World Bank Reprint Series: Numiber 150 Hollis B. Chenery Interaction~s between Industrialization and Exports Reprinted with permission from A merican Economic Review, vol. 70, no. 2 (May 1980), pp. 281-87. CHANGES IN TRADE SHA RES AND ECONOMIC GROWTH Interactions Between Industrialization and Exports By HOLUIS B. CHENERY* Sustained economic growth requires a odology emphasizes interrelations on the de- transformnation of the structure of produc- mand side, which tend to be neglected in tion that is compatible with both the evolu- other approaches. Attention is focused on tion of domestic demand and the opportuni- the effects of early or late development of ties for international trade. This transforma- manufactured exports, which is a major tion normally involves a substantial rise in source of the differences among strategies. the share of industry and-except for a few specialized mineral producers-a shift away from dependence on primary exports to- I. The Transformation of Production ward manufactured goods as a source of and Trade foreign exchange. There is considerable evi- dence that success in developing manufac- The structural transformation of develop- tured exports is critical to this process, and ing countries is characterized by a period in ccarversely that continued emphasis on im- which the rising share of manufacturing in port substitution will ultimately lead to a GNP approaches that of primnary production slowing down of growth. and a significant portion of manufactured Despite the amount of attention given to goods begins to be exported. Countries that alternative strategies of trade and develop- have reached this stage have been alterna- ment since the work of I. M. D. Little, Tibor tively described as "semi-industrial" or Scitovsky and M. F. CT. Scott, there has "newly industrialized."' Depending on the been little attempt to examine the under- criteria used, there were between twenty and lying relationships in quantitative terms. A twventy-five such countries by 1970. fuller understanding of the various mech- The present sample consists of seven anisms that have been posited requires that of the sixteen principal semi-industrial the internal and external aspects of industri- countries identified by Bergsm an: Korea, alization be examined together in a frame- Taiwan, Colombia, Turkey, Yugoslavia, work that brings out the several interactions Mexico, and Israel (in ascending order of among them. 1960 per capita income)?2 Japan and Nor- This paper is drawn from a comparative way, which had largely completed the trans- study of sources of industrial growth in formation of production by 1960, are added selec ted semi-industrial countries that have for comparative purposes. The sample was followed policies ranging from the extremes selected primarily on the basis of the availa- of export promotion to import substitution. bility of input-output data covering fifteen The core of the analysis is a set of input-out- years or more. Table 1 gives selected struc- put a.ccounts that permits changes in the tural characteristics for the nine countries, structure of demand, trade, and production to be analyzed in comparable terms over 'Joel Bergsman identifies sixteen significan~t semi- periods of fifteenI to twenty years. 1 will industrial countries, ten of which are the subject of the recent OECD study of "The Impact of the Newly compare the main effects of different types Industrialising countries." of trade and development strategy on in- 2The remaining nine countries identified by Bergs- d us trial growth and structure. The meth- man are: Egypt, the Philippines, Brazil, Portugal, Hong Kong, Singapore, Greece, Argentina and Spain. Mar- ginal cases include India, Uruguay, Chile, South *World Bank. Africa, Thailand and Malaysia. 281 282 AMERICAN ECONOMIC ASSOCIA Tf ON MA Y 1980 TABLE 1-INDIcATORS OF STRaUCrUR AND GROwrH Per Capita GNP Share of GYDP Average Value-Added Population Level Growth Manufactured in (millions) (US$1970) Ratea Exportsa Exports* Industrya,b Group A Korea 1955 22 131 - 1.6 0.2 13.1 1963 27 149 1.6 4.9 1.2 16.9 1973 33 323 8.0 31.8 24.3 29.7 Taiwan 1955 9 199 - 8.3 1.4 23.6 1963 12 252 3.0 17.6 6,2 28.6 1973 15 513 7.4 51.6 38.3 43.8 Israel 1955 2 950 - 11.5 4.9 31.6 1963 2 1429 5.2 21.4 11.2 35.5 1973 3 2374 5.2 28.3 15.1 36.7 Norway 1955 3 1244 - 40.7 10.4 35.2 1963 4 2168 7.2 39,0 12.6 33.1 1973 4 3179 3.9 43.4 19.1 30.4 Group B Yugoslavia 1955 18 329 - 6,6 2.0 41.7 1963 19 510 5.6 15.6 11.8 40.8 1973 21 813 4,8 22.3 12.5 41.4 Japan 1955 89 500 - 10.7 9.1 26.5 1963 97 992 8.9 9.3 7.4 40.8 1973 108 2349 9.0 10.3 8.8 42.5 Group C Colombia 1955 13 285 - 12.4 0.2 19.3 1963 17 309 1.0 11.9 0.5 23.2 1973 23 415 3.0 15.5 3.7 24.9 Turkey 1955 24 264 - 5.2 0.2 16.9 1963 30 319 3.7 5.9 0.3 19.0 1973 38 461 3.8 8.1 1,5 24.5 Mexico 1955 31 424 - 16.7 3.8 26.8 1963 40 513 2.4 10.4 1.7 27.2 1973 56 719 3,4 9.2 3.0 31.1 Source: Sce Kubo and Robinson. aShown in percent. bIndustry includes manufacturing and construction. A. The Sources of Growth The model is based on the following accounting balances for each sector: An input-output model is used to provide a consistent framework for the analysis of (1) X= u,( W + D)+ E, growth and structural change. The same twenty-three-sector classification is used for (2) M,=m( +D) each country, which leads to a comparablem( +D) decomposition of output growth in each sec- where X is total output, D is domestic final tor into the direct and indirect effects of demand, W is intermediate demand, M is increases in domestic demnand, exports, and imports, E is exports, m5 is the share of import substitution. For this purpose import imports in total supply, and u1 is the substitution is defined for each sector by the domestic share. Assuming that W, = aX reduction in the share of total supply that is the level of output can be expressed by provided by imports. the solution to the corresponding Leontief VOL. 70 NO. 2 T RADE SH AR ES A ND ECONOMIC GROH TH28 model as direct effects of these policy differences on (3) X4=2 ry( uD +E) production are shown most cleaily by changes in the share of manufactured ex- The coefficients rj are the elements of the ports, which are given in Table 1. Since the inver'se of a Leontief domestic matrix in sample illustrates a wide variety of develop- which the coefficients (u,ag) represent the ment patterns, there is little difficulhy in amount supplied from domestic sources. dividing the countries into three groups on Equation (3) makes it possible to solve for this basis. Group il: Countries with high or the increase in output of each sector, AX,, in rapidly rising manufactured exports: Korea, terms of increases in internal and external Taiwan, Israel, Norway. Group B: Inter- demand in all sectors (XD~ and å E3) and mediate cases: Yugoslavia and Japan. Group changes in two sets of parameters (Aug and C: Co~untries with low manufactured ex- a).The solution for AX, can be expressed por's: Colombia, Turkey and Mexico. as the sum of four factors:3 In the two intermediate cases, manufac- tured exports rose rapidly before 1960 but (a) The expansion of domestic demand in maintained a relatively high and stable all sectors (DD): share of GDP thereafter. 11 The four sources of growth of all manu- 3 uVf'Dk facturing for these three groups of countries are given in Table 2. The subperiods are five (b) The expansion of exports in all sectors to ten years, depending on the availability (EE): of input-output data. They extend from the mid-1950's to the early 1970's, except for r JåAEJ Japan where it was possible to make ap- proximate calculations for the prewar (c) Import substitution in all sectors (IS): period. In all countries except NorwNay, the data cover part of the initial period of im- 3 r du~ (D/2+ W4/)) port substitution, which is particularly not- J able in the analysis of Korea, Taiwan, and (d) Technological change (TC): Colombia. Thereafter the patterns diverge A~X~ substantially. 2 r 2 d XZIn the four countries in Group A, the j k growth of manufacturing is increasingly due The effects of trade policy are shown by to the continued expansion of exports, terms (b) and (c), export expansion and which accounts for 50 percent or mnore of import substitution. When there is no the total increase in output. In Korea and change in import proportions or in input- Taiwan, export expansion led to a rapid output coefficients, the last two terms acceleration of industrial growth; but in vanish and sectoral growth is determined Israel, Norway, and Yugoslavia the demand only by increases in internal and external effects of export expansion were largely demands. offset by import 1liberalization. The countries in Group C are typical of a B. The Role of Trade larger group (which includes India, h>razil, Chile, Uruguay, and Argentina) whose de- Trade and development strategie9 are velopment strategy has been based on im- of ten characterized by a spectrum varying port substitution for several decades (see from inward to outward looking or from my book with Syrquin, Table 16). The de- "import substituting" to "export led." The composition of the sources of manufactur- 3This formulation is discussed in the paper by ing growth shows that export expansion was Moises Syrquin and myself. The superscripts refer to the smallest of the four factors, accounting time periods. for less than 10 percent of the total increase. 284 AME RICA N ECONVOMIC ASSOCIA TION MA Y 1980 TAntsiL 2--SOURCES OF GR0WTH 1N MANUFACTUiNo OurPUsr Average .Percent of Total Annual Domnestic Changes in Growth Demand Export Import Input-Output Period Rate Expansion Expansion Substitutior. Coefficients Group A Korea 1955-63 10.4 57 12 42 -11 1963-70 18.9 70 30 0 0 1970-73 23.8 39 62 -3 2 Taiwan 1956-61 11.2 35 28 25 12 1961-66 16.6 49 44 2 5 1966-71 21.1 35 57 4 4 Israel 1958-65 13.6 62 27 13 -2 1965-72 11.3 71 49 -37 17 Norway 1953-61 5.0 65 36 -16 15 1961-.69 5.3 51 58 -19 10 Group B Yugoslavia 1962-66 16.6 74 25 -5 6 1966-72 9.1 72 38 -22 12 Japan 1914-35 5.5 70 33 5 -8 1935-55 2.8 71 --7 15 21 1955-60 12.6 76 12 -3 15 1960-65 10.8 82 22 0 -4 1965-70 16.5 74 18 - 1 9 Group C Colombia 1953-66 8.3 60 7 22 11 1966-70 7.4 76 5 4 15 Turkey 1953-63 6.4 81 2 9 8 1963-68 9.9 75 5 10 10 1968-73 9.4 71 16 -2 15 Mexico 1950-60 7.0 72 3 11 14 1960-70 8.6 86 4 11 -l 1970-75 7.2 81 8 3 8 Source: See Table 1. Colombia, the least industrialized of this doing so, some of the differences in trade group, illustrates the typical pattern of de- policies will be noted. clining effects of import substitution with no The seven developing countries all experi- offsetting rise in export effects. Af ter import enced some degree of balance-of-payments substitution is largely completed, manufac- disequilibrium during the 1950's, reflected in turing growth cannot exceed that of do- foreign exchange shortages and quantitative mestic demand and therefore tends to de- restrictions (QR?s), and exacerbated by over- cline until there is a cha nge in trade policy. valued exchange rates. The trade and ex- change-rate regimes of four of these II. Trade-Development Sequences countries-Korea, israel, Colombia and Turkey-.have been compared to a larger The previous section established a rough sample by Anne Krueger. While Korea and grouping of countries based prinmarily on Israel show progressive liberalization and the role of manufactured exports in the reduction of QRs by the early 1960's, structural transformation of the economy. I Colombia and Turkey maintained high will now examine the differences in develop- levels of protection and import substituting ment-trade sequences at a less-aggregated policies for most of the period. Both the level to ascertain the extent to which they latter had intervals of liberalization and ex- vary among countries and industries. Before port expansion in the late 1960's. VOL, 70 NVO. 2 TR ADE SH A RES AND ECONOMIC GROWT H 285 9 0 - ' .0 0 - t0 - O - to 1. ovt 20 o 10 too 20o 300 200 100 0 1000 lt 2000 lOt0 2000 3000 totrto'tt 60 ttl 0 21 0 3t10t54 2009 t 119tO 1t 4 1 l2012 2 0 1104 202 t45 01 • - Gro th ate(Pe ent eothO Rafe 005er0ent0 tiop ooris(otupoto O toot est taoupot 0 l'oOtrOIsonp 16 89 RB' 11 lo8 1atns4 x s210 t ' 4 x a t1o ms a 01 1msaDmn -E tyotepn m l mu httto 111 Un ni \lean i-i pr a en i m o w e Taiwan followed Olooaot t seqenc siia ttahiey Lgtidutyinldsetr strategy wit Sreivl OF ees f L potec- stage 2h Sthre prncpa soures of growth extndsummay,hth coprison toicie othe woodas inrouts,utcas (b) Tabe2ay idsown countries in orsm.I Group A(lsJpn)atvl (cheiures, me2, andt3rorm ec.) and (c)s Kaored xporainced trae erey m960's fuilye i wichts.ad n aco roo thos 190.I Group C, Mxicoriminated aas toutwrarLoin deeopenes wie Grna- moderatematorm ofsimf orttsdbstitutionnflhinerycl ypicanlymp evesuopsiattaoneaadirepydlris oftfraeg wisteaticeto orevesian of andec stagexprt The aherna source s of growth, a Ine whichmadte possibe hoigher goth indcased in oTable (as Ti Table 2)s earler hon coutesn morouward-lusokipn) tradean inmoresnouce 2,ind forh indutry ofwhese faveopent pohice thimp 19s wie scortsubttto solysgiiati The indicareyth diferees In tadditon the Afirstri, a setoreshowsace atei ofpfoeint ssuence amon seoreasan, andur- machinr. ase soffrcences aso persit was teen branches of manufacturing in our mod- less-aggregated levels. There is a corre- els have been aggregated into three groups: sponding change in the pattern of exports in 286 A MERIC A NECONOMIC ASSOCIATION MAY 1980 timing, the inward-looking countries of Group C indicate the effects of the exhaus- /11 \tion of import substitution possibilities in all 1 sectors. The failure to develop manufac- - tured exports (except on a modest scale in \~ light industry) has led to the decline of the ;X ' rate of growth shown in both light and heavy industry, but not yet in machinery. Even though Colombia, Turkey, and 1"*x.i~ 1 ~ ~Mexico are relatively large countries and urowt Rateenenexpansion of the domestic market has not offset this failure. A more detailed analysis *~0~ 1shows these countries lagging particularly in J~2Smachinery and metal products, sectors in -which the countries in Group A have had 1'N.un above average growth. 40/ In summary, the general features of the \ inward-looking pattern carry over to each of 20 1 \ the major sectors, as in the case of the 0 on export-led strategy. While the opportunities for import substitution persist longer in ;; i ~ IL~ ~ 1 ~--~ heavy industry and machinery than in light 4004 1000H40 industry, its ultimate decline is similar. " " "II. Concluding Remiarks FIGURE 3. SOURCES oF~ INCRAES I MACHThERY (PBRCENT) This paper illustrates an approach to the analysis of structural change in which inter- nal and external factors are treated together each country (not shown) with a growing in an interindustry framework. The method component of heavy industrial products. is adapted from techniques used in develop- Despite these differences in degree, the ment planning and takes advantage of infor- major impression from these comparisons is mation collected for this purpose. It can be a general similarity in the trade-develop- extended to encompass production func- ment sequences of each sector in the three tions and factor use by sector in order to developing countries in Group A, After an provide a more complete analysis of the initial period of strong import substitution, sources of growth and structural change. export expansion became the major source The effects of trade policy on industrial of industrial growth in Korea and Taiwan in structures that are revealed by this analysis each major sector, and also led to an accel- are quite striking. Although import substitu- eration of growth in each. In Israel, how- tion is an important feature of early stages ever, exports predominated only in light in- of industrializa tion in all developing dustry. countries, it can be accelerated or retarded A final phase of import liberalization in by trade policy. The later stage of expansion each sector is shown by Norway, Israel, and of manufactured exports is more susceptible Yugoslavia. In this phase export expansion to policy influence and is shown to have a is partly offset by increased imports, so that large effect on the subsequent course of the rate of sectoral growth is determiined industrial development. primarily by domnestic demand. This methodology can also be used to Inward-Looking Sequences. Although explore the effects of the balance-of-pay- there are some significant differences in ments conistraint on the pattern of develop- VOL. 70) NO. 2 TR ADE SHA RES AND ECONOMIC GROWTH 287 ment. In this context the development of Industrialized Countries," staff work. manufactured exports appears even more paper no. 351, World Bank 1979. important as a source of foreign exchange H. B. Chenery and D. B. Keesing, "The Chlang- than as a source of demand because it pro- ing Composition of Developing Country vides one of the principal means of exploit- Exports," staff work. paper no. 314, ing comparative advantage and of avoiding World Bank 1979. balance-of-payments bottlenecks. Hllis B. Chenery and Moises Syrquin, Patterns To complete the linkage between industri- of Development 1 950-1970, London 1975. alization and export growth, it would be ____and ___, "A Comparative Analy- necessary to examine the changes in com- sis of Industrial Growth," in R. C. 0. parative advantage that result f'om the ac- Matthews, ed., Measurement, H-ist'ory and celeration of growth and learning by doing Factors of Economic Growthi, New York in successful export-led strategies. This pro- 1979. cess, which is explored in my paper with Anne 0. Krueger, Foreign Trade Regimes and Donald Keesing, has been a major factor Economic Development: Liberalizaionz behind the growing share of the semi- A ttempts and Consequences, New York industrial countries in world markets for 1978. manufactures. Their success in exporting Y. Kubo and S. Robinson, "Sources of In- manufactured goods has in turn contributed dustrial Growth and Structural Change: to more rapid industrial growth in a cumu- A Comparative Analysis of Eight Coun- lative process. tries," mimeo., World Bank 1979. 1. M. D. Little, Tibor Scitovsky, and M. F. G. REFERENCES Scott, Indus/ty and Trade in Some Develop- ing Countries: A Comparatuire Study, Ox- J. Bergsman, "Growth and Equity in Semi- ford 1970, Reprinted from THE AMERICAN ECONOMIC REVIEW © The American Economic Association THE WORLD BANK Headquarters: 1818 H Street, N.W. Washington, D.C. 20433, U.S.A. European Office: 66, avenue d'Iéna 75116 Paris, France Tokyo Office: Kokusai Building, 1-1 Marunouchi 3-chome Chiyoda-ku, Tokyo 100, Japan The full range of World "Jank publications, both free and for sale, is described in the World Bank Catalog of Publications, and of the continuing research program of the World Bank, in World Bank Research Program: Abstracts of Curreni Studies. The most recent edition of each is available without charge from: PUBLICATIONS UNIT THE WORLD BANK 1818 H- STREET, N.W. WA SHINGTON, D.C. 20433 U.S,A