|
Limping towards a Ditch without a Crutch: The Brave New World of Tanzanian Cotton Marketing Cooperatives
CDR Working Paper Subseries no.iii. 98.18, December 98
Contents
Sources: Registrar of Cooperatives (1991-94); Regional Cooperative Development offices (1997) In a few of the 16 villages visited during the fieldwork in 1997, the PS now covered another village besides that it was located in. Where it had been determined that PSs should be amalgamated, decisions to headquarter a PS in one village rather than another were usually based on which village had the larger number of godowns. It was unclear whether there was much in the way of local consultation in this exercise. Subsequently, where PSs covered more than one village, membership from the village where the PS was not headquartered was usually very low and in one case non-existent. All but one of the PSs surveyed for this study was solely engaged in agricultural marketing. The exception was a PS in a remote part of Shinyanga Rural district, which was running the village’s only posho mill and shop as well as buying cotton. In the larger villages visited therewas usually at least one formally organised group or society (usually a women’s group) pursuing other economic concerns. Some of these called themselves cooperatives, but few or none appeared to be registered as such. Of the three main RCUs in the WCGA in 1991, one (Mara Cooperative Union (MCU)) was liquidated in 1996. In 1992 it had owned a rice mill, 11 retail shops and a bed cover factory, as well as three ginneries and a transport fleet (Coopers and Lybrand, op. cit.). In 1994 MCU sold a majority share in one of these ginneries (Ushashi). The circumstances surrounding this sale were highly controversial and will be returned to. Otherwise however, MCU appears to have resisted efforts at ‘downsizing’ prior to its demise. In 1992 Nyanza Cooperative Union (NCU), the regional union covering Mwanza, owned 4 oil mills, 2 rice mills, 29 retail shops, 3 hostels, a bag factory, an ox-cart factory and 4 farms, besides operating 9 solely-owned and one jointly-owned [ Manawa ginnery, jointly owned with the Cotton Board.] ginnery plus a transport fleet (Coopers and Lybrand, ibid.). By 1997 a new oil mill was under construction, but three of the solely owned ginneries were mothballed. Shinyanga Regional Cooperative Union (Shirecu) owned 4 oil mills, a rice mill, 12 retail shops and a maize farm in 1992, besides its 6 ginneries and transport fleet (Coopers and Lybrand, ibid.). By 1997 the retail shops had been closed, but little else had changed, except that in 1994 Shirecu had started a seventh ginnery as a minority partner in a joint venture [ Mwanhuzi ginnery, jointly owned with Farai Ltd and the Cotton Board.]
b. PS membership and the principles guiding it According to the 1991 Cooperative Act, membership of PSs should cease to be mandatory and free of charge, and instead be voluntary and confined to those buying a share (normally costing Tsh 500 (around US$1)). This returned the principle governing PS membership to its pre-1970s content. At that time, around 60 percent of WCGA cotton farmers were members of PSs (McLoughlan, op.cit) [ Interestingly, McLoughlan claimed that this number would have been much higher but for restrictions on entry placed by officers wanting higher shares of society earnings.] . Officials at the Department of Cooperatives headquarters in Dar es Salaam guessed that, in 1997, the proportion of cotton farmers who had taken up membership of the newly-voluntarised PSs was likely to be less than 10 percent. The figures held regionally were widely held to be unreliable, and no surveys had been carried out to establish real membership levels. In 11 of the 16 villages visited it was possible to obtain apparently accurate figures of both PS shareholders and numbers of (nucleated) households (kaya). Almost all the households in these villages grew cotton. In all there were 7,117 households and 1,901 PS shareholders (assuming only one household member was likely to join a PS, this corresponded to 26.7 percentof all households). In the 3 villages with records in Mara, ‘representatives’ of 36.3 percent of households were members. In the 5 villages in Mwanza the proportion was 34.6 percent, and in the 3 villages in Shinyanga 15.2 percent. Membership varied between 3.5 percent and 56.6 percent of households, depending on the village. A membership of 26.7 percent of households corresponds to one of approximately 10.7 percent of all adults. [ TADREG (1997) found PS membership levels of 26 percent of all Tanzanian cash crop growers in a national survey in 1997. However, it found much higher membership levels amongst cotton growers than those reported here, namely 59 percent. This may reflect confusion amongst respondents to the TADREG survey concerning what membership entailed. In a pilot exercise for my own survey, respondents were simply asked whether they were ‘members ’ of a PS. Almost all replied positively. When the question was rephrased to whether they were shareholders (the current legal requirement for membership), the results above were generated instead.] Who were these members ? A survey of 67 cotton farmers conducted in these same 11 villages at the same time as the cooperative study suggests a resurrection of the membership pattern described by Migot-Adholla for the period prior to mandatory membership (Table 2). Table 2: Incidence of PS membership by social category
Source: own interviews (1997) The farmer survey conducted in parallel to this study incorporated a stratified sample of owners of oxen, oxen hirers and hand cultivators. Almost every study of farming systems in the WCGA over the last thirty years has identified ownership of oxen (and its usual covariant, ox-ploughs) as the most important single indicator of local rural wealth and status. The farmer survey confirmed this pattern: owners of oxen typically cultivated 3-4 times more land, had six times more uncultivated land, and sold 3-4 times more cotton annually than hand cultivators. This category was more than twice as likely to be members of PSs than were hand cultivators [ The high level of PS membership amongst the rural population as a whole implied in Table 2 is a reflection of the deliberate over-representation of oxen owners and oxen hirers in the sample.] The reasons why individuals (particularly better-off peasants) joined PSs will be returned to below, but can at this point be summarised as a mixture in gaining access to certain resources and a vague feeling that PSs might offer some collective protection against abuses by private traders of the position of power which market liberalisation was thought likely to give them. Less well-off peasants traditionally always occupied a more marginal relation to PSs, and in many cases had been both subjected to private exploitation by PS leaders and discriminated against in thedistribution of subsidies to productivity gains. In general this group was less fearful of private traders than better-off peasants.
c. Affiliation of PSs to Unions and formation of new Unions In the post-1991 situation the possibility has emerged for PSs to affiliate to any CU they wish, or to none. Of the 8 villages visited in all in Mwanza region, all PSs were affiliated to NCU. Of the 5 villages visited in all in Shinyanga, PSs in 4 villages were affiliated to Shirecu. In the remaining village the affiliation was also to Shirecu, but this was disputed by a group of villagers who in effect operated as a second PS, affiliated to a new CU. In the 3 villages visited in Mara, the PS was not affiliated to any CU in 2. In the third, affiliation was disputed between 2 new CUs and in practice there were again two PSs. Each of the cases of disputed affiliation and non-affiliation will be briefly outlined. Both the Mara cases of PS non-affiliation had arisen in the wake of the liquidation of MCU in 1996. MCU was succeeded by four new CUs, each predominantly based on one of Mara region’s four districts. In the district where these two villages were both located (Bunda) two of the new CUs were contending for PSs’ affiliations. In one village the PS leaders considered that neither of the new CUs had a sufficiently good record of paying levy to PSs to justify affiliation: ‘that would be the main basis of the choice. We will see what happens in other villages. There is no rush’. In the other, located in the far west of the district, the PS’s leaders considered that neither of the two new CUs was sufficiently ‘local’. Unlike Shinyanga and Mwanza, Mara is highly ethnically segmented and the leaders of this PS were waiting for fellow-tribesmen to form yet another CU [ To be called Namusuna.] , ‘which would represent the interests of western Bunda’. Preparations for its formation were apparently already in train. One of these PSs was buying on an agency basis on behalf of a private buyer while the other had simply rented out its godowns to private buyers for the whole season. Levels of PS non-affiliation were quite high in Bunda district (a commonly-quoted figure was that 22 of the 46 PSs there were not affiliated to any CU). Elsewhere in Mara the proportion was probably smaller, but still significant. In Mwanza and Shinyanga I heard of no cases of non-affiliation. The village in Shinyanga where the PS’s affiliation was disputed was the epicentre of a new CU basing itself on Meatu and Mawsa districts (Meatu and Maswa Cooperative Union (MEMACU)). The PS joined MEMACU when the latter was founded in 1996, but at its 1997 annual general meeting reaffiliated to Shirecu (see below). The PS leadership was still split, however. Shirecu claimed 288 members in the village and MEMACU 258, mostly but not exclusively the same people. The local Shirecu leadership had taken control of three of the village’s godowns while the MEMACU leadership controlled the other two. Villagers were free to decided which organisation they should sell to. The village in Mara where the PS’s affiliation was disputed was the epicentre of a new CU basing itself on Bunda district generally, which had sprung up as a breakaway from Mara CU in 1995. A number of Mara CU loyalists in the village never accepted this decision. After the demise of Mara CU they demanded affiliation to another new CU, Mara Farmers’ Cooperative Union (MAFACU), whose leadership and orientation resembled Mara CU’s. The two organisations claimed 68 and 50 members in the village respectively, Once again these were mostly, but not entirely, the same people. By 1997 disputed PS affiliation was quite a common phenomenon in several parts of the WCGA. In Mara Region there were a number of such cases in Musoma Rural district as well as in Bunda. In Mwanza region, there were several such cases in Geita and Sengerema districts, where the breakaway Mweli Farmers’ CU was based, and in Shinyanga region there were cases in Kahama, Bukombe, Maswa and Meatu districts. Disputed affiliation was made more likely because of the nature of most PS constitutions. These allowed affiliation to a CU on the basis of a vote by a simple majority of members at a quorate general meeting, but disaffiliation only on the basis of a two-thirds majority. By 1997 there were no fewer than nine CUs operating in Mara, Mwanza and Shinyanga regions, of which 2 were formed in 1984 and the rest in 1995-96 (see Table 3). Mostly the 7 new ones had been started up by groups of 5-10 PSs from specific districts applying for and gaining registration (or provisional registration, at any rate) from the Registrar. Table 3: CUs in Mara, Mwanza and Shinyanga regions, 1997
Key: (1) Regional Cooperative Department estimates (2) CU’s own figures Sources: Regional Cooperative Departments, own interviews The grounds (and strategies) for forming the new CUs will be discussed in a subsequent section. All claimed to have expanded steadily since their formation, both within their ‘home’ district and then into neighbouring ones.
d. Autonomy and accountability within CUs From the 1970s onward, critics of cooperatives in Tanzania spanning almost the whole spectrum of organisational life outside the ruling party claimed that these organisations were simply state agencies. Unions lacked internal democracy, while at the same time they treated PSs simply as branches, through which instructions were relayed to the peasantry [ Amongst other places, this critique is expressed in the manifesto of one of the breakaway CUs in Shinyanga (MEMACU, 1997).] . But while there was indeed no internal democracy in the CUs and while their popular functions were indeed in abeyance, it is less clear that PSs ever acted as mere ‘relay stations’ for government and CU commands. An essential difference between PSs and local branches of the ruling CCM party, for example, was that PSs’ relations with CUs were inter alia ones of economic exchange. Moreover, the exchanges in question were financed without a great deal control by third parties - namely the National Bank of Commerce and the Treasury. This meant that while PSs were the weaker partner within a commercial relation with CUs, there were still various opportunities for PSs to exploit the element of mutual dependence implied by every exchange relation, as well as incentives for private interests within both CUs and PSs to engage in illicit forms of collaboration in order to divert government money for private purposes. In this context, a significant element of CU-PS relations revolved around summoning -or trying to exorcise- ‘hewa la pamba’ (ghost payments, see above). In the wake of the post-1991 reforms, PSs came to enjoy more autonomy from CUs. The reformers believed that the consequence would be an internal democratisation of CUs, with CU leaderships becoming accountable to more politically-aware PSs and their members. However, the PSs surveyed in this study tended to interpret their enhanced autonomy primarily in terms of reduced central accountability in their own economic decision-making. And where PSs did assert themselves politically, in the arenas of CU committee meetings and annual general meetings, it was mainly to obstruct CU managers carrying out corporate changes whose implementation was also at the heart of the reformers’ agenda. Increased PS autonomy from CUs became expressed in two rather different ways, depending on local conditions. In Mwanza region - or at least in central and eastern Mwanza region- private buyers were many and by 1997 mostly had already been operating for a couple of seasons. They had used this time to establish independent chains of buying posts operating from rented houses in most larger villages. In any case, in most villages private buyers had no alternative but to organise their own buying posts, for very few PSs had more than a single godown, and all were constitutionally contracted to reserve one godown for purchases on behalf of Nyanza CU. In central and eastern Mwanza, PSs came under extremely strong price competition. To try to retain local market share, a very common PS strategy was to localise price-setting decisions. NCU’s opening price in Mwanza had been set by the organisation’s annual general meeting in April 1997, in line with a consideration of the anticipated world market price and the organisation’s costs. Within a few days of the season opening at the end of June it became obvious - at least in central and eastern Mwanza - that this price was well below that which private buyers were willing to pay. NCU’s flexibility to modify its price was restricted constitutionally as well as financially, but in the end it was agreed that in certain ginnery zones where competition was greatest (including the ones surveyed) a higher price could be offered to those delivering very large quantities of cotton to a PS. However, on hearing of this, almost all the PSs within these zones which still had cash began purchasing cotton at the new, higher price irrespective of the unit quantity delivered. In other ginnery zones, where a premium was introduced only for peasants bringing cotton direct to the ginnery instead of to a PS, all PSs in villages near to ginneries began to offer the same premium. This situation was repeated in at least one of the new CUs (Victoria CU) in Mara region, where the level of competition was similar to eastern Mwanza. After the CU found its opening price was not competitive and revised it upwards, a number of affiliated PSs took this as a signal to set their own prices - in this case 10-15 percent higher than the new, stipulated level. In Shinyanga, the situation was rather different. In the first place, there were fewer competitors, at least until the last weeks of the season when the crop in Mara and eastern Mwanza was exhausted and there was a movement southwards by buyers. Secondly, as will be described, there was a degree of collusion between buyers - including Shirecu. Thirdly, a very large number of Shirecu PSs owned more than one godown - in some cases up to five - as a result of the Union’s earlier highly extravagant building programmes. In this context, the Regional Commissioner for Shinyanga held a meeting with the region’s three largest private buyers prior to the opening of the season and requested them to buy through the ‘underutilised’ infrastructure and manpower of the Shirecu PSs. 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 PSs were enthusiastic and embraced the opportunity to act as agents for private buyers whenever they were approached. At every ginnery zone headquarters I visited in Shinyanga I was told that all Shirecu-affiliated societies with more than one godown had sub-contracted or rented out all additional ones. Only in one of the villages I visited was there a breakaway CU-affiliated PS with more than one godown, and here too the ‘extra’ godown had been sub-contracted. ‘Sub-contracting’ differed from renting in that the PS received a weight-based levy rather than rent, and was responsible for supplying workers for the buying operation.
e. Financial ‘viability’ and independence Since 1991-95 both PSs and CUs have been supposed to attain financial ‘viability’, which in the current Tanzanian context means ceasing to be dependent on public organisations higher up the financial hierarchy forgetting or forgiving their debts, or lending them money at non-market interest rates. As already stated, in order to prepare the CUs for implementation of the provisions of the 1991 Cooperative Act, a so-called ‘once and for all’ clean-up of their balance sheets was undertaken in 1991-92. This relieved rather than eliminated RCU debt however, for its main provision was to convert existing overdrafts and accumulated interest arrears into non-interest bearing loans. Debt forgiveness was not implemented until mid-1995, when as part of the preparations for cotton market liberalisation, the Treasury assumed all responsibility for RCUs’ outstanding debts to the domestic banking system (Business Times, 6 October 1995). Of the three main RCUs in the WCGA, Mara CU failed miserably to recover any economic independence. Although according to the newspaper report just referred to, it benefitted in 1995 from the debt forgiveness operation like other RCUs, by 1 March 1996 its debts once more stood at Tsh 13bn (US$ 20m). Amongst the political cognoscenti in Musoma town (its former headquarters) this debt had two causes- firstly, a high level of outright theft by Union leaders, and secondly the 1995 Tanzanian General Election. Little can be said of the former, except that the ‘total lack of financial control’ in MCU identified by Coopers and Lybrand in 1992 was clearly never rectified. As for the latter, it can be pointed out that during 1995 Mara became a political cockpit. The home region both of the ‘father of the nation’ (former President Nyerere), of a large group of powerful figures in Tanzanian government, military and business circles who rose to influence during and soon after his rule, and of a number of nationally prominent figures in the leading opposition party (NCCR [ NCCR started life in the early 1990s as a coalition of forces, including radical students, campaigning for thoroughgoing constitutional reform in Tanzania. By 1995 it acquired as its leader the most prominent populist politician from the ruling party CCM (and popularly esteemed former Minister of Home Affairs). Ideological differences between NCCR and CCM, never strong, now dwindled and their articulation was replaced by accusations and counter-accusations of corruption. At the same time, many politicians associated with CCM but who had failed to win nomination to run as parliamentary candidates for the ruling party crossed to NCCR (and other parties) with the intention of gaining nomination and entering parliament on the coat-tails of the reputation of NCCR ’s leader.] ), Mara saw perhaps the most hotly contested group of constituencies in the country outside of the capital. Already during the primary elections for nomination of CCM candidates deep personal, political and ethnic divisions manifested themselves and these steadily deepened as the election proper grew nearer. The leadership of Mara CU and the Mara region CCM organisation overlapped considerably. The Mara CU Chairman, who stood as the CCM candidate in Musoma Rural, was also the CCM Regional Treasurer. It was widely asserted that not only his lavish campaign, but those of other CCM candidates, were financed from the MCU budget from the primary stage to the election proper. The relation of Mara CU, CCM and NCCR will be returned to later. While Mara CU fell into the hands of the Receiver, NCU and Shirecu survived, at least until 1997. As already indicated, neither can be said to have done so through major restructuring, however. Although both had become a little more specialised in their range of operations than in the early 1990s, and while Nyanza CU’s (but not Shirecu’s) employment has been cut back (from about 1,400 to about 880), both were structurally still the same kind of organisation as before the reform. The main difference was that they had now become mainly dependent on short-term importer finance rather than long-term government borrowing. In the case of Shirecu, 70 percent of crop purchasing capital was being obtained through importer pre-finance in 1997, at interest rates of 12-15 percent. Another 10 percent was obtained through a National Bank of Commerce overdraft at normal corporate rates (21 percent) and 20 percent was obtained from the Union’s own resources. Nyanza CU finance was obtained roughly in the same way, with the exception that none seems to have been derived from own resources. Interest on importer pre-finance was normally repaid in the form of giving a discount on delivered cotton. The Department of Cooperative Development in Dar es Salaam had since 1991-92 pursued a policy of encouraging CUs, particularly the large ones, to set up their own banks. These banks were supposed to initially attract capital from affiliated PSs and then from other investors, prior to lending working capital to CUs’ marketing arms on a fully commercial basis. In the country as a whole, four such banks had been set up by mid-1997, including one in Kilimanjaro, but only one had got off the ground in the WCGA. This was in Biharamulo (Bukoba) - outside the areas covered by this survey. NCU was in the process of setting one up, but despite having lent affiliated PSs considerable sums to buy shares it had still raised only about a quarter of the minimum capital requirement. None of the new CUs in the WCGA had managed to set up banks either, nor did they display any interest in doing so. All were almost completely dependent on importer or local buyer pre-finance, although on more unfavourable conditions than the big RCUs. The big RCUs had two relative advantages in their dealings with importers. Firstly, since they owned their own ginneries and transport and had agents in Dar es Salaam, they could sell lint on f.o.b. terms. This implied retaining the higher share of value added during the ginning phase, and gaining a margin of a few US cents/lb through transport to the port of exit. Secondly, they could guarantee lint in sufficient quantity to sell by tender, which meant that larger importers had to bid against each other to obtain supplies (thus raising the price obtained a little further, or reducing the agreed rate of interest). The new CUs 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. They could not gin at all without paying a substantial ginning fee to a private ginner, because they owned no ginneries of their own. Most of them felt that they had a claim to RCU ginneries within their home area, whose title deeds were usually still in the names of the district CUs which had been compulsorily amalgamated into RCUs some decades before.However, the one new CU which had tried to pursue such a claim in court had been unsuccessful. The Registrar of Cooperatives told me that legislation would probably have to be altered before breakaway CUs could obtain a division of RCU property generally, and that such legislation was actually contemplated, but that ‘the new CUs had been reluctant to delay their formation until it was on the statute book, as I advised them to’. This point will be returned to. In most cases, the nature of the contracts between the new CUs and private buyers were much the same as those between these buyers and the PSs belonging to Shirecu. A private buyer would advance working capital on an interest-free basis against a seasonal, or more likely monthly purchasing target, and set the producer price. The CU would deliver unginned seed 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 didn’t own a ginnery) the ginning itself. The only advantage to the private buyer of contracting with the new CUs rather than individual PSs was a saving on the time (and possibly some of the money) which setting up and supervising a PS-based buying network would have entailed. In two cases, the new CU was organising the ginning and the private buyer was buying on a ex-ginnery spot price basis rather than on a pre-set ex-buying post one. In these cases pre-finance was being advanced at a fixed interest rate without the private buyer setting the seed cotton producer price. The CUs were also responsible for organising transport in such cases. Theoretically such contracts should have given CUs considerably higher mark-ups. But the big RCUs were not contract ginning for their newer rivals, and spare ginning capacity in the private sector was hard to find. Where it could be found, it entailed the new CUs having to wait until the private ginner had ginned all his own cotton, which meant that the new CUs fell behind in their delivery schedules and hence incurred high interest charges on the pre-finance. Moreover, while the real cost of ginning to the ginnery owner was in the range of Tsh 20-60 (ca. US cents 3-10)/kg lint, actual ginning fees were in the range Tsh 140-160/kg lint. CU margins in these cases were therefore only a little higher than where it was the private buyer who organised the ginning. The 6 new CUs dealing with cotton [ Tarime CU was dealing only with coffee.] listed in Table 3 above between them had contracts with seven private buyers. Six of these buyers had contracts with one CU each and one had contracts with four. A majority of the contracts with the new CUs were entered into by the small category of private exporters who at the time owned no ginnery of their own. The two largest of those concerned were both foreign-owned, one owned ultimately by the Swiss private company Reinhart and the other by the Singapore-owned Olam Ltd. Reinhart were at an advanced stage with plans to open their own ginnery in western Mwanza, and Olam were also giving this possibility strong consideration. In the medium-term at least this implied a probable reduction of dependence on the new CUs as sources of cotton. On their agency sales of seed cotton, the new CUs were obtaining Union levies of between Tsh 8.6 and Tsh 10/kg seed cotton (plus levies in the same range which were passed onto their affiliated PSs). Since they managed to collect between only 200,000 and 5.3m kg of seed cotton each (see below), the new CUs’ corporate revenues for 1997-98 season therefore fell in a range of US$ 3,225 to US$ 73,500. Out of these had to be met quite substantial expenses. None employed less than three permanent staff and one employed 22; all were renting office space. There was clearly no opportunity for them to even begin to accumulate the level of working capital necessary to establish an independent presence on the market. Although in the short-term Nyanza CU and Shirecu were much better off, it was clear that they were also balancing on the edge of a precipice. As competition was increasing, especially for Nyanza CU, the producer price was rising steeply and cotton was becoming physically harder to obtain. This translated both into lower (or in some cases negative) margins in relation to pre-season contract prices, and higher interest charges on pre-finance and overdrafts. Leading representatives of both Unions openly expressed doubts about their organisations’ capacity to survive, and more than one well-informed local observer felt that 1997-98 season could well be NCU’s last. Because of their dire financial situation, the older (1984 vintage) CUs had at last begun to clamp down on outstanding debts from affiliated PSs. Some PSs in Mwanza had simply not been advanced any funds for buying in 1997-98 season, on the grounds of having deficits stretching back over several seasons. On the other hand, as already indicated, PSs’ ability to exercise financial and organisational autonomy from CUs was simultaneously increasing. Hence, while CUs were making a few gestures in the direction of financial restraint, PSs were taking advantage of their new freedom by exercising no restraint. Most had 3-4 (and sometimes up to 6) permanent employees, even though their activities never lasted longer than 5 months a year, and continued in addition to pay honoraria in one form or another to ‘committeemen’. Contracts with private buyers, where they could be obtained, were greeted as means of staving off the need to reduce expenditure. No PS which I visited had used surpluses obtained from such contracts to start independent savings and credit activity. In fact, none had even used these contracts to obtain non-cash benefits from private buyers which would have helped to secure PSs’ future, such as repairs to leaking godowns.
f. From economic monopsony to ‘competitive performance’ During the days of cooperative monopsony, CUs bought on credit and determined producer price levels as a post-cost and post-margin residue. The opening of the cotton market for private participation obliged them to try to buy on a cash basis after 1995, and to offer competitive prices. In 1997-98 season NCU and Shirecu were able to buy continuously on a cash basis throughout Mwanza and Shinyanga regions for the first three or four weeks of the season. Thereafter all areas got cash on an intermittent basis for three or four more weeks, and PSs were obliged to buy on 2-3 days’ credit. After this time (mid-August 1997) most Nyanza CU PSs received little or nofurther cash and had to buy entirely on credit. A few Shirecu PSs continued to be supplied with cash on a regular basis after mid-August, but not the majority. On the other hand, the great bulk of the harvest had already been purchased by this time. Cash supply to PSs affiliated to the new CUs was considerably worse than this. In most cases it seems to have been intermittent from the outset and usually to have stopped completely after weeks 6-7. All but five or six (of 20 or so) private buyers also experienced liquidity problems during the season. The exceptions were the 4 or 5 largest private companies (which included two multi-national corporations) and another multinational which pulled out of the WCGA after about 8 weeks, in order to concentrate its resources on buying cashew in south-eastern Tanzania. The levels of working capital required to remain in the market on a continuous basis throughout the season were huge. Simply to buy enough seed cotton for ginning the comparatively modest target of 10,000 bales of lint required almost US$ 1 million, without expenditure on buying operations, transport and ginning. Many private buyers with their own ginneries began the season with a cotton buying capital of perhaps US$ 160,000, and then had to interrupt buying several times while waiting for payment for cotton they had already ginned. This meant that, after the first two or three weeks of the season, the old RCUs were rarely confronted with simultaneous competition from a large number of private buyers. Normally however they did face simultaneous competition from at least 3 or 4. Competition was severest in eastern Mwanza, where there were a lot of private buyers and where price war occurred. Even when Nyanza CU PSs had cash, their prices were on average 10-15 percent below those of their competitors. Shirecu’s prices were much nearer to those prevailing in Shinyanga generally, and Shirecu was generally more liquid than Nyanza CU. Prices offered by the new CUs were also below market rates, and these organisations were considerably less liquid than both Nyanza CU and Shirecu. Official returns from private buyers and CUs to the Cotton Board are used by the Board to compute overall market shares for seed cotton purchases. At least as regards private buyers the accuracy of these figures is doubtful, since the returns which they are based on are also used in order to calculate liability for certain taxes. Nevertheless, since private buyer reporting behaviour has probably not changed much over the years they may be useful for identifying trends over time. CUs have accounted for a steadily decreasing share of purchases countrywide since the first year of liberalisation in 1994-95. That year they accounted for 85 percent of seed cotton purchases. In 1995-96 this dropped to 70 percent and in 1996-97 to 60 percent. For the WCGA in 1997-98 their share was officially 52.6 percent, and for the ginnery zones which included the villages in this survey, 47.7 percent (figures for sales until 28 August, 1997, when almost all cotton had been purchased). All the Unions except Shirecu badly missed their pre-season targets (Table 4). Nyanza CU’s market share, computed on the basis of these official returns, fell from 21.9 percent in 1996/97 to 15.4 percent in 1997/98. Shirecu’s increased marginally to reach 26.9 percent. The combined collection of the post-1995 CUs was 6.2 percent. Table 4: Cooperative Unions’ purchasing targets and results, WCGA, 1997-98 season
Key: * up to 28 August, 1997 Source: own interviews Nyanza CU and Shirecu only succeeded in reaching the levels of purchases reported in Table 4 as a result of developments in the western parts of Mwanza and Shinyanga, areas not covered by this study. In these areas there was a much larger crop and fewer competitors than in the eastern and central parts of the WCGA in 1997/98. Table 4 shows that the new CUs which were operating in the western districts (Mweli Farmers’ and Kahama CUs) were correspondingly more successful than the others, whose catchment areas were the eastern districts and Mara. An impression of the range of local variations in market situation, even within the eastern part of the WCGA, can be gathered from Table 5, which summarises the ‘performance’ of each of the 19 cooperative buying operations in the 16 villages visited. In Mwanza, even where local competition was absent, large numbers of villagers (and even of PS members) sold to private buyers outside their village, while in all villages visited of those selling to PSs were also selling part of their crop to private buyers. Shirecu PSs obtained higher local market shares on average, but given the number of peasants selling to them these shares were still rather low, again suggesting a pattern of double selling. In general, PSs could only obtain local market shares of more than half in fairly exceptional circumstances, i.e. where the PS nearly matched the going market price, where cash supply continued up to at least week 9, and where there were three or less private buyers competing with it. However, these were necessary rather than sufficient conditions, for there were also cases where all were present without PSs obtaining a market share in this range. Table 5: Competition and competitive performance, surveyed WCGA PSs, 1997-98 season
Key: No. PBs in village: number of private buyers buying in PS’s ‘home village’ N sellers: estimated number of villagers selling to PS ‘market share’: purchases recorded by PS in relation to purchases recorded by clerks of all buying operations in village in which PS was situated * variation expresses different points of time during season ** including NCU buying post at Manawa ginnery *** including MAFACU, operating a private buying post ¤ personal estimate ¤¤ records of one private buyer missing Source: own interviews
4. Local-level social and political processes in the WCGA In this section the discussion moves away from what has happened to cooperatives in the WCGA since 1991-95 to how this pattern of developments has arisen. Attention will be shifted to social, and more particularly political relations in the WCGA. Here, a series of inter-related underlying processes can be distinguished. Firstly, political and economic liberalisation have resulted in a certain form of ‘empower ment’ of some previously subordinate groups and a certain form of disempowerment of some formerly dominant ones. Peasants in general have been ‘empowered’ vis-a-vis local leaders, while PS officers have been ‘empowered’ in relation to CU ones. In addition to the ‘disempowerment’ of local leaders and CU officials which this implies, loss of power to effect certain political out comes has also been experienced by regional-level leaders. But the type of ‘empowerment’ which has occurred has largely been negative or passive or, to use a more familiar distinction, it has been appropriated as ‘freedom from’ - i.e., as a release from certain obligations - rather than as ‘free dom to’, i.e., as constituting a platform for collectively bringing about structural change. The main exception to this tendency has been where the ‘negative empowerment’ of PS officials has become linked with struggles (presumably long-standing) within Regional CUs, giving rise to the formation of breakaway CUs. By contrast, ‘empowerment’ of peasants has remained negative, not least because no forces emerging at other levels (not least political parties) have evinced any interest in their mobilisation. A second process has been for some well-established solidarities to weaken or dissolve as a result of the emergence of a new constellation of economic incentives on the one hand and the centrally-directed shift toward a less monolithic national political system on the other. In particular the ‘traditional’ solidarities between on the one hand, the different levels of the state system (village, District, Region) and on the other, the corresponding levels of the cooperative ‘movement’ have been severely attenuated, not least because private cotton companies have become available as sources of income for the state and because different levels of the state apparatus appear to be effectively free to make their own revenue-collection arrangements. A third process has been the eclipse of the largely ideologically-defined focuses of political mobilisation during the single party experience (for technocratically-defined ‘socialism’, against a combination of ‘traditionalism’ and ‘profiteering’ at home and imperialism abroad) by a new set of focuses. This new set has three interwoven reference points, none of them ideological in the normative sense: differential accessing of public resources, ethnicity/sub-ethnicity, and ‘leadership qualifications’. However, these components are sometimes combined with more re cognisable ideological focuses and in certain exceptional circumstances even subordinated to them. This process embraces mobilisations of both inter- and intra-party kinds, and those by other organisations in the public sphere. A fourth process is for the over determination of certain forms of political conflict, in particular ones between persons and organisations claiming to represent specific Districts andthose claiming to represent others, as well as between those claiming to represent specific Districts and those claiming to represent broader Regions. A majority of these processes can be detected by sketching out four sets of socio-political relations which form the parameters of changes in the cooperative sector: those between PS officers and peasant cultivators; those between CU leaders and PS officers; those between Regional CU leaders and leaders of breakaway CUs; and those between cooperative leaders at all levels, the political authorities, other political actors and the ‘private sector’. Discussion of each set of relations will take as a point of departure the social characteristics of the actors involved. Part of the conclusion of this paper will be concerned with these processes’ explanation.
a. Primary Society officers and peasant cultivators The nature of the now voluntarised membership of PSs, or at least the significant over-representation in their ranks of owners of oxen, has already been mentioned. In most villages, middle-aged male owners of oxen seemed to be even more prominent amongst PS leaderships, just as they were in village government, CCM (or, if an opposition party was locally dominant, in that), and Sungusungu [ For more on Sungusungu see Conclusion, below.] . This does not mean that the same individuals were always involved in all four sorts of organisation, however, for I came across some village governments which had taken decisions directly contrary to the interests of the leaders of ‘their’ PSs. More generally, there seemed to be a high turnover in PS leaderships, caused by their frequent failure to be popularly re-elected. PS leaderships comprise a lay committee led by a chairman and a treasurer, and a secretary, who is a full-time employee. I got the strong impression that secretaries were very often the better educated (again male) younger relatives of PS lay officers. Intuitively, it might seem obvious that the two major issues facing WCGA PS leaders in relation to local peasant cotton cultivators would be recruitment of PS members and attraction of cotton growers to sell through societies. Actually, active recruitment efforts were the exception rather than the rule, and in all but one of the Districts studied (Maswa), most PS leaderships made few visible efforts to attract sellers. While membership of cooperatives was clearly now voluntary in all the villages visited, it was considerably less evident whether it was genuinely open, or whether all members gained access to the same things. In a number of places I was told by agricultural extension officers that ‘villagers here don’t really know they can join. They knew they used to have to be members, but now a lot think that the PS has become some kind of private company’. Of the 19 societies listed in Table 5, recruitment campaigns had been held by only six. Four of these cases occurred in villages where two CUs were actively competing against each other. In the fifth the village government had campaigned for peasants to join ‘in order to root out a corrupt leadership’; only in one was a membership drive conducted by PS leaders themselves for non-factional reasons. While I never heard complaints that membership was being deliberately restricted, some PSs were offering services or payments (unauthorised at Union level), for which there would not have been enough resources to make to all members had membership been open. For example, a num ber of societies were making loans to members to cover school fees, or sometimes much more substantial items of expenditure. In other PSs, officers told me that they were making ‘second payments’ to members some years, ‘whenever we hear that the Union has done okay’. ‘Second payments’ is the local term for members’ dividends, distributed in cooperatives the world over in years in which a surplus had been attained. None of the Unions in the WCGA had achieved a surplus in the 1990s, yet these payments seemed to be quite widespread [ They had been made in a majority of societies in Nassa ginnery zone (Mwanza) in 1994 and 1995, for example .] . The one service which it appeared that PS members as a category normally benefited from, i.e., which was not either restricted to a circle around the PS leadership or available to everyone who sold to a PS, irrespective of whether or not they were a member, was the provision of insecticide on credit. Table 6 (based on the farmer survey parallelling this study) shows that PS members ‘normally’ used insecticide with considerably greater frequency than non-members but that in 1997/98 season, when it ceased to be available on credit, members stopped or reduced using it to a greater extent than non-members. Table 6: Insecticide use and membership of Primary Cooperative Societies
Key: * refers to those continuing to use insecticide Source: own interviews The importance of PS membership for credit-based insecticide supply was especially great in Mwanza; in Shinyanga insecticide had not been widely distributed for some years, either on credit or for cash. Amongst the 18 Mwanza peasant cultivators surveyed who normally used it, 11 were PS members and seven non-members. Seven of the 11 members stopped spraying in 1997-98, but none of the non-members did so. As already indicated, except where there was competition within a particular village between an established and a breakaway CU, PSs’ efforts to maximise cotton purchases were at best uneven. In Shinyanga, where it was the policy both of regional and village government to encourage private companies to organise their buying effort through PSs, leaders of most PSsappeared to be quite unconcerned about whether they bought more for Shirecu or for the private company. The same was also true in some villages in Bunda district, Mara. Only in Mwanza was the major CU involved in what normally would be interpreted as open and systematic competition with private buyers in general, and only here could be found anything resembling locally-devised tactics or strategies deployed to help PSs obtain a good market share. One of these tactics, namely purchases at above the CU-authorised price, has already been mentioned. In a few places this was supplemented by the offer of free transport to the buying post for those bringing large amounts of cotton, or ‘special prizes’ for the largest sale during a particular week or month. These tactics were borrowed by PS leaders from amongst the range improvised in the same localities by the agents of the private companies. Less transparent tactics were also in play. In relation to non-members, and in a context where there were widespread popular suspicions about the reliability of private buyers, PS leaders in Mwanza wherever possible reinforced the feeling that private buyers might not return at the end of the season to provide free seed for the following year’s planting. In order to obtain seed from a buyer one needed a receipt showing a sale of cotton to that buyer. A very large number of peasants therefore sold at least some cotton to PSs on the grounds that the latter could not run away from their obligation to provide seed. In relation to members, the main tactics applied seem to have been a mixture of threats and calling in of favours and obligations, familial and otherwise. It was my impression that the threats employed typically played on many members’ confusion about whether the PS was a state institution or not. For those thinking it was some kind of state institution, the belief seemed common that members were legally obliged to sell their cotton through it. In the farmer survey which accompanied this study, 36 (54 percent) of 67 peasant cultivators interviewed stated that they sold cotton to a CU through a PS [ The TADREG (1997) survey mentioned above contained a question asking respondents to state their preference for private or cooperative buyers (not quite the same as asking who they sold to). 49 percent of cotton cultivators stated a preference for cooperatives and only 26 percent for private buyers (25 percent expressed no preference or replied ‘don ’t know ’).] . 21 of the 31 PS members sold through PSs, as against 15 of the 36 non-members. Respondents were asked the main reason why they chose to sell to a particular buyer. The question was open-ended and responses were classi fied later. All those who sold to private buyers answered (literally) either ‘because of the price’ or ‘because there was no alternative cash buyer in the village that day’. Table 7 shows the re sponses of those selling to CUs via PSs to the same question. The Table distinguishes the answers of PS members and non-members. In some villages in Mwanza PS members selling to private buyers were threatened with expulsion from the PS, and one of the peasants I interviewed in this village actually had just been expelled. Other than cutting off this individual from future access to credit for insecticide (if indeed this ever became available again), it is unclear what this entailed; perhaps expellees withbad loans would be taken to court. This was becoming a widespread method for other kinds of debt collection in the WCGA, although it was unclear how effective it was. Table 7: Main reason for selling cotton via a primary cooperative society, 1997/98
Source: own survey Given that cooperative societies in the WCGA have never operated according to a pure ‘Roch dale’ model - although for a period they did predominantly serve popular interests - and given that their post-1984 voluntarisation has unfolded mainly from above, it is perhaps inevitable that their reconstitution at local level should be more or less exclusively as institutions of the upper stratum of the rural population. What is perhaps more surprising is that loyalty to them amongst this group (and, indeed, a wider population) has been strong. The exact basis of this loyalty is not clear, but taken together with responses to other questions in the survey (including some already mentioned) it seems mainly to have involved a combination of a wish to preserve personal access to certain services, a strong belief that private traders could not be trusted, and a vague fear of reprisals. Seemingly, the poor were not receiving the services referred to and nor did they have much expectation of obtaining them elsewhere. The only important dimension of private trader trust worthiness could be tested on the spot by whether they had cash. Amongst the remainder of the population, including a majority of those who cultivated with their hands [ Corresponding to the over-representation of oxen-owners amongst PS members, and of PS members amongst those selling to PSs, hand cultivators were over-represented amongst those selling to private buyers. 13 of 21 hand cultivators interviewed sold their cotton to private buyers, as opposed to 12 of 29 oxen-owners and 4 of 17 oxen- hirers.] , there was simply general relief that it was no longer compulsory to sell through PSs, or at least - as several peasants told me - ‘now when we sell to them, at least we don’t have to pay a bribe to get cash for our cotton’.
b. Cooperative Union leaders and Primary Society officials CU officials in Tanzania occupy a completely different world from PS leaders. The horizon of most PS leaders is normally the village and exceptionally the CU Ginnery Zone committee or the District council; the main points of reference for officials of the large RCUs in the WCGA are the different elements of the Regional and District Administrations, figures in the cooperative ‘movement’, the Ministry of Agriculture and the National Bank of Commerce nationally, Members of Parliament from the Region, the Region’s large-scale transporters, and the Tanzanian agents of the main international importers. RCU leaders (especially in Mwanza) normally have some kind of tertiary education, PS leaders have normally attended only primary school. It is hardly surprising that their assumptions are also far apart. Two established sites of interaction between RCU and PS leaders have already been mentioned: one collusive (hewa la pamba, i.e. collusion to defraud) and one conflictual, namely the ‘obstruction’ of RCU reform by CU committeemen, who in turn are drawn from those PS leaders who make it onto Ginnery Zone committees [ See above, footnote 4.] . A third, again conflictual, site of established interaction can also be mentioned. This is the reverse of hewa la pamba, namely conflict over the reconciliation of RCU and PS accounts, and alleged favouritism in this process, in the form of RCU’s differential treatment of various ginnery zones, and of different PSs based in them. Conflicts both around these issues and others have increased steadily since 1991-95, as RCUs have felt themselves under increasing pressure to balance their books, as they have in consequence sought to enforce PSs’ financial accountability to them, and while PSs’ opportunities to exercise autonomy have increased simultaneously. Seen from the side of PS leaderships, the upshot has been the inflicting of an ever-length ening train of slights and injustices. RCUs have unilaterally reduced or stopped making cotton purchase advances for a variety of reasons, from PSs being deemed to have ‘unacceptable’ levels of debt to ginnery zones being deemed to have too little cotton left to make such advances safe to issue. They have likewise allowed some PSs to pay more for their cotton than others. They have stopped advancing inputs on credit to most societies and stopped advancing cash to pay hon oraria to the committeemen of others. Finally, they have simply withheld levy from certain so cieties, on the grounds of their negative balances with the RCU. Seen from the side of the RCU, PSs were becoming increasingly undisciplined and/or ‘disloyal’. In Mwanza the complaint was mainly about discipline, particularly on price, in Shinyanga about disloyalty, in the form of rent ing out Shirecu property and PS services to private buyers. Nobody in the RCUs whom I talked to had any clear idea of how these problems, as they saw them, could be overcome. The spontaneous inclination was to discipline PSs, ‘but if we were to discipline all those who had been disloyal we would lose 80 percent of our societies’. Moreover ‘it’s not just PSs that are betraying us. Often they have been encouraged to do so by village gov ernment and even people higher up who should know better’. Resources from the CooperativeDepartment in Dar es Salaam to improve member education were not seen as part of the solution, even though for RCU leaders the word ‘uneducated’ was a standard prefix for ‘PS leaders’. In fact, existing member education packages ‘are maybe part of the problem: after the members get them, we get told to go to hell’. Meanwhile, almost no PS leader I interviewed had a clear conception of the extent of crisis facing the RCUs. When asked ‘what can cooperatives do differently in order to survive ?’, PS leaders all answered that RCUs should return to free input distribution, ‘give all members second payments every year’ [ ‘People are really broke around April when these payments used to be made. It would make a tremendous difference ’.] , and so on. When given these answers by PS leaders, I would normally try to explain why I thought that they were impractical. PS leaders then always replied in one of three ways. The first was to insist that RCU leaders should simply work harder, ‘then we will be saved’. The second was to signal fatalism in one form or another- usually, ‘then we will die’. The third, which was found amongst the youngest and most articulate, was to say that this was regret table but of no great concern to themselves personally as they had already made their own ar rangements for future income generation. If PSs continued to be based on (albeit sectional) forms of solidarity within villages, this was not visibly replicated in any feelings of solidarity or recognition of wider responsibilities by local PS leaders with regard to the defence of Unions. PS leaders tended to view Unions mainly as em ployers, while Union leaders on the other hand viewed PS leaders as unruly commission agents. | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
