There was general consensus that local socio-cultural norms and conditions had a major and direct influence and any proposal for SFGA formation should be based on an analysis of the local situation, making sure that it fitted into the local socio-economic environment. Collective activity was possibly more of a necessity in rural, remote areas than in the proximity of urban areas, although a major point here was that "collective action" was potentially useful to all disadvantaged groups in improving their situation; however, a significant corollary was that improvement in the condition of some could be viewed as a challenge to the status quo, and hence a threat to other, potentially more powerful, groups. In particular, the likelihood of resistance or economic or physical reprisal by more powerful forces had to be assessed. Land tenure could be a significant factor here. Building on traditional group-help practices could be very effective, such as in Indonesia, where the society was comparatively homogenous, but less so when SFGA formation would cut across traditional groupings, such as the biraderi in Pakistan, with the risk of competition or conflicting loyalties.

However, the general assumption that village traders, middlemen, commercial agri-business and large-scale farmers were automatically to be considered "bad" was not always justified. The negotiation of mutually beneficial working relationships between SFGAs and other commercial actors should be possible.

Understanding the kind of social capital that exists in an area, as well as its social parameters (ethnicity, religion, caste, education, etc.), was important, and essential in any SFG and SFGA promotion activity.

Problems arose in reconciling the SFGA concept with the established pattern of farmers’ groups in a society, particularly where farmers’ groups were used to channel services or for political purposes. The informal nature of an SFGA could be both boon and bane: it provided flexibility but precluded some activities, in particular in connection with obtaining formal credit.

Without exception, successful and sustainable SFGs and SFGAs were built on two cornerstones: (1) they satisfied members’ felt-needs first, not the needs of outsiders; and (2) they generated net positive benefits for their members, i.e., for each member, the economic and social benefits of participation in the group must outweigh the costs of that participation - without this there simply would be no participation.

Where traditional organizations existed, the formation of SFGAs was only justifiable if they complemented existing organizations. This complementarity came from the fulfilment of the functions that existing organizations could not perform, either because SFGAs responded to a different type of need, or because the existing organizations were inadequate structures for new functions, etc. With this approach of complementarity, it was possible to incorporate the SFGAs more easily and effectively in the institutional context by minimizing conflicts of interest.

To some extent, forming an SFGA was simply an extension of the logic behind forming its constituent SFGs. It could also be a reflection of changing realities in people’s circumstances. As their local socio-economic context changed, so they felt a need for new tools to deal with the world, and might need to adapt to different ways of doing business.

It is clear that stimuli for the initiation of SFGAs come from many directions. Top-down encouragement had been applied where a line ministry, donor agency or other entity saw an advantage for easier implementation of its activities, or if its political agenda required it to promote such associations. Bottom-up stimuli came either from SFG members themselves or from those associated with them, such as GPs. Horizontal stimuli came from other SFGs or the example of other local SFGAs. All these stimuli could be reinforced in various ways, either to meet a challenge or to profit from economies of scale. Experience indicated that for long-term sustainability, SFGAs should be initiated and largely financed from inside, with GPs or similar facilitators providing guidance. If SFGAs were promoted by outsiders, then more tangible technical support or incentives would be needed. However, outsiders often had their own agendas, and fixed targets and timeframes. Thus, in Indonesia, SFGAs were externally promoted as a matter of policy, but in fact received little in the way of effective GP input, and were slow to establish independence, while similar SFGAs in India flourished with effective GP input. In Zambia, some groups formed to gain access to credit for operations, although that was not the only reason. In fact, only about 10% of the 400 Zambian groups had external loans.

It was generally agreed that GPs or others involved in promoting SFGA formation should - in theory - be trained in the concepts and requirements for SFGA operation, so as to be able to advise wisely. However, in practice, field-experience was limited, and there was clearly a need for guidelines and associated manuals and training materials on the topic of SFGA formation.

Groups emerged and evolved when people had common problems and united to search for solutions. More often than not, the objectives were to fill in certain voids not catered for by mainstream development.

Outsiders could help a group identify its problems, and might also offer possible solutions, but if these outsiders initiated the formation of SFGAs, experience showed that the sustainability of the SFGA was more difficult than when the insiders of the local community initiated formation. The reasons being that there was always great pride in ownership in all that was "homegrown" and each individual felt more obliged to protect the well-being of the group by demonstrating solidarity. They were also aware that the decisions they took, the implementation of those decisions and the consequences would be their responsibility.

There was also a tendency for there to be too high expectations and hopes if the initiators were outsiders, and this could create some dependency, resulting in group members not maximizing on their own capacities.

A fundamental step was to study the local socio-economic context, and to plan tactically: focus on economic objectives first, then move to social and policy-related objectives when stronger and wiser, and cooperate with others rather than confronting them.

It was also possible to change the attitude of power-holders through example and to convince them that there was a positive-sum solution ("win-win") rather than a zero-sum ("win-lose") outcome.

SFGA formation process

The reasons were many and varied. SFGAs in Zambia had played a key role in information exchange, coordination of services, and conflict resolution between groups, something that single SFGs could not do, but their functions were both social and economic. In India, the reasons for forming SFGAs varied from area to area and from one social category to another. Among the scheduled castes/dalits, they were formed primarily to establish social identity and to work against untouchability, and, secondarily, to obtain governmental assistance for undertaking income-generation activities. SFGAs among the scheduled tribes were formed primarily to protect the tribal land and counter exploitation of tribals by non-tribal moneylenders and petty businessmen. Farmer groups among the backward castes were formed primarily to cater for agricultural credit needs.

Economic reasons for forming SFGAs included group purchasing of inputs, communal grain storage, coordinated planting and selling of harvests, and group savings activities. Social reasons included increasing power, fighting for justice, improving a sense of community and common purpose.

As the role of peasant organization for land reform had shown, recognizing contradicting interests as such, and creating grass-roots mobilization to bargain for just and legal reforms, which change the existing status quo peacefully, could actually strengthen collective solidarity and lead to considerable self-reliance of poor peasants. Experience showed that for groups of poor peasants, having a "common enemy" could be a strongly motivating and solidifying force. Overcoming injustice (e.g., lack of land tenure) or exploitation (merchants, moneylenders) was often a more strongly felt motivation than just economic growth, though it did imply more risks.

Groups emerged and evolved when people had common problems and united to search for solutions. More often than not, the objectives were to fill voids left unfilled or not catered for by mainstream development. Group formation from the inside was need driven; group formation promoted by outsiders could be for a range of reasons, ranging from idealism, through altruism and hard economics, to bureaucratic control and political power. The question should be asked whether SFGA formation was promoted for the benefit of the promoter or of the SFGs!

There could be an ideological conflict when the establishment of SFGAs made them part of the competitive market economy: the conflict between whether the SFGA should pursue a "community development" versus a "commercial development" objective was often more a problem for the support actors than for the SFG members. Many development-oriented organizations (most NGOs) initiated SFGs with the objective of empowerment and equity. Much had been achieved in this respect. However, the same organizations often had problems in integrating these groups into the mainstream (often competitive and capitalist) economy, for two reasons:

• many development-oriented organizations found it difficult to see SFGAs adopting the same kind of rules and strategies as the competitive market economy, which the support organization originally perceived as one of the reasons for the problems justifying SFGA formation; and

• while an organization might be good at facilitating the formation of SFGs and SFGAs, that same organization might not have the qualifications to support change towards a business-like SFGA organization.

Among some external promoters, an often-stated - but seldom achieved - goal in forming SFGAs was to achieve organizational autonomy (although other external promoters - rhetoric aside - seemed more interested in promoting dependency, while yet others focused on social mobilization purposes or reduction in the cost of service delivery). SFGAs could not rely eternally on support from donor agencies through NGOs or other sources. Although such assistance had to extend over a long period to have an impact, it should end, and would definitely have to end one day. If farmers were not engaged in economic activities to generate resources, and mechanisms introduced within SFGAs to capitalize and use some of their services or income to finance growth, then the sustainability of the whole development process was at risk. The role of development activists was therefore to assist them to improve their outputs through targeted human resources development initiatives that would address the issue of rural leadership development, knowledge acquisition and skills development. It meant developing the financial mechanisms needed to mobilize group savings and capital, and to gain access to external credit. It also meant developing the organizational capacities to manage SFGA operations and services such as inputs supply and provision of services. Participatory and coherent methodologies were needed to enable SFGA members to achieve their development aspirations, based on principles of local control, cooperation among SFGAs and self-reliance. Such methodologies were not available in an easily accessible form.

Ultimately, the reason for SFGA formation would be that the perceived benefits outweighed the costs. In this context, "benefits" covers not only direct measurable economic returns, but also indirect, non-individualized factors, including social capital, benefits in kind, and altruistic community development considerations.

Opinions differed, and the answer was probably that it depended on the local circumstances. Many SFGAs were formed specifically to address a particular issue, and, if successful, might expand into additional relevant areas, or fade away having completed or achieved their aims. Under other conditions, an SFGA might be formed to pursue multiple activities. There were clearly a plurality of approaches to social organization. Once SFGs had identified the common ground to form SFGA, then an effective SFGA was likely to be formed. Because the reasons could be more than one, the SFGA could address multiple issues, but the general consensus of the conference was that focusing on too many objectives simultaneously often led to management problems. Certainly, many of the successful SFGAs were those focusing on core livelihood needs, and the term "core livelihood needs" could well have included a number of disparate but inter-related activities. Thus, from a single target, they developed to deal with a number of associated - often complementary - tasks, but with maturity and growth would probably be better specializing in their core competences, for several reasons:

1. concentrating on too many issues at once, or trying to provide too many services to member groups, could create problems for the SFGA;
2. it could disperse energies;
3. it could lead to conflicts between those member groups wishing more emphasis given to one issue or service and those wishing more stress on another; or
4. it simply might be more difficult to manage and control multiple activities rather than a single one - at least from an accounting and financial control point of view.

In Zambia and Sri Lanka, this problem was partly solved by setting up function-specific committees within the SFGA to manage the various SFGA activities, but the level of success was unclear.

It helped to start with a more narrow focus, where benefits were quick, tangible, and relatively likely, and then to expand on that foundation. But beware of overconfidence. Some SFGAs started with a narrow focus, but, stimulated by one or two successes, quickly expanded to many activities, too quickly, resulting in a sort of "organizational meltdown" and consequent failure. Diversification of activities and services should be done in a careful manner.

There was a need to differentiate why SFGs were established in the first place. Excluding those formed just because of external incentives, it depended on the local situation. Three reasons stood out:

• Empowerment  A marginalized portion of a local society joined into groups. Most of these SFGs were promoted by social intermediaries (often NGOs). The first objective of SFGA formation was the further empowerment of that portion of organized society. This was most relevant in highly differentiated societies where there was little participation of marginalized people in development activities. Many NGOs in Nepal, for example, saw themselves as social intermediaries. Training and awareness building was the first major initiative being addressed. Most of these groups ran a high risk of being dominated by the local promoter.

• Access to savings and credit  Savings and credit groups were another common form of SFGs. Many outside institutions used savings and credit as the first entry point in SFG formation and used group savings mobilized as a test of the level of commitment of people joining the group. This was most relevant in areas of poor infrastructure and a poorly developed rural banking system. However, too many groups gave too much emphasis to access to external credit as the main stimulant to group formation, with less attention paid to the savings side. Such approaches might stimulate rapid SFG formation but provided little hope for sustainability.

• Production and marketing  Commodity-focus groups were formed around a common interest in generating improved income through the joint production and/or marketing of a commodity (coffee, soybean, handicrafts, etc.). This was most relevant in areas where new commodities were taking off and where middleman competition was weak, or where benefits from production suffered from being disconnected from markets (middlemen, information, etc.). These groups ran a risk of forming around an offered service, such as subsidized input supply, transport, etc.

In all cases, the final goal for development in the rural area was a better livelihood, which also meant income-generation. This required meeting all three goals to the greatest extent possible. The SFGs formed for one purpose could not necessarily automatically take over or agree upon another purpose. Thus, the formation of SFGAs might take very different forms and was a dynamic process depending on people's interests.

SFGAs never saw their development problems in disaggregated, sectoralized form, but rather as a mix of social, economic and cultural factors. A major issue here was that development assistance agencies were sector minded: one agency was interested only in emergency assistance, without thinking of what should follow when the emergency receded; another would say that it was interested only in programmes not in projects; others would come with stimulating messages on cassava, rice, cotton, coffee, etc., undermining the accompanying measures that, if put together, would effectively change society for the better. Obviously, one single agency could not address all issues and provide all assistance needed, but dialogue, linkages, networking, technical coalitions and coordination of these links could help develop appropriate mechanisms to face such situations. SFGAs were useful institutional mechanisms for developing such links. In the days of globalization and rapid changes in all spheres of human life, acquisition of technical skills was the key for the survival of small-scale producers once they grouped themselves into an SFGA.

The contributions to Question 6 tended to demonstrate situation-dependence. If the issues all concerned more or less the same people within a small locality, then choosing to set up several SFGAs, each focusing on a different issue, would probably result in there being overlapping memberships by single individuals, or an SFG belonging to several SFGAs at the same time. That could be problematic, but need not be. A typical farmer in a developed country might belong to numerous organizations in their community: the local church or mosque, a marketing co-op, a social club or society, etc., each providing different functions. That was the sort of "social capital" that Putnam and others talked about. Such a process of organizational de-concentration and specialization might even lead to more efficient activity control, accountability and improved performance.

A very important aspect of Zambian PPP group development was that PPP groups transformed from a "reactive" to a "pro-active" attitude towards development, with important help from locally-trained people (mostly women) acting as GPs. Rather than waiting for assistance of all forms - especially credit - groups were stimulated to identify what they could achieve themselves, taking control of their own situation.

There was awareness in PPPs of a need for a build up of the economic capacity of SFGAs: what was missing was the knowledge of how to do it! Hence, the co-financing of SFGA activities by the member groups had been a problem in the absence of a built-in strategy and mechanism for financing and sustaining SFGA activities. This was one very good reason why guidelines for efficient functioning of SFGAs were needed.

The pre-conditions listed in [TP-50] were considered absolutely necessary for success. Sometimes, other priority needs of the small farmers, such as solving land ownership problems, legal restrictions in access to subsidies and credit, were decided at district or lower administration levels, and were best tackled only through higher-level organizations (i.e., SFGAs).

Problems often began when group promoters, eager to meet predefined quantitative targets, pushed ahead with SFGA formation before a sound foundation of self-reliant SFGs had been established. A premature start for an SFGA could be very detrimental for the constituent groups. The question then became: "When were SFGs mature enough for an SFGA? What were possible indicators?" One answer was "When SFGs have accumulated sufficient savings of their own and management skills to operate without the continued presence of a GP." Another significant indicator was that some neighbouring SFGs had already had informal meetings and engaged in some inter-group activities, e.g., for obtaining inputs, storage, processing, marketing, etc.

Most important of all, of course, was that candidate groups were really willing to constitute an SFGA, and that they were prepared to pool their capabilities, facilities and some of their savings for this purpose.

It was obvious that small groups of 8-15 people could do very little to change the course of events in society, but the right moment to form inter-group associations was when the group members themselves felt the need for it and had the capacity to do so. Remember that one needed to demonstrate something to inspire others. There was no one right time, as so many things had to come together. However, seeing was believing.

Question 8. Who can be a "member" of an SFGA? A group, an individual or both? Should the rights and duties differ?

Some participants felt that the SFG should be represented at SFGA level by its office bearers.

The constituent SFGs were obviously members, but whether membership should be exclusive to them, or whether the SFGA could be an independent body formed by the SFGs but open to non-SFG persons or other entities, was unclear. The division between cooperating SFGs and formal cooperatives blurred frequently in discussion, and the formal/informal parameter was ignored.

There was little direct discussion on this point; and the limited references implied an understanding of the term SFGA broader than the conference’s definition. However, it was felt that identifying the most appropriate representational mechanism - with its implications for "membership" categories - was important, and often a key factor in ensuring SFGA sustainability.

Question 9. How should SFGA leaders be chosen so that they do not lose contact with and accountability to the base? What are the advantages and disadvantages of "shared" leadership?

The election of SFG leaders by SFG members seemed straightforward, but, at higher levels, the base group members might not personally know about the capabilities of the members of other SFG s, possibly leading to the development of partisan attitudes. Probably this type of behaviour was inevitable in any type of organization. The mechanism adopted should try to discourage the formation of an upward-looking leadership elite.

The question had to be considered carefully due to the potential for internal inconsistency. Some said that an SFGA representative should not be an SFG office bearer. Other felt that SFGA office bearers should be elected by the totality of SFG members. Problems arise as many SFG members had little knowledge of potential candidates, and there was some debate over the merits or otherwise of electioneering. The simplest representational pattern was indirect selection, with the SFGs selecting representatives by consensus, and these persons then represent members from below at higher levels, rather than being selected by some large group where the bonds of accountability are attenuated because people do not know each other as intimately, and cannot observe on a day-to-day basis how faithful the leadership is being to the group and its ideals.

Selection by consensus was better rather than election (secret ballot or not), since when there were elections, this tended to divide people (If you know which 9 people voted for you, you also know which 7 others voted for me, and you will not feel as responsible to the latter as to the former; if all 16 agreed that you would be the representative, you are beholden to everyone.).

The organizational process did not necessarily need to commence by choosing representatives (leadership), but could start with doing activities on an informal, ad hoc basis, to learn which members were the most dedicated to group improvements, and which had the most skill. By "working first and organizing second," you could identify some new, appropriate and committed leadership, rather than if one started with selection (elections), and invariably "the same old leaders" were chosen again.

A special feature of the "shared leadership system" practised in many PPP and SFDP projects at group level, at least for an initial period, was that of frequently rotating leadership, which allowed each SFG the opportunity to identify the capabilities and skills of the individual small farmers, so that the most capable and skilled people would be selected for appropriate positions on a rotating basis. However, shared leadership at higher levels could lead to discontinuities in SFGA management. In Indonesia, it was observed that SFGA leadership very rarely changed, because often the chairperson was also a formal leader, such as a village head. Transparency needed to be built into the management function, whether the leadership was shared or not. It was very difficult to make group members realize the importance of continued vigilance, which was crucial for efficiency of SFGAs. This is precisely why an understanding of the socio-political situation and its influence is so crucial.

Shared leadership was a good way to give people a chance to demonstrate or acquire their leadership skills, but untried leaders had the potential for disaster. Also, in rural societies, the concept of elections could be foreign. In cases where, for instance, there was only one literate person, then that person would be a strong candidate. Again, voting could sometimes be divisive, so it was better to select leaders by consensus.

One means of getting the kind of leadership needed for SFGAs to be successful was for each group come up with its own "job description" for leadership positions, stating what is expected of the person (some degree of education, an even temper, no partisanship, no bad qualities like drinking too much alcohol, etc.). This narrowed the pool to persons who met consensual criteria, and, as a result, anyone selected was well aware of the expectations of their peers.

SFGA leaders should be chosen from known good negotiators who genuinely knew and had the interests of the constituent SFGs and their members at heart - they might deviate later but members had to take that chance. These leaders were important because they kept open the channels between the grass-roots membership and the traditional authorities and government. They should be able to convince the powers-that-be that the SFGAs were there to help and unite.

Several participants stressed the necessity for ensuring transparency in all SFGA dealings, with accountability to all SFG members, yet there was little discussion of suitable mechanisms to ensure such accountability. It was suggested that this could be a topic for future discussion.

The "factors" had to be considered in relation to the local context. This would vary with each place, activity, or purpose of the organization. Avoid rigid formulas. The social context would determine the probable success or otherwise of mixed versus single-sex groups, and the likelihood of women getting a fair hearing in discussion or being able to participate equally in administration and leadership. To promote equality might require dual approaches: women-only groups for activities, and sensitization training for the men in mixed groups or SFGAs. To promote the more active involvement of women in SFGA activities implied that they were also active at SFG level. This could well require assistance from GPs to instil in women skills associated with leadership, mixed-meeting participation, how to approach male officers, etc. Confidence building was an essential element in such skill development.

Certain facilitating actions might have to be taken to assist the participation of women in SFGA meetings, such as child-minding arrangements. Certainly, the participation of women would only be seen when they were convinced that the benefits from participating in the association outweighed the opportunity costs of loss of time devoted to other activities. The role of the GP as facilitator could here be crucial.

There was evidence that women-only SFGs had greater ability to generate internal consensus and to minimize female "free-riders" among their members compared to mixed groups, but the evidence did not yet extend to SFGAs.

Homogeneity of membership - implying that all members had similar socio-economic characteristics - was a considerable positive factor in establishing SFGA solidarity. That homogeneity might also apply to having needs in common, or simply to needing unity to survive in isolated conditions.

Women also faced numerous kinds of repression which were gender-based and against which they were powerless unless they were organized to pressure the state for change, or resist their oppressors. Women often needed to organize in order to address both their day-to-day economic needs and constraints and to improve their social and/or legal status. Experience showed that the more successful organizations able to enhance women's capacity to meet their needs effectively were women's own organizations. It was only after women had been able to consolidate their position through their organization that they were able to effectively participate in wider, mixed, organizations, and influence their functioning. A good example of this could be found in West Bengal in the women's forest committees. Where such committees had been formed without the women having gone through any process of struggle in organizing themselves for their needs, their participation in forest committees had been weak and ineffective. However, where the reverse has happened, women had been effective whether through joining mixed committees or by forming exclusive women’s committees.

Addressing coverage, a critical mass was required in the SFGA in order to be able to generate income greater than costs from its operations and thus be successful and sustainable. Geographical separation was also a factor. In areas where population density was low and scattered, as in many areas in Western Zambia, distances limited the number of groups or the size of the SFGAs to those who could easily be serviced by that SFGA. If insiders formed SFGAs, it was in the absence of one major problem: distance between SFGs, with the attendant difficult communications between people, the enormous efforts needed to transport goods and people to and from markets, to and from collective bargaining events, and so forth.

Homogeneity was paramount at the SFG level; the smaller the number of members the better, so that everyone could understand the process, become really involved in it and then "own" it. Groups could be mixed, but - at the lowest level - women and other social groups in society should be given the chance to create separate organizations so that they could address their specific problems. Inter-group associations would then provide the mechanism for putting community global interests together.

The introduction of gender issues into the training for GPs and thence for SFGA group leaders and members should help ensure better balance in group composition and activities.

The general balance seemed to weigh on the side of GP as facilitator rather than as instigator or active promoter of SFGA formation. The role of the GP in the formation of the SFGA should be that of promoting inter-group contacts and cooperation linkages, i.e., encouraging an environment conducive to SFGA formation, and explaining the advantages and disadvantages that SFGA formation could have for their groups. The GP could facilitate at the meetings for the formation of the SFGAs. Others felt GPs should be both initiators and facilitators.

GPs should be selected carefully from socially-sensitive persons and they should be always aware that they were there only as temporary support, and that they would have to gradually withdraw from the SFG and SFGA. Success assessment should be based on SFGA performance once GP support had been withdrawn, not on numbers of SFGs and SFGAs formed in a particular period. At the same time, the groups should have some assurance that the GP would remain alongside them for the time it took to become established.

The problem of training GPs was highlighted, where poorly-trained promoters did not distinguish between their role as facilitator and initiator or where a rapid turnover of GPs lengthened the SFGA establishment process, or both. Another parameter in the training conundrum was "Where do you find the GP trainers with field experience in SFGA formation?".

The "How" of Question 11 was less addressed. Raising awareness of the potential costs and benefits of association amongst SFGs was possible, as well as following up on spontaneous inter-group activity, and providing advice upon request. Overall, the initiation of an SFGA should be needs driven, and that "need" must be an intrinsic recognition of problems and the possible routes towards solutions.

Creating an enabling environment was seen to be important, and here the GP could help mediate with outsiders - whether private or public sector.

As an overall tenet, it might help for the development agency to think in terms of "providing a service to develop the potential for choice in the target group." This might not be as simple as it sounded, but did force a full-perspective view of a particular situation. Investing in a particular development programme might be of no use if other factors were not in place that supported this element of choice. Also, the exploitation of choice was dependent on knowledge. Training, education and organization were important factors, but interaction with outside agencies, communication development and information dissemination (particularly market information) should not be overlooked. Two other very important dimensions related to independence or autonomy of action and financial resources and other assets. People with higher levels of financial reserves and assets generally had more freedom of choice.

GPs should provide appropriate information to the SFGA to support its dialogue and lobbying with external forces and institutions. GPs also had an important role to play in mediating in power struggles and assisting in conflict management within the SFGA or within or between SFGs, but eventually methods are needed for the SFGAs themselves to handle such situations. The GP should work tirelessly towards building up a sense of local ownership of the process being put in place.

SFGA organization and management

There was general consensus that internal resource mobilization should be encouraged to facilitate SFGA development: the problem was that it was not always being implemented. Such contributions should come primarily from member groups; external resources should only be resorted to as a temporary supplement to what the SFGAs could mobilize from their member groups, and ideally within the context of a well-defined, full-self-sufficiency strategy. This was considered to be the best method, since groups were more effective at raising income together at group level than at the SFGA level. Several activities could be organized in different groups, and monitored by the SFGA through reports and actual field visits. Internal resource mobilization should be a key emphasis area in the development of SFGAs. Equity or savings mobilization must start from day one. The initial priority should be given to internal resource mobilization. Support from aid agencies should be in the form of supplemental loans to match SFGAs own capital to expand their activities: outright grants too often resulted in dampening of the initiatives of SFGAs. Emphasis should be placed on internal resource mobilization.

Regrettably, among questions left unanswered were the "How?" questions regarding the best way to mobilize internal resources at SFGA level, to ensure transparency in accounting, and to safeguard, protect and manage funds mobilized.

Member groups with adequate savings at their disposal certainly had the potential to play a more vital role in SFGA decision-making and in the financing of SFGA’s joint activities because they were already convinced of the advantages of mutual help, and had mobilized internal resources to help finance their own self-help activities. The self-reliance and sustainability of groups was directly proportional to the extent of their saving ability. When individual groups with strong savings programmes failed to finance bigger projects, they often realized the need for an inter-group association. They needed some technical as well as financial help from other groups and showed greater interest in forming an alliance. A minimal availability of some additional, reasonably priced, external resources could be a positive factor in this case, but there was consensus that too much external financial assistance did more harm than good.

External financing should only be used as an instrument to induce internal resource mobilization, possibly through "matching loans" [repeat "loans"] that were not more than two-thirds of the amount put up by the SFGA, i.e., the SFGA should have a saved capital equal to 60% of the total cost of a proposed project in order to apply for a loan equal to 40% of the total cost. Such loans should be offered only to selected SFGAs with a good repayment record or whose constituent SFGs had good repayment records. Regrettably, such fiscal rigour was rare in the aid world: more common were loans that were a multiple of savings - double, triple or even greater multiples - and thus outweighed the savings element, overwhelmed the borrower’s ability to service the loan and led to financial problems. Such leveraged loans were not considered prudent banking practice in the developed world, so why should they be condoned in the name of "Development"? Resource mobilization was not that difficult if the people recognized the purpose and ultimate use of that activity. However, it had been found (in Pakistan), where community savings funds were established from the unequal contributions of many community members, the internal lending of these jointly held funds to individual community members was not generally approved. The primary reason was that the savings of each individual in a village organization varied. If this amount was lent out, the risk to different individuals was going to be different. There were serious questions of what villagers considered equitable treatment involved in this. In contrast, when savings were raised in equal amounts by villagers for maintenance of infrastructure, such as a micro-hydel, there were no problems in using it collectively. Thus, the way internal savings were mobilized at group and inter-group level did matter. Moreover, the way in which external funds were mobilized and linked to internal savings also mattered. There was a need to help develop the capacity in communities to raise resources both internally and externally.

Although not discussed further, the point was made that non-cash savings should receive greater recognition in resource assessments: in rural areas, grain and livestock represented very significant savings in kind.

There was no consensus among participants. Some related points were raised:

• A service cost should be included in the transaction cost (such as fertilizer supply), and a part retained by the SFGA. As the confidence of the member groups grew, the member groups could be charged more directly for services received.

• The ability to stimulate self-reliance through membership dues was crucial for an SFGA. It was an essential tool for mobilizing SFGA funds, both ex ante and post facto. Regular payment of dues was an important acid test of SFGA sustainability and group solidarity. If members continued to pay dues, it implied that they perceived the SFGA as delivering a value-added element that was of benefit to them.

• Fees and charges should reflect the costs involved in providing services. There should be volume and early purchasing discounts, where appropriate, but no cross-subsidization of activities.

While membership dues could cover incidental running costs of the SFGAs, they would seldom be sufficient to cover larger, growth-related expenses.

Ploughing back a part or all of the profits from savings on group purchases or group marketing was one possibility.

Accumulative Saving and Credit Associations (ASCAs) were another more useful and comparatively advanced method of saving at small-group and inter-group levels. Individual SFG member or SFG contribution were collected and deposited in a joint account of the group in a commercial bank. A part of this amount could be borrowed by an SFG member, or an SFG in the case of an ASCA established at SFGA level, to be repaid in three to four instalments with some predetermined fixed interest rate or service charge. However, such informal saving methods were complicated to operate at SFGA level. It was easier if all constituent SFGs contributed a fixed amount every month (or every meeting) and the pooled amount was kept in a joint account. This pooled fund could act as a basis for internal lending to finance projects of individual member SFGs.

As SFGAs grew, there was often a need for longer-term member contribution or equity outlays, which conferred some right of ownership and control over the SFGA and the right of limited compensation in the event of withdrawal. Putting in place an acceptable financial system at SFGA level required confidence building amongst SFGs and rigorous transparency in all financial dealings at SFGA level. This was particularly so if SFGAs were to be promoted on the basis of their relying on internal resources.

Poverty being a multifaceted problem, participants agreed that SFGs and SFGAs needed to be geared to meet a multitude of member needs. Then only would the poor feel the benefits.

When SFGAs started out, there was a strong argument for allowing them the opportunity of addressing a number of issues or activities of interest to the affiliated member groups. With experience, as they matured, they might wish to concentrate on fewer activities, because trying to concentrate on too many issues at once, or trying to provide too many services to member groups, could create problems for the SFGA as:

• it could disperse energies;
• it could lead to conflicts between those member groups wishing more emphasis given to one issue or service and those wishing more attention paid to another; or
• it simply might be more difficult to manage and control multiple activities rather than a single one - at least from an accounting and financial control point of view.

Suppose an SFGA were involved in providing marketing, savings and credit, and input purchasing services to its member SFGs, and also suppose that the SFGA incurred running costs in providing each of these services and might, in the particular case of providing marketing and input purchasing, have some working-capital requirements: then what mechanisms could the SFGA use to raise the necessary revenue and working capital to cover those outlays? Should all these cash needs be covered by charging all member SFGs a global service fee to cover all services, or should only the users of the particular service be charged in proportion to their use of that service?

There was the dilemma of simplicity (all bore an equal share of all costs) versus equitability (each paid according to use). If there were a desire for transparency in transactions, then the complications of cost-centre-based accounting would be needed, but that had the advantage that cross-subsidization of loss-making activities would show up more clearly.

Was it better to encourage the SFGA’s management to handle multiple activities (services) altogether or was it better to adopt a compartmentalized service approach, in which the SFGA management established a set of special technical service committees each one responsible for managing one particular SFGA service? The latter seemed more practical. There was always the possibility that non-core SFGA services could eventually be "spun off" as the SFGA matured, leading to the creation of new SFGAs (possibly with separate leadership but maybe with some overlapping SFG membership).

Functional Service Committees progressed well for some time in Sri Lanka. Continued support from line agencies/service NGOs seemed to be a factor that determined sustained interest of the committees. When they found that the anticipated support was not forthcoming, it was natural to expect some loss of interest on the part of the committees.

Saving and credit groups might evolve into a cooperative banking system. As they evolved in this direction, the specific skills needed for this process needed to be supported. The same groups should not be asked to initiate a common production or marketing goal.

SFGAs might have different functional development patterns depending on the circumstances:

• Path A: Single-purpose SFGA (e.g., credit and savings) > co-op-type formalized credit union or bank.

• Path B: Multi-purpose SFGA (e.g., mix of credit and savings, input procurement, marketing of produce) > a complex, multi-faceted organization, requiring greater management capacity.

• Path C: Multi-purpose SFGA > spin off activities into associated single-purpose SFGAs.

Production and marketing groups should form cooperative production and marketing associations, and consider hiring technical staff to support the members.

Savings and credit should be separated from production and marketing to avoid demand for credit occurring simultaneously (e.g., before coffee harvest) by all members, and the failure of a crop (e.g., because of a hailstorm or pest) would not cause the immediate crash of the bank-SFGA.

The key was a clear vision for their objectives by SFG-members and assistance in differentiating between specific functions.

A critical determinant of an SFGA’s ability to handle multiple tasks effectively was the availability of skilled manpower. If the skills were not available, then they would have to be acquired, and here the aid agencies could help through human resources development. This would argue for taking an in-depth look at the organizational and functional evolution of a sample set of SFGAs and their related socio-economic settings to identify common patterns of development.

For SFGAs to be effective, there should be rules, regulations and procedures - just as there were for the groups - to govern basic functions and activities. The rules were often determined by the legal status of the SFGA: whether it was formal or informal. Some SFGAs might remain informal for some time. These rules could be informal or formal, in written or oral form. Oral rules - reinforced by mnemonic devices such as songs, poems or repetition - were obviously more applicable to situations of poor general literacy and numeracy.

There was need for a system of keeping of accounts and keeping of minutes, with decisions arrived at, etc.

The basic principle that an SFGA should be fully accountable to its member groups was of utmost importance. An SFGA should represent groups and not become a sort of semi-autonomous executive body. This then implied rules to ensure the necessary transparency of SFGA transactions and accountability of SFGA leaders to the base membership. The stage of development at which an SFGA should prepare a set of written rules depended on various factors, most of which were location-specific, such as levels of experience and of literacy, SFGA objectives and activities desired, and so on. However, a basic pre-determinant would be that most of the members of an SFGA had expressed a strong need for written rules and regulations. The preparation of written rules was usually considered a time-consuming, trial-and-error democratic process. It might start with some draft key rules and procedures, to be extended gradually as and when the need arose, until a final set was evolved. The introduction of written rules pre-supposed literacy; if literacy levels were poor, then the first step ought to be to introduce literacy training.

An SFGA’s rules, whether written or unwritten, should be based on local values and social norms.

If an SFGA engages in commercial activities, such as wholesaling, then it would almost certainly have to abide by local legislation, including accounting and taxation systems, which de jure required some adherence to written rules and contracts.

This Question stimulated considerable discussion, although the term "M&E" seemed variously to refer to external M&E of SFGA development or to internal controls for SFGA management purposes. The preconceptions of participants coloured attitudes towards M&E: whether the stimulus for SFGA formation was intrinsic or extrinsic. This was a seen as a key element in SFGA operations, but one where little practical experience seemed to be available. Most M&E proposals put forward in the conference were oriented from the point of view of external actors wishing to assess the success or failure of projects. For M&E to be a sustainable element in SFGA activities, it would have to, above all, serve internal information needs, not the reverse.

In Zambia, it had been the practice for the SFGAs to have quarterly meetings and an annual general meeting for reporting, monitoring and evaluating activities. Functional SFGA committees were often established to monitor the constituent groups, and normally visited the groups at least twice a year, or more often upon request or when there was a problem. Simple books of accounts - showing assets, savings in kind, cash receipts, expenditures and balances - were kept and reported at the end of each activity calendar year. Physical stock taking of assets and other kinds of savings, such as crops, were conducted. Guidelines were available for promoting and supporting SFGs^[3], but so far there were no similar handbooks at the SFGA level.

Methodologies for M&E of SFGA development were proposed based on two approaches, both of which were for use by outsiders supporting SFGA establishment.

The extrinsic group-promotion camp had the most to say. Their approach was that while a number of questions and dilemmas remained not covered, it would be extremely useful to get a quantitative overall view on how many (in total and per country) of SFGs and SFGAs could be considered either:

• successful,
• unsuccessful, or
• doubtful.

Such data could give a comparative perspective and could indicate why in some areas or countries projects did better than in others.

Some reflection on the cultural, social, economic and political context of each country or area would be useful to highlight the - often crucial - influence, or not, of "the risk of economic or physical reprisal taken by more powerful forces" [TP-7]; "fear" of "disturbing the existing status quo" [TP-13]; competition, or not, with "more formal associations" [TP-13]; and the effects of recent trends towards "privatization of markets and downsizing of public sector programmes" [TP-3]. Through this, the influence of these factors on national policies and local power-relations, and hence on collective self-help action and, more specifically, on SFGA development, could be better assessed. The land tenure situation in each project area was part of this complex, and so was the influence of merchants and moneylenders.

This led on to a third parameter: "How to assess/quantify/evaluate the process of social capital formation and SFGA development?" The whole area of assessing quasi-intangible outputs was one that presented difficulties all round - the economists tried to ignore it as they could not quantify it easily. How did one place a value on social capital? What were the units with which one could measure social solidarity?

In contrast to the foregoing, proposals for appropriate tools for internal self-assessment of SFGAs were few. There were two relevant FAO publications, but they dated from 1988 and 1989 and represented an initial attempt at the task. There was a clear need for an updated guide to internal M&E for SFGAs that reflected knowledge gained in the subsequent decade.

The key proposals were that the SFGAs use simple bookkeeping systems, and develop a habit of regular meetings where the issues and finances would be discussed, and that they keep some basic account books. This was one area where outsiders could help, and this point should be stressed in the proposed manual on SFGA development