Hitotsubashi University Repository

Title Intensity Analysis of World Trade Flow

Author(s) Yamazawa, Ippei

Citation Hitotsubashi Journal of Economics, 10(2): 61-90

Issue Date 1970-02

Type Departmental Bulletin Paper

Text Version publisher

URL http://doi.org/10.15057/8041

Right

INTENSITY ANALYSIS OF WORLD TRADE FLOW

By IPPEI YAMAZAWA*

I. Introduetion

There has been an increasing interest in the interdependence among economies in the world. This is partly because advanced economies are forced to expose themselves interna-tionally in the globa] movement of trade liberalization in the post-war era on one hand and the structural adjustments in global division of labour required to promote more efflcient economic aid to countries in the 'South' on the other. Furthermore, it is also caused by the increased necessity for the harmonization of cyclical policies among major economies in the present crisis of world monetary system. An increasing number of studies in world trade flow, traditionally the most important aspect of interdependence among economies, have been

published recently.l This paper measures the interdependence among countries based on world trade matrix

and analyzes factors affecting it. Traditional theory of international trade tells us that trade

is determined by the difference in comparative advantage structures between countries. But many traditional theories were worked out in two-country models, and it has been pointed out by many writers that various other factors are important in determining trade in many country models, which will be given some empirical evidence in this paper.

Two methods have been developed to analyze the world trade flows; gravity model and trade intensity index. In gravity model, trade between two countries is mechanistically de-termined by gross national products of exporting and importing countries and economic dis-tance between the two. The GNP of an exporting country represents the size of its supply capacity and that of an importing country its total dernand. The volume of trade between the two countries tends to increase if the GNP of either country increases, and tends to de-crease, if the economic distance between them (measured in terms of transport cost) increases. If this relationship holds between any pairs of countries, country i's export to country j is

expressed as

* Lecturer (K~shi) of Economics. Hitotsubashi University. The author has benefited from intense discussion with Professor Kiyoshi Kojima of Hitotsubashi Uni-

versity and Dr. Peter D. Drysdale of Australian National University. He is also indebted to Miss Masako Kohno and the computer center at the Institute of Business Research. Hitotsubashi University for stati-

stical work. l In addition to the literatures referred to in this paper, see the following survey article of world trade

models. Taplin, G. B. "Models of World Trade",IMF Staff Papers, Vol. 54, No. 3 (Nov. 1967). Inter-dependence among economies was one of the main theme in the conference on the linkage of econo-metric models of individual countries held in Tokyo September 1969 by Japan Economic Research Center.

62 HITOTSUBASHI JOURNAL OF EcoNoMICS [February t'

where Yi, Yj be the GNP's of the two countries, Dtj be the measure of economic distance between them, and a, p, r, 6 be positive constants.2

Trade intensity analysis, on the contrary, concentrates the structure of departures of actual

trade flows from trade fiows estimated in gravity model. The index of intensity of country i's export trade with country j is defined by

Lj~~X~Xi /X~:LIX:. (2) where X~' (~ ~ Xtj), X.j (= ~ Xsj) and X.. (~ ~ ~ Xij), represents the total export of countr

i, total import of country j, and the total volume of world trade respectively.3 It is easily proved that, in a simplified gravity model where bilateral trade is solely determined by the GNP's of a pair of countries, Iij is always equal to unity.4 That is, Iij equals unity when the value of trade is proportional to the GNP's of the two countries, it exceeds unity when the trade becomes more intensive between the pair of countries, and it falls short of unity when trade becomes less intensive between countries i and j. High intensity of trade refiects such various factors as the strong complementarity in comparative advantage structures between

2 Tinbergen applied the equation (1) to the trade flows among 42 countries and obtained the following results.

log Xij= -O. 6627+1. 0240 Iog Y~+0. 9395 Iog YJ (, 6802) (. 0270) C. 0269)

-O. 8919 Iog Di, (. 0455) R2= . 8094

Tinbergen J., Shaping the World Econo"ry, New York, 1962. 3 Brown A. J., Applied Economics, Aspects of the World Econo'ny in War and Peace. London, 1947.

Kojima K., Sehai Keizai to Nihon Boekl (World Econo'ny and Japan's Foreign Trede). Tokyo, 1962. Both Brown and Kojima defined intensity of trade as

Xij / X,j Itj*j~X~. / X..-X.i

In theory, country j's import share in total world import minus country t's export may be more appro-priate reference for the geographical distribution of country i's exports than j's import share in total world trade unadjusted (X.j/X..). However, this adjustment of denominator not only complicates the calculation procedure but also makes it impossible to derive the clear-cut relationships among various indexes .

The weighted sum of li, for all import markets amounts to unity in either formulation,

X.j X.j )

~ ( .. , ~ .t Itj = 1 X X..-X Ilj* 1 j(~1) j(~i) The degree of overestimation resulting from the use of the formula (1) is proportional to the exporting

country's share in total world trade (Xi./X..) and will not differ much among countries except between US and small exporters.

d The simplified gravity model is represented by Xij =a Yt; Yjr

Since Xt. =aYi;;~ Yjr

X.j=aYjr:~t Yip

and X..=a(~:8 Yte)(~j Yjr),

the intensity of trade is always equal to unity Lj = l.

1970] INTENSITY ANALYSIS OF WORLD TRADE FLOW 63 the pair of countries, smaller geographical and psychic distances between them, and mutually favorable trade agreements between them, and low intensity the countrary situations. In this sense trade intensity analysis and simplified gravity model are complementary with each other, which may be compared to price and income analysis of world trade flows.5

Various factors mentioned above are reflected in the value of intensity of trade. If the

effects on trade intensity of the degree of complementarity in comparative advantage structures are separated from the effects of other factors, it enables us to identify traditional trade-determin-

ing factors and their overtime changes. Dr. P. Drysdale has developed the intensity analysis by decomposing trade intensity into two main components; the "commodity bias" or the "degree of complementarity" which mainly reflects the complementarity in comparative advan-tage structures between exporting and importing countries, and the "degree of special country

bias" which sums up the effects of other factors.6 One aim of this paper is to reformulate Drysdale's decomposition of trade intensity more

rigorously and to associate the intensity analysis with the studies of comparative advantages. The following four sections are devoted to the theoretical or methodological arguments of trade intensity analysis, and an alternative decomposition of trade intensity is represented in

Section 5. Another aim is to produce an application of the theory to the matrix of world trade flow

at three periods. Here the world-wide tendency of lv converging to unity is depicted, which is related to changes in the global division of labor caused by industrialization of primary good exporters on one hand and to the breakup of traditional trade ties and the formation of new ones on the other. This empirical testing is complementary to that of Drysdale's to the extent that the latter analysis gives detailed consideration of the two components of trade

intensity between Japan and Australia at individual commodity level whereas the former gives the whole structure and its overtime change of the intensity of trade among various countries in the world. Section 6 explains statistical data and procedures and the last three sections describe the application. At the end is attached an appendix which clarifies some technical

characteristics of trade indexes.

II. Two Components of Intensity of Trade

A country's patterns of exports to and imports from the world are principally determined

by its structure of comparative advantage and disadvantage vis-~-vis the world. Assume a homogeneous commodity be traded in a world where both transport costs and artiflcial impedi-

ments to trade are negligible. Then the country i's export of commodity h to country j is expected to be the product of country j's total import of this commodity multiplied by the share of country i in the world trade of the same commodity, which is expressed as

= .,~ h) ~ X 5 The concept of trade intensity i5 used for the prediction of world trade flow in Uribe. P. de Leeuw,

C. G. and Theil, H. "The Information Approach to the Prediction of Interregional Trade Flows," Rev.

of Econ. Stud., Vol. 33 (1966). 6 Drysdale P. D., Japanese Australian Trede, 1967 (Unpublished Doctoral Dissertation).

64 HITOTSUBASHI JOURNAL OF EcoNoMlcs [February where Xijh is the expected value of country i's export of commodity h to country J which is rewritten as

X~ijh Xt "~ . X. jh

~ The expected value of total exports from country i to country j is defined as the sum of expected values of all commodities.

The expected intensity of trade is obtained by replacing the expected value of trade for the actual one in the formula (2).

C - Xij /X.f

The divergence between the expected value of trade and the actual one defines the degree of special country bias.

_ Xaj Bsj- X~ij

_ Xtj _ 1

(7) ~ X~sj h LVl(\X_Xi!f ) B;j

where Bij is the degree of special country bias in each commodity trade (B$j't=Xijh/X~ijh) and

Bif turns out to be a weighted harmonic mean of Bij . The first line of equation (7) gives a decomposition of trade intensity into two components

which is the basic formula for our analysis.

III. Determinants of Complementarity

What economic implications are attached to the expected intensity of trade and what factors determine its value? It is rewritten as follows.

~Xijh IX,j Cij= hXi. / X..

_ Xt"e X.f X..

~( )

~ t' ' .j ' X..h

~( h f = ~ )( ,~)( '2 .h X. Xt X. X.j X.

, ) X X X X X ~( ::h th' =h ) ,} Xl

w here

S Xi '} . X. '} sh= / :. • j't= ! / :.

1970] lNTENSITY ANALYSIS OF WORLD TRADE FLOW 65 Sih and Rj are the share of commodity h in i's total exports and ,i's total imports respectively

both divided by commodity h's share in world total trade.7 They measure the degrees of country i's export specialisation and country j's import specialisation in commodity h respec-tively. Since their weighted average over all commodities always take a constant value of unity,

~( X..h )Si'~=~1( X.. )Rjh=1 (11) X. . X. .

each of them takes value around unity. Sth of over (under) unity implies that country i ex-ports commodity h more (less) intensively than the world average, and the higher (lower) the value of Sih the stronger (weaker) is country i's export specialization in commodity h. Simi-larly, the higher (10wer) the value of Rf}, the stronger (weaker) is country j's import special-

ization in commodity h. The vector of Ssh over all commodities,

2 , Si'e) (Sil,Si ••-shows the structure of export specialization of country i, which reflects country i's structure

of comparative advantage. Sih of higher values indicate commodities with the production of which the country has strong comparative advantage, while those of lower values indicate those of weak comparative advantage. This also applies to the vector of indexes of import specialization. However the structure of import specialization is affected not only by the struc-

ture of comparative disadvantage but also by protective commercial policies much more than

that of export specialization. Table I gives the numerical example of a hypothetical case of two countries A and B ex-

porting to country C. As is shown in Table I (b), both exporting countries, A and B, have the similar patterns of export specialization with the highest value in commodity IV and the lowest one in commodity I. Country C has the structure of import specialization which match

the structures of export specialization of its trade partners. Although countries A and B have the similar patterns of export specialization, the depar-

ture of the indexes of export specialization from unity is half as large for each commodity in

country B as in country A. That is, A has more concentrated structures of export speciali-zation than B, or B has more diversified one than A. The degree of concentration or diver-sification is affected by such important aspects of comparative advantage as size of a country,

skewed resource endowments, and so on. They are measured in terms of standard deviations

of specialization indexes from their mean, unity.

= o(S1) X. . (Si'h - 1)

X ~~( • -=

) 2

a(Rj) X. . (R3h - l) X

As is shown in the right-end column of Table I (b), the standard deviation of country B is

half of that of country A.

7 A country's structure of export specialization is principally determined by its structure of compara-tive advantage. In theory a country's comparative advantage is represented by comparative cost before trade is opened, but in empirical studies variables of export performance are frequently used to reveal its comparative advantage, since in international trade statistics price data are less reliable than value data.

See Balassa B. , 'Trade Liberalization and "Revealed" Comparative Advantage', The Manchester School oj Econonac and Socral Stedres Vol. 33. No. 2, May 1965, pp. 91 123

66

TABLE (a)

HITOTSUBASHI JOURNAL OF ECONOMICS [February 1 . AN EXAMPLE OF THE CALCULATION OF THE DEGREE OF COMPLEMENTARITY

Commodity Compositions of Country A's and B's Exports, Country C's Imports, and World Trade.

(b) Relative Share Indexes and Their Standard Deviations.

(c) Degrees of Complementarity. CAC=cov (SA, Ra)+1=1. 2692, CBC=cov (SB, Ro)+1=1. 1346

cov (SA Rc) O. 2692 _ ~' _ _ _ O. 1346 rAa . . O 8946 rBa . . O. 8946 ,1(Sl)xa(Ro) ~ O 6928xo 4343 ~ ' ' ~ = O 3464xo 4343

Covariance of the indexes of country i's export specialization and those of country fs im-port specialization is defined as

cov(St Rj)~~Llr( X.. )(Sih-1)(Rjh-1)

' X.. L( ,f =h )

X. X' Si'} • Rjh - 1 =Ctj-1

or

C,j = cov (Si, Rj) + I (13) This gives a clear economic meaning to the expected intensity of trade, Cij. If country i's structure of export specialization matches country j's structure of import specialization closely, that is if indexes of i's export and j's import specialization are positively correlated (cov(St,

Ri)>0), Cij takes a value over unity, while they match poorly, or they are negatively cor-related (cov(Si, Rj)<0), Cti takes a value under unity. If they are independent (cov(Si. R/)= O), Cij equals unity. In this sense, the hypothetical intensity of trade measures the degrlee of complementality in the specialization structures between trade partners.

The degree of complementarity is not only affected by the degree of match of the special-

ization structures of exports and imports, but also by the degree of concentration or diver-sification in them. A country with highly concentrated structure of export specialization tends

to have high complementarity in its export trade than another country with the similar but more diversified structure of export specialization. Thus if we calculate the correlation co-efficient between the specialization structure of exports and imports, we obtain the measure of the degree of match of the two structures neutral from the degree of concentration or

1970] INTENSITY ANALYSIS OF WORLD TRADE FLOW 67 diversification.

(F(Si) • c(Rj) In the hypothetical case of Table l, the degree of complementarity of A's export trade

with C is twice as high as that of B's export trade with C, whereas they have the same cor-relation coefiicient, which implies that the difference in CAC and BBC is solely attributed to the

difference in the degree of diversification in the structure of export specialization between A

and B.8.9 Regression analyses of indexes of export specialization on variables or their proxies of the

alleged determinants of comparative advantage such as capital-labor ratio, skilled-unskilled labor ratio. R and D variables, scale economy variables attached to the production of indi-vidual commodities,

h, ) ) S f(( h, 1'b= ) ( K Ls can verify the theories of comparative advantage.lo

Overtime changes of structures of export specialization can be explained by changes in such determinants of comparative advantage as mentioned above by means of regression anal-ysis. An alternative is to regress a variable which characterizes the overtime change in the structure of export specialization on such broad factors as the rate of total investment (1), the

rate of increase in education (b), and the rate of wage increase (~,) all of which cause changes

in comparative advantage. For example, the ratio of average specialization indexes of two commodity groups one of which represents simple labor intensive commodities and the other high technology intensitive commodities may be used for that pulposes.

i= Sil =f(1 w e ) (16) S

Although structures of import specialization are also affected by other factors than compara-tive advantage, they generally show contrary but less concentrated structures than those of export specialization.

8 The effects of rtj and those oi c(Si) and a(R/) are distinguished by Drysdale in his explanation for a decline of 30 to 40~; in total commodity bias in Japan's export trade with Australia in the post-war period.

"At first sight it appears strange that total commodity bias fell since the structure of Japanese exports grew rather more like the structure of Australian imports between 1953 and 1963. An overriding con-sideration was that the composition of both trade flows came to resemble the compositions of world trade more closely than before. Hence there was less reason to expect Japan to export relatively more to Australia than to other countries because of their uniquely similar trade structures." Drysdale P., op. cit. p. 164.

The first statement refers to the increase in rJA (although the contrary evidence is shown in Table 5 and the second to the decline in a(SJ) and a(RA).

9 Linneman introduces to his gravity model analysis of world trade flow a similar concept to our ri!' However he has not given its clear economic implication.

See Linneman, H., An Econo'netric Study OJ International Trede Flow, Amsterdam, 1966. IQ A similar attempt was made in Gruber W. H. and Vernon R. The Technology Factor in a World

7rede Matrix, mimeographed 1969. However, they used as the dependent variable country i's share in world trade of commodity h (Xt.h!X.,h) instead of our relative share index. Sih.

68 I{ITOTSUBASHI JOURN'AL OF ECoNohllcs [February

IV. Determinants of Special Country Bias

The intensity of trade between a pair of countries is also affected by other factors than

those which determine the commodity composition of each country's global trade. They are put aside from the hypothetical model of homogeneous commodity trade. They include trans-port cost, discriminatory tariffs and other import restrictions, product differentiation within commodity classes, and other international economic relations than trade such as capital move-ments and economic cooperations.

They are supposed to affect all commodities in a bilateral trade to the same extent, so that the degree of complementarity between a pair of countries is solely determined by the global structures of export and import specialization and is neutral from other factors. This is not necessarily the case in reality. Transport cost tends to depress imports of bulk com-modities from the distance more than those of other commodities, so that the global com-modity composition of the exporter does not refiect precisely its structure of export speciali-

zation relevant to the importer. Similarly, preferential tariff arrangements may discriminate against non-mernber sources of supply in imports of manufactures more than in those of raw materials. Product differentiation is more important in finished manufactures than in semi-finished ones. Therefore, the degree of special country bias is not necessarily the same for all commodities but vary around the overall special country bias in each bilateral trade, which

is a weighted harmonic mean of special country bias of individual commodities. The size and structure of divergence vary in each bilateral trade, and we can, for the

first approximation, regard them as random disturbance around the overall bias and indepen-dent from the overall degrees of complementarity and special country bias. Thus we can separate two groups of factors affecting the intensity of bilateral trade. One is the structure of comparative advantage of exporters and importers modified somewhat by their commercial policies, which determines the degree of complementarity between two countries. The rest of the factors constitutes the second group which mainly determines the degree of special country bias in bilateral trade.

Traditional theories have been forcussing on the first group of factors as the determinants

of the pattern of international trade and occasional mention have been made for factors of the second group. This is partly because most traditional theories of international trade have

been worked out in two country model in which no special country bias can be introduced until recently discriminatory effects of customs union or free trade area is explored in three

country model. It is interesting to see, from the viewpoints both of theory and policy in international economics, how much the second group of factors affect trade patterns. .

It is important to note that a weighted harmonic mean of special country biases of a country ~vith its all trade partners is always unity.

~(Xvij~ Dl ~ ~ - X~i j

~. . A*. ) Dij

That is, if a country's overa]1 special country biases are larger than one with some countries,

they must be smaller than one with other countries. Therefore, the structures of overall special country biases in bilateral trades between all possible pairs of countries in the world

69 INTENSITY ANALYSIS OF WORLD TRADE FLOW 1970]

throw light on the study of major determinants of special country bias . The degree of special country bias in country i's export trade with country j is regressed

on various factors which characterize ecocomic relationships between the two countries. Eco-nomic distance is measured in terms of average transportation cost between the two countries (Dij) ; the effects of preferencial trade arrangements such as common market, free trade area,

and Common Wealth Preferences is introduced in dummy variable form (P,j) whose value is either one if such preferencial arrangement exists between two countries or zero if not.

If a product of country i is preferred by country j to the products of other countries in the same commodity class, the hypothetical volume of country i's export of this commodity to country j (X~ijh) understates the true one based on country i's structure of comparative advantage relevant to country fs market, so that the special country bias for this commodity will be overestimated. The closer is the living standard of the two countries, the more their

products tend to be preferred each other in many commodity classes. Thus the difference in per-capita incomes between the two (AYij) is introduced as a proxy variable to represent the effect on overall special country bias of proximity or familiarity through product differentiation.

Drysdale suggested that the increased investments in Australia by U.S. since mid-1950's and by Japan in recent years tend to increase imports of capital goods and related commodities

from these countries, resulting in higher special country bias in the export trade of these countries with Australia. This is more likely to be the case for distribution of economic aids by advanced countries among LDC's since tied loans are the most used form of giving aids. A variable quantifying the intensity of economic relationships other than trade is introduced

to represent these factors, for example

Aij= Kf'/ K..

where Kij is the flow of capital or aids from country i to country j.

Above arguments are summed up in functional form as

AYij J;ij=f(Dij, Pij, AYij, Aij, •••) (19) Dij and A Yij tend to affects Bij in negative direction, whereas Pij and Aij in positive direc-

tion. Overtime changes of special country bias also refiect changes in some of these underly-

ing factors and overtime analysis of Bij contribute to the study of the determinants of special

country biases, since some of them are more easily quantified in their changes than their

absolute levels. It is interesting to note that symmetrical movements are expected between B,j and Bji,

special country biases in bilateral trade of both directions, since most determinants of special country bias tend to affect bilateral trades in both directions in similar ways, which is likely

to be the case for economic distances, proximity in living standard or familiarity, and reci-

procity in trade arrangements,

V. Alternative Formidation of Complementarity

Another formula of the degree of complementarity is derived from an alternative assump-tion of world trade flow. Assume, in stead of a homogeneous commodity, that the product of each exporter in a commodity category is differentiated from each other's product, so that

70 HITOTSUBASHI JOURNAL OF EcoNOMlcs [February a country's import of the commodity is composed of various exporters' products and each exporter's share in it is determined by the preference pattern of the impbrter.11

Thus an exporter's performance in different import makets should be considered to reflect its comparative advantage relevant to each market, whereas its export to the world as a whole is merely the average of its performance in different markets. The structure of country i's specialization in a particular market is expressed in index form by

Ssj Xij'~ X. '~ 20) 'e=

Xij X while the country's export specialization to the world as a whole (Sih) is merely an average of individual SiJ weighted by the geographical composition of the country's export,12

Sih ::: Vl( Xij )Sij

. h 21) Xi j

Degree of complementarity is defined in this alternative model by the sum of products of Sijh and Rjh weighted by the share in world trade of each commodity, which equals the covari-ance of Sij'~ and Rj added by one.

-~(X.. ) C~j* = Stjh ' Rj X. .

= cov (Si,h, Rjh) + I (22) If the structure of i's export specialization in fs market matches that of fs import speciali-zation, the potential volume of i's export to jtends to be larger than those of its competitors, so that the potential intensity of i's export trade with j in the world with neither transport cost nor artifical trade impediments (C;j*) will be over unity. In the reverse case Cij will be under unity.

Degree of special country bias is defined by the gap between the actual and expected

11 ommodity classifications detailed enough to satisfy the homogeneity assumption are not available in reality, so that the model of differentiated products will be more practical in this sense. For the use of this assumption in the studies of international economics, see Johnson, H. J. "The Internatronal Competr tive Position of the United States and the Balance of Payments Prospects for 1968." Rev. of Econ, and Stat. Vol. 46, N. l, (Feb. 1964), p. 23-25.

Each exporter is as competitive as in each importer's market in the model of homogeneous products, while in the model of differentiated products, an exporter is more competitive in some markets than it is in others. These are two alternative sets of assumptions on world trade flow and are distinguished from each other also in Oom, V. D., "Models of Comparative Export Performance" Yale Econ. Essays, Vo]. 7, No. 1. (Sep. 1967).

lz ijh may be defined alternatively as i's export performance of commodity h in fs market divided not by the share of commodity h in world trade, but by that in J's import trade.

Xif~ / X.jh (i) Stjh= Xij / X.j

However. Ctj with Sijh in equation (i) in the place of Sijh in equation (22) turns out to take the constant value of unity for all cases,

Xh CiJ ) =

( ". Sif}.R!~

X.

_ X.,h VX_i~i / X..h X_ij_'}V~:ii~X. '~ ~( Il ej / X.. ) /

X.j )L\ X.j ~ .. /\ X /

Xif~ 1 =

~h( =-Xi j

1970] 1NTENSITY ANALYSIS OF WORLD TRADE FLOW 71 values of trade intensity.

Bij*= Cij*

Thus we have obtained an alternative formula for the decomposition of lij. Values of Sij'~ are independent from such factors as distance and artificial trade impedi-

ments provided that these factors affect the trade of all commodities uniformly, and, therefore,

C:~j* will be independent of these factors. And the difference between C;j and Cij* comes

out from alternative definition of commodity category. In order to distinguish between the two models, consider the following cases. Suppose

i's productivity in the production of commodity h improves so that i's export specialization in

commodity h is strengthened. Country i's export to country j which has strong import specialization in the same commodity tends to increase and so does th~ intensity of i's export

trade with j. Since both S.'~ and Sijh mcrease in this case, the increase in li, will be explained

by higher complementarity in either model. On the other hand, in stead of over-all increase in productivity of commodity h, a change

from low quality products to high quality one in the production of commodity h occurs. Export specialization in high-quality h commodity will increase, whereas that in low-quality

one will decline, so that i's export specialization in h-commodity as a whole remains constant. Country i's export to country j with strong preference for high-quality h commodity tends to increase, Ieading to higher intensity of i's export to j, whereas i's export to country k with

strong preference for low-quality h commodity tends to decrease, resulting in lower intensity

of i's export trade with k. This change is reflected in the export specialization in individual markets (increase in S,jn

and decrease in Sikh), but not in export specialization to the world as a whole (constant Sih).

Thus the degree of complementarity of i's export trade with j will, ceteris paribus, increase, and that with k will decline in the alternative model, while both of them remain constant in the initial model, so that the changes in I,j and Ith are explained by changes in C,j and Cih in the former model, but in the latter they have to be explained, with constant Cij and Cih,

by changes in Bij and 1~ik, that is, changes in special country biases. Thus in the initial model, there always exists the tendency of underestimation of Ci,

which are eliminated partly at least in the alternative model. However, only the initial model

is applied in the empirical study of this paper since additional data are required for the com-

putation of the alternative model.

VI. Statistical Problems and Procedures

Calculations are based on trade statistics taken from U.N.'s Monthly Bulletin of Statistics (Special table E published in March issue every year). It provides world trade matrix of six-

teen regions for years 1955-67 in U.S, dollars and commodity composition of global exports and imports of each region according to six commodity categories. Matrix of trade flows among fifteen regions (USA. Canada, Latin America, EEC, UK. EFTA excluding UK. Other Western Europe, Eastern Europe including USSR, South African Republic, Other Africa, Japan, West Asia-Asian part of Mid East, Other Asia, Mainland China and other Asian Communist countries, Australia and New Zealand) are constructed for three-year averages at

72 HITOTSUBASHI JOURNAL OF ECONOMICS [February three periods 1955-57, 1960-6_2, and 1965-67. The commodity composition of global exports and imports of each region for the same three year-averages at three periods are calculated under six commodity classes (Standard International Trade Classification Section O and 1, 2 and 4, 3, 5, 7, 6 and 8).13 The data has both merits and demerits, the effects of which should be taken into accounts in the interpretation of the results of calculations.

(1) Country groups are taken as units instead of individual countries except five countries,

US, CAN, UK, S AFR, and J each of which makes a unit by itself. This tends to mitigate irregular trade figures inherent to small, primary exporting countries. Furthermore, this makes it possible to cover the whole world trade and to conclude about the over-all structures of Isj' Cij and Bij. However, country grouping tends to make even structures of export and import specialization and to under-estimate both rij and a resulting in under-estimates of Cij (to be exact their departure from unity) and thus in under- or over-estimate of Bsj'l4

(2) All indexes of Sih, Rj'~, Cij, Iij. BiJ are ratios of ratios and tends to be neutral from

cyclical or irregular movements of absolute trade figures. And three-year averages with five year durations between each period are appropriate to eliminate short term fluctuations and to look to the secular changes of trade relationships.

13 The statistics before 1961 are shown under different regional classification and are adjusted to that of after 1961. The sixteenth region, the rest of the world which consists of Carribian and Pacific islands, is excluded. Figures in world trade matrix include SITC Section 9 (miscellaneous transactions and com-modities n. e. s.) which is excluded from data of commodity composition. These omissions prevent the data from being fully exhaustive of world trade, but both of them are listed as residuals and suffer con-siderably from statistical errors and the inclusion of them doesn't seem to add much to our analysis.

14 If a country construct a unit by itself, zero should be attached to the expected value of trade within the unit.

Xit:~0. Thus the sum of expected value of trade over all world trade defined by the formula (4) tends to under-state the world total

Xi. hX. jh Xi . '} - =

~ ~ XifL= X..'} ~ X,fL X. J~ i ,(i~j) i J(~t) i ~(L\ ~..; )~ X.j'l ~t

Xi.h = . (X~.h-X,jh) Xn =~X h(1- X.ih

i.

X..h

X..'~- Xt.'~'X.jh =

i X..h

;X..h[1 ~( )] X,.h V X.jh ~ ..h j~ X..h '

This understatement will be adjusted by defining the expected value of trade divided by the factor in parenthesis.

(Xi jh)* = Xt jh ' 1

[ a )] Xt.h V X.th '

~(

l- X..h A X..!~

In our analysis, only five units need this adjustment, which does not seem to alter our results consider-ably.

1970] INTENSITY ANALYSIS OF WORLD TRADE rLOW 73 (3) Six commodity classification covers all important commodities traded, which is re-

quired to derive unbiased estimates for the degree of complementarity. However, each com-modity category is often too broad to catch important changes of commodity specialization.

Let a commodity category c consists of two sub-commodities a and b. Then the index of export specialization in commodity c is the weighted average of those of the two subcom-

modities. Sic= Xi.a+X,,b X..a+X b

Xi . X. . (X.. )Sa+(X..a) ib (24)

X. . Xc The excess of th~ variance of country i's export specializations under the commodity classifi-

cation with the commodity c separated into the two sub-commodities ((r2) over that under the aggregated commodity classification (a2), others being equal between the alternative classifica-

tions, is X.. (S,b-1)2-(X.. ) rl 2-6,2=(X.. )(Si"-1)2+( (Sic _ 1)2

X. . X. . * X. . X a.X b (25) = " - Si"-Stb)2 X. . • X. .c ib, (Ti2>a,2. That is, so long as a commodity category is composed ~ Therefore, so far as Si"~S

of sub-categories with different values of export specialization, the degree of diversification in trade specialization calculated from aggregated commodity classifications tends to under-

estimate its true value. On the contrary, the degree of complementarity calculated from the aggregated commodity

classification (C^ij) tend to over- or under-estimate the true values of complementarity (C,j).

Their difference is represented by

^

( X.,aX.. )(Sia_Sib)(RfL-Rjb), (26)

Ci, - Cij - X. .X. .e

If the specialization of both country i's export and country j's import are stronger in one sub-commodity than in the other of the commodity c, C^tj falls short of Cij. If on the con-trary they are stronger in different sub-commodities, C^ij exceeds Cij. Since the former is likely to be the case between countries with contrary specialization structures whereas the latter

is the case between countries with similar specialization structures, the degrees of comple-mentarity estimated from aggregated commodity classification tend to under-estimate the de-

partures of their true values from unity. Since the value of lij is not affected by the degree of aggregation in commodity classifica-

tion, then with given value of lij the under-estimation of C~j due to aggregation tends to over-estimate Bij when both B,/ and Cij are over or under unity, while it tends to under-

estimate Btj in other cases.

VII. Complementarity among Fifteen Regions

Table 2-6 give the calculations of trade indexes in II-IV among fifteen regions of the world

at three periods. This and the following two sections are concerned with the global structures of C,j, ~ij and Itj and their overtime changes. In this section the structure of export special-

ization are explored for each region at three consecutive periods, which is summed up to the analysis of the structure of C,f and its overtime changes. It will be shown that the g_ rowing

74 HtrOTSUBASHI TOURNAL OF EOONOMICS [February

industrialization of traditional primary exportors has made the structure of trade specialization

more diversified in most countries and caused the tendency of C ' ij convergmg to unity. In the next section an off-hand analysis depicts some factors affecting the structure of Bij

and the symmetricity of their effects. The tendency of Bij converging to unity is pointed out and associated with the reformation of the existing trade ties. The structure of Cij is super-imposed by that of Bij to produce trade intensity relationships among fifteen regions in the last section. It will be interesting to note that Bij dominates the structure of Itj in one of

eight and that there exists a strong tendency of 1 ij converging to unity. These propositions on the global structure of trade flow and its overtime changes are derived by means of con-tingency tables, but they are to be supplemented by more rigorous regression analyses described

in 111 and IV.

To begin with the commodity composition of world trade (the first three rows in Table 2a and 2b), primarjr goods (SITC Section 0+1, 2+4, 3) and manufactured goods (5, 7, 6+8) are balanced in 1955-57 (.4910 vs .5090). Each category in the former group has declined and each of the latter has increased steadily in the following decade. (The combined share of the manufactures is .5623 in 1960-62, and .6294 in 1965-67).

There are two groups distinguished clearly from each other by the structures of export and import specialization. One is the group of such industrial exporters (Group I) as US, UK, EEC, and Japan, and the other, the group of primary goods suppliers (Group P), consists of the rest of the regions except O EFTA and E EUR. The last two regions are classified as the first group in the structures of their export specialization but as the second group in those

of their import specialization.

Typically Group I regions have indexes of export specialization over unity in manufac-tures but those of under unity in primary commodities, while Group P regions have contrary structures of export specialization. But the details differ within groups according to stages of

development. US has strong export specialization in categories 7 and 5 but rather weak one in category 6+8. UK and EEC are strong in all three categories, and Japan has the strogest in 6+8 but her export specialization has just risen above unity in categories 7 and 5 by 1965-67. Since export specialization in primary commodities are considerably weak both for Japan and U.K. but not very weak for US and EEC, the latter two regions have more diversified structures and smaller standard deviations than Japan and UK.

Group P regions have strong export specialization in one or two categories of 0+1, 2+4, and 3, but they are divided into two groups according to how much they are industrialized in their export specialization. CAN, S .AFR, OTH ASIA, CHN MX, O W EUR constitute Group Pl whose export specialization in category 6+8 reached the level of unity and import specialization in the same category is weaker and even below unity. On the other hand LAT AM,

OTH AFR, W ASIA, and ANZ constitute Group P2 whose export specialization in category 6+8, Iet alone categories 5 and 7, is weak and they have to depend mainly on the export of primary commodities. Structures of export specialization are more concentrated and the standard deviations are greater in C;roup P2 regions.

Structures of import specialization both in Group I and Group P regions are contrary to those of their export specialization and are generally more diversified and have much smaller standard deviations than the latter. This may well be explained by the similarity of demand patterns in comparison with skewness of resource endowments among countries on 9ne hand,

1970] INTENSITY ANALYSIS OF WORLD TRADE FLOW 75 and protective commercial policies on the other.

Structures of export specialization of O EFTA and E EUR become similar to those of Group I but their relative shares do not deviate much from unity in any categories and are characterized by much diversified structures. On the contrary their import specializations are

close to those of Group P with low values in primary commodities and over unity values in manufactures. These intermediate characteristics result from the fact that the two regions are mixtures of various countries and constitute somewhat self-sufficient regions.

There appear in Table 2 distinct overtime changes in the structures of export and import specialization. For US, UK and EEC export specialization have weakened and import special-ization strengthened in all categories of manufactures. As these categories are increasing their shares in world trade, this tendency reflects that the exports from these countries are getting behind of the world export and their imports are growing faster than the world totals.

This implies in relative sense that these countries are losing their competitive power or com-parative advantage in all categories of manufactures (in all three broad categories at least) in

world market.=5 On the otherhand in categories 0+1 and 2+4 their export specialization increases and

import specialization declines. This trend in commodity categories of declining shares in world trade appears to reflect protectionistic agricultural policy and substitution of raw materials

by synthetic ones. These growing diversification in the structure of export and import special-

ization are represented clearly by distinct declines in standard deviations of exports and imports.

Japan, on the contrary, although weakening in the exports of categories 6+8, is gaining rapidly her share in categories 7 and 5, which reflects her increasing competitive power in world market. Import specialization in primary commodities and category 6+8 are increasing but those of categories 7 and 5 stay still, which is contrary to overtime changes in other Group I regions. O EFTA shows the similar changes to those of Japan and E EUR follows

the rests of Group I countries. For Group P regions export specialization increases in all categories of manufactures and

import specialization decreases in category 6+8. This seems to reflect the industrialization of

primary exporting regions, especially import substitution and export promotion in category 6+8 in them. (Such an advanced country in this group as Canada has followed the line of overtime change similar to Japan in which export specialization decreases in category 6+8 but it increases drastically in category 7.) There appear no distinct overtime changes in either

export or import specialization in primary commodities. Thus the tendency toward more diversified structures of export and import specialization are depicted for group P regions.

Standard deviations of export specialization (a) increase consecutively over the decade in

two regions, (b) show upward trends in one regions, (c) show downward trends in three, and (d) decrease consecutively in eight regions.*6 Similarly those of import specialization are (a)

*5 Increase in import specialization in manufactures may be attributed partly to the increase in mutual trade of differentiated manufactures among industrial exporters. It will not be ascertained until structures of trade specializations are studied in more disaggregated commodity classification.

*6 These four types of overtime changes are defined as follow. Let x*, x,, x* be values of x at consecutive periods 1, 2, and 3.

x*<x2<xs (a) xl<x3~x,} (b) x2~xl<x3 x2~x3<xl} (c) x3<xl~x2 x*<xz<x* (d)

76 HITOTSUBASHI JOURNAL OF ECONOMICS [February in one, (b) in none, (C) in five, and (d) in nine.

Degree of complementarity between a pair of regions is represented by correlation co-efiicients rij. Group I regions and Group P regions are distinguished from each other by the sign of rij; rij rs negative between a pair of regions within the sarne group but it is positive between a pair of regions belonging to different groups. O EFTA and E EUR has negative r,j with Group I regions and positive rij with Group P regions in their export trades, while they have the contrary structures of r,j in their import trades. Especially, UK and EEC of Group I and ANZ, OTH AFR, and S AFR of Group P show clear structures of export and import specialization characteristic of their own group, apart from their degrees of concentration, which reveal in high absolute values of rij both in export and import trades,

Since the structure of export sp~cialization and that of import specialization of a region are

negatrvely correlated m general (that rs r,,<0) there appears symmetncrty m the sign of ri, and rji; out of 630 pairs r,j and r,j have opposite signs in 72 pairs, most of which are associated with either O EFTA or E EUR whose correlation coefiicient within regions is positive (r,i>0).

The symmetricity is also depicted in overtime changes in r,j and rji. (See Table 7a.) Out of 105 pairs rij and r,i move in the same direction in 62- pairs (toward zero in 36 pairs and

depart from zero in 26 pairs). Symmetrical movement is significant at 5~ Ievel and the coefficient of association is 0.3684.

Degree of complementarity (C.J) is determined by the degree of match (r,j) and that of diversification (a(Si), a(Rj)) of the structures of export and import specialization. That is, the

essential structure of Cij is determined by that of rij, but the departure of Cij from unity is affected by standard deviations. Great departures of Cij from unity are depicted in such countries with relatively high absolute values of r*j and great standard deviations in export

or import specialization as UK, J among Group I and ANZ, OTH AFR, LAT AM, OTH ASIA, S AFR among Group P. On the contrary Cij are close to unity in such countries as OTH EFTA, EEC, US, E EUR with small riJ values and small standard deviations.

Although symmetricity between C*j and Cji results from that between r*j and rji, the symmetricity in their movements over the decade is not statistically significant (see Table 7b). In 60 out of 105 pairs both Cij and Cji move toward unity, but this is mainly because there exists a strong tendency for Cij to converge to unity. Out of 225 cases (a) Cij converges ccnsecutively to unity in 108 cases, (b) tends to converge to unity in 66 cases, (c) tends to diverge from unity in 34 cases, and (d) diverges consecutively from unity in 17 cases. The ratio of convergence ((a)+(b)) is 0.7733, which is statistically significant at 1~ Ievel, that is,

the null hypothesis that there exists no tendency for Cij to move in particular direction is rejected at the risk of l~ possibility of error.

The tendency of Cij toward unity is explained not only by the decline in both ,1(S) and ,7(R) but also by the tendecy of r,j toward zero (see Table 7c). The association between the two tendencies is significant at 1~ Ievel and the association coefficient is 0.8783. That is, the tendency for the degree of complementarity to move toward independence level is not only explained by the growing diversification in both export and import specialization of most regions, but also by the fact that patterns of trade specialization become similar between Group I regions and Group P regions. This tendency has been caused by the decrease of export specialization and increase of import specialization in manufactures (especially in category

6+8) on Group I regions' side and the contrary changes in the structure of trade specializa-tion in the same categories of Group P regions' side, that is, rapid industrialization of tradi-

1970] INTENSITY ANALYSIS OF WORLD TRADE FLOW 77 tionally primary goods exporters in category 6+8 and the corresponding decrease in com-petitive power of industrial exporters in the same commodity category.

VIII. Speclal Country Btases among Fifteen Regrons

An off-hand analysis of the figures in Table 5 reveals some characteristics of the structure

of B-'J .

Firstly, high values of B,j are depicted in such intra-regional trade as in LAT AM. EEC,

OTH EFTA, E EUR, OTH AFR, W ASIA, ANZ and trades between neighboring countries such as US and CAN, Regions in Europe (EEC, OTH EFTA, O W EUR and UK), OTH AFR and S AFR, and Asian regions (J, OTH ASIA and CHN MX). Secondly, trades bet-

ween an industrial exporter and its raw material suppliers are characterized by high B • such "! '

as between UK and its former dependents CAN, ANZ and S AFR, between EEC and OTH AFR, between US and LAT AM, between J and ANZ. OTH ASIA, W ASIA, in which there exists no high special country biases among raw material suppliers with each other. Thirdly, political relationships give rise to high special country biases as are shown between E EUR

and CHN MX, and between J and US. Overtime changes of special country biases give light to their determinants. Distinguished

increases in Bij are found in CAN-US trade, intra-ANZ trade, intra-LAT AM trade, intra-EEC trade, intra-EFTA trade, and intra-OTH AFR trade, all of which are explained by economic integration or other new trade agreements. Similarly increase in Bij in trades between US and ANZ, J and ANZ, CAN and J, and US's export trade with OTH ASIA and J's export trade with US seem to refiect the increasing intensit f ' tr -Pacific trade On y o m a . the contrary, Bij declines in trade between ANZ and UK, CAN and UK, EEC and OTH EFTA, which refiect the breakdown of the British Common Wealth and the establishment of EEC. An unambiguous trend is found in a rapid decline in Bij between E EUR and CHN MX, and a rapid increase of each's trade with the rest of the world; trades of CHN MX, with

ANZ, J, OTH ASIA, W ASIA, OTH AFR, and LAT AM, and trades of E EUR with OTH ASIA, LAT AM, OTH AFR, J, CAN, and UK.

Above observations suggest the symmetricity between BiJ j* and their movements. and B The conflict case in which erther B*J and Bji exceeds unity while the other falls short of unity occurs in less than one in seven and the symmetricity is significant at 1~ Ievel (the associa-tion coefacient is .9419. See Table 7d). Similarly, the symmetric movement of BiJ and Bji (both increase or both decrease) is significant at 1~ Ievel (the association coefficient is .5501.

See Table 7c). These statistical evidences support our hypothesis of the reciprocity in the effects of determinants of special country biases.

Is there any tendency for BiJ to move in any particular direction ? It does not make any sense to count the number of increase and that of decrease in Bij since they tend to be balanced under the equation (17). Convergence of Bij toward unity is depicted in 132 cases out

of 219, that is, 60.27~, which is large enough to reject at 1~ Ievel of significance the null

hypothesis that there exists no particular tendency of overtime changes. The convergence toward unity of special country biases seems to reflect the break up of traditional trade blocs

and the strengthening of new trade ties over the world.

78 HITOTSUBASHI JOURNAL OF ECONOMICS [February IX. Intensity of Trade among Fifteen Regions

The structure of lij is the structure of Cij multiplied by that of Bij. Namely the high complementarity between Group I regions and Group P regions and low complementarity within each group multiplied by high or low country biases resulting from geographical close-ness or other relationships produce the structure of intensity of trade among fifteen regions.

Each industrial exporter has high intensity of trade with particular Group P regions which is located in its neighbourhood or from which it obtains its raw materials, where high Cij and

high Bij reinforce each other (US-CAN, LAT AM; UK-S AFR, ANZ; EEC-OTH AFR). It has, on the contrary, very low intensity of trade with other industrial exporters which are located in distance, where low Cij and low Btj reinforce each other (J-EEC, UK). Low in-tensity of trade is common among Group P regions where low Cij and low Bij reinforce each other (ANZ-LAT AM-CAN-OTH AFR-S AFR-OTH ASIA). Trade intensity of about unity is obtained in such intermediate cases with high Cij ofiset by low Bij as between an industrial

exporter and primary exporter not closely associated to itself (J-S AFR, LAT AM; EEC-ANZ, S AFR). Such intra-regional trades as in LAT AM, EEC, OTH AFR, OTH ASIA. W ASIA, and ANZ are characterized by intensity of trade of more than unity resulting from low Cij dominated by high Bif'

It appears that the structure of B,j dominates that of lij, since the divergence of Cij from

unity is much smaller than that of Bij; most of Cij are included in the range between 0.5 and 2.0, whereas more than a half of J~ij (336/657=.5114) are located outside the range, and thus dominate lij. The conflict case of Bij and C,j being on the opposite side of unity occurs

in 275 out of 657 cases (that is, 41.85~), out of which in 34 cases (12.4~) Cij dominates lij. (Note that this is independent from the estimation bias of Cij and Bij since both of them tend to be either overestimated or underestimated to the same extent) On the other hand, overtime cbanges of Cij conflict with those of B*j in 215 cases out of 438 (e.i. 49,08~) and Cij dominates the overtime changes of lij m 52 cases (e.i. 24.18%). The dominance of Cij s small m determ nmg the value of I*j, but rt has been strengthened twrce m determining the

overtime changes of liJ. There appears a strong tendency for I*j to converge to unity; (a) in 86 out of 219 cases

lij converge consecutively to unity, (b) it tends to converge in 51 cases, (c) it tends to diverge

from unity in 37 cases, and (d) it tends to diverge consecutively from unity in 45 cases. The tendency of convergence to unity is significant at 1~ Ievel. However, its association with the tendency for Cij to converge to unity is not significant (See Table 7f). Thus the convergence of special country biases causes the intensity of trade to coverge to unity.

Above analyses are tentative and need to be supplemented by more rigorous analysis suggested in 111 and IV, but several propositions follow from them,

(1) Although the use of broad commodity classification have forced us to make even the structures of export and import specialization and to underestimate the degree of complemen-tarity, it has not missed the basic structure of international division of labour and its overtime

changes. C;j is dominant in determining the value of lij only in one of eight, while it is twice as dominant in determining their overtime changes.

(2) Factors of special country bias are more important in determining intensity of trade

among regions than have been expected in traditional theory of international trade. This implies the importance of such factors as space, preferential trade arrangements, and so on,

1970] INTENSITY ANALYSIS OF WORLD TRADE FLOW 79 in determining the pattern of trade in many country world, which encourages us to introduce

space factor into the theory of international trade,17) (3) The convergence both of C;~j and Bij toward unity reflects two important structural

changes of world trade flow; industrialization of primary goods exporters and the breakup of traditional trade relationships. Consequent convergence of lij toward unity will provide

valuable implications for policy purposes.

AppENDIX: EXTREME VALUES OF TIIE TRADE INDEXES

De~ree of aggregation in country groupings and commodity classification affect the values

of liJ, C,j. Sth and Rjh, through limiting their maximum values. li/ takes the minimum value of O when X.J=0 and the maximum value of the smaller of

(X../Xi.) and (X../X.j) since Xij cannot exceed Xt. and X.j (Xij~~Xa" Xtj~X.j).

( X. . X. . ) O;~1ij~min X~. ' X.j

That is, the larger is either exporting or importing country, the smaller the maximum values

Similarly, since Xi.'~ cannot exceeds Xi. and X.. (O~Xi.h;~Xi.. O~~X;.";~X.."), St ,*

limited by the smaller value of (X. ./Xi.) and (X../X.j).

o~ Sil~~ min( X. . X. . )

- - ~. ' X..h The larger is the exporting country's share, and the larger is the share of the commodity in the world trade, the smaller the index of export specialization is. Similar constraints are

imposed on the index of import specialization.

(X.. X.. ) O;~Rjh;~min X.j' X.,'*

Cij takes the minimum value of zero when either Sih or RJh or both of them are zero for all commodities (Sin.Rjh=0 for all h). The maximum value of Ctj depends upon the

maximum values of (;(Rj) and c(Si) since C., I = cov(St. Rj);~ ~/cr2(RJ) 02(St)

The maximum values of variance of export or import specialization are obtained when the trade specialization index of a commodity with the smallest share in world trade takes a posi-

tive value and those of the other commodity are all zero, ( X.. )

max 02(Si) = max a2(Rj) = min -1 X. .h

Thus the maximum value of cov(Rj, Si) is (, X.. )_

max cov (Si. RJ) = ~/max a2(Si) max 02(Rj) = min 1

X. ,h

Then the range of values of CiJ is ( X.. )

O;~Cij~~min X..'}

That is, it is not independent from the degree of aggregation in commodity classification.

IT Some factors of special country biases have been introduced explicitly into the theory of interna-tional trade by H. G. Johnson in his Wicksell Lecture, Comparative Cost and Com'nertial Policy Theory

for a Developing World Economy. Stockholm, 1968.

[February HITOTSUBASHI JOURNAL OF ECONOMICS 80

1955-7 60-2 65-7 INDEXES OF EXPORT SpECIALIZATION TABLE 2a

a(Si) 6+8 7 5 3

2+4 0+1

O. 4201 O. 3978 O. 3932

O. 4980 O. 5523 O. 4181

O. 9581 1. 0301 1. 0393

O. 4725 O. 4510 O. 4043

O. 7164 O. 6633 O. 5702

O. 3755 O. 3457 O. 3262

O. 7361 O. 6845 O. 6142

O. 2821 O. 1688 O. 1627

O. 6362 O. 7940 O. 7192

O. 9396 O. 9059 O. 8816

O. 9301 O. 7971 O. 7048

2. 1994 2. 4281 2. 5181

O. 7891 O. 8127 O. 7563

O. 8112 O. 7456 O. 7817

1. 1172 1. 1349 1. 1521

O. 2622 O. 2726 O. 2837

O. 8549 O. 7192 O. 6510

1. 3012 l. 1846 O. 9469

O. 2986 O. 2896 O. 3836

l. 5026 l. 3639 1. 2782

1. 4156 1. 2218 l. 1943

1. 3008 1. 2901 1. 2748

O. 7621 O. 8811 1 , 0440

O. 8453 O. 9755 O. 9676

1. O016 O. 9027 1. 1594

O. 6286 O. 5786 O. 6598

2. 4810 2. 1748 l. 8435

O. 1417 O. 1717 O. 1983

O. 6819 O. 8230 1. 0972

O. 8537 1. 7118 1. 2514

O. 2371 O. 2703 O. 3459

O. 1949 O. 2296 O. 2589

1. 8028 l. 5949 1. 5300

O. 3561 O. 3824 O. 8190

O. 0069 O. 0127 O. 0300

1. 3049 1. 3876 1. 3040

2. 0384 1. 9281 1. 6890

O. 9968 1. 0739 1. 0437

O. 2776 O. 3711 O. 4577

1. 4811 1. 2187 1. 2014

O. 2723 O. 1769 O. 2063

O. 0270 O. 0166 O. 0220

O. 9513 1. 0946 1. 3231

O. 0171 O. 0187 O. 0213

O. 0620 O. 0887 O. 1267

O. 0373 O. 0636 O. 0655

O. 0890 O. 0964 O. 1358

O. 0519 O. 0601 O. 0868

1. 3638 l. 4821 1. 3468

1. 0221 O. 5559 O. 5215

O. 1943 O. 2081 O. 2627

1. 6740 1. 5338 1. 5166

1. 5706 1. 5121 l. 4216

l. 3088 1. 2935 1. 2764

O. 2671 O. 3541 O. 5458

O. 7774 O. 8282 O. 7639

O. 8962 O. 6233 O. 5136

O. 1657 O. 1946 O. 1896

O. 8505 O. 7801 O. 9701

O. 0623 O. 0567 O. 1023

O. 2654 O. 2425 O. 2693

O. 5050 O. 3457 O, 5194

O. 1778 O. 2174 O. 3693

O. I142 O. 1024 O. 0966

O. 7268 O. 3819 O. 3681

O. 2366 O. 4019 O. 5294

2. 2193 2. 7007 2. 5009

O. 7175 O. 5862 O. 4572

O. 4371 O. 3606 O. 2863

O. 0454 O. 0812 O. 1215

O. 1396 O. 1938 O. 1689

l. 2670 1. 2356 1. 1664

O. 1496 O. 1608 O. 3907

O. 0671 O. 8325 2. 3535

O. 0274 O. 0421 O. 0344

7. 1101 8. 1786 8. 6933

O. 8030 O. 7164 O. 6829

O. 1111 O. 0951 O. 1447

O. 0819 O. 2379 O. 2905

O. 1832 O. 1586 O. 1283

O. 7204 O. 8922 O. 9494

1. 6425 1. 9949 l. 9968

1. 0239 l. 2773 l. 4834

O. 3523 O. 3593 O. 4040

O. 2092 O. 2371 O. 2487

1. 1850 1. 0335 l. 0311

1. 7649 1. 6541 1. 4982

O. 9094 O. 8424 O. 9660

2. 0587 2. 4768 2. 4433

2. 3943 2. 4417 2. 2096

O. 2642 O. 2409 O. 2008

O. 4206 O. 3386 O. 3299

2. 4193 2. 5359 2. 4616

2. 2617 1 . 8070 2. 0082

2. 8364 2. 8170 2. 9582

O. 1933 O. 1764 O. 1637

O. 7157 O. 9506 1. 0335

1. 0770 1. 1223 1. 0751

2. 4257 2. 4154 2. 6466

O. 6103 O. 5676 O. 6076

O. 3178 O. 3319 O. 4066

O. 9008 O. 8586 O. 8324

2. 0314 2. 1020 2. 0713

O. 7124 O. 8164 O. 7649

1. 2586 1. 5089 1. 4108

1 , 9388

2. OOOO 1 . 7290

O, 3526 O. 3754 O. 2355

O. 3451 O. 3049 O. 2997

1. 3459 1. 5009 1. 5602

1. 6316 1. 1416 1. 9575

1. 9764 2. 3784 2. 6479

WOR SHAR

US

CAN

LAT AM

EEC

UK

OTH EFTA

O W EUR

E EUR

S AFR

OTH AFR

J

W ASIA

OTH ASIA

CHN MX

ANZ

81 lNTENSITY ANALYSIS OF WORLD TRADE FLOW 1970]

INDEXES OF IMPORT SpECIALIZATION 1955-7, 60-2, 65-7 TABLE 2b

O. 4092 O. 3357 O. 2280

O. 4825 O. 3901 O. 4519

O. 5017 O. 4995 O. 3969

O. 2653 O. 1578 O. 1347

O. 6267 O. 5425 O. 4399

O. 5269 O. 3197 O. 2436

O. 2437 O. 2996 O. 2145

O. 2153 O. 1576 O. 1497

O. 4066 O. 5075 O. 5267

O. 4863 O. 3353 O. 3082

O. 9018 O. 8805 O. 8765

O. 3345 O. 2720 O. 2208

O. 2200 O. 1675 O. 1799

O. 6814 O. 1761 O. 3539

O, 4803 O. 4133 O, 4352

O. 2622 O. 2726 O. 2837

1. 1304 1. 1559 1. 2673

1. 1772 1. 0326 O. 9305

1. 0833 O. 9122 O. 8574

O. 8520 O. 9823 1. 0427

O. 6137 O. 7579 O. 8963

1. 4814 l. 1946 l. 1734

O. 9015 O. 8893 O. 9068

O. 8300 l. 0764 O. 9450

1. 1800 l. 3067 1. 0728

1. 4454 1. 2098 l. 0666

O. 3039 O. 2977 O. 4148

1. 3323 1. 1759 1. 1363

l. 2258 1. 0595 O. 9593

O. 9255 O. 7660 O. 8086

1. 4225 1. 2736 1. 0702

O, 1949 O, 2296 O. 2589

O, 2769 O. 4774 O. 7831

1, 7846 1. 6180 l. 7211

1 , 7673 1. 7312 1 , 4782

O. 6528 O. 7747 O, 7889

O, 2193 O. 3568 O, 4666

1, 4710 1. 3452 1, 1950

1. 3712 1, 4603 1. 2733

1. 1580 1, 1350 l, 1986

1. 5730 l, 6072 1, 7366

1. 5004 1, 2253 l. 3299

O. 3129 O, 4991 O, 3500

1. 1797 l, 2110 1, 1634

1. 0337 1. 0783 1. 0871

2. 2226 1. 1761 O, 9582

1, 5319 1. 3922 l, 5450

O. 0519 O. 0601 O. 0686

O. 5551 O. 5118 O. 4900

1. 0774 1. 0946 O. 8355

1. 7602 l. 7563 l. 7763

O. 8904 O. 9714 O. 9858

O. 5467 O. 6214 O. 7111

1. 6227 1. 3220 l. 2326

l. 4806 1. 4324 l. 3739

O. 7021 O. 7279 O. 8828

1. O116 l. 2640 l. 1699

1. 2836 1. 1976 1. 1529

O. 9468 l. 0163 O. 7658

1. 0575 1. 0368 O. 9856

1. 4309 1. 3463 1. 3275

1. 6057 1. 0378 1. 7142

1. 1807 l. 3574 1. 3646

O. 1142 O. 1024 O. 0966

O. 9512 1. 0984 O. 9606

O. 9502 O. 8609 O. 7708

O. 8119 O. 7245 O. 6695

l. 0652 1. 0395 1. 0799

O. 9599 1. 0397 1. 1619

O. 9044 O. 8596 O. 8462

1. 0754 O. 9136 O. 8893

O. 7954 O. 7172 O. 6866

O. 5933 O. 7144 O. 6128

O. 7477 O. 8107 O. 6489

O. 9699 1. 4716 2. 0185

l . 0268 O. 9631 O. 7960

O. 7465 O. 7426 O. 7213

O. 6011 O. 8820 O. 2715

O. 9515 1. 0222 O. 8084

O. 1832 O, 1586 O. 1283

l, 2886 1. 2711 1. 0292

O. 4570 O, 5101 O, 5742

O. 3701 O. 3924 O, 5249

1. 4530 l, 2795 1. 1917

l, 5536 l, 4376 1. 4049

O, 3748 O. 5966 O, 6289

O. 7577 O. 7045 O. 8754

l. 3591 1. 1253 1. 0717

O. 3650 O. 4808 O. 4283

O. 2456 O. 2815 O. 3631

2. 8001 2, 8651 2. 8876

O. 3566 O. 4106 O. 4970

O. 6883 O. 7392 O. 7220

O. 7095 O. 9807 1. 1902

O. 4343 O. 5062 O. 5274

O. 1933 O. 1764 O. 1637

1. 4270 l. 3046 1. 0937

O. 4915 O. 6341 O. 5179

O. 6169 O, 6321 O. 7326

1. I122 l, 0558 1. 0681

l. 9318 l , 9079 1. 7312

O, 3536 O, 5843 O. 6748

O, 8150 O. 7403 O. 7353

O, 9317 O, 8504 O. 9588

l, 0167 O, 2783 O. 3141

O. 6790 l, 0711 1. 0048

O. 9628 O, 7802 1. 0589

O. 9463 O. 9923 1. 0257

O. 9890 1, 0718 1, 1774

O. 2162 1. 2050 1, 3786

O. 4065 O, 3759 O, 3470

WOR SHAR

US

CAN

LAT Alvl

EEC

UK

OTH EFTA

O W EUR

E EUR

S AFR

OTH AFR

J

W ASIA

OTH ASIA

CHN MX

ANZ

[February HITOTSUBASHI JOURNAL OF ECONOMICS

INDICES OF TRADE INTENSITY TABLE 3

82

OTH EFTA O W EUR UK EEC LAT AM CAN US

O. 9502 l. 0448 O, 8958

O, 1067 O, 1645 O. 1673

O, 5979 O, 5174 O, 7362

O. 5161 O. 5422 O. 5104

O. 3063 O. 2800 O. 2097

O. 5242 O. 4082 O. 4300

O. 5647 O. 6943 O. 8292

1. 7611 l. 8620 1. 5814

O. 8532 O. 9455 O. 8300

O. 7303 O. 8378 O. 7400

O. 3629 O. 3401 O. 2608

O. 8261 O. 8238 O. 8147

2. 7942 2. 6890 2. 6975

O. 5325 O. 6057 O. 6738

l, 1516 1, 2015 2. OO21

4. 1528 4. 3911 4. 9228

l. 4381 l. 3327 1. 3113

2. 0961 l. 9451 1. 7756

1. 1858 1. 4384 l. 5929

2. 2970 2. 0577 l. 7521

1. 4282 l. 4287 1. 6212

1. 9601 1. 9888 2. 4124

O. 6610 O. 6616 O. 6801

l. 0199 O. 8333 O. 6921

1. 3781 O. 9473 1. 0727

O. 1864 O. 3209 O. 8257

1. 0384 l. 0188 1. 1559

O. 5678 O. 4459 O. 4393

O. 2227 O. 2422 O. 2918

1. 6449 1. 5690 1. 8095

2. 4664 2. 6216 2. 5659

O. 3664 O. 3473 O. 4177

1. 5757 l. 6143 l. 7599

O. 6756 O. 7534 O. 7657

1. 7126 l. 5278 l. 2538

O. 8418 O. 8430 O. 6900

O. 5897 O. 7301 O. 6519

O. 8242 O. 7269 O. 7003

O. 3385 O. 5533 O. 6140

O. 1970 O. 2290 O. 2262

l. 1503 1. 3455 O. 9643

O. 4145 O. 7918 O. 6971

O. 7017 O. 4992 O. 3781

O. 9620 l. 0899 O. 9785

O. 5855 O. 4911 O. 5575

O. 3009 O. 3522 O. 4280

O. 5066 O. 3805 O. 4129

3. 2916 3. 4962 4. 3876

2. 3541 2. 0759 1. 9753

O. 2895 O. 3801 O. 3472

1. 3452 1. 1832 l. 1069

O. 3122 O. 3150 O. 3284

O. 9538 O. 8241 O. 7338

O. 6668 O. 5922 O. 8796

O. 1642 O. 5117 O. 6685

O. 0812 O. 0909 O. 0864

O. 1856 O. 2383 O. 2736

O. 0704 O. 1113 O. 1220

O. 0231 O. 0293 O. 0633

O. 2824 O. 3626 O. 3106

O. 1101 O. 0727 O. 1682

O. 1716 O. 4361 O. 3485

O. 1780 O. 1791 O. 2039

O. 1888 O. 1840 O. 2764

O. 1412 O. 0919 O. 1201

l. 3097 1. 7083 l. 6989

1. 4245 1. 5299 l. 2393

O. 2199 O. 4477 O. 5101

2. 3531 2, 0138 l. 8052

O, 2213 O, 2192 O, 2293

1. 3980 1. 1895 l, 1673

O. 0426 O. 1174 O. 1063

O. 9006 l. 1361 O. 9549

O. 2150 O. 2321 O. 2954

O, 1208 O. 2405 O. 3585

O. 1236 O. 1150 O. 1849

O. 5119 O. 6602 O. 5904

4 2413 3 3900 3 1096

O. 6215 O. 4600 O. 4392

O. 2290 O. 2451 O. 3877

O. 9811 O. 7243 O. 5615

O. 2656 O. 2905 O. 2236

O. 0221 O. 5879 O. 8582

O. 0679 O. 1038 O. 2084

O. 1793 O. 5317 O. 2789

O. 2698 O. 3429 O. 3561

O. 0648 O. 0590 O. 1923

o. 3049 o. 3922 o. 3540

5. 0091 5. 0571 5. 1588

3. 7104 3. 4132 2. 7492

O. 5eo5 O. 6399 O. 6427

O. 6035 O. 8017 O. 9901

O. 6274 O. 5721 O. 6563

O. 7841 O. 7411 O. 7855

O. 0576 O. 0489 O. 0658

O. 6244 1. 1725 O. 8734

O. 8019 O. 7886 O. 6274

1. 8337 2. 4173 2. 5212

O. 6258 O, 6356 O. 4236

1. 2678 1. 4059 1. 5718

O. OOOO O. 0133 O. 0132

O. 6227 l. 0621 l. 0611

US

CAN

LAT AM

EEC

UK

OTH EFTA

O W EUR

E EUR

S AFR

OTH AFR

J

W ASIA

OTH ASIA

CHN MX

ANZ

83 lNTENSITY ANALYSIS OF WORLD TRADE FLOW 1970]

1955-57 60-62 65-67

ANZ CHN MX OTH ASIA W ASIA J

OTH AFR S AFR E EUR

O. 5780 1. 0673 1. 5380

O. 6715 l. OOOO O. 8991

O. 0562 O. 1088 O. 0560

O. 4316 O. 4268 O. 4259

5. 5047 4. 4358 3. 9019

O. 5993 O, 5835 O. 5379

O. 2397 O. 2167 O. 2878

O. 0289 O. 0336 O. 0432

O. 4243 O. 7155 O. 5305

O. 0236 O, 1681 O, 1731

O. 9198 1. 6440 2, 0445

2, 0180 1. 5416 1. 3690

l. 5603 1. 6065 1. 4946

O. 1747 O. 2794 O. 6522

2. 3003 2. 6471 2. 7599

O. OO11 O. OOOO O. OOOO

O. 0227 1. 2653 1. 1143

O. 0311 O. 7836 l. 2462

O. 2766 O. 3940 O. 5730

O. 9904 1. 4263 l. 6463

O. 2234 O. 2515 O. 3156

O. 0450 O. 0759 O. 1018

O. 7929 l. 0876 1. 2217

O. Il50 O. 1600 O. 1202

O. 1002 O. 1028 O. 1852

1. 9052 2. 3019 l. 9524

O. 2165 O, 3826 O. 5742

O, 5555 O, 2356 O, 4196

O, 3097 O, 1542 O. 2400

O. 7029 O. 4831 O. 4197

1. 6196 l. 4455 1. 1243

O. 4059 O. 3777 O. 3350

1. 1751 l. 0589 O. 9642

2. 2093 2. 0543 l. 9476

O. 8100 O. 7706 O. 7877

O. 9617 l. 1087 1. 0268

1. 1422 l. 0541 l. 1323

O. 2432 O. 2003 O. 2067

O. 3660 O. 6370 O. 6696

O. 1031 O. 1213 O. 1005

O. 1221 O. 1415 O. 1400

8. 5673 5. 2533 3. 2107

O. OOOO O. 2202 O. OOOO

O. 5279 O. 7533 O. 9620

O, 1047 O. 2644 O. 2594

O. 2086 O. 3287 O. 4133

O. 2359 O. 4338 O, 2400

O. 9020 l. 9470 O. 8155

O. 4645 O. 4752 O. 6165

O. 1735 O. 2862 O. 0898

O. 2582 O. 3113 O. 3218

O. 1808 O. 2455 O. 2522

O. 1640 O. 1388 O. 1962

2. 1895 l. 7817 l. 3634

l. 9670 l. 8415 l. 6618

O. 8155 O, 6901 O. 7840

1. 3847 O. 5016 2. 7577

O. 2072 O. 2503 O. 2700

l. 2446 1. 3365 1 3848

O. 4421 O. 4290 O. 3185

4. 4858 4. 2970 3. 9208

1. 0056 O. 9734 O. 9062

O. 8823 O. 7646 O. 6720

1. 6894 l. 4239 1. 3643

5. 1311 3. 6891 3. 1574

O. 0378 O. 2415 O. 3531

O. 7040 1. 7899 2. 2574

O. 4646 O. 4432 O. 6956

O, 4244 O. 5049 O. 7158

O. 3286 O. 4498 O, 7290

3. 7750 3. 3419 3. 8370

2. 1351 2. 4566 3. 8962

O. 9266 l. 0656 1. 6997

2, 5452 1. 4224 1. 6558

O. 8773 O. 9979 O. 8877

o. 3849 2. 7200 3. 5096

3. 9129 3. 7462 3. 5429

2. 2017 2. 6945 4. 4956

O. 8995 1. OI04 1. 2992

1. 0699 O. 9967 l. 0408

O. 1622 O. 2839 O. 9861

O. 3608 O. 5202 O. 7022

3. 1311 3. 0161 3. 3134

1. 9508 O, 7363 3, 0153

3. 7468 4, I138 3. 9729

O. 5404 O. OO15 O. 7923

O. 3867 O. 4968 l. 2963

O. 0099 O. 1667 O. 2204

l. 1218 1. 1055 l. 2001

O. 8299 O. 6946 O. 6476

O. 1020 O. 1606 O. 1468

O. 7329 O. 8253 O. 8762

3. 8888 3. 9648 4 3718

O. 0235 O. 0694 O. 0599

O. 0889 O. 1130 O. 2916

O. 2019 O. 4592 O. 6250

O. 7196 O. 7836 O. 7304

O. 1450 O. 1392 O. 2514

O. 0686 O. 0507 O. OOOO

O. 3127 O. 3254 O. 3351

O. 1855 O. 2492 O. 2902

O. 5712 O. 5110 O. 5120

1. 9243 l. 2874 1. 7662

7. 4529 6. 0834 6. 1659

O. 1802 O. 0660 O. 0159

1. 7200 l. 4477 l. 3051

1. 0740 1. 2313 1. 2485

l. 7978 l. 6205 l. 3006

O. 4757 O. 5063 O. 5789

O. 0542 O. 2218 O. 2518

O. 2361 O. 1871 O. 1808

O. 7094 O. 7915 O. 4961

O. 0248 O. 0709 O. OOOO

O. 2337 O. 6805 O. 9024

O. 2178 O. 4642 O. 5947

8. 2507 5. 6970 2. 5955

O. 2826 O. 2423 O. 2417

84

US

CAN

LAT

EEC

UK

OTH

oW

AM

EFTA

EUR

E EUR

S AFR

OTH

J

AFR

W ASIA

OTH

CHN

ANZ

ASIA

MX

Us

(o (o

(o

1

1

1

1

l

o o o

o o o

l O

1

1 1 1

o o o

1 1 1

1

1

1

o o

1

1 O

1

l 1

l l l

1 1 1

8368) 8890) 9344)

1451 1383 0295

2468 2167 0549

8938 8983 9798

7597 8079 9521

OI05 9795 0096

2410 1686 0711

9021 9635 9902

1871 188 1

0877

3002 238 1

0680

9741 9692 03 58

0121 1168 9831

2430 2142 0972

2659 1950 llO1

3042 2449 0563

HITOTSUBASHI JOURNAL OF ECONOMICS [February

TABLE 4. DEGREE OF COMPLEMENTARITY :

CAN LAT AM EEC UK OTH EFTA O W EUR

1. 1748 1. 1050 1. 1160

(o. 8485) (o. 8348) (o. 9227)

O. 6628 O. 7119 O. 6498

1, 1770 l, 1449 1. 1209

l 3294 l 2481 1 2237

1. O115 1. 0343 l . 0305

o. 7029 o. 7956 o. 8097

1. 1012 l. 0505 1. 0576

o. 7625 o. 7445 o. 7564

o. 6049 o, 6775 o. 6797

1. 2031 1 . 1356 l. 1694

o. 8933 o. 8420 o. 7648

O. 6647 O. 7171 O. 7263

O. 6614 O. 8108 O. 7143

O. 5484 O. 6415 O. 6374

l 1905 1 1595 l 120C

o 8477 o 7847 o 9052

O. 6697 O. 6486 O. 7161

1. 1929 1. 1821 l. 1247

l. 3408 1. 3088 l. 1944

1 0248 l 0507 l 0425

O. 7116 O. 7561 O. 8562

l. 0858 1. 0499 l. 0207

o. 7640 o. 6953 o. 7829

o. 6030 O. 6088 o. 7007

l. 1729 1. 1185 l. 1172

O. 7813 O. 7184 O. 6894

o. 6385 o. 6442 o. 7455

o. 6600 o. 7346 o. 7725

O 5500 O 5870 O 7170

o o o

1

1 1

1

1

l

o o o

o o

o o o

1

1 1

o o o

l l 1

1

1

l

o o o

l

1 1

1

1 l

l l 1

1

l 1

9178 9620 9608

0810 0747 0314

1410 0889 088 1

8917 9430 9631

8257 9071 9337

9929 9907 9928

1417 O008 0484

9630 9813 9867

1369 1108 08 1 l

2141 1274 0984

8644 9462 9570

0883 0492 0797

9-038 l 212 089 3

1854 0796 0834

2627 1427 1010

O. 8056 O. 9011 O. 9217

1. 1495 1. 1759 l. 0894

1 4949 1 4617 1 4125

o. 7522 o. 7922 o. 8532

(o. 5859) (o. 6746) (o. 7695)

O. 9681 O, 9400 O. 9483

1. 4186 1. 3338 1. 2237

o. 8855 o. 9338 o. 9506

l. 2709 1. 3169 l. 2292

1. 5202 1. 4508 1. 3324

O. 6742 O. 7601 O. 7830

l . 0538 l. 0907 1. 1782

1 3994 l 3488 1 2703

1. 4188 l. 1832 1. 2848

1. 6017 l. 5413 1. 4390

l

1 1

o o o

o O o

1

1 1

1

1

1

1

1

1

o o o

1

1 1

o o o

o o o

1 1

l

o o o

o o o

o o o

o o o

1377 0610 025 1

9048 8882 9362

6000 7227 7760

2392 1390 0946

3482 2023 1290

0493 0483 0406

6728 8203 8992

0608 03 28 0164

7857 8026 8785

5699 7205 8045

3534 1737 1504

8526 8501 8518

6463 7728 8544

6634 9019 8713

3711 6681 7437

1 l

o o o

o o o

1 l 1

1

l

1

O

1 1

o o o

1

1

l

o o o

o o o

l 1

l

1 O O

o o o

o o o

o o o

0911 09 66 0617

9084 8707 9622

8846 8036 8336

0661 0974 0597

1333 1767 0998

9869 0197 0190

8573 8463 9033

0498 0331 0180

8812 8210 8928

8149 7756 8536

O019 0415 05 60

0294 8961 8877

8450 7969 8727

8355 8262 8 665

8089 7646 8 448

85 INTENSITY ANALYSIS OF WORLD TRADE FLOW 1970]

1955-57 60-62 65-67

ANZ CHN MX OTH ASIA W ASIA J

OTH AFR S AFR E EUR

l. 1337 1. 0554 1. 0876

O. 8956 O. 8559 O. 9017

O. 6389 O. 6674 O. 5993

l. 2070 1, 1704 1, 1495

1, 3224 l. 2430 1. 2338

l. 0333 1. 0454 l. 0508

O. 6990 O. 7462 O. 7837

l. 0715 1. 0472 1. 0459

O. 7891 O. 7369 O. 7589

O. 6031 O. 6386 O. 6586

1. 3135 1. 2174 1. 2097

O. 8956 O. 9743 O. 7967

O. 8749 O. 7123 O. 7178

O. 6810 O. 8738 O. 7135

O. 5160 O. 5515 O. 5675

l. 2760 1 . 0490

l. 0755

O. 8303 O. 9709 l. 0434

O. 4787 1. 0333 l. 0160

1. 0293 1. 0363 l. 0435

l, 1904 O. 9850 1, 0062

l . 4285 1. OI04 O, 9978

1. 0517 O. 9911 1. 0503

O. 9994 O. 9588 O. 9651

O. 8942 O. 9159 O. 9589

l. 0928 1. 0559 1 0397

l. 0516 1. 0289 l. 0244

O. 9249 O. 8870 O. 9361

O. 8879 O. 8821 O. 8821

O. 8287 O. 9260 O. 8891

l . 1071 1. 0709 1. 0515

1. 0420 l. 0332 1. 0217

1. 1360 l. 0994 l. 0706

l. 1728 l. 1314 l. 0963

1. 0033 1. 0186 1. 0268

1. 2452 l. 2960 l. 2428

l. 2567 1. 3581 l. 5583

O. 6557 O. 6935 O. 6928

1. 1309 1. 0441 1. 0679

O. 9011 O. 8731 O. 9131

O. 6824 O. 8521 O. 8087

O. 6516 1. 0283 l, 1182

1, 1469 l, OOIO O. 9609

O, 7665 O, 9962 1. 0796

O. 9439 O. 9815 1. 0127

O. 9871 O. 9980 O. 9896

O. 9567 O. 9433 O. 9462

O. 8483 O. 9028 O. 9680

1. 0171 l. 0231 1. 0054

O. 8443 O. 8190 O. 8937

O. 5295 O. 6206 O. 5870

l. 0096 O. 9073 O. 8621

1. 3900 1 , 2063

1, 1406

1 , 2202 1, 1228 1, 1034

1. 3261 l. 1577 1. 1521

1. 0474 1. 0361 1. 0389

O. 5751 1. 0193 O. 9688

1. 1363 O. 9198 O. 9503

O. 6041 O. 9011 O. 4089

O. 8984 O. 9163 O. 9053

1. 1556 l. 0649 l. 0291

O 7793 O. 7786 O. 7635

O. 7541 O, 7948 O, 8446

1, 2173 1. 1359 l, 1137

O. 9819 O. 9463 O. 8172

O. 9357 O. 9513 O. 9812

l. 4477 1. 4897 l. 4085

l. 6568 1. 5725 1. 6674

O. 7460 O. 9079 O. 9381

l. 0482 l. 0189 1. OllO

O. 7898 O. 8017 O. 8332

O. 6405 O. 9896 1. 0610

(O. 6355) (O. 9289) (1. 1170)

O. 5680 1. 0617 l. 2138

O. 8825 O. 9153 O. 9355

O. 9263 O. 9734 O. 9801

O. 8501 O. 9180 O. 9710

O. 7657 O. 8025 O. 8819

O. 7941 O. 9151 O. 9265

O. 6697 O. 7506 O. 8333

(O. 5020) (O. 5173) (O. 5100)

1. 0835 1. 4485 1. 9000

1. 6520 l. 5593 1. 4709

O. 6253 O. 7670 O. 7574

1. 3359 1. 1667 1, 1407

O, 7347 O. 8181 O, 6851

1 5717 1 1604 l 2926

1 8654 1 6903 l 6848

O. 6486 O. 7704 O. 8113

O. 6938 O. 9118 O. 8673

O. 5219 O. 7272 O. 7725

1, 1273 l, 1003 1, 1214

O. 9033 O. 8423 O, 8949

O. 8006 O. 5375 O. 5172

1. 1454 l, 2129 l, 1778

1, 2519 1, 3227 1, 2887

1. 0173 l. 0217 l. 0483

1 . 0255 l. 0258 1. 0182

O. 9674 O. 8865 O. 9136

l. 0365 1. 0813 l. 0666

O. 8620 O. 7215 O. 7665

1 . 0435

l. 0551 l. 0576

O. 9457 1. 0215 1. 0207

O. 9748 l. 0400 1. 0476

1. 0175 1. 0345 1. 0237

(O. 8359J (O. 7092) (O. 7171)

O. 7575 O. 5729 O. 5758

l. 1931 1. 2780 1. 2593

1. 0565 O. 9959 O. 9976

1. 0167 l. 0039 l. 0090

1. 0652 1. 0188 O. 9896

O. 6305 O. 712O O. 6264

1, 1006 O. 9809 O, 9268

O. 9177 1. 0601 1. 0307

O. 8502 O. 7650 O, 7351

O. 7249 O. 6641 O. 6687

O. 7818 O. 8605 O. 6719

O. 6939 O. 4977 O. 5122

l. 0900 1. 0059 O. 9788

1 . 0799

l. 0350 O. 9807

1. 1480 O. 9915 O. 9993

[February l{ITOTSUBASI-II JOURNAL OF ECONOMICS

CORRELATION COEFFICIENTS OF EXPORT TABLE 5

86

OTH EFTA O W EUR UK EEC LAT AM CAN US

O. 8899 O. 8114 O. 7316

-O. 7552 -O. 7817 -O. 4218

-O. 4941 -O. 6364 -O. 7461

O. 6220 O. 4798 O '2623

- . 3631 -O. 6330 -O. 6269

-O. 7924

- . 8418 -O. 8848

-O, 7381 -O. 4588 -O. 4525

O. 4789 O. 5870 O. 4863

O. 8240 O. 8262 O. 9020

-O. 7373 -O. 6059 -O. 7412

O. 6136 O. 8575 O. 5580

O. 5547 O. 5473 O. 6290

O. 9036 O. 8033 O. 7688

-O. 6096 -O. 7803 -O. 5711

- . 6870 -O. 6829 -O. 6881

O. 8623 O. 6770 O. 6528

-O. 6306 -O. 7668 -O. 4091

-O. 7293 -O. 7168 -O. 7456

O. 5743 O. 7207 O. 6888

O. 7633 O. 8894 O, 8156

-O. 1440 O, 1904 O. 2726

O 9608 O, 9641 O 9605

O. 9226 O. 9537 O. 9289

O. 2492 O. 4375 O. 5111

-O, 8366 -O, 8496 -O 8255

-O. 9223 -O, 9042 -O. 9187

-O. 1359 -O. 3199 -O. 3608

-O. 8638 -O. 8019 -O. 6786

-O, 9170 -O. 8882 -O. 8631

-O. 0715 -O, 1712 -O. 1639

O. 8135 O. 8086 O. 7727

O. 9483 O. 9319 O. 8591

O. 1316 O. 2935 O. 3281

O. 7765 O. 8234 O. 6621

O. 9528 O. 9588 O. 8682

O. 0635 O. 2546 O. 2074

-O. 7951 -O. 7495 -O. 7336

O. 7249 O. 6544 O. 5152

-O. 7665 -O. 7523 -O. 6947

-O. 8435 -O. 8213 -O. 6740

O. 4093 O. 6089 O. 4151

-O. 6392 -O. 7777 -O. 6935

o. 9073 o. 8989 o. 8279

-o. 6476 -o. 7232 -o. 6909

o. 6794 o. 7357 o. 7243

O. 7258 O. 6556 O. 5847

-O. 4950 -O. 7037 -O. 6083

O. 8113 O. 8844 O. 8374

-O. 7807 -O. 7135 -O. 5900

O. 6061 O. 5923 O. 3205

-O. 7393 -O. 7682 -O. 7605

-o. 8363 -o. 7656 -o. 6857

o. 7437 o. 7680 o. 7834

-o, 7738

- . 8247 -o. 7494

-O, 8083 -O. 8267 -O, 7739

O. 0085 O. 1738 O. 3705

O. 0549 -O. 1428 -O. 2078

-O. 8686 -O. 9650 -O. 9104

O, 7212 O. 6815 O. 8761

-O. 1272 -O, 1931 -O. 2416

O 8832 O, 9173 O, 8570

-O, 5590 -O. 5548 -O. 6998

O. 0390 O. 0688 O. 1608

O. 8590 O. 8916 O. 8286

-O. 5498 -O. 4281 -O. 4533

O. 1514 O. 1284 O. 2349

-O. 8421 -O. 8645 -O. 8554

O. 3705 O. 2976 O. 4192

-O. 1982 -O. 2321 -O. 3107

- . 8714 -O. 9124

- . 8039

O. 4526 O. 4359 O. 5319

-O. 1005 -O. 1668 -O. 2066

-O. 8061 -O. 8342 -O. 7848

-O. 8318

- . 7779 -O. 7963

-O. 7017 -O. 6924 -O. 6277

- . 8507 -O. 8744 -O. 7906

-O. 7875 -O. 4117 -O. 6760

- . 8985 -O. 9146 -O. 9131

O. 8076 O. 7910 O. 8122

O. 8237 O. 4529 O. 8281

O. 8529 O. 8792 O. 8661

O. g736 O. 9451 O. 8769

O. 8617 O. 6772 O. 7923

O. 8863 O. 7967 O. 6505

-O. 9129 -O. 8764 -O. 8477

-O. 8354 -O. 7127 -O. 7333

-O, 8027 -O. 7285 -O. 6188

-O, 8805 -O. 8921 -O. 8009

-O. 8649 -O. 6506 -O. 8087

- . 8376 -O. 8097 -O, 6965

-O. 9496 -O. 8314 -O. 7320

O. 7122 O. 7460 O. 3101

O. 6294 O. 6266 O. 2316

-O. 5493 -O. 6717 -O. 2195

- . 8199 -O. 8629 -O. 3685

O. 0684 -O. 1773

O. 1290

O. 8001 O. 7338 O. 5079

-O. 848~ -O. 6441 -O. 2658

O. 7187 O. 7056 O. 5348

O. 7806 O. 7830 O. 3385

-O. 0682 -O. 1152

O. 2229

O. 0134 O. 1433

-O. 0295

O. 7526 O. 7851 O. 5639

O. 8012 O. 7791 O. 6176

O. 6654 O. 6429 O. 2144

US

CAN

LAT AM

EEC

UK

OTH EFTA

O W EUR

E EUR

S AFR

OTH AFR

J

W ASIA

OTH ASIA

CHN MX

ANZ

87 1NTENSITY ANALYSIS OF WORLD TRADE FLOW 1970]

1955-57 60-62 65-67 AND IMPORT SpECIALIZATION

ANZ OTH ASIA CHN MX W ASIA J

OTH AFR S AFR E EUR

O. 6625 O. 5375 O. 5123

-O. 4364 -O. 6315 -O. 5404

-O. 7846 -O. 7811

- . 8858

O. 9122 O. 9139 O. 8496

O. 9371 O, 8863 O, 9423

O 1851 O 3179 O 3580

-O. 8513 -O. 8971 -O. 8093

O. 5281 O. 6775 O. 6489

-O. 6904 -O. 8016 -O. 7703

-O. 8795 -O. 9651 -O. 8897

O. 7019 O. 6599 O. 6836

-O. 0988 -O. 0256 -O. 1855

-O. 8578 -O. 8563 -O. 8574

-O. 8187 -O. 4094 -O. 8422

-O. 9020

- . 9560 -O. 8626

O. 9641 O. 7005 O. 5428

-O. 5001 -O. 2994

O. 2938

-O. 7983 O. 1839 O. 0436

O. 3170 O. 5449 O. 6159

-O. 0058 -O. 4451 -O. 4644

-O. 5019 -O. 4874 -O. 2201

O. 3674 O. 2672 O. 2812

-O. 4509 -O. 7519 -O. 6927

-O, 3499 -O. 4208 -O, 5139

-O. 4520 -O. 2114 -O. 3218

O. 5459 O. 6087 O. 6626

o. 6406 o. 3307 o. 5602

-o. 4083 -o, 6853 -o, 6745

O. 5915 -O. 1887

O. 0438

O. 8777 O. 0892

-O. O110

O. 2023 -O. 1469

O. 4361

O. 8930 O. 7401 O. 5462

O. 6795 O. 6379 O. 5025

O. 5084 O. 5730 O, 3696

O. 8604 O. 8108 O. 7907

O, 7213 O. 7281 O 7645

O 0264 O 1985 O 3718

O. 2971 O. 3948 O. 6128

-O. 8078 -O. 7719 -O. 8670

-O. 7282 -O. 6496

- . 8263

-O. 6816 -O. 4282 -O. 5973

O. 7455 O. 4974 O. 5865

-O. 4776 -O. 5628 -O. 4774

-O. 5117

- . 8846 -O. 8820

O. 1914 O. 3471 O. 8206

O. 2377 O. 2967 O. 2919

-O. 1581 -O. 6991 -O. 5551

-O. 6945 O. 2347 O. 5439

O. 7641 O. 0336

-O. 6787

-O. 5386 -O. 0277

O. 3128

-O 3468 -O 1615

O I155

-O 2083 -O 0731 -O 3581

-O. 6161 -O, 5220 -O, 2362

O 1821 O, 5037 O, 1508

O 0285 -O 3045 -O 4824

O 5874 O 3423 O 2612

O 9583 O 8122 O 8301

O 9361 O 7092 O 8654

O 2599 O 3115 O 3873

O. 7567 O. 9301 O. 8351

O. 8648 O. 9587 O. 9615

O. 2393 O. 4635 O. 3877

- . 5335 O. 3027 O. 3422

-O. 1634 O. 4398 O. 5584

O. 2165 O. 6336 O. 4859

-O. 3099 -O. 4264 -O. 4159

-O, 6635 -O. 1209 -O, 1001

O, 2151 -O. 5717 -O. 1992

-O. 2638 -O. 2313 -O. 6632

-O. 4917 -o. 5518 -O. 5974

O. 7603 O. 48eo O. 2300

-O. 4562

- . 5443

- . 5221

-O. 7318

- . 8378 -O. 6693

-O. 7822

- . 8325 -O. 7979

O. 6986 O. 6266 O. 7304

-O. 2527 -O. 3280 -O. 1322

O. 7803 O. 7003 O. 6480

-O. 7094 -O. 4012 -O. 3272

O. 3517 O. 3349 O. 2204

-O. 4612

- . 8018

- . 7219

o. 3792 o. 6442 o. 6724

-O. 0247 -O. 0813 -O. 3286

O. 7750 O. 7177 O. 8636

-O. 5937 -O. 6876

- . 7932

O. 0421 O. 2098 O. 4077

-O. 6794 -O. 7447 -O. 7526

-O. 8199 -O. 7669 -O. 8929

O. 7427 O. 6236 O. 6476

-O 6794 -O 7216 -O 7469

-O 6348 -O 9290 -O 9135

O 5105 O 6872 O 6985

O 3563 -O 0382 -O 0261

O 2758 O 1492 O 3690

O 4761 O 1505

-O 0967

-O. 6682 -O. 0727

O. 2280

-O. 6594 -O. 5416

O. 4230

-O. 5674 O. 3088 O. 5243

-O. 6768

- . 6223 -O. 4745

-O. 4129 -O. 2133 -O. 1415

-O. 6099 -O, 4311 -O, 1399

- . 8874

- . 8934 -O. 7074

-O. 7589 -O. 4186 -O. 4257

-O. 8837 -O. 8078 -O. 6550

O. 9161 O. 7816 O. 7102

O. 7813 O. 2443 O. 4271

O. 8589 O. 6907 O. 6780

-O. 2480

- . 2233

- . 4057

-O. 9158

- . 8422 -O. 8093

-O. 7763 -O. 3527 -O. 5509

-O. 4131 -O. 2337 -O. 2817

O, 4973 -O 1338 -O 5544

-O 4110 O 4786 O 2910

-O 3136 -O 6138 -O 7023

-O 8799 -O 7164 -O 6405

-O. 8573 -O. 8144 -O. 8317

-O. 6614 -O. 3686 -O. 7969

-O. 6783 -O, 8722 -O, 8039

O. 5299 O. 0467

-O. 1877

O. 4577 O. 2979

-O. 1656

O. 6154 -O. 0480

- . 0044

[February HITOTSUBASHI JOURNAL OF ECONOMICS

DEGREE OF SpECIAL COUNTRY BIAS TABLE 6

88

OTH EFTA O W EUR UK EEC LAT AM CAN US

O, 8708 O. 9527 O. 8437

O. 1175 O, 1889 O. 1739

O. 6759 O. 6439 O. 8832

O. 4536 O. 5110 O. 4979

O. 3386 O. 3152 O. 2240

O. 8737 O. 5649 O. 5542

O. 7010 O. 7706 O. 8996

l. 5320 1. 5835 1. 4515

O. 5707 O. 6468 O. 5876

O. 7957 O. 8710 O. 7702

O. 33571 O. 3164 O. 2528

O. 7240 O. 7565 O. 7487

2. 3470 2. 3189 2. 4084

O. 6282 O. 7718 O. 7444

1. 7196 1. 8527 2. 7959

3. 5347 3. 9737 4. 4111

1. 3489 l. 9-144

1. 2373

1. 8495 1. 6529 1. 6144

1. 2016 1. 4105 1. 5631

1. 8535 1. 8065 1. 6006

l. 0593 1. 1883 1. 4358

1. 8680 1. 8970 2. 3182

O. 8788 O. 8352 O. 7971

1. 1897 O. 9846 O. 7662

1. 3126 O. 9169 1. 0538

O. 2115 O. 3909 O. 9249

1. 5434 1. 2423 1. 2856

O. 5352 O. 4317 O. 4322

O. 2835 O. 3025 O. 3322

1 6993 l 6692 1 9083

1, 7385 l, 9654 2. 0967

O. 4137 O. 3719 O. 4395

1. 7671 1. 7120 1. 8275

O. 8182 O. 8306 O. 8202

1. 7249 l. 5422 1. 2629

O. 7056 O. 7131 O. 6135

O. 4398 O. 5578 O. 5458

O. 8042 O. 6918 O. 6717

O. 5108 O. 7772 O. 9450

O. 1673 O. 2000 O. 2018

O. 8653 1. 0780 O. 7880

O. 5088 1. 0209 O. 8166

O. 7003 O. 4793 O. 3580

O. 9345 1. 2164 1. 10?_4

1. 0274 O. 6816 O. 6931

O. 2223 O. 3000 O. 3720

O. 5943 O. 4476 O. 4848

2. 5899 2. 6547 3. 5694

1. 5485 1. 4308 l. 4824

O. 4295 O. 5001 O. 4434

1. 1782 1. 1050 1. 0558

O. 3242 O. 3210 O. 3328

O. 8389 O. 7418 O. 6787

O. 9371 O, 7833 1. 0275

O, 1512 O, 4874 O, 6549

O. 1062 O. 1308 O. 1103

O. 1834 O. 2304 O. 2655

O. 1002 O. 1399 O. 1507

O. 0210 O. 0279 O. 0599

O. 3343 O. 4551 O. 3559

O. 1317 O. 0880 O. 1941

O. 2121 O. 5705 O. 4125

O. 2754 O. 2317 O. 2386

O. 2846 O. 2040 O. 3172

O. 2999 O. 1375 O. 1615

1, 2427 1, 5662 1. 4418

1. 0179 1, 1342 O. 9756

O, 1550 O, 3784 O. 3970

1. 9372 1. 7861 1. 6434

O. 2561 O. 2317 O. 2396

1. 2845 1. 1337 1. 0811

O. 0706 O. 1929 O. 1518

O. 7678 1. 0157 O. 8547

O. 2752 O. 3230 O. 4285

O. 1585 O. 3230 O. 4739

O. 2044 O. 1698 O. 2720

O. 4254 O. 5814 O. 5048

2. 6480 2, 1993 2. 1609

O. 5162 O. 4103 O. 4031

O. 1932 O. 2270 O. 3578

O. 7769 O. 6339 O. 5100

O. 4161 O. 4510 O. 2999

O. 0335 O. 8003 1. I111

O. 1235 O. 1769 O. 2907

O. 2008 O. 6315 O. 3647

O. 4059 O. 4782 O. 4903

O. 0980 O. 0728 O. 2692

O, 5561 O. 6115 O, 5555

4. 3741 4. 4424 5. OI06

2. 9759 2. 8053 2. 6001

O. 6272 O. 7124 O. 6560

O. 7945 O. 9924 1. 0399

O. 6209 O. 5841 O. 6500

O. 6318 O. 6399 O. 7333

O. 0639 O. 0508 O. 0664

O. 5260 O. 9869 O. 8029

O. 6168 O. 6369 O. 5874

1. 8826 2. 4943 2. 4340

O. 6183 O. 5691 O. 4310

1. 0199 1. 1579 1. 4325

O. OOOO O. O117 O. O119

O. 4774 O. 8531 l. 0045

US

CAN

LAT AM

EEC

UK

OTH EFTA

O W EUR

E EUR

S AFR

OTH AFR

J

W ASIA

OTH ASIA

CHN MX

ANZ

89 lNTENSITY ANALYSIS OF WORLD TRADE FLOW 1970]

1955-57 60-62 65-67

ANZ CHN MX OTH ASIA W ASIA J

OTH AFR S AFR E EUR

O. 5099 l. O112 1. 4140

O, 7498 1. 1755 O. 9972

O. 0880 O, 1630 O. 0935

O. 3576 O. 3646 O. 3705

4. 1624 3. 5685 3. 1623

O. 5800 O. 5582 O. 5119

O. 3430 O. 2904 O. 3673

O. 0270 O. 0321 O. 0413

O. 5378 O. 9711 O. 6991

O. 0391 O. 2633 O. 2628

O. 7002 1. 3503 1. 6900

2. 2534 1. 5822 1. 7186

2. 3121 2. 2554 2. 0824

O. 2566 O. 3198 O. 9142

4. 4585 4. 8001 4. 8639

o, o009 o, oooo o. oooo

O. 0273 l. 3032 1. 0679

O. 0650 O. 7583 l. 2265

O. 2323 O. 4000 O. 5694

O. 9622 l. 3763 O. 5775

O. 2235 O. 2623 O. 3271

O. 0503 O. 0828 O. 1062

O. 7539 1. 0570 l. 1925

O. 1244 O. 1804 O, 1285

O. I129 O, 1165 O. 2100

2. 2991 2. 4860 2. 1960

O. 1516 O. 3786 O. 5754

O. 5282 O. 2377 O. 3995

O. 4753 O. 1501 O. 2146

O. 6432 O. 4575 O. 4037

1 . 4629 1. 3498 1. 0691

O. 3896 O. 3656 O. 3279

1. 0344 O. 9631 O. 9006

l. 8836 1. 8157 1. 7765

O. 8074 O. 7564 O. 7672

O. 7723 O. 8554 O. 8262

O. 9088 O. 7761 O. 7266

O. 3709 O. 3754 O. 2984

O. 3236 O. 6101 O. 6270

O. 1144 O. 1389 O. 1101

O. 1790 O. 1661 O. 1732

7. 4697 5. 2480 3. 3414

O. OOOO O. 2210 O. OOOO

O. 9179 O. 7390 O. 9930

O. 1110 O. 2693 O. 2561

O, 2114 O. 3293 O. 4176

O. 2466 O. 4599 O. 2536

l. 0634 l. 1599 O. 8425

O. 4566 O. 4644 O. 6132

O. 2055 O, 3494 O. 1005

O. 4877 O. 5017 O. 5483

O. 1791 O. Z706 O. 2926

O. Il80 O. 1151 O. 1720

l. 7943 l. 5868 1. 2355

l. 4833 1. 5905 1. 4424

O. 7786 O. 6660 O. 7546

l. 2185 O. 5454 2. 9019

O. 3428 O. 2778 O. 6005

1. 9429 1. 3506 l. 3051

O. 4922 O. 4682 O. 3518

3. 8817 4. 0350 3. 8097

l. 2905 l. 2503 l. 1870

1. 1700 O. 9620 O. 7956

1. 3877 1. 2535 1. 2250

5. 2261 3. 8985 3. 8637

O. 0404 O. 2539 O. 3599

O. 4862 l. 2015 1. 6026

O. 2804 O. 2818 O. 4171

O. 5690 O. 5564 O. 7631

O. 3135 O. 4414 O. 7210

4. 7799 4. 1687 4. 6055

1. 9705 1. 6958 2. 0506

1. 4819 1. 3894 2. 2444

1, 9051 l, 2191 l. 4515

1, 1942 1. 2198 l. 2958

O. 6776 2. 5618 2. 8914

4. 4340 4. 0931 3. 7874

2. 3769 2. 7683 4. 5869

1. 0581 l. 1006 l . 3380

l. 3973 l. 2421 1. 1803

O. 2043 O. 3103 1. 0644

O. 5387 O. 6931 O. 8426

1. 8952 l. 9342 2. 2526

1. 2412 O. 6345 2. 3326

2. 0085 2. 4337 2. 3580

O. 8333 O. 7808 O. 9766

O. 5575 O. 5449 l. 4948

O. 0190 O. 2293 O. 2854

O. 9950 1. 0046 l. 0701

O. 9188 O. 8248 O. 7237

O. 1274 O. 2989 O. 2840

O. 6398 O. 6804 O, 7439

3, 1062 2, 9974 3, 3921

O. 0231 O. 6797 O. 0571

O. 0867 O. IIO1 O. 2863

O. 2087 O. 5180 O. 6841

O. 6942 O. 7246 O. 6848

O. 1683 O. 1930 O. 3280

O. 0657 O. 0481 O. OOOO

O. 3307 O. 3185 O. 3283

O. 1903 O. 2383 O. 2770

O. 5613 O, 4939 O, 5001

1. 8214 1. 2927 1. 7705

7. 3301 6. 0594 6. 1109

O. 1692 O. 0648 O. 0161

2. 2709 2. 5273 2. 2668

O. 9001 O. 9635 O. 9914

2. 8516 2. 2763 2. 0764

O. 4322 O. 5162 O. 6246

O. 0591 O. 2093 O. 2443

O. 2777 O. 2447 O. 2459

O. 9786 1. 1920 O. 7420

O. 0189 O. 0824 O. OOOO

O. 3368 l, 3675 1, 7621

O, 1998 O, 4614 O, 6077

7. 6397 5. 5042 2. 6468

O. 2462 O. 2444 O. 2419

90

(a)

TABLE 7. Symmetric

HITOTSUBASHI JOURNAL OF ECONOMICS

TESTS OF INDEPENDENCE Movements of rij and rji.

IN Two WAY

(b) Symmetric Movements of

(c) Convergence of

Cij and

Cij and rtj.

Cji.

(d) Symmetricity between Bij and

(e) Symmetric Movements of Bij

Bfl .

and

(f) Convergence of Cij and lij.

Bji.

CONTINGENCY TABLES

u = 3.5385 (significant at 5%

Q = 0.3684

u = 0.0046 (not significant)

u = 48.9792

(significant at 1% Q = O. 8783

u = 170.03

(significant at 1% Q = O.9419

u = 8. 7577

(significant at l% Q = O. 5501

u=01.3698 (not significant)

Note: The variable u is defined as 2 2

_ ~~~~::M.j!M)] ~~ u - t. ' M. j/M ~

i=1 j=1 and is approximately chi-square distributed of degree of freedom one for large value of n the null hypothesis that two classifications are independent from each other. The variable the coefficient of association between the two classifications

MllM22-Ml2M21 Q= Ml IM22 + Ml 2M21 For more deta]Is see Mood, A.M. Introduction to the Theory of Statislics 1950, Chap. 12.

level)

level)

level)

level)

under Q is