intra-industry trade and inter-industry specialization as concurrent sources of international trade...

22
Intra-lndustry Trade and Inter-Industry Specialization as Concurrent Sources of International Trade in Manufactures By Antonio Aquino Contents: I. Introduction and Summary. -- 11. Notion and Sources of Intra-Industry Trade. -- III. Measurement of Intra-Industry Trade. -- IV. Empirical Performance of Different Indices of Intra-Industry Trade. -- V. A Heterodox Relationship between Export and Import Patterns, the Case of Trade in Manufactures in 1972. -- VI. Inter-Industry Specialization in Manufactures over the Period 1951-1974. I. Introduction and Summary I n pure theory international specialization is typically identified as the only source of international trade, and practically the entire body of normative trade theory is based upon this belief. It is however fairly evident, from an even superficial look at trade figures, that very substan- tial trade flows are not related to any meaningful notion of specialization. The most striking examples are provided by trade in passenger motor cars: France exports Renault to Germany, Germany exports Volkswagen to France, and Italy exports Fiat to France and Germany in exchange for Renault and Volkswagen. It does not seem however to make much sense to say that France specializes in Renault, Germany in Volkswagen and Italy in Fiat. The distinction between trade which is the outcome of a genuine proc- ess of specialization (inter-industry trade) and trade which consists in the exchange of almost identical commodities (intra-industry trade) is important for a variety of reasons. First, the pattern and intensity of intra-industry trade, being connected to not yet sufficiently explored structures of consumer preferences, rather than to an interaction between commodity characteristics and country attributes, are much more difficult Renuzrk : This paper is based on Chapter 8 ot a thesis approved for the Ph. D. in economics at the London School of Economics in May 1977. I wish to thank Alasdair Smith for his invaluable help while carrying out the research. The research could be carried out thanks to the award of an "European Fellowship," administered by the British Council. I wish also to thank the referees of this journal whose criticism led to extensive revisions of an earlier draft. Weltwirtschaftliches Archiv Bd. CXIV. i,~

Upload: antonio-aquino

Post on 23-Aug-2016

212 views

Category:

Documents


0 download

TRANSCRIPT

Intra-lndustry Trade and Inter-Industry Specialization

as Concurrent Sources of International Trade

in Manufactures

B y

Antonio Aquino

C o n t e n t s : I. Introduction and Summary. -- 11. Notion and Sources of Intra-Industry Trade. -- I II . Measurement of Intra-Industry Trade. -- IV. Empirical Performance of Different Indices of Intra-Industry Trade. -- V. A Heterodox Relationship between Export and Import Patterns, the Case of Trade in Manufactures in 1972. -- VI. Inter-Industry Specialization in Manufactures over the Period 1951-1974.

I. Introduction and Summary

I n pure theory international specialization is typically identified as the

only source of international trade, and practically the entire body of normative trade theory is based upon this belief. I t is however fairly

evident, from an even superficial look at t rade figures, tha t very substan- tial trade flows are not related to any meaningful notion of specialization.

The most striking examples are provided by trade in passenger motor cars: France exports Renault to Germany, Germany exports Volkswagen to France, and I ta ly exports Fiat to France and Germany in exchange for Renault and Volkswagen. I t does not seem however to make much sense to say that France specializes in Renault, Germany in Volkswagen and I ta ly in Fiat.

The distinction between trade which is the outcome of a genuine proc- ess of specialization (inter-industry trade) and trade which consists in the exchange of almost identical commodities (intra-industry trade) is important for a variety of reasons. First, the pat tern and intensity of intra-industry trade, being connected to not yet sufficiently explored structures of consumer preferences, ra ther than to an interaction between commodity characteristics and country attr ibutes, are much more difficult

Renuzrk : This p a p e r is b a s e d on C h a p t e r 8 ot a thes is a p p r o v e d for the Ph. D. in e c o n o m i c s a t the L o n d o n School of E c o n o m i c s in M a y 1977. I wish to t h a n k A l a s d a i r S m i t h for his i nva luab l e help whi le c a r r y i n g o u t the research . The r e sea rch cou ld be ca r r i ed o u t t h a n k s to the a w a r d of a n " E u r o p e a n Fe l l owsh ip , " a d m i n i s t e r e d b y the Br i t i sh Counci l .

I wish also to t h a n k the referees of this j o u r n a l whose c r i t i c i sm led to ex t ens ive r ev i s ions of a n ear l ier d ra f t .

Weltwirtschaftliches Archiv Bd. CXIV. i,~

276 Anton io Aquino

to predict, and probably much more strongly influenced by random fac- tors, than the pattern and intensity of inter-industry trade. Second, the price elasticities of imports and exports from a given country are likely to be much greater for intra-industry than for inter-industry trade. Third, the welfare gains of trade are likely to be much greater for inter-industry than for intra-industry trade.

The main purpose of this paper is to provide a contribution to a bet ter assessment of the empirical relevance of intra-industry trade and inter- industry specialization as concurrent sources of international trade in manufactured commodities. To this purpose, after a few thoughts about the theoretical nature of the phenomenon of intra-industry trade, in Section I I I it is shown that the indices proposed by Grubel and Lloyd [1971 ; 1975] to measure its empirical relevance are quite unreliable when- ever the country's overall trade is substantially imbalanced and some different measures of intra-industry trade are proposed which do not seem to be affected by these shortcomings. Then, with reference to the pat tern of trade in manufactures in 1972, an assessment is made of the size of the difference between the G-L indices and the measures proposed here. This difference turned out in some cases to be as high as 94% and 825/0. In Section V is investigated the relationship between export and import pat- terns of manufactures, with the surprising (at least for orthodox trade theory) finding that when a country exports relatively much of a com- modity it will usually also import relatively much (and not relatively little) of the same commodity. Finally the paper investigates on an annual basis the changes in the extent of inter-industry specialization in manufac- tures over the period 1951--1974 . The results show that the extent of inter-industry specialization in manufactures has been quite limited over that period, and has even declined over time.

II. Notion and Sources of Intra-Industry Trade

The definition of intra-industry trade as a country's simultaneous ex- port and import of the "same" commodity presents a fundamental ambi- guity stemming from the fact tha t the identification and measurement of the phenomenon heavily depends upon the degree and the kind of homo- geneity of the commodities included in each statistical group.

If the purpose is to identify the "systematic" portion of trade -- i. e., the portion of trade which is the outcome of a real process of speciali- zation stemming from the interaction between differences in commodity characteristics and differences in country attributes --, the kind of homo- geneity which matters is the one connected to the commodity character- istics exerting a major influence upon the pat tern of comparative advan-

Int ra-Indust ry Trade and Inter- Industry Specialization 277

tage. As far as manufactures, to which the analysis of this paper is mainly devoted, are concerned, inter-industry differences in "technology-inten- sity" (or skill-intensity, or research-intensity) have emerged from the im- pressive amount of empirical research carried out in the last two decades 1 as by far the most powerful factor of international specialization. On this basis intra-industry trade should be defined as the simultaneous export and import by the same country of commodities having identical "tech- nology-intensity."

It is fairly evident that if one is interested in this notion of intra- industry trade, its measurement on the basis of the available statistical classifications is subject to both an upward and a downward bias. The upward bias stems from the heterogeneity of the commodities included in each statistical group, even at the finest level of disaggregation, from the point of view of their "technology-intensity." The downward bias is relat- ed to the fact that commodities having identical technology-intensity are often included in different statistical groups.

As examples of statistical groups which are highly heterogeneous from the point of view of technology-intensity even at the finest level of disag- gregation one can mention the group 7142 of the SITC (the United Nati- ons Standard International Trade Classification System) which includes the simplest calculating machines together with the most sophisticated calculating and accounting systems, and the group 7242 of the SITC which includes portable radios together with highly sophisticated professional re- ceivers z. Examples of products with almost identical technology-intensity but considered in completely different sections of the SITC are textiles (SITC 65) and clothing (SITC 84) for low-technology products, and syn- thetic plastic materials (SITC 58) and scientific instruments (SITC 861) for high-technology products.

A rigorous analysis of the phenomenon of intra-industry trade would then require a thorough reclassification of products to obtain groups homo- geneous inside and heterogeneous between them from the point of view of the commodity characteristics which most matter. This is something far

i See, for example, Hirsch [i974] , alld the works quoted there.

This phenomenon has become more and more important as the spreading of tariff provisions for offshore assembly (OAP) has s t imulated firms, part icularly mult inat ionals , to localize in different countries the various phases of vert ical ly integrated industries accord- ing to the particular factor requirements of each phase. According to Finger US OAP im- ports from developing countries increased from 1969 to 1972 at an annual rate of 60% and West German OAP imports from these countries by 36% per year. Ir~ 1972 such imports made up io% of West German and 22% of US manufactured imports from developing countries [Finger, 1975].

I t is evident that if different phases of an industry are characterized by different factor requirements the resulting trade should be included in inter-industry specialization.

18"

278 A n t o n i o A q u i n o

beyond the scope of the present research, but this kind of considerations should be kept well in mind while evaluating the results of any analysis of intra-industry trade and inter-industry specialization.

Turning to the sources of intra-industry trade, the first clear insights into the nature of this phenomenon are due to Linder [1961, p. lO2] who perceived the strong potential for intra-industry trade between countries with similar income levels, arguing that intra-industry trade is similar to the trade flows between different regions of the same country. Along simi- lar lines argued Gray I who showed how product differentiation, combined with increasing returns to scale at least over some part of the supply curve, can explain intra-industry trade (two-way trade in Gray's termi- nology).

But Hufbauer and Chilas [1974] showed that intra-industry trade is much more important when considering different countries (United King- dom, France, West Germany, the rest of Western Europe, Canada, United States, Japan) than when considering the four major regions of the United States. This finding confirmed the authors' belief that the structure of tariffs is the main source of intra-industry trade. In particular Hufbauer and Chilas argued that the tariff reductions implemented in the last two decades mainly consisted of reciprocal concessions industry by industry and favored intra-industry trade over inter-industry specialization.

As a mat ter of fact, the experience of recent years in several industrial countries has shown that national governments have been very often in- clined to employ various instruments -- from substantial subsidies to direct intervention -- in order to avoid the effects of the even limited measures of trade liberalization: the dismantling of the less competitive sectors and the transfer of the factors so released to the more competitive ones. The justifications usually put forward for this refer not only to the short-run difficulties of the changes but also to the political importance of maintaining an integrated productive structure, i .e. , to preserve some productive facilities in all manufacturing sectors.

Such a behavior of national governments, in presence of scale econo- mies and product differentiation, can probably explain a good portion of intra-industry trade. In presence of scale economies the rate of tariff or subsidy required to assure the survival of the sectors involved decreases perhaps quite rapidly until they reach a fairly high level of activity. Once the level of production has been fixed at the level which minimizes the cost of assuring the survival of the sector, if products are, or can be, differ- entiated, and to some extent almost all manufactures can be differentiated, it is more profitable to sell them on the international markets rather than

a Gray [x973]. - - See also Davies's comment [x977] and Gray's reply [I977].

Intra-Industry Trade and Inter-Industry Specialization 279

exclusively on the home market, even in presence of substantial flows of imports of very similar products.

III . M e a s u r e m e n t o f I n t r a - I n d u s t r y Trade

Grubel and Lloyd [1971 ; 1975] have made the most serious effort so far to devise a rigorous measure of intra-industry trade. The basic index proposed as a measure of intra-industry trade in country j's foreign trade in commodity i is the following:

(Xlj "~- i i j ) - - I X i j - - i i j i B l | = (Xij -~- Mii) �9 I 0 0 (I)

Where Xii and Mij stand for the values of country j's exports and imports, respectively, of commodity i.

Bij measures intra-industry trade (the numerator of the fraction) as a percentage of j's total trade in commodity i. Its value ranges between zero (when either Xii or Mij is zero so that there is no intra-industry trade in commodity i) and IOO (when Xij -- Mij, so that all trade in commodity i is intra-industry trade).

As a summary measure of intra-industry trade in country j, at any given level of industry-aggregation, Grubel and Lloyd proposed a weighted average of the values of Bij, with weights given by each industry's share in j's total trade:

(Xi. j + Mi]) X (Xij q- i i | ) - - Z lXi. j - - i l l ] __ i i

Bj ' = Zi Bit Z (Xij q- Mii ) X (Xij q- Mij ) �9 IOO (2) i i

Similarly to Bit, B i expresses j's total intra-industry trade as a percentage of its total trade. Grubel and Lloyd [1975] correctly pointed out that B i is a downward biased measm'e of intra-industry trade whenever j 's total trade is not balanced. To correct for this bias Grubel and Lloyd proposed the following adjusted summary measure of the proportion of intra- indus- try trade in country j's total trade:

(xij + - xix j - Mijl C j ~ _ i i

- - Y. (Xi.~ + Mij) - - I X ~ j - - Y. Mijl " IOO = B j / ( l - k ) (3) i i i

1 This index is denoted as Bi by Grubel and Lloyd [1971 ; I975]. Since the use of the in- dustry suffix i can be misleading in an index referriug to a country, it was thought preferable to use here the correct country-suffix j in reporting their formulas.

Denoted as ~i by Grubel and Lloyd [x97x; I975]; see also note to Bj.

2 8 0 A n t o n i o A q u i n o

where k is j's overall trade imbalance as a proportion of j's total t rade

(k = I xij - ~. Mij / X (Xij + Mij ) ) i i i

From their text [Grubel and Lloyd, 1975, p. 22], and since they did not think it necessary to correct the elementary index Bij, it is evident that Grubel and Lloyd believed that the bias of the summary measure Bj arises in the process of obtaining it as a mean of the values of Bij (implic- itly considered unbiased). But this does not seem to be correct. If j 's total trade is imbalanced, B j is a downward biased summary measure of intra-industry trade just because Bij is a downward biased measure of intra-industry trade in each commodity. This because one cannot possibly maintain that the overall imbalance has not an imbalancing effect on the single commodities' trade flows and then recognize that the imbalancing effect appears at a highest level of industry-aggregation. I t is true that there is no reason to expect the imbalancing effect to be equiproportional in each single industry, but on average the imbalancing effect on each industry's trade must be equal to the overall imbalance. In the absence of any information about inter-commodity differences in the strenght of the imbalancing effect the best one can do is then to assume tha t it is equi- proportional in all industries and equal to the overall imbalance.

The most straightforward way of correcting for the overall imbalance at the elementary level, i. e., before calculating the values of ]~ij (under the assumption that the imbalancing effect is equiproportional in all indus- tries) requires an estimate of what the values of exports and imports of each commodity would have been if total exports had been equal to total imports. Denoting with a superscript "e" these theoretical values of ex- ports and imports, these can be obtained as follows:

1 1 Z i (Xij + Mi- i) ~ iE (Xij + Mii)

Xi~ = X i j ; M~j = i i j (4t Z X~. i Z M ij i i

1 I t can be easily verified that Zi xi] = 2;i Mi~ = ~ i E. (Xij + Mij).

�9 Applying formula (I) to the values of X~j and M~ one can obtain a measure of the proportion of intra-industry trade in j's trade of commodity i, purged of the imbalancing effect of the overall imbalance in j's trade. We shall call Qij this new index. A weighted average of the values of Qij relative to the various commodities, with weights given by each commod-

In t r a - Indus t ry Trade and In t e r - Indus t ry Specialization 281

ity's share in j's total trade, gives our corrected summary measure of the proportion of intra-industry trade in j's total trade:

Z (Xij + Mi i) - - X iX~.~ - - M~I Qj i i

. . . . . . - - l O O (s)

i

since X (Xi~ + M~) ~ Z (Xij + Mij ) i i

Qj has a great advantage over both Bj and Cj. From (2) and (3) one can see that both Bj and Cj depend upon the value of the expression

]Xi j - - Mij ! . This expression has the peculiarity that for any subset of i

commodities for which Xij ~< Mij for all i or Xij t> Mij for all i its value is constantly equal to ] X Xij -- X Mij I whatever the values of Xij and Mij,

i i provided that one of those constraints is respected 1. This means tha t both B~ and Cj assume identical values in situations in which it is evident that the importance of intra-industry trade is substantially different. The following example illustrates this possibility.

Chemicals Textiles Machinery

Total

Case I Case 2

Xi j Mij X i j Mij

20 io io IO io 5 40 5 40 20 20 20

70 35 7 ~ 35

In case I it is evident that there is no inter-industry specialization between chemicals, textiles and machinery as the ratio of exports to im- ports is the same for all industries; in this case all trade is then intra- industry trade. In case 2 the situation is altogether different. The country now reveals a clear tendency to specialize in textiles with respect to chemi- cals and machinery; in this case only a proportion of total trade is intra- industry trade. If however the indices proposed by Grubel and Lloyd [1971 ; 1975] were to be applied mechanically to these situations one would get quite a surprising answer: the same values would be obtained in case I and in case 2, both for Bj and for Cj. I t can be easily verified that Bj would be

1 This seems to suggest, arid the empirical analysis of Section IV will confirm it tha t while Bj is a downward biased measure of in t ra - indus t ry trade Cj is an upward biased measure of int ra- industry trade.

282 A n t o n i o A q u i n o

equal to 66.66 and Cj to Ioo both in case I and in case 2. Both indices would so detect the same amount of intra-industry trade in two situations so substantially different. A much more sensible answer is obtained by using Qj. Its value would be in fact equal to Ioo in case I (no inter-indus- t ry specialization hence all trade is intra-industry trade) and to 57.I in case 2 (42.9% of trade stems from inter-industry specialization and the remaining 57.I% is intra-industry trade). In the next section we shall verify that such large differences between Bj, Cj, and Qj are much more than a theoretical possibility. We shall find in fact for the real world some differences even much larger.

The above example can also be used to illustrate the need for correct- ing the imbalancing effect of the overall trade imbalance not only at the aggregated level but also when measuring intra-industry trade in the single industries. If this is not done the results of the analysis by industry (using Bij ) are inconsistent with the results of the aggregated analysis (using Cj). In case I of the above example, for instance, Bij would indicate that only 2/3 of j's trade in each industry is intra-industry trade; using Cj as a summary measure of the importance of intra-industry trade in j 's trade we would reach instead the conclusion that all of j's trade is intra- industry trade.

Grubel and Lloyd perceived the existence of the peculiarities of CJ above pointed out. They did however fail to identify their exact nature and believed that in any case they did not apply to Bj. 1 It is however clear from the previous analysis that Bj has exactly the same shortcomings as C~. 2

The l~l mentioned at the end of the quotation in footnote I is an in- dex derived by Grubel and Lloyd from Michaely [x962 ] and defined as follows:

1 Xij Mij

l~i : 1 - - ~ i~ y~ Xi j ~, Mi j (6) i i

Contrary to Grubel's and Lloyd's belief l~i is immune from the shortcom- ings common to Cj and Bj. In fact, substituting (4) in (5) it can be easily shown tha t F i is equivalent to our summary measure Qj .3

i " I f a country has a large imbalance of trade such that exports for most industries exceeded imports, or vice versa, it would be better to use the unadjusted measure Bi [Bj] in preference to Ci [Cj] or Fi "~Grubel and Lloyd, x975, p. 27].

This is also evident from the fact that Bj ~ Cj/(1--k) ; see formula (3)-

s The equivalence between Fi and Qi was suggested to me by Alasdair Smith.

I n t r a - I n d u s t r y Trade and I n t e r - I n d u s t r y Specia l iza t ion 283

IV. Empirical Performance of Different Indices of Intra-Industry Trade

The three final sections of this paper are devoted to the empirical analysis of intra-industry trade and inter-industry specialization in manu- factures. Most of this is based upon a disaggregation of trade flows of manufactures in 25 classes of products 1 (listed in Table 2). In my judge- ment such a disaggregation allowed us to reach a fairly good degree of homogeneity of the products included in each class from the point of view of their technology intensity~, while keeping at a manageable level the work of data gathering; this constraint was particularly stringent since all countries exporting a significant amount of manufactures (listed in Tables I and 4) were included in the analysis and part of this spanned over a period of 24 years.

One can see from Table 2 that the classes of products were not identi- fied mechanically, on the basis, for example, of the number of digits of the SITC's groups or the share of each class in tofal trade. Classes of products representing a very large share of total trade and corresponding to most of a two-digit group or a one-digit section of the SITC (65 except 6516; 71 except 714, 715, 717; 8 except 86) were considered together with very specific products representing a very small share of total trade and corres- ponding to four-digit subdivisions of the SITC (7241 and 7242). The identification of the classes of products was done in fact on the basis of the results of a much more disaggregated analysis (based upon 135 manu- factured products) of the technology intensity of manufactures [Aquino, 1977, Chapter io]. This analysis had clearly shown that while in some cases it was necessary to reach a high level of disaggregation in order to obtain a good degree of homogeneity, in other cases this could be obtained for much larger groups of products.

In this section is presented an assessment of the magnitude of the differences between B~, Cj and Qj with reference to trade flows of manu- factures in 1972. In the first three columns of Table I are given, in the order, the values of Bj, Cj, and Qj. Column 4 gives the percentage differ- ence between Qj and Bj and column 5 the percentage difference between Qi and Cj.

The values in column 4 show that Bj is in most cases a substantially downward biased index of intra-industry trade. As expected the size of the bias is particularly high for the countries with a large imbalance in total trade of manufactures -- Brazil (94%) and Japan (82%). On the other hand, the results confirmed that C~ is an upward biased measure of

x Only the ana lys i s of i n t e r - indus t ry spec ia l iza t ion over the per iod ~ 9 5 i - - i 9 6 2 was carried out a t a s l igh t ly more aggrega ted level (x8 indus t r ies l i s ted in Table 3).

* As stressed in Sect ion I I , th is is the k i n d of homogene i t y mos t r e l e v a n t for our purposes-

284 A n t o n i o A q u i n o

I

,~'Z~ ~ ' ~

~ ~ o~

1 - - _ - -

~ea

I

~ l ~ ' ~ j ~ i ~ l ~ ~ . ~ . . j . ~ ; ~

~ . ~ o o ~o ~ =~ ~ ~ ~ % ~ ~. ~ .

[ g l [ ~ 1 " ~ . . . . . . I " ' ' " ~

. �9 " , . ; ~ �9 * ~ . . . . ~ �9 -,..; ~ . . . . . . .

f 1 r

I

o.~

Intra-Industry Trade and Inter-Industry Specialization 285

intra-industry trade. The size of the bias was again found fairly high for most of the countries, and particularly for Greece (59%), and Japan, Brazil, Mexico (38%).

Judging from the values of Qj (column 3), intra-industry trade as a proportion of total trade in manufactured commodities, once its measure has been purged for the overall trade imbalance, ranged in 1972 from 22.9% in the case of India to 87.4% for France. For most of the leading industrial nations this percentage was greater than 700/o : United King- dom 81. 9 , Netherlands 78.7, Sweden 76.3, West Germany 76.0, Austria 75.0, Canada 73.5, I taly 72.3, Denmark 70.3, Belgium 7o.1. Somewhat smaller was the value of the index for the United States (57.3)-

A similar kind of analysis, this time by industry, is presented in Table 2. Here Qi is compared only with the unadjusted index B i because only this was used by Grubel and Lloyd [1971 ; 1975] to measure intra-indus- t ry trade as a proportion of world trade in each industry. Moreover, while Grubel and Lloyd calculated it as a simple arithmetic average of the values of the elementary index Bij , here both B i and Qi have been calculated, similarly to Bj and Qj, as weighted average (this time across countries) of Bij , and Qij, with weights equal to the share of country j in world ex- ports plus imports of industry i. Here again the difference between Qi and B i turned out to be fairly high, though not as high as for the analysis by country. As for the level of intra-industry trade, when measured by QI, it accounted in 1972 for more than 50% of total trade in most manu- factured industries. The percentage was lower only for radio and television broadcast receivers, and for aircraft.

In evaluating these results one should however take into account the fact that this is an analysis of intra-industry trade inside the manu/acturing sector. Consequently the adjusted measures Cj and Qj have been calculat- ed by correcting for each country's imbalance in total trade of manufac- tures. In so far as this is larger than the imbalance in total trade in all commodities, both the absolute value of Cj and Qj, and the difference between Qj and Bj and Cj are likely to be greater than when considering trade in all commodities. The main theoretical reason for isolating in this analysis trade in manufactures is that the factors of inter-industry special- ization inside the manufacturing sector are different from the factors of specialization between manufacturing and the other sectors.

2 8 6 Antonio Aquino

.~ go

.0 ..= .,,:

g ' ~ ~ , . . , ~a

,0

0

B

. . ' ~ -

I v".l

i

} i

,, o ~

g

I

I I I I I I I

"~1 I I I ' l l I I ~ ' " I

I I I I

" 44~ 6 ~ & 4 ~ 4~K6 ~-4~-4d "~ "

.<.,.6 ~ ~ , A g . �9 ,4 c; 4 - d g.,4 ,,~ ~ &. ' . 6 ~ ,q , ,6 & . 4 . 4 4-

o m ~

o ~ ' ,

Int ra-Iudustry Trade and In ter - Indus t ry Specialization 287

V. A Heterodox Relationship between Export and Import Patterns, the Case of Trade in Manufactures in I972

The literature on intra-industry trade has now obliterated the ortho- dox picture of a world in which a country either exports or imports a com- modity according to the country having in it a comparative advantage or disadvantage. The general feeling inherent in this literature seems how- ever to be that intra-industry trade is something which makes the pat- tern of trade flows much more dimmed a reflection of a process of inter- industry specialization which remains all the same the fundamental de- terminant of the pattern of trade. More specifically the general belief in international economics seems to remain that if a country exports rela- tively much of a certain commodity, it will import relatively little of it. As far as I know only I,inder [1961 , pp. 9 I, i38 J challenged this general belief, but apart from a sophisticated but inconclusive empirical test carried out by Hufbaner [197o], Linder's hints have not up to now been pursued any further.

This problem has been resumed here with a very simple test. The em- pirical nature of the relationship between export and import patterns has been investigated here by estimating, for each country, the elasticities of the share of each industry in the country's total exports of manufactures (normalized by the share of the industry in world exports of manufac- tures) 1 with respect to the industry's share in the country's total imports of manufactures (normalized by the industry's share in world imports of manufactures). A similar analysis was performed by industry.

The results of the analysis, reported in columns 6-- 9 of Table I and columns 4--7 of Table 2, are quite surprising. Particularly in the analysis by country (with elasticities estimated across industries), the values of the elasticity (Table I) turned out to be much more often positive than nega- tive. This means that relatively high values of exports are much more often associated with relatively high than with relatively low values of imports. This turned out to be true particularly for Canada, France, Netherlands, Portugal, Switzerland, India, Singapore, Korea Republic and Hong Kong, for the analysis by country (Table I, columns 6--9) and for office machines, radio broadcast receivers, passenger motor cars, and parts for motor vehicles, for the analysis by industry (Table 2, columns 4--7). The only significant exceptions to this pat tern turned out to be Japan among the countries, and miscellaneous manufactures except in- struments among the industries.

From the results of the analysis by industry it looks as if it is very unlikely that these results are just the outcome of the heterogeneity of

1 This normalized export share is equivalent to the Balassa index defined in Section VL

2 8 8 A n t o n i o A q u i n o

product classes. International trade in the products of different phases of vertically integrated industries, stimulated by the spreading of tariff pro- visions for offshore assembly, can perhaps contribute to explain in part this result. The fact however that for industries such as office machines, radio broadcast receivers, passenger motor cars and parts for motor ve- hicles, export shares and import shares turned out to be significantly positively rather than negatively correlated clearly indicates the possibility that the pattern of home demand, in accordance with Linder's suggestive hints, has a much stronger and completely different impact upon the pat- tern of trade flows in manufactures than it is predicted by the orthodox theory of international specialization.

VI. Inter-Industry Specialization in Manufactures over the Period 195 I - - 1974

The results of Section V, according to which, at least as far as trade in manufactures is concerned, the relationship between export and import patterns is much more often positive than negative, cast substantial doubts upon the notion itself, let alone the measure, of inter-industry specialization. Nor can these doubts be resolved in any easy way, as could be the at tempt to measure inter-industry specialization by combining ex- ports and imports data to obtain one notion or the other of net exports.

With all these doubts well in evidence we present in this last section of this paper an analysis year by year, of the extent of what is tradit ionally called "inter-industry" specialization in manufactures over the period 1951--1974. As a measure of the intensity of inter-industry specialization in a country or an industry (in a given year) we use here the standard deviation for that country or industry of the Balassa index of"revealed comparative advantage" or "export performance" applied here to trade in manufactures alone 1. For country j and commodity i this is defined as follows:

Ei/X Eij i

% = X i,/X x Eij ' ioo

Where Eij stands for j's exports of commodity i, and only manufactured commodities are considered.

I t is intuitively obvious why the standard deviation of the index of specialization is a measure of the extent of inter-industry specialization.

1 The reason for using this index here, as a l t e rna t i ve to the c o m p l e m e n t to ~oo of Qj and Qi, was tha t , be ing based on ly upon expo r t values , the s t a n d a r d d e v i a t i o n of the Ba l a s sa index halves the work of da t a ga ther ing . This was fa i r ly i m p o r t a n t in an ana lys i s of in ter- i ndus t ry specia l izat ion spann ing over 24 years .

lntra-Industry Trade and Inter-Industry Specialization 289

If, for example, the commodity-composition of a country's exports were the same as the commodity-composition of world exports, Sij would be equal to IOO for all industries and its standard deviation would be zero, so correctly reflecting the fact that in that country there would not be any inter-industry specialization. When a country specializes in some of the industries, the value of the index becomes greater than IOO for these industries and smaller than IOO for the remaining industries; the greater the extent of inter-industry specialization, the greater the spread of the values of the index around IOO, the greater the value of the standard deviation.

Since inter-industry trade if appropriately measured would be the complement to one of intra-industry trade, any acceptable measure of inter-industry specialization should be strongly correlated, (inversely) to a measure of intra-industry trade. The statistical relationship between the standard deviation of the index of inter-industry specialization and our summary measures of intra-industry trade (Qj for countries and Q ~for industries), was investigated with reference to 1972 pat tern of exports and imports of manufactures. The values of the standard deviation for that year, both unweighted and weighted by the industry's or country 's share in world exports of manufactures, by country and by industry, are given in the last two columns of Tables I and 2. As expected, the linear corre- lation turned out to be significantly negative in all Cases.

In particular the coefficient of correlation between the weighted stand- ard deviation and Q was - . 744. for the analysis by country and --. 856 for the analysis by industry.

The values of the weighted standard deviation of the index of inter- industry specialization for the two subperiods 1951 --1962 and 1962--1974, by industry and by country, are given in Tables 3, 4, 5 and 6. The general impression given by the values in these tables is that the extent of inter- industry specialization in these industries has been rather low over the whole period 1951--1974, with a tendency towards a further reduction, particularly during the subperiod 1951--1962.

The only industries for which the standard deviation turned out to be greater than the mean value of the index (ioo) in most years of the period were: television broadcast receivers, radio broadcast receivers, aircraft, ships and boats. Among the countries this happened for Ireland, Greece, Norway, Portugal, Switzerland, Mexico, India, Korea Republic, Hong Kong. Particularly low was the standard deviation of the index of special- ization for France (declining from 57 in 1953 to 25 in I968 ), West Ger- many (declining from 49 in 1951 to 29 in i974), and the United Kingdom (ranging between 24 and 4o).

290 A n t o n i o A q u i n o

l

o

~ " ~ ~ ~ ~ �9

l I I I I t 1 1 1 1 I l l

~ ' ~ ' ~ ~ "

1 1 1 1 I t l i l l I l l t

~ o ~

.,9~

.~ II I �9 ~., o'1 ~ ,

o ,~~

-~~ ~ o ,...1 0

' n 0 ~o" ~

~-~

.~ , ~.~, ~ i~ ~

~.~

G

og

: : : : : : : : : : : : : : : : : : ~

:~ : : : : : : : : : : : : : :~

~ E ~ ~ o ~ ~ . ~ ~ ~ . ~ . ~ 0 ~ 0 , ~ ~ 0 ~ , ~ 0 ~ m ~ , ~ ~ 0 ~ 0

I n t r a - l n d u s t r y T ra de a nd I n t e r - I n d u s t r y Special izat ion

t~

O3

I I I I I [ 0 I I I I I I I

I I I I l l t l t l l l II

1

w ~ M w M M ~

~d

..~

r

i i i i i i i i i i i i i i i i

2 9 I

Weltwirt~.haftliches Arehiv Bd, CXIV, x9

292

&

.R

1 lob

M

M

Antonio Aquino

�9 " " ~ 4 a ~ ~ ' a ~ ' 4 4 4 " ~ a ~ I t l l l 1 I I 1

�9 " " , ~ ; " " ~ ' ~ " ~ - ~ " ~ - ~ " ~ A ~

I 1 1 1 1 l II I

o ~ ~

. ~ o . . . . . o . . . . . . ~

I

I n t r a - I n d u s t r y T r a d e a n d I n t e r - I n d u s t r y S p e c i a l i z a t i o n ~ 9 3

C~

P.

I

M

~ ~ ' " " ~ 4 . " o ; ~ "

I 1 1 1 1 1 11 I I I l l l l l I

. . . . . . 2 ~ o

~ - ~ o ~ ~oo ~ ~ - ~ ~ o~oo ~

g . . . . . . . ~ e . . . . . . . . . . . . . . .

. . . . . . o . . . . . . . . . o ~ ~

�9 ~ . . . . . ~ �9 . . ~ . . . . . . . . . . . . ~ . �9 ~ . . . . . ~ �9 . ~ . ~ . . . . . . ~ ~ . . . . ~ H

z9e

294 A n t o n i o Aqu ino

The low values and the decl in ing t r e n d found for th is measure of the e x t e n t of i n t e r - i n d u s t r y specia l iza t ion is a conf i rmat ion of the resul t s ob- t a i ned in the analys is of i n t r a - i n d u s t r y t r a d e b y Grnbe l and L l o y d [1975] and conf i rmed also in Sect ion IV of th is paper . The e x t e n t of i n t e r - i n d u s t r y spec ia l iza t ion be tween m a n u f a c t u r i n g indust r ies , p a r t i c u l a r l y as far as the l ead ing indus t r i a l count r ies a re concerned, has been fa i r ly l imi t ed over the pe r iod 1951--1974, and , moreover , in spi te of the d i s m a n t l i n g of t r a d e bar r ie rs and the resu l t ing s t e a d y increase in i n t e rna t i ona l t r a d e the re has been a t e n d e n c y t owards a fu r the r r educ t ion in the e x t e n t of special i- za t ion.

The (increasing) he te rogene i ty of p roduc t classes can pe rha ps cont r ib - u te to exp la in ing this phenomenon , and the same can be sa id of govern- m e n t in te rven t ions to coun te rac t the decl ine of e s t ab l i shed indus t r ies . I t looks however un l ike ly t h a t these two factors a lone can exp la in comple te ly , or even a subs t an t i a l pa r t , of this phenomenon . Also in the l ight of the re- sul ts of Sect ion V we m u s t t hen conclude t h a t the t r a d i t i o n a l v iew t h a t i n t e rna t iona l t r a d e is essen t ia l ly the ou tcome of a process of spec ia l iza t ion s t e m m i n g from the in t e rac t ion be tween c o m m o d i t y charac te r i s t i c s and coun t ry a t t r i b u t e s does no t seem to be a n y longer an accu ra t e desc r ip t ion of the real world.

References

Aquino, Antonio, Technical Progress and International Specialization in Manu/ac- tures, I95I--~974, Ph.D. diss. at the London School of Economics, 1977.

Davies, Robert, "Two-Way International Trade: A Comment", Weltwirtseha/ttiches Amhiv, Vol. 113, 1977, pP. 179--I81.

Finger, J. M., "Tariff Provisions for Offshore Assembly and the Exports of Devel- oping Countries", The Economic Journal, Vol. 85, London, 1975, PP. 365--371.

Gray, H. Peter, "Two-Way International Trade in Manufactures: A Theoretical Underpinning", Weltwirtscha/tliches Amhiv, Vol. lO9, 1973, pp. I9--39.

- - , "Two-Way International Trade: Reply", ibid., Vol. i 13, 1977, pp. 182--I84.

Grubel, Herbert G., and P. J. Lloyd, "The Empirical Measurement of Intra-Industry Trade", The Economic Record, Vol. 47, Melbourne, 1971, pp. 494---517 .

- - , a n d - - , Intra-Industry Trade, The Theory and Measurement o/ International Trade in Differentiated Products, London, Basingstoke0 1975.

Hirsch, Seer, "Capital or Technology ? Confronting the Neo-Factor Proportions and Neo-Technology Accounts of International Trade", Weltwirtscha]tliches Archly, Vol. 11o, I974, PP. 535--563.

Hufbauer, G. C., "The Impact of National Characteristics and Technology on the Commodity Composition of Trade in Manufactured Goods", in: Raymond Vernon (Ed.), The Technology Factor in International Trade, Universities-l~ational Bureau Conference Series, 22, New York, London, I97 o, pp. 145--231.

Intra-Industry Trade and Inter-Industry Specialization 295

Hufbauer, G. C., and J. C. Chilas, "Specialization by Indust r ia l Countries: E x t e n t and Consequences", in: Herbert Giersch (Ed.), The International Division o/ Labour, Problems and Perspectives, In ternat ional Symposium, Ttlbingen, 1974, pp. 3--38.

Linder, S. B., A n Essay on Trade and Trans/ormation, Stockholm, I96I.

Michaely, M., Concentration in International Trade, Contributions to Economic Anal- ysis, 28, Amsterdam, I962.

Organisation for Economie Co-operation and Development (OECD), Statistics o/ Foreign Trade, Series C, Paris, various issues.

Organisation for European Economic Co-operation (OEEC), Foreign Trade Statistical Bvlletins, Series IV, Paris, various issues.

United Nations (UN): Commodity Trade Statistics, New York, var ious issues. Yearbook o/ International Trade Statistics, Hew York, various issues.

* S

Z u s a m m e n f a s s u n g : Intra-industr iel ler Handel und inter industr iel le Speziali- sierung als Ursachen des internat ionalen Handels mi t Industr ie~rzeugnissen. - - Der FIauptzweck dieses Aufsatzes ist es, zur besseren Einsch~.tzung der empir ischen Rele- vanz yon In t ra-Handel und industrieller Spezialisierung als gleichzeitige Quellen des internationalen Handels mit Industr ieerzeugnissen beizutragen. ~ a c h einigen (3ber- legungen zur theoret ischen l~atur des Ph~nomens des In t r a -Hande l s wird gezeigt, dab die Indizes, die Grubel und Lloyd zur Messung seiner empir ischen Relevanz vor- geschlagen haben, dann ganz unzuverliissig sind, wenn sich der Gesamthande l des Landes in einem erheblichen Ungleichgewicht befindet. Es werden andere MaBe fiir den In t ra-Handel vorgeschlagen, die n icht mi t diesen M~ngeln behaf te t zu sein scheinen, und die groSen Unterschiede im empir ischen Verhal ten der verschiedenen Indizes werden an Hand des Handels mi t Indus t r iewaren im Jahre I972 gezeigt. Dann wird die Beziebung zwischen der Expor t - und der I m p o r t s t r u k t u r fllr Indu- striewaren untersucht mit dem liberraschenden Ergebnis, dab ein Land dann, wenn es verh~ltnismliBig groBe Mengen eines Gutes exportiert , im al lgemeinen auch ver- hiiltnismiil3ig viel davon importiert. Schliei31ich un te r such t die Arbeit auf Jahres- basis die Ver~nderungen des AusmaBes der inter industr ie l len Spezialisierung im Zeitraum 1951--1974 . Die Ergebnisse zeigen, dab das Ausma0 der inter industr ie l len Spezialisierung in diesem Zeitraum recht begrenzt war und sogar im Laufe der Zeit abgenommen hat .

R 6 s u m 6 : Le commerce intra- industr iel et la sp6cialisation inter- industr iel le comme des sources concurrentes du commerce in ternat ional en produi ts manufac tu - tiers. - - Le bu t essentiel de cet article est de contr ibuer ~ une d6terminat ion meil- leure de l ' importance empirique du commerce intra- industr iel et de la sp6cialisation inter-industrielle comme de sources concurrentes du commerce in ternat ional avec des produits manufacturiers . Apr~s quelques r6fiexions sur la na tu re th~orique du ph~nom~ne du commerce intra-industr iel nous d6montrons que les indices propos6s par Grubel et Lloyd pour mesurer son impor tance empirique sont toujours enti~re-

296 A n t o n i o A q u i n o Int ra-Indust ry Trade and In ter - Indust ry Specialization

m e n t d o u t e u x si le commerce to t a l du p a y s es t s u b s t a n t i e l l e m e n t d i sba lanc6 . N o u s p r o p o s o n s que lques mesu res d i f f6rentes d u c o m m e r c e i n t r a - i n d u s t r i e l q u ' e l l e s n e s e m b l e n t pa s ~tre influ6 p a r ces d 6 f a u t s e t n o u s d $ m o n t r o n s les g r a n d e s d i f f6rences de la p e r f o r m a n c e e m p i r i q u e des ind ices d i f f$rents c o n c e r n a n t le r6seau d u c o m m e r c e avec des p rodu i t s m a n u f a c t u r i e r s en I972. P u i s n o u s a n a l y s o n s la r e l a t i on en t r e les r6seaux d ' e x p o r t a t i o u e t d ' i m p o r t a t i o n des p r o d u i t s m a n u f a c t u r i e r s avec le r 6 s u l t a t 6 t o n n a n t que si u n p a y s expor t e r e l a t i v e m e n t b e a u c o u p d ' u n p r o d u i t il n o r m a l e m e n t v e u t auss i i m p o r t e r r e l a t i v e m e n t b e a u c o u p du m ~ m e produi t . F i n a l e m e n t , le pap i e r a n a l y s e su r une base annue l l e les c h a n g e s de l ' 6 t e n d u e de la sp6c ia l i sa t ion i n t e r - i ndu - s tr ie l le en p rodu i t s m a n u f a c t u r i e r s en cour s de la p6r iode 1951- -1974 . Les r 6 s u l t a t s d 6 m o n t r e n t que l ' 6 t endue de la sp6c ia l i sa t ion in t e r - indus t r i e l l e en p r o d u i t s m a n u f a c - tu r i e r s 6 ta i t assez l imit6e en cours de ce t t e p6r iode e t m ~ m e d6c l ina i t en cour s du t emps .

R e s u m e n : Comercio i n t r a - i n d u s t r i a l y especia l izaci6n i n t e r - i n d u s t r i a l como fuen te s c o n c u r r e n t e s del comerc io i n t e r n a c i o n a l en m a n u f a c t u r a s . - - E1 p r i n c i p a l p rop6s i to de este a r t ieu lo es c o n t r i b u i r a u n me j o r e n t e n d i m i e n t o de la r e l e v a n c i a empt r i ca del comerc io i n t r a - i n d u s t r i a l y de la especia l izac i6n i n t e r - i n d u s t r i a l como fuen t e s concu r r en t e s del comerc io i n t e r n a c i o n a l en b ienes m a n u f a c t u r a d o s . Despu~s de a l g u n a s ref lexiones acerea de la n a t u r a l e z a te6r ica del f e n 6 m e n o del comerc io i n t r a - i n d u s t r i a l se m u e s t r a , que los Indices p r o p u e s t o s por G r u b e l y L l o y d p a r a med i r su r e l evanc i a empl r ica son poco conf iables en el caso q u e el comerc io t o t a l de u n pals estA s u s t a n c i a l m e n t e desequ i l ib rado . Se p r o p o n e n m e d i d a s d i fe ren tes del comerc io i n t r a - i n d u s t r i a l que pa recen no ser a f e c t a d a s p o t e s t a s l im i t ac iones y se m u e s t r a n las g r a n d e s d i ferencias en el desempef io empir ico de los d i s t i n t o s Indices con respec to al p a r t 6 n de comerc io en m a n u f a c t u r a s de 1972. E n s e g u i d a se i n v e s t i g a la re lac i6n en t r e los p a t r o n e s de e x p o r t a c i 6 n e i m p o r t a c i 6 n de m a n u f a c t u r a s con el r e su l t ado so rp renden te , que c u a n d o u n pa i s e x p o r t a r e l a t i v a m e n t e m u c h o de u n b ien t a m b i 6 n i m p o r t a r ~ r e l a t i v a m e n t e m u c h o del m i s m o bien. F i n a l m e n t e , et a r t i cu lo i nves t i ga sobre u n a bas i s a n u a l Ins c a m b i o s en l a m a g n i t u d de la espec ia l izac i6n i n t e r - i ndus t r i a l en m a n u f a c t u r a s p a r a el per iodo 1951 - - I 974. Los r e s u l t a d o s m u e s t r a n , que la m a g n i t u d de la especia l izaci6n i n t e r - i n d u s t r i a l en m a n u f a c t u r a s h a s ido bas - t a n t e l i m i t a d a a t r a y , s de es te per lodo e i ne l u so h a dec l inado a t r av6s del t i empo .