: ' G2,6/fiFr /VcC6.,-0' POLICY RESEARCH WORKING PAPER 0~~~~~~~~~~~~~~~~~~~~~~t4 ,fA ' .-~~~~~~~~~~~~~~~~~~~~~~~~~~~~ . -'" Technical and Marketing - ;2muw~~~~~~eKifnvinmet - Support Systems -or Successful S mall and Medium-Size Enterprises l in Four Countries - -- _ -- Brian Levy with Albert Benry, Motoshige Itoh, Lins Kim,= Jeffrey Nugent, and Shzcjiro Urata _ Ihe World Bank a Policy Researc Department Fimance and P:iae Sector Development Diisioi - ; Decmber 1994 -_-_X___ IPOLICY RESEARCH WORKING PAPER 1400 Summary -findings Studies of successful small and medium-size enterprises technological capability. Demand for collective (SMEs) and their marketing and technical support mcchanisms tended to be greater when technological systems were undertaken for Colombia, Indonesia, requirements of production were comple,' or when the Japan, and the Republic of Korea. Three to four endowments of privare technological netwo rks in certain subsectors were examined in each country. Thc sample countries or industries were weak. worldwide amounted to 445 firms. Broad-based collective technical support facilitates the Mechanisms to support export marketing varied across emergence of an information-rich environment for firms, countries and subsectors How they varied depended and may be worth pursuing in many settings. Examples greatly on whether SMEs operated within well- of such support include: deveoped private networks. When market peneation * Sponsoring courses on specialized topics. begins, nansacrion costs are high and collective Facilitating the use of cxpert consultants (either narketing support can be importanL As marl -ts direcdy, by making a consultant available to a broad "thicken," initiatives by foreign buyers become more array of firms, or indirectly, by providing financial important. Gcercally the most effective collective support for the use of consultants). marketing support was of the kind that can be provided * Promoting information-sharing among fiirms. more cffcctively by decentralized organizations - such Countres that already have strong broad-based as industry associations or local governments and colleccive support and that are moving into chanbers of commerce (to support firms' participation in technologically more advanced activities might consider trade firs, for example) - than by central government "high-intensity" support, but should proceed with institutions. caution. Private mechanisms were more important than collective mechanisms for helping firms improve their - Ths paper -a product of the Finance and Private Sector Development Division, Policy Research Department -is part of a larger effort in the department to examine the impact of proactive intervention on SME -performance. Copies of the paper are available free from the World Bank, 1818 H Street NW, Washington, DC 20433. Please contact Daniele Evans, room N9-055, extension 38526 (43 pages). December 1994. Il Poliy ReMb Wo*iug Pae Series dissemwcs the fzns of work in progress to encomuge tt exchnge of iWs about deopment issu An ob)cthse of the sets is toga rhe finngs out qiddy, cvn 'the estatinare kss tha fuiy pol she . papsmaea" the names of th ators and soud be sed and eied accordigly. The frimdtgs interprtaton, and condusions ar dth aut hs' own and should nos auributed to the Word BDank, its Executive Board of Directors, or any of its mmnbercountrkc Produced by the Policy Research Dissemination Center Technical and Marketing Support Systems for Successful Small and Medium-Size Enterprises in Four Countries Brian Levy with Albert Berry, Motoshige Itoh, Linsu Kim Jeffery Nugent, and Shujiro Urata- Table of Contents 1. Research Background Research Issues and Methodology ............ Some Background Characteristics of Subsectors, Entrepreneurs and Firms ...... 6 11. Export Marketing and its Support Systems .................. ..................... a Leading Mechanisms of Export Marketing ............. 9 Cumulative Processes and the Role of Collective Marketing Support ............... 17 The Delivery of Collecfive Marketing Support ...... ......................... 19 Public Policy towards Export Marketing ..................... 23 IIl. The Acquisition of Technolgoical Capability and its Support Systems ..... ............ 25 Leading Sources of Technological Capability .............................. 26 Collective Technical Support Its Demand and Use..... 30 The Delivery of Collective Support. 32 Public Policy Towards the Acquisition of Technological Capability . .38 IV. Some Commons Themes ..3.. ... ................ 35 ; ~~Tables: Table 1: The Sample of Countries and Subsectors ....................................- 5 Table 2: The Growth of Fims ................................. -a Table 3: Education and Experience of Entrepreneurs .............................. 6b Table 4:- The Impact of Country Characteristcs and Industrial Organzation on Mechanisms of Export Markefing - A Summary of Country Experiences .9a Table 5: Marketing Contacts Among Rattan Fumiture Exporter ..................... lla Table 6: Exports by Korean SMEs in Four Subsectors. . 12a Table 7: Korea -The Reblfive Usefulness of Extemal Marketing Support . 12b Table 8: Colombia - The Relative Usefulness of Extemai Marketing Support. 15a Table 9: Indonesia - The Relative Usefulness of Extemal Markedng Support .. .... X a Table 9a:- Changes in Private Versus Collective Marketing Support Over Industry and Firm Life Cycles in Seven Subsectors .a. Table 10: Collective Marketing Support What Works? .............................. 20a Table 11: The Relative Usefulness of Export Marketing Services of Industry Associations and National Export Agencies in Korea . . - 21a Table 12: Leading External Sources of Technological Upgrading 27a Table 13: Mechanisms of Extemal Technical Support in Four Countries - An Overview . 29a Table 14: The Usefulness of Broad-Based Collective Technical Support in Selected SubsectorsandCountres 32a Table 15: Utiliation of Korea's Leading Specialist Technology Institutions in Four Subsectors ...... 36a E Fgures----- Figure 1: Use of Collective Marketng Support ....... 17a Figure 2: Use of Collective Technical Support ...........................-.-.31a in Fouroui by, Brian Levy with contributions from Albert Berry, Motoshige Itoh, Linsu Kin, Jeffrey Nugent and Shujiro Urata This paper analyzes comparatively the results of country sudies of successful small and medium enterprises (SMEs) and their support systems m Indonesia, Japan, Korea and -Colombia. The objective of the research was to shed empirical light on the policy question of what should be the appropriate role of goverment in the development of marketi and technological support systems for SMEs.]l We begin by describing thfe research issues and methodology, and some background features of the counties. subsectors and firms in which field research was conducted. Thereafter, we present in turn for each of the dtree support stems the major results of the research, and their implictions for policy. We conclude by highlightig some general themes that emerge from the analysis of individual support systems. 1: RESEARCH BACKGROUND I 1: Research Issue and Methodology- Lawn. Two presumptions underlie the decision to focus the research on government and SME support systems. The first presumption is that the micro-environment of private, govem l and NGO (non-governmental organization) market and istional supports external to the firm is an important detrminant of SME success.2j Large finms can use non-market medhanisms to internalize within the firm many of the functoal tchnical and marketing skils that that they need to conduct business. SMEs, by contrast, lack he rsources and in-house f1 The research examined also fiancial support sstems in Japan and Korea. The resuts are reported in the individual country studies. 2- This idea is implicit in the literatre which ghights the role of the e eur in large firms in developing counties as a "gap filler. See Jones ain Sakong (1980); Kilby (1971); and Leibenstein (1963, 1968). For the link between ffie development of private and public support systems and SME parcipaton, see Jacobs (1969); Pior and Sabel. (1984); Ranis and Schive (1986); Rhee (1988); and Levy (1991). 2 functional diversificiation of their larger counerparts, and depend more on exteral sources to acquire technological and marketing capability. Acquiring the resources they need involves transactions costs. It follows that the less developed are technical and marketing support systems, the higher will be the transctions costs to SMEs, with a corresponding impainnent in their competitive position. So one key goal of the research is to learn what are the characteristics of well-functioning :upport systems.-. The second presumption is that the social reums of well-functioning SME support systems - and, more broadly, of SME success - exceed the private returns to individual firns. This presumption can be defended via four distinct lines of reasoning. First1 and most filiar, is the argument that there exist rm and other market failures associated with the provision of technical and marketing support to SMEs.4/ Second, a proliferation of rapidly-grwing SMEs holds the promise of providing a seedbed for the emergence of dynamic and efficient arge-scale national firms, and consequently a more flexble and competitive domestic economy.51 Third, there can be important political economy advantages of broadening the base of private sector pauicipation. A broader private base imples less concentrtion of economic power, which can rcduce the risk of the emergence of a mutualy beneficial, but socially - unproductive, rent-seekiqg relaionship between business and goverunent. Moreover, broad- 1i Note that this goal is substantially narrower than that of learing what are the determinants of SME success (which could inclide a myriad of policy and human capital variables), with important implications for the design of the research. 4- For marketing and technical support, the public good character of informtion implies that the socially optimal supply of information exceeds what would be supplied privately. Collectve action problems could inhibit firms from organizing to jointly supply 'intra-industry" public goods. Technically, this argument can be viewed as one of 'inter-temporal spillovers Investments today which expand the number of successful firms make the economy more capable in the future for resonding flexibly to shocks, and ting advantage of new (currently unknowable) opportnities as they emerge See Bruton (1985, 1989, 1993) for a concepta discussion which makes the case that it is socially desirable for developing country governments to pursue interventions which lead to a more flexible economy. - 3 based participation in the private sector can ease social tensions, especially in countries where there are strong divisions between the economicaUy dominant group, and other groups in society. Fourth, there is some evidence that a substantial role for SMEs contributes to a more equitable distribution of inconle.A/ Given these four lines of reasoning, the corollary follows that if government can intervene in a cost-effective manner to improve the prospects of SME success beyond what would be achieved in a wholly private marketplace, it should do so. Taken together, the two presumptions shift the terrain of cebate as to whether governments should intervene to strengthen narketing and technological support systems from an ideological to an empirical one. Surprisingly, while there is ample documentation of failed atempts at intervention, there has been remarkably little empirical research as to whether collective (governmental, business association and NGO) interventions have been impoItant for the successful development of support systems.2f The research summaried here is the fruit of a susined effort to fill this gap. Mefibgoggx- Our starting point for analysis of the role of collective interventions was with the SMEs themselves. We sought to learn from stuctred field inteviews how these firis took care of their marketing and technological tasks, to what extent they drew on external supports to undertake these taslcs, and what was the relative utilization and usefilness (as perceived by the finms themselves) of private and collective supports. Our goal in adopting this approach, rather than studying directly specific imtventions made by colective istuons, was to anchor any broader judgements as to the role of government in an ue ng of the dynamics of the private marketplace itself. fit See, for example the analysis in Fei, Ranis and Kuo (1979) of the relation between the evolution of income distrbution in Taiwan and the role of SMEs. - - U For some studies of the role of technical and mar support, see Hogan. Keesing and Singer (1991); Schmitz and Musyck (1993); Cortes, Berry and Lshaq (1987); Berry and Mazumdar (1991); and Nugent (1990). Rhee, Ross-Larson and Pursell (1984) used a... methodology similar to the present one to examine the acquisition of technical and maketing. capability by large firms. 4 How we selected our sample of firms was crucial for the broader validity of our policy judgements. The key decision was to target the empirical analysis on successes: successful countries, successful subsectors, successful firms. The underlying presumption was that the economic institutions identified by firms as important for successful outcomes are themselves at least reasonably efficient.BI The research followed the lead of other analyses,2/ and used direct or indirect participation in export markets as the primary criterion of success (although, as will be seen, in conducting the research it was not possible to adhere rigidly even to this criterion of export-orientation). While in general the presumption that the.institutions identified by firms as efficient were indeed so seems reasonable to us, the reader can continully test any policy judgements we draw from our evidence by posing questions as to the cou uribcua. Key questions to keep in mind include: Given that the SMEs surveyed inleed appear efficient is it probable that, in the absence of some co'1ective institution identified as impormtant, their performance would have worsened? Alternatively, should there be reason to qution whether the. surveyed firms indeed are economically efficient, is it nonetheless probable that any collective institution deemed to be making a positive contriLution would still be socially useful in an environment populated by. economically efficient firms? In principle, variations across subsectors in their tecihologies and patterns of indusrial. .orgnization might influence the role of collective supports. Similarly,. their role could be affected by -variations across countries in the levels of development of their private sectors, and .the manner of their integration into the world economy. To account for these variations, we sampled across a broad range of countries and subsectors. . / This presumption is standard in the economic analysis of institutions. See, for example, - - - -Oliver Williamson's (1985, 1975) analyses of transactions costs, and Douglass North's (1981) 3nalysis of the relatonship between property rights and economic growth. While at the limit * - applications of the presumption can degenerate into tautologies, with careful design it is possible to formulate empirically testable propositions. 92/ For example, Pack and Westphal (1986), Krueger (1984), and the World Bank's recn (1993) study of the East Asia miracle. 5 -:lbe samW. Ideally, the sample of countries and subsectors would have been large enough to make it possible to analyze cross-country variations in support systems for a particular subsector, as well as cross-sectoral variations in support systems widtin individual countries. In practice, as Table I summarizes, budget limitations restricted the research to four countries; 34 subsectors were examined in each country, with the sample worldwide amounting to 445 firms. Details of sampling decisions with respect to subsectors and firns are contined in the individual country studies. While all four countries presently are judged to be relatively "successful" examples of development, they were intentionally chosen to encompass a wide spectrum of development experiences. They vary radically in 1991 per capita incomes - from $610 for Indonesia, to $1,260 for Colombia, $6,330 for Korea and $26,930 for Japan. Two of the countries are located in North-East Asia, one in South-East Asia, and one in Latin America. Unavoidably, given the wide variety of levels of development (and resource endowments), and the. desire to focus on export-oriented subsectors, it was not possible to achieve much overlap in the subsectors studied across countries. Yet the costs of limited overlap are less than they might first appear. For one thing, even within seemingly identical subsectors, there are substantial variations across countries in both the specific products produced and in the patterns of iudustrial organization.j/ For another, as the analyses of marketing and technical support systems will reveal, it is not product-type per se that is the crucial basis for comparson. For marketing support systems, what turns out to be crucial is the distinction between subsectors in which SMEs are embedded in pre-existing private networks, JQ/ To cite one example, Levy (1991) compares the footwear export industries of Korea - and Taiwan. Not only are there differences in their mixes of products (Taiwan produces relatively more plastic footwear, Korea relatively more leather), even when the product is identical (e.g. running shoes with leather uppers) there have historically been radical differences in their paterns of industial organization, with large firms dominant in Korea and SMEs in Taiwan. 5a TABLE 1:THE SAMPLE OF COUNTRIES AND) SUBSECTORS (NUMBER OF FIRMS SAMPLED IN BRACKES) INDONESIA (91) COLOMBIA (125) KOREA (122) JAPAN (107). Garments (34)" Garments (47) Woven Textiles (42) Woemet!sin _____________ _ _ _ _ _ _ _ _ _ _ _ _ _ ~~ ~~~~Fukui-1(33). Rattan Furniture (33) Automotive Automotive Components (20) Copnsin Ohta _____________ _____________ ____________~~_ (34) Carved Woodetn Eectranic Furniture in Jepara Conmponets (20) (24)_ _ _ _ __ ___ Machfiney (44) Factory Automaion _______ ~~~~~~~~~(40)_ _ _ _ Leather Produt (3)Silverware in Tsubame (40) 6 and activities where no such networks exist; six of fte subsectors selected fit into the former category, and seven into the latter. For technical support systems, a crucial distinction is between simple, craft-based and complex, engineering and science based technologies. 1.2 omne Backgound Characteristics of Subsectors. Entrereneurs and Firms Here we highlight three sets of characteristics of the firn samples: the rates of growth of the firms surveyed; the education and experience of entrepreneurs; and the relation between the fimns surveyed and pattems of subcontracting in their respective subsectors. Not surDrisingly, given our focus on success, the Indonesian, Colombian and Korean firms sampled all have enjoyed very substantdal expansion since stal-up.1l/ As Table 2 shows, the start-up size of the Colombian firms sampled was very nmaii, with 76 percent of them having 20 or fewer employees, and only 10 percent 50 or more employees. By the time of the survey, however, these proportions had virtually reversed, with 54 percent in excess of 50 employees, and only 14 percent with fewer than 20. The averge start-up size of firms sampled in Indonesia was larger; even so, 48 percent of them commenced with fewer than 50 employees, and only 23 percent with more than 250. But by survey time only 9 percent still had fewer than 50 employees, while 56 percent had 250 or more. As for Korea, Table 2 documents substantial firm growth, though less than in the other countries. However, for Korea in particular these employment data subsanmially underestimate the grorith of firms, since Korean wages have risen very rapidly and firns have responded by substituing capital for labor, while achieving very rapid increases in viue added. As Table- 3 shows, the levels of education of these successful SMEs are stiingly high. In both Korea and Colombia, over 80 percent of the entrepreneurs sampled bad completed at least some university; and, at 75 percent, the proportion was almost as high for Indonesia's garment and rattan entrepreneurs. Half of the Korean sample had undertaken their university studies in the field of science or engineering. And prior to startng their firms, over 90 percent .11,1 The Japanese firims are substantially older, and we do not have comparable data. 6a TABLE 2: THE GROWTH OF FIRMS % FIRMS WITH NO MORE THAN % FIRMS WITH A *LARGE" A "SMALLO NUMBER OF NUMBER OF E)APLDYEESW' EMPLOYEES' START-UP CURRENT START-UP CURRENT. INDONESIA 48% 9% 23% 56% COLOMBIA 76% 14% 10% 54% KORAd .81% 34% 4% 36% NOTES: a! 'SmallI is defined differently for dhe three countries. For Indonesi'a and Korea it is 50, and for Colombia 20 employees b/ 'Large" is defined differently for thie three countries. For Indonesia it is'at least 250, for Colombia it is 50, and for Korma it is 100 employees. Cl xcudngwoven txie,wher employment increases were limited as a result of a shift to automatic loans. 6b TABLE 3: EDUCATION AND EXPERIENCE.OF ENTREPRENEURS PERCENTAGE OF ENTREPRENEURS SAMPLED WITH:- ______ HIGH ~~SOME UNIVERSITY IEDUCATION' PRIOR SCHOOL OF WCTEXEENEIN EDUCATION- RELATED OR LESS BUSINESS COUNTRY AT LEASr GRADUATE A SCIENCE OR SOME - ENGINEERING PRIOR_:_-__ INDONESIA __OR_LESS_ .:BUNS : * Ganents 'and 25% 75% . Rattan - Iepara -7i..--:71 29 - COLOMBIA 17 83 19% (ALL) ______ FKOREA (ALL) 13 87. .11 50% 91% 7 of the Koreans had experience in a related business, almost 60 percent of them in a management position. A substantial majority of the Japanese sample also reported having had prior experience in a related business. The exceptions to the overall pattern of high education are revealing. The export entrepreneurs from Jepara, Indonesia are shown in Table 3 to be substantially less educated tha in the other countries and subsectors; uniquely among the subsec-tors studied, their source of competitive advancage lies in unusually well-developed wood-carving skills. Also, although not shown in the table, the owners of the Japanese SMEs samnpled, espec-ially those in their fifties or older, tend to have no more than high school education. However, as the diLscussion Of marketing supportsystems below will highlight, they appear to face fewer marketing demands tha the other entrepreneurs, inoar as they operate within a well-developed network of inter- firm relatins. Finally, subcontracting relations are ubiquitous in all four countries. As thie cotuntry~ case describes in detail, all of the Japanese firms surveyed are embedded in vertical, multi-tier prodution and marketing relations. In Korea, 82 percent of the firms surveyed subcontract, out either specific- production tasks or fabrication of the entire product; the corresponding figures for Indonesia and C-olombia are 76 percent and 67 percent. What is remarkablc about the Korean experience in particular is the rapidity of change in the role of subcontracting, in industrial production. In 1973, only 25 percent of all Korean SMEs earned any-revemue from subcontracting; by 1990, subcontracting was a source of revenue for 70 percent of Korean firms, and fulfly 90 percent of these earned over 80 percent of their revenue in this manner- For all of the pervasiveness of subcontracting, and for all that its expansion (as in Kore) doubtless multiplies the opportunities available to small business, the results of the: Indonesian and Colombian firm surveys sound a note of' caution as to whether subcontracting can serve straightforwardly as a stepping-stone to SME success mn export markets. Fewer than 14 percent of the Colomobian SME exporters surveyed, and fewer tha 22 percent of the hidoneian SUE exporters, had themselves had prior experience as subcontractors prior to becomning direct exporters. Considered together with the evdneeneucto lees, these 8 data point to the possibility of substantial dualism among SMEs, with an educated minority successfully "graduating" to become direct exporters, and the majority more-or-less permanently engaged in the international marketplace, if at all, only indirectly as subcontractors. II: EXPORT MARKETIG AND lTS SUPPORT SYSTEMS Even with favorable market conditions, exports do not proceed spontaneously, but through the medium of institutions, and via exchanges which involve both ecx ante and ecx post transactions costs. Ex ante transactions costs comprise the costs of search by buyers and sellers to identify a party with whom to transact Ex post transactions costs comprise the costs to one party that result from non-performance by the other. SME success in export markets thus requies the presence of mechanisms which reduce the costs of search to SME suppliers and to their buyers, and which can reliably signal the reputaion of suppliers to putative buyers (and vice versa). A central goal of this stLdy is to leam what was the role of collective marketing support systems - how extensively they were used by SME exporters, how much value SMEs ascnrbed to them, and what specific lknds of collective support were most highly valued by SMEs. Yet firms do not make their decisions on the use of collective marketing support in a vacuum. There also exist a range of private channels which reduce the tasactions costs to firms of participating in export markets - trading intermediari, search efforts by foreign buyers, opportnities for indirect exporting via subcontracting, and the presence of other prc-enistg private networks that can ease entry into export markets. Where they exist, these private networks represent substite channels which reduce SME demand for collective support. Additionally, SMEs eager to export can make direct efforts of their own to win orders - although collective support Cif it Is availale) is likely to complement, rather than substitute for, such efforts. IbThe next subsections analyze comparatively for Indonesia, Japan, Korea and Colombia the supply of private mechanisms of export support, and the impact of their availalit on the 9 use of collective support. Thereafter, we describe in some detail the specific kinds of collective marketing support have proven useful in the sample countries, and the specific institutional mechanisms that have been most effective in delivering them. I1. 1: 1hadingMchanisms of Eport Marketing The field research uncovered very substantial variation in the mechanisms of export marketing across countries and subsectors. Taken together, three propositions appear to sumnarize much of this variation: * Variations in industrial organization infuence the m anisms of export marketing. In some instances, SMEs are embedded in well-developed, pre- existing private networks, with low tasactions costs of entering export markets, and with virtually no role for collective support. In others, the challenge of linking into the iternational marketplace is more formidable, involves higher transactions costs to SMEs and calls for distictive m g mechanLisms. -ntertemporal. variations within individual iadustdes influence the mechanisms of export marketing. In the initial period of market penetration of a new export industry, transactions costs are high and domestic-resources play an important role. Subsequently, markets "thicken", inidatives by foreign buyers become more important, and the transactionm costs of export entry by SMEs declie. (A parallel pattern is evident over the life cycle of firms, regardless of the tining of first export entry.) * Variations across counties in the extent of their engagement in the international marketplace influence the mehanisms of export marketing. The challenges and mechanisms of export marketing are very different, and the transactions costs of penetrating these markets are lower, for SMEs active in countries that are already "on the map" as export suppliers, than for those active in countries whose export drive is still in its infimcy. Table 4 offers capsule summaries of the experiences of the four countries studied in relation to these propositions. The experiences of Japan and Indonesia iustrae the impact on marketing patterns of industrial organization and related pre-existing networks, and those of Korea and Colombia (plus Indonesia again) illustrate both the evolution of marketing mechanisms from the pioneering export phase onward, and the impact on this evolution of a country's profile in the global marketplace. Japan and Indonesia: The'Power of Pre-existing Private Networks :a. TABLE 4: THE IMPACT OF COUNTRY CHARACTERISTICS AND INDUSTRIAL ORGANIZATION ON MECHANISMS OF EXPORT MARKETING - A SUMMARY OF COUNTRY EXPERIENCES INDUSTRIAL COUNTRY'S EXPORT ORGANIZATION PROFILE- JAPAN No direct marketing; SMEs Not Applicable embedded in dense, pre- existing private networking INDONESLA Extended Chinese communitky Nascent reputation as export confers marketing advantages supply source on non-Pribumi Firms. Pribumis use substitLte mechanim-s KOREA Emergent network of export Strong reputation as exporter traders (inter-temporal, simplifies direct export Dimension) marketing for SMEs (subsubent to initial market . . . ~~~~~enty)- COLOMBIA Limited pre-existing networks Limited exposure to export marketplace complicates . ______::____.___-_ marketing for SME exporters 10 In different ways, the Japanese and Indonesian studies highlight the power of pre- exisdng private networks in facilitating SME participation in the export marketplace. Jpan. Across the gamut of Japanese industry, the marketing of SME products is taken care of virtually automatically by pre-existing private production and trading networks, with no direct interface between the firm and its foreign buyer, and virtually no role for collective support The three subsectors examined in the Japan case study illustrate the wide variety of patterns of industrial organization that yield this outcome. Multi-ter. subcontracting relations among Japanese SMEs engaged in the production of automobile components in the town of Ohta comprise a by-now familiar example of a pre- existing private network. Fuji Industries (assembler of the Subarn automobile) is located in the region, and half of Obta's first-der subcontractors supply parts .o the company. Sales to Fuji Industries account for 43 percent of sales of these firms, with vitualy all other sales going to other automobile assemblers in Japan. Dependence on a dominant parentfin grew for the tehnically less-sophisdcated second-tier and lower-tier subcontractors, with 75 percent of sales of second-tier subcontractors going to their parent fErms. In Japan's silverware and textile subsectors, the pre-existing private networks comprised interactions with trading houses and withi mput suppliers. Tsubame's silverware exports have long been channelled tbrough the country's export trading houses; indeed, as of the time of our survey these trading houses continued to account for 80% of export sales. And SME. weavers of synthetic textiles in Fukui-Ishilkawa are embedded in center-satellite relations to an even larger degree than their silverware counterparts: about 85% of textile weaving takes place under 'service fee contracts" in which trading houses or yar-producing chemical companies supply yarn to SME weavers, which is woven and dyed, and then retumed to the yam supplier for subsequent sale. In all, indirect exporting through pre-existing private networks was so pervasive in Japan that it proved infeasible to systematically sample firms as to the relative value of different kinds of marketing support for direct exports. Indonesia. In Indonesia, the impact of another kind of pre-existing network is highlighted by sharp differences in the mechanisms of export marketing of pribumi and non- pribumi (Chinese) finns. These differences appear to be the result of linkages among an extended Chinese community that transcend national boundaries and provide an informal private network into which Indonesia's entrepreneurs of Chinese origin - but not their pribumi counterparts - can straightforwardly connect. As Table 5 suggests for rattan furniture exporters, pre-existing linkages translate into a superior ability on the part of non-pribumi firms to make initial export contacts outside of Indonesia: 63% of initial export contacts of non-pribumi Indonesian finis were made outside the country (via a business trip abroad, or a friend or agent based abroad), as compared with only 15% for pribumi firms. As the table shows, the disparity in the role of contacts made abroad continues even as fims become established in export markets. In garments, too, 53 % of the most recent export contacts for non-pribumi firms - but none for pribumi firms -were made abroad-42/ Lacking both contacts abroad and ready access to buyers visiting Indonesia, pnriumi garment and rattan furniture firms made use of substitute mechanim. Thus over half -the pribumi firms in Table 5 (but under 20 percent of the non-pribui) used collective- mechanisms to link up with export markets. Korea. Colombia (and Indonesia again: Penetrating the Exportn Marketlace de Novo Unike Japanese and non-pnbumi Indonesian firms, SMEs in Korea and Colombia generally lacked access to pre-existing prvate networks, as did Indonesia's pribumi SMEs.- Consequently, empirical analysis of these settings holds the promise of showing how marketing mechanisms start-up and evolve as specific new export activities take hold, with resultant declines in tie transactions costs of export entry by SMEs. Additionally, the radicaly different export profiles of Colombia on the one hand, and Korea and Indonesia on the other, provides. 121 Note, though, that despite extensive search it was only possible to identify six pribmi. garment firms for inclusion in the sample. The third Indonesian case, that of carved furniture exports from Jepara, does not shed any useful light on the role of pre-existing networks; all but one of the firms surveyed were pnbumi, and they were not embedded in any pre-existing network. lla TABLE 5: MARKETING CONTACTS AMONG RATrAN FURNrrURE EXPORTER INITIAL CONTACTS CURRENT- CONTACTS Pribumni NonProbumi Prium fNon~Pribumi PRIVATE__ _ _ _ _ _ _ _ _ _ In County 4 3 13 13 Abroad 210 4 26 COLLECTIVE 7 3 1s 6 TOTAL 13 16 3-5 45 -~ .- --- 12 -an. opportunity to explore how much inpact a.country's broader profile in theintemnational marketplace has on the development of export marketing mechanisms. Korea The way in which Korean SMEs have been incorpoated into that country's industrial economy is different from the Japanese patten. Whereas Japanese indusrial organization is characterized by a dense network of interaction between large and small firms along the lines described'above, Korea's industrial success in the 1960s and early. 1970s was built on its giant industrial conglomerates, with SMEs playing an increasing role only afier 1977. Moreover, the ties between large and small firms have been looser in Korea than in Japan, with a correspondingly greater extent of direct participation in exports among Korean SMEs1.' As Table 6 shows, large firms tended to be the export pioneers, but direct SME exports expanded rapidly in all four subsectors studied. While in the woven textile and electronic parts subsectors exports were substantidal aready in 1983, in the remaining two subsetors the move by SMEs into export markets took place only in the lat half of the 1980s. For woven teffiles the pioneering phase of SME exports actually predates the export start-Wp of most of our sample firms. Table 7 details for each subsector the relative importnce of four mecnisms of export marketing at various points in time.- For Korea, export 'sLoflmtigfl refers to manufacture against specifications for export by traders, fl to production of components for assemblers who export finisheaL products. As Table 7 confirms, in all periods marketing of woven textiles . is dominated by a network of specialist export traders/subcontractors with headquarters or branch offices clustered in the area of Taegu, the location of almost 80 percent of Korea's woven textile firns. As noted, woven teffile SMEs already had established a position in export 131 In the automobile and electronic components subsectors, the majority of firms surveyed were embedded in subcontracting relations with large assemblers,Sales to assembler-principals accounted for over 90% of sales for 14 of the 20 Korean auto parts firms surveyed, and for 9 of 20 electronic parts firms. However, even in these sectors over 60 percent of surveyed firms :-manuactued at least some products direcdy for export markets; these direct exports were the focus of the analysis of export marketing in the Korean case study. 12a TABLE 6: EXPORTS BY KOREAN SMEs IN FOUR SUBSECTORS (MILLION OF USS) 1- ~ ~~~ -. . T , . . WOVEN AUTO PARTS ELECTRONIC METAL CUTTING TEXTILES jPARTS |EQUIPMENT 19 83 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ SME Export Value - 349 4 199 21 |(USMilions) _ _ SME Shares of 21.3% 9.6% 19.5% Total Exports -1988. SMEExport Value 1,454 83 134% 68 = (USsKllions) - SME Shares of 42.8% 25.9% 22.9% Total Exports I991 SMEExportValue 3,453 - 184 2380 123 (USMMIIions) SME Shares of 60.6% 44.1% 24.8% Total Exports - = 12b TABLE 7: KOREA - THE RELATIVE USEFULNESS OF EXTERNAL MARKETING SUPPORT (ALL FIRMS I LEAST USEFUL; 5 = MOST USEFUL) DIRECT BUYER SUBCONTRACTING COLLECTIVE - MARKETING_ INITIATIVE PRINCIPAL - i SUPPORT - Woven Textiles * Start-up, Pioneers- 2.4 3.2 4.2 1.2 | Start-up, Followers 2.5 2.4 4.0 1.3 urQ t, All 2.5 -2.8 3.7 - 1.3 Auto Parts : _ - . - Start-up, Pioneer 2.5 1.0 3.3- 2.2_ Start-up, Followers 1.5 3.8 2.2 - 2.0. Current, Alt 3.0 - 3.6 2.8 1.3 Electronic Parts - ___-_-_::- Start-up, Pioneers 2.0 1.8 2.8 3.3 |Strt-up. Followers- .6 j 1.8 3.0 2.4 Current, All 3.4 3.9 .1.6 1.5 Factory Automation - Start-up, Pioneers 3.2 2.7 2.8 1.5 Start-up, Followers 2.3 3.6 -2.2 2.0 Current, All 4.0 3.6 2.0 - * 2.0 13. markets by the late 1970s, and there is very little variation across periods in the relative importance of the various export marketing mechanisms. The export marketing patterns for this subsector parallel more those of Japan and non-pribumi Indonesian firmns than they do those of the remaining three Korean (and the Colombian) subsectors. For the rest, the Korean survey results illustrate vividly how export marketing channels evolve - how SME exports get going in individual subsectors, and how the transactional environment for export marketing "thickens" as the number of SME exporters proliferates. Table 7 organizes the subsectoral data on the relative importance of various marketing mechanisms into three time-related subgroups - the time of the field survey; and firms' initial period of entry into export markets, with initial export entry in turn disaggregated between pioneer SME exporters and SMEs that moved into international markets only once the industry had established its role as an exporter. As the table shows, the initial imnpetus for exports in the three subsectors came primarily from resources already within the country - from export traders, from coilective marketing support institutions, and from exporting firms themselves. Traders/subcontractors played the leading pioneer role for auto parts SMEs, and the second most inportant role in electronic parts and factory automation SMEs. Their prominent early role can be interpreted as an important spillover for Korean SMEs from being active in a country with a strong export profile, insofar as it was the country's earlier export success. (typically by larger firms) that brought these traders/subcontractors onto the scene. Collective marketing support leads among pioneer electronic parts SME exporters, and has a significant subsidiary role in auto parts. Direct marketing efforts by firms provided the critical initial impetus in the factory automation subsector, and also scored high among autoparts pioneers. In aU three subsectors the role of direct marketng declines for the export start-up of follower exporters, but picks up again for current exports. Over time, initiatives of foreign buyers come to the fore. Buyer initiatives are identified in Table 7 as the least important mechanism for initial export entry by pioneer SMEs in the auto and electronic parts subsectors, and second least important for factory automation. But information appears to spread rapidly as Korean firms come on line as a primary international 14 supply source for yet anotherproduct. Indeed, in the later periods initiatives by foreign buyers emerge in all three subsectors as the leading mechanism for facilitating participation in export markets. In all, then, the evidence from Korea on export marketing suggests that "export entry begets export entry", that a cumunlative process can take hold.14 In all three subsectors for which we have data, local actors (presumably prompted by numerous export incentives) led the pioneering phase of SME exports. Their activity took place against the backdrop of Korea's strong overall reputation as a reliable and cost-effective supplier of exports. Once exports of a new product were underway, foreign buyers came onto the scene, and the momentum of sustaining the export drive shifted increasingly into their hands, with complementary efforts by increasingly committed export suppliers. Throughout the process, the transactions costs of linling into export markets progressively declined, with corresponding increases in opporunities for export entry by SMEs. Colokmbia. Like their Korean (but unlike their Japanese or non-pnbumi Indonesian) counterparts, Colombian SMEAs lacked access to pre-existing private networks to link into the export marketplace. The Colombian garment industry experienced an export surge in the 1970s with exports in 1980 amounting to over $110 million already in 1980, too early for the pioneenng export phase to be depicted adequately by our firm sample. The export take-offs in the leather products and machinery subsectors came over a decade laterIjf and so offer evidence of the pioneering phase of export marketing of new products. As was observed for Korea domestic resources played a leading role in the pioneering export efforts of Colombian firms in the leather and machinery subsectors. Direct marketing - 4/ Hirscbman (1958) was a pioneering study of the role of cumulative processes in economic development. Levy (1989) modelled formally the role of cumulative processes in the development of export markets. :1.5/ Garment exports subsequently declined to a low of $40 miWion in 1983, but expanded later in the decade to reach $449 million by 1991. In the machinery. subsector, too, exports peaked at $39 million in 1980, subsequently declined to $12 million, but then expanded in recent years to reach $102 million by 1992. Leather product exports got underway only in the 1980s, rising from $14 million in 1984 to $106 million by 1991. 15 efforts of firms emerge in Table 8 as the leading mechanism of export market penetration for pioneer firms in both subsectors. Collective marketing support plays a signiflcant subsidiary role for pioneer exporters, as well as a continuing useful (though diminishied) role as the subsector matures. Aside from this central similarity, there also are some striking differences between Colombia and Korea in the marketing mechanisms of pioneer exporters, differences which plausibly result from the divergent export profiles of the two countries. For all that Colombia has the reputation of being one of the few Latin American countries to have maintained a relatively liberal trade regime for a sustained period of time, until very recently Colombian industrial policy placed nowhere near the emphasis on exports as did Korea's, and Colombian firms have had nowhere near the sustained presence in manufactures export markets as have the Koreans. One consequence of this limited history which is evident in Table & is the virtual absence of trading and subcontracting ties as a mechanism of export market.penetration for - Colombian SMEs; by contrast, such ties were of central inportance for pioneering Korean A second and somewhat more puzzling difference is in the role of foreign buyers: despite Colombia's lower export profile, foreign buyers are relatively more important at the pioneering phase for Colombian than for Korean SME exporters. One plausible resolution of this apparent paradox may be that, by comparison with their Korean counterparts, Colombian firms historically have been less interested in international opportunities at the early stages of export market development, and so have left relatively more of the (modest in absolute terms): early running to foreign buyers. Further, whereas in Korea the relative importance of direct marketing efforts and initiatives by foreign buyers shifted sharply in favor of foreign buyers as the export drive deepened, in Colombia direct marketing initiatives play the leading role even in the current period. Plausibly, this tardier responsiveness of foreign buyers to emerging opportunities to source from Colombia is a consequence of the lower profile of the country in. the global marketplace. Put differently, the evidence suggests that the cumWulative process of ncreasing participation by foreign buyers, declining transactions costs of export marketing, and 15a TABLE 8: COLOMBIA - THE RELATIVE USEFULNESS OF EXTAERAL MARKETING SUPPORT (ALL FIRMS 1 = LEAST USEFUL; 5 = MOST USEFUL) -. -1 - _ _ _ I _ _ _ I _ _ _ _ __ I _ _'' _ _-i _ _.-.- .DIRECr B YER SUBCONTRACTING COLLECTIVE * MARKETING INITIAITVE PRINCIPAL SUPPORT- Leather Products ._-_._ ._ . Start-up, Pioneers 4.4 : 3.5 1.8 2.6 Start-up, Follower. 3.5 3.7 2.3 1.8 Current, All 3.9 3.4 1.4 .2.5 Garments ._- _- __ - ___ -_ -_:. Start-up, Pionsers 3.7 3.9 1.4 3.0 Start-up, Followers -3.7 3.6 1.2 1.3 Curret, All 3.8 3.6- 1.1 2.4 MachLinery _ - . -_._____:_-__ Start-up, Pioneers 3.9 3.0 1.0 2.5 Start-up, Followers 4.2 3.6 1.3 1.9 CUrrent, A1l L3.8 3.2 . 1.2 I 2.2 16 increasing participation by SMEs has not yet taken hold in Colombia with anything like the vigor that we observe for Korea. -donesia (again). For all three ndonesian subsectors, our surveys yielded evidence on the evolutdon of export channels. As Table 9 shows, relative to both Korea and Colombia, initiatives by buyers turn out to be unusually dominant in all time periods. For the pioneering periods, the explanation for buyer dominance vanes across subsectors. For garments, the presence of pre-existing networks among the extended Chines community along the lines nated earlier provides sufficient explnation, since the vast majority of that industry's garment exporters are Chinese in background. By contrast, most of the wooden fumiture exporters in the sample were pnibumi, based in the traditional woodcarving center of Jepara, who took advantage of the area's abundant supply of carving skills to produce reproduction antique furniture for export markets. Paralleling ihe pattern hypothesized for Colombia, Jeparan producers had not initially set their sights beyond the domestic market. So, - - as the case study details, and as is signalled by the score for buyer initiatives im Table 9, the initial impetus for exports came from outsiders. For rattan funiture, it is necessary to distngish between the "prehistory and the acuala period of export take-off. in the 1970s and early 1980s, rattan furniture exports were pioneered by a small number of predominandy prbumi entrepreneurs. They are the finns grouped in the "pioneer" category and, as can be seen in Table 9 and consistent with the discussion earlier, were disproportionately dependent on collective support for the initial export efforts. However, take-off of rattan fritmre exports came only in the latter 1980s, at about the --- time the Indonesian govermnent banned the export of raw and semi-processed ratan. The majonty of new exporters during this phase were Chinese (many of them former traders in raw rattan) who, as in garments- were able to rely on their extended network to penetrate export markets. Thus, as Table 9 shows, at export start-up non-pribumi "follower" firns identified buyer initatives as their most important source of support. To be sure, some of the "follower" -entrants were pibumi. However, they lacked access to the network and hence, as is evident in .~~~nmswr neieti -16a TABLE 9: INDONESIA - THE RELATIVE USEFULNESS OF EXTERNAL MARKEnNG SUPP(RT (ALL FIRMS 1 LEAST USEFUL; 5 = MOST USEFUL) I DIRECT - BUYE | SUBCONTRACTING COLLECTIVE ._____ |MARKETING__| INiTIATRVE | PRINCIPAL SUPPORT Start-up, Pioneers 3.4 4.1 2.3 | _2.1_._. Start-up, Followers 3.1 3.4 2.5 2. :_Curet, All 3.3 _3.9_1_7 1.8 Carved Wooden Fumnifte from Jepama _ - ._. -Start-up, Pioneers 1.8 3.4 2.4 - 2.0 Current, Al 1.3 4.7 1.9 1.7 Rattan Furiture, All - . -_: Start-up, Pioner 2.5 3.5 1.0 4.8 Start-up, Followes 3.4 3.7 . 1.4 2.8 Crent, All j 3.1 4.3 1.1 3.1 Rattan Furniture, Non-Pribumi Start-up, Pioneers : , - - Start-up, Followers 3.0 4.2 - 1.1- 2.6 Current, All f 2.8 4.5 1.1 2.9 Rattan Furiture, Pribumi .n;n._._X Start-up, Pioneers 2.5 . 3.5 1.0 4.8 Start-up, Followers 4.1 2.9 1.8 3.1 Current, Al 3.6 4.1 1.1 3.3 17 Table 9, depended disproportionately on their own efforts and collective support to penetrate export markets. Tuming to the high scores for buyer. initiatives in the current period, while in the early 1980s Indonesia barely had any presence on intemadonal markets for manufactures, manufactures exports subsequently expanded at an astonishingly rapid rate to reach $14.7 billion by 1991. By that time, like Korea but unlike Colombia, the country had achieved a high export profile and foreign buyers were flocking in to take advantage of its low production. costs. Hence the dominant role of buyer initiatives in the current period evident in Table 9 for all three subsectors. . . .~~~~~~~inr 11.2: Cunulative Processes and the Role of Collective Marketing Suport Using a simple demand and supply framework, we summarize the previous discussion in a way that highlights the sources of variation in the demand and. use of collective marketing support across countries and subsectors. The three propositions highlighted earlier account for much of the observed vzriation. The first proposition was that, insofar as private and collective markeing mechanisms are substittes, the presence of pre-exig private networks at the inital stages of expansion into export markets is an importnt deterinaniof the demand for collective marktii'g suppeot. Heuristcally, DlDl in Figure 1 represents those countries, subsectors and fms which lack access to pre-existing private networks and hence might demand quite substantial collective marketing support, while D2D2 re-presents countries, .subsectors and firms which enjoy access to pre-existing networks and so have less demand for collective supporclj/ In seven subsectors reviewed above, private networks already were in place as of the time of export entry of the firms surveyed: "'silverware, auto parts and textiles in Japan; IW Fonnally, Qd = f(Xp,X2.....Xn), where Qd is the demand for collective support, Xp is the cost to te firm of using that support, and X2-Xn are other determinants of demand. As the figure shows, we make the standard assumption as to the negative relation between price and quantity demanded. In the text, we explore three of potentially many non-pnce deteminants of demand for collective support. 17a Figure 1 Dl Si D2\ - -/ .. ,Dl D2 Use of Collective Marketing Support 18 * nonpribumi garment and rattan exporters in Indonesia; * Korea's textile SMEs in the industrial district of Taegu; and * Colombia's garment SMEs. By and large, SMEs in these subsectors had relatively less use for collective support than did SMEs in the remaining seven subsectors, where pre-existing private networks were underdeveloped: * pribumi rattan and firniture exporters in Indonesia;. * Korea's auto parts, eletrnic parts, and factory automation SMEs; and * Colombia's leather and machinery SMEs. Evidence from these latter seven subsectors offers strong support for a second proposition highlighted earlier, namely that intertemporal variations within individual industie influence the mechanisms of export marketing. Table 9A presents in summary form some key patterns that were evident in Tables 7-9. The data are normalized, with the average usefulness scores for "buyer initiative" set at one for each subsector. Each subsector's average score for collective marketing support is recalculated as a ratio of the "buyer initiative score". Averaging these normalized scores across subsectors (within counties), and across all subsecors and countries yields the patterns evident in the table. Table 9A highlights the powerfil cumulative processes at work in the evolution of the mechanisms of export marketing. The usefulness of collective marketing relative to buyer initiative is highest for an industry's export pioneers as they sta-up their export efforts, and then subsequently declines as foreign.buyers come onto the scene and the transactional environment for export marketing 'thickens", with a corresponding fall in the transactions costs of using entirely private mechanisms of export marketing. As Table 9A shows, in general this pattern is evident not only over the industry life cycle (compare the relative scores at entry for export pioneers and export followers), but also over the firm life cycle (compare the current scores versus the scores at entry for all firms). The decline in use of collective mechanisms can be illustrated as a continuing leftward shift of the demand schedule, DIDI in Figure 1. 18a TABLE 9A: CHANGES IN PRIVATE VERSUS COLLECTIVE MARKETING SUPPORT OVER INDUSTRY AND FIRM LIFE CYCLES IN SEVEN SUBSECTORS RELATIVE USEFULNESS SCORES: COLLECTIVE SUPPORT/BUYER INITIATIVE Pioneer Firms, Follower Firms, All Firms, |Entry Entry Current All Countries and 1.2 0.7 0.6 SubSectors KOREA 1.5 0.8- 0.4 INDONESIA 1.0 0.8a 0.6 COLOMBIA 0.8 0.5 0.7. al For Jepara, this score was estimated as the midpoint between the Pioneer and Current Scores. 19 The powerful empirical pattern.evident in Table 9A should not be interpreted as implying that collective support, though evidently helpful in the initial stages, is a necessaly catalyst for the take-off of a new export industry. Indeed, as is evident in Tables 7-9, only i-'- two of the seven subsectors - electronic parts in Korea, and rattan exports from Indonesia by pribumi-owned firms - did collective efforts comprise the leading source of support for export pioneers. In three subsectors - factory automation in Korea, and leather products and machinery in Colombia - direct marketing efforts by firns themselves, were crucial to the initial penetration of export markets. In one subsector (Korean auto parts) traders played the crucial catalytic role. And in the final subsector, exports of carved wooden furniture from Jepara, notwithstanding the overall pattern evident in Table 9A, buyer initiave comprised the crucial catalyst. The third proposition highlighted earlier was that inter-country variations in the extent of engagement in the export marketplace influence export marketing mechanisms. The relevant- comparisons here are between Colombia with its relatively low export profile on the one hand, and Korea and Indonesia with their higher profiles on the other. Colombia's lower export profile, and consequenly weaker visibility to foreign buyers, might be hypothesized to correspond with a higher demand for collective support - DIDI in Figure 1, versus D2D2 for Korean and Indonesia. Surprisingly, Korea's higher export profile than Colombia's did not translate into more foreign initiative in the pioneering phase of SME exports in a new industry, wit a correspondingly lower demand for domestic resources, collective or otherwise. But, as Table 9A summarizes, subsequent to the initial phase of export entry, the predicted differences were evident, with a sharper relative decline in the usefulness of collective marketing supports in Korea than in Colombia. These differences can perhaps be interpreted as signalling the failure us far of Colomnbia to "lock-in" its presence in the international marketplace, and set in motion the cumulative process of increased participation by both foreign buyers and domestic suppliers which has proven so crucial for Korea's economic success. 11.3: The Delivery of Collective Marketing Support 20 In al four country studies, private mechanisms of export marketing support turn out to be more important than collective ones. Yet, at least some firns, subsectors and countries benefit quite substantially from the presence of collective marketing support. So, given a policy decision to support SMEs, the question arises as to how to provide such support. We answer this question in two steps. First, we examine what specific kinds of collective support have proven most useful to SMEs. Thereafter, we examine what kinds of institutions have proven most effective in delivering this support- Collective support: what works? The Colombian, Idonesian and Korean field surveys collected detailed information from firms as to the relative usefulness of different types of collective marketing support for their export efforts. Table 10 reports the results for each specific type of support that was used by at least 20% of the firms sampled in an individual eubsector. The results are strikiingly similar across countries. First, finms ascnbe a high value to participation in trade fairs at home and abroad as a means of penetrating export markets. Trade fairs emerged as the leading or second most valued coliective source of export marketing support in seven of the nine subsectors for which data are reported in Table 10. Moreover, as we shall see in the analysis of tehnical support systems, trade fairs also turn out to be an important source of technological learning. Second, only in the garment and textile industry (but within that industry systematically in all three countries) do frms identify infornation on export opporunities, rather than participation in trade fairs, as themost important source of marketing support only in the garment and textile industies. This result reminds us that international trade is conducted in different ways for different industries: for some, high profile fairs are the high points of the trading season, for others an ongoing search by individual buyers and sellers for trading partners matters more. Third, SMEs ascnbe very limited value to direct intoduction of buyers by official export agencies. Efforts along these lines played virtually no role in either Colombia or Indonesia. In Korea, the electroics parts subsector provides a partial exception to the general pattern even though it is noteworthy that recipients of this kind of support (every single one 20a TABLE 10: COLLECTIVE MARKETING SUPPORT: WHAT WORKS? A: AVERAGE USEFULNESS FOR ALL RESPONDENTS B: SERVICE USERS AS PERFECTAGE OF RESPONDENTS C: AVERAGE USEFULNESS FOR USERS ONLY (I = LEAST USEFUL; 5 MOST USEFUL) TRADE FAIR TRADE FAIR INFORMATION ON INTRODUCE ABROAD HOME EXPORT 1 BUYERS __ __ __ __ __OPPORTUNMES COLOMBIA ._-:_-._:-_ Leather A 2.7 3.1 1.8 B 53% 68% 41%. C 4.3 4.1 2.9 . Gannent A 2.2 2.5 2.5 1.7 B 40% 49% 60% 34% C _4.0 3.9 3.5 3.1 Machinery A 2.8 - 2.8 1.9 1.5 . - - B 57% - 64% 41% 32% C 4.1 3.9 3.2 2.6 INDONESIA . . . _._ -_____ Rattan Furniture A 3.0 2.7 1.9 1.9 B 58% 76% 33% 52% C 4.3 3.2 2.4 2.8 Garments A 1.8 1.9 2.0 1.7 B 34% 50% 53% 28% C 3.3 2.8 2.9 3.4 KOREA _ _ _ _ _ _ _ _ _ _ _ _ Wovea Textiles A - - 1.8 B - - 32% C - - 3.5 - Auto Parts A - - T 1.5 B - - 25% C - - 2.9 Electronic Parts A 1.6 2.0 .1.7 2.9 B 20% -40% 27% 100% c *3.,9 3.5 3.5 .2.9 Factory Automation A 2.4 2.2 2.4 1.8 B 50% 50% 57% 43% C 3.9 3-5 3.5 2.9 .21 of the sample finnsl) ascribe it only moderate value. The Korea Trading Company played an exceedingly limited role. For all that its explicit mandate was to serve as the "general trading company' for SMEs, orny 4 of 72 direct Korean exporters in the sample actually made their export sales through its channels! Collective support: who should provide? Any suggestion that governments should support the collective. provision of marketing support has to confront the reality that the institutional capability to deliver such support is weak in most developing countries. Indeed, the developing world is littered with failed export support programs and 'white elephant' export institutions. The hopeful message of the present study is that there may be instituional alternatives to elaborate export agencies. The empirical evidence highghts the benefits of interventions with a "light touch' that provide firms with the wherewithal to find buyers for themselves, rather than attempt to substitute for efforts by putatve exporters. Moreover, the evidence suggests that the delivery of export marketing support might usefully be decentralized, tailored to the specific realities of individual marketplaces so as to be able to respond to the enormous diversity of players and market mechnisms across subsectors. All three country studies offer empirical support for the proposition that delivery of export marketin support to SMEs should be decentralized..Z Korea's flagship export agency, KOTRA, has developed elaborate .export information systems, multiple tools. for promoting national exporters, and extensive international networks of offices. Yet as Table 11 sihws, aside from the electronics subsector, there is virtually no difference in the extent of support. provided by KOTRA and by industry associations.jl/ Of 72 direct exporters included in the sample, 29 reported using KOTRA's services, and 23 reported receiving support from their 12/ In Japan, too, industry associations and chambers of commerce. were the. primary direct channels of export marketing support for those few firms who sought to market directly. - B! - It is worth noting explicitly that no judgernent is intended here as to the nett benefits. to Korea of KOTRA, whose goals went well beyond providing support specifically for SMEs. The point is rather that a KOTRA-like agency is not necessary for the effective provision of SME support. 21a TABLE 11: THE RELATIVE USEFULNESS OF EXPORT MARKETING SERVICES OF INDUSTRY ASSOCIATIONS AND NATIONAL EXPORT AGENCIES IN KOREA Average Score Number of users Average Score .____________ : _________ ._________ _ l (all firms) . . (Users) _. _ . - | Number Number of KOTRA Associa-. KOTRA Associa- KOTRA Associa- of Firms Exporters tion tion tion Woven 42 31 1. 15 8 - 9 3.4 3.2 Textiles Auto Parts 20 14 1.2 1.3 4 4 1.8 2.5 Electronic 20 15 1.7 1.2 8 I 2.8 5.0 Parts Factory 40 12 1.5 1.5 .9 9 3.4 3.1 Automation X . L L :- TOTAL 122 72 _ _- 1 29 .. 3.0 3.1 22 trade association. Moreover, the average rating of the value of the services of the two sets of suppliers was virtually identical. The Indonesian experience offers three lessos - one cautionary, one niixed and one hopeful.- The cautionary lesson concerns the role of centralized export institutions. Like many other developing countries, Indonesia erected in the 1970s two elaborate export institutions - NAFED, whose prime responsibility is to promote trade fairs with Indonesian participation, and the Iternational Trade Promotion Center, an international network- of commercial offices located in Indonesian embassies abroad. Yet by 1985, fairs aside, the perception was widespread that these insftituons were of limited usefulness. The mixed lesson concerns the role of industry associations. Parts of Indonesia's furniture and garment associations emerged in the firm surveys as useful sources of collective support. Yet in general Indonesia's industry associations tend to be exceedingly weak. They are not professionally staffed; they are rarely accountable to their members; they tend to be' captured by powerful players in their industries; and their umbrella organization lacks credible autonomy from government. As the country study makes clear, it will be no simple matter to re-orient a substantial number of the country's industry associations to play a constructive - promotional role ....... The hopeful lesson concerns the experience of the Export Support Board, created in 1985 in response to the perception of filure of the exisdng export institutions. The ESB represents an intriguing departure from the institutional model adopted earlier in that its role was to finance marketing support, not deliver it directly: it provided partial subsidies to nascent exporters who sought to participate in trade fairs abroad, to develop export promotional material, and (as discussed farther under the rubric of technical support systems) to employ technical consultants from abroad. Its complement of professional staff never exceeded 10 persons, yet its services were more widely used, and were rated more highly among the firms surveyed, than those provided by the elaborate, central export edifices. 23 -The performance of Colombia's national export agency, PROEXPO (created in 1967), in providing direct marketing support to firns also has been less than impressive.2/ Relatively few of the Colombian firms which used collective support (Table 10), reported that it came from PROEXPO. The industry associations, by contrast, show considerable promise in this area, especially those in the leather and (more recently) garments industries. Working closely witi their member firms, these industry associations are developing the sort of sector-specific knowledge and skills which cannot realistically be expected from general purpose agencia like PROEXPO. A successful hybrid form which is beginning to take hold is for PROEXPO and other public sector agencies to work collaboratively with industry associations - with the public agencies providing some funding to help organize fairs and assist visits abroad by potential exporters. 11.4: public Policy towards Expor Markting. For decades, countries have used high profile national export agencies to support their export efforts. Compartive empirical analysis in Indonesia, Japan, Korea and Colombia suggests that, viewed from the SME perspective, and judging by the experience of successful exporters, initiatives along these lines may be misplaced. For one thing, private networks tum out to be the dominant means whereby SMEs enter into cxport markets. For anotcb*r, for the subgroups of firms for whom collective marketing support has proven to be useful, it has generaLly been of a kind that can more effectively be provided by decentralized organizations such as industry associations, or local govermnents and chambers of commerce. Whether these decentralized organizations indeed provide useful support to SMEs appears to depend in part, however, on whether tiey are dominated by large finns, or whether SMEs have a significant voice. A number of levels of government plausibly could play a usefully role in facilitating the development of decentralized organizations capable of sponsoring participation in fairs and - offering SMEs other forms of effective marketing support. 12/ Although note that its primary focus has been on the provision of credit for exporters, an activity which it apparently has undertaken impressively. 24 Turning specifically to national authorities, aside from facilitating a supportive business and incentive enviroment for exporting, is their appropriate role in supporting export marketing? One task is to "market the country" - to raise international awareness of the potential advantages of doing business, including export business, with the emergig export nation.2Q/ What emerges from t'lis study is that a case can also be made to encourage through partial subsidies - as opposed to directly deliver - the use of export marketing supports. Consideration could also be given to using the lure of financial support to encourage the development of decentralized delivery of export marketing senrices - to facilitate the establishment of subsector specific industry associations that are responsive to SME concerns, to encourage those that function primarily as lobbying organizations to develop the capability to provide export marketing and other services to their members, or to encourage the emergence of private providers of export marketing services, including for-profit sponsors of trade fairs. Finally, looking beyond specific collective services, public policy may have some role to play in nurturing the private networks that have emerged in this comparative study as the bedrock of export marketing efforts. Although not studied in any depth, the case studies uncovered tanalizng hints of pr-active goverment efforts to promote such networks: public support for the development of private trading companies in Korea and, earlier, in Japan; Japanese legislation to protect the iterests of smal firms embedded in subcontractng relationships; actions in Korea to designate specific lines of business for subcontracting rather than vertically integrated production. To be sure, as is documented by both the individual case studies and the volrmniinous literatre on the determinants of inter-firm contacting arrangementss,2J/ the development of these networks has largely been the result of spontaneous- private forces, often deeply rooted in national history and culture. Yet, insofar as incentives can encourage subcontracting or entry by export traders (both large and small) the opporunities 2Q/ For a detailed discussion of how to proceed with this task, see Wells and Wint (1990). 2 L1/ - For one key contrbution, see Williamson (1985). 21. For key co.tn 25 for SMEs to participate in export- markets would be substantially enhanced, with significant social returns. --m: Developing new products and acquiring new production capabilities, and improving upon those that a firm already has at its disposal, are crucial to success in export markets and in all four countries most of the SMEs surveyed reported substantial technical improvements. One important influence on technical effort and perfonnance is the extent to which policies expose firms to domestic or international competition, and thereby act as a spur to improve capabilities. Another important influence comprises education policy, including importantly education in science and engineering, and its impact on the supply of technologically-usefil human capital. Also influendtal are management processes within finns and their impact on on- the-job technological learning (including through reverse engineering). The present study focuses on a fourth influence, namely the micro-environment of private, governmental and NGO market and institutional technological supports external to the finr, on which it can draw as it seeks to build technological capability. A central goal is to evaluate the impact of collective technical supports. In principle, their impac-t could be substantial: inter-firm spillovers of knowledge are pervasive, so individual firms do not capture for themselves all the benefits of their investments in knowledge, and in the absence of solutions to classic problems of collective action may rationally underinvest. However, the challenges of collective provision can also be substntial, since for information to be useful, ik may need to be tailored to the specific needs of an individual industry, if not an individual fum. Moroever, as with export marketing, there exists a variety of external private mechanisms which support a firm's technological effort. Consequently, we embed our analysis of collective technical support within the broader context of the range of private and collective mechanisms which can aid technological effort. Three broad categories of external technological support can be distinguished.- The first category comprises technological learing that occurs as a byproduct of a firm's transactions categor copie:e-crn 26 with its buyers and suppliers. The second category comprises technological learning that is facilitated by a firm's being embedded in an "information-rich" environment, one replete with other firms engaged in similar activities, with a menu of courses that address its specific business problems, and with access to a network of specialized consultants. The final category comprises "high-intensity" technological support, including formal technology transfer agreements with other private finms engaged in a similar line of business, and sustained joint work with specialist technology institutions. Note that while the channels for leaniing in the first category are entirely private, in the latter two categories the channels might be either private or collective. We first review comparatively the empirical results from the four country studies as to- which sources proved most useful in helping finns to acqWire technological capability. Thereafter, we evaluate in some depth the specific role of collective technological support. M. 1 Leading Sources of Technological Capability Five of the subsectors (garments in Indonesia and Colombia, both fiurniture subsectors in Indonesia, and leather products in Colombia) use craft-based technologies; two (woven textiles in Korea and Japan) use simple, engineering-based technologies; and five (machinery in Colombia, automotive components in Korea and Japan, and electronic products and factory automation equipment in Korea) are examples of more complex, engineering-based technologies.2=/ While craft-based and engineering-based activities pose distinct technological challenges, each can be met in distinctive ways. Indeed, as the comparative review which follows will highlight, and for reasons that are closely analogous to those identified earlier for export marketing, there turn out to be distinctive national patterns in the external mechanism of technological support for both craft- and engineering-based activities. Japan. As Table 12A reveals, the leading external sources of technological capability are remarkably similar for Japan's silverware, woven textile and auto parts SMEs, despite very substantial subsectoral differences in the technological challenges. Paralleling export marketing, - 22/ Japan's silverware subsector is a hybrid, with strong roots as a craft-based industry but with increasing application of science-based technologies to -sustain competitive advantage. 27 vertical relations with large finms emerge as a crucial source of external capability in all dtree subsectors: chemical companies and (to a lesser extent) trading houses are crucial sources of technological capability for woven textile SMEs; and silverware and auto parts SMEs receive important technologicial support from their immediate principals in the multi-tier subcontracting chain. More broadly, the Japan study highlights the power of inter-firm flows of information within industrial districts as an important source of technological capability. All three subsectors dominated the economic base of a narrow geographical area - silverware in Tsubame city, synthetic woven textiles in Fukui-Ishikawa prefecture and auto parts in Ohta city. Vertical linkages represent one example of inter-firm information flows in these industial districts. In addition, Table 12A identifies similar firms and public agencies (organized, as discussed further below, at the district level) as useful sources of capability. And equipment suppliers - many of which are themselves located within the specialized, product-centered industrial districts - emerge in all three subsectors as an external source of technological capability second only to subcontracting principals. Indonesia. Interactions with the intemational marketplace emerge in Table 12B as the most important source of technological capability for Indonesian SMEs. While both pnbumi and non-pribumi firms have thus gained technologically from export expansion, Table 12B also points to distinctive advantages enjoyed by non-pnbumi fins. In both the rattan and garments subsectors, expatriate employees play a disproportionately important role as a source of technological capability for non-pribuin finns vis-a-vis their pribumi counterparts. While half of the non-pnbumi users recruited expatriates from Hong Kong and Taiwan, both Chinese centers, no_pnbumi users drew expatriates from these sources. Pribumi firms rely disproportionately on substitute channels of technical leaming. For furniture finns (both rattan producers and those from Jepara), technical literature has played an important role. As for pribumi gdrment firms (whose role in the garment industry is an exceedingly small one), Table 12B suggests that they have not reaped any direct technological benefits from Indonesia's export success, but instead rely on the local diffusion of technological capability by private 27a TABLE 12: LEADING EXTERNAL SOURCES OF TECHNOLOGICAL UPGRADING (AVERAGE USEFULNESS SCORE FOR ALL FIRMS - BRACKETS: I - LEAST IMPORTANT; 5 - MOST IMPORTANT) I____________ PRIVATE SOURCES COLLECTIVE SOURCES I LEADING I SECOND THIRD SOURCE PUBLIC I INDUSTRY . ______.___.j SOURCE ISOURCE ' AGENCY ASSOCIATION A: JAPAN ._._._. _._ Silverware Subcontracting Equipment Supplier Similar Firms (1.7) 1.8 1.4 Principal (2.5) (1.8) _ : Woven Textiles Subcontracting Equipment Supplier - 2.0 1.0 Principal (4.0) (2.9) Auto Parts Equipment Supplier Subcontracting' Similar Firms (1.6) 1.5 1.4 .______________ _ -(2.8) Principal (2.7) . . . ' . . . B: INDONESIA Rattan Buyers (3.6) Expatriate (3.4) Similar Firms (2.3) 1.0. 1.7 (NonPribumi) .' Employees , ._._._ ._ . .Rattan (Pribumi) Buyers (3.6) Technical Expatriate (2.1) 1.5 2.1 literature (2.6) Employees Garments Buyers (3.0) Expatriate (3.0) Technical 1.0 . 1.0 (NonPribumi) Employees Literature (2.2) . . . . . Garments Similar Firms (3.8) Subcontracting Equipment (3.0) 1.0 1.0 (Pribumi) (3.3) Jcpara Buyers (3.2) Technical Subcontracting 1.1 1.7 Literatur. (2.0) (1.8) ._._.__ C: KOREA Woven Textiles Similar Firns (2.4) Buyers (2.0) Equipment 1.4 1.2 ._____________ _ .Suppliers (1.8) Auto Parts Subcontracting Formal Technology Equipment 2.6 1.1 Principal (3.7) Transfer (2.6) Suppliers (2.4) Electronic Parts International ' Equipment Subcontracting 1.5 1.2 Exhibition (3.0) Suppliers (2.3) Principals (2.0) . . Factory International Buycrs (1.8) Foreign 2.1 1.2 Automation Exhibition (2.2) Professionals (1.7): Formal Technology Transfer (1.7) . D: COLOMBIA ' Leather Foreign Buyers Equipment Similar Firms (2.4) 1.1 T 2.1 - a-n (2.7) Suppliers (2.6) Garments Equipmnent Technical Similar Firms (2.1) 1.3 2.0 Suppliers (3.0) Literature (2.3) _ . - | . - _ . -Machinery Technical Equipment Local Buyers (2.5) 2.0 1.6 Literature (3.9) Suppliers (2.9) _ _ . . * 28 participants in the industrial districts where much. of Indonesia's garnent export industry is located.2N Kwsa. In both Japan and Indonesia the leading external mechanisms of technological capability are those with which SMEs interact as a by-product of their core business activities. By contrast, as Table 12C summarizes, only among Korea's auto parts SMEs do vertical. linkages with procuring principals emerge as anywhere near as important a source of technological capability as they are in Japan.241 Instead, for the three subsectors with relatively complex technological requirements (that is, all but woven textiles), a disproportionately large number of leading extemal mechanisms involve the transfer of technological capabilities from abroad, with the transfer requiring conscious technological effort by firms. Thus: * international exhibitions are the most important external source of capability in both the electronic parts and factory automation subsectors; * formal technology transfer agreements are among the three most important private external sources in both the auto parts and factory automation subsectors; and * moonlighting over weekends by Japanese engineers is an important source of capability for the factory automation subsector. Also noteworthy is support from public agencies, which is the second most imporant external support in both the auto parts and factory automation subsectors. Overall, Korea's strategy of technology acquisition in the three complex subsectors appears strikingly purposeful, both at the firm and the industry level. Viewed against the backdrop of the Japanese experience, these purposeful efforts to learn from abroad can be interpreted as a substitute for the relative weakness of vertical inter-firm relations as a channel for technological learning. 231 Garment production for exports is concentrated in Bandung (where it is the dominant industry) and in Jakarta, both in West Java. 24/ The relatively weak performance of subcontracting as a channel of leaming for Korea vis a vis Japan has been observed by Porter (1990); see also Kang and Park (1990) and Joo Hoon Kiin (1991). 29 Even for woven textiles, where technology is simpler, it is noteworthy that vertical relations with buyers emerge in Table 12C as nowhere near as important for Korean SMEs as they are for SMEs in comparable industries in the other countries. Rather, horizontal linkages among firms in the textile industrial district of Taegu are most useful. Colombia. By comparison with Japan and Indonesia, Colombia's SMEs receive strikingly little support from buyers downstream. While in Korea, too, the evidence pointed to weak vertical ties, unlike Korea, Colombian filns do not appear to have purposefully used substitute private mechanisms to learn from abroad. Rather, the overall picture in Table 12D is one of rather ad hoc pattems of technological learning. This overall sense of technological isolation is moderated somewhat by evidence in the table of horizontal inter-firm flows of information among garment and leather firms, and of some significant technological support role on the part of industry associations (on which more below). Overview. In all four countries, private rather than collective mechanisms emerge as the leading external sources of technological capability. Yet, underlying this aggregate similarity there are substantial cross-country (and crosssectoral) differences in mechanisms of technological acquisition. While our Inmber of subsectoral observations is limited (especially given the need to distinguish between craft-based and engineering-based activities), we nonetheless find it heuristically useful to identify in Table 13 distinctive country patterns of tchnological acquisition. Paralleling export marketing, a striking feature of Table 13 is the influence of country "endowments" on the mechanisms of technological acquisition by SMEs. In Japan, strong vertical and horizontal inter-finn relations drive the technology acquisition process. In Indonesia, the crucial drivers are international linkages, both longstanding ones within an extended Chinese community and the more recent inflow of foreign buyers. In Taegu, Korea (which apparently is something of an exception on the Korean scene) horizontal linkages within an industrial district play an important role. Where endowments are limited, the challenge of technological acquistion is a formidable one, and as Colombia's machinery industry illustrates, the consequence can be 29a TABLE 13: MECHANISMS OF EXTERNAL TECHNICAL SUPPORT IN FOUR COUNTRIES - AN OVERVIEW CRAFT-BASED ENGINEERING-BASED ACTIVITIES ACTIVITIES JAPAN Vertical inter-firm relations Vertical inter-firm relations :_______________________ -plus industrial district plus industrial district INDONESIA Linkages through international marketplace ._:_ _ KOREA Horizontal flows within Activist Technology industrial district strategies at firm and _______ _______ ______ _______ _______ ______ industry level COLOMBIA Limited industrial district Ad Hoc Technological :____A ____: ________-__ phenomenon Learning 30 technological isolation and ad hoc learning. Yet the experiences of both Korea's engineering- based SMEs, and (to a lesser extent) Colombia's craft-based leather and garment SMEs suggest that it is possible to successfully surmount this challenge via activist strategies at both die firm and collective level, It is to this collective dimension of the technology acquisition process to -which we now trm. III.2: Collective Techical Support: Its DeMn=d and Us Private and collective mechanisms of technical support can be viewed as partial substitutes. SMEs might be hypothesized to initially meet their technological needs internally, as a by-product of their core business activities, or through existing business contacts. Only if they still perceive25/ unmet technological needs are they likely to demand collective technical support. Thus, building on the previous section: * Demand for collective support is likely to be higher the more complex are the technological requirements of production - lowest for craft-based technologies, higher for simple engineering technologies, and higher stil for complex engineering technologies. D* emand for collective support is likely to be higher the weaker are the country- and industial organization-related "endowments" of private technological networks available to a firm. A straightforward extension of these hypotheses is that, within a given subsector, there can be variations in the technology-related endowments of individual SMEs, with the demand for collective technical support greater for poorly-endowed firms.Zg 251 There is abundant evidence that many technologically primitive firms do not perceive their own limitations. See, for example, evidence from Sri Lanka and Tanzania in Levy (1993). 2Xt Viewed in relation to the first hypothesis, it might be argued that the relevant variable is technological complexity relative to a firm's internal capabilities, and thus that demand for collective technical support will be greater for firms with weaker capabilities. Viewed in relation to the second hypothesis, SMEs within a given subsector might plausibly vary in their. access to private technological networks, and hence in their demand for colective technical support. 31 Evidence in the country studies enables us to group our country-subsector observations to correspond with the three schedules of demand for collecdve technical support depicted in Figure 2: 'Demand is lowest (i.e. corresponds with DiDl) for SMEs in the six Japanese and Indonesian subsectors studied, and for Korca's woven textile SMEs. Their technologies are relatively simple, and they have good access to private technological supports (among Indonesian firms, especially so for the non- pribumi). * Demand is somewhat higher (i.e. corresponds with D2D2) in Colombia's two craft-based subsectors. While technologies are relatively simple, their access to private technological supports is weaker than for their Indonesian and Japanese counterparts. * Demand is highest (i.e. correspords with D3D3) for Colombia's machinery SMEs, and for SMEs in Korea's three engineering based subsectors. Additionally, within individual subsectors, smaller and less-well-connected fims might be hypothesized to be on a higher demand schedule than their larger counterparts. As will be discussed furtiher below, there also are subsuntial cross-country (and cross- sectoral) variations in the supply capabilities of collective instutions. Consequently, there can be formidable identification problems in distngushing between supply-side and demand-side explanations for cross-country (and cross-sectoral) differences in the use of collective technical- support.. Demand-side variations alone seem sufficient to account for two of the empirical paterns observed. First, as is evident in Table 12C, mi Korea collective tehical support was valued most highly for the technologically-complex automotive parts and factory automation subsectors, and least highly for the technologically-straightforward woven textile subsector, a pattem which is consistent with the first hypothesis presented above. Second, both the Colombian and Indonesian case studies present evidence that within individual subsectors collective technical support was more highly valued by firms which were "marginal' at start- p. Thus,- in Indonesia's rattan and wood furniture bsectors, colective technical supports were utlized disproportionately by smaller pnbi firms with less-educated entrepreneurs, - :, -ffthough even for this group the role of colective supports has been modest. In Colombia 31a Figure 2 - D3 ,S1 Si< :< D3. S2 D2 S3 Dl Use of Collective Technical Support 32 colective support was more highly valued by the smaller leather firms - with the avenge usefulness score declining monotonically from 3.0 for the smallest firms to 1.0 for the largest. In both lndonesia and Colombia, the field surveys uncovered substantial subsequent success, .including export success, by these initially "marginal" firms. On the supply-side, evidence presented in the next subsection suggests that it is reasonable as a first approximation to distinguish between three supply schedules: * SISi, corresponding to Indonesia, whose institutional capabilities to supply .support appear weakest among the four countries studied; * S252, corresponding to Colombia, which appears to have an intermediate level of collective supply capability; and * M3M3, corresponding to Korea and Japan, where the institutional capabilities for collective supply are strongest. As is illustrated in Figure 2, taken together the above propositions as to supply and demand IMPly that, among the countries and subsectors studied, the use of collective technical support will be highest for SMEs in Korea's three engineering-based subsectors, and will be lowest for Indonesian finns (especially the non-pribumi).22/ By and large, this pattem is evident in Table 12. M--.3: The Delivery of Collective SuDport Colective technical support takes on one of two very different forms. It can be "broad- based", and work to facilitate the emergence of an "information-rich" environment. Alternatively, it can try to promote "high-intensity" technological learning by supplying technical inputs direcdy to firns. While it is reasonably straightforward to distinguish between the supply of "broad-based" and of "high-intensity" support for Colombia, Indonesia and Japan, the two tpes of support are provided jointly in Korea. Consequently, the Korean experience is presented separately. 22/ The relative importance of collective technical support among the remaning seven subsectors depends on the magnitudes of shifts in demand and supply schedules, and hence cannot be depicted unambiguously. 32a TABLE 14: THE USEFULNESS OF BROAD-BASED COLLECTIVE TECHNICAL SUPPORT IN SELECTED SUBSECTORS AND COUNTRIES COUNTRY-SUBSECTOR USERS/ AVERAGE NUMBER OF FIRMS SCORING I . ~~~~~~~RESPONDENTS SCORE AMONG;. __________________. _ USERS 3.-4. OR 5 4 OR 5 INDONESIA : ._._-_.__ RATTAN Courses - 15/33 Industry Association 15/33 -3.1 10 5 Consultnts 9/33 3.0 5 4 JEPARA Courses 12/24 Industry Association 13/24 2.2 5 3 COLOM BIA _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ LEATHER PRODUCTS Courses 21/34 Industry Association 17134 3.2 1- 7 GARMENTS Courses 32142 . Collective Providers 20142 3.4 9 MACHINERY Courses 27144 . Public Technology Agency 16144 3.8 7 Product Standards Agency 12/44 3.5 8 JAPAN TSUBAMEfSILVERWARE Technical Cemter 18/36 2.5 10 4 Industry Association 13135 2.2 6 2 FUKUI/TEXTILES Technical Center 14/29 3.0 8 7 Industry Association 3/30 3.7 2 2 OHTAIAUTO PARTS Technical Center 6/33 3.8 5 4 Industry Association 5130. 3.2 2 2 33 Broad-based collective sport in noesi. Colomia AnB p. "Broad-baed" collective support works to enhance the overall availability of usable informaifon, leaving firms to judge what infonnation sources might be most useful, and how they might be adapted to a firm's specific needs. Examples include: sponsoring courses oa specialized topics; facilitating theu use of specialized consultants (either directly by making a consultant available to a broad array of finns, or indirectly by providing financial support for the use of consultants); and promoting information-sharing among firms. Supports along these lines are relevant across all levels of technological complexity, and the institutional demands of provision are relatively light. As Table 14 summarizes, such support was used -and reported to be quite useful -in Indonesia, Colombia and Japan. In Indonesia, just under half (27 of 57) of the finns surveyed in the rattan and carved wooden furniture subsectors attended courses that focused on specific technical subjects relevant to their businesses (e.g. wood-drying and finishing). Virually all of these were offerings by the fmuniture industry association (ASMINDO) or by independent non- governmental organizations, and about half of the participant gave a score of the or higher to collective support as a source of technological capabilit. Use of technical consultants was confined to less than twenty percent of the firns. However, half of the users gave scores of four or five to the support provided by consultants - and eighty percent of the users reported that the costs of using consulting support were at least partially subsidized by collective providers. There was more diversity in Colombia in the sources of broad-based. collective support, althoLgh courses predominated. In leather products, the collective lead has been taken by a dynamic industry association, ASOCUEROS, which has off-ers a wide variety of courses, sponsors technical consultants, and in general encourages networldng and information-sharing among firms. Half of the 34 leather finrs received technical support from.ASOCUEROS, and seven firms (o,yer twenty percent of the sample) gave this support a usefulness score of 4 or 5. In the garment industry, over dthe-fourths of the finns sampled had taken technical courses relevant to the business -wiih the leading poviders being the national training organization, 34 SENA, the economy-wide small business association, ACOPI, and the garmnit industry association. The non-electrical machinety industry is technologicaly more complex dun the others and, perhaps reflecting this differmce, specialist technology agencios rat than industy associations emerge as the leading collective providers. However, even for this subsector, courses (rather dn high-intesity support) emerge as dte dominant source of coUective support, with SENA the leading provider.2&! In Japan, tchnical centers under the umbrella of local govemenmts are the primary providers of broad-based collective technical support. Their role was most significant in two of the three subsectors surveyed,29t where (as Table 14 shows) about half the rms surveyed used, their services, with 56% of these conferring upon them a usefulnes score of 3 or higher. While in the past the technical centers also provided high-intensity support, at present their effors- appear largely to be confined to sponsoring courses. providing test and inspection services, and providing assistn on caummon technical problems that aise from normal business practices, including a modest nmmber of firm visits. A further filmctioh oa Japan's collective technical support system - difficult to measure but, by all accounts, very useful - is as a node in an iformationu-ch networtl The Japan silverware case study includes one especially vivid example of how ths network functions. A Tsubame entrepreeu came up with the idea of making silverware with handles whose shape could adjust to fit the needs of individual disabled users. The collective support network (in this instance the local business asscation) helped him identify a range of semia and exhibitions held throughout Japan which plausibly could provide b with useful informaton. Evetally he came upon a company which was developing precisy the mtrial he needed for his 2-1 -Intrestinly, 27% of the Colombian machinery firms sampled (12 of44j also repod support from Columbia's product standards certification agency - with an aerage usefulness score among users of 3.5. There was no evidence of sundards agenies playningany usefil role in Indonesia. These agencies appantly are important in both Korea and Japan - however, ffieir roles were not eamined systematically in the Korean and Japanese case studies. - 221 - In the third subsef r - auto parts in Ohta - only 18% of firms sureyed (6 of 33) used the center; however, the average usefulness score for these six users was 3.8.; 35 product. However the company was part of a large group. Ordinarily, such a company would have had little incentive to work with a small Tsubamne firm. Again, however, collective support providers (in this case a public research center) intervened to help open doors. The product was successfully developed,. has won several awards for innovation, and is being marketed successfuly. Overall, a conisteint feature in Indonesia, Colombia and Japan was that broad-based collective support was most effectively delivered by decentralized institutions - either by industry associations, independent non-governmenta organizations, or by local governments in specialized industral districts. The record of centralized institutions in deliverin services was more uneven. Financing, though, sometimes was provided by center institutions - the Export Support Boad (for technical consultant) in Indonesia, the central government mn Japan, and PROEXPO in Colombia. This pattern of central finacing and decentraized delivery reflect the wide diversity across activities in the kind of information that is useful, and consequendy the need for deliverers of broad-based collective technical support to be close to - and imfmila with the needs of - reasonably homogenous client groups. Hig-inenitytecnial uport in Colombia. Indonesia and Japan. TIhe goal of high- intenity collective support is to meet those specific technological needs of finns .which are not adequately addressed through other channels. Demand for support along these lines is likely to emerge only at relatively substantial levels of technological complexity. Simiilarly,- the insitLtional demands of supplying usefu high-intnsity support are likely to be substatial.- the collective agency supplying such support needs to have more competence on very specific problems (at quite sophisticated levels of technology) than do the firms themselves. In neither Indonesia nor Colombia did sample firmns report benefiting on any significant scale from public programs to provide direct high-intensity technical assistance. Indonesia's Minisry of Industy has in place a network of technical extension services which in principle might have been intended to provvide such support, but which in practice suppliedvirtually nothiing of benfit to thie surveyed. firmns.-The only Colombian public agency~ identified'as 36 directly providing technical support on any significant scale was SENA. However, as we have seen, virtually all of this support was broad-based, complementary to SENt's primary mission as a vocational training institute. (SENA presently is in the midst of a radical restructuring of its training function, involving a substantial shift to decentralized mechanisms of delivery.) During earlier periods, Japan's technical centers indeed appear to have provided substantial high-intensity support. The Fukui technical center, for example, played an important role between 1910 and 1930 in introducing the technology for rayon production into the area. And back in -the 1700s Tsubame's local government invited producers from Tokyo to teach nail production to village manufacturers, an initiative which was the forerunner of the silverware industry which dominates the area today. Presently, the role of these centers is predominantly the broad-based one outlined above although, as the example of handle 'Anovation illustrated, in a modest way all three technical centers continue to support the efforts of leading firms. The case of Korea. By contrast with the other three countries, the Korean firms reported reeiving substantial benefits from dh country's network of state-owned support institutions. Table 15 summrze the survey results on the utilization of collective technical support in Korea. Sixty-eight percent of the firms sampled use at least one type of collective support, with the utilization rate ranging from a low of 26% for woven textiles, to 85 % and 90% for electronic and auto parts respectively, to a high of 93% among factory automnation firms. The table distinguishes among four types of collective technical support. Provision of technology information can be viewed as "broad-based" support, and joint and contract technology development as "high-intensity" support, with technology assistance and training falling in between. Collective support is most common for purposes of training, and least common for joint contract technology development. Support for joint technology development appears to be valued most highly. Unlike the patterns in Colombia and Indonesia, the finms which use the Korean system tend to be disproportionately large and export-oriented, suggesting that Korea's collective tecbnical support is rather more on the "high-intensity" end of the specum. The larger role for high-intensity support corresponds to-the higher proportion in the Korean sample of activities requirng complex gineing tecnologies. TALEL&l UTILIZtATION OF KOREA'SLEADING SPECIALIST TECHINOLOG Y INSTITUTIONS IN FOUR SUBSECTORS ___ IVOVENV TEXTIES AUTO PARTS (m -20) ELECTRONIC PARTS FACTORY (n u l) (i=20) AVTOAMATON TECHNOLOGY LEAiDING USE LEADING USE LEADING USE LEADING USE ALL INS TITUTION (BY LEADING USE)__ _ _ _ _ _ _ _ _ _ _ _ _ _ __ _ _ __ _ _ __ _ _ _ _ __ _ _ NVAI. AVE. INO. SEC. NUM. I A VE. JNO. SEC. NUMI. 1A yE. NO.SE. NUAf- JA VE. NO. SEr N&f A VE. BEA SCORE _USaE DER SCORE USE BER SCORE jUSE HERR SCORE__ USEBE SCR TECHNOLOGY INFO. I ASSISTANCE _______ ___ Safi PROMIOTION 4 .3 0 6 2.8 3 5 3.8 2 a 2.4 6 34 3.2 COAIPANY _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ INDUSTRIAL 0 -0 5 4.2 0 I 0 4 3.s 0 10 3.9 ADvANCEMfENT ADAIIN1STRATION SVBSECTOR.SPECIFC 9 If.1 I 2 4.0 5 3.6 1 3.6 30 3.s 1NST1TU TIONS__ _ _ _ _ _ _ _ _ _ _ __ _ _ _ _ _ _ _ _ _ _ TRAINING - - - . --- KOREA SrANDARIJ 0 10 4.5 2 731 011 3.7 2 26 3.6 ASSOCIATlION2 43 1841 13564 3. CENVTER ___ ------ JOINT AND) CONTRA CT TECIIN0LOG1CAL DEVELOPMENT_________ KOREA ACADEiY OF 0 0 3 4.3 3 2 45 1 3.0 7 19 4.0 INDUSTRIAL TECIINoLoor KOREA INSTITUTE OF 0 0 2 41.5 a 2 3.0 1 4 4.3 1 10 4.3 TEChINOLOGY NUAIBER OF FIRMS (11) 4.0 (18) 3.8 1 (17) 3.5 37 3.5s (a3) 3.7 USING ATLEAST ONEI II OF TlE ABOVE (AND I II A VERAGE SCORES..I I 37 As Table 15 reveals, a central feature of the Korean technology support system is its institutional pluralism. Support is provided by six leading parastatals, plus subsector-specific institudons. For woven textiles, the leading specialist institution is the Korea Textile Industries Technology Institute, which provides a fill range of broad-based support, and for factory .automation it is the Korean Association of Machinery Industry, one of the most well-developed trade associations in Korea, and for electonic parts it is the Korea Electronics Technology Institute. -By contrast with the pattem for the other three countries, no single institution reaches more than 30% of the firms surveyed (and two service fewer than 10%), but viewed as a single system with multiple components the overall coverage is impressive. Moreover, the usefilness scores (as rated by users) are relatively high for all institutions: on a five-point scale, none averaged below 3. Indeed, the average score was 4 for the two institutions which provided the most intensive support. A. firther coast with the other counties is the predominant role of cetalized parastatals in delivering technical support. Even so, there is debate within Korea as to whether this centralized delivery mechanism remains appropriate, and in recent years a number of industry-specific technology support ishtutions have been established.3Q/ Moreover, comparative empirical research on parastatal behavior has revealed Korea to be an outlier in its ability to establish and sustain a relatively efficient parastatal sector. Further, in none of the other countries studed was the delivery of technical support by centralized parastatals. anything near as successful as in Korea. For countries that lack an overall record of strong performance by parastatals, an effort to establish a high-intensity network of collective technical support along Korean lines would appear to carry substantial risks. 32/ The present parastatal-driven support infrastructure resulted in part because of a general predisposition towards centralization inKorean society, and in part because all but one of the parastatals initially were established for other purposes and only after some years broadened their missions to include technical support for SMEs. Increasingly Korea is experimentng with more dece approaches tD providing support, with KOTMTI a leading example. 38 m.4: Public Policy Towards the Acquisition of Technological Capability As with export marketing, a central lesson of the comparative research is that private mechanisms play the leading role in the efforts of firns to acquire technological capability. Consequently, the priority task for public policy is to ensure that the business environment facilitates - rather than obstructs - the private-to-private flow of information. Key measures here might include: * an openness to expatriate workers and technology transfer from abroad; * investment in human capital, including engineering education; and a spatial policy which nurtures the emergence of urban industrial districts. Yet once a friendly business enviromnent is in place, the question remains as to whether there might be pro-active interventions at the micro level which can accelerate technological upgrading by fms. The comparative analysis of four countries suggests that there are - that there exist identifiable categories of firms, subsectors and countries for which the benefits derived from collective technical support are substantial. The analysis uncovered two distinctive approaches to collective technological support - broad-based and high-intensity support. The goal of broad-based support is to contribute to an information-rich environment, and its delivery mechanisms characteristically are decentralized, with the role of central government limited to that of partil financier. It can be a useful aid to the technological efforts of firms of all levels of technological complexity. By contrast, the goal Of high-intensity support is to offer direct technical assistance to firms, typically by parastatal providers. Its role becomes increasingly salient as industrl development begins to encompass technologically complex activities. While it is possible that - if successfil - the retums from high-intensity support can ber very substantial, on all other counts the advantages would appear to lie with broad-based support. Costs are likely to be lower, since unike high-intensity s provision of broad- based support does not reqlire erection of an elaborate physical and personnel infiastructure. Being institutionally less elaborate, the organizational demands of broad-based support are fewer, and the risks of organizational failure lower. And being decentalized in delivery, 39 broad-based providers are more likely to gauge accurately what kinds of technical support finms would indeed find useful. In sum, the country studies imply that 'broad-based" collective technical support may be worth pursuing in many settings. Countries that already have in place a well-function system of broad-based collecdve support, and are moving into technologically more advanced acdvities could consider also the option of "high-intensity" support, but in doing so should proceed with caution. IV: SOME COMMON THEMES We conclude by highlighting three key themes which cut across the analyses of technological and marketing support systems. First, the leading source of support comes from private channels - from buyers and traders, from similar firms, suppliers and subcontracting principals, and from the detenmined efforts of SMEs themselves. It follows that the first order of business for SME support policy is to ensure that the private marketplace can work, that liberal rules goveem the in ational flow of technical and marketing resources. Second, the benefits of private support mechanisms are not available equally to all, but vary with an SME's "endowments": whether SMEs are embedded in pr-exidng private inter- firm or community-based networks; whether firms are pioneers in a new activity or one of many participants in an already mature subsector, whether the country in which they are active has a high profile in the export marketplace. Collective market and technological supports tum out -to be used and valued dispropzrtionately by less well-endowed - though subsequently successful - firms. Third, the record of delivery of collective thnial and marketing support is a chequered one, but some promising new approaches appear to be coming to the fore. The most promising interventions are those with a "light touch": their delivery mecbanisms generaly are decentralized; and their goals are to support, rather tha supplant, the private marketplace - to co-finance SMB efforts to tap the maxketplace for new marketing and technical resources, to I.~~~~~~~~~~~~~~~~~~ I .40 make available additional sources of technical information, and to facilitate communication between firms, including between SMEs and international buyers. In all, the empirical results suggests that, consistent with the emphasis of the previous decade, the overali business and incentive environment is the most important determinant of the effectiveness of marketing and technological support systems for SMEs. However, the results also caution against complacency, against the presumption that a liberalized private marketplace will be sufficient to secure industrial development. References Arrow, Kenneth J. (1962) "Economic Welfare and the Allocation of Resources for Invention" in NBER The Rate and Direction of Inventive Activity (Princeton: Princeton University Press) Berry, Albert and Dipak Mazumdar (1991) "Small-Scale Industry in the Asian-Pacific Region" Asian-Pacific Economic Literature, Vol. 5, No. 2, pp.35-67. Cortes, Mariluz, Albert Berry and Ashfaq Ishaq (1987) Success in Small and Medium-Scale Enterprises: The Evidence from Colombia (Oxford University Press for the World Bank Fei, John, Gustav Ranis and Shirley Kuo (1979) Growth with Equit: The Taiwan Case (New York: Oxford University Press) Hirschman, Albert (1958) The Strategy of Economic Development (New Haven: Yale University Press) Hogan, Paul, Donald B. Keesing and Andrew Singer (1991) "The Role of Support Services in Expanding Manufactured Exports in Developing Countries" World Bank Economic Development Institute, Seminar. Series Jacobs, Jane (1969) The Economy of Cities (New York Random House) Jones, Leroy P. and Sakong I1 (1980), Govermment. Business and E rnburship in Economic Development: The Korean Case (Cambridge: Harvard University Press) Kang, Moon Soo and Chun-Kyung Park (1990) "Industrial Structure Control" (in Korean), Seoul, Korea Development Institute Kim, Joo Hoon (1991) Liberalization. Subcontracting and Reorfanization (Seoul: Korea Development Institute) Krueger, Anne (1980) 'Export-led Industial C-rowth Reconsidered", Paper presented for the Eleventh PAFTA Conference, Seoul, Korea Kilby, Peter (1971) "Hunting the Heffalump" in Peter Kilby (ed.) Entrepreneurshig and Development (New York: The Free Press)4 Leibenstein, Harvey. (1963) Economic Development and Economic Growth (New York: John -Wiley and Sones) -(1968) "Entrepreneurship and Development" American Economic Review 58 Levy, Brian (1993) "Obstacles to Developing nigenous Small and Medium Enterpises: An Empirical Assessment", World Bank Economic Review. Vol. 7, No. I -(1991) "Transactions Costs, the Size of Firms and Industrial Policy: Lessons from a Comparative Analysis of the Footwear Industry in Korea and Taiwan" Journal of Development Economics. 34, pp. 151-178. (1989) "Export Traders, Market Development and Industrial Expansion", Williams College Research Memorandum Number 114, March. Little, Ian M.D., Dipak Mazumdar and John M. Page Jr. (1987) Small Manufacturing Enterprises, (Oxford U. Press for The World Bank) Marsden. Keith (1984) "Services for Small Firms: The Roles of Government Programmes and Market Networks in Thailand" International Labour Review 123, MarchlApril pp. 235-249 North, Douglass C. (1990) Institutions. Institutional Change and Economic Performance (New York: Cambridge University Press) Pack, Howard and Larry Westphal (1986) "Industrial Strategy and Technological Change: Thery vs Reality", Journal of Development Economics Vol. 22 Piore, Michael J. and Charles F. Sabel (1984) The Second Industrial Divide (New York: Basis Books) Porter, Michael E. (1990) The Comoetitive Advantage of Nations (New York: The Free Press, 1990) Pyke, Frank and Wemer Sengenberger (eds.) (1992) Industrial Districts and Local Economic Regeneration (Geneva: Intemational Isitute for Labour Studies) Ranis, Gustav and Chi Schive (1986) "Foreign Investment and Economic Development in Taiwan" in Walter Galenson (ed.) Trade Policy and Economic Development in Taiwan Rhee, Yung Whee (1988) "The Catalyst Model of Development: Lessons From Bangladesh"s Success with Garment Exports", Indusry Development Division, World Bank, mimLo -, Bruce, Ross-Larson and Garry Pursell (1984), Korea's Competitive Ed,e: Managgnste Entry into World Markets (Baltimore: John Hopkdns U. Press) Scuhmitz, Hubert and Bernard Musyck (1993) "Industrial Districts in Europe: Policy Lessons for Developing Countries? Institute of Development Studies, Discussion Paper Number 324 Steel, William, and Leila Webster (1990) "Ghana's Small Enterprise Sector: Survey of Adjustment Response and Constraints', Industry Series Paper Number 41, World Bank, Industry and Energy Deparunent Stiglitz, J. E. (1989) "Imperfect Information and Economic Development" American Economic Review papers and proceedings Webster, Leila (1991) "World Bank Lending for Small and Medium Enterprises: Fifteen Years of Experience" World Bank. Discussion Paper Number 113 Wells, Louis T. Jr. and Alvin G. Wint f'990) "Marketiig a Country": "Promotion as a Tool for Attracting Foreign Investment" World Bank Group Foreign Investment Advisory Service, Occasional Paper 1 Williamson, Oliver E. The Economic Institutions of Capitalism (New York: The Free Press, -1985) (1975) Markets and Hierarchies (New York: The Free Press) World Bank (1991) "World Bank Support for Small and Medium Industry in Selected Countries" Operations Evaluation Department, Report No. 9530 World Bank (1993) The East Asian Miracle: Economic Growth and Public Policy (Washington, D.C.: Oxford University Press) Policy Research Working Paper Series Contact Title Author Date for paper WPS1384 Fiscal Decentralization and David Sewell November 1994 G. Langton Intergovemmental Finances in the Christine I. Wailich 38392 Republic of Albania WPS1385 Fiscal Federalism Dimensions of Tax Robin Boadway November 1994 C. Jones Reform in Developing Countries Sandra Roberts - 37754 Anwar Shah WPS1386 EU Bananarama ill Brent Borrell December1994 G. Ilogon 33732 WPS1387 Fiscal Decentralization and the Size Jaber Ehdaie December 1994 C.Jones of Government: An Extension with 37699 Evidence from Cross-Country Data WPS1388 Does Voice Matter? For Public Samuel Paul December 1994 B. Moore Accountability, Yes 35261 WPS1389 Ruble Overhang and Ruble Shortage: Patrick Conway December 1994 L. Suki Were They the Sane Thing? 33974 WPS1390 Regional Integration and the Baltics: Pirita Sorsa December 1994 J. Ngaine Which Way? 37947 WPS1391 Where in the World is Population Jeff Kling December1994 S. Fallon Growth Bad? Lant Pritchett 38009 WPS1392 Some Economic Consequences of Jean-Paul Azam December 1994 C.Jones the Transition from CMI War to Peace David Be,an 37699 Paul Coirer Stefan Dercon Jan Gunning Sanjay Pradhan WPS1 393 The Interface of Trade, Investment, J. Luis Guasch December 1994 J. Troncoso and Competition Policies: Issues and Sarath Rajapatirana -37826 Challenges for Latin America WPS1394 Macroeconomic Reform and Growth Lawrence Bouton December 1994 R. Marlin in Africa: Adjustment in Afiica Christine Jones 39026 Revisited Miguel Kiguel WPS1395 A Typology of Foreign Exchange Janine Aron December 1994 R. Luz Auction Markets in Sub-Saharan Ibrahim Elbadawi 39059 Africa WPS1396 Foreign Exchange Aucton Markets Janine Aron December1994 R.Luz In Sub-Saharan Africa Dynamic Ibrahim Bbadawi 39059 Models for Auction Exchange Rates Policy Research Working Paper Series Contact Title Author Date for paper WPS1397 Are Private Capital Flows to Url Dadush December 1994 J. Queen Developing Countries Sustainable? Ashok Dhareshwar 33740 Ron Johannes WPS1398 The Cost of Air Polution Abatement Raymond S. Hartrnan December 1994 E, Schaper David Wheeler 33457 Manjula Singh WPS1399 How Important to India's Poor is the Martin Ravalilon December 1994 P. Cook Urban-Rural Composition of Growth? Gaurav Datt 33902 WPS1400 Technical and Marketing Support Brian Levy with December 1994 D. Evans Systems for Successful Small and Albert Berry, Motoshige Itoh, 38526 Medium-Size Enterprises in Four Linsu Kim, Jeffrey Nugent, Countries and Shujlro Urata WPS1401 Colombia's Small and Medium-Size Albert Berry December 1994 D. Evans Exporters and Their Support Systems Jose Escandon 38526 WPS1402 Indonesia's Small and Medium-Size Albert Berry December'1994 D. Evans Exporters and Their Support Systems Bran Levy 38526 WPS1403 Small and Medium-Size Enterprise Motoshige Itoh December1994 D. Evans. Support Policies in Japan, Shujiro Urata 38526 WPS1404 The Republic of Korea's Small and Linsu Kim December 1994 D. Evans Medium-Size Enterprises and Their Jeffrey B. Nugent 38526 Support Systems