THE COOPERATIVE REFORM PROCESS IN TANZANIA AND SRI LANKA

Introduction

Since the ex-colonial countries gained their freedom, there has been a constant search for a development 'magic formula' of sustained economic growth with poverty reduction. The recent emphasis has been on group-based lending to poor people (in particular to women) in the expectation that this would enable them to trade their way out of poverty, but problems with the sustainability of NGO-led lending, within the context of the World Bank's emphasis on the need for financial deepening (2007), have led to a renewed interest in savings and credit co-operatives. The Bank's Development Report for 2008 shifts the emphasis to improving the output of rural economies, the organizational vehicle for which is recognized to be a farmer association. In both cases, the solution being proposed is some kind of co-operative. The statistics of co-operative development are impressive; globally, 800 million people are members of co-operatives and 100 million are employed by them. There are several typesconsumer, industrial, housing, agricultural, fishing, credit, health and so on -and they form an important, but often overlooked, part of the institutional map in the economies of developing nations. A recent study of 11 countries in Africa has estimated that around seven percent of the population are members, and finds the numbers of cooperatives are growing (Develtere et al. 2008).

In the now-developed countries, co-operatives evolved autonomously according to the ebb and flow of social movements and economic comparative advantage, but in the developing countries cooperatives were promoted by colonial and then nationalist governments as a way of modernizing traditional economies (Birchall 1997). Nearly always, and nearly everywhere, they were controlled and regulated by governments and not allowed to become truly memberowned. In general, members came to see them as quasi-governmental agencies that provided useful services but did not belong to them (Develtere 1994). Having been propped up by government patronage, during the period of structural adjustment following the cold war many of them disappeared and their apex federations collapsed. This experience has led to scepticism about the value of co-operatives, but a reform process is under way to rehabilitate them and make them into what they should always have been -member-owned and controlled businesses. The problem is that there is no obvious alternative organizational vehicle for reducing poverty, particularly among rural populations Simmons 2009, see also Wanyama et al. 2009), and it is likely that donors will want to continue to develop them in the run up to the Millennium Development Goal finale in 2015. Three questions then arise that we want to explore in this article:

This article draws on data from our recent study of cooperatives in Tanzania and Sri Lanka. 1 We carried out a sample survey of 475 co-operatives of various types in three regions of each of the two case study countries, and undertook semi-structured key informant interviews with 30 people in Tanzania and 28 in Sri Lanka. We analyzed relevant policy documents, financial reports and other written materials, and wrote a history of co-operatives in each country. At the international level, we interviewed key informants from 18 organizations and several specialist consultants (Birchall and Simmons 2008). Here we focus on the reform process, and draw mainly on the documentary analysis and key informant interviews in Sri Lanka and Tanzania. 2

2

Why do co-operatives need reforming?

Why do co-operatives need reforming? To answer this question, we need to provide a short history of co-operatives in developing countries. There are three distinct periods; the colonial period, the nationalist period after countries gained their freedom in the 1950s and 1960s, and the period of structural adjustment and market liberalization that began in the late 1980s and takes us up to the present day. Of course, there is a pre-history of mutual aid organizations; in Africa they took the form of savings and credit societies and mutual work groups (Muenkner and Shah 1993), but the colonial powers did not build on these forms but ignored them, introducing European 1 UK Economic and Social Research Council funded project, RES-155-25-0077: The role and potential of co-operatives in the poverty reduction process, in the ESRC's Non-governmental Public Action Programme; the project lasted for 2.5 years from 2006-8 and was rated 'outstanding'. 2 The key informants were selected in a purposive sample that was determined in advance so as to ensure a wide range of views and minimize the danger of being 'captured' by any one viewpoint. We identified 12 kinds of organization that had to be covered and set targets for minimum numbers to be interviewed in each category: the co-operative ministry and its registry/commission (3 people); the national co-operative federation (2 people); specialist co-operative federations (6 people); co-operative colleges/institutes (2 people); Northern co-operative development agencies (3 people); multi-lateral donors (2 people); bilateral donors (2 people); international NGOs (2 people); national/local NGOs (2 people); private business federation (1 person); a women's research centre (1 person); a research centre for poverty reduction (1 person). We achieved 28 interviews in Sri Lanka and 30 in Tanzania. models derived from their home countries' co-operative legislation. There were a variety of promoters: white settlers, indigenous groups, aid agencies, missionaries and nationalist politicians (Birchall 1997, Develtere 2008, but by far the most important influence was that of colonial administrators. When we talk of path dependency in contemporary co-operative sectors, the dependency really begins with the legal and administrative structures imposed by the colonial powers.

The colonial period

The approach of each administration was different. Develtere and his colleagues suggest that there were four traditions that they label the unified model, the social economy tradition, the social movement tradition and the producers' tradition, corresponding roughly to the British, French/Spanish, Belgian and Portuguese traditions. However, a careful reading of the history in each country shows that, while the emphasis may have been different, the approach of the colonialists was uniformly authoritarian. Only in those countries not affected by colonialism did a fifth model of a 'self-generated movement' emerge (Develtere et al. 2008, Ch. 1). The British approach was to develop a single co-operative 'movement' based on a unified cooperative law derived from the British Industrial and Provident Society law and the Raiffeisen co-operative banking model. They set up an elaborate three-or four-tier structure of primary societies, district and provincial unions and national federations. While paying lip service to the idea of autonomous co-operatives, in practice governments began to regulate the sector quite heavily as soon as it began, setting up a powerful co-operative department headed by a registrar who had much wider powers to intervene in the affairs of co-operatives than had the British Registrar of Friendly Societies on which the system was modelled. This 'British-Indian pattern' was copied throughout the British empire (Muenkner and Shah 1993).

The French did not set up multi-tiered structures, nor did they focus their efforts through one co-operative registry. Co-operative development was done by general administrators rather than specialized personnel, and each type of co-op was under the supervision of a different ministry; agricultural co-ops under the ministry of agriculture, credit co-ops under the ministry of finance, and so on. However, they did try to impose their own ideal model, the Societe C 2010 The Authors Journal compilation C CIRIEC 2010 Indigene de Prevoyance (SIP) that had the same kind of function as large, multi-purpose agricultural co-ops (Muenkner 1989). Develtere describes them as 'centrally administered frameworks that had no relation to the existing traditional structures' (1994, p 46). In 1947, a French co-operative law was extended to the colonies that made possible the founding of autonomous co-operatives. However, in 1955 a decree replaced this and introduced a supervising authority similar to the British Registrar (Muenkner and Shah 1993). The government interest was still in harnessing mutual societies to centrally-designed rural development schemes (Develtere 1994). Again, as in the British system, the basic pattern applied universally to all French colonies.

The Belgian approach was based on legislation in their home country that defined a hybrid of municipal and co-operative enterprises (see Lambert 1963). After the Second World War, they began to develop more along the British-Indian system, creating special co-operative departments and introducing large agricultural co-ops. As in the French colonies, they were paternalistic, topdown structures; provincial governors fixed the prices co-ops would pay their members for produce, appointed education committees and advisors, while the district commissioner appointed co-operative management (Muenkner and Shah 1993). The Portuguese approach was to set up agricultural co-ops for the export trade, but they were 'merely functional appendages of rural extension work of semi-public agencies' such as cotton and coffee marketing boards (Develtere 2008, p 10).

We can see these processes at work in more detail in our case study countries. During the colonial period, in both Sri Lanka and Tanzania small co-operative sectors grew up, promoted by government officials and highly regulated. In Sri Lanka the emphasis was on rural credit and distribution of consumer goods, while in Tanzania it was on the organizing of cash crops for export. In Sri Lanka, cooperatives began in 1906 with the formation of a credit society along similar lines to the movement begun in India. A law of 1911 laid down the basic rules, while government inspectors kept a close watch on the sector and used it to give loans to farmers. Growth was slow and societies remained small; by the 1930s there were 600 societies, with a membership of 240,000. During the Second World War, food shortages and rationing led the government to organize an islandwide network of consumer co-ops. By 1945 there were over 4000 with a membership of over a million people, and as one historian describes it 'More than half the population were being clothed and fed through Co-op stores' (Jayaweera 1987). Then in 1947, just before C 2010 The Authors Journal compilation C CIRIEC 2010 independence, there began a drive for forming agricultural production and sales societies, which made rapid progress in dairying, fisheries, small scale plantations, and cottage industries.

In Tanzania, the initiative originally came from coffee growers, who in 1925 formed the Kilimanjaro Native Planters Association to market their coffee crops. Colonial government officials encouraged it but became alarmed by conflicts between the association, European planters and local chiefs. The financial crisis in 1929 was a pretext for government to close it down and promote its own marketing coop sector that was integrated with the chieftaincy and guided by the state (Gibbon 2001). A law of 1932 regularized this, but because of the lack of democracy it was deeply unpopular, and in 1936 the British military had to quell protests against it. Cotton co-ops proved to be more popular; these were formed from the late 1940s onwards out of campaigns by nationalist politicians against unpopular Asian agents. The colonial government promoted them and by 1958, 275 societies were taking 83% of the cotton crop; they were then given a marketing monopoly. The Government was uninterested in member education; co-ops were seen simply as a good way of organizing the export trade in coffee and cotton.

The nationalist period

After independence, all the English speaking countries of Africa continued to follow the British-Indian pattern. Most of the Frenchspeaking countries also continued to follow the 1955 Act that had promoted parastatals. It was a period of 'extravagant praise and great expectations for co-operatives' (Laidlaw 1978: 64), and governments began to take an active role in promoting them. Nationalist leaders had grown up in the co-operatives, and so had a high regard for them (Spaull 1965). Aid agencies were also keen to promote them and did so persistently over the next 30 years until structural adjustment programmes forced a rethink. Co-operatives were helped by some ambiguity over their meaning; they could be 'read' as either a means towards a market society or towards a socialist society, and in both cases were seen as part of a project of modernization.

For co-operative members almost everywhere it was business as usual, but with new officials in charge. Where genuine co-ops developed, 'the ruling party or the state quickly brought them under their control to create jobs for officials and to use them for political ends' (Muenkner and Shah 1993: 16). A dependency relationship soon reasserted itself, with co-ops dependent on handouts from the state, and with political parties using them as an organizing base. Everywhere there was a rapid growth in co-operatives, which in some countries meant a brief flowering of independent societies. For instance, in Sri Lanka the new government started a drive to create agricultural fishery and industrial co-operatives. By 1956 there were 995 agricultural societies which gained a predominant position in supply of fertilisers. Smallholders in the export sectors (rubber, coconut and tea) formed co-ops for inputs, marketing and processing, while dairy co-ops were also formed on a small scale, selling to the government milk board. By the mid 1950s there were 75 different types of co-op, single purpose but reflecting the needs of the members. However, they were small and often not viable, with one village having seven or eight different types (Rajaguru 1996). In Tanzania after independence there was a rapid expansion of rural co-ops in a wide range of sectors. The first five year plan saw a drive to place all marketing of crops under control of co-ops, which were seen as the main vehicle of the government's effort to modernize the economy. By 1965 over 20 types of crops were being marketed through 1287 primary co-ops, and they controlled over 80% of agricultural production and marketing (Banturaki 2000).

Rapid growth led in many cases to weak co-operatives that were badly managed and sometimes corrupt. Inevitably, governments began a period of intensification of control, culminating in some countries in the complete takeover of the sector. For instance, in Sri Lanka the multi-purpose societies had proved a disappointment; they were not really multi-purpose, retail sales were inadequate, and there was inefficient and dishonest management. With a lack of member education, they had become simply distributors of rationed commodities, and only a few had created a real development role based on identification of members needs. In 1970, 5818 MPCS societies were amalgamated into 371, each covering 10-20 villages. As Rajaguru put it 'Co-operatives were born from the pen of the Registrar and not from voluntary association ' (1994, p 6).

In Tanzania also, rapid expansion of the sector was associated with a decline in efficiency. By 1966 there were rising complaints, and a Presidential commission of enquiry was set up to investigate charges of nepotism and corruption. The Commission urged the government to expand co-operative education, strengthen control over the movement, and increase the powers of the Registrar to fire incompetent and corrupt leaders. All these were enacted in a law passed in 1968, while 16 district and regional co-operative unions were taken over, their committees dismissed and government personnel put in. At this time, government interference was seen as beneficial, the argument being that government was using cooperatives to achieve the political aim of socialism, an aim which overrode any claims on the part of co-operative members to autonomy. Under the Arusha Declaration of 1967 co-ops were seen as 'instruments' for implementing the policy of socialism, and from 1969 onwards they were fitted into Ujamaa villages as multi-purpose cooperative societies.

Takeover by governments did not achieve the required results, but the answer to disappointment was seen to be even more centralized control (Hussi et al. 1993). For instance, in Sri Lanka, under an Act of 1970 co-op boards were to consist of 15 people, only five of whom were elected by the members; in many cases hardly any were elected and there was little interest among members to stand. Then in 1972 another act increased the Registrar's powers even further; the government's view was that large sums of its money were in co-operative hands and with 46% of the societies either defunct or loss making it had to take control. However, political patronage and corruption continued. In Tanzania, in 1976 President Nyerere declared that co-operatives could not cope with his 'quick march to socialism', as they were capitalist organizations. All cooperative unions were dissolved and replaced with crop authorities that were required to do marketing directly from the villages. There was resistance; the unions failed to die a natural death and had to be forcibly closed by the police. The decision to abolish them was not accepted by many in government and, after a crisis in rural production that was blamed on the poor performance of the crop boards, in 1984 Nyerere changed his mind. The unions were reestablished, but only 'to serve primarily political interests' as a mass organization under the direction of the ruling party. As one historian describes it 'Heavy government involvement and manipulation had systematically eroded and diminished the poverty reduction potential of co-operatives' (Sizya 2001: 6). By 1989 the co-operative unions' debts had risen to 40 billion shillings; the government confessed publicly that 87% of these were the result of its own policies.

Elsewhere, the pattern was similar. In Zimbabwe, the ZANU government kept the bureaucratic structure inherited from the British, but filled most of the committee places with party members. In Chad, Madagascar and Tunisia, co-operatives became 'heavy bureaucratic organizations' dominated by government or political parties, in Senegal multi-purpose co-ops were promoted as local agents of a state marketing board, and in Zaire the co-operative and mutual movement were linked to one national trade union which was a satellite of the single party in power (Develtere et al. 2008).

In 1966, the International Labour Organization passed a Cooperatives (Developing Countries) Recommendation which called for governments to set up one central body that would be the instrument for developing co-operatives, but without affecting their independence and the voluntary nature of membership. Other international agencies saw government control as a temporary phase that would eventually lead to democracy. But by then, as Alex Laidlaw put it, the co-operative 'movement' had gained three masters: the civil service, government ministers and local politicians. The old colonial system had been taken over and then used vastly to increase control over co-operatives (Laidlaw 1978).

By the mid-1970s, there was a growing awareness that 'the poor had not been reached' (Verhagen 1984: 4). In 1975 a UNRISD report declared 'rural co-operatives have seldom achieved the development goals set for them by economic and social planners' and 'they bring little or no benefit to the masses of poor inhabitants' (1975: 10 and ix). The International Co-operative Alliance held an experts' consultation, which found that co-operative principles such as voluntary membership were being thwarted (ICA 1978). Cooperatives were a very imperfect embodiment, and occasionally a caricature, of the ideals of co-operation. Yet the experts agreed that 'theoretically and in the long run' co-ops could solve the problems of development. Support continued, but they were seen merely as implementing agencies for externally-generated projects. As well as being 'gov-operatives' they had become 'don-operatives' (Braverman 1991).

At the same time, there were some developments that pointed towards a different kind of future. The most important of these was the different trajectory of growth taken by savings and credit cooperatives. In the mid-1990s there were more than 5400 of them in Africa with a market share of 1.6%. They built on the traditional forms of saving such as tontines and savings clubs, and because they were small tended to be outside of government and party control (Muenkner and Shah 1993). In Asia, average growth during 1985-90 was nearly 8% in membership and 17% in savings, and by 1995 there were 14,500 unions affiliated to the World Council of Credit Unions (WOCCU), with over six million members. Again their market penetration was low at 0.34%, but this was to be expected if the poorest people were benefiting from them (Birchall 1997: 184-5 They varied in size from the informal credit circles called 'hui' groups in Vietnam to the Grameen Bank in Bangladesh.

Here is a more detailed example. In Sri Lanka, the old savings and credit co-ops (called 'Sanasa societies') were small enough to escape the attention of government. However, the establishment of the multipurpose co-ops affected them badly, as people visited the MPCSs for their rice ration and used the co-operative rural banks that were attached to them; between 1964 and 1978 the number of Sanasa primary societies dropped from over 4000 to 1300(Hulme et al. 1996. Then, in 1977 a community activist, Dr Kiriwandeniya, seeing the potential in these societies as a base for a potential social movement, reactivated the society in his home village. He then organized a seminar to publicize the results, and the movement began to grow. Between 1980 and 1985 the number of primary societies almost doubled and 19 district unions had been formed. Efforts were made to incorporate lower income groups and women. By the mid 1980s the movement faced serious constraints on further growth due to lack of technical and managerial capacity, and there began a long association with international NGOs that provided assistance. However, while the higher tier organizations became more dependent on grant aid, the primary societies remained financially independent. The secret of Sanasa's success was that the leadership consciously avoided contact with politicians, and it became a norm that people active in party politics could not be leaders.

Another positive development was -at long last -a serious attempt by some agencies to invest in co-operative education. Previously, the emphasis had been on training of government department staff and managers of co-op federations, but now the ICA and ILO began programmes that targeted managers, board members and ordinary members. Another development was the beginning of a process of parallel promotion and strengthening in which local coop leaders and Northern co-operative development agencies began to work together. This was particularly important in the credit sector, where Desjardins International (DID), the Canadian Cooperative Association (CCA) and WOCCU began to work in long-term programmes with credit movements in several countries (MacPherson 1999). Finally, there was a shift of emphasis among donors towards semi-cooperative forms that were able to circumvent the governmentregulated official sector and develop independently. The UN's Food and Agriculture Organization was particularly active in developing alternatives to traditional farmer co-ops, calling them producer associations and using the new participatory development methodology C 2010 The Authors Journal compilation C CIRIEC 2010 (Rouse, 1996). The International Labour Organization's ACOPAM programme worked in six countries in Sub-Saharan Africa, promoting over 2000 grassroots organizations benefiting 80,000 people. Also using a participatory methodology, the programme set up selfmanaged cotton markets, village grain banks, savings and credit schemes, village irrigation schemes, women's groups and so on. Before it ended in 2000, the programme helped to draw up new cooperative laws in these countries to create a favourable environment for these new forms of association (Birchall 2004).

The period of structural adjustment and co-operative reform

During the 1980s, the infrastructure that had sustained the 'cooperative' sector began to fall apart under the impact of a structural adjustment policy that forced many governments to privatize and deregulate their economies. Agricultural co-ops that had operated in a protected environment were not in a good condition to face the rigours of market liberalization. In 1994, a report on Sub-Saharan Africa summed up the challenges: low business efficiency was 'the rule rather than the exception', there was a weak capital base, heavy indebtedness and limited credit-worthiness. There was a low level of preparedness for the changes in their environment, and with members being disillusioned it was difficult to see from where the pressure for change would come (Birgegaard and Genberg 1994: 3-4). A report published by the World Bank in 1993 was hopeful that new associative forms would develop and that the old co-ops would transform themselves. However, it reported that some African governments were still exerting control and providing subsidies; their mental attitudes and the vested interests involved were highly resistant to change. The report also criticized the World Bank for not being clear about what it wanted to achieve in working with co-ops, and for not paying attention to the need to strengthen them as institutions in the long term. Governments were still using marketing co-ops in the old way, compelling them to be purchaser of last resort for crops, and subjecting them to legal restrictions that were not placed on other private sector businesses. Member participation was still discouraged through control by local officials and political leaders. As a condition of disbursement of credit, the Bank had supported the passing of new co-operative laws in several countries, yet co-operative leaders were saying that even the new laws implied too much government control (Hussi et al. 1993: 9). Under the impact of budget cuts, introduction of market prices, C 2010 The Authors Journal compilation C CIRIEC 2010 and the collapse of the parastatal marketing boards, many cooperatives disappeared. Many more continued at a survival level, while some, under strong leadership and good management, began to prosper. Wanyama describes it this way: 'liberalization has produced a sieve for sifting the grain from the chaff in the co-operative sector' (Wanyama et al. 2009: 386).

Again, our case studies can provide more detail on this period. In Sri Lanka, the MPCSs were cushioned by their role as main distributor of rationed commodities and welfares, but other sectors experienced a steep decline. In 1978 there were 270 agricultural co-operatives, but by 1984 they had declined to 100, and it was not until the late 1980s that several new, smaller farmer co-ops would emerge to meet their members' needs. Textile weavers' co-ops experienced intense competition, and by 1994 they were down to 29 societies. By the early 1980s only 78 of the 450 listed fishery societies were running at a profit. The credit co-operative sector proved to be an interesting exception. However, in 1985 Sanasa came under intense political pressure to become the rural delivery mechanism for a USAID funded 'million houses programme' by which people could secure loans for housing improvement. The number of societies more than doubled, but many were formed just to get the loans. In 1988, just before elections, the ruling party forgave loans to poorer people and widespread defaults resulted. Sanasa was insulated from the losses, but it was left with thousands of societies that were weak or defunct, and since then the priority has not been further growth but consolidation and development of national apex institutions such as the Sanasa Development Bank and the Education Campus. In 2005 there were 8440 societies, with 858,000 members (Sanasa 2005).

In Tanzania, the impact of liberalization was disastrous. Cooperatives were in a weak position at the start of the trade reform process, and had no breathing space to adjust, and so private traders took over much of their business. They had inherited structures and attitudes that put little emphasis on membership, and they lacked professional management (Sizya 2001). Some primary societies began to operate as agents of private traders and rent out facilities to them. District unions operated facilities that, under liberalization proved to be unviable, such as cotton ginneries, oil mills, transport and hotels (Gibbon 2001). Government forgave 35 billion shillings of the 40 billion of debts the cooperatives had piled up, but this still left five billion to be paid back to the banks. All of this was exacerbated by increasing misappropriation by dishonest managers and committee members. Between 1994 and 2000 there were 262 C 2010 The Authors Journal compilation C CIRIEC 2010 cases of dishonesty waiting for police or court action (Government of Tanzania 2000). For instance, Mara Co-operative Union collapsed, owing to a high level of outright theft, and the underwriting of the expenses of one candidate in the 1995 general election (Gibbon 2001). At Kagera Co-operative Union (KCU) in the coffee growing area, after much evidence of mismanagement, members wanted to vote out the committee but regional government leaders kept it in power, and a government report was not made available for several more years. Added to this, there was factionalism among committee members because two main opposing groups were 'each fighting for a bigger slice of the co-op pie' (Banturaki 2000: 79).

A reforming act of 1991 should have helped. It recognized the ICA principles of autonomy and independence for co-operatives, and allowed primary co-ops to choose whether or not to belong to their district union. However, Gibbon says this voluntarization of co-operatives 'mainly tended to consolidate the space for the pursuit within them of private interests ' (2001: 396). It also set in train disintegrative tendencies whereby primary societies withdrew support for unions and set up their own competing unions, or became directly subordinated to private companies. In some areas, notably the tobacco, cashew and cotton growing areas, many collapsed; the number of primary societies decreased from nearly 9000 in 1990 to just over 4000 in 1994. The picture was not uniformly bad. Coffee unions began to benefit from fair trade arrangements with European buyers, and managed to hang on to market share; KNCU, for example, lost 20% of its coffee purchases to private traders in 1996, but in 1999 still had 70% of the market. New farmer associations began to be formed to take the place of the defunct co-ops, and a national level Tanzania Farmers Network (MVIWATA) was set up to represent them. The Sacco sector continued to grow, and by 1999 there were over 900 primary societies. The number of co-operative unions increased from 27 to 45, four new apexes were registered, including in 1993 the Tanzania Federation of Co-operatives (TFC), and in 1996 the new Kilimanjaro Co-operative Bank began doing business.

3

What is the co-operative reform process?

The reform process began in the 1990s, and was given impetus and direction by three developments at the international level. First, in 1995 the ICA provided a new identity statement and a new set C 2010 The Authors Journal compilation C CIRIEC 2010 of principles emphasizing the autonomy and independence of cooperatives (Birchall 1997: ch. 7). This was the international lead that co-operative promoters had been waiting for. Second, in 2001, the United Nations produced new guidelines aimed at providing a supportive environment (UN 2001). It called for laws that would safeguard the autonomy of cooperatives, provide light regulation, and ensure a 'level playing field' with other types of business, and declared boldly:

This process should have as its purpose the early and complete disengagement by governments from the internal affairs of cooperatives (UN 2001: 18).

Third, in 2002 the International Labour Conference adopted a Recommendation (No. 193) on the promotion of cooperatives (ILO 2001). Like the UN Guidelines, it drew explicitly on the work done by the ICA to reformulate the cooperative principles. It declared that promoting cooperatives as 'tools' of development is wrong; assistance has to be given to their members -to create income-generating activities, gain access to markets, improve their own social and economic well-being -while respecting their autonomy.

Since 1990, new cooperative laws have been enacted and old ones repealed in many countries. Fifteen sub-Saharan African states have revised their laws; others will follow. In India, in 1995 an Act in Andra Pradesh introduced, for the first time, the concept of cooperative autonomy and self-reliance; ten other Indian states then passed similar laws. Fiji, Indonesia, Jordan, Malaysia, Mongolia, Nepal, the Philippines, Thailand and Viet Nam have new laws, and several Latin American and Caribbean states are also revising theirs. Co-operators can now appeal to these international standards for guidance and to the ICA and the ILO for help (the latter has helped the governments of over 60 countries with drafting new laws and guidance on cooperatives). A new, more sophisticated understanding of the relationship between governments and cooperatives is emerging that is beginning to undo some of the mistakes made in the past.

The destination of the reform process can be inferred from the UN and ILO documents:

1. an autonomous co-operative sector, self-regulating and free to go in whatever direction the members choose, subject to cooperative principles 2. light regulation of the process by government, which registers and (in cases where co-operatives are corrupted) deregisters C 2010 The Authors Journal compilation C CIRIEC 2010 them just as it would any other form of limited liability company or association 3. an enabling environment for co-operatives that provides equal treatment with other types of business 4. a partnership between government and co-operative sectors in which supports are given so that the full potential of the form is realized

How does the reform process reach this destination? The first task of government is to pass a new co-operative law that reflects 'those features which best distinguish co-operatives from other forms of business organization' (Henry 2002: 8). These include the identity principle that establishes ownership by members, and the principle of member promotion that establishes that the rewards come to members as users rather than as investors of capital. Once co-ops are defined, the law should go on to provide rules about the management of co-ops, a surveillance committee, incompatibility criteria for board members, rules in external and internal financing, on education and training of members, boards and employees, on regular audit, on settlement of disputes and so on.

However, there are some practical problems in this legal reform process. First, the UN emphasizes the need for one consistent national policy, while the ILO recommends decentralization of policy and regulation to regional and local levels. We believe that, in the light of our case studies, this latter view is problematic as cooperative sectors that are permeated by local political interests and over-regulated by bureaucracies would much prefer a consistent, national-level regulatory body.

Second, the aim cannot be just to set co-operatives free of regulation, since freedom may lead to lack of transparency, corruption or the registering of 'pseudo-cooperatives' for short-term gain. The aim is to preserve the integrity of the legal form, to free cooperatives from unnecessary regulation, while providing a supportive environment. The concept of 'earned autonomy' is crucial here, since the consensus among our key informants is that there should be a grading system in which co-ops are more or less highly regulated depending on their performance on key variables such as degree of transparency, accuracy of accounting, inclusiveness of governance and so on.

Third, there is the question of whether one should aim for one general law or specialized laws for different co-operative sectors. The consensus seems to be that a general law is better as it establishes C 2010 The Authors Journal compilation C CIRIEC 2010 the membership-based form of business across sectors first, before applying particular regulations to each sector. However, in a situation in which vested interests are preventing a new law from being passed, it may be preferable to go ahead with a law driven by one sector; we recommend this for the savings and credit sector in Sri Lanka. Finally, even if a country achieves a reformed co-operative law there will be difficulties putting it into practice. Officials may see it as applying only to new co-operatives while co-op members may simply be unaware of it; an educational campaign is needed if it is to have any effect. Also, in some countries such as India and Sri Lanka there is devolved government and provincial governments have their own co-operative laws. If these remain unreformed, the effect of the national law may be minimal and only apply to national level co-operatives. Then there is the need to rewrite the byelaws of individual co-operatives, so as to guarantee good governance and a clear definition of the rights of members over their elected boards (Henry 2002).

The law on its own cannot force a co-operative sector to become democratic and a strategy is needed for turning unreformed co-operatives into new ones. This has two aspects. On the one hand, as Shah says of co-operatives in India, 'What is needed is a proactive, deliberate process of turning over government controlled co-operatives to their members ' (1996: 288). On the other hand, also needed is a strategy for business survival and growth so that reformed co-operatives can survive in the market place. Shah provides a useful scheme for organizational redesign (1996). There are three sub-systems that need to be strengthened: membership, governance structure and the operating system, and these are linked to three types of effectiveness: patronage cohesiveness, governance effectiveness and operating effectiveness. Getting the design of cooperatives right is fundamental to ensuring their success. However, this shows what should be done but does not show how to get there. Shah suggests a fourth design principle which he labels 'Secure, retain and continually nurture member-allegiance'. We would argue that this is not really a design principle but a reform strategy for sustainable change over time. Shah identifies important stages, such as a successful launch or relaunch of the co-op, creation and nurturing of credible expectations among members, an early demonstration of the advantages of being a member, and gradual growth of trust. This suggests a participatory approach such as we have developed in an 'umbrella strategy' for member development in large consumer co-ops in the UK (Birchall and Simmons 2004).

All of this assumes that the reform process is linear and that there are no vested interests trying to stop it. Such interests are not as strong as they used to be; the economic weakness and marginality of co-ops means there are fewer incentives to defend their position. Nonetheless, we can expect resistance from some quarters. Our key informants emphasized the need for a widespread and sustained programme of member education, particularly in situations where boards and managers are resisting reform. The key resource here is a co-operative college or NGO that can deliver adult educationtype programmes among members using grant funding from government.

4

Why has the reform succeeded in some countries and not others?

In this section we draw on accounts of the reform process in each of our case study countries to explain why reform has succeeded better in some countries than in others. Table 1 summarizes the elements of the reform process in Tanzania and Sri Lanka.

Table 1

operative Reform and Modernisation Programme. UNITED NATIONS ECONOMIC AND SOCIAL COUNCIL, 2001, Co-operatives in social development: report of the Secretary-General, New York. UNITED NATIONS RESEARCH INSTITUTE FOR SOCIAL DE-VELOPMENT, 1975, Rural Co-operatives as Agents for Change; a research report and a debate, Geneva: UNRISD. VERHAGEN K., 1984, Co-operation for Survival: an Analysis of an Experiment in Participatory Research and Planning, Amsterdam: Koninklijk Instituut voor de TropenUN 2001. WANYAMA F., DEVELTERE P. and POLLET I., 2009, 'Reinventing the wheel? African cooperatives in a liberalised economic environment', Annals of Public and Co-operative Economics, 80:3 361-392. WORLD BANK, 2007, Finance for All? Policies and pitfalls in expanding access, World Bank Policy Research Report: Washington. WORLD BANK, 2008, World Development Report 2008: Agriculture for Development, Washington. processus de réforme coopérative en Tanzanie et au Sri Lanka Cet article relate les résultats d'uneétude des coopératives au Sri Lanka et en Tanzanie au cours d'une période de trois ans. L'article pose trois questions: pourquoi les secteurs coopératifs ont-ils besoin de réforme; quel est le processus de réforme coopérative; et pourquoi les réformes ont elles eu du succès dans certains pays et pas d'autres ? L'article propose un bref historique des coopératives en trois phases: la période coloniale, la période post coloniale nationaliste et la période de libéralisation des marchés. Il indique que le contrôle exercé par les gouvernements coloniaux s'est accentué sous les gouvernements nationalistes, les coopératives devenant des parastataux. La libéralisation a permis un effort soutenu de la part des agences internationales pour réaffirmer la nature spécifique des coopératives en tant qu'entreprises possédées par leurs membres. Néanmoins, les coopérativesétaient mal préparéesà s'ajusterà un marché concurrentiel età la régulation du gouvernement. De nombreuses coopératives ontéchoué, certaines ontété corrompues et quelques autres sont C 2010 The Authors Journal compilation C CIRIEC 2010 devenues de entreprizes réellement contrôlées par les membres. L'article propose une analyse documentée et des interviews très informatifs fournissant un bilan sur le processus de réforme en Tanzanie et au Sri Lanka. Il apparaît que le processus n'est pas terminé et est souvent contesté. Genossenschaftsreformprozess in Tansania und Sri Lanka Dieser Beitrag berichtetüber Ergebnisse einer dreijährigen Studië uber Genossenschaften in Sri Lanka und Tansania. Es werden darin drei Fragen aufgeworfen: Warum bedürfen Genossenschaftssektoren der Reform, was ist der Genossenschaftsreformprozess, und warum war die Reform in einigen Ländern erfolgreich und in anderen nicht? Der Beitrag bietet eine kurze Geschichte der Genossenschaften in drei Phasen: der Kolonialperiode, der post-kolonialen nationalistischen Periode und der Periode der Marktliberalisierung. Es wird gezeigt, dass die Kontrolle, die von kolonialen Regierungen ausgeübt worden war, unter nationalistischen Regierungen verstärkt wurde, wodurch die Genossenschaften einen parastaatlichen Status bekamen. Die Liberalisierung brachte eine nachhaltige Bestrebung internationaler Organisationen, den besonderen Charakter der Genossenschaften als Unternehmen im Eigentum der Mitglieder wieder zur Geltung zu bringen. Jedoch waren die Genossenschaften schlecht darauf vorbereitet, sich einem Wettbewerbsmarkt und der Aufhebung staatlicher Regulierung anzupassen; viele scheiterten, einige wurden korrumpiert, und nur ein paar von ihnen wurden wirklich zu mitgliederkontrollierten Organisationen. Der Beitrag stützt sich bei der Berichterstattungüber den Reformprozess in Tansania und Sri Lanka auf dokumentarische Analysen und Interviews mit Schlüsselpersonen. Er kommt zu dem Ergebnis, dass der Prozess noch nicht abgeschlossen und häufig umstritten ist.

In Sri Lanka, the need for reform is obvious. In our interviews with key informants we asked how independent the MPCSs are from government. The latest Co-operative Statistics show that most hold regular elections and that a large proportion of members vote. In 2002, 136 MPCSs were due to hold elections and 120 did so (some were affected by the war in the North). Voting for boards is indirect, with delegates elected from the branches; 9692 delegates were eligible to vote, and 6815 did so, making it a 70% turnout (Cooperative Department 2002). Despite these impressive statistics, all of our key informants describe the MPCSs as government controlled.

Here are some typical comments:

The problem with that is when you have a lot of money in the co-operatives, the politicians jumped in. Politicians who are in the party in power control co-operatives.

You cannot become chairman of a co-operative in a particular village if you are not aligned to a member of the ruling party

When asked if MPCSs hold regular elections, one informant replies Technically they should, but I think it's ignored by the Registrar, particularly if a favoured group is controlling the society, they just leave them alone. Unlike the consumer arm of the MPCSs, the rural banks are doing quite well. They invest mainly in government bonds and lend mainly to urban businesses, so are criticized for recycling money from rural to urban areas. However, there is a worse problem; their lack of separation from the other parts of the business puts the savings of rural people in jeopardy. There is too much temptation to use the savings deposited to bail out the ailing consumer business. If depositors were to lose confidence, they might collapse. One of our informants confirms this:

Yes, they misuse the banks' money for consumer business, which is not a profitable business. They are always losing a little bit. With careful study, you find very little of the lending is for productive investments. They lend for housing construction or consumer activities, or pawnbrokery, not for industry or agriculture. Unfortunately, they behave like agent of urban elites, the collecting agent for the urban banks, in the consumer business for the importers of Colombo, so they have really forgotten their mandate. They should be an agent of the rural people, their members.

Despite the need for change, legal reform has not happened; all co-operatives are still registered under an Act of 1972Act of , amended in 1983Act of and 1992. There is a Ministry of Co-operatives, under which is the Co-operative Development Department. Co-ops are regulated at national, provincial and district level. There is a problem of overlapping national and provincial laws, and district commissioners do not apply them consistently. Co-ops are subject to annual audit, they have to gain permission to do anything new, and at the district level primary co-ops are subject to micro-regulation; at one primary society we were told they have to gain permission to write a cheque for more than 2000 rupees. As one of our respondents puts it:

Too much law, too much commissioners, too much ministers C 2010 The Authors Journal compilation C CIRIEC 2010

The 1972 Act brought in more regulation. On the one hand it is not enough to deal with serious abuses; one respondent explained:

You see what you could do as a Co-operative Commissioner or registrar, you can have inquiry, and then you report back to the co-op board. But who are the board members -the people who done all the errors, mistakes. And then they put it to the general assembly. Members are intimidated. You can't come; they will kill you if you go to that meeting, right. So then it usually gets passed, and nobody will be punished. The general body will say "Nothing wrong with this co-op" but you lost everything that you had.

On the other hand, genuine co-operative activity needs lighter regulation if it is to succeed. It needs a grading system that acknowledges good governance with freedom to develop. Also, we were told the power of the provincial commissioners should be produced in favour of a central regulatory body, because at the local level there is sometimes collusion between the regulators and the regulated.

In 2001, a presidential commission was appointed to report on the state of the co-operative sector. It reported back and drafted a new co-operative law, but the report was shelved. One member of the commission comments:

We followed the new Co-operative Act of India, Philippines, and some other countries. We studied and, according to our environment, we changed a lot. We took most of the stuff

from ILO Recommendation and UN guidelines and Co-operative Ministers' Conference recommendations. A lot of inputs we gathered and we try to localize those things. I think it's a very good move, and new thinking. It's a completely new law

Unfortunately it has still not gone to the Parliament.

In Tanzania, reform also began with a Presidential Commission, but it did not stop there. In 2000 a Presidential Special Committee of experts identified the constraints to the revival, strengthening and development of co-operatives: inappropriate structures, a weak capital base, poor management and theft, an inappropriate policy and legal environment, weak co-operative support institutions, confinement of the form to a few sectors, and lack of education and training. Its recommendations were all followed up. In 2001 the Department of Co-operatives was upgraded to a separate ministry of Co-operatives and Marketing, and in 2002 a Co-operative Development Policy was promulgated, with the purpose of 'enabling co-operatives to get back on to the development path and at the same time become more C 2010 The Authors Journal compilation C CIRIEC 2010 responsive to the needs of their members' (2002 p 4). The mission was to evolve co-operatives that are member based, work for betterment of members needs and of the communities in which they are situated, operate competitively, and reflect concern for present and future members. The policy demanded legislation and an implementation strategy, and so in 2003 a new Co-operative Societies Act was passed, and followed up by a rewriting of the co-operative society rules. A new code of conduct required that before people can stand for a board they have to have served three years as an active member and have a minimum of a secondary education. They have to disclose their personal property for each year of service on the board, and are limited to three terms of three years before they have to stand down. A similar code applies to staff, who have to have minimum qualifications and fixed term contracts. Both are liable for losses if found to be negligent. In SACCOs, a supervisory committee is required that has the internal capacity to supervise, rather than just relying on the Co-operative Department and the audit body, COASCO, to prevent misuse of funds. A mutual fund for members has been launched to protect primary societies against defaults on loans.

Implementation of these reforms is through a Co-operative Reform and Modernization Programme (CRMP) put together in 2004 by a team of 20 experts from government, NGOs, the apex federation, and academics, with assistance from the ILO. It represents a wide consensus of opinion about how to transform the new policy and supporting law into workable strategies and plans. It is closely linked to other government vision statements such as the Development Vision 2025, and to programmes such as the poverty programme MKUKUTA, and the agricultural sector development program, ASDP. The objective is:

to initiate a comprehensive transformation of co-operatives to become organizations which are member owned and controlled, competitive, viable, sustainable and with capability of fulfilling members' economic and social needs (CRMP 2004: iv).

The task is recognized to be large and urgent. The report noted 'the scale of embezzlement of funds and fraudulent practices in co-operatives has reached alarming proportions', fuelling public mistrust. There were delays in handling cases by police and courts, poor prosecution and corruption. The main problem was the district co-operative unions. In 2004, they had accumulated debts of Tshs 48.3billions (including 2.63bn owed to farmers, 2.69bn owed to employees, and 2.26bn to primary societies). They were regarded by C 2010 The Authors Journal compilation C CIRIEC 2010 banks as not creditworthy. The reform process was to promote good governance and accountability, putting the 2003 law into practice. It would deregister district unions that fail to change, and require all co-operatives in selected regions to call a special general meeting and elect a new board. The process would be monitored by the Registrar who would vet would be leaders.

The process has begun, but has yet to be extended to the whole country. The Registrar says the new elections have acted as a clear signal to the old guard and to members that things are changing. Members have reacted enthusiastically; in one case 1000 members turned up to the special general meeting. The reform process is to Empower members through member education projects. This refers to the rolling out of a pilot MEMCOOP project which, from 1996-2000 was active in the Kilimanjaro Region, and has now spread to a limited extent elsewhere. It is a capacity building programme through adult education, giving members the practical tools to evaluate the performance of their co-operatives (as a result of the project, 12 primaries left the district union, and took their members' business to a final buyer, more than doubling the price their members received). After primaries have taken control of the business process, they will determine what kind of secondary cooperatives they need; these will be a derivative of joint action, strengthening the business of the primaries. Opportunities will be available for improving coffee quality, searching for better markets, and negotiating for better credit terms. The reform process aims to transform co-operatives into viable economic enterprises. This means a process of determining the business needs of members, preparing development plans, implementing them, monitoring and evaluation. It will modernize Saccos, through improving capitalization, strengthening internal controls, increasing the range of products, improving outreach, improving knowledge of best practices in microfinance, and so on. It is recognized that a national co-operative bank is needed. Another aim is to reduce indebtedness of co-ops, through use of government export credit guarantees, a new method of financing co-ops using warehouse receipts as collateral, and a seed capital fund.

act as assistant registrars. Implementation of the CRMP may suffer if their role is not reviewed. The main training facility at Moshi University College (MUCCOBS) is the key to rolling out the member education programme, but serious capacity building is needed for training and research. The audit body, COASCO also needs capacity building, and help will be needed from several NGOs and from cooperative sectors in the North. A National Co-operative Advisory Council has been set up to advise government and provide a pool of experts. The CRMP estimates that $13.7 million are needed to carry it out. The government has promised TSh15 billions for first phase, and the write off of TSh23.8 billions of debts to central and local govt institutions. A report for donor agencies has been drawn up and 54 action lines itemized, but there is a need to mobilize development partner interest.

Why has there been so little reform in Sri Lanka and so much in Tanzania?

The contrast between Sri Lanka and Tanzania needs some explaining. In Sri Lanka, the new draft law has not been presented to parliament because of vested interests. One interviewee comments:

No, because here if a government minister gets hold of something, he or she will not let it go. These are power tools.

Another says

It didn't get into practice. They are not prepared to enact something that loses them power More specifically, the new law needs to be submitted by the current Commissioner, and he is reluctant. One interviewee explains After the changes the power of the Commissioner is reduced, so these bureaucrats are not willing to do that Another reason for the lack of willingness to change is ignorance of the international trend towards an autonomous co-operative sector. One interview puts it this way:

Unfortunately, recent commissioners and secretaries have not been aware of co-operative philosophy, the current thinking of co-operative movement in the world. They are just administrators, and they try to prevent change. Even some of the co-operative leaders, they're not willing to change, think the existing thing is better Another reason is that co-operative leaders do not want to lose the relationship with government, so they feel threatened by the new law. We interviewed the person who is blocking the new law and he explained that the existing act is needed for existing MPCSs, and that a new act is only needed for new national-level cooperative businesses, allowing joint ventures with the private sector. He foresees that in 20 years' time the 1992 Act will be obsolete but not until then.

Why is there no effective lobby for change? With the exception of SANASA, the apex organizations that should represent the cooperative sectors are weak. In 1994, a historian commented 'Their capacity to influence opinion at the higher levels and to influence decision making for co-operative development has been minimal' (Rajaguru 1994: 95). Not much has changed. One of our interviewees says:

There is no voice from the Co-operative Movement. The leaders meet the politicians and sometimes meet the president, and they want support from the government. They are begging for government support, that's the way co-op leaders behave in this country (apart from Sanasa). And the government also thinks they are a servant; if government needs some assistance to deliver some service to the rural masses, then they use co-ops as a vehicle. That is the attitude and mentality and that's how it operates here in Sri Lanka. Even politicians have realized the value of co-ops, but co-op leaders don't know how to show their strengths; even with the national budget, no proposals are come from the co-ops Part of the problem with the apex organizations is their lack of technical capacity. When we visited the National Co-operative Council they complained that they have no graduates working for them and only one computer in the entire building. Unlike Sanasa, they have no capacity for working with international NGOs and obtaining donor funding; they do not speak the right language or meet the right criteria. This is illustrated by the experience of an Israeli NGO that has brought funding from co-operatives for the Tsunami relief effort. After trying to work with the Institute and then the NCC for a year and 'getting nowhere', the Israelis had to shift to a partnership with Sanasa in order to carry out a programme of business training for co-operatives in the affected region. However, when we visited NCC had just hired two young graduates on short term contracts to manage a partnership with the Italian apex organization Legacoop.

The Co-operative Commissioner is another possible source of change, as is the Institute of Co-operative Management, but our interviewees complain that frequent changes of personnel and a lack of consensus about the need for change undermine the reform process. Another source of change is a microfinance (MF) bill being put through parliament which will give the Central Bank the power to monitor MF institutions. It is thought that the new MF Act will not apply to co-operatives, but in the future, if the rural banks are not managed properly they may come under the Act. Another possible source of pressure for reform is the buoyant and highly respected SANASA movement which, if it started taking the lead, and if it came together with related movements such as the Sarvodaya village development societies, could affect change in the financial sector.

What about the prospects for reform from below? There are some growth points in a few MPCSs that have branched out into new areas and renewed their democratic base. Some of the agricultural co-operatives in tea and coconut production are also models of good practice and are vital for the whole economy of their area. As one commentator puts it:

If SANASA society is doing this nicely, then others say "My god, that people are doing there, that so and so is leader, so and so is chairman, they are getting pride, people are respect to them. Why not we?" However, one respondent told us that in the MPCSs there is no energy for the kind of emulation that occurs in more lively sectors because the district commissioner has too much control: Why has the reform process gone so far in Tanzania? One reason is that there are no vested interests at national level prepared to defend the agricultural co-operative unions; their performance during the 1990s was so poor, and corruption so widespread, that a consensus emerged at the highest level for reform. Agricultural marketing co-ops are recognized as a vital part of the mix of measures needed to revitalize the rural economy, and so there is no choice but to reform them or substitute something similar. Another reason is that in development policy a lot of reliance is being put C 2010 The Authors Journal compilation C CIRIEC 2010 on the SACCO (savings and credit) movement. We were told that in two years from 2004 to 2006 their membership grew from 250,000 to 600,000 and new ones are being registered every week. The President wants to use them to deliver a promise of a 'billion shillings' of loans to each region. Yet they remain small and vulnerable and urgently in need of technical and managerial strengthening. From a politician's point of view, they cannot be allowed to fail.

6

What are the prospects for co-operatives in Sri Lanka and Tanzania?

In Sri Lanka, unless there is a change of personnel at the top it is likely that efforts at reform will continue to be blocked. The rural co-operative banks (RCBs) may eventually be delinked from the MPCSs, as part of a different reform process concerning the microfinance sector. It seems likely that pressure from the World Bank will continue and that the RCBs will become regulated by the Sri Lanka Central Bank. If this happens, the weaknesses of the consumer arm of the MPCSs will be revealed. However, it is unlikely that they will be allowed to collapse, since they provide the only network by which government can distribute essential goods in a crisis; government has recently recognized this by setting up a new wholesale society. Yet without basic reforms in the relationship between members and boards, boards and politicians, managers and district commissioners, it is difficult to see how their performance can be improved. They are, as one key informant put it, 'an unmanageable network'. SANASA societies will eventually escape regulation as co-operatives, either by sponsoring their own law or being taken under micro-finance law. New co-ops will most probably register as companies.

In Tanzania, the prospects for a co-operative revival are much better. Unlike in Sri Lanka, there are no legal or regulatory impediments to the development of savings and credit co-ops, and they are expected to continue to grow. The reputation of agricultural co-operatives is poor, and so in some parts of the country farmer associations are being promoted instead, registered as companies and with their own national federation. Yet in the coffee growing areas the district unions have been going through a process of revival, and with direct links with purchasers in Europe for organic and fair trade coffee their prospects are brighter than they have been for a long time.

Our key informants express several reservations about the progress of the reform process (CRMP). First, it is being led by government, and so is not independent. One of our informants sums up:

Before it can reform the co-operative sector, government has to reform itself It is trying to use a system built up to regulate co-ops to promote them. An education programme included training for co-op inspectors, but this did not work as at the time the inspectors saw no reason for it; their negative attitude towards co-operatives was too ingrained. The task of promotion has to be separated from that of supervision, and decentralized to a much more local level. Second, there are concerns about the capacity of the Tanzania Federation of Co-operatives (TFC) to drive the reforms forward. It has been described as a 'hanging structure' because most of the apex bodies that would be members have collapsed. Of the four apexes in coffee, cotton, cereals and tobacco only the last one remains. The TFC needs restructuring, but there is suspicion that the government will 'keep the empire going' rather than turning it into a lean structure responsive to primary co-ops.

Third there is the question of funding. The CRMP is getting $300,000 dollars a year from the government's general budget. There was a plan for a donor conference to support the program, but it was never followed up. There has been a shift in policy among donors towards support to the general budget of the finance ministry, so the CRMP is dependent on its annual allocation. However, one of our key informants commented that the budget allocation is enough and is supplemented by resources for promotion of co-ops in the cashew, cotton and tobacco areas where the co-operative sector had died out. Another budget has been provided to train a hundred agents per year for five years at Moshi College to work with primary co-operatives.

Fourth, the two main types of co-op face different pressures. In the agricultural sector, potential co-op members have become very suspicious of the whole idea. One of our informants put it this way:

Now the problem is . . . people who have already been bitten by a snake, they see anything they suspect -it's a snake

The loyalty of farmers, some of whom still have not been paid for crops sold to a co-op some years ago, will take a long time to get back. Linked to this is the question of governance in the two key coffee unions, Kagera (KCU) and Kilimanjaro (KNCU). There has been reform, and both unions are making an impact on farmers' incomes through fair trade. However, a careful watch needs to be made to ensure they remain well governed and managed in a 'lean' way so that the maximum of benefit goes back to the farmers. Dissatisfaction with the pace of reform in KNCU has led to 12 primary societies leaving and starting to do their own marketing direct with private buyers. It may be better if the old district unions are allowed to collapse, as they have done in Uganda, so that primary societies can organize their own marketing agency. In the SACCOs sector, there is pressure to deliver the government's plan to lend TSh21 billions (a billion to each region). This may lead to default on loans as people see the money as a hand out, just as they did in Sri Lanka with the Million Housing Programme.

The fundamental question is whether attitudes can change. Government has to learn to let go, to allow co-ops to be free to develop on their own, making their own mistakes and learning from them the hard way. Co-op boards need to learn that government is not going to bail them out when they get into debt. Member participation is needed at the very lowest level if the reform is to succeed.

Conclusion

This article has provided a summary of the history of cooperatives in developing countries in three phases, and has illustrated these with case studies of Tanzania and Sri Lanka. Then it has provided a detailed description of the co-operative reform process. Drawing on interviews with key informants, it has identified reasons why this process has so far succeeded in Tanzania but failed to get started in Sri Lanka, and has analyzed the prospects for continued reform in the future. There are two ways of looking at the reform process. It can be seen as a package in which all the elements have to be in place and the pace of change has to be carefully managed from above. Or it can be seen as a bottom-up process of member education and empowerment, in which apex organizations and government have the task of sweeping away obstacles and allowing members to pursue their own interests. They do this through their primary societies that deliver tangible economic benefits and support secondary co-operatives only when it is in the interests of their members to do so.

How typical are Sri Lanka and Tanzania of other countries that are engaged in a co-operative reform process? It is tempting to see C 2010 The Authors Journal compilation C CIRIEC 2010 them as two ends of a continuum with other countries being somewhere in between, but without doing more case studies we cannot draw this conclusion. We know that there are other possibilities that lie outside such a neat continuum. One scenario, provided by Uganda, is that apex co-operatives collapse and primary co-operatives rebuild their systems from the ground up (Mrema 2008). We might call this reform through structural collapse. Another, provided by Vietnam, is that a reform process is officially completed but in some regions the role of government remains pervasive and the 'new-style' cooperatives have not really become autonomous (Fforde et al. 2001). We might call this pseudo-reform. Our aim now is to compare many more countries on the elements of the reform process, in a rapid scanning process that will not depend on so much in-depth fieldwork but on expert consultations in each country. We hope that this will result in a heightened consciousness among policy-makers at national and international levels of the nature of the process. More speculatively, it ought to be possible for international bodies such as the International Labour Organization and International Co-operative Alliance to provide a grading system, similar to that being developed by the World Bank for good governance. In this way, a process of emulation might begin in which governments and cooperative sectors vie with each other to show they are further down the road of reform.

TANZANIA GOVERNMENT, 2004, Co-