rjis paper is prepared for staff use and is not for publica tion. The views are those of the 0 authors and not necessarily those of the Bank. INTRJNATIONA.L BANK FOR RECONSTRUCTION ANM DEVELOPMENT Bank Staff Working Paper No. 164 October, 1973 Protection and levelopment This paper was written for presentation at the Rehovot Conference on Economic Growth in Developing Countries, held in Israel in September 1973. Its purpose was to summarize the debate on the "pros and cons" of protection as an instrument of industrialization policy in developing countries for policy makers. The paper argues that while there are strong arguments for moderate levels of protection as an instrument of industrialization, in the past many countries have used protection excessively with counterproductive economic and social effects. This does not, however, mean that some measure of protection, varying with a country's condition, may not be desirable. Japanese experience suggests that under suitable conditions protective instruments can con-tribute to industrial growth. Protection cannot be viewed in isolation from other development policies. A strategy designed to improve the. standard of living of lower income groups would also require relatively low levels of protection, avoiding the excess costs encountered in the past. Prepared by Helen Hughes Economics of Industry Division Development Economics Departnent Suwimary and Conclusions Recent years have seen substantial studies of import substitution and protection in developing countries, and stidies now being undertaken in addition analyse the effects of export subsidies on manufactured exports. Most of these studies, however, deal with countries and periods characterized by import substitution "at all cost" and by excessive protection. The lessons which can be drawn from Japan's experience where high protection coincided with rapid industrial development, have largely been ignored. Yet protectionist arguments are still very strong in developing countries and they find the Japanese model very attrac&ive. The case for protection rests on "infant industry", diversification and balance of payments, employment and "trade retaliation" arguments. In addition, industrialization often has a high value in itself. The costs of protection are principally reflected in the misallocation of resources, inefficiency in the conduct of business, unfavorable balance of payments effects, and poor employment results. Many of these effects, however, are due to excessive protection, the maladministration of the relevant policy instruments and associated measures such as capital subsidies and tax holidays, rather than necessarily to protection as such. The products for which most developing countries have provided the highest effective protection in the past are, furthermore, those associated with an overall development strategy which favors upper income groups. A development strategy focused on the needs of the lower income groups would be likely to stimulate an industrial structure v,which would need relatively low protection, and hence would avoid the costs which have in the past been associated with high protection. -all- - -> -1 - - - -.. Z4! 1,@ . . i: .- - - , - Pa: : -- w.M .... . -.>. ;s*:J;.e... L- .er Protection and Deveopmnent The dangers of import substitution and protection in developing countries have received coniiderable attention in recent years. The evidence, however, is rather one sided. The empirical studies in the main deal with countries and periods characterized by import substitution "at all cost" and by excessive protection. 1/ Japar, the outstandiing example of rapid and successful development, has been largely ignored, although a barrage of protectionist instruments which culminated in an almost total prohibition of competing imports (and of foreign investment) was an integral part of the policy mix of the 1950s and 1960s rapid growth period. Studies of protection, moreover, have tended to be fuzzy about the objectives of development, to regard trade policy as the dominant, if not the only policy factor in development, and to neglect issues of policy implementation. The cost of protection of exports - of export subsidies - is only now beginning to receive attention. 2/ It is therefore not surprising that while there has been a reaction against the extreme import substitution ideology of the 1950s 1/ The three principal groups of studies are, of course, those organized under the auspices of OECD and smmarized in I.M.D.Little,4 M.Scott and T. Scitovsky, Zidustry and Tade in Some Developin Countries, those carried out with IBRD financing and reported in B.Balassa and Associates, The Structure of P-rotection in Developin Countries, 1971, and those now being carried out by the National Bureau of Economic Research, with initial results reported by J.N.Bhagwati and A.D.Kruger in "Exchange Control Liberalization and Economic Development", American Economic Review, MaY 1973. 2/ B. Balassa's current comparison of trade policies in six semi- industrialized countries includes an evaluation of export subsidies; the NBER project covers export promotion instruments, and other studies are under way at the Kiel IEnstitute of World Economics. -2- and early 1960s, protectionist views persist. Developing countries are keenly interested in the Japanese development "model", and some have successfully begun to adapt it to their needs. The infant industry argument for some measure of protection is still strong in "least developed" countries, although those which have made some progress in industrialization, and those which may be described as semi-industria:lized, are moving from protection for import substitution to subsidies for exports. The emphasis on the social objectives of development, and particularly on employment, is strengthening the employment argument for protection. In sum, a review of the policy implications of the protection - free trade debate is timely. 1/ I. The Case for Protection The justification of protection dates back to nineteenth century attempts by French, German and American entrepreneurs to catch up with those British manufacturers who had either outpaced them, or were about to do so. 2/ The "infant industry" argument has remained the core of the protectionist case, repeated, fortissimo, as one country after another sought to begin manufacturing. The economic justification of providing public subsidies to offset an entrepreneur's costs on account of future 1/ F.Diaz-Alejandro, "Trade Policies and Economic Development", and W.M. Corden, "The Costs and Consequences of Protection: A Survey of Empirical and Theoretical Work", (Papers contributed to the Conference on Research in International Trade and Finance, Princeton, March 1973) have reviewed the "state of the art" as economists perceive it; this paper is therefore confined to policy issues. 2/ 'The protection argument is usually, though not necessarily, applied to manufacturing production. The same case can be, and often is, made for the production of primary commodities such as sugar and rice. 33 - social gains remains basically unchanged, but recent experience has brought some revision of thought on "first best" instrumeDts of policy. A subsidy which would meet the costs of "infancy" directly is in practice likely to be so costly to collect and administer, that a less direct instrumaent of assistance, usually a tariff, may turn out to be the "first best" solution. In practice, in any case, political feasibility not economic optimality is the final determinant in the choice of instrunents. Governmients may be generally unduly conservative in judging what is politically feasible: unfortunately, representative government and political feasibility appear to be inversely related. Typically then, the "infant industry" case leads to tariff protect ion. Further arguments have been added to the "infant" industry case over the years. The first, dating back to the 1920s and 1930s, but greatly strengthened when it was revived after World War II is the "terms of trade" case for diversification and the protection of the balance of payments. The argument was stated principally in import substitution terms. 1/ Countries which were dependent on one or two agricultural exports with an elastic supply and inelastic demand experienced declining terms of trade in the 1920s and 1930s. They had their imports of industrial goods interrupted during World War II, and feared a repetition of the 1930s' experience after the wartime boom. Balance of payments considerations and the need for 1/ R. Prebisch, Towards a Dynamic Development Policy for Iatin America, New York, 196l7, and S. Macario, "Protectionism and Industrialization in latin America", Economic Bulletin for Latin America, March 1964. Import substitution has subsequently been more closely defined and measured. -4- industrialiL.ation were thus linked in theory; they were to become raore closely linked in practice through import licensing and other quantitative trade controls as balance of payments problems reasserted themselves in the late 1940s,. Diversification through industrial development seemed also to be relevant to the solution of growing unemployment problems and the low living standard in developing coutries. Whereas earnings in agriculture and agricultural countries were low, earnings in industry and in industrialized countries were high, promising rapid increases in wage earnings. 1/ The "trade retaliation" argment for protection began to evolve in the 1930s, gaining momentum in the 1950s and 1960s. While some developing countries began to industrialize at the turn of the nineteenth century and during World War I, the first widespread "third world" wave of industrialization came in the 1930s, coinciding with the industrialized countries' move toward protectionism. At the time of the next and principal move toward industrialization, in the post World War II period, monetary and trade restrictions were widespread. There has been very considerable monetary liberalization since the 1950s, but many primary product markets are still restricted, and trade liberalization in industrial goods has occurred principally in goods traded among industrial countries. Developed country tariff protection against key products in which developing countries might be expected to have a comparative advantage - processed raw materials and labor intensive goods such as textiles and footwear - has been retained, and in many instances tariffs have been bolstered by 1/ Employment/wage arguments for protection also date to the 1920s. M.C. Samu.elson succintly summarized the case in "The Australian Case for Protection Re-exarined", Quarterly Journal of Economics, November 1939; see also W.F.Stolper and P.A.Samuelson, "Protection and Real Wages", Review of Economic Studies, 1941. quantitative restrictions. Recently even sub-contracted labor intensive exports such as electronic components and assemblies have come under threat of protection in the United States. The much vaun4tid, but rather hollow genieralized preference schemes do not include "sensitive" products. 1/ During most of their industrializing period the developing countries have thus been faced with an international economy which has been far from free trade. The international economy is, moreover, not competitive. fluperfect competition and product differentiation are characteristic of many manufactured product markets. Strong oligopoly groups were beginning to take a commanding role in the international economy before World War II, and in the 1950s and 1960s they considerably extended their power. The developing countries have noted that Japan'!s emergence as a major industrial power with its own 'multinational' corporations has been the principal broadening influence. Discriminatory pricing followed imperfect competition. Developing countries have long been familiar with variatioLs in f.o.b. prices by country of destination, according "to what the market would bear". 2/ Such pricing practices made the evaluation of comparative advantage difficult, and led to temporal price instabilities which greatly magnified the difficulties of planning. It is true that some developing countries were able to benefit by buying "marginally" priced goods, in times of excess capacity bot;h for their own consumption and as inputs for exports. However, 1/ D.Wall, "The Commonwealth and the Generalized System of Preferences", A Study prepared for the Commonwealth Secretariat, 1972,(mimeographed). 2/ Colombia, in an examination of pharmaceutical prices, is however one of the few countries to have undertaken a systematic study of developed country pricing policies. -06 in recent times of boom and scarcity, they have found that their inputs costs have risen disproportionately, and sanetimes inputs are simply not available. Domestic consumers and the volume of exports have suffered. In addition to "marginal" pricing moreover, there have been clear cases of classical "dumpingtn, that is the temporary lowering of prices to prevent import substitution and establish dominance in a market to raise prices. The oligopolistic nature of the international economy is particularly marked in relation to developing countries' ability to enter manufacturing. There are two principal obstacles to entry. Multinational corporations were frequently able to establish a consumer market for their branded differentiated product before a country's industrialization was begun. Internationally advertized products continue to enjoy an advantage in most national markets, and developing cnuntries are no exception. Local manufacturers of soap, toothpaste, and food products usu-ally have great difficulty in penetrating such a market. Second:ly, the oligopolistic na-ture of industry is particularly marked in the market for sophisticated technology. In a number of fields such technology can only be obtained as part of a package deal with capital and market participation by a multinational corporation, so that independent entry by local producers is difficult unless a government takes very strong protective measures. Internal communications, shipping and insurance services are no more competitive than manufacturing. Shipping freight rates are notoriously set by regional car-Lels. The building up of developing country shipping fleets, however costly and inefficient in its early stages, is a protectionist reaction against cartelization. Japan had a long struggle -7- before it was accepted by the shipping "conferences". Since the 1930s developed countries have been building up complex export subsidy systems which further distort international trading. Nationalist neo-mercantilism, and "marginal pricing" for export markets gained respectability when GATT was persuaded to make drawbacks of indirect taxes a legitimate form of export subsidy. This introduced some extremely dubious concepts into international trade practice. The proposition that indirect taxes fall on consuners rather than producers, whereas direct taxes fall on producers, was given credence. 1/ Substantial credit subsidies to exporters by leading industrialized and European socLalist countries further diversified "international prices". Finally, the strongest argument for protection, largely unaffected by the evolution of economic thought, is also the oldest. A measure of self sufficiency is desirable for its own sake, as a symbol of newily acquired or asserted independence. Industrialization always was - and still is - a consumption good with a high marginal utility to the body politic. The case for subsidizing exports, particularly of expor'u of manufactures, is essentially an extension of the "infant industry", diversification and balance of payments, employment and "trade retaliation" arguments. There is no analytical distinction between producing for domestic consumption and for exports. Similarly the dichotomy between "linward" and "outward" orientation is not concerned with import substitution veras exports, but rather with the difference betwleen maintaining 1/ The implication that consumers of imported goods do not have to bear the socia.l costs of producing them has curious implications for international resource allocation. It suggests that all consumers will be better off if they always buy imported rat,er than locally produced goods -8- international standards of efficiency in production and ignoring such standards. It is interesting to note that the political and administrative difficulties of paying direct cash subsidies to producers have been overcome in many instances with regard to exports. Whereas it has not yet been possible to evolve practical forms of direct subsidy to producers of import u-abstituting products, 1/ there are many instances of cash or near cash payments to exporters of manufactured products. In some of these cases at least part of the subsidy may be passed on to consumers of other countries. The mercantilist strains underlying nationalist thinking are very st.rong indeed. II. The Case Against Import Substitution and Protection The central argument against protection is concerned with the distortions it creates in the exploitation of comparative advantage, and hence in the allocation of resources. The greater the distortions, the higher the social costs of protection. The development of effective protection, domestic resource cost and other shadow price measurement techniques has provided some basis for comparing the extent of distortions amongst productive activities and countries. 2/ The measurement concepts are limited by their static and partial equilibrium nature. There are 1/ A direct subsidy to labor training has long been advocated as a "first best" alternative to tariffs and other forms of protection. There have been suggestions for tariff aum subsidy arrangements to overcome the budgetary problems of pure subsidy arrangements. These proposals have not been found practicable to date. 2/ Paradoxically however the notion of an "international price" has become more and more elusive just as economists evolved these methodologies and methods of individual project evaluation. -9- very considerable data problems, which mean that the resulting estimates have to be interpreted 'very broadly indeed, mere1y as orders of magnitude. As measurement is in terms of market prices, the reference firm is the marginal firm, for this is the firm that determines domestic prices. This usually means that reducing protection, even substantially, would not mean eliminating an industry, though the marginal firm, or firms could be forced out of business. 1/ A higb rate of protecticn thus does not necessarily mean that an industry is inefficient per se, but rather that it is using resources inefficiently, not necessarily merely, or even largely as a result of trade policies. However, an industry which might under liberal trade and other, more optimal conditions enjoy a comparative advantage, might thus require a very high rate of protection. It has been well established thax6 protection, and particularly high prctection, leads to inefficiency in the conduct of business. In the absence of foreign competition, domestic competition may be too weak to ensure competent entrepreneurial and managerial performance. Low levels of production typical of early industrialization are frequently characterized by monopoly and oligopoly., and even if there is co±nipetition, it is likely to be weak in comparison to external competition because all the competing firms are "infant" firms. 2/ Protection has other;i costs. Contrary to expectations, import substitution with high protection, has tended to exacerbate rather than improve the balahce of payments. The high costs and prices of highly protected manufacturing have tended to discriminate against traditional and 1/ c.f. A. Kreuger, "Some Economic Costs of Ecchange Control: the §larkish Case", Journal of Political Economy, October 1966. 2/ H. Leibenstein, " Allocative Efficiency vs. X-efficiency", American Economic Review, June 1966. - 10 - other non-manufacturing exports, hindering their expansion. High prices have tended to limit the domestic market, limiting import substitation by raising the costs of backward integration into intermediates and capiual goods. Balance of payments management problems grew because imports of intermediates could not be reduced or cut off in lean times, as imports of non-essential final products had been in pre-import substitution days, because urban employment was dependent on imported inputs. High costs made the export of manufactures difficult if not impossible, limiting opportlunities for export income earnings, particularly in labor intensive products in which developing countries were likely to have a comparative advantage. The case against import substitution "at all cost" was recognized in the 1950s and early 1960s when Israel, Taiwan and Korea began to subsidize manufactured exports to offset both the relatively high domestic costs of produaction and the higher profits to be made under protection which encouraged manufacturers to produce only for the domestic market. Singapore joined the stream, but having had little protection, it was able to use a fairly simple, low cost export incentive package. Pakistan was the most spectacular, if not most successful, convert to protection for exports. By the beginning of the 1970s subsidies for exports were commonplace amongst countries which had previously foLlowed high protection principally for import substitution. Protection has had social as well as economic costs. High costs and high prices limited the domestic market, reducing the potential for industrial growth and hence for employment irn industry. Employment creation was in any case more limited than it might have been because protectionist instruments such as import duty rebates on capital goods stimulated capital intensive production methods. Import licensing and foreign exchange control implementation has tended to favor large and capital intensive producers. Foreign and local funds have been often available to privileged, generally large scale borrowers at low, and sometimes at negative interest rates, through development banks. Other forms of incentives associated with protection - income tax- holidays, subsidized land and public utilities - have also tended to favor large scale and capital intensive producers. Multinational corporations benefited particularly. Small producers, who would have tended to use higher domestic imports and more labor inteansive production methods, were placed at a disadvantage. Attracted by the multipurpose possibilities of such instruments as tariffs and import licensing, many countries have created complex, unwieldy administrative systems. Most countries are not satisfied with the trade protection afforded manufacturing industries by trade policy measures, but subsidize manufacturers further by giving them income tax holidays of varying duration, and by subsidizing credit, land and public utilities. The high costs of these policies have now been well established. They include a further layer of protection to industry, a stimulus to capi-tal intensiveness, and in the case of foreign investment, frequently transfers of revenues to the government of the lending country. The benefits are known to be small. Yet developing countries throughout the world continue to compete with each other in the extent and duration of such give-aways. 1/ 1/ See E. Vernon, o nty at Bay, 1971 and H.H{ughes ,"Assessment of Policies Towards Direct Foreign Investment in the Asian-Pacific Area", in P. Drysdale, ed.,, Direct Foreign Investment in Asia and the Facific, 1972. - 12 - Ihe manufacturers' opportunities for profits through the manipulation of the "rules of the game" often exceed those which may be earned in the course of business. Access to government offices has accordingly become an important locational factor, contributing to the tendency towards geographic concentration in large urban centers. The multinational corporation with its superior managerial,legal a3id other resources, which at times include political pressure through its own government, is usually in a very strong position to benefit from such "rents", and does so accordingly. It pays entrepreneurs to spend more time in Ministers' waiting rooms than on the shop floor. The developed countries' industry lobbies are sonn imita-ted and extended, albeit usually on a smaller scale and on a more personal basis. Subsidies for exports have similar effects. They are sometimes charged through high domestic prices, taxpayers contribute to cash subsidies, or public services and irnvestment are reduced by revenues foregone. In the East Asian countries the subsidies have been concentrated on labor intensive, low skill and low capital "cheap labor" industries and this has tended to stimulate the growth of local employment albeit in low productivity, low wage occupations. Singapore followed more enlightened policies seeking to build skill industries as soon as a stop gap low wage industry program gave it breathing space to do so, and Korea is now planning on these lines. In Latin America, however, export subsidies have tended to go to industries which are relatively capital intensive. Multinational corporations have frequently been best able to exploitJ export subsidies, directly by taking advantage of incentives, and by transfer pricing which maximizes tax benefits. - 13 - In some countries strong trade unions have been able to share in the economic "rents" of protection through relatively high wages and fringe benefits. To the extent that such a share indeed came out of profits, and where extended family systems distributed such earnings fairly widely, this was probably a more efficient form of redistribution of profits than company income taxes might have been. High wages, however, also contributed to capital intensiveness; where they were additional to high profits, they raised prices, reduced international competitiveness and hence employment opportunities. The case against import stubstitution oriented protection stated, many tiresome questions remain. Japan's experience is the most obvious. Protection became important in the 1930s 1/ and it only began to be dismantled in the late 1960s, 2/; aided by strong cultural 'natural' protection, much still remains. On a smaller scale, but consciously following the Japanese "model", Iran showed steady and rapid growth with high protection from the early 1960s. The contribution of manufacturing to growth was more important than petroleum until 1970, domestic prices of rnass consumption goods were close to "interna-tional" prices by 1971, and with the first import substitution phase well established, manufactured exporbs were by then growing at 40 per cent per annum. 3/ 1/ H.B. Chenery, S. Shishido and T. Watanabe, The Pattern of Japanese Growth, 1914-1954", Econometrica, January 1962, estimated that some 38% of the increased share of industrial output in total output was due to import substitutLon between 1914 and 1954 (p. 118). 2/ R. Komiya, "Japan's Non-Tariff Barriers to Trade in Manufactured Productsr , in H.E. English and K.A.J. Hay, eds., Obsbacles to TBade in the Pacific Area, Ottawa, 1972. 3/ J. Hansen, "Industrial Developmenb in Iran, 1960-1970", forthcoming. - 14 - The Japanese experience needs to be compared in detail with other cases of "protection" to sift out those elements which lead to "success" from those which spell disaster. It is clearly possible to combine protectionist policies with internal competitiveness and significant, levels of exports, and to administer even complex protective measures effectively so that a highly protected economy can at the same time be very sensitive to international economic trends. Thus it seems that it is not protection as such, its level or the variations in protection among industries which are the crucial factors, but rather the way in which various protectionist and other developmnent policies are combined and administered which determines the outcome of a protectionist developmenb strategy. III. Trd PoiyOtos3o Deloigounltries 'The range of policy options for developing countries is so broad, at least conceptually, that trade policy options for "development" can only be discussed in very general terms. A country's size, geographical situation, natural resource endowment, cultural heritage, level of economic development, administrative capability and political effectiveness are one set of determining circumstances. A small country's import substi ution opportunities are much more limited Whan a large country's, partly because of resource considerations, but principally because of the limited size of its market. Small countries thus typically turn to exports of manufactured goods before large countraies at a comparable level of industrial development. 1/ Geographical proximity to 1/ Latin America has been an exception, see C.F. Diaz-Alejandro, "Some Gnaracteristics of Recent Export Expansion in Latin America", July 1973 (mimeographed), p.9. large markets is similarly an advantage for export orientation, though good shipping services can do much to offset distance. A combination of long distance and poor communications with world markets is usually a strong deterrent to a manufactured export orientation. A good natural resource base, particularly of readily exportable raw materials, broadens a country' s import substitution options. It removes or reduces balance of payrnents and resource mobilization constraints, lengthening the possible purely import substitution period. It increases the cost a country may be able to tolerate to enjoy the advantages of protection. It usually means the development of a relatively goodi physical and social infrastructure and enough commercial devolopment to ease a country's entry into manufacturing. The exploitation of natural resources may however be of an enclave nature, with a failure to develop a countrywide inLrastructure, combined with high wage expectations. Te latter, on its own, may not be critical, but it imposes constraints-on a country's industrialization options. A country's mores - its cultural heritage in the broadest sense - are factors in. the options for industrial development. The rapid develop- ment of East Asian countries in contrast even to their Southeast Asian neighbors illustrates the importance of Schumpeterian factors in industrial development and of the sociological factors influencing them. Political effectiveness and administrative capacity are important limiting factors. A country can only effectively use tVe trade policies. its administrative capacity can handle. Administrative capacity tends to vary inversely with a country's size, so that large countries, in spite - 10 - of their large potential domestic markets, in general have greater difficulties in administering complex protectionist policies than small countries. Smuggling can become a very useful safety valve in such circumstances and general "rules of the game" are likely to be more effective than attempts at close administration. A country's level of industrial development itself influences its trade policy options. Only a handful of countries, well developed in o-ther respects, and well situated, have been able to develop industries without recourse to somie protebtion. Hong Kong in fact is almost unique. Its long history as a manufacturing entrepot, shipping manufactured goods out of China, dates back to the early nineteenth century, and it became a manufacturing center in the 1920s. Ih the late 1940s it benefite4 from an exceptional influx of entrepreneurs, capital and skilled labor. Continued influx of refugees from Chi-na maintained a pressure on wages. A competent (but undemocratic and colonial) administration ensured good infrastructure and low taxes. A marginal purchaser of industrial inputs, it was able to take advantage of marginal pricing, and it was a, marginal suipplier to world markets. Nevertheless, by the mid-1900s it was facing export restrictions. Its manufacturers' reaction was to move rapidly to higher quality, more selective products, with higher returns to capital and labor. 1/ This can however be explained better in cultural than in trade policy terms. The experience of Hong Kong has been very difficult to duplicate, even in a city as similar as Singapore. In all other countries a measure of protection seems to be necessary to nurture manufacturing skills to the 1/ R. Hsia, "Hong Kong Textile Exports; A Case Study of Voluntary Restraintus", English and Hay, op.cit. - 17 - point where export possibilities can be encouraged. The problem is to determine such le-vels and to use such instruments of protection and export subsidy that initial stages of industrial development do not inhibit later industrial development and the ability -to move from import substitution to exports, 1/ and to focus on socially desirable development objectives. Past industrial development, with only a few exceptions, has catered to the development expectations of the upper income groups. The principal incentives to import substitution were not given to mass consumption goods, but to products used by middle and upper middle class consuners. Such products were more highly priced than they would have been without import substitution, but the purchasers were subsidized by high capital, salary and in some cases wage returns of the typical enclave industary sector. AA alternative strategy, focused on the production of mass consumption goods for the poor would be likely to have a quite different employment and income distribu-tion effect. In general mass consumption goods such as low quality textiles, pots and pans, building materials and simple agricultural implements require less protection than durable consumer goods and automobiles as a first layer of import substitution. Technology is in general simpler, economies of scale are less important, and there is more room for small enterpreneurs and for regional development within a country. Mul-tinational corporations do not have a comparative advantage in these product groups. Bicycles bus and truck assembly is less capital 1/ High protection for impcrts may be offset by high export subsidies but the side effects of such policies are generally very undesirable socially. - 18 - intensive than motor car assembly, providing more scope for employment, and it is better suited to relatively simple backward integration industries. Developing countries starting out on an industrialization path could begin with relatively low levels of protecticn, and those already industrializing could substantially reduce their levels of protection by shifting their product mix from a largely urban, upper and middle class orientation to one caterinig to the rural population and the urban poor. A shift in goverinment expenditures toward labor intensive public works, particularly in rural area,s, and to low income site and service and housing projects, could stimulate largely domestically based construction material industries. Cement blocks, bricks and drainage pipes can be made largely of local materials, and their bulk ensures them a measure of natural protection against imports and effectively encourages decentralized development. Such an industrial strategy, integrated with agricultural policies which raise rural productivity- and incomes, and other policies aimed at social welfare, would minimize potential conflicts between employment creation and higher incomes for lower income grops and growth. The conflicts between the interests of the onall upper income groups and the mass of the population, at least in the short run, are however, considerable. Multinational corporations also form a political interest group, particularly in countries in which they have already made investments. Changes in product mix would not only affect immediate import substitution possibilities, but also the export opportunities. The bulk of developing country manufactured expor ts have thus far gone to developed countries. For most labor intensive products demand is relatively - 19 - inelastic, and imports are already li:nited by protectionist attitudes. It is true that the export product mix from developing countries is growing rapidly., 1/ and that a greater variety is making import restrictions more difficult, but some of the "new" exports involve heavy subsidies. Because of their high mutual trade barriers, developing countries have until now failed to benefit from intra-industry trade which has dominated trade growth among industrialized countries in the last twent;y years. 2/ Intra-industry trade is generally more self- adjusting because workers in an industry can move fairly readily into another branch of the same industry, and whole industry groups are not affected. Developing countries would have much to gain from increasing their elficiency through competitiveness in products suited to their level of development. Such trade would be likely to stimulate product and production process innovatinn, broadening the choice of techniques, and making adaptation to scarce capital and low cost labor easier. The greater efficiency following from mutual trade would provide a better basis for market penetration to developed country markets for these need not be neglected in stimulating trade among developing countries. The degree of attention to developed country markets wo,iLd depend on each country's estimate of its developed market prospects. Within these general positions there is again scope for considerable maneuvering. 1/ H.B. Chenery and H.Hughes, "I[dustrialization and Trade Trends: Some Issues for the 1970s", in H.Hughes, ed.,Prospects for Partnership: Industrialization and Trade Policies in the -197s (forthcoming). 2/ B.Balassa, "Tariff Reductions and fRade in Manufactures among the ILduslarial Countries", American Economic Review, June 1966. - 20 - For all but the two exceptional city states - Hong Kong and Singapore - the issue now is thus not one of protection versus free trade. Barriers to their exports are as much a reality as their own distorting protectionist import substitution and export policies. Diefficient productive capacity has been established, but much of it is depreciated, and the problem is for the most part not how to eliminate it altogether, but what is worth retaining, and how to minimize the cost of production in such i.ndustries. 1/ Given the .,ast and present biases against agriculture because of the subsidies to industry, the problem is not how to reverse the treud, but by which policies and instruments can agricultural and industrial development become mutually reinforcing. Agreed that in the past industrial development has with minor exceptions been pursued at the cost of employment and the low income groups, the issue is not growth or employment, but how can both these objectives be fulfilled. Concluding Remarks Given the increasing complexity of industrial catching up some "infant industry" protectionism by developing countries was inevitable. Combined with the developed countries' restrictive practices, and their unwillingness to deal with the employment and income distribution adjustments which a rational international economy would require, the pressures which led to the excesses of developing countries' protectionist policies become understandable. It is generally agreed that for most developing countries reducing protection, and simplifying the protectionist instruments bolstering 1/ See S. Guisinger and D.C. Meade, "Foreign Exchange Saving as a Measure of Economic Efficiency", 1973, (mimeographed) for an elaboration of this point. -21- import substitution, would be likely, in conjunction with improvements in other development policies, to lead to sounder, more rapid and socially better oriented development. It is, however, by no means proven that high protection and complex protectionist instruments are per se inimical to rapid or socially oriented development, and the costs of export subsidies have to be borne in mind in turning toward an export orientation. For each country, in each circumstance, the optimal policies in trade and other areas which will reconcile conflicting needs and interests are unique. Moving toward such policies will take time, for policy changes have social and economic costs. Economic project evaluation techniques can be used to compare the costs and benefits of likely new projects under varying assumptions about trade liberalization and the social weights to be attached to various development objectives. It must be remembered however that such measures, and trade policies generally can only be a part, and often only a small part, of the development framework. Mhe "rules of the game" of the international economy will continue to exert an important effect on the trade policy choices open to developing countries. The rich, industrialized countries have created the "rules", they dominate trade and monetary arrangements, and they can bear the costs of change more easily than the developing countries. Theirs is the key role in improving the international trading framework.