WPS8101 Policy Research Working Paper 8101 Bureaucratic Blockages Water, Civil Servants, and Community in Tanzania Julia Bailey Development Research Group Impact Evaluation Team June 2017 Policy Research Working Paper 8101 Abstract How do civil servants in district water and sanitation the private sector and community organizations. Using departments address problems of water access in rural qualitative research from two of these water and sanita- communities in Tanzania? What are the bureaucratic pro- tion departments, this report shows that the institutional cedures they follow? How do the bureaucratic procedures and bureaucratic contexts in which civil servants work around formulating budgets, managing money, and inter- redirect their attention away from maintaining existing acting with communities impede or enhance their ability infrastructure and towards building new water projects. to manage water projects? This report addresses these and The focus on new projects corresponds to their efforts related questions by examining the social, economic, and to answer the objectives of higher levels of government. political contexts in which Tanzanian civil servants in the Improving water access depends on the shared efforts of water sector work. This research focuses on civil servants civil servants and community groups to maintain existing employed by water and sanitation departments in district projects. Civil servants’ focus on new projects therefore offices, where infrastructure projects are initiated and man- poses a problem to ensuring that they work commu- aged by engineers and technicians in coordination with nity organizations and maintain existing water projects. This paper is a product of the Impact Evaluation Team, Development Research Group. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://econ.worldbank.org. The author may be contacted at Julia.bailey@mail.mcgill.ca. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team Bureaucratic Blockages: Water, Civil Servants, and Community in Tanzania By: Julia Bailey, DIME Research Consultant Development Impact Evaluation (DIME), World Bank 1818 H Street, NW Washington, DC 20433 USA Email: Julia.bailey@mail.mcgill.ca Keywords: Government bureaucracy, the state, civil servants, water, rural Tanzania JEL Classficiation: Not applicable List of Acronyms Big Results Now (BRN) Central Management Team (CMT) Community-Owned Water Supply Organizations (COWSOs) Department for International Development (DfID) District Executive Director (DED) District Water Engineer (DWE) Government of Tanzania (GoT) Ministry of Finance and Planning (MoFP) Ministry of Water and Irrigation (MoWI) National Water Policy (NAWAPO) Payment by Results Program (PbR) President’s Office for Regional and Local Government (PO-RALG) Results Based Aid (RBA) Rural Water Supply Program (RWSP) Rural Water Supply and Sanitation Program (RWSSP) Tanzanian Shilling (TZ) Ward Executive Director (WEO) Water Sector Development Program (WSDP) Village Executive Director (VEO) Village Water Committee (VWC) 2 Table of Contents Introduction 4 Document Layout 8 From Supply to Demand: De-centralization in Tanzania’s Water Sector 9 Maji Safi and Ngare Sidai 15 Research Methods and Approach 21 Institutional Landscape 28 Bureaucracies and Budgeting 37 Bureaucratic Barriers: Feeling ‘Stuck’ in the Office 41 Maintenance and Monitoring: The Symbolic Capital of New Infrastructure 50 Conclusion 59 Bibliography 64 3 Introduction The role of civil servants in the Tanzanian water sector has transformed over time. In the post-colonial period (1961-1980), the state was the sole provider of public resources and water infrastructure (Cleaver and Toner 2006). This monopoly aligned with the government’s socialist approach to the provision of services and national development more broadly. With the help of the private sector and community-based institutions, however, the state over the past 36 years has moved towards a more decentralized delivery of services to communities. To this end, local government is now the avenue through which water projects are managed, monitored, and delivered in rural villages. Local governments coordinate their activities with Community Owned Water Supply Organizations (COWSOs), which are responsible for maintaining infrastructure by collecting and managing tariffs from water users. COWSOs take ownership of projects after they are registered with the state, and are expected to use the tariffs they collect to budget for infrastructure repairs once projects have been constructed. If these organizations require help from the state to fulfill their mandate, they must ask for support from employees of local government who respond when necessary. This institutional arrangement between local government and COWSOs forms part of a demand-based approach to water-service delivery. Rather than relying solely on the state, communities must actively participate in developing and managing water resources through this approach. While employees of district water and sanitation departments are supposed to work closely with COWSOs to maintain infrastructure, in reality, they manage infrastructure with other employees of government institutions and with the private sector. Their focus towards these actors is the result of the bureaucratic procedures involved in their work, including budgeting, monitoring projects, and organizing contracts with the private sector. The 4 bureaucratic procedures involved in undertaking these activities divert their attention away from the needs of water-users and towards the objectives of higher levels of state. These objectives include building new water projects rather than maintaining existing infrastructure. The state’s focus on building new infrastructure significantly challenges water access in rural areas of the country, where 46% of the 74,000 water points are not functioning at any given point in time (Coville and Rogger 2015). Moreover, a full 25% of all water points break down within two years of being constructed. Consequently, nearly 7.5 million people – 30% of Tanzania’s rural population – live without reliable access to clean water. While the country has made significant strides by expanding the number of water projects in rural areas, the lack of attention to maintenance by local governments severely limits the impact of these investments in the water sector. In order to address the problem of maintenance, the Department for International Development (DfID) proposes directing funds towards a Payment by Results Program (PbR). The PbR program is a form of ‘results-based-aid’ (RBA) that provides financial incentives to Local Government Authorities (LGAs) to expand rural access to water by disbursing funds in response to a set of pre-determined, measurable results. These results include increasing the number of functioning water points by continuing to build new infrastructure and by maintaining existing projects. To understand the impact of the PbR program, DfiD, along with the Ministry of Water and Irrigation (MoWI), is working to collect data on the number of functioning water points in the country. To collect this data, the MoWI are asking district water departments to count the number of functioning and non-functioning water points in their districts. These numbers are then submitted to the MoWI on a monthly basis. This focus on collecting and maintaining accurate data on the number of water points is a result of how the PbR program is 5 designed: disbursement of funds depends on measurable outcomes, and outcomes can only be measured once data on existing infrastructure have been collected. DfiD’s PbR program promises to address many of the financial limitations district offices face in extending services to rural communities. However, as this report details, finances alone cannot answer the many challenges that civil servants face in delivering water to rural communities. Namely, employees in district offices encounter a number of constraints related to government bureaucracy. Government bureaucracy in Tanzania, while having the capacity to create lines of transparency between levels of the state, can also cripple civil servants’ ability to fulfill their roles and responsibilities to rural communities. This problem with bureaucracy is certainly not unique to Tanzania. As Matt Hull (2012) outlines in his discussion of government bureaucracy in Pakistan, the bureaucratic procedures that civil servants follow distill individual agency across a large collective of employees, making it difficult for individuals to act on their responsibilities. Similarly, Akhil Gupta (2012) illustrates how state policies and programs in India aim at addressing chronic poverty are subverted by the very bureaucratic procedures set by the state. As he explains, civil servants become occupied by the procedures of government and, as a result, are insulated from problems of chronic poverty. Hull and Gupta’s points speak to the conditions that employees face in district water and sanitation departments in Tanzania. The observations described in this report are based on findings from qualitative fieldwork in two district water and sanitation offices in Tanzania. The author worked with District Water Engineers (DWEs) and technicians. DWEs lead district water and sanitation departments and are primarily responsible for managing water projects and addressing community needs. Technicians support these engineers by working as liaisons between DWEs and rural communities. In describing how they aim to fulfill these responsibilities, I focus on the activities they engage in 6 as part of their jobs. These activities involve a number of bureaucratic procedures, including those around formulating and circulating budgets, communicating and engaging with rural communities, and managing and monitoring infrastructural projects. By tracing the institutional linkages that these activities create between district water and sanitation departments and the state, I show how civil servants become oriented towards the demands of higher levels of government. This focus poses a problem for managing water infrastructure under the current demand based approach to service delivery in Tanzania. This approach requires civil servants to work closely with community institutions, yet they face a number of barriers in both visiting and communicating with water-users. These obstacles to working with water-users include the very processes of government that civil servants are mandated to follow as employees of the state. This conclusion has implications for understanding state-functioning by turning attention away from the categories of bad/good governance that are often used to describe the provision of state services, particularly in the developing world (Sharma and Gupta 2006). Rather than drawing on binary categories, I consider how state procedures impede civil servants’ efforts to complete their roles and responsibilities to rural water-users. I focus primarily on the procedures that deter them from addressing problems of maintenance with COWSOs. As I outline below, maintenance is not a technical issue, but rather a political problem that involves a number of institutional linkages and bureaucratic procedures across multiple levels of government. This report offers an entry point to considering the intersection between water access, maintenance, and bureaucracy by considering the institutional and bureaucratic dynamics of local government. 7 Document Layout The first part of this report outlines the history of policy in the water sector of Tanzania, the field-site context, and the methodological approach to conducting fieldwork. Here, I describe how civil servants’ roles and responsibilities have changed as the Tanzanian government shifted from a supply to a demand-based model for the delivery of public services. As I outline, civil servants currently act as facilitators between state and non-state institutions when it comes to water projects, as a result of this shift. After describing the history of the water sector in Tanzania, I introduce the reader to the two districts where this research took place: ‘Maji Safi’ and ‘Ngare Sidai’. While there are many environmental and economic differences between the two districts, the institutional dynamics and bureaucratic processes within the water and sanitation departments are similar. The proceeding section describes the research methodologies that inform the findings presented in this report, including participant observation and open- ended interviewing. The remainder of the report outlines my research findings. I begin by describing the ‘institutional landscape’ in which civil servants in district water and sanitation departments work in order to highlight their relationships to other state and non-state-institutions. As I illustrate, employees of district water and sanitation departments engage primarily with institutions, including NGOs and the private sector, that operate at the same level of government or higher. I call these connections ‘horizontal’ and ‘vertical’ linkages, respectively. Their engagements with these institutions means they work less closely with institutions that are situated hierarchically below district offices, including COWSOs. I then move on to discuss budgeting at the district level in order to show how the attendant processes reinforce the aforementioned institutional dynamics. When devising budgets, civil servants work more directly with state institutions than 8 with COWSOs even though the government shares the cost of maintaining infrastructure with communities. Following this section, I describe the dynamics of district offices where civil servants find themselves ‘stuck’ in the office. I explain their inability to travel to rural areas in relation to a set of ‘micro-bureaucratic’ processes, particularly those around communication, which inhibit their ability to work closely with water users. I then show why civil servants focus more on new projects than on existing infrastructure by examining how the political and cultural environment in which they work impedes maintenance, monitoring, and evaluation. These findings show that civil servants lack the material and political resources to perform the full scope of their jobs, especially maintaining existing water points. The institutional and bureaucratic landscape in which they work likewise insulates them from the communities they serve. These factors result in a situation where civil servants have limited avenues for learning about the needs and challenges of water users. Consequently, decisions about how to manage infrastructure become informed by flows of information and funding from higher levels of the state, rather than by the communities that consume and manage water. This situation represents a significant problem for water service delivery in rural Tanzania. Employees of local government are supposed to maintain infrastructure in coordination with communities under the current demand-based approach. However, infrastructure is not being maintained. This report sheds light as to why this is the case. From Supply to Demand in Tanzania’s Water Sector I was one of the first water engineers in Tanzania. There were only 120 of us who were sent to Bangalore for school in an engineering program. When we returned home after receiving our degrees, we worked at the Ministry of Water and in water departments at regional offices. Things have changed though. We now have an engineering school at the University of Dar Salaam and technical colleges. This means we have many more engineers in the 9 country, so projects can be run by the districts, rather than by the regions. We are doing everything ourselves. – Mr. Mbombe, Regional Water Engineer Mr. Mbombe was in his sixties, nearing retirement. He had worked in the Tanzanian water sector for over 30 years and was employed in a regional water department when we spoke. In our discussion, Mr. Mbombe explained to me how policies in the water sector shifted over time – from when the colonial government brought in foreign expertise and engineers to design and construct water projects, to when regional level water departments ran projects immediately after independence in 1961. Post-independence, Mr. Mbombe explained, the Tanzanian state invested more and more funds to train Tanzanians as engineers and technicians by investing in engineering departments and technical schools in its national universities and colleges (Mbombe, April 7th 2016). The expansion of technically trained staff within the country meant that district water and sanitation departments staffed by Tanzanians could run projects on their own. Today, DWEs and technicians help design, implement, and manage water projects throughout the country. The dramatic shift from having water projects run by foreign experts to district officials is the result primarily of changing state policies that align with global trends in the provision of water services. In many countries in the developing world, policy-makers see building institutions as necessary for delivering water resources (Grover and Krantzberg 2013; Koppen et al. 2007). This emphasis on institution building shifts the focus away, to some extent, from the technical aspects of building and maintaining infrastructure. Furthermore, the need to coordinate activities across institutions raises questions around water-governance – that is, which state or non-state institutions are responsible for which aspects of water-service delivery? Increasingly with the shift from centralized to de-centralized development planning that is occurring globally, and in Tanzania (Cleaver and Toner 2006), states are assigning community institutions with the 10 role to manage the day-to-day operations of water projects (Grover and Krantzberg 2013; Koppen et al. 2007). In development parlance, governments are encouraging ‘participatory development’ and ‘community ownership’. During Tanzania’s late colonial period (1940-1950s), the state provided 75% of the funding needed to build and maintain water infrastructure (Cleaver and Toner 2006). Local authorities provided the remaining 25%. Most of the projects that were initiated through these funds focused on improving existing systems of irrigation and building large-scale projects for commercial agriculture (Cleaver and Toner 2006; Therkildsen 1988). As a result, very few rural communities had access to piped water for drinking. In order to address this problem, the federal government took control of water services upon gaining independence in 1961, and promised to provide free water to Tanzanians in rural areas (Therkildsen 1988). To fulfill this promise, Julius Nyerere, the first president of Tanzania, instituted socialist approaches to water service delivery as part of a broader program that he labeled Ujamaa (1961-1980). Ujamaa, the Swahili word for ‘ ‘brotherhood’ or ‘socialism’, was used to promote various state-driven development projects. These projects included covering the entire cost of supplying water to rural areas. Through Ujamaa, rural Tanzanians were able to access and consume water for free. In 1971, the government launched a 20-year Rural Water Supply Program (RWSP) aimed at providing access to safe, clean drinking water within a walking distance of 400 meters from every household in the country by 1991 (Cleaver and Toner 2006; Jiménez and Pérez-Foguet 2010). Implementing this policy was highly centralized, as the federal government replaced local authorities with their own representatives in committees at the district and village level. Although the number of Tanzanians with access to clean water increased from 12% to 46% of the population between the 1960s and the early 1980s (Cleaver et al. 2006), the projects 11 constructed during this period were not maintained. According to Mr. Mbombe, it was not until the 1990s that the Tanzanian government began directing funds to the maintenance of existing infrastructure. This marked a shift from the state’s previous focus on building new infrastructure. As part of this shift, the state also tried to encourage water users to contribute to maintenance costs through a series of policy reforms. In 1991, a new National Water Policy was adopted that established Village Water Committees (VWCs) (Cleaver and Toner 2006; Sokile et al. 2003). VWCs were placed under the management of village governments, where there were also committees for health, education, and planning. The aim of VWCs was to get village government more involved in managing infrastructure by monitoring existing water projects and regularly reporting to district water and sanitation departments. VWCs were also mandated to ensure that village government funds were directed towards repairing existing infrastructure. Since these committees were formed, there have been subsequent efforts to further decentralize water projects through the creation of COWSOs. Unlike members of VWCs, who manage water infrastructure in coordination with village governments, COWSOs can legally own infrastructure independently of the state. According to the law, they are independent from the state once they are registered with local government. While 10% of Tanzanian villages have COWSOs (Coville 2016), the majority of villages in Tanzania still manage their water supplies through their village governments. Where COWSOs exist, they manage water infrastructure with local government through the guidelines provided by the 2009 Water Supply and Sanitation Act. This act stipulates that COWSOs address ‘minor maintenance’ repairs, whereas the state must address ‘major maintenance’ problems. Water policy in Tanzania continued to evolve through the 2002 National Water Policy (NAWAPO). NAWAPO is the policy framework that encourages communities to take a more 12 active role in service delivery (Cleaver and Toner 2006). NAWAPO requires villagers to contribute money or physical labor to capital investments, and to take responsibility for operating and maintaining existing infrastructure. The policy also emphasizes the benefits of having the private sector (e.g. consultants and contractors) assist with designing, constructing, and monitoring of new projects with local government. Within this network of actors, employees of district water and sanitation departments are responsible for supervising local water-user groups and initiating major repairs on water infrastructure when necessary. They also monitor the status of existing infrastructure, solicit input from the people they represent, and coordinate their activities with NGOs, the District Executive Director (DED), the President’s office for Regional and Local Government (PMO-RALG), and the MoWI. Conversely, community-based institutions work more closely with village executive officers, ward officers, and relevant members of the district offices, including water technicians, community development officers, ward committees, registrars, and district water engineers (DWEs). So, while COWSOs are responsible for managing infrastructure and user-needs, civil servants are mandated to respond to the needs of these institutions by providing technical support. While the relationship between COWSOs and local government is delineated in federal legislation, the relationship between these two institutions remains unclear. This is because the 2009 Water Supply and Sanitation Act does not specify what counts as ‘minor’ or ‘major’ maintenance. The lack of clarity in the 2009 Act between these terms means that civil servants can define and decide what counts as a major repair. With little funds to address repairs, local government depends on COWSOs to secure finances to address breakdowns on existing infrastructure. The lack of state finances provided for repairs is likely linked to the cultural and 13 political emphasis on new infrastructure, manifested in state-policies that focus on building new projects. The Water Sector Development Program (WSDP) I and II are examples of programs that focus primarily on building new infrastructure. The WSDP I also creates upward lines of accountability between district water and sanitation departments and the MoWI. WSDP I mandates that district governments initiate water projects to service a minimum of 10 villages. These projects require district water and sanitation departments to work with consultants and contractors, who are hired to design, manage, and build new projects. The WSDP I also mandates employees of district water and sanitation departments to report their progress to the MoWI through a computerized information management system and through water point mapping exercises (GoT 2007). As mentioned above, these water point mapping exercises aim to capture the number of functional water points within a district in order to improve planning, budgeting, and financial management. WSDP II continues these initiatives, but encourages districts to focus on the needs of district as a whole, rather than focusing their attention on the new projects that address issues of water access in 10 villages alone (GoT 2007). During the time of this research, districts continued to follow WSDP I’s initiative to provide water points in 10 villages and therefore were primarily focused on realizing these projects in their day-to-day activities. As this report details, the state’s efforts to decentralize public service provisioning have not been fully realized in the water sector of Tanzania. While communities are mandated to play a larger role in managing infrastructure than under the supply-based model of delivering public services, they are not being fully supported by the state. The lack of support COWSOs receive is perpetuated by the current pressures on local government to meet demands by institutions at 14 higher levels of government, where the focus tends to be on constructing new projects. This focus towards government and not towards communities is reinforced through a number of institutional and bureaucratic dynamics that are described below. Before describing these dynamics, I outline the field sites and methods that inform this research. Maji Safi and Ngare Sidai There are 169 district offices in Tanzania. These offices are located in small towns or regional centers, and contain within them a number of government departments— health, education, community development, agriculture, and of course, water and sanitation. Each of these departments is accountable to their respective ministries. The district executive director (DED) coordinates development programs and budgetary processes across these different departments. In addition to government employees, district councils, composed of elected members from each ward, participate in decision-making within district offices. District water and sanitation departments are required to employ at least one head engineer (the DWE) who coordinates the activities of the department, and at least five technicians. District administrators help technical staff type reports and file letters and other government documents. I worked in two district water and sanitation departments during my fieldwork in Tanzania from September 2015 to June 2016. I refer to these districts pseudonymously as Maji Safi and Ngare Sidai in order to respect the anonymity of research participants. These two field- sites were selected because they provide strong comparative case studies of more general trends in the local water sector of Tanzania. District water and sanitation departments are heterogeneous institutions as a result of the distinct political, social, environmental, and economic contexts in which employees work. Furthermore, DWEs have different methods of organizing both staff and day-to-day activities in their offices. Thus, in selecting the two district offices in which I worked, 15 I did not try to identify a district that was more or less representative of the 169 districts in Tanzania, but considered a number of external factors and logistical concerns that would facilitate comparative research. While I outline the differences between Maji Safi and Ngare Sidai below, it is important to note that both these districts follow the same national legislation and are situated within similar institutional contexts. Moreover, the bureaucratic procedures these departments follow in creating linkage across different levels of the state and government are similar, even though they are adopted in subjective ways. Therefore, while the findings of this report are based on research from two districts only, the trends I describe below likely describe realities in districts in other areas of Tanzania as well. Furthermore, I triangulated my findings from these districts by visiting other districts in the region and by discussing my observations with regional engineers and with representatives from the MoWI and with PO-RALG. I am therefore confident that the findings and observations in this report speak to larger trends across local government in Tanzania as a whole. Maji Safi and Ngare Sidai are located in two different areas of the country with different climatic, environmental, and economic conditions. Maji Safi has a much higher annual rainfall than Ngare Sidai, ranging from 600mm in the region’s lowlands to 1200mm in the highland plateaus. While there are areas in the district with seasonal droughts, where rainfall levels that drop below 600mm, the district nevertheless receives significantly more rainfall than Ngare Sidai, where rainfall levels range from less than 500mm in lowland areas to 900mm in the highlands. Since 2007, however, Ngare Sidai has experienced extended periods of drought where rainfall levels dropped to between 79.3 to 179mm in the lowland areas. 16 In addition to rainfall levels, the ecology of these two districts is dissimilar. This has a major effect on livelihoods. Maji Safi supports agriculture, and most residents of the district earn a living from growing maize, sorghum, sweet potatoes, cotton, sunflower, cassava, and banana. In addition to small scale farming, some communities rely on pastoralism, which involves raising small numbers of chickens, sheep, and cows. Many residents who live close to the major town in the region work as wage laborers on large-scale commercial farms. By contrast, the land in Ngare Sidai is semi-arid, making it difficult to grow food without extensive irrigation systems. To my knowledge, these irrigation systems do not exist in the district. As a result, most residents of Ngare Sidai earn an income from raising goats, sheep, cows, and for some families camels as well. Historically, pastoralists in Ngare Sidai migrated cows, sheep, and goats during the dry season to highland areas, and then returned them to lowland areas in the wet-season, creating patterns of transhumance. Within the last 40 years, however, pastoralists in Ngare Sidai have begun to establish permanent homesteads (enkang/boma) year round, even during the dry season. If rainfall levels drop significantly, herders migrate livestock to where there is available pasture, sometimes traveling to distant locations to also find water, while women and children stay behind in their permanent settlements. These differences in climate, ecology, and livelihood each influence how water infrastructure is built and maintained. In Maji Safi, water projects are largely gravity schemes, wherein water is directed from the tops of hills to storage tanks below, and is then funneled into a series of water points in one or more villages. As well, local communities use boreholes that run on diesel generators and on electricity where it exists to pump water from underground reserves. Small, shallow wells, as well as hand-pumps, are also used to deliver water to villages. These projects are designed to provide clean drinking water to two to four villages with 2,000- 17 7,000 people. However, due to ongoing immigration into the district, these projects may service as many as 7,000-12,000 rural inhabitants. In contrast to Maji Safi, Ngare Sidai relies largely on boreholes to meet its water needs. To construct boreholes, contractors sometimes drill over 200 meters down into the earth to access water. These boreholes run on generators, rarely electricity, and need to be replaced after 10 to 15 years. In addition to boreholes, Ngare Sidai relies on some gravity schemes constructed close to hills to provided water to villages below. Due to the lack of permanent water sources in the district, water in Ngare Sidai is a scarce resource to which many residents lack access. Unlike in Maji Safi, where residents often obtain drinking water from local rivers and through collecting rainwater, in Ngare Sidai residents rely more heavily on built infrastructure. This dependence means that residents will often have to walk up to 20 kilometers one-way to access water when infrastructure in broken. Because the need for water is so acute in the district, many civil servants describe the district’s priorities as being centered on water provision above education and health. In Maji Safi, discussions with civil servants suggest this is not the case. While water-access in that district is considered important, improving access to health and education are higher priorities. Though both districts are subject the same federal legislation and bureaucratic procedures, the internal structures of the district offices in Ngare Sidai and Maji Safi are strikingly different. In Maji Safi, there are seven water technicians and one water engineer. Three of the technicians are stationed at the division level, which in Tanzania includes several wards. Each of these technicians has access to motorbikes and is expected to travel to the central district office once a month for staff meetings. The technicians who work out of the central office have different roles: the head technician directly assists the DWE; another technician is responsible for liaising with COWSOs; and the remaining two technicians are responsible for sanitation and 18 for infrastructural repairs. In addition to these staff members, the office has an administrative assistant. The water and sanitation department in Ngare Sidai has one head DWE who works with two junior engineers, who were transferred to the district in March 2016 due to previous staffing shortages. In addition to these three engineers, there are two technicians: a plumber who performs basic repairs, and a technician who works as a liaison between the district office and local villages. An administrative assistant also works in the office, along with a security guard.1 In addition to these differences in staffing, the district offices in Maji Safi and Ngare Sidai work with different institutions at the community level. In Maji Safi, there are 26 registered COWSOs, whereas in Ngare Sidai there are around four COWSOs, but they are not registered with the local government. This means they cannot own the infrastructure they manage. Village governments still manage the majority of water projects in Ngare Sidai despite recent policy reforms. These differences are not unique to these two districts. In some districts in Tanzania, COWSOs are well established and function much like district water departments. For example, in Hai district, located further east of Ngare Sidai, COWSOs are confederated and are legally 1 These differences in staff are common in Tanzania, where district water departments are heterogeneous in terms of staff numbers and how they are organized. From my observations and discussions with DWEs in other districts of the country, most are constrained by the number of staff. Ngare Sidai, though, was an extreme example of under-staffing. Until March 2016, the district only had one engineer, three technicians, and one administrative assistant. Most district offices have anywhere from five to twenty technicians on staff. Maji Safi’s district water and sanitation department, while neither under- nor over-staffed, was unique in terms of how the DWE organized the technicians, assigning them to divisions where they lived. While other districts considered assigning technicians to work within divisions in order to ensure they worked closely with communities and monitored the status of infrastructure in those areas, most technicians in Tanzania work out of central offices. When asked why this is the case, the DWE in Maji Safi described the difficulties in trying to persuade technicians to work at the division level where there is often a lack of amenities in comparison to the towns where many district offices are situated, including Maji Safi’s. 19 able to build, expand, and repair existing water infrastructure. In Karatu, a district in the northwest of Tanzania, COWSOs collected enough tariffs from water users to purchase vehicle transport, hire employees, buy physical supplies for infrastructure (such as pipes), and maintain their own head offices. In Maji Safi, registered COWSOs are able to expand existing water projects and pay for repairs, making them generally more self-sufficient than other COWSOS in the district and country. There are also differences between COWSOs within Maji Safi. While more established COWSOs work to enhance their capacity to expand and maintain existing infrastructure, newer COWSOs that have only recently obtained legal status are not yet functioning independently of the village government (which, as mentioned above, used to be in charge of managing water infrastructure before the NAWAPO). In Ngare Sidai, most villages organize their infrastructural repairs through non-state institutions (customary institutions, the private sector, NGOs) and through the village government. In addition to these differences, the district water and sanitation departments in Ngare Sidai and Maji Safi have different relationships to the PbR program. Maji Safi is eligible for the PbR pilot project, which will test the efficiency of the PbR program in eleven districts. These pilot districts are selected on the basis of whether they accurately and regularly report the number of water points that exist, and which function, to the MoWI. Maji Safi is eligible because of the high quality of data provided by the district to the MoWI, relative to other districts in the country, combined with its relatively low levels of water-point coverage. Ngare Sidai is not eligible for the PbR program because of the lack of reliable data the district has provided to the MoWI. As I discuss in the methodology section below, this difference affected my relationship with the DWEs and regional staff, who in Maji Safi knew about the PbR program and my role in an 20 evaluation that involves the World Bank and DfiD. In Ngare Safi, neither the engineers in the district office nor at the regional level had heard of the PbR program. As a result, the PbR program was not a topic that we discussed at length. Research Methods and Approach The research in this report draws on a number of different qualitative research methods, including participant observation and in-depth interviewing, in order to understand how civil servants perceive their roles in ensuring water access for rural Tanzanians. While conducting fieldwork, I paid particular attention to how civil servants engage with local communities and assess the need for building new infrastructure versus repairing existing projects. To address these concerns, I spent the first two months of research primarily in Dar es Salaam, where I met with members of the water sector, including NGOs, the MoWI, and employees of the World Bank who are involved in water projects. I also visited two district offices in the Coastal Region and met with local COWSOs. Before commencing research in the district offices of Ngare Sadai and Maji Safi, I also attended a three-day workshop run by the MoWI that focused on training DWEs on issues of governance. I attended this workshop in order to become familiar with how DWEs are trained on issues of governance and broader policy objectives in the water sector. The fieldwork component of this project was conducted in Maji Safi and Ngare Sidai for a total of three months. I spent seven weeks in Maji Safi and four weeks in Ngare Sidai, where I participated in meetings and traveled to field-sites with technicians from district water and sanitation departments. I also met with COWSOs and village water committees and members of 21 the district council. In addition to these activities, I conducted interviews with civil servants who work in regional offices and in other departments, including the department of community development. Along with these interviews, I had numerous conversations with water engineers who work in other district offices and with members of civil society. While all the facets of my research proved valuable, interviews with district employees, regional engineers, water users, and members of civil society provided particularly important insights into the ways in which Tanzanians access and manage water resources. The topics discussed in these interviews were wide-ranging, but in general I sought to gain an understanding of the following issues:  The historical shifts in Tanzania’s water sector, particularly around issues of governance and maintenance.  The career trajectories of civil servants, including their professional histories and ambitions for the future.  The organizational structure and hierarchies within district water and sanitation departments.  The demands and responsibilities placed on district water and sanitation departments to coordinate and communicate their activities with PO-RALG and the MoWI  The ways in which district water and sanitation departments collect and evaluate data concerning the status of water infrastructure and community-based institutions. In addition to these interviews, I mapped the institutional dynamics of district water and sanitation departments as well as the lines of communication between district water departments and local communities. By institutional dynamics, I mean the network of actors across institutions and departments within and outside of the district office who are engaged in the 22 bureaucratic processes of managing water points. Tracing these relationships involved mapping out the processes – as ideal types – that district water and sanitation departments engage in to fulfill their roles and responsibilities. Through this work, I sought to understand the lines of communication between the district and other institutions by considering the types of documentation that flows in and out of district offices. This work involved investigating which institutions DWEs report to and how they communicate with different levels of government. While cellular phone calls, text messages, and emails were all forms of communication employed by the actors I studied, I focused on letters and hard-copy documents, which in Tanzania are the official means of communication between government departments. In doing so, I drew on the methodological approach of anthropologists who have also studied state and non-state institutions by focusing on paperwork. As Hull (2012: 22) decisively states, “[I]f you want to understand bureaucratic activities, follow the paper”. In his ethnography of the Central Development Authority (CDA) in Islamabad, Pakistan, Hull demonstrates the methodological and analytically importance of tracing the ‘documentary regime’ of the institution. The documents he ‘follows’ include files, reports, letters, maps, petitions, and office manuals. Hull’s focus on the documentary regimes of a state institution is a methodology that Harper (1997) also employs in his ethnography of the International Monetary Fund (IMF). Namely, Harper looks ‘Inside the IMF’ to understand the intersection between documents and digital technologies. By tracing the ‘document career’, he shows how graphic artifacts travel outside the offices where they are produced and become embedded across a network of institutions. Harper’s approach to following documents across institutions highlights the importance of using documents to trace institutional linkages and the resulting political landscapes they create. 23 I draw on the methodological approaches of these authors to understand the bureaucratic procedures and institutional landscape of local government, but also to address issues of ‘representation’. Representation involves asking how government employees of various bureaucratic institutions formally communicate their work and their needs (Sharma and Gupta 2006). These forms of representation include, for instance, procedures for promotion, organizational charts, requests for funding, and personnel files. In the last section of this report, I discuss how water-point mapping exercises can be one mode of representation that the state uses to address the needs of the MoWI. Anthropological notions of the state also inform this research. The ‘Anthropology of the State’ looks at the actions of individuals within state-institutions and considers the power- dynamics that surround their routine tasks, beliefs, and social interactions with others. This focus on the daily activities of state-actors moves away from understandings of the state as a pre- constituted entity (Sharma and Gupta 2006). As Sharma and Gupta (2006) observe, although states may resemble each other structurally, the forms they take in specific locations differ considerably. Furthermore, the differences in the political and social dynamics within different branches and levels of the state make it difficult to define the state according to a single ‘ideal- type’ (Weber 1922).2 Rather than trying to depict the state as characterized by a single attribute, anthropological approaches to understanding the state focus on the contradictions and complexities that exist across different branches of government. This approach to understanding the state begins by looking at the ‘cultural constitution’ of the state (Sharma and Gupta 2006:11): 2 Weber (1922) developed the term ‘ideal type’ as an analytical construct that allows the researcher to ascertain and measure the similarities and differences between cases. He developed an ideal type of a Western bureaucracy. 24 that is, the social actions of individuals and groups as they navigate through the institutional, cultural, and political landscapes they face in their professional lives. In discussing the cultural constitution of the Tanzanian state, I draw on the professional experiences and perspectives of civil servants at multiple levels of government. I also consider their relationships with non-state actors, including NGOs. While NGOs are meant to be ‘non- governmental’ organizations, they nevertheless operate in similar ways to the state by managing water service provisioning in rural areas. They do so by building projects and training COWSOs. By working with NGOs, but primarily with civil servants, I saw the state at multiple levels and intersections. Through this layered perspective of the state, I am able to outline the contradictory aspects of Tanzanian government. Most especially, the government’s attempt to de-centralize water service provisioning through a set of activities that centralize power and decision-making at higher levels of government. My entry into Tanzania’s bureaucracy was through the very documents I studied – letters. I was able to gain permission to conduct research in Maji Safi and Ngare Sidai through letters that PO-RALG provided to in September 2015. These letters were also circulated within the regional secretariat. Once I had approval from the district office and regional secretariat, I was, in theory, required to travel with a letter from the District Executive Director (DED) when I visit a village government. These letters communicated to village governments that I had permission to ask questions of them concerning water infrastructure, governance, and maintenance. In practice, I was able to visit most village governments without written approval. In addition to these letters, two research assistants, as well as numerous NGOs, facilitated this research. In Maji Safi, I worked with a research assistant named Joy (an engineering intern in the department of water and sanitation) for my first ten days who the DWE appointed to work 25 with me. Through Joy, I was able to access government files, including letters written by COWSOs, monthly meeting minutes, and correspondences between the district office and other stakeholders (contractors, NGOs, and village governments). Joy also helped me organize field- visits and facilitated my participation in important events, such as ceremonies for celebrating new infrastructure that were organized by the regional government. After the first week, however, she went on personal leave. As a result, I had to rely more heavily on the DWE, who facilitated my interviews and travels to field sites. During these activities, either the DWE or a member of his office accompanied me. This meant that the DWE facilitated my interviews with members of rural communities by, for example, ensuring that I traveled with technicians when I left the office. This made it difficult to triangulate my findings. However, I was able to observe, very closely, how civil servants interact with members of communities and other stakeholders in the water sector. This proved important to understanding the dynamics between district water and sanitation departments and COWSOs, as well as other relevant parties. In Ngare Sidai, I worked with a research assistant named Noel who was not an employee of the district offices, but instead a resident of the district who had conducted research on many different topics (HIV/AIDs, Land Rights, Pastoral Governance). She was well informed on a broader set of issues that affected water users, as she herself was one. With Noel’s assistance, I was able to travel to meet with NGOs, local community members, and village governments. Through these activities, I gained a stronger understanding of how local communities viewed civil servants at the district level. Thus, Ngare Sidai offered a viewpoint of district offices from the perspective of communities, whereas Maji Safi provided me with a stronger understanding of the water sector from the perspective of civil servants in local government. I see these two 26 different perspectives as complimentary rather than conflicting, since they allow me to compare the opinions and experiences of two key groups of actors. While this research benefited from these various activities and opportunities, I also faced many constraints and barriers. Since I had a limited amount of time in which to conduct this research, my focus was on the district water and sanitation departments rather than on institutions such as NGOs, departments of community development, and regional offices, which also play important roles in the water sector. As well, since I was not able to conduct research during the full cycle of budgetary processes, I was unable to follow these processes as they occurred in real-time. Instead, I relied on secondhand accounts and anecdotes from civil servants, who described their experiences of formulating and finalizing budgets. In addition to these limitations, I faced barriers accessing information and sometimes participating in the daily routines of the district offices. This was particularly true in Ngare Sidai. Unlike in Maji Safi, where the DWE provided copies of previous budgets and PO-RALG reports, and where I was invited to attend meetings with NGOs and the regional office, in Ngare Sidai I was not privy to these documents or events. Furthermore, when I did conduct interviews with the DWE, his responses lacked detail and depth. It was clear from these interviews that he was reluctant to share information with me. With the help of my research assistant, however, I was able to triangulate his responses through interviews that I conducted within the district offices and with members of local communities. The DWE’s reluctance to assist me in my research was undoubtedly linked to my status as an outsider and an employee of the World Bank. During my research in Ngare Sidai, there was an ongoing investigation into financial discrepancies concerning a water project, which civil servants discussed as being a ‘World Bank’ project. One member of the district water and 27 sanitation department was implicated in this investigation. While I routinely explained that I was not involved in the investigation, I suspect that the controversy likely limited my access to the activities of the district water and sanitation department. While my position as a researcher for the World Bank, rather than as an independent researcher, likely hindered by ability to gain access in Ngare Sidai, in Maji Safi I suspect that my links with the World Bank and DfiD worked to my advantage. The DWE and engineers at the regional level were aware of the PbR program and were therefore willing to support my research. The disadvantage was that many of the responses from district staff focused on problems of funding, when in fact other issues also placed significant constraints on civil servants in performing their jobs. To account for this, I probed and challenged responses to my questions. In sum, these methods and approaches to research inform the findings and conclusions that I describe in the sections that follow in this report. Institutional Landscape The term “institutional landscape” refers to a number of institutions that work together and alongside district water and sanitation departments to participate in water service delivery. Below is a chart of the different institutions that district water and sanitation departments in Tanzania communicate and coordinate their activities with in order to monitor and manage water-access in rural areas. 28 COWSOs are situated at the bottom of this chart to highlight their relative proximity to water-users and infrastructure. COWSOs are mandated to play a pivotal role in ensuring that projects are maintained through collecting tariffs from water-users and monitoring the status of water points. While different COWSOs have different legal abilities, they generally have a better understanding of the status of water points than district water and sanitation departments. As such, they are important institutions which, at a minimum, can be responsible for reporting back to the district on the number of functioning water points in a region and for advocating for government help when they need technical support and resources from the state. While some 29 COWSOs report directly to the district by writing letters to the District Executive Director (DED) who then forwards them to the DWE, most COWSOs report directly to ward governments on a quarterly basis. Ward-level governments act as channels of communication between villages, COWSOs, and district offices. The civil servants that work at the village and ward levels include a Ward Executive Officer (WEO). Village Executive Offices (VEOs), along with COWSOs, meet with WEOs and elected councilors every three months to discuss development issues. Representatives from various committees, including committees for education and health, also attend these quarterly meetings. In Ngare Sidai, where there are no registered COWSOs, village governments are responsible for ensuring that village funds are directed towards repairing water infrastructure. To do so, they work with customary institutions – ‘age-sets’ - that have historically and still today manage communal resources, including water infrastructure, which I explain further below. Briefly, age-sets are composed of men who are divided according to age and are responsible for managing communal resources. Councilors from wards also attend meetings within the district, where they report on water-related issues to the DED and relevant department heads, including head engineers in water and sanitation departments. If, for example, there is a problem with a water project, DWEs send technicians to the project either to provide technical advice or to perform a simple repair. In the case of repairs, the village government or COWSOs, where they exist, are responsible for covering the cost of the technician’s transport and per diems, if he or she is required to leave the office for over three hours. District water and sanitation departments are also required to support COWSOs when they face problems securing funds, managing resources, or mediating the politics of managing water infrastructure in the villages they serve. These politics include 30 problems with collecting tariffs from water users, but also from other state institutions such as local prisons or health clinics. As well, COWSOs face difficulties managing water infrastructure independently from village leaders who sometimes seek informal power by controlling water infrastructure. The private sector works closely with district water and sanitation departments as well, even performing some of the roles and responsibilities of DWEs. The main actors from the private sector are consultants and contractors. Through procurement processes, districts hire contractors to build new projects and to repair existing infrastructure. Their contracts also often include clauses for establishing COWSOs and training local technicians to manage new infrastructure. Contractors also must work with consultants, who are employed to design projects with district water and sanitation departments and to liaise between contractors and DWEs. In this role, they monitor the status of new projects, facilitate site-visits with civil servants, and report back to the district on the status of ongoing projects. During the time I was in Tanzania, however, this arrangement changed as a result of budgetary constraints at the ministerial level. As a result of a lack of funds to pay these consultants, many are being let go from projects. To fill their roles, regional water departments are being asked to assist district water departments in monitoring the construction of new projects. In addition to ward and village governments, COWSOs, and the private sector, district water and sanitation departments have ‘horizontal’ and ‘vertical’ linkages with institutions that exist at higher levels of government to community.3 By horizontal linkages, I refer to institutions that work with these departments that are neither situated hierarchically ‘below’ or ‘above’ them 3Ferguson and Gupta (2002) use the terms ‘verticality’ and ‘encompassment’ to discuss the spatial dimensions of the state. What they call ‘verticality’ and ‘encompassment’, I describe here as vertical and horizontal linkages between district water and sanitation departments and state and non-state institutions. 31 in terms of legal and political power. These include non-governmental institutions that districts work with, as well as other departments within the district. In both Ngare Sidai and in Maji Safi, NGOs help train and establish COWSOs and local community organizations. They also build new infrastructural projects. To do so, many NGOs work with district water and sanitation departments by asking for permission to work in the district and by eliciting technical and institutional support for their projects. The support district and water sanitation departments provides takes many forms, and includes efforts to create COWSOs, to provide technical advice on new infrastructural projects, and to assist with the hiring of private contractors. While NGOs are situated neither hierarchically above district water and sanitation departments nor below, there are complicated power dynamics between the non-governmental sector and district water and sanitation departments. District water and sanitation departments often need to abide by the demands and conditions set by these institutions. For example, in Maji Safi, the DWE wanted to work with a local NGO that required communities to raise a significant portion of the funds needed to build infrastructure as part of the ‘demand-based approach’ to service delivery. Within the institutional framework provided by the federal government, the DWE in Maji Safi complained that demanding resources from villagers was too difficult for him to initiate. As he explained, many communities expect to receive water for free from the state, making it difficult for his department to extract significant revenues from villages in order to initiate projects. In addition to these constraints, the DWE in Maji Safi occasionally had to significantly adjust his approach to managing a water project by complying with the demands of NGOs. For example, in one instance he agreed to form two COWSOs for a water project funded by a foreign NGO. The NGO’s idea was to have one COWSO in each of the villages serviced by a single 32 water project. The DWE engineer opposed this approach, based on a previous experience he had with two COWSOs that were formed to manage one project that supplied water to two villages. Working separately, he explained, the two COWSOs ran into a number of both financial and logistical problems. Based on this experience, he encouraged the NGO to support a unified COWSO, with sufficient finances and equipment to maintain the project. The NGO did not take this advice, and as a result the DWE agreed to form two COWSOs for one project, despite his reservations. While NGOs can wield a significant amount of power over district water and sanitation departments, they too have to ask for permission from local government to train COWSOs and to construct new projects. To do so, they write letters explaining their objectives to the regional secretariat and the DED. The DED forwards these letters to the DWE, who choses whether or not to meet with the NGOs and help coordinate their projects. In reality, DWEs often face challenges trying to coordinate with NGOs, complaining that they often build projects without informing the district. This poses a problem for district officials since it means they are not always aware of new infrastructural projects that are being constructed and that will soon need to be maintained. District water and sanitation departments also work with other departments in the district, including the department of finance and planning and the department of community development. Most especially, the department of community development plays an important role in helping the district water and sanitation department work with local communities and creating COWSOs. They do so by employing a ‘registrar’ who registers these institutions as legal entities in coordination with the DED and the district water and sanitation department. To be formally registered, COWSOs are required to submit an application to the district, which includes a registration fee and an outline of their constitution. COWSOs must also have a bank 33 account number, which they disclose to the district. Once the district receives an application from a COWSO, the DED coordinates a series of steps with the registrar and the DWE to register these COWSOs as legal entities. District water departments answer to the DED on a weekly basis, as the DED plays a major coordinating role between district departments and elected councilors. The relationship between civil servants, including those in water and sanitation departments, and elected officials is often tense. This is a byproduct of the different roles and responsibilities of these actors. Civil servants perceive councilors as irresponsible when they promise their constituents resources and services that the government cannot provide in order to gain political support. For example, during the 2016 national elections, many office-seekers promised voters free water. This promise directly contradicted the current policy approach which dictates that rural Tanzanians must pay for the water they use. Civil servants also complain that councilors fail to understand the needs of the district as a whole, leading to situations where they advocate for resources to their home- areas without considering their village’s needs in relation to other needs of the district. In addition to COWSOs, NGOs, the private sector, and other department heads in the district, DWEs report back to PO-RALG, the DED, and to a number of committees within the district administration. PO-RALG and the DED are a ‘vertical linkage ’, situated hierarchically above district water and sanitation departments. The role of PO-RALG is to coordinate district water and sanitation departments by offering technical support and by collecting data from reports which district water and sanitation departments submit to the region on a quarterly basis. In these reports, engineers list the number of functioning water points, the status of COWSOs, (the amount of money they have in their accounts, their major challenges, and their legal status), new projects being undertaken, and budgetary information (how much money they have at their 34 disposal versus how much money they have released for new projects). To discuss the findings from these reports, the regional water department hosts meetings with DWEs from each of the districts within the region. During these meetings, DWEs present the opportunities and challenges they face in their districts, while the regional engineer discusses issues that affect the region as a whole. PO-RALG then compiles these reports and submits the data contained in them to the MoWI. These data include the number of water points, COWSOs, and ongoing projects that exist across the region. To varying degrees, COWSOs, village and ward governments, the DED, departments for finance and planning, and community development, NGOs, MoWI and PO-RALG are the main institutions that district water and sanitation departments work with. However, there are a number of other institutions to which employees of a district and water sanitation department report to through ongoing meetings and formal reporting processes. These reporting processes include submitting written reports. The DWE reports to the committee for security, which a district commissioner chairs and which coordinates activities with regional security committees. She or he also reports to a Local Authorities Accounting Committee, as well as a number of standing committees within the district council (1.finance, administration and planning, 2. education, health and water, 3. economic affairs, work and environment, 4. HIV/AIDS). These committees within local government require the DWEs to provide frequent updates on the status of ongoing projects, their department’s expenditures, and the contracts they have with the private sector. DWEs also meet with the Central Management Team (CMT) on a weekly basis. The CMT is composed of department heads in the district. District water and sanitation departments are accountable to other departments and committees in the district and to higher levels of government. This is the result of a number of 35 formalized institutional practices that exist in districts, including regular meetings between DWEs and DEDs, PO-RALG, and formal procedures for reporting back to these state-actors and institutions. Conversely, DWEs are not required to have scheduled meetings or to report back to COWSOs or VWCs where they exist. This discrepancy means that the lines of communication between lower levels of government and water-users are directed upwards, thereby centralizing decisions at higher levels of the state. The budgetary processes that occur within district offices exemplify this point. As the graph below outlines,4 budgetary processes filter decision-making upward. This leads to a situation where district water and sanitation departments are focused more on the bureaucratic processes of the state and less on community-dynamics around water- access and infrastructure. 4 The graph on the budgetary procedures is a simplified representation of the budget process because it does not represent the full range procedures and institutions involved in creating budgets. Rather, the graph shows some of the institutions and procedures involved in creating budgets from the perspective of civil servants in district water and sanitation departments. By looking at budgeting from their perspective, budgets devised at the ward-level are not represented in this chart. 36 Bureaucracies and Budgeting Budgeting processes for water begin at the district water and sanitation departments, rather than at the local level with COWSOs. COWSOs have their own budgets and demand resources from the district when needed, including when they require major infrastructural repairs. Ultimately though, it is the DWE who decides whether or not to meet the demands of COWSOs in annual water budgets. The budgets that district water and sanitation departments devise have columns that account for a number of costs, ranging from fees for consultants to the costs of office maintenance. There are also staff-related costs, which include per diems and costs associated with staff training. District water and sanitation departments also must account for the 37 costs of fuel and repairs for vehicles. Once the district has budgeted for their costs, the budgets move across the district office and then upwards to higher levels of government. Within the district office, a committee, composed of elected councilors, reviews the water and sanitation budgets. Within this realm of decision-making, I was told, political interference occurs when civil servants propose seemingly ‘objective’ solutions to water problems, while councilors use their power to further their political aims. Councilors do so by trying to direct funds to their villages, even when other communities have comparatively fewer water points and less water infrastructure. As one engineer told me of the budgetary negotiations, “it’s too political. There are technical solutions to these problems, but it’s all about politics when you work in government”. Statements such as this highlight a dilemma when planning budgets: while civil servants see issues of water access as being driven by technical factors, their jobs require them to also manage the ‘politics of water’. By ‘politics of water’, I refer to attempts by elected leaders to secure support from certain communities and individuals by promising water infrastructure. One way this was done during my time in Tanzania was by promising ‘free water’ to constituents and by directing projects to home constituencies. From the perspective of civil servants, these promises make it difficult for them to follow mandates from government legislation from higher levels of the state. As a result of these tensions, there are delays during the budgetary process as councilors and civil servants compromise and negotiate the contents of these documents. For example, DWEs are sometimes forced to alter their plans for where to build new infrastructure. As one DWE told me from the Coastal Region, “moving a project from one village that needs a new borehole to another village that also needs a borehole within a powerful councilor’s area. In the end, the technical solution always wins, but we make compromises”. While I cannot comment 38 on whether the technical solutions proposed by DWEs always ‘win’, I do suspect that DWEs find ways to negotiate and navigate political tensions by strategically moving projects to the areas where powerful councillors reside, albeit to villages where projects do not exist or where they need to be expanded. In essence, these strategic moves by DWEs to manage both the demands of powerful councilors and the needs of water-users across the district depend on their ability to understand the politics of water as well as the technical aspects of water access. Once the committee has made an agreement, the budget is passed to the finance committee that exists at the district level. The finance committee reviews the budget and afterwards forwards it to the district general assembly. At the general assembly, the budget is debated and can even be rejected, although many civil servants told me this rarely occurs since two previous committees have already reviewed the document. From the general assembly, the budget travels to the MoWI and finally to the Ministry of Finance (MoFP). From the MoFP, the budget is debated by Parliament before it is finalized. Often there are delays between when the budget is finalized and when districts receive money for their water and sanitation programs. For example, the MoWI allocated 10 million TZ for maintenance that was scheduled to be received in November by district water and sanitation departments. However, by May of 2016, the department in Maji Safi had yet to receive these funds. The delays in the time between when decisions over funds are finalized and when local governments receive money are not unusual. Instead, they are a result of bureaucratic processes that revolve around money and funding at multiple levels of the state, including within district offices. There are different types of funding that enter into districts, including funds specifically allocated for certain water projects and funds that are given to district offices to use at their discretion. Furthermore, NGOs that work within districts provide money to local governments to 39 initiate their projects. Funds that are specifically for district water and sanitation departments move through a separate set of channels than those provided by NGOs. Funds that enter into the district by external institutions, such as NGOs, travel through a number of institutions within each district. These institutions are principally responsible for budgeting and planning projects in coordination with the finance committee, district planning officer, DED, DWE, CMT, as well as a procurement officer and tender board. If, for example, the district received funding from an NGO for a water project, members of the CMT need to meet to discuss how to spend the funds. Once the members of the CMT have agreed on how to use the money allocated to district water and sanitation departments, the district begins to plan how to disperse the money to the relevant parties, including contractors and private companies. This involves an extensive procurement process. To initiate the procurement process, the DED and the district’s procurement officer create tenders for projects. For large-scale projects, multiple tenders are often required. For example, if the project requires pipes, cement, and human labor, separate tenders need to be advertised for each of these items. While the tenders are being circulated, the DED and the procurement office create a ‘tender board’, which acts as an evaluation committee for the procurement process. Members of this committee are paid a ‘sitting fee’ and sometimes per diems for their time when they need to attend meetings outside the office to discuss the tenders. During these meetings, a minimum of three members needs to be in attendance before any decision regarding the tender can be made. Importantly, the DWE cannot be a member of this tender board. After the committee has approved applications for the tenders, the approved applications are reviewed by the DED, who can decide whether or not to veto the tender board’s decisions. If the DED decides to approve the board’s decision, she or he writes a letter to the CMT to inform 40 department heads the status of the procurement process. If the money required for the project exceeds 50 million TZ shillings, then the contract needs to be signed by a public prosecutor. Once signed, the funds are put into to the district sanitation and water account, which is managed by the DWE. The time between when district water and sanitation departments initiate the budgeting process (around October) and when money arrives at the district office is significant, often exceeding one year. These delays are the endless topic of discussion at the district, where discussions I had with civil servants about budgets focused overwhelmingly on the timing of funding. As many district-level civil servants described, initiating new projects and planning for repairs, as well as going about the day-to-day activities of district water and sanitation departments, depends on when funds get deposited in district water and sanitation bank accounts just as much as the actual amounts received. The section below discusses the implications of these delays on employees’ everyday work-routines. Bureaucratic Barriers: Feeling ‘Stuck’ in the Office As this section shows, civil servants often get ‘stuck’ in their offices as a result of problems with planning and communication across different levels of the state, including delays with budgets. As mentioned above, technicians offer support to DWEs, particularly in helping rural communities with problems of maintenance. Yet despite this mandate, technicians often find themselves ‘stuck’ in the office, removed from the communities they serve. In explaining why this is the case, I begin by describing the secondary effects of the bureaucratic processes around budgeting, and then describe the ‘micro-bureaucratic’ processes around communication between district staff and rural communities. Through a discussion of these processes, I argue 41 that there are a number of barriers between technicians and the communities that cannot be explained by funding constraints alone. Discussions around budgets focus on issues of timing: when districts receive funds, when they can pay contractors, and when they can plan field-site visits. Importantly, technicians have very little influence in these concerns, and funding delays often circumscribes their day-to-day activities. For example, the department of water and sanitation office in Ngare Sidai did not have any power during the time I was there (May-June), as a result of delays receiving funds from the district. When I asked why these funds were delayed, the administrator, Ester, explained that the budget for electricity was too low. As a result, the office could not pay its bills beginning two months prior to my arrival. To receive additional funds for this from the district, she told me that: We need to write a letter requesting the amount of money we need for more electricity to the DED. The DED then sends our request to the finance committee who checks to see if there is money available. To budget for the electricity, we also need to provide the previous receipt to the finance department to check to see when it was paid and when the electricity ended. Once we make an arrangement with the finance committee, they can give us the money to pay for electricity. Technicians did very little to expedite this process, explaining that they needed to wait for the DED to sign the forms needed to receive funds from the finance committee. The problem of electricity had a ripple effect in the district office. During the month I was there, for example, the lack of electricity meant employees were forced to migrate to the main building, where most of the district department’s offices were located, in order to charge their phones or use a computer. As a result, the employees of the department were geographically fragmented and disjointed, with technicians, engineers, and the DWE working in separate offices that were spread across a large 5-story building. Meanwhile, the office for the department of water and sanitation remained vacant during large periods of the day. This was, as 42 a technician named Juma told me, bad for both morale and for organizing activities. As he explained: “I never see the DWE because he’s always in the main office. I come to work and there is sometimes nobody here and I don’t know what to do”. To add to Juma’s point, the absence of employees in the office also meant that water-users did not have a central place to come to demand services from local government. During the time I was in Ngare Sidai, I never once saw a water-user in the office asking for help from the district. Juma’s sense of confusion over what to do on a daily basis was certainly not unique to his experiences working in local government. Many civil servants lack a strong understanding of what to do during long periods of time during the day as a result of the lack of resources at their disposal and the lack of agency afforded to technicians. As one lead technician explained to me, named Peter: “you know, there are many things we would like to do, but we can’t. We need to wait until we get funds to repair our vehicle. We don’t have any money, so we don’t have transport. So we are here with nothing to do”. Peter continued by describing how civil service was different from the private sector, attributing his inability to plan his day to more general attributes of working for government. As Peter described: “In the private sector you need to hustle, there are deadlines and people are serious about those deadlines, so you can’t be bored. But in the civil service, you spend your time waiting because nothing happens according to the deadlines you set. So you just wait, bored”. Like so many other civil servants I met, the inability of technicians to control the tempo of their work environments meant that waiting was part of their daily work routines. Rather than being idle, civil servants had different strategies for filling wait time. In Maji Safi, civil servants worked during the day on tasks that were unrelated to their primary jobs, but which serviced the district more broadly. For example, many men spent their mornings and even 43 their afternoons cutting grass outside the office and grooming the few plants that decorated the grounds. This job was mandated by the central government as a way to reduce the need to hire more staff, but became part of some technicians’ efforts to meaningfully fill their days with work. This was especially the case for one technician named Bombe whom I came to know in Maji Safi. Bombe worked as a technician for a private company for over twenty years before joining the civil service. When I asked him why he chose to work for the state, Bombe explained, like so many other employees of the district, that the state-employment provided him with job and security and a regular wage, unlike in the private sector where income depended on consumer demand and therefore the conditions of the market. Bombe chose to work for the state in order to secure a regular income for his family. Every morning Bombe cut the lawn, even coming to work on weekends to make sure the grounds were maintained. For him, waiting to go into the field to fix a pipe or to advise a COWSO provided him with the opportunity to perform ‘little tasks’. Performing these ‘little tasks’ formed part of his strategy to fill his days with meaningful activities. Many other civil servants I met also filled their time by performing odd jobs, such as helping to write reports, attending meetings, and socializing with employees from other departments in the district. They also talked about current events in the news and debated a number of decisions made by the then new president. These hours spent precariously ‘busy’ and sometimes idle formed part of the social life of the district offices. In district offices, civil servants often had to find ways to manage boredom and their desire to actively be engaged and busy during the day. While civil servants waited for resources, so too did communities. As many communities are dependent on help from technicians to repair water infrastructure, the lack of face-to-face 44 interaction with technicians deterred them from wanting to work with the state to manage water projects. For example, a community in Ngare Sidai that I met with needed help repairing a project with a borehole and a series of water tanks. The village was relatively accessible compared to other locations in the district, as it was located only 15km from the district offices along a tarmac road. The village wrote a letter to the district asking for technical advice on the broken-down project. As one member of the village government explained to me, it took the district almost two months before they sent a team of technicians to review the cause of the breakdown. Since most of these villages depended on the infrastructure to access water (there were no permanent streams or rivers in the village), the time it took the district to respond to their demands was seen as unreasonable. As a result, for subsequent repairs, the VWC hired a technician from the closest large city – over 150km away – to travel to provide technical advice. As I heard from other VWCs and members of civil society in Ngare Sidai, many other villages also worked more closely with the private sector and with NGOs as a result of the district’s inability to promptly respond to their needs. Having heard the frustrations of community members and the complaints of civil servants who felt ‘stuck’ in the office, I often asked the junior water engineers in Ngare Sidai why they did not travel to meet with village governments. As one engineer explained to me: “We don’t have access to a vehicle today and we would have had to plan it ahead of time to arrange transport, per diems for the driver, and fuel costs. We have not contacted the people in the village government. I don’t even know who to call to meet.” As this quote details, there were both financial and logistical issues involved in traveling to rural communities. Many state-employees 45 feel financially impaired, particularly when per diems are not available5, but also socially and politically constrained in their efforts to organize site-visits to existing projects. Civil servants are impeded by the lack of social contact they have with members of the communities they serve. This lack of contact has much to do with the bureaucratic procedures required to arrange visits. To formally arrange a visit with a rural community, village governments and community representatives have to be informed ahead of time that a meeting will take place, typically at least one week in advance. Writing and sending formal letters is how this task is done. While mobile phones are also used to confirm these meetings, letters are the ideal form of communication precisely because they can be physically documented, unlike a telephone conversation. Yet, though letters materially represent lines of communication between parties, they do not necessarily create substantive ties between civil servants and the communities they serve. In fact, they can even create social distance by simplifying more fluid interactions and relationships. Take the example of a visit I attended with the technician who was charged with assisting a COWSO in NS or MS. During the visit, we went to the village office to sign the logbook to show that we were in the village for a meeting. Afterwards, we traveled to the COWSO’s office, 5 While I do not think lack of funds alone explains why civil servants are ‘stuck’ in the office, per diems is an important financial constraint to discuss when considering the movement of civil servants from offices to rural areas. To travel to communities, civil servants are entitled to per diems if they are away from the office for more than three hours. Their per diem payment depends on their rank within the government. For example, DWEs are paid higher per diems compared to technicians. Furthermore, the amount of money paid in per diems to civil servants increases if they are required to be away from their workplace for longer than one working day (7:30 am – 3:30 pm). As well, drivers who are employed by district offices are also paid per diems, albeit at a lower rate than other department employees. As a result, the cost of traveling to communities can be significant for a district water and sanitation department, since DWEs travel to sites with a driver and often two to three technicians. While these costs are significant, many civil servants view per diems as a necessary supplement to their salaries, which they describe as being significantly less than what they could earn in the private sector. 46 where all ten members were waiting to speak with the technician. The COWSO was having difficulties collecting funds from a local prison, which required water from a project built three years earlier. While the COWSO successfully collected funds from water-users, the prison refused to pay due to a dispute about metering. The technician told the COWSO to write a formal letter to the prison, explaining that the COWSO will cut off its water supply if it does not pay its bills. He also suggested writing to outside donors for help expanding their existing project. As he said: “I’m just a little person in the district, I can’t do much, but you can write a letter to the DED, or the DWE asking for funds”. Looking at me, he continued: “Why don’t you write the World Bank”. As these suggestions indicate, the technician did not see himself as someone who could solve the COWSO’s problems, but instead saw himself as an advisor who could direct the members of the COWSO to appeal to more powerful institutions by way of letters. These letters, while having the potential to create linkages between institutions, in this situation worked to prevent civil servants from meaningfully assisting the COWSO. At this very same meeting with the COWSO, the technician and I took photos of the members of the COWSO before departing. While saying goodbye, the chairman of the COWSO and the technician continued to discuss problems with water access and funding, as well as technical aspects of managing the existing infrastructure. The conversation about the water project in question eventually led to a discussion about farming and family life. At the end of their conversation, the chairman of the COWSO asked the technician for a copy of the picture, encouraging him not to forget to send the images over his phone. The technician responding by promising to remember to send the photos along, after which the chairman jokingly said: “You’ll remember, because I’ll write a letter to the DED requesting the pictures. So you’ll hear from me 47 soon. Then you can write me a letter back, detailing when they will be received.” As this interaction suggests, letters also serve to formalize more fluid exchanges that exist between technician and communities. Members of rural communities can also use written letters to document their needs and demands. While having the potential to act as a tool of advocacy, letters depend on the capacities of water-users to spend time writing, formulating, and sending these documents to the state. From my observations living and working in rural parts of East Africa, this effort is difficult for water-users who are often women. In many African contexts, women are responsible for collecting water for their households’ daily consumption along with performing many other tasks associated with childcare, farming, and taking care of domestic animals. Letter writing, while potentially offering them a powerful tool to demand services from the state, depends on their ability to write and send formal documents to district offices. Since many rural water users lack the resources and time to write letters demanding for services, it is important that technicians frequently visit communities in order to familiarize themselves with the conditions of water infrastructure and local institutions. Regular site visits to communities are particularly important in a context where civil servants work in areas of the country with which they are sometimes unfamiliar. In order to minimize potential conflicts of interest, employees of district water and sanitation departments move to places where they are not from. As a result, they are often unfamiliar with local cultural practices and social institutions. This was made apparent to me in Maji Safi where civil servants often described the differences between where they were stationed and their home communities. For example, one technician described his home region as being ‘more development’ compared to the district where he worked. As he explained: “They don’t like development here, they are very into their traditions”. When I asked him which traditions he was referring to, he responded vaguely by 48 suggesting that in general “they just don’t like to pay for anything, including water.” Similarly in Ngare Sidai, civil servants described the local community in the district as distinct compared to communities in the rest of Tanzania. As one civil servant suggested: “When I go to some areas of the district, I feel like I’m not even in Tanzania. It’s like I’m in another time.” When I asked him why, he suggested: “People just live differently here, it’s like another country.” These comments reflect the lack of social and cultural familiarity between civil servants and the communities they serve. Writing letters and establishing formal lines of communication do not necessarily allow civil servants to familiarize themselves with the customary institutions that often manage infrastructure and water resources. Understanding the dynamics of these customary institutions is particularly important to addressing issues of maintenance. While COWSOs and VWCs are mandated to address issues of maintenance, it is not unusual, I suspect, for rural communities to draw on customary institutions to raise funds to repair water projects. This certainly was the case in Ngare Sidai, where age-sets collected livestock from water users to sell and pay for problems with boreholes. As mentioned above, age-sets divide men according to their age and take responsibility for managing communally owned resources, including water. In Ngare Sidai, like most pastoral communities in the region, men between the ages of fifteen and eighteen organize finances for repairing infrastructure by collecting livestock from homesteads according to the number of livestock families owned and are able to provide. While women also contribute to the finances needed to pay for water infrastructure and repairs, they have less control over their family’s cows, sheep, and goats. As a result, they contribute with their limited financial earnings from selling beadwork and milk. According to many community-based NGOs, these age-set institutions and community-financial schemes help finance infrastructure repairs even though they did not follow 49 the institutional structure of a COWSO. Understanding how these institutions work and how they can be supported as an alternative to COWSOs can help create closer institutional links between civil servants and communities. Maintenance and Monitoring: The Symbolic Capital of New Infrastructure This next section looks at how civil servants fulfill their responsibility to maintain and monitor existing infrastructure. As I show, the policy and cultural environment in which employees of district water and sanitation departments work cause them to privilege building new infrastructure rather than rehabilitating existing projects. This emphasis on ‘the new’ is related to the ‘symbolic capital’ that infrastructure provides to state and non-state institutions. By ‘symbolic capital’, I am suggesting that infrastructure works as a symbol of the state to rural communities that represents the state’s presence in rural areas. Infrastructure is a source of ‘symbolic capital’ because it has more power to represent the state’s efforts to supply services to communities than maintaining projects. While water point mapping exercises could provide civil servants with data to mobilize political support for maintaining infrastructure, the information they gather is used to inform higher levels of the state only. As a result, civil servants experience frustrations with water point mapping, which they perceive as simplifying complex issues involved in monitoring and managing infrastructure. As the engineer in Maji Safi described to me in detail, at least five projects in his district needed government funds in order to be rehabilitated. Despite knowing the number and nature of each water point that was not functioning in the district, he had not completed a single major repair since being transferred to the district as the head engineer in 2014. He, like so many civil servants I met, talked about his inability to address major infrastructure repairs in relation to the 50 policy environment in Tanzania. This policy environment was described in the Historical Context section above, but is worth reiterating here in order to understand how civil servants’ understand their roles and responsibilities in relation to the shift from a supply- to a demand- based approach to managing infrastructure. As was outlined, water service provisioning in Tanzania is now de-centralized, which means that once the state builds new infrastructure, communities must then manage water projects through COWSOs. This places local government officials in a situation where they feel more responsible for ensuring infrastructure is built than maintained. As one regional water engineer explained to me: “Why would we maintain projects? There is no money for this. The national government already spent money on maintenance in the 1990s, but now we are building new projects.” The privileging of new projects is also apparent in the WSDP I program, which requires district offices to initiate water projects that service at least ten villages. This mandate has an effect on the day-to-day routines of the district water and sanitation departments in both Maji Safi and Ngare Sidai. Rather than working with communities, DWEs coordinate activities with consultants and contractors by regularly visiting new projects and by managing their contracts. For example, in Maji Safi, the DWE traveled to the six projects under construction in his district once a month. Similarly, in Ngare Sidai, the DWE spent most of his time visiting sites where projects were being built or meeting with consultants and contractors in his office. Consequently, DWEs work as project managers, largely in coordination with the private sector, rather than with local communities who must become future owners of these projects through COWSOs. In addition to the WSDP I initiative, many civil servants described the ‘Big Results Now’ (BRN) initiative as overly focused on new infrastructure. The BRN initiative was launched nationwide in 2013 to improve the provision of public services. The initiative is borrowed from 51 the Malaysian model of development and focuses on six areas of the economy, including water. Much like the DfiD’s PbR program, the BRN forms part of an ongoing results-based aid (RBA)6 program in order to increase the number of Tanzanians who have access to clean water. To this end, the BRN initiative tries to expand water projects in areas where villages are not connected to dams or other sources of fresh water. This is in addition to the program’s other principal goals of rehabilitating and replacing degraded infrastructure. Through these initiatives, the BRN initiative has succeeded in improving public service provisioning in the water sector. Compared to the previous pace of increasing water access for 300,000-500,000 people per year, in 2013/14 2.36 million rural Tanzanians gained access to clean water through projects initiated through BRN (MoWI 2014). However, though acknowledging the successes of the BRN initiative, many civil servants complain that the current BRN framework places too much emphasis on increasing access by building new infrastructure, rather than maintaining existing projects7. As one regional engineer told me: The Big Results Now is a good idea, but it was too ambitious. Now, we are dealing with delays in projects and it’s taking district engineers a long time to sort out the projects. A lot of them are sitting half finished because there are not any funds to keep them going. They should have focused on rehabilitating the old projects. ‘Big Results Now.’ It sounds good, but it’s placing pressure on building new projects, rather than maintaining existing infrastructure. There has been no planning, only ambition. For this engineer, ‘Big Results Now’ translates into building new infrastructure even though the BRN initiative makes provisions for new infrastructural repairs as well. 6 Results-based aid (RBA) is a development program that disburses grants or loans in response to pre-defined results. 7 Indeed, in 2013-2014 over 8,458 new water points were constructed in Tanzania. Only 20 new points were established through extending existing programs, and 974 water points were repaired to a functioning level. 52 The focus on new infrastructure can be understood in relation to the wider cultural politics around water infrastructure. Infrastructure for delivering water in Tanzania serves many purposes beyond simply providing rural communities with access to water. New projects also provide the state, for instance, with symbolic capital. The ability of infrastructure to serve as a source of state power is illustrated by Daniel Rogger’s (2016) findings on the correlation between political campaigns and infrastructure development. As Rogger shows, the Tanzanian state builds more water projects during an election cycle than at other times. As well, more projects are built in districts where there are strong rivalries between political parties. The infrastructure that the state builds relies on high-cost technologies, meaning that projects are difficult to maintain. As these findings suggest, new projects are more politically attractive than paying the recurrent costs of project maintenance. The state’s focus on new projects is further perpetuated by ceremonies and celebrations that occur after a project has been constructed. During the ceremonies that I attended, members of government traveled to a project in a rural community, where they typically cut a ribbon or unveiled a plaque on a newly built tank or borehole. These events to celebrate new projects are followed by speeches, wherein members of government discuss the importance of paying for water and creating COWSOs to manage repairs. These ceremonies not only publicize the state’s efforts to deliver water services to communities, they also promote the current demand-based approach to managing public services in rural areas. Far less attention is devoted to efforts to rehabilitate infrastructure. For instance, roadside signs advertise new projects by listing the donors, contractors, and state institutions involved in building new projects; such signs, however, never advertise who is responsible for maintaining them. 53 The state is certainly not the only institutional actor that perpetuates the cultural politics around new infrastructure. NGOs also encounter these politics when trying to address questions of water access. Take the example of an NGO I observed while working in Maji Safi. The members of this overseas NGO travel to the area once a year to manage the six water projects the organization had originally built. Some of these projects were not being maintained. While the traveling members of the NGO wanted to direct funds for maintenance, their overseas donors and patrons refused. When I asked why this was the case, one member of the NGO explained: We have multiple donors for our projects. We try to fundraise as much money as we can. So, if one donor builds a project, the other donor doesn’t want to maintain a project they didn’t build. They want their own projects. They want to be able to say ‘we built a borehole that provides water for 3,000 Tanzanians. As these remarks suggest, new infrastructure serves as a way for NGOs to demonstrate their capacity to deliver services to rural Tanzanians. Maintaining infrastructure, by contrast, leads to complex propriety issues between donors. These issues of ownership make investing in new infrastructure more appealing to overseas charitable institutions, acting as a visible symbol of their efforts to support rural communities. The lack of attention paid to repairing infrastructure means that civil servants depend on COWSOs to maintain projects to a much greater extent than the current policy directs them to. As mentioned above, the state is responsible for addressing major repairs. The problem, however, is that there is no existing definition of what counts as a major repair. As a result, employees of district water and sanitation departments have the power to determine when they are responsible for assisting COWSOs. Often, civil servants argue that COWSOs are responsible for most of the costs of maintaining infrastructure, asserting that these institutions have the money to address these issues. As one civil servant told me: “The contractors are responsible for 54 repairs up to 2 years after they finish building the project. This gives COWSOs 2 years to collect money. They have money”. While many contractors are responsible for repairs on water projects for 6 months to 2 years after the project has been constructed, COWSOs are sometimes unaware of the contractual arrangements made between local government and private companies. This was the case with one COWSO I visited in the Coastal Region, where the COWSO did not know which private contractor built the project they managed and whether they were still under contract to address breakdowns. This was problematic, since one project contained a borehole that failed to pump water into nearby tanks. While COWSOs are sometimes able to secure funds from water-users to repair infrastructure, this is overwhelmingly not the case in many rural areas. This deferral of responsibility to COWSOs became apparent in the dozens of conversations I had with civil servants over the course of my research: when I mentioned maintenance to a civil servant, he or she almost always began discussing COWSOs. This was true even when I asked about major repairs. COWSOs and maintenance are terms that are used almost synonymously in discussions concerning sustainability. Furthermore, the degree to which the state depends on COWSOs to manage new infrastructure is apparent in monitoring and evaluation questionnaires that I saw in Maji Safi. In one questionnaire produced by PO-RALG, COWSOs are graded and ranked according to whether they have a constitution, whether they hold regular meetings, and whether their leadership consists of men and women. The PO-RALG template also evaluates COWSOs on whether and how they plan to operate and maintain water projects, and whether they have funds in their accounts for regular repairs. There are also columns that address issues of problem solving, budgeting, and reporting back to the district. Importantly, this monitoring and evaluation questionnaire is relatively simple compared to the guidelines for reporting set by a donor. 55 The document that was created by a donor to monitor the status of COWSOs was far more elaborate, with many more columns and conditions for success. To provide just a few examples, this document asks whether COWSOs had a ‘vision’ and a ‘mission’ – as in, do they have a plan for expanding existing projects and ensuring their sustainability. The document also assesses whether COWSOs have an effective ‘organizational structure,’ including members with defined leadership roles and clear procedures for making decisions. In terms of technical knowledge, COWSOs are asked whether or not they have technical staff and if they actively oversee water system operations. The document also asks about the frequency with which COWSOs report back to districts, and if they seek out opportunities to participate in training related to technical and governance issues. In terms of financial management, the document asks if water points are metered, if they follow procurement procedures for hiring contractors, and if COWSO staff follow set protocols for ensuring these procedures are transparent and strictly followed. There are also columns that address issues of community involvement by asking if ‘water user satisfaction’ is consistently tracked and analyzed through processes for collecting and addressing formal complaints, such as through a community outreach program. These different demands were assessed and then ranked using a point system, with the result used to determine whether COWSOs are ‘minimally operational’, ‘partially operational’, or ‘fully operational’. Very few COWSOs are ‘fully operational’. Despite being relatively absent in rural areas, civil servants presume that these institutions are responsible for maintaining infrastructure. This incongruity is not the result of a lack of willingness on the part of civil servants to address issues of major repairs, but rather the result of a lack of resources. These resources are not just financial, but also political and cultural, as described above. Civil servants work within a political and policy environment where new infrastructure is valued as a source of symbolic 56 capital, not just as a resource that provides communities with access to water. The cultural value assigned to new projects makes it difficult for DWEs to advocate maintaining existing projects. As the DWE in Maji Safi told me: “Unless donors say: ‘you can only use this money for maintenance’, I guarantee that funds will be used to build new infrastructure. Somewhere along the way, politics get involved and you find yourself building another project.” As this remark suggests, DWEs feel subject to the political pull of building new projects, even though they are aware that maintaining existing infrastructure could better service rural communities. Water point mapping could provide the information needed to effectively advocate a change in focus within local government. Water point mapping exercises require districts to report on the number of water points that are functioning within the district on a monthly basis to the MoWI. These data are collected by ward governments and sent to district officials, who then forward their findings to the MoWI. Once the MoWI has collected the data, they circulate within higher levels of the state, rather than being shared with districts to inform their programs and progress. For example, water point mapping could tell the district how they compare to other districts in the region, or could be used to show the percentage of water points that work or that need to be rehabilitated. While DWEs potentially know this information from collecting data and sharing information with other DWEs in their regions, the concentration of data provided from water point mapping exercises throughout the region could be used a as representational tool to advocate for more funds for rehabilitating projects. By representational tool, I am suggesting that not unlike a report, the data from water point mapping can be used to create a picture of the problems with water infrastructure in districts. Water point mapping is potentially an affective tool for districts to understand how well they are maintaining water points compared to other districts in the country. However, through water point mapping exercises, data flows upwards to 57 inform the MoWI, meaning that it is not used to evaluate the status of water access by local government. Currently, the data are merely a monitoring tool for the MoWI. As a monitoring tool that informs the MoWI rather than local governments, water point mapping is sometimes viewed by DWEs as being redundant exercises. DWEs have less formal strategies for gathering information on the status of water infrastructure. For example, the DWE in Maji Safi frequently stopped people he saw on the side of the road to ask them questions about water projects in their villages. These efforts to monitor infrastructure were embedded in their field-site visits to new projects. Relative to information gathered from these conversations, water point mapping exercises require DWEs to synthetize complex problems into numbers. As the DWE in Maji Safi told me: It’s difficult to count water points. At any one point the status of infrastructure can change, yet we have to judge what counts as a functioning or non- functioning water point. So if it works for two weeks and then doesn’t work for the next two weeks, does it function? If there isn’t any water because people have put in private connections to their homes, draining public supplies, does infrastructure not function? Sometimes private users cause infrastructure to break down at public water points. So, it’s a problem with project design, not a problem with maintenance. It looks like a problem with maintenance when we map out water points, but really it’s not. Projects aren’t being used for what they were built for – public use. Pointing to the template for reporting water point functionality, the DWE added: “Anyways, we have to fill out these forms because they want to know numbers. They want us to give them simple answers to complex problems” (Emphasis added). As these comments suggest, the DWE of Maji Safi saw data collection as being a task that informed other institutions, offering simplified representations of more complex problems related to infrastructure. Water point mapping can also be used as a representational tool for districts in their efforts to appease higher levels of the state. In Maji Safi, this was not the case. The DWE there tried and struggled to collect accurate data. Nevertheless, water point mapping exercises provide 58 district offices with a representational tool that can be used to also strategically disguise the challenges they are facing in their districts. For example, in Ngare Sidai, the DWE reported that almost all water points in his district are functional, while through conversations I had with village governments I found that at least one-third to one-half of water points needed to be rehabilitated. This discrepancy between what the DWE reported and what I gathered from conversations with village governments can be understood through Pritchett et. al’s (2014) conceptualization of ‘isomorphic imagery’.8 Pritchett et al. (2014) use this term to describe how governments use representational tools, such as water point mapping exercises, in order to imitate a ‘well-functioning state’. In the case of water point mapping, districts can report on the number of functioning points in order to fulfill their obligations to the MoWI without accurately informing the state about the conditions of water infrastructure in their districts. The MoWI is currently using data verification exercises to identify misreporting. However, the institutional arrangements and flows of information will continue to act, I suspect, as a disincentive for districts to engage in data collection activities. If DWEs could use the data from water point mapping exercises to bolster their case in discussions over budgeting within the district, then perhaps they would be more motivated to accurately report on their progress and findings from monthly updates to the MoWI and DED. Conclusion This report has looked at the institutional, political, and bureaucratic environment in which civil servants work in order to understand how they address issues of water access in rural Tanzania. I have discussed the formal state-procedures around budgeting, coordinating field- 8See Holvoet et al. 2015 for a description of how village governments participate in ‘isomorphic imagery’ in their efforts to create annual budgets. 59 visits, and managing projects, and emphasized how these processes insulate civil servants from the communities they serve. This poses a problem under the current policy approach to water delivery in the country, wherein local government is charged with working with institutions to ensure rural villages have access to clean water. While COWSOs are primarily responsible for minor repairs, districts are responsible for major rehabilitation projects. As the research has detailed, however, many projects are not being adequately maintained. This is the result of a lack of coordination between institutions at the local level. In many places, COWSOs are supposed to exist but often do not. Even where they do, civil servants pay insufficient attention to maintaining existing infrastructure and working with these institutions. Rather than seeing this discrepancy as a ‘failure’ on the part of civil servants, I have discussed how the conditions of their workplaces and the policy environment under which they operate deter them from working in communities to address the maintenance of water infrastructure. The factors that act as barriers between local government and rural communities do not relate to funding alone, but included challenges posed by the broader institutional environment in which civil servants work. Despite efforts to decentralize service provisioning in recent years in Tanzania, civil servants remain confined – both physical and institutionally – to the district offices where they work. In these offices, many of their day-to-day obligations are to other state employees who work within the district or at the regional and even national level. Budgetary procedures, as I discussed, as well as water point mapping exercises, offer examples of how civil servants’ attention becomes directed to the objectives of institutions that are horizontally and vertically located within the state. Moreover, I have discussed how civil servants work more closely with the private sector in their efforts to manage new projects. Their roles and 60 responsibilities in managing these relationship mean that they pay insufficient attention to community institutions and the local dynamics that surround water infrastructure. The current demand-based approach to water service delivery only reinforces the separation between the state and local communities. The expectation that communities will come together to request greater water resources assumes that members of rural communities have the capacity and linkages to the state to make their voices heard. From my experiences in rural areas, women are the primary users of water points—they are the ones in their households responsible for gathering water for drinking, bathing, and cooking. Women are also responsible for a number of tasks that address the immediate needs of their households. These responsibilities undoubtedly draw them away from being able to travel and actively engage with civil servants at the district level. Furthermore, different literacy levels between women and men in rural Tanzania makes the burden of having to write letters to demand state services more difficult for them. Therefore, one should not assume that those who access water most directly, mainly women, have the capacity to demand changes in water provision from the state. Yet, civil servants use the logic of demand-based service provisioning to wait for communities to report problems with water access, rather than pre-empting problems by maintaining infrastructure in rural villages. This disjuncture between reality and expectations creates a gulf that intensifies the institutional and bureaucratic barriers that already exist between civil servants and rural water users. Key questions need to be asked to overcome the barriers that block water access in rural areas: 1. How can civil servants and communities delineate their respective roles and responsibilities in coordination with one another? 61 2. How can clearer and more accessible lines of communication be established between civil servants and community-based institutions? 3. How can the roles and responsibilities of district water and sanitation departments be increased to ensure they address issues of maintenance? My approach to addressing these questions has involved considering how everyday activities, on a small-scale, can be altered to ensure that civil servants are given adequate tools and support to meet their responsibilities and feel engaged by their jobs. My propositions for meeting these objectives are as follows: Decentralize decision making within district water and sanitation departments by allocating specific responsibilities to technicians, so that they can complete tasks independently of the DWE. Station technicians at the division level where they have more regular and frequent contact with rural communities and their leadership without having to go through tedious bureaucratic procedures. Provide technicians with financial incentives to work at the division level by redirecting the costs of per diems and travel required for field-visits. Include COWSOs in discussions over budgets with district water and sanitation departments. Facilitate regular meetings by having a member of district water and sanitation departments attend quarterly ward meetings, which COWSOs also attend. Reverse the lines of reporting so that information not only flows upward to higher levels of government, but also downward to district water and sanitation departments and community- based institutions. Celebrate Maintenance by publically advertising efforts to rehabilitate existing projects, such as placing signs at key locations that inform water users who to contact if infrastructure breaks down. 62 These suggestions are minor measures to address larger problems and should therefore be considered partial solutions to a much wider set of issues described in this report. Most especially, they will not fully resolve the gulf between communities and the state that exists under the current demand-driven approach to service delivery. This approach, while potentially empowering communities to act as service providers, also provides opportunities for the state to deflect their roles and responsibilities to rural Tanzanians. To mitigate this risk, one must reverse the ‘policy narrative’ around water service provisioning. By policy narrative, I refer to Emery Roe’s (1994) description of policies being narratives, not unlike a story with a beginning, middle, and end. The end is a set of problems that will ensue without some form of external intervention. Currently in the Tanzanian water sector, policy narratives stress the importance of communities as the solution to current problems in the water sector, particularly around issues of maintenance. This focus on communities as service providers seeks to push forward a larger agenda associated with decentralization and participatory development in Tanzania, which has emerged since the 1990s (Green 2014). Rather than viewing decentralization and participatory development as being the transference of rights and responsibilities to communities, more creative ways need to be considered to establish closer links between communities and the state. Civil servants in local government can be empowered to create these links if the state provides them with the political and institutional support needed to focus their activities and obligations downwards to rural water users, rather upwards to the central state. 63 Bibliography Cleaver, Frances and Anna Toner 2006 The Evolution of Community Water Governance in Uchira, Tanzania: The Implications for Equality of Access, Sustainability and Effectiveness. NARF Natural Resources Forum 30(3): 207–218. Coville, Aidan and Daniel Rogger 2015 Evaluation Concept Not Template: Building a Supportive Environment for Operation and Maintenance in the Tanzanian Rural Water Supply Subsector: Evaluating the Payment by Results (PbR) scheme. Washington: World Bank. Coville, Aidan 2016 Understanding the Drivers of rural water maintenance in Tanzania. Development Impact Evaluation (DIME). Washington, DC: World Bank. Ferguson, James and Akhil Gupta 2002 Spatializing States: Toward an Ethnography of Neoliberal Governmentality. American Ethnologist 29(4): 981. Goldin, J., C. Sneddon and L.M Harris 2013 Introduction: interrogating hegemonic discourses in water governance. In Contemporary Water Governance in the Global South: Scarcity, marketization and participation, eds. M. Harris, J. Goldin, and C. Sneddon, NYC: Routledge. Pp. 1-11. Government of Tanzania 2007 Ministry of Water, Water Sector Development Programme (2006-2025). Programme Implementation Manual. Dar es Salaam. Grover, Velma I. and Gail Krantzberg 2013 Water Co-Management. Boca Raton, FL: CRC Press, Taylor & Francis Group. http://dx.doi.org/10.1201/b14591, accessed September 26, 2016. Greene, Maia 2014 The Development State: Aid, Culture & Civil Society in Tanzania. Rochester NY: James Currey Press. Gupta, Akhil 2012 Red Tape: Bureaucracy, Structural Violence, and Poverty in India. Durham, N.C.: Duke University Press. http://public.eblib.com/choice/publicfullrecord.aspx?p=1173034, accessed October 1, 2016. Holvoet, Nathalie, Liesbeth Inberg, Katrien Van Aelst, Elisabeth Lulu and Yona Matekere 2015 Local Water Sector Governance in Tanzania: Mapping Monitoring and evaluation actors, activities, and use in two villages of Mzumbe Ward (Mvomero District, Morogoro Region). 64 Institute of Development Policy and Management. Preliminary Findings. University of Antwerp: Belgium. Hull, Matthew 2012 Government of Paper: The Materiality of Bureaucracy in Urban Pakistan. Berkeley: University California Press. Harper, Richard 1997 Inside the IMF: Documents, Technology and Organisational Action. New York: Routledge. Jiménez, Alejandro and Agustí Pérez-Foguet 2010 Building the Role of Local Government Authorities towards the Achievement of the Human Right to Water in Rural Tanzania. NARF Natural Resources Forum 34(2): 93–105. Koppen, B. C. P. van, Mark Giordano and John Butterworth 2007 Community-Based Water Law and Water Resource Management Reform in Developing Countries. Wallingford, UK; Cambridge, MA: CABI. Mombe, J. 2016 Interview. Dar es Salaam, Tanzania. April 7th, 2016. MoWI 2014 Water. Available from: http://www.pdb.go.tz/documents/Achievement/Tanzania%20BRN%20AR%20Popular%20Versi on%20-%20Water%20FA%20A3.pdf. Accessed: July 2016. Pahl-Wostl, Claudia 2015 Water Governance in the Face of Global Change: From Understanding to Transformation. http://public.eblib.com/choice/publicfullrecord.aspx?p=4178439, accessed September 26, 2016. Perret, S, Stefano Farolfi and Rashid M Hassan 2006 Water Governance for Sustainable Development. London; Sterling, VA: Earthscan. http://site.ebrary.com/id/10566958, accessed September 26, 2016. Pritchett L., Woolcok, M. and Andrews, M. 2010 Capability Traps? The Mechanisms of persistent implementation failure. Working paper 234. Washington, D.C.: Center. Sharma, Aradhana and Akhil Gupta 2006 Introduction: Rethinking Theories of the State in an Age of Globalization. In The Anthropology of the State: A Reader. Eds. Sharma Aradhana and Akhil Gupta. Malden MA: Blackwell Publishing. Pp. 1-43. Sokile, C.S, J.J Kashaigili and R.M.J Kadigi 2003 Towards an Integrated Water Resource Management in Tanzania: The Role of 65 Appropriate Institutional Framework in Rufiji Basin. JPCE Physics and Chemistry of the Earth 28(20): 1015–1023.. Richards, Tony et al. 2008 Water Supply and Sanitation Sector Reforms in Kenya, Tanzania, Uganda, and Zambia: Challenges and Lessons Learned. Roe, Emery 1994 Narrative Policy Analysis : Theory and Practice. Durham: Duke University Press. Rogger, Daniel 2016 How Politics Influence Water Infrastructure Provision. Development Impact Evaluation (DIME). Washington, DC: World Bank. Therkildsen, Ole 1988 Watering White Elephants?: Lessons from Donor Funded Planning and Implementation of Rural Water Supplies in Tanzania. Uppsala: Scandinavian Institute of African Studies. Weber, Max 1963 Max Weber: Selections from His Work. New York: Crowell. 66