RETURNENG NATEQZALS:
I}V1ESI.J Place in book drop to

LJBRARJES remove this checkout from

‘uunzyuu-L. your record. fifl§§ will
be charged if book is
returned after the date
stamped be1ow.

 

 

 

 

I .
W.

Jim 18 “1999

 

 

 

CLASSICAL COMPARATIVE COST THEORY: TIME SERIES
AND CROSS-SECTIONAL EVIDENCE FOR THE

UNITED STATES. JAPAN. AND EUROPE

by

Andrew Soiocha

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
For the degree of

DOCTOR OF PHILOSOPHY

Department of Economics

I985

'l

'7')

V

gxgcaxux

ABSTRACT
CLASSICAL COHPARATIVE COST THEORY: TIME SERIES
AND CROSS-SECTIONAL EVIDENCE FOR THE
UNITED STATES. JAPAN. AND EUROPE
by

Andrew Solocha

The purpose of this dissertation was to investigate the
comparative cost advantage of manufacturing industries in
the United States. Japan. and Europe between 1967 and l980.
Since it was not possible to derive accurately all the costs
of production. the internal ranking of industries by
comparative advantage was approximated using the inverse of
unit labor cost and its components. labor productivity and
labor compensation.

The results for 1967 and 1980 indicated that the United
States held a comparative advantage. as approximated by unit
labor costs. in the Following industries: agricultural
manufacturing. paper, nonferrous metals. publishing,
industrial chemicals. plastics. office and computing
equipment. and professional goods. It held a comparative
disadvantage in traditional low technology products, wood
products. furniture. pottery. glass. rubber. iron and steel.
and most transport equipment.

The three major industries in the U.S. economy which
experienced a reduction in international competition between
1967 and 1980 were iron and steel. transport. and rubber. It

*was seen that these significantly increased their unit labor

costs relative to other manufactures. causing them to fall
sharply in the internal ranking. This was reflected in the
large trade deficits they incurred. In contrast. the iron
and steel. transport and rubber industries in Japan and
Germany were able to maintain their internal ranking.

In 1967 and 1980. Japan held a comparative advantage in
glass. iron and steel. nonferrous metals. transport
equipment. and industrial chemicals: a comparative
disadvantage in tobacco. food products. traditionally low
technology products. and professional goods. Japan's
comparative advantage remained relatively stable between
I967 and 1980.

In 1980. Germany held a comparative advantage in
tobacco. food products. wood products. pottery. and
professional goods; a comparative disadvantage in paper.
nonferrous products. metal products. rubber. industrial
chemicals. and machinery. Italy held a comparative advantage
in leather. forest products. and pottery. It displayed a
comparative disadvantage in tobacco. nonferrous metals.
rubber. plastics. and electrical machinery. Great Britain
held a comparative advantage in industrial chemicals.
plastics. tobacco. printing. and other nonmetal products.
Its comparative disadvantage lay in wearing apparel.
furniture. pottery and china. iron and steel. and transport

equipment.

To Janet and Anya

ii

ACKNOWLEDGEMENTS

The completion of this study depended on the assistance
and cooperation of many people. First. I would like to thank
my wife for her patience and encouragement. When writing a
disseration. patience is certainly a scarce commodity.
Second. I would like to thank my family for their support
(in more ways than one) throughout my graduate studies.

I am deeply appreciative of the guidance and
encouragement which I recieved from my Dissertation Advisor.
Professor Mordechai E. Kreinin. His attention to difficult
questions and prompt reading of the materials made the
disseration easier to complete.

Many thanks go out to the Economics Department for
their helpful suggestions and the use of the micro
computers. I valued the help from the fine librarians at
M.S.U.. who were instrumental in my research. Finally. I
must thank Elizabeth Johnston for editing this work.
although I am fully responsible for the errors which may
remain.

Although it appears to be our comparative disadvantage.

let there be peace on earth!

CHAPTER
ONE -

TABLE OF CONTENTS

INTRODUCTION 00..........OOOOOOOOOOOIOOOOOO.

TWO/K THE CLASSICAL COMPARATIVE COST THEORY . .....

THREE -

FOUR -

A.

,B.

C.
D.
E.
F.
G.
H.

I.
H.

The Ricardian Model ........ ......... ....
Limits to the Terms of Trade ............
Introduction of Money ...................
Limits to a Sustainable Exchange Rate ...
More Than Two Goods .....................
Unit Labor Costs ........................
Transportation Costs ..... ...............
Incomplete Specialization and Variable

Costs of Production .....................
Ricardian vs. Heckscher-Ohlin Models ....
Conclusion ..............................

REVIEW OF THE EMPIRICAL LITERATURE .........

A.
B.
C.

D.
E.
F.
G.

H.

MacDougall's Original Formulation .......
Tests of the Comparative Cost Theory ....
Additional Tests for Other Pairs

of Countries ..... ......... .............
Evaluation of the MacDougall Approach ...
Extension of the MacDougall Approach ....
Justification of Labor Costs ............
Comparative Cost Theory

Reformulated ....... .................. ...
Summary .................................

DISAGREGGATED-CROSS SECTIONAL EVIDENCE FOR
THE UNITED STATES ..........OOOOOCOOOOOOOOOO

A.
B.
C.

D.
E.
F.
G.

Data ....... ... ........ . ..... . ...... .....
The Models ........................ ......
Analysis of U.S. Manufacturing

Industries ..............................
Low Technology Industries ...............
Medium Technology Industries ............
High Technology Indutries ...............
Conclusion ............... ........ .......

iv

PAGE
I

6

30

30
33

39
42
45
49

SI
53
57

58
64

66
66

72
75

CHAPTER PAGE

FIVE - CROSS-SECTIONAL AND TIME SERIES EVIDENCE
FOR THE UNITED STATES. JAPAN. AND EUROPE ... 77

A. Data .................................... 78
B. The Cross-Sectional Models .............. 80
C. Time Series Evidence for the

United States ........................... 82

D. Pooling and Time Effects .... ............ 85
E. Conclusion .... ......... . ................ 89
SIX - ANALYSIS OF INDUSTRIES ..... ................ 95

A. Agricultural Manufacturing .............. 95
B. Traditional Low Technology Industries ... 97
C. Forest Products ......................... 98
D. Nonmetal Products ....................... 100

E. Medium Technology Industries .. .......... 101

F. Iron and Steel ..................... ..... 102

G O Nonferrous Meta I s O I O O I O O O O O O O O O O I O O O O O O 103

H. Metal Products ............ . ..... ........ 104

I. Rubber Products ......................... 105

J. Transport Equipment ....... . ............. 106

K. Publishing ..................... . ........ 107

L. Energy .................................. 107

M. High Technology Industries .............. 108

N. Industrial Chemicals ........ ............ 108

O. Other Chemicals ....... ....... ........... 109

P. Plastics ........ . ....................... 110

0. Machinery ............................... 111

R. Electrical Machinery .. ....... ........... 112

S. Professional Goods ... ......... . ..... .... 112

T. Summary ...................... ...... ..... 113

SEVEN - CONCLUSION ................................. 117
APPENDIX A ............ ........ ... ...... ... ........... 124
APPENDIX B ........ ........ ...... ........ ..... ....... . 139
APPENDIX C .............. . ............................ 158
BIBLIOGRAPHY ... ...................................... I64

LIST OF TABLES

TABLE PAGE
2.1 Introduction of a Multicommodity

Framework .................................... 15
4.1 Unit Labor Costs and Net Trade Positions

Of U.S. IndUStrIes’ 1967 and 1980 cease-coo... 60
4.2 OLS Results from the Cross-Sectional Data .... 65
5.1 U.S. Unit Labor Costs and Trade

Positions. 1967 and 1980 ..................... 90
5.2 Japanese Unit Labor Costs and Trade

Positions. 1967 and 1980 ................ ..... 91
5.3 German Unit Labor Costs and Trade

Positions. 1967 and 1980 ........ ...... . ...... 92
5.4 U.K. and Italian Unit Labor Costs and Trade

pOSItIonSFOr1980.........................o. 93

vi

LIST OF FIGURES

FIGURE PAGE
2.1 Unit-Value Isoquant ............ ........ ...... 24
2.2 Factor Endowments and Factor Prices .... ...... 25
5.1 Correlation without Time Effects .... ......... 86

5.2 Correlation with Time Effects ................ 86

vii

CHAPTER I

INTRODUCTION

The main objective of this dissertation is to
investigate the comparative advantage of manufacturing
industries in the United States. Japan. and Europe during
the period 1967 to 1980. A nation’s ranking of industries by
comparative advantage can be obtained from the inverse of
the production costs. Since it is difficult. if not
impossible, to derive accurately the actual costs of
production. unit labor costs are used as an approximation.

Most previous studies comparing unit labor cost and its
components. labor productivity and labor compensation.
across countries have limited their scope to one industry at
a given time. The problem with this approach is that a
country's labor compensation is expressed in a domestic
currency and must be converted to a common currency for
comparison by employing exchange rates. However. these are
often managed by central banks and deviate from their long—
run equilibriums. causing the analysis to be invalid. To
correct for these distortions. an approach developed by
Kreinin (1981. 1982. 1984. 1985) was extended to encompass
“the performance of unit labor costs in each industry
v-elative to the national manufacturing average. Therefore.

c:omparative cost advantages "are embedded in the total

economy”. shifting the focus to relative rather than
absolute unit labor costs (Kreinin. 1984. p. 42).

An important objective of this dissertation is to
construct internal rankings of industries approximating a
country"s comparative advantage at different times. 1967 and
1980. within each country. The exchange rate between the
countries separates the internal rankings into export
industries. nontraded goods. and import-competing
industries. The performance of unit labor costs is analyzed
to explain shifts in the rankings over time.

Comparative advantage plays an important role in
determining a country's trade flows. In light of the large
UzS. current account deficits recently registered.
industries which lose their comparative advantage face
severe international competition and often seek trade
restrictions. Since factors affecting a country's
comparative advantage can have important policy
implications. close examination of the performance of
relative unit labor costs is required.

Chapter II develops the theoretical foundations of the
comparative cost theory. beginning with the Ricardian
hypothesis. Money wages. a multigood framework.
transportation costs. incomplete specialization. and
variable costs of production are incorporated inlorder to
construct a more realistic model. These additional factors
provide an added dimension to the theory without adversely
affecting the results derived from the simple Ricardian

hypothesis. Differences and similarities between the

comparative cost theory and the He'ckscher-Ohl in model are
also examined. and anleffort is made to reconcile the two
theories.

The empirical literature which tests the Ricardian
hypothesis. beginning with the study by MacDougall. is
reviewed in Chapter III. Challenges to the Ricardian
hypothesis by Bhagwati. Kreinin. and Stern as well aa
supporting evidence for it by Daly and Stryker are
presented. Related material by Kreinin which does not
directly test the Ricardian hypothesis but which is based on
comparative cost theory is also reviewed in this chapter.

Chapter IV investigates the effect of unit labor costs
on the trade flows of 139 U.S. manufacturing industries
categorized by the Standard Industrial Classification (SIC)
system at the two-. three-. and four- digit level for 1967
and 1980. The first part develops an internal ranking for
these years and formulates a cross-sectional model
determining the effect of unit labor costs on trade
performance. The second part arranges the industries into
low. medium. and high technology categories and analyzes the
performance of their unit labor cost and its components.

Cross-sectional and time series evidence for the United
States. Japan. Germany. the United Kingdom. and Italy is
examined in Chapter V. A ranking of 28 manufacturing
industries. classified by International Standard Industrial
Classification (1510). is constructed for 1967 and 1980. and

a cross-sectional model is developed to determine the effect

of differential unit labor costs on trade performance. The
second part analyzes time series data for individual
industries to determine the effect of changes in labor
productivity'and compensation on the trade performance of
individual industries. In addition..all time series and
cross-sectional data combined with dummy variables are
"pooled” into one equation to remove the biases of trade
discrimination and exchange rate changes over time on trade
performance.l

Chapter VI integrates the statistical analysis of the
previous chapter and examines the performance of unit labor
costs and trade flows of individual industries. The
industries are grouped into low. medium. and high technology
categories. consistent with Chapter IV. as a basis for
analysis.

The final chapter summarizes the empirical findings of
the dissertation and offers a broader interpretation of the

main findings with regawd to methodology and economic

policy.

Eggtngtes

IThe ”pooling” or panel data models are based on models from
Kmenta (1972) and Judge et al. (1980).

CHAPTER II

THE CLASSICAL COMPARATIVE COST THEORY

This chapter develops the theoretical foundations of
the classical comparative cost model for ranking
industries by unit labor costs. The examination begins with
the Ricardian hypothesis. the restrictive assumptions of
which are relaxed by introducing money wages. a
multicommodity framework. transportation costs. and
incomplete specialization. In addition. similarities and
differences between the classical comparative cost model and
the Heckscher-Ohlin theory are highlighted. Within this
framework. comparative cost theory attempts tolanswer the
fol lowing three questions:

1. what is the basis for trade. and what goods should a
country export and import?

2. On what terms are the traded goods exchanged?

3. Should disturbances occur in the trade patterns.
what forces will bring about an adjustment?

This chapter wi l l concentrate on question 1. although the
other two questions will receive minor consideration.

The pure theory of international trade differs
structurally from positive theories in the kinds of
questions asked rather than in the kinds of assumptions
made. As Viner (1937. p. 437) has stated. ”the classical

theory of international trade was formulated primarily with

a view to its providing guidance on«questions of national
policy mu. Recognition of its welfare analysis orientation
is essential to the understanding and the appraisal of the
classical doctrine." Thus. a clear understanding of
comparative cost theory is important in formulating policy

fOr industries concerned with international competition.

The Riggrdign Mgggl

The first clear statement of the comparative cost
theory was developed by David Ricardo.l TheRicardian
model assumes:

1. Two countries and two commodities.

2. One factor of production - homogeneous labor.‘2

3. Labor is completely'mobile within a country» and all
occupations pay the same wage rate.

4. Labor is immobile between countries. so that factor
endowments cannot change.

5. Cormodities move freely between countries without
transportation costs.

6. The average productivity of labor is different in
the two countries.

7. Constant unit labor costs.

Ricardo assumedthat one of the two countries could
produce both goods using less labor. thus having an
absolute advantage in both goods. In his example. England is
able to produce a unit of cloth by employing 100 hours of
labor. or to produce a unit of wine by employing 120 hours
of labor. Portugal is able to produce a unit of cloth by
employing 90 hours of labor. or to produce a unit of wine

by employing 80 hours of labor.

Portugal is more efficient than England in the
production of both goods since Portugal can produce each
good with less labor than England. One English worker is
worth 90/100 of a Portuguese worker in the production of
cloth. and 80/120 of a Portuguese worker in the production
of wine..Although English workers are less efficient than
their Portuguese counterparts in the production of both
goods. England's degree of lower efficiency is smaller in
the production of cloth. Alternatively. Portugal's degree of
superiority is greater in the production of wine.

Thus as Ricardo has pointed out. although Portugal has
an absolute advantage in the production of both goods.
Portugal has a comparative advantage in the production of
wine. while England has a comparative advantage in the
production of cloth. The degree of comparative advantage can
be summarized in the following inequalities: 90/100 >
80/120 (comparing the same commodity in different
countries). or 90/80 > 100/120 (comparing different

commodities in the same country).

Limits :9 the Terms 9f Trggg

Starting from autarky. each country can increase the
production of one commodity only by switching some labor
from the production of the other commodity. 3 The cost of
cloth is the amount of wine that can be produced by the
labor necessary to produce an additional unit of cloth. In
the Ricardian example. a domestic exchange ratio (relative

price) in terms of relative costs will be established in

England of 1 unit of wine for 1.2 units of cloth. and in
Portugal of 1 unit of wine for 0.88 units of cloth. These
domestic cost ratios define the limits to the terms of
trade.

Taking account of the domestic exchange ratios. it
would be profitable for Portugal to engage in international
trade if 1 unit of wine could be exchanged for more than
0.88 units of cloth. For England to profit. less than 1.2
units of cloth must be exchanged for 1 unit of wine. Thus.
the limits to mutually'beneficial trade must lie between
(L88 and 1.2 units of cloth for 1 unit of wine. The domestic
cost ratios of the two commodities in the two countries
determine the limits to mutually beneficial trade (Kreinin.
1983). within these limits. each country should specialize
in the production of the commodity in which it has a
comparative advantage and obtain the other commodity through
international trade.

The classical comparative cost theory states that once
free trade is establ ished. the terms of trade wi l I settle
within these limits set by comparative costs. Both Mill and
Marshall concluded that. within these limits. the forcesof
reciprocal demand would determine the equilibrium terms of
trade. Thus. the burden of adjustment to a change in tastes
would fall primarily on changes in relative prices withlonly

secondary effects on responses in supply. 4

I0

W

The first significant expansion of the Ricardian model
was the introduction of money wages by Taussig (1927).
He elaborated on the three types of situations:
absolute cost. equal cost. and comparative cost. The
Ricardo-Mi ll model is easily elaborated in money terms since
labor is assumed to be the only input and the only
requirement is the money wage rate. As did the earlier
writers. Taussig uses two countries and two products.

The equal cost situation is 31/32 - PI/bz or al/bI a
az/bz. where:
a. - labor cost of one unit of good A in country 1:
a2 a labor cost of one unit of good A in country II:
b. a labor cost of one unit of good 8 in country I: and
b2 2 labor cost of one unit of good 8 in country II.
Since the cost ratios are the same. the domestic terms of
trade are the same in both countries. thus providing no
basis for trade.

when the condition 31/32 < 1 < PI/bz holds. the
principle of absolute cost advantage is displayed. With a. <
a2 and b. > b2. country I has an advantage in product A.
while country 11 has an advantage in product 8. If there is
to be trade between the two countries. a1 ' ”I < a2 ' H2 and
b1 ' H. > b2 * H2. The nelative wage ratio. w,/w2. can thus
fluctuate within the limits set by thelmoney costs of the
products. but not including w,/w2 - 32/31 or Nl/HZIBZ/bl.

If one country is superior in both commodities. but by
different proportions (a./a2 < bl/PZ <1). then we satisfy the

comparative costs principle. In order for country I to

II

export product A. a. * H. < a2 * H2. therefore. 1 > a!/a2 <
142/141. as with the absolute costs advantage example. This
allows the wages to fluctuate within limits and still
permit trade. For country II to export product 8. b1 * ”I >
b2 '1 H2. and by cross-multiplying. the result is 1 > b‘lbz >
HZ/Hl. cu- 1 > HZ/H.. Hl/HZ > 1. Thus. the wage ratein
country I. which may be considered the technologically
superior country. must be higher than in country 11 if there
is to be bilateral trading. The proportion of wages. wl/wz.
must be at least as high as the ratio bz/b1. but no higher
than aZ/a1.

Hhen NZ > H‘. a disequilibrium condition occurs. and
one country is cheaper. in money terms. in the production of
both goods and thus exports both. To correct for this
situation. the classical economists relied on Hume's specie-
flow theory as an equilibrating device. Since country I will
be exporting more because of the wage differential. it will
be acquiring more foreign currency and gold. The increase in
the country’s monetary base will increase its money supply
and drive up wages and prices. At the same time. country 11
will lose its currency and gold. thus contracting its money
supply. This will drive down wages and prices until a! ' H.
a a2 ' H2: after this process of adjustment. normal trade
conditions would be restored.

while the choice of the money wage is arbitrary. the
ratio of money wages between the two countries lies between
an upper and lower limit. as explained above. Only the

choice within these limits is arbitrary. The exact ratio is

IZ

determined by relative demand conditions. subject to the
constraint that total value of exports must equal the total
value of imports.

In examining the relationship between absolute
advantage and wage rates. Taussig finds that "those
countries have high money wages whose exported conmodities
command a good price in the world's markets". (Angel. p.
105). Following Senior and Mill. Taussig makes the range of
money incomes dependent on the terms of international demand
and the relative productivity of labor. In addition.
competitive markets are assumed in order for the marginal
productivity of labor to equal the wage rate.

Within this framework. the exporting industries in each
country are the primary determinant in setting the general
rate of money incomes and wages. But the exporting
industries set only nominal wages and not real wages.
Becauselof labor mobility. the money wage rate set by the
exporting industriesiwill become. under perfect competition.
the rul ing wage rate in the country at large (nonexporting
as well). In nonexporting industries. prices will vary
with changes in wages and incomes. as in the exporting
industries. but import prices will not reflect these
changes. Because of international trade. incomes would

rise. leading to lower real prices of imports.

imit ustai able Exch n e Rate
The limits to the exchange ratio can be derived from

the limits to the commodity terms of trade.5 The latter is

I3

expressed as an upper and lower 1 imit where mutual 1y
beneficial trade occurs. If both commodities in each country
are assigned prices in their own domestic currency. the
limits to the commodity terms of trade are converted into
the limits to a sustainable exchange rate.

Within the limits specified. the precise exchange rate
is determined by reciprocal demand subject to the balance-
of-payments equilibrium. If the exchange rate were outside
these limits. one country could undersell the other in both
commodities. This would lead to an external imbalance which
would require an exchange rate adjustment. Thus. the
exchange rate must reflect the cost/price relationship

between the two countries.

More Than Two 5

Frank Graham (1923) employed the classical model and
made it more realistic by incorporating a multicommodity and
multicountry framework. This section will discuss Graham's
contribution to trade theory in a multicommodity framework.

Graham. in the classical tradition. dealt with a single
input -labor- and constant costs of production. He felt.
however. that concentration on two countries and two
commodities severely limited the scope of the Ricardian
analysis. Thus he states that ”the classical theory of
international values seems to be open to grave objections
which. whi le they do not subvert its foundations.
nevertheless call for a substantial modification of its

conclusions”. (1923. 99. 54-5). Within a Ricardian two-

I4

country. two-conmodity model. adjustments in international
demand are not found in changes in production (supply).
Instead the adjustment process is a reciprocal demand
phenomenon. reflected by changes in trade volumes and the
terms of trade. Graham's emphasis. in contrast. is directed
toward changes in supply. involving shifting of productive
resources and altering the composition of trade. With
additional countries and comnodities. adjustments to changes
in demand incorporate shifts in production. along with the
possibility of a change in commodity composition. These
shifts in output and alterations in trade COMpOSItIOfl and
direction would tend to minimize the changes in the terms of
trade and thus are a long-run phenomenon.

In a rigorous fashion. let the number of units of
labor-cost needed to produce a unit of the goods A. B. .....
F in country I be denoted by 31' b‘.u.. f1 and a2.
b2.....f2 in country 11. If R is the exchange rate. then (a,
" W1 " R) < (a2 ' W2) holds for any commodity A that
country I exports.

A good will be exported only if its supply price (money
cost) is lower than in the foreign country. For any
comnodity B that is inported by country I. the condition (b.
" W. " R) > (b2 ' W2) holds. Then al/az < [Hz/(W! " R)].
where WZ/(W. ' R) is the ratio of money wages.

Thus. 51/32 < b‘/bz means that country I enjoys a
comparative advantage over country 11 in production of

commodity A. We can then arrange the various goods in the

15

order of comparative advantage of country 1 over country 11
in goods A. B. C. D. and F in the fol lowing way:
31/52 < Pl/Pz < cllcz < d1/d2 < e‘lez < f‘lfz.

We can then draw a line dividing the commodities which
country I exports from those it imports by the quotient
"2/("1 ' R0. To determine the exact position of the dividing
line. we>must introduce reciprocal demand. subject to the
balance-of-payments constraint.

This principle is illustrated in Table 2.1. For trade
to be possible. country 11 will export product A. and
country I will export product F. At least three of the
remaining four products could also be traded. depending on
the ratio of money wages. W2/(W‘i‘R). If Wz/(WI’R) :- 1.2 is
chosen. country 11 will export products A and 8. since in
country I compared to country II. the input cost of these
goods is proportionally greater than the proportion of the
money wage ratio. Wz/(Wl'R). The money cost of product C is
equal in both countries. and thus this product wi l i not be
traded. With the same analysis. country I wi l I export

products D. E. and F.

Table 2.1. Introduction of a Multiconmodity Framework

 

Labor Cost Per Unit

 

Goods Country 1 Country 11 1/11
A 14 10 1.4
B 13 10 1.3
C 12 10 1.2
D 11 10 1.1
E 10 10 1.0
F 9 10 0.9

I6

Since product C wi l i be produced by both countries in
order for there to be a balance in trade. the total money
value of country 11's exports of A and 8 must equal the
total value of its imports in products 0. E. and F from
country I. If this state of equilibrium is upset by a shift
in demand in country I1 away from its domestic goods
(A.B.C) to foreign goods (D.E.F.). a disequilibrium in the
balance of payments will occur. To finance these additional
imports. country 11 must export an apprOpriate amount of
product C. Since full employment of a constant amount of
labor is assumed. there must be a reduced consumption of
goods A. B. and C in country 11 to supply additional
exports.

In moving from the initial to the new equilibrium.
there does not have to be any change in the production of
any of the six products. The entire adjustment process could
consist of country 11 consuming only a portion of product C
and exporting the rest. meanwhile importing more of products
0. E. and F. Throughout this process of adjustment. all
prices and the terms of trade would remain unchanged. If the
terms of trade were to move against a country because of
excessive exports of one commodity. it would export other
commodities in which its comparative advantage was less and
stop the adverse movement of the terms of trade.

Graham viewed the terms of trade as normal ly equal to
the cost ratio in one of the countries. with at least one

commodity (intermediate) produced in common by more than one

I7

country. This intermediate commodity could serve as a link
between the two countries' cost structures. In a three-
commodity. twp-country example. each country could produce
the commodity in which it had the greatest comparative
advantage as well as the intermediate commodity. Labor costs
in each country would then link the price of the
intermediate commodity to each of the other two commodities.
A change in world demand would then cause a reallocation of
labor in each country between the two goods it produces
without any change in the terms of trade.

Although the multicommodity framework was first
designed for comparing the cost ratios of producing the same
commodity between different countries. Viner (1937) and
Takayama (1972) pointed out that comparing the cost ratios
of different commodities within the same country produces
the identical pattern of comparative advantage. This was
illustrated in the first Ricardian example. 6 Thus. it is
possible to obtain an internal ranking of all industries by
their cost ratios within each individual country. yielding a

ranking by comparative advantage.

nit r Co t
All commodities produced by a country can be ranked
internally in the order of their domestic costs of
production. 7 These are defined as the ratio of total
compensation per unit of input to productivity per unit of
input. Each country exports the commodities in which its

comparative advantage in terms of production costs is

18

greatest and imports those commodities which rank lowest
(highest production costs). The cutoff point between what is
exported and imported depends on reciprocal demand
considerations subject to the balance-of—payments
equilibrium.

This ranking system can have important implications for
economic policy; If an industry"s costs of production are
high relative to other industries. the industry’would be
ranked low in terms of comparative advantage. For an
industry to improve its ranking by comparative advantage. it
must reduce factor compensation and improve factor
productivity.

Although there are many factors of production. a
ranking of industries based on labor costs of production is
important. Takayama (1972) states that if fixed coefficients
of production areiassumed. comparative labor costs can be
meaningful even in a multifactor model. Since all factors
are used in fixed proportions. they can be represented by
only one factor. labor. Thus. a fixed coefficients
production function can be represented by a labor theory of
value.

The conditions under which a fixed coefficients
production function and a labor theory of value display the
same properties were first developed by Samuelson (1951).
These conditions. which hold true even if each industry has
a choice of several methods of production are as fol lows:

1. All methods available to any one industry have only

one and the same output (the product of that
industry).

I9

2. All methods in the technology have among their
inputs one and only one scarce primary commodity
(labor). drawn in from the outside. and this
commodity is the same for all methods and
industries.

An elementary exposition presented by Koopmans (1953)
states that any efficient nonnegative bill of goods (net
outputs) can be sustained by only one price constel lation
(by one ray of prices ka. kp2..... kpn. where k > 0). Given
this price constellation. even though many processes of
production may be available to each firm or industry. only
one such process will actually be used.Thus. the only costs

that need to be considered within this framework are labor

costs.

Transggrtgtion Costs
The introduction of transportation costs can lead to a

third class of goods. nontraded goods. which enter only into
domestic trade. To obtain the conditions for nontraded
goods. ‘51:12 is denoted to represent the real cost in units
of labor for transporting commodity A from country 1 to
country'II and at21 for transportation from country 11 to
country I.

If the country exporting the commodity pays the cost of
transportation. commodity'A will be exported only if
(a. + atlzvaz < WZ/(Wl ' R) and imported if 31/(32 + at21)
< WZ/(Wl ' R). But if W2/(W‘ ' R) lies between (a. +
at12)/32 and (a, + at12)/a2 such that a./(a2 + at21) <
WZ/(W‘ ' R) < (a1 + at12)/°2' commodity A will not be

exported or imported unless the difference in its cost of

20

production between the two countries exceeds the cost of
transporting it from one to the other.

The export performance of a country now depends on its
comparative costs of production and its transportation
costs. This would cause the international division of labor
to be carried out to a lesser degree than with zero

transportation costs.

Incgmglete Specialization and Variable Cgsts gf’Pnguction
Classical theory held that partial specialization could

not exist. As Mill (1909. p. 589) stated. "cost of carriage
has one effect more. But for it every commodity would (if
trade be supposed free) be either regularly imported or
regularly exported. A country would make nothing for itself
which it did not also make for other countries.” Even if no
transportation costs and constant costs are assumed. it is
possible that only one country would specialize completely.
while the other country produces both commodities. As Graham
has pointed out. this would occur when one country was
economically small and the other large. Here. the first
country could not by itself meet the total demand of both
countries for the good in which it specializes. Thus. the
large country would have to produce both commodities to
satisfy its demand. being incompletely specialized. while
the small country is completely specialized.

The assumption of constant unit cost can also be
relaxed. Increasing costs (diminishing marginal returns) and

decreasing costs (increasing marginal returns) can be

21

examined in a comparative cost framework without negating
the general results of the model.

To relax the assumption of constant costs. other
factors of production must be introduced. The assumptions of
the labor theory of value can easily be replaced with the
opportunity costs: assuming the availability of all its
resources. a country can shift some of them from one line of
production to another. 8 Then the price ratio (or rate of
exchange) between commodities can be obtained from their
opportunity costs. The opportunity cost of steel is the
amount of texti le that must be given up to get an additional
amount of steel regardless of whether it were labor.
capital. or some other factor that were shifted from textile
to steel production. Its Opportunity cost is the amount of
steel that must be sacrificed for an additional quantity of
textile.

In a constant cost model. there is a constant
rate of substitution in production between any two
commodities. In a multifactor model with diminishing
marginal returns. the substitution in production of the two
commodities becomes limited so that the possibility of
incomplete specialization emerges as the general result of
the model. In addition. the division of labor will be
carried out less than under constant costs. since the
comparative disadvantage at the margin diminishes and
finally disappears.

Graham takes this one step farther and states that

22

because trade leads to the development of special ization.
the combined output may be smaller than if there were no
trade at all. This condition would occur when a country's
exports are produced under diminishing returns. while non-
traded and import competing goods are produced under
increasing returns. This is provided that the increase in
total money values of production in one country is less than

the decrease in the other country.

i H k h r I

Jones (1979) argues that the sharp distinction often
drawn between the Ricardian comparative cost and the
Heckscher-Ohlin models can disappear when a multicommodity.
multicountry framework such. as Graham's. is employed. The
Ricardian and Heckscher—Ohlin models are distinguished from
each other by the assumption concerning a commonly'shared
technology. The Ricardian model assumes that the average
productivity of labor is different between countries beeause
they do not share the same common technological knowledge.
If countries do not share the same technology. then factor
prices will not equalize as in the Ricardian two by two
model. The Heckscher-Ohlin model assumes that countries
share a common technology. and thus factor prices do
equalizeu However. as will be shown in the analysis. this is
only a necessary and not a sufficient condition for factor
price equalization. By employing a multicocunodity framework.
it is demonstrated that factor prices need not equalize.and

thus the sharp distinction of the two models fades.

23

The Ricardian and Heckscher-Ohlin models have two
significant distinguihing features which are not dependent
on the number of commodities and countries considered. The
first is that the Ricardian model completely ignores
differences in factor intensity. Homogeneous productive
resources are converted into outputs at constant costs. so
that in the Ricardian model it is necessary to concentrate
on only one factor of production - labor.‘The Ricardian
model also assumes that countries differ in the average
productivity of labor in the same occupation without
specifying the cause. whereas the Heckscher-Ohlin model is
based on an assumption that knowledge of the best productive
techniques is available»tolall countries.

Jones concentrates on the role of factor endowments in
determining production and trade patterns by considering two
smal l countries which share access to the same technology
but differ in their endowments of capital and labor. Unit-
value isoquants are constructed for five commodities for
these countries. Each unit-value isoquant shows all
combinations of capital and labor that can produce 31 worth
of output of that particular good. The country's capital-
labor endowment ratio would dictate the location of
production along this composite unit-value isoquant and the
country’s wage/rent ratio. These unit-value isoquants are

represented in Figure 2.1.

24
Figure 2.1. Unit-value Isoquant

CAPITAL: L.) \_

I

 

LABOR

The fol lowing information about possible production
patterns for this small country can be derived from
Figure 2.1. (I) Some commodities will never be produced by
this country because its technology in these commodities is
inferior to that prevailing elsewhere in the world. This is
illustrated by commodity 5. (2) Certain techniques for
producing some comnodities wi l i never be observed regardless
of the community's endowment base. This is illustrated by
point G.

If a ray from the origin representing the country's
capital/labor endowment ratio cuts the locus between points
0 and E in Figure 2.1. then this country would specialize in
commodity 3. It would then export commodity 3 in exchange
for al 1 the other commodities. If the economy had a higher
endowment proportion of capital. so that a ray from the
origin to the endowment point cut the locus between E and

F. the country would produce both commodities 3 and 4. and

25

import commodities 1.2. and S. In this way. the pattern of
production depends on factor endowments.

Figure 2.1 can be used to illustrate the wage/rent
ratio that would correspond to any factor-proportions. The
factor price ratio is shown by the slope of the inner locus
ABCDEF at the point where it crosses the factor endowment
ray. Figure 2.2 traces out the relationship between factor
endowments and factor prices for a given set of world

prices.

Figure 2.2 Factor Endowments and Factor Prices

WAGES/I
RENTS

 

 

 

 

w
\

 

CAPITAL/LABOR

At a low capital/labor ratio. the wage/rent ratio would
also be low. and the country would specialize in commodity
I. If the country"s capital/labor endowment ratio increases.
it would now produce both commodities 2 and 1. For local
variations of the capital/labor ratio near this production

point the country remains incompletely specialized in 2 and

26

1. and factor prices are equalized at this level.
corresponding to a two-by-two model.

From this analysis. it can now be determined whether
two small countries facing the same set of world prices and
sharing the same technological knowledge will have their
factor prices equalized. As illustrated in Figure 2.2. even
when two countries share the same technology and are
incompletely special ized in two commodities. their factor
prices need not be equalized. This would occur when one
country produces commodities 2 and 3. and the othercountry
produces commodities 3 and 4.

Although production functions may be identical. as
assumed by the Heckscher-Ohl in model. the techniques
actually adopted to produce commodity 3 will differ between
countries because the wage/rent ratio is higher in the
relatively capital-abundant country. The difference between
countries in comparative labor productivities that is
asserted by assumption in the Ricardian model can be deduced
from differences in relative factor endowments in a
Heckscher-Ohlin model. Thus. there is very little difference

between the two models in a multicommodity framework.

anclugion
This chapter developed the theoretical foundationsof

the classical comparative cost model beginning with the
Ricardian model. The assumptions of the latter were relaxed
with the introduction of money wages. a multicommodity

framework. transportation costs. and incomplete

27

specialization to construct a more realistic model. With the
relaxation of the assumptions. Jones demonstrated that the
sharp distinction often drawn between the Ricardian and
Heckscher-Ohl in models disappeared when a multicommodity.
multicountry framework was employed. since factor prices
need not equalize. The difference between countries in
comparative labor productivities that is asserted by
assumption in the Ricardian model can be deduced from
differences in relative factor endowments in a Heckscher-
Ohlin model.

Viner (1937) pointed out that comparing the cost ratios
of different commodities within the same country was valid
if the ratios between costs are examined..All commodities
produced by a country can be ranked internally in the order
of their domestic costs of production. Each country would
export the commodities in which its comparative advantage in
terms of production costs is greatest and import those
commodities which rank lowest (highest production costs).
This provides the theoretical foundations for ranking
industries in the subsequent chapters.

The empirical tests of the comparative cost theory are
reviewed in the next chapter. beginning with MacDougal l's
analysis of the Ricardian hypothesis. Challenges to the
Ricardianihypothesis and the subsequent relaxation of its

assumptions are also examined.

28

F Qtnotgg

1The question of whether Ricardo can claim the credit
for establishing the theory has been debated in the
literature. The controversy stems from the fact that Colonel
Torrens published an analysis of comparative advantage in

1808. The Economists Bgfute . and in 1815. An Essay m the

Extgrml an Trage. while Ricardo's Pringiglgs appeared in
1817. However. Viner (1937) believes that despite Torrens'

earlier publication. Ricardo is entitled to the credit for
being the first to place an euphasis on the theory. the
first to place the theory in an appropriate setting. and for
obtaining a general acceptance of it by economists.

In a similar vein. Chipman (1965) states that although
Ricardo was influenced by Torrens's Extgrml Qgrn Mg.
Ricardo added a great deal to the theory. He also states
that both Torrens and Ricardo contributed in essential ways
to the development of comparative advantage. but Torrens
should be given credit for the principal discovery.

2The early comparative cost theories relied heavily
on the labor theory of value. Haberler (1936) showed that
the deductions obtained from using the labor theory of value
do not depend on the validity of its assumptions. Thus it is
possible to discard the labor theory. along with its
unrealistic assumptions. without having to discard the
results obtained from it.

3This section is based on Kreinin (1983). chapter 11.

4In contrast. Graham (1923) objected to the normal
equilibrium position lying between the cost ratios and
cal led this a "l imbo" ratio. Graham demonstrated conditions
under which the terms of trade would 1 ie at one of the
limits set by comparative advantage. This would occur when
the two countries were of different economic sizes. and the
economic importance of the two products were not the same
in each country. Thus the demand of country A for country
8 products would not be the same as the reciprocal demand.
Since these two demands will not be balanced. the terms of
trade will be forced to one of the limits set by
comparative costs. The appearance of a balance inside
these limits would only be a slight coincidence.

5This section is based on Kreinin (1983). chapter 11.

6If we consider two countries. m and n. each producing
goods 1.2."..i. then their respective unit labor costs can
be represented as lim and lin. The degree of comparative
advantage can be expressed as the ratio
[(l Avail ))/(i lAvoil )] -
[(I:mél1nfl'(£b9(lgag/Avoil’2)] - C. The industries in each
coungry are examined relatIve to the manufacturing average.
If C > I. then country m has a comparative advantage in the

29

production of good i. If C < 1. the country n has the
advantage. Thus. the unit labor costs are first ranked
internally and then international comparisons are
made. See Deardorff (1980) and Kreinin (1982) for the
justification of using the manufacturing average as the
denominator.

7This section of the reformulated comparative costs
theory is based on Kreinin (1983). chapter 11.

BViner (1937. p. 490) is opposed to the linking of the
labor theory to the classical theory of comparative
advantage. ”Except for Ricardo. none of the classical
expounders of the doctrine of the comparative costs. with
the relatively uninportant and partial exception of James
Mil l. was an exponent of a labor-cost theory of value."
Viner also states that ”the association of the comparative-
cost doctrine with the labor cost theory of value is a
historical accident.”

CHAPTER I I I

REVIEW OF THE EHPIRICAL LITERATURE

Classical trade theory states that international
specialization is determined by comparative advantage based
on relative productivity differentials among countries. It
does not attempt to explain why the differences occur:
rather. the assumption is that they exist and make trade be
possible and beneficial. Relative differences in the wage
structure and capital-labor ratios of some industries may
distort the relative rankings determined by productivity
differentials. Nevertheless. labor productivity captures the
greatest portion of the variations. and hence the classical
economists believed that additional factors were not
sufficiently important to cause significant changes in the
pattern of international trade as determined by relative
labor productivity differences.

The classical comparative cost theory was tested
empirically several times. beginning with MacDougall. but
these tests had limited success. A review of this empirical

l iterature fol lows.

MacDougal 1'; Original Formulgtign

Using Rostas's productivity data for twenty-five
British and U.S. industries. MacDougall (1951.1952) tested

selected aspects of the classical comparative cost theory.

30

31

The empirical results were based on the Ricardian labor
theory of value. Assuming a two-country world. Great Britain
and the United States. each country will have a comparative
advantage and export those goods for which its relative
output per worker. (U.S./U.K.). exceeds its relative money
wage rate in the respective industries. However. bilateral
trade flows could not be uti l ized because the bulk of U.S.
and U.K. exports in 1937 went to third markets. which became
the basis for the tests.

According to pre-World War II statistics from the
Ministry of Labour and the Department of Labor. U.S. weekly
wages in manufacturing industries were on average double
those in Great Britain. Where U.S. output per worker was
more than twice that of the British. the United States was
expected to dominate the export market. and where the U.S.
output per worker was less than twice that of the British.
Britain was expected to dominate the export market. The
empirical results indicated that 20 of the 25 industries
(covering 97 percent of the sample by value) conformed to
theoretical expectation. MacDougall found a strong inverse
relationship between relative wage costs per unit of output
(relative wages/output) and relative exports. and a positive
relationship between relative output per worker and relative
exports. When relative output per worker was plotted against
relative exports on a double logarithmic scale. the
regression line had a lepe of almost four and the resulting

R2 was 0.64. This indicates that a one percent difference in

32

relative output per worker for a commodity was associated
with 3-4 percent higher exports for the country with the
greater productivity level.

In the traditional Ricardian framework. with perfect
markets and homogeneous products. the exports of each
country should go to the other. Where either country had a
comparative advantage. however small. the country would
capture the whole export market. Instead. the majority of
U.S. and U.K. exports in 1937 went to third markets. with
little trade between themselves. In industries where Britain
had the comparative advantage in 1937. U.K. exports to the
United States were only a fraction of one percent of total
LLS. consumption in that commodity. Similarlyu in industries
where the United States had the comparative advantage in
19373 LLS. exports to Britain were more substantial. but
still only a small percentage of total British consumption
in that commodity.

One explanation is that the assumptions of the
Ricardian model do not hold in the real world. Instead. with
imperfect markets. each country obtained the larger share of
the third market where it had the comparative advantage. and
each countrytended to capture a larger share of the third
market the greater its comparative advantage. As theory
would suggest. this was the expected result when the
assumptions are relaxed. confirming the labor theory of
comparative costs.

The existence of high tariffs is a second explanation

for the lack of penetration of the U.S. and U.K. domestic

33

markets. In 1937. U.S. tariffs fully offset Britain's
comparativeladvantage in almost every commodity with the
exceptions of cement and coke. where transportation costs
are high. and footwear. where the British advantage was
smal l. U.S. tariffs were much higher on commodities where
Britain had a comparative advantage than on other
commodities. British tariffs fully offset U.S. comparative
advantage in paper and glass. while they only partly offset
tLS. comparative advantage in machinery. motor cars.
wireless sets and values. pig iron. and tin cans. But by
1950. MacDougall (1962) found a general reduction in U.S.
tariffs. In particular. this was most noticeable for the
commodities in which the United States had a comparative
disadvantage. In contrast to the 1937 LLS. tariff rates. the

1950 rates seldom offset Britain's comparative advantage.

Tests of the r tiv C st Th r

Balassa (1963) and Stern (1962) continued and updated
MacDougall's original work. They examined the relative
importance of productivity. wages. and capital costs in
determining the pattern of exports for selected
manufacturing industries in the United States and Great
Britain. In addition. Stern examined the relative wage
structure and export performance of the two countries in
1959 to determine whether changes in comparative advantage
had occurred.

Paige and Bombach”s I950 productivity estimates for 44

manufacturing industries in the United States and the United

34

Kingdom were utilized by Stern. Balassa's sample covered
only 28 industries. representing 43.1 percent of the
manufacturing output in Britain and 41.4 percent in the
United States. The productivity'data were measured as net
output per worker: gross output minus purchased inputs other
than labor. divided by total employees. The relative wage
and output per worker measures are man-year figures. while
the employment figures include direct and indirect labor.

Net cost ratios were included as an explanatory
variable to measure capital costs and profit per unit of
output in addition to the unit labor costs. Thus. net cost
ratios could provide some indication of relative resource
productivity and could be important if the assumptions of
the labor theory of value were relaxed.

In 1950. the U.S. and U.K. mutual trade was still a
smal l percentage of the total trade of each. Since most of
the U.S. and U.K. trade was with third countries. Stern
treated the rest of the world as a third country. examining
total U.S. and British exports in addition to their mutual
trade.

Balassa. in contrast. excluded mutual trade between the
United States and Britain since the differential tariff
rates would distort trade patterns. Instead. the effect of
1950 relative productivity differentials on relative export
performance for 1951 in third countries was examined. Still.
corrections for the discriminatory effects of commonwealth

preferences and other distorting factors were not accounted

35

for.

In examining the impact of productivity differences on
export performance. the elasticity of substitution between
U.S. and British exports of the same commodity was assumed
to exceed unity. Thus. a positive relationship was expected
between relative productivity differentials and export
performance with the resulting regressions in logarithmic

form as fol lows:

.Belesse

log EI/EII . -1.76 + 1.59 log 91/911 R2 . 0.74; (3.1)
(0.18)

Stern

log 51/511 = -0.68 + 1.27 log PI/PII R2 - 0.19. (3.2)
(0.43)

where E is the export performance. P is productivity. I is
country one. and 11 is country two. To provide consistent
reporting. all regression results providing R have been
converted to R2.

The coefficients in both equations were significant
and have the correct signs. The difference in the magnitude
of the coefficients and R2 between the two equations is due
to Balassa's smal ler sample size. By excluding industries
which may be dominated by demand factors or are nontraded.
he was able to obtain a higher R2. In addition. Stern found
that relative wage differentials between the two countries
had narrowed between 1950 and 1959. When comparing the two
periods. each country experienced export gains in the

industries in which it had a comparative disadvantage. This

36

wasattributed to relative productivity increases in the
comparative disadvantage industries for each country.

In an attempt to improve the regression results. wage
ratios were added as a determinant of export performance.
The resulting regression results are:

fiauesse

log EI/EII = -5.16 + 1.46 logPI/PII + 1.25 logWI/WII. (3.3)
(0.33) (0.57)
R2 . 0.77
Stern incorporates wage ratios into unit labor costs by
dividing the output per worker ratios into the wage ratios.
The results are as follows:

Stern

log 51/511 = 0.01 -1.40 log y R2 s 0.18: (3.4)
(0.59)

where y is unit labor costs.

These results are similar to the ones obtained in
equations 3.1 and 3.2: however. although the wage variable
in equation 3.3 was significant at the 95 percent level. it
had the wrong sign. Despite the statistical significance of
log WI/WII. Balassa states that no definite relationship
between wage rates and relativerexport performance can be
established since the R2 did not differ significantly
between equations 3.1 and 3.3. 1 His reason is that
productivity differences were not offset by higher wages in
high productivity industries: instead. productivity
differences were the primary factors in determining export
performance. One reason wage ratios might not be an

important factor in explaining relative export performance

37

is that causation runs in both directions between the
variables. While lower wages could lead to improved export
performance. the latter may also cause wages to increase in
an industry.

Another reason wage ratios and export performance might
not be correlated is that wage rankings of industries are
similar among countries. A similar proposition. first
advanced by'Taussig. states that wage rankinglare similar
among countries because there is little competition between
the labor forces of various industries. Interindustry wage
differentials are determined by the disutility and
regularity'of'work. the required skill levels.land other
factors which are uniform in all countries: thus. wage
differentials should be smaller than productivity
differentials.

In empirical tests of this proposition. Caves (1960).
Kravis (1956b). Lary (1968). Lebergott (1947). and Balassa
(1963) have shown that the rankings of industries by hourly
earnings of workers are similar. Kravis compared the ranking
of average hourly earnings in 20 manufacturing industries
for Japan and the United States obtaining a coefficient of
rank correlation of 0.82: he concluded that ”for most
industries. international differences in productivity are
greater than international differences in wages" (1956b. p.
68). Lary analyzed average annual wages in 13 industry
groups for 11 countries. The ranking of industries from low
to high wages was found to be similar in the seven developed

countries (the United States. Canada. Sweden. Australia. the

38

United Kingdom. Germany. and France). but among the four
developing countries (Mexico. Japan. Brazil. and India) the
ranking was less consistent. Stanley Lebergott found that in
the post-World War 11 years. interindustry wage
differentials were almost identical among the United States.
the United Kingdom. and Canada. while differing slightly for
Sweden. This was brought out by Balassa. who obtained a
coefficient of variance of 37.1 for productivity ratios and
10.7 for wage ratios.

The last relationship explored is between relative
export performance and net cost ratios. where net costs
refer to per unit cost and profits in addition to labor

costs. The results are as fol lows:

Balassa

169 51/511 = 6.16 - 1.59 log NI/NII R2 = .50: (3.5)
(0.30)

£159.10

log 51/511 . 0.01 -1.41 log NI/NII R2 - 0.13: (3.6)
(0.59)

where NI/NII - relative unit costs. The results are also
similar to equations 3.1 and 3.2 but with a higher standard
error and lower coefficient of determination.

These results confirm the classical theory. indicating
that the relative export performance of the United States
and the United Kingdom fol lowed establ ished lines of
comparative advantage as suggested by labor productivity.
The introduction of additional variables seemed to modify

the results slightly. while differences in capital costs per

39

unit of output did not seem to have any significant effect

on export performance.

Additional Tests for Other Pairs of Countries

Kreinin (1969) provides a critical analysis of the
comparative cost theory as formulated by MacDougall. He
expands the analysis by examining cross-sectional data for
three pairs of countries other than the United States and
the United Kingdom. Canada and Australia. Canada and the
United Kingdom. and the United States and Canada were
examined for the years 1948 and 1950-1951. Time series data
were also examined for the United States and the United
Kingdom over the period 1958-1965.

Intracountry trade was negligible in the cross-
sectional analysis for Canada and Australia because they
serve different markets due to the distance involved and
other traditional factors: thus. exports to third markets
had to be considered. At the official exchange rate
prevailing in 1950. average earnings per employee in
Canadian manufacturing industries were double those in
Australia. According to the Ricardian hypothesis. Canadian
export volume should exceed Australia's in industries where
the productivity index islebove 200 percent. andlCanadian
export volume should fall below Austral ia's where the ratio
is less than 200 percent. However. more than half of the 20
industries examined performed contrary to theory.

Relative wage costs were derived by dividing the

productivity ratio into the wage rate ratio. Relative

4D

exports were then correlated with relative wage costs. with
the expectation of a negative coefficient. However. this
relationship was insignificant. Kreinin attributed this
to the lack of variation in the interindustry wage structure
in the two countries and in the ranking by comparative
advantage. An additional factor is that the dependent
variable. relative export. is limited by the fact that
Canada and Australia have few common markets. Thus.
extension of MacDougall’s analysis to this pair of countries
is questionable.

In the analysis for Canada and the United Kingdom.
relative exports were not found to be correlated with either
relative costs or price variables. However. the regression
was severely limited by its small sample size. and no
conclusive results are possible.

The UAL and Canadian results indicated that a large
share of their industries. total trade was intracountry.
Thus. a study of this pair of countries provides a close
approximation of the simple Ricardian two-country model.
Unlike that model. incomplete specialization due to
increasing costs. product differentiation. and trade
barriers prevented post-trade commodity prices from being
equalized internationally. These factors also caused trade
in each commodity group to flow in both directions. Taking
this into consideration. intracountry exports were found to
be correlated with labor productivity and labor costs. The

results expressed in logarithmic form are as follows:

41

x. = 0.37 + 5.50 yl R2 = 0.43; (3.8)
(1.76)

x. = 1.37 - 6.68 y2 R2 = 0.38: (3.9)
(2.24)

where x. is intracountry exports. VI is the labor

productivity ratio.land Y2 the labor cost ratio.

These results adhere more closely to the Ricardian
hypothesis than do the previous results. Nevertheless. this
sample deleted one fourth of the industry groups: when all
21 industries were included. the R2 dropped to»0.2.1flwile
the parameters became statistically'insignificant.

Time series data was also examined for the United
States and the United Kingdom in 25 industries for the
period 1958-1965. Changes over time in relative labor
productivity were correlated with changes in relative export
performance to third markets. with the expectation of a
positive coefficient. Only six industries exhibited
significant positive relationships. while eight exhibited
negative relationships. and eleven exhibited none. Three of
the six which showed strong positive relationships -
leather. pulp. and glass - are characterized by homogeneous
commodities. The problem is that with only eight annual
observations there are not enough degrees of freedom on
which to base any definitive conclusions.

The MacDougal l—Stern analysis was then integrated by
examining the common commodities in both studies. Changes
over time for relative labor productivity were then
correlated to changes in relative export performance for the

United States and the United Kingdom over the years 1937.

42

1950. and 1959. The results indicate that one-third of the
sample did not COHFOWflitO expected behavior. Because this
analysis is based on one or two degrees of freedom. no
definitive conclusions can be reached.

Finally. time series data was examined for automobiles
in six EurOpean countries for the years 1955. 1960. and
1965. When changes in relative labor productivity were
correlated with relative export performance for sets of
these European countries. Kreinin found little or no support
fer the Ricardian hypothesis.

Kreinin concluded that his evidence did not support the
simple Ricardian hypothesis. Rather. factors other than
supply. such as product differentiation and demand.
determine the pattern of trade. For the case of
differentiated commodities. labor productivity may only set
the limits to the terms of trade. whereas demand conditions
determine the exact terms of trade. The major constraint to
using labor productivities for differentiated commodities is

that the limits to trade may be wide.

Evaluation‘gf the MacDougall Ageroach
Bhagwati (1969) and Stern (1975) were critical of the

 

MacDougall-Balassa-Stern (M-B-S) approach because these
early empirical tests (including Stern’s) were too
simplistic. Since M-B-S assumed labor was the only
significant factor of production. measures of comparative
labor productivity were employed as a proxy for comparative

costs. Bhagwati and Stern cal l for a general framework for

43

determining trade patterns and are skeptical of the simple
Ricardian approach. finding it difficult to adapt the one-
factor Ricardian model to a multifactored world.

Stern and Bhagwati state that treating labor
productivity as an exogenous variable and as the most
important determinant of comparative cost differences is
limiting: therefore. concentration on labor productivity
instead of other factor productivities is purely arbitrary.
If human capital and other capital inputs into natural
resources are considered. comparative capital productivity
could be significant. and a multifactor model should be
adopted. Thus. tests of the Ricardian hypothesis should be
based on intercountry differences in total factor
productivity. not just labor. Nevertheless. because of the
increasing importance)of'multinationals in international
trade. capital inputs have become mobile among countries and
its price equalized. If there is little variation in capital
prices. differential labor costs remain the important
factor.

Bhagwati and Stern state that since the Ricardian model
emphasizes comparative differences in production functions.
tests of the Ricardian model should be developed in terms of
comparisons of production functions. The expected results
are that a country"s exports would be concentrated in its
relatively most efficient industries and its imports in its
least efficient ones. Along these lines. Minhas (1962) has

compared the isoquants of several countries and found that

44

except for a pure scale factor. all the isoquants had
essentially'the same parameters. Bhagwati and Stern conclude
that Minhas has devised a method for testing the Ricardian
hypothesis in a more general manner. Stryker (1968) attempts
to incorporate this approach in testing the comparative cost
model. The problemiwith this method lies in the choice of
the production function to estimate. Since different
production functions exhibit different properties. the
choice of any particular type wil l influence the results.
Thus. this more genenal approach needs further theoretical
development.

Bhagwati states that a sufficient assumption for the
Ricardian hypothesis to hold is for the labor productivity
ratios to be a monotonically increasing function of the
price ratios. In testing this assumption. Bhagwati found
that the linear regression between labor productivity'and
price ratios fit poorly. In addition. the linear
relationship between comparative unit labor costs and export
price ratios had a poor fit. From these results. Bhagwati
concluded that both relationships are insignificant. 2 When
relative inter-country wage rates were included as a
separate and additional explanatory variable. a slight
improvement over the simple correlation between export price
ratios and labor productivities alone was found.

Bhagwati (1969. pp. 22-23) states that his results
"cast sufficient doubt on the usefulness of the Ricardian
approach” and that there is no evidence to support the

Ricardian hypothesis since "labor productivity'.“. is not a

45

datum in the sense that production functions are)‘

Even if an empirical relationship is established
between relative exports and variations in relative labor
productivity. the fundamental question of what determines
comparative cost differences cannot be answered. The reason
is that the Ricardian model simply assumes the existence of
comparative cost differences. Thus. Bhagwati and Stern
conclude that the Ricardian model is of limited importance

in explaining the composition of international trade.

Extension of the MacDougallgApprgggh

J.D. Stryker (1968) extends the M—B-S analysis by
incorporating capital inputs. labor inputs. differences in
technology. economies of scale. and relative factor prices
into a comparative cost model. thus meeting some of the
criticism of Bhagwati and Stern. Time series and cross-
sectional data are examined for the United States and Canada
for the period 1949—1962.

The first part of Stryker's study develops a two-
country. partial equilibrium model which examines some of
the factors underlying international price competition. The
model is:

0 = ALGKB + R: (3.10)
where O is the quantity of output produced: L is the labor
input: K is the capital input: R is the raw material input:
A is a proportionality parameter: a and 8 are the

elasticity coefficients of value added with respect to labor

and capital (V = O - R. where V is value added). Value added

46

is in constant dollars: labor input is in man-hours of

production and nonproduction workersl the capital VBFIBDIB

was estimated as the deflated book value of the gross
capital stock.

The assumptions of the model are as follows:

1. The manufacturing production function is subject to
constant returns.

2. Labor and capital are homogeneous.

3. Raw materials are obtained from outside the
manufacturing sector and are initially supplied by
domestic sources.

4. The production of raw materials is subject to
diminishing returns.

5. Perfect competition exists in all markets.

6. All firms are in long run equilibrium so that a single
set of factor prices faces each within a country.

7. The parameters of the production functions differ
between countries.

8. All firms engage in average cost pricing.

The supply of manufacturing commodities is a positive
function of the ratio of foreign to domestic prices for the
commodity. and although scale effects in manufacturing can
change the slope of commodity supply curves. the analysis is
fundamentally’the same.

The model expressing the relationship between relative
export growth in the United States and Canada and the

relative growth of labor productivity for a cross-sectional

47

analysis of 24 industries is as fol lows:
[oi/X)”. - (me... = a + b[((.v/L)/(V/L))us - (3.11)
((Q/L)/(V/L))ca1..

Value added and labor data were collected on an
establishmentbasis. while capital stock data were collected
from firms. Since the two are not always classified in the
same industry. problems in data compatibility may occur. In
some cases where large firms own plants in several
industries. part of the value of capital stock reported for
a particular industry exists in establishments classified in
another industry.

Since tariffs can affect the exports of two countries
differently at a given time. exports in three separate
markets are examined. The first dependent variable is the
difference between the relative rates of change of exports
to third markets: the second is the difference between the
relative rates of growth of reciprocal exports. The latter
examines the effects of average cost changes on both exports
and imports. since exports from the United States to Canada
are the same as imports of Canada from the United States and
vice versa. The third dependent variable is the difference
between relative rates of growth of total exports.

While tariffs affect the exports of two countries
differently at a given time. changes in relative cost
advantages will not be affected as long as the tariff rates
are not changed during the period considered. Changes in
external demand do not affect this analysis since a change

in world demand is reflected in a change in international

48

prices. and this affects all countries equally; However.
international prices are not the sole factor for
international competition. and thus product differentiation
may be an important factor for international competition
which may modify the results.

The results of equation 3.11 are as fol lows:

 

Dependent variable a b R2

Third-market exports -0.022 1.975 0.153
(0.871)

Reciprocal exports -0.019 1.715 0.128
(0.823)

Total exports -0.024 1.927 0.207

(0.728)

The results indicate that the coefficients and the
coefficients of determination are significantly different
from zero at the 0.05 level. although the Rz’s are low.
Thus these results indicate that labor productivity'is an
important factor in determining the pattern of trade.

An attempt is then made to determine whether increases
in the amount of capital employed in production may have
offset changes in labor productivity. The findings show
that the relationship between changes in exports and changes
in labor productivity have not been offset by changes in the
capital stock. This is due to capital being mobile between
the United States and Canada. thus equalizing its cost.

The role of technological innovation is also examined
in an attempt to determine the underlying cause of

differences in productivity of the two countries. The

49

findings are that the increases in productivity and the
growth of exports over time are related to the differing
rates of technological progress taking place within
manufacturing industries.

Stryker's evidence suggests that changes in the
structure of'LLS. and Canadian exports were related to
changes in average labor costs for the period examined.
Within this framework. labor productivity is a useful proxy
for total factor productivity. however. the differences in
labor productivity are not explained by the comparative cost
model. Stryker attributes these differences in productivity

to differences in technological innovation.

Jggtificgtion of Labor Costs

In examining labor’s importance as a factor of
production and a major component of total cost. labor's
share of national income must be determined. Dennison (1967)
found that labor income was almost 80 percent of net
national income for the United States. Canada. and
northwestern EurOpe. Labor shares this high can only occur
if high proportions for labor income characterize a majority
of the individual industries. If labor costs are a large
part of value added. differences in either relative wage
rates or relative labor productivities will play an
important role in the structure of trade.

In an analysis of the steel and motor vehicle
industries. Kreinin (1984. pp. 41-42) concentrates on the

labor cost of production because of ”its overwhelming

50

contribution to total production costs.” Labor costs were
estimated to constitute 46 percent of total production costs
for the motor vehicle industry. In addition. that industry
is composed of multinationals which transfer capital in the
production of components” partsw and final assembly’among
many nations so as to minimize costs. In this way. labor
costs determine plant location» Thus. an emphasis on wage
and labor productivity differences for countries by industry
appears to be justified.

The question of whether the differences in output per
worker are reflected in export prices and international
trade is an important one. Rostas’s 1937 data on
productivity levels in the United States and the United
Kingdom are a measure of the gross output of an industry in
real terms. In contrast. Paige and Bombach’s 1950 data are a
measure of the net value added in an industry in real terms.
The export price ratios used in all the studies were unit
values of individual commodities. which are conceptually!
closer to measures of gross output rather than to net value
added. Studies employing the 1937 data displayed a close
correlation between export price ratios and labor
productivity while studies employing 1950 data showed little
or no correlation. From the way the data are constructed. a
close correlation between prices and net value added could
occur only'if the commodity and industry definitions were
the same and the relationships between output and material
purchases fromlother industries were the same between the

two countries. Thus previous attempts to establish a

SI

correlation between export price ratios and labor
productivity ratios failed due to the lack of conceptual and
statistical compatibility in the data. and they do not

provide evidence against the Ricardian hypothesis.

ngegggtive Cost Theory Refgrmulated

Anderson and Kreinin (1981) and Kreinin (1982.1984.I985)
analyzed the motor vehicle and steel industries by examining
the cost advantage of an industry relative to all other
manufacturing industries within the same country instead of
between countries. This approachlallows industries to be
ranked internally by the inverse of their production costs
equivalent to a ranking by comparative advantage. Unit labor
costs. consistinglof labor compensation and productivity.
were employed as an approximation of total production costs.
If an industry experiences a rise in labor compensation. or
a fall in productivity relative to the national
manufacturing average. it will fall in ranking by
comparative advantage.

Anderson and Kreinin confined their study to the United
States over the period 1957-1977. while Kreinin in the three
later articles examined the United States. Japan. and the
industrialized European countries (German automobiles. and
steel and iron only for the rest) over the period 1964-1980.
The results indicated that between 1957 and 1977. labor
compensation in the~LLS. iron and steel industries increased
faster than in all manufactures. but labor productivity

increased at a rate below the manufacturing average. causing

52

a fall in the ranking. Over .the period 1964-1980. unit labor
costs in U.S. iron and steel grew faster than the
manufacturing average. with the difference between the two
widening. Thus. U.S. steel fell in ranking by comparative
advantage because of relatively high labor compensation and
relatively low labor productivity.

In contrast. Japan's iron and steel industry was able
to keep its unit labor costs in line with the manufacturing
average and maintain its ranking by unit labor costs. German
iron and steel also improved its internal ranking with
increases in labor productivity and decreases in
compensation relative to the manufacturing average. French
and British iron and steel lost ground by comparative
advantage in a manner similar to the United States.

U.S. motor vehicles improved their ranking between 1957
and 1967. While labor productivity increased rapidly after
1967. compensation increased even faster. causing the motor
vehicle industry to fall in ranking after 1967. Other
factors relating to managerial decisions. such as product
mix (large verses small cars). also played a role in the
motor vehicle industry's deterioration.

The Japanese motor vehicle industry was able to keep
unit labor costs in line with the Japanese manufacturing
average and maintain-its ranking. German motor vehicles
were able to increase labor productivity at the same rate as
all manufactures between 1962 and 1966. After 1967. labor

productivity increased at a rate below that of al 1 German

53

manufacturing. with a pronounced moderation after 1974.
Labor compensation kept pace with the manufacturing average.
causing unit labor costs to increase after 1975. Therefore.
Kreinin states that the German motor vehicle industry

appears to be losing its comparative advantage.

m

The work of MacDougall. Balassa. Stern. and Stryker
confirms the comparative cost theory. Using 1937 data.
MacDougall feund a strong inverse relationship between
relative U.S. and U.K. wage costs per unit of output and
relative exports. and a positive relationship between
relative output per worker and relative exports. Balassa and
Stern's results indicated that the 1950 relative U.S. and
U.K. export performance followed established lines of
comparative advantage as suggested by labor productivity.
The introduction of differences in capital costs per unit of
output did not have any significant effect on export
performance. Stryker's time series evidence suggested that
changes in Udi.and Canadian exports were correlated with
changes in labor productivity. and that the latter is a
useful proxy for total factor productivity. although its
differences are not explained by the comparative cost model.
Stryker attributed these differences to technological
innovation.

In contrast. Kreinin concluded that his evidence did
not support the simple Ricardian hypothesis. and that

factors other than supply. such as product differentiation

54

and demand. determined the pattern of trade. For
differentiated commodities. labor productivity sets the
limits to the terms of trade. which may be wide. whereas
demand conditions determine the exact terms of trade.

Bhagwati and Stern were also critical of the M-B-S
approach because it was too simplistic. and they call for a
multifactor framework in determining trade patterns. In
tests of the Ricardian hypothesis. Bhagwati found a poor fit
between 1950 relative labor productivity and price ratios
and the relationship between comparative unit labor costs
and export price ratios. When relative intercountry wage
rates were included as an explanatory variable. a slight
improvement was found.

The question of whether the differences in output per
worker are reflected in export prices was addressed by Daly.
Since earlier studies measured export prices and labor
productivity on a different basis. their failure to
establish a correlation between export price ratios and
labor productivity ratios was due to the lack of conceptual
and statistical compatibility in the data: they did not
provide evidence against the Ricardian hypothesis.

By reformulating the comparative cost theory to examine
internal changes in unit labor costs. Kreinin has developed
an effective tool which yields information useful in making
policy decisions. Industries concerned with import
competition can decrease their unit labor costs by
decreasing labor compensation. increasing productivity. or

both. to improve their ranking by comparative advantage and

55

trade performance.

The next chapters incorporate the Kreinin approach in
the analysis of manufacturing industries in the United
States. Japan. Germany. the United Kingdom. and Italy. The
performance of each industryms unit labor cost and its
components are examined in relation to the manufacturing

average to determine the effect on its internal ranking.

56

Footnotes

1Simply because the wage vafiable has the wrong sign
and does not "add much” to the R . is not a valid reason to
declare it insignificant. See Schmidt (1976) and
Johnson(1972).

2Goodness of fit is not the correct criterion for
judging whether a relationship is significant. See
Schmidt(1976) and Johnson (1972).

CHAPTER IV

DISAGREGATED CROSS-SECTIONAL EVIDENCE FOR UNIT

LABOR COSTS IN THE UNITED STATES

This chapter investigates the effect of unit labor
costs on the trade flows of U.S. manufacturing industries
for 1967 and 1980. The first part develops an internal
ranking of 139 industries for these years and formulates a
cross-sectional model for determining the effect of unit
labor costs on international trade. The second part arranges
the industries into low. medium. and high technology
categories and analyzes the performance of their unit labor
costs.

The unit costs of production is the inverse of
comparative advantage and is the criterion for the internal
rankings. The lower the unit costs of production. the higher
is an industry"s ranking by comparative advantage. Unit
costs of production can be approximated by unit labor costs.
which in turn are defined as the ratio of total labor
compensation to labor productivity; Value added per employee
is used as a proxy for labor productivity.

Pol icy implications can be derived from the internal
ranking of industries. If an industry’s level of labor costs
is high relative to other industries. it would be ranked low

by comparative advantage. For the industry to improve its

57

58

ranking. it must cut unit labor costs through a reduction in

labor compensation or an increase in productivity.

Data

 

The industries are categorized by the Standard
Industrtal Classification (SIC) system at the two-. three-.
and four-digit leved. LLS. data for total compensation.
value added. and total employees were obtained from various
issues of two Bureau of the Census publications. Annual
Survey gf Manufactures and the Censug 9f Mgnufgctures for
1967. 1972. and 1977.

This chapter departs from past studies by using the
total compensation of employees instead of merely wages.
Total compensation is a more comprehensive measure of the
cost structure among industries and is the better specified
variable. Besides wages. it includes social security
payments and employee fringe benefits. and among these are
paid vacation time. health care. paid holidays. and personal
absences. Total employees include production as well as
clerical and administrative workers.

LLS. data for export and import values by two-. three-.
and four-digit SIC breakdown were collected from various
issues of the Bureau of the Census. UL; Commodity Exgorts
gag Imggrts a; Related 29 Outgut. Data for export and import
values by eight-digit SIC breakdown were drawn from the LLS.

Department of Commerce. U.S. Imgorts. Consumgtion _Qy Wgrld

 

Areas. FT210. and DAL Exgorts Domestic Merchanise. FT610.

 

The unit labor costs and net trade positions of U.S.

59

industries for 1967 and 1980 are displayed in Table 4.1. The
complete tables. including the components of unit labor

costs. are presented in the Appendix.

TABLE 4.I
U.S.

6D

Industries.

Unit Labor Costs and Net Trade Positions of
1967 and 1980

 

Industry 1980 1967
Unit a X-M b Unit a X-M b
Labor Costs Labor Costs

Meat Products 0.562 -766.1 0.612 -428.9
Dairy Products 0.370 -266.9 0.465 22.5
Preserved Fruits 0.379 N.A. 0.386 1.6
Grain Mill Products 0.265 2714.9 0.294 515.3
Bakery Products 0.488 -90.8 0.532 -13.1
Sugar. Confectionery 0.294 -2214.7 0.363 -624.4
Fats and Oils 0.303 3143.1 0.389 -48.5
Beverages 0.320 -1946.2 0.356 -477.8
Misc.Foods Kindred 0.297 237.4 0.321 381.6
Tobacco Products 0.220 991.6 0.216 129.6
Weaving Mills. Cotton 0.615 181.8 0.630 —28.2
Weaving Mills. Manmade 0.562 212.0 0.627 20.8
Weaving. Finished. Wool 0.554 ~81.9 0.553 -88.0
Narrow Fabric Mills 0.607 36.2 0.621 5.0
Knitting Mills 0.571 90.2 0.593 11.8
Textile Finishing 0.524 N.A. 0.645 N.A.
Floor Covering Mills 0.463 7.2 0.420 -45.7
Yarn and Thread Mills 0.610 41.2 0.586 -39.3
Miscellaneous Textiles 0.535 -40.3 0.560 -273.2
Men’s and Boys’ Suits 0.626 -278.7 0.681 -20.7
Men’s and Boys' Furnishing 0.531 -1249.0 0.609 -92.7
Women's and Misses' Out. 0.562 -1013.9 0.600 -89.6
Women's. Children’s Under. 0.543 -84.2 0.561 1.5
Hats. Caps. and Millinery 0.560 -2264.6 0.676 -12.4
Children's Outerwear 0.554 -100.2 0.607 -226.5
Fur Goods 0.491 -10.4 0.545 1.8
Misc. Apparel and Accs. 0.575 -497.8 0.612 -28.1
Misc. Fabricated Textile 0.574 176.7 0.594 35.7
Sawmills and Planing Mill 0.621 ~1026.5 0.535 173
Millwork. Plywood. Struc. 0.643 -335.2 0.637 -272.2
Wood Containers 0.602 13.1 0.658 -169.6
Wood Buildings.Mobile Home 0.594 107.6 0.622 1.0
Misc. Wood Products 0.557 -445.8 0.587 -63.6
Household Furniture 0.583 N.A. 0.603 N.A.
Office Furniture 0.506 N.A. 0.528 N.A.
Public Bldg..Related Furn. 0.585 N.A. 0.631 N.A.
Partitions and Fixtures 0.598 N.A. 0.607 N.A.
Misc. Furniture and Fixt. 0.541 N.A. 0.605 N.A.

Pulp Mills

Paper Mills.Building Paper

Paperboard Mills
Converted Paper
Paperboard Cont. Boxes

Misc.

Building Paper.Board Mills

Newspapers
Periodicals

Books
Misc.

Publishing

61

Table 4.1 continued

Commercial Printing
Manifold Business Forms

Greeting Card Publishing
Blankbooks and Bookbinding

Printing Trade Services

Industrial
Plastic Materials.

Drugs

Soaps.
Paints.

Cleaners.

Inorganic Chem.

Syn.

Toilet

Allied Products
Industrial Organic Chem.

Agricultural Chem.

Misc.

Chemicals

Petroleum Refining
Paving and Roofing Mat.

Misc.

Petroleum.Coal Prod.

Tires and Inner Tubes

Rubber.

Reclaimed Rubber

Rubber 9

Plastic

Plastic Footwear

Hose

Fabricated Rubber

Misc.

Leather Tanning

Boot.

Footwear.

Plastics

Shoe Cut Stock

Leather Gloves
Luggage

Handbags and Purses
Leather Goods

 

Expt. Rubber

0.383
0.467
0.481
0.418
0.597
0.605

0.527
0.295
0.370
0.420
0.617
0.427
0.362
0.545
0.624

0.352
0.403
0.306
0.201
0.366
0.290
0.220
0.348

0.158
0.433
0.319

0.558
0.679
0.459
0.872
0.634
0.514

0.549
0.547
0.512
0.731
0.503
0.589
0.509

0.2
-2595.8
1293.9
480.1
134.1
-51.8

-26.5
161.7
198.4
N.A.
92.1
4.3
4.2
N.A.
7.3

454.9
4219.5
674.9
536.5
213.4
4260.3
2365.9
513.6

-11538.0
12.4
348.3

-746.3
-489.7
-0.2
117.5
N.A.
N.A.

43.3
-72.7
-2348.3
-46.9
-222.2
-444.6
-43.9

0.418
0.534
0.396
0.465
0.606
0.506

0.582
0.372
0.385
0.515
0.647
0.476
0.454
0.640
0.722

0.306
0.389
0.260
0.195
0.416
0.374
0.311
0.441

0.234
0.458
0.323

0.496
0.667
0.611
0.607

N.A.
0.538

0.652
0.596
0.605
0.680
0.568
0.643
0.631

-805.4

208.4
86.9
21.9
N.A.

-0.6
72.6
73.5
-I4.7
52.4
N.A.
-0.2
-2.1
N.A.

739.1
625.2
232.3
105.8

50.9

38.7
111.4
209.3

-550
11.9
2.0

-II.8
-81.5
2.4
N.A.
N.A.
N.A.

-2601
-O.4

-171.5

-27.3
-10.9
-62.1

-0.9

62

Table 4.1 continued

Flat Glass

Glass

Products of Glass
Cement. Hydraulic
Structural Clay
Pottery
Concrete.
Cut Stone
Misc. Nonmetallic Miner.

Gypsum. Plaster

Blast Furnace. Steel
Iron. Steel Foundries
Primary Nonferrous Metals
Secondary Nonferrous
Nonferrous Rolling
Nonferrous Foundries
Misc. Primary Metal

Metal Cans

Cutlery. Handtools
Plumbing. Heating
Fabricated Struc.
Screw Machine. Bolts
Metal Forgings. Stamp
Metal Services
Ordnance

Misc. Fab. Metal

Engines. Turbines

Farm. Garden Machinery
Construction Mach.
Metalworking Mach.
Special Industry Mach.
General Industry Mach.
Office. Computing Mach.
Refrigeration

Misc. Machinery

Electric Distributing
Elect. Industrial
Household Appliances
Elec. Lighting. Wiring
Radio. TV Receiving
Communication Equip.
Electronic Components
Misc. Elec. Equip.

-5380.6
115.1
-941.1
-2672.5
1306.4
N.A.
4.3

N.A.
35.4
3.7
N.A.
-375.3
N.A.
N.A.
N.A.
-45.2

2025.2
344.7
N.A.
N.A.
379.4
2676.1
N.A.
1730.9
N.A.

264.8
1019.5
N.A.
198.2
-3022.4
87.3
696.0
912.1

0.539
0.511
0.515
0.353
0.626
0.661
0.490
0.667
0.528

0.580
0.718
0.337
0.510
0.494
0.640
0.613

0.483
0.508
0.549
0.609
0.582
0.637
0.608
0.605
0.564

0.583
0.534
0.560
0.621
0.606
0.585
0.485
0.500
0.625

0.547
0.570
0.489
0.482
0.514
0.699
0.623
0.545

-29.9
56.3
0.3
-10.2
5.5
-79.2
3.5
-13.3
52.8

-864.3

48.0

N.A.
10.3
N.A.
23.1

N.A.
50.1
50.7
209.9
-15.3
201.0
N.A.
-31.2
111.3

376.3
146.1
N.A.
226.3
N.A.
415.4
N.A.
362.2
N.A.

150.8
207.0
59.3
N.A.

-397.6

299.1
225.9
22.1

 

63

Table 4.1 continued

 

Motor Vehicles 0.708 -13882.7 0.512 -65.8
Aircraft. Parts 0.598 N.A. 0.715 1737.3
Ship. Boat Building 0.760 526.0 0.806 56.8
Railroad Equipment 0.545 51.0 0.623 131.3
Motorcycles. Bicycles 0.543 -1516.8 0.573 —99.5
Guided Missiles. Space 0.637 N.A. N.A. N.A.
Transportation Equip. 0.605 404.7 0.605 -16.8
Engineering. Scien. Inst. 0.508 N.A. 0.583 132.8
Measuring. Controlling 0.576 N.A. 0.579 277.1
Optical Instruments 0.501 202.7 0.631 -29.2
Medical Instruments 0.483 682.0 0.489 113.3
Ophthalmic Goods 0.543 -174.3 0.554 -11.2
Photographic Equip. 0.314 612.3 0.375 192.9
Watches. Clocks 0.527 -941.5 0.595 -125.6
Jewelry. Silver 0.476 -1329.6 0.545 -122.2
Musical Instruments 0.614 36.6 0.667 ~28.1
Toys. Sporting Goods 0.418 -1311.4 0.513 -117.3
Pens. Office. Art Goods 0.444 122.5 0.533 27.4
Costume Jewelry 0.542 -378.1 0.526 -83.3
Misc. Manufactures 0.558 83.3 0.568 N.A.
All Manufactures 0.543 0.471

 

 

(a) Unit labor costs are defined as
compensation/value added.

(b) Units for net trade are millions of dollars.

64

The Models

This section examines the effect of unit labor costs on

the trade performance of manufacturing industries in a cross

sectional model for 1967 and 1980. A double log form of the

equation is used to obtain elasticities and because the

Ricardian hypothesis traditionally'has been tested that

way; The equations are as follows:

X/M‘ , A1 + 42 L, + 01.

log(X/M)i = 81 + 82 log(L). + U2.
Where
X/M = trade performance:
L = unit labor costs

(total compensation/value added);
U1.U2 = error terms: and
i = industries (1.....139).
The coefficients are
A1.BI = intercept terms:

A2.B2

sensitivity of trade performance to

(4.1)
(4.2)

changes in

unit labor costs. The expected sign of 82 is

negative.

65

Table 4.2 OLS Results from the Cross-Sectional Data

1980 Results:

X/M I 6.70 - 8.34 L (4.11)
(2.18) (4.17)

R?- . 0.04 F-stat . 4.01 N s 112 0-11 stat = 2.21

109(X/M) a -1.22 - 1.60 109(L) (4.21)
(0.36) (0.48)

R2 = 0.09 F-stat . 11.10 N =112 o-w stat ..- 1.93

1967 Results:

X/M a 16.54 - 21.20 L (4.12)
(6.31) (11.58)

R2 = 0.03 F-stat . 3.35 N = 115 D—W stat - 2.17

log(X/M) = - 0.90 - 1.51 Iog(L) (4.22)
(0.45) (0.63)

R2 . 0.05 F-stat - 5.68 N - 115 0-w stat . 1.82

 

The statistical results indicate that unit labor costs
were>an important factor in explaining the trade flows of
the U.S. during the two years examined.Although the R2 was
low inlall regressions. the relationship between unit labor
costs and trade performance was significant as indicated by
the F-statistics. In addition. the coefficients of the
explanatory variables were significant and of the correct
sign.

The low R2 are not a cause for concern since all the
regression statistics were significant. Considering the
multitude of factors affecting international trade flows. a
low R2 was not unexpected since only one explanatory

variable was used - unit labor costs. The goal of the

66

analysis was not to obtain high R2. but instead to generate
significant coefficients of the correct sign. In addition to
these results. an alternate specification. based on

Deardorff"s model. is tested in Appendix C.

Analysis of U.S. Manufacturing Industries

 

This section analyzes the unit labor costs. which are
presented in Table 4.1. of low. medium. and high technology
manufacturing industries in the United States. The breakdown
of industries into these three categories is based on
Department of Commerce classifications.

Each industryfls level of unit labor costs and its
components. labor compensation and productivity. are
examined in relation to the manufacturing average. If an
industry’s unit labor costs increases relative to other
manufacturing industries. its trade position is expected to
suffer. Since this study concentrates on supply side
considerations. all deviations in an industry's trade
position cannot be explained. Nevertheless. supply factors

can explain a large portion of the trade flows.

ng Technology Inggstries

The low technology industries have been grouped into
four major categories: agricultural-based manufacturing.
traditional low technology'products. forest products.(and
nonmeta 1 products.

As shown in Table 4.1. the United States held a strong

comparative advantage in 1967 and 1980. as approximated by

67

unit labor costs. in the agricultural-based manufacturing
industries. which consists of food products. beverages. and
tobacco. They are generally characterized by high labor
productivity and low compensation relative to all
manufactures. The exception is the meat products industry.
which displayed relatively low labor productivity levels and
ranked 90 in 1980.

From 1967 to 1980. all (LS. agricultural-based
manufacturing industries were able to strengthen their
position with decreases in unit labor costs. These
reductions stemmed from increases in labor productivity and
decreases in compensation relative to the manufacturing
average.

Despite lower unit labor costs. its trade position did
not improve uniformly. Although grain mil l products. fats
and oils. and tobacco products showed a significant
improvement. the dairy products. sugar. and beverage
industries experienced a marked deterioration.

The United States helcia comparativeldisadvantage in
the production of traditional low technology products. which
encompass texti le. wearing apparel. leather products. and
footwear. These industries have been grouped together
because of similar cost structures and trade positions.
Their cost structures are characterized by low levels of
labor productivity and compensation relative to the
manufacturing average and their trade flows are heavily

regulated by multi- and bilateral trade agreements.

68

The United States was able to reduce its unit labor
costs from 1967 to 1980 in the wearing apparel. leather
products. and footwear by reducing labor compensation
relative to all manufactures. In textiles. unit labor costs
in the weaving mi ll and textile finishing industries were
lowered. while those in wool weaving. yarn. thread. and
floor covering mills increased.

The wearinglapparel and footwear industries suffered
trade deficits in 1967 and 1980. Except for children's
outerwear products. the deficits increased between these two
years even though unit labor costs decreased. Since supply
considerations do not explain the trade patterns. trade
restrictions and demand Factors must be the prime
determinants of trade flows in these commodities.

The forest products category consists of wood products.
furniture. and paper products industries. All of these incur
high transportation costs. which constricts trade flows and
causes many of the products to be labeled nontraded goods.

In 1967 and 1980. the United States held a comparative
disadvantage in the manufacture of wood products and
furniture. The exception was office furniture. which ranked
52 in 1980. The cause for the low rankings were the low
levels of labor productivity relative to all manufactures.
Since many of the wood and furniture products are non-
traded. changes in their unit labor costs are expected to
have little effect on trade flows.

As shown in Tables 4.1 and 11-8 in the Appendix. the

United States held a comparative advantage in the

69

manufacture of paper products. with the exceptions of
paperboard (ranking 108) and building paper (ranking 114).
The high rankings were caused by high labor productivity
relative to all manufactures. Except for paperboard mills.
the paper products industries were able to reduce unit labor
costs between 1967 and 1980. In response. their trade
position improved. The exception was paper mills. which
despite lower unit labor costs experienced a large increase
in its trade deficit.

The nonmetal products classification consists of
pottery and china. glass. and other nonmetal products.
Generalizations regarding this classification are difficult
because'of'the diversity'of the industries.

The United States possessed a comparative disadvantage
in the manufacture of pottery and china and glass products
because of high unit labor costs. But whereas these were
increased in the glass industry from 1967 to 1980. they were
reduced in the the pottery and china industryu The latter
recorded trade deficits in both years and despite reduction
in unit labor costs from 1967 to 1980. its trade deficits
mounted. Since its trade position was not responsive to
supply considerations. demand factors are the major forces
in determining trade flows.

The other nonmetal products industries exhibited widely
varying unit labor costs. Unit labor costs were low for
cement. concrete. and nonmetallic minerals and high for

structural clay and cut stones. Since nonmetal products are

70

subject to steep transportation costs. which cause them to
be labeled nontraded goods. changes in unit labor costs in
these industries would have little effect on their trade

performance.

Medium Technolggy Industries

The medium technology industries are defined as iron
and steel. nonferrous metals. metal products. energy.
rubber. transport equipment (except for aircraft and guided
missiles). and publishing.

The United States held a comparative disadvantage in
the manufacture of rubber in 1967 and 1980. The exception
was reclaimed rubber production. where low unit labor costs
prevailed. Labor productivity was above average in the tire
and reclaimed rubber industries. below average among the
rest of the industries. The rubber industry's ranking
experienced a significant reduction from 1967 to 1980
because of a decrease in labor productivity and an increase
in compensation relative to all manufactures. The escalation
of its unit labor costs led to a rapid deterioration in the
industry's trade deficit.

In 1967. the United States held a comparative
disadvantage in iron and steel foundries. and there were
"moderate" unit labor costs in blast furnace steel
production. By 1980. the latter was in a position of
disadvantage because of escalations in labor compensation

relative to all manufactures. These sharp increases aided in

71

turning a small trade deficit in 1967 into a large deficit
in 1980.

The United States held a comparative advantage in the
primary and secondary nonferrous metals industries because
of high labor productivity. and a comparative disadvantage
irinonferrous foundries because'of low labor productivity’
relative to all manufactures.

From 1967 to 1980. the primary nonferrous metals and
rolling industries increased. while the secondary nonferrous
metals decreased their unit labor costs. The response of
trade flows to these changes was not consistent with the
comparative costs theory. and thus demand is the important
determinant of trade flows.

In 1967. the transport equipment category displayed
consistently high unit labor costs with the exception of
motor vehicles. where these were low. However. by 1980 unit
labor costs had increased significantly in this sector so
that the entire category possessed high unit labor costs.
The sharp increase was reflected in trade position of the
motor vehicle industry. In 1967 it registered a small
deficit. but by 1980. this had swelled to record
proportions.

The metal products industries have a heterogeneous cost
structure. The manufacture of metal cans. plumbing
materials. and heating equipment involves low'unit labor
costs while the opposite is true of screw machines and metal
forgings. Overall these industries reduced their unit labor

costs from 1967 to 1980. but for those for which dataare

72

complete. their trade positions deteriorated. Thus. demand
factors are the most important determinants of metal
products’ trade flows.

The United States held a comparative advantage in the
publishing industries except for commercial printing (ranked
123) and printing trade services (ranked 124). Al 1 sectors
reduced unit labor costs from 1967 to 1980. but since these
industries generate low trade volumes. analysis of the
category difficult.

The United States appeared to exhibdt a comparative
advantage in the energy industries that consists of
petroleum refining and coal products. Because this category
is overwhelmingly dominated by demand. which may distort the
ranking. only a brief mention is made here.

The high rankings for the energy industries stem from
the extraordinarily high values for labor productivity
relative to the manufacturing average. But. despite
displaying low unit labor costs. petroleum refining recorded
large trade deficits in 1967 and 1980. Clearly. demand
outweighed supply in the determination of trade flows
because of the special position of petroleum in developed

economies.

T h l n i
The high technology industries are defined as

industrial chemicals. plastics. electrical machinery. office

machinery. turbines. aircraft. and professional goods.

73

The United States held a comparative advantage in the
production of industrial chemicals in 1967 and 1980. These
sectors are characterized by high labor productivity
relative to all manufactures. They were able to maintain
their low unit labor costs from 1967 to 1980. with the
agricultural chemical industry improving its ranking.
Because of low unit labor costs. industrial chemical
industries registered trade surpluses in 1967 and 1980. Due
to a decline in its unit labor costs. agricultural chemicals
significantly increased its trade surplus.

The United States held a comparative advantage in the
manufacture of plastics in 1967 and 1980. This strong
ranking is attributable to the high level of labor
productivity in plastics relative to all manufactures. The
industry sustained trade surpluses in both years because of
low unit labor costs. With reductions in labor compensation.
the plastics industry was able to decrease its relative unit
labor costs and increase its trade surplus in 1980.

In 1967 and 1980. the United States held a comparative
advantage in the manufacture of office and computing
equipment and engines and turbines had moderate unit labor
costs. These industries possessed high levels of labor
productivity relative to the manufacturing average. In
addition. each was able to decrease its unit labor costs
from 1967 to 1980. Although trade data were not available
for the office and computing equipment industry. the turbine
industry registered a trade surplus in 1967 and 1980. With

reductions in its cost structure. it was able to improve its

74

trade position in 1980.

The electrical machinery industry encompasses products
with varied unit labor costs. These are moderate in
electrical distributing. industrial equipment.
communications equipment. and electronic components. low for
radio and TLV. receiving equipment. All were able to reduce
unit labor costs from 1967 to 1980. Except for radio and
T.V. equipment. the United States recorded a trade surplus
in electrical machinery for 1967 and 1980. In conformity
with the comparative costs theory. all industries except for
the one noted. were able to improve their trade surpluses.

In 1967. the United States held a comparative
disadvantage in manufacture of aircraft. The low ranking was
caused by the high labor compensation relative to the
nanufacturing average. By 1980. aircraft reduced their unit
labor costs with significant decreases in relative labor
compensation and productivity. Although aircraft registered
a surplus in 1967. trade data was not available for 1980.
‘thus further conclusions are not possible at present.

The United States possessed a comparative advantage in
'the production of professional goods. The exception was
Uneasuring and controlling instruments industry. which ranked
99 by unit labor costs. The high levels of labor
ilroductivity contributed to the strong ranking of these
industries. This ranking was preserved from 1967 to 1980.
\Nith the engineering. scientific. medical. and optical

H1struments industries reducing their unit labor costs.

75

Professional goods registered a trade surplus in 1980. With
reductions in their unit labor costs. the optical and
medical instruments industries were able to improve their

trade positions.

Conchgign

The statistical results indicated that unit labor
costs were an important factor in explaining the trade flows
of the United States during the two years examined. Although
theR2 was low'inlall regressions. the coefficients were
significant and had the correct sign.

The results of the rankings by unit labor costs for
1967 and 1980 indicated that the United States held a
comparative advantage in manufacture of agricultural
manufacturing products. paper products. nonferrous metals.
publishing. industrial chemicals. plastics. office and
computing equipment. and professional goods. The United
States held a comparative disadvantage in traditional low
technology products. wood products. furniture. pottery.
glass. rubber. iron and steel. and most transport equipment.

Iron and steel. rubber. and transport were the only
industries to decline significantly in ranking from 1967 to
1980. This was reflected in the large trade deficits they
incurred. In contrast. the aircraft. agricultural chemicals.
and professional goods industries decreased their unit labor
costs and resulted in improvements in their trade position.
But all industries did not conform to comparative cost

theory. The trade position of some industries was not

76

affected by supply but instead dominated by demand factors.
trade restrictions. and transportation costs. Nevertheless.

supply considerations did explain a large portion of the

trade flows.

CHAPTER V

CROSS-SECTIONAL AND TIME SERIES EVIDENCE FOR THE

UNITED STATES. JAPAN. AND EUROPE

This chapter investigates the effect of unit labor
costs on trade flows for the period 1967-1981 in the United
States. Japan. and three European countries: Germany. the
United Kingdom. and Italy. The first part examines unit
labor costs and trade positions of manufacturing industries
and develops a cross-sectional model correlating them at two
points. 1967 and 1980. The second part analyzes time series
data to determine the effect of changes in labor
productivity and compensation on the trade performance of
individual industries. In addition. all the time series and
'cross-sectional data. combined with dummy variables. are
"pooled" into one equation to remove the biases of trade
discrimination and exchange rate changes over time on trade
performance.

The only consistent unit labor cost data available for
Japanese and European manufacturing industries are
categorized by the three-digit International Standard
Industrial Classification (ISIC). The ISIC breakdown
provides data for 28 manufacturing industries. and it is

more aggregated than the SIC breakdown for the United States

77

78

used in the previous chapter. The problems stemming from the

aggregation are discussed in the data section.

Data

 

German. Japanese. Italian. and British data for labor
compensation. value added. and total employees at the three-
digit ISIC breakdown were obtained from the United Nations.
Yearbogk 9: Industrial Statistics. Italian. British. and
German data for 1967 were incomplete. Additional statistics
were drawn from the Statistisches Bundesant. Statistisches
Jahrbush ffi; gig Bundesregublik Degtschland.

Japanese. German. Italian. and British data for export
and import values by two- to four-digit Standard

International Trade Classification (SITC) breakdown were

collected from the United Nations. Commodity Trade

Statistics. The OECD exports are from 1:95.18 gy Commogities
Market Summaries: Exggrts. series C.

All data were rearranged to conform to the three-digit
ISIC breakdown. A table of concordance from SITC to ISIC
breakdown is provided in the United Nations. Proceedings gfi

the ____.United Netioos swarm so Leeds and Less-Jem-
Additional information is contained in Farhang Niroomand's
thesis: Determinants gt Q3 Strgcture gt i._.i_._&_‘)y Foreign Lag;
1Q Manufactures: 1963-1280.

A major problem with the ISIC data is its
presentation. Low technology industries. such as glass.
leather. and footwear. where trade flows are small (partly

because of trade restrictions). are presented in a

79

disaggregated form. The trade restrictions are in the form
of tariffs and quotas. which are governed by multi- and
bilateral agreements.

In contrast. high technology industries. such as
machinery and transport equipment. where large trade flows
occur. are highly aggregated. For example. the machinery
industry consists of such diverse categories as lawn and
garden equipment. refrigeration. and laundry equipment. as
well as electronic computing equipment. turbines. and
oi lfield machinery. Lawn and garden equipment and similar
categories employ low to medium technologies. while
electronic computing uses high technology.

The same problem is true for transport equipment. which
consists of such medium technology industries as motor
vehicles. ship building. and railroad equipment. as well as
the high technology industries of aircraft; guided missiles.
and space vehicles. The aggregation of medium and high
technology industries makes definitive conclusions
impossible because of diverse cost structures. Nevertheless.
consistent trends and broad conclusions may still be
obtained.

Tables 5.1-5.4. presented at the end of the chapter.
show unit labor costs and net trade positions of
nanufacturing industries in the United States. Japan. and
three European countries for 1967 and 1980. The former
represents the first year examined in the time series
analysis. while the latter is the last year for which ISIC

data were available for all the countries. The complete

80

tables. including the breakdown of unit labor costs. are

presented in the Appendix.

The Cross-Sectional Models

This section examines the effect of unit labor costs on
trade performance in a cross-sectional model. Four
alternative specifications. including the double log form.
of the comparative cost model were developed. The use of
export shares has been employed in the ”revealed comparative
advantage" literature and is used in equations 5.3 and 5.4
as an alternative specification of trade performance. The
equations are judged on their ability to generate
significant coefficients of the correct sign. and the R2 are
not an integral part of the analysis. The equations whose

complete results are presented in the Appendix are as

follows:

(X/N)m s A1n + 111an + 111”: (5.1)
log(X/M).n 1: BIn + anlog(L)1n + U2": (5.2)
(Xs).n . Cln + can,n + 03"; (5.3)
log()(s).n . 01n + D2nlog(L)'n + U4“: (5.4)

The additional variables are:

Xs - country's share of total OECD exports
(exports/total OECD exports):

i a industries 3 1.....28:

n scountries s 1..”.5 sUnited States.Germany.
JapaanHted Kingdom. and Italy.

Considering the distorting effects of differential

trade restrictions in a cross-sectional model. most of the

81

models performed well. The dependent variable expressed as a
ratio between exports and imports. as in equations 5.1 and
5.2. obtained significant coefficients of the correct sign
for unit labor costs for the United States. Germany. and
Great Britain. Italy generated correct signs. but they were
not significant. Japan consistently obtained a positive sign
regardless of the specification for 1967 and 1980. with a
significant coefficient for I967; The implication is that
trade restrictions and demand considerations were more
important than supply in explaining trade flows in the two
years for Japan.

The dependent variable expressed as an industry's
export share of total OECD exports. as in equations 5:3 and
5.4. did not perform well. Japan and Germany generated
positive coefficients: Italy’s were insignificant. Since
only exports were explicitly represented in this variable.

import-competing industries are underrepresented. Thus this

variable is not as well specified as the ratio of exports to
imports.
The equations in log linear form did not perform

significantly different from the standard form. They do
yield elasticities and provide an indication of the
responsiveness of a country"s trade flow to changes in unit
labor costs.

In addition. the Spearman rank correlation coefficients
between unit labor cost rankings and export share rankings
were estimated. The advantage of nonparametric tests is that

they are less restrictive in their assumptions: however.

82

their power is low in comparison with parametric tests.
Because of this. the Spearman rank correlations did not
provide any additional definitive insights and their results
are not reported here.

The conclusion is that the examination of unit labor
costs provides important information for policy makers who
are concerned with the trade balance of a particular
industryu That balance'can be improved either by lowering

labor compensation or by increasing labor productivity.

Time Series Evidence for the United States

This section analyzes the effect of changes in labor
productivity and compensation cw1LLS. trade performance. The
absolute and relative versions of the comparative cost
theory are tested in the models below. The absolute version
(equations 5.5 and 5.6) correlates changes in the components
of unit labor costs with trade performance. By contrast. the
relative version (equations 5.7 and 5.8) correlates changes
in the components of unit labor costs as a ratio of all
manufactures with trade performance. Both versions are
estimated in double logarithmic form to obtain elasticities
for the explanatory variables. Unit labor cost could not be
used as an explanatory variable because it requires too many
restrictions on estimation and makes interpretation of the
results impossible. 1

Equations 5.7 and 5.8 are closer to the comparative

cost theory as reformulated by Kreinin and the g griori

83

expectations are that equations 5.7 and 5.8 wi l l generate
more significant coefficients of the correct sign. The
equations whose complete results are presented in the

Appendix are as follows:

(X-M)1t 8 A1: + AZP”: +A3W‘t '1' U1:8 (5.5)
log(X/M)‘t = 81. + BZlog(P)‘t + BBlog(W)‘t + U2.: (5.6)

C3[W.t/(sumW‘)t] 'i' U3‘3
log[(X/M)‘t/(sumXflsumMfit) 8 01. + (5.8)
021°g[P.t/(sumP.)tl +
D3log[W‘t/(sumW‘t/(sumW.)t] + U3‘.

The additional variables are:

P = labor productivity per employee in each
U.S. industry (value added/employee):
W = labor compensation per employee in each

U.S. industry (total compensation/employee).

The additional coefficients are:

A2.82 a sensitivity of export performance to changes
C2.02 in labor productivity: the expected sign of
82 is positive:

A3.83 sensitivity of export performance to changes
CB.D3 in labor compensation: the expected sign of
B3 is negative.

Productivity per employee in each industry is expected
to be positively correlated with export performance since
increases in productivity improve an industry's ranking by
comparative advantage. Compensation per employee is expected
to be negatively correlated with export performance because
increases in compensation erode an industry's ranking. The

magnitudes of the coefficients are interpreted as

elasticities and may be useful for policy decisions.

B4

In general. equations 5.5 and 5.6 performed poorly. as
expected. in explaining U.S. trade performance. Equation 5.5
yielded the expected signs for labor productivity in nine
cases. but only three were significant. The expected sign
was produced in twelve cases for compensation. half of which
were significant. Equation 5.6 yielded the expected signs
for labor productivity in twelve cases. five of which were
significant. and in eleven cases for compensation. five of
which were significant. Both equations general ly produced
high st.

When estimating relative changes in labor productivity
and labor compensation. as in equation 5.7. the results
performed markedly better. Equation 5.7 yielded the expected
signs for labor productivity in nineteen cases. seven of
which were significant. The expected sign was produced in
twenty—two cases for compensation. ten of which were
significant. The logarithmic form did not perform as well as
equation 5.7. but it did yield better results than either of
the absolute VEPSIODS. Equation 5.8 yielded the expected
signs for labor productivity in half cases. half of which
were significant. and in fifteen cases for compensation.
seven of which were significant.

A statistical problem present in all the equations is
the existence of multicol linearity. since labor compensation
is a function of labor productivity. these two variables are
correlated. Nevertheless. although some individual industry
equations were sensitive to multicol linearity and produced

inconsistent results. it was not a restricting problem in

85

general. Despite these statistical difficulties. the
relative version of the comparative cost theory performed

well.

Pool ing Models and Time Effects

Stryker was one of the first to use a time series
approach within a comparative cost framework. His method. in
a limited way. solved the problem of trade restriction
biases. If these restrictions are constant between periods.
changes in trade performance are not influenced by
differential trade restrictions among industries.
Unfortunately. trade restrictions did change within the
period studied. and adjustments had to be made to correct
for the distortions.

This section attempts to solve the problem with the
use of a panel data model. which pools all time series and
cross-sectional data along with dummy variables into one
equation. The dummy variables portion can correct for
distortions over time and thus are cal led time effects in
the literature. 2 The time effects variables in the panel
data model captures changes in differential trade
restrictions among industries and exchange rate changes over
time without distorting the relationship between the
explanatory variables and trade performance. The importance
of time effects in modeling trade flows is illustrated in
Figure 5.1 and 5.2. using unit labor costs as an explanatory

variable. ceteris ggribus.

86

Figure 5.1. Correlation without Time Effects

X/M

O
/

m

o .. .. .. -- .. -- .. .. -- .. .. .. ..

I
l
l
I
l
I
I
W
/

> O- - -- -- --
/

\ Unit Labor Costs
\

Figure 5.2. Correlation with Time Effects

X/M

/

,/
//
/
//

O

o -- -- .. .. -.. .. ..

t2.

-/t1.

 

Unit Labor Costs

In time I. X/M :- AB = OE. labor costs a 0A. In time 2. X/M
a: AC = 00. unit labor costs a CA. where X/M I: trade

performance. t a time periods.

87

If a decrease in trade restrictions or exchange rate
occurs in time period 2. trade performance (X/M) would
increase from OE to 00 without a change in unit labor costs.
If a regression without the time effects (Figure 5.1) were
to be estimated. an upward bias would occur in the slope
coefficient. If time effects are included (Figure 5.2). a
change in trade restrictions would be totally'reflected in a
change in the intercept. not the slope. Thus this process
would yield a lower slope coefficient. and. as expected.
lower trade restrictions would be responsible for the
increase in trade performance of OD-OE.

The models are:

109mm)”: = 01 + (:2 169(9)“: + 03 16901)”; (5.9)
+ Q1 v,t + + ij.t + mt:
log(X/M)it :- 01 + 0216;;(13)it + (:3169111)it + uzt. (5.10)

Theladditional variables and coefficient are:
J g t“! 3 lees-’14:

V

time effects to capture changes in trade
discrimination: set of t-I dummy variables.
Vit - { 1 if i s t. 0 if i at ):

Q = coefficients for the set of dummy variables:
adjustment to the intercept term.

The panel data models were estimated with OLS. The time
effects in equation 5.9 were assumed to be fixed. which is
the least restrictive assumption. and the estimates of the
coefficient are consistent since t is large. 3 The results

for equation 5.9 are as fol lows:

88
logX/M = -3.89 + 1.11 IogP - 0.43 109W - 1.93 VI - 1.72 V2
(0.53) (0.21) (0.40) (0.44) (0.42)

- 1.71 V3 - 1.69 V4- 1.50 V5- 1.28 V6- 1.02 V7
(0.41) (0.40) (0.39) (0.38) (0.37)

- 1.03 V8 - 0.99 V9 - 0.95 V10 - 0.73 V11
(0.36) (0.35) (0.35) (0.35)

- 0.45 V12 - 0.33 V13 - 0.23 V14
(0.34) (0.34) (0.34)

(Standard errors in parentheses)
R2 = 0.20 F-stat .. 6.45 0-11 stat . 2.75 N . 420.
The signs of both the explanatory variables are as
expected. but the coefficient for total compensation per
employee is not significant at the 95 percent confidence
level. In addition. the coefficients of the dummy variables
decrease in value over time. One explanation is that trade
restrictions havelbecome less influential on'trade flows:
the second is that trade flows have trended upward over
time.‘4
The results of equations 5.10 without the dummy
variables are as fol lows:

IOQX/M c - 1.79 + 1.22109P - 0.93 logW
(0.33) (0.19) (0.26)

R2 s 0.14 F-stat = 34.30 0-w stat . 2.42 N s 420.
All the signs were correct. and the explanatory
variables are significant. The main difference between
equations 5.9 and 5.10 is that W is not significant when
dummy variables are included and is significant when they
are omitted. One explanation is that wages are correlated
with the time trend: since time trends are eliminated with

the panel data model. wages become insignificant. Another

89

explanation is that since cross-sectional variations drive
the panel data model with time effects. and cross-sectional
variations in labor compensation between industries are

small. the coefficient would be insignificant.

Conclusion

The statistical results obtained in this chapter
confirm the importance of unit labor costs as a determinant
of international trade patterns. The cross-sectional model
produced significant results of the correct sign for all the
countries except Japan. which generated a positive sign for
unit labor costs in 1967 and 1980.

For the time series model. correct signs for labor
compensation and productivity were produced for two-thirds
of the industries when using the relative version of the
comparative cost theory. Finally. the panel data model
removed the biases of changes in differential trade
restrictions and exchange rates and generated positive
results based on all available data for the U.S. economy
over the period 1967 to 1981.

This evidence establishes the importance of considering
the performance of unit labor costs in the formation of
policy for industries suffering from international
competition. In order for an industry to improve its ranking
and strengthen its trade position. it must increase labor
productivity and decrease compensation. The performance of
the unit labor costs of individual industries based on the

empirical evidence gathered here is presented in the next chapter.

90

TABLE 5.1 U.S. Unit Labor Costs and Trade Position.

1967 and 1980

 

 

 

 

 

 

INDUSTRY UNIT LABOR COSTS a X—M b
1967 1980 1967 1980
Food Products 0.384 0.377 —376.4 3673.2
Beverages 0.344 0.352 -565.9 -2422.2
Tobacco 0.187 0.220 151.0 1051.7
Textiles 0.536 0.553 -740.4 -1462.7
Wearing Apparel 0.558 0.555 -708.8 -6634.0
Leather and Products 0.570 0.550 -174.7 -984.7
Footwear 0.549 0.515 -260.5 -2763.4
Wood Products 0.574 0.589 -506.6 -817.4
Furniture. Fixtures 0.555 0.569 -76.1 -722.5
Paper and Products 0.455 0.478 -573.6 -985.7
Printing. Publishing 0.498 0.494 184.7 630.2
Industrial Chemicals 0.293 0.461 1665.7 10689.7
Other Chemical Products 0.365 0.279 694.8 2639.1
Petroleum Refineries 0.203 0.158 -626.5 -12952.3
Petroleum. Coal Product 0.368 0.389 N.A. 481.6
Rubber Products 0.483 0.624 -47.0 -992.6
Plastic Products n.e.c. 0.485 0.514 60.1 859.1
Pottery. China. etc. 0.595 0.603 -96.0 —677.6
Glass and Products 0.461 0.573 -7.4 -65.4
Nonmetal Products n.e.c. 0.448 0.482 19.1 125.8
Iron and Steel 0.522 0.714 -1519.0 -9475.1
Nonferrous Metals 0.444 0.511 1305.8 -3356.3
Metal Products 0.516 0.565 642.3 2096.3
Machinery n.e.c. 0.511 0.541 4161.6 25205.7
Electrical Machinery 0.526 0.540 445.2 —999.6
Transport Equipment 0.538 0.656 1232.6 -634.3
Professional Goods 0.452 0.456 529.5 2648.6
Other Industries 0.501 0.495 -481.3 -4143.9
0.471 0.543

All Manufactures

 

(a) Unit labor costs are defined as

compensation/value-added.

 

(b) Units for net trade are millions of dollars.

91

TABLE 5.2 Japanese Unit Labor Costs and Trade Position.

1967 and 1980

 

INDUSTRY

 

 

 

UNIT LABOR COSTS a X-M b

1967 1980 1967 1980
Food Products 0.304 0.336 126.78 -762.69
Beverages 0.198 0.224 -4.13 -206.56
Tobacco 0 0 -6.68 -57.87
Textiles 0.371 0.439 1239.88 2762.71
Wearing Apparel 0.418 0.526 319.65 -1049.07
Leather and Products 0.383 0.456 36.71 0.48
Footwear 0.393 0.463 97.45 -215.23
Wood Products 0.391 0.444 2.48 -7953.21
Furniture. Fixtures 0.427 0.418 16.61 -73.28
Paper and Products 0.320 0.359 -44.54 -716.54
Printing. Publishing 0.375 0.403 -0.24 1.72
Industrial Chemicals 0.187 0.254 234.80 1577.97
Other Chemical Products 0.188 0.193 -161.08 -887.15
Petroleum Refineries 0.109 0.084 -749.39 -6427.01
Petroleum. Coal Products 0.267 0.204 N.A. -12089.54
Rubber Products 0.322 0.391 107.74 1743.82
Plastic Products n.e.c. 0.338 0.370 48.24 117.61
Pottery. China. etc. 0.453 0.466 94.95 517.76
Glass and Products 0.254 0.298 42.40 127.55
Nonmetal Products n.e.c. 0.301 0.319 82.78 714.41
Iron and Steel 0.299 0.271 903.31 14560.40
Nonferrous Metals 0.282 0.272 -483.88 -2569.42
Metal Products 0.379 0.407 383.49 3541.10
Machinery n.e.c. 0.359 0.399 308.36 16330.09
Electrical Machinery 0.312 0.372 964.11 19035.22
Transport Equipment 0.316 0.379 1573.19 32114.97
Professional Goods 0.417 0.445 292.51 4749.72
Other Industries 0.355 0.414 546.77 1917.62
All Manufactures 0.319 0.355

 

(a) Unit labor costs are defined as

compensation/value-added.

(b) Units for net trade are in millions of dollars.

92

TABLE 5.3 German Unit Labor Costs and Trade Performance.

1967 and 1980

 

 

INDUSTRY UNIT LABOR COSTS a X-M b

 

 

 

1980 1980 1967

Food Products 0.290 -1696.34 -332.30
Beverages 0.317 -602.36 ~68.30
Tobacco 0.060 247.63 20.98
Textiles 0.641 -898.08 94.97
Wearing Apparel 0.498 -5444.13 -I78.34
Leather. Footwear. etc. 0.534 -2768.74 -83.11
Wood. Furniture. etc. 0.471 -1689.58 -71.50
Paper and Products 0.580 -1847.18 -329.28
Printing. Publishing 0.585 891.65 102.07
Industrial Chemicals 0.538 11004.29 1763.83
Petroleum Refineries 0.061 -10217.00 155.59
Rubber Products 0.592 172.69 47.27
Plastic Products n.e.c. 0.549 645.39 54.24
Pottery. China. etc. 0.441 95.29 48.53_
Glass and Products 0.531 151.27 72.89
Nonmetal Products n.e.c. 0.405 270.87 56.94
Iron and Steel 0.519 4818.90 1024.66
Nonferrous Metals 0.824 -1185.28 -445.50
Metal Products 0.626 3230.25 678.89
Machinery n.e.c. 0.625 25470.81 3800.80
Electrical Machinery 0.560 5441.31 1203.89
Transport Equipment 0.578 19460.68 2276.49
Professional Goods 0.461 1784.72 403.80
Other Industries 0.481 -194.74 305.01
All Manufactures 0.495

 

(a) Unit labor costs are defined as compensation/value-added

(b) Units for net trade are in millions of dollars.

93

 

 

 

 

 

TABLE 5.4 U.K. and Italian Unit Labor Costs and Trade
Performance for 1980
INDUSTRY UNIT LABOR COSTS a X-M b
U.K ITALY U.K. ITALY

Food Products 0.469 0.447 -2452.32 468.18
Beverages 0.303 0.360 1067.11 718.59
Tobacco 0.320 0.749 569.82 -314.65
Textiles 0.704 0.559 -559.71 -51.87
Wearing Apparel 0.729 0.582 -979.32 3786.70
Leather and Products 0.637 0.481 0.13 524.04
Footwear 0.674 0.586 -518.61 3478.14
Wood Products 0.581 0.483 -2208.82 -2075.83
Furniture. Fixtures 0.641 0.493 -101.26 2043.22
Paper and Products 0.594 0.499 -2755.50 -1056.91
Printing. Publishing 0.517 0.570 441.97 427.60
Industrial Chemicals 0.461 0.475 1722.71 -2537.75
Other Chemical Products 0.350 N.A. 3265.02 N.A.
Petroleum Refineries 0.082 0.032 85.90 -2890.81
Petroleum. Coal Products 0.361 N.A. -103I.18 N.A.
Rubber Products 0.590 0.618 376.53 133.80
Plastic Products n.e.c. 0.551 0.756 -23.15 475.44
Pottery. China. etc. 0.661 N.A. 256.07 N.A.
Glass and Products 0.592 N.A. 15.80 N.A.
Nonmetal Products n.e.c. 0.392 0.471 39.82 2189.95
Iron and Steel 0.843 0.498 -1074.85 -336.01
Nonferrous Metals 0.584 0.615 -I657.47 -2326.37
Metal Products 0.615 0.543 1044.13 2318.24
Machinery n.e.c. 0.567 0.543 8324.39 6238.49
Electrical Machinery 0.590 0.614 402.01 -109.44
Transport Equipment 0.730 0.638 525.14 -912.72
Professional Goods 0.595 0.541 166.36 -891.99
Other Industries 0.538 0.547 -321.38 1764.24
All Manufactures

(a) Unit labor costs are defined as

0.553

compensation/value-added.

(b)

0.537

Units for net trade are in millions of dollars.

94

Fggtnotes

lIf unit labor costs were to be used in equation 5.5
and EMS, one restriction would applys that the coefficient
of W be negative. But in the estimation of 5.7 and 5.8.
three restrictions are necessary. causing the expected sign
for unit labor costs to be positive and making the results
tenuous.

2For more on ”pooling" or panel data models. see Kmenta
(1972) and Judge at al. (1980).

3The fixed effects model does not make any specific
assumptions about the distribution of the time effects. In
addition. correlation between the time effects (changes in
exchange rates and differential trade restrictions) and the
explanatory variables does not bias the resulted The "within"
estimator used in the regression analysis of the panel data
with time effects (OLS on P. W. and V) is consistent as n or
t becomes large. See Judge at al. (1980. pp. 328-31. 336-
38).

‘The hypothesis that V) = v2 - = v14 can be
tested using a joint F test comparing the restricted and
unrestricted residual sum of squares. It is obvious in the
present case that the hypothesis is rejected.

CHAPTER V1

ANALYSIS OF INDUSTRIES

This chapter integrates the analysis of Chapter V by
examining the performance of unit labor costs for individual
manufacturing industries in the United States. Japan.
Germany. Italy. and Great Britain. The industries are
categorized by low. medium. and high technology. based on
U.S. Department of Commerce definitions and similar to the
classification used in Chapter IV.1 Each industry’s level of
labor productivity. labor compensation. and unit labor costs
is examined as a ratio to the manufacturing average for each
country. The relative unit labor costs can then be compared
between countries to determine comparative advantage. and
industries with a comparative advantage are expected to have

a trade surplus.

Agriculturgl Manufgcturing

Tables 5.1-5.4 show that all the countries ranked high
by unit labor costs in food products. The LLS. ranking was 7
in 1967 and 5 in 1980. Japan ranked 10 in both 1967 and
1980. during which time its costs of production remained
stable. The food products industry in Germany and Italy
ranked 3 by unit labor costs in 1980. while Britain ranked 8.

From 1967 to 1980. thetLLS. food products industry

strengthened its position by increasing labor productivity

95

96

and decreasing compensation relative to all manufactures. In
response. the U.S. trade position improved. U.S. food
products registered a trade deficit in 1967 but recorded a
surplus in 1980. Given the time series evidence in Chapter
V. LLS. food products"trade balance is sensitive to changes
in unit 'labor costs. The conclusion is that the U.S. food
products industry improved its trade balance by reducing its
unit labor costs.

Each of the countries examined ranked high in the
beverage industry category. The LLS. and Japanese industries
ranked 4 by unit labor costs in 1967 and 1980. Germany
ranked 4 in 1980. while Britain and Italy ranked 2. Still.
U.S. beverages registered trade deficits in 1967 and 1980.
Although the industry decreased its unit labor costs by
increasing relative productivity» its trade position
deteriorated between the two periods. The time series
evidence suggests that the U.S. beverage industry's trade
position was not sensitive to changes in unit labor costs.
Thus. demand factors and trade restrictions were responsible
for the deterioration in its trade balance.

The United States. Germany. and Britain held a
comparative advantage in the tobacco industry. The LLS.
tobacco industry ranked 1 by unit labor costs in 1967 and 2
in 1980. Germany ranked 1 in 1980. and Britain ranked 3. All
of these high rankings due to high productivity levels.
resulting in large trade surpluses for each country. In

contrast. the Japanese and Italian tobacco industries are

97

very small. Japan ranked last by unit labor costs in both
1967 and 1980. while Italy ranked last in 1980. There was no
value added in the Japanese tobacco industry. and the
average wage per employee was high relative to the

manufacturing average.

Traditional Low Technolggy Industries

Tables IV-B to VIII-B show'that the traditional low
technology industries are characterized by low levels of
labor productivity and compensation relative to the
manufacturing average within each country. The United
States was able to reduce its unit labor costs and improve
its ranking in each of the traditional low technology
industries because of reductions in labor compensation
relative to all manufactures. Nevertheless. neither the
United States. Japan. Germany. nor Britain ranked high by
unit labor costs in textiles. wearing appanel. or leather.
Italy exhibited a comparative advantage in the leather
products industry. which ranked 6 in 1980.

While no country ranked high in footwear. the United
States enjoyed a decided cost advantage. This was primarily
because of the low level of compensation in U.S. footwear
relative to all manufactures. Although the United States
recorded trade deficits in all the traditional low
technology industries. performance varied among industries.
In textiles. the United States recorded trade deficits in
1967 and 1980. and by keeping its relative unit labor costs

stable between the two years. the industry's trade deficit

98

remained constant.

U.S. wearing apparel and leather sustained large trade
deficits in both 1967 and 1980. In contrast. U.S. footwear
registered only a small trade deficit in 1967 but a large
one in 1980. Although U.S. wearing apparel. leather. and
footwear were able to lower their relative unit labor costs
from 1967 to 1980. their trade position deteriorated. Since
the trade positions of these sectors did not improve with
lower unit labor costs. trade restrictions and demand
factors were responsible for the distortion of these

industries' trade patterns.

Forest Products

Tables 5.1-5.4 show that Germany and Italy held a
comparative advantage in the manufacture of wood products
and furniture. while the United States and Japan displayed a
comparative disadvantage. U.S. wood products ranked 27 by
unit labor costs in 1967 and 24 in 1980. Japan ranked 22 in
1967 and 1980. The German wood products and furniture
industry ranked 8 by unit labor costs in 1980: Britain's
ranked 13 and Italy's 7.

The strong positions of Germany and Italy in wood
products was due to their high labor productivity.
Nevertheless. both generated large trade deficits in wood
products for 1980 despite their low unit labor costs. But
since there is relatively little trade in wood products.
changes in its unit labor costs have little effect on trade

flows. U.S. wood products sustained trade deficits in 1967

99

and 1980. Despite lowered unit labor costs. the wood
products trade balance did not change significantly.

Italy held a comparative advantage in the furniture
industry. resulting in a trade surplus. In contrast. the
United States. Japan. and Britain held a comparative
disadvantage. Their high unit labor costs were caused by low
labor productivity relative to all manufactures. The U.S.
furniture industry ranked 24 by unit labor costs in 1967 and
22 in 1980. Japan ranked 26 in 1967 and 20 in 1980.
Britain’s ranked 22 in 1980. and Italy's 8.

The furniture industry is characterized by low labor
productivity relative to the manufacturing average. The
United States. Japan. and Britain had high levels of labor
compensation relative)to»productivity. resulting in high
unit labor costs. The U.S. furniture industry suffered trade
deficits in 1967 and 1980. although the industry was able to
lower its relative unit labor costs between 1967 and 1980.
its trade position deteriorated over the period.

Tables 5.1-5.4 show that the United States displayed a
comparative advantage in the paper industry. and because of
increased productivity its cost advantage over the other
countries widened. The U.S. paper industry ranked 11 by unit
labor costs in 1967 and 9 in 1980: Japan ranked 13 in 1967
and 11 in 1980. Germany's industry ranked 18 in 1980.
Britain’s 18. and Italy's 16. The U.S. paper industry
registered trade deficits in both years despite its
comparative advantage. Since relative unit costs were

reduced during this period with no improvement in the paper

100

industry”s trade position. demand had to play the major role

in determining trade flows.

Nonmetal Products

Germany and Italy held a comparative advantage in the
pottery and china industry. while the United States and
Japan had a comparative disadvantage. The U.S. pottery and
china industry ranked 28 by unit labor costs in 1967 and 25
in 1980: the comparable figures for Japan were 26 and 27.
The industry in Germany ranked 6 in 1980. in Britain 23. and
in Italy 4. Although labor productivity in this sector was
relatively low in all the countries. German and Italian
labor productivity was high relative to its compensation.
This resulted in low unit labor costs and a strong trade
surplus for both countries.

The U.S. pottery and china industry recorded trade
deficits in 1967 and 1980. Despite reducing relative unit
labor costs between the two years. the pottery industry’s
trade position did not improve. Since the industry's trade
position was not responsive to supply considerations. demand
factors and trade restrictions were the major forces in
determining trade flows.

Japan displayed a comparative advantage in the glass
industry. which ranked 5 in 1967 and 8 in 1980. The high
ranking was due to high productivity relative to the
manufacturing average. Japan was able to translate its
relative cost advantage in glass manufacturing into a strong

trade surplus.

101

The U.S. glass industry's unit labor costs increased
between 1967 and 1980. thus reducing its ranking from 12 in
1967 to 23 in 1980. The U.S. glass industry recorded trade
deficits in both years. and in conformity with the
comparative cost theory. its trade position deteriorated.

The United States. Japan. Germany. and Britain ranked
high in other nonmetal products. U.S. nonmetals ranked 9 by
unit labor costs in 1980 and 10 in 1967: for Japan the
ranking was 9 in both years: Germany ranked 5 in 1980.
Britain 6. U.S. nonmetal products recorded trade surpluses
in 1967 and 1980. Since its relative unit labor costs
remained stable during this period. the industry”s trade

position also remained stable.

Medium Technology Industrigg

 

Medium technology manufacturing has been defined as the
iron and steel. nonferrous metals. metal products. energy.
rubber. transport equipment. and publishing industries. A
problem exists with the aggregation of categories. For
example. the transport equipment classification consists of
medium1and high technology industries. whcin means that only
broad conclusions can be drawn. Another problem is the
inclusion of the energy classification. dominated by demand
factors which lead to extraordinarily high values for labor

productivity. thus distorting the analysis.

102

Iron and Steel

The U.S. iron and steel industry has experienced
increasing pressure from international competition over the
past twenty years (Kreinin. 1984). The main factors
responsible for the decline in UJL competitiveness were
rapid escalation of labor compensation relative to the
manufacturing average and slow adaptation of new
technologies. leading to declines in relative productivity.
Both factors increased iron and steel’s relative unit labor
costs. thus reducing its internal ranking and international
competitiveness.

An indication that new technologies were not adopted in
the iron and steel industry are estimates that less than 30
percent of U.S. steel was continuously cast in 1983. vs. 90
percent of Japanese steel. The continuous casting process
turns out a single strand of molten metal for simultaneous
rol 1 ing. thereby el iminating ingot casting. reheating and
trimming.

Imported tonnage accounted for 22 percent of total LLS.
steel sales in 1982. Although most of this was imported from
Japan and Europe. an increasing amount coming from such
Third World producers as Brazil. South Korea. Mexico.
Taiwan. and Trinidad and Tobago. After European and
Japanese producers agreed in 1982 to limit steel exports to
the United States. developing countries quickly increased
their exports.2

Tables 5.1-5.4 show that Japan and Germany held a

comparative advantage in iron and steel production. whi le

103

the United States and Britain had a comparative
disadvantage. The U.S. ranked 19 in 1967 and last in 1980:
the figures for Japan 8 and 6. Germany ranked 6 in 1980.
Britain ranked last. and Italy 9.

The Japaness iron and steel industry was able to
strengthen its position because of increases in labor
productivity while maintaining its wage structure. The U.S.
industry. in contrast fel l drastical ly in ranking by unit
labor costs. Despite a slight increase in labor
productivity. it experienced large increases in labor
compensation relative to all manufactures. U.S. iron and
steel registered trade deficits in 1967 and 1980. Since its
relative unit labor costs increased significantly. its trade

position also deteriorated rapidly.

Nonferrous Metals

An important factor affecting this sector is increasing
government ownership of nonferrous metals industries. Over
the past decade. the major Third World copper producers --
Chile. Zambia. and Zaire -- have nationalized production. It
is estimated that 55 percent of the world's output is now
produced by government corporations compared to 33 percent
ten years ago. The government enterprises are not motivated
by profit but instead are concerned with domestic
employment. foreign exchange. and the demands of
international lending agencies. They often ignore economic
conditions in the copper markets and keep producing in order

to maintain employment and foreign exchange. In the process

104

they depress metals prices and export unemployment.

The private aluminum industry has also been hurt by
high levels of production from government-owned smelters in
Venezuela and Bahrain. By 1985. Alcoa estimates that
government smelters will account for 50 percent of the
industry"s output compared to 15 percent in 1960.

Japan held a comparative advantage in the manufacture
of nonferrous metals. while Germany and Italy had a
comparative disadvantage. U.S. nonferrous metals ranked 8 in
1967 and 13 in 1980: Japan’s ranking was 7 in both years.
Germany ranked last in 1980. Britain 14. and Italy 20.

Tables V-B and X-B show that Japan widened its
advantage in this sector by increasing relative labor
productivity and decreasing labor compensation. In the U.S.
industry. relative unit labor costs remained stable between
1967 and 1980.1L5u nonferrous metals recorded consistent
trade deficits in 1967 and 1980: since unit labor costs did
not change. the industry's trade position also remained

stable.

Metal Products

The metal products category is predominantly composed
of medium technology industries. except for ordnance. which
because of high levels of R30. has been classified as high
technology. However. ordnance products constitute a smal 1
share of the overall category and should not bias the

results.

105

Although none of the countries ranked high in metal
products. the United States and Italy had the lowest
relative unit labor costs. thus yielding a comparative
advantage. Rankings for the U.S. industry were 18 in 1967
and 21 in 1980. Comparable rankings for Japan were 20 and
18. Germany ranked 22 in 1980. Britain 20. and Italy 12.

Tables IV-B and IX-B show that the U.S. metal products
industry slightly decreased its unit labor costs by
increasing labor productivity and decreasing compensation
relative to all manufactures. The U.S. industry registered
trade surpluses for 1967 and 1980. Since its relative unit
labor costs did not significantly change over the period.

its trade position also remained stable.

Rubber Products

Although no country ranked high in rubber products. the
United States. Germany. and Italy displayed a comparative
disadvantage. The U.S. industry ranked 13 by unit labor
costs in 1967 and 26 in 1980: in Japan. 14 and 15 for these
years. Germany ranked 20 in 1980. Britain 16. and Italy 21.

Table 5.1 shows that the U.S. rubber industry
experienced a significant reduction in ranking by unit labor
costs between 1967 and 1980. The causes were the large
decreases in productivity per employee and rises in
compensation relative to all manufactures. This drop in
ranking was reflected in the industry’s trade position.
Although there were trade deficits in 1967 and 1980. these

had increased by 1980. The higher unit labor costs of the

106

rubber industry were responsible for its trade balance

deterioration.

Transport Egyipment

The transport category is aggregated. consisting of both
medium and high technology industries. Since the individual
industries in this group are heterogeneous. consistent
conclusions about individual ones are impossible. but broad
conclusions about the category can still be drawn.

The motor vehicle industry has requested trade

restrictions to stem the tide of imports from Japan and

Europe. In response. the Japanese have extended their
"voluntary" limit on U.S. exports through March 1985. The
new passenger-car limit. starting April 1. 1984. is 1.85
million vehicles compared to 1.68 million annually the

previous three years. which signifies that the U.S. industry
will continue to enjoy strong nontariff protection. This
resulted in U.S. automotives experiencing high average
compensation per employee. poor quality standards. and
inefficient plant. However. a slow improvement in
international competitiveness may be under way (Kreinin. 1984).
Tables 5.1-5.4 show that Japan held a comparative
advantage in the manufacture of transport equipment. The
U.S. industry ranked 22 in 1967 and 27 in 1980. Japan's
rankings were 12 and 14 respectively. Germany ranked 17 in
1980. Britain 27. and Italy 22. Although the U.S. transport
industry improved its relative labor productivity.

compensation per employee increased even faster. This

107

resulted in higher unit labor costs and a lower ranking: it
registered a trade surplus in 1967 and a trade deficit in

1980. consistent with the comparative cost theory.

Publighing

The United States and Britain displayed a comparative
advantage in the printing industries. The LLS. rankings were
15 by unit labor costs in 1967 and 11 in 1967. Comparable
Japanese ranking were 19 and 17. The figure for Germany was
19 in 1980. for Britain 9. and for Italy 16. The U.S.
printing industry was able to improve its ranking with cuts
in labor compensation. The time series evidence suggests
that its trade balance was sensitive to changes in unit
labor costs. Thus in conformity with comparative cost
theory. the U.S. printing industry was able to increase its

trade surplus between 1967 and 1980.

Energy

The energy industries consists of petroleum refining
and petroleum and coal products. All the countries studied
appeared to exhibit a comparative advantage in this sector.
but the category is overwhelmingly dominated by demand
factors which distort the analysis.

The U.S. petroleum refining industry ranked 2 by unit
labor costs in 1967 and 1 in 1980. Japan ranked 1 in both
years. Britain and Italy ranked 1 in 1980 and the German
petroleum refining and other petroleumland coal products
industries ranked 2 by unit labor costs in that year. The

United States. Japan. and Germany each recorded trade

108

deficits in the petroleum refining industry for 1980.
despite its high ranking. Thus. demand was the primary
determinant of trade flows.

The United States. Japan. and Britain ranked high in
petroleum and coal products. The U.S. ranking by unit labor
costs was 6 in both years. Japan ranked 6 in 1967 and 3 in
1980. Britain ranked 5 in 1980. U.S. petroleum and coal
products registered a trade deficit in 1967 but a trade
surplus in 1980. In conformity with the comparative cost
theory. the U.S. industry reduced its relative unit labor

costs and improved its trade position.

High Technology Industries

In the high technology manufacturing category are
industrial chemicals. other chemicals. plastics. machinery.
electrical machinery. and professional goods. A severe
problem with this classification is the aggregation in most
of the industries. especially industrial chemicals.
machinery. and electrical machinery. They are composed of
both medium and high technology industries. which distorts

the unit labor costs of the overall category.

Ingustrigl Chgmicglg
Although the category is highly aggregated. except

for industrial organic chemicals (medium technology). the
sector employs high technology. Thus. general conclusions

can be reached.

109

Tablee.5u1 and 5.2 show that while the United States
and Japan have had a comparative advantage in industrial
chemicals. Japan’s advantage appears to be disappearing. To
obtain stronger conclusions. disaggregated data must be
observed for Japan. but unfortunately these have not been
published. The U.S. industrial chemical industry ranked 3 by
unit labor costs in 1967 and 8 in 1980. Japan ranked 2 in
1967 and 5 in 1980. The German figures. which includes al 1
chemical industries. was a ranking of 14 in 1980 and its
wage structure has remained stable since 1967.

According to Tables IV-B and IX-B. both the U.S. and
Japanese industries increased their relative unit labor
costs from 1967 to 1980. This has stemmed from decreases in
labor productivity and increases in compensation relative to
all manufacturing. Nevertheless. the high technology segment
of industrial chemicals. examined in Table 4.1. maintained
its costs structure. LLS. industrial chemicals registered
trade surpluses.in I967.and 1980.*widening the surplus in

the latter year.

Other Chemigalg

Other chemicals consists of the categories drugs.
soap. paints. and miscel laneous chemicals. Aside from the
drug industry. this category employs medium technOIOQY. but
since drugs are the most important segment. the whole
category is included here as a high technology industry.

Tables 5.1. 5.2. and 5.4 show that the United States.

Japan. and Britain achieved a high ranking in other chemical

110

industries. The U.S. ranking by unit labor costs was 5 in
1967 and 3 in 1980. Japan ranked 3 in 1967 and 2 in 1980.
Britain ranked 4 in 1980. Although Japan’s other chemicals
kept unit labor costs constant between 1967 and 1980. the
0.5. industry decreased unit labor costs substantial ly. This
drop was owing to large increases in labor productivity and
decreases in compensation relative to the manufacturing
average. U.S. other chemicals registered trade surpluses in
1967 and 1980. By reducing unit labor costs. the industry":
trade position improved between the two years in conformity

with the comparative cost theory.

Plastics

The plastics industry is characterized by low labor
productivity relative to all manufactures within each
country. While no country ranked high in plastics. the
United States held a comparative advantage and Italy
a comparative disadvantage. U.S. plastics ranked 14 in 1967
and 1980: Japan ranked 15 in 1967 and 12 in 1980. Germany
ranked 15. Britain 11. and Italy last in 1980.

Tables IV-B and IX-B show that LLS. plastics decreased
unit labor costs between 1967 and 1980 by decreasing
compensation per employee relative to the manufacturing
average. Since the U.S. industry also registered trade
surpluses in both years. the improvement in its trade

position conforms with the comparative cost theory.

Machinery

The aggregated machinery category consists of both
medium and high technology industries. The former farm and
garden. metalworking. and refrigeration machinery. The
latter are engines and turbines. construction. and office
and computing machinery. However. even the last segment is
not homogeneous. containing medium technology calculating
and accounting machines and typewriters as well as high-
speed computers. Thus. consistent conclusions are impossible
to reach about the individual industries.

Although neither the United States nor Japan held a
clear comparative advantage in the machinery industry.
Germany had a clear comparative disadvantage. This was due
to Germany's low level of productivity and high level of
compensation relative to al 1 manufactures. U.S. machinery
ranked 17 by unit labor costs in both 1967 and 1980: Japan
ranked 16 in 1980 and 17 in 1967. Germany ranked 21 in 1980.
Britain and Italy 12.

The U.S. machinery industry decreased its unit labor
costs between 1967 and 1980. This was achieved by increases
in labor productivity and decreases in compensation in
comparison to al I manufactures. The industry registered a
small trade surplus in 1967 and a large one in 1980. By
reducing unit labor costs. the machinery industry"s trade
position improved over the period in conformity with the

comparative cost theory.

112

Electrical Machinery

The highly aggregated electrical machinery category
consists of medium and high technology industries. The
former are household appliances. electrical lighting. and
radio and TV receiving equipment. The latter are electrical
distributing and communication equipment and electronic
components” The dichotomy'within this sector makes strong
conclusions impossible.

None of the countries ranked high in electrical
machinery. although the United States had the lowest unit
labor costs. thus yielding a comparative advantage.lLS.
electrical machinery ranked 20 by unit labor costs in 1967
and 16 in 1980: Japan ranked 11 in 1967 and 13 in 1980.
Germany ranked 16 in 1980. Britain 15. and Italy 19.

In U.S. electrical machinery. unit labor costs
decreased between 1967 and 1980 owing to increased labor
productivity and decreased compensation relative to all
manufactures. Nevertheless. the trade surplus in 1967 had
changed to a deficit by 1980. Although the U.S. industry
lowered relative unit labor costs between the two years. its
trade position deteriorated. Since supply factors improved.

demand factors were responsible.

Prof sio al s

The professional goods category consists primarily of
engineering and scientific devices and medical instruments.
With few exceptions. the entire category'is classified as

high technology.

113

Tables 5.1-5.3 show that the United States and Germany
held a comparative advantage in the manufacture of
professional goods while Japan had a comparative
disadvantage. U;S. professional goods ranked 10 by unit
labor costs in 1967 and 7 in 1980. The respective Japanese
rankings were 24 and 23. Germany ranked 7 in 1980. Britain
19. and Italy 11.

The strong U.S. position stems from high levels of
relative labor productivity. whereas Germany’s is derived
from low relative compensation levels. Japan. in contrast.
displayed very low levels of labor productivity relative to
compensation. The U.S. professional goods industry decreased
unit labor costs between 1967 and 1980 with large increases
in labor productivity and moderate decreases in compensation
relative to all manufactures. The UJL industry recorded
trade surpluses in both years and. by moving up in ranking.
was able to improve its trade position over the period. in

conformity with the comparative cost theory.

Summary

The analyses of individual industries indicates that in
1967 and 1980. the United States held a comparative
advantage in the manufacture of food‘products. tobacco.
paper. printing. industrial chemicals. plastics. and
professional goods. It held a comparative disadvantage in
iron and steel. rubber. transport equipment. wood products.
and glass. The industries which experienced a reduction in

international competitiveness between 1967 and 1980 were

114

iron and steel. transport equipment. glass. and rubber.

In 1967 and 1980. Japan held a comparative advantage in
glass. iron and steel. nonferrous metals. transport
equipment. and industrial chemicals: a comparative
disadvantage in tobacco. food products. traditional ly low
technology products. and professional goods. Japan’s
comparative advantage remained relatively stable between
1967 and 1980.

In 1980. Germany held a comparative advantage in
tobacco. food products. wood products. pottery. and
professional goods: a comparative disadvantage in paper.
nonferrous products. metal products. rubber. industrial
chemicals. and machinery.

In 1980. Italy held a comparative advantage in leather.
forest products. and pottery. It displayed a comparative
disadvantage in tobacco. nonferrous metals. rubber.
plastics. and electrical machinery.

In 1980. Britain held a comparative advantage in
industrial chemicals. plastics. tobacco. printing. and other
nonmetal products. Its comparative disadvantage lay in
wearing apparel. furniture. pottery and china. iron and
steel. and transport equipment.

Knowledge of a country's comparative advantage at a
particular time and of the changes over time are useful to a
policy maker. In order for an industry whose ranking is low
and which is suffering from a trade deficit to improve its
position. it must decrease unit labor costs by reducing

labor compensation. increasing productivity. or both.

115

Although all industries do not respond quickly to changes in
supply. these do explain a large portion of trade flows and
must be considered in the formulation of international

policy decisions.

1 16
Footnotes

1For a detailed breakdown of industry categories see
Davis (1982). Kelly (1977). and Boretsky (1982). all from
the U.S. Department of Commerce.

2Background material for the medium and
high technology industries was obtained from Forbes. January
1984.

CHAPTER VII

CONCLUSION

The purpose of this dissertation was to investigate the
comparative cost advantage of manufacturing industries in
the United States. Japan. and Europe between 1967 and 1980.
Since it was not possible to derive accurately all the costs
of production. the internal ranking of industries by
comparative advantage was'approximated using the inverse of
unit labor cost and its components. labor productivity and
labor compensation.

Although most relevant studies have measured an
industry"s unit labor costs across countries at one time.
they were limited in scope and subject to distortions
stemming from deviations in the long-run equilibrium
exchange rate. The distortions can be corrected for in an
approach. developed by Viner and Kreinin. which examines an
industry’s unit labor costs relative to the national
manufacturing average. Thisiallows all commodities produced
by a country to be ranked internally by their domestic unit
labor costs. Each country'would export the commodities in
which its comparative cost advantage is the greatest and
import those ranking the lowest (highest production costs).
with the exchange rate separating the two. Thus. comparative

cost advantages "are embedded in the total economy” and

117

118

become an effective tool on which to base policy decisions
(Kreinin. 1984. p. 42).

Increases in an industry’s unit labor costs relative
to the manufacturing average reduce its internal ranking and
international competitiveness. Since comparative advantage
is important to the determination of trade flows. and in
light of the large U.S. current account deficits registered
recently. industries subject to acute import competition can
alleviate their situation by decreasing unit labor costs.
Therefore. increasing labor productivity. decreasing labor
compensation. or both. become important to improving
internal ranking and strengthening an industryms trade
position without relying on trade restrictions.

The theoretical foundations of the classical
comparative cost model were developed beginning with the
Ricardian hypothesis. The assumptions of the latter were
relaxed with the introduction of money wages. a
multicommodity framework. transportation costs. and
incomplete specialization to construct a more realistic
model. It was demonstrated that the sharp distinction drawn
between the Ricardian and Heckscher-Ohlin theories
disappeared with the relaxation of the assumptions. The
differences between countries in comparative labor
productivities that were asserted by assumption in the
Ricardian hypothesis were deduced from differences in
relative factor endowments in the Heckscher-Ohlin model.
Therefore. these two theories do not provide conflicting

viewpoints. but instead are interdependent and are both

119

important to understanding international trade flows.

MacDougall. Balassa. Stern. and Stryker's empirical
results confirmed the comparative cost theony. MacDougall
found a strong inverse relationship between 1937 relative
U.S. and U.K. wage costs per unit of output and relative
exports and a positive relationship between relative output
per worker and relative exports. Balassa and Stern's results
indicated that the 1950 relative U.S. and U.K. export
performance followed established lines of comparative
advantage as suggested by labor productivity. The
introduction of differences in capital costs per unit of
output did not have any significant effect on export
performance. Stryker's time series evidence suggested that
changes in U.S. and Canadian exports were correlated with
changes in labor productivity and that the latter is a
useful proxy for total factor productivity.

In contrast. Bhagwati. Stern. and Kreinin concluded
that their empirical evidence did not support the simple
Ricardian hypothesis. and that other factors were important
in determining the pattern of trade. Although Bhagwati's
tests produced a poor fit between 1950 relative labor
productivity and price ratios and the relationship between
comparative unit labor costs and export price ratios. the
variables were measured inconsistently. Thus. as Daly
stated. failures to establish a correlation between export
price ratios and labor productivity ratios stem from the

lack of conceptual and statistical compatibility in the data

120

rather than provide evidence against the Ricardian
hypothesis.

The statistical results in this dissertation generated
from the cross-sectional and time series data confirmed the
importance of unit labor costs as a determinant of
international trade patterns. With the exception of Japan.
which generated a positive sign for unit labor costs in 1967
and 1980. the coefficients generated by the cross-sectional
models were significant and of the correct sign. The
coefficient regressions from these equations were low.
which was not unexpected considering the multitude of
factors affecting international trade.

For the time series model. correct signs for labor
compensation and productivity were produced for most of the
industries when using the relative version of the
comparative cost theory. Finally. the panel data model
removed the biases of changes in differential trade
restrictions and exchange rates and generated positive
results based on all available data for the U.S. economy
over the period 1967 to 1981. Thus. the empirical evidence
reveals that unit labor costs are an important determinant
of international trade flows and must be considered in the
development of policy for industries suffering from import

,competition.

The results for 1967 and 1980 indicated that the United
States held a comparative advantage. as approximated by unit
labor costs. in the following industries: agricultural

manufacturing. paper. nonferrous metals. publishing.

121

industrial chemicals. plastics. office and computing
equipment. and professional goods. It held a comparative
disadvantage in traditional low technology products. wood
products. furniture. pottery. glass. rubber. iron and steel.
and most transport equipment.

The three major industries in the UAL economy which
experienced a reduction in international competition between
1967 and 1980 were iron and steel. transport. and rubber. It
was seen that these significantly increased their unit labor
costs relative to other manufactures. causing them to fall
sharply in the internal ranking. This was reflected in the
large trade deficits they incurred.1ricontrast. the iron
and steel. transport and rubber industries in Japan and
Germany were able to maintain their internal ranking. Thus.
the increases in international competition for these
industries resulted from changes in unit labor costs within
the United States and not from the extraordinary performance
of foreign competitors.

In contrast. the aircraft. agricultural chemicals. and
professional goods industries decreased their unit labor
costs. which improved their trade position. But all
industries did not conform to comparative cost theory. The
trade position of some industries was not effected by supply
but instead dominated by demand factors. trade restrictions.
and transportation costs. Nevertheless. supply
considerations did explain a large portion of the trade

flows.

122

In 1967 and 1980. Japan held a comparative advantage in
glass. iron and steel. nonferrous metals. transport
equipment. and industrial chemicals: a comparative
disadvantage in tobacco. food products. traditional ly low
technology products. and professional goods. Japan’s
comparative advantage remained relatively stable between
1967 and 1980.

In 1980. Germany held a comparative advantage in
tobacco. food products. wood products. pottery. and
professional goods: a comparative disadvantage in paper.
nonferrous products. metal products. rubber. industrial
chemicals. and machineryu Italy held a comparative advantage
in leather. forest products. and pottery. It displayed a
comparative disadvantage in tobacco. nonferrous metals.
rubber. plastics. and electrical machinery. Great Britain
held a comparative advantage in industrial chemicals.
plastics. tobacco. printing.iand other nonmetal products.
Its comparative disadvantage lay in wearing apparel.
furniture. pottery and china. iron and steel. and transport
equipment.

This evidence establishes the importance of considering
unit labor costs in the formation of policy for industries
experiencing international competition. Knowledge of a
countrst comparative advantage at a particular time and the
rate of its change are useful to policy makers. An industry
‘with low ranking which is suffering from a trade deficit can

strengthen its position through decreased unit labor costs:

by reducing labor compensation. increasing productivity. or

123

both. Thus. internal responses to international competition
are vital. without reliance on trade restrictions. to

solving an industry's trade problems.

APPENDIX A

TABLE I-A

Concordance Between the Three-Digit International
Standard Industrial Classification (top number) and
the U.S. Standard Industrial Classification

 

311/2 313 314 321 322 323 324
201 2082 21 22 2311 3111 3131
202 2083 239 232 3151 314
203 2084 233 3161
204 2085 234 317
205 2086 235 3199
206 236
207 2371
2087
209
331 332 341 342 351 352 353
24 25 26 27 281 283 2911

282 284
286 2851
287 289
354 355 356 361 362 369 371
295 3011 3079 326 3211 3241 331
299 3021 322 325 332
3031 3231 327
3041 3281
3069 329
372 381 382 383 384 385 390
333 34 35 36 37 38 39
3341
335
336
339

124

125

TABLE II-A

Concordance Between the 1980 Three-Digit International
Standard Industrial Classification (top number) and
the Standard International Trade Classification

 

311/2 313 314 321 322 323 324
012 11 122 26 84 61 851
014 65 831
035
037
046 331 332 341 342 351 352
047 ------------------------------
048 24 821 251 892 51 53
056 63 64 52 541
058 562 55
061.2 58 572
062 591 598
071.2 592
072
073 353 354 355 356 361 362
091 ------------------------------
098 334 32 233 893 666 664
335 62 665
341
369 371 372 381 382 383 384
661 67 68 69 71 76 78
662 72 77 79
663 73
667 74
75
385 390
87 894
88 895
896
897
898

126

TABLE III-A

Concordance Between the 1967 Three-Digit International
Standard Industrial Classification (top number) and
the Standard International Trade Classification

 

311/2 313 314 321 322 323 324

012 11 122 266 84 61 851

013 267 831

032 65

046

047 331 332 341 342 351 352

048 ------------------------------

052 243 821 251 892 51 53

053 63 64 521 541

055 561 55

062 581 571

071.3 599

072.2

072.3 353/4 355 356 361 362 369

073 ------------------------------

091 321 231.2 893 666 664 661

099 332 231.3 665 662
341 231.4 663

62
371 372 381 382 383 384 385
67 68 69 71 72 73 B6
812

390

891

894

895

896

897

127

TABLE IV-A

OLS Results from Equation 5.1

 

Country A1 42 R2 N o-w
U.S.
1980 5.67 -8.42 0.25 27 1.50
(1.49) (2.91)
1967 17.14 -27.70 0.18 27 1.13
(6.18) (11.70)
Japan
1980 1.956 5.186» 0.008 26 1.93
(4.534) (12.038)
1967 -43.80 168.45 0.21 25 1.99
(23.55) (69.02)
Germany
1980 1.72 -0.66 0.013 23 1.64
(0.64) (1.22)
Britain
1980 3.2694 -0.0035 0.19 27 2.05
(0.0480) (0.0015)
Italy
1980 4.19 -1.41 0.0005 23 2.54

(7.75) (13.89)

 

 

 

Country

128

TABLE V-A

0L5 Results for Equation 5.2

 

BI 82 N D-W
U05.
1980 -I.68 -1.96 0.19 27 1.92
(0.62) (0.80)
1967 -2.17 -2.80 0.26 27 1.55
(0.72) (0.94)
Japan
1980 1.63 1.29 0.04 26 1.29
(1.39) (1.30)
1967 5.41 3.57 0.33 25 2.13
(1.22) (1.06)
Germany
1980 ~0.036 -0.219 0.03 23 1.07
(0.264) (0.294)
Britain
1980 6.70 -1.07 0.12 27 1.72
(3.70) (0.57)
Italy
1980 -0.86 -1.99 0.05 23 2.93

(1.82)

129

TABLE VI-A

OLS Results for Equation 5.3

 

Country 1 C1 C2 R2 N D-W
U.S.
1980 31.36 -38.94 0.15 27 1.89
(9.59) (18.73)
1967 47.54 -51.44 0.16 26 2.15
(12.65) (24.05)
Japan
1980 0.027 0.173 0.03 26 1.55
(0.072) (0.191)
1967 -0.062 0.483 0.20 25 1.83
(0.068) (0.200)
Germany
1980 8.51 12.48 0.15 23 1.79
(3.42) (6.46)
Britain
1980 0.2432 -0.00025 0.27 27 1.93
(0.0480) (8.31'10-8)
Italy
1980 14.82 -7.64 0.004 23 1.88

(14.89) (26.69)

130

TABLE VII-A

OLS Results for Equation 5.4

 

Country DI 02 R2 N D-W
U.S.
1980 -1.23 -2.60 0.03 27 0.43
(2.23) (2.87)
1967 -9.45 -0.93 0.08 26 2.15
(0.50) (0.66)
Japan
1980 -2.15 0.73 0.03 26 1.13
(0.86) (0.80)
1967 —0.34 2.06 0.26 25 1.57
(0.84) (0.73)
Germany
1980 2.82 0.24 0.11 23 1.73
(0.13) (0.15)
Britain
1980 4.08 -1.04 0.17 27 1.82
(2.91) (0.46)
Italy
1980 0.88 -1.72 0.08 23 2.16
(0.82) (1.29)

 

131

TABLE VI 1 I-A

OLS Results for Equation 5.5

 

Industry 41 42 43 R2 D-W

Food Products -4I49.71 -339.19 1270.63 0.86 0.78
(996.34) (267.61) (756.31)

Beverages 226.42 -67.53 82.95 0.98 2.57
(78.99) (19.23) (55.29)

Tobacco -222.51 7.16 20.17 0.99 1.19
(24.11) (3.27) (13.80)

Textiles ~641.90 -136.69 202.70 0.38 1.23
(217.01) (252.84) (457.33)

Wearing 3580.20 ~408.05 ~236.83 0.99 1.84
Apparel (510.70) (245.87) (499.43)

Leather and 413.16 -20.03 -62.81 0.91 1.54
Products (175.36) (67.28) (134.31)

Footwear 2177.28 134.72 -691.61 0.95 0.99
(301.67) (99.75) (206.47)

Wood Products -80.71 -248.17 378.07 0.50 1.56
(325886) (87.83) (157.88)

Furniture. 468.96 61.14 -I92.84 0.94 1.36
Fixtures (105.93) (75.70) (139.60)

Paper and -38.78 65.45 -I91.20 0.44 1.40
Products (272.49) (67.76) (149.00)

Printing. 240.45 109.83 -210.72 0.87 1.06
Publishing (134.03) (35.18) (78.09)

Industrial -3483.81 1.37 455.38 0.87 0.97
Chemicals (929.97) (104.58) (445.74)

Other Chemical -266.14 85.33 -190.29 0.90 1.31
Products (268.14)

(205.99)

(73.19)

I32

 

Petroleum
Refineries

Petroleum.
Coal Products

Rubber Products

Plastic

Products. n.e.c.

Pottery.

China. etc.

Glass and

Products

Nonmetal
Products

Iron and
Steel

Nonferrous

Metals

Metal
Products

Machinery

n.e.c.

Electrical
Machinery

Transport
Equipment

Professional

Goods

Other

Industries

3498.82
(1693.00)

- 235.21
(22.78)

473.44
(222.53)

-522.65
(85.41)

223.38
(16.11)

112.80
(134.97)

-108.75

2998.99
(837.11)

(659.78)

(327.66)

-10564.53
(2985.61)

~200.7B
(870.77)

3068.32
(1393.00)

-934.97
(264.60)

902.34
(640.24)

-9.74
(22.53)

0.32
(3.27)

24.41
(50.96)
-45.35
(38.79)

-28.83

-10.40
(47.08)

-48.85
(30.82)

-I79.45
(181.76)

-15.66
(113.52)

-66.27
(195.47)

-1130.88
(999.07)

(301.03)

-605.89
(407.95)

-93.33
(96.98)

(218.42)

-386.74
(184.00)

26.64
(8.48)

~121.96
(84.20)

164.04
(78.96)

-3.07
(19.05)

12.82
(79.49)

106.81
(66.18)

-78.93

-64.82
(221.97)

218.33
(358.31)

3522.94
(1988.64)

732.75
(590.86)

791.54
(617.71)

372.69
(222.35)

519.13
(484.34)

0.86

0.97

0.76

0.89

0.997

0.06

0.89

0.31

0.70

0.95

0.33

0.35

0.91

0.92

1.51
1.14

0.91

0.91

2.16

0.98

1.94
1.97
1.24
0.91
1.46
1.33
1.49

0.87

133

TABLE IX-A

OLS Results for Equation 5.6

 

 

Industry 81 82 83 R2
Food Products 0.043 -I.836 2.532 0.88
(0.529) (0.832) (0.937)
Beverages -3.344 -2.737 3.975 0.94
(1.139) (1.299) (1.361)
Tobacco -2.703 3.534 -3.590 0.35
(3.038) (1.864) (1.782)
Texti les -2.485 0.474 0.430 0.68
(2.022) (4.365) (4.562)
Wearing -1.301 —2.562 2.897 0.06
Apparel (0.865) (2.995) (3.452)
Leather and -2.162 0.714 -0.463 0.25
Products (0.563) (1.995) (2.313)
Footwear -5.482 6.159 -6.829 0.71
(0.734) (2.092) (2.391)
Wood Products -0.785 -1.009 1.424 0.40
(0.534) (0.857) (0.904)
Fixtures (2.278) (5.655) (6.026)
Paper and -1.503 0.981 -0.816 0.48
Products (0.378) (0.622) (0.680)
Printing. -0.579 7.093 -8.467 0.71
Publishing (0.633) (1.926) (2.195)
Industrial 1.547 -0.044 -0.158 0.26
Chemicals (0.646) (0.634) (0.691)
Other Chemical 2.318 0.121 -0.628 0.76
Products (1.940) (1.650)

(1.606)

1.13

0.89

0.72

1.27

0.65

134

 

Petroleum
Refineries

Petroleum.

Coal Products

Rubber Products

Plastic

Products. n.e.c.

Pottery.
China. etc.

Glass and
Products

Nonmetal
Products

Iron and
Steel

Nonferrous
Metals

Metal
Products

Machinery
n.e.c.

Electrical
Machinery

Transport
Equipment

Professional
Goods

Other
Industries

0.941
(0.838)

15.187
(5.453)

1.823
(1.558)

-0.169
(0.898)

-I.038
(0.659)

2.443
(2.822)

0.535
(1.187)

0.908
(1.631)

-3.128
(2.482)

1.267
(0.866)

2.387
(0.519)

1.079
(0.614)

1.992
(0.835)

2.334
(0.627)

-0.404
(0.783)

0.344
(0.748)

-3.910
(1.790)

-1.203
(1.896)

-0.365
(1.457)

-2.319
(2.082)

-2.642
(3.281)

0.112
(1.706)

-0.820
(2.167)

0.168
(3.099)

0.570
(1.919)

-1.014
(1.459)

(1.793)

-2.088
(1.246)

-I.270
(0.942)

-0.412
(2.134)

-1.381
(1.154)

1.867
(1.306)

0.477
(1.785)

0.701
(1.538)

2.486
(2.232)

2.382
(3.006)

-0.295
(1.710)

0.248
(1.980)

0.589
(3.216)

-1.018
(2.018)

0.736
(1.606)

1.353
(1.990)

1.752
(1.187)

0.961
(0.989)

0.397
(2.428)

0044

0.63

0.39

0.27

0.09

0.05

0.26

0.62

0.31

0.41

0.74

0.04

0.56

1.87

0.89

0.81

0.77

0.68

0.90

1.51

1.08

0.85

0.56

0.99

1.27

0.88

 

135

TABLE X-A

OLS Results for Equation 5.7

 

 

industry ' c1 02 03 R2 o-w
Food Products 5603.59 697.50 -7112.52 0.54 1.84
(1733.79) (524.27) (1921.69)
Beverages 324.53 -373.39 182.94 0.13 1.45
(504.04) (302.13) (585.72)
Tobacco -0.22 0.72 -1.41 0.31 2.15'
(0.28) (0.35) (0.64)
Textiles 6.13 10.90 -IB.58 0.17 2.03”
(4.30) (9.81) (13.17)
Wearing 0.56 12.81 -12.12 0.06 2.16"
Apparel (2.55) (19.64) (15.54)
Leather and -92.98 -288.59 377.41 0.20 1.41
Products (101.03) (376.23) (266.18)
Footwear 0.40 -6.70 5.07 0.22 1.80“
(0.88) (4.01) (3.86)
Wood Products -40.67 238.26 -166.52 0.29 1.58
(296.46) (175.14) (509.90)
Furniture. -0.16 -2.32 2.21 0.04 1.78“
Fixtures (0.74) (3.58) (3.80)
Paper and 480.29 -7.40 -448.75 0.17 1.35
Products (314.67) (127.95) (333.61)
Printing. 38.34 294.83 -317.82 0.32 1.52
Publishing (85.49) (231.70) (166.90)
Industrial 3.24 -1.58 0.73 0.05 1.91'
Chemicals (6.63) (3.01) (8.76)
Other Chemical -0.31 2.74 -4.59 0.13 1.64“
Products (2.14) (9.76)

(9.55)

 

Petroleum
Refineries

Petroleum.
Coal Products

Rubber
Products

Plastic
Products

Pottery.
China. etc.

Glass and
Products

Nonmetal
Products

Iron and
Steel

Nonferrous
Metals

Metal
Products

Machinery
n.e.c.

Electrical
Machinery

Transport
Equipment

Professional
Goods

Other
Industries

136

12.73 0.94
(6.55) (0.46)
0.42 -0.05
(0.29) (0.13)
10.93 1.23
(161.85) (109.03)
199.37 -106.81
(176.20) (232.80)
52.52 222.11
(I49.78) (371.85)
15.53 1.83
(21.54) (10.49)
92.72 -58.37
(58.06) (29.94)
58.52 1137.40
(1231.77) (1181.25)
1133.59 308.69
(806.62) (270.24)
-0.41 4.86
(7.14) (7.43)
1707.37 14531.67
(13323.14) (7505.40)
-1.44 0.86
(1.62) (1.77)
-0.13 2.75
(1.88) (1.33)
589.28 614.49
(774.54) (501.88)
2.01 3.81

(3.11)

' Sample size =196B-1981

-10.64
(5.22)

-0032
(0.27)

-16.79
(184.21)

-127.83
(353.45)

-245.25
(372.37)

-16.80
(13.78)

-33.62
(66.40)

-953.88
(579.30)

-1301.57
(683.24)

-3.80
(11.38)

-14412.98
(12273.64)

0.75
(1.57)

-3096
(1.03)

-1217.16
(1065.93)

-6.29
(6.03)

0.16

0.34

0.20

0.27

0.05

0.25

0.08

0.39

0.12

1.40”

2.42*

1.07

1.39

1.00

1.99’

1.50.

1.17

2.07“

137

TABLE X I -A

OLS Results for Equation 5.8

 

 

industry 01 02 03 R2 D-W

Food Products -1.70 2.84 -12.56 0.37 0.77
(0.66) (1.82) (5.31)

Beverages -2.76 -3.54 17.63 0.83 1.60
(0.53) (1.70) (2.45)

Tobacco 1.62 1.04 —1.25 0.10 1.04
(1.57) (1.97) (1.42)

Texti les -7.15 -4.55 -9.50 0.89 1.63
(0.65) (2.93) (4.38)

Wearing -2.14 -0.10 0.02 0.002 1.22

Apparel (0.49) (1.56) (1.34)

Leather and -2.44 -I.39 -0.92 0.44 1.25

Products (0.52) (1.62) (1.35)

Footwear -3.70 2.24 -3.27 0.29 0.96
(0.50) (1.40) (1.52)

Wood Products 0.90 -2.73 9.21 0.22 0.99
(0.85) (1.55) (5.13)

Furniture. -1.59 3.30 -6.36 0.28 0.50

Fixtures (0.61) (4.04) (4.74)

Paper and -0.79 1.20 4.78 0.50 0.67

Products (0.13) (0.88) (2.24)

Printing. 0.86 3.41 -I.27 0.42 1.73

Publishing (0.11) (1.77) (1.32)

Industrial 1.50 -0.43 -0.62 0.22 1.01

Chemicals (0.56) (0.76) (0.99)

Other Chemical -2.05 3.88 4.36 0.58 1.40

Products (0.86) (1.43) (3.65)

Refineries (3.37)

(0.75)

(0.80)

-"

Petroleum,
Coal Products

Rubber Products

Plastic
Products. n.e.c.

Pottery.
China. etc.

Glass and
Products

Non-metal
Products

Iron and
Steel

Nonferrous
Metals

Metal
Products

Machinery
n.e.c.

Electrical
Machinery

Transport
Equipment

Professional
Goods

Other
Industries

3.99
(1.18)

-0.75
(0.17)

-0.27
(0.30)

-1.58
(0.30)

0.12
(0.15)

-0.13
(0.09)

-0.36
(0.45)

-3.74
(1.14)

0.39

-0.01
(0.25)

‘0.17
(0.13)

0.50
(0.18)

0.51
(0.07)

-1.14
(0.46)

138

-3.60
(3.25)

1.77
(1.08)

0.35
(1.80)

0.91
(0.64)

-1.61
(1.92)

1.53
(1.87)

0.52
(2.42)

0.65
(3.05)

1.46
(1.96)

-3074
(1.15)

-I.43
(1.13)

-0.46
(0.60)

-0.55
(0.95)

-1.96
(2.11)

-0.62
(5.58)

1.19
(1.98)

-3076
(2.77)

0.10
(0.07)

-0.22
(2.70)

-3090
(4.05)

-2.30
(1.50)

19.00
(8.01)

5.40
(3.35)

9.35

2.13
(1.15)

-1.46
(0.57)

3.69
(1.86)

0.65
(1.54)

0.11

0.34

0.27

0.25

0.06

0.08

0.17

0.33

0.66

0.66

0.94

0.77

0.92

2.62

0.56

1.26

1.34

1.33

1.16

1.33

APPENDIX 8

TABLE I-B

Classification of Two-. Three-. and Four-Digit
SIC Manufacturing Industries Into
Low. Medium. and High Technology
Categories

 

Low Technology Inggstries

20 - 26. 31. 32. 39

Medium Technolggy Industries
27. 284. 2851. 286. 289. 29. 30. 33. 341-347.
349. 352-356. 358. 359. 363. 364. 369. 371.

372. 3743. 3751. 379

filgh Technology Inggstries
281-283. 287. 348. 351. 357. 361. 362.
365-367. 372. 376. 38

This classification system is based on 1L5. Department
of Commerce publications. There are no classification
systems which can truely capture an industry's use of
technological production methods. These industries are
defined in a broard manner with many diverse segments so

that technological labels on a specific industry can only

139

140

have limited applications. The methodology of assigning
technology labels to industries is found in Davis (1982).

Boertsky (1982) and Kelly (1977L

141

TABLE II-B

_U.S. 1980. Two-. Three-. and Four-Digit
SIC Comparative Cost Rankings

Industry Value/ Comp/ Unit Rank
Employ Employ Value

 

 

(I) (2) (3)

Meat Products 30.84 17.34 0.562 90-93
Dairy Products 50.08 18.54 0.370 19-20
Preserved Fruits 40.81 15.45 0.379 21
Grain Mill Products 80.18 21.25 0.265 5
Bakery Products 39.27 19.15 0.488 44
Sugar. Confectionery 61.03 17.96 0.294 7
Fats and Oils 66.42 20.12 0.303 10
Beverages 68.95 22.08 0.320 14
Misc. Foods Kindred 52.70 15.63 0.297 9
Tobacco Products 105.99 23.34 0.220 3-4
Weaving Mills. Cotton 21.92 13.49 0.615 121
Weaving Mills. Manmade 25.05 14.08 0.562 90—93
Weaving. Finished. Wool 25.46 14.10 0.554 81-82
Narrow Fabric Mills 20.58 12.50 0.607 116
Knitting Mills 20.42 11.67 0.571 94-95
Textile Finishing 27.50 14.41 0.524 61
Floor Covering Mills 31.10 14.38 0.463 38
Yarn and Thread Mills 20.03 12.22 0.610 118
Miscellaneous Textiles 28.52 15.26 0.535 68
Men’s and Boys' Suits 19.32 12.09 0.626 125
Men’s and Boys' Furnishings 17.78 9.44 0.531 65
Women's and Misses' Out. 16.47 9.25 0.562 90-93
Women’s.Children’s Under. 17.22 9.35 0.543 71-73
Hats. Caps. and Millinery 16.56 9.28 0.560 88-89
Children's Outerwear 15.73 8.71 0.554 81-82
Fur Goods 35.48 17.43 0.491 45
Misc. Apparel and Accs. 17.64 10.14 0.575 98

Misc. Fabricated Textile 22.71 13.04 0.574 97

_—
'-

Sawmills and Planning Mills
Millwork.
Wood Containers

Wood Buildings.Mobile Homes
Wood Products

Misc.

Household Furniture

Office Furniture

Public Bldg.. Related Furn.
Partitions and Fixtures
Furniture and Fixt.

Misc.

Pulpmills
Papermills.
Paperboard Mills

Converted Paper
Paperboard Cont. Boxes
Building Paper.Board Mills

Misc.

Newspapers
Periodicals

Books
Misc.

Industrial

Drugs

Soaps.
Paints.

Misc.

Petroleum Refining
Paving and Roofing Mat.
Petroleum.

Misc.

Tires and Inner Tubes
Plastic Footwear
Reclaimed Rubber

Plastic Hose
Fabricated Rubber
Plastics

Rubber.

Rubber.

Misc.

p1WO°d9

Building Paper

Publishing
Commercial Printing
Manifold Business Forms
Greeting Card Publishing
Blankbooks and Bookbinding
Printing Trade Services

Inorganic Chem.
Plastic Materials.Syn.

Cleaners.
Allied Products
Industrial Organic Chem.
Agricultural Chem.
Chemicals

Coal Prod.

142

25.07
24.87
17.72
26.72
25.53

21.13
36.46
25.58
28.55
29.47

79.80
59.24
59.76
44.81
32.22
33.36

30.94
71.98
48.80
37.82
28.70
44.38
48.09
26.20
30.87

76.62
64.34
78.16

104.40

57.13

101.20
102.68

61.54

215.65

52.15
69.54

46.74
17.09
41.00
28.29
28.56
30.92

0.621
0.643
0.602
0.594
0.557

0.583
0.506
0.585
0.598
0.541

0.383
0.467
0.481
0.418
0.597
0.605

0.527
0.295
0.370
0.420
0.617
0.427
0.362
0.545
0.624

0.352
0.403
0.306
0.201
0.366
0.290
0.220
0.348

0.158
0.433
0.319

0.558
0.679
0.459
0.872
0.634
0.514

123
129
112
108

84

101-102
52

104

109-111
69

22

39

41

25-27
108
114-115

62-63
8

19-20
29
123
31
17

74-76
124

85-86
133

35
139
127

56-58

 

143

 

Leather Tanning

Boot. Shoe Cut Stock
Footwear. Expt. Rubber
Leather Gloves

Luggage

Handbags and Purses
Leather Goods

Flat Glass

Glass

Products of Glass

Cement. Hydraulic
Structural Clay

Pottery

Concrete. Gypsum. Plaster
Cut Stone

Misc. Nonmetallic Miner.

Blast Furnance. Steel
Iron. Steel Foundries
Primary Nonferrous Metals
Secondary Nonferrous
Nonferrous Rolling
Nonferrous Foundries
Misc. Primary Metal

Metal Cans

Cutlery. Handtools
Plumbing. Heating
Fabricated Struc.
Screw Machine. Bolts
Metal Forgings. Stamp
Metal Services
Ordnance

Misc. Fab. Metal

Engines. Turbines

Farm. Garden Machinery
Construction Mach.
Metalworking Mach.
Special Industry Mach.
General Industry Mach.
Office. Computing Mach.
Refrigeration

Misc. Machinery

29.47
19.63
20.35
10.23
25.46
16.98
20.58

48.69
39.73
29.77
69.72
28.88
27.88
38.03
22.28
44.45

45.61
32.37
75.01
48.03
47.00
29.84
41.53

67.33
36.06
34.82
33.58
33.35
35.91
27.45
34.15
36.59

52.33
45.43
47.29
39.03
35.38
39.08
47.60
39.24
29.42

16.17
10.74
10.42

7.48
12.82

9.99
10.48

29.83
22.34
17.44
29.05
16.65
16.80
18.82
13.56
20.47

32.52
23.15
31.15
21.86
24.52
19.51
20.70

28.18
19.64
17.49
19.18
19.30
23.49
15.35
22.43
19.49

27.96
23.34
24.36
22.34
21.17
21.88
22.92
20.75
18.48

0.549
0.547
0.512
0.731
0.503
0.589
0.509

0.613
0.562
0.586
0.417
0.583
0.603
0.495
0.609
0.461

0.713
0.715
0.423
0.455
0.522
0.654
0.498

0.418
0.545
0.502
0.571
0.579
0.654
0.559
0.657
0.533

0.534
0.514
0.515
0.572
0.598
0.560
0.482
0.529
0.628

79-81

77

55
138

51
106

54

119

90-93
105

24
101-102
113

47
117

36

135
136
30
34
60
130-131
48

25-27
74-76
50
94-95
100
130-131
87
132
66

67
56-58
59
96
109-111
88-89
42
64
126

144

 

Electric Distributing

 

19.45

 

37.84 0.514 56-58
Elect. Industrial 36.12 20.05 0.555 83
Household Appliances 38.56 17.81 0.462 37
Elec. Lighting. Wiring 35.40 17.49 0.494 46
Radio. TV Receiving 43.11 18.08 0.419 28
Communication Equip. 39.89 23.29 0.584 103
Electronic Components 33.90 18.61 0.549 78-80
Misc. Elec. Equip. 37.08 20.35 0.549 78-80
Motor Vehicles 42.73 30.27 0.708 134
Aircraft. Parts 47.58 28.45 0.598 109-111
Ship. Boat Building 26.67 21.79 0.760 138
Railroad Equipment 49.34 26.90 0.545 74-76
Motorcycles. Bicycles 31.26 16.99 0.543 71-73
Guided Missiles. Space 52.06 33.16 0.637 128
Transportation Equip. 30.94 18.72 0.605 114-115
Engineering. Scien. Inst. 40.74 20.70 0.508 53
Measuring. Controlling 35.06 20.18 0.576 99
Optical Instruments 42.29 21.17 0.501 49
Medical Instruments 36.14 17.46 0.483 43
Ophthalmic Goods 28.54 15.49 0.543 71-73
Photographic Equip. 86.96 27.32 0.314 12
Watches. Clocks 28.85 15.20 0.527 62-63
Jewelry. Silver 31.58 15.05 0.476 40
Musical Instruments 24.00 14.73 0.614 120
Toys. Sporting Goods 33.88 14.16 0.418 25-27
Pens. Office. Art Goods 35.07 15.57 0.444 33
Costume Jewelry 23.69 12.83 0.542 70
Misc. Manufactures 27.34 15.25 0.558 85-86
All Manufactures 37.48 20.37 0.543
(1) Units of average value added per employee are in

thousands of dollars.

(2) Units of average compensation per employee are in

thousands of dollars.

(3) Unit labor costs are defined as
compensation/value-added.

145

TABLE III-B

U.S. 1967. Two-. Three-. and Four-Digit

SIC Comparative Cost Rankings

 

Industry Value/ Comp/ Unit Rank
Employ Employ Value
(I) (2) (3)

Meat Products 11.45 7.01 0.612 104-105
Dairy Products 14.96 6.96 0.465 28-29
Preserved Fruits 13.81 5.33 0.386 18
Grain Mill Products 25.78 7.59 0.294 5
Bakery Products 13.23 7.04 0.532 52
Sugar. Confectionery 21.10 7.67 0.363 13
Fats and Oils 15.02 5.84 0.389 19—20
Beverages 21.70 7.73 0.356 12
Misc. Foods Kindred 21.52 6.91 0.321 8
Tobacco Products 27.06 5.85 0.216 2
Weaving Mills. Cotton 8.01 5.04 0.630 115
Weaving Mills. Manmade 8.49 5.32 0.627 114
Weaving. Finished. Wool 10.25 5.67 0.553 66
Narrow Fabric Mills 8.18 5.08 0.621 107-108
Knitting Mills 7.96 4.72 0.593 85
Textile Finishing 9.63 6.21 0.645 124
Floor Covering Mills 13.75 5.77 0.420 24
Yarn and Thread Mills 7.94 4.65 0.586 83
Miscellaneous Textiles 10.98 6.15 0.560 68-69
Men's and Boys' Suits 7.71 5.25 0.681 134
Men's and Boys’ Furnishings 6.33 3.86 0.609 101-102
Women's and Misses' Out. 7.78 4.66 0.600 89
Women's. Children's Under. 7.64 4.29 0.561 70
Hats. Caps.and Millinery 6.93 4.68 0.676 132
Children’s Outerwear 6.87 4.17 0.607 97-99
Fur Goods 13.79 7.52 0.545 60-62
Misc. Apparel and Accs. 7.21 4.42 0.612 104-105
Misc. Fabricated Textile 8.43 5.01 0.594 86
Logging Camps 9.85 5.27 0.535 56
Sawmills and Planning Mills 8.28 5.27 0.637 119-120
Millwork. Plywood. Struc. 9.78 6.43 0.658 127
Wood Containers 7.62 4.74 0.622 109
Misc. Wood Products 9.08 5.33 84

 

0.587

 

146

 

 

Household Furniture 8.90 5.37 0.603 90
Office Furniture 13.58 7.16 0.528 50-51
Public Bldg.. Related Furn. 10.34 6.52 0.631 116-118
Partitions and Fixtures 12.23 7.42 0.607 97-99
Misc. Furniture and Fixt. 10.24 6.20 0.605 91-94
Pulpmills 22.10 9.24 0.418 23
Papermills. Building Paper 16.83 8.98 0.534 54-55
Paperboard Mills 22.52 8.93 0.396 21
Misc. Converted Paper 15.18 7.06 0.465 28-29
Paperboard Cont. Boxes 11.62 7.04 0.606 95-96
Building Paper. Board Mill 15.70 7.95 0.506 40
Newspapers 12.46 7.25 0.582 78-79
Periodicals 23.62 8.79 0.372 14
Books 20.39 7.84 0.385 17
Misc. Publishing 13.42 6.91 0.515 47-48
Commercial Printing 11.93 7.71 0.647 125
Manifold Business Forms 16.00 7.61 0.476 30
Greeting Card Publishing 62.05 28.18 0.454 26
Blankbooks and Bookbinding 9.45 6.05 0.640 121-122
Printing Trade Services 12.88 9.30 0.722 138
Industrial Inorganic Chem 31.16 9.53 0.306 6
Plastic Materials. Syn. 27.57 10.72 0.389 19-20
Drugs 34.55 8.99 0.260 4
Soaps. Cleaners. Toilet 40.56 7.92 0.195 1
Paints. Allied Products 19.95 8.29 0.416 22
Industrial Organic Chem. 17.08 6.39 0.374 15
Agricultural Chem. 22.01 6.85 0.311 7
Misc. Chemicals 17.97 7.93 0.441 25
Petroleum Refining 44.47 10.4 0.234 3
Paving and Roofing Mat. 17.20 7.88 0.458 27
Misc. Petroleum. Coal Prod. 26.80 8.64 0.323 9
Tires and Inner Tubes 19.67 9.75 0.496 38
Rubber. Plastic Footwear 8.42 5.61 0.667 129-131
Reclaimed Rubber 13.76 8.41 0.611 103
Fabricated Rubber 12.29 7.46 0.607 97-99
Misc. Plastics 11.80 6.35 0.538 57
Leather Tanning 10.40 6.78 0.652 126
Industrial Leather Belting 11.67 6.44 0.552 65
Boot. Shoe Cut Stock 7.97 4.75 0.596 88
Footwear. Expt. Rubber 7.67 4.64 0.605 91-94
Leather Gloves 5.76 3.92 0.680 133
Luggage 8.69 4.93 0.568 72-73
Handbags and Purses 7.57 4.87 0.643 123
Leather Goods 7.46 4.70 0.631

116-118

Flat Glass
Glass
Products of Glass

Cement. Hydraulic
Structural Clay

Pottery

Concrete. Gypsum. Plaster
Cut Stone

Misc. Nonmetallic Miner.

Blast Furnance. Steel
Iron. Steel Foundries
Primary Nonferrous Metals
Secondary Nonferrous
Nonferrous Rolling
Nonferrous Foundries
Misc. Primary Metal

Metal Cans

Cutlery. Handtools
Plumbing. Heating
Fabricated Struc.
Screw Machine. Bolts
Metal Forgings. Stamp
Metal Services

Misc. Wire Products
Misc. Fab. Metal

Engines. Turbines

Farm. Garden Machinery
Construction Mach.
Metalworking Mach.
Special Industry Mach.
General Industry Mach.
Office. Computing Mach.
Refrigeration

Misc. Machinery

Electric Distributing
Elect. Industrial
Household Appliances
Elec. Lighting. Wiring
Radio. TV Receiving
Communication Equip.
Electronic Components
Misc. Elec. Equip.

 

 

147

9.54
7.03
6.95
8.79
6.35
6.48
7.18
5.96
7.69

9.56
8.12
8.71
8.05
8.43
7.63
9.13

9.15
7.55
7.34
7.74
8.13
8.55
6.47
6.72
7.75

9.32
8.00
8.49
9.38
8.45
8.60
8.49
7.61
7.81

7.82
7.70
7.34
6.79
6.27
9.30
6.74
7.60

0.539
0.511
0.515
0.353
0.626
0.661
0.490
0.667
0.528

0.580
0.718
0.337
0.510
0.494
0.640
0.613

0.483
0.508
0.549
0.609
0.582
0.637
0.608
0.605
0.564

0.583
0.534
0.560
0.621
0.606
0.585
0.485
0.500
0.625

0.547
0.570
0.489
0.482
0.514
0.699
0.623
0.545

58
43
47-48
11

113
128

36

129-131

50-51

77

137

10
42
37

121-122
106

32
41
64

101-102

78-79

119-120
100

91-94
71

80-81
54-55
68-69

107-108

95-96
82
33
39

112

63
74
34-35
31
46

135
110-111

60-62

 

148

18.48

Motor Vehicles 9.47 0.512 44
Aircraft. Parts 14.12 10.10 0.715 136
Ship. Boat Building 10.07 8.12 0.806 139
Railroad Equipment 13.92 8.67 0.623 110-111
Motorcycles. Bicycles 11.35 6.51 0.573 75
Misc. Transport Equip. 10.06 6.09 0.605 91-94
Engineering. Scien. Inst. 13.39 7.80 0.583 80-81
Measuring. Controlling 13.34 7.73 0.579 76
Optical Instruments 13.64 8.61 0.631 116-117
Medical Instruments 14.33 7.01 0.489 34-35
Ophthalmic Goods 11.19 6.20 0.554 67
Photographic Equip. 26.14 9.80 0.375 16
Watches. Clocks 11.16 6.64 0.595 87
Jewelry. Silver 12.25 6.68 0.545 60-62
Musical Instruments 9.55 6.37 0.667 129-131
Toys. Sporting Goods 10.40 5.34 0.513 45
Pens. Office. Art Goods 11.88 6.33 0.533 53
Costume Jewelry 10.23 5.39 0.526 49
Misc. Manufactures 10.95 6.23 0.568 72-73

 

 

(1) Units of average value added per employee are in

thousands of dollars.

(2) Units of average compensation per employee are in

thousands of dollars.

(3) Unit labor costs are defined as
compensation/value-added.

149

TABLE IV-B

U.S. 1980 ISIC Comparative Cost Rankings

 

UNIT RANK

 

INDUSTRY VALUE/ COMP/
EMPLOY EMPLOY LABOR
COSTS
(I) (2) (3)
Food Products 47.14 17.76 0.377 5
Beverages 62.89 22.15 0.352 4
Tobacco 105.99 23.33 0.220 2
Textiles 23.45 12.97 0.553 19
Wearing Apparel 17.23 9.57 0.555 20
Leather and Products 21.82 12.01 0.550 18
Footwear 20.30 10.45 0.515 15
Wood Products 25.83 15.21 0.589 24
Furniture. Fixtures 24.61 13.99 0.569 22
Paper and Products 46.05 21.99 0.478 9
Printing. Publishing 35.14 17.36 0.494 11
Industrial Chemicals 58.27 26.88 0.461 8
Other Chemical Products 79.61 22.19 0.279 3
Petroleum Refineries 215.65 34.00 0.158 I
Petroleum. Coal Products 57.76 22.47 0.389 6
Rubber Products 34.47 21.51 0.624 26
Plastic Products n.e.c. 30.92 15.88 0.514 14
Pottery. China. etc. 27.88 16.80 0.603 25
Glass and Products 38.17 21.89 0.573 23
Nonmetal Products n.e.c. 40.89 19.70 0.482 10
Iron and Steel 41.51 29.62 0.714 28
Nonferrous Metals 47.00 24.01 0.511 13
Metal Products 35.82 20.25 0.565 21
Machinery n.e.c. 41.25 22.32 0.541 17
Electrical Machinery 37.26 20.11 0.540 16
Transport Equipment 43.22 28.37 0.656 27
Professional Goods 45.28 20.63 0.456 7
Other Industries 29.62 14.67 0.495 12
All Manufactures 20.37 0.543

(1) Units of average

(2) Units

37.48

 

value added per
thousands of dollars.

of average compensation per
thousands of dollars.

(3) Unit labor costs are defined as
compensation/value-added.

employee are

employee are

in

in

Japanese 1980 Comparative Cost Rankings

 

INDUSTRY

150

TABLE V-B

RANK

 

(1) Units of average value added per employee are

million yen.

(2) Units of average wage per employee are in

million yen.

(3) Unit labor costs are defined as
compensation/value-added.

VALUE/ WAGES/ UNIT

EMPLOY EMPLOY LABOR

COSTS

(1) (2) (3)

Food Products 5.76 1.94 0.336 10
Beverages 11.30 2.53 0.224 4
Tobacco 0 4.63 0 28
Textiles 4.16 1.83 0.439 21
Wearing Apparel 2.33 1.23 0.526 27
Leather and Products 4.19 1.91 0.456 24
Footwear 4.38 2.03 0.463 25
Wood Products 4.33 1.93 0.444 22
Furniture. Fixtures 4.89 2.05 0.418 20
Paper and Products 7.42 2.67 0.359 11
Printing. Publishing 7.70 3.11 0.403 17
Industrial Chemicals 14.72 3.75 0.254 5
Other Chemical Products 17.26 3.33 0.193 2
Petroleum Refineries 51.76 4.34 0.084 1
Petroleum. Coal Products 15.67 3.20 0.204 3
Rubber Products 7.09 2.77 0.391 15
Plastic Products n.e.c. 6.23 2.30 0.370 12
Pottery. China. etc. 4.27 1.99 0.466 26
Glass and Products 10.45 3.11 0.298 8
Nonmetal Products n.e.c. 7.66 2.45 0.319 9
-Iron and Steel 13.88 3.76 0.271 6
Nonferrous Metals 11.58 3.15 0.272 7
Metal Products 6.12 2.49 0.407 18
Machinery n.e.c. 7.67 3.06 0.399 16
Electrical Machinery 6.58 2.45 0.372 13
Transport Equipment 8.27 3.14 0.379 14
Professional Goods 5.36 2.38 0.445 23
Other Industries 5.16 2.13 0.414 19

All Manufactures 7.20 2.55 0.355

151

TABLE VI-B

German 1980 Comparative Cost Rankings

 

 

 

INDUSTRY VALUE/ COMP/ UNIT RANK
EMPLOY EMPLOY LABOR
COSTS
(1) (2) (3)
Food Products 97.6 28.3 0.290 3
Beverages 107.8 34.2 0.317 4
Tobacco 551.3 33.0 0.060 I
Textiles 39.8 25.5 0.641 23
Wearing Apparel 40.1 20.0 0.498 10
Leather. Footwear. etc. 42.7 22.8 0.534 13
Wood. Furniture. etc. 63.5 29.9 0.471 8
Paper and Products 54.1 31.4 0.580 18
Printing. Publishing 60.2 35.2 0.585 19
Industrial Chemicals 75.5 40.6 0.538 14
Petroleum Refineries 825.9 50.0 0.061 2
Rubber Products 55.2 32.7 0.592 20
Plastic Products n.e.c. 53.3 29.3 0.549 15
Pottery. China. etc. 56.7 25.0 0.441 6
Glass and Products 59.2 31.4 0.531 12
Nonmetal Products n.e.c. 84.7 34.3 0.405 5
Iron and Steel 67.3 34.9 0.519 11
Nonferrous Metals 42.5 35.0 0.824 24
Metal Products 52.2 32.7 0.626 22
Machinery n.e.c. 57.8 36.1 0.625 21
Electrical Machinery 60.1 33.7 0.560 16
Transport Equipment 66.3 38.3 0.578 17
Professional Goods 67.3 31.0 0.461 7
Other Industries 53.3 25.7 0.481 9
All Manufactures 67.8 33.6 0.495

(1)
thousand marks.

(2)
thousand marks.

(3)

 

Units of average value added per employee are in

Units of average compensation per employee are in

Unit labor costs are defined as compensation/value-added

152

TABLE VII-B

U.K. 1980 Comparative Cost Rankings

 

INDUSTRY

VALUE/

 

COMP/ UNIT RANK
EMPLOY EMPLOY LABOR
COSTS
(I) (2) (3)

Food Products 11.46 5.37 0.469 8
Beverages 21.18 6.42 0.303 2
Tobacco 21.67 6.94 0.320 3
Textiles 6.68 4.70 0.704 25
Wearing Apparel 5.35 3.90 0.729 26
Leather and Products 7.50 4.78 0.637 21
Footwear 7.12 4.80 0.674 24
Wood Products 9.62 5.59 0.581 13
Furniture. Fixtures 9.32 5.97 0.641 22
Paper and Products 10.50 6.24 0.594 18
Printing. Publishing 13.40 6.93 0.517 9
Industrial Chemicals 17.12 7.89 0.461 7
Other Chemical Products 17.55 6.15 0.350 4
Petroleum Refineries 114.12 9.35 0.082 I
Petroleum. Coal Products 19.38 7.00 0.361 5
Rubber Products . 10.52 6.21 0.590 16
Plastic Products n.e.c. 10.46 5.76 0.551 11
Pottery. China. etc. 7.64 5.05 0.661 23
Glass and Products 11.48 6.80 0.592 17
Nonmetal Products n.e.c. 16.44 6.45 0.392 6
Iron and Steel 7.75 6.53 0.843 28
Nonferrous Metals 11.33 6.62 0.584 14
Metal Products 9.75 6.00 0.615 20
Machinery n.e.c. 11.09 6.29 0.567 12
Electrical Machinery 9.83 5.80 0.590 15
Transport Equipment 9.05 6.61 0.730 27
Professional Goods 9.50 5.65 0.595 19
Other Industries 9.17 4.93 0.538 10

All Manufactures 10.89 0.553

(1) Units of average
thousand pounds.

(2) Units of average
thousand pounds.

(3) Unit labor costs

compensation/value-added.

are defined as

value added per employee are in

compensation per employee are in

153

TABLE VIII-B

Italian 1980 Comparative Cost Rankings

VALUE/ COMP/ UNIT RANK

 

 

INDUSTRY

EMPLOY EMPLOY LABOR

(I) (2) (3)

Food Products 30.96 13.83 0.447 3
Beverages 43.39 15.64 0.360 2
Tobacco 13.15 9.85 0.749 23
Textiles 19.43 10.86 0.559 15
Wearing Apparel 15.92 9.27 0.582 17
Leather and Products 21.96 10.57 0.481 6
Footwear 15.42 9.04 0.586 18
Wood Products 22.14 10.69 0.483 7
Furniture. Fixtures 22.11 10.89 0.493 8
Paper and Products 28.90 14.43 0.499 10
Printing. Publishing 29.70 16.94 0.570 16
Industrial Chemicals 34.20 16.25 0.475 5
Petroleum Refineries 601.11 19.42 0.032 1
Rubber Products 24.52 15.14 0.618 21
Plastic Products n.e.c. 16.96 12.82 0.756 24
Pottery. China. etc. 26.86 12.65 0.471 4
Iron and Steel 29.81 14.85 0.498 9
Nonferrous Metals 22.52 13.86 0.615 20
Metal Products 23.53 12.77 0.543 12-13
Machinery n.e.c. 26.19 14.21 0.543 12-13
Electrical Machinery 21.83 13.41 0.614 19
Transport Equipment 21.06 13.44 0.638 22
Professional Goods 27.19 14.72 0.541 11
Other Industries 20.16 11.03 0.547 14
All Manufactures 24.65 13.25 0.537

(1) Units of average
hundred thousand

(2) Units of average
hundred thousand

(3) Unit labor costs

value added per employee are in

lira.

compensation per employee are in

lira.

are defined as
compensation/value-added.

154

TABLE IX-B

U.S. 1967 Comparative Cost Rankings

 

INDUSTRY VALUE/ COMP/ UNIT RANK
EMPLOY EMPLOY LABOR
COSTS
(I) (2) (3)
Food Products 15.58 5.99 0.384 7
Beverages 19.95 6.87 0.344 4
Tobacco 27.07 5.07 0.187 1
Textiles 8.79 4.71 0.536 21
Wearing Apparel 7.25 4.05 0.558 25
Leather and Products 8.78 5.00 0.570 26
Footwear 7.69 4.22 0.549 23
Wood Products 8.88 5.10 0.574 27
Furniture. Fixtures 9.27 5.15 0.555 24
Paper and Products 15.27 6.95 0.455 11
Printing. Publishing 13.93 6.94 0.498 15
Industrial Chemicals 26.78 7.84 0.293 3
Other Chemical Products 22.42 8.19 0.365 5
Petroleum Refineries 44.30 8.97 0.203 2
Petroleum. Coal Products 19.43 7.14 0.368 6
Rubber Products 14.45 6.98 0.483 13
Plastic Products n.e.c. 11.79 5.71 0.485 14
Pottery. China. etc. 9.77 5.81 0.595 28
Glass and Products 14.32 6.60 0.461 12
Nonmetal Products n.e.c. 14.55 6.52 0.448 9
Iron and Steel 15.04 7.85 0.522 19
Nonferrous Metals 16.62 7.37 0.444 8
Metal Products 13.38 6.90 0.516 18
Machinery n.e.c. 14.80 7.57 0.511 17
Electrical Machinery 13.22 6.96_ 0.526 20
Transport Equipment 15.35 8.26 0.538 22
Professional Goods 15.83 7.15 0.452 10
Other Industries 10.76 5.39 0.501 16
All Manufactures g14.18 6.68 0.471

 

 

Units of
thousands

(1) average value added per employee are in

of dollars.
(2)

Units of average compensation per employee are in

thousands of dollars.

(3) Unit labor costs are defined as
compensation/value-added.

155

TABLE X-B

Japanese 1967 Comparative Cost Rankings

 

WAGES/ UNIT

 

INDUSTRY VALUE/ RANK
EMPLOY EMPLOY LABOR
COSTS
(I) (2) (3)
Food Products 1.29 0.39 0.304 10
Beverages 2.21 0.44 0.198 4
Tobacco 0 0.91 0 28
Textiles 0.99 0.37 0.371 18
Wearing Apparel 0.71 0.30 0.418 25
Leather and Products 1.24 0.47 0.383 21
Footwear 1.17 0.46 0.393 23
Wood Products 0.99 0.39 0.391 22
Furniture. Fixtures 0.97 0.42 0.427 26
Paper and Products 1.62 0.52 0.320 13
Printing. Publishing 1.73 0.65 0.375 19
Industrial Chemicals 3.77 0.71 0.187 2
Other Chemical Products 3.18 0.60 0.188 3
Petroleum Refineries 7.86 0.86 0.109 1
Petroleum. Coal Products 2.00 0.53 0.267 6
Rubber Products 1.50 0.48 0.322 14
Plastic Products n.e.c. 1.33 0.45 0.338 15
Pottery. China. etc. 0.88 0.40 0.453 27
Glass and Products 2.28 0.58 0.254 5
Nonmetal Products n.e.c. 1.57 0.47 0.301 9
Iron and Steel 2.55 0.76 0.299 8
Nonferrous Metals 2.40 0.67 0.282 7
Metal Products 1.36 0.52 0.379 20
Machinery n.e.c. 1.63 0.59 0.359 17
Electrical Machinery 1.58 0.49 0.312 11
Transport Equipment 1.96 0.62 0.316 12
Professional Goods 1.26 0.52 0.417 24
Other Industries 1.17 0.42 0.355 16
All Manufactures 1.59 0.319

0.51

 

(1)
million yen.

(2)
million yen.

(3)

Units of average value added

per employee are in

Units of average wage per employee are in

Unit labor costs are defined as wages/value-added.

156

TABLE XI-B

German 1967 Comparative Cost Rankings

 

INDUSTRY

 

 

OUTPUT/ WAGES/ UNIT

EMPLOY EMPLOY LABOR

COSTS

(1) (2) (3)

Food Products 113.6 10.8 0.095
Beverages 80.9 12.1 0.149
Tobacco 264.7 9.7 0.037
Textiles 39.8 9.2 0.232
Wearing Apparel 30.7 7.4 0.240
Leather and Footwear 34.0 8.1 0.239
Wood Products.Furniture 42.8 10.4 0.242
Paper and Products 49.7 11.0 0.222
Printing. Publishing 32.5 12.0 0.370
Industrial Chemicals 71.2 14.3 0.201
Petroleum Refineries.Coal 430.3 16.7 0.039
Rubber Products 41.3 13.0 0.316
Plastic Products n.e.c. 41.6 11.8 0.283
Pottery. China. etc. 25.3 9.9 0.390
Glass and Products 37.5 11.5 0.306
Nonmetal Products n.e.c. 50.0 12.2 0.244
Iron and Steel 50.6 12.4 0.246
Nonferrous Metals 67.3 12.4 0.184
Metal Products 39.8 11.5 0.288
Machinery n.e.c. 38.3 12.3 0.321
Electrical Machinery 36.0 11.5 0.319
Transport Equipment 48.9 13.6 0.277
Professional Goods 27.2 10.3 0.381
Other Industries 31.7 8.9 0.280
All Manufactures 49.7 11.6 0.233

(1) Value added data were not available for Germany in 1967.
Instead. gross output was substituted in an attempt to
obtain a ranking system. Units of average gross output
per employee are in thousand marks.

(2) Units of average wage per employee are in thousand marks.

(3) Unit labor costs are defined as
compensation/gross-output.

157

Table XII-B

Comparative Rankings of U.S..
and Italian Manufacturing Industries in 1980

Food Products
Beverages

Tobacco

Textiles

Wearing Apparel

Leather and Products
Footwear

Wood Products
Furniture. Fixtures
Paper and Products
Printing. Publishing
Industrial Chemicals
Other Chemical Products
Petroleum Refineries
Petroleum. Coal Product
Rubber Products
Plastic Products n.e.c.
Pottery. China. etc.
Glass and Products
Nonmetal Products

Iron and Steel
Nonferrous Metals

Metal Products
Machinery n.e.c.
Electrical Machinery
Transport Equipment
Professional Goods
Other Industries

Japanese.

German. U.K..

 

 

JAPAN GERMANY U.K. ITALY
3 8 3
4 2 2
I 3 23

23 25 15
10 26 17
13 21 6
- 24 18
8 13 7
- 22 B
18 18 10
19 9 16
14 7 5
- 4 _
2 1 1
_ 5 _
20 16 21
15 11 24
6 23 4
12 17 -
5 6 -
II 28 9
24 14 20
22 20 12-13
21 12 12-13
16 15 19
17 27 22
7 19 11
9 10 14

APPENDIX C

An Extension and Empirical Test of
Comparative Advantage

This section extends and tests empirically Deardorff“s
(1980) formulation of comparative advantage. Heideveloped a
general model of comparative advantage whose conclusion is
that there must exist a negative correlation between a
country"s autarky prices and its pattern of net exports.
This is demonstrated in his theorem below.
if Pa’Tni o, (l)
where Pa is a vector of autarky prices and T is net exports
for n industries. then
coritxl.x2) _<_ 0 as xliixz $0.

The assumptions of the model are (I) the production
possibilities set. given its technolOQY. is closed. convex,
and bounded from above: (2) a country”s preferences can be
represented by a family of community indifference curves
exhibiting the property of local nonsatiation: (3) producers
and consumers behave competitively to maximize the value of
net output and utility of consumption: and (A) no trade
subsidies exist.

The significance of this model lies in the assumptions
that were not made. namely that: (1) utility and production

functions are not assumed to be differentiable: (2) utility

158

159

functions need not be homothetic: and (3) neither functions
are assumed to be identical across countries. Thus.
countries can differ in tastes. technologies. and factor
endowments.

Deardorff extends the model by examining the ratio of
autarky to world prices.

Pi = Pal/PWJ j=1.u.n: (2)
where P8 and Pw are autarky and world prices respectively
and j is the industries. Next consider a country’s net
exports valued at world prices.
ej = iji'Tnj. (3)
Then assuming balanced trade.

Zej = pwj'Tnj = e. (A)

By examining the product f)*e. we obtain
f'e = illDEIJ/ia‘"J - ”"3”3n = zpaji‘rnj = Pa'Tn.

Thus. from the theorem in equation 1.
cor(f’.e) g 0. (5)
This is Deardorff"s first corrollary.

We can now extend the model by examining a competitive
market where PhJ = MCJ. PhJ is the domestic price of good j.
If we define a production function employing two factors.
capital-K. and labor-L.

05 = f(KJ.LJ). (6)

Then by allowing the production function to be
differentiable.

MC.j s wJ/MP‘J = rJ/MPkJ. (7)

Multiplying through to obtain factor prices and factor

160

costs.
MPIJ.MCj = NJ and MPIJ.LJ.MCj = WJTLJ‘ (8)
Npkj-ch = r3 and Mij'Kj'MCJ = rj*KJ. . (9)

By adding the two equations to derive total factor

costs.

MCj*(MP.j*Lj + Mij*Kj) = wj*Lj + rj*Kj. (IO)
and

MCj = (wj*Lj + rJ*KJ)/(MP‘1*LJ + MPkJ*KJ). (ll)

Multiplying through by QJ/QJ.

M.. = i’ .* P * ’ .
CJ _(wJ LJ + rJ Kfi/Q.) [Qj/(MP.J*LJ + 14ka an. (12)
This simplifies into equation 13. where MC = ZMCJ.
MC = AC/(function coefficient) = Ph. (13)

and Ph is the vector of home prices.

If the function coefficient (fc) is greater than one,
then increasing returns to scale hold. Similarly if fc =1.
constant returns to scale hold. and if fc < I. then
decreasing returns.

If. as an approximation. we assume that Ph = Pa. then
substituting equation 13 into equation 5 yields.
cert(Ach/Ac")*(fc"’/fc").e) g 0. (14)

This formulation provides a method of testing trade
theories of scale economies within a comparative cost
framework. Due to the unavailability of world functional
coefficient data. the empirical testing of this hypothesis
required the assumption that fcw be equal to fch. This
assumption is not unreasonable since the H-0 hypothesis
assumes both identical production functions across countries

and constant returns to scale. while the Ricardian model

161

assumes the latter. Thus. compared to the standard trade
theories. the assumption that fcw = fch is more flexible.
Although economies of scale must be equal in the same
industry worldwide. they do not have to be equal to one and
can vary among industries.

The average costs of an industry is approximated by
relative unit labor costs. These are defined as an
industry"s unit labor costs relative to average in all other
industries. In turn. unit labor costs consist of labor
productivity per employee and labor compensation per
employee.

Since this dissertation examines five countries. the
United States. Japan. Germany. the United Kingdom. and
Italy. a weighted average of these will approximate world
costs. The weights are value added as a percentage of world
value added, thus reflecting the relativelimportance of a

country"s unit labor cost in an industry. The weights are as

follows:
(vji/ §\Vji)/ é‘ivji/ é‘vi). (15)

where VJ. is the value added of industry j in
country i for m countries and n industries.
The sign of equation 14 can be derived from the sign of
the covariance between the vectors where.
cor(Uh/Uw.e) = cov(Uh/U",e) (16)
[ ..Z'EBFZBQTI"JQ...1 i» - . 5

Since the denominator is nonnegative. the correlation

and covariance must have the same sign. Although only the

162

,covariance matrix is needed for testing the Deardorff model.
both will be presented in Table C1 with net exports
specified both as exports minus imports and as a ratio.

The results of the covariance matrix yielded the
expected sign in eight of the ten cases. For the two cases
where the wrong sign occured. the correlation matrix
indicated a weaker magitude compared with the alternate
specification of net exports. Considering the multitude of
factors determining trade flows. these results strongly
support the extension of the Deardorff model.

The conclusion is that comparative advantage. based on
relative unit labor costs. is an important determinate of
net exports. This extension of Deardorff’s model uses less
restrictive assumptions than the standard trade models.
provides a comparative coSt framework for testing scale
economies. and fits the data well. Thus. it should be
considered in further studies of the determinants of

international trade flows.

163

Table C1

Covariance and Correlation Matrices Between
Net Exports and Relative Unit Labor Costs

 

 

 

 

 

for I980.

Covariance Matrix

Country X-M X/M
Germany 276978.8 -0.04185
Italy -31.8994 -0.00773
U.S. -162.164 0.02656
Japan -315279 -0.22299
U.K. -66.891 -0.01674
Correlation Matrix

Country X-M X/M
Germany 0.16276 —0.21068
Italy -0.14355 -0.12525
U.S. —0.27454 0.15516
Japan -0.23181 -0.27515
U.K -0.26251 —0.18855

 

 

BIBLIOGRAPHY

Allen. William. International Trade Theory: Hume tg
Ohlin. New York: Random House. 1965.

Anderson. Richard. and Kreinen. Mordechai. "Labor Cost and
U.S. Comparative Advantage in Steel and Motor
Vechicles". The World Economy. June 1981.

Angell. James. The Theory gfi International Prices.
Boston: Harvard University Press. 1926.

"Annual Report on American Industry." Forbes. January 2.
1984. pp. 99-243.

Balassa. Bela. "An Empirical Demonstration of Classical
Comparative Cost Theory". The Review 9f Economics and
Statistics. August 1963. pp. 231-38.

. 'Revealed' Comparative Advantage Revisted: An
Analysis of Relative Export Shares of the Industrial
Countries. 1953-1971”.The figgchester School. 1977.

 

 

 

pp.327-44.

. ”Trade Liberalization and 'Revealed'
Comparative Advantage" The flgnchester School. May 1965.
pp. 99-123.

Bhagwati. Jadish. ”Pure Theory of International Trade:
A Survey". In: Jadish Bhagwati (ed.). Trade. Tariffs.
gag growthl Essays lg Interngtiongl Economics.
London: Weidenfeld 5 Nicolson. 1969. pp. 3-94.

Boretsky. M.. ”The Threat to U.S. High Technology
Industries:Economic and National Security
Implications." Draft. International Trade
Administration. U.S. Department of Commerce. March 1982.

Caves. Richard. and Ronald Jones. World Trade and Payments:
Ag Introduction. third ed.. Boston: Little. Brown and
Company. 1981.

Chipman. John. ”A Survey of the Theory of International

Trade: Part 1. The Classical Theory". Econometrica. 33.
July 1965. pp. 477-519.

164

I65

Daly. J. ”Uses of International Price and Output
Data". In: D.J. Daly (ed.). International Comparisons
gfi Prices and Output. Studies in Income and Wealth.
Vol. 27. New York: Columbia University Press for
National Bureau of Economic Research. 1972.

Davis. L.A.. "Technology Intensity of U.S. Output abd Trade."
Office of Trade abd Investment Analysis. U.S.
Department of Commerce. February 1982.

Deardorff. Alan. "The General Validity of the Law of
Comparative Advantage”. Journal gfi Political Economy.
October 1980. pp. 941—57.

Denison. Edward. Why Growth Rgtes Differ-Postwar
Experience 1g Nine Western Countries. Washington. D.C.:
Brookings Institution. 1967.

Haberler. Gottfried. The Theory 9f Interngtiongl Tradez
with its Applicgtions 59 Commercial Policy. (translated
From the German by Alfred Stonier and Frederic Benham).
New York: The Macmillan Co.. 1936.

Graham. Frank. "The Theory of International Values Re-
examined". Quarterly Journal 9f Egonomics. November
1923. p. 54-86.

Johnson. J. Econometric Methogg. second ed.. New York:
McGraw-Hill. 1972.

Jones. Ronald. "’Two-ness’ in Trade Theory: Costs and
Benefits". International Trade: Essays 1g Theory.
Amsterdam: North-Holland Publishing Co.. 1979.

Judge. George. et al.The Theory and Practice gfi
Econometrigg. New York: John Wiley and Sons. 1980.

Kelly. R.K.. "The Impact of Technological Innovation on
International Trade Patterns." Office of International
Economic Research. U.S. Department of Commerce.
December 1977.

Kmenta. Jan. Elements gf Egonometrics. New York: Macmillion.
1971.

Koopmans. Tjalling. "Maximization and Substitution in Linear
Models of Production". Igput-Ogtput Relations (edited
by the Netherlands Institute). Leiden: H.E. Stenfert
Kroese N.V.. 1953. pp.99-109.

Kravis. Irving. "'Availability' and Other Influences
of the Commodity Composition of Trade". Journal 9f
Political Economy. April 1956. pp. 143-55.

166

. ”Wages and Foreign Trade". The Review gf

 

Economics and Statistics. February 1956. pp. 14-30.

Kreinin. Mordechai. International Economics: A Policy
Aggroach. fourth ed.. New York: Harcourt. Brace.
Jovanovich. Inc.. 1983.

"Labor Costs and Comparative Advantage in
Motor Vechicles and Steel: Further Analysis and

 

Extensions". unpublished paper. Michigan State University.

1983. pp. 1-20.

"The Theory of Comparative Cost-Further
Empirical Evidence". Economia Internagjonale. November
1969. pp. 662-73.

 

. ”United States' Comparative

Advantage in Motor Vechicles and Steel". included as
Chapter 6 in Brazer and Laren (eds.) Michigan's Fiscal
ggg Economic Structure. Ann Arbor: University of
Michigan Press. 1982. pp. 150-176.

 

. "Wage Competiveness in the U.S. Auto and
Steel Industries". Contemporary Poligy Issues.
January 1984. pp. 39-50.

 

Lary. Hal. Imgorts gfi Manufactures from Less Developed
Countries. New York: NBER. 1968.

Lebergott. Stanley. "Wages and Structures". The Review gfi
Economics and Statistics. November 1947. PP. 274—85.

MacDougall. Donald. "British and American Exports: A Study
Suggested By the Theory of Comparative Costs. Part I".
The Economic gournal. December 1951. pp. 697-724.

. Monica Dowley. Pauline Fox. and

Senta Pugh. "British and American Productivity Prices.
and Exports: An Addendum”. Oxford Economic Pagers.
October 1962. pp. 297—304.

 

Marks. Stephen. "Comment on Kreinin". Contemporary Policy
Issues. January 1984. pp. 51-53.

Mill. John Stuart. Principles 9f Political Economy.
edited by W.J. Ashley. London: Longman’s Green. and
Co.. 1909.

Minhas. Bagiche. "The Homohypallagic Production Function.
Factor Intensity Reversals. and the Heckscher-Ohlin
Theorem". Journal 9f Political Economy. April 1962.
Pp. 138-56.

167

Niroomand. Farhang. "Determinants of the Structure of U.S.
Foreign Trade in Manufacturingi963-I980." Diss. Michigan
State University. 1983.

Parry. Thomas. ”Trade and Non-trade Performance in U.S.
Manufacturing Industry-Revealed Comparative Advantage".
The Manchester School. 1975. pp. 158-72.

Ricardo. David. Qg the Principles gf Political Economy. and
Taxation. in The Works ggg Correspondence 9f David
Ricardo. Vol. 1. ed. by Piero Sraffa. Cambridge. U.K.:
Cambridge University Press. 1951.

Samuelson. Paul. "Abstract of 8 Theorem Concerning
Substitutability in Open Leontief Models". Activity
Analysis 9f Prodgction ggg Allocgtion (edited by
Koopmans). New York: John Wiley. 1951. pp. 142-146.

Schmidt. Peter. Econometrics. New York: Marcel Dekker. inc..
1976.

Stern. Robert. “British and American Productivity and
Comparative Costs in International Trade". Oxford
Economic Pagers. October 1962. PP. 275-296.

 

- — . "Testing Trade Theories". In: Peter Kenen
(ed.). Interngtiongl Trade and Fingnce: Frontiers for
Research. New York: Cambridge University Press. 1975.
pp. 3-49.

Stryker. J. Dirck. "The Sources of Change in Export
Performance: The United States and Canada". in Kenen
and Lawrence. eds.. Igg. Ogen Economy: Essays gg
International Trade and Finance. New York: Columbia
University Press. 1968. pp. 150-74.

Takayama. Akira. International Trade: gg Aggroach g9 the
Theory. New York: Holt. Rinehart. and Winston. 1972.

Taussig. Frank. International Trade. New York: The
Macmillan Co.. 1927.

 

-------------- . Princigles gfi Economics. Vol. I.
second ed.. New York: The Macmillian Co.. 1919.

Torrens. Robert. Ag Essay gg the Exterggl Corn Trade.

reprint. Clifton. N.J.: Augustus Kelly. 1972.

Viner. Jacob. Studies 1g the Theory gf International
Igggg. New York: Harper 8 Row. 1937.

"liliiillWilli?)