Protection and Non-Competing Groups

Author(s): Bertil Ohlin

Source: Weltwirtschaftliches Archiv, 33. Bd. (1931), pp. 30-45
Published by: Springer
Stable URL: http://www.jstor.org/stable/40416834
Accessed: 29-08-2016 02:04 UTC

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted
digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about
JSTOR, please contact support@jstor.org.

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
http://about.jstor.org/terms

Springer is collaborating with JSTOR to digitize, preserve and extend access to Weltwirtschaftliches
Archiv

This content downloaded from 128.143.23.241 on Mon, 29 Aug 2016 02:04:05 UTC
All use subject to http://about.jstor.org/terms
 Protection
and Non-competing Groups
By
Professor Dr. Bertil Ohlin
Stockholm

I.

is often said that ideas and theories do not precede and call forth
events, but on the contrary, things happen first and find their
explanation and justification afterwards. It also happens, however,
that the justification fails to appear and that, nevertheless, the economic
development continues as it has started. Such has been the case with
the protectionist movement of the last half century, which has gone
from victory to victory in the world of politics and is still growing rapidly,
whereas the attempts to justify such a policy - to create a theory of
protection - have all failed rather decisively.
Recently a book has appeared1 which is intended to fill this gap
and to explain "why and how, in all countries and at all stages of their
evolution, it is advantageous to protect certain branches of industry
through protective duties or state subvention" (p. 5).2 This theory of
protection is said to yield "scientific criteria" as to which industries ought
to be protected and which not, how great protection in the former
industries ought to be, and how long the protection should last (p. 22).
Let me try to outline very briefly the main contents of the new
theory. The first step towards its construction is an observation of the
fact that certain industries are much more productive than others. The
net production of an industry is obtained if from the "value added by
manufacture" - as used in the American statistics - one deducts the
costs of depreciation and repair of the capital instruments. This net
production is the real national income: "le gain national réalisé par une

1 M. Manoï leseo, Théorie du protectionnisme et de l'échange international. (Biblio-

thèque internationale d'économie politique.) Paris 1929.
2 References in parenthesis throughout this article are to pages in the original work.

This content downloaded from 128.143.23.241 on Mon, 29 Aug 2016 02:04:05 UTC
All use subject to http://about.jstor.org/terms
 Bertil Ohlin, Protection and Non-competing Groups oj

branche quelconque de la production'1 (p. 34). It may well be great in

a certain industry, although the profits of the capitalists there are small. -
In order to compare the productivity of various industries it is necessary
to put the "net production" in relation to the quantity of labour and
capital used in each industry. In other words, the output per worker
and the output per unit of capital in each industry must be established.
P P
If for a certa
p T C
geometric mean
in this industry
ance with the h
As statistics are
industries, mea
with figures for
following count
Bulgaria, very d
per worker is mu
(p. 61). Figures
the population
of the national
produce 80 % °f
agriculture are
"Il en résulte qu
avantage relati
industrielles" (p
The conclusion
may be highly
at higher prices
from abroad. For
contribute mor
industries whic
It is not necessa
thing to pay w
and to buy at hig
To explain whe
certain formulas.1 Consider two countries A and I. The relation between
the quantity of productive factors used in I and that used in A to produce
a certain quantity of agricultural goods is q, and the corresponding relation
for a quantity of manufactured goods is qv If both q and qx are greater
1 Readers who dislike formulas can spring over the next two and a half pages of this
paper and read only the last section of part one.

This content downloaded from 128.143.23.241 on Mon, 29 Aug 2016 02:04:05 UTC
All use subject to http://about.jstor.org/terms
 32 Bertil Ohlin

than unity, country A is superior

is larger than qv then - indicates a

culture. Assuming A to be small an
to be fixed, A gets the whole benefit

exportation of agricultural goods

than it could get by producing th
All this is just as explained in the R
productivity in manufacturing indus
is K times greater than in agricultur
unit of productive factors is K times
for assuming differences in profits, t
factures must exceed the correspondin
portion. Therefore, agricultural good
manufactures in proportion to the qu
their production would require. For a

one obtains only -== as much manufac

of equal productivity, i. e., uniform f

q q
of the former sort
qi Jvqi
factures as it would have obtained by p
It gains from international trade only if

unity, i.e., if - is larger than K. "If the

agriculture y-^y is smaller than the intrins

industries [abroad2], then ... the latter have
In this there is not, as Mr. Manoïlesco
conflict with the classical theory. Assume t
superiority, i. e., greater productivity, for
higher prices paid in I to the productive
in agriculture will then reduce the competi

enable manufacturing industries in A to ex

1 This assumption is the basis of the whole of

upon below. In his criticism of Ricardo he makes ass
countries at the same time reap the whole bene
consequence (vd. p. 130-138).
2 The word "abroad" is inserted by me. The author has defined K as "le rapport des
productivités entre l'article industriel et l'article agricole dans le pays industriel" (p. 144).
8 cf. F. W. Taussig, International Trade. New York 1927. p. 46.

This content downloaded from 128.143.23.241 on Mon, 29 Aug 2016 02:04:05 UTC
All use subject to http://about.jstor.org/terms
 Protection and Non-competing Groups ^o

The question becomes more complicated only in so far as there is an

intrinsic superiority Kv i. e., higher prices for the factors of production
in manufacturing industries than in agriculture in A also, not only in I.
The necessity of paying higher factor prices in these industries in A
than in agriculture, serves as a handicap against the former and may
make them unable to compete with imported manufactures, even if
production in A would have brought a greater quantity of them than
the use of the same factors in the production of agricultural goods for
export to pay for imported manufactures will do. Through commerce

A obtains =p- as much manufactured goods as it would have done, if

Kqi
the same factors had been used in manufacturing industries at home.
Thus, A gains from having its own manufactures, if the value of
this expression is smaller than unity. However, manufactures will be
q i
unable to com
Kqi Ki
it may well fulfil this con
trade may bring a loss in
q 1 q K
of a loss is: iy^r~y^r or K> >=
ivqi Ki qi Ki
international trade, if the comparative su
smaller than the intrinsic superiority in I
between the indices of intrinsic superiority
The greater the intrinsic superiority of ma
the chances that trade brings a loss and, h
manufactures alive through protection in A
This seems to be the logical conclusion of
Unfortunately, he makes the mistake of mixi

that A looses from trade, if Kj>- . In othe

elusion quoted above in the following way: If
of A in agriculture is smaller than the intrinsi
in A, then the latter have a full right to ex
Now, let q represent the average superio
production - i.e., superiority with regard
which represent the whole of A's product
in the production of any individual article.

it still holds that if - < K, it is advantageo

at home. From this he draws the followin
parative superiority of a country A with
a certain article on the one hand, and the pr
Weltwirtschaftliches Archiv Bd. XXXIII. 3

This content downloaded from 128.143.23.241 on Mon, 29 Aug 2016 02:04:05 UTC
All use subject to http://about.jstor.org/terms
 34 Bertil Ohlin

on the other, is smaller than the int

then it is more advantageous to pro
(p. 159 sq.).
In order to express his ideas in monetary terms, the author proceeds
to assume that "dans les deux pays la même somme d'argent coûte le
même effort national", in other words that the price level for the produc-
tive factors is equal in all countries (p. 160). His conclusion then becomes:
"The more the productivity with regard to a certain article exceeds the
average productivity of a country, the more the domestic price of this
article has the right to exceed the foreign price" (p. 161), and yet production
at home is advantageous. This is perhaps more clearly expressed through
an example. "A car load of foreign coal can be obtained for 6,000 lei in
Roumania, whereas Roumanian coal of the same quality costs 7,500 lei
per car load. However, the annual output per worker in the coal industry
is about 75,000 lei, whereas the average productivity for all industries
in Roumania is about 30,000 lei per worker and year. Under these circum-
stances it is more advantageous for Roumania to produce the coal than
to import it" (p. 161). To produce 100 car loads one needs 10 workers
during a year, but to produce export goods at the value of 600,000 lei
to pay for the import of 100 car loads one needs 20 workers during a year.
Evidently, it is profitable to support industries which have high produc-
tivity. If the latter is more than three times as high in a certain indus-
try as in other industries, a duty of 200 %» which makes the prices of
products from the former industry three times as high as import prices,
is advantageous.

II.

This I think is the main content of Mr. Manoïlesco's new theory of

protection. Let me examine it critically but briefly.
As far as the last example is concerned, one may well ask, why the
average productivity of all Roumanian industries is regarded as represent-
ative for the productivity in the export industries. If there is a certain
expandable export industry, where the output exceeds 60,000 lei per
worker and year, then the importation of foreign coal is profitable
according to the author's own assumptions. - Secondly, the assumption
that the price level for productive factors is equal everywhere is so far
from reality, that conclusions built on this assumption can have but

1 i. e., the superior productivity of the industry producing this article. The author
here again makes the mistake to put the intrinsic superiority in A (Kt) instead of in
I (K); cf. above.

This content downloaded from 128.143.23.241 on Mon, 29 Aug 2016 02:04:05 UTC
All use subject to http://about.jstor.org/terms
 Protection and Non-competing Groups oc

little interest. Money wages in the United States are much

ten times as high as in Roumania.
For some further criticism let us turn back to the first stag
development of the theory. Consider first the index of product
is the total output to be put in relation only to the quantity
and capital, and not to the quantity of natural resources ? This d
the author seems to have felt himself, for in a foot-note (p.
adds the following statement: "In comparisons of agricultura
the form ^TC could fruitfully be replaced by the form ^TS o

yCTS , where S is the surface of the land employed by a cer

of agriculture." This, however, does not help him very far
reason is a geometric mean to be used? And, what is a unit o
It is highly significant that Mr. Manoïlesco in all his illustr
the output per worker as an index of productivity and makes no
to use the formula.1
In brief, I fail to see that this formula can be used to measure the
productivity of various industries in a country. It is better to stick to
the criterion used in orthodox economics: the industry which is able to
pay the highest prices for the productive factors is the most effective one.2
Of course, this does not enable one to put all industries on a scale of
productivity, but it does divide the industries in two classes: those which
are able without artificial support to pay the existing factor prices, and
those which are not able to do so and, therefore, can exist only if part
of these prices is paid from some outside source.
It goes without saying that the output per worker does not pro-
vide any test as to productivity, as the quantity of other productive
factors used per worker is widely different in different industries. The
high value output per worker in the chemical industry is chiefly due
to the fact that relatively much more capital is invested than in most
other industries.
A much more useful indication of productivity is obtained simply
through figures for the height of wages in various industries. As the rate
of interest does not vary much between industries in the same country,
such an index of productivity in money terms is misleading chiefly in
so far as the price paid for natural resources of identical qualities is different
or one industry realises a higher profit than another one. Of course, one
would have to consider also the differences in the quality of labour used,

1 Mr. F. Palander, Kiel, has read the proofs of this paper and made suggestions,
which have led to certain corrections and amplifications.
2 For qualifications to this statement see below.
3*

This content downloaded from 128.143.23.241 on Mon, 29 Aug 2016 02:04:05 UTC
All use subject to http://about.jstor.org/terms
 36 Bertil Ohlin

e. g. the fact that female labour plays

than in most others.
However, such an index of productivity in monetary terms would
tell very little as to the "desirability" of the existence of the one industry
or the other. A low nominal wage may be a high real wage, if the costs
of living are low; e. g., farm labourers obtain food much cheaper than
workers in manufacturing industries. Besides, the commodity prices
obtained by a certain industry may be much lower than the potential
import prices for similar goods. A rise in its sale prices would enable it
to pay higher wages and to climb some step higher up on the scale of
productivity.
Curiously enough Mr. Manoïleso says nothing even about the differ-
ences in costs of living, which are so largely responsible for the lower
nominal wage level and value output per worker in agriculture. Neither
does he observe, that the high value output of certain industries in certain
countries is due to the fact that protection enables them to sell their
goods at high prices. Still less, of course, does he consider such less obvious
parts of the effects of duties as their tendency to raise the prices of the
productive factors in terms of money.
I leave now the usefulness or lack of usefulness of the author's index
of productivity, and turn to his conclusion that a manufacturing industry
"has a right to exist" in an agricultural country, even if it produces
goods at higher prices than the import prices of similar foreign goods,
assuming that its superior intrinsic productivity more than makes up
for this "commercial" inferiority. From the assumptions made this may
be true enough. But the author forgets to draw the other conclusion
which also follows, namely that agriculture has no right to exist
in his agricultural country! As he has assumed constant costs and
fixed prices on the world market, it would pay to produce only manu-
factured goods and to import agricultural products. If he had drawn this
conclusion, he would perhaps had come to doubt the fruitfulness of the
assumptions he has made.
All this may seem damaging enough for a theory of protection which
is to be valied for all countries and all times. Yet the fundamental
criticism remains to be delivered.
It is assumed that productive factors can move from industries with
low to those with high productivity. The benefit from protection comes
from the fact that it causes such a transfer. Why then does not such
a transfer take place without protection ? If a worker obtains a much
higher wage in the Roumanian coal industry than in most other Roumanian
industries, why do not those employed in the latter offer themselves to
the coal industry at somewhat lower wage rates than the present ones ?

This content downloaded from 128.143.23.241 on Mon, 29 Aug 2016 02:04:05 UTC
All use subject to http://about.jstor.org/terms
 Protection and Non-competing Groups oy

If they did, the Roumanian coal industry would be renumerative without

import duties. - The lower productivity in certain industries than in
others must mean lower prices paid to the productive factors employed,
for certainly nobody would suggest that there is only a difference in profits.
What prevents the underpaid factors from turning to industries that pay
better? Why does such a transfer take place only when the state puts
on import duties or gives other aid and not otherwise?
Evidently, the gain in ' 'production* ' which according to Mr. Manoïlesco
in many cases more than balances the loss from "commerce", when duties
are put on, has nothing to do with protection as such. He can attribute
this gain to protection only because he has made the tacit assumption,
that a transfer of productive factors from less productive
to highly productive industries is impossible without protec-
tion! Let us assume, on the contrary, that the productive factors are
mobile. They will flow over to the industries which pay them the highest
rewards. If the costs of production there are too high for competition
with foreign industries, the prices of the productive factors will fall -
but not so low as before in the less effective industries - and thereby the
competitive power will rise. Thus, the more effective industries can exist
without protection.
In the actual world the prices paid for the productive factors in a
country differ widely from one industry to another, and these differences
do not tend to disappear. The real problem, therefore, is this: What are
the causes of such differences? Does their existence, i. e., the use of the
productive factors which they imply, involve a "waste" or "loss" in some
way or other ? If so, is protection a good way of bringing about a more
effective use of the factors? Neither of these questions is analysed by
Mr. Manoïlesco. He does not discuss the first one and regards it as self-
evident that the answers to the other two are in the affirmative. As a
matter of fact they all stand in need of close analysis. A few observations
on them will be made below.

III.

Why do certain industries pay higher prices for the factors of pro-
duction than others in the same country ? Naturally, the qualities of the
factors used must be equal, if there is to be any meaning in speaking
about such differences in factor prices. For this reason it is not worth
while, from our present point of view, to give much attention to the prices
of natural resources; they are differently situated and must remain so
and this is analogous to a great difference in quality. Nor is it necessary
to say much about the differences in interest as they are comparatively
unimportant. It is the differences in wages which above all should be

This content downloaded from 128.143.23.241 on Mon, 29 Aug 2016 02:04:05 UTC
All use subject to http://about.jstor.org/terms
38 Bertil Ohlin

analysed. The fact that profits in

than in others also deserves attent
In a small country with good co
population like Denmark, the ann
with years of experience is about
lodgings is estimated in money. Th
while a worker in the manufactur
11 - 12 kr. per day. In Sweden th
provide food and lodgings for the
4 kr. per day, while manufacturing
income exceeding 3,700 kr. per y
other hand, workers in the same ind
may receive only 8 - 9 kr. in sma
The chief cause of such differen
the height of the costs of living.
Sweden are concerned, the local d
industry are so calculated in the
trade unions as to give compensat
rents, food prices etc. The low mone
due to the cheapness of living. They
farms, and pay low rents. As far as t
in the hands of a farm labourer is
in a manufacturing city in countr
Do nominal wage differences betw
inequalities in the costs of living,
duction in the national income ? Wo
in industries which are able to pa
is of course in the negative. One is i
incomes in terms of commodities,
the contrary, mean a loss if money
low in the places where costs of li
manufacturing industries. Produc
cheap country districts would be
the advantages offered by the latt
of feeding and housing the work
districts. - Regard the workers' f
It is right to consider the local d
material just as much as other cos

1 Why is it that profits and savings are s

than in agricultural countries ? With one or
from the manufacturing nations.

This content downloaded from 128.143.23.241 on Mon, 29 Aug 2016 02:04:05 UTC
All use subject to http://about.jstor.org/terms
 Protection and Non-competing Groups on

be uneconomical not to take into account the extra costs of

this raw material in the cities.
On the other hand, if the wage differences are greater than the extra
costs of supplying the workers with the goods and services he consumes,
then the national income in terms of goods is lowered. Factories which
ought to be placed in cities with good harbours etc. are driven out in
the country districts by the too high wage bills. And if we compare
agriculture and manufacturing industries: the development of the latter
is kept back. Workers who could be employed in manufacturing industries
at a real wage somewhat lower than the one paid there at present but
higher then the agricultural wage, are retained in agriculture. The only
condition under which no loss results from such a situation is if in manu-
facturing industries a somewhat lower wage than the present one would
cause a corresponding loss in the effectiveness of the workers, while
a rise in the agricultural wage would fail to raise the effectiveness of
the farm labourer. Certainly, there is no reason for assuming this to be
the case in countries where the standard of living among the manufactur-
ing workers is so high as in the Scandinavian countries. In countries, on
the other hand, where manufacturing wages are much lower than in the
former countries, the farm wages are so low that it is most probable
that an improvement in the food and lodgings of the agricultural population
would increase its effectiveness.
Thus wage differences which exceed inequalities in the costs of living
are undesirable, if one wants a high national income in terms of goods.
Which are the causes of such differences ? Consider first the case of different
manufacturing industries. That labour of the same sort is paid better in
the one than in the other industry, even when both are situated in the
same place or in places with equal costs of living, is no doubt due chiefly
to the influence of trade unions. They prevent the man with the lower
wage from offering his labour to the high wage industries at a cheaper
rate than that paid to the workers there. The workers are artificially
divided in non-competing groups. The high wage industry, where the
marginal productivity of labour is relatively great, is kept back, while
the low wage industry develops more than it would have done under
conditions of uniform wage rates. The national income in terms of goods
and services is reduced.1 It depends upon the circumstances in each

1 If the prices of the commodities are not fixed by conditions in other countries,
certain qualifications to the reasoning are necessary. If the low- wage industry sells its goods
cheaper than the prices at which they could be imported but exports nothing, an increase
in its prices would enable it to pay higher wages without much reducing its sales, the wage
bill being a small part of the total costs. Thus, the extra development owing to low wages
may be small.

This content downloaded from 128.143.23.241 on Mon, 29 Aug 2016 02:04:05 UTC
All use subject to http://about.jstor.org/terms
40 Bertil Ohlin

particular case, whether this influen

between various industries is great o
wages in an industry will have greate
demand for its products. For this rea
in export industries will reduce the n
much more than its practice in hom
paid to workers in the trade union m
commodity price, the decline in sales
is likely to be small.
Turn now to the relation between
industries. In most countries the stan
latter is higher than for those emplo
statement needs a qualification. It ass
depends only upon the quantity of th
consumes. This is a narrow concept,
the things a farm worker gets for no
with nature, etc., and which ignores suc
in the pleasure and satisfaction obtain
would compare the conditions of a cit
farmer and forget the "imponderabilia"
which fall to the latter and more or
disposal for clothes, movies, etc. In
imponderabilia weigh less than with
are not wholly absent. They are just
no objective measure is available for jud
attached to them in the personal bala
another.
The most practical way of dealing
seems to be to use the more formal
concepts, which takes into account o
are bought. The differences in real wag
facturing industries then depend par
which certain individuals have for th
mobility reduces the national incom
but it is obvious that the maximum n
"desirable" under these conditions. O
clined to say - also such statements
individual has a free choice between a h
industries and a lower real wage in ag
then the loss in individual and nation
the increase in wellbeing which is not
income concept.

This content downloaded from 128.143.23.241 on Mon, 29 Aug 2016 02:04:05 UTC
All use subject to http://about.jstor.org/terms
 Protection and Non-competing Groups aj

However, if the difference in real wages, in the opinion of many a

cultural workers more than outweights their preference for farm wor
then the lack of mobility is partly due to other causes. Personally, I h
little doubt that this is the case in many countries at present. F
workers, perhaps also some poor independent farmers, would lik
obtain work in manufacturing industries but cannot do so. The l
industries chiefly attract new labour during times of good busin
conditions when unemployment among their old workers is slight
most European countries, however, there have been no good busi
conditions in the last years. If a farm labourer succeeds in obtaining w
in a factory and then a reduction of the staff proves necessary some t
later, the likelihood that he as the newcomer will be dismissed is
great. The natural tendencies of employers to keep those who ha
served long as well as the policy of the trade unions tend to bring
this result. Thus, the burden of unemployment is more or less concentr
in the former farm worker, and his average annual income falls b
that of his fellow workers and may not exceed what he obtained in ag
culture with its more permanent employment. Such experiences natur
make other farm workers less willing to attempt a transfer to manufac
ing industries. In this way farm workers and factory workers ar
certain extent made to non-competing groups. Besides, the polic
mighty trade unions in some countries tend to bring about the
result. With different means they pursue a "closed shop" policy.
To the extent that the difference in real income is due to other
circumstances than the free choice of the workers, there is nothing to
offset the reduction in national income which ensues, unless account
is taken of "non-economic" circumstances, like the importance of being
self supporting in war times. Thus, unless there is some special reason
for acting otherwise, an increased mobility and a certain transfer of
workers to manufacturing industries is "desirable".
In my opinion, the case for a transfer is strengthened, not weakened,
if other things than the free choice of individuals is considered. Many
of them prefer to remain poor in agriculture because they are used to
the conditions here and imagine that they would not like the factory
work and city life. Man is a conservative animal. As a matter of fact
many of them would soon get used to the new conditions and would
consider the higher standard of living as factory workers as a great gain,
being still less inclined to go back to farm work than they are now to
leave it. Why is one to place more weight with their psychological reactions
before the change than after? The thing to compare is their life before
and after. If they are equally happy or unhappy with their work in both
situations, the higher standard of living in factory work represents a gain.

This content downloaded from 128.143.23.241 on Mon, 29 Aug 2016 02:04:05 UTC
All use subject to http://about.jstor.org/terms
 42 Bertil Ohlin

The belief they hold before the change

and life is then not an offsetting ele
In conclusion, farm workers and fact
form non-competing groups in many E
in real wages lead to a reduction in t
desirable unless some special view p

IV.

It may be pointed out in this connection that an economist like

Taussig, who like all orthodox analysers of international trade is so
interested in the questions of /'gain* ', says nothing about the loss which
comes from the existence of non-competing groups.1 He analyses a case
where the wages in the linen industry are 50 % higher than in the wheat
growing industry in the United States as well as in Germany, whereas
the wage level is in both industries lower in the latter than in the former
country. As the United States have a comparative advantage in wheat
production, they will exchange this commodity for German linen. Thus,
the United States export goods produced with relatively low wage labour,
and import goods produced with relatively expensive German labour.
The product of one American labourer is exchanged for that of one German
labourer, although the average wage level is much lower in Germany.
Under these assumptions the United States will produce nothing but
wheat and Germany nothing but linen. Taussig fails to see that this
means a net loss to the former country. The situation is illustrated by
the following table:2

Wages and Produce of 10 Days' Labour in the

United States and Germany (Dollars)

State e, . Wages per Total ~ . , w Produce Domestic Supply

State e, .
United States.... 1.00 10.00 20 wheat 0.50
„ „ .... 1.50 15-00 20 linen 0.75
Germany
„

The outcome w
at a price of $ 0
unit. Thus, every
However, the sam

1 Taussig, loc. cit., p. 43 sqq.

8 ibid., p. 47.

This content downloaded from 128.143.23.241 on Mon, 29 Aug 2016 02:04:05 UTC
All use subject to http://about.jstor.org/terms
 Protection and Non-competing Groups ao

of wheat for export could produce 20 units of linen, or 5 units

the quantity obtained from Germany. In other words, if Amer
workers are allowed to take work in the wheat growing in
wage of $ 1. 00 per day - which Taussig's reasoning assum
no American labourer is allowed to take work in the linen
a lower pay than $ 1.50 per day, the outcome is a serious r
the American national income.
Taussig's conclusion that "trade will develop as it would if prices
within each country were governed by labor costs alone"1 is false. It is
true that the United States will export wheat and import linen, but
the terms of exchange will be different. Had there been no non-competing
groups, anything from 1 to 1.50 units of linen would have been given
by Germany in exchange for one unit of wheat, more or less in accordance
with the conditions of demand. At present, assuming the 50 % wage
discrepancy to be fixed, anything from 0.67 to 1 unit of linen will be given
in exchange for one unit of wheat.2

V.

Can protection reduce the losses which follow from the existence
of non-competing groups, i. e. bring about a use of the factors of production
which resembles what it would have been in the absence of such groups ?
The answer is under certain conditions in the affirmative. In the case
above the United States would gain from an import duty on linen, which
prevents the import of this commodity from Germany. This duty would
have to be as high as 50 % to be effective, in spite of the small difference
in the supply prices on the table. For the scale of supply prices would
be raised in the United States and reduced in Germany, until the supply
price of wheat were as high in the former as in the latter country, e. g.
$ 0.60 per unit, and thus the inducement to export it to the latter country
disappeared. In Germany the supply price of linen would also be $ 0.60,
whereas in the United States the 50 % wage discrepancy would bring
it up to $ 0.90. Thus, the American linen industry would "need" an
import duty of 50 %• Of course, an American export duty of 50 % on
wheat would also make international trade stop. Evidently, these duties
counter-balance the effects of the non-competing groups in the United

1 Taussig, loc. cit., p. 48.

a It may seem peculiar that the terms of exchange can be different in the two cases,
as the United States exchange wheat for linen in both of them. It is, however, quite natural
for the non-competing groups mean a change in the conditions on which each country could
produce the import good for itself and a change in the wage level in the German ex-
port industry.

This content downloaded from 128.143.23.241 on Mon, 29 Aug 2016 02:04:05 UTC
All use subject to http://about.jstor.org/terms
 44 Bcrtil Ohlin

States, but not of those in Germany. T

the wages in the German wheat pro
15 units of wheat per 10 workers. Un
no trade if either (1) there are no non-
the United States, or (2) there is a w
import duty on linen or an export du
duties makes up for the "artificial" c
Let us leave these very simple num
problem in more general terms. If a
competing groups, duties might aim at
of industries as would have existed,
An industry which produces for the ho
power reduced through the forcing
see this power restored by means of an
in a similar situation as to labour
also, to be able to develop in the sam
mobile labour.
It goes without saying that the situation will not be exactly what it
would have been under other labour conditions. A cash bonus corresponding
to the extra labour costs would bring the situation more close to a ''normal"
one, than a system of duties can do. The import duty will raise the com-
modity prices, reduce sales and the number of workers employed, compared
with the "normal" situation. Thus, the wage level in other parts of
the labour market will tend to be depressed, and some industries will
develop more than they would otherwise have done. Yet, the fact remains
that a combination of non-competing groups and duties makes the line
of industries more like the "normal" one than if wage discrepancies
exist but no duties.
It does not follow that the existence of artificial non-competing
groups is a sufficient argument for protection. The more natural remedy is,
of course, to increase the labour mobility and do away with the watertight
labour compartment. Among measures which increase the mobility of
labour, one can mention a construction of the unemployment relief,
which makes it easier for a worker to support the relatively great unemploy-
ment during the first period of his work in a new industry. Protection
is likely to work the other way, i. e. to increase the wage discrepancies,
as it keeps a larger number of workers employed in an industry, where
wages are forced up, than would otherwise be the case. In other words,
it increases the temptation to pursue the policy further which is, partially
at least, the cause of the evil.
1 The post-war Australian tariff policy, which runs counter to all common sense, might
use this argument as a small, very small justification.

This content downloaded from 128.143.23.241 on Mon, 29 Aug 2016 02:04:05 UTC
All use subject to http://about.jstor.org/terms
 Protection and Non-competing Groups a 5

If protection is given for the reasons here mentioned, it must be only

to the high wage industries. The less well situated workers, thus, will
have to pay to these who are better off. Above all protection of a low
wage industry like agriculture is out of the question. Needless to say,
a policy of this sort has nothing in common with the policy of protection
pursued almost everywhere at present.
The existence of more or less non-competing groups of labour strength-
ens the case for temporary, not for permanent duties and only in very
special situations, e. g. in agricultural countries which have facilities for
certain manufacturing industries, like the textile industry, but because
of small labour mobility are unable to get a sufficient supply of labour
for such work unless a substantially higher real wage is paid than in
agriculture. So high a wage the new industries cannot pay in the be-
ginning. Protection may hasten their development by enabling them
to pay this wage and thus to overcome the friction due to small labour
mobility. For all practical purposes this comes to the same as the
infant industry argument for industrially young nations, which is,
however, usually based on the effects of protection in changing the
quality of labour rather than on its ability to cause a transfer of labour.
As a matter of fact both these aspects of the question are or can be
important.

This content downloaded from 128.143.23.241 on Mon, 29 Aug 2016 02:04:05 UTC
All use subject to http://about.jstor.org/terms