Home       About us   Issues     Search     Submission Subscribe   Contact    Login 
Conservation and Society
An interdisciplinary journal exploring linkages between society, environment and development
Conservation and Society
Users Online: 674 Home Print this page Email this page Small font sizeDefault font sizeIncrease font size

Previous articleTable of Contents Next article

Year : 2007  |  Volume : 5  |  Issue : 2  |  Page : 202-231

Where Community-Based Water Resource Management has Gone Too Far: Poverty and Disempowerment in Southern Madagascar

International Studies Programme, California State University, Long Beach, USA

Correspondence Address:
Richard R Marcus
International Studies Programme, California State University, Long Beach 1250 Bellflower Blvd, Long Beach, CA 90840-4605
Login to access the Email id

Source of Support: None, Conflict of Interest: None

Rights and Permissions
Date of Web Publication26-Jun-2009


Madagascar has struggled with the question of decentralisation for more than three decades. Since coming to power in 2002, President Marc Ra­valomanana has both reformed and accelerated this process, granting new roles and responsibilities to regional and community leadership. This politi­cal path is consistent with shifts in natural resource management in the 1990s, notably in the water sector. We thus see the role of the national gov­ernment diminishing in favour of resource management at the community level. This paper explores the impact of increased responsibility for water management and decision making in the southern district of Ambovombe­Androy. The assumption is that this sort of decentralisation leads to empow­erment at the local level and improves accountability, civic engagement and equity. Unfortunately, in the case of Ambovombe, 'local empowerment' quickly translates to 'you're on your own'. 'Decentralisation' quickly trans­lates into state disengagement. To avoid this, a finer relationship between state and local institutional relationships and responsibilities needs to be ex­plored. Only once we understand what a community is, and what its capacity can be, can we figure out what responsibilities it needs to take on to ensure the efficacy of a state that tends to be at best inefficient and at worst preda­tory.

Keywords: Madagascar, Ambovombe, water, CBNRM, community, power, decentralisation, governance

How to cite this article:
Marcus RR. Where Community-Based Water Resource Management has Gone Too Far: Poverty and Disempowerment in Southern Madagascar. Conservat Soc 2007;5:202-31

How to cite this URL:
Marcus RR. Where Community-Based Water Resource Management has Gone Too Far: Poverty and Disempowerment in Southern Madagascar. Conservat Soc [serial online] 2007 [cited 2020 Aug 11];5:202-31. Available from: http://www.conservationandsociety.org/text.asp?2007/5/2/202/49230

   Introduction Top

AMBOVOMBE-ANDROY is a poor district in Southern Madagascar. 'Poor' is too generic. For 8 months out of the year there is no water. No water for agri­culture. No water for drinking. As a result, the usual impediments to devel­opment are rarely discussed. When there is no water all other activities lose meaning. As a result, Ambovombe-Androy is a district at the margins of an oft marginalised country. The question is what is being done about it. Na­tional policies are consistent with international norms that support Integrated Water Resource Management. Thus, the role of the national government is diminishing in favour of civic group resource management at the local (com­munity) level. This sort of decentralisation is intended to empower the local population to improve accountability, civic engagement and equity. It ad­dresses the greater responsiveness capacity of local authorities to local popu­lation needs while improving efficiency, equity and local 'ownership' of the governing process (Rondinelli et al. 1989; Manor 1999; Crook and Manor 1998; Blair 2000; Ribot et al. 2006). Decentralisation is thus an innately par­ticipatory action that is consistent with the deepening of democratic norms. Yet, for these norms to be actuated, citizens must not only be able to vote and hold office but also hold the power to act collectively (Blair 2000).

Unfortunately for the case of Ambovombe, 'local empowerment' quickly translates to 'you're on your own'. 'Decentralisation' quickly translates into state disengagement. In this sense the state actually undermines the decen­tralisation process. In a multinational study, Ribot et al. (2006: 1864-1865) determine that the incomplete process of decentralisation sees states grabbing power by limiting the types of power that are transferred and choosing local institutions that serve central interests. As a general rule Madagascar appears to reflect both of these patterns. Yet, it presumes a vested need for sectoral power by the state in the region for neopatrimonial or other ends. In the case of Ambovombe-Androy the process appears closer to marginalisation. The need for central power in the region's water sector is minimal as long as there is no infrastructure. The lack of water is the power. This paper seeks to apply a finer lens in looking at this relationship between state and local institutional actors and their responsibilities.

Travelling west on Route Nationale (RN) 13 from the southern port city of Tolagnaro starts with a visual feast. More than 1500 mm of rainfall each year brings lushness to the area as the lowland rice paddies blend towards the 4000 ha of humid forest. Warm, humid winds engulf the region. After about 30 km the road begins to wind up into the Anosy mountains. The mountains provide an orographic effect, effectively blocking the ocean winds and creat­ing a microclimate. Looking east of the mountains the humid zone is dotted with what remains of primary rainforest. Descending down the westbound road into the Amboasary-Sud district, rainfall averages drop to 600 mm per year and the flora begins to transition into dry-deciduous forest. Wet rice paddies give way to commercial sisal (Agave sisalana) farming. By the time the road reaches the town of Amboasary, only 70 km from Tolagnaro, the dura­tion of the starving season increases from 2-3 months to 4 months. The levels of development decrease two-fold and the level of marginalisation increases. [1] The saving grace for Amboasary is the Mandrare River. While irrigation in­frastructure is limited, its banks and alluvial soils have ensured that domestic water needs are met for the 184,000 inhabitants (official figure) of the district and that 70 per cent of the population remains in the agricultural sector.

We continue west on RN 13, crossing the Mandrare's ailing steel bridge. The volcanic ridge delves into a hydrologically closed basin as we enter Am­bovombe-Androy district. Rainfall averages drop by half in the 36 km it takes to reach the district capital of Ambovombe. The forest is gone and in its place unconsolidated sandy soils give rise to the beginning of the unforgiving Spiny Forest ecosystem. The percentage of population engaged in agriculture drops by 20 per cent. Rice cultivation is replaced by manioc and sweet potato. Irri­gation infrastructure (generally for rice and manioc) drops from 49 per cent of farms in Amboasary to zero in all but one Ambovombe commune. Survival is based on the management of 14,000 heads of cattle spread out among the population [Figure 1].

Even worse than the lack of irrigation water in Ambovombe is the lack of wa­ter for livestock and domestic consumption. Rainwater harvesting is limited to wealthy small holders, generally those who inherited the houses built by French settlers during the colonial era. With so little rain, and most of it arriv­ing within a 3-month period, collection facilities are costly and difficult to maintain. There are no rivers. Groundwater resources are scant. As detailed below, limited resources and extreme demand drives prices to unobtainable levels. 'Development' becomes a tattered shroud for poverty management. Those who live on less than a dollar a day are considered poor. By this meas­ure over 85 per cent of the population is poor, compared to between 65 and 75 per cent in Amboasary. [2] But this measure has little meaning. Amboasary is still poor. Forty-nine point one per cent of Madagascar is poor. Three-hundred and twenty-three million people in sub-Saharan African are poor (UNDP 2004). It is a figure so aggregate as to mask vast differences in poverty. In the case of Ambovombe, it masks intra-community class structures, and, more important herein, the stark difference from most of the rest of Madagascar that there is little hope within the population and little action on the part of the na­tional or international community to move beyond humanitarian services to poverty alleviation, let alone 'development'. According to a local mayor (in a June 2005 interview):

'In our commune, the problems are numerous and it is not possible to cite them all here. But we can summarise them like this: our commune, and our region of Androy, is marginalised by the central government. Even the president of the republic doesn't want to visit. There are conse­quences of this marginalisation. We are the poorest part of the country of Madagascar. Moreover, the rain doesn't fall here and that aggravates the problem of water and chronic famine is irresolvable. For our commune our fundamental problem, which we often say to the state, is the absence of water.'

The development focus is quite narrow: augment the water supply through improved institutional management at the local level. [3] However, there is little water supply to manage and efforts to augment supply are routinely scuttled.

In this paper, I use the case of Ambovombe-Androy to argue that interna­tional water policy norms that advocate decentralising water management and decision making to the local level do hold the potential for positive outcomes, but they can also increase, rather than decrease, water scarcity. Community organisations can be highly effective but only where certain conditions are met. They are effective at turnout or primary delivery point level (Rosegrant 2002)-when the regularity of the resource is assured by the state (or the state in partnership with regional government and/or the private sector) and the definition of community, its rights, responsibilities and capacities, reflect that of the local population. In contrast to dominant critiques of global approaches such as sustainable livelihoods (Chambers and Conway 1992), resource impe­rialism (Adams 2003) and free markets (Barlow and Clark 2002), my finding is that political decentralisation associated with resource management local­isation and implemented as part of an untested perception that budding de­mocracies are enhanced by local empowerment, has led to a disengagement of the state, its economic resources, and its management capacity. Rural commu­nities are suddenly faced not with increased opportunities for institutional par­ticipation, transparency, or oversight, but with having to pay exorbitant costs for water delivered in irregular intervals. [4] They must manage the resource at these irregular intervals or rise to the challenge of managing the resource at the supply level. People in Ambovombe are savvy, but ill-prepared to carry out these imposed municipal functions and unable to raise the level of user fees or community taxes necessary to fund infrastructure development. As a result, systems are neither complex nor adaptive [5] and decision making at the local level is limited to market opportunities at hand. Donor-funded projects support 'appropriate' technologies that prop up overpriced and unsustainable water delivery markets and mine water without proper scientific review. The result is the concomitant manufacturing of economic hardship and water re­source scarcity. I thus argue that the universal decentralisation of water does not improve participatory (in some circles considered 'democratic') access as much as it undermines effective governance. A new, more nuanced view of state responsibilities and community participation that ensures both resource perpetuation and not just participation but viable choices by community members is now overdue.

   International Water Policy Norms and the State Top

In 1977, The United Nations held the first Conference on Water in Mar del Plata, Argentina. The Action Plan that followed (1983) concluded that '…... relatively little importance has been attached to water resources systematic measurement. The processing and compilation of data have also been seri­ously neglected'. The 1987 Brundlandt Report (Our Common Future) helped us further define the problem. Bringing to currency the term 'sustainable de­velopment', it argued that development must meet the needs of the present without compromising the ability of future generations to meet their own needs. Sustainable Development challenged a growing notion that resources should not be consumed and instead agreed with the Lockean notion that the Earth was there for us but with the caveat that it must remain intact for our progeny. With water singled out, the global norm that it should be consumed, but not depleted, was launched.

'Sustainable' water use was further codified in the 1992 Dublin Confer­ence. The same year, at the Earth Summit in Rio de Janeiro, advocates of sus­tainable development argued that 'all peoples, whatever their stage of development and their social and economic conditions, have the right to have access to drinking water in quantities and of a quality equal to their basic needs' (UNDESA 2002). This fundamental commitment to doing what it takes to ensure water resource sustainability has been reiterated in international fora repeatedly over a decade.

If we are agreeing to consume water despite the recognition that there is a growing crisis of scarcity then we have to decide who gets to make decisions about that process? Is it the role of government, in which case we should discuss mechanisms and bureaucratic modalities for water governance or, is it up to the individual as regulated by civil society? Writers considering sustainable devel­opment (UNDESA 2002; Chambers 1997; World Bank 2003), common pool re­source management (Korten 1986; Berkes 1989; Poffenberger 1990; Western and Wright 1994; Ostrom et al. 2002; Deitz and Ostrom 2003), and global norms in water management policy (WMO 1992; OECD 1998; Kemper et al. 2003) all have come to point in the same direction: (1) all resources, especially water re­sources, need to be managed, (2) decision making across multiple levels fosters sustainability, and (3) communities, especially rural communities, have a greater interest in managing resources sustainably than state or corporate managers. Communities are more aware of local environmental processes, and thus can be mobilised to manage the resource within multi-level natural resource use re­gimes. In the water sector, these assumptions codified in the Dublin Accords are reflected in attempts at complementing centralised physical infrastructure with lower cost community-scale systems, decentralised and open decision making, water markets and equitable pricing, application of efficient technology, and en­vironmental protection (Gleick 2003). [6] The four `Dublin Principles' read (1992):

Principle No. 1 Fresh water is a finite and vulnerable resource, essen­tial to sustain life, development and the environment

Principle No. 2 Water development and management should be based on a participatory approach, involving users, planners and policy mak­ers at all levels

Principle No. 3 Women play a central part in the provision, manage­ment and safeguarding of water

Principle No. 4 Water has an economic value in all its competing uses and should be recognised as an economic good

Consistently, the enhanced role of local communities in construction, owner­ship, management, administration, operation and maintenance of water sup­plies is considered as a good thing. Not often considered, however, are the dynamics of the communities themselves. How well does the state function at the community level? What is the relationship between local government and civil society? What is the role of equity of stakeholders in the decision­making process? If participation is determined by the right to participate and income is treated as a constant then the ability to participate is not a factor.

What is the relationship between sustainability and equity in accessing the commons and how is this affected by institutional structures?

Given its ability to engage diverse cultural norms through a universal sus­tainable development tenet, the idea of multi-level governance or polycentric governance institutions is normatively appealing. Yet there is little explora­tion of its empirical validity; where it is explored, it is generally in an attempt to maximise multiple actor regime effectiveness. As Ostrom (1999) has elo­quently argued, in order for multi-level resource use designs to be successful multi-level systems they must be viewed as complex adaptive systems in which there is no dominating central authority. Linkages need to be made be­tween international, state, and local interests in order to determine who gets to make what decision about water resources and why they get to make those de­cisions.

Agrawal and Ribot (1999) discuss this problem of linkage between the state and local interests as a relationship gap between actors, power and institu­tions. When actors do not share goals for conserving resources and are un­equally powerful, as is commonly the case, institutions are significant for two reasons. On the one hand, they denote some of the power relations (Foucault 1983: 222, 224) that define the interactions among actors who created the in­stitutions; on the other they also help to structure the interactions that take place around resources. Once formed, institutions exercise effects that are in­dependent of the forces that constituted them. While they characterise interna­tional actors as working within such state thinking, it is not an intellectual stretch to consider international actors as an integral part of the power rela­tionship, particularly in cases like Madagascar where international and state policies are virtually fused.

In a different work, Agrawal (1999) finds that there are processes not cap­tured by spatial allocation. The narrow geographic focus of managers at local levels makes them inappropriate managers for geographically larger re­sources. This appears to be the case in Ambovombe where water supply is and will increasingly be tied to geographically broader supply areas; the level of scarcity within the district and broader (deconcentrated) region mandates ne­gotiated interactions across localised and even regionalised governing units. As the state level governing authority in the water sector cedes responsibili­ties to lower levels that lack the capacity or jurisdiction to secure water, and the state disengages, the likely outcome is not a power struggle-it is a power vacuum in which there is no authority capable of addressing the dire water re­source needs that are both longstanding and rapidly growing.

To be fair, the problem is recognised by the people who implement policy in the field. For instance, the World Bank funded a 'pilot project' for water in Southern Madagascar called PAEPAR (Projet d'Alimentation en Eau Potable et d'Assainissement en milieu Rural). The project exemplifies the decentral­ised approach that dominates international water policy. The articulated pro­ject objectives were to: (1) 'endow the public sector, the private sector and the communities with institutional means and appropriate technical materials so that the supply of water and sanitation in rural zones is assured in an effica­cious, economical and permanent manner' and (2) 'to bring communities to participate in the preparation, execution and management of the installations of potable water, and contribute to costs of maintenance'. The project litera­ture goes on to delineate, in order, the manner of local participation and the relationship of the local groups it creates to government and international in­stitutions across levels. Presumed is that the answer to the water debacle in southern Madagascar lies at the community level and in local participation. This significantly limits the types of water supply that can be investigated, as larger supply projects that transect multiple communities require higher levels of management. As discussed below, the Bank's project is taken by all in­volved, including those who funded it, to be largely unsuccessful. In a 2004 interview with one World Bank official in Madagascar, for instance, the au­thor asked: if this model has failed then what is the alternative? He replied: 'I'm eventually going to think about [that]'. What is important here is that this official identified the local approach as a 'fractured' approach that under­mines the end goal (sustainable water resource use while augmenting supply) but one that the Bank is unlikely to move away from in the near future. [7] Nearly 3 years hence his response appears to be accurate in this setting as the PAEPAR project was not renewed and new World Bank efforts in the water sector, integrated into the 2007-2010 Country Assistance Strategy, focus on ensuring the viability of the (urban and periurban) water utility and expanding irrigation in the high-production rice bowls.

The Malagasy state is following the universal policy trend in land, forest and water management. This brings up two questions: how does the decen­tralisation take place and what is the nature of governance over that process? Here Robin Mearns' (2004) discussion of decentralisation in Mongolia is in­structive. Mearns (2004: 133) argues that 'contemporary Mongolia is charac­terised by a mosaic of formal and informal institutions, including the results of new experiments in policy making seen alongside the remnants of old ar­rangements and patterns of behaviour'. The institutional reforms introduced by the state are multitudinous and overlapping. 'While the stated intentions of many of the reforms in question are to promote social inclusion and environ­mental justice, their practical outcomes currently have precisely the opposite effects' (134). In some cases, she argues, the decentralisation is consistent with both democratic decentralisation norms and local institutions. In many they are not, thereby causing an imbalance of power between actors that un­dermine local capacity for holding leaders accountable. The critical distinc­tion here is between decentralisation ('formal transfers of power to actors and institutions at lower levels within a political-administrative and territorial hi­erarchy' Agrawal and Ribot 1999) and democratic decentralisation ('power and resources are transferred to authorities representative of and downwardly accountable to local populations' (Ribot 2002)). Mongolia's decentralisation, consistent with international norms, often fails to meet the democratic litmus test and the lack of accountability in the process leaves but a thin institutional hull incapable of serving vital management functions. Moreover, there is an ongoing crisis of fiscal centralisation.

Madagascar is faced with the same challenge. There is a similar mosaic of informal and formal institutions at the local level. Democracy at the macro­level may be growing horizontally but decentralisation is failing to allow for downward accountability to local institutions in a vertical manner. The prob­lem in Madagascar is made more complex by the overlapping deconcentration process. Where decentralisation transfers power to the local level, deconcen­tration creates regional and/or local institutions that are vertically accountable upwards to the central administration. The twenty-two regions in Madagascar, created by President Marc Ravalomanana in 2004, supersede the district power while adding to the (decentralised) commune-level responsibility. Communes, the most local level of governance, are thus accountable to their local populations but responsible to the regional leadership appointed by the central government. [8] In Madagascar the creation of deconcentrated entities is thus an institutional effort to undermine the accountability requisite for decen­tralisation to succeed. The lack of fiscal decentralisation apparent in the Mon­golia case is even more alarming in Madagascar. The central government collects 97 per cent of total revenues. The decentralised level, communes, ac­count for 2 to 3 per cent of revenues with a significant urban bias in distribu­tion. Deconcentrated entities, regions, account for approximately 10 per cent of revenue expenditure. While responsibilities for development might be shift­ing downwards, financing for it is not creating a significant fiscal gap (World Bank 2004).

Turning to the water sector, there is a Ministry of Decentralization in Madagascar, but that is to ensure compliance with decentralisation of political institutions. Natural resources fall either under the Ministry of Water and For­est or the Ministry of Energy and Mines with water falling under the latter. As put by Herivelo Rakontondrainibe of the Malagasy Directorate of Water and Sanitation, Ministry of Energy and Mines in September 2003: 'now we have a policy, we have a strategy, and we know the importance of community in­volvement, and of the need to decentralise…' Madagascar's Water Code, written in 1998, states as its primary point that water should be universal, well managed, and not a free good. Water is necessarily in the public domain, but must be managed and conserved by assigning a value. The management, dis­tribution, organisation, and financing can be public or private, but it must in­clude a cost recovery mechanism and must involve local participation. It also follows the trend which began with Dublin that water is not treated as a whole. Drinking-water policy and infrastructure would be separated from ag­ricultural and industrial policy and infrastructure. From a delivery perspective this makes sense. From a resource management perspective, however, it ap­pears somewhat strange since the resource supply is the same, the population of the country is largely agrarian, and the country's economy is primarily ag­ricultural. The agricultural sector is completely omitted from the primary goals of the document. Chapter 3, Section 1 of the Water Code does ulti­mately address irrigation needs. Yet even this section merely describes the process of obtaining irrigation rights (via the National Water and Sanitation Authority) and does not say anything about management, organisation, policy, or financing of water for agricultural purposes. It would appear that the Water Code is inconsistent with the Millennium Development Goals that President Marc Ravalomanana enshrined in the Madagascar Action Plan (the hallmark of his development efforts) in March 2006: reduce by half the proportion of people without sustainable access to safe drinking water by 2015 (goal 7), reduce by half the proportion of people living on less than a dollar a day (goal 1), and reduce by half the proportion of people who suffer from hunger (goal 1).

The last of these in particular appears problematic. Agricultural water pol­icy that does exist is not well codified, but it does have certain objectives. These are to maintain irrigation infrastructure, reduce the state contribution to irrigation, help the state disengage from the management and maintenance of irrigation, and increase infrastructure investments from rural users. Consis­tent with donor policies, it is attempting to do this by establishing community­level water users associations (parallel to the commune-level government) and giving management responsibilities to them (Dinar and Subramanian 1997).

There is an empirical question here about state functionality. Despite sig­nificant reform that brings in community voice, there is systemic failure of in­stitutions to sustainably deliver sufficient water. Does the state provide an enabling environment where the diverse stakeholders can engage at the com­munity level or is it merely devolving responsibility? Is it that the relevant population does not accept the state, successfully rebuffs it, or otherwise fails to be captured by it? [9] Does the state itself engage in resource exploitation for the primary purpose of its own empowerment (Scott 1999) thus mandating its marginalisation in favour of community decision making? Is state disengage­ment about local empowerment or about self-seeking leaders ridding them­selves of the responsibility of expensive water sector development? Is the state a (lame) leviathan responsible for the equitable development of water re­sources?

   The Case of Ambovombe Top

'This is the remotest area from Tana [the capital]. People don't see really what's happening here, so most of what the programming for this region is, isn't really implemented because it's very remote and people don't really know exactly what's happening, they don't know. So they can take advantage' (Male resident, Ambovombe town).

Ambovombe-Androy is a district (fivondronana) of about 184,000 inhabi­tants within the Tulear province. Divided into seventeen communes, there is a fair amount of economic and livelihood diversity. [10] A majority (69 per cent) of the population (author's survey) claim farming as their primary occupation and none claim herding as their primary occupation. This can be attributed to a distinction, common throughout southern Madagascar, between the notion of farming as a labour pursuit and herding as a social pursuit even if it is for economic ends (Rakotoarisoa 1998). The distinction between pastoral and ag­ricultural activity is thus artificial even though in Ambovombe pastoralism dominates the economic and social milieu (Heurtebize 1986). Ambovombe is one of Madagascar's poorest districts, and is one of the three districts in southern Madagascar that suffer from extreme water scarcity. Ten of the country's eighteen communes that regularly experience severe food scarcity are in Ambovombe district. Associational life is weak, and concomitant social capital is low (Marcus 2000). Despite donor and government efforts, only one commune (Antanimora Sud) claims the existence of a regularised water user's group.

The extreme east of the district benefits from the Mandrare River and the extreme north of the district ascends the northern boundary of the basin to where rainfall is, if not plentiful, adequate. Three communes suffer from wa­ter scarcity, and have coping mechanisms that may be inadequate, but func­tion well enough that food, access to markets, or education are of a higher concern. For twelve of the seventeen communes, the majority of the district's population, water dominates daily concerns, household labour and the local economy.

The hyphenation in Ambovombe-Androy comes from its distinct ethnic identity. Tandroy is one of Madagascar's eighteen ethnic groups and Ambo­vombe is the capital of Androy. Amboasary-Sud district to the east is pre­dominantly of Tanosy ethnicity and there is modest social tension between the two groups for both historic- and resource-based reasons.

The Tandroy people have long lived with water shortage. The word 'An­tandroy' itself means 'people of the thorns' in reference to the Spiny Forest in which most of the population lives. There are western accounts of water con­cerns dating back to Etienne de Flacourt 400 years ago (with more notable de­scriptions at the onset of the French colonial period). This has resulted in a history of diversification in both agricultural production and animal hus­bandry, high migration trends (Heurtebize 1986), and the use of raketa (cactus fruit) for meeting animal (and sometimes human) water consumption needs. The problem today is significantly magnified. Population pressures through­out Androy and neighbouring Androsy and Mahafaly regions have limited migration options. Some seek alternative employment in regional cities or the capital but jobs are few and far between (Heurtebize 1986). The use of raketa was a source of conflict during the colonial era (Kaufmann 2001). At times these conflicts were severe enough to obviate its usage. Environmental change and increased consumption in recent years has significantly reduced the raketa availability even as the current debate is whether its production should be increased as a socioeconomic good or curtailed as an ecologically damag­ing invasive species (Middleton 2002). As a result, when the people of Am­bovombe claim they have decreasing water access their concerns tend towards two buckets per day per household-about 5 l per person (1/100th of average US water consumption). The World Health Organization view is that below 20 l per day is water poverty. The United Nations argues that below 50 l per day is water poverty.

Part of the decline in water availability has to do with climatic shifts that have led to a decrease in rainfall. The country as a whole is expected to see water availability decrease by nearly two-thirds by 2025. Following from Gleick (1993), Madagascar withdrew 41 per cent of its 40 km 3 /year renewable water resource. Estimates today range as high as 60 per cent of its renewable water resources. Madagascar already is a country in water stress by diverse measures (UNEP 2002b; Falkenmark 1989). With only 45 per cent water sup­ply coverage (WaterAid 2006) and a population growth rate of 2.8 per cent Madagascar is at risk for entering into water scarcity. Even more challenging, the supply is poorly distributed (UNEP 2002a; WaterAid 2006). As a result, some areas, particularly the south of the country, already see high levels of water stress while others, such as the northeast, are among the wettest places on earth. Madagascar has suffered along with East Africa in El Nino-Southern Oscillation effects and since 1968 there have been decreases in rainfall (UNEP 2002a). There is some evidence that climate change is having a positive effect on water drainage areas in some parts of Madagascar and a negative effect in others (de Wit and Stankiewicz 2006) and further evidence of a disproportio­nate change in climate in southern Madagascar (Vallet-Coulomb et al. 2006).

The larger concern, however, is the dramatic population growth as a whole. The cultural value and social status based on family in Androy, and the value of sons for herding, buoys a growth rate significantly higher than the national average.

'Most holes are dry here. Water's rare…so the price is 750 (francs], but that is not something that you just can find anywhere. You have to try hard to access a bucket of water because there's none anywhere. If there is like a small quantity in town, everyone just comes here and buys it and then it's not, not sufficient for everyone' (Male resident, Ambovombe town).

Water in Madagascar has historically been the responsibility of the Depart­ment of Water within the Ministry of Energy and Mines. The 1980s saw a seesaw imbalance between market incentives and wanting water resource in­frastructure. This led to the regularised need of World Food Programme (WFP) intervention. This forced the Malagasy state to recognise the urgency of the water crisis in the south and it created the Alimentation en Eau dans le Sud (AES). The AES operates exclusively in the south of the country, operat­ing parallel to the Department of Water within the ministry. It was created by presidential decret 86-241 in August 1986 with the idea of organising statutes for water management in the south. It is the only public water supply interest in the region. It has an operating budget of Malagasy Ariary (MGA) 496 mil­lion (US$ 271,483). There is no provision in the AES budget for investment. Ambovombe town, home to about 46,000 people, is the dusty district capital. There are a small number of water-harvesting facilities and several publicly owned boreholes. The AES has several other borehole facilities in town, but most are too saline for human consumption. The remaining water needs, which are plentiful, are met in part through private boreholes. One part of town in particular has had success in digging freshwater wells and there is no shortage of profiteering by their owners. The AES rate for water from boreholes, when it exists, is 30 MGA per bucket. In April or May, just after the rainy sea­son, private water may cost 1000 MGA per barrel (500 l) or 50 MGA per bucket (15 l). By September those same providers commonly charge 300 MGA per bucket. In late 2003, 500 MGA per bucket was not unheard of [Figure 2] and [Figure 3].

The problem is that in May there is some water in rural areas left over from the rains. By September that water is long since dried up. People dig in the sand to fill buckets one cup at a time and, ultimately, are forced to travel to Ambovombe town to fill a water bowser (a barrel on the back of an ox cart). This strains Ambovombe town's meagre water supply.

'The rainy season runs through December, January, February, perhaps up to June. There is often rain, so the need for water is less. But during the dry season from July to September, October, November, when it's dry, then the need for water from, from Amboasary more often. So we must rely on them because [here] it's dry' (Female resident, Ankare­mena village).

The WFP, among others, has tried to create alternatives. In the 1990s, the WFP helped create storage facilities in rural areas. Unfortunately, most of the time the water was not plentiful enough for it to last more than an extra month, and the committees built to maintain the storage areas were ineffective at doing so. Further, the practice of digging for water in Ambovombe leads to very high turbidity. Clay-like solids in the water require removal and the wa­ter requires purification. Solids are not necessarily harmful in themselves, but in this mineral region mica or other contaminants can cause illness. Solids can also carry pathogens ( Escherichia More Details coli, Cryptosporidium, Giardia lamblia, etc.) and they make it difficult to purify the water to reduce diarrhoea, cholera and other diseases. The purification process requires boiling, and fuelwood is very dear. As a result, well-intentioned programmes like that of the WFP bring about only a small augmentation of low-quality water.

In 1993 the Government of Madagascar, through the AES, entered into a cooperative agreement with the Government of Japan for the amelioration of water supply in southern Madagascar. The strategy was curious from the start. Two dozen water trucks were purchased by the Japanese government and a gravity pump-fed water filling station was built alongside the Mandrare River. [11] The AES was put in charge of managing and maintaining the new equipment and intervening in the Ambovombe water market. Each of the trucks held 6 m 3 of water and could fill, drive to Ambovombe, discharge, and refill four times a day for a total of 576 m 3 per day. This is a scant contribu­tion to the realised need of the population, but nonetheless critical [Figure 4].

The administrative process set up is that each fokontany (the lowest admin­istrative unit in the Malagasy system of governance) is to ask its commune for water. The Ambovombe communes tender a request for water to the AES. The cost was set at 15,000 MGA for a commune, 24000 MGA (about $48 in 1993) for a private request and has remained constant (about $13 today). According to information from focus group discussions with community members, the system was corrupted from the start, as drivers would hold the water hostage demanding a gift for the delivery. As the years passed the AES fell short of its maintenance responsibilities. By 2004, only six of the twenty-four trucks were still in operation. [12] According to the AES regional operations director there is a 'problem of the tyres'. He explained that it is difficult to get replacement tyres for the trucks in this area. The author tried to clarify:

'Just the tyres-is this the only problem with the other trucks'? 'Yes, just the tyres.' 'If I were to buy new tyres for the trucks…?' 'Then we would be very happy.'

'When [the AES water truck project] started it was running well; it was good, but that was back when it started. Then people started to be cor­rupted and started taking parts of the truck to sell. Like taking wheels, tires and sell them to other private individuals so things like that. I am not very confident that the government would be good to run the pro­ject. Maybe it would be better if it was like private, if it was like pri­vately run…[the project] should be closely inspected to make it run properly or else just people get corrupted easily…the people here are corrupted and also the mentality of people wants this money, they steal the money so it never reaches its destination' (Male resident, Ambo­vombe town).

The author was able to account for six other trucks in various states of disre­pair; four in the AES repair yard and two pieced out for private household wa­ter storage within Ambovombe town. Community members throughout the district accuse the AES central office of being corrupt. Specifically, the accu­sations are that officials sold off trucks in part or in whole for their own profit even before the period of disrepair began. This is wholly unsubstantiated, but clearly there is more of a problem than obtaining new tyres [Figure 5].

The smaller number of trucks means each truck's cargo has a higher pre­mium. Water pump station officials report that each truck comes two to three times per day (not four). The process of tendering a request to the AES seems to have completely broken down. Communities report that a truck arrives every 1 or 2 months. It is generally half full, the other half having been sold off en route to wealthy landowners with storage facilities. The expected gift to the driver has risen to equal the cost of the tanker itself. Presuming that six trucks make three runs per day and the water was to make it to destination that would mean 108,000 l or less than a litre per person.

While corruption is always a controversial discussion with communities and local officials, price is not. The pricing problem is universally agreed upon. One hundred twenty thousand francs is a tremendous sum in the district, one that community members feel they can pay but with hardship. However, that works out to only 20 francs per litre. It is not enough to meet cost recov­ery goals. It is not even enough to pay for the petrol for the delivery, let alone considerations of infrastructure maintenance, administrative costs or infra­structure growth. Community willingness to pay testing is still pending, but it appears that 400 francs per bucket (about 26 francs per litre) will be consid­ered close to the higher sum. There is reason to question this local perception as need drives people to pay the 24000 MGA demanded by the driver in addi­tion to the 24000 MGA for the water. It is likely the AES can raise its prices, and it probably should, but even at double the price it would not be enough to rectify the price gap.

There are strong advocates within the AES for private vendor markets. This has in fact been tried. However, private vendor markets can only work if ei­ther (1) they can charge enough to justify the business or (2) the government subsidy is high enough to compensate for the low pricing. The population cannot afford the real cost of trucked water. Where should the money for the water come from? Some officials argue that Antananarivo should be subsidis­ing water in southern Madagascar. However, tax-based redistribution is unlikely to work. The Malagasy tax structure is such that, even by African standards, there is high share of customs-based taxes and revenue. Household­and community-level taxes are not well collected. The capital thus does not have the resource base to subsidise water through direct tax redistribution. It would have to be from the country's meagre general funds. Private vendor markets will only exacerbate the problem. The problem here is that the deliv­ery mechanism itself is too expensive. [13]

The AES, and more importantly the Ministry of Energy and Mines, recog­nises the shortcomings of this system and appears to consider it a band-aid. The primary intellectual efforts are going towards increasing community par­ticipation. Communities need to be involved in supply, facilities management and cost recovery. This is the heart of the approach used in the AES's collabo­ration with the World Bank and the Projet d'Alimentation en Eau Potable et d'Assainissement en milieu Rural (PAEPAR). The PAEPAR has a US$ 17.3 million budget. The primary activity has been groundwater exploration. Spe­cifically, a World Bank-funded technical advisor determines where it is most likely to find a fresh water supply and then works with the AES and the com­munity to dig a borehole. The average cost is about US$ 9600 per hole. Where fresh water is found a water users' group is set up in the commune to manage the borehole and ensure that community members pay for it. Price fluctuates with supply but can range from 100 francs per bucket (6 or 7 francs per litre) to 1500 francs per bucket (100 francs per litre).

Project staff report that community groups are equally problematic. Where they are parallel to the commune they are often thought of as challenging au­thority. Where they are merged into the commune they often lose the support of the population. Collecting maintenance fees has proven especially prob­lematic. This local-level corruption appears associated more with the AES than with the Bank in that it permeates Bank and non-Bank-funded projects alike. In one village, for instance, the mayor appoints the water users' group. However, the group runs parallel to the commune and is not subservient to it. The population therefore cannot hold the members of the group accountable for their actions. It collects 2000 MGA per household per month for mainte­nance but does not actually deliver water.

The problem is the taps. The village has a pump system that circulates wa­ter to three delivery points. Obtaining water is simply a matter of opening the tap and filling a bucket. Unfortunately, the tap itself is broken. At the time of study it had not been functioning for more than a year? Why not? Surely a new tap costs less than even 1 month of what is collected monthly for mainte­nance. The committee says that it is the job of the AES to fix the tap. They gave the money to the AES but the AES has not done it. The AES says that no request was ever made and no money was ever paid [Figure 6].

In some ways this is a luxury problem. Of the more than 600 boreholes dug, only a small portion has yielded fresh water. Many have been dry, most have been saline. Often the latter helps stave off cattle woes, but humans do not find it fit to drink. Community leaders and some long-time expatriates chuckle at the inefficacy. Hydrological studies are wanting, but they know that communities have themselves tried to mine water for decades. The tech­nological jump provided by the World Bank allows for flow control and more careful bores that reduce saline intrusion, but it is far from a panacea. Even the World Bank technical advisors and project leaders realise its shortcom­ings. The district cannot rely on groundwater.

So why dig boreholes if they are so unlikely to yield potable groundwater? A community group can manage a borehole. It is entirely within the bounda­ries of the commune. The World Bank will only fund projects that can result in community participation and commune-level management. Bringing water from outside of the district is thus effectively a transboundary challenge re­quiring management at a level above the commune. The problem is that the district is predominantly comprised of a hydrologically closed basin and the water within the basin is largely saline-there is not enough water in the basin for the population. Either the population needs to be moved out of the basin to the water, or the water needs to be moved to it.

The question then is: 'What method of obtaining water is cost effective for a poor population, sustainable, easily managed, and does not require signifi­cant regular economic inputs from the Capital?' The answer is 'a pipe'. The reticence to such an approach is well summed by Gleick [14] (2003): 'substan­tial, often unanticipated social, economic, and environmental costs. Tens of millions of people have been displaced from their homes by water projects over the past century'.

Gleick's concerns are well founded. As he discusses (Gleick 2002), much of the world went through decades of trying to build its way out of its water difficulties with tremendous negative impacts on sustainability of the resource as well as the environment. Now water resources need management. To ac­complish that efficiently, central bureaus need to give way to multi-level insti­tutions. As demand is rising we must consider markets and pricing, efficiency and equity, to ensure sustainable water use. But this does not describe Ambo­vombe. Ambovombe has not been overbuilt. The understandable fears of the over-consuming west are being transferred to the south, creating an unreason­able burden on poor local populations with no physical water resource access.

A pipe requires management. Following Agrawal's (1999) aforementioned finding with the Raika population in India, there are diverse Antandroy insti­tutions and strategies that must be taken into account. Yet, the insightfulness of Agrawal's findings are that while heavy-handed state-led social engineer­ing (in this case sedentarisation) produce deleterious effects, market forces (often embedded), social networks, geography, political boundaries, and other macro-level constraints on local decision making also need to be considered. Local organisations are well placed to take the management mantle on some things and not others, often determined by matters of scale. In the water sector Rosegrant et al. (2002) represent a consistent view fast coming to dominate the literate: community associations can be valuable but only below the turn­out. Water pricing must account for larger geographic areas even as this effec­tively leads to price subsidies for rural areas from urban areas. Water is most cost effective by scale; even standpipes are not cost effective if they do not ensure regularised supply. As they note:

'Specification of the base rights, base charges, and efficiency prices in such a scheme will be politically difficult, but appears feasible. The es­tablishment of base water rights would increase the political feasibility of water pricing by formalizing existing water rights, rather than being seen as an expropriation of these rights. With efficiency prices paid on only the marginal demand above or below the base right, non-punitive incentives are introduced. The reliance on water user associations to manage water "below the turnout" improves local accountability, transparency, and the flexibility of water allocation. Information costs would be reduced because local irrigators with expert knowledge of the value of water would bear the costs and generate the necessary infor­mation on the value and opportunity costs of water below major turn­outs'.

It also requires a significant initial investment but the maintenance and deliv­ery costs are low. This is not revolutionary thinking. There is in fact a 141 km pipeline in southern Madagascar, funded by the Government of Japan in 1999, that starts at the Menarandra River (near Ampotaka) goes west to the mining town of Tsihombe (district neighbouring Ambovombe to the east) and then turns south to the coast, ending at the resort village of Cap Saint Marie. Evi­dently, where there is an exogenous reason for investment the money is found. There are significant problems with AES's administration of the pipeline that diminishes its value. However, that is not a natural resource supply question; it is an accountability problem and an indication that either the AES needs to clean up its act or stop administering it. It still helps meet the need of the peo­ple in Ampotaka, Tsiombe, Cap Est and in between. Water delivered at station costs 10 MGA per bucket (about 0.6 MGA per litre).

There is discussion in the regional AES office, the regional Ministry of En­ergy and Mines office, and of course within the population that this is exactly what Ambovombe needs. Indeed, at least one study has already been con­ducted (concluding that it would actually be cheaper to bring water from the Manamholo River than the Mandrare even though it is three times further be­cause its high altitude would allow for gravity pumping alone) and the AES has its own estimates. Local officials have their own economic prognostica­tions. In 2004, JICA embarked on a multi-year study to assess the feasibility of a pipeline to Ambovombe. At the time of writing, that report was still forthcoming. JICA's leading consideration appears to be extending the exist­ing pipeline from Tsihombe, projecting 2015 as the year for beginning con­struction. The World Bank's 'Integrated Growth Pole' for the south, based in neighbouring Anosy, is predominantly concerned with development through mining. Plans for road rehabilitation in Androy appear favourable but, despite the Bank's own assessment that water is a great regional concern requiring supply development, this initiative only refers to the JICA study and not to any dedicated sectoral effort.

Why a pipeline has not been built, and appears unlikely to be built in the near term, remains unclear (despite significant efforts by the author to get to the root of the matter). Local leaders, and the AES, argue that there is not enough money. This appears unlikely given both the World Bank's efforts in the region and the current unspent funds for water in the south governed by the African Development Bank. Other responses by policy makers imply an ideological rationale: transregional pipelines are inconsistent with decentrali­sation efforts. Others imply a calculated power relationship. In 1901, then French Governor Joseph Gallieni asked General Hubert Lyautey to tour Madagascar to assess the needs and weaknesses of the population. Of Ambo­vombe he wrote: 'One walks in a full forest, but a forest of underwater dreams, a forest of trees without leaves, of large euphorbia which are named fanitholosse, made out of large sinister stumps…[sic] there is no water, it rains almost never' (Lyautey 1935: 133). He went on to suggest that he who controls the water controls the population. The most likely reason intimated, is that Ambovombe is just not important enough to the central government to warrant significant efforts. Unlike Tsihombe to the east and Amboasary to the west it is not a resource-rich district. There is significant racism against An­tandroy rooted in pre-colonial Merina expansionism and Merina integration into French administrative efforts and this likely further marginalises Ambo­vombe. That Madagascar has a populist president who won every province but Tulear (where Ambovombe is situated), and Ambovombe did not support the president is not a good harbinger for the region being brought into the admin­istrative fold.

   What Should the State-Local Relationship Look Like? Top

The policy direction is clearly towards management at the expense of engi­neering, including supply considerations such as a pipeline. Clearly the AES itself is testimony to why returning to the days of yore when centralised facili­ties had absolute control over funding and direction is not an option. Yet the reasons for the AES direction have less to do with the improved water deliv­ery and environmental ethic than limiting the state's responsibility to pay for and deliver water. The state, and specifically its department in charge of water in Ambovombe, is disengaging. That might be overstating the point as it was never engaged in the first place. In contrast to land and forest policy where the state has taken control of large tracts and increased regulation dramatically over the past decade, [15] even while decentralising certain governance features to the commune level, the state is using the trend towards decentralised water man­agement as an opportunity to reduce its responsibilities in the water sector.

A middle path first requires 'more comprehensive water policies addressing the broader array of human perceptions, meanings and values' (Burmil et al. 1999). It then requires the management and human capital and the state en­gagement and investment in delivery resources. This will serve us well in dis­cussing what the state and the community might do if both want to ensure the supply of water in Ambovombe. First, it is necessary to ascertain community desires before giving the community responsibilities (Chambers and Conway 1992). We need to know more about the dynamics of each community-and how communities are delineated-before assigning community responsibili­ties. We need to know who (leadership type) the community will listen to spe­cifically about water management, what cost recovery schemes are considered acceptable, the type of labour inputs and trade-offs a community is willing to undertake, and what the enforcement mechanisms and penalties for non­compliance are. Second, as Gleick (2002) articulates, water is different from other resources and we need to consider it as such. It is an immediate need. If a household cannot pay then it is not morally acceptable to cut it off entirely (that is why Gleick advocates block pricing for water with a large first level block). Whether a household should be cut off due to non-compliance with the rules for reasons other than ability to pay need to be addressed.

The conventional wisdom appears to be that management groups can be formed at the basin level. However, in this case the water source is outside the basin and the policy trend is to create water users groups at the commune level. Unfortunately, it may not consistently be the commune that serves as the best community-level institution. In some cases smaller units, such as the fokontany or hamlet, may be more appropriate. In other cases kin identities may lead to organisation that makes more sense. This is possible because the role of the community is not to manage water but to ensure accountability of the institution that is managing the resource. It is a political function. The ad­ministrative function at the community level may be in local distribution and fee collection, but the responsibility has to stop there, as one commune cannot force another to deliver water effectively.

Challenges to the decentralisation mechanism have proliferated. For in­stance, an independent review of community-based natural resource manage­ment, funded by the U.S. Agency for International Development, looked at 342 transfers of management and responsibility from the state to the resource user groups at the commune level, seventy-six of them in Tulear province (USAID 2004).

The study found that there were both positive and negative outcomes of the decentralisation process. For instance, some participants benefited economi­cally, but it created economic disparities with only a fraction of society bene­fiting from decentralisation. There was greater recognition of resource values, but a lack of willing social organisers. There was an improvement in the inte­gration of local rules (dina) and national laws, [16] but social conflicts often re­sulted over new rules and institutional functions in accessing resources. And, there was improved government transparency but an inability to reinforce rules (especially across communes).

As water decentralisation is following in the footsteps of forest, marine and land decentralisation in Madagascar, a lot can be learned about what we may expect from the emerging water user group model. The findings in this study support a growing consensus that institutions are important; local involvement in resource management is not a panacea for resource management challenges (Campbell et al. 2001; Ostrom 2001; Poteete and Ostrom 2004). What does the review say should be done? First, increase the coherence of information in the process. That is, ensure that the state and international donors increase, not decrease, their local presence in terms of communication. Second, there should be enhanced legislation governing local finance handling as well as an increased presence of extra-commune mediation. Third, there is a governance void above the commune level that must be rectified. This requires the estab­lishment of critical inter-institutional relationships across levels. Fourth, re­source valorisation needs to be integrated into the politics of development to ensure that there is broader benefit and conservation restrictions are clearer. Fifth, greater recognition of the role of the fokontany and inter-fokontany rela­tionships is necessary. There needs to be a mediating force that collaborates with the communes to ensure inter-fokontany relations over resources. Finally, there needs to be a system of evaluation of the transfer process to determine if each transfer under decentralisation is succeeding or failing and why.

The report, while not conceptually complete, provides a valuable starting place for disaggregating state-local relations. There are functions that the state is better at and functions the community is better at. What these are is still under investigation. However, some preliminary observations can be made. In the case of water, the province has little authority and less capacity. The state is clearly charged, in the Water Code, with the ultimate responsibility of meeting water needs. Thus, the state, with international assistance, is the only existing mechanism for financing regional water supply enhancement and management. In some cases the national electricity and water company, JIRAMA (Jiro sy Rano), can provide that function instead of the AES, but it is still a state-level function. The state, with its partners, needs to focus on modest infrastructure enhancement.

Community-level water users groups are necessary. However, their univer­sal design needs to be revisited. In some cases 'community' may be better served at the fokontany or some other level. The power of the water user's group needs to be able to vary by place in its relationship to the commune government. In some cases it may make more sense for the commune leader­ship itself to undertake local governance functions. Most importantly, the primary goal of local water user's groups needs to be a political, not opera­tional, one. Operational needs are inter-district. The local population knows when there is corruption or other problems of resource governance. They need to have the authority to do something about it. To handle the responsibility of ensuring propriety of officials, as they have been handed, they need to be granted the power to censure officials who have done something wrong. At present they neither elect regional state water officers nor hold any judicial sway over them.

   Conclusion Top

In the beginning of this paper I asked: How well does the state function at the community level? What is the relationship between local government and civil society? What is the role of equity of stakeholders in the decision­making process? What is the relationship between sustainability and equity in accessing the commons and how is this affected by institutional structures? The answers to these questions are somewhat provisional as the research is ongoing. Yet, it is clear that the state functions poorly at the community level. New reforms may ultimately help structure relationships across levels but for now they appear to obfuscate more than illuminate paths to successful deconcentration. The responsible agency, the AES, is under-funded, over­committed, and desperately in need of agency reform. Local government and civil society relationships are complex and marked by high variation. In some communes there is little demarcation. Civil society leaders hold office. In oth­ers, the mayor, despite being an elected official, commands little respect. In virtually all cases the relationship between community-based natural resource management institutions and the commune are fractious.

Community-based resource management institutions have benefited some people at the local level, but they have served to divide the community at the same time. In this sense stakeholder equity has worsened, not improved. Fi­nally, it is clear that there are net winners and losers as individuals at the local level gain power through decentralisation. The institutions across scales that are seeking sustainable water resource use have undermined equity in access­ing the commons even while undermining existing social norms, classes and kinship relations.

The answer is not a return to centralised management. In the case of water there was little centralised management to begin with so it would imply the creation of new centralised management. Antananarivo has shown little con­cern over Ambovombe. Why should it gain power to rectify Tandroy woes when it has not shown any interest in doing so in the past? The answer lies in better understanding the state-local nexus. Institutionally, the decentralization process requires a more clearly articulated relationship between decentralised and deconcentrated levels of governance. In the current schema responsibility is decentralised while funding is kept centralised or deconcentrated. This needs to be rationalised. The decentralisation process would be more success­ful if the state sought to determine what it can do best (presumably with inter­national assistance), and what the community can do best. This is a capacity question as opposed to a question of power. In so doing it must also consider what a community is and remain open to the relative nature of communities by district, commune, fokontany, and kin group as determined by organic growth and historicity rather than decree from the top-down. Where under­stood as both instruments of organisation and institutional arrangements communities can be well placed to manage water resources within their boundaries. Yet, they have limited capacity to address the need to augment supply. Most importantly, Ambovombe's communities need to be viewed as the complex mosaic of relationships that both enhance and detract from the power of the state in a dynamic fashion. Only once the state accepts the diver­sity and power inherent in the community and engage it, can it hope to see a completed decentralisation process with net local gains.


This paper is part of an ongoing research project on water governance and decen­tralization in Madagascar and Kenya. In 2002, I returned to the area for the first time since my (1997-1998) dissertation fieldwork. The goal was to home in on the key concerns of community members apparent in my earlier work. In 2002, 2004 and 2005 I conducted interviews with local leaders and focus groups in all but two of the district communes, and in 2005 I conducted a district-wide survey (n = 521) stratified by commune and fokontany. In 2004 and 2006 I interviewed water sector leaders in government, non-government organizations and donor or­ganizations. Research funds were provided by Yale University and The Univer­sity of Alabama in Huntsville faculty development grants. These were sufficient to also fund the fieldwork for a doctoral student in Anthropology from the Uni­versity of Tulear, Emeline Anjarantena, whose 6 months of fieldwork in 2005 on changing patterns of household decision making in Ambovombe contributed to my thinking on this subject.[67]

   References Top

1.Adams, William M. 2003. Decolonizing Relationships with Nature. In: Decolonizing Nature: Strategies for Conservation in a Post-colonial Era. (eds. William Adams and Martin Mulli­gan), Earthscan, London.  Back to cited text no. 1    
2.Agrawal, A. 1999. Greener Pastures: Politics, Markets, and Community Among a Migrant Pas­toral People. Duke University Press, Durham.  Back to cited text no. 2    
3.Agrawal, A. and Ribot J. 1999. Accountability in Decentralization: A framework with South Asian and West African cases. Journal of Developing Areas 33:473-502.  Back to cited text no. 3    
4.Barlow, M. and T. Clark. 2002. Blue Gold. The New Press, New York.  Back to cited text no. 4    
5.Berkes, F. 1989. Common Property Resources: Ecology and Community-Based Sustainable De­velopment. Belhaven Press, London.  Back to cited text no. 5    
6.Bhatia, R. and M. Falkenmark. 1993. Water Resource Policies and the Urban Poor: Innovative Approaches and Policy Imperatives,' Water and Sanitation Current, UNDP-World Bank Water and Sanitation Program, Washington, DC, USA.  Back to cited text no. 6    
7.Blair, H. 2000. Participation and accountability at the periphery: Democratic local governance in six countries. World Development 28(1): 21-39.  Back to cited text no. 7    
8.Burmil, S., T.C. Daniel and J.D. Hetherington. 1999. Human values and perceptions of water in arid landscapes. Landscape and Urban Planning 44: 99-109.  Back to cited text no. 8    
9.Campbell, B.M., J.A. Sayer, P. Frost, S. Vermeulen, M. Ruiz Perez, A. Cunningham, and R. Prabhu. 2001. Assessing the performance of natural resource systems. Conservation Ecol­ogy 15 (2):22.  Back to cited text no. 9    
10.Chambers, R. 1997. Whose Reality Counts?: Putting the First Last. Intermediate Technology: London.  Back to cited text no. 10    
11.Chambers, R. and G.R. Conway. 1992. Sustainable Rural Livelihoods: Practical Concepts for the 21st Century. Discussion Paper 286. Brighton: University of Sussex: Institute of Develop­ment Studies. UK.  Back to cited text no. 11    
12.Collignon, B. and M. Vezina. 2000. Independent Water and Sanitation Providers in African Cities. Full Report of a Ten-Country Study. Water and Sanitation Program, Washington, DC, USA.   Back to cited text no. 12    
13.Crook, R.R. and J. Manor. 1998. Democracy and Decentralisation in South Asia and West Africa. Cambridge University Press, Cambridge.  Back to cited text no. 13    
14.de Wit, M. and J. Stankiewicz. 2006. Changes in Surface Water Supply Across Africa with Pre­dicted Climate Change. Science 311: 1917-1921.  Back to cited text no. 14    
15.Dietz, T., E. Ostrom and P.C. Stern. 2003. The struggle to govern the commons. Science 302(5652): 1907(6).  Back to cited text no. 15    
16.Dinar, A. and A. Subramanian. 1997. Water Pricing Experiences: The International Perspective. World Bank Technical Paper No. 386.  Back to cited text no. 16    
17.Droy, I. 1991. La rehabilitation des petits perimetres irrigues a Madagascar: une reponse aux impor­tations de riz? Cahiers des Sciences Humaines 27(1-2): 117-127.  Back to cited text no. 17    
18.Falkenmark, M. 1989. The massive water scarcity now threatening Africa-Why isn't it being ad­dressed? Ambio 18: 112-118.  Back to cited text no. 18    
19.Foucault, Michel. 1983. The Foucault Reader. (ed. Paul Rabinow). Pantheon, New York.  Back to cited text no. 19    
20.Galvan, Dennis and Richard R. Marcus. 2003. Sustainable Development in Sub-Saharan Africa: Re­source Management through Institutional Syncretism in Madagascar and Senegal. In: Achieving Sustainable Development: The Challenge of Governance Across Social Scales (eds. Hans Th. A. Bressers and Walter A. Rosenbaum). Praeger, Westport CT and London.  Back to cited text no. 20    
21.Gezon, Lisa L. 1997. Institutional Structure and the Effectiveness of Integrated Conservation and Development Projects: Case study from Madagascar. Human Organization 56(4): 462-470.  Back to cited text no. 21    
22.Gleick, Peter H. (Ed) 1993. Water in Crisis: A Guide to the World's Fresh Water Resources. Ox­ ford University Press, New York.   Back to cited text no. 22    
23.Gleick, Peter H. 2002. The World's Water 2003-2003: The Biennial Report on Freshwater Re­sources. Island Press, Washington, DC, USA.  Back to cited text no. 23    
24.Gleick, Peter H. 2003. Global Freshwater Resources: Soft-Path Solutions for the 21st Century, Science, 302: p. 1524-1528.  Back to cited text no. 24    
25.Henkels, D.M. 1999. Une vue de pres du droit l'environnement Malgache. African Studies Quar­terly 3.2: 30-46.  Back to cited text no. 25    
26.Heurtebize, G. 1986. Histoire des Afomarolahy (Extreme-Sud de Madagascar). Centre National de la Recherche Scientifique, Paris.  Back to cited text no. 26    
27.Hyden, G. 1980. Beyond Ujamaa in Tanzania: Underdevelopment and an Uncaptured Peas­antry. University of California Press, Berkeley.  Back to cited text no. 27    
28.Kaufmann, J.C. 2001. La Question des Raketa: Colonial struggles with prickly pear cactus in southern Madagascar, 1900-1923. Ethnohistory 48(1-2): 87-121.  Back to cited text no. 28    
29.Kemper, K., A. Dinar, W. Blomquist, G. Alaerts, M. Diez and A. Bhat. 2003. Integrated River Basin Management and the Principle of Managing Water Resources at the Lowest Appropri­ate Level-When and Why It Does (Not) Work in Practice? World Bank Research Project. http://lnweb18.worldbank.org/ESSD/ardext.nsf/18ByDocName/SectorsandThemesRiverBasinManagementIntegratedRiverBasinManagementProject.  Back to cited text no. 29    
30.Korten, D. 1986. Community Management: Asian Experience and Perspectives. Kumarian Press, West Hartford, CT.  Back to cited text no. 30    
31.Kull, C. 2004. Isle of fire: The political ecology of landscape burning in Madagascar. University of Chicago Press, Chicago.  Back to cited text no. 31    
32.Lyautey, H. 1935. Lettres du Sud de Madagascar: 1900-1902. Librarie Armand Colin, Paris.  Back to cited text no. 32    
33.Manor, J. 1999. The Political Economy of Democratic Decentralisation. Directions in Development Series. World Bank, Washington, DC, USA.  Back to cited text no. 33    
34.Marcus, R.R. 2001. Seeing the forest for the trees: Integrated conservation and development projects and local perceptions of conservation in Madagascar. Human Ecology 29(4): 381-397.  Back to cited text no. 34    
35.Marcus, R.R. 2000. Cultivating Democracy on Fragile Grounds: Environmental Institutions and Non­Elite Perceptions of Democracy in Madagascar and Uganda. Ph.D. Dissertation. Department of Political Science, University of Florida.  Back to cited text no. 35    
36.Maro, P.S. 1990. The impact of decentralization on spatial equity and rural development in Tan­zania. World Development 18(5): 673-693.  Back to cited text no. 36    
37.Mearns, R. 2004. Decentralisation, rural livelihoods and pasture-land management in post­ socialist Mongolia. European Journal of Development Research 16(1): 133-152.  Back to cited text no. 37    
38.Middleton, K. 2002 Opportunities and Risks: A Cactus Pear in Madagascar. ISHS Acta Horticul­turae 581: IV International Congress on Cactus Pear and Cochineal.  Back to cited text no. 38    
39.Mody, J. 2001. Management of river basin systems through decentralization. World Bank Litera­ture Review. Review/$FILE/RiverBasinProjectLiteratureReview.pdf.  Back to cited text no. 39    
40.Olmstead, S. 2003. What's price got to do with it? Environment 45(10): 22-35.  Back to cited text no. 40    
41.Organization for Economic Cooperation and Development (OECD). 2003. Improving Water Management: Recent OECD Experience. OECD, Paris.  Back to cited text no. 41    
42.Organization for Economic Cooperation and Development (OECD). 1998. Water Management: Performances and Challenges in OECD Countries. OECD, Paris.  Back to cited text no. 42    
43.Ostrom, E. 2001. Decentralization and Development: The New Panacea. In: Challenges to De­mocracy: Ideas, Involvement and Institutions (eds. K. Dowding, J. Hughes and H. Mar­getts), pp. 237-256. Palgrave Publishers, New York.  Back to cited text no. 43    
44.Ostrom, E. 1999. Coping with the tragedies of the commons. Annual Review of Political Science 2: 493-535.  Back to cited text no. 44    
45.Ostrom, E., T. Dietz, N. Dolsak, P.C. Stern, S. Stonich and E.U. Weber. 2002. The Drama of the Commons. National Research Council, Washington, DC, USA.  Back to cited text no. 45    
46.Poteete, A.R. and E. Ostrom. 2004. In pursuit of comparable concepts and data about collective action. Agricultural Systems 82: 215-232.  Back to cited text no. 46    
47.Poffenberger, M. 1990. Keepers of the Forest: Land Management Alternatives in Southeast Asia. Kumarian Press, West Hartford, CT.  Back to cited text no. 47    
48.Rakotoarisoa, J.-A. 1998. Mille Ans d'Occupation Humaine Dans le Sud-est de Madagascar: Anosy, une ile au milieu des terres. l'Harmattan, Paris.  Back to cited text no. 48    
49.Razanabahiny, V.L. 1995. (Convention entre members de communautes villegoises) Son oppor­tunite ou non dans la conservation de la nature. Cas de la Reserve Naturelle Integrale d'Andohahela- Tolagnaro. Antananarivo: Universite d'Antananarivo. Ecoloe Normale Superieure, Filiere Lettres Malgaches (Anthropologie). Memoire de C.A.P.E.N.  Back to cited text no. 49    
50.Ribot, J.C. 2002. Democratic Decentralization of Natural Resources: Institutionalizing Popular Participation. World Resources Institute, Washington, DC, USA.  Back to cited text no. 50    
51.Ribot, J.C., A. Agrawal and A.M. Larson. 2006. Recentralizing while decentralizing: How na­tional governments reappropriate forest resources. World Development 34(11): 1864-1886.  Back to cited text no. 51    
52.Rondinelli, D., J.S. McCullough and R.W. Johnson. 1989. Analyzing decentralization policies in developing countries: A political economy framework. Development and Change 20(1): 57-87.  Back to cited text no. 52    
53.Rosegrant, M.W., X. Cai and S. Cline. 2002. World Water and Food to 2025: Dealing with Scar­city. International Food Policy Research Institute and Colombo, Washington, DC, USA; In­ternational Water Management Institute, Sri Lanka.  Back to cited text no. 53    
54.Saleth, R.M. 1999. Evaluating Water Institutions and Water Sector Performance. World Bank, Washington, DC, USA.  Back to cited text no. 54    
55.Scott, J.C. 1999. Seeing Like a State: How Certain Schemes to Improve the Human Condition Have Failed. Yale University Press, New Haven.  Back to cited text no. 55    
56.World Meteorological Organisation (UN WMO). 1992. International Conference on Water and the Environment, Dublin/Ireland: The Dublin Statement on Water and Sustainable Devel­opment. United Nations WMO, Geneva.  Back to cited text no. 56    
57.World Bank. 2000. The World Development Report 2000. Oxford University Press, New York.   Back to cited text no. 57    
58.World Bank. 2003. World Development Report 2003: Sustainable Development in a Dynamic World. World Bank, Washington D.C.  Back to cited text no. 58    
59.World Bank. 2004. Decentralization in Madagascar. World Bank, Washington, DC, USA.  Back to cited text no. 59    
60.UN Department of Economic and Social Affairs, Division for Sustainable Development (UNDESA). 2002. Earth Summit: Agenda 21. agenda21/index.htm  Back to cited text no. 60    
61.UNEP. 2002a. Africa Environmental Outlook: Past, Present, and Future Perspectives. UNEP, New York.  Back to cited text no. 61    
62.UNEP. 2002b. Global Environment Outlook. UNEP, New York.  Back to cited text no. 62    
63.United Nations Development Programme. 2006. Human Development Report. United Nations Development Programme, New York.  Back to cited text no. 63    
64.U.S. Agency for International Development/RESOLVE. 2004. Evaluation et perspectives des transferts de gestion des ressources naturelles a Madagascar. Rapport Final Phase 1.' Un­published report, July 2004.  Back to cited text no. 64    
65.Vallet-Coulomb, C., F. Gasse, L. Robison, L. Ferry, E. Van Campo and F. Chalie. 2006. Hydro­logical modeling of tropical closed lake Ihotry (SW Madagascar): Sensitivity analysis and implications for paleohydrological reconstructions over the past 4000 years. Journal of Hy­drology 331(1-2): 257-271.  Back to cited text no. 65    
66.WaterAid. 2006. WaterAid's plans in Madagascar: 2005-2010. URL: .uk/uk/what we do/where we work/madagascar/3256.asp.  Back to cited text no. 66    
67.Western, D. and R.M. Wright. 1994. Natural Connections: Perspectives in Community-Based Conservation. Island Press, Washington, DC, USA.  Back to cited text no. 67    


  [Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5], [Figure 6]


Previous article Next article
    Similar in PUBMED
   Search Pubmed for
   Search in Google Scholar for
 Related articles
    Access Statistics
    Email Alert *
    Add to My List *
* Registration required (free)  

    International Wa...
    The Case of Ambo...
    What Should the ...
    Article Figures

 Article Access Statistics
    PDF Downloaded2207    
    Comments [Add]    

Recommend this journal