Contesting Institutional Engineering for Decentralized Natural Resource Governance in Malawi

In line with 1993/1994 political change from autocratic to democratic rule in Malawi, centralized natural resources management has been replaced by decentralized approaches. Decentralized natural resource governance, however, requires transfer of responsibility and control over resources to locally elected actors. Using Domasi and Njala irrigation schemes, this study explores the processes and outcomes of institutional engineering that was considered prerequisite for the establishment of local governance in Malawi. The study findings reveal that decentralization is predominantly a political activity; hence, the transition to decentralization cannot be a matter of just passing democratic legislation. Consequently, mere institution of democratic structures does not automatically lead to achievement of democratic governance as was theorized.


Introduction
This article undertakes a critical review of Malawi's experience with decentralized natural resources management as a new orthodoxy strategy to achieving sustainable governance of natural resources and institution of local democratic governance (Adhikari, 2001;Ribot & Larson, 2005). It is motivated by the fact that since 1980s, devolution of responsibility and control over natural resources from central government to user groups has become a widespread policy in many developing countries, including Malawi (Meinzen-Dick, Raju, & Gulati, 2000;Shah, van Koppen, Merrey, de Lange, & Samad, 2002;Zuka, 2013). According to Ribot and Larson (2005), at least 60 countries now claim to be decentralizing natural resource governance for economic, social, and political reasons. Founded on the rationality of citizen participation and democracy, devolution policies have not only challenged the morality of centralized systems of resource governance as theorized by Hardin (1968) but also questioned its very efficiency, effectiveness, and sustainability. Democracy literature, in particular, presents decentralization as a tool to achieving the trinity goals of good governance, development, and poverty reduction. The drive for decentralization is also strengthened by disappointing record of centralized development planning and justified on the basis that political governance cannot be separated from economic governance (Bradhan, 2002;Lee, 2002). In general, devolution policies contest the state's incentives and capability to manage local natural resources efficiently, sustainably, and equitably (Agrawal, 2001;Meinzen-Dick et al., 2000). Inter alia, empirical evidence revealed that state management of natural resources has been characterized by limited responsiveness to the needs of local communities, increased social and economic inequality among local communities, destruction of indigenous natural resource management knowledge, and lack of downward accountability in resource use (Lee, 2002). Thus, participatory school of thought particularly advocates reducing the size of government by decentralizing management of common pool resources to local communities (Meinzen-Dick et al., 2000;Restrepo, Vermillion, & Munoz, 2007).
Current policy shifts toward decentralized democratic natural resource governance in developing countries are, however, implemented against a long legacy of autocratic rule from both state and traditional actors (Mamdani, 1996). According to Mamdani (1996), since independence, management of natural resources was grabbed by the state and traditional authorities in what he calls "bifurcated" approach (see also Blaikie, 2006). Thus, institutional engineering to emancipate local communities from the autocratic tendencies of the state as well as undemocratic traditional actors has been the primary task in the process of establishing community-based natural resources management (CBNRM; Blaikie, 2006;Meinzen-Dick et al., 2000;Ntsebeza, 2005). However, while theoretical claims predict a successful decentralized natural resources management, empirical evidences from both local and international CBNRM projects have highlighted mixed outcomes. While some are more optimistic (Agrawal, 2001;Katz, 2002;Ostrom, 2000), others strongly argue that its outcomes remain disappointing (Blaikie, 2006). To what extent then do the theoretical claims about the success of decentralized natural resources hold in practice? In others words, has institutional engineering failed to institute democratic natural resource governance?
Employing the lens of state-society relations, especially formal-informal institutional interaction model, this study responds to these questions by examining the challenges associated with the process of crafting institutional framework and governance structures for decentralized natural resource governance. Using Domasi and Njala Irrigation schemes as case studies, the study's crucial questions to be addressed are as follows: What were the guiding principles in crafting rules to guide resource use in the two irrigation schemes? To what extent has the current attempts at democratic natural resource governance managed to emancipate local communities from the control of the state and traditional actors? and What institutional forces explain the condition of decentralized resource governance at the two sites? In general, therefore, the study intends to explore the adequacy of the formulated democratic institutions in achieving the objectives of good natural resource governance as measured by the following variables: • • Access and equity: fair distribution and allocation of resources-irrigation plots and water resources. • • Transparency and accountability: the imperative to make elected leaders answerable for their public behavior as well as responsive to the entity from which they derive their authority. It includes the distribution of power and status among local farmers as evidenced by participation in decision making and in elections, observance of resource users' rights, and respect for resource users' democratic views. • • Predictability and rule of law: the existence of laws, regulations, and policies to regulate society and their fair and consistent application to resource use. • • Conflict resolution: air handling and resolving of conflicts and disputes over resources among resource users and other surrounding local and national entities and actors.
There is a growing literature on decentralized natural resource governance internationally and locally (see Blaikie, 2006;Ferguson & Mulwafu, 2004;Zuka, 2013). However, most of these studies have concentrated on the historical evolution and outcomes of CBNRM and not the theoretical analysis guiding the evolution. In particular, the theoretical dilemma and outcomes of the interaction between, first, state and society, and formal and informal institutions after policy shift have not been adequately explored. This gap makes this study unique.

Conceptualizing Decentralization
Decentralization is a complex process with numerous and at times confusing definitions (Adamolekun, 1991;Smoke, 2003). Broadly defined, decentralization is the transfer of political, administrative, and legal authority, or responsibility for planning, decision making, resource allocation, and administering public functions from the central government to its field, district, and regional administrative units; semiautonomous and parastatal organizations or even nongovernmental organizations (NGOs) or voluntary associations of people; and private entities (see Adamolekun, 1991;Smoke, 2003). Generally, there are four forms of decentralization classified based on their objectives, namely, political, administrative, spatial, and market (Cohen & Peterson, 1996;Smoke, 2003). Political decentralization, which is largely the focus of this article, is concerned with transfer of decisionmaking power to lower level of government units or to citizens or their elected representatives. A good example of political decentralization is devolution, which refers to transfer of governance responsibility for specified functions to subnational levels outside the direct control of the central government (Adamolekun, 1991). Administrative decentralization, divided into deconcentration and delegation, is the hierarchical and functional distribution of powers and functions between central government and noncentral government. Deconcentration is the transfer of administrative responsibility for specified functions to lower levels within the central government bureaucracy (Tambulasi & Kayuni, 2007). In other words, it is the redistribution of government administrative responsibility within central government ministries or agency and involves sharing central government with offices outside the national capital. In a deconcentration, local staffs do not possess any authority to make decision on their own, but are under the technical supervision and control of the central government. Delegation is the transfer of government decision making and responsibilities to organizations and institutions, which enjoy varying degrees of autonomy vis-à-vis central government ministries such as parastatal or public corporations (Adamolekun, 1991;Cohen & Peterson, 1996). Spatial decentralization entails efforts at reducing excessive urban concentration in a few large cities by promoting regional growth poles that have potential to become centers of manufacturing and agricultural marketing (Cohen & Peterson, 1996). Market decentralization focuses on shifting the responsibility of providing public goods from the state to private firms, community groups, cooperatives, and NGOs. It is broadly perceived as the creation of conditions that allow goods and services to be produced and provided by market mechanisms sensitive to the preferences of individuals. This form of decentralization has become prevalent with global establishment of liberalization and privatization policies (Adamolekun, 1991).

Merits and Demerits of Decentralization: Theoretical and Empirical Perspectives
Decentralization is, inter alia, justified based on its perceived administrative, political, and economic benefits. Administrative objectives are mainly concerned with the best approach to organizing development activities to achieve administrative convenience and efficiency. The former centers on making appropriate decision, whereas the latter is about best resource allocation. The argument is that local governments are better placed to achieve administrative convenience and efficiency than central government because they are closer to local communities; thus, they can be more sensitive in the identification of development priority needs. From an economic perspective, local governments are regarded to be in a better position to raise local revenue to implement local projects as local communities will be motivated to fund and monitor their own development cause (Adamolekun, 1991;Smoke, 2003). The major political objective of decentralization is the achievement of good and democratic governance. Good governance signifies creation of proper law and order, elimination of unnecessary government interference, and establishment of corrupt free public administration as a means of achieving socioeconomic progress (Grindle, 2004;Sharma, 2007). Democratic governance, on the contrary, refers to the recognition of political, civil rights, and citizen participation as basic values and development end in themselves (Abdellatif, 2003;Elahi & Danapoulos, 2004;Waheduzzaman, 2010). Suffice to mention that, depending on their aims, different actors vary in their orientations to these two concepts, preferring to use either good governance or democratic governance.
The economic objective of decentralization centers around three things, namely, economic efficiency, resource mobilization, and service and product quality. The understanding is that economic efficiency is increased as citizens become committed to development plans they have been involved in identification. Resource mobilization, on the contrary, is improved as people become motivated and willing to contribute to local development programs. Finally, quality of service and product increases as a result of competition and choice among local government units. The assumption is that citizens would choose where to locate themselves to take advantage of the quality and cost differentials (Adamolekun, 1991;Smoke, 2003). It should, however, be pointed out that the administrative, political, and economic objectives of decentralization can hardly be fully achieved in isolation. It is also strongly argued that implementing devolution on its own strengthens top-down political dominance. This recognition, therefore, calls into question the kind of decentralization that can be implemented to achieve decentralized natural resource governance.
Theoretical arguments for decentralization notwithstanding, increased literature has shown that decentralization is not a panacea to all challenges of economic development. First, decentralization has inherent theoretical flaws in accounting for its benefits, especially when applied to developing countries. Decentralization also has the potential of promoting a sense of locality or regional belonging resulting in heightened tribal and neopatrimonial tendencies (Adamolekun, 1991;Smoke, 2003). In Malawi, this outcome is faulted on its potential to cement already hostile ethnic and regional groupings prevalent in the country; and continue to provide a platform for tribal neopatrimonial tendencies. Second, conditions within developing countries do not allow full enjoyment of the benefits provided by decentralization. For instance, high poverty levels in most countries limit preference differentiation, making the question of administrative efficiency irrelevant. Equally important, the economic objective of quality service provision is defeated by the fact that costs in developing countries are not related to competing local governments, and citizens do not really have choice of jurisdiction (Adamolekun, 1991;Smoke, 2003). In addition to these two theoretical and conditional blocks, decentralization suffers from design and implementation problems such as budget deficits, fiscal irresponsibility, and local elite capture of the development agenda (Smoke, 2003). The crucial question for this study is the extent to institutional engineering for decentralized resource governance has provided answers to escape from these shortfalls.

Objectives of Decentralized Natural Resource Governance
Decentralized natural resource governance has generally four main objectives, namely, as a means to achieving decentralization and local empowerment, as a means to reducing government expenditure on resource management, as a means to instituting sustainable resource use, and as a means to reducing rampart poverty among local communities (Blaikie, 2006;Lee, 2002). Decentralized natural resource governance as a means to achieving decentralization and local empowerment is mainly supported by international organizations and financial institutions that perceive participation as a sure way to local empowerment. To this end, processes of decentralizing natural resource governance have involved radical institutional reform in line with the principles of democracy and decentralization. Political decentralization involving establishment of institutions of elected representatives is seen as a precondition for achieving local democratic governance (Ribot, 2004). Democracy literature, in particular, presents decentralization as a good tool to achieving the trinity of good governance, development, and poverty reduction. This view is reinforced by the unsettled dichotomous views regarding the role of traditional and formal institutions in promoting local democracy. Some scholars consider traditional institutions as antithetical to the development of democracy (Mamdani, 1996), whereas others argue that traditional institutions can successfully coexist with democratic forms of government. There is also increased consensus among scholars that the main impetus toward decentralization came from the fiscal crisis of the state (Blaikie, 2006). Global economic crisis especially in the early 1980s resulted in governments failing to meet the costs of monitoring resource use and maintaining resource facilities.
Decentralized natural resource governance as a means to instituting sustainable utilization of local resources rose from a growing realization that state management of natural resources has been unsuccessful (Blaikie, 2006;Zuka, 2013). It is generally designed to achieve resource sustainability through social and economic incentives to local users who have for long time been blamed for resource depletion and destruction. Rather than being new, decentralized resource governance is a modern attempt to revive the traditional institutionalized mechanisms for managing and conserving the natural environment (Kellert, Mehta, Ebbin, & Lichtenfeld, 2000). The new approach is uphold by the evidence that local people have the capacity to manage their common pool resources efficiently, sustainably, and equitably (Katz, 2002;Ostrom, 2000). It is enforced by the theoretical shifts in the game theories and democratic governance, though practice is largely driven by the latter (Agrawal, 2001;Blaikie, 2006).
Last, there are widespread claims, especially among development practitioners, that the main objective of decentralized natural resource governance is poverty reduction of the local people. This objective has gained ground with the present understanding that poverty does not only relate to lack of physical assets but also social and political capacity (Lee, 2002). In other words, the approach improves the local people's well-being physically, socially, and politically through improving their access to resources, social interaction, and good local governance, respectively. Thus, the drive for decentralization is justified on the basis that political governance cannot be separated from economic governance (Bradhan, 2002;Ribot, 2004). There is, however, a growing skepticism about the real objectives of decentralized natural resource governance. A number of scholars argue that decentralized natural resource governance is a ploy to pacify local communities into accepting the broader aim of protecting local resources from which they would continue to be excluded (Blaikie, 2006;Ferguson & Mulwafu, 2007;Musumali, Larsen, & Kaltenborn, 2007).

Case Study Background Information
Historically, Malawi inherited and continued with strong state intervention into natural resources management as backed by the colonial legacy that doubted the ability of local communities to efficiently manage their resources (Ferguson & Mulwafu, 2004; see also Fairhead & Leach, 1995). Thus, most natural resources, including forests, wild life, and water resources, were brought under direct state control. In the case of irrigation water management, all formal public irrigation schemes were strictly put under state management. However, a plethora of factors, including financial constraints, rendered state management of public irrigation schemes inefficient and ineffective (Blaikie, 2006;Veldwisch, Bolding, &Wester, 2007;Zuka, 2013). Thus, in line with global political moves toward decentralization and public participation in resource management, the Government of Malawi (GoM) promulgated policy legislation for community management of irrigation of public irrigation schemes. The National Irrigation Policy and Development Strategy (GoM, 2000) and Irrigation Act (GoM, 2001), in particular, reflect this goal by "promoting full ownership of public irrigation schemes by the beneficiaries through their legally constituted local organizations that will oversee all matters related to operation and maintenance of these schemes" (GoM, 2000, p. 7). Likewise, the Irrigation Act (2001) vests legal authority in a group of small-scale farmers to own, use, and maintain the schemes. The formulation of management structure in the irrigation sector was aimed at decentralizing and restructuring water resource governance to cut off dictates of the state and its agencies, especially local chiefs. The course taken was based on the understanding that in Malawi, and many other countries in Africa, traditional authorities have been historically undemocratic, unaccountable, and autocratic agents of the state (see Ferguson & Mulwafu, 2004;Mamdani, 1996;Ntsebeza, 2005). Thus, local chiefs were barred from active involvement in the management of irrigation scheme. Instead, the schemes are managed by individuals elected among the community members. This view was premised on the idea that institutional arrangement that is led by elected leaders would lead to institution of democratic governance.
Thus, the most debatable aspect of institutional engineering for decentralized natural resource governance in Malawi is whether the exclusion of the state and traditional leaders for elected leaders have resulted in the establishment of democratic natural resource governance. In Malawi, local chiefs have been the state's trusted agent and have been actively involved in natural resources management (see Blaikie, 2006;Ferguson & Mulwafu, 2004). During one-party autocratic rule, in particular, local chiefs were actively involved in the management of public irrigation schemes. Thus, this study also addresses the unsettled debate of whether traditional institutions are antithetical to development of democracy (Mamdani, 1996) or support democracy (United Nations Economic Commission for Africa [UNECA], 2007). It should be pointed that the context for decentralized natural resources management in Malawi reveals a gamut of local and external actors, namely, elected leaders, traditional leaders, local communities, and the state. Thus, the new institutional arrangement presents a potential conflict between the state and traditional leaders, on one hand, and new elected Water User Association (WUA) leaders, on the other.

State-Society Relations and Formal-Informal Institution Interaction Model
It is against the backdrop that regime change in natural resources management changes the state-society relations that this study adopts the theory of state-society relations and formal-informal interaction model. State-society relations is defined as interactions between state institutions and societal groups to negotiate how public authority is exercised and how it can be influenced by people. They are focused on issues such as defining the mutual rights and obligations of state and society, negotiating how public resources should be allocated and establishing different modes of representation and accountability. (Department for International Development [DFID], 2010, p. 15) Broadly, there are two main assumptions within state-society relations. The first assumption views state and society as dichotomous and mutually exclusive, whereas the second assumption perceives the relations as an aggregation of the inherently unproblematic spheres (Seller, 2010). According to Seller (2010), state is an "institutionalized collective power superordinate to other organizations that is sovereign vis a vis other states, autonomous or distinct from the rest of society, and identified socioculturally with a national collectivity" (p. 4). Society, on the contrary, is defined as a condition of living in association or an organized community (Mitchneck, 2011). The relevancy of this framework to this study is that institutional engineering restructures arrangements under which these relations operate, thus potentially bringing tension and negotiations between the two actors.
The two major assumptions guiding state-society relations are explored further within the formal-informal institutional interaction model. According to Helmke and Levitsky (2004), formal-informal interaction model explains the nature and outcomes of interaction between formal and informal institutions. Institutions are rules of the game or humanly devised constraints that shape human interaction in society (Knowles, 2006). Formal-informal interaction model focuses on the nature of relationships between formal and informal institutions and their outcomes. In the case of this study, the relationships are between elected WUA leaders, on one hand, and the traditional and religious leaders, on the other hand.
The formal-informal institution interaction model resonates very well with institutional context created in instituting decentralized natural resource governance in the irrigation sector in Malawi. From the perspective of development practitioners, existence of democratic management structure project a positive-sum game with regard to achieving local democratic governance (GoM, 2000(GoM, , 2001. This view is, however, theoretically simplistic. Although institutional theory can help understand the outcomes of decentralized natural resource governance, it is limited in a historical context where neither formal nor informal institutions exists in isolation (Mamdani, 1996). Again, there is a strong belief that many rules of the game that structure and guide people's life are informal institutions that are created, communicated, and enforced outside officially sanctioned channels (Helmke & Levitsky, 2004;Knowles, 2006). Thus, formal-informal interaction model is critical to explaining interaction between the formal and informal institutions and also state-society actors. The application of this model is empirically supported by Meinzen-Dick et al. (2000), who noted that collective action for natural resource management does not necessarily require a formal organization.
Informal institutions are defined as socially shared rules, usually unwritten, that are created, communicated, and enforced outside officially sanctioned channels, whereas informal institutions are rules and procedures that are created, communicated, and enforced through channels widely accepted as official. In the case of natural resource governance, this mainly relates to state recognized constitutions, laws, regulations, and organization rules (Helmke & Levitsky, 2004;Knowles, 2006). Generally, formal-informal interaction model bridges the dichotomous functional-dysfunctional view about informal institutions (see Ntsebeza, 2005). The model proposes four typologies of informal institutions based on convergence and divergence of institutional outcomes and effectiveness of formal institutions, namely, complementary, accommodating, substitutive, and competing. Complementary informal institutions combine effective formal rules and convergent outcomes and fill the formal institutional gap. Accommodating informal institutions combine effective formal institutions and divergent outcomes by creating incentives to alter the outcomes of formal rules. Substitutive informal institutions combine ineffective formal institutions and compatible outcomes. Competing informal institutions combine ineffective formal rules and divergent outcomes, producing competing informal institutions such as clientelism, patrimonial, clan politics, and corruption (Merry, 1988). This model is relevant in the explaining outcomes of natural resource governance in areas experiencing power and responsibility transfer from government to communities. In Malawi, devolution of responsibility from government to local farmers has been associated with establishment of formal rules and regulations that are formally recognized by the state (Ferguson & Mulwafu, 2004). However, informal actors are having a lot of influence at a local level (Zuka, 2013). Therefore, as theorized in the Helmke and Levitsky, the success of the community management of irrigation schemes depends on the complex interaction between formal and informal institutions within the broader state-society relations.

Method
This study was conducted in two smallholder irrigation schemes of Domasi and Njala in Southern Malawi. Domasi scheme is 500 ha while Njala scheme is 45 ha. Both irrigation schemes grow rice during rainy and dry seasons. The two areas were considered appropriate for this study because of two reasons. First, both of them were under state management and have devolved management responsibility to farmers following the promulgation of the 2001 Irrigation Act. Thus, they present good examples of natural resources under elected local actors. Second, their institutional arrangement fit into the analytical design for assessing decentralized natural resource governance as they have both formally formulated local democratic constitutions. There are 1,560 and 240 irrigation farmers at Domasi and Njala irrigation schemes, respectively.
This study adopted both quantitative and qualitative approaches to investigate the processes and outcomes of institutional engineering for decentralized natural resource governance in Malawi. Quantitative data were used to explain quantitative trends such as number of plots owned by farmers. The qualitative approach was employed in collecting data relating to examining the processes of crafting institutional framework for decentralized natural resource governance; and assessing performance challenges associated with this framework. Qualitative approach is also ideal in examining the complex institutional interaction between state and society; and between formal and informal institutions. In this study, formal institutions refer to a set of rules and regulations guiding decision about resource use, allocation, and control as outlined in the WUA constitution. Informal institutions, on the contrary, refer to resource governance rules, regulations, and practices outside the WUA constitution. Most of these informal institutions are deeply rooted in local tradition (see Kambewa, 2005). Thus, WUA leaders were grouped as formal actors, whereas traditional and religious leaders grouped as informal actors.
The primary data were collected through key informant interviews, focus group discussions (FGDs), and household survey. Key informant interviews were held with purposively selected respondents from WUA leaders, local chiefs, faith leaders (Sheikhs and Christian Pastors), and government officials. A total of eight FGDs were conducted with irrigation farmers at both Domasi and Njala irrigation schemes. Key informant interviews and FGDs were specifically designed to assess the interaction between irrigation farmers, traditional leaders, WUA executive members, and government officials. The discussions were also designed to assess institutional performance in terms of regulating access to resource and its efficient use. The study surveyed and interviewed a total of 106 and 50 households at Domasi and Njala, respectively. These households were selected using simple random sampling, and semistructured questionnaires were used to collect data. The household survey was aimed at assessing farmers' mode and equality of access to irrigation plots and water resource. Furthermore, the study employed participant and field observation in which a number of WUA meetings were attended to understand the nature of interaction between WUA leaders, local chiefs, and farmers.

Institutional Framework for Irrigation Scheme Management: Crash of Interest
Both Domasi and Njala Irrigation schemes are managed by WUA comprising different committees, namely, the Executive, Discipline, Health, Water, Finance, Auditing, Marketing, and Environment. According to the constitutions of the two irrigation schemes, WUA leaders are elected at a general meeting every 3 years and that every legitimate member of the association is allowed to stand for any post except for the posts of chairperson, secretary, and treasurer, which are reserved for people who know how to read and write. However, the findings of this study revealed that the process of electing leaders has been marred by a lot of irregularities largely emanating from the unconstitutional agreements between local chiefs and the elected WUA leaders. First, elections are not regularly conducted as the last one was held in 1999 and at a time of this study, farmers doubted the probability of holding elections soon. This status quo is uphold by the support from traditional chiefs and religious leaders who argue that elections are not needed as they do not have problems with the current leadership. Second, the process of scrutinizing candidates for the top WUA positions is less democratic. Instead, unconstitutional relations between WUA executive members and local chiefs have created a parallel committee for vetoing out individuals thought to be a threat to their local relationship. Generally, local chiefs and WUA executive take advantage of the ambiguous constitutional requirements for individuals willing to run for WUA elected positions, namely, good behavior; able to read and write for the positions of president, secretary, and treasurer; and trustworthy and a committed local farmer. The first step in electing leaders involves individual's declaration of interest to stand for election. The names of interested individuals, together with a list of farmers supporting their intent, are then forwarded to outgoing WUA executive who scrutinizes their names with the help of local chiefs to see whether they qualify. After this exercise, outgoing WUA executive releases the names of those who are going to compete for the top positions. All other positions are also supposed to be elected during the general meeting. This procedure seems to be democratic on face value.
However, consultations with farmers revealed that local chiefs and influential people have a lot of influence on the choice of leaders. For instance, consultations revealed that some of the candidates are rejected by traditional and religious leaders on the premise that they do not have the interests of the local people at heart. The study revealed that failure to get endorsement from the local and religious leaders is enough reason for individuals to lose elections as local farmers fear to be associated with persons not approved by their local chiefs. In particular, farmers fear being labeled acting against the local chiefs and those in authority. For instance, farmers revealed that during the elections held in 1999, individuals who contested on position of president were just dictated to irrigation farmers by the local chiefs. In this case, those people who contest and get elected for top positions are generally individuals favored by the local chiefs and, therefore, elites within this social system (see also Ntsebeza, 2004). This situation is ironically passively accepted by the state. According to the National Irrigation Act, the state through District Councils is responsible for holding elections. However, the District Council is not monitoring the operations of the two schemes arguing there is no need to intervene if there are no problems. This situation underscores the need for clearly highlighting the role of the state in enforcing institutional adherence in the process of achieving decentralized resource governance (see Edigheji, 2007). The findings from this study, therefore, demonstrate that while existence of democratic structures have the potential to stimulate democratic practice, state-society relations and formal-informal interaction play a great deal in determining the ultimate outcomes.

Myths and Realities Surrounding Equitable Access to Natural Resources
The blurred positive-sum interaction between formal-informal institutions at the two sites is made clearer when resource distribution (irrigation plots and irrigation water) is considered. In theory, irrigation plot and irrigation water are supposed to be equally shared a priori adoption of community management. However, unclear constitutional stipulation on important issues such as number of plots each farmer is legally allowed to hold has resulted in high perceived cases of neopatrimonialism. For instance, while 85% of irrigation farmers at Domasi hold less than four plots of irrigation land (each plot is 0.25 ha), there are about 15% of the farmers who hold more than five to eight plots and 2.8% of the farmers hold more than eight plots. Most of the people holding more than eight plots at Domasi are leaders in WUA executive, and local chiefs and their relations. Consultations with farmers also indicated that head irrigation plots, which receive a lot of irrigation water, are usually allocated to WUA executive leaders, local chiefs, and their relations, while local farmers are allocated tail irrigation plots. The local farmers argued that "there is favoritism in plot allocation. Plots closer to water are usually given to individuals related to WUA executive members and local chiefs." It is worth pointing that regime change in irrigation management has not changed the pattern of plot ownership (Ferguson & Mulwafu, 2007).
This study found that while irrigation plot access disparities have a long history in the one-party rule, they are currently perpetuated by social and political power relations. In particular, plot and water allocation is highly influenced by social and power relations that exist between WUA executive members and local chiefs. Farmers reported that relationship with local chiefs and WUA executive highly increases individual's chances of acquiring a plot under "the pretext of respect for our community leaders." The social relation refers to the individual farmer's relationship with the chiefs while the political relates to farmers' relationship with individuals in the WUA executive. With regard to irrigation management, the influence of local chiefs emanates from inherent authority of controlling community social life and social standing in the societies. It should also be pointed out that there is weak boundary between social and political leaders referred to herein above. This is so as there is either strong blood relation between influential WUA executive members and the local chiefs or the local chiefs themselves are WUA executive members. At Njala, for instance, the vice president of WUA is the group village head, whereas at Domasi, the president of WUA is a brother to one of the most influential chiefs in the area. It is, therefore, unlikely that inequalities in irrigation plots ownership at Domasi and Njala can be addressed by merely putting in place democratic structures. In the case of Domasi and Njala, compromise between competing formal (WUA executive) and informal (local chiefs and religious leaders) institutions is reproducing undemocratic tendencies within the democratic structures resulting in clientelism and patrimonialism (see also Mamdani, 1996;Merry, 1988;Ntsebeza, 2005).
Equally important, decentralized irrigation scheme arrangements have not resulted in efficient management of the water resource. This is so as water is a mobile resource, and entitlement is effective only when it is captured within the users' resource system. In the case of Domasi and Njala, farming activities upstream diverts a lot of water resulting in less water reaching the two schemes. This is despite both WUAs paying for water use withdrawal rights. This scenario underscores the role of the state in regulating resource use among resource users spread over a vast geographical area and with different organizational arrangement (see Seller, 2010). According to the consultations conducted, institutional gap between Domasi and Njala irrigation schemes, on one hand, and water users upstream, on the other hand, makes it difficult for water users to negotiate water use and address silting up of the high stream. Findings from this study demonstrate that the major problem facing water use regulation between the two schemes and farmers upstream originates from the fact that institutional engineering was limited to geographical area covering the scheme. This approach ignored the fact that management of mobile resources requires institutional framework that addresses the resource system. Thus, decentralized resource governance requires state-society synergy, which would allow both the state and society to mutually benefit from each other by reducing the cost of enforcing rules (Edigheji, 2007).
Decentralized democratic natural resource governance is heralded as a means for poverty reduction. However, the scenario at Domasi and Njala paints a doubtful picture on, first, the extent to which decentralized natural governance can achieve the goal of poverty reduction and, second, the extent to which democratic institutional structure can be relied upon in addressing resource inequalities. This is so as the absence of the state has weakened enforcement of law, and illegal ties between elected actors and traditional leaders have weakened the checks and balances within the system. In the present scenario, ordinary farmers cannot have no room to hold elected leaders accountable. The challenge is that the traditional leaders have long established social and cultural authority over the affairs of the people in the area. Thus, to establish their authority, local democratic institutions need not only the support of local informal institutions but also policy support from the state. The current dilemma is that relations between informal institutions and formal democratic institutions quench off public democratic consciousness cultivated through the creation of democratic institutions. For instance, interviewed farmers in both areas indicated that they have lost trust in both traditional leaders and elected WUA executive.

Deceptive Mechanisms for Transparency and Accountability
Empowerment of local communities is one of the major objectives of decentralized natural resource governance. This involves instituting democratic resource governance characterized by resource users' participation in decision making, observance of resource users' constitutional rights, and transparent and accountable management system. Measured against these variables, there is a lack of democratic resource governance at Domasi and Njala. In terms of resource users' participation in resource governance, 68% and 96% of the farmers interviewed during household survey indicated that they did not take part in the formulation of rules governing resource use. Instead, they argued that the rules were formulated by the WUA executive, and members were just informed about the rules. In terms of transparency and accountability, farmers lamented that WUA executive is not presenting regular financial and audit reports of WUA finance as provided for by their constitutions. In fact, at both sites, WUA executive could not produce even a single financial audit report.
The trend of natural resource use governance at Domasi and Njala reveals that the process of institutional engineering was captured by a few powerful local elites. Ironically, institutional engineering at both sites was externally driven by the state and international agencies to enforce democratic practice. In general, establishment of formal institutions was perceived to be prerequisite for successful decentralized resource governance. Thus, the focus of regime change was on formulating written constitutions for the two schemes. However, as earlier indicated, most farmers were only informed of the crafted constitutions. Lack of adequate farmers' participation in the formulation of the constitution did not only result into most farmers being ignorant of the constitutional provisions but also sowed seeds of resentment against the constitution. These findings reveal that mere creation of democratic arrangements was the main theoretical flaw that was made at inception. In particular, rolling back the state and local informal actors was considered a proper mechanism to avoid their involvement.
However, findings of this study reveal that WUA executive and local chiefs have created power and benefit sharing relationships that are not provided for in the constitutions. The influence of the local chiefs on the management of irrigation schemes is also enhanced by the fact that, unlike WUA leadership, traditional authorities form part of the district development framework as they are members of the district development planning system. To this end, they are able to negotiate matters relating to natural resource management with the state in their favor. For instance, the state has still maintained the role of local chiefs in customary land management. Decentralized natural resource governance, therefore, requires a holistic approach where all institutions interacting with formally created governance institutions need to be not only democratic but inclusive (see Manor, 2011). This institutional arrangement has the potential to allow informal institutions to evolve into a democracy as public consciousness for democracy will not be suffocated by rent-seeking tendencies. For instance, Shackleton, Campbell, Wollenberg, and Edmunds (2002) found that exclusion of traditional leaders from conservation committees in Namibia was counterproductive, resulting in conflict and delays until the chiefs were incorporated into committees (see also Bergstrand, 2003). In the case of this study, while decentralized resource governance can perceived to be stable at face value, it is characterized by a lot of reciprocity and exchange of favors that go against the principles of a democracy. This is as leaders (including local chiefs) have become accountable to government and not local people (see also Manor, 2011). A good example of these favors is a case in which a good market at Domasi was informally but effectively restricted by WUA executive to WUA executive members and local chiefs. Equally important, local chiefs reported attending WUA executive trainings and conferences not meant for them. It is unlikely that these tendencies can be corrected without outside actors (preferably state) monitoring.

Compromised Mechanism for Conflict Resolution
The other rationale for devolving management of natural resources comes from the idea that decentralized natural resource governance reduces conflicts as resource users collectively monitor resource use and compliance to rules. Decentralization also makes resolution of conflicts easier as users employ local values and norms of cooperation. Yet, FGDs at Domasi and Njala reveal that the two sites experience a number of conflicts, including farmers' failure to turn up for collective work at the scheme, illegal water diversions in the scheme, and farmers' failure to clean canals. All these conflicts point to the failure of collective action, monitoring, and persuasion. This article argues that these outcomes are as a result of failure to institute collective action and conflict resolution. This is so as institutional engineering was implicitly guided by the view that putting in place democratic arrangements and rolling back the influence of the central government and its agents were all that were needed to establish decentralized natural resource governance (see Zungu, 1996). This was a misplaced premise as it undermined not only the influence of informal institutions on formal institutions but also the need to involve the state in enforcing democratic rules. In the case of this study, WUA formal institutions were created parallel to existing informal management structures controlled by local chiefs. This arrangement created potential conflict between local chiefs and WUA executive, making it difficult for the latter to resolve conflicts. This is confirmed by the responses of local chiefs who indicated that they are not happy with this arrangement. Instead, local chiefs argue that it is unfair to leave them out of irrigation management as they have a long history in natural resource management (see also Fairhead & Leach, 1995). One clear example of their role is that while WUA executive leaders have the mandate to discipline farmers, they cannot do so without informing and getting the consent from local chiefs. Some offenders are also not disciplined because they are connected to influential local chiefs. Revelations from Domasi and Njala point to the idea that decentralized natural resource governance loses a lot by neglecting traditional structures and customary law and enforcement of laws from the state. Mawaya (2009) argued that formal recognition of local bylaws and rules can be useful in enforcing compliance. This cannot, however, be achieved without formally recognizing local chiefs.

Conclusions and Recommendations
This article has demonstrated that the question of instituting decentralized democratic natural resource governance in areas with a history of strong state management is a complex one. This is largely so as institutional performance of the decentralized structures depends not only on the negotiated competing interests between the local elected and traditional leaders but also on the policy implementation support from the central government. Evidence from Domasi and Njala irrigation schemes demonstrates that the promise of establishing decentralized democratic natural resource governance through democratic arrangements is not automatic. Thus, this study underlines the need for understanding the dynamic interaction between state and society actors (formal-informal) before formulating or designing institutional arrangements for establishing democratic natural resource governance. In particular, understanding state-society synergy and the convergent and divergent outcomes resulting from the influence of traditional actors on elected actors is important in designing institutional arrangements for democratic governance.
The shift to decentralized natural resource governance is ripe and essential taking into account the shortfalls in centralized natural resources management in Malawi, in particular, and in Africa, in general. There are currently high theoretical expectations that, unlike the failed state's fencesand-fines approach, decentralized natural resource management presents the best alternative to managing these resources. However, outcomes of decentralized irrigation scheme management at Domasi and Njala reveal that the positive contribution of democratic institutional structure, while hard to refute, is challenged by a number of factors. First, contrary to the dominant view of rolling back the state, findings from Domasi and Njala demonstrate that there is a need to adopt state-society synergy that is built on trust and reciprocity. This approach is necessary if the state has to enforce local actors' compliance to natural resources democratic policy stipulations such as holding elections for office bearers. Second, findings from this study suggest that the transition from centralized to decentralized natural resource governance requires institutional engineering that can allow democratization of local institutions (both formal and informal). To this end, the future and success of decentralized natural resource governance depends on the extent to which informal institutions will be allowed to evolve together with formal institutions rather than being left out (see also UNECA, 2007). This article, therefore, concludes that decentralization is predominantly a political activity; hence, the transition to decentralized natural resource governance cannot be a matter of just passing democratic legislation. Equally important, the mere institution of formal democratic structures does not automatically lead to achievement of democratic governance as was theorized.

Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding
The author(s) received no financial support for the research and/or authorship of this article.