Decentralization and rural food security: some theoretical and empirical relationships

The subsidiarity principle implies devolving power and resources to the lowest possible (adequate, efficient, sustainable) organizational level, and may thus be utilized within the framework of any of the above forms of decentralization. In this spirit, the World Food Summit Plan of Action concludes by stating that the Plan is to be implemented "at the lowest possible level at which its purpose could be best achieved" (FAO, 1996b, p. 43).

This article focuses on devolved decentralization. It is, however, beyond its scope to assess in detail which level of decentralization may indeed be the most appropriate for all the different areas of government intervention on food security grounds.

A PROBLEM STATEMENT

The planning processes that drive efforts at local area development are influenced directly or indirectly by socio-economic linkages between all kinds and sizes of urban agglomerations and the countryside that surrounds them. The quest for food security has not yet been examined in the light of rural-urban linkages issues and the recent trends of increased political, institutional and fiscal decentralization, and thus been supplemented with qualitative and quantitative data resulting from field work. That is to say, although the relationship between the level of decentralization and extent of food security in rural areas will ultimately remain an empirical question, the academic literature has not yet provided a theoretical backdrop against which to verify hypotheses (this point is also made by Parker, 1995). An analytical framework that may be applied to the study of rural-urban linkages would allow policy-makers to address some questions from the point of view of the present discussion, such as how can decentralization contribute to the promotion of self-sustaining local development and food security, for example through strengthening rural-urban socio-economic linkages?

The standard, project-oriented approach taken by many development agencies is stranded in this theoretical void; much of the experience derived and accumulated from development projects thus remains untapped in that it is of not much use in informing empirically the relationship between decentralization, poverty and food security. A more comprehensive analysis would draw on some practical applications, including FAO's experience with initiatives such as the People's Participation Programme and the Special Programme for Food Security (which is still in its early stages).

As is related in World Bank (1995, p. 14):

"(...) The degree and different types of patterns of decentralization have not been described and measured in a consistent way across experiences or over time, so that at best only an anecdotal characterization of the decentralization of rural development and rural service delivery programs can be made. Without consistent description and measurement of the patterns of decentralization, it is not even possible to assess the issue of whether greater decentralization in some form is associated with greater success in rural development and rural service delivery, or whether it results in better targeting of the poor and reduced poverty levels."

PERSPECTIVES ON RURAL-URBAN LINKAGES AND THE NEED FOR A "SPATIAL" APPROACH

The perspective taken by analyses of rural-urban linkages is one way of approaching the many-sided relationship between rural food security and decentralization. Rural-urban exchanges have been classified by Rondinelli (1981) in seven broad categories: physical, economic, population, technological, social, service and political linkages. More recently, Gould (1988) discusses such linkages in terms of resource flows and interactions; the vast body of work produced by Hazell and Röell (1983) and Hazell and Haggblade (1991) has focused on economic farm-non-farm linkages. Reardon and Stamoulis (in press) provide an innovative conceptual framework of spatial analysis for rural-urban linkages research by examining the intersectoral effects of rural income diversification on decentralized economic development.

How can food security be sustained financially in the long term? Food security is a function of the general income of households. If the income accruing (in cash or kind) from any one component of the activity mix of a rural household can be increased, this may add to improved national food security. Self-sustaining agricultural development may be difficult to achieve locally without rural income diversification. There is evidence that a "virtuous circle" of reinvestment of non-farm income in agriculture, leading in turn to more non-farm income (through spinoff effects) and further reinvestment, can be set in motion (see Evans and Ngau, 1991). Specific agricultural production decisions can often be traced back to measures of (non-farm) income diversification, which may be the safest and most sustainable means to ensure a sufficient level of agricultural productivity and hence medium-term food security.

A most important aspect relating to the potential of decentralized development is an improved income distribution. To the extent that decentralization measures lead to decentralized development by facilitating the diversification of rural households' income portfolios, they may contribute to poverty alleviation, rural development and food security. Without interrelated policies, there is no reason to assume that decentralization will by itself lead to lower income groups' obtaining more work and/or higher incomes, or to being better able to cope with risk. If carried out in a participatory fashion, the restructuring of rural institutions and local governments under decentralization will create more and/or wider political spaces, and will help win legitimacy, for some of the marginalized stakeholders in rural development issues.

Smallholder farmers may be unable to afford production inputs or access to marketing structures as subsidies are severely cut. In a number of countries the domestic private sector has not had the time or incentive to provide these services on a commercial basis, making imported inputs very costly, and affordable only to large-scale farmers and small farmers with access to non-farm sources of income.

Where small farmers can dispose of incomes that are not covariant with weather (and other) risks associated with agricultural production, food shortages resulting from cropping shortfalls may be more easily overcome. The provision of information on cropping forecasts and income-generating opportunities (on possibilities for agricultural and non-agricultural diversification, niche markets, etc.) on the part of local governments is one policy component that could be tied to programmes which facilitate local institutions' access to credit and decentralized financing. In this regard, in the agricultural sector rural local governments should collaborate with farmers' organizations as well as informal structures, particularly in areas that are more difficult and costly to access, and with the private sector in areas of less limited commercial potential and with a higher degree of organization. Such collaboration should also allow for longer planning horizons and make easier the coordination of rural development - the Achilles heel of the integrated rural development projects of the 1970s.

To a certain degree, the nature of economic resource flows - for example, between the farm and the non-farm sectors - gives an indication of the extent of market efficiency. Efficient markets are necessary for the food security of the increasing number of (mostly urban) households that are not producers of food for their own consumption. Within the framework of initiatives to promote greater decentralization, the lowering of the transaction costs of net buyers and sellers of food in the market (purchasing and marketing difficulties) are typically addressed by infrastructure construction programmes commissioned by national governments. However, what is more interesting for the relation between decentralization and food security in the broader sense is that such issues are also addressed indirectly by initiatives such as policy support to intermediate cities.

Central place theory (Christaller, Weber) and economies of scale are important elements to consider in rural-urban linkages analysis. The principle of subsidiarity should thus be applied with a view to promoting rural-urban linkages to central services and functions that are better carried out at least at the lower level of intermediate cities. These would be more suitable locations for agricultural research centres and extension services. Economies of scale suggest that the level at which such services can be run efficiently and remain sustainable will much depend on the size of the country, population density and distribution, etc., taking into account regional disparities. The vicinity of intermediate cities to both agricultural and non-agricultural enterprises could help foster more concerted action with respect to more integration of rural development, thereby taking into account the comparative advantages of the public and the private sectors.

To model development outcomes, a "space-based" approach to rural-urban linkages would integrate the effects of "exogenous" off-farm income earnings (e.g. from businesses in local rural towns or migration remittances) that are used to buy farm sector products and local, non-tradable off-farm sector products. This type of income is further used in capital investment in all sectors of the rural economy for capital formation. Thus, exogenous changes in agricultural productivity and/or increases in output volume owing to the creation of new local markets (and hence a shift to higher-valued products that increase farm income) may generate sustained rural and national economic growth through input and output linkages and multiplier effects. Past analytical efforts have mostly ignored the full extent to which these effects have an impact on food security and rural welfare, beyond the immediate area under investigation.

Such neglect has arisen out of the interaction of various factors, such as the practical implications of the lack of consultation and cross-fertilization among the disciplines that form part of the social sciences, as well as the "urban bias" of education and policy-making. For administrative structures, this has meant the involvement in the rural space of a vast number of disarticulated ministries and line agencies that are only restrained stakeholders in rural development. Agro-industries, as in the case of Zamora (see Box), are typically victims of such grey areas of institutional amenability. In many countries the ministry of finance holds disproportionate political and decision-making power, sometimes overriding the development priorities of other actors. Very few countries have established a ministry (or a government department or agency) of rural development, that tries to decentralize decision-making and coordinate develop-ment initiatives and cooperation, thereby actively encouraging rural-urban linkages.

In line with the approach proposed here, a Working Group on Rural Development in South Africa (Working Group on Rural Development, 1993, p. 30) postulated that: "In order to pursue the best balance between people issues (mobilization and institution building), consumption services and production activities, and to stimulate linkages between these, a thorough understanding is needed of the socio-economic forces that operate in the area or region under consideration. In South Africa comparatively little is known about intersectoral linkages and the type of positive interaction that could be generated between developed and developing areas in the interests of rural development. Proven technology that could be used to increase production particularly of non-agricultural activities in rural areas is not readily available. There is a great need not only to interpret and collate existing information in terms of integrated rural development objectives, but also to promote institutions that contribute towards a better understanding of integrated rural development."

DECENTRALIZATION AND RURAL INSTITUTIONS HISTORICALLY: THE URBAN BIAS LEGACY

By institutions, this article refers to: a significant and persistent element (such as a practice, a relationship, an organization) in the life of a culture that centres on a fundamental human need, activity or value, occupies an enduring and cardinal position within a society and is usually maintained and stabilized through social regulatory agencies (Webster's 3rd New International Dictionary).

In many developing countries a strong urban bias prevails in policy interventions. This bias refers to primary cities, the loci of central governments, and not usually to other conurbations such as small or intermediate cities. Urban bias is, moreover, often reinforced by a sectoral (industrial) bias that works much to the disadvantage of the agriculture sector, which continues to be taxed heavily and be regarded as lagging behind other sectors in socio-economic development. Respectively, such a policy bias was first discussed by wider audiences following the work of Lipton (1977) and Bates (1981), at least in the sense that these authors paved the way for later assessments of (development) outcomes of policy bias from the point of view of the political economy - beyond what has typically been the field of enquiry of sociology and (to a lesser extent) economic geography.

This is in contrast with the policies of the rich countries, which (can afford to and are under pressure to) engage in a "rural bias" of policy interventions. Likewise, since the 1970s, Taiwan Province of China and the Republic of Korea, and also mainland China and India as well as, more recently, South Africa, have shifted their policies from a pro-urban to a more pro-rural disposition. Some others have also done so, not least to try to capture rural votes as democratization advances worldwide.

The low-income/urban policy bias versus high-income/rural bias pattern is explored by Hayami and Anderson (1986), who trace it back to a theory of supply and demand for agricultural protection as a country develops and food becomes relatively less important in household expenditures; and by Moore (1993), whose analysis focuses more on the political relationships between peasant farmers and the state and the economic role of political institutions. The approach of these authors ensures that, although the distinction between what is considered "urban" or "rural" is increasingly blurred by rural-urban linkages, migration, local income diversification, etc., some basic elements of the dichotomy still hold true for the present discussion.

Let us place the concept of food security in the following socio-political context provided by Moore: the notion of "development" has as such increasingly come to be central to popular understandings of the role of the state. Thus, the legitimacy of state governments is becoming very dependent on their production of development and assurance of food security. At the same time, in poor rural areas governments generally dispose of only few means beyond the state institutions formally dedicated to agricultural development to uphold political cadres outside the state apparatus, and so precedence is thus given to political considerations. Moore (1993, p. 120) concludes that:

" The consequences for economic performance do not lie in the realm of intersectoral transfers of resources, but in the effect on the economic performance of agriculture of its being subject to the attentions, nominal or real, of so many agencies and officials dedicated in principle to agricultural development."

This phenomenon seriously undermines the food security functions that local governments can and should engage in, and agricultural and rural development at large. The supply-driven nature of most development assistance, the "oversupply" of line and agricultural development (governmental) agencies, together with their political agendas and fragmented and indeterminate institutional mandates, and the disarticulation between these many actors in the rural space all contribute to their slow responsiveness in times of rural food insecurity, and to the remarkably limited success of their attempts at increasing local resource mobilization and organization in rural areas in general.

SOME SELECTED ISSUES IN THE SUSTAINABILITY OF SUPPLY AND ACCESS TO FOOD

Let us now discuss some issues in the sustainability of supply and access to food that relate to the domain of rural institutions and organizations, and view them from the policy bias perspective laid out above. The three issues (technological change, food prices and information) will all point out the importance of context-specific evaluations of decentralization measures through a food security lens.

It is indispensable that governments in regions that are poor or marginalized from international markets, or with insufficient comparative advantage in food production, do not act in short-sighted autarkic ways. They should not place too much emphasis on their region's complete self-sufficiency in food by preventing (potentially beneficial) food trade with other regions. The concept of food "sovereignty" appears to be a more logical and realistic discourse at the national level, unless it involves a nation's claim to total self-sufficiency at all times.

The conflict of interests between small farmers and the state has typically resulted from the farmers' need for profits and the state's desire to keep food prices low, to fuel industrialization through the opportunity of employing relatively cheap labour. Although there have been substantial fluctuations, the prices of farm sector products relative to other goods have, since 1900, declined by half, and this decline has most recently been accelerating. It has occurred over a period of time during which population has increased threefold, and demand for agricultural produce fourfold (Binswanger et al., 1985, quoted in FAO, 1996c). Few would question the crucial role of research in respect to more appropriate technologies in agriculture if we are to advance towards FAO's objective of achieving and sustaining "food for all".

Technological change

The importance of enhancing farmers' participation in production issues has increasingly been recognized. Extension services may provide ample opportunities for cost-efficient participation. Agricultural research has often been neglected in favour of extension, which is mostly cheaper (also in terms of staff costs) and provides more opportunities for political patronage (because agricultural extension creates more government jobs and is inherently mobile and flexible). If extension is at the expense of research, it may respond neither to the farmers' needs nor to the future needs of the increasing number of net buyers of food.

FAO's Extension, Education and Communication Service finds that: "Given the importance of linkages for successful extension, an important issue is the effect decentralized extension has on linkages to research - does it result in a decoupling of research-extension linkages, weakening local access to knowledge that only exists in central locations, or does it actually improve linkages, not only at the local level but also at regional and national levels? In the absence of political commitment for central government to support `devolved' extension workers and researchers, and unless traditional extension-research linkage problems are addressed, it seems unlikely that decentralization by itself will result in improved integration of extension and research functions. On the other hand, if decentralization improves farmers' participation and feedback by giving farmers, and especially small-scale farmers, more of a voice in setting agendas, then the outcomes will be positive, including the development and dissemination of technologies that more closely fit farmers' needs."

Although they may be treated as a variable that is largely endogenous to the processes of rural development, technological changes and innovations should not be left to market forces alone. "The capacity to advance knowledge in science and technology is itself a product of institutional innovation" (Ruttan, 1989, p. 1375). Public institutions and organizations should be strengthened that foster and structure people's participation (including the rural population's negotiating capacities in the political arena) and political competition. The conflict between smallholder agriculturists and their respective national governments has in certain instances been ameliorated by technological change, since technological innovation can lead to rising farm profits without necessarily lowering food prices. The lessening of some of the friction among the interest groups usually creates more political space, which can constitute an important point of entry for decentralization. In the food security equation, technological change refers typically to the physical and economic access to food and the sustainability of its supplies locally.

FOOD PRICES

Safety nets ensure that households that are excluded from market participation do not become food-insecure. The safety net concept is urban-biased in itself, mostly a result of urban poor households typically being more dependent on food purchases than rural poor households. Rural safety nets could be designed as an integral part of the national decentralization process to incorporate private and non-economic elements such as networks of cereal (or rice) banks and grain storage facilities. Through decentralized cereal banks, farmers and their rural communities do not have to buy grain in the market at the end of the dry season when prices are highest. Such banks can operate at the interregional level, purchasing grain cheaply at harvest time, then storing it in a system of local facilities and selling it at cost when needed. They serve to re-establish food security in the immediate term following emergency cropping shortfalls. Whether or not private traders and entrepreneurs are more effective at interregional arbitrage, they may play an important complementary role with respect to cereal banking and storage, thus contributing to sustainable overall food security.

In the least-developed countries, urban food consumers are not only more influential politically, but they are few in number relative to rural food producers. If obliged to subsidize one group, governments may conveniently choose to subsidize the smaller group. The debate about "getting the prices right" has for the most part neglected the importance of informal food markets to respond to biased policies. Adjustment of distorted food prices through the promotion of free market orientation and private food enterprise is necessary but not sufficient for achieving short-term food security in a number of low-income food-deficit countries (LIFDCs). Problems of (food) market failure revolve around, and/or are reflected in, imperfect and asymmetric information, risk and uncertainty, incompletely specified property rights, externalities, collective goods, economies of scale and natural monopolies (FAO/SDA, in press). To reduce market imperfections and demand constraints, intermediary institutions could provide linkages between local governments and the private sector, and absorb some of the risks and difficulties that neither of the two can manage by themselves.

Non-governmental organizations (NGOs), formal public institutions and some of the informal institutions that have come to exist in rural areas should be strengthened, not weakened, to allow for greater and quicker responsiveness to local needs and a greater differentiation of locality-specific policy interventions. In the food security equation, food prices typically refer to the economic access to food and the sustainability of its supplies locally.

INFORMATION

For a better targeting of food security programmes, improved quantitative and qualitative information on beneficiaries is needed, especially longer time series of consistently measured data. Although in many countries a general feeling of disillusionment with and wariness of government prevails, and may worsen the well-known shortcomings of survey questionnaires designed to assess levels of food security at the household level (e.g. recall errors, underreporting of general income), such information may be less difficult and costly to compile through service providers that are part of local government infrastructure. On the part of local governments, the political implications of their involvement in targeting the food-insecure also need to be taken into consideration, as local government politicians may misuse targeting programmes to win votes.

At the local level, first-hand information exists on current and potential resources (human and natural), on the likelihood and character of food crises and on the environmental situation, and a way must be found to articulate local needs and inform policy-makers swiftly. This implies a monitoring function of rural local government bodies, covering issues such as harvest shortfalls, shifts in the local power structure to the disadvantage of the food-insecure, civil unrest, conflicts, disputes, etc., which may severely affect food security negatively at the local level. Local governments may respond to strengthen the management of cereal (rice) bank networks where necessary, provide regulations that suit the specific resource endowments and socio-economic and cultural specificities of different areas, and contribute to the mediation and resolution of conflicts affecting food security. They have easier access to inside information on who in a rural community is qualified and suited to carry out certain tasks that can make governance more effective and less costly, be it for conflict resolution or the maintenance of a village's feeder road.

New organizations and rural institutions may be created or existing ones modified to improve the flow of information (the "soft" infrastructure) both from central government to local bodies, but especially vice versa, through the participation and inclusion of key individuals in rural communities. Information should be shared in a two-way process between rural people and the different tiers of government. Information flows and indeed all transactions at the interfacing level must thereby become as transparent as possible, also to avoid an "appropriation of decentralization" by local (possibly in conjunction with urban) élite groups. In the food security equation, information issues refer typically to the physical and economic access to food and the stability and sustainability of supplies.

To the extent that the processes of decentralization allow lobbying on behalf of food-insecure and other disadvantaged groups, these groups may be able to make themselves heard and improve their situation. To facilitate greater participation on the part of these groups, policy-makers should take into account the different levels of social capital and institutional and implementation capacity, and organize themselves in the most localized manner to ensure quick and adequate responses to immediate local needs.

In the longer term, prevailing levels of social capital, organizational and institutional capacity should be assessed systematically, and ways and methods found to maintain or possibly improve these. If such policies are to become realistic options, the opportunity cost of devoting scarce government resources to purposes such as the strengthening of social capital may need to be examined carefully on a case-by-case basis. The latter is reflected in longer-term patterns of collective action that meet and build solidarity even in the absence of specific tasks to conduct (Goulet, 1989). Joint community action around what is considered an urgent issue may in that sense be regarded not as the organizing, but rather as the mobilizing of resources. Priority may have to be given to local institutional development, particularly in areas tormented by civil conflict and/or characterized by a high degree of social atomization and stratification. Since it creates organizational and implementation capacity at the rural community level, local institutional capacity-building is in itself a strategy for successful decentralization.

Decentralizing governments are likely to be facing severe shortages of qualified personnel for (old and new) local government functions, which require a wide range of practices and skills. Within the broader national framework of institutional restructuring, schemes for the retrenchment of state employees can be specifically targeted for bridging this gap. Such schemes should be proactive and could focus on retraining bureaucrats of the national government and administration units for the needs of future local government services. Irrespective of whether or not the decentralization process is already advanced enough to redeploy the ex-civil servants immediately in rural local governments, the retraining programmes may be designed flexibly enough to provide additional training in some of the skills that may find a wider application and become useful in the interim period.

A WIDER ROLE OF LOCAL GOVERNMENTS IN THE FRAMEWORK OF INSTITUTIONAL RESTRUCTURING

In relation to local government authorities, an urban bias is manifest in the strong urban orientation in the terms of reference of local governments. Webster's 3rd New International Dictionary in fact defines municipality as "a primarily urban political unit having corporate status and usual powers of self-government". To date, the "culture of local government" is mostly shaped by the many problems related to large cities, for the handling of many of which local governments are responsible. However, typically such problems are quite different from the scenarios that local governments find themselves confronted by in less populated rural areas.

Along with the consolidation of a new policy orientation based on increased participation of civil society, democratization and the fundamental principles of a market economy, a general trend towards the restructuring of public institutions and the progressive disengagement of the state has come to be perceived as necessary.

While it is in the majority of cases still early to evaluate the effectiveness of institutional restructuring exercises, the following general areas of consensus seem to be emerging:

1. Restructuring exercises are usually carried out as a result of nationwide fiscal crises.
2. They should be carried out swiftly so as not to give opponents time to organize obstruction.
3. They are more likely to be successful where and when a robust and lively central state apparatus is in place.
4. They should be carried out in a cross-sectoral fashion, with the aim of ensuring coherence in rationalizing the governance machinery, which implies the need for an authority to coordinate the overall restructuring process.
5. Where there are comparative advantages, tasks should be transferred to the private sector.
6. In accordance with the principle of subsidiarity, the efficiency of public institutions appears to be a function of the physical and relational vicinity with the clientele they are expected to service (a differentiated universe of individuals each with their own needs, not the amorphous public that governments have traditionally serviced), thus implying that - as part of the overall restructuring process - responsibilities need to be decentralized to the appropriate levels of government, where the delivery of services is most cost-effective and financially feasible and sustainable.

These points should be related to the following proposal; that:

1. government decision-making processes be articulated around an organic horizontal structure of concentric circles rather than a superimposed and "synthetic" pyramidal chart, as has been the case in the past;
2. local government instances be viewed as self-contained units with substantive financial and technical autonomy (for all functions for which the principle of subsidiarity has been established);
3. at the national level, the functions of government be limited to those that are truly of a national nature.

The United Nations Capital Development Fund (UNCDF) provides a definition of what should be considered the "viability" of local government units as "(...) the combination of a size (in terms of population served) and a capacity (in terms of local human resources and appropriate organization and procedures) that enables them to make planning decisions and assume full or joint responsibility for providing a reasonably wide range of local infrastructure and services. The size of the area and population covered by a local government unit is associated with the range of its actual (legally sanctioned) or potential (applicable through decentralization legislation) responsibilities for the provision of services. This relates not only to the nature of the different services and the economies of scale in their production but also to the degree to which a particular service lends itself to joint provision, i.e. to the sharing of responsibility for its provision among different levels of government and community-level institutions".

Point d) above is especially significant when it comes to decentralization exercises. Service problems in both rural and urban areas are increasingly arising not because of a lack of institutional capacity, but rather from the absence of an accepted institution that could play a strategic, coordinating role in service delivery. Point e) is most important for creating the enabling environment necessary for ensure food security in the long term. With respect to points f) and ii), ideally, most but not all rural services and responsibilities should be transferred to local government authorities (e.g. to the Panchayati of India or the Municipio of Latin America), vesting in them line responsibilities pertaining to the different sectors of government. Pareto-efficient delivery of services implies that these be controlled and financed at the scale where there are no major spillover effects (FAO/SDA, in press).

This points directly to one of the most important policy implications in respect of the relationship between local governments and decentralization. It has become evident that decentralization is a complex process that is not easily sustained over time. In a number of countries the process is still pending implementation or has been carried out halfway and "got stuck", although clear and extensive decentralization legislation may have been passed. The real impact of development interventions is naturally felt much more at the local level than at the regional and national levels, where it is less visible. As the benefits of local government measures materialize locally, civil backup to the decentralization process is likely to follow. Civil support of decentralization policies and of the people promoting such policies may create enough momentum to sustain decentralization over time and "push" the process through certain problematic bottlenecks.

Policy implications of rural local government bodies acting as intermediaries

Local governments should try to find new channels of consultation and communication at all institutional levels, strengthening especially bodies that may successfully carry out an intermediary role between the different stakeholders in rural development (particularly the bodies more easily accessible to rural women and female-headed households), and act as neutral fora of exchange of information and points of view. In Brugmann (1994, p. 138), a local "intermediary institution" (e.g. an ad hoc stakeholder planning group) is defined as "an institution which has a governmental mandate to undertake tasks for a local government outside the existing institutional constraints upon that local government. The institution is intermediary both in its facilitative role between government and other sectors/actors, and in its evolutionary role as a step in the process of reforming local government".

The restructuring of rural institutions in general and local governments in particular may build new bridges to draw on external support for development issues where necessary or useful, and foster other new forms of alliance and partnership in the rural areas. Such new forms may be the result of the increased flexibility that local government services can, by their nature, dispose of - for example with respect to the types of relationships and contractual arrangements that they devise. These may include informal, performance-based agreements that comprise a mix of private and public sector involvement. For example, Tendler (1997), who is quoted in FAO/SDA (in press), finds that in Ceará, Brazil, a local farmers' organization provides a particular public extension service officer with a list of queries and problems to solve within a given time span. The farmers' organization pays for the extension officer's transportation costs, which guarantees regular visits. This arrangement contains some of the incentives normally associated with the private sector.

As their historical users, people in the rural communities themselves have naturally the most information on the availability and use of local resources for food production. But as village environments have been changing rapidly over recent years, they are less likely to have complete or adequate information about the constraints on local resources for food production, which may be related to natural resource degradation and depletion. It could be said that, on the whole, the information gap between farmers and specialized government agencies has been narrowing (see FAO, 1996a). It should be a task for local intermediary institutions or, as part of their monitoring function, a challenge for rural local governments, to bridge this gap.

Thereby, external support may be brought in where necessary to assess, conceive and implement jointly viable solutions to food insecurity caused by problems of natural resource degradation. With a view to improving agricultural productivity and longer-term food security, external support may be sought through decentralized government channels to help identify large unrealized economic production potential. For a given land parcel under the agro-ecological conditions at hand, this may be based on the best cropping alternative to an existing monoculture, for example. Local governments may thus be able to link operationally the findings of research on superior cropping alternatives to the improvement of rural infrastructure in order to maximize the positive spatial impacts that a modification or upgrading of the physical ("hard") infrastructure will have on agricultural input costs, product prices, etc. (Schultink, 1992).

"To act as neutral fora of exchange of information and points of view" is of course easier said than done. The encouragement of dialogue locally is often made difficult since such dialogue may be subject to inertia, numerous specific problems and socio-cultural and political constraints. It may be helpful to organize and mobilize the stakeholders in local development around an issue that provides strong enough incentives for dialogue, for example around a topic of common interest such as an undesignated but earmarked grant to be disbursed to a rural geographic area, or the digging of contour trenches. Possibly, the goodwill necessary to win the trust of the "beneficiaries" of rural development may be introduced by making available a small, discretionary fund that is not tied to the bureaucratic requirements of regular budget procedures. The discretionary fund may provide a limited number of unconditional grants, which may serve as a basis for rural communities to obtain required additional funding from other sources.

If existing institutions are targeted for such objectives, successful mediation and support to get organized will, among other things, depend on their credibility, a function of their past effectiveness and historical legacy. In any case, it seems advisable that most intermediary institutions be given clear terms of reference and a mandate for only a limited period, after which they should themselves be exposed to a performance evaluation by the national authority in charge of coordinating the institutional restructuring process. On the other hand, an underlying legal mechanism could be established that permits the renewed creation of suitable intermediary institutions.

It is obvious that what has been termed the local "dependency syndrome", prevalent in some rural communities and resulting from past paternalistic, top-down approaches to rural development, must be changed. There is a danger of local governments continuing, or even increasing, dependency, given their vicinity to and continuous presence among the stakeholders in local development. With increased funding being made available to local government units, this danger may be exacerbated, unless it is made clear that their main priorities lie in capacity-building and mediation, and that they view these tasks as an immediate objective.

In sum, policy-makers ought to be cautious and realistic about what rural local governments can or should do. They must be clearly aware of the risks of decentralization, and of the substantive financial outlays required for effective decentralization measures and institutional restructuring. At least in the short term, not too much should be expected from such policy interventions, and the danger of a rush to "place all the eggs in the local government basket" must be avoided.

DECENTRALIZATION AND RURAL FOOD SECURITY

The relationships and correlations between rural food security and decentralization need to be explored much further. This article has pointed to the need for a comprehensive analytical framework that may be used to test various hypotheses and provide a backbone for a more systematic and consistent collection of empirical data (along the lines of Parker's "soufflé" theory of decentralization - Parker, 1995). The historical determinants of socio-political relations in the agricultural economy and the legacy of "urban bias" need to be an integral part of such a framework that could lead to a tentative typology of decentralization processes. This would facilitate the task of drawing some preliminary conclusions at a time when most countries are decentralizing to some degree - under different circumstances, in distinct ways and not necessarily for the same reasons (therefore, few appear to consider food security specifically) - and informing policy-makers on how to design partnership-based services strategies for rural local governments.

National governments need to decentralize in order to gain access to local-level environmental information, improve resource management and information flow on local food availability and markets and accomplish greater equity, productivity and cost-effectiveness in the delivery of services to sustain food production and distribution. Although the characteristics of rural food security are for the most part complex and locality-specific, this does not automatically mean that all agricultural and rural institutions should devolve the services that sustain food security in one way or other to the local level. The application of a spatial approach to food security issues would ensure not only that these are seen in their larger, intersectoral context, but also that the institutional level that links the local to the national arenas does not get overlooked. It is there that a greater degree of articulation between the different actors in rural development, between ministries and line agencies (through institutional restructuring and strengthening) may be conceived.

The severe difficulties that farmers and many small rural communities experience in attempts at organizing themselves collectively, particularly in sparsely populated areas, are well known. What may be needed for serious policy reorientation are not more "conventional" agricultural support institutions (rather, fewer more decentralized and efficient ones), but a variety of rural political institutions to complement them. These need to be institutions representative of rural interests, such as farmers' and women's organizations, and they must also address the crucial issue of sharing costs and achieving economies of scale. Where they exist, donors, international development agencies and especially NGOs should seek collaboration with local government units to promote new forms of rural alliance and partnership. New formulas could make use of, and increase, the potential flexibility inherent in informal relationships and agreements: they may reveal themselves to be crucial elements for the successful integration of rural food security into the decentralization agenda.

BIBLIOGRAPHY

Bagnasco, A., ed. 1988. La costruzione sociale del mercato. Bologna, Italy, Il Mulino.

Bates, R. 1981. Markets and states in tropical Africa: the political basis of agricultural policies.

Binswanger, H., Mundlak, Y., Yang, M.C. & Bowers, A. 1985. Estimates of aggregate agricultural supply response from time series of cross-country data. EPDCS Working Paper No. 1985-3. Washington, DC, World Bank.

Brugmann, J. 1994. Who can deliver sustainability? Municipal reform and the sustainable development mandate. Third World Planning Review, 16(2).

Conyers, D. 1994. Decentralisation and development: a review of the literature. Public Administration and Development, 4: 187-197.

Crook, R. & Manor, J. 1994. Enhancing participation and institutional performance: democratic decentralization in South Asia and West Africa. Washington, DC, ODA.

Evans, H.E. & Ngau, P. 1991. Rural-urban relations, household income diversification and agricultural productivity. Development and Change, 22: 519-545.

FAO. 1996a. Halting degradation of natural resources: is there a role for rural communities? By J.-M. Baland & J.-P. Platteau. Rome.

FAO. 1996b. Rome Declaration and World Food Summit Plan of Action. Rome.

FAO. 1996c. Growth theories, old and new. Economic and Social Development Paper No. 136. Rome.

FAO. 1997. Decentralization and rural development: FAO's experiences. By R. Paniagua. Rome.

FAO/SDA. Decentralization and rural development: The role of the public and private sectors in the provision of agricultural support services. By L.D. Smith. Rome, FAO. (in press)

Gould, W. 1988. Rural-urban interaction and rural transformation in tropical Africa. In D. Rimmer, ed. Rural transformation in tropical Africa, chap. 6. Athens, Ohio, USA, Ohio University.

Goulet, D. 1989. Participation in development: new avenues. World Development, 17(2): 165-178.

Hayami, Y. & Anderson, K. 1986. The political economy of agricultural protection. East Asia in an international perspective. Allen & Unwin.

Hazell, P.B.R. & Haggblade, S. 1991. Rural-urban growth linkages in India. Indian Journal of Agricultural Economics, 46(4): 515-529.

Hazell, P.B.R. & Röell, A. 1983. Rural growth linkages: household expenditure patterns in Malaysia and Nigeria. IFPRI Research Report No. 41. Washington, DC, IFPRI.

Lipton, M. 1977. Why poor people stay poor: urban bias in world development.

Moore, M. 1993. Economic structure and the politics of sectoral bias. Journal of Development Studies, 29(4).

Parker, A. 1995. Decentralization: the way forward for rural development? World Bank Policy Research Working Paper No. 1475. Washington, DC, World Bank.

Reardon, T. & Stamoulis, K. Relating rural agro-industrialization, intermediate cities and farm-non-farm linkages. (in press)

Rello, F. 1996. Intermediate cities and rural development: the case of Zamora, Michoacán (Mexico).

Rondinelli, D. 1981. Government decentralization in comparative perspective: Theory and practice in developing countries. International Review of Admininstrative Science, 47.

Ruttan, V. 1989. Institutional innovation and agricultural development. World Development, 17(9).

Schultink, G. 1992. Evaluation of sustainable development alternatives: relevant concepts, resource assessment approaches and comparative spatial indicators. International Journal of Environmental Studies, 41: 203-224.

Tendler, J. 1997. Good government in the tropics. Baltimore, USA, The Johns Hopkins University Press.

UNCDF. Local Development Funds: promoting decentralized planning and financing of rural development. UNCDF Working Paper. (in press)

Working Group on Rural Development. 1993. Guidelines for rural development in South Africa. Pretoria, Government of South Africa.

World Bank. 1995. Decentralized rural development and enhanced community participation: a case study from northeast Brazil. Washington, DC.