|
King Cotton under Market Sovereignty: The Private Marketing Chain for Cotton in Western Tanzania, 1997/98
CDR Working Paper Subseries no ii.98.17
Contents
Abstract This paper examines the emergence of a private sector marketing chain for cotton in Tanzania in the period 1994/95-1997/98, based on field work conducted between June and September 1997. It embodies a description of the background to the current marketing system, cotton’s changing place in the Tanzanian macro-economy, the division of labour within the marketing chain, the relation between the cotton chain and intersecting chains, and earnings and profits of different types of participants. Alongside this discussion is one of the character of the social groups occupying different positions in the chain’s division of labour and the nature of the social relations between them. The paper concludes with a discussion of general issues of market development and regulation for agricultural commodities, on the basis of a comparison between the cotton chains in Tanzania and Tamil Nadu, India.
1. Introduction In the late 1940s and during the 1950s and 60s, cotton formed the basis of the commodification of African agriculture in the Tanzanian part of the Lake Victoria zone. After two decades of steady decline, the 1990s have finally witnessed signs of significant revival. This revival has been associated with the opening up of the sector for private participation at all levels, coupled with the apparent disintegration of the cooperative ‘movement’ which had dominated the sector for all but a few years since the late 1950s (Gibbon, 1998a,b). On the other hand, by 1997 serious questions were already being asked about the sustain ability of the recent private sector boom. In particular, they were being raised about the quality reputation of the Tanzanian export crop and about the spread of cotton diseases within the main growing areas. These questions feed into debates concerning the long-term sustain ability of econ omic growth based on a poorly regulated ‘free market’, especially where raw materials are involved. This paper examines these questions, but in the course of a somewhat different focus, namely what the cotton marketing chain in Tanzania (as well as its international extensions) tell us about the economic and social characteristics of spontaneously emerging marketing chains and private enterprise generally in deregulated but still predominantly marginalised economies. The conceptual framework adopted here is derived from recent studies of global and national commodity chains by, amongst others, Gereffi (1994), Harriss-White (1996) and Bernstein (1996). In these studies, commodity chains are analysed with respect to sequences of production and circulation in ‘core’ and related chains, the division of labour accompanying them, the distribution of earnings and profits entailed and the forms of economic and social organisation through which market power is constituted and reproduced. In this paper, the latter are approached as an outcome of the interaction between a given set of economic and social structures on the one hand and strategies employed by specific groups of actors and actresses onthe other. A secondary framework will also be employed, for discussing the relation of the cotton chain to more general processes in the current world economy. This arises out of recent critiques of ‘globalisation theory’ by Ruigrok and van Tulder (1995) and by Hirst and Thompson (1996). This study of the private cotton chain was conducted over a four month period during the 1997-98 buying season [ Unlike other crop seasons in Tanzania, the cotton season ’s start is dated from the beginning of marketing not the beginning of production. Hence crops sold during the 1997-98 marketing season would have been planted between November 1996 and January 1997.] and is part of a wider project of the author, examining emerging marketing chains in post-liberalisation Tanzania. Additional studies were conducted of three Tanzanian fisheries (prawn, Nile Perch and dagaa) during 1995-96 (Gibbon 1997 a,b,c) and of cotton cultivation and cooperative-based marketing (Gibbon 1998 a,b). The private cotton chain study was based on interviews with 59 private sector owners and operators carried out in Mwanza, Shinyanga and Mara regions around the southern shore of Lake Victoria - the country’s main cotton-growing area [ Two interviews were also conducted with international brokers in Liverpool, England.] . These interviews were complemented by 36 conducted with government staff at various levels, and by 101 conducted either with cooperative ‘movement’ staff or with peasant cultivators. Interviewees amongst private sector owners and operatives encompassed proprietors and managers of private ginneries, field supervisors and buying post operators for a variety of private buying companies and managers of Tanzanian textile companies. With a few exceptions, interviews followed the lines of individual or corporate oral ‘business histories’, using check-lists inspired by that found in the appendix to Harriss-White (1996) and revised after experience in the earlier fieldwork referred to. A number of different check-lists were used, whose design was tailored to different roles in the chain. Interviewees were chosen on the basis of seeking to interview all proprietors or managers of privately owned, joint-venture operated or privately-leased ginneries in the larger towns of the Lake zone and rural parts of five administrative districts within it [ Magu and Misungwi districts in Mwanza region, Shinyanga rural and Maswa district in Shinyanga region and Bunda district in Mara region. The research assistants for the study were Mr H Ndunguru and Mr S Makipesile (Mwanza), Ms P Maganga and Ms A Mlele (Shinyanga) and Mr A Masawa (Mara).] , field supervisors from the major private buying companies, and an operative from every private (and cooperative) buying post within 11 villages situated within the five districts. Interviews with proprietors and managers were all held in English; those with other categories were usually held in Kiswahili with questions and answers translated back and forth from and into English. Interviewees were told that the study was an academic one and that the author was based at the University of Dar es Salaam [ During the fisheries studies interviewees were informed that they were being interviewed as part of a background study of private sector development for DANIDA. The form of introduction in this study was changed precisely to see if this may have biased respondents ’ relations to the author. Surprisingly, if it did bias them it did so only in the same way as when I introduced the study as an academic one, for I noticed remarkably little difference in respondents ’ reactions to these two different forms of introduction.] . The paper opens with a sketch of the international, national historical and local socio-economic backdrops to the private cotton chain in Tanzania. This is followed by a discussion of its macro-economic significance. Subsequently the paper turns to a description of the division of labour within the chain and of some of the commodities intersecting with it locally as related inputs, services and by-products. Next comes a discussion of the social character of the actors and actresses in the chain and the structures through which they interact. This is supplemented by information on earnings and profits in some nodes of the chain. In conclusion the paper turns to certain general questions concerning market development and regulation.
2. The international cotton trade During the mid-1990s around 15m ton of cotton lint was being traded internationally each year. Most of this volume was actually cotton which was being re-traded, for primary exports by the 70-odd producing countries themselves stood at only about 6.5m tons. The latter volumes had themselves been rising steeply, from figures of 3.9m tons in 1975 and 4.2m tons in 1985 (Heijbroek and Husken, 1995). The leading world producer of cotton lint is China and - highly unusually for a primary commodity- the leading exporter is the US. Both of these trends have become more marked since the sharp decline of production in the former Soviet Union, although Uzbekistan remains the second largest exporter (followed by CFA-zone African countries considered as a group) The leading importing countries are in Asia, which over the period 1975 to 1995 increased its share of world textile production from 40% to 65%. Within Asia itself the most important importers are becoming Indonesia and Thailand (Heijbroek and Husken, op.cit.). Although demand for textiles has steadily increased over the last twenty years, and while cotton textile’s share of all textile production has stabilised at just under half, the international cotton market is nevertheless not particularly stable. This lack of stability is induced by large and unpredictable fluctuations in the Chinese, Indian and Pakistani harvests [ China normally imports around 0.8m tonnes lint/year, but domestic production has been increasing and domestic consumption falling (especially in the wake of large devaluations by its south-east Asian competitors in the wake of the Asian financial crisis). In 1998 China announced plans to export 0.3m tonnes; the prices at which the first 0.2m tonnes were sold at gave rise to claims of dumping by US growers ’ groups. The same year, India announced an intention to import 90,000 tonnes as a result of late rains; normally it imports only about 5,000.] . In years where these harvests are good there is a glut of cotton on the world market and prices fall; in years where they are bad there is a shortage and prices may rise steeply. Overall however, exports have increased faster than import demand over the last decade, as more and more consuming countries have promoted domestic production for their own mills. Consequently, there has been a real fall in the world price, from an average of US$ 1542/ton (f.o.b.) in 1983-85 to US$ 1472/ton (f.o.b.) in 1993-95 (current prices, computed from FAO Trade Yearbooks). Neither of these prices represents an improvement on the current f.o.b. price of the 1960s. Cotton is probably the most globalised of all primary commodity trades. On the one hand almost all tropical and many temperate countries export the crop, while on the other a large majority of all countries have a textile industry at least partly dependent on imports. Probably because of this fragmentation of supply and demand, control over this trade is far less concentrated than for most primary commodities. Moreover, it is primarily located not at the processing or distribution stage (as with cocoa and coffee) but at the international brokerage one. Brokers normally ‘accumulate cotton from different sources to achieve the (grades) desired, bridging the differences in quality, timing, place and volume from the different sources and for the specific demands of the different textile industries (Heijbroek and Husken, 1995: 14). This function has become increasingly important as the leading producers have sought to introduce shortened lead times and desires to reduce stocks, and this in turn has been one factor contributing to the evolution of a futures market as brokers seek to lay off or hedge risk [ Futures trading is the trading of options to buy specific volumes of specific crops which will appear on the market at specific times, as opposed to physically buying cotton itself.] . Cotton grading is based on staple length and tensility, cleanliness, colour, residual trash content and gin preparation. Based on a matrix generated by combining these axes, an intricate international pricing system has been generated. This is complicated further by seasonal premiums and discounts. Traditionally, cotton has been bought and sold on the basis of bale-by-bale specification, although the larger international traders are said to be striving to overcome tradition and convince some buyers to purchase large amounts on an undifferentiated basis within a given range (Kneen, 1995). Heijbroek and Husken (op.cit., 49) list the top 29 international cotton brokers. There are four important groups: US-based, Europe- (mainly Liverpool-) based, Japanese and Australian. The top 10 largest brokers control 30-40% of all trade; of these six are US-based, two European and two jointly based in the US and Europe. The two largest are Dunavant, a privately-owned company based in Memphis, and Hohenberg/Ralli, an alliance between the US and British subsidiaries of Cargill - again a US-owned private company, but this time one which is amongst the world’s two or three largest agri-business multinationals. 3. The cotton trade in Tanzania Cotton was grown in what is now Tanzania prior to colonialism, but the now universal American Upland variety was only introduced by the Germans in the late 1890s. On the coast the Germans used physical force to promote production but, in what was to become the main growing area around Lake Victoria, fixed prices and incentives to local leaders were employed instead. The first Lake zone ginnery was built in 1906 on Ukerewe. During the first phase of British rule (1918-49) the main emphases of government ‘development policy’ were food self-sufficiency and settler (or government-scheme based) exportenclaves. Cotton was never a settler or a government crop, and (lint) cotton production never exceeded 8,000 tons/year. After 1949 policy turned to the promotion of peasant export crop production (cf., Bryceson, 1990: 124). The Department of Agriculture developed a large field staff [ Staff numbers increased from 875 in 1950 to 2231 in 1959-60 (World Bank, 1961: 104)] and a number of large-scale agricultural development projects were launched. Easily the biggest of these was the ‘Sukumaland Development Scheme’, located in the Lake zone, under which large tracts of bush were cleared and water points established in what were to become the major cotton-growing districts of Geita, Maswa (including Meatu) and Biharamulo. Cotton production expanded during the 1950s under the impact of these measures, coupled with rising local producer prices (reflecting a virtual doubling of the international price of cotton during the Korean War), a resumption of physical compulsion, in the form of bye-laws specifying minimum cotton acreages (cf Magoti, 1984, 64-65 and Guyashi, 1988, 40-54 on Maswa and Mara respectively), and the government-led spread of agricultural marketing cooperatives (see Gibbon, 1998b). Production increased to an average of 29,000 tons/year (lint) in the second half of the 1950s and 41,000 tons/year in the first half of the 1960s. Evidence emerged of peasant investment in improved means of production and of very high levels of commercialisation. [ According to Fuggles-Couchman (1964, 25), writing in 1960, ‘the use of the ox-plough in Lake, Northern and Western Regions has increased enormously since 1945, leading to larger acreages being cultivated...In 1960 traders in Shinyanga district sold 4,000 ox-ploughs. In north Mara over 15,000 ox-ploughs are now in use and in Nzega 4,000 are employed. ’ Ruthenberg (1964: 329) estimated that, shortly after independence, amongst cotton cultivators in Shinyanga region sales accounted for up to 69 percent of total agricultural output.] During this period around 90% of all national seed cotton production became concentrated in the Lake zone, a pattern which has remained. At first, cooperatives competed with private buying agents to purchase seed cotton from peasant cultivators. However, both paid the same price fixed by the Cotton Board. Likewise, from the late 1950s cooperatively-owned ginneries began to compete with existing privately-owned ones but also in a context where all exports were conducted through a parastatal agency buying lint from ginneries at fixed prices. Private sector participation was overwhelmingly (80% in the 1940s, according to Honey (1982: 500)) Asian. The largest Asian cotton proprietor was Vilhaldas Haridas, an Indian based in Uganda, who at one time owned 19 ginneries and two oil mills around the Lake (Honey, 1979). Asian private agents were legislatively excluded from seed cotton purchase in 1959 and from ginning in 1964, when cooperative monopolies were enforced as part of a drive by the last colonial, and later the first independent, government to promote agricultural ‘modernisation’ through this vehicle. Nationally, Primary Cooperative Societies increased in number from 457 in 1961 to 1533 in 1966 (Saul, 1974) and now as a rule became involved in the supply of integrated input packages. In parts of the country and for crops where there was no history of cooperatives, their forced-pace spread led to chaotic results, but in the Lake zone the lintproduction trend continued to rise up to and including the first half of the 1970s, when it reached an average of 71,200 tons/year. The second half of the 1970s saw the beginning of a serious decline in cotton production, which was to last until the early 1990s. This decline almost certainly had its initial roots in the linked phenomena of villagisation and dissolution/reconstruction of cooperatives along new lines. Villagisation entailed the forcible change from a ‘homestead’-based rural social and economic geography to a nucleated village-based one, with attendant disruption of production and local agro-ecological irrationalities [ Increased walking time to fields, increased local competition for both agricultural and grazing land, etc.] . At village level, changes to cooperatives entailed a replacement of single-function, member-controlled agricultural marketing societies by multi-purpose, village government controlled and compulsory membership organisations. Cooperative Unions were meanwhile initially (1973) sidelined by, then later (in 1976) formally abolished in favour of, Crop Authorities with monopoly input supply - and later marketing - functions (the export function remained with the Cotton Board). By 1981-85 production had fallen to an average of 46,800 tons/year (lint). Also contributing to this collapse was a fall in producer prices, starting according to Lele, van der Walle and Gbetibouo (1989) in 1970. At least after 1980, this was associated with a long-term decline in the prices (and returns to labour) of cotton relative to maize and rice- two crops grown in the Lake zone for both consumption and sale, giving rise to crop switching [ Closer to Mwanza town also to horticultural crops and grams .] (Dercon, 1993; URT/QEH, 1994). After it became clear that there was widespread political as well ‘economic’ opposition to the Crop Authorities, Regional Cooperative Unions were revived from above in 1984. The circumstances of their rebirth were unfavourable. The country was experiencing an acute economic crisis associated with severe shortages of foreign exchange and domestic credit, the processing and transport infrastructure inherited by the RCUs had become badly run down, and the new institutions suffered badly from corruption and lack of financial control. By the beginning of the 1990s, most of the major cotton-purchasing RCUs were technically insolvent and a pattern had become entrenched of buying seed cotton only on credit. Donor interest in supporting the cooperative sector had meanwhile expired, while from a different angle the Tanzanian political liberalisation process which was set in motion in 1991 had begun to lead to a partial separation of the state from so-called ‘mass organisations’. These factors provided a background to the cooperative reforms of 1991-92, under which agricultural marketing cooperatives were legally reconstituted as one functional type of cooperative amongst others, Primary Societies and Cooperative Unions were reconstituted on a voluntary basis and Unions’ balance sheets were ‘cleaned up’ to allow them to become viable self-financing entities. Restruc turing of cooperatives long these lines was not immediate, but gathered pace from around 1994-95. A comprehensive liberalisation of the Tanzanian cotton sector began to be implemented from the same time. Donors, led by the World Bank, had been pressing for change since the late 1980s and a few limited reforms were undertaken as early as 1990-91 [ RCUs became economically responsible for export of the crop, although in practice all continued to use the Cotton Board as an agent in this respect until 1994-95. A new category of joint venture ginneries independent of the RCUs was also designated, with the target that these should gin 50 percent of the crop (in fact only one or two were actually formed, although two others built with Dutch assistance as joint ventures were eventually opened as solely private enterprises). In 1992-93 the government-set producer price was given an ‘indicative ’ status and in 1993-94 RCUs were allowed to set their own producer prices (Undolle, 1997: 16).] . The Cotton Regulations of 1995 (URT, 1995) specified that henceforth anyone could open a cotton buying post provided that they bought a license and paid taxes, paid sellers in cash, undertook cotton grading, operated with inspected scales and displayed price lists. Anyone could also open a ginnery provided that they bought a license and paid taxes, agreed to certain reporting procedures and reserved 15% of seed collected for redistribution to cultivators. Anyone, including ginners, could moreover export cotton provided that they bought a license, paid taxes and followed certain reporting procedures. Price controls at all levels were abolished, and the Cotton Board would start its own purchasing arm to act as a buyer of last resort in areas where there was a shortage of private buyers. The Regulations also specified the introduction in 1996-97 season of an export procedure incorporating a Central Tender System supervised by the Board, ostensibly to ‘improve transparency’, i.e. to prevent under-invoicing (URT, 1995, op.cit.); the system’s implication was that sale by private treaty would be outlawed. The first generation of private ginners objected strenuously to this proposal on the grounds that it would prevent them raising importer finance for crop purchase and that it would introduce delays to exporting which would disqualify Tanzanian cotton from the ‘market-window’ premium accruing to cotton exported between July and August (see below). After a meeting between private ginners and ‘government officials’ in May 1996 the Central Tender System plan was dropped and has not been resurrected [ In 1995 the Tanzanian Parliament also appears to have passed resolutions concerning the regulation of the agricultural export trade, under which all exporters ’ sales revenues should be paid into forex bank accounts held in Tanzania itself, and that the producer price should never fall below 60 percent of the export price. It is not clear whether these resolutions were ever gazetted .] . The first private four ginneries were opened in 1994, and another four opened the next season as it became clear that RCUs did not welcome the prospect of contract ginning on behalf of private exporters. By the beginning of the 1997-98 buying season in June, 1997 there were a total of 14 working ginneries under exclusively private ownership, four which were public-private joint ventures [ In addition there was a joint venture between different public organisations (Manawa Ginnery Company, Mwanza region).] and two public ones which were exclusively leased to private buyers. This compared with 15 to 18 functioning cooperatively or publicly owned ginneries. In addition to the private, semi-private and lease holding ginners, there were five to eight licensed independent private buyers selling on to ginners or exporting on their own account after contract ginning. Allthe new private ginneries and most of the other private buyers were working exclusively in the Lake zone. The advent of private buyers ensured that producers could normally obtain cash for the cotton that they grew, and (as a result of competition) that they systematically began to obtain higher - though not spectacularly higher - shares of the world market price than previously. In the five years before liberalisation, the producer price averaged 46.8% of the world market price (itself a major improvement on all but one quinquennial since the early 1970s [ The lowest producer price share of the world market price appears to have been in 1976-77 when it was only 18 percent.] ). In the four seasons since, the average has been 51.3% [ computations based on Gillham et al (1995) and own data.] . There is some indication of a production response, though again not a spectacular one. Officially marketed lint production averaged 55,160 tons/year in the five years prior to liberalisation, while in the four years since it has averaged about 62,500 tons/years. On the other hand, this latter figure is almost certainly depressed by private ginners’ underdeclaration (see below). 4. The local setting The Lake zone (or, officially, the ‘Western’) cotton growing area (WCGA) extends from around Musoma town in Mara region on the south-eastern shore of Lake Victoria to Biharamulo town just inland from the lake’s south-western corner, i.e. across Musoma rural, Bunda, Magu, and Kwimba districts to Mwanza town, and then beyond it westwards to Sengerema, Geita and Biharamulo districts. Inland of the Lake it extends well over 100 kilometres south-east to include Serengeti, Bariadi, Maswa and Meatu districts, well over the same distance due south to include Misungwi and Shinyanga rural districts and parts of Tabora and Singida regions, and the same distance south west to include Kahama and Bukombe districts. The crop from the eastern side of this zone is mostly Type 1 (long staple); staple lengths decline further south and west, where the crop is classified as Types 2 and 3. Apart from within 20-30 km of the Lake shore itself, the Lake zone is basically a flat, treeless semi-arid steppe interrupted by large and spectacular granite outcrops. There are visible signs of soil degradation in most places. Day temperatures range between 15 and 30°C and, by Tanzanian standards, relative humidity is low. Rainfall is bimodal and generally averages 500-700 mm/year. Locally, rains are highly unpredictable. Droughts occur on average two years in ten. Cotton is the main cash crop grown throughout the zone. Maize, rice, sweet potatoes, sorghum, cassava and groundnuts are all also cultivated in significant quantities, but of these only rice is predominantly also a cash crop. In much of the zone, particularly as one moves inland from the Lake (where rainfalls are higher) cotton is the only viable cash crop. The only agriculturalactivity rivalling it in importance is cattle-keeping. In the more arid parts of the zone, most households own at least a few cattle and non-ownership is a sign of poverty. Some herds are extremely large, numbering hundreds. Nearer to the Lake cattle-ownership is less widespread, and owning any cattle at all is an indicator of wealth. Throughout the zone, an important axis of social differentiation is ownership of oxen and ox-ploughs. In the farmer survey conducted in the eastern part of the zone as part of this same research, ox-plough owners possessed an average of 27 acres of land, of which they cultivated 16.5 while hand cultivators possessed an average of 6.1 acres, of which they cultivated 4.5. On average, ox-plough owners sold more than three times more cotton than hand cultivators. But average holding sizes varied considerably from one village to the next as, depending on how closely to each other villages had been sited in the early 1970s, some had large land areas at their disposal and some had much smaller ones. There are few sizeable towns in the zone. The mass of the population lives in small, sprawl ing, low density villages with perhaps a school, a couple of kiosks and a posho mill. The largest building is normally the godown of the Primary Cooperative Society. Homesteads are mud-brick, low, thin, flat-roofed structures. Most are thatched while a few have bati (corrugated iron) roofs. Houses of wealthy villagers are distinguished by being longer than the average and by having stone foundations. Cattle-keeping households normally have their bomas next to the house. In most districts, besides the district headquarters town, there were normally a dozen or so larger villages with populations of 700 households or more. Near to many of these villages there would be a weekly or fortnightly market, and along their main street would be found several general stores, eating places and guest houses. It was here where most private buying posts were usually found, in rented guesthouses or former shops along the main street. Interspersed in this otherwise basically featureless landscape was the occasional ginnery. The Cooperative Union ginneries, mostly built in the 1950s or early 1960s, were normally situated in the larger villages. Ginning was carried out in corrugated iron shed structures usually 10-15m high and 100m long. Normally, 20 or so roller gins would stand on the floor of the shed in two lines at 5 metre intervals. Roller gin technology, based on belt-driven gears, has hardly changed for a century. The gins would be fed with seed cotton by hand from a gallery, running the length of the shed. Separated seed was normally carried away by a conveyor running under the floor, while labourers picked up the ginned cotton and shifted it to a bailing press. Near to the ginning shed would be a series of godowns for both seed and ginned cotton, usually dispersed around a rutted yard littered with broken-down vehicles (and often groups of drivers and mechanics sitting around disembodied gearboxes). The new ginneries were often found in the open countryside; their shimmering new struc tures could often be seen at a great distance. Many of them operated with the less labour-intensive saw ginning technology, which also required bigger ginning sheds to house the machines. There were several concentrations of new ginneries, the most notable being on the border between Bunda district (Mara) and Magu district (Mwanza) and on the eastern outskirts of Mwanza town, where there were three ginneries within a few hundred metres of each other at Nyakato. Besides cotton and rice, the two main cornerstones of the Lake zone economy in the 1990s have been fishing and artisanal mining. Fishing, almost exclusively for Nile Perch and dagaa (freshwater sardine) drew labour from all over the Lake, although in Tanzanian waters physical operations were mainly concentrated in the latter’s south-west corner. When I conducted fieldwork on these fisheries in 1996, there were probably about 3,500 artisanal vessels engaged in the Tanzanian Nile Perch fishery and about 2,000 engaged in the dagaa one. In all there were about 20,000 more or less full-time fishermen. In addition there were perhaps another 7-8,000 persons who were involved on a full-time basis in factory processing (Nile Perch only), artisanal processing and trade. Almost certainly, considerably larger numbers than this were engaged in artisanal mining. Artisanal gold mining is found mostly in Geita and Kahama districts, where according to Chachage (1995) there are somewhere between 100,000 and half a million persons living in mining camps. Artisanal diamond mining was also visible in many areas, including the lengthy perimeter of Anglo-American’s Mwadui holding in Shinyanga Rural and Maswa districts, and in the courtyard of the only hotel in Shinyanga town with piped water in its rooms there were regular informal auctions taking place, with dealers from Europe buying from a procession of local brokers. Interviewing in the villages of the Lake zone, my impression was that youths from nearly every village had been drawn to the goldfields. Fishermen by contrast tended to be drawn almost exclusively from villages close to the shoreline. 5. Cotton and the national economy Together with coffee, cotton has since the 1950s been Tanzania’s main export crop. Production, export and domestic consumption figures are provided in Table 1. As can be seen, there are inconsistent series of figures both for total marketed lint production, exports and domestic consumption. The Cotton Board itself appears to publish separate inconsistent series. In addition, there are inconsistencies between both series of figures for total marketed production when recorded in its own right and the figures obtained by adding up the series for exports and domestic consumption (the sum of the latter two series is almost invariably lower than the former). Notwithstanding these data-related problems, besides the recovery in total production in the 1990s as measured by series (ii) of the total production volume figures, the most noticeable trend concerning volumes has been the decline in domestic consumption. This corresponds to the virtual collapse of the domestic textile industry, which began early in the 1980s but worsened considerably in the early 1990s. Rated capacity fell from about 350m sq m in 1989 to 277m sq m in 1996 (due to closures) while production fell from 93m sq. m in 1980 to 60m in 1983, 71m in 1989 and only 28m in 1996 (de Valk, 1996; TMAT, 1997). The causes of this collapse have been the subject of a bitter debate between local manufacturers, academics and others, which will not be entered into at this point. Production probably fell further in 1997, although additionalclosures were partly offset by the reopening of the country’s largest textile mill (Friendship/Urafiki) under new ownership. Table 1: Production, export and domestic consumption of cotton lint (tons), Tanzania 1946-98
Sources: Production: (i) Gillham et al, 1995 (ii) Min of Agriculture, Crop Development Section (iii) Economist Intelligence Unit Exports: (i) Gillham et al, 1995 (ii) Bank of Tanzania Domestic consumption: (i) Gillham et al, 1995 (ii) Cotton Board Problems concerning consistency of data affect also those on the value of Tanzania’s exports, including its cotton exports. However, a clear basic pattern is evident, whereby cotton forms a declining but still considerable share of Tanzanian exports. As Table 2 (based only on the Economist Intelligence Unit series) shows, revenues from cotton comprised an average of 19.9% of export revenues in the period 1989-93, but in the four subsequent years this fell to 17.4%. Table 2: Cotton exports and total exports (million US$), Tanzania 1990-98
Source: Economist Intelligence Unit, Tanzania Country Reports In the Lake zone there are probably approaching half a million cotton cultivators, the great majority producing less than 0.5 ton seed cotton/year. Total cash income to cultivators is probably somewhere between USD 60 and 90m/year, an annual injection of cash which should be making a profound difference to the region’s economy (but which in drought years like 1997 was probably mostly spent on buying food). The Table also shows that the decline of cotton’s share of export revenues since 1995 basically reflects the latter’s considerable absolute expansion after a long period of stagnation, for cotton revenues have climbed. Direct formal employment in the cotton sector was substantial, but not on a scale equivalent to fisheries. The 900-plus Primary Cooperative Societies in Mara, Mwanza and Shinyanga regions alone had about 2,700 permanent employees, and Cooperative Unions in these regions about a further 1,750 permanent employees (most working at ginneries and oil mills). Private ginners and other private buyers employed probably about 400 permanent staff. In addition to this figure approaching 5,000 persons with permanent employment in the cotton trade, there were about 1,000 seasonal workers in the Regional Cooperative Union ginneries, 2,400 in the private ginneries and about another 2,500 working at private buying posts, giving a total of approaching 6,000 temporary workers in all. While cotton liberalisation has thus seen a considerable expansion in employment in the sector, around 90% of this expansion has been in the form of temporary and casual jobs. 6.The private marketing chain Within Tanzania the division of labour within the cotton chain currently incorporates the functions of cultivation, marketing of seed cotton (cooperative and private), ginning (cooperativeand private), and domestic and export trading of ginned cotton. Marketing (particularly of seed cotton) is differentiated not only between cooperative and private channels, but also to a great extent between different types of operation within these channels. The procedure used here will be to first identify the different private economic functions involved, and then in a subsequent section describe their occupants and their social inter-relations. The cooperative marketing system is described at length in Gibbon (1998b). The cotton chain as described here is also intersected at various points by related chains: those for farm inputs, transport, local storage capacity, ginning machinery and cotton seed by-products such as seed oil and cake. This group of chains will be described in a later section. Cultivation All but a tiny fraction of cotton cultivation is performed and managed by peasants. The exception is a handful of small demonstration plots under the control of Cargill, one of the main private ginners; production on these is not economically significant (so far, at any rate). Contemporary Tanzanian cotton farming systems are described in detail in Gibbon (1998a) and will be only briefly recapitulated here. On the eastern side of the Lake zone cotton growing area where the author conducted a survey in 1997, peasants had just under half of their land under cotton irrespective of whether they cultivated by plough or by hand. This proportion was significantly higher than those reported in surveys carried out in the same area shortly before liberalisation. Cash payments and small in creases in real producer prices appeared to have encouraged a producer response in the second season of liberalisation, although there was no evidence that this had been subsequently ampli fied. Considerable evidence was found of differentiation amongst cotton producers, according to whether cultivators owned or did not own oxen. On average, oxen owners had considerably higher concentrations of assets, cultivated areas under cotton and cotton sales than either ox-hirers or hand cultivators. However, oxen owners were a much more heterogeneous group than the other two categories, and aside from their method of land preparation their farming methods differed little from other categories. The dominant farming methods for all groups closely resembled those described in the farming systems literature for the area ever since the 1960s. Rotation of cotton with other crops was practised by a majority both of ox- and hand-cultivators. Planting mainly occurred in December. Hand cultivators all cultivated by using the highly labour-intensive method of making matuta (high ridges split in two at the beginning of each season and rebuilt in the previous season’s furrow). Ox-cultivators in Mwanza and Mara regions cultivated mainly by planting seed in rows behind the plough; in Shinyanga a majority planted by broadcasting. A large majority of peasants in both categories weeded their crop three times during the course of the season. Hand cultivators predominantly used household labour and labour from communal work groups, oxen owners used household labour and labour hired by the piece. A major change had however occured in input use, even in relation to the situation immediately prior to liberalisation. While chemical fertiliser has hardly been used at all for cotton cultivation in the Lake zone since the 1960s, insecticide has been used very widely. Levels of use had fallen sharply since liberalisation. Recalling their ‘normal’ use of insecticide prior to 1994/5, 73% of respondents sprayed their cotton at least once, with an average number of sprayings of 2.6/year. In 1997/8 season only half of all respondents sprayed their cotton, with an average number of sprayings of 2.2/year. Amongst those using insecticide, volumes used per acre of crop sprayed were also well below recommendations. A comparison of cotton yield data for the Lake zone obtained from surveys conducted in recent years (including my own), with yield data from the late 1960s appears to indicate an approximate halving of yields over the intervening 30 year period [ In the late 1960s cotton yield figures in the range of 660 kg/ha to 798 kg/ha were reported (Meertens, Ndege and Eserinck (1995), Gillham et al (1995), Anthony and Uchenda (1975)). According to my own survey, yields between 1995-96 and 1997-98 season averaged 420 kg/ha.] . This must reflect to some extent a decline in natural soil fertility, for in the late 1960s possibly a majority of the area under cotton had only fairly recently been virgin land. On the other hand, it almost certainly also reflects the gradual unravelling of the enforced system of high- (or at least higher-) non-labour inputs, which was established in the last decade of colonial rule and extended into the first decade or two of independence. By contrast, the deployment of labour inputs appears to have remained fairly constant, or even to have been slightly intensified. Private seed cotton marketing This discussion of the private marketing system takes as its point of departure the private marketing of seed (i.e., unginned) cotton, rather than the more familiar starting point of private ‘primary marketing’. The reason for this is that a considerable part of the private marketing of seed cotton incorporates primary marketing agencies which are formally public in character. The variety of private seed cotton purchasing systems, and the tendency of many purchasers to straddle more than one system, is illustrated in Table 3. This summarises information on 21 of the 24 private buyers operating in the WCGA in 1997-98. The most common methods for buying seed cotton were buying through own buying posts (own-organised or which had been organised by village government), buying from commissioned agents, and buying through commissioned District Cooperative Unions. This does not mean that a majority of privately procured seed cotton was obtained by these methods however, for Table 3 provides no indication of either the extent of individual companies’ buying operations or the volumes that they procured through specific channels. The different methods will be discussed in turn, in the order they are listed in Table 3. The section concludes with a discussion of illicit purchase of seed cotton. Table 3: Private seed cotton marketing systems, Tanzania WCGA, 1997-98 season
Key: Own bps (i): buying posts organised with own employees Own bps (ii): buying posts organised through village governments Cmsd agents: commissioned agents Cmsd PSs: commissioned primary cooperative societies Rntg f. PS: renting premises from primary cooperative societies DCUs: commissioned District Cooperative Unions PCs: commissioned private company Source: own interviews Purchasing through buying posts with own employees: This method involved companies setting up a network of buying posts in rented guest houses or stores on the main streets of larger (and some also smaller) villages, after first conducting local negotiations over payment of levy to village government. Company proprietors or their agents made sure that the buying posts were equipped with weighing scales and stationery and found salaried cashiers and weighing clerks, watchmen and sometimes others to staff them. Cashiers and weighing clerks were almost always selected centrally and brought in from outside the villages in which they worked, although watchmen were almost always local. Additional temporary local labour for loading was hired on a casual basis. Buying posts were normally supplied with purchasing capital by supervisors once or twice a week, who also visited them every couple of days to reconcile accounts and check purchased volumes, and who organised transport between the buying post and the ginnery. All of the five companies using this buying method in 1997-98 were ginners who had been operating already for at least a couple of seasons. One, who had only a few buying posts organised anyway, ceased buying cotton after only a few days when its owner realised that prices were rising out of its reach; they spent the rest of the season contract ginning. In most cases this method of purchasing appeared to be considered a somewhat expensive source of obtaining additionalvolume. Generally it was thought that only where high volumes could be guaranteed were the heavy overheads entailed by this system justified. Three of the four companies using it for the full season had no more than 50 buying posts each organised on this basis, mostly found in areas where production was very high. Only for one company was this method of obtaining seed cotton the exclusive or even the main one used by the buyer. This company, one of the two largest private ginners in the country, had invested considerable time and effort in designing a unique seed cotton purchasing system based on it. Although this system entailed far higher overheads than those used by others operating own-employee buying posts (see below) its authors claimed that it was nonetheless superior to all others. The company operated around 225 buying posts on its basis. According to the analysis of this company, the real problem faced by most buyers using own-employee buying posts was not high overheads but rather high rates of employee fraud. This fraud took two forms. The first and most obvious was the tendency of cashiers to abscond with cotton purchasing capital. The prevalence of such fraud was freely admitted by most buyers. The owner of the company which closed its buying operation only a few days into the 1997-98 season reported that two seasons before his Operations Manager had been killed in a car crash, leaving a ‘hiatus of control in the field’, in which buying post employees had run away with a total of Tsh 13m (then around US$ 250,000). The owner of another commented that it was ‘a known fact that in Tanzania one cannot trust any employee with Tsh 100,000 (US$ 160) and my own experience confirms this’. Probably much more debilitating than employee abscondment or direct theft of purchasing capital was what is know in the trade as ‘shrinkage’. ‘Shrinkage’ is the private sector version of hewa la pamba, ‘pioneered’ for years in the cooperative sector. It comprises the systematic remittance to suppliers of purchasing capital of volumes of purchased seed cotton lower than those recorded as having been purchased. Shrinkage is normally brought about by a combination of means including adulteration of cotton with foreign material (sand, stones), inclusion of moistened cotton and the more or less skilful falsification of purchasing records. In instances where grading occurs, it also includes deliberate misgrading. Apart from the company claiming to have mastered the own-employee buying post system, all the other buyers using this method admitted shrinkages in the range of 5-15% of the value of purchasing capital advanced. Purchasing through buying posts operated by village government-appointed workers: This seed cotton purchasing system worked much as that just discussed, except that village govern ments had been asked by the companies concerned both to select and to act as guarantors for buying post employees. In addition, most had been asked to find buying post premises. The thinking behind this was to pass onto village government the costs of any employee abscondment with purchasing capital. All the seven companies using this buying system were ginners. For four of them, 1997-98 was their first season and for two others it was only their second. The owners of five of these six companies referred to their lack of substantive previous experience with purchasing cotton ontheir own account as a (if not the prime) reason for choosing this method. For a majority it was the only purchasing method employed. While none of these companies had suffered employee abscondment with purchasing capital, all without exception suffered from the problem of shrinkage, which was at the same level as that reported for most of the first group. Besides this, the proprietors of several companies complained that their dependence on village governments brought costs as well as benefits. Although the experience was generally reckoned to be a positive one, the rents they were being charged to hire premises seemed to be generally higher than market ones, levies sometimes became inflated, village governments often selected employees ‘we would not have chosen ourselves’, and ‘we often found ourselves being asked for favours which others were not required to supply’. On the other hand, enjoying a good relation with village governments could help protect buyers from having to implement parts of the Cotton Regulations (see below). Purchasing through commissioned agents: This purchasing system involved buyers identifying suitable local independent agents who would buy on their behalf, take care of most practical matters and deliver to the ginnery gate, either against a guaranteed ‘special price’ or against a fixed commission. Several different forms of this system could be identified. Firstly, some independent agents worked on their own capital, others on purchasing capital supplied by the ultimate buyer, and some on a combination (or sequence) of the two. Secondly, the range of the practical matters contracted out to them varied considerably, most notably with regard to issues like appointment of cashiers and organisation of transport. Thirdly, commissions themselves and the forms in which they were paid also varied. Five of the seven companies using this purchasing method were ginners. Of the other two, one was reselling seed cotton to other buyers and the other was exporting after contract ginning. More importantly, companies utilising this method for purchasing seed cotton generally displayed a greater familiarity either with Tanzanian market conditions or with commodity trading generally (or both) than others. Three of the companies using this method were either owned or co-owned by Africans (half of the six in all with this ownership characteristic). All except one of the seven could be described as having considerable experience in commodity trading, either in Tanzania or internationally. Commissioned agents were normally either already known to those commissioning them or recruited on the basis of inquiries and/or recommendations. Usually they were residents of the larger villages in the main producing areas. A common ‘qualification’ was that they had to own property and have some capital of their own. Property might take the form of a registered land title, a substantial house, a lorry or a shop or business. Two companies required that they had guarantors. Some companies also expected them to have some experience of the cotton crop. Commissions varied according to how risk was shared, and to what range of tasks were contracted out to commission agents. At one end of the spectrum were companies which provided agents only with licenses to buy cotton and with a guideline buying price. Agents worked on their own capital, organised a number of their own buying posts determined by this level, organisedand their own transport and were paid a commission of around Tsh 50/kg (US$0.08) for seed cotton delivered to the ginnery gate. At the other end of the spectrum was a Singapore-based general commodity trader who decided where buying posts should be located, obtained licenses for them, conducted negotiations with village governments, set prices and weekly buying targets, provided most, if not all of the capital which agents were to use, organised transport, and provided and paid its own cashiers for each post. Agents working on this basis were normally asked to coordinate only around five buying posts, and were receiving more like Tsh 20/kg (US$ 0.03) for seed cotton delivered to the ginnery gate. At this latter end of the spectrum, the agent had little personal choice over what kind of buying system he or she should operate and the only real variation between agents consisted in how much of their own capital they added to the company’s. At the other end however, agents had the freedom to decide how they would collect seed cotton. In practice, most subcontracted direct purchasing to sub-agents. In such cases however, sub-agents rarely worked on their own capital (which typically was not large enough to establish a serious local presence) or organised their own transport. Markets in Tanzania are characterised, inter alia, by very low levels of integration. Prices for the same commodity vary enormously from one part of the country to another, with only an indirect relation to transport costs. Consequently, a standard profit-maximising strategy amongst traders is arbitrating, i.e., ‘leapfrogging’ existing nodes (wholesaling) or sites (wholesale markets) of integration in order to establish a personal presence at both points of production and consump tion. Although seed cotton markets are amongst the most highly integrated in the country, this phenomenon also emerged in them. Towards the end of the 1997-98 buying system a procession of sub-agents buying in areas up to 100 km from the Mara-Mwanza regional border had begun appearing at private ginneries in this area with hired lorries loaded with cotton, in order to leapfrog their own patrons and sell direct at ginnery gates. Conversely, one of the ginners they were selling to was beginning to plan how these machinga [ Machinga is the collective noun applied in Dar es Salaam to the lowest stratum of street hawkers (i.e., the group working completely on the capital of others). It is said to derive from a village in a ‘backward ’ part of southern Tanzania, from whom the first generation of these hawkers (who appeared in Dar es Salaam only in the early 1990s) supposedly hailed.] , as he derisively referred to them, could be directly (rather than only indirectly) integrated into his buying systems for the following season. Most of the companies using commissioned agents were not at all anxious to deal with machinga however. I was frequently told that ‘the whole point of having a good agent is that you can trust him (sic [ See below on agents and gender.] ) with Tsh 5m’ (US$ 8,000) [ Another mentioned Tsh 7m (US$ 11,200) .] . Only very rarely would a carefully-selected agent abscond, whereas ‘an employee will generally run away at the first opportunity’. Secondly,all those using both agents and other methods of seed cotton purchase (or who had used other methods of cotton purchase earlier) expressed the conviction that use of agents was the most reliable method for combining volume with low overheads: ‘these guys can find cotton where nobody else can. They can price discriminate and they have high manoeuvrability’. Purchasing through commissioned primary cooperative societies: The methods of cotton seed purchase described so far were most common in Mwanza and Mara regions, where a number of private ginners and buyers had been operating for the whole period since 1994-95 and where there was still the largest concentration of them in 1997-98. Another distinguishing feature of the situation in Mwanza region was that in most villages there was only one cotton godown (warehouse), almost invariably belonging to the village primary cooperative society. These godowns had been built by the Regional Cooperative Union (RCU), and primary cooperative society membership of the RCU was generally thought to imply an obligation on the society to deliver seed cotton exclusively to its RCU. In Shinyanga the situation was different in important respects. In the first place, there were fewer private buyers and ginners. Secondly, as will be touched on later, there was a degree of collusion over price between buyers. Thirdly, a very large number of village primary cooperative societies owned more than one godown- in some cases up to five - as a result of the Shinyanga RCU (Shirecu)’s earlier highly extravagant building programmes. In this context, sometime in the recent past the Regional Commissioner for Shinyanga had held a meeting with the region’s three largest ginners at which he had requested them to buy through the ‘underutilised’ infrastructure of the Shirecu-affiliated primary societies. Whether Shirecu was consulted is unclear. When I tried to discuss this with their headquarters staff I received no reply, although lower down the management chain at ginnery zone level there was frustration and annoyance. On the other hand, village governments and primary societies were enthusiastic and embraced the opportunity to act as agents for private buyers whenever they were approached. At every ginnery zone headquarters which I visited in Shinyanga I was told that all Shirecu-affiliated societies with more than one godown had reserved one for Shirecu and sub-con tracted or rented out the rest. The same situation prevailed in respect of primary societies in Shin yanga which had left Shirecu to join breakaway district-based cooperative unions (see below). Elements of the same situation prevailed in Mara, although for different reasons. Here the RCU had been liquidated in 1996 due to a huge accumulation of debt. By this time two breakaway Cooperative Unions concerning themselves with cotton had already been formed, based mainly in Bunda and Musoma Rural districts. A third, which endeavoured to cover the whole of Mara region, was formed shortly afterwards. By 1997-98 season most, but by no means all, primary societies were affiliated to one or other of the three new Unions. Many of these societies, like those in Shinyanga, owned more than one godown. Some had decided to dedicate one godown to buying on commission for private buyers, while reserving another for supplying ‘their’ Union. Some of the unaffiliated societies had also decided to buy on commission for private buyers, considering that the latter would be been more reliable payers of levy than thevarious competing Unions. In one case I came across in Mara, a village primary cooperative society was acting as agent for both a private buyer and for a Cooperative Union to which it was not affiliated. Generally however, the phenomenon of primary societies buying on an agency basis seemed rather less widespread in Mara than in Shinyanga. Contracts between private buyers and primary societies had certain basic similarities. Invari ably the primary society would provide not only a godown but also its officers as buying post staff. The private buyer would advance purchasing capital to them and organise transport to the ginnery. The primary society would receive a levy by weight for seed cotton delivered to the gin nery. In Shinyanga and Mara this levy was Tsh 10/kg, the same as being offered by Shirecu and the Mara Unions. In Tabora it was Tsh 12/kg. In most cases it was agreed that the salaries of the buying post staff should be paid out of the levy, but since levy was not paid until the end of the season, the companies usually paid posho (a negotiable living allowance) to those staffing the posts. In a few cases the companies were actually paying one or two salaries per buying post, in addition to the levy - or at least, this was how the primary societies concerned had understood the relation [ Some societies had demanded that companies also directly paid their committeemen (i.e., lay officers who ( inter alia ) undertook occasional supervisory work at buying posts), but as far as is known no company actually agreed to this .] . The phenomenon of hewa la pamba which plagued the Cooperative Unions arose out of arrangements whereby primary societies received advances against their own estimates of what they would collect, and whereby levy was paid on cotton recorded as collected from societies and delivered to ginneries. Typically, discrepancies arose firstly between estimates of cotton collected and cotton actually collected and secondly between cotton said to have been dispatched from the society and cotton actually delivered at the ginnery. Some buyers using primary cooperative societies claimed to have largely avoided hewa la pamba, since the advent of private buyers meant that the social basis for the collusion which the phenomenon rested on had been removed. But others admitted that shrinkage constituted probably a more serious problem than with own-organised buying posts: ‘you can’t stop the secretaries making ghost payments’. The use of primary societies had one clear advantage: it eliminated almost all the costs associated with setting up one’s own buying operation and it assured reasonable political relations with local authorities. But otherwise overheads were the same as for own buying posts and levels of loss were almost certainly higher. Purchasing at rented primary society godowns: This form of seed cotton buying involved private buyers simply renting a village cooperative primary society godown as a buying post, while supplying their own personnel to staff it. It appeared to be confined to Mara region. I personally encountered an instance of it involving an unaffiliated primary society and heard of others. For the private buyer, the method closely resembled buying on an own buying-post basis. The significance of its existence lay primarily in what it implied for the primary societiesconcerned. For all purposes these had ceased to perform any recognisably cooperative functions, nor were they part of any recognisable larger cooperative structure. Ironically, in the instance I encountered, the primary society concerned was renting one of its two godowns to a private company and another to a district-based Cooperative Union. Purchasing through District Cooperative Unions: Between 1995 and 1996 seven new district-based Cooperative Unions (DCUs) were set up in the WCGA, six of which were exclus ively concerned with marketing cotton. Four were set up as breakaways from functioning RCUs; the other two were set up in the wake of the liquidation of Mara Cooperative Union. By mid-1997 these six Unions were claiming a total of 249 affiliated primary societies, as against the 725 claimed by the two remaining large RCUs, Shirecu and Nyanza (Mwanza). Relevant government agencies’ estimates put total number of DCU-affiliated societies at nearer 130. The reasons for these breakaways are examined in detail in Gibbon (1997b) and will not be repeated here. None of the new DCUs had managed to acquire ginneries in the process of breaking away; even in Mara, Mara Cooperative Union’s remaining ginneries were taken over by the Receiver and leased to private buyers rather than being inherited by the DCUs. With neither large numbers of affiliated primary societies nor ginning capacity, nor with agents in Dar es Salaam, the new DCUs were at a considerable disadvantage vis-a-vis the large, well-established RCUs. They could neither add value through own ginning, nor increase their margins through transporting to the port of exit, nor guarantee cotton in sufficient quantities to justify a tendering process for overseas importers (and hence obtain purchasing capital on advantageous terms). They thus had recourse to contracting their buying services to local private exporters, in much the same way as these exporters contracted with RCU primary societies. Private buyers would advance working capital on an interest-free basis against a seasonal, or more likely monthly purchasing target, and set the producer price. The DCU would deliver cotton against this target and at the end of the season receive a levy based on a fixed sum/kg for whatever cotton was actually delivered to the ginnery nominated by the private buyer. Normally the private buyer was responsible for organising transport to the ginnery and (where he or she didn’t own a ginnery) the ginning itself. However, in two cases the DCU was organising the ginning and the private buyer was buying lint rather than seed cotton, on an ex-ginnery spot price basis rather than on a pre-set ex-buying post one. In this case pre-finance was being advanced at a fixed interest rate without the buyer setting the seed cotton producer price. The DCUs were also responsible for organising transport in these cases. In all cases the DCUs also handled payment of all local (as opposed to national) taxes. These were paid to them over and above the levy payments which will be described in a moment. The six DCUs between them had contracts with seven private exporters. Six of these exporters had contracts with only one DCU and one had contracts with four. A majority of the contracts were entered into by exporters who at the time owned no ginnery of their own. The two largest of these were both foreign-owned, one by the leading Swiss-based cotton broker Reinhart and the other by the Singapore-based general commodity trader Olam. Reinhart were at anadvanced stage with plans to open their own ginnery in western Mwanza and Olam were also giving the possibility strong consideration, implying a probable reduction of dependence on the DCUs as sources of cotton in the medium-term. One of the five Tanzanian-owned companies using this buying method was in its first year of operation and also clearly had an intention of building its own ginnery. On their agency sales of seed cotton, the DCUs were obtaining Tsh 8-15/kg in own levy, plus Tsh 8.7-10/kg for the affiliated primary society. Taken together, this was roughly the same as the lower range of commissions paid to independent private agents. Working with DCUs was one of the cheaper alternatives for private buyers, particularly in respect of start-up and main tenance costs. It took little or no time to initiate relations with DCUs, and monitoring operations could be conducted in a simple and concentrated form. Moreover, farming out to them the respon sibility of paying local taxes incurred considerable savings in time and harassment. The disad vantage of working through them lay in DCUs’ own highly restricted coverage and the fact that in Shinyanga and Mwanza the RCUs had seemingly embarked on a war to the death against them. This entailed the RCUs concentrating major efforts on both out-competing and internally under mining the DCUs. As a result, although in 1997-98 season the DCUs referred to obtained a re spectable proportion of the declared WCGA seed cotton crop - 6.2%, as against 42.3% collected by Nyanza and Shirecu and 47.8% collected by private buyers without going through DCUs [ A further 3.7 percent was collected jointly by RCUs and DCUs in the peripheral regions of the WCGA, i.e., Kagera, Tabora and Singida .] , DCUs were in the process of losing affiliated societies and members back to RCUs. Purchasing via private companies: There was only one instance in the WCGA of a private buyer purchasing seed cotton from another private company. The company making the first line purchases was based in Mara and was one of the minority which as yet had no ginnery. Part of the cotton it was purchasing was being contract ginned at local ginneries for export sale, but part was being sold as seed cotton to a foreign-owned company which was leasing a ginnery in Mara. This company’s European parent was also the main customer for the purchasing company’s lint exports. The local branch of the foreign-owned company was advancing local currency for the purchasing company’s seed cotton collection while its parent was advancing purchasing capital in forex to cover both seed cotton purchase and ginning charges. The details of the contract are unknown, but an agent of the local branch of the foreign-owned company complained that its local supplier had ‘eaten’ 20% of the purchasing capital advanced to it. Illicit purchasing of seed cotton: While, as will be seen, the seed cotton buying methods described above did not necessarily entail following the Cotton Regulations in important respects, they all involved licensed purchasing. Alongside licensed purchasing (and licensed purchasers) was another, occasionally overlapping, category of unlicensed purchasing and purchasers. The most important category of unlicensed buying in 1997-98 season was buying which took place between around April and June 23 1997, when the buying season officially opened. Cotton planted in November would often be pickable by late April/early May, although it was still likely to be premature. This did not effect cotton quality (as opposed to quantity) but it did effect seed quality; seed from prematurely picked cotton was not usable for planting or for extracting oil. Supply for pre-season purchases was distress sales by badly-off farmers. In 1997-98 this situation was common in those areas where drought coincided with cotton being the only cash crop. Pre-season purchases were very common indeed in Magu district, Mwanza, where most local officials believed that one quarter to one half of the entire crop had been sold prior to the season. The further one ventured from Magu, the less prevalent the practice seemed to have been. Demand for pre-season purchases came from a mixture of shopkeepers in the larger villages, big farmers and big livestock owners (categories where there was considerable overlap) [ In one case I encountered in Misungwi district the pre-season buyer was not a trader but the Village Executive Officer, using the village ’s lorry ‘and his official power ’.] . Around Kabila village in eastern Magu around seven pre-season buyers were operating, all of whom were farmer-trader-shopkeepers. They were openly buying at their own shops in the village, but also at a series of homesteads away from the village and from lorries parked on the public highway. One, who was also involved in the long-distance grain trade [ When I interviewed him he had just returned from Makumbaku, 2,000 km away, where he had been buying maize.] , had been buying illegally for three seasons. Like the others in the village he was operating mainly on his own capital (this season, Tsh 1.7m (US$ 2,720)), employing a weighing clerk (who was a relative), ‘scouts’ to tour remoter homesteads ‘to tell the peasants to bring their cotton in to me’ and casual labour for loading. In his first year of operation, he resold to the village primary society when the season opened; in 1996-97 he had hired lorries in Magu town and taken the cotton into Mwanza to sell at the gate of the town’s biggest ginnery. This season his volume didn’t justify the hire of a lorry and he had arranged for one of the licensed private buyers in the village to pick up the crop. The village government and a ward officer had been cut into his operation, for Tsh 10,000 (US$ 16) and ‘a tip’ respectively. The pre-season buyers were generally buying at between Tsh 80-100 (US$ 0.13-0.16)/kg and reselling at around Tsh 200 (US$0.32) when the season opened. A few were paying Tsh 130-150/kg. While sales to them clearly fell into the ‘distress’ category, and while in eastern Magu they were nicknamed wagalagaji - a name normally used to refer to capital-poor cattle brokers and carrying connotations of sharp practice [ This arose from the fact that wagalagaji habitually worked on mali kauli , literally ‘goods on word of mouth ’. Mali kauli entails a trader agreeing a minimum price with a seller, taking the goods on credit for an agreed period and then retaining whatever in addition to this sum the goods could finally be sold for. It is extremely common in African- controlled Tanzanian trade.] - there appeared to be little public resentment against pre-season buyers. Indeed, I heard a series of stories (not from buyers themselves) about suchbuyers being arrested but then released after the police were surrounded by peasants swearing that all the cotton in the buyers’ possession came from the buyers’ own plots. Most pre-season buyers stopped buying as soon as the season opened, but a few continued either in licensed or unlicensed guises. In both cases they served as the (official or unofficial) commissioned agents of the two ginneries, located in Mwanza and Musoma towns respectively, who relied most on this method of seed cotton purchase. (The ginnery in Musoma town was widely said to have been buying cotton at the ginnery gate in 1997-98 well before the season opened officially.) Besides buying in periods when licenses were not valid (i.e., before the season) or operating without a license at all, the other most common licensing infraction was buying with a license, but beyond the area which the license covered. A lot of such infractions involved buyers (almost invariably commissioned agents and/or machinga) buying from the back of a lorry, but a few involved them buying from fixed premises. Unlike the first of these infractions, the last two were normally punished. Ginning By 1997 there were 21 privately- or joint venture- owned or leased ginneries in the WCGA, with a total capacity of around 4,500 bales/day. This compared to the 15-18 ginneries in the WCGA under the exclusive control of RCUs, with a total rated capacity of around 3,840 bales/day (but an actual capacity considerably below this). Assuming a conservatively short ginning season length of 138 days, total capacity in the sector had thus reached something like 150,000 tons lint/year, more than double the declared harvest in 1997/98 and at least 50% higher than any harvest which might be expected in the medium term. Still, there were plans for new private ginneries: a further five were actually under construction during the fieldwork period. Table 4 summarises basic information on those already operational at the end of the 1997/98 season. Geographically, the private ginneries were spread more or less evenly between Mara, Shinyanga and Mwanza regions, although private ginning capacity in Shinyanga was around one third higher than in either of the other regions owing to the larger average size of ginneries there. Private ginneries in Mara were on average older than in the other regions, as half of them were former RCU property either leased to private companies or in which the latter were involved in a joint venture capacity. Table 4: Ginning company information, Tanzania WCGA, 1997-98 season
NB: companies N, P and T from Table 3 neither owned nor leased ginneries; company V, which appears in this Table but not Table 3, was not buying in the areas covered by the fieldwork. Key: n/a: not applicable n/k: not known (no interview conducted) Yr coms’d: year plant commissioned Locat’n: location (Ma= Mara, Mw= Mwanza, Sh= Shinyanga, Si= Singida, T= Tabora) O/ship: ownership status (JV = joint venture, public/private or public/public) Plant: (RG= roller gin; SG= saw gin) Capacity: actual (a) or rated (r) bales/day Fixed inv: fixed investment (in US$; r= reported; e=author’s estimate); excludes investment in oil mills and transport except where marked ** W/force:F=full-time, C=casual; *ginnery site only, otherwise not distinguished Activity: OG= ginning own cotton; CG= contract ginning Sources: own interviews supplemented by information from Tanzania Cotton Lint and Seed Board, Mwanza Zonal Office The two companies engaged in leasing arrangements in Mara had entered them for different reasons. One was a foreign company which was ‘testing the water’ in preparation for building and operating its own ginnery, while the other was one of the few Tanzanian African-owned enterprises active in the sector. Prior to liberalisation and in its earlier incarnation as an import-export firm, this company had supplied Mara RCU with new gins for one of its ginneries (actually, there was nothing wrong with the old ones, which were simply placed in storage). When Mara RCU was liquidated still owning the company for the gins, it entered cotton marketing andleased the ginnery from the Receiver in an attempt to recover its losses. This company was paying US$ 36,000/year for a two-year lease. Three of the five joint venture companies involved partnerships between private companies and RCUs, and a fourth had also done so before the RCU in question (Mara) was liquidated. All of the joint ventures excluding the one involving Shirecu (Company C) involved ‘distress sales’ by the RCUs in question. For example, in the case of the Singida venture, a senior RCU official was quoted as saying ‘a retreat (to Company Q) was necessary to avoid auctioning the assets’ (Business Times, 23 June 1996). Even the Shirecu joint venture had an involuntary element, since it arose from a Dutch government cotton sector rehabilitation project to which privatisation conditionalities became attached. The joint venture in Mwanza which involved Nyanza RCU arose from the same project, but became an all-public sector affair [ With the Tanzania Investment Bank and the Cotton Board .] when the prospective private partner withdrew. There are two long-established ginning technologies, roller ginning and saw ginning. In the case of roller ginning, leather-covered rollers revolve in contact with a fixed metal blade. Lint is fed into the junction of the roller and the blade and pulled through the gap, which is too narrow for seed to pass through. The movement is helped by a beater blade. Double roller gins (the predominant type found in Tanzania) have two moving blocks or rollers and reciprocal knives. With saw gins, circular saws revolve on an axle at high speed, tearing lint away from a roll or pile of seed cotton by pulling it through closely spaced ribs. The saws project through bars, which are spaced to prevent defibred seed from going forward with the lint, and lint is gathered off the teeth of the saws (either by blasting or brittle-brushing) before these re-emerge beyond the bars to engage more seed cotton. In both technologies, the seed falls into a collecting box, sack or conveyor and the lint is transported physically or mechanically to a baling press (for details see Prentice, 1972: 123 et seq.). Roller gins are individually very cheap (new, ca. US$ 5,000-20,000, depending on the manufacturer), and comparatively simple to install and operate. Their maintenance and labour requirements are high, but spare parts are generic and easily fabricated locally. They are gentler on cotton fibre and hence more appropriate for ginning extra long staple cotton. Certain longer-staple roller ginned lint (including most of that produced in Mara and eastern Mwanza) commands an export premium of up tp US$ 0.05/lb. The generally-acknowledged disadvantages of roller gins is their low productivity (the 342 roller gins listed in Table 4 had an average unit capacity of only 5.6 bales/day) and their high labour requirements. Saw gins are individually very expensive (new, up to US$ 0.5m). They are comparatively difficult to install and their operation has a higher skill content. They are less maintenance-inten sive but their spares cannot be easily fabricated locally. They tend to disturb the cotton fibre and to break the fibres of longer staples [ At least, if the machines are run at full speed] . Saw ginned cotton commands no premium, but the savingsarise from its lower labour content and saw gins’ higher productivity more than compensates for this. The 33 saw gins listed in Table 4 had an average unit capacity of 79.4 bales/day. Some saw ginners impressed on me a further advantage of their technology in the contem porary Tanzanian context. The relevant aspect of this context was the declining quality of purchased seed cotton (see below). Given the common absence of grading, and consequent common presence of discolouration, foreign matter, etc., the fact that ‘with roller ginning cotton comes out looking much as it goes in’ had turned from an advantage to a disadvantage. Conversely, saw ginning’s tendency to disturb the fibre meant that ‘some things don’t show up as badly as they otherwise might do’. A majority of the new private ginners nonetheless chose roller ginning, probably on the basis of its technical simplicity (most were new to the cotton trade, see below) and its low fixed investment costs. For the seven saw ginners for whom some kind of information on fixed investment costs were available, the latter averaged US$ 1.6 m in all (although with a very high mean deviation). Average fixed investment costs for the nine roller ginners for who information was available was only U$ 0.7m (with a much lower mean deviation). When all private ginners are examined together with respect to capacity, fixed investment, employment, and capacity utilisation, a clear three-way pattern of stratification emerges (Table 5). Table 5: Stratification of Tanzanian private ginneries, 1997-98 season
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
