|
The Development of Small and Medium-Sized Enterprises in Developing Countries - The Case of Zimbabwe
CDR Working Paper 98.5, May 1998
Contents
Abstract This paper looks at the small enterprise/informal sector from a macro-perspective: Does the sector expand or contract, and when and why does it expand or contract. Macroeconomics tend to focus only on the formal sector and therefore generally has little to say about it. The small enterprise literature has more to say, but provides many different and often conflicting interpretations and conclusions about development trends. These are difficult to test, because almost all our knowledge about small enterprises and the informal sector is based on cross-sectional data or at best a few (in fact very seldom more than two) comparative cross-sections, which are in fact seldom strictly comparable. Models and theories of small enterprise development are therefore inferences from cross-section data and not really based on solid empirical knowledge. This becomes a problem when we attempt to draw conclusions about the effect of events such as structural adjustment, because we cannot know whether the observed changes are a result of short-run policies or long-run structural changes. In an attempt to overcome this problem the paper investigates the changes in the structure of the small enterprise sector in Zimbabwe over two periods of time three comparative cross-sections and compare the Zimbabwe data with data from Kenya and Botswana. One of the current debates about small enterprises concerns their ability to graduate into medium-sized enterprises. In the last part of the paper we discuss on the basis of data from Zimbabwe the constraint to such a graduation.
This paper is an attempt to look at the small enterprise/informal sector from a macro-perspective: Does the sector expand or contract, and when and why does it expand or contract. Macroeconom ics tend to focus only on the formal sector and therefore generally has little to say about it. The small enterprise literature has more to say, but provides many different and often conflicting interpretations and conclusions about development trends. These are difficult to test, because almost all our knowledge about small enterprises and the informal sector is based on cross-sectional data or at best a few (in fact very seldom more than two) comparative cross-sections, which are in fact seldom strictly comparable. Models and theories of small enterprise development are therefore inferences from cross-section data and not really based on solid empirical knowledge. This becomes a problem when we attempt to draw conclusions about the effect of events such as structural adjustment, because we cannot know whether the observed changes are a result of short-run policies or long-run structural changes. A second problem is that until recently most empirical studies have been based on fairly small samples, often less than 100 enterprises and seldom more than a few hundred. By necessity most studies have therefore also focused on only a part of the small enterprises (e.g. a specific sector (manufacturing more often than trade), the smallest, the medium-sized, the urban, the rural, the poorest, or the most successful) or treated the small enterprises as if they were a homogeneous group. One of the results of this has been endless discussions about the definitions of small enterprises and the informal sector, rather than about the diversification and development of the sector. Another result has been that different theories and models of small enterprise development have been seen as alternative rather than complementary. Only recently surveys with much larger samples have been carried out in a number of African countries, and in some countries even two consecutive surveys, which permit a more desegregated analysis and in some cases also a more detailed analysis of the dynamics. We shall attempt to make such an analysis of the small enterprise dynamics in section 3 based on data from Zimbabwe on small enterprises at three points in time. This one-country analysis hides, however, the large national differences existing in the structure of small enterprise sectors, and in section 4 we compare the small enterprise sectors in Zimbabwe, Kenya and Botswana. As a starting point we shall in section 2 outline four different theoretical approaches to small enterprise or informal sector development, which result in rather different expectations of the development of small enterprises. We shall use the discussion of these different approaches as a basis for the empirical analysis made in sections 3 and 4. One of the important current debates concerns the development of the medium-sized enterprises, whether in the upper end of the informal sector or in the lower end of the formalsector. Is the middle missing? Is it possible to make small enterprises mature into medium-sized or large enterprises? What are the constraints to such a process? On the basis of the more dynamic framework outlined in sections 2 and 3, we shall discuss these questions in section 5.
2.Four Theoretical and Empirical Explanations of the Development of the Small Enterprise/Informal Sector Small enterprises develop in between agriculture and the formal, primarily urban economy and thus depend on the development of both. However, most theories of small enterprise development tend to focus either on the urban informal sector and relate it to the short-run swings in the formal urban economy or to focus on rural non-agricultural activities and relate them to the long-run structural changes in the rural economy resulting from processes of commercialisation, industrialisation and urbanisation. The different theoretical explanations also differ in what they see as the major driving force, namely either an excess labour supply which cannot be absorbed in the formal sector and is therefore forced into the informal sector and small enterprises in spite of poor pay, or a demand for goods and services which is not satisfied by the formal sector (or which it supplies at prices which are unattractive to part of the consumers) and therefore makes up a potential market for small enterprises. The four different theoretical explanations presented below combine these two characteristics in different ways: I Labour Supply and the Urban Informal/Small Enterprise Sector The small enterprise sector is assumed to develop in response to the growth in unemployment and to function as a place of last resort for people who are unable to find employment in the formal sector. Therefore the urban informal sector will grow in periods of crisis when the formal sector contracts or grows too slowly to absorb the labour force, in spite of its inefficiency and the generally low pay it affords its participants. However, in periods when formal employment grows the small enterprise sector is assumed to contract again and thus develop anti-cyclically to the formal economy. A large part of the literature has taken this position in interpreting the growth of the informal sector after structural adjustment (see e.g. Brand et al. 1995, Meagher and Yunusa 1996 and Daniels 1994). Although there are undoubtedly a certain element of truth in the explanation, studies based on this theory of small enterprise development suffer from a number of empirical problems: Firstly, empirical studies pursuing this argument tend to limit their sample to the poorest and smallest enterprises and eliminate the more "modern", successful, capital and knowledge intensive enterprises beforehand, and therefore imply a certain element of tautology when they argue that small enterprises are inefficient and afford their owners low earnings. Secondly, due to lack of time series data, growth estimates are often based on information on the year of establishment of the enterprises interviewed in cross-section surveys. However, given the short life of many small enterprises, such data are likely to give a completely wrong picture of the growth in the number of small enterprises, because we cannot know how many enterprises have gone out of business before the time of the survey and, therefore, do not know whether the registered enterprise-establishments are a result of rapid replacement of enterprises or of real growth in the number of enterprises. Thirdly, even if a growth in the number of small enterprises can be identified in the period after structural adjustment, the argument would require that the informal sector did not grow, or at least grew more slowly before adjustment , and without time series data this is impossible to show and as we shall see in section 3 it may in fact be doubtful. Fourthly, if the informal sector is to function as a place of last resort, it must be easily accessible. However, many studies show that this is only the case for a small part of the informal sector. It is sometimes argued that small informal businesses concentrate on trade because this is supposed to require less capital and knowledge than production. While it may be true that production requires more investment capital than trade (when we talk about very small-scale producers even that may be doubtful), small-scale trade is likely to require more working capital in order to secure a certain income than small-scale production, partly because the value added is lower for the trader than for the producer, and partly because in small-scale production the customer will often be required to pay for the materials in advance, while the small-scale trader will have to give credit (even more often than large formal retailers). Therefore there is a strong limitation to the extent to which the small enterprise sector can function as a place of last resort during crises. Scoones et al. (1994) have studied the development of non-agricultural activities during the drought in a drought-stricken area in Zimbabwe. He shows that while many people started to operate small non-agricultural activities in the beginning of the drought, most had to stop again as the crisis tightened, partly because they had eaten their working capital, and partly because their customers had nothing more to buy for. Therefore, the informal sector plays only a limited role as a place of last resort. That place is rather reliance on family networks and patrons. II Output Demand and the Urban Informal/Small Enterprise Sector The argument in this theory is that a prerequisite for the development of small enterprises, whether formal or informal, is that there is a market for their products and services. Therefore the small enterprise sector will tend to develop cyclically concurrently with the economy as a whole. However, the small enterprises will also develop in competition with the large enterprises in the formal sector, and their development will be constrained by formal sector monopolies. Structural adjustment policies which limit such monopolies and attempts "to create a level playing field" will therefore be an advantage for the small enterprises, because it may allow them to conquer market shares from the large ones. Proponents of structural adjustment tend to base their arguments on this theory. Empirical studies based on this theory tend to focus on the upper end of the informal sector, often the manufacturing enterprises and the larger, more resourceful and successful small enterprises, which have a potential to grow into the formal economy (see e.g. the different sampling procedures in the two Nigerian studies made by Meagher and Yunusa (1996) who adhere to a labour supply theory and Dike (1997) who adhere to a output demand theory). They often show a special interest in the possibilities for strengthening the small enterprises through mutual networking or clustering or through the creation of linkages between small and large enterprises, such as sub-contracting or franchising. However, due to low levels of trust in the business community and poor infrastructure such arrangements have in general been little developed in the African countries (see e.g. Liedholm and Mead 1991, Grierson and Mead 1995, and McCormick and Pedersen 1996), and are unlikely to develop in crisis situations where the large enterprises operate at a low capacity utilisation. Instead enterprises appear to integrate vertically as much as possible in order to be independent of the unstable environment (Trulsson 1997). This clearly reduces the efficiency of the large enterprises and the market access of small enterprises to services and inputs. Based on the Gemini data, Daniels (1994) shows that enterprises in high-profit sectors tend to behave as expected by the output-demand theory , while enterprises in low-profit sectors rather follow the labour supply theory. III Commercialisation of the Rural Areas, Urban Migration and the Development of Rural Non-agricultural Activities and the Urban Informal Sector a.Monetisation, Commercialisation and Urbanisation of the Rural Areas and Growth in Non-Agricultural Activities Most African countries are still predominantly rural with a large share of agricultural production being non-monetised subsistence production. However, as a result of increasing commerciali sation and urbanisation they are gradually being drawn into the money economy. This increasesthe demand for cash and urges people to engage in cash-earning non-agricultural activities, at the same time as increasing cash among people creates a growing market for rural non-agricultural goods and services. Therefore, even if the economy as a whole stagnates, increased monetisation may in itself lead to an increased rural market primarily for non-agricultural activities, but also to some extent for the informal and formal sectors in the urban areas, which supply the rural areas. This is the stand of the theory of agriculture-led industrialisation (Mellor 1976, Adelman et al. 1989, and Haggblade and Hazell 1989). Contrary to the output demand theory, which linked the informal sector to a cyclical urban economy, this theory links the informal sector to the long-run rural development and the generally irreversible trends of de-agrarianisation and urbanisation (See e.g. Bryceson and Jamal 1997). It is supported by the large share of the informal sector which in most low-income countries are located in the rural areas. b.Rural Labour Surplus, Urban Migration and the Development of the Rural and Urban Informal Sectors Parallel to the labour surplus theory presented above, which related labour surplus to formal sector employment, at least part of the new literature on de-agrarianisation relates the development of rural non-agricultural activities to the rural surplus labour, which either supplements agricultural production with non-agricultural activities or migrates periodically or permanently to the urban areas (Bryceson 1996 and Bryceson and Jamal 1997). The effect of such a theory would be pretty much the same as of the theory of commercialisation of the rural areas, namely a continuous growth in the small enterprise/informal sector, although it would tend to focus more on the development of the urban informal sector as a result of rural-urban migration. IV Industrialisation and Small Enterprise Development. According to both traditional Marxist and classical economic theory, small enterprises are expected to be ousted by "modern", large-scale industry and thus gradually disappear as a result of industrialisation and economic development. These expectations were based on empirical findings that small enterprises tended to be less efficient than large ones, because the large enterprises achieved either scale economies or monopoly powers. Many studies of small African enterprises in the 1960s and 1970s expected this development (see e.g. Müller 1980), and after independence most African governments (and many donors) tended to see the small enterprises as a result of a stagnating formal industry and a sign of under-development to be suppressed. Clearly some sub-sectors of small enterprises, like tailoring, do compete with industrial products (whether nationally produced or imported) and have been forced to reduce their level of activity or restructure. However, in general small enterprises have disappeared neither in theindustrialised nor in the developing countries. The problem with the argument and most of the empirical studies of scale economies is that the small and large enterprises are assumed to be doing the same thing in direct competition. However, the small enterprises generally do not do the same as the large ones. They tend to find niches in the input or product markets where the large enterprises cannot exploit their scale economies and therefore cannot compete. Or they specialise in distribution to peripheral or low-income areas where the distribution costs of the large-scale sector are prohibitive. Therefore many enterprises specialise in trade. Even those small enterprises which we traditionally classify as producers mostly sell their products retail and often use a larger part of their time and earn a larger share of their profit as retailers than as producers. Where the small enterprise would not be able to compete with the industrial producer it may be able to compete with the local retailer selling the industrial product. There will of course be an indirect competition between the formal and the informal producers, but not a complete subordination as assumed by the theory of petty commodity production. Therefore Müller (1997) could find that Tanzanian blacksmiths whose doom he had prophesied in the 1970s were still in operation and innovating in 1994.
3.A Composite Theory of Small Enterprise Development. The Development of Small Enterprises in Zimbabwe The expectation to small-enterprise development of each of the four theories presented above are shown diagrammatically in Figure 1. The four theories of small-enterprise development have generally been seen as alternative explanations, although they are usually applied to partly different segments of the small enterprise sector. Therefore we shall argue here that the different explanations should rather be seen as complementary theories, depicting simultaneous processes which influence different segments of the small enterprise sector with different weights. We would thus expect different parts of the small enterprise sector to develop differently over time. Taking the different theoretical explanations at face value we would expect: -that activities with easy access functioning as a place of last resort would expand during crisis and contract again during an upturn in the economy (theory I in Figure 1); -that market responsive activities would expand during an upturn in the economy and contract during crises (theory II in Figure 1); -that activities developing as a result of long-run processes of commercialisation and monetisation in the rural economy would expand both during crises and during an upturn in the shorter cycles of the economy (theory III in Figure 1); -that especially manufacturing activities operating in direct competition with the formal sector would contract both during crises and an upturn in the economy (theory IV in Figure 1); Such a composite theory would be in correspondence with the findings of Pedersen’s (1997) analysis of small businesses in small rural towns in Zimbabwe that even during the deepest crisis in 1992 some sectors benefitted from the crises at the same time as others suffered. To investigate such a composite theory ideally requires time series of desegregated data for the development of the small enterprise/informal sector. Such data are hard to come by. However, in Table 1 we present data for the sectoral composition of the rural non-agricultural activities in Zimbabwe in 1986, 1991 and 1993, respectively. Here the first period 1986-91 represents a period of relative economic progress, while the second period 1991-93 was a period of economic crisis, where the introduction of structural adjustment measures and a serious drought coincided in 1992, and led to rapid price hikes and dramatically falling buying power for the majority of the population. All the data are based on a complete count of the number of small enterprises (employing 0 to 50 persons) at all addresses in a stratified random sample of rural and urban enumeration areas. The expanded figures should thus in principle represent a complete picture of the population of small enterprises. In practice there has of course been some differences in the sampling procedures especially between the first and the two later data sets (see the note to Table 1), however, we still believe that the data are sufficiently comparable to give a crude indication of trends. Unfortunately similar data are not available for the small urban enterprises in 1986, but only for the years 1991 and 1993. These data are also shown in Table 1. They show that the urban small enterprises only make up 30 per cent of the total. They will be analysed in more detail below. Small Rural Enterprises in Zimbabwe In comparing the three percentage distributions of small rural enterprises in Table 1 it should be remembered that the total number of small enterprises has increased dramatically during the three years, both because the rural population has increased by around 3 per cent a year and because the average number of small enterprises per household has increased. Helmsing (1991) cites a survey from 1983/84 showing that 15 per cent of all rural households were operating non-farm activities then, while the GEMINI study (McPherson 1991) found that around 40 per cent of all rural households operated at least one non-agricultural activity. This corresponds to an annual change Table 1. Sectoral distribution of small rural and urban enterprises. Signs between the
Note. The 1991 and 1993 data are drawn from the GEMINI studies (Daniels 1994) and the 1986 data from Helmsing (1991). All three data sets are based on household interviews in a random selection of rural areas in the county. There are however a number of differences in the sampling procedures used in the GEMINI and Helmsing studies, which means that a comparison can, at best, give a crude indication of trends. For instance, while Helmsing’s study covers only the rural areas proper and is based on simple averages, the GEMINI study includes the small rural centres and is based on population weighted averages. Exclusion of the small towns from the 1991 and 1993 data would narrow the gap between the 1986 and 1991 data a little, but without changing the structure of the data. in the percentage of rural households with non-agricultural activities of about 3-3.5 per cent and an annual increase in the number of non-farm activities of about 6-6.5 per cent, compared to 7.1% in the period 1991-93. Thus from 1986 to 1991 the number of rural non-agricultural activities is likely to have increased by 50-100 per cent, while it increased by 12 per cent between the 1991 and 1993 surveys. Thus only sub-sectors for which the percentage figure for 1991 is at least 50 per cent below the 1986 figure are likely to have experienced an absolute decline, and those with increasing percentage figures are likely to have increased by more than 50 per cent in absolute terms. A comparison of the pattern of change in the main sectors shows that the direction of change is the same in the two periods, i.e. a relative contraction of manufacturing and construction and an expansion of trade and services. This speaks in favour of a structural theory of type III or IV. However the speed of change seems to be larger during the crisis period, which proponents of the labour supply theory might see as a sign in favour of type I. If we look in more detail at the sub-sectors within the manufacturing sector, the upper part of Table 1 shows that the small manufacturing activities are highly concentrated in a few subsectors: beer brewing (8.2 per cent in 1991), dressmaking and tailoring (5.8 per cent), knitting (9.2 per cent), crocheting (7.5 per cent), grass and cane products (20.0 per cent) and wood carving (5.1 per cent). In total these six subsectors comprise 56 per cent of all the enterprises. Although the manufacturing activities as a proportion of all rural non-farm activities went down in both periods, this pattern does not hold for all sub-sectors. Only tailoring and dressmaking, knitting, wood carving, and tin smithing went down during both periods, dropping from 32 per cent to 16 per cent of all non-farm activities. Especially tailoring and dressmaking, which dominate the group of sub-sectors, have decreased rapidly probably in both relative and absolute terms during both periods. On the other hand, knitting and wood carving have decreased more moderately and have probably experienced a small absolute increase especially during the first period. These sub-sectors are met with less severe competition from formal industry and imports than tailoring, because they more often produce a quality which is high compared to the industrial substitutes. They also have access to tourist and export markets. That especially tailoring and dressmaking are hit by the long-run structural decline depicted in theory IV corresponds to our expectations. It has been argued that the opening for imports of second-hand cloth as part of the structural adjustment should have hit tailoring and dressmaking hard. The figures hardly support that view as the sub-sector appears to have contracted even more rapidly before the structural adjustment than after. The sub-sectors crocheting, carpentry and other manufacturing have grown in relative terms in both periods, from 13 per cent in 1986 to 22 per cent in 1993, thus corresponding to theory III. That "other manufacturing" has grown in both periods corresponds well with our expectations, as we would expect a gradual diversification of the economy as part of the structural transforma tion of the economy. It is more surprising to find crocheting here, which is the most important subsector in the group. It grew especially rapidly during the second period, probably because crocheting work has been used for bargaining in the informal import trade from Botswana and South Africa, which has increased rapidly after the initiation of the structural adjustment programme (Brand et al. 1995). Food processing, grass and cane products and brick making expanded their share of production during the upturn and contracted during the downturn. This could be an indication of a cyclical market orientation a la theory II. However, for both food processing and grass and cane products, the contraction during the drought is likely to be caused by lack of raw materials rather than lack of market. The drought relief provided for the rural areas came in the form of mealy meals which had already been processed in the large urban mills, and therefore led to growth in retail trade rather than in milling. Finally, pottery, welding and shoe works and repair have contracted during the upturn and expanded during the downturn. They thus seem to follow the labour supply theory of type I. This is, however, not very logical as neither welding nor pottery can be said to be easy entry sectors. Both require considerable technical knowledge and welding in addition relatively large investments. Shoe repair could better be considered an easy entry sector, but it only grew very slowly during the crisis period. Small Urban Enterprises in Zimbabwe For the small urban enterprises we only have data for the crisis period 1991-93 (see Table 1). These data show that the small urban enterprises make up less than 30 per cent of all the small enterprises, which is still an over-representation relative to the total population, of which only 25 per cent is urban. The number of small urban enterprises grew by only 3.6 per cent per year during the period (the number of people working in the small urban enterprises appear to have increased even less (Daniels 1994)), which is, however, less than half of the growth in the number of small rural enterprises (7.1 per cent), and probably also less than the increase in the urban population during the period. This does not appear to support the hypothesis that increased urban labour supply during the crisis should have led to a growth in the number of small enterprises (theory I). Although there has been a tendency to contraction, especially in the public formal sector in many African countries as a result of the structural adjustment, this has in most cases led only to a stagnation in formal employment but seldom to a large absolute decline in the formal urban employment. Therefore the number of formal employees actually retrenched is small relative to the national labour market. It is of course more important in terms of the urban labour market, but even here it is so little that only empirical studies based on very large samples will be able to register it. Studies of small enterprises after structural adjustment therefore generally do not findretrenched employees among the small-scale entrepreneurs (see e.g. Brand et al. 1995). The large number of urban youth and rural migrants entering the urban labour market every year, whether there is a crisis or not, is much more important for the urban unemployment. And the falling real wages in the formal sector is likely to be a much more important reason for employees or their wives to start a small enterprise than retrenchment; but that fall in real wages is not just a result of structural adjustment, but started much earlier. The data in Table 1 on the sectoral distribution of the small urban enterprises show that the urban enterprises are much more concentrated in a few sectors than are the rural enterprises. The urban enterprises are especially concentrated in petty trade, which comprised 28.6 per cent of all the urban enterprises in 1991 growing to 35.6 per cent in 1993, and within the manufacturing sector in tailoring, dressmaking, knitting and crocheting, which together comprised 49.5 per cent of all the enterprises in 1991, but had fallen to 46.4 per cent in 1993. However, of the four sub-sectors only tailoring and knitting went down, while both crocheting and dressmaking increased rapidly. Thus in the urban areas there has been a concentration of small enterprises into low income easy access sectors during the crisis, which might speak in favour of the labour supply theory. The dominance of these sectors already before the crisis indicate, however, that this growth in the low income sector has not been specific to the crisis period but is rather a continuation of more long-run processes of urban migration and the increasing participation of women in the urban labour market which has taken place especially since independence. This also corresponds to the findings of Brand et al. (1995) that most of the small enterprises run by women had been established well before the crisis in 1992. These results indicate that there is a great diversity in the development of the small enterprises, by sectors and otherwise. But the results also indicate that the development of the small enterprises is linked primarily to long-run (and basically irreversible) structural processes, such as urbanisation and commercialisation in the rural areas. This is in contrast to most of the current policy debate, which tends to discuss the small enterprise development in a context of relative short-term events such as the structural adjustment, droughts and cycles in the urban economy. Clearly such events have an impact on the small enterprises, but the impact is likely to differ from activity to activity, and the small enterprise sector as a whole is likely to increase whether there is a crisis or not, although it may not be the same sub-sectors which are growing in different periods. Since the 1970s, one of the main goals of rural development policies has been to increase agricultural labour productivity and rural standards of living in order to limit rural-urban migration and growth in the urban informal sector. The fallacy of this argument is, however, that a population surplus will be created in the rural areas, whether agricultural labour productivity increases or not (The exception to this is where land productivity increases without a correspond ing increase in labour productivity as it can be the case in irrigation schemes especially when theymake more than one harvest possible). If agricultural labour productivity does not increase, the population will have to leave the rural areas because they cannot be fed; if productivity increases, they will have to leave because there is no work for them. This is the reason why it is so difficult to stop the process of urbanisation; it is also the reason why the informal sector tends to grow whether there is a crisis or not, either in the form of rural non-agricultural activities or in the form of an urban informal sector. However, the success of agriculture does make a difference for the structure of the informal/small enterprise sector. Successful small-scale agriculture requires an efficient system of produce traders, farm input and consumer goods suppliers and other agricultural services which create a market for small and medium-sized enterprises in the rural areas. Therefore successful agriculture creates a basis for rural non-agricultural activities and the development of small rural towns (Pedersen 1997). An agriculture with low productivity primarily lead to urban migration and increases in the urban informal sector in the large towns. Just as agriculture, whether it is successful or not, tends to create a rural population surplus which results in urban migration and increased urbanisation, the large-scale, formal sector tends to generate small urban enterprises whether it is successful or not. The formal sector enterprises which have developed in most African countries have tended to develop into vertically integrated, closed and self-sufficient enterprises which rely on their environment as little as possible. The reason for this has been poor infrastructure and unreliable local input and service suppliers. But the closedness of the formal sector has also prohibited such small and medium-sized suppliers from developing and thus reduced the efficiency of the formal sector itself. This low efficiency has, however, permitted an informal sector to develop in parallel to the formal sector, and at least partly operate in its own parallel network. Opening the formal sector by externalising activities will open new possibilities for the small enterprises, but it will also increase the formal sector efficiency and therefore reduce the space for inefficient small enterprises. It is therefore likely to change the structure of the small enterprise sector and increase it’s sophistication, but not necessarily it’s aggregate size.
4.National Differences in the Structure of the Small Enterprise Sector The above analysis indicates that the changes we are able to observe in the few empirical data we have are due to long-run structural changes rather than short-run cycles and policies. However, it does not say much about what these long-run structural changes are. In a literature survey on macro analyses of micro enterprises in developing countries, Liedholm and Mead summarize that "macro-level empirical evidence indicates that as aggregate per capita income increases there isa systematic pattern of evolution of small and medium-sized enterprises towards larger firms based in larger localities, producing more modern products". Of course, such aggregate empirical evidence over the full range of development levels fits well to the classical and Marxist economics, which expect the importance of small enterprises to diminish in the course of development. It also corresponds to Adelman’s generalised model of agriculture-led industrialis ation. However, it is too simplistic to take account of differences in small enterprise efficiency, linkages between the small and large enterprises, and changing roles of women entrepreneurs, differences in the level of education in the labour force and other socio-economic differences. Therefore it is unlikely to be useful in understanding the differences between African countries covering a limited range of development levels. To show this, we have in Table 2 compared the size and sectoral composition of the small enterprise sectors in Zimbabwe, Kenya and Botswana. Contrary to what the "aggregate empirical evidence" leads us to expect the table shows that the size of the small enterprise sector is twice as large in Zimbabwe as in Kenya even though the per capita income is almost twice as high. This is also surprising because Kenya is known as having a policy which, at least on paper, has been favourable to the informal sector while Zimbabwe’s policies have clearly favoured large industry. A closer look at Table 2 reveals that the difference between Zimbabwe and Kenya is almost entirely due to a very large number of clothing enterprises and grass and cane weavers in Zimbabwe. If they are left out of the distribution, the number of enterprises per 1,000 inh. is very close to that in Kenya. From a sectoral perspective this very large informal clothing sector in Zimbabwe is difficult to explain as Zimbabwe has a much larger and more efficient large-scale clothing industry than Kenya, and a clothing retail sector which is dominated by large retail chains. Thus while the clothing industry in Zimbabwe make up close to 20% of the total manufacturing employment, it employs less than 5% in Kenya. This must mean that there in spite of the larger size of Zimbabwe's clothing industry there still remains a large clothing market which the formal sector is unable to reach. The many small clothing enterprises in Zimbabwe means that two thirds of all the small enterprises are manufacturing enterprises against only one third in Kenya. Correspondingly, retail trade makes up two thirds of the small enterprises in Kenya but only one third in Zimbabwe. However, in absolute number of enterprises per 1,000 inh. the sector is of approximately the same size in the two countries. Given the importance of large retail chains in Zimbabwe, this is surprising , but must be an outcome of the larger per capita income in Zimbabwe which still leaves a market for the small retailer in spite of competition from the chain stores. As a result of the large number of small clothing enterprises, a much larger share of the small enterprises are women-owned in Zimbabwe than in Kenya. This could be a result of the higher per capita income and a more monetised economy, which presses more women to earn cashincomes. However, the large number of small clothing enterprises is not a result of the structural adjustment policies and the drought, because their number was even larger in 1991 than in 1993. Although the average size of the small enterprises is about the same in Zimbabwe and Kenya, the many small clothing enterprises mean that the share of one-person enterprises is much larger in Zimbabwe (78 per cent) than in Kenya (57 per cent), and the share of enterprises with 2-5 persons correspondingly smaller (19 per cent against 42 per cent). On the other hand, there are also more of the largest group of small enterprises (6-50 persons working) in Zimbabwe (3 per cent) than in Kenya (1 per cent). Surprisingly the large number of small clothing enterprises in Zimbabwe also means that the small urban enterprises in Zimbabwe are smaller than the rural ones. This is contrary to Kenya and also to our theoretical expectations, because market access tends to be better in urban than in rural areas. Thus for Kenya Hosier (1987) has shown that while rural enterprises primarily produce-to-orders more urban enterprises produce in larger batches for the market. In Table 2 we also compare Zimbabwe to Botswana, which has a per capita income almost three times that of Zimbabwe. The Botswana data show that the size of the small enterprise sector is much smaller than in both Kenya and Zimbabwe. Thus, although the share of women-owned enterprises is larger than in Zimbabwe, the number of women entrepreneurs per 1,000 inh. is much smaller, and the pressure on women to earn a cash income appears to be smaller. Furthermore, the women are not concentrated as strongly in the clothing sector as in Zimbabwe, but appear to be spread more on different sectors. Probably due to the large share of women entrepreneurs, there are more one-person enterprises (66 per cent) than in Kenya, but not as many as in Zimbabwe. On the other hand, the number of enterprises in the largest group of small enterprises is larger (7 per cent) than in both Kenya and Zimbabwe, which however still means that, in absolute terms, there are fewer of these medium sized enterprises per 1,000 inh. than in both Kenya and Zimbabwe. These findings might be seen as an indication that the share of the largest group of small enterprises increases with per capita income as both the economic theories and the "aggregate empirical evidence" indicate. However, this is not the case when we look at the absolute number of enterprises per 1,000 inh. Also the share of one-person enterprises appears to depend more on labour market structures which are not clearly linked to per capita income but rather to different social structures, such as the entry of women on the non-agricultural labour market and the structure of the educational system, which tend to be of a long-run nature and not easy to change in the short run. We would see this as an indication that the small enterprise sector cannot be explained by simple one-dimensional models, but must be linked to the development of both the formal urban economy and the rural agricultural economy and to the development of both labour markets and the structure of the demand for small enterprise products. Table 2
5.Development of the Missing Middle as a Result of Structural Adjustment One of the arguments made in much of the newer small enterprise literature is that the middle size enterprises, i.e. those enterprises between the large and formal and the small and informal, are missing, because the large have been favoured by the economic policies and the small have been constrained. Therefore, it is argued, we must create an enabling environment or a level playing field by dismantling the formal sector monopolies. This should permit small enterprises to grow larger in a more even competition with the large formal enterprises. However, many small enterprises serve small spatial or product markets, and as long as they earn a living there is no real reason why they should grow in size, especially because the success of one enterprise is likely to be the death of others. Therefore the objective of an enterprise policy should not be to increase the size of the enterprises, but to increase productivity and develop new markets. Summarising the earlier economic literature on the missing middle, Fafchamps (1994) explains the missing middle as a result of an u-shaped cost-curve: Small enterprises benefit from the economies of being small (e.g. low management costs) and invisible (e.g. evading taxes and labour legislation), and large enterprises benefit from the economies of scale, while medium-sized enterprises suffer from not being invisible and yet have not achieved economies of scale. Therefore growth in size will not necessarily increase efficiency. However, the concepts of small- and large-scale economies are rather nebulous, and based on new institutional economics. Ferrand (1997) has substituted them with the concept of institutional discontinuities, which seems more fruitful. One of the difficulties of medium-sized enterprises is that the large formal sector enterprises tend to operate in their own closed production and distribution network or institutional setting to which the medium-sized enterprises do not have access, while the small enterprises operate in a local family-based network, which is insufficient for the medium-sized enterprise. Of course, the structure of the formal sector institutional setting varies from country to country and from sector to sector. In some cases it has been operated as parastatal monopolies; in other cases the large private enterprises base their operation on services and infrastructures obtained within their own national or multinational corporation and not accessible to others. Where marketed services exist, it is often operated within networks of minority entrepreneurs, such as Asians in East Africa, whites in Zimbabwe and Lebanese or Greeks in West Africa, which may be difficult to access for small or medium-sized black African entrepreneurs. The extensive recent literature on the capital market in developing countries supports the claim that the medium-sized enterprises tend to be caught between the formal and informal capital markets. They seldom have access to the formal banking system, and the lending capacity of thenetwork of family and friends seldom allow small enterprises to expand into medium or larger size. Recently NGOs and donors have with some success attempted to reach the small and medium-sized enterprises through special credit programmes, such as lending guarantees and group lending, though it is still only a very small fraction of the small enterprises which have benefitted from these programmes. However, the dual system of services is probably even more serious for the medium-sized enterprises in the case of trading organisations. Most small enterprises sell directly to the final consumer and thus earn the wholesale and retail profits as well as the producer profit. If it is to expand its market beyond the local area, it will have to give up part of its profit to a wholesaler or middleman. However, this will only be attractive if it is able to increase its productivity and find an efficient and trustable wholesaler. Studies in Asian countries increasingly show that traders have often have played an important role in the development of rural industries (see e.g. Weijland’s (1994) studies from Indonesia). In African countries such wholesale trade is, however, generally neither well-developed nor very efficient. Thus the GEMINI data indicate that in both Zimbabwe and Botswana only a fraction of a percent of the small enterprises are wholesales, while it is somewhat more in Kenya. Lack of physical and institutional infrastructure has generally contributed to the inefficiency of wholesale organisations. The non-existence of production standards and a trustable legal system of contract enforcement have meant that long-distance trade almost entirely takes place either within the large formal organisations which are able to control their internal transactions or in the form of direct person-to-person transactions and person-carried commodity flows. Examples of such person-carried trade links are frequent in the literature on African trade, see e.g. Gibbon (1997) on fish trade between Tanzania and Zaire and Sørensen (1997) on grain trade in Uganda. In the absence of institutional trust, the individual trust is also reduced. Without institutional backing only wealthy persons may be trusted, because poor people may, no matter how honest they are, easily by incidents beyond their power be brought in situations where they will not be able to fulfil their obligations. Therefore, like many others, Trulsson (1997) finds in a detailed study of small-to-medium-si zed manufacturing industries in northwestern Tanzania that the enterprises tended to integrate vertically as much as possible in order to avoid dependency on an unpredictable environment, rather than expanding the market horizontally. A study of small industrial firms in Gutu, a small Zimbabwean town (10-20,000 inh.), also shows that this lack of trading organisations is a major constraint to their development, and in addition some of the strategies they follow in their attempt to reach the market are presented (Pedersen 1997). From the mid-1980s to the mid-1990s, eight small-to-medium-sized industrial firms with 15 to 35 employees each had been established in the town: a bakery, a scotchcart producer cumautorepairer, a soap factory, two clothing factories, a commercial grain mill, an oil mill, a bookbinder cum printer. There are of course also smaller producers, a large number of retailers and some wholesales in the town. The different market strategies of the eight medium-sized enterprises are outlined below: -The bakery is a branch of a large national firm. It has its own vans and distribution system reaching out to retailers in the town, the rural areas of the district and the neighbouring district, but is met with competition from bakeries in neighbouring districts trying to sell bread in Gutu; it has also recently got competition from a small bakery in Gutu centre primarily producing bread and cakes primarily for customers in the town centre. -The scotchcart producer produces for the local market only. He sells directly to the farmers. The market is, however, very unstable and mostly limited to the after-harvest period in good years. Therefore he produces only periodically. He competes with larger producers in Harare and Bulawayo by delivering a stronger, but rougher product, which is also cheaper. -The soap factory produces for both the local and the national market (though primarily the southeastern part). It was established some years before the structural adjustment and had then, as other small soap producers, difficulties getting tallow (the most important raw material), but always managed to get it either locally or through an import permit. It has been able to deliver to some of the large national grocery wholesales. It has recently moved its production to Masvingo, the nearest larger town, at least partly in order to get closer to the market. -The grain mill was established at the time of the structural adjustment hoping to benefit from the liberalisation of the grain trade. Before the structural adjustment the milling market was shared between a few very large urban based milling companies, which also operated large grocery wholesales, and a large number of small service mills, which milled people’s own grain for a fee. As it was only permitted for the Grain Marketing Board to move grain across district borders, most of the small service mills were located in the rural areas. To operate smaller commercial mills profitably was not possible, because the mills had to buy their grain from the marketing board which sold at a fixed price all over the country, and thus indirectly subsidized the big mills which on the open market would have much larger collection costs than a small local mill. Thus a small roller mill set up in Gutu in the early 1980s by one of the large milling companies never became profitable. It was sold to an African businessman around 1990, but closed a few years later. The large urban mills served the urban market, but as the Grain Marketing Board generally did not sell raw grain in the food deficit areas, theyalso had a de facto monopoly in serving the grain deficit areas. When the grain trade was liberalised, people in both the urban areas and in the grain deficit areas started to buy raw grain. As a result the number of small service mills increased rapidly especially in the urban areas, while the capacity utilisation of the large mills declined. The new mill established in 1991 was much smaller than the old one. Like similar mills established in a number of other district service centres at the same time it used a technology which permitted it to be operated both as a commercial and a service mill. They hoped to be able to compete with the large mills on the commercial market for mealy meal, but for a number of different reasons many of them have in practice survived primarily as service mills. Firstly, they have not had capital enough to buy grain directly from the farmers and therefore have not been able to free themselves from the marketing board and benefit from the trade liberalisation. Secondly, they have in a number of cases been unable to get grain from the marketing board because the large mills bought all the grain in the depots in areas where they had local competitors. Thirdly, and probably most importantly, it has been difficult for them to get into the market apparently partly because the large mills cum grocery wholesales have threatened not to deliver groceries to local retailers selling locally milled grain at a lower price than their own, and partly because they have offered credit which the small mills could not afford. As a result the small mills which have done best are those which either have their own distribution system, like the one in Gutu which is operated by a businessman also operating a supermarket and a wholesale serving the rural shops in the district, or have been able to secure contracts to deliver to local institutions, such as schools and military establishments. -At the same time as the oil mill was established in Gutu similar mills were established in a number of other district service centres. Although they have entered a market without an excess capacity, they have had very much the same difficulties both getting raw materials and getting into the distribution system. Although there is a small but increasing production of sun flower in Gutu district which could serve the small mill, this production was hard hit during the drought years, and it is also doubtful whether the local oil mill has the sufficient capital to secure a stable link to the oil seed producers. In Gutu it is the same businessman who operates the grain and oil mills. This is potentially important, as it will make it easier to combine the by-products of the two productions in a production of stockfeed. However, when we visited Gutu last this had not happened yet. -The oldest and smallest of the two clothing factories was established in the mid-1980s. It produces primarily work clothes for the industrial market, but also school uniforms. The largest of them opened in 1992, just as the economy was hit by structural adjustments and drought. At the time its plan was to produce women dresses for the chain stores whichdominate the clothing market in Zimbabwe. However, this soon proved to be impossible, partly because the small factory cannot deliver large enough batches to cater for the demand of the large retail chains and partly because it went into a serious crisis resulting in a large excess capacity. Due to the dominance of very large clothing factories and retail chains operating their own internal wholesales, open clothing wholesales are little developed. At the same time it was difficult for it to get into the local retailers because so many of them are chain stores which do not buy locally. To reach the consumer market both factories have therefore opened their own retail shops and also have salesmen travelling in the rural areas. However, they cannot live of the local consumer market alone. The large factory has therefore also entered the market for work clothes. This is a niche market which is not large enough to be of interest for the large industry, but permit the small factories to operate in sufficiently large batches to be profitable. Thus the two clothing factories have both managed to survive at a time when the clothing industry in Zimbabwe has generally been hit by massive crisis and extensive enterprise closures. -Finally, the bookbinder cum printer primarily caters for a local government institutional market. He started rebinding old schoolbooks for the schools, but now he also prints exercise books for the schools and formulas for the administration. He has also started a small chalk production and prints t-shirts for the schools’ sports teams. These examples show some of the difficulties small enterprises meet when they try to expand from a local retail market into a larger production. They show that markets are not something which is just there. Markets have to be developed, and the examples show some of the strategies that the small and medium-sized enterprises use to overcome their problems: -Diversification in order to level swings in the individual markets (the scotchcart producer); -Vertical integration to avoid instabilities in input or output markets (the mills); -Develop one’s own distribution system (the bakery, the mills and the clothing factories); -Develop institutional and niche markets (the book binder, the grain mill and the clothing factories). However, except for the last option, these strategies tend to be counterproductive because they divert resources away which could have been used to improve the production system and expand into larger scale. In line with the argument developed in the beginning of the paper, where we described the small enterprise sector as developing in the interregnum between the urban formal sector and agriculture, we can see the medium-sized enterprises as developing in the interregnum betweenthe large-scale, urban formal sector, agriculture and the small-scale informal sector. This indicates that there may be three different types of medium-sized enterprises: -Enterprises developing in the rural areas and out of agriculture and the rural informal sector. This is typically the successful rural trader who accumulates and grows into medium-size but seldom has invested in manufacturing; in our Zimbabwean case the scotchcart producer who sells his product directly to the farmers is the only one of the enterprises which can be said to have sprung out of a rural basis;
-Enterprises developing between agriculture and the large-scale processing industries now after the market liberalisation. In our Zimbabwean case this is the grain and oil mills and the only area in which rural traders invest; -Enterprises developing between the formal and informal sectors. In our Zimbabwean case these enterprises have all been established by people with a basis in the urban formal sector rather than in the informal sector. Thus, although structural adjustment has undoubtedly opened new opportunities for the development of small-to-medium-sized enterprises, these opportunities are not likely to be fully utilised unless the institutional gap existing between the small-scale informal institutions and the often closed infrastructures and services in the formal sector is bridged. In the absence of new open access institutions for the medium-sized enterprises, the dismantling of the large-scale monopolies is likely to benefit only the small enterprises catering directly to the retail market. | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
