratio occur under the thrust of variations in the factor-price ratio,
the overall capital/labour ratio and the relative rate of technical
advance in the two industries. Any theory attempting to explain
the basis of international trade must always commence with the
theory of resource allocation and production in a closed economy.
This task having been accomplished in the previous chapters, we
are now in a position to pinpoint the factors that determine a
country's pattern of trade. The issue is what goods a country will
export and import. Stated differently, is it possible to predict a
country's configuration of exports and imports just by examining
the characteristics of a closed economy? Seeking a clear-cut answer
to this query constitutes the subject-matter of this chapter.
The traditional answer to the question of why a particular country
exports a particular commodity is simply that, owing to the handi-
work of nature or man, it is able to produce that commodity at a
lower comparative cost than the rest of the world, usually taken to
represent a single entity in the community of nations. International
trade occurs because countries stand to benefit by the exchange of
goods that are produced at dissimilar relative costs and hence at
different relative prices. It is desirable, therefore, to go deeper and
unearth the causes responsible for inter-country disparities in the
comparative cost or the commodity-price ratios. The reasons,
quite ostensibly, must be sought in the international differences in
factor endowments, tastes or levels of technology, because, as
established in the preceding chapter, these are the forces which
influence the commodity-price ratio in autarky. However, before
we get involved in a detailed examination of the theories identifying
themselves with these forces, it is imperative that the necessary
conditions preventing the occurrence of international trade be
expressly spelt out.
The Basis of International Trade 47
3.2 Conditions for the Absence of Trade
If countries were identical in all respects in the absence of the
opportunity to trade, the commodity-price ratios would be similar
everywhere and there would be no incentive to trade even if such an
opportunity was available. Obviously, this situation requires the
fulfilment of the following conditions:
I. Production functions are similar internationally.
2. Factor endowments are the same internationally.
3. Tastes are similar in all countries.
4. Production functions are neo-classical.
5. All markets are universally characterised by perfect
competition.
This set of conditions is sufficient to generate inter-country
identity on both the demand and the supply side, resulting thereby
in similar autarky prices and, as a consequence, precluding any
possibility of mutually profitable trade. Clearly, then, the relaxation
of any of these five requirements could provide a basis for inter-
national trade. In other words, there are five different ways in which
an 'international economy' could be simulated by the trade
theorist. However, from the viewpoint of propositions concerning
the pattern of trade, the two most widely known simulations are
those of Ricardo [11] and Heckscher-Ohlin [3, I 0], who,
respectively, formulated their theorems in terms of relaxing con-
ditions (I) and (2). Needless to say, if one of these conditions is
relaxed in order to forecast the pattern of trade, the remaining
four assumptions must be retained if the trade pattern is to follow
the prediction under all circumstances.
Figure 3.1
50 Studies in the Pure Theory of International Trade
x,
Figure 3.2
52 Studies in the Pure Theory of International Trade
Figure 3.3
54 Studies in the Pure Theory of International Trade
Figure 3.4
t It was not necessary to make this assumption in the simple Ricardian formula-
tion because factor supply and demand conditions played no role in the determina-
tion of the autarky price ratio.
58 Studies in the Pure Theory of International Trade
not be pursued here, but its significance will become clear later when
we examine the validity of the Heckscher-Ohlin theorem.
R'
R
R'
R
p 0 p
Figure 3.5
relationship between p and w in terms of the R' R' curve under the
condition that k 2 > k 1 at all w, whereas the second quadrant
depicts the positive relationship between p and w in terms of the
RR curve drawn for the case where k 2 < k 1 at all w. Owing to the
international similarity of production functions, RR and R' R'
curves describe the nature of the relationship between p and w in
both countries. Owh exceeds Ow1 to reflect the inter-country factor-
endowment relationship expressed in (3.7), so that if
k 2 < k1, then Ph > p1 (3.9)
which implies that the home country will export the capital-
intensive commodity X1 and import the labour-intensive commodity
x2 from the foreign country, and if
(3.10)
and the home country will export x2 and import xl, but x2 is now
capital-intensive relative to X1 . Thus, whatever the factor-intensity
relationship between the two commodities, the relatively capital-
abundant country exports the relatively capital-intensive commodity
and imports the relatively labour-intensive commodity from the
60 Studies in the Pure Theory of International Trade
0 ph pf ph p
Figure 3.6
Figure 3.7
s
S'
kh
kf
s
S'
p ph Pf 0 P'h pi p
f
Figure 3.8
x,
Figure 3.9
Figure 3.10
factor. Going back to Fig. 3.9, let us suppose that the home con-
sumption pattern is strongly biased towards the first good. Then at
the price ratio reflected by DE (and AB), the home consumption
point would lie on DE to the left of C; but so long as this consump-
tion point lies between P and C, it is easy to see that the point Sh
would lie on HH' to the right of P, so that the H.O. theorem would
continue to hold. However, if the home consumption point at the
commodity-price ratio reflected by DE lies to the left of P, the point
Sh would lie on HH' to the left of P, and the relationship between
the autarky price ratios would be reversed, so that the H.O. theorem
would not hold. Thus, in the absence of international similarity of
consumption patterns, the H.O. theorem may not be valid even if
all other requirements are satisfied. As argued before, for the inter-
country similarity of consumption patterns it is not sufficient that
the community preferences be identical; it is also necessary that in
each country they be homothetic. The significance of this de-
sideratum in the case where national incomes differ has been
demonstrated by Romney Robinson [12]. The tenor of his
argument is quite simple. If the community preference map of each
country is identical but not homothetic, the income elasticities of
demand for the goods differ, but the countries will still consume the
two goods in the same proportion at a given commodity-price
ratio if the national incomes are the same everywhere. Otherwise
the proportion in which the two goods are consumed by each
country at any price ratio will differ even if their community
indifference curves are non-intersecting; and this, as we have seen
above, could seriously compromise the validity of the Heckscher-
Ohlin theorem.
3.5 A Comparison
In the last two sections we have examined in detail the Ricardian
and the Heckscher-Ohlin theories explaining a country's pattern of
trade. This places us in a strong position to combine and collate
the various issues involved in the establishment of the two
theorems.
Perhaps the only similarity between the two theories consists in
the obvious fact that both tend to explain the basis of trade in terms
of international disparities in comparative costs. Otherwise the
two theories are poles apart, because of the following points of
difference :
68 Studies in the Pure Theory of International Trade
Figure 3.11
so thatt
Stability of Equilibrium
How is one to ensure that the international trade equilibrium
schematised in Fig. 3.12 is stable? For example, if the home offer
curve was given by OH' instead of OH, there would be three points
of intersection and hence three possible equilibria, E, E' and £".
However, it can be shown that equilibrium at E is unstable if the
offer curves are given by 0 F and 0 H', but stable if the home offer
curve is given by 0 H. The demonstration consists in differentiating
the balance-of-payments (B) equation
B=E1 -pE2
t Since each point on the offer curve represents by definition a certain level of
export supply or import demand, at every point p 2 E2 = p 1E 1 • However, the latter
equality should not be misconstrued as the equilibrium condition at this point. It
merely suggests what a country is willing to import in return for some exports at
certain terms of trade. The equality becomes an equilibrium condition only when
one country's import demand is exactly matched by the other country's willingness
to meet that demand at certain terms of trade.
72 Studies in the Pure Theory of International Trade
H
E,
Figure 3.12
-dB = E2 [ -p
dp El
·dE
-1- -p -dE2
dp E2 dp
-1 J= E2 (a 1 +ah-1)
where E2 > 0, and a1 and ah are respectively the foreign and home
elasticities of demand for imports. Evidently, stability is ensured if
a1 + ah > 1, that is, if the sum of the two elasticities of import
demand exceeds unity.t Reverting to Fig. 3.12, it is now easy to see
why E is. unstable and E' and E" stable when the two offer curves
are given by OF and OH'. For example, with E', if the terms of
trade shift in favour of the home country from OE' to OZ, its
balance of payments runs into a deficit of D 1D 2 in terms of the
imported commodity, which in turn means that dB/dp > 0. How-
ever, with E, a similar favourable shift in the home country's terms
of trade to 0 W brings about a balance-of-payments surplus of
D 1 D 3 , so that dB/dp < 0. Here, then, the equilibrium is unstable.
The stability of E" can be established analogously. This analysis
simply echoes Marshall's view [8] that an unstable equilibrium
point must be flanked by points of stable equilibria, and that the
number of equilibrium points must be odd. It follows then that a
unique (single intersection) equilibrium must be stable.
It is possible to derive the stability condition in terms of only one
commodity and in the process obtain more information. Since the
export of a commodity by one country must in equilibrium equal
the import demand of the other country,
E1 =and E 2 = E 1f
Elf
where Elf and E1f are respectively the quantities imported and
exported by the foreign country. The balance of payments in terms
of one commodity can be written as
B =E 1 -Elf= (X1 -D 1 )-(Dlf-Xlf)
= (X1 +X1f)-(D 1 +Dlf) = x 1 -d1
where x 1 and d 1 are respectively the world supply and demand for
the first commodity. Any discrepancy between x 1 and d 1 causes
disequilibrium and sets into motion changes in the terms of trade.
We are interested in finding conditions which cure this situation of
disequilibrium. Stability in terms of the world market of the first
commodity requires that dB/dp < 0, that is, a rise in the world
relative price of the first commodity causes an excess supply, or a
rise in B, and conversely. Differentiating B and remembering that
with initial trade equilibrium x 1 = d 1 , we get
dB x1
-d = -(bl +c~)-
'P p
where b 1 = -(p/x 1 )(dxddp) and c 1 = (p/d 1 )(ddddp) are respect-
ively the elasticities of world supply and demand for the first
commodity. Similarly,
B = (D 2 +D 2f)-(X2 +X2f) = -(x 2 -d2 )
74 Studies in the Pure Theory of International Trade
and
dB x2
dp -(b 2 +c 2 ) -
p
where b 2 = (p/x 2 )(dx 2 /dp) and c 2 = -(p/d2 )(dd2 /dp). Stability
conditions expressed in terms of the second commodity again
require that dB/dp < 0, that is, a fall in the world relative price of
the second commodity must lead to a situation of excess demand,
and vice versa. It may be observed that, in the absence of decreasing
costs and Giffen goods, dB/dp < 0 in terms of both markets.
However, if either the opportunity costs are decreasing or the goods
are of the Giffen variety, the equilibrium may be unstable. It is
interesting to observe that, in the presence of positively sloped
supply curves, the necessary condition for instability is that the
commodities be Giffen goods. Unless otherwise specified, the
equilibrium point will be assumed to be stable.
Figure 3.13
The Basis of International Trade 75
Factor-Price Equalisation
One striking property of the Heckscher-Ohlin model is that under
certain conditions factor prices in the free trade equilibrium get
completely equalised in the trading countries. Despite undertones
of unrealism, the so-called factor-price equalisation theorem has
held remarkable fascination with trade theorists, if only because
the theorem is in fact valid under some admittedly stringent con-
ditions. Although the controversy over the theorem has raged for
more than twenty years, the underlying logic of the theorem is
very simple.
We have already shown that in a two-commodity, two-factor,
constant returns to scale model, there is a unique relationship
between the commodity-price ratio and the wage/rental ratio,
provided factor intensities are non-reversible and both commodities
are produced. Now under free trade and the absence of transport
costs there prevails only one commodity-price ratio in the trading
countries, so that if production functions are internationally
similar, there will be only one wage/rental ratio (ro) prevailing in
both countries, which in turn implies that each commodity will
utilise the same capital/labour ratio everywhere. It follows then
that the marginal product and hence real reward of each factor
must under free trade be the same in both countries.
It is thus seen that the underlying logic of the factor-price
equilisation theorem is very simple. Yet the long controversy that
this theorem has sparked belies this impression. There are two
issues here, which, in view of the earlier debate, deserve examination
in detail. What are the roles played by (i) incomplete specialisation
and (ii) the non-reversibility of factor intensities in the attainment
76 Studies in the Pure Theory of International Trade
K
Figure 3.14
Figure 3.15
t For the sake of illustration, suppose that the isocost line is tangential only to one
unit isoquant and that the other unit isoquant lies above it. Evidently, then, only
one commodity will be produced, because the unit cost of the other commodity
exceeds its price.
t This was established in section 2.2 of the previous chapter.
78 Studies in the Pure Theory of International Trade
t Suppose the endowment ray of the home country coincides with OC and that
of the foreign country coincides with ODin Figs. 3.14 or 3.15. Then both countries
will be completely specialised but factor-price equalisation will still prevail, for the
wage/rental ratio in each country will be given by the slope of AB.
The Basis of International Trade 79
above applies regardless of the number of commodities and factors.
Suppose that there are r factors and n commodities, something of
which is produced in the world, though not necessarily in each
country. Given a vector of n world equilibrium commodity prices,
what are the conditions which will ensure that each of r factors of
production receives the same reward in each country?
As before, given the price vector, commodity units can be so
chosen that all price ratios will be equal to unity. Corresponding
to these unit product-price ratios, we can define n unit isoproduct
surfaces, which may or may not possess a common tangent isocost
plane. If a common tangent plane does not exist, factor-price
equalisation, as before, is impossible. If such a plane exists,
equalisation is possible but not necessary. For' factor prices to be
equalised internationally, it is necessary that the endowment ray of
each country lie within the cone of diversification specified by n
points of tangency. For otherwise the relative factor prices
associated with the common tangent plane may not be consistent
with the full employment of all factors.
What about the uniqueness of the common tangent plane? In
the two-by-two case, this uniqueness implies the non-reversal of
factor intensities. In the n-by-r case, where it may be extremely
difficult, if not impossible, to define relative factor intensities, one
of the necessary conditions for uniqueness is that n = r. If n =I= r,
uniqueness is impossible, for then the number of equations does not
match the number of unknowns. However, we have shown in the
two-by-two case that uniqueness is not a necessary condition for
the international equalisation of factor rewards. This is also true
with the n-by-r case. All that is necessary and sufficient to ensure the
equalisation of factor prices in the two countries is that each
country's endowment ray be flanked by the cone of diversification
defined by the points of tangency between the common tangent isocost
plane and the unit isoproduct surfaces. This rule applies to all cases
defined by n ~ r.
REFERENCES
[1] Bhagwati, J., 'The Proofs of the Theorems on Comparative Advantage',
Economic Journal, LXXVII (Mar 1967) 75-83.
[2] Chipman, J. S., 'A Survey of the Theory of International Trade: Part 3, The
Modem Theory', Econometrica, XXXIV (Jan 1966) 18-76.
80 Studies in the Pure Theory of International Trade
[3] Heckscher, E. F., 'The Effect of Foreign Trade on the Distribution of Income',
in Readings in the Theory of International Trade, ed. H. S. Ellis and L.A. Metzler
(Philadelphia: Blakiston, 1949).
[4] lnada, K., 'A Note on the Heckscher-Ohlin Theorem', Economic Record, XLIII
(Mar 1967) 88-96.
[5] Johnson H. G., 'Factor Endowments, International Trade and Factor Prices',
Manchester School of Economic and Social Studies, xxv (Sep 1957) 270-83.
[ 6] Kemp, M. C., The Pure Theory of International Trade (Englewood Cliffs, N.J. :
Prentice-Hall, 1964).
[7] Leontief, W. W., 'Domestic Production and Foreign Trade: The American
Capital Position Re-examined', Proceedings of the American Philosophical
Society, XCVII (Sep 1953) 332-49.
[8] Marshall, A., The Pure Theory of Foreign Trade (London: London School of
Economics and Political Science, 1949).
[9] Mill, J. S., Principles of Political Economy with Some of their Applications to
Social Philosophy, 3rd ed. (London: Parker & Co., 1852).
[ 10] Ohlin, B., Interregional and International Trade (Cambridge, Mass.: Harvard
U.P., 1933).
[II] Ricardo, D., On the Principles of Political Economy and Taxation (London: John
Murray, 1817).
[12] Robinson, R., 'Factor Proportions and Comparative Advantage', Quarterly
Journal of Economics, LXX (May 1956) 169-92.
4 Gains from Trade
recommendable if, in the new situation, gainers can more than com-
pensate the losers, so that everyone can be better off than before.
The question of whether the compensation should actually take
place is an ethical one for which the economist has no pronounce-
ment. The important point is that some 'desired' type of income
distribution can always be maintained if the state is ready to adopt
a policy of lump-sum transfers. In this spirit we assume throughout
our analysis of the gains from trade the existence of a community
welfare function which is a function of the quantities of various
products consumed by the community and which possesses proper-
ties that are parallel to individual utility functions. t Under this
approach the desired level of income distribution becomes inherent
in the choice of the utility function. Most of the assumptions made
in Chapter 2 are retained, except that, in the interest of generality,
we assume any arbitrary number of commodities and factors of
production. Thus production functions are still assumed to be
homogeneous of the first degree, all goods are non-inferior and fac-
tors are fully employed and inelastically supplied. The supply curves
for all commodities are upward-sloping; all markets are perfect and
factors of production are fully mobile internally but immobile be-
tween countries ;t furthermore, unless otherwise specified, the
country under analysis is assumed to be a small country. This last
assumption implies that international prices are unaffected by the
policies followed by the country under question, simply because the
trade volume of this country constitutes a tiny fraction of the total
volume of world trade in any commodity. The foreign offer curve
in the two-good setting then becomes a ray from the origin, so that
whatever the position of the home offer curve, the terms of trade are
unaltered.
Under this setting the economy can be described by the following
equations:
(4.1)
t The exact conditions under which the community utility function enjoys the
properties of the individual utility functions have been established by Samuelson [II].
t This assumption is not necessary and will be relaxed in the chapter on inter-
national investment. Furthermore, the assumption of inelastic factor supplies is
made only for simplifying the exposition.
Gains from Trade 83
(4.3)
(4.4)
(4.5)
(4.6)
(4.7)
(4.8)
and
Xt+PX2+p3X3+ .. ·PnXn = Dt+PD2+p3D3+ .. ·PnDn (4.9)
t For the sake of expository convenience only three factors are explicitly shown in
the production function. However, the number of factors, as we shall see later, is
immaterial.
84 Studies in the Pure Theory of International Trade
t For simplicity, we derive this expression for the three-good, three-factor case;
the extension to arbitrary numbers of goods and factors, not necessarily equal, follows
the same procedure.
Differentiating (4.2) totally, we have
dX1 = a1dK1+b 1dL 1+e1dN1•
Similarly, from (4.6)-(4.8),
dK 1 +dK2 +dK 3 = dL 1 +dL 2 +dL 3 = dN 1 +dN2+dN3 =0
because K, L and N are all inelastically supplied. Taking this and (4.3)-(4.5) into
account, the expression for dX1 can be written as
dX 1 = a 1 dK 1 +b 1 dL 1 +e 1dN 1 = -p[aidK 2 +dK 3)+b 2(dL 2+dL 3)
+eidN2 +dN3)]
-p[dx2 +; dx3J.
(4.12)
t Note that this situation does not imply monopoly or imperfect competition at
home.
86 Studies in the Pure Theory of International Trade
x,
Figure 4.1
Figure4.2
DP 1
Figure 4.3
(4.16)
t This is a more general version of Kemp's theorem and is free from any trade-
restricting connotations.
Gains from Trade 91
l dU dD 2
-·- = ep-- (4.18)
U1 de de
and
u; .----;};
dU
=
dX2
-sp~. (4.19)
t It is worth noting here that in the discussion of the consumption tax or the pro-
duction subsidy, it is not necessary to specify whether a commodity is exported or
imported. Thus whether D 2 in (4.18) and X 2 in (4.19) are respectively the consump-
tion and the production of the importable or the exportable good is immaterial to the
theorem in question. The theorem permits easy extension to other forms of inter-
vention in free trade, including the consumption tax or subsidy and the production
tax or subsidy on the exportable good.
92 Studies in the Pure Theory of International Trade
FP'
Figure 4.4
Figure 4.5
Gains from Trade 93
and (4.10) by
94 Studies in the Pure Theory of International Trade
I dU dE2 dE3
--=tp-+t3P3- (4.22)
ul dt dt dt
H'
Figure 4.6
Figure 4.7
s = t, ds = dt, and
that is to say, we compare those rates of the tariff and the production
subsidy which result in an equal rise in the output of the second
commodity.
Corden's theorem can be demonstrated simply by subtracting
(4.19) from (4.17), so that
dD 2
G = tpdt < 0
in the absence of inferior goods, for then a rise in the tariff rate would
lead to a decline in the local demand for the importable good.
The two-commodity case permits a simple geometrical demon-
stration of this proposition. In Fig. 4.8, P is the initial production
point and P' the production point as a result of the introduction of
the tariff which results in a domestic-price ratio indicated by the
slope of DP. An equivalent production subsidy which also shifts the
production point from P to P' furnishes a welfare level of u•. How-
ever, with the tariff, the welfare level is given by U,, which lies below
u•. Hence the tariff is shown to be inferior to the equivalent produc-
tion subsidy. The reason is not far to seek. The introduction of a
98 Studies in the Pure Theory of International Trade
FP
x1
Figure 4.8
FP
x1
Figure 4.9
Figure 4.10
Gains from Trade 101
u:I dU
dp
[ dp 3 ]
= - £2 + £3 dp .
Now if all terms of trade alter in the same direction, there is no
problem. For example, if dp 3 and dp have the same sign, dU/dp is
unambiguously negative. The trouble arises when dp 3 and dp have
different signs, so that the sign of dU/dp is not predictable in the
absence of information concerning the extent of the shifts in each
terms of trade and the initial volumes of the two imports. It is worth
pointing out, however, that the sign of dU/dp may be indeterminate
not because of the presence of more than two traded goods, but
because the two terms of trade may not change in the same direction
and it may then be impossible to speak of an improvement or a
deterioration in the terms of trade.
so that
dD 1 dD 2 d£ dp
-+p(l+t)- = t p -2- E2 - .
dt dt dt dt
I dU dD 1 dD 2 d£2 dp
- - = -+p(l+t)- = tp--£2-
U1 dr dr dr dr dr
or
(4.24)
t Assuming, of course, that the relative price at which B offers its goods is lower'
than the tariff-inclusive price of the goods which the home country imports in the
absence of the customs union from country F.
104 Studies in the Pure Theory of International Trade
where E1 = (t/ E 2 )(dE2 /dt) is the total home elasticity of demand for
imports with respect to the tariff,t and AP = (Ijp)(dpjdt) is the pro-
portionate change in the terms of trade as the home country switches
its trade from country F to country B.t The home country's agree-
ment to form a customs union implies that dt < 0. The signs of E1
and AP, however, depend on the characteristics of the union. In
general, dp is negative with the trade-creating union, but positive
with the trade-diverting union. However, in our example, where the
small home country initially trades with country F, the trade
creation can be represented by a simple removal of the tariff, with
no accompanying change in the terms of trade. This suggests that
with trade creation dt < 0 and AP = 0, but with trade diversion
dt < 0 and AP < 0. Like the sign of AP, the sign of E1 is also deter-
mined by the nature of the customs union. Under trade creation,
where dp :::; 0, the repeal of the tariff must lead to a rise in the
demand for the importable good. In other words, d£2 /dt and hence
E1 are negative.§ Now for the customs union to be beneficial to
welfare, dU > 0. It is at once clear from (4.24) that with AP = 0 and
E1 < 0, the trade-creating union leads to a rise in social welfare.
The sign of £ 1 is not so apparent when the union is of the trade-
diverting type. Since AP < 0 under trade diversion, (4.24) suggests
that the necessary condition for dU > 0 is that £ 1 < 0. Lipsey
postulated a Ricardian economy where any amount of profitable
trade results in complete specialisation in the exportable good.
Under Lipsey's setting, then, dX2 /dt = 0, so that dE2 jdt reduces to
dD 2 /dt. Thus, if £ 1 is to be negative, dD 2 jdt must be negative. This
t £, is the total home elasticity of demand for imports because it represents the
effects on the home demand for imports of a simultaneous change in both 1 and p
that occurs as a result of the formation of customs union.
t The reader may be reminded here that the change in the terms of trade resulting
from the formation of the trade-diverting union is not because the home country
comes to possess natural monopoly power in trade and ceases to be a small country,
but because of the exogenous and once-for-all shift in the terms of trade consequent
upon the home country's decision to form the union with Band thus accept the given
terms at which B will offer its goods.
§This can be seen immediately by differentiating £ 2 = E 2 (1,p) totally with
respect to 1 to obtain
d£2 o£2 o£2 dp
-=-+-·-
dl a1 ap d1
so that with o£2 /ol < 0, o£2 /op < 0 and dp/dl ~ 0 under trade creation, d£2 /dl and
hence £, are negative.
Gains from Trade 105
means that the formation of the customs union must give rise to a
decline in the relative price of the importable good facing the con-
sumers in the home country. For the unfavourable movement in the
terms of trade that accompanies the trade-diverting union is bound
in the presence of complete specialisation to lower the value of home
production (or national income measured in terms of outputs), and
this factor tends to make dD 2 negative and dD 2 /dt positive. For
dD 2 to be positive, therefore, the necessary condition is that the
trade-diverting union contribute to a decline in the relative price
of the second good. This latter condition can also be seen to be
necessary if the trade diversion is ever to lead to an improvement in
welfare. For if dD 2 > 0 and hence dD 2 /dt < 0, E1 < 0, and if
JE,J > JAPJ, dU may be positive. However, if the substitution effects
on the consumption side are ignored, dD 2 must be negative, dD 2 /dt
and hence E, must be positive, and with A, < 0, dU must be nega-
tive. t In other words, if the goods are consumed in a given propor-
tion, trade diversion must lead to a loss in welfare. This is Lipsey's
interpretation. He accuses Viner of implicitly assuming that the
consumption pattern is constant, for otherwise trade diversion may
also contribute to an improvement in welfare.
Lipsey's critique, however, is open to a serious objection in that he
postulates a Ricardian economy and ignores the substitution possi-
bilities on the production side. Let us then examine these issues
when dX2 =F 0. If the formation of the union results in a decline in
the relative price of the importable good (which, as we have seen
Figure 4.11
is also the only supplier of the home imports. Now suppose the home
country and B form a union which eliminates the importation of the
F goods by any union partner. The new terms of trade facing the
home country are then given by FPb which are less favourable to the
home country than those reflected by FP, ensuring that the union
is trade-diverting. If, following Lipsey, we ignore the substitution
effects on the consumption side, the consumption point in every
case must lie on a given ray from the origin. Suppose this ray is
given by OR and the initial consumption point lies on it. If the sub-
stitution effect on the production side is also ignored, then the
production point remains unchanged in spite of the shift in the
relative prices from DP to FPb. Under· this setting, the new con-
sumption point is given by C' which is clearly inferior to the initial
point C. This is how Lipsey interprets Viner. He argues that if
substitution is allowed on the consumption side, the new consump-
tion point will be different from C' and the level of welfare attained
108 Studies in the Pure Theory of International Trade
will be higher than that associated with C'. In the diagram, FPb is
drawn tangential to the original indifference curve at Cb. It follows
immediately that, by consuming at Cb, the home country can attain
the original level of welfare even with the customs union. Clearly,
then, if the terms of trade between the home country and B were
slightly more favourable than those reflected by FPb, the customs
union would result in a gain in welfare.
If substitution is permitted on the production side also, the pro-
duction point will shift from P to P' where FP~ (parallel to FPb)
touches the transformation curve. If substitution on the consump-
tion side is ignored, the new consumption point C" is superior to the
initial point C in terms of welfare. The welfare gain, of course, would
be higher if commodities are not consumed in a given proportion.
But the point is that even if this is conceded, substitution on the
production side itself may nullify Lipsey's assertion that with a given
consumption pattern trade diversion necessarily leads to a decline in
welfare. The proposition of general validity is that welfare with a
trade-diverting union may decline, rise, or even remain unchanged.
A necessary condition for trade diversion to be welfare-improving
is that the union terms of trade (FPb) must be intermediate to the
domestic-price ratio (DP) and FP. Likewise, a sufficient condition
for trade diversion to tie detrimental to welfare is that the union
terms of trade are such as to raise the output of the importable good
or leave it unchanged. This can be verified easily from the diagram.
In the course of our analysis, the reader may have noticed the
intimate relationship that exists between the propositions estab-
lished in this section and Kemp's theorem that a higher tariff is
inferior to a lower tariff. We have already shown that Kemp's
theorem may not hold when more than one import is subject to the
tariff. It follows then that our customs union results also cannot be
extended beyond the narrow domain of two traded goods.
REFERENCES
[1] Bhagwati, J., 'The Gains from Trade Once Again', Oxford Economic Papers,
XX (July 1968) 137-48.
[2] - , 'Customs Unions and Welfare Improvement', Economic Journal, LXXXI
(Sept 1971) 580-7.
[3] -,and Srinivasan, T. N., 'Optimal Intervention to Achieve Non-Economic
Objectives', Review of Economic Studies, XXXVI (Jan 1969) 27-38.
[ 4] Corden, W. M., 'Tariffs, Subsidies and the Terms of Trade', Economica, XXIV
(Aug 1957) 235-42.
Gains from Trade 109
[5] Gehrels, F., 'Customs Unions from a Single Country Viewpoint', Review of
Economic Studies, XXIV (Jan 1956) 61-4.
[6] Johnson, H. G., 'Optimal intervention in the Presence of Domestic Distortions',
in Trade, Growth and the Balance of Payments (Chicago: Rand McNally, 1965).
[7] Kemp, M. C., The Pure Theory of International Trade and Investment (New
Jersey: Prentice-Hall, 1969) chap. 12.
[8] Krueger, A. 0., and Sonnenschien, H., 'The Terms of Trade, the Gains from
Trade and Price Divergence', International Economic Review, VIII (Feb 1967)
121-7.
[9] Lipsey, R., 'The Theory of Customs Unions: A General Survey', Economic
Journal, LXX (Sep 1960) 496-513.
[10] Samuelson, P. A., 'The Gains from International Trade', Canadian Journal of
Economics and Political Science, v (May 1939) 195-205.
[II]--, 'Social Indifference Curves', Quarterly Journal of Economics, LXX (Feb
1956) 1-21.
[12] Viner, J., The Customs Union Issue (New York: Carnegie Endowment for
International Peace, 1950).
5 The Theory of Nominal Tariffs
Because of its long chronicle, the theory of nominal tariffs has earned
a pivotal place in the literature on the pure theory of international
trade. Historically, the main concern of this theory has been with the
implications of tariffs on the imports of final products, and it is only
recently that the trade theorist has begun to explore the implications
of tariffs on the imports of intermediate products, despite their
existence in practice for centuries. The fast-proliferating literature
concerned with the latter question is surveyed in the chapter on
effective protection; the present chapter is concerned solely with the
issues arising from the imposition of nominal tariffs.
The theory of nominal tariffs figured prominently in the classical
writings of Mill [11], Marshall [8] and Graham [1, 2], among
others, and as with the gains from trade, the theory of tariffs
branched directly out of the controversy over free trade and pro-
tectionism. Some of the normative aspects of tariffs were examined
in the previous chapter. Here our main concern will be with the
positive analysis of tariffs. Although the analysis of tariff dates
back to the classical tradition, interest in the nominal tariffs has
been kept alive by two pioneering contributions, one by Stolper
and Samuelson [13] and the other by Metzler [10], whose basic
analytical framework was cast in the modern Heckscher-Ohlin
mould. The seminal results of Metzler have been more recently
challenged by Sodersten and Vind [12], but Jones [ 4] has effectively
demonstrated the spurious nature of their critique. The analysis in
this chapter parallels closely the illuminating discussion provided
by Jones.
pounded in the previous chapter is not suitable for the issues dis-
cussed in the present chapter.
In the presence of tariffs, each country's demand for imports
depends on the international terms of trade, p, and the level of its
tariff. As before, £ 2 = D 2 - X 2 is the import demand by the home
country for the second commodity, and E 1f = D 1f-Xlf is the
foreign country's import demand for the first commodity. Let lh
and If be the tariffs imposed respectively by the home and the foreign
country and let 7;. = (I +I h) and 1f = (1 +If). Then
E2 = E 2(p,7;,) (5.1)
and
(5.2)
Furthermore,
(5.3)
and
(5.4)
where Ph = pi;, and pf = p/1J are respectively the domestic-price
ratios in the home and the foreign country, Y is national income in
the home country and lf is national income in the foreign country.
For our purposes it is the change in the real income affecting the
change in demand that is relevant. If the change in social welfare is
an index of the change in real income, then for the two-commodity
case we may write
dU
- = dY = dD 1 +phdD2 (5.5)
ul
which expresses the change in the home real income in terms of the
first commodity and is free from any cardinal utility restrictions.
Similarly,
(5.6)
In the absence of tariffs, the budget constraint, as in the previous
chapter, simply shows that the value of consumption equals the
value of production. However, in the presence of the tariffs the
complications arising from the distribution of tariff proceeds have
also to be taken into account. Here there are several alternatives.
112 Studies in the Pure Theory of International Trade
The government may just collect tariff revenue without spending it,
in which case the value of consumption expressed in terms of
domestic prices is still restricted by the value of production. On the
other hand, the government may spend the tariff proceeds in the
manner that private consumers would; this procedure also causes
little trouble. However, if the government's spending proclivities are
different from those of the private sector, the analysis becomes quite
complicated, for now we also have to consider separate demand
functions for the government. The simplest way of handling the
problem arising from the disposal of tariff proceeds is to assume that
the government hands the tariff revenue back to the private sector
in a lump-sum fashion. In this last alternative, which is the one
adopted here, the value of consumption expressed in terms of
domestic prices is restricted by the value of production as well as
tariff proceeds.
Another issue that deserves consideration here concerns the way
the tariff revenue is to be expressed. For the home country which
imports the second commodity, the tariff revenue simply equals
Elf(plf-pl) = tfplElf
(5.7)
The Theory of Nominal Tariffs 113
and
(5.13)
(5.14)
(5.15)
(5.16)
Pt = P' + Tt
in (5.13), we derive
(5.17)
(5.18)
t The substitution effect on the production side represents the change in the output
of the commodities that accompanies the change in commodity prices. This substitu-
tion effect depends, among other thirrgs, on the elasticities of factor substitution in
the sectors. For further discussion of this effect, see Jones [ 4].
The Theory of Nominal Tariffs 115
and
Ah = (sh+eh) (5.19)
(1-mhth/T,)
Since th/T, < 1, and for non-inferior goods 0 < mh < 1, the de-
nominator of(5.17)-(5.19) is positive; (5.18) furnishes the factors
that influence the elasticity of demand for imports as we move along
the offer curve. In view of the construction of the offer curves
described in Chapter 3, it is not surprising to find that ah depends
upon the substitution effects on the consumption and the produc-
tion side as well as the marginal propensity to consume importables;
in addition, the initial level of tariff also affects the magnitude of the
elasticity of the import demand. The expression (1-mhth/T,), as
Jones ([4] p. 420) has observed, is similar to the Keynesian type of
multiplier, and this indeed is the case. Let us call the term
1/( 1 - mh th/T,) the tariffmultiplier. If initially there is no tariff, so that
th = 0, this multiplier phenomenon disappears. In the presence of
an initial tariff, however, this multiplier, as with the Keynesian
system, has a multiple effect. The major difference is that, unlike in
the Keynesian framework where the multiple effect is on national
income, the multiplier in the barter-trade model has a pronounced
effect on the demand for imports. It affects both ah and Ah. For
example, suppose that with a tariff already existing there is an exo-
geneous increase inp so that £ 2 declines. Now a decline in £ 2 lowers
the community's receipts from the tariff proceeds, which leads to a
further decline in the demand for imports, and so on. Therefore the
final decline in £ 2 equals the initial decline in £ 2 multiplied by the
tariff multiplier 1/(1-mhth/T,), and similarly for the term Ah. For a
given p and an initial tariff, any further rise in the tariff serves to
lower the import demand by raising the domestic production of the
importable good as well as by lowering its domestic consumption.
These two effects show themselves in the numerator of (5.19). The
same multiplier effect appears once again. The initial decline in the
import spending serves to lower the community's receipts from the
tariff proceeds, which in tum leads to a further decline in the import
demand and so on, so that eventually the tariff multiplier again has
a cumulative effect on Ah. If initially there is free trade, the intro-
duction of the tariff at constant p causes no change in the national
income, as is evident from (5.16) where with p* = 0 and T, = 1, dY
also equals zero. It may be noted that this is why there will then be
116 Studies in the Pure Theory of International Trade
no income term in (5.19). t The income term (mh) does appear when
th > 0, but only via the tariff multiplier.
Enough has been said on the factors determining the demand for
imports by the home country. By following essentially the same, but
a little more involved, procedure, an analogous expression for ET1
can be obtained :t
EfJ = p*[ef+sf+mf/Tf]- TJ(ef+sf). (5.20)
(l-m1 t1 jT1 )
The next step involves the differentiation of the market adjustment
condition (5.9), so that
p*+Ei = ET. (5.21)
Substituting (5.11) and (5.12) in (5.21) then furnishes
A Tj-AhT:
p* = 1 (5.22)
a1 +ah-l
The denominator of (5.22) is positive for the foreign trade market
to be stable. Hence the relationship between p and the tariff is
determined by the signs of A1 and Ah, both of which, in the absence
of inferior goods or a positive initial tariff, are positive. It is now a
simple matter to observe that, unless one of the a1 and ah is infinity,
an increase in the tariff rate by any country results in an improve-
ment in its terms of trade. With the home country, for example, an
increase in its tariff (Tt > 0, Tj = 0) will lead to a decline in p and
hence to a shift in the terms of trade in its favour. This result is
attributable to the existence of the negative relationship between
E2 and r;, when terms of trade are kept constant. From (5.17) it is
apparent that with p* = 0, E!/Tt < 0 irrespective of whether
initially the tariff is positive or zero, provided, of course, that inferior
goods are non-existent. Since at constant terms of trade an increase
in the tariff results in a decline in the demand for imports, the inter-
national relative price of the importable good must eventually
decline to restore equilibrium in the foreign trade market. However,
if the exportables are inferior in social consumption or if the im-
t This merely confirms our result derived in the previous chapter, that with initial
free trade and given terms of trade, /aissez-faire is the optimal policy.
t It may be recalled here that ET = £!1 , and that e1 , s1 and m1 carry the same mean-
ing in the foreign country as their counterparts in the home country.
The Theory of Nominal Tariffs 117
H'
\
\ H
I
I
I
JE'
8 Q
Figure 5.1
118 Studies in the Pure Theory of International Trade
(5.24)
Figure 5.2
Figure 5.3
commodity (its importable good) should be less than the home marginal
propensity to consume the first commodity (its exportable good).
A further comprehension of these results may be gained from an
examination of Figs. 5.2 and 5.3. Fig. 5.2 depicts the case of an
elastic foreign offer curve, OF, whereas in Fig. 5.3 OF is inelastic in
the neighbourhood of initial equilibrium pointE and the post-tariff
equilibrium pointE', which is generated by the intersection of OF
and the tariff-distorted home offer curve OH', with the pre-tariff
home offer curve given by OH. The post-tariff terms of trade are
given by the slope of OE', and at the new international prices the
foreign country offers 0 R of the second good in exchange for RE'
of the first commodity; the home producers offer RA, so that the
difference AE' is collected by the government as tariff revenue in
terms of the first commodity. The rate of tariff in terms of the first
commodity therefore is AE'/E' R. The domestic-price ratio in the
home country can be computed simply from the formula that
or
122 Studies in the Pure Theory of International Trade
which is nothing but the inverse of the slope of the dotted ray OA. In
Fig. 5.2, where the foreign offer curve is shown to be elastic, OA is
less steep than OE, which indicates the initial terms of trade. This im-
plies that the domestic-relative price of the exportable commodity
has declined as a result of the introduction of the tariff. By contrast,
in Fig. 5.3, where the foreign offer curve is drawn inelastic in the
neighbourhood of equilibrium points, OA is steeper than OE, show-
ing thereby that the domestic-relative price of the exportable good
(importable good) has risen (declined) in the post-tariff equilibrium.
Thus the diagrams demonstrate the necessary condition for the
validity of the Metzler paradox, namely, that the foreign import
demand should be inelastic.
For the given foreign tariff ( Tj = 0), the home welfare is maximised
It the coefficient of p* equals zero. The optimum tariff (t 0 ) formula
is then given by
(5.27)
Figure 5.4
Note that the optimum tariff is not the one that is associated with the
most favourable terms of trade. For example, if OH' intersected OF
at Q, the terms of trade would be more favourable than those
furnished by G; but the concomitant level of welfare would be
suboptimal.
So far we have assumed that the foreign country takes no action in
response to the tariff imposed by the home country. The more
realistic situation is where both countries attempt to reap gains from
their monopolistic positions in trade. When each country is ready to
retaliate against the imposition of the tariff by the other, the optimum
tariff argument loses much of its appeal. Both countries obviously
cannot gain by imposing tariffs, and if retaliatory measures con-
tinue, both may lose in the final equilibrium. In all likelihood both
countries would then be better off with the free trade policy,
although, as Johnson [3] has shown, one country under certain
conditions may be able to benefit by restricting its trade at the
expense of the other even if the latter takes retaliatory measures.
modity, the difference AE' being the tariff revenue collected by the
government. Let us now consider an export tax imposed at a rate
equal to AE'/RE'. The foreign importers now receive RE', the home
exporters offer RA in exchange for AB and the customs revenue in
terms of the first commodity again equals AE'. Hence the extent of
the shift of OH is the same in both types of taxes. Evidently, then,
the terms of trade, imports, exports, domestic production and con-
sumption are the same in both cases; the only difference is that in
one case the tax is paid by the importers, whereas in the other it is
paid by exporters.
A simple algebraic proof of Lerner's symmetry theorem is given
as follows. Let te be the export tax imposed by the home country, so
that
given, of course, that th = te. This implies that the home outputs
are also the same in both cases. However, the demand for each
commodity depends not only on relative prices, but also on real
income. Hence for the symmetry theorem it is necessary to show that
real incomes are also the same under both taxes. Note that, in the
case of the tariff, real income expressed in terms of local prices is
given by
Y = X1 +PhXz +thpEz
whereas that in the case of the export tax is given by
It is interesting to note that the tariff revenue in the two cases looks
different, but is in reality the same. The apparent difference is attri-
butable to the choice of the numeraire. Therefore, after adjusting for
the different absolute prices, thp£ 2 = teEd(!+ te). Hence Y = Y,,
which, with similar relative price structure, implies that production,
consumption and imports are the same under both types of taxes.
128 Studies in the Pure Theory of International Trade
REFERENCES
[I] Graham, F. D., 'Some Aspects of Protection Further Considered', Quarterly
Journal of Economics, XXXIX (Feb 1925) 324-30.
[2] - - , The Theory of International Values (Princeton U.P., 1948).
[3] Johnson, H. G., International Trade and Economic Growth (London: Allen &
Unwin, 1958) chap. ii.
[4] Jones, R. W., 'Tariffs and Trade in General Equilibrium: Comment', American
Economic Review, LIX (June 1969) 418-24.
[ 5] Kemp, M. C., The Pure Theory of International Trade (Englewood Cliffs, N.J.:
Prentice-Hall, 1964).
[6] --, The Pure Theory of International Trade and Investment (Englewood Cliffs,
N.J.: Prentice-Hall, 1969).
[7] Lerner, A. P., 'The Symmetry between Import and Export Taxes', Economica,
III (Aug 1936) 306-13.
[8] Marshall, A., The Pure Theory of Foreign Trade (London: London School of
Economics and Political Science, 1949).
[9] - - , Money, Credit, and Commerce (London: Macmillan, 1923).
[!OJ Metzler, L. A., 'Tariffs, the Terms of Trade, and the Distribution of National
Income', Journal of Political Economy, LVII (Feb 1949) 1-29.
[II] Mill, J. S., Principles of Political Economy (London: Longmans, Green, 1909).
[12] Siidersten, B., and Vind, K., 'Tariffs and Trade in General Equilibrium',
American Economic Review, LVIII (June 1968) 394-408.
[13] Stolper, W. F., and Samuelson, P. A., 'Protection and Real Wages', Review of
Economic Studies, IX (Nov 1941) 58-73.
6 Economic Expansion and
the Terms of Trade
6.1 Introduction
The analysis of interrelations between economic expansion and the
terms of trade cannot boast of a long, uninterrupted history. Here
and there one may come across a few scattered references to terms
of trade in studies of the incidence of technological innovations on
the growing country's net gain from growth in the works of Mill
[18], Edgeworth [9], Bastable [2] and Ohlin [19]; of the desirability
of the policy of protection in the wake of technical improvements
in what is known as the German tariff controversy ;t of the celebrated
transfer problem sparked primarily by Thornton [22] and Hume
[14] and subsequently resurrected by Taussig [21] and Keynes [17]
(in the famous debate over the German reparations problem);
finally, of the impact of international capital movements featuring
in the works of Fanno [10], Iverson [15] and Ohlin [19]. But the
avowed application of the tools of general equilibrium analysis to
the diagnosis of the behaviour of the terms of trade consequent upon
growth is only a post-war phenomenon. Whatever ingenuity the
problem of shifts in inter-country commodity prices elicited from
the economist in the pre-war era was, with a few exceptions,t inci-
dental; it arose not because of an interest in the application of
economic theory to the problem of economic development, which
itself never figured prominently in neo-classical writings, but was
due to the economist's preoccupation with some· other issues of
current importance wherein considerations of the terms of trade
were unavoidable.
Like many other fields of economics, trade theory has welcomed
the dent made in it by the theory of economic growth pioneered by
Sir Roy Harrod. t The evolution of this branch of theory has been
further accentuated by two other burning issues: (1) the problem of
international liquidity, of which the apprehensions concerning the
long-run shortage of dollars in the world markets constituted but
one phase, for since 1958 the dollar paucity has turned into the
dollar glut; and (2) the question of the secular deterioration in the
terms of trade of the countries producing primary products.
Until the middle of the 1950s it was quite fashionable among
trade theorists to argue that the dollar difficulties of the post-war
period constituted a legacy of the pre-war years; that there was in
fact a long-run tendency for the U.S. balance of payments to run into
surplus, only to create cumulative difficulties for the non-dollar
world. A variety of reasons for the tendency were advanced. Balogh
[1] and others took the faster growth of the U.S. productivity than
that in other countries as granted, with the result that the increasingly
cheaper U.S. exports were displacing the exports of other countries.
Hicks [13] suggested, in addition, that innovations in the U.S.
tended to be 'import-biased', that is, they were concentrated in those
goods which competed with American imports.
So overwhelming was the economist's obsession with the secular
shortage of dollars until even the late 1950s that Keynes's sole dis-
sension that 'in the long run more fundamental forces may be at
work, ... tending towards equilibrium', so that 'the chances of the
dollar becoming dangerously scarce in the course of the next five to
ten years [wereJ not very high' ,:j: proved only a distant cry in the
wilderness. From the strength of hindsight Keynes once again proved
right, although the affirmation of his prophecy came almost a decade
after his demise. As of 1950, the U.S. balance of payments has dis-
played chronic deficits; the entire situation has been so transformed
that strong equivocations are now cast in some quarters as to the
very stability of the dollar. The post-war international monetary
crises, climaxed by the recent Smithsonian accord leading to the
t To be sure, Harrod [12] did not address himself to the trade problems arising
from growth; but he did succeed in opening new vistas for the trade theorist to
explore. His references to growth stemming from technical progress and capital
accumulation may be deemed partly responsible for the assiduity with which the
economist has engaged himself to unravel the interrelation between growth and the
terms of trade.
t Quoted from Clement et a/. ( [ 5] p. 352). The bracketed expression is theirs.
Economic Expansion and the Terms of Trade 131
H
R -----------
Figure 6.1
(6.5)
1. (6.6)
must also exceed mh, so that p must fall, or to say the same thing in
different words, the home terms of trade must improve.
From (6.4) three possibilities emerge, to wit, mh ~ zh. These are
depicted by Fig. 6.2. Growth is represented by an outward shift of
the transformation curve from H H' to GG'; changes in the outputs
and the demands for X 1 and X2 are evaluated at unaltered terms of
trade, as may be evident from the fact that the pre-growth terms of
trade given by FP are the same as those in the post-growth situation
given by FP' (FP is parallel to FP'). As a consequence of growth, the
production point shifts from P toP'. Now if the consumption point
shifts from C to C', that is, if PP' is parallel to CC', zh equals mh. The
home offer curve does not shift in this case, so that the terms of
trade will remain unaltered. If the consumption point moves to any-
where towards the left of C', say C, zh falls short of mh, so that the
terms of trade will move against the home country; if to anywhere
towards the right of C', say C", zh exceeds mh and the terms of trade
will switch in favour of the home country.
FP 1
x,
Figure 6.2
136 Studies in the Pure Theory of International Trade
so that
(6.10)
The first two terms in the right-hand side of (6.10) represent the
change in the value of production resulting from growth at constant
terms of trade, and hence equal o YjoG. Therefore
dY _ oY E dp
dG- oG- 2 dG. (6.11)
(6.12)
Since fJ Y/oG > 0, it is clear from (6.12) that for the home real
income to decline as a result of growth
(6.14)t
(6.15)
x,
Figure 6.3
(2.37)
140 Studies in the Pure Theory of International Trade
and
n _ ).,L1K* -).,K1L* (2.38)
2 - IA.I
where, from (2.17) and (2.17*),
I)., I = L1L2
A.L1A.K2-).,K1A.L2 = A.Ll-A.Kl = A.K2-).,L2 = LK (k2-kl).
To begin with, let us assume that only the labour force grows, while
the capital stock remains constant, so that L* > 0, K* = 0 and
hence the growth agent (G) is identified by L. With this in mind,
(2.37) and (2.38) can be simplified to
ax 1 XtK2
(6.16)
aL L 1L 2(k 2-k 1)
and
ax2 -X2Kt
aL (6.17)
= L 1L 2(k 2 -k 1)"
The change in national income at constant terms of trade, with
p = l initially, is then given by
ay axl ax2 x1K2 -X2K1 w
-=-+-= =- (6.18)t
aL 8L 8L L1L2(k2 -kt) Pt
so that
(6.19)
t Solving the two price equations in Chapter 2 and remembering that p = I ini-
tially, we obtain
Economic Expansion and the Terms of Trade 141
then, with p initially unity, it means that the relative share of capital
in the first commodity exceeds that in the second commodity, that is,
the first commodity is capital-intensive relative to the second. With
the denominator of (6.19) positive, zL > 1; on the other hand, if
this denominator is negative, that is, if the first commodity is labour-
intensive relative to the second, zL < 0. Thus it is clear that the
economic expansion resulting from labour growth alone is ultra-
biased. It is ultra-import-biased if the importable good (the second
good) is labour-intensive, and ultra-export biased if it is capital-
intensive relative to the other commodity. The terms of trade facing
the growing home country will eventually improve in the former
case but deteriorate in the latter case. The reason follows directly
from the Rybczynski theorem, according to which an increase in the
supply of labour alone results at constant terms of trade in a rise in
the output of the labour-intensive commodity at the expense of the
output of the other commodity. Hence if the labour-intensive com-
modity happens to be the importable good, labour expansion is
ultra-import-biased; ifthe exportable good, it is ultra-export-biased.
Similar results can be obtained if it is capital which is the expand-
ing factor while the labour force remains unchanged. In other words,
economic expansion is necessarily ultra-biased if only one factor
experiences an increase in supply.
The geometric exposition of the effects of an increase in the supply
of one factor on the two outputs at constant commodity prices was
presented in Chapter 2 in terms of the box diagram. Here we utilise
an alternative diagrammatic technique which turns out to be of
great help in the ensuing discussion of technological change.
Consider Fig. 6.4, where AB gives the wage/rental ratio, x 1 and
x 2 are the unit isoquants of the two commodities with their points
of equilibrium at v and s respectively, and point E indicates the
factor-endowment ratio (K/L). For the sake of simplicity, the dia-
gram shows that the aggregate output in the economy equals one
unit of the first commodity (Ov) plus one unit of the second com-
modity (Os), and that since the unit isoquants lie on the same isocost
line (AB), the commodity-price ratio equals unity. It is worth point-
ing out that a good grasp of the way this diagram is drawn is essential
for comprehending the present and subsequent analysis. For
example, note that the level of the two outputs is ascertained by
drawing from E lines parallel to Ok 1 and Ok 2 , the slopes of which
indicate the capital/labour ratios in the two industries. Point E,
142 Studies in the Pure Theory of International Trade
K
L
Figure 6.4
Xi= X!= L*
because it may be recalled that
These results will gain further clarification from Fig. 6.5, which
is constructed on the same principle as Fig. 6.4. It can be observed
that if factor growth follows the growth path sE, parallel to k 1 ,
there will be no change in the output of X 2 ; this depicts the case
where dK/dL = k 1 . For example, if the factor-endowment point
shifts from E to G, the output of X 1 rises from Ov to On, without
any change in the output of X 2 (equal to Os). It follows, therefore,
that if factor growth were to follow a path steeper than sE, sayER
144 Studies in the Pure Theory of International Trade
/(
Figure 6.5
(here dK/dL > k 1 ), the output of X 1 would again rise to On, but
that of X 2 would decline from Os to Oc. Similarly, it can be easily
demonstrated that if factor growth followed the growth path EP
(i.e. dKjdL = k 2 ), the output of X 2 would rise without any change
in the output of X 1 ; if it were to follow a path flatter than
vE(dK/dL < k 2 ), the rise in the output of X 2 would be accompanied
by a decline in the output of X 1 • Finally, if the growth path followed
by the factors is OE, that is, if the factor-endowment point shifts
from E to Q, the output of both commodities will rise in the same
proportion; for the output of X1 rises to On and that of X 2 toOk
and it can be easily seen that sk/Os = vn/Ov.
The impact of growth in both factors on the home terms of trade
can be determined as before. It may be observed that only those
cases admit of a priori conclusions where dK/dL ~ k 1 , or k 2 ~
dK/dL; in other cases results will depend on many additional factors
such as the magnitude of the home marginal propensity to consume
importables, etc.
Economic Expansion and the Terms of Trade 145
w* = (6.20)
(6.21)
and
(6.24)
(6.26)
L
Figure 6.6
X 2 to k'1 and k2. The next step, of course, is to draw from E lines
parallel to the new factor-intensity rays, so that En is parallel to k2
and Ec to k'1 • It may be observed that the output of X 2 rises from
its pre-expansion unit level of output (now equal to Om) to Oc,
whereas the output of X 1 declines from Oq to On. In other words,
neutral technical progress in X 2 is ultra-import-biased. A similar
diagram can be constructed to show that a neutral improvement in
xl alone will be ultra-export-biased.
Intensive-Factor-Saving Technical Progress
Suppose now that the improvement in X 2 , the labour-intensive in-
dustry, is labour-saving. Evidently, this is a case of intensive-factor-
saving technical progress. With technical progress in the second
industry tending to bring about a greater saving in the use of labour
than capital, B!. 2 > B't 2 • The implications for X 1 are evident from
a glance at (6.25); with 101 < 0, the output of X 1 declines even more
than was the case with the neutral improvement in X 2 alone, for
now an additional negative term iOIA.K 2 A.L 2 (Bt 2 - B't 2 ) is added to
L
Figure 6.7
150 Studies in the Pure Theory of International Trade
Since k 2 < k 1 and B! 2 < B1 2 , both k 2 /k 1 and B! 2 /B1 2 are less than
one. The sign of X! is likely to be positive if (i) the capital/labour
ratios in the two industries are markedly different, and/or (ii) the
saving in the use of capital due to technical progress in X 2 is not
much greater than the corresponding saving in the use of labour,
that is, B! 2 / B1 2 is close to unity. On the other hand, if the
capital/labour ratios in the two industries are sufficiently close so
Economic Expansion and the Terms of Trade 151
K
L
Figure 6.8
Figure 6.9
REFERENCES
[I] Balogh, T., The Dollar Crisis: Causes and Cure (Oxford: Blackwell, 1949).
[2] Bastable, C. F., The Theory of International Trade, 4th ed. (London: Macmillan,
1903).
[3] Bhagwati, J., 'Immiserizing Growth: A Geometric Note', Review of Economic
Studies, xxv (June 1958) 201-5.
[4] - - , 'Growth, Terms of Trade and Comparative Advantage', Economia
Internazionale, XII (Aug 1959) 393-418.
[5] Clement, M.D., Pfister, R. L., and Rothwell, K. J., Theoretical Issues in Inter-
national Economics (Boston: Houghton Mifflin, 1967).
[6] Corden, W. M., Recent Developments in the Theory of International Trade
(Princeton: International Finance Section, Princeton University, 1965).
[7] --,'Economic Expansion and International Trade: A Geometric Approach',
Oxford Economic Papers,vm (Sep 1956) 223-8.
[8] Dietzel, H., 'The German Tariff Controversy', Quarterly Journal of Economics,
XVII (May 1903) 365-416.
[9] Edgeworth, F. Y., 'On a Point in the Pure Theory of International Trade',
Economic Journal, IX (March 1899) 125-8.
[10] Fanno, M., Normal and Abnormal International Capital Transfers (Minneapolis:
Univ. of Minnesota Press, 1939).
[II] Findlay, R., and Grubert, H., 'Factor Intensities, Technological Progress, and
the Terms of Trade', Oxford Economic Papers, XI (Feb 1959) 111-21.
[12] Harrod, R. F., Towards a Dynamic Economics (London: Macmillan, 1948).
[13] Hicks, J. R., 'An Inaugural Lecture: The Long-Run Dollar Problem', Oxford
Economic Papers, v (June 1953) 117-35.
(14) Hume, D., Writings on Economics, ed. E. Rotwein (Madison: Univ. of
Wisconsin Press, 1955).
(15] Iverson, C., Some Aspects of the Theory of International Capital Movements
(Copenhagen: Levin & Munksgaard, 1936).
[16) Johnson, H. G., 'Effects of Changes in Comparative Costs as Influenced by
Technical Change', in R. Harrod and D. Hague (eds), International Trade in a
Developing World (London: Macmillan, 1963) chap. 4.
(17) Keynes, J. M., 'The German Transfer Problem', in Readings in the Theory of
International Trade (Philadelphia: Blakiston, 1950).
[18] Mill, J. S., Principles of Political Economy, ed. Sir W. J. Ashley (London:
Longmans, Green, 1909) bk 111.
[19] Ohlin, B., Interregional and International Trade (Cambridge, Mass.: Harvard
U.P., 1933).
[20] Prebisch, R., 'Commercial Policy in the Underdeveloped Countries', American
Economic Review, XLIX (May 1959) 251-73.
[21] Taussig, F. W., International Trade (New York: Macmillan, 1927).
[22] Thornton, H., An Enquiry into the Nature and Effects of the Paper Credit of
Great Britain (London: Allen & Unwin, 1939).
[23] Wagner, A., 'Agrarian State versus Manufacturing State', in F. W. Taussig
(ed.), Selected Readings in International Trade and Tariff Problems (Boston:
Ginn, 1921).
7 Intermediate Products:
The Inter-Industry Flows
Until now we have assumed that goods are produced with the help
of primary factors only, but since in practice much of the production
activity would come to a halt if there were no intermediate
products - goods which are produced to be used as inputs in other
goods - this assumption, although a convenient one, is very
unrealistic. There is hardly any justification for this assumption,
even though the bulk of trade theory has ignored the presence of
material inputs which constitute a very large proportion of the
total volume of world trade. t It is only recently that trade theorists
have come to recognise the importance of intermediate products in
the production process, but even here the general tendency has been
to defend the neglect of the treatment of material inputs in the earlier
literature. In his book on trade theory, for example, Kemp
([7], p. 148), citing Vanek [10], defends the neglect of the
incorporation of intermediate goods in earlier trade theory. In
some respect, this neglect may be justified. Many properties of the
general equilibrium model presented in the previous chapters carry
over to the model with intermediate goods. However, there are
some crucial differences which have not been recognised before.
The analysis of such similarities and differences is the subject-
matter of this chapter.
Within the discussion on intermediate products, the earlier
analyses have been disproportionately concentrated on the implica-
tions of inter-industry flows, goods that serve the dual role of
intermediate as well as final products. However, it cannot be
denied that there exist several types of goods which are used not at
all for consumption, but are produced solely to be used as inputs
in other final goods. These may be called 'pure' intermediate
products. In this chapter we are concerned only with the inter-
and
where cij• as before, denotes the direct requirement of the ith input
per unit of the jth good. Full employment of the primary factors
implies that
(7.3)
and
(7.4)
It is convenient, however, to express the full-employment conditions
in terms of net outputs only. Solving for X12 and X21 from the
following two equations:
C12 = X12/(x2+X21)
156 Studies in the Pure Theory of International Trade
and
we obtain
X 12 -_ C12(x2+C21x1) an
d X _ C21(x1+C12x2)
21 - ·
1-c12 • C21 1-c12 · C21
Substituting these in the full-employment equations, (7.3) and (7.4),
we get
(7.5)
and
Since cl2c21 < 1 by assumption, and since ICI = eLl CK2- CK1CL2•
the sign of IR I is the same as the sign of Iq. t The sign of Iq reflects
the factor-intensity ranking of the two commodities in terms of
'apparent' or net production coefficients, whereas the sign of IRI
expresses the factor-intensity ranking in terms of 'true' or gross
production coefficients. The conclusion is unmistakable. In the
inter-industry flow model, the gross and the net factor-intensity
rankings of commodities are identical. In other words, if a
commodity appears to be, say, capital-intensive relative to the
other, then it will remain capital-intensive even when account is
taken of the amounts of the primary factors embodied in the
production of those quantities of a goo.d which are used in the pro-
duction of the other. This identity suggests that at least those
theorems, usually derived in the absence of intermediate products,
that depend exclusively on the inter-industry factor-intensity
ranking will hold without any qualitative modification even in the
presence of inter-industry flows. As will be shown later, this indeed
turns out to be the case. But first we attend to the properties
associated with the transformation curve in the present model.
(7.17)
(7.19)
(7.20)
and
(7.21)
160 Studies in the Pure Theory of International Trade
where A.ii denotes the proportion of the total endowment of the ith
primary factor used directly and indirectly in the jth sector (for
example, A.Ll = RL 1x 1/L), and ()ii stands for the total distributive
share of the ith primary factor in the jth sector (for example,
()Ll = RL1wfp 1 ). As usual, in obtaining (7.20) and (7.21) use has
been made of the cost-minimising condition (7.16). There is yet
another way of writing (7.20) and (7.21), one which for some
problems turns out to be more useful. Differentiating (7. 7) and
(7.8), we get
(7.20*)
and
(7.21*)
where Pii is the net relative share of the ith factor (i = L, K, 1, 2) used
in the production ofthejth product; for example, PL1 = CL1wfp 1 •
Evidently, then, PL1 +pK 1 +p 21 = 1, and so on. Solving (7.20) and
(7.21) simultaneously, we obtain
w• = (OK2PT-OKlP!>!IOI (7.22)
r• = (- OL 2 PT + OLlp!>fiOI (7.23)
and
(w*-r*) = -(p! -pVJIOI (7.24)
where
It can be readily seen that the sign of IOj is the same as the sign of
jRj, so that if k 2 ~ k 1 , jOj ~ 0.
Each direct input-output coefficient CiJ• and thus each total
input-output coefficient Rii, is a function of the prices of productive
inputs. Therefore, we may write
assumed all along, there are certain restrictions on u{k which are
sufficient to ensure the positive signs of tx and {3. This complicated
and lengthy task is taken up in the appendix. Returning to (7 .18)
and (7 .19), let
and
* (A.K2/'L +AL21'K)( * *)
x1 = IA.I w -r (7.27)
and
(7.28)
where
and
* _ -(A.KlYL +A.LlYK)( * _ *)
Xz - IA.I·IOI Pl Pz . (7.30)
G'
Figure 7.1
equal C12 and C21 , that is, the slope of EC12 intersecting 0 X1 at
A, and that of EC21 intersecting OX2 at B, represent, respectively,
the requirement of X1 per unit of X 2 and of X 2 per unit of X1 ; OA
is then the net output of X1 and OB the net output of X 2 , and these
two furnish a point E' that corresponds to E on the gross
transformation curve TT'. In other words, AX of X1 and BY of
X2 serve the role of material inputs in producing net outputs equal
to OA and OB. By repeating such construction for other points on
TT', we derive all the points on the net transformation tt' except
its corner points, t and t'. For example, t' is obtained by drawing
from the origin a ray OC', parallel to EC21 and intersecting TT'
at C'; drawing C't' parallel to EC 12 then furnishes one of the
Intermediate Products: The Inter-Industry Flows 165
• p* (}Kj(pf-pt) (7.31)
w - j = IOI
166 Studies in the Pure Theory of International Trade
and
• • -(hj(p!-p!)
r -pi = IOI (7.32)
(7.33)
and
*_ A.L1K*-A.K1L*
(7.34)
x2 - IA.I
With xl capital-intensive relative to x2, IA.I < 0. A rise in the
supply of, say, capital alone then raises the net output of the
capital-intensive commodity, x 1 , and lowers the net output of the
other commodity, x 2 , to show that the Rybczynski theorem con-
tinues to hold in the presence of material inputs. However, the
theorem holds only qualitatively, because the magnitude of the
changes in the net outputs is different in the presence of inter-
mediate goods. In the case where only the labour force is growing,
for example, we may write from (7.33) and (7.34) that
dx 1 cK2+CK1c12
(7.35)
dL CL1CK2- CL2CK1
and
dx 2 -(CKl + cK2c21)
(7.36)
dL eLl cK2- cL2cK1
Given that the denominator of (7.35) and (7.36) is negative, the
Intermediate Products: The Inter-Industry Flows 167
decline in the net output of X 1 and the rise in the net output of X 2
resulting from labour growth alone will be greater in the presence
of intermediate goods (so that cl2 and c21 are positive) than
would otherwise be the case. In other words, the economic
expansion resulting from labour growth becomes more ultra-
biased in the presence of inter-industry flows. This is because,
with the increase in the output of one commodity, additional units
of output of the other commodity must be withdrawn from final
consumption, a factor that vanishes in the absence of intermediate
products. It follows then that the magnitude of the changes in the
terms of trade consequent upon the increase in labour supply will
also be more pronounced in the presence of inter-industry flows.
Thus we see that the Stolper-Samuelson and the Rybczynski
theorems are unaffected qualitatively by the incorporation of inter-
industry flows. It is now a simple matter to show that the magnifi-
cation effect and the fundamental duality that were found to exist
between these two theorems also remain unaltered. A simple proof
consists in the observation that equations (7.31)-(7.34) are exactly
the same as the corresponding equations (2.35)-(2.38) in Chapter 2,
except that j),l and 101 in each chapter are different in magnitude
but not in sign.
(7.41)
and
*
* () * --P2+(1-C12C21)
() L2w+x2r p 12 T* (7 42)
.
where Pii and ()ii are defined as before. If commodity prices are
kept constant, (p!- p!) = 0 so that, by subtracting (7 .42) from
(7.41), we obtain
(7.43)
Our conclusions in this section, just like the ones derived in the
section on the transformation curve, depend chiefly on the signs of
YL and YK· Recalling that both YL and YK are positive, and given our
assumption that C12 C21 < l, it can be seen that xt > 0 and
x~ < 0. In other words, the output of the commodity enjoying
the technical improvement rises and that of the other commodity
declines when terms of trade are kept constant. Thus neutral
technical progress in any commodity is ultra-biased even when
inter-industry flows are present.
FP 1
Figure 7.2
,,
Figure 7.3
172 Studies in the Pure Theory of International Trade
labour and the other good. Each of Of2 and 0/1 are drawn under
the constraint of full employment of labour, which in Melvin's
words implies that 'all possible production points must be convex
combinations of points on/1 and};' ([9] p. 142). In Fig. 7.3, inputs
are measured along the negative axes and outputs along the
positive axes; M 1 and M 2 are the points where/1 and/2 respectively
approach their maxima. 1,-et AB be the common tangent drawn to
/ 1 and};. Then M 2 AHH'BM 1 is the net transformation curve. In
autarky, or in the case where the imported good is used only for
consumption, the transformation curve is given by the segment
HH'. If the home country is a small country and faces a given
world-price ratio different from that given by the slope of HH',
then, of course, the home country will benefit by the introduction
of trade. It also follows, from the argument presented above in the
two-primary-factor case, that the gains from trade will be higher
if the world-price ratio is such as to take the production point to
the negative quadrant. However, if the equilibrium world-price
ratio is obtained through the interaction of the offer curves of the
trading countries, it is not necessary that the free trade price ratio
will differ from the autarky price ratios of both trading partners.
This follows directly from an examination of Fig. 7.4, where OKH
H'
K'
Figure 7.4
Intermediate Products: The Inter-Industry Flows 173
and OGF are the home and foreign offer curves of the Ricardian
model if the imported goods are not used as inputs. Thus K and G,
as in Chapter 3, indicate the points of complete specialisation in
the home and the foreign countries respectively. Point K, for
example, conforms to point H' in Fig. 7.3. However, if the
imported goods are not only used for consumption but also for
production, the points of production are no longer limited to either
H or H'; as suggested, they extend to the negative quadrants. This
means that the kinks in the offer curves in Fig. 7.4 also occur at
higher volumes of trade. Now if the foreign offer curve extends
only to OG 'F' and the home offer curve undergoes a large shift to
OK'H', the free trade equilibrium will then occur at A instead of
at £, which is the point of intersection between the two offer curves
in the absence of trade in intermediate products. The reader may
notice a clear difference in the two situations. If intermediate goods
are not traded, the free trade price ratio (0£) differs from the
home autarky price ratio (0 K), whereas in the presence of trade in
intermediate products the free trade price ratio and the home
autarky price ratio do not differ, as is the case in Fig. 7.4. In the
former situation the home country definitely benefits from trade,
as is described in Fig. 7.3, where FP is the world-price ratio
corresponding to OE, production moves from the self-sufficiency
equilibrium point S to H', consumption to C', and social welfare
improves from U0 to U 1 • In the latter situation, where the autarky
price ratio reflected by the slope of AB remains unchanged,
production moves from S to any point on H' B to the south-east
of H' (but it cannot go beyond B because the price ratio is
unchanged), consumption remains at S, again because of the
absence of any change in autarky prices, and welfare is unchanged
at U0 , which lies clearly below U1 , the welfare level that could be
attained if the imported good was not an intermediate good. In
Fig. 7.3 the actual production point in free trade is given by B, the
net output of X 1 equals OP, the import of X 2 equals GS, part of
which, SC, is used for consumption and the other part, CG, in the
production of X 1 • This is how Melvin shows that the introduction
of trade in intermediate goods may not be gainful to both countries.
The converse of this result is also true. In Melvin's words, 'a situa-
tion which results in a gain for only one country when there is trade
in consumption goods only, may very well result in gains for both
countries when trade in intermediate goods is allowed' ([9] p. 151).
174 Studies in the Pure Theory of International Trade
H'
Figure 7.5
7.6 Appendix
Since each production coefficient is a function of input prices, we
can write Cij = Cij(w,r,pJ(i = L,K,I,2; j = 1,2; i =I=}). Differen-
tiating this totally, and remembering that p 1 does not affect the
input choice in the first commodity, we have
C!1 = PLluhw* + PK1u£Kr* + Pz1ubp~
c:l = PLl(JiKw*+PKl(Jhr*+PzlUkzp~ (A7.1)
Ci1 = pLlubw*+PKlu}czr*+pzlU~zP~·
Intermediate Products: The Inter-Industry Flows 175
Similarly,
C!2 = PL2aiLw* + PK2uiKr• + P12uf.,p!
c;2 = PL2uiKw• + PK2uicKr• + p12ufop! (A7.2)
Ci2 = PL2ui, w• + PK2afor* + P12uLp!
(A7.4)
where the lJ's are, as before, the gross share of the relevant primary
factor. If all factor prices change in the same proportion, the
production coefficients remain unchanged. Thus if w• = r* =
pf = p~, C;j = 0. This implies that
(A7.5)
Cft = -(P2t(}K2ub+PKtuiK)(w*-r*)
c:1 = (P21(}L2ub+pL1uiK)(w*-r*) (A7.11)
q't = (pL1lJK2ub-PKteL2ub)(w*-r*)
and
Cl2 = -(pl2(}Ktuf.t +PK2uf.K)(w*-r*)
c:2 = (Pt2(}L1uh +PL2UEK)(w*-r*) (A7.12)
Ct2 = (PL2(}Ktuf.t-PK2(}L1u'it)(w*-r*).
Hence
• _ dRLI _ CL1Clt +C21RL2C!1 +C21(CL2Cl2+C12RL1Ct2)
RL l- -
RLl eLl+ cL2C21
PL1Ctt +p21lJL2Ct1 +P2t(PL2Cl2+P12lJL1Ct2)
=
Intermediate Products: The Inter-Industry Flows 177
Rt2 = - (P+P12~)(w*-r*)
'1L
• _ <~+P21P>( •
R K1- *)
QK W -r
• _ <P+P12~)( w • -r *)
R K2-
'1K
where
~ = PL1P210i:2ub+PK1P21(}f_2ub+PL1PK1u£K
P = PL2PuOi.1uf.1 + PK2PuOl1ui.1 + PL2PK2u1K
QL = PL1 +PL2. P21• QK = PK1 +PK2. P21
'1L = PL2 + PL1 • P12• '1K = PK2 + PK1 • Pu·
0'LL -PK1)
1 -- (- -P21)
- 0'1LK + (- - 0'1L2·
PL1 PL1
Substituting for ulK from (A7.7) and for ub from (A7.9), we get
1 -
O'LL- 1 ( PK10'KK
-2-
2 1 +2 PK1P210'K2
1 + P210'22.
2 1)
PL1
178 Studies in the Pure Theory of International Trade
1
(TLL =
1 [
-2 (A7.13)
PKl
PL1
We have previously argued that under the assumption that pro-
duction functions exhibit constant returns to scale and are quasi-
concave, the 'own' partial elasticities of substitution are negative,
which implies that the quadratic form on the R.H.S. of (A7.13)
must be negative definite. This constraint is satisfied if (i) uh < 0,
(ii) u1 2 < 0, and (iii) ul::Ku1 2 - (ul:: 2 ) 2 > 0. The first two conditions
have already been assumed to be valid, while the third may be
written as
Remembering that
a:= P21PL1ei_2ub+P21PK10l2ul::2+PL1PK1uh
we can now write that
or
or
or
(A7.14)
Intermediate Products: The Inter-Industry Flows 179
REFERENCES
[I) Allen, R. G. D., Mathematical Analysis for Economists (London: Macmillan,
1938).
[2) Batra, R.N., and Pattanaik, P. K., 'Economic Growth, Intermediate Products,
and the Terms of Trade', Canadian Journal of Economics, IV (May 1971) 225-37.
[3) Casas, F. R., 'The Theory of Intermediate Products, Technical Change and
Growth', Journal of International Economics, II (May 1972).
[4) Chipman, J. S. 'A Survey of the Theory of International Trade: Part 3, The
Modern Theory', Econometrica, XXXIV (Jan 1966) 18-76.
[5) Ethier, W., 'Input Substitution and the Concept of the Effective Rate of
Protection', Journal of Political Economy, LXXX (Jan-Feb 1972) 34-47.
[ 6) Guisinger, S. E., 'Negative Value Added and the Theory of Effective Protection',
Quarterly Journal of Economics, LXXXIII (Aug 1969) 415-33.
[7) Kemp, M. C., The Pure Theory of International Trade and Investment
(Englewood Cliffs, N.J.: Prentice-Hall 1969) chap. 7.
[8) McKinnon, R. 1., 'Intermediate Products and Differential Tariffs: A Generaliza-
tion of Lerner's Symmetry Theorem', Quarterly Journal of Economics, LXXX
(Nov 1966) 584-615.
[9) Melvin, J. R., 'Intermediate Goods, the Production Possibility Curve, and the
Gains from Trade', Quarterly Journal of Economics, LXXXIII (Feb 1969) 141-51.
[10) Vanek, J., 'Variable Factor Proportions and Inter-Industry Flows in the
Theory of International Trade', Quarterly Journal of Economics, LXXVII
(Feb 1963) 129-42.
[II) Yates, P. L., Forty Years of Foreign Trade (London: Allen & Unwin, 1959).
SUPPLEMENTARY READINGS
[12) Warne, R. D., 'Intermediate Goods in International Trade with Variable
Proportions and Two Primary Inputs', Quarterly Journal of Economics,
LXXXV (May 1971) 225-36.
8 Pure Intermediate Products
These and some other interesting points emerge vividly from the
subsequent discussion, which also highlights the need for a full-
fledged and separate analysis of pure intermediate products.
(8.2)
and
(8.3)
or
eLl cL2(k1 -k2)+ cMl cLMcL2(kM-k2)
+ CM2CLMCL1(kt -kM) ~ 0 (8.7*)
because (k 1 -kM) > (k 2 -kM), which in tum implies that the first
and the last two terms in (8.7*) have identical signs. From all this
discussion, the following lemma may be derived:
where ()ij is the gross relative share of the ith primary factor in the
jth final good. As in the previous chapter, the expressions for Rt
(i = L,K;j = 1,2) can be obtained by differentiating
cij = Cij(w,r,pM)
(i = L,K;j = 1,2)
and
-(a+ e) (w*-r*)
(}Ll
- (P+b) (w*-r*)
(}L2
• l_- -
RK (a+e)(.
( ) - w -r
*)
Kl
where
tx = PLtPKtulK + PiMPLtPMtulM+ Pf.MPKtPMtulM > 0
f3 = PL2PK2uf.K + PiMPL2PM2uf.M + PiMPK2PM2uiM > 0
e = PLMPKMPMtu'tK > 0
b = PLMPKMPM2u'tK > 0.
so that
p 1 X1 +P2X2 = w(L 1 +L 2)+r(K1 +K2)+pMM
= w(L 1 +L 2+LM)+r(K1 +K2+KM) = wL+rK.
Differentiating this totally,
(p 1dX1 +p 2dX2)+(X1dp 1 +X2dp 2) = Ldw+Kdr. (8.14)
However, from (8.1*) and (8.2*) and (8.12) and (8.13),
Ldw+Kdr = X 1 [RL 1 dw+RK 1 dr]+X2[RL 2dw+RK 2dr]
= X 1 dp 1 +X2dp 2.
where
YL = -(A.LIR!, +A.L 2R! 2)/(w*-r*) > 0
and YK = (A.KtR1, +A.KzRlz}/(w*-r*) > 0.
It is clear that Xt /(p!- p!) > 0 and Xi(P!- p!) < 0, because both
lA. I and 181 have the same sign as IRl:
RLIX! RLzXz
A.Ll A.L2
L L
IA.I = = x,x21RI
RKIX! RK2xz LK
AK! AK2
K K
RLlw RK 1 r
eLl 8K!
Pt Pt
and 181 RL2w
=~IRI.
P1P2
RKzr
8L2 8K2
P2 P2
All this suggests that the transformation curve is strictly concave to
the origin. t
t For further details concerning the shape and the properties of the transformation
curve in the presence of pure intermediate products, see Batra and Casas [I].
188 Studies in the Pure Theory of International Trade
• A.LlK* -A.K1L*
and X2 = jA.j
If factor intensities are defined in the gross sense, jA.j, which has the
same sign as jRj, has a definite sign; if they are specified in the net
sense, then again !A.! (or jRj) has an unambiguous sign only if the
net and the gross factor-intensity rankings are identical. It follows
then that the Rybczynski theorem is necessarily valid only if factor·
intensities are defined in the gross sense; if they are specified in the
net sense the theorem may not be valid. For example, with L* > 0
and K* = 0, Xi > 0 and X! < 0 ifjA.j > 0 or if the second industry
is capital-intensive relative to the first in the gross sense. However,
if X1 is labour-intensive relative to X 2 in the net sense, that is, if
k 1 < k 2 , but it is capital-intensive in the gross sense, so that
k~ > k;, the Rybczynski theorem does not hold, because then the
output of the apparently labour-intensive commodity will decline
as a result of a rise in the supply of labour alone when terms of trade
are kept constant.
In the similar fashion, the Stolper-Samuelson theorem may not be
valid if the gross and the net factor-intensity rankings are not
identical. This can be seen by solving (8.12) and (8.13) to obtain
and
Since again !O! (or jRj) may not have the same sign as (k 1 -k 2 ), the
Stolper-Samuelson theorem may not be valid in terms of the net or
the apparent factor-intensity rankings.
w* - r * = T*w· (8.18)
Solving (8.15) and (8.16), and using (8.18) and the expressions for
Rt presented in the preceding section, we obtain
X*
I
= [l + A.K2YL1..1.1·1+..1.L2YK]
01
T* > 0
and
From (8.26) it is clear that the relationship between wand pis unique
only if jej has an unambiguous sign. If jOj < 0, then with wh > wf
it is clear that
Ph> PJ
where Pi is the autarky-relative price of the second good in terms of
the first in the jth country (j = h,f). The home country will then
export X1 and import X 2 . On the other hand, if jej > 0, then with
(J)h > (J)f
Pure Intermediate Products 195
(8.27)
and
*_ BLip! -8L2p!
(8.28)
r - IBI
Similarly, from (8.13) and (8.11) and (8.27) and (8.28),
* * (8L2- PLM)p*
PM-P2 = IBI (8.29)
= rpL2PLM (kM-kl)
w
where, in obtaining (8.29) and (8.30), we have made use of the fact
that PLM+PKM = ()Ki+()Li = I (j = 1,2) and that IBI = ()Kz-{}KI
= ()LI -{}L 2 • If wh > w1 and k 2 > kM, then from the Heckscher-
Ohlin theorem we should expect the home country to export X 2 and
import M. For this pattern of trade to take place, we require that
* *
pz-PM < O
w*
Figure 8.1
t At constant prices, the increase in the total value of domestic output resulting
from the importation of EM units of the intermediate good is given by
ilX1
[- ax2]
+ p - EM.
oEM oEM
This can be easily seen to be equal to PMEMIP 1 in terms of the first commodity, and
hence to PMEM/p 2 in terms of the second.
Pure Intermediate Products 199
Figure 8.2
8.8 Appendix
With u{k denoting the partial elasticity of substitution between fac-
tors i and k in the jth sector, and pij the direct share of the ith factor
per dollar of the jth good, the differentiation of cij = Cij(w,r,pM)
yields
(A8.1)
Substituting for Ct 1 , CtM and CZ, 1 from (A8.9) and (A8.10) then
gives
(A8.12)
1
aLL=
1
-2-
(2 1 2 1
Px10"xx+ Px1PM10"xM+PM10"MM ·
2 1) (A8.16)
PLl
Since ah < 0, the quadratic form on the R.H.S. of (A8.16) must
be negative definite, implying that
akxa1M-(akM) 2
or
alx > -pM1alMakM
----~------~~
(pLlalM + Px1 akM)"
Substituting this in the expression for ex, we find that
or
PM1 (
1 +
1 1 )2
!X> ( 1 ) PKMPL10"LM-PLMPK1(1KM .
PLlO"LM Px10"KM
Since (pL 1alM+Px 1aicM) = -pM 1a1M > 0, it is readily seen that
ex > 0. Similarly, it can be shown that f3 > 0.
REFERENCES
[I] Batra, R.N., and Casas, F. R., 'Intermediate Products and the Pure Theory of
International Trade: A Neo-Heckscher-Ohlin Framework', American Economic
Review, LXIII (June 1973).
[2] Casas, F. R., 'Pure Intermediate Products, Factor Intensities and Technical
Progress in the Theory of International Trade', Southern Economic Journal,
XXXIX (July 1972).
SUPPLEMENTARY READINGS
[3] Batra, R. N., and Singh, R., 'The Intermediate Products and the Two-Sector
Growth Model', paper presented at the Econometric Society Meetings, Detroit
(Dec 1970).
[4] Khang, C., 'A Dynamic Model of Trade between the Final and the Intermediate
Products', Journal of Economic Theory, I (Dec 1969) 416-37.
[5] Kuo, C., 'A Two-Sector Growth Model with an Intermediate Product in an Open
Economy', Ph.D. dissertation, Department of Economics, Univ. of Western
Ontario (Apr 1972).
9 The Theory of Effective
Protection
t See, for example, Barber [2], Balassa [!], Basevi [3], Grube! and Lloyd [8] and
Travis [ 17] among countless others.
The Theory of Effective Protection 203
S'
lz
I
I
I
I
I
I
----,T
N I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
0 E G w F
Figure 9.1
t If all factors were available at constant factor prices, the supply curves for the
goods would be infinitely elastic. If one factor has an elasticity of supply less than
infinity, the supply curves for goods will have the identical elasticity only if that factor
is used in fixed proportions. If it is not, the rise in the unit cost would be less than the
rise in the factor price, and the supply curves for goods and the factor would possess
different elasticities.
The Theory of Effective Protection 207
where ()Q 2 = pQCQ 2 /p 2 is the share of the imported input in the total
cost of the imported final good, and
_ pz(I+tz)-pz _ Apz d _ PQ(I+tQ)-PQ _ ApQ
t2 - - - an tQ - - .
Pz Pz PQ PQ
If many imported inputs are used in the production of an importable
commodity, then the E.R.P. formula for thejth industry is given by
n
tj-
j; I
I eijti
ej = ---n--
1- I
j; I
eij
where ()ii is the share of the ith imported input in the total cost of the
jth good. It can easily be seen that this formula measures the pro-
portionate change in the value-added that occurs as the economy
switches from a free trade situation to one involving tariffs. Let v~
be the value-added per unit of output in the second industry when
tariffs are introduced. Then
and
so that
v~- v 2 t 2 - ()Q 2 tQ
e2 = -- = .
v2 I-{}Q2
The derivation of the E.R.P. formula from Fig. 9.1 serves to show
that the sign of the formula is linked with the direction of the change
208 Studies in the Pure Theory of International Trade
in the output. For example, if the input tariff, ta, is sufficiently large
and raises the supply curve beyond S' to an extent that increased the
unit production cost above the increase in p 2 , the E.R.P. would be
negative. But then the output of X 2 would also decline. Hence, under
the partial approach, there is a one-to-one relationship between the sign
of the E.R.P. and the direction of the shift in the output of the protected
commodity. A positive change in the output implies a positive E.R.P.,
and vice versa.
This conclusion continues to hold even when we allow substitution
among inputs, which tends to lower the increase in the unit produc-
tion cost because the possibility of variablity in production coeffi-
cients permits savings in costs by giving the producers the oppor-
tunity to economise on the use of inputs with increased prices.
Indeed, the higher are the substitution elasticities among inputs, the
smaller would be the upward shift in the unit cost. In Fig. 9.1, the
supply curve shifts only to the dotted line S" when substitution
among inputs is permitted. The output of X 2 now equals OW and
the net stimulus to output is given by AR > AC. Hence, other things
remaining equal, the substitution among inputs gives rise to a higher
E.R.P. Under these conditions
v2 = P2(1+t2)-[Ca2PaO+ta)+Pa~Ca2J
so that the E.R.P. formula is given by
, t 2 -8a 2(ta+q2)
e2 = (9.2)
l-8a 2
0 G
Figure 9.2
librium model where one of the two final goods is produced with the
help of two primary factors, K and L, and two intermediate products,
one of which is domestically produced but not traded, and the other
which is imported but not locally produced. The main difference
between this model and the production models used so far is that
in the presence of traded inputs the effective factor supplies become
variable, for the traded inputs can be imported (or exported) at
world prices which are given under our small country assumption
that we maintain throughout. This is true in spite of the constancy
of the supply of domestic primary factors which is generally assumed
in production models without traded inputs. Thus, in the present
model, the prices of some inputs are exogenously given but the
demand for them is endogenously determined, in direct contrast to
the models used up to the previous chapter, where factor supplies
were exogenously given but factor prices were dependent variables.
This difference is crucial to the existence of paradoxical results that
have appeared in some general equilibrium analyses attempting to
explore the consequences of the E. R.P. for resource allocation.
Unless otherwise specified, we assume throughout our analysis
that there are three domestically produced commodities: X 2 , the
importable good, X 1 , the exportable good, and M, the non-traded
intermediate product, which is produced solely to be used as an input
in the production of the importable good. A fourth commodity, Q,
is imported from abroad and used as an input again in X 2 but not in
X1 . The material inputs could also be utilised in X 1 , but this would
unnecessarily complicate the analysis without adding anything new
to the exposition. There is perfect competition in all markets, pro-
duction functions exhibit constant returns to scale and diminishing
marginal rates of substitution, and two primary factors, labour (L)
and capital (K), are inelastically supplied and fully employed. World
prices of all traded commodities are constant, but domestic prices
of the importables as well as the non-traded input may be altered
owing to the imposition of tariffs.
Let Cii represent the amount of the ith factor per unit of the jth
product (i = L,K,M,Q;j = 1,2,M). With full employment,
(9.3)
and
(9.4)
212 Studies in the Pure Theory of International Trade
As regards the price equations (9.5) and (9.6A), the factor intensities
are specified by the rows of the C matrix, even though it does not
contain CQ 2 ; the reason for this asymmetry is attributable to the
fact that, although CQ 2 enters into the determination of p 2 , it absorbs
no proportion of domestic primary factors, because Q is produced
abroad. Since it is the proportion of the domestic primary factors
used in the production of each commodity that determines its factor
intensity, the use of Q by X 2 causes no modification to the factor
intensities.
As in the previous chapter, the presence of the domestically pro-
duced intermediate good (M) makes it necessary to introduce a dis-
tinction between the net and the gross capital/labour ratio in the
second industry. For analytical convenience, we assume that X 2 is
capital-intensive relative to X 1 in both the gross and the net sense.
This implies that
where in this chapter A.ii equals the proportion of the ith factor used
directly in the production ofthejth good, and (}ii is the relative share
of the ith input in the jth good. Substituting (9.15) in (9.14) yields
and
= xlx21CI
KL
and
IOI = (}Ll (OK2 + (}KM(}M2)- (}Kl (OL2 + (}LM(}M2)
= ~ [CL1(CK2+CKMCM2)-CK1(CL2+CLMCM2)]
P1P2
=~ICI·
P1P2
The Theory of Effective Protection 215
Evidently, the signs of jA.j and jOj are the same as the sign of jej,
which implies that both IA.I and 101 are positive if X 2 is capital-
intensive relative to X 1 in the gross sense. The determinant of 0 can
be written in yet another manner which highlights the role played
by the imported input. Let t/IL 2 = OL2 + OLMOM 2 be the gross share
of labour and t/JK 2 = OK 2+0KMOM 2 be the gross share of capital in
the second industry. Since factor shares in any industry add up to
unity,
and
so that
jOj = OL1t/IK2-0K1t/1Lz = OLlt/JKz-(1-0Ll)(l-OQz-t/IK z)
= t/JKz-0K 1(1-0Q2) (9.17)
or
(9.18)
w* = -w
eKl < •
P2-
eQ2PQ*) = eKl
-we2
(l -
eQ2 ) (9.19)
and
r
* -- eLl < *
TOT e *) eLl (1 - eQ2 )·
P2- Q2PQ = TOf e2 (9.20)
,.,.
'--M2 -0
- L20'LMW
2 *+0 K20'KMr
2 *+0 M20'MMPM
2 *+0Q20'MQPQ
2.
C!M = 0LMui"Lw*+0KMui"Kr*
qM = OLMui"Kw* + OKMO'~Kr*
The Theory of Effective Protection 219
where the (}'s are as defined and u{k as usual is the partial elasticity of
substitution between factors i and kin thejth industry. Substituting
these in (9.11), we obtaint
A.LlXT +(A.L2 +A.LM).f!
= /h(w*-r*)+/Ja 1 (w*-p~)+/Ja 2 (r*-p~) (9.22)
where
PL = A.Ll(}Ktuh+A.L2(}K2ulK+A.LM(}KMufK
+ AL2 (}M2 lJi:MulM + ALM(}K2 (}LMO'i:M
Pat = A.L2(}Q2ula+A.LM(}LM(}Q2u'ita
and
Pa2 = A.LM(}KM(}Q2 u'ita·
Performing analogous substitutions in (9.12) yields
AKtXT + (A.K2 + A.KM).fl
= -PK(w*-r*)+/Ja 3 (w*-p~)+/Ja 4 (r*-p~) (9.23)
where
PK = A.Kl(}Lluh+A.K2(}L2ulM+A.KM(}LMufK
+ AKM(}KM(}L2 ulM + AK2 (}M2 lJfMui:M
Pa3 = A.KM(}LM(}Q2u'ita
and
Pa4 = A.K2 lJa2 ui:a + A.KM(}KM(}Q2 u'ita·
Subtracting (9.23) from (9.22) and performing some manipulations,
we derive
IA.I(X!-X!) = (PL +PK)(w*-r*)+A.L20Q2uia<w*-po)
+ A.K2(}Q2ui:a(p~- r*) + lJa2 (A.LM- A.KM )ui,.a(ptt- P~). (9.24)
From (9.19) and (9.20) we get
(9.25)
t As in the last two chapters, the expressions for c,7 can be simplified by noting
that, if all input-price ratios were constant, Cij = 0. This is what has been done in
obtaining the following equations.
220 Studies in the Pure Theory of International Trade
-8LtP!+(8Ll-l/JL2)pa
pa-r• = IOI (9.27)
Subtracting pa from both sides of (9.15) and using (9.26) and (9.27),
we obtain
* * ((JKM(]Ll -{JLM(]Kl)P!
PM-PQ = IOI
( ((JLM(]Kl- (JKM(]Ll)- ((JLM(JK2- (JKM(JL2)]pa
(9.28)
+ IOI
With this last equation we have collected all the ingredients necessary
to examine the implications of effective protection for the relative
outputs of the two final commodities. One of the basic differences
between the effective protection and the nominal protection model
is that, unlike the latter, there are several ways of conferring protec-
tion on a final commodity when tariffs on imported inputs are per-
mitted; for example, X 2 can be protected either by imposing an
output tariff, or by lowering the input tariff (or by subsidising the
imports of the input), or by any combination of the two such that
(p!- 8Q 2Pa) > 0. To begin with, we consider the two polar cases
which involve the imposition of the output or the input tariff. In a
model where only imported intermediate goods are allowed to exist,
Jones [11] has demonstrated that when protection is conferred
purely through output tariffs, the relative output of the protected
commodity must rise even when substitution is allowed between
primary factors and imported intermediate inputs; if protection is
provided through the other extreme way of lowering the input
tariff, the relative output of the protected good may or may not
rise, depending on whether or not the imported input is a better
substitute of one primary factor than of the other. We shall now
examine these results in the context of our model which allows for
the existence of two intermediate inputs, one non-traded and the
other imported from abroad.
Consider first the case where p~ is zero but p! > 0, so that pro-
tection is provided by the simplest way of the output tariff. Actually,
this case is no different from the conventional model of nominal
protection where intermediate inputs are ignored. First, we have to
The Theory of Effective Protection 221
ALM-AKM =; (k-kM) ~ 0 if k ~ kM
t There are two reasons why we make this assumption. First, it contributes to sim-
plicity, and second, we want to make sure that some paradoxical results derived below
do not depend on the special case where some factors are complementary to each other.
222 Studies in the Pure Theory of International Trade
to obtain
).A.)(X!-Xi)
2 2 e2(1-0a2)
= -[(fJL +fJK)+(AL20KtO'LQ+AK20LlO'KQ)OQ2] )0)
2 e 2(1-0a 2)
- [0Q2(ALM-AKM)(0Kl -OKM)aMQ] )0)
- OQ2 [(.A.L2ara- .A.K2aia) + (.A.LM- .A.KM)a~a]P~· (9.24A)
It can be seen that introducing t 2 alone can lead to the lowering of the
relative output of x2 if
t If A.KM > A.LM• then k > kM. If IlK, > IIKM• the share of capital in the first industry
exceeds that in the second industry, which implies that k 1 > kM, as can be seen from
the fact that
and
where w = wfr.
Thus, if k > k M > k 1 , both expressions in the parentheses on the right-hand side of
(9.29) are positive. The same result holds if k < k 1 < kM.
The Theory of Effective Protection 223
which, with p~ < 0 and (}L 1 > t/IL2 for 1e1 > 0, requires that
(}K 1 < t/JK 2 . If the non-traded good does not exist, t/IK 2 reduces to
(}K 2 . Hence the necessary condition for the 'perverse' possibility to
arise is that the relative share of capital in the first industry be less
than that in the second industry, which is, of course, within the
realm of possibility, because it does not, as we have shown above,
conflict with the positive sign of 1e1.t When a non-traded material
input is introduced, but a~Q = 0, then the necessary condition
involves the comparison between the gross share of capital t/IK2 in
the second industry using the intermediate goods with the share of
capital in the first industry which does not utilise any intermediate
good. This discussion leads to the following generalised theorem
which applies to Jones's result as well as ours:
Theorem 9.2. Given that a~Q = 0, the necessary condition for the
relative output of the protected final good to decline, when
effective protection is granted by subsidising imports of the
E"
(L/CL2)
...........
E
fL/Cf2J
(K/ChJ
---
G
(KjCK2)
8 F' F x1
I !
(L/C[1J (LjCL1J
Figure 9.3
The Theory of Effective Protection 225
with the input coefficients furnish the lines EF and GH which are
~rawn by joining the points that determine the maximum output of
one commodity, with zero output of the other. For example, OE =
LjeL2 is the maximum output of X 2(with zero output of X 1 )t and is
obtained by employing the entire supply of labour in X 2 • Along EF
labour would be fully employed, along GH capital would be, and at
their intersection at J there would be full employment of both factors.
The diagram is drawn in such a way as to reflect our assumption that
X 2 is capital-intensive relative to X 1 . This can be verified by com-
paring the slopes of EF and GH, which are, respectively, given by
eLI;eL2 and eK 1 jeK 2 , so that eK 2 IeL 2 > eK 1 jeLI. At the produc-
tion point given by J, the output of X 1 equals OB and that of X 2
equals JB, and the ratio between the two outputs, X 2 /X1 , is given
by the slope of OJ.
If there were no imported inputs, then eii would be related only
to w and r and a rise in r and a decline in w consequent upon the
grant of protection to X 2 would lead to a rise in eLl and eL2 and a
decline in eK 1 and eKz· This follows from the assumption of linearly
homogeneous production functions. Let the prime on an input
coefficient denote its magnitude in the new situation involving
tariffs. Then
e~i > eLi and eii < eKi U = 1,2).
With given supplies of K and L, all the four points E, F, G and H in
Fig. 9.3 would move along the axes either up or down. For example,
with ef. 2 > eL 2 , pointE would move down to, say, E'. Similarly,
G would shift toG', Fto F' and H to H', and the new full-employment
equilibrium would be established at J' where the dotted lines E' F'
and G' H', representing, respectively, the labour- and capital-
constraint lines in the new situation, intersect. It may be observed
that X 2 and the (X2 /X 1 ) ratio have risen and X 1 declined in the new
situation. This is how nominal protection in the usual two-good,
two-factor model raises the output of the protected commodity at the
expense of the output of the unprotected commodity.
Let us now suppose that an intermediate good (Q) is also imported
in addition to X 2 • If the effective protection is granted only by intro-
ducing a nominal tariff on X 2 , the conditions of the model do not
t Assuming for the time being that each good could be produced with the help of
only one factor.
226 Studies in the Pure Theory of International Trade
tAn examination of(9.24) confirms the view that (9.31) is the necessary condition
for the validity of this paradoxical result.
228 Studies in the Pure Theory of International Trade
H' x,
Figure 9.4
The Theory of Effective Protection 229
t The analysis of this chapter, like most recent general equilibrium studies that have
attempted to evaluate the E.R.P. concept in terms of its predictable effect on resource
allocation, is far from encouraging to the use of effective tariffs. The villain, as
suggested above, is not the non-traded input, not even the concept of the effective
tariff, but the imported input whose availability at constant world prices makes the
effective factor supply variable when it is substitutable with the domestic inputs.
Jones [11], Ramaswami and Srinivasan [14] and Corden [6] have provided some
basis for salvaging the E.R.P. concept by demonstrating the necessity of biased
substitution effects for the existence of the paradoxical resource-allocational implica-
tions. Our analysis, however, destroys even that little hope, for here, even if substitu-
tion effects are unbiased, the relative output of the effectively protected commodity
may still decline. Witness here the fact that in (9.29) even if
so that there are no biased substitution effects, inequality (9.29) could be satisfied to
produce the paradox.
The Theory of Effective Protection 231
whence
(9.33)
or
dX 1 (p 2 -pQdQ/dX2 )
(9.34)
dX 2 = Pt
which shows that theM RT in the present model equals the ratio between
the marginal value-added in each final commodity. t Coming now to
the demonstration of the optimality of free trade, let us consider the
following system of equations:
U= U(D 1 ,D 2)
E 1 = (P2/P 1 )E2 + (PQIP 1 )Q (under balanced trade)
D 1 = X 1 -£ 1 = X 1 - [(P2/P 1 )E2 +(pQ/pdQ]
and
D2 = X2+E2.
Differentiating the utility function totally with respect to the E.R.P.
(e 2 ), we obtain
(9.35)
This means that the slope of the transformation curve HH' at any
point in Fig. 9.5 reflects not (P2/P 1 ) but v 2 /p 1 . Suppose Pis the free
trade production point; then AB, whose slope equals v2 /p 1 , may be
conceived as the budget line facing the producers. The equilibrium
on the consumption side, however, is reached when the MRS =
(p 2 /p 1) > (v 2 /p 1). Since X 1 does not use the imported input, the
Figure 9.5
234 Studies in the Pure Theory of International Trade
dV
de 2
= VI [tQ PQ dQ + t2P2 {dD2-
p 1 de 2 p 1 de 2 de 2
dX2}]
or
(9.36)
t Since X 2 uses the imported input for which the foreign country must be paid, the
value of consumption in terms of X2 must be less than the value of its production.
t This can be seen by drawing a non-intersecting social indifference curve (not
drawn) touching the dotted lineAE.
The Theory of Effective Protection 235
or
dY(l- T2mh) = - pdX2 (1- OQ 2 )e 2 + 1QpQX2 dCQ 2 /p 1
Since dX2 /dt 2 > 0 and oD 2 jop(l + t 2 ) < 0, the optimal effective
tariff is negative.
The economic explanation of these results is quite simple. As
shown in Chapter 4, the introduction of the output tariff gives rise
to a distortion on both the production side and the consumption side
and thus to the production and the consumtpion loss. However,
the production loss can be mitigated by introducing an input tariff
which tends to lower the output of the importable goods and take
it closer to the free trade output level. The production loss is com-
pletely eliminated if the tariff on the input is equated to t 2 /0Q 2 , so
that e2 = 0, provided CQ 2 is constant. The reason lies in the fact that
when the imported input is used in fixed proportions, the E.R.P. and
the output of X 2 are positively related, so that when e 2 = 0, the
output of X 2 is the same as what it would have been under free trade,
with the result that the production loss resulting from the tariffs
238 Studies in the Pure Theory of International Trade
REFERENCES
[I] Balas sa, B., 'Tariff Protection in Industrial Countries: An Evaluation', Journal
of Political Economy, LXXIII (Dec 1965) 572-94.
[2] Barber, C. L., 'Canadian Tariff Policy', Canadian Journal of Economics and
Political Science, XXI (Nov 1955) 513-30.
[3] Basevi, G., 'The U.S. Tariff Structure: Estimate of Effective Rates of Protection
of U.S. Industries and Industrial Labour', Review of Economics and Statistics,
XLVIII (May 1966) 147-60.
[4] Casas, F. R., 'Optimal Effective Protection in General Equilibrium', American
Economic Review, LXIII (Sep 1973).
[5] Corden, W. M., 'The Structure of a Tariff System and the Effective Protective
Rate', Journal of Political Economy, LXXIV (June 1966) 221-37.
[6] --,'The Substitution Problem in the Theory of Effective Protection', Journal
of International Economics, I (Feb 1971) 37-58.
[7] Ellsworth, P. T., 'Import Substitution in Pakistan: Some Comments', Pakistan
Development Review, VI (autumn 1966) 395-407.
[8] Grube!, H. G., and Lloyd, P. J. 'Factor Substitution and Effective Tariff Rates',
Review of Economic Studies, XXXVIII (Jan 1971) 95-104.
[9] Guisinger, S. E., 'Negative Value Added and the Theory of Effective Protection',
Quarterly Journal of Economics, LXXXIII (Aug 1969) 415-33.
[10] Johnson, H. G., 'The Theory of Tariff Structure with Special Reference to World
Trade and Development', in Trade and Development (Geneva: Institut Univer-
sitaire des Hautes Etudes Internationales, 1965).
[II] Jones, R. W., 'Effective Protection and Substitution', Journal of International
Economics, 1 (Feb 1971) 59-82.
[12] Leith, J. C., 'Substitution and Supply Elasticities in Calculating the Effective
Protection Rate', Quarterly Journal of Economics, LXXXII (Nov 1968) 588-601.
[13] --,'The Effects of Tariffs on Production, Consumption and Trade: A Revised
Analysis', American Economic Review, LXI (Mar 1971) 74-81.
The Theory of Effective Protection 239
(14] Ramaswami, V. K., and Srinivasan, T. N., 'Tariff Structure and Resource Allo-
cation in the Presence of Factor Substitution', in J. N. Bhagwati eta/. (eds.),
Trade, Balance of Payments, and Growth (Amsterdam: North-Holland, 1971),
[15] Ruffin, R. J., 'Tariffs, Intermediate Goods, and Domestic Protection', American
Economic Review, LIX (June 1969) 261-9.
[16] - - , 'The Welfare Implications of Effective Protection', in H. G. Grube! and
H. G. Johnson (eds.), Effective Tariff Protection (Geneva: General Agreement
on Tariffs and Trade and Graduate Institute of International Studies, 1971)
chap. 5.
[17] Travis, W. P., 'The Effective Rate of Protection and the Question of Labour
Protection in the United States', Journal of Political Economy, LXXVI (May/June
1968) 443-81.
SUPPLEMENTARY READINGS
(18] Anderson, J., and Naya, S., 'Substitution and Two Concepts of Effective Pro-
tection', American Economic Review, LIX (June 1969) 607-11.
[19] Corden, W. M., The Theory of Protection (Oxford: Clarendon Press, 1971).
[20] Ethier, W., 'Input Substitution and the Concept of the Effective Rate of Pro-
tection', Journal of Political Economy, LXXX (Jan/Feb 1972) 34-47.
10 Factor Market Imperfections
earn the same reward in both industries, but that the wage rates
differ. The different rates of return in the two industries can be
analysed without any difficulty. The crucial factor that prompts this
analysis concerns the differential between factor-price ratios, irres-
pective of whether or not both factor prices differ in the two indus-
tries. The rest of the assumptions are the same as those made at the
beginning of Chapter 2. Thus product markets are still perfect, pro-
duction functions exhibit constant returns to scale and diminishing
returns to factor proportions, factors are perfectly mobile and factor
prices perfectly flexible, and so on.
and
(10.12)
The next step involves the introduction of the elasticity of sub-
stitution in each sector, so that
(). =
q.-cr.
J J (} = 1,2).
1 wt-r*
C! 1 = -OK 1a 1(w!-r*)
C11 = eL1()1(wf-r*)
C!2 = - [OK2u2(w!- r*) + eK2()2~·]
and
and
(10.15)
and
I() I = -w -r · CLl CL2(k
1
2- rxk 1 ) (10.16)
P1Pz
where ki = Ki/ Li is the capital/labour ratio in the jth sector. At the
same time
(10.17)
and
(10.18)
The main feature of the undistorted two-sector model is that IA.I
and 101 have the same sign, for then rx = I. Now the sign of IA.I
indicates the ranking of industries in terms of the factor intensities
in the 'physical' sense, for here the comparison between the employ-
ment of the physical units of the factors in the two sectors is
involved; the sign of I() I on the other hand reflects the ranking of
industries in terms of factor intensities in the 'value' sense, for here,
as is evident from (I 0.18), we compare the relative share of any
factor in the two sectors in determining whether or not an industry
is capital- or labour-intensive relative to the other. This may also
be seen directly by using (10.5), so that from (10.16)
in the first industry exceeds that in the second. This, however, may
no longer be true in the distorted model. For example, if tx > 1, the
sign of (k 2 -txk 1 ) and hence of 101 may be opposite to the sign of
(k 2 -k 1) and IA.I when X 2 is capital-intensive relative to X 1 in the
physical sense. From (10.5), tx > 1 implies that w 1 < w2 , that is,
the differential is paid by the second commodity. With labour being
paid a premium in the second industry, it is not difficult to visualise
a situation where the relative share of labour in the second industry
may exceed that in the first, so that 101 < 0, even though X 2 employs
a greater proportion of capital than labour, i.e. A.K 2 > A.L 2 •
What if tx < 1, that is, w 1 > w2 • Here lA. I and 101 must possess the
same sign when k 2 > k 1 . The premium to labour is now paid by the
first industry which is also labour-intensive in the physical sense.
From this we conclude that if an industry pays the differential on
its intensive factor (in the physical sense), IA.I and 101 have the same
sign, that is, the factor intensities in the physical sense cannot get
reversed in the value sense.
The picture is further blurred by the fact that, as first demonstrated
by Johnson [I OJ, the physical factor-intensity relationships may also
change sign in the presence of the differential, even though in its
absence they are non-reversible. This point can be illustrated with
the help of the box diagram presented in Fig. 10.1, where X 1 in the
absence of distortions is shown to be labour-intensive relative to x2
Figure 10.1
Factor Market Imperfections 245
101 will again have the same sign. In the diagram, for example, k 2
becomes less than k 1 for IX < a: 0 ; then k 2 may also be less than a:k 1 •
Thus factor intensities may get reversed in both the physical and
the value sense.
The distinction introduced above between factor intensities in the
value and the physical sense is very important. It is clear from (10.6)
and (10.7) that the relationship among the physical variables of the
system, such as commodity outputs, factor endowments, etc., is
influenced by the factor intensities in the physical sense, i.e. by the
sign of IA.I, whereas the relationship among the financial variables,
like factor and commodity prices, is regulated by the factor inten-
sities in the value sense, i.e. by the sign of 101. If IA.I and 101 have
opposite signs, the usual positive link between financial and physical
variables established in the previous chapters will be reversed.
(10.20)
and
• •
w* -r* = P1 -p2 (10.21)
1 IOI
remembering that, with a:* = 0, wt = w!.
As expected, 101 determines the relationship between commodity
and factor prices. If lA. Iand 101 are both positive, then it is clear that
a rise in the relative price of the first commodity raises the relative
and the real reward of its intensive factor, labour, in both industries
(because wt = w!) and lowers the relative and the real reward of the
other factor, capital. The Stolper-Samuelson theorem continues to
hold in the presence of the given wage differential. However, if
101 < 0 when IA.I > 0, the Stolper-Samuelson theorem is no longer
valid in the presence of the factor market distortion. The reason
once again lies in the fact that the sign of lA. I has nothing to do with
Factor Market Imperfections 247
v 1dL 1 +z 1dK 1
v 2 dL 2 +z 2 dK 2
p[(v 2 dL 2 /cx) + z 2 dL 2 ]
-f3p (10.22)
v 2 dL 2 +z 2 dL 2
where
248 Studies in the Pure Theory of International Trade
x,
Figure 10.2a
0 H'
Figure 10.2b
Factor Market Imperfections 249
(10.24)
physical sense rises and that of the other industry falls. With
k 2 > k 1 , and jA.j > 0, the rise in the output occurs in the second
commodity at the expense of the output of the first, even though to
begin with it was the relative price of the first commodity that in-
creased. In other words, if jA.j and jej have dissimilar signs, the
price-output response is negative or 'abnormal', and the long-run
supply curves are negatively sloped.
(10.25)
k! < k!.
k! = Ch -q, = u 1(w!-r*)
and
If the elasticities of substitution are the same in the two sectors, factor intensities
can never get reversed; if they are constant, there can at most be one reversal; if they
are variable, there may be more than one reversal points like R; finally, the necessary
condition for this reversal is that u 2 < u 1 . On the other hand, the precise con-
ditions for the reversal with a given w 1 can be obtained by differentiating k 2 jak 1 to
obtain
k!-a*-k! = a*(u 2 -l).
Thus fork 2 /ak 1 to decline, it is necessary that u 2 + I. For further discussion of these
points, see Herberg et a/. [9].
252 Studies in the Pure Theory of International Trade
Figure 10.3
t What this really means is that the sign of jej is reversed after point Rand so is
the relation between w 1 and pat a given :x.
! It is worth noting that the discussion of the Heckscher-Ohlin theorem illustrated
b) Fig. 10.3 is identical with the corresponding discussion in Chapter 3 in terms of
Fig. 3.6. The only difference is that the factor-intensity reversal in the earlier chapter
was caused by the multiple intersections of the unit isoquants of the two com-
modities. whereas now it arises from the existence of the differential.
Factor Market Imperfections 253
If in Fig. 10.3 each country's w 1 lies on the same side of R', as is the
case given by the pair Ow 1h and Ow 11 ,
Ph< PJ
so that the home country will export the capital-intensive com-
modity X 2 and import the labour-intensive commodity X 1 • The
Heckscher-Ohlin theorem continues to hold. However, if each
country's w 1 lies on different sides of R', as is described by the pair
Ow'1 h and Ow 11 ,
p~ > PJ
so that the home country will export X 1 and import X 2 , thereby
contradicting the Heckscher-Ohlin theorem. But this result is by no
means necessary even if each country's w 1 lies on different sides of
R'. This can be visualised if the home country's w 1 was to be slightly
below Ow'lh but above R'.
Coming now to the physical definition of'relativefactor abundance,
the Heckscher-Ohlin theorem cannot be disproved if IX is the same in
the two countries. This is because the Rybczynski theorem, as will be
shown later, continues to hold in spite of the differential. Whether
or not the factor intensities get reversed in the value sense makes
little difference to the discussion of the Heckscher-Ohlin theorem in
terms of the physical definition, for, as stated before, the definition
of factor intensities in terms of distributive shares does not control
the relationship between commodity outputs and factor endow-
ments. Of course, if rx is not the same in the two countries, or if one
country has no distortions in factor markets, then the factor inten-
sities in the value sense will also play a role in determining the vali-
dity of the theorem in question. As we show in the subsequent dis-
cussion, the level of output in each commodity at given commodity
prices becomes a function not only of the country's factor endow-
ments but also of IX, and the effect of the latter on the output levels
is determined by the signs of both I)-I and 101. This is how the factor
intensities in the value sense participate in determining whether the
theorem under question continues to be valid in the presence of the
differential. If IX is the same in the two countries, the effect of the
distortion on the two outputs is also the same, which means that
the sign of 101 is of no consequence to the validity of the theorem.
Thus we conclude that the Heckscher-Ohlin theorem continues
to be valid in terms of the physical definition of relative factor
254 Studies in the Pure Theory of International Trade
k* _ C* -C* _ a 1 [(pj-p!)+8L21X*]
1 - Kl Ll - IOI (10.30)
and
k * _ C* -C* _ Uz[(p'j-p!)+OL11X*]
2 - K2 L2 - IBI (10.31)
(10.32)
and
()'Llf3K+ )..Klf3L)(p'j- p!) + BIX*
X!= (10.33)
IA.I·IBI
Factor Market Imperfections 255
where
and
B = eLlu2(A.xlA.L2eK2 +A.LlA.K2eL2)+0L2A.LlA.KlO"l > 0.
The conclusions that emerge from (10.26)-(10.33) are categorical
for a small country facing given terms of trade, so that (pT- p!) = 0.
The implications of a rise in tx for real factor rewards, capital/labour
ratios and the two outputs at constant commodity prices are the
same as those of a rise in the relative price of the first commodity
at the constant wage differential. The reasoning is straightforward.
A rise in pifp 2 alone directly serves to benefit the first industry
relative to the second. A rise in tx alone tends to move the differential
against the second industry and thus indirectly benefits the first. In
both cases the relative gain accrues to the first industry; hence a
rise in tx or pifp 2 alone generates similar repercussions in the
economy.
In our discussion in section 10.2 on factor intensities, we showed
that the presence of the differential may cause a reversal in the factor
intensities in both the physical as well as the value sense. Following
Magee [15], we are now in a position to show that factor intensities
in the physical sense can never get reversed if either tx or p( = p 2 /p 1 )
alone gets altered. Consider Fig. 10.4, which retains some features
of Fig. 10.1. Let us begin with the case where tx = 1 and the efficiency
locus is given by 0 1 A0 2 , A being the actual production point. A
decline in tx serves to move the efficiency locus towards the diagonal
and, for some tx < tx 0 < 1, moves it to the opposite side (see Fig.
10.1). However, from (10.30) and (10.31), a decline in tx, with 101 > 0,
leads to a decline in the capital/labour ratio in both industries if the
commodity prices are constant. All this implies that a fall in tx will
move the production point to somewhere above 0 1 A0 2 , below 0 1 A
(so that k 1 declines) and above 0 2 A (so that k 2 declines). On the
contrary, the rise in tx will move the production point to somewhere
below the undistorted locus 0 1 A0 2 , below 0 2 A and above 0 1 A.
These considerations suggest that the distorted efficiency locus is
bounded from above by the two rays 0 1A and 0 2 A when com-
modity prices are unaltered. The dashed curve 0 1 CAB0 2 in Fig.
10.4 is one such curve where the new production point is given by
either B or C, depending on whether tx* ~ 0. It is evident from the
256 Studies in the Pure Theory of International Trade
Figure 10.4
diagram that the distorted contract curve cannot cross the diagonal
if the terms of trade are constant, whatever the level of a below
unity. In the limit there will be complete specialisation in X 2 , but
factor intensities in the physical sense will never get reversed. This
result derives directly from the fact that when commodity prices
are fixed, a change in a moves the capital/labour ratios in the same
direction. For this reason, the sign of 101 is unimportant in this
matter. For the same reason alone, if a is given, a change in com-
modity prices, which also moves the capital/labour ratios in the
same direction, can never cause a reversal in the physical factor
intensities. However, if commodity prices and a are changing simul-
taneously, k 1 and k 2 may move in the opposite direction, in which
case it is possible for the contract curve to move to the opposite
side of the diagonal. tIt is worth pointing out here that factor inten-
sities may still get reversed in the value sense, even if only one of a
and p is altered.
The implications of a shift in a for the real factor rewards have
important and interesting policy prescriptions for trade unions. To
t From (10.30) and (10.31) it is clear that k 1 and k 2 move in the opposite direction
only if (8L2a•- p*) and (8L 1a*- p*) possess conflicting signs. Evidently, the necessary
condition to ensure this result is that
Factor Market Imperfections 257
be specific, assume that the second sector is the unionised sector and
tx > 1 so that w 1 < w 2 • In other words, suppose that the producers
in the unionised sector pay a premium to labour over the labour
force employed in the non-union, first sector. Suppose further that
the union in the second sector seeks to secure an increase in this
premium, thereby increasing the value of tx. When the repercussions
on the rest of the economy are taken into account, however, the
union may end up with a lower real wage rate for its members. As
(1 0.26) and (I 0.27) suggest, with (p!- p!) = 0, the real wage rate
will decline in both sectors if I01 is negative, that is, if the unionised
sector pays a higher share to labour than the non-union sector and
hence is labour-intensive in the value sense. The reason should be
obvious from (10.30) and (10.31). With 101 < 0, a rise in tx induces
substitution of labour for capital in both sectors; resulting in a
decline in k 1 and k 2 which in turn leads to a general decline in the
wage rate and a rise in the return to capital. The policy prescription
is then unmistakable. If the union discovers that the non-union sector
pays labour a lower relative share, it should seek a reduction in the
inter-industry wage differential. The union policy could prove self-
defeating if it always pressed for higher wages.
The change in the wage differential affects not only the factor
rewards, but has repercussions for the two outputs as well. Here
again the identity of the signs of JA.I and 101 plays the crucial role.
Specifically, a rise in tx raises the output of X 1 and Jowers that of X 2
at constant commodity prices if 1.?.1 and 101 possess the same sign.
The converse is true if 1..1.1 and 101 have opposite signs.
I dU dY dX 1 dX 2 dp
- - = - = - + p - - E2 - .
ul dtx dtx dtx dtx dtx
258 Studies in the Pure Theory of International Trade
From (I 0.22),
axl _ f3p aX2 and aX 1 -f3p a;z.
aa. aa. ap
Using these,
dY
da.
= P ax2 (I-/3)+ dp
aa. da.
[P ax2 (I-{3)-Ez]·
ap
(10.34)
x,
~<1.0(>1
Figure 10.5
or
whence
dp P o£2/orx.
(10.35)
drx. E 2(a1 +ah-1)
where, as before, a1 and ah are, respectively, the foreign and home
elasticities of demand for imports. For stability, a1 +ah > I. Now
so that
o£2 oD2 ay ax2 ay oX2
p - = p - · - - p - = mh--p-
orx. oy arx. arx. arx. arx.
ax2
= p orx. [ mh(I - p)- 1]
and a shift in the same direction, and thus each shift strengthens the
ramifications of the other.
The implications of a change in the differential for welfare deserve
special consideration when terms of trade are variable, for there the
inspection of (1 0.34) makes it clear that even if the factor intensities
in the physical and the value sense are divergent, so that (dp/da) < 0,
the change in the differential does not give rise to unambiguous
shifts in community welfare, because of the ambiguity involved in
the sign of [p(aX 2 /ap)(l- {:1)- £ 2 ]. This is rather surprising because
the latter expression displays the effect of a change in the terms of
trade on welfare, which effect was demonstrated to be unambiguous
in Chapter 5. However, this premise, a detailed discussion of which
is postponed to the next section, turns out to be invalid in the
presence of distortions. The full impact of a change in the differential
on welfare can be divided into two components: (1) the effect of a
change in a at constant terms of trade, plus (2) the effect of a change
in p at a given a weighted by the magnitude of (dp/da). These com-
ponents may move in the same or the opposite direction, so that the
sign of (dY/da) is uncertain. To obtain a better perspective, let us
rewrite (10.34) as
0
Figure 10.6
utility function totally with respect to the rate of tariff. The equations
involved in this procedure are U = U(D 1 D 2 ), D 1 = X 1 -E 1 ,
D 2 = X 2+E2,pE2 = E 1 ,ph =p(I+t 2)and(dXtfdX2) = -fJp(1+
t 2 ). Using these equations and following the procedure used many
times before, we obtain
0 H'
Figure !0.7a
Figure 10.7b
266 Studies in the Pure Theory of International Trade
level, U 2 , lies above the free trade level of welfare given by U1 • Thus,
we conclude that if the price-output response is positive, free trade
may be inferior to no trade, provided the differential is paid by the
import-competing industry.
Needless to say, the necessary condition for free trade to be inferior
to no trade in the presence of the differential and the negative price-
output response is that the differential is paid by the producers of
the exportable good.
These results can be explained in terms of the familiar concepts of
production and consumption losses (or gains) arising from the inter-
vention in the free inter-country flow of goods and services. t As the
rate of tariff is increased, the consumer has to pay increasingly high
prices for the importables, and this leads to a consumption loss
irrespective of the presence or the absence of the wage differential.
The consumption loss is maximised when the tariff wall is high
enough to be prohibitive. In the case of production there are two
possibilities, depending on whether the differential is paid by the
producers of X 1 or X 2 • When the differential is absent, the intro-
duction of the tariff gives rise to a decline in the output of the export-
able good and hence to a decline in the country's specialisation or
to a production loss, given that the price-output response is positive.
In general, the output structure is biased against the industry which
pays the differential, for at the prevailing market prices the output
of that industry would be higher if factor markets were undistorted.
Therefore, if the output of the industry suffering from the differential
rises consequent upon a policy shift, the economic inefficiency de-
clines in spite of the constancy of the wage differential, because the
output structure moves closer to what that structure would be in the
undistorted economy. This loss of inefficiency, then, constitutes a
production gain. The converse is true when the output of the industry
paying the differential declines. On balance, the policy switch from
free trade to no trade may give rise to the production gain or loss,
depending on which industry pays the differential. In terms of Fig.
I 0. 7A, the shift from free trade to no trade generates a production
loss as the output of X 1 , the industry paying the differential, declines,
and this loss adds to the consumption loss, so that free trade is un-
ambiguously superior to no trade. In Fig. l0.7B, on the contrary,
such a policy shift results in a production gain, and if this gain out-
Figure 10.8
x,
FP
FP 1
Figure 10.9
270 Studies in the Pure Theory of International Trade
_I dU = dY = oX 1
U1 dG dG oG
oX2 [ oX2 (l-p)-E
+ P oG + P op 2
JdGdp (10.40)
Figure IO.IOa
H 1 G'
Figure IO.IOb
272 Studies in the Pure Theory of International Trade
and
(10.42)
which are, of course, exactly identical with (2.37) and (2.38) obtained
in Chapter 2, which shows that the Rybczynski theorem continues
to be valid in the presence of the wage differential. From these
equations,
so that
oY oX1 oX2 X1K2 -pX2K1
- = - +p - = --=-=--=-=--~----=.--__,..: .
oL oL oL L1L2(k2 -kl)'
growth occurs owing to the rise in the supply of capital alone, then,
from (10.41) and (10.42),
so that
oY pL1X2 -L2X1
oK L1L2(k2 -a.kl)"
Now, from (10.3)-(10.5),
r pL 1X 2 -a.L 2X 1
P1 L1L2(k2-a.k1)
whence
oY r (k 2 -a.k 1) (pL 1 X 2 -L 2X 1)
oK = P1. (k2 -kl) (pL1X2 -a.L2Xd
which makes it clear that (o Y/oK) will be negative even if the
physical and the value factor intensities are identical, provided
(pL1X2-L 2X 1 ) and (pL 1X 2 -a.L 2X 1)"have opposite signs. Thus
the conclusion is unmistakable. For growth to be immiserising with
unchanged terms of trade, the reversal of the physical and the value
factor intensities is necessary only if output expansion occurs because
of an increase in the supply of the factor that earns a premium in one
sector.
Having demonstrated that growth may be immiserising in the
presence of distorted factor markets, even when terms of trade are
constant, a simple inspection of (1 0.40) reveals that this immiserisa-
tion may be reinforced as a result of the growth-induced improve-
ment in the terms of trade, or that if(oYjoG) was positive, (dY/dG)
may be negative in spite of, or rather because of, the terms of trade
improvement fostered by growth and so on. t
t For further discussion of this resu1t, see Batra and Scully [2].
274 Studies in the Pure Theory of International Trade
of this thesis argue that the cause of poverty in the developing nations
lies in the deterioration in their terms of trade which started in the
latter part of the nineteenth century and persisted until the beginning
of the Second World War. Participants on the other side of the
debate have strongly challenged the empirical and analytical sound-
ness of this hypothesis.
We have no intention of getting involved in this still mooted issue.
But participants on both sides of the debate have displayed at least
tacit consensus on the fact that the deterioration in the terms of
trade by itself is a bad thing, that it leads to a decline in the rate of
growth of income per se. However, we have shown that such a de-
terioration is no longer directly related to the rate of growth if factor
markets are imperfect. To the extent that factor markets in the
underdeveloped countries are very likely to be imperfect, it is neces-
sary first to decide whether or not a deterioration in the terms of
trade is necessarily harmful before arguing for it or contesting it as a
sufficient explanation for ubiquitous poverty in the underdeveloped
world.
Figure I 0.11
switch from no trade to free trade at the given world prices reflected
by the slope of FP takes the consumpti on point to C, but under the
postulated conditions leaves the production point unchanged at S.
However, the immobility of factors does not interfere with the
superiority of free trade over no trade as U 2 lies above U1 • This
occurs because, although production gain is zero, the consumpti on
gain remains.
Consider now the case where factor immobility is accompan ied
by factor-pric e rigidity; the switch from no trade to free trade would
cause unemploy ment of some factors employed in X 2 as a result of
the decline in its relative price from that given by the slope of DP to
one furnished by the slope of FP. As a consequen ce the output of X 2
will decline, although that of X 1 will remain unchanged as unutilised
factors cannot move from X 2 to X 1 . In other words, the consumpti on
gain in the presence of factor-pric e inflexibility will be accompan ied
by the production loss, and if the latter is large enough, free trade
may be inferior to no trade. This is the possibility depicted in Fig.
10.11, where free trade takes production toP, consumpti on to C'
and welfare to U0 which lies below the autarky level, U1 • It is this
Factor Market Imperfections 277
x,
8
Figure 10.12
278 Studies in the Pure Theory of International Trade
Initial Unemployment
Both Haberler and Johnson assume full employment in the absence
of trade in spite of inflexibility of factor prices, and unemployment
of a fraction of given factor supplies is caused in their framework
only by the introduction df trade. But why should not the rigidity
of factor prices result in some unemployment even in the absence of
trade, or for that matter in the initial situation of free trade? The
analysis of rigid factor prices will not be complete unless this con-
sideration is taken into account.
In the presence of unemployment, the initial production point
will lie inside the transformation curve. For the sake of brevity, we
consider only the case involving a shift from free trade to no trade,
because in the other case of moving from no trade to free trade the
results are qualitatively unchanged. Consider Fig. I 0. I 3, where P'
is the production point that would have prevailed if there was full
employment at free trade prices; the actual production point, how-
ever, is given by P, consumption is at C and welfare under free trade
at U 2 • The consumption loss resulting from the imposition of a
prohibitive tariff leads to the level of welfare given by ul, with p
now being the production as well as the consumption point. At the
no-trade price ratio given by DPthe output of X 1 will decline, so that
there will be an increase in unemployment of factors specific to X 1 ,
whereas the output of X 2 will rise, which it can because of the exist-
ence of a pool of unemployed factors specific to it. As a consequence,
the new production point will lie in the rectangle BPRT. Evidently,
then, the level of no-trade welfare may lie above or below the free
trade level or even remain the same, depending on the final produc-
tion point. Fig. 10.13 pictures the case where the self-sufficiency
equilibrium is at Sand the autarky level of welfare, given by U3 ,
lies above the free trade level U 2 . It is here that the conclusion is
substantially different from the case where there is full employment
initially.
Factor Market Imperfections 279
x,
r-------------------------------~pt
DP 1
0 8 T
Figure 10.13
REFERENCES
[I] Batra, R. N., and Pattanaik, P. K., 'Factor Market Imperfections and Gains
from Trade', Oxford Economic Papers, XXIII (July 1971) 182-8.
[2] ---,and Scully, G. W., 'The Theory of Wage Differentials: Welfare and Im-
miserizing Growth', Journal of International Economics, 1 (May 1971) 241-7.
[3] Bhagwati, J. N., 'Distortions and Immiserizing Growth: A Generalization',
Review of Economic Studies, XXXV (Oct 1968) 481-5.
280 Studies in the Pure Theory of International Trade
[ 4] Bhagwati, J. N., and Ramaswami, V. K., 'Domestic Distortions, Tariffs and the
Theory of Optimum Subsidy', Journal of Political Economy, LXXI (Feb 1963)
44-50.
l5J --,and Srinivasan, T. N., 'The Theory of Wage Differentials: Production
Response and Factor Price Equalization,' Journal of International Economics,
I (Feb 1971) 19-35.
[6] Haberler, G., 'Some problems in the Pure Theory of International Trade',
Economic Journal, LX (June 1950) 223-40.
[7] Harberger, A. C., 'The Incidence of the Corporation Income Tax', Journal of
Political Economy, LXX (June 1962) 215-40.
[8] Herberg, H., and Kemp, M. C., 'Factor Market Distortions, the Shape of the
Locus of Competitive Outputs, and the Relation between Product Prices and
Equilibrium Outputs', in J. N. Bhagwati eta/ (eds.), Trade, Balance of Pay-
ments, and Growth (Amsterdam: North-Holland, 1971 ).
[9] - - , - - , and Magee, S. P., 'Factor Market Distortions, the Reversal of
Relative Factor Intensities, and the Relation between Product Prices and Equi-
librium Outputs', mimeographed (1970).
[10] Johnson, H. G., 'Factor Market Distortions and the Shape of the Transforma-
tion Curve', Econometrica, XXXIV (July 1966) 686-98.
[II]--, 'Optimal Trade Intervention in the Presence of Domestic Distortions',
in Trade, Growth and the Balance of Payments (Chicago: Rand McNally, 1965).
[12] --,'A Note on Distortions and the Rate of Growth of an Open Economy',
Economic Journal, LXXX (Dec 1970).
[13] Jones, R. W., 'The Structure of Simple General Equilibrium Models', Journal
of Political Economy, LXXII (Dec 1965) 557-72.
[14] Lewis, W. A., 'Economic Development with Unlimited Supplies of Labour',
Manchester School of Economics and Social Studies, XXII (May 1954) 139-91.
[15] Magee, S. P., 'Factor Market Distortions, Production, Distribution and the
Pure Theory of International Trade', Quarterly Journal of Economics, LXXXV
(Nov 1971) 623-43
[16] United Nations, Department of Economic Affairs, Relative Prices of Exports
and Imports of Underdeveloped Countries (1949).
SUPPLEMENTARY READINGS
[17] Batra, R. N., and Pattanaik, P. K., 'Domestic Distortions and the Gains from
Trade', Economic Journal, LXXX (Sep 1970) 638-49.
[18] --.and--, 'Factor Market Imperfections, the Terms of Trade, and Wel-
fare', American Economic Review, LXI (Dec 1971) 946-55.
[19] --.and Casas, F. R., 'Factor Market Distortions and the Two-Sector Model
of Economic Growth', Canadian Journal of Economics, IV (Nov 1971) 524-42.
[20] Hagen, E. E., 'An Economic Justification of Protectionism', Quarterly Journal
of Economics, LXXII (Nov 1958) 496-514.
[21] Magee, S. P., 'Factor Market Distortions, Production and Trade: A Survey',
Oxford Economic Papers (forthcoming).
[22] Pearce, I. F., 'The Theory of Wage Differentials: The n x n Case', Journal of
International Economics, I (May 1971) 205-14.
11 Product Market
Imperfections: The Theory of
Monopoly in General
Equilibrium
The price equations now include monopoly profits. Let C,i stand
for monopoly profits per unit of output in thejth sector. Then
(11.3)
and
(11.4)
MR.= P·
J J
(1-_!_)
6j
t Freedom of entry may be restricted by assuming that either all production takes
place under government franchise, patent laws protect the monopolist, or the
monopolist has such an efficient technology that no other producer can possibly com-
pete with him.
Product Market Imperfections 283
C.= TR.-
J
TC.) = p.X.-(rK+wL)
J J J J
"J X X
J J
= Pi~-pJI-(1/t)][MPKiKi+MPLiLi]
(11.5)
~
because, from Euler's theorem, MPKiKi+MPLiLi = ~· Under
competitive conditions ei = oo, so that the excess profits are zero.
With monopoly, however, C,.i > 0, and this brings us directly to the
problem of determining ei in the context of our general equilibrium
system. This indeed is the key to the solution of the system. Perhaps
the main reason for the neglect of monopoly in the traditional
models has been the difficulties encountered in specifying func-
tional forms for ei in the context of general equilibrium. Melvin and
Warne have provided an ingenious way out by assuming a homo-
thetic social utility function, one implication of which is that the
elasticity of demand for each product becomes a function of the
commodity-price ratio and nothing else. In this chapter we assume
that the social utility function is of the constant elasticity of substi-
tution (C.E.S.) variety. In a closed economy,
whence
xl _ p(l +p-/II1D) _ (1-flao) - p"D
x2- (l+pflan) -p -
284 Studies in the Pure Theory of International Trade
and
(11.8)
where
ei = _a~ .Pi (j = 1,2).
opj ~
These expressions confirm the fact that, under the homothetic
utility function, ei is determined solely by p. Furthermore, since the
numerator and denominator of (11.7) and (11.8) differ only by u0 ,
the magnitude of ei is determined only by the magnitude of u 0 .
Specifically, ei ~ I, if u0 ~ 1.
2 (k 2 -k
I L 1 L L·K
IA.= 1)
=(A.Ll-A.Kt)=(AKz-ALz) (11.15)
and
• _ (PL+PK)( •- *) (11.20)
UoPc - IA.I w r
e• - -
P2 UoPp
2 {ian
• - -A • (A 0) (11 21)
1 - [l + p:an] [l + UoP:an] Pp - 1Pp 1 > ·
t Equations (11.21) and (11.22) simply show that as the monopolist, facing one of
the demand curves derived from the utility function, raises his price, the elasticity of
demand for his product rises and vice versa. It is for this reason that e; is shown to be
related to the producers' price ratio rather than the consumers' price ratio, although
on the consumption side e; is related top,.
Product Market Imperfections 289
p2u2 p-flav
e* - D P p* - A p* (A 2 > 0). (11.22)
2 - [1 +pp flan][I +UnPp flav] P - 2 P
Pc
Figure 11.1
290 Studies in the Pure Theory of International Trade
tion that lA. I and 101 are positive, so thatpP is shown to be a decreasing
and p, an increasing function of w. The equilibrium commodity-
price ratio is Pe• and corresponding to it the equilibrium wage/rental
ratio is we. It is manifestly clear from the diagram that there is only
one wage/rental ratio at which the producers' and consumers' price
ratios are identical. It must be remembered, however, that the
monopoly equilibrium exists only under the constraint of an ex-
ceeding unity, for otherwise pP will not be positive.t
x,
45°
0
Figure 11.2
292 Studies in the Pure Theory of International Trade
xl =I
x2 .
Thus, when the two elasticities of demand are equal, the output of
the two goods is also the same. This rather unusual result is attribu-
table to our particular assumption made for simplifying the cal-
culations, namely, (a/b) = 1. In general, the output ratio cor-
responding to the point where e 1 = e2 is determined by the value
of the parameters a, band {3. Draw a 45° line from the origin which
intersects the production possibility locus at E. Then along any
point on the ray OE, e 1 = e2 • One characteristic of the production
point given byE is that at this point theM RT equals the commodity-
price ratio. Thus if the social indifference curve were to touch HH'
at E, the autarky equilibrium under monopoly would be identical
to that under competitive conditions. However, E is one of an
infinity of possible production points and there is no a priori reason
to believe that community preferences will be such as to lead to the
competitive solution, especially when the production and utility
functions are independent. Let an indifference curve Ue pass through
E. Then it is evident that atE, MRT = -pP > MRS= -Pc· A
movement away from E in either direction creates a divergence be-
tween the M RT and pP" As we move along HH' towards H, X1 rises
and X 2 declines, so that the price ratio facing the consumers rises (as
can be confirmed from ( ll.l 0)) and hence e2 rises and e1 declines.
This in turn leads to a rise in oc above unity, and from (11.26) the
Product Market Imperfections 293
w*-p! =
-[(1-err2)eKI +eK2err!AI +eK!err2A2] ( * *) (11.27)
lei P2 -p~
-[(I-erri)(}K2+eK2err!AI +eK!err2A2] ( * *)
w*-p! = lei P2- P! (11.28)
Here a rise inp leads to a decline not only in X1 but also in e 1 , so that
once again we cannot predict how the rise in p will affect the real
income of the monopolist in the first industry. Thus we arrive at the
conclusion that a change in the relative commodity prices exerts a
determinate influence on the real incomes of the primary factors but
not on those of the monopolists.
The same is true of the relative returns of the monopolists, which
are now given by
7t2
-=p-•-
x2 el
7t1 xl e2
because with a rise inp, (X2 /X1) rises but (ede 2 ) declines, so that the
final result is indeterminate.
FP
Figure 11.3
x1
Figure 11.4
300 Studies in the Pure Theory of International Trade
x,
H' F'
Figure 11.5
Product Market Imperfections 301
in the home country and the latter in the foreign country. It is also
evident that ph, the autarkic relative price of the second good, which
is reflected by the slope of U 1 at Sh, exceeds pf, the corresponding
price in the foreign country, indicated of course by the slope of U 1
at Sf. Furthermore, in autarky e2 h > e2f and e 1h < e1f, where the
subscripts h and f refer as usually to the home and the foreign
country. With Ph > pf, the home country will import the second
commodity and export the first. Returning now to the original ob-
jective for which Fig. 11.5 was introduced, it is easy to see that as Ph
and pf move closer to each other when trade is introduced, (e 2 je 1h
and (e 2 /e 1 )f also shift towards each other and at some point are
equalised. Such a point is given byE in the diagram, and the slope of
U1 at E furnishes the world-price ratio FP. Indeed, if the world
elasticity ratio is known, the free trade price ratio can be computed
directly from (II. 7) and ( 11.8), because then
e2 aD+pPan
el I +aDpPan·
which reveals that, once the world elasticity ratio is known,p can be
determined immediately because f3 and aD are given parameters.
This formula also suggests that aD can be equal neither toe nor to
lje, for otherwise p will be either zero or infinity. Furthermore, in
order to avoid the possibility of a negative p, aD is such that it exceeds
or falls short of both e and 1/e.
11.7 Appendix
Some of the equations presented in the text without proof will now
be derived.
Consider a utility function with a constant elasticity of substitu-
tion, av, where av = 1/(1 + {3), -I < {3 < oo. With two sectors in
the economy producing two goods, xl and x2, the utility function
can be represented as
(All.l)
Maximising, we obtain
az
- ~ (aX!fl +bX:Zfl)-lffJao( -af3X!lfao)-ply = 0 (All.3)
oX1
az
-!{3 (aX!fl +bX:Zfl)-11/lao( -bf3X21/ao)-P2Y = 0 (All.4)
ax2
az
Y -p1X1 -p2X2 = 0. (All.5)
oy
Product Market Imperfections 303
(All.6)
(All.7)
whence
P2[l +crv(P2)-/J"n(Pt)/J"n]
P2 +(p2)"D(Pt)fJ"D
so that
Similarly,
(12.1)
and
(12.2)
306 Studies in the Pure Theory of International Trade
where, as before, cij is the amount of the ith input divided by the
jth good (i = L,K; j = 1,2,n), for example, CLn = (Ln/Xn). The
price equations, on the other hand, now become
CL1w+CK 1 r = p 1 (12.3)
CL2w+CK2r = P2 (12.4)
and CLnw+ CKnr = Pn (12.5)
where Pi is the price of the jth good. As before, the input-output
coefficients can be solved from the expressions relating them to
factor prices. But we still have one too many unknowns, for there
are five equations and six unknowns, X 1 , X2 , Xn, w, rand Pn· The
system can be given a determinate solution, however, by introducing
another equation concerning the demand-supply equality for Xn.
Then
(12.6)
which shows that the local supply of the non-traded good equals its
demand, which in turn depends on p, q and Y, where p and q are,
respectively, the relative prices of the second and the non-traded
good in terms of the first, and where Y is national income. This gives
us a completely determinate system of six equations and six un-
knowns, given that the production coefficients can be solved from
the factor prices.
The solution of the model with the non-traded good proceeds by
first solving for w and r from the two price equations (12.3) and
(12.4) in terms of the input-output coefficients and p 1 and p 2 , which
are determined internationally. The values of w and r so obtained
can be plugged in (12.5) to solve for Pn in terms of the production
coefficients of all three commodities. Since production coefficients
themselves are functions of wand r, we get a determinate solution
for factor rewards, the price of the non-traded good and capital/
labour ratios in all three goods, once the values of the parameters
p 1 and p 2 are internationally determined. Note that this solution is
obtained independently of demand conditions and factor endow-
ments. If the country in question is a small country and hence faces
given prices of traded goods, the demand conditions and factor
supplies exert no influence on factor prices and prices of non-
traded goods. If in addition we assume that production functions
are similar internationally, wand r will also be the same in the two
countries. This follows clearly from the solution of (12.3) and
Non- Traded Goods 307
x,
Figure 12.1
308 Studies in the Pure Theory of International Trade
X _ CL2L(k2-k)+CL2CLit(kn-k2)Xn (12.7)
1 - ICI
X _ CL1L(k-k1)+CL1CL1t(k1-kn)Xn (12.8)
2- ICI
where, as usual, k1 is the capital/labour ratio in the jth good
(j = 1,2,n) and ICj is the determinant of the matrix of production
coefficients involved in X 1 and X2 , that is, ICI = C" 2CL 1-CL2CK1·
If commodity prices are constant, (12.3)-(12.5) show that factor
prices and production coefficients will remain unaltered. But this is
not sufficient to ensure that X 1 and X 2 will also be unaltered so long
as Xn is positive and all capital/labour ratios differ. For then the out-
put of one or both goods will vary with the variation in Xn even
though commodity prices were invariant. Thus when the number of
goods exceeds the number of factors, any level of commodity out-
puts is consistent with a given set of prices. The production possi-
bility surface is then described by straight lines (as in Fig. 12.1) and
each line is a locus of possible output configurations for the three
commodities corresponding to a given set of commodity prices. No
wonder, then, that when all three goods are traded, the system is
indeterminate, for the specification of international prices is insuffi-
cient to determine the output levels. The characterisation of one of
the goods as non-traded, however, solves the problem. For suppose
that the output of X, is fixed at some level furnished by domestic
demand; then X 1 and X 2 can be easily obtained from (12.7) and
(12.8). Thus there is no indeterminacy if one of the three goods is
non-traded, although the production possibility surface continues
to be a ruled surface.
A richer understanding of these issues emerges from an examina-
tion of Fig. 12.2, which is constructed on the basis of an ingenious
geometrical technique discovered recently by Melvin [5]. Fig. 12.2
comprises two boxes, one given by 0,.0 2 and the other by 0 1 0 2 •
There are three origins, each representing one commodity; for On
represents X,, 0 1 represents X 1 , and so on. The output level of each
good is given by the distance of the production point from its origin.
Suppose the output of Xn is fixed at the level 0 1 0,., then the remain-
Non-Traded Goods 309
Figure 12.2
to obtain
A.LIXi+A.L2X! = L*-(A.LIC!t +A.L2Ct2)-A.Ln(Ct,.+X:) (12.9)
(12.11)
(12.12)
(12.13)
and where, in obtaining (12.11 )-(12.13), use has been made of the
cost-minimising condition OLiCti+ OKic:i = 0, with O's as before
denoting the relative share of the relevant factor. In (12.14), En 2 is
the cross-elasticity of demand between the domestic and the second
good, En is the price elasticity and 'In the income elasticity of demand
for the non-traded good. From (12.11) and (12.12)
whence
q*
p* =wJOnJ (12.15)
t This was accomplished in Chapter 2, and the expressions for Cl1 and Cl1 derived
there remain unchanged in the present model because each good is still produced with
the help of two factors.
312 Studies in the Pure Theory of International Trade
to the traded goods. Suppose that all goods are gross substitutes, so
that (X"* jp*) is positive. We still have to reckon with the sign of lA. I
and IA.n 2 l. Suppose IA.I > 0, so that k 2 > k 1 ; if IA.n 2 l is also positive,
so that k" > k 2 , then (X.*fp*) may not be unambiguously negative.
On the other hand, if the domestic and the second good are com-
plements, so that (X"*/p*) < 0, then with both 101 and IA.n 2 l > 0,
(Xt /p*) is unambiguously negative and the price-output response
in the first industry is normal. Similar considerations apply to the
price-output response in the second industry. The gist of this dis-
cussion is that, in the presence of purely domestic goods, the price-
output response of the traded goods is no longer predictable. The
outcome depends not only on the substitution effects on the produc-
tion side, but also on the nature of the demand for three commodi-
ties as well as their relative factor intensities.
This result generates far-reaching ramifications for several results
which we derived in the chapter on the theory of nominal tariffs.
First, the offer curve may not only be negatively sloped at some
volume of trade, but may also bend back towards the origin. In other
words, the price elasticity of demand for imports may no longer be
negative. This factor tends to create instability in the model. Second,
the introduction of the tariff may result in a rise in the import
demand at constant terms of trade, and hence eventually cause a
deterioration in the terms of trade of the tariff-imposing country. t
t It is true that, in the presence of the domestic good, the demand for the import-
able good is also influenced by the price of X. which may rise or decline as a result of
the tariff, and this factor, absent in the usual two-good model, may strengthen or
attentuate the resultant decline in the demand for the importable good. Nevertheless,
in the absence of the categorical relationship between the output of X2 and the tariff-
induced rise in its local relative price, it is no longer certain that the rate of tariff and
the demand for imports are negatively related when world prices are kept constant.
For this reason, tariffs may or may not cause an improvement in the terms of trade
of the tariff-imposing country. For further discussion, see Komiya [ 4].
Non- Traded Goods 313
w• = (12.18)
• (}Ll
r = TOTP2• (12.19)
X* _ (2K2L* -2L2K*)+(2L22Kn-2K22Ln)Xn*
(12.20)
1 - 121
(12.21)
I21Xf = [2K2(1-'7n(JL2Ln)+'7n(JL2L22Kn]L*
- [ AL2(l- '1n0K...1.Kn) + '7n0K...1.K2ALn]K*
IA.IX! = [2Lt0-'7n()K2Kn)+'7nOKA.Kl2Ln]K*
- [2Kl(I-PJn()LA_Ln)+PJn()L2LlAKn]L*
Now
t See the expression for (l/U1 )(dUfdt) in Chapter 4. For a full discussion of gains
from trade in the presence of non-traded goods, see Batra [I].
316 Studies in the Pure Theory of International Trade
or (12.27)
where w1 is the wage rate in the jth sector and where rx1 ~ l is a
constant. With no wage differential, rx 1 = rx 2 = rx 3 . In the presence
of the differential, however, the following possibilities arise:
l. rx 1 = rx 2 < rx,., and the differential is paid by the non-traded
good.
2. rx 1 = rx, < rx 2 , and the differential is paid by the import-
competing good.
3. rx 2 = rx,. < rx 1 , and the differential is paid by the exportable
good.
In addition to these, there are six more possibilities if the differential
exists among all the industries. These are:
rx 1 ~ rx 2 ~ rx,. and w1 ~ w2 ~ w,;
rx 1 ~ rx,. ~ rx 2 and w1 ~ w, ~ w2 ;
rx 2 ~ rx 1 ~ rx,. and w2 ~ w1 ~ w,.
For expository purposes, we rewrite the full-employment equations
as
L 1+L 2+L, = L and K 1+K2+K, = K.
Our first task here is to obtain an expression concerning the relation-
ship between the outputs of all three industries, something similar
to the slope of the transformation curve in the two-good case.
Totally differentiating (12.25) and dividing through by dX2 , we
obtain
dX1 FK1dK1 +FL1dL1
(12.28)
dX2 = FK2dK2+FL2dL2.
From the full-employment equations
dL 1+dL 2+dL, = 0 and dK 1+dK2+dK, = 0.
Using these and (12.26) and (12.27), (12.28) can be written as
dX1 = -p[FK2(dK2+dK,)+(rxdrx2)FL2(dL2+dL,)J
dX2 FK2dK2 + FL2dL2
(12.29)
Non- Traded Goods 317
where ~ = FK2dKn+(a.n/a.2)FL2dL,. ~ l if
y FK2 dKn + (a. tfa. 2)FL2 dLn
dU
U1 = pdX2(l-fJ)+qdX,.(I-y). (I2.3I)
Let us now consider the case where factor markets are distorted
but the domestic goods do not exist, so that (12.31) becomes
dU = pdX2(I- {3).
ul
Here {3 reflects the divergence between the marginal rate of trans-
formation and the international price ratio. As expected, the optimal
policy in the presence of the differential requires a change in X 2 such
that the divergence between the MRT and the given international
price ratio is eliminated, so that the value of {3 to producers comes
to equal unity. The second-best policy in other words requires the
imposition of production tax-cum-subsidy on the second andjor
the first commodity. The first-best policy of course consists in the
elimination of the wage differential to the producers by means of the
grant of a factor tax-cum-subsidy, such that (rxtfrx 2 ) and {3 are
equated to one. None of these results is new, but they have been
reproduced to present a contrast to the results to be derived below.
It is noteworthy here that these results are not modified even if the
non-traded good exists, so long as y = I or rx 1 = rx., so that wages
are the same in the exportable and the purely domestic-good
industries.
Next, consider the case where the non-traded good exists and the
wage differential is paid by either X 2 or X •. In the first case
rx 1 = rx. < rx 2 , which means that y = I but {3 < I. Here again the
second-best policy calls for a production tax-cum-subsidy to be
granted to the traded goods. In the second case where the differen-
tial is paid by the non-traded good, rx 1 = rx 2 < rx. and y < 1 but
{3 = I, so that the change in welfare reduces to
dU = qdX.(l- y).
ul
Here the second-best policy calls for a consumption subsidy to be
granted to the domestic good, such that the divergence to the pro-
ducers between - (dXtfdX2 ) and p[l + (qjp)(dX./dX2 )] is eliminated,
that is to say, the effective value of y to the producers is equalised to
unity. Note that this policy does not apply to the traded goods. Since
the distortion is in the domestic production sector of the economy,
the thrust of the cure should also lie in removing the inequality be-
tween -(dXtfdX2 ) and p[l +(qjp)(dX./dX2 )], just as in the two-
good model the second-best optimum is achieved by changing the
output of X1 and X 2 in such a way that the difference between
Non- Traded Goods 319
q (l-y) dX2
(12.32)
p(l - /3) = dX"
320 Studies in the Pure Theory of International Trade
in which case free trade alone is the second-best policy and the im-
position of the production tax-cum-subsidy is not needed. Although
this result is purely of academic interest, for only by coincidence will
y and f3 be such as to satisfy (12.32), yet the condition does suggest
an interesting possibility. In any case, if the difference between the
left- and right-hand sides of(l2.32) is small, the gain to be derived
from the introduction of the production tax-cum-subsidy may not
be more than negligible and may not be worth the administrative
headaches and costs. A necessary condition for (12.32) to be satis-
fied is that (dX2 /dX,) be negative, because both y and f3 exceed unity
when the differential is paid by X1 producers. The chances of(l2.32)
being satisfied improve if the wage differential exists among all three
industries; nor is it necessary any more that (dX2 /dX,) be negative,
because the differentials may be such that y ~ I when f3 S I. The
essential point is that there exists a certain configuration of wage
differentials which will satisfy (12.32). Nevertheless, in all the cases
discussed in this section, the first-best policy requires the grant of
factor tax-cum-subsidy so that the differential to the producers is
eliminated.
An interesting policy device to attain the second-best welfare sug-
gests itself from this discussion. Suppose the level of the inter-
industry factor-price differential itself depends on government
policy. For example; Johnson [3], referring tQ Harberger [2], has
suggested that the inter-industry factor-price differential will occur if
the government imposes a corporation income tax (which is a tax on
the use of capital in the corporate sector alone) that drives a wedge
between the gross rewards of capital in the corporate and the non-
corporate sector. If the government must impose the corporation
income tax in order to raise revenue, etc., then our argument sug-
gests that it should be imposed in such a way as to satisfy condition
(12.32). Although y and f3 represent the wage differentials among
the three industries under the assumption of similar reward of
capital in all industries, the essential nature of our analysis is un-
modified in the case where the rewards of capital differ among in-
dustries but the wage rates are similar. Depending on the situation,
the level of the tax on the use of capital could be determined in the
following manner:
(i) The entire revenue can be raised from taxing the use of
capital alone in the first commodity. If the other two in-
Non- Traded Goods 321
REFERENCES
[I] Batra, R. N., 'Non-Traded Goods, Factor Market Distortions and the Gains
from Trade', American Economic Review (Sep 1973).
[2] Harberger, A. C., 'The Incidence of the Corporation Income Tax', Journal of
Political Economy, LXX (June 1962) 215-40.
[3] Johnson, H. G., 'Factor Market Distortions and the Shape of the Transformation
Curve', Econometrica, XXXIV (July 1966) 686-98.
[4] Komiya, R., 'Non-Traded Goods and the Pure Theory of International Trade',
International Economic Review, VIII (June 1967) 132-52.
[5] Melvin, J. R., 'Production and Trade with Two Factors and Three Goods',
American Economic Review, LVIII (Dec 1968) 1249-68.
13 International Investment
but
and (13.1)
For the sake of illustration, let us assume that the home country
imposes a tariff (t11) on the imports of the second good, so that
p, = p(l +t,)
where, as before, p is the terms of trade and p, the domestic-price
ratio in the home country in terms of the first good. However, in
the presence of the tariff the capital market cannot remain in
equilibrium. For example, let
MPK 111 = MPK 11 (13.2)
that is, let the return to capital in the first industry be the same in
the two countries. However, when exchanged for the equivalent of
the second good, MPK 111 = p,MPK211 and MPK 11 = pMPK21 , so
that in view of (13.2)
(13.3)
Since, in the presence of the tariff, p,. > p, it follows from (13.3)
that
Figure 13.1
(13.6)
(13. 7)
and
(13.8)
where (13.5) and (13.6) have been described several times before;
(13.7) specifies the production constraint which takes into account
the amount of home country's capital(/) invested abroad; and (13.8)
describes the balance-of-payments constraint in the presence of the
foreign investment, that is, the value of imports equals the value
of exports plus the earnings on foreign investment which equal rii,
where 'I is the foreign rate of return on capital expressed in terms
of the second commodity. If the home country is a net creditor, I is
positive; it is negative if the home country is a net debtor. Using
(13.5), (13.6) and (13.8), (13.4) can be written as
(13.9)
and
(13.11)
where
(} = 1,2).
(13.12)
(13.13)
(13.14)
What is interesting in the case of (13.14) is not only that the optimal
tariff formula is different in the presence of international investment,
but that there also arises the possibility that the optimal tariff may
be zero or even negative, for llJ and y1 may possess any sign. If the
home country is a net creditor, lli > 0; YI on the other hand depends
on whether X 1 is capital- or labour-intensive in the foreign country.
If X 1 is capital-intensive abroad, YI > 0, and if it is labour-intensive,
YI < 0. This follows from the Stolper-Samuelson theorem, but only
if the foreign country is incompletely specialised. Thus we conclude
that in the case of incomplete foreign specialisation the optimal tariff
may be of any sign and may even be zero, even if the country in
question possesses monopoly power in international trade. The reason
for this surprising result is not far to seek. In the presence of foreign
investment two terms of trade are involved, namely,p and ri, and as
a result of the tariff p and ri may or may not move in the same
direction, depending on whether the first good is capital- or labour-
intensive abroad. Now the home income rises owing to a decline in
p, but declines owing to a decline in 'I· Thus if the two effects cancel
out, the optimal tariff is zero. Nevertheless, in the presence of
complete foreign specialisation when YI = 0, the optimal tariff
reduces to the standard formula derived in Chapter 5.
330 Studies in the Pure Theory of International Trade
and then to
(13.16)
and
MPK 2 h = MPK 21 (1-rh).
But rh = MPK1h/Ph and r1 = MPK 11 /p, so that
MPK 11 (l-rh) pr/1-rh)
rh = = .
Ph Ph
International Investment 331
On the other hand, if the home country is a net debtor, and imposes
a tax r1 on the earnings of foreign capital, the capital market is in
equilibrium if
and
MPK 2h(l-r1 ) = MPK 21 .
Using rh and r1 , however, this becomes
rh(l-r1 )(l+th) = r1 . (13.18)
Substituting for rh from (13.17) and (13.18) in (13.16), the two
optimal rates of taxation are given by
and
(13.19)
and
0 +Ji.JYJ)(flfyf-afyi)
'J = . (13.20
a1 [JJ.1 y1 -yi(1 +JJ.1 Y1 )]
As before, the signs of rh and r1 depend on the signs of y1 and Ji.J
which in turn depend on the inter-industry factor-intensity relation-
ships in the foreign country and whether I ~ 0. To begin with,
suppose that the optimum tariff is zero, that is, 1+ Ji.JYJ = 0. This,
from (13.19) and (13.20), means that the optimal taxes on foreign
332 Studies in the Pure Theory of International Trade
REFERENCES
[I] Gehrels, F., 'Optimal Restrictions on Foreign Trade and Investment', American
Economic Review, LXI (Mar 1971) 147-59.
International Investment 333
[2] Jones, R. W., 'International Capital Movements and the Theory of Tariffs and
Trade', Quarterly Journal of Economics, LXXVII (Feb 1967) 1-38.
(3] Kemp, M. C., 'The Gain from International Trade and Investment: A Neo-
Heckscher-Ohlin Approach', American Economic Review, LVI (Sep 1966)
788-809.
[4] MacDougall, G. D. A., 'The Benefits and Costs of Private Investment from
Abroad: A Theoretical Approach', Economic Record, XXVI (Mar 1960) 13-35.
[5] Mundell, R. A., 'International Trade and Factor Mobility', American Economic
Review, XLVII (June 1957) 321-37.
[6] Negishi, T., 'Foreign Investment and the Long-Run National Advantage',
Economic Record, XLI (Dec 1965) 628-32.
[7] Rakowski, J. J., 'Capital Mobility in a Tariff-Ridden International Economy',
American Economic Review, LX (Sep 1970) 753-60.
14 International Trade
in a Dynamic Economy
t See, for example, Bardhan [1,2], Kemp [3] and Khang [4].
International Trade in a Dynamic Economy 335
( 14.1)
p 2 D2 = sY = s(rK+wL). (14.2)
£ 2< · > = s ( -r K +-
w )
L = sX-
1
(14.4)
mm P2 Pz P
(14.6)
Figure 14.1
\
\
\
\
\
\
\
\
\
\
0~------~~~----~~----~'T---------------~p
\
\
\
\
\
\
~------------~--H'
\
\
\
~----
Figure 14.2
* _ -p
e2- •[d2
- (} +-
x2 ·----.
x!J
1 (14.7*)
e2 e2 p
where e! = £!- L* and x! = X!- L*. Thus the relationship
between the excess demand function and the terms of trade changes
little when it is stated in per capita terms, provided the factor sup-
plies are constant. The uniqueness of the momentary equilibrium
340 Studies in the Pure Theory of International Trade
s
e2(min) = -X 1
p
(14.9)
and
e2(max) = -(1-s)Xz (14.10)
where w is the wage/rental ratio. In the case of complete specialisa-
tion, the sign of oe 2 jok is apparent from one glance at (14.9) and
(14.10). Evidently,
oe2(min) - s ox 1 0 (14.11)
~-pak>
and
8e2(max) = -(I-s) OXz < 0 ( t4.12)
ok ok
because (ox)ok) > 0. The case of incomplete specialisation, how-
ever, is not susceptible to such straightforward results. Differenti-
ating (14.8) partially with respect to k and remembering that wand
r are constant with a given p, we have
oe2 sr ox2
ok = P2 -8k (14.13)
International Trade in a Dynamic Economy 341
(l4.t5)
op
= (14.16)
ok
342 Studies in the Pure Theory of International Trade
Similarly,
(14.17)
K
* =D-
2 sr (w+k)
=-·-- (14.18)
K p2 k
(14.20)
w
Q* = (r*-p*)+--(w*-k*) (14.21)
2 (w+k) ·
344 Studies in the Pure Theory of International Trade
• •
r -p 2
(hzp*
=-~-()-I an
d
w• = w·
-p*
Q* w ]p*
K* =
[
() LZ - k +w rei - k +w w k*.
Now
w
-k-- = {)L = {)l{)Ll +{)2{)L2·
+w
Using this relationship and the fact that
1{)1 = {)Ll- {)L2
These three equations confirm the fact that Q* /kJ, QJik* and
QJikJ are all negative under the capital-intensity condition.
International Trade in a Dynamic Economy 345
(b) The Home Country Specialised in the Capital Good, but the
Foreign Country Incompletely Specialised
If the home country is specialised in the production of capital goods,
then p 2 D 2 = SP2X2 , so that
K* = D 2 sX2 sx 2
K K y·
Moreover, x 2 = l/CL2 and k = CK 2 /CL2. Therefore
s
K*=-
CK2
whence
s
Q = K*-L* = --L*. (14.29)
CK2
However, since the foreign country continues to be incompletely
specialised,
(14.20)
Q* Q*
K*
but k* = 0
J
because CK 2 does not change with a change in k 1 alone. From the
analysis in Chapter 2 we know that
c:.2 = ()L2(1'2(J). and c:.2- Ct2 = k* = (1'2(1).
so thatt
Q*
k* = -{}L2 < 0. (14.30)
00. (14.31)
t The reader may be reminded here that in the absence of technical progress Ct.J
is the same as At, the notation used in Chapter 2 for a change in the input-output
coefficient.
t It can easily be seen that the foreign pattern of specialisation plays no role in
determining the sign of (dk1 jdk)a=o when the home country is specialised in the
capital good.
International Trade in a Dynamic Economy 347
0=0
0 A k
Figure 14.3
t Here again, one can see that the pattern of home specialisation plays no role in
determining the sign of (dk1 /dk)a~o· when the foreign country is specialised in the
cap1tal good. Thus we see that whenever a country is specialised in the capital good,
the level of specialisation in the other country is immaterial in determining the shape
of the Q = 0 and Q1 = 0 curves.
348 Studies in the Pure Theory of International Trade
I
------a,=o
0 k
Figure 14.4
International Trade in a Dynamic Economy 349
0=0
o,-o
0 k
Figure 14.5
350 Studies in the Pure Theory of International Trade
REFERENCES
[I] Bardhan, P. K., 'On Factor Accumulation and the Pattern of International
Specialisation', Review of Economic Studies, xxxm (Jan 1966) 39-44.
[2] - - , 'Equilibrium Growth in the International Economy', Quarterly Journal
of Economics, LXXXIX (Aug 1965) 455-64.
[3] Kemp, M. C., 'International Trade and Investment in a Context of Growth',
Economic Record, XLIV (June 1968) 211-23.
[ 4] Khang, C., 'A Dynamic Model of Trade between the Final and the Intermediate
Products', Journal of Economic Theory, I (Dec 1969) 416-37.
[5] Oniki, H., and Uzawa, H., 'Patterns of Trade and Investment in a Dynamic
Model of International Trade', Review of Economic Studies, XXXII (Jan 1965)
15-38.
SUPPLEMENTARY READINGS
[6] Atsumi, H., The Long-Run Offer Function and a Dynamic Theory of Inter-
national Trade', Journal of International Economics, I (Aug 1971) 267-99.
[7] Bardhan, P. K., and Lewis, S., 'Models of Growth with Imported Inputs',
Economica, xxxvii (Nov 1970) 373-85.
[8] Inada, K., 'International Trade, Capital Accumulation, and Factor Price
Equalization', Economic Record, XLIV (Sept 1968) 322-41.
[9] Johnson, H. G., 'Trade and Growth: A Geometrical Exposition', Journal of
International Economics, I (Feb I97I) 83-IOI.
(10] Stiglitz, J. E., 'Factor Price Equalization in a Dynamic Economy', Journal of
Political Economy, Lxxvm (May-June I970) 456-88.
[II] Vanek, J., 'Economic Growth and International Trade in Pure Theory',
Quarterly Journal of Economics, LXXXV (Aug 1971) 377-90.
Author Index