Small enterprises under adjustment in Senegal

INDUSTRY AND ENERGY DEPARTMENT WORKING PAPER INDUSTRY SERIES PAPER No. 55 Small Enterprises Under Adjustment in Senegal March 1992 The World Bank Industry and Energy Department, OSP  SMALL ENTERPRISES UNDER ADJUSTMENT IN SENEGAL Ron Parker and William F. Steel March 1992 Industry Development Division Industry & Energy Department World Bank  ACKNOWLEDGEMENTS The data for this study were collected by Charbel Zarour and the Centr Africain t'Etudes Superieures en Gestion (CESAG), sponsored by the Industry and Energy Division of the Sahelian Department in the World Bank's Africa Regional Office. Larry Salmen, Sangone Amar, and Maurice Clerc helped gready in designing and carnying out the surveys. Data processing was provided by Patricia Bibes and Narayana Poduval, and word processing by Wilson Peiris and Vivian Cherian. The authors are grateful to Fran;ois Laporte, Salomon Samen, Yi Shao and participants in a World Bank seminar for helpful comments.  Table of Contents Page No. ExEClUIV E3 Sun UMMvIAR Y .. . .................................. i I NlIODUCTriON .............. ... . ............................. 1 ANALYMCALFRAMEWORK ......................................... 2 Dualism and the Dynaniic mle of SSEs ............................................... 2 Overcoa ning financial market impmrfnctions ......................................... 3 Structunra adjustment to overcotne policy distrtiorns ................................ 3 Experiences of African SSEs under adjustment ...................................... 4 THIE SENEGAiJLESE IDUSTERLA CONTEXT ............................................. 6 Industrial history and performance ........ . ............................ 6 Profile of the SSE sector ..................................., 8 Support for SSEs .................................. 10 The need for industial adjustment . . ............................ 11 iTfiE N2PI AND rIMS EFFEC'rS .................................. 12 Measures ...............-.......................... 12 Itnplemnentation ........ . . ......................... 13 Expectei inmpact ...................................... IS Inmpact on manufacturing . . . ............................... 16 Hlywpotheses to be examnined . ............ ... .................. 17 THE SURVY AND SAMiPLE CHI{ARAC SER FtICS .................................. 19 The sample .......................... Entrepreneurs ........................................... 20 Subsectors ................................... 22 Markets ................................... 23 New firms ..................................... 23 RESPONSES UNDER ADJUSTMENT ................................... 25 Output response ................................... 25 Business practices ............................................ 27 Keys to finn growth . .. ............................ 28 Constraints on operations and investmnent .................................. 31 Finalcee .. ................................ 36 Prospects .................................. . 38 CONCIUSIONS ....................................... .. .. 40  REFERENCES ANNEXES: Annex 1: Exchange Rate Fluctuations, 1980-1990 Annex 2: Census of Industl Firm Closings, 1986-89 Annex 3: Output of Key Industrial Products, 1980-89 Annex 4: Distribution of Survey Responses Table 1: Entrepreneurs' characteristics Table 2: Product markets Table 3: Characteristics of new and pre-NPI firms Table 4: Changes in prmAuction, business practices and competidon Table 5: Characteristics of firms by production trend Table 6: Changes since 1986 by production trend-industry Table 7: Distribution of flexibility factors hy producdon trend Table 8: Major problems in operadons by production trend-industry Table 9: Changes since 1986 by production trend-commercial Table 10: Major problems in operations by production trend-commercial Table 11: Constraints on operations, detailed Table 12: Constraints on operations by category Table 13: Constraints on new investment Table 14: Access to formal cadit by production trend and size Table 15: Access to formal credit by nationality Table 16: Expectations for the future by production trend and size Annex 5: Statiscal Notes  EXECUTIVE SUMMARY Senegal, like other Afiican countries, has been undergoing reforms aimed at introducing a more competitive structure to industry and enabling the private sector to play an increasing role in growth and industrialization. Private sector responses have been difficult to measure, however, since most indigenous private firms are small in scale. This study addresses the gap in knowledge by a firm-level survey of small-scale enterprises in comparison with both microenterptises and larger-firns (21 firms in each category). The evidence suggests that some small finns are dynanic agents that can help generate industrial growth. Liberalization policies in Senegal have led to difficulties for the most visible actors: large- scale firms that had benefited from past protection against imports and local competition. These companies have complained about the detrimental effects of reforms. This study shows that small enterprises may have compensated to some extent for the decline of large firis. Despite the weak demand growth in Senegal since 1986 and no substantial change in export incentives, a substantial share of small fimns have managed to increase output and profits. In addition, a high proportion of small-scale enterprises (SSEs) were post-NPI creations, suggesting a relatively high rate of new investment at this level. The survey found that firms were making changes in their operations to adjust to the more open environment for business. In particular, some firms were taking advantage of easier access to imported inputs, and those that did so generally increased output. Changes in product mix and new equipment purchases also aided some firms in adjusting, though in some instances these actions were the belated responses of large fms with falling sales. Small firns made adjustments in their operations at high rates relative to firms of other sizes. Nearly 60 percent of SSEs changed their product mix, bought new equipment or inereased their imported share of raw materials. Over three quarters sought a bank loan, suggesting that these firms saw sufficiently profitable opportunities to take on debt. The study indicated that internal factors were probably more indicative of a finm's prospects than the external environment. Expanding firms were at least as likely as contracting ones to be found in markets in which comnpetition had increased. Despite complaints that import liberalization was squeezing out Senegalese firms, expanding companies also were more likely to be in competition with imports. This fmding suggests that the NPI is encouraging some fims to orient themselves towards producing internationally competitive goods. Though few firms were primarily exporters, a substantial proportion (40 percent) exported at least some of their product (themselves or through an intennediary), and both expanding and new firms were more likely tc export Almost all new fims in the samplF were SSEs. Most new frms had a strong external orientation compared to other sample firms. Those that were not exporters or in import-competing markets were oriented towards specialized market niches. This response of new investment in SSEs is encouraging from the standpoint of potential for future growth. Despite the finding that factors internal to the frm were associated with growth, finns were most likely to cite external factors as constraints on their operation, especially competition and demand. Among seco.."-i constraints, finance, raw materials and business environment were seen as problems by SSEs. Specific areas of high concern were credit for raw materials and taxes. Studies in other countries have found access to finance to be an overwhelming constraint for small businesses. This study indicated that finance was not as widespread a problem for finns in Senegal, except for small firms with investment opportunities. In particular, frms that had increased output subsequent to the NPI but were not planning future expansion cited absence of finance as their pincipal constraint. Other firms were less interested in credit, pardy because dtcy previously had access to finance but also because the economy was stagnating.  INTRODUCTION This study looks at private manufacturing and commercial enterprises in tin context of Senegal's Nouvelle Polidque Industrielle (NPI) enacted in 1986. It examines changes in firms' operatons subsequent to the refonn process, seeking to uncover and explain variations in impact amnong different types of firs. Observations from other African countries have identified fimn size as a potentally important factor in analyzing the nature and extent of responses to reform. Though results differ somewhat from country to country, a substantial concentration of entrepreneurial dynamism in a post-adjustment environment has typically been located in small-scale enterprises (SSEs). Therefore, this study compares Senegalese SSEs' responses to those of large firms and microentexprises, using data gathered from a 1990 survey, in order to assess the differential impact of adjustment policies. In a developing context, and in Africa in partcular, size plays an imnportant role in industrial characteristics and pspects. Most large scale African fimns descended from colonial enterpises or were created by the state, often with foreign financial backing. These finns have generally benefited from high protection, monopoly power, and direct allocation of inputs and credit-to the exclusion of SSEs. Extensive regulation of the economy also inhibited the development of formal SSEs by imposing high costs on firms large enough to attract the notice of regulatoty and tax authorities.1 Other factors associated with size, such as economies of scale, access to technology, and skill requirements, generate differences in operating styles, capabilities and constraints. As a result, post-reform responses to new incentives and regulatriy chainges may be expected to vary with firn size. Senegal presents an interestng case forexamining finns' responses under adjustment since it has an old and relatvely well-developed industral sector deriving from its position as supplier to the colonial French West Afrcan market. The prmary objectves of this study are to validate the findings in other countries that show a relatvely high degree of entrepreneurial dynamism in the small scale sector after adjustment and to determine the constraints which inhibit firms' ope. tions, investment and growth. This investigation will be an important step in developing a strategy for Senegalese industrial assistance, focusing efforts on those firms and subsectors whose adj; .unent responses appear most promising. Section II sets out the analytical framnv'work. Sections m and TV provide the background on Senegal's economy and reform package, in the process setting out hypotheses to be tested by the empirical data. Sections V and VI provide the results of a survey of industrial and commercial firms in Dakar and surrounding regions to examine the adjustment responses of the firms and the, factors that may have influenced them. I lhe point at which these costs become binding varies acoss counties, r tion subsecrss, and indvidua fims. Since microenterprises are geneally considered to be "informal" (i.e, opting outside the official sphere) and medium-scale enteprises "fomnal". SSEs can be considered as a transiional category. Definitions of categones overlap in different studies: micoeterpnse defrnitions range from I up to 9 employees and th lower limit for medium or large sca;e may be anywhere from 30 to 100. For analytical convenince, SSEs ae defimed in ftis paper as having from 10 to 49 workes  ANALYTICAL FRAMEWORK Empirical evidence from other countries justifies a focus on the modem small scale sector in Africa. These studies suggest that small enterprise development addresses a variety of important objectives in a country preparing for rapid industrial growth. In most African countries examined, SSEs have been the most responsive to opportunities created by structural adjustment reform packages. Dualism and the Dynamic Role of SSEs In the theoretical literature, development is often presented as a process of overcoming dualism.2 The "missing middle" is a tern sometimes used to describe dualism with respect to the distibution of firm size.3 Compared to industrializing countries in Asia, for example, African countries are observed to have a disproportionately low number of firms in the intermediate (i.e., SSE) range between self-employment (or microenterprises) and larte industrial plants.4 Filling this gap can be argued to be a normal, if not necessary part of the industrialization process, for several reasons. 'E ficiency Empirical evidence indicates that industrialization initially -nvolves a rapid shift from household and artisanal production into SSEs. Although subsequent growth comes more from large firms (some of which originated as SSEs), a substantial share of SSES may remain competitive because they serve dispersed local markets, provide differentiated products with low scale economies for niche markets, or specialize narrowly as subcontractors for larger fms. (Anderson, 1982; Anheier and Seibel, 1987; Cones et al, 1987; Liedholm and Mead, 1987; Little, et al, 1987; Nanjundan, 1987; Staley and Morse, 1965.) Employment Objectives SSEs offer the possibility of achieving both growth and equity objectives because they are relatively labor-intensive (compared to large industries) and they involve the application of modern technology (unlike most self-employment), albeit at a low level. Thus, expansion of the SSE sector is consistent with employment and income distribution objectives while allowing for sustained productivity increases through improvements in technology. (Staley and Morse, 1965; Steel and Takagi, 1983.) Development of Capabilities Perhaps the most important dynamic role for SSEs is the gradual building up of the entrepreneurial and managerial base needed for efficient investment in and management of large- scale industries over the longer term. Developing human capital is especially important if 2 See Gerald Meier, Leading Issues in Economic Development (New York: Oxford University Press, 1984), pp.149-175. 3 Peter Kilby, "Breaking the Entrepreneurial Botleneck in Late Developing Countries: Is there a Useful Role for Government?" Journal ofDevelopment Planning, (New York: United Nations), pp.221-249. 4 Hettige, H., W.F. Steel and J.A. Wayem, The Impact of Adjustment Lending on Industry in African Countries, (Washington, D.C.: World Bank, Industry and Energy Department Working Paper, Industry Series No. 33. June 1991). 2  industrialization is to be indigenous and sustainable. African countries have been notably unsuccessful in their attempts to by-pass this process through foreign and state involvement in large industries. (Bolton, 1971; Bruton, 1990; Kilby, 1988.) Overcoming Financial Market Imperfections In the 1970s, recognition of the potential role of SSEs led the World Bank and other development agencies to create special programs of assistance targeted towards small and medium enterpnses (World Bank, 1978). Lack of access to credit was idendtfied as a major constraint on SSE growth. This problem was addressed largely through lines of credit for SSEs and assistance to development finance institutions in setting up SSE units or creating specialized lending institutions. Results from several countries indicate that SSEs have the potential to be dynamic agents in a developing country context, though controversy exists with regard to how extensive the impact of SSE growth can be. A study of Bank lending to small and medium industries (SMIs) in five non-African countries found generally positive results: ...lending projects have been successful in promoting SMI expansion and employment creation at good economic and financial rates of return.... The considerations that prompted the Bank to move into this area in the 1970s-general evidence that SMI could be relatively efficient while also creating substantial employment-receive some confirmation from the estimated... economic rates of return on subprojects. (World Bank, 1991b, p.78.) Caution must be exercised against an overly optimistic assessment of SSEs, however, since success varies across countries. Financial services play a key role in small enterprise credit. A study of 33 World Bank SME projects for which post-completion assessments were available revealed that a well-developed financial system is critical to the sustainability of the lending programs. This finding has particularly strong implications for Africa, where financial infrastructure is generally weak. Africa had the lowest repayment rate (61 percent) of any region, and credit channeled through commercial banks and development finance institutions typically did not reach the small scale sector in Africa, rather being focused on relatively large subloans ($83,408, on average) at the upper end of the eligible size range. No African SME lending program was judged "sustainable", using a 90 percent repayment rate criterion (Webster, 1991, pp. 10-11). Future efforts must give particular attention to overcoming the financial market distortions that have inhibited SSEs' access to credit under previous schemes. Structural Adjustment to Overcome Policy Distortions Several arguments can be advanced that the difficulty in expanding SSE lending in Africa is in part attrbutable to policies that work against SSEs and tend to maintain a highly dualistic fum size distribution. First, in many African countries, including Senegal, the state has sponsored the development of large industries under policies designed to inhibit the emergence of direct competition. Second, most countries have evolved extensive state controls over the allocation of resources, particularly foreign exchange and credit, whose scarcity was often aggravated by maintaining fixed or subsidized prices. Access to those resources-and the scarcity rents associated with them-has generally been limited to large enterprises and those with political connections. To exist, smaller fims either had to depend on large firns for access to inputs or pay a premium in parallel markets. 3  Third, overegulation of the private sector has inhibited the normal process by which firms gain in productivity and economies of scale as they grow. Extensive controls and regulation often reflect mistrust of the private sector, which can stem from socialist suspicion of private profit, concern about domination by commercially powerful non-indigenous minorities, or fear of political opposition from a growing middle class. These regulations impose high costs on SSEs and induce many entrepreneurs to remain informal to avoid them, thus creating a banier to "g-aduation." Fourth, many governments have maintained a hostile attitude toward infornal economic activities. In some countries, particularty in East Africa, legal bariers originated in colonial efforts to limit participation by Africans in the modern economy. State actions against "illegal" informal enterpnises are a not uncommon manifestation of the desire of governments to modernize or to force people to register and pay taxes. These manifestations add to the riskiness of small business operations. Under conditions such as these, an adjustment program that includes reduction of the role of the state, liberalization of controls, and increasing reliance on the market and the private sector would be expected to facilitate expansion of the SSE sector. Assuming that some profitable investments were repressed under controls and regulations, expansion of the SSE sector should be forthcoming even without special assistance measures-which may, nevertheless, be appropriate to accelerate the process. Experiences of African SSEs under Adjustment In Africa, evidence is sketchy on the performance of SSEs, especially in countries undergoing structural adjustment programs. The studies available provide indications of a relatively high degree of dynamism in the sector, but performance varies by country. In addidon, several studies point out that SSEs are heterogeneous and their prospects vary widely from firm to firn and among different subsectors. On the whole, however, studies found a high proportion of dynamic firms among SSEs.5 In finn-level surveys in four African countries (Ghana, Malawi, Tanzania, Mali), SSEs exhibited dynamic responses under adjustment, as measured by a relatively high percentage of firms increasing output, purchasing new equipment, increasing share of imported inputs, and/or changing product mix after policy reforms. The experience in Ghana after the implementation of the Economnic Recovery Program is typical: ...there is evidence of considerable entrepreneurial initiative [among SSEs] in changing product mix and seeking market niches that have opened up under the new exchange rate regime. Some SSEs have adapted to changing demand by producing specialty products, custom-made items, or low-cost substitutes for imported goods.6 Additional adaptive behaviors found among SSEs in other studies include subcontracting anrangements, absorption of labor and capital from declining large firms, increased employment, and a high rate of new firm start-ups. Credlt for expansion or upgrading was a frequently 5 See Dawson, 1991 C(anzania and Ghana), p. 1; Bagachwa, 1991 (Tanzania), p. 28; and Mumbengegwi, (Zimbabwe) 1991, p.25 for variation in impact among different firms. The following section refers to the above thee studies as well as Steel and Webster, 1991 (Ghana); Lessard, 1991 (Mali); Ikiara, 1991 (Kenya); and Frischtak, 1990 (Malawi). 6 Steel, W.F. and L Webster, 1991. Swall Enterprises under Adjustment in Ghana, (Washington, D.C.: World Bank, Industry and Finance Series, World Bank Technical Paper Number 138.), p.ix. 4  mentioned constraint for SSEs, as were ac.ess to raw materials, business environment, competition, and low demand. In contrast to SSEs, microenterprises have found themselves ill-equipped to respond to the new policy environment. In particular, individual microenterprises face weak demand for their products because of low barriers to entry in the category, increased import competition, low income per capita, and policies that restrain aggregate demand. In several studies, microenterprises rated high on characteristics which were typically ;enalized under a structural adjustment program: few linkages, demand mosdy from the poor, little technological enhancement, low barriers to entry and intense comnpetition.7 On only one measure-use of local inputs-did microenterprises rate high on factors favored by structural adjustment. In contrast, modern SSEs rated high relative to microenterprisos on all characteristics favored by policy reform except the latter. The historical context may have implications for post-reformn performance. For example, Ghanaian SSEs did much better than T4nzanian firms, which were particularly underdeveloped due to a hostile state attitude before the reform process. In Zimbabwe, the small-scale sector was limited to transport and retail services, the legacy of colonial policies restricting indigenous manufacturing, black urbanization, and residential integration. As a result, the study was not overly optimistic about Zimbabwean SSE prospects in an environment of major reform These studies indicate that SSEs' responses under adjustment depend on several factors. In pardcular, the reforrn package must be successful in easing regulations that have hampered SSEs in the past. Even if liberalization is successfully achieved, a positive response may not materialize quickly if past restrictions have debilitated the SSE sector. In addition, the effects of increased domestic and import competition are important; variables such as access to technology, ability to diversify, and flexible use of inputs may determine whether a firm succumbs or rises to the challenge. Finally, access to credit and inputs needed to expand or upgrade is critical. 7 See Dawson, 1991 and Bagachwa, 1991. 5  THE SENEGALESE INDUSTRIAL CONTEXT Like most African countries. Senegal's colonial economy was based on supplying raw materials and consuming imported finished goods. Though it is now among the most urbanized of African countries, with 34 percent living in towns or cites, the rhythm of Senegal's economy still moves with the rual areas. Its major cash crop remains groundnuts, whose fortunes are affected by climatic conditions. A poor crop leads to a decrease in purchasing power for the 70 percent of the population engaged in agriculture and has repercussions throughout the economy. Extended droughits in the l-980s, exacerbated by other extemal and internal factors, began a period of economic crisis from which the country has yet to emergr,. Overall GDP growtk ;' Senegal since independence has been poor. In fact, Senegal lags behind all other forner Fret .olonies. With population increases approaching 3 percent per year, Senegal's meager 2.3 percwit annual GDP growth since independence has meant a decline in real living standards. Indu,n"rial History and Performance After the Second World War, Senegalese industries expanded rapidly, driven by their role as a source of provisions for the colonial French West African market of about 20 million people. Between 1950 and 1960, Senegal's industries were ahead of those of most other African countnes, featuring diversified production and exports of groundnut derivatives. Most industries, however, were foreign-owned and the business environment inhibited the activities of local private entrepreneurs and investors. The newly independent state in 1960 inherited an extensive industrial and commercial infrastructure, with the predominant activities being fish conserves, groundnut oil refineries and other food industies, textles, and water and elec'ic supply. In addition, Senegal's membership in UMOA (Union Monetaire Ouest Africaine) provided the advantages of a convertible cuirency and, as a result, access to imported inputs, equipm.nt and manufactures. Strategy Following independence, however, the West African market block was broken up as France's former colonies developed their own industries to satisfy intemal needs. Senegal found itself with a larger industrial sector than warranted by its small internal market, one-tenth the size of the West African market. In addition, the established industrial base with large capacity fostered monopolies and inhibited new entrants. As a result, the average age of Senegalese manufacturers is 24 years old, twice the corresponding figures for Cameroon and Cote d'Ivoire. Senegal responded to this situation by protecting its old companies, using high tariff and quantitative barriers and favoring import substitutes. These policies led to the development of an non competitive industrial sector: almost all subsectors had only one or two major companies. I addition, the Government tried to attract foreign investors and promote export-oriented large-scale projects with ta; breaks, subsidies, price controls, and other incentives. These efforts resulttd in omplicated and inefficient procedures and a difficult environment for manufacturers and investors. External factors, including recurrent droughts, stagnant internal demand and the oil crisis, further limited growth. Senegal's attempts to promote exports failed primarily as a result of two factors. First, high trade protection, with nominal tariffs averaging 25 to 95 percent in 1985, and quotas on over 150 products prompted entrepreneurs to concentrate on more profitable import substitution. Second, Senegalese factor cost were non competitve due to wage, prcing and exchange rate policies. 6  The Government's response to low growth rates, particularly in the mid-1970s and early 1980s, was interventionist policies. The number of parapublic industrial enterprises tripled from 9 to 27 from 1972 to 1986.8 The problems associated with the interventionism were not immediately apparent, however, due to the relatively strong performance in the midst of a groundnut and phosphate boom in the mid-1970s. After 1977, however, the performance of public enterprises deteriorated drastically and severely drained public resources. A widely-accepted diagnosis noted that heavy institutional and management controls stifled parapublic operations, that these firms lacked managers with sufficient experience and technical skills, and that the sector often served a social function rather than aiming at profit maximization. The NPI sought to :3verse the effects of previous policies and create an environment more favorable to the private sectu: Growth Senegal's industrial production increased by 4.5 percent per annum between 1960 and 1985, much slower than in Cote d'Ivoire or Cameroon.9 This expansion barely met the needs of a population growing at 2.5 percent yearly over the same period. Because other sectors of the economy grew more slowly, however, industry's share of GDP rose from 17 percent in 1960 to 29 percent by 1985. Manufacturing comprised nearly two-thirds of the industry total and 19 percent of GDP in 1985, up from 12 percent in 1960. Most manufacturing growth is attributable to import substitution, though there were modest gains in export development in the late 1960s. The services sector continued to dominate the economy at 52 percent of GDP in 1985, down from 59 percent in 1960. Aggregate rates conceal high and low growth periods of industrial production as well as substantial variation among the sector's components. Whereas manufacturing drove industrial growth in the 1960s, it began to lag behind after 1970. Manufacturing grew by 6.1 percent per year between 1960 and 1970 while the category "other industries" (mining, construction and utilities) hardly grew at all. Between 1970 and 1982, however, manufactures grew at only 3.3 percent per year while other industries increased by 11.7 percent yearly. From 1982-85 both grew only slightly, at rates of 0.3 percent and 1.3 percent per year, respectively. Employment Employment in formal manufacturing more than doubled from 1-960 to 1985, an increase of over 3 percent a year. Almost all growth, however, took place from 1960 to 1975. Large companies generally utilized capital-intensive technologies, in part because Senegalese labor laws made it expensive for modern sector firms to hire workers and difficult to fire them. As a result, manufacturing output growth did little to ease the excess labor situation. Estimates in the 1980s placed urban unemployment at nearly 20 percent.10 Despite a gradual shift in employment from agriculture to modern manufacturing after independence, industry's share remained small. By ttic mid-1980s, 75 percent of the labor force continued to work in agriculture, down from 83 percent in 1965, while industrial employment contributed 10 percent, up from 6 percent Increased industrial sector employment after the early 8 World Bank, 1989a, p.64 and World Bank, 1989b, p.6. 9 Industrial statistics reported in this section include manufacturing, construction, mining and utilities. In discussing the survey results, the term "industry" is used principally to refer to manufacturing. 10 Estimates are from Farrell, et al., 1989; Terrell and Svejnar, 1989. 7  1970s came mainly from the urban informal and small scale sectors, as modern manufacturing employed less than 2 percent of the labor force as of the mid- 1980s. State participation After independence, the Govemment sought ownership of productive activities fur several reasons: (i) to control important national resources (groundnuts, phosphates), infrastructure and essential services; (ii) to promote development where private initiative was judged to be weak (tourism and cotton); and (iii) to increase management and employment opportunities for Senegalese. State participation started slowly in the 1960s and early 1970s, but accelerated after the mid- 1970s as high phosphate and groundnut export revenues and international borrowing fueled state acquisition of majority stakes in previously foreign-owned companies.l1 According to a USAID-profile, industrial production in 1987 was 87 percent private and 13 percent parastatal, representing 24.5 and 3.5 percent of GDP, respectively.12 State participation is estimated at about 30 percent of manufacturing value added. 3 The state's interests in the four 'argest groundnut oil processing plants and in two phosphate mining companies account for two-thirds of its holdings. Given the poor performance of public sector enterprises, the Government sought to sell off its shares in industry, with the exception of "strategic" activities (groundnut oils and phosphates), as part of its structural adjustment program. Subsector composition The structure of the industrial sector has not changed substantially since independence, still being dominated by import substitutes and the processing of local raw materials for export. Few intermediate production plants for goods downstream or upstream of major industries have been established. Agro-industries dominate the formal industrial sector, accounting for some 40 percent of manufacturing value added, followed by chemicals, textiles and metals and mechanical industries. Manufacturing production is heavily concentrated, with the largest 140 enterprises accounting for 95 percent of measured manufacturing production and the largest 40, 80 percent. Informal industrial activities are concentrated in artisanal products and construction. Most manufacturing involves consumer goods rather than intermediate or heavy industries. The indigenous private sector is skewed towards commercial rather than manufacturing activities. Profile of the SSE Sector At independence, nearly all non-artisanal private companies were foreign-owned or controlled. Most of these large firms continued under foreign domination, though the owners generally added little fresh investment into the Senegalese economy. Initially, new activity came from the Lebanese, who developed small-scale companies, taking advantage of the absence of a Senegalese small-scale private sector. In the 1970s, Senegalese businessmen began to create 11 State involvement also included minority shareholding in many enterprises which, unlike those under majority contrl, were not necessarily subject to direct govemment oversight. 12 See USAID, 1990a, p.8. 13 This figure is the most widely-used and includes mining. (World Bank, 1989a and IFC, 1985, annex 2.) Please note that large discrepancies surfaced, which could not be resolved. World Bank 1989b, p.6, for example, shows state paticipation in manufacturing and mining at only 6 percent of GDP. 8  modem small- and medium-scale industrial firms, although usually not in direct competition with the large scale sector. Senegal's small-scale sector is difficult to measure for lack of comprehensive statistics and consistent definitions among various estimates, which have placed the number of registered SSEs at anywhere between 250 and 1000.14 The most comprehensive Government censuses of small and medium industries counted 200 units in the countty in 1974 and 479 in 1984 in the Cap Vert region alone, suggesting a substantial increase. The more recent figure represented 90 percent of all industrial establishments counted in the region, 59 percent of them Senegalese-owned. The principal recorded industries were mining, food and fishing industries, textiles and clothing, metal working and machinery, paper and printing, wood working and furniture, and construction.15 Given that :here was little indigenous production outside the artisanal sector (mostly informal) before independence and that many SSEs created since are registered (somne firms with as few as 10 employees are listed in official industrial statistics), these figures probably give a credible idea of the sector's pre-NPI size. It was small compared to an estimated 50,000 microenterpdises (Van Dijlc, 1986, p.30) and contributed at most a modest percentage to industrial sales, since the largest 140 manufacturers accounted for 94 percent of modem sector turnover. The market dominance of large scale companies left litde room for modem, small-scale entities. The small-scale sector is composed primarily of Lebanese and Senegalese entrepreneurs. Most French fims are in the medium/large categories, though the French assisted some Senegalese in creating SSEs under a mid-1970s program of "industrial cooperation" for SSE promotion. Lebanese and Senegalese businessmen to some extent face substantially different prospects, because of uneven access to resources and levels of skills and experience. Lebanese SSEs largely arose to fill the early post-colonial vacuum in the sector. Generally, their operations focused on personal control, flexibility and short term profitability, conditions which discouraged entry into large-scale enterprises. They tended to favor commercial activities, but also took advantage of opportunities that arose in small-scale manufacturing such as "paints, building, hardwares, plastic products, readymade garments, towels, bakery products, packing materials and fabricated items like doors, windows, grills, handicraft items and furniture for local consumption."16 A close relationship with Senegal, including the same religion, similar conceptions of the role of the family, and a commitmnent to the country (most are more or less permanent residents) allowed Lebanese firms to integrate themselves into its economic fabric. The Lebanese work closely with the indigenous infonnal sector, which they use as sales outlets and low-cost suppliers. This relationship has sometimes brought on charges of exploitation.17 Senegalese SSEs, particularly those involved in manufacturing, gained their impetus from Government efforts in the late 1960s to promote indigenous entrepreneurship. Their activities initially concentrated in real estate and tourism, but soon broadened to incorporate a variety of 14 See EDI-1, 1986, p.18 for the upper limit and Van Dijk, 1986, p.30 for the lower. Another estimate counted 600 modem industrial enterprises of which 120 were 'modemn, factory scale entities,' suggesting around 480 small scale enterprises (Berg, 1990. p.1 11). 15 Zarour. Charbel. "Constraints on the Senegalese Banking System that Limit Credit to Small and Medium Enterprises" (Report for USAID, not dated). 16 See EDI-I,. 1986, p. 18. 17 See OECD, 1988. 9  productive operations. More recently, the Mouride religious sect has played a role in providing a pool of savings for SSE start-ups, especially in commerce and services, but also in agro-industnes and fishing."' Generaly, these companies do not export due to limited finance, poor marketing capabilities and lack of an international quality product. Support for SSEs Several agencies created in the 1970s have promoted the development of indigenous SSEs, though results have been mixed due to the persistence of monopoly rights for many large firms and a cumbersome policy environment which inhibited firms' flexibility. The Government aided SSEs primarily through three institutions: SONEPI (Socidt6 Nationale d'Etudes et de Promotion Industrielle), established in 1969 to promote SSEs; SONAGA (Soci&t6 Nationale de Garantie et d'Assistance au Commerce) founded in 1971 to assist SSEs with credit guarantees; and SOFISEDIT (Societd FinanciRre Sdndgalaise pour le Developpement de l'Industrie et du Tourisme), founded in 1974 to provide SME finance, but skewed towards medium-scale fims. In addition, SSEs benefited from some fiscal incentives, such as exoneration of duties and taxes during start-up. The Government set up a dozen funds managed by SONEPI, SONAGA and SOFISEDIT to assist SMEs in obtaining credit by guaranteeing, subsiding or supporting loans. Though subsidized loans have generally failed elsewhere in Africa, a follow-up study found that repayment y recent university graduates receiving concessionary loans under a SONAGA scheme was satisfactory as of 1985. On the credit side, however, banks were generally reluctant to provide funds to SSEs, even with Govemment backing, due to a lack of confidence in the entrepreneurs' managerial experience.19 A number of projects were created or assisted under the various schemes. Between 1969 and 1986, SONEPI conducted 445 feasibility studies which resulted in 142 projects. In the "domaine industriel"20 created by SONEPI, 20 production units operated in 24 sheds in 1986; SONAGA assisted 150 graduates between 1971 and 1986 in a variety of activities and project performance were judged "satisfactory" by one analysis, though the criteria were not clear. SOFISEDIT, however, was virtually non-functional by 1983 due to a shortage of funds.21 These schemes, though supporting a few modest efforts, did not have a wide impact. Tley generally sought to aid SSEs within the same highly regulated environment that has inhibited the large scale sector, often with similar consequences. Procedures were cumbersome, with months or even years passing between project conception and implementation. Entrepreneur selection was often not careful enough. For example, SONEPI, a promotional agency, went out in search of entrepreneurs rather than allowing the process to be demand-driven. As a result, funds often went to inexperienced or incapable recipients. Factors external to the firm also limited project perfonnance and widespread use of the facilities, including small markets, competition from imports, the non-availability of or delays in 18 See Bano, 1990, pp.40-1; and OECD, 1988, pp.43-4. Savings generated through informal systems known as tonines! are aely used for productive works. 19 EDI-I, 1986. 20 Industrial sites where SSEs afe concentrated and provided special exoraions, privileges, and assiance. 21 This judgment was made in EDI-1. 1986, pp.14-5. 10  obtaining finance and raw materials, and procedural bottlenecks. The economic crisis of the early 1980s starved the credit funds of cash. In addition, competition from the informal sector was strong since it did not incur the extra costs of attaining formal status. A revealing attitude was that "imany owners of small enterprises [viewed] SONEPI as a tax office and therefore (feared that al visit to such an organization may invite tax problems at a later stage."22 At 12 percent of GDP in 1990, Senegal's aid inflow is high, with some 30 public sector donors and over 100 NGOs active in the country. Some of these contributions may affect small- scale manufacturers, for example, a USAID loan program to aid small companies producing agricultural goods, furniture, and metal products. The French have played a leading role in SME promotion by setting up agencies in the mid-1970s to accompany investments undertaken in the framework of their industrial cooperation policy. These included PROPARCO, a subsidiary of the Caisse Centrale de Cooperation Etconomique, which takes minority shareholdings in industrial projects and occupies a seat on the boards of directors of participating frmns. These efforts bore little fruit, and French investment has been on a downward trend since the mid- 1970s. Some studies have noted cultural barriers to the development of a modern small-scale sector, asserting that certain Islamic tenets, communal practices, and colonial inheritances do not jibe with a spirit of entrepreneurship.23 Charging interest is forbidden by the Koran and, as in other parts of Africa, Senegalese tend to view individualism as "a form of deviance." Citizens in good financial standing are expected to share their resources with less fortunate relatives and friends; business earnings and assets are not distinguished in this redistribution system. In addition, the French bureaucracy left an orientation towards creating good civil servants ("fonctionnaires") to implement national development plans, which may sometimes conflict with individual business interests. The Need for Industrial Adjustment In the mid-1980s, Senegal needed z program to revive industrial prospects in the context of structural adjustment efforts. The fresh approach would requir measures to stimulate the industrial sector by opening it to both extemal and internal competition, while at the same time easing supply side constraints and improving incentives and opportunities. The obstacles manufacturers' faced were substantial: a cumbersome regulatory frmnework; highly non compedtive markets; the lack of industrial and entrepreneurial traditions; demand constraints tied to a real effective exchange rate favoring imports; a small intemal market; the deflationary effects of structural adjustment; and high factor costs for industry, particularly labor and utility. At the same time, however, Senegal had many factors in its favor which were atypical of African countries: political stability; a relatively modern infrastructure; a fairly well-developed banking structure; efficient service enterprises for industry; a high-level commitment to a reduced government role; and a leading strategic and political position in West Africa.24 22 EDI-I, 1986, p. 13. 23 See Courcel 1988. Arund 85 pecent of Scnegal's population is Musli- 24 World Bank. 1989a, pp.68-9; IFC. 1985, pp.54; and Van Dij,k 1986, p.32. 11  THE NPI AND ITS EFFECTS The NPI represented an effort to stimulate industry's contribution to three objectives of the structural adjustment program: more optimal use of resources; an improved-environment for tanslating incentives into firm-level responses; and increased opportunities for investment. The main instruments for achieving these objectives were, respectively, opening the sector to competition, reducing input and labor constraints on firm flexibility, and strengthening agencies and programns designed to promote industry. The first objective was primarily addressed in the NPrs main body, whereas the last two were subsumed under "accompanying measures." Measures The specific policy changes were put in place with varying timetables for their implementation. The most important measures are listed below.25 bfts= IQ impve compddve environment: * Elimination of quantitative import restrictions over a three year period; * Reduction and standization of tariffs; * Elimination of exonerations and special tariffs for individual enterprises (unless covered by "conventions speciales");26 * Elimination of almost all "prix mercuriales" except those used to counteract dumping;27 * Review of old "conventions speciales" and freeze on the granting of new ones; * Decontrol of industrial prices except five products granted monopoly privileges under 'conventions speciales"; * Simplification of procedures for acquiring import/export licenses. Accompanying measures to facilitate responsiveness: * Study ways and means to reduce energy, gas and telecommunications costs; * Revision of the Labor Code to ease hiring and firing restrictions. 25 Berg. 1990, p. 20. Berg concluded that "most of the hard adjustment decisions that Senegal needs to make were put off in the 1980s," suggesting that a change in the exchange rate or a reduction in labor costs were key to the competitiveness of Senegalese enterprises. Moreover he spoke of a policy environment of low sanctions for poor performac, weak local initiative, and wateed-down recommendations. Bank staff have expressed concern that the negative tone of the report's conclusions was not consistent with its detailed analysis and that many of the reporWs cridcisms had already been identified and were being addressed in SAL IV. Furhm othed inhibiting factors were outside the control of the structumal adjustment package. 26 Conventions speciales are special privileges or incentive packages granted by the Government to large fums in stategic sector In 1990, 15 enterprises were still operating under such agreements. 27 Prix meruria" are reference prices for setting tariffs which are used to combat underdeclaration of the value of import 12  Accompanying measres to increase investment oortunides: * Revision of the country's Investment Code, including the establishment of a one-stop investment office;28 * Establishm,ent of a restructuring finance faciFty financed by the World Bank; * Strengthening of export credit and insurance agencies; * Increase in the export subsidy and simplification of procedures, including change in the formula for calculating subsidies (based on local value added rather than local content). The NPI focused on giving private entrepreneurs an appropriate environment to create or expand operations. In particular, it was hoped that recent college graduates, laid off public sector workers, existing entrepreneurs with young firms, and others with education or financial resources would seek opportunities to invest in relatively modern SSEs. This development would be important to offset the decline of some large-scale firms as a result of decreased protection and state divestiture. Implementation Before analyzing the NPrs impact, it is necessary first to evaluate the extent to which it was carried out and could, therefore, be expected to produce the intended responses. Implementation issues center around the Government's failure to carry out many of the accompanying measures, the extent of import liberalization in view of the partial relapse on tariff rates and customs procedures, and the sequencing and speed of the adjustment process. Although studies differ as to the NPI's effectiveness, they generally indicate that reforms increased competition from imports for most industrial subsectors, particularly in the early stages before backsliding on tariff reforms. Import competition increases came pnmarily as a result of the abolition of quota restrictions and the liberalization of import licenses. Fimns received little support in the adjustment process, however; neither export incentives nor input costs were substantially improved, and as a result local fms could not reduce prices to enable them to better compete in world markets. Competitive environment It is clear that a large degree of import liberalization occured. Quantitative restrictions were lifted on schedule for many products, though some remained for important inputs such as sugar, cement and flour, forcing manufacturers to pay well above world prices. The number of fiums and individuals holding import-export licenses from the Ministry of Commerce rose from 2,669 befoe the reform to 9,391 by September 1989, a nearly fourfold increase. The result was a 46 percent increase in merchandise imports from 1986 to 1990 (in U.S. dollars). Domestic compettion was improved by successful decontrol of industrial prices and libralizadon of distibuton channes for some subsectors. In addition, future "convention speciales" were prohibited, although existing arrangements persisted and continued to distort the economic environment. The net result was that industrial competition increased substantially but unevenly primarily as a result of low priced goods coming from countries which had experienced major depreciations. This change should have had the desired effect of weeding out some non viable firms, depending upon the activit, and pronpting other inefficient firms to restructure to increase productivity. 28 A one-stop finance window, or "guichet unique" is a single office empowered to secure al govermat appvals needed under the investment code. 13  Although tariff reforms were implemented on schedule, they were partially reversed by the Government in 1989-ostensibly to counteract fraud but also to raise revenues in the midst of a fiscal crisis. This reversal, which was resisted by the World Bank, primarily consisted of a five percent increase in the base customs duty, applicable to almost all goods.Y9 In addition, other instruments were reintroduced to counteract fraud and give special preferences, such as t'valeurs mecuriales", "minimum de perception", and "codes de precision", generally of a protecdonist character.30 Special tariffs were levied on textiles, matches and batteries. The net result of all these changes was to leave in place a relatively complex tariff structure of widely varying protection and to give large discretionary power to customs officials at the border. In some instances, these measures discouraged official imports, with precipitous declines in the recorded inflow of cigarettes (84 to 18 tons) and green tea Q41 to 83 tons). Nevertheless, customs receipts did increase from 74 billion CFA francs in 1988-89 to 92 in 1989-90, though it must be pointed out that they also increased during the NPrs early stages from 1985-86 to 1987-88.31 Firm Response and Incentives Measures aimed at improving incentives and the environment for firms to respond were not fully carried out. Input cost reductions and labor reforms were studied but implemented incompletely or not at all. Electricity costs were reduced by 8-18% in July 1986, but this fall was subsequently moderated by increases in subscription fees. Water costs decreased slightly while fuel and telecommunications actually increased as a percentage of production costs. Salaries in the modern sector rose, and the ratio of salary to output went from 8 percent in 1985 to 10 percent in 1988. A new labor law eliminated the requirement that employers hire only through the official Labor Exchange Office, but measures to allow unlimited recourse to short-term contracts were defeated by union pressure. Despite further progress on labor legislation in late 1989, substantial restrictions remained. The policy reforms have not succeeded in improving the position of Senegal's manufacturers on export markets, especially considering the large real depreciations achieved by other developing countries such as Ghana and Nigeria. Senegal's membership in the West African Monetary Union prevents it from using the exchange rate to modify the relative price of tradables versus non-tradables. Despite Senegal's relatively low inflation rate, the real effective exchange rate appreciated from 1985 to 1990 by 20 percent (see Annex 1 for details). This figure contrasts with a nearly six-fold depreciation of the Nigerian naira and smaller but important depreciations in many other Sub-Saharan African countries and low-income countries as a whole. There was little substantial alleviation of factor prices, and Senegal's international competitiveness deterorated as lower-priced goods began to enter its markets and compete with locally-produced goods.32 The refonns aimed at supporting exports, such as the establishment of a duty drawback system and 29 The ending of quota restrictions may have facilitated the dumping of surplus goods on Senegals markets. In addition, fraud in the form of underinvoicing and underbilling may have increased as cost-cuuing measures in an increasingly competitive import trade. No longer protected, pre-existing importers complained of "unfair' competition. 30 A "minimum de perception" is a minimum specific tax collected per unit, supposedly to eliminate fraud achieved through underinvoicing, but often set well above the ad valorem tariff rate applied to world prices. "Codes de precision" are customs classification codes which allow officials to reclassify products within the same category for greater precision in protecting local products. 31 These figures are from a fortcoming AF51E report on the NPI. 32 The figures from this and the preceding paragraph come from a forhcoming AF51E report on the NPI. 14  changes in export subsidies coverage and calculation, although implemented, were not sufficient to overcome the improved position of Senegal's international competitors.33 The Govemment of Senegal implemented investment code reforms and established a one- stop investment window to improve the regulatory constraints faced by frims. At the same dme, however, in practice, entrenched bureaucratic habits continued to slow down investmnent and other industrial sector activities. In addition, special conventions persisted for 12 firms, creating roadblocks for others. The 1987 investment code did make the legal framework more transparent and automatic and eliminated preferential incentives to specified types of business. The number of projects approved by the facility nearly doubled from 114 to 219 from 1988 to 1990. The industrial restructuring finance facility established in 1988 was slow in starting up (no loans had been concluded by mid-July 1989) since at the time the private formal banking sector was in disarray. In addition, one analysis noted that the strict and sophisticated financial standards set by the facility would effectively exclude the majority of firns from consideration, including, undoubtedly, smaller enterprises.34 Expected Impact On balance, implementation problems resulted in only the liberalization of import and intemnal competition being effectively carried out. Apart from consumers, only new entrants into import trade, firms using imported inputs for which restrictions had been lifted, and firns able to take advantage of niche markets created by the withdrawal of inefficient firms stood to gain. Particularly susceptible to the negative effects would be large-scale companies producing import substitutes whose survival would depend on whether they could quickly become intemationally Dompetitive. Subsectoral analyses predicted that textile companies would rrobably be the hardest hit by foreign and informal sector competition. Food industry results would be mixed, with fish canning and flour milling activities predicted to survive but biscuit and processed milk products manufacturers being threatened by subsidized European goods. Mechanical engineering and metal conversion companies were also expected to experience diverse impacts, with activities requiring substantial scale economies or capital intensity being hardest hit. Many chemical companies were in a position to deliver positive results (Courcel, 1988; Boston Consulting Group). It must be kept in mind that firm level performance may differ from that of the sector. For example, while the commercial sector would likely expand after import liberalization, individual trading firns which previously exercised monopoly power would be hurt by increased competition. Similarly, the microenterprise sector would probably grow due to the abundance of unemployed people seeking activities with easy entry, but individual microenterprises would find it hard to grow because new entrants would increase competition. The NPrs impact would be substantially muted, however, by the failure to implement a strong adjustment program, particularly on the export side. In fact, Senegal's actual and potential exporters found themselves at an increasing disadvantage compared with the rest of the world, which was maldng inroads into their own markets. At the samne time, this competition should have some beneficial long-term aspects in exposing Senegalese frmns to the realities of the international marketplace and thus encouraging adjustments to improve efficiency and production standards. Nevertheless, the lack of effective input cost reduction measures, particularly for labor and udlides, and the suspension of export subsidies meant that firms would have little support in improving internaional competitiveness. 33 Expon subsidies were suspended in 1991 as a result of budget difficules. 34 Berg. 1990. p.134. 15  Impact on manufacturing 35 Estimates of the growth of outpt,t in large-scale firns since 1985 vary, but generally suggest weak performance in recent years. Early data indicated that aggregate industrial anti manufacturing production each increased by a total of 14 percent from 1985 to 1989; a mor recer.t esdmate, however, revised the manufacturing share of GDP for 1988 to 1990 downward by about 6 percentage points versus the mid-1980 peak levels, implying a 33 percent decline in manufacturing output if other components of GDP have remained constant and if the higher share figures for earlier years were accurate.36 Though the picture remains incomplete, some companies clearly have suffered. Annex 2 lists 53 firms that closed from 1986 to 1989. Output of key industrial products was down 34 percent from 1985 to 1989, with particularly large declines in shoes, tobacco products, galvanized iron sheets, cotton thread, cloth, metallic cans and car assembly (see Annex 3). Most of the decline (22 percent) took place from 1988 to 1989. Some analyses have suggested that production has shifted towards the medium, small and informal sectors, but this trend is hard to measure. Both import and domestic competition have played a role in company failures. CAFAL, which manufactures matches from local wood, was challenged by Swedish and Chinese matches that flooded the market after the lifting of quotas. Despite a 40 percent ad valorem duty and investment in new technology, CAFAL was unable to adjust, in part because it could not lay off excess workers. The company is currendy in danger of bankruptcy. Local producers arose to challenge the large cotton textile company, ICOTAF, whose monopoly on the import of printing ink had made it the only Senegalese manufacturer of African prints. After the NPI, ICOTAFs monopoly was eliminated and several small manufacturers entered. This development is expected to increase productivity in the subsector and expand the production base in the country. Import liberalization occasionally had a perverse effect. Some large French firmns with subsidiaries in Senegal found that they could close down unprofitable local plants and take advantage of the new openness to imports. Using Paris-based sales teams, these firms cut costs substantially and still sold efficiently in Senegal's small market. In addition, some large French firms are gradually abandoning their manufacturing or assembly activities to concentrate on trading. An analysis of labor and capital movements from six large companies that closed down after the NPI's implementation found that much of the assets of four of them were absorbed by modern sector enterprises. These were typically new entities in the same subsector as the old, but restructured or streamlined to adapt to new market conditions. The tracking of displaced workers discovered that 29 percent had found new positions, almost equally split between the irformal and modem sectors. Many modem sector rehires were in the new companies spawned from the old. About 40 percent of informal sector rehires were working in productive activities, though an 3S According to a fonthcoming country economic update, Bank staff have serious concerns about the accuracy of aggregate figures, particularly recent data, most of which is based on projections. In many insrances the past data itself may have been an estimate not based on actual survey data, or based on infomation covering only major activities. Hence, they may be more reliable as indicators of trends than as absolute values. These deficiencies highlight the need for periodic sample surveys. 36 World Bank Country Briefs (1991) shows the manufacturing share of GDP at 12 to 13 percent in 1988 through 1990, below the mid-1980Ws peak of around 18 to 19 perc nt of GDP. Though there have been some increased data collection problems in recent years and the figures do not reflect gains that may have occurred among smaller enterprses, they probably do accisawely reflect a decline in the recorded, large-scale sector. 16  almost equal proportion (35 percent) entered petty commerce.37 Finally, the study noted three innovative projects started by displaced workers, in tanning shoe production and general mechanics. The entrepreneurs' prospects to develop from artist - production into moden SSEs were enhanced by their experience working in modem sector fi.ms in similar activides.38 The study also tracked displaced workers with plans for reinsertion in the labor market, who overwhelmingly (84 percent) envisioned projects in micro-scale retail trade and hoped for assistance from NGOs (46 percent) or the banking system (38 percent).39 Aggregate data on industrial production are not sufficiently comprehensive to reveal any shift in output from large-scale firms to smaller firms. Evidence of a movement towards smaller scale production is supported, however, by a post-NPI USAID study that places 30 percent of industrial sector GDP (8.5 percent of total GDP) in the informal sector versus 57 percent (16 percent of total GDP) in the fornal private sector (USAID, 1990). Industrial sector employment was found to be nearly 84 percent informal (286,000 workers, 12 percent of the labor force). Demand has shifted somewhat to lower-priced products due to decreases in purchasing power, particularly among the poor and middle income population. (Barro, 1990, p.26.) This development favors the informal sector. A 1988 USAID survey found that 19 percent of productive and service-oriented microenterprises had been established in the previous two years. An additional 26 percent had been established in the previous three to five 3 mars.A Hypotheses to be examined Though the NPI's negative effects, particularly for large firmns, are very noticeable, some analyses have stressed the positive aspects of the shifts and responses induced by its success in introducing competition to many Senegalese industrial markets.41 The decline of some large- scale firns was expected. The critical question for this study is whether some of the hoped-for positive responses occurred among SSEs, whose less complex technology and management structure make them likely to be more flexible in the short run than larger finns. By decreasing state involvement in the industrial sector and opening up the economy, structural adjustment programs directly harmed inefficient, large companies that previously were protected from competition and supported with subsidies and favorable legislation. The microenterprise sector, with low barriers to entry, was flooded by laid-off workers fromn large- scale firms, making it hard to raise productivity in individual enterprises, despite an expected demand shift towards less expensive goods as prices of large-scale manufactures were liberalized. 37 Since there is no precise definition, it is unclear whether the study's author equates 'informal with "microenteprise" (1-9 workers). Some of those working in productive infonnal sector activities may, in fact, be included in what is here defined as the small scal sectr (1049 wors). 38 Issa Bano, 1990. Impact de la Nouvelle Polidque lndusirielle sur le Secteur IrforielICoraute Socio-culewtc et Ajustement, (Consultanfs repont to the World Bank-unpubished). 39 Petty commerce became a prime field for reinsertion after ethnic clashes in 1989 led to the explsion of Mauritanians, who until then had dominated the subsector. See Ron Parker 1991, "The Senegal-Mavriti Conflict of 1989: a Fragile Equilibrium", The Journal of Modern African Studies, 29, 1, pp.155-71, for the economic, political and social effects of the conflict 40 Zarour (1989). lTese figures reflect two phenomena: entrepreneus acquiring existing establishments and apprentices going into business for themselves. 41 Some analysts (see Desnot, 1988; Bandt, 1989; Judet, 1989) deaied the suddenness of the reform process whereas Berg, 1990, concluded that, on the whole, reform was neessay and benefic for indury. 17  In additon, most microenterprises may not be "dynamic" in orientation, since their typically low- income owners have few resources and limited educational backgrounds, and use profits largely for subsistence. In the context of the NPI, however, it is possible that some former modern sector workers could launch micro-scale activities which would not have such a subsistence orientation. The preceding analysis suggests that, as in other African countries under adjustment, the primary hope for entrepreneurial dynamism in Senegal in the short term would come from SSEs, which might be able to probe market niches that opened up as previously protected, inefficient companies reduced output. Several hypotheses follow frorm the analysis of how adjustment and the NPI were carried out in Senegal: * Growth and new investment at the firm level would concentrate in SSEs, since micro enterprises would be constrained by new entry and large-scale firms would be harned by the loss of their previous protection and monopoly posidons. Firns producing import substitutes would be hardest hit since few positive measures have been implemented to improve their ability to compete. * Subsectoral effects would vary widely. * Export response would be weak. The commercial sector would benefit from import liberalization, but individual firms may face difficulties from increasing competition. - Finms that use imported inputs which were liberalized would benefit. 18  THE SURVEY AND SAMPLE CHARACTERISTICS In assessing the implications of Senegal's NPI, it is important to examine rtsponses at the level of the firm. In the liberalized environment, the ability to compete becomes the cmcial determinant of survivability. Initially, firns may respond by cutting costs or changing ptoduct lines in ways that lay the basis for future competitiveness but do not immediately raise outputL Aggregate production figures, therefore, may n ask -nicro-level adjustment. Furthermore, aggregate data for small and microenterprises are unavailable or unreliable at best. Tlis section is based on a survey intended to permit comparisons of responses across different size categories and other firm characteristics. The Sample The survey was designed to assess the response of Senegal's private sector following the NPI's introduction. It focused in particular on SSEs to investigate their potential relative to other sectors and to avoid duplication with a concurrent USAID study, which covered large-scale firms. The approach was to do in-depth furm-level analysis and to concentrate on a limited number of key subsectors as a complement to the USAID study's broader quantitative approach. In addition, the commercial sector was included to test the hypothesis that it may have benefited from the liberalized environment to the detriment of productive activities. The survey was confined to the Cap Vert region, which includes Dakar, where the NPrs effects were likely to have been most keenly felt. The on-site interviews were conducted during May and June 1990 by two local research groups, Centre Africain d'lgtudes Supdrieures en Gestion (CESAG) and a consulting firm headed by Charbel Zarour. For man'ifacturers, the intent of the survey was to select representative finms spread relatively evenly among a range of size groUDs in order to compare and contrast firns by size. The implicit assumption was that distinctions t -t among firms of different sizes. Hence the sample was chosen to provide sufficient pool of firms within each size group so that meaningful comparisons could be made; it is not necessarily representaive of Senegalese manufacturing firms as a whole, which are skewed heavily towards microenterprises on a per unit basis. Within each size group and subsectors, firms were selected at random, although it was impossible to be statistically representative in a strict sense for lack of an up-to-date census of smtll enterprises. Hence the analysis focuses on differences among groups rather than on averages for the sample as a whole. Three size groups were designated for manufacturers: "micro" (1-9 regular employees), 'small" (10-49) and "medium/large" (50+) consistent with the government's classification system.42 A total of 63 manufacturing firms were selected, equally divided among the three siz-, groups and ranging from 1 to 290 workers. Five subsectors were represented: food, textiles/sht:es, metal products, automotive (including repair shops) and chemicals. Microenterprises surveyed were mostly textile/shoe companies (71 percent), with metal and food firms making up the rest. The other size groups had more even distributions spread among all five subsectors and with no subsector comprising more than 50 percent of the size group (see Figure 1). Annual sales of 42 Firms were classified according to their size at the time of the interview for consistency, since not al fins provided information for 1986. The one exception, a failed chemicals firm, was classified by its size before shutting down operations. Of 52 firms providing information on employment in 1986 or at stlt up (for new firms), only 6 (12 percent) had changed categories by 1989. One microenterprise graduated into the SSE group and two SSEs graduated into the 50 plus size category, while three medium/large fums feU into the SSE group. 19  industrial sample firms is estimated at 47 billion CFA francs, which is equivalent to 9 percent of recorded 1988 industrial sales.43 FIGURE I SUBSECTORAL COMPOSITION OF SIZE GROUPS All Firm Mico 8SE* jervg 1_416 1 10% %4 _1ssj_41 7 % %