Journal of Social Development in Africa (1999), 14,2,21-38 Informal Security Systems in Southern Africa and Approaches to Strengthen them through Policy Measures* FRANZ VON BENDA-BECKMANN+ AND RENATE KIRSCH* Background In the framework of a GTZ-financed policy measure, a regional seminar was held in Malawi on The Capacity of Social Security Systems in Southern Africa - Conditions, Constellations and Socio-Political Relevance in September 1995. It was attended by participants from five countries: Malawi, Mozambique, Tanzania, Zambia and Zimbabwe.++ The seminar had two major objectives. One was to contribute to a better understanding of the complexity of social security systems. Social security systems had been defined loosely as the totality of organisational forms through which goods and services are provided to people in situations or periods of need and distress, in which they cannot achieve 'security' by their own, individual means. This encompasses a wide variety of services provided by state agencies, churches, NGOs, kinsmen and neighbours, and is subject to different forms of legal regulation, statutory rules, customary and religious rules as well as self-regulation. While these partial subsystems by themselves are already rather complex, the complexity of social security systems is increased by the manifold interrelationships and mutual dependencies between elements of the different partial systems. One aim of the seminar therefore was to enhance the participants' understandings of these complex constellations of social security systems, to provide analytical means for description and analysis. This contribution is a slightly revised version of the paper presented by Renate Kirsch at the Social Policy Advisory Services GTZ conference on Social Policy, held in Oberursel, Germany, from 23 - 27 June 1997. This paper was translated by AnitaDeppe, Alpine Business, Lilongwe, Malawi. The paper was based on F von Benda-Beckmann, R Kirsch and J Frcibcrg-Straub (1997) The capacity of social security systems in Southern Africa: Conditions, constellations and socio-political relevance, Vol 1, Eschborn: GTZ. Franz von Benda-Bcckmann is professor of agrarian law and legal anthropology at the Agricultural University Wageningen, The Netherlands. Advisor to the Poverty and Social Policy Division at the Ministry of Economic Planning and Development, Government of Malawi. ++For a full report on the seminar, see F von Benda-Beckmann, R Kirsch and J Freiberg- Straubl997. 22 Franz von Benda-Beckmann & Renate Kirsch The other major objective of the seminar was to identify policy options that might increase the capacity of social security systems in Southern Africa with special emphasis on the poor. Given the severely limited capacity of social security systems, and their even decreasing capacity under the present socioeconomic changes, the seminar aimed at looking at options to strengthen the capacity of social security systems in their totality and suggest fruitful ways in which the different elements of these complex systems could be interlinked. The seminar sought to achieve this on the basis of a comparative view of the different social security systems existing in Southern Africa. It was intended to reach a more realistic estimation of the coping capacity of these systems by getting to know the constellations and conditions of different social security systems. In this contribution, we shall discuss the major themes discussed and conclusions reached during the seminar. First of all, we present the conclusions drawn with the question of how to conceptualise social security. Thereafter, we shall introduce the major results of the analysis of social security systems in Southern Africa that were reached during the seminar. We shall conclude with the recommendations of the seminar and their meaning for the GTZ. Theoretical discussions on perspectives and definitions When designing the seminar, the coordinators had faced the problem of defining the term "social security." What exactly was to be understood by social security in developing countries, and in which way was it to be demarcated from other areas of social and economic policy? The approach taken for the preparation was to argue for a functional rather than an institutionalist or policy-driven conceptualisation of social security, and to emphasise the multidimensional character of social security categories.1 During the conference, no conclusive attempts were made to come to a coherent set of concepts of social security systems or their component elements. But there was a wide consensus that the ILO (1984) definition of social security was too restrictive, and therefore a wider conception of social security was to be preferred. There was general agreement that social security, for academic and policy purposes, should be seen to include formal, government regulated, public social security as well as informal and/or traditional social security.2 In their papers and statements, most participants used concepts such as 'informal,' 'traditional' or 'indigenous' loosely, often acknowledging the unsatisfactory nature of these 1 Forafurtherelaborationofafunctionalapproachtosocialsecurity.seePartsch, 1983 and F and K von Benda-Beckmann, 1994. For the conference, the approach was introduced in the background paper by Freiberg-Strauss and Kirsch, 1995. This approach has been adopted in earlier regional seminars organised by the GTZ. See Getubig and Schmidt, 1992. The definition of Getubig (1992) was referred to by some participants. Informal Security Systems in Southern Africa 23 concepts, yet having no more adequate conceptual alternatives at hand. Midgley (1995) had pointed out that not too much energy should be spent on clarity of definition. In his view, the current state of discussion in this area is sufficient for the development of concepts. Most politicians and researchers by now have a clear picture of the differences between government programmes in the field of social security and the institutional answers to the conditions of the insecure prevailing in indigenous cultures. However, contrary opinions voiced at the seminar pointed to the frequent cliche descriptions of traditional and informal social security systems on the basis of which state policy measures are initiated. Midgley's position therefore appeared to be too optimistic to many participants, as it could lead to a situation where the performance of conventional policy measures and traditional social security systems would not sufficiently be taken into account. The need for further conceptual clarification, therefore, was repeatedly emphasised. There was a clearly expressed need to overcome the use of dualistic/ dichotomous categories in describing and analysing the complexity of social security systems. Multidimensional categories During the discussions it became quite clear that, if one retains the dualistic Categories, one remains bound by their one-dimensional character. One-dimensional here means, that institutions or even sectors of social security systems are distinguished as formal/informal or modern/traditional by reference to one dominant set of criteria. The problem with this procedure is (1) that more than one dimension of similarity/difference becomes involved in the conceptual discussions, and (2) that similarities and differences in these dimensions usually do not coincide in a way that would allow one to form consistent dualistic categories which encompass all these dimensions. Looking at the papers and discussions at the seminar, four different dimensions were used, often implicitly, but not systematically elaborated. It seems useful to spell these dimensions out here. These are: 1) The type of normative or legal framework that defines rights and obligations for the provision and receiving of (social security) goods and services. 2) The source of the provision, such as kinsmen or neighbours, members of a self-help group, a state agency, a religious organisation, or an NGO. With indicating the source, one simultaneously brings into view the social relationships between providers and recipients of social security goods and services. 3) The historical dimension which allows to distinguish differences in the extent to which social security mechanisms are rooted in (local people's or governments') traditions, and those which have emerged during more recent developments. 24 Franz von Benda-Beckmann & Renate Kirsch 4) The constellation of actual social relationships of persons, the ways in which (different categories of) people are embedded in complex webs of social security relationships with different provisioning agents and subject to legal regulation stemming from different sources. These sets of criteria are useful to characterise, and differentiate between institutions of social security. However, they should be distinguished as separate sets of criteria and should not be merged into one grand dichotomy. Social security regulated by the state is not necessarily provided by the state, nor is it necessarily 'modern' in the historical sense. Distinctions which emphasise only one dimension, like thesource of normative regulation, and generalise from there with respect to the other dimensions, are misleading and lose their value for the purposes of description and analysis. Functional vs institutionalist approaches Conventional approaches view social security from an institutional perspective, ie, social security being provided by public institutions which render certain services. The function of these institutions have been differentiated from the ordinary socioeconomic organisation. While in earlier definitions of social security, such public institutions were more or less identified with the statutory provisions regulated or provided by the state, more recent conceptualisations also include informal and traditional systems. But also where such wide definition include nonstatutory social security, the normative framework and objectives of the social security institution as such remain part of the definition/perspective. The disadvantages of such an approach for a comparative conceptualisation of social security are obvious.3 What social security is, then comes down to what Parrott quotes from an expert: "Social security in any country is what the Department of Social Security in that country happens to be doing at the time" (1992:381). Obviously, such approaches do not provide a sound basis for comparative perspectives on social security systems or policies. What is required is a more abstract circumscription of social security as a field of problems to which different actual systems and policies can be related. A functional approach, in contrast, departs from an abstract definition of social (in)security as a field of social problems, and looks from there at all empirical institutions, social relationships and practices that have a social security function. It thus allows a much better perspective that links the social security conditions of 3 Parrott has succinctly pointed out the problems the now dominant 'institutional' definitions of social security bring with them (1992:381). The ILO uses one concept, the United Nations System of National Accounts uses another one. The European Community which relies on the ILO to carry out all the technical work connected with the harmonisation of social security systems, adopts yet another definition. Informal Security Systems in Southern Africa 25 people to relations and institutions as well as to policy. It sees individuals as being involved in a set of multiple relationships with a variety of (potential) social security providers, and thus having 'social security mixes.' It also includes social relationships which are not normatively defined or institutionally differentiated as social security, that fall outside an institutionalist analysis. Moreover, such an approach also allows a better perspective on the complexity of social security systems and the interdependence of their components. The linkages and interdependence of security systems are often neglected in conventional approaches. Civil servants, for instance, are usually part of both statutory and non-statutory social security systems. If they do not receive their own pension, or if their pension is not sufficient, they can nevertheless indirectly benefit.from social security based on kinship relationships. Civil servants who gain pension rights often try to nurture social relations in their families in rural areas for many years because these networks have a non-monetary security function in old age. On the other hand, village relatives of civil servants in the rural areas are also related to the statutory system in an indirect way through their civil servant relatives. These types of connections show that different forms of security systems are indirectly linked to one another. The policy implications of conceptual discussions Last but not least, it was explicitly stated during the conference that definitions are politically important because they delineate the domain of social practice and policy one is concerned with. Adopting a functional approach to social security, for instance, and conceiving of social security as a field of problems, yields a completely different picture of social security policy than if one departs from an institutional or policy definition of social security. In the latter cases, what social security is, is defined by those laying down the normative framework of institutions or policies. If the functional approach is adopted, on the other hand, any policy of influence for that field is a relevant social security policy, whether policy makers call it social security policy or not. Such normatively or terminologically (policy terms) defined social security will always be only one part of social security policy. For these reasons, further conceptual clarification needs to remain on the agenda for future meetings. Complex Relationships but Weakened Capacities - Changing Social Security Systems The five country papers4 and the analytical contributions provided a wide ranging overview of the different problems to be discussed at the seminar. They highlighted the social, economic and political developments which had led to the present conditions in the field of social security. These conditions are extremely unsatis- < Dava, 1995; Ishengoma, 1995; Kaseke, 1995; Kutengule, 1995; Mukuka, 1995. 26 Franz von Benda-Beckmann & Renate Kirsch factory for major parts of the population in all spheres of social security, whether regulated by the state government or subject to non-state forms of social organisation. The papers also showed that despite the many problems shared by all five countries, there were also a number of country-specific problems such as the great extent of urbanisation in Zambia or the heritage of a recent civil war in Mozambique. The country papers make it quite clear that these conditions present an immense challenge for all those who see it as their responsibility to initiate changes that would improve social security conditions, to state governments, national and foreign NGOs as well as to foreign or international donor agencies working in the field of social policy. In this section we shall present some significant aspects of social security, as they emerged in the papers of and discussions with the African participants. Social security was addressed by the participants mainly in connection with social problems and social change. Security is mainly offered through traditional and informal systems. For the majority of the population in Southern Africa they are the only source of security. State forms of security are perceived as additional measures, but their significance is regarded to be rather low. Insurance systems mainly cover formal employment only, and remain therefore out of reach for the majority of the population. Social assistance programmes, on the other hand, are considered to be insufficient, inaccessible due to administrative hurdles and unreliable in the sense of their legally-binding character. However, as mentioned earlier, it is normal for most people in Africa to be involved in a plurality of social security arrangements. With the exception of a few very remote communities, most people aim at being involved in various types of security arrangements. They create their own individual mix of diverse security services, be it as provider or beneficiary. The interrelatedness of these systems ensures that persons who do not have direct access to state-run security facilities might get it indirectly through their relatives. Effects on one form of security is therefore not limited to that system alone, yet they will involve more and more persons and will, in turn, have an impact on all other systems of social security. Socioeconomic change The continuing importanceof traditional and informal social security systems were emphasised throughout the seminar. However, it also became clear that one should not glorify and romanticise these systems, neither for the past nor for the present. They have their own constraints and limitations; moreover, they have changed and continue to change. The changes are visible in many respects: in the constellation of social relationships on which social security provision is based; in the increasing asymmetries in social security relations of (mutual) help; in the resources available for social security transfers; in the willingness to provide these resources; and in the structure of authority which can monitor and enforce social security rules in traditional law. Informal Security Systems in Southern Africa 27 Migration, urbanisation and monetarisation were frequently mentioned as explanations for changing security systems. Migration, for instance, often means that network structures of family clans cannot be maintained. In urban areas, new forms of social networks emerge. Membership in social security networks does no longer come about in the first place through kinship relations. What counts now is the fulfilment of assistance services, mutually agreed upon on the basis of personal relationships, be it in the workplace, in the neighbourhood or in the church community. These more recent forms of local social security differ from the older traditional ones with respect to the length of time for which assistance is obligatory, and with respect to the underlying rationale for giving assistance. While in kinship relationships there is a life-long obligation to provide solidarity assistance, modern networks have to rely on the assistance being given on a more voluntary reciprocal basis. In newly established self-help groups, the capacity for redistribution is often very limited. These groups will only exist as long as their members fulfil their obligations. These processes of change cannot clearly be judged as being either positive or negative. Through the interdependence between social security mechanisms they acquire a rather contradictory character: the involvement in self-help initiatives means security for those who cannot or do not want to rely exclusively on traditional kinship systems. The resources flowing into these new networks are withdrawn from traditional systems, and can thus weaken the latter even more. Rotating saving and credit associations or income-generating activities, for instance, are promoted as innovative and modern forms of social security and market integration. They have a social security component in terms of social contacts within the group and additional self-help activities (eg, emergency funds or commitment agreements with regard to mutual assistance). However, financial and social obligations within these groups tie up resources. A women investing her money and time in a savings club has less opportunities to redistribute her resources within the original kinship system. Equally contradictory are the mechanisms according to which people gain access to these new institutions, or not. Keebet von Benda-Beckmann points out in her contribution to the seminar (1995) that these newly emerging non-statutory organisations have a different rationality of including or excluding members than traditional security systems would do. There is a risk that people who had been relatively secure, being integrated in the more conventional kinship systems of social security, remain excluded from the modern forms of security. It is to be feared that this affects in particular those persons who are not in a position to render a contribution to the strengthening of the self-help capacity of the group. The introduction of new regulations concerning the access to and distribution of goods and services signifies new chances for one individual, but excludes others. 28 Franz von Benda-Beckmann & Renate Kirsch Each external intervention (however grass-root or self-help oriented it might be) is therefore prone to weaken existing traditional forms. The conclusion to be drawn from this analysis is that policy measures should be as little disruptive of traditional mechanisms as possible. Offers for assistance should be designed in such a manner that they can be used by the affected people as and when needed. Unequal relationships Traditional security systems have the reputation to be rather egalitarian and sustainable. Yet a number of instances were presented that make clear that this is in part glorifying existing cliches. African participants pointed at limits of these systems; some of which are even said to have an exploitative or 'predatory' (Ngwira, 1995) character, such as katapila in Mozambique (a money lending system with arbitrary interest rates) or ganyu in Malawi (a casual labour system which is mainly used by households in times of food shortages). The security aspect of ganyu lies in a certain obligation of wealthy community members to support the needy by employing them on a temporary basis. However, the certainty to actually find work as a casual labourer is in sharp contrast to the extremely low level of remuneration. Job seekers have to work more often and longer periods of time as casual labourers to be able to secure their subsistence. Ganyu is mostly seasonal and peaks at times of high demand for agricultural labour. The labourer neglects his/her own fields, whereas the hirer benefits since he/she uses the labour input productively. According to empirical surveys, ganyu has become the most common survival strategy for food insecure households in Malawi, since traditional transfers in the form of gifts and remittances no longer happen (Ngwira 1995). The redistributive capacity of the traditional system has been weakened in such a way that security can no longer be guaranteed. The loss of this security mechanism is compensated by selling of manpower at exploitative market conditions. Similar problems of increasing social and economic asymmetries within social security arrangements came up when issues of gender relations were discussed. Participants described the heavy burden which women have to carry in the productive and reproductive sphere. There usually is an unequal division of labour which forces women to shoulder the burdens of subsistence agriculture, child care and domestic labour. Through new social and economic developments, women become even more vulnerable. Also the negative impact of inheritance law in patrilineal societies were mentioned. The husband's family or lineage claims all goods which the couple had acquired together during their marriage without assuming the responsibilities for the widow. Women's organisations attempt to Informal Security Systems in Southern Africa 29 increase the awareness of these problems and propagate that husbands should make testaments making their wife the heir to the property. The discussions also showed some ambivalence in the participants' appreciation of some gender-related aspects of social security systems. The problem of polygynous marriages is such an instance. Polygynous marriages (and or widow remarriage or widow inheritance) can be seen as an element of traditional social security systems. For it provides close social relationships with social security rights and obligations to persons who otherwise may be, or remain excluded from such relations. On the other hand, it may force women into relationships of dependence and subservience which they no longer find acceptable. It was also noted during the discussions that especially in patrilineal societies, traditional legal rights of women are lesser and weaker than those of men, giving them less access to resources - material or immaterial - with which they could support themselves independently. Also women's position after divorce and in widowhood were discussed as particularly difficult situations where the traditional mechanisms that used to support women have weakened or are regarded as no longer legitimate. Widow inheritance, for instance, or re-marriage in general was seen as a sometimes desperate way to continue to receive male support both morally and materially. Problems associated with age differences were also discussed. It was mentioned that Tanzanian youth in rural areas claim that they perceive the traditional social labour exchange system rather as a burden than a benefit. It is dubious whether the traditional security system would also function if they would require assistance from other community members. They regard themselves as service providers maintaining the system and not as beneficiaries of it. They have the feeling that their manpower could be used more productively instead of always giving it for community activities. Doubts about the capacity of existing systems, the subjective feeling of being exploited and a stronger tendency towards individualisation mark the cornerstones of this socio-psychological area of tension. For how long these security forms, where advantages are selectively being seized, but the provision of services is questionable (as shown in different constellations by the two examples), can still be maintained, mainly depends on the economic, political and social power of the partners in these relationships. For purposes of political intervention level, it can be concluded that further investigation, of the social powers is required which mainly contribute to maintaining traditional security arrangements. Is there a balanced or unequal relationship between service provision and benefit? How do the relationships of dependency look like, and how strong is society's acceptance of existing and possibly unequal systems? 30 Franz von Benda-Beckmann & Renate Kirsch Conditions for social security So far, factors have been mentioned that mainly relate to relational and institutional mechanisms of social security services. Apart from that, there are other circumstances which weaken the conditions for social security arrangements. This concerns mainly the availability of resources for social security redistribution which are being eroded in most Southern African states. Access to land and land rights are a good example for this. Land being the basis of subsistence and agricultural production in general, is one of the most important sources of social security. As per Mozambican law, for instance, land is communal property. Access to land is regulated through traditional local authorities. The peace process and market liberalisation have led to significant changes in the public sphere, and, for a short period, left behind some kind of legal vacuum with regard to land. Land concessions had been given for 99 years to international consortia. Male and female farmers have lost their land almost without resistance. After the civil war there were no longer functioning community institutions or authorities capable of representing the interests of the affected people within the legal system. One of the most important resources for securing a livelihood and for the continuation of security provision had collapsed. The significance of land and land rights as components of social security is largely ignored by most governments in Southern Africa. Governments intending seriously to promote informal social security systems, will sooner or later face issues of redistribution of resources, such as land utilisation. Strengthening the resource basis of the rural population will probably lead to broader and more sustainable improvements than any expansion of public social security services might be able to achieve. Such examples furthermore show that government policy decisions and legislation can have serious effects on the capacity of security networks. In many cases, unintentional effects hamper the service capacity, since social security is often not perceived as an issue by the relevant sector policies. Further examples in this context are the Special Crop Act in Malawi which forced farmers to sell their produce at fixed prices to a state-run marketing organisation. The policy of low prices forced farmers to alienate the basis of their subsistence production instead of accumulating resources. Laws which prohibit street vending had been interpreted as restricting individual survival strategies, which particularly for women are a source for generating income. Structural adjustment policies The effects of structural adjustment measures in Africa were also the subject of heated discussions. Structural adjustment often leads to retrenchment of public Informal Security Systems in Southern Africa 31 servants, as well as to the introduction of cost-sharing schemes in the health and education sector. These measures often have effects on most social groups and their security arrangements. Restricted access to public services with a security component (such as social welfare programmes, health or (indirectly) education facilities) is a reality, particularly for the poor. This results in increasing dependence on informal and traditional security systems. Retrenched public servants, for instance, also lose their access to state-run security services. These persons often belong to those who are in a position to channel resources from the state-run system to traditional security arrangements. Instead of providing resources, they now have to be called on by themselves, thus doubling the burden on informal systems. These examples show how the resource base for social security is reduced through state policies. They also show that state-run institutions do not perceive the existing linkages and interdependencies between the various forms of social security. The policy measures are directed at making changes in the state regulated public domain; the impact on the complex social security system as a whole remains disregarded and is not scrutinised. Aspects and Consequences for the Promotion of Informal Social Security in the Framework of Policy Measures The seminar aimed at analysing the present situation and identifying starting points for policy measures that could contribute to the strengthening of social security systems. The major topics discussed were: What are the prerequisites and framework conditions of the promotion of social security? Who is responsible for the provision of social security services? What is the duty of the state and its administration? In what way can improved targeting of the target group be guaranteed? What needs to be taken into account when integrating and linking formal and informal security systems? The results of the discussion will be presented in the following. Social security: fragmented policy measures and integral experiences It follows from the functional approach to social security that social security policy has to be seen in a much more comprehensive way than would appear from the conventional normative frameworks for policy interventions in this area. Social security has rather to be understood as a cross-sectoral task affecting almost all spheres of political life. This statement has two consequences: (a) Analysis and policy measures cannot be restricted to the conventional areas of social security policy. To a large extent, they have to include also 32 Franz von Benda-Beckmann & Renate Kirsch social, economic and political prevailing conditions in general. Social security should be regarded as one component of a comprehensive poverty reduction strategy, instead of being limited to social policy in a narrower sense. As mentioned above, access to productive resources and employment is an equally important prerequisite for strengthening social security systems for the poor. (b) As a consequence, line ministries will have to assess in which way their policy measures might have an impact on, and possibly negative consequences for, the social security of the people. This will require considerable reorientation of policy makers. For in any public administration, policy measures intended to have an impact on social security are fragmented according to administrative and ministerial divisions of labour. Problems of employment, housing, education and health policy are being dealt with differently, by different state institutions, that often pursue contradictory policy objectives. The problems that are being addressed in such a fragmented way, however, come together in the social relationships of affected individuals, families and communities. Being unemployed and homeless, not being able to pay school fees or medical bills, are integral aspects of social insecurity of any one person (or group, for that matter), although policies may differentiate and construct them as diffcrentcategories. The conclusion isobvious: The affected people's understandings of their own security and insecurity have to be grasped by policy makers, and subsequently have to be translated into concrete measures (key words: processoriented and integrated programme design). Such a demand seems to be difficult to implement. The departmental fragmentation of the government apparatus appears to be insurmountable, and tensions amongst various sector policies are well-known. Measures aiming at increasing production for market exchange or to stimulate the marketability of resources may easily conflict with social security objectives. From a social security perspective, one can therefore conclude that it is safer for poorer households to remain engaged in subsistence production.because this sphere can guarantee them a minimum of food security in the medium term, whereas market production which depends on prices and monetary incomes would be the greater risk for these households. What on the one hand looks like an optimal allocation of resources, may weaken the resource base of the poorer segments of the population. These tensions are inevitable and cannot be eliminated through the rhetoric of integrated development measures and policies alone. But if policy makers are aware that they are engaged in policies relevant to social security, even if social security is not their policy domain, then there is hope that they become more Informal Security Systems in Southern Africa 33 sensitive to integrate social security aspects with other objectives of their policies. This also needs to lead to a reconsideration of the role of state and customary laws for social security. The earlier discussion has shown that important policy domains had been disregarded during the current debate, and that they go beyond the boundaries of the framework of social policy in a narrower sense. During the discussions at the seminar, family, marriage, divorce and inheritance law as well as land law and property law were identified in connection with social security. A number of different fields of law are of significance to social security even if they are not, at first glance, associated with social security or social security law. It was also pointed out that while lawyers distinguish between state and customary law as different, and often unrelated legal systems, it is the totality of existing legal systems which contribute to the strengthening, or weakening, of social security. Elements of different legal systems often become combined in people's reality, or played off against one another. The significance of legal pluralism in terms of social security still largely is a blind spot in the analysis. The role played by legislation with respect to social security should, therefore, be given more attention. Strengthening non-state social security While much attention was given to state regulated and provided social security, the importance of informal and traditional social security mechanisms was also strongly emphasised in the policy discussions. Among the participants there was general agreement that traditional social security systems, however weak they might have become, will still be indispensable carriers of social security for most people in the future. In view of the present inability of states to provide sufficient means to take care of the social security problems of the poor and needy, the strengthening of these systems should be an important objective of social security policies, especially in attempts to improve the social security conditions of the poor. How non-state social security should be strengthened or linked to state regulated systems was not elaborated in detail; too much depends on the specific context of the problems to be addressed. But generally it can be said that such strengthening demands a widening of the room for manoeuvre of people, networks and groups. People must have the social, political and economic room to establish self-help activities in the field of social security. However, an important and related conclusion was also that the stimulation and development of new forms and organisations of social security always bring with them the danger of newly emerging patterns of in- and ex-clusion. Keeping in mind that most social relationships upon which non-state regulated forms of social security rest are multifunctional, new forms of pooling and redistributing resources for dealing 34 Franz von Benda-Beckmann & Renate Kirsch with risk and contingencies will generally be more exclusive and therefore weaken the social security function of those relationships which are not included in the new self-help organisation. Moreover, whether one thinks of people's self-help organisation or state programmes aiming at increasing the self-help capacity of local people, for instance through improved opportunities for income generation, there will always be many people such as the old, infirm, handicapped or destitute who simply do not have the mental and/or physical abilities to participate in such activities. These are the people most likely to be excluded from new forms of social security. They are also the people from whom it cannot be expected that they can contribute, in kind or in money, to the financing of their social security. One important conclusion therefore is that social security policies and projects must always be aware of the likelihood that the most vulnerable and needy persons become the major victims of new developments in social security, and that steps must be taken to prevent this. This means that any intervention, whether by government agencies or NGOs, must be extremely careful not to disrupt existing social security relationships by new boundaries of inclusion/exclusion. The other important conclusion was that policy measures aiming at strengthening non-state forms of social security should not be solely preoccupied with organisational forms. For a real strengthening will usually only be possible with the help of additional resources. It was repeatedly stated that social insecurity cannot be overcome by improving organisational structures and institutions. More important is to improve the social, economic and political conditions for the successful functioning of the mechanisms of social security. The division of responsibilities and tasks for improving social security Social security involves a number of social actors (families, kinsmen, self-help groups, NGOs, churches, state institutions). Who should be responsible for which type of service provision, and how could a good co-operation be guaranteed? It was a unanimous conclusion that active participation in structured attempts to improve the conditions of social security was not the responsibility of only the state. In the fields of policy making as well as implementation, various types of non-governmental organisations have an important role to play, and an important role needs also be played by local communities, community leaders and common people. The state's role was seen as particularly important in the policy field. For only states have, or at least can have, the nationwide organisation which makes it possible to generate resources and to redistribute them inter-regionally and interlocally. Moreover, given the fact that the field of social security is affected by a number of different state policies, only the state itself has the possibility to change and integrate such cross-sectoral policies in a structured and coherent manner. Forsaking state responsibility to market forces, through the dominance of an Informal Security Systems in Southern Africa 35 incrementalist policy style of African governments combined with an anti-welfare mentality, was regarded to be one of the main reasons for preventing innovative problem-solving approaches. Moreover, given the unequal distribution of resources, it has to be expected that the market and commercially provided social security services will not be available to the poor. While the role of the state in providing resources was strongly emphasised, the state administration as providing agent on the other hand received very critical attention. For most participants, state agencies were regarded as being incapable of reaching the poor and reacting adequately to the difficulties involved with dealing with the poor. It was stated that the state's perspective is often limited to macro-level interventions. Most participants expected that NGOs and churches, supported by the state, could play an important role in channelling resources to appropriate persons without a gender bias. Persons more familiar with and sensitive to local people's lifeworlds and understandings of insecurity such as well-trained members of NGOs or social workers can have an important role as intermediaries between the state institutions and resources on the one side and the people on the other. The flow of communication of ideas that inform and guide policy and implementation should not be a one-way, "top-down" traffic. A participatory and "bottom-up" approach to social security was seen to be an extremely important precondition for a better social security policy. However, the potential of NGOs should also not be overestimated. Their production of collective goods and services is systematically selective (see Glagow, 1995). Inclusion of one category of beneficiaries always means exclusion of others. External intervention and attempts to involve NGOs may further weaken traditional forms of social security systems (cf Keebet von Benda-Beckmann, 1995). NGOs also tend to target "their own" people, which can result in the duplication of services and also in progressing a distance between communities. It is very difficult to co-ordinate the work of NGOs, as each wants to protect its own agenda and justify its existence, operations and provision of employment. NGOs thus are not all altruistic organisations, as is often claimed, but very much follow the interest of others, of donors, their own members or their clients. Moreover, the decision-making structure of NGOs often is in the hands of only a small group of active members. Many newly-founded NGOs do not even have links yet to grassrootorganisationsbut are initiatives of urban-biased individuals who are still trying root their NGO in some segment of the society. NGOs thus were by no means romanticised during the seminar. Yet it was generally thought that they could provide the urgently required link with the people. They have the ability to provide social protection at micro-level. Policies therefore should encourage and support the role of NGOs. 36 Franz von Benda-Beckmann & Renate Kirsch Conclusions: Demands on Policy Finally, some general implications for state policies towards the non-state sector of social security were spelt out in the discussions: I. A formalisation of traditional or informal support systems by the state was strongly disapproved as an option by most participants during discussion. II. State policies as well as policies of other social actors, like NGOs should therefore do their best to uphold or improve the capacity of these mechanisms. III. Whenever introducing new policies, governments, the state, NGOs and other social actors should take apreventiveperspective, considering the probable influences of their policies on social security systems. Policies that further erode social security systems should be avoided. The individualisation of land rights was given as an example of how traditional systems of check and balances that organise and ensure access to land to almost all community members - one of the most important resources of social security - can be harmed. IV. State decentralisation policies would also have positive effects on traditional support arrangements. Policies encouraging the development of markets, infrastructure and employment opportunities in the rural areas would make migration to urban areas, in search for wage employment opportunities, less urgent. V. A political and legal framework should be given to those forces which render a contribution towards strengthening the solidarity capacity of the society (eg through self-help, co-operative or trade union movements to enable them to flourish. Here again, decentralisation of government mechanisms, including transfer of financial and political authorities to regional and local administrative authorities, would undoubtedly have a positive effect on this process. Recommendations The main recommendations made at the seminar pertained to the following aspects of the social security problematic. * Social security policy should be reconsidered in the light of the discussions and conclusions of the seminar. More in particular, policies aiming at strengthening non-state forms of social security should be rethought and promoted. Informal Security Systems in Southern Africa 37 * On the basis of the general conclusions of the seminar, country-specific workshops should be held to work out the implications for social security policies at national and regional levels. * New approaches to social security require much more detailed insight into thecomplex interrelations between the differentelements of the totality of social security regulations, institutions and practices. This needs much more systematic empirical research as well as further theoretical and methodological refinement. * To test, and work with the insights derived from new research and theoretical reflection, action research projects should be set up to accompany and monitor new policy measures and projects in the field of social security. Here, particular attention should be given to projects that aim at integrating state and non-state forms of social security. * Committees should be set up for the exchange of experiences between implementing NGOs and the state administration in the area of poverty alleviation and social policy. * More efforts must be made to improve the training of African researchers, government and NGO staff involved in policy formation and implementation. The training should encompass both the strengthening of analytical capacities as well as the translation of analytical insights into pragmatic projects. The issue of targeting (whom? why? and how?) should have an important place in these efforts. * Foreign and international donor agencies, such as the GTZ, can and should play, an important role in facilitating these processes. The financial, scientific and organisational resources upon which they can draw can be of great assistance. * The GTZ could and should play an important role in supporting a range of activities in the five countries. In particular, the GTZ, in cooperation with governmental and non-governmental organisations as well as academic institutions, should: - help to provide materials for the training of researchers, social workers, NGO and government officials working in the field of social security; - help to organise national workshops on social security policy; and - research into crucial issues of social security and set up carefully selected action research projects as a basis for social policy advice to governments. 38 Franz von Benda-Beckmann & Renate Kirsch References Benda-Beckmann, F & K von (1994) "Coping with insecurity," pp 7-31 in F von Benda-Beckmann; K von Benda-Beckmann; H Marks (eds) Coping with insecurity: An 'underalP perspective on social security in the Third World, Special issue, Focaal 22/23. Benda-Beckmann, K von (1995) "Legal pluralism, social security and gender," Contribution to the seminar. Benda-Beckmann, F von; R Kirsch & J Freiberg-Strauss (1997) The capacity of social security systems in Southern Africa: Conditions, constellations and socio-political relevance, Vol 1, GTZ, Eschborn, Section 401. Dava, G K (1995) The importance of social security systems in a context of social and economic hardship. The case of Mozambique. Country study, Internal GTZ paper. Freiberg-Strauss, J & Kirsch, R (1995) A review on the issue of social security in developing countries: Approaches and their relevance for social policies, GTZ, Internal paper. Getubig, IP (1992) "Social security and the poor," in IP Getubig & S Schmidt (eds) (1992) Rethinking social security: Reaching out to the poor, APDC & GTZ, Getubig, IP & Schmidt, S (eds) (1992) Rethinking social security: Reaching out Kuala Lumpur/Eschborn: pp 1-17. to the poor, APDC and GTZ, Kuala Lumpur/Eschborn. Glagow, M (1995) "Local NGO's in the process of urban poverty alleviation," Contribution to the seminar. International Labour Organisation (ILO) (1984) Introduction to social security, International Labour Office, Geneva. Ishengoma, A K (1995) "The capacity of social security systems in Southern Africa: Tanzania experience," Internal GTZ paper. Kaseke, E (1995) "Social security in Zimbabwe: A country study," Internal GTZ paper. Kutungule, M (1995) "Social policy and poverty in Malawi: A review of its capacity to cater for the poor," Internal GTZ paper. Midgley, J (1984) Social security, inequality and the Third World, John Wiley & Sons, Chichester. Midgley, J (1995) "Options for interlinking statutory and traditional social security systems," Contribution to the seminar. Mukuka, L (1995) "Zambia country study paper: The capacity of social security systems in Southern Africa: Conditions, constellations and social political relevance," Internal GTZ paper. Ngwira, A (1995) "An appraisal of social security in rural food security: A case of central Malawi," Contribution to the seminar. Parrott, A (1992) "Social security: Does the wartime dream have to become a peacetime nightmare?" in International Labour Review, 131: 367-386. Partsch, M (1983) Prinzipien und Formen sozialer Sicherung in nichtindustriellen Gesellschaften, Duncker und Humblot, Berlin. Schubert, B (1995) "Targeting approaches and their relevance to social security," Contribution to the seminar.