2.1. Study area

The systems provided water to around 60 % of the population from 10 rural settlements in the municipality of Riofrío, Valle del Cauca, Colombia. Figure 1 shows the study area and the location of the settlements served by the WSSS studied. The territory is mountainous located on the Andes range, with altitudes between 1000 and 1900 m (Corporación Autónoma Regional del Valle del Cauca [CVC], 2012). A basic socioeconomic characterization was carried out within this research. Systems" subscribers were families with an average of three members. The main demographic characteristics were low fecundity, a higher male population, and the preponderance of people about 40 years. Regarding education, access decreased from 100% coverage for children enrolled in primary to 22 % for the young between 18 to 24 years. Only 16% of the population completed secondary school. 17% of homes were built with inadequate materials, mainly lack of floor, and 95% of households had access to improved sanitation. People's livelihoods were dependent on small scale coffee farming and livestock, which provided income levels generally below the national legal monthly minimum wage (246 USD in 2013, considering an exchange range of 2400 COP/ 1 USD). A more detailed socioeconomic characterization of the studied area can be found in ^{Dominguez (2014}).

Figure 1 Study area and location of the settlements served by the WSSS 

The study area was selected because a Non-Governmental Organization (NGO), the Departmental Committee of Coffee Growers (DCC), developed a project with interventions on natural resources conservation and farmers" wellbeing in this region. The study of WSSS was considered an integral part of the wellbeing of these communities. The selection of this location for intervention by DCC considered several criteria, including: area, proportion of coffee to the area, number of coffee farms, and community acceptance of the project. However, there were no special criteria for selecting WSSS, since the four studied systems were all the collective systems in the study area.

2.2. Data collection and analysis

3.1. Managerial aspects

The studied collective systems covered approximately 60 % of the inhabitants in the region. The uncovered population had individual systems. This lack of universal coverage is typical of rural regions around the world, where providing adequate water supplies to small, remote and scattered rural communities is a challenge not easily met (^{Swatuk & Kgomotso, 2007}). Despite the global progress with the MDGS from 2000 to 2015, the proportion of the population with access to improved sources in the rural areas remain substantially lower, compared to the urban population (^{Bain, Wright et al., 2014}; WHO & Unicef, 2014). In Colombia, low coverage of water in rural areas is partly explained by the low capacities for collecting information, planning, and decision-making in municipalities with low and scattered population, and that receive few transfers from the national government. Furthermore, local governments tend to prioritize water projects in urban areas, due to technical, and political reasons (Asociación Colombiana de Ingeniería Sanitaria y Ambiental [^{Acodal], 2014}; Departamento Nacional de Planeación [^{DNP], 2014}).

The number of subscribers in the systems varied in a range from 16 to 560. In all systems there was some arrangement of community organization responsible for the administration, in line with the general managerial tendency of water provision in rural regions worldwide (^{Mahmud et al., 2007}). In Colombia, it is estimated that around 90% from all the water providers in rural areas are community-managed (^{DNP, 2014}), representing about of 12,000 organizations supplying water to approximately seven million people (^{Bernal et al., 2014}; ^{Smits et al., 2013}). WSS₁ was the only system where the small number of users (16) and difficulties between them to agree on managerial issues (e.g. tariffs) prevented them to constitute a formal organization.

The case of WSS₁ is consistent with the challenges highlighted by van den Broek and Brown (2015) for community management, arguing that rural communities are heterogeneous, there are individualism, differences in wealth, gender, etc. that prevent them to organize effectively around collective action.

In general, the systems fulfilled some legal requirements such as having Statutes and Internal Rules for the Board. However, none of these organizations was registered with the regulatory agency of the State for public services. The managers argued fear to be privatized as the reason to avoid fulfilling this requisite (formalization). The requirement to force community-based organizations to fulfil the same standards of large operators represents a condition of inequality and is one of the factors that prevent them to formalize (^{Dupuits & Bernal, 2015}). In Colombia, it was estimated that by 2013 only 13% of the communal organizations were registered with the government (DNP, 2014).

WSS₃ and WSS₄, two from the four systems, also had permissions granted by the Environmental Authority to abstract water, and paid water fees. This case reflects the situation in Colombia, where only 33 % rural providers have the legal right to abstract water (DNP, 2014), and fits the theory of common pool resources which argues users are more likely to follow rules formulated by themselves, in contrast to those that come from government agencies or external elites (^{Ostrom, 2000}).

Table 1 shows some financial aspects from the organizations. Concerning to financial aspects, tariffs were participatively set in annual system users" assemblies. The tariffs set were low and flat except for WSS4, which had differential rates. In this system, when severe damages occurred, resources were obtained from recovery of the default rate. The organization made investments, including replacing pipes, building storage tanks, and improving the office. WSS3 managers believed the resources from tariffs were sufficient for O&M, but scant if major repairs or expansions were needed. To address the lack of resources, the board carried out O&M activities involving users. For example, gatherings to plant trees at the headwaters, using trees from local species provided by subscribers from their own farms were developed. Additionally, contributions in money and labour were requested to customers when damages occurred. In WSS2, managers believed the money raised was enough because no personnel were hired. Resources from tariffs were only for small investments and repairs. In WSSi there were no tariffs. Therefore, when damages occurred, conflicts arose among users because there were no mechanisms or resources to deal with problems. Lack of resources for O&M has been identified as one of the main problems that can lead to failure of community-managed systems (van den Broek & Brown, 2015). Although the Colombian State recognises subsidies on the water tariffs for the rural population in the lower strata (one, two and three from six categories), these organizations ignored the procedures to materialize access to these resources and translate this benefit to the users.

Table 1 Financial aspects of community organizations 

¹Values for 2012; N.A.: Not Applicable

Source: Own elaboration

Financial aspects in these community organizations show that the abilities and resources among the committees were different, being WSS₄ the most organized system with staff, tariff structures, investments, and revenues. In contrast, WSS₁ had no organization or resources available for undertaking the most basic tasks.

Except for WSS4, systems did not obtain enough resources to ensure compliance with legal, financial, technical and managerial standards of service provision. Thus, management of water systems is particularly challenging in rural settings since typically resources are insufficient (^{Strauch & Almedom, 2011}).

Most systems had an inadequate tariff structure, insufficient to cover O&M costs. The contribution of users to at least the full cost of O&M through fees is widely recognized as a key factor for sustainable water services (^{Mandara et al., 2013}). The common pool resources theory indicates that the establishment of a fair system in which user inputs are equivalent to the benefits they obtain contribute to system sustainability. In contrast, sustainability is at risk if some users get higher benefits, paying few costs, which could get some users unwilling to follow the system rules (^{Ostrom, 2000}). However, in some of the studied systems poverty, seasonality and unreliability of income, make difficult for people to contribute financially. Even in developed countries, limited revenues due to low population densities, poverty, geographically large service areas, and rate structures which fail to include the full economic costs of the systems make small rural supplies to provide sub-optimal services (^{Wedgworth et al., 2014}).

Table 2 presents information related to staff availability and their training. WSS₁ did not have a caretaker. One user usually performed system labours without support from others. In WSS₂, in the absence of caretaker, the organization established three groups of nine users to perform monthly maintenance tasks. Occasional and emergency repairs were also made under this modality. In WSS₃, a caretaker was hired under a salary that failed to meet legal standards. The organization was interested in providing training to the caretaker, and to be certified on plumbing by the National Learning Service. However, the caretaker was illiterate. In WSS₄, the secretary and caretaker had higher levels of education compared to other systems" collaborators. Additionally, they were the only staff from the researched systems, who received the current legal minimum wage and benefits. In all the systems, funds were not available to train personnel. This lack of capacity development may be a possible factor influencing the weaknesses mentioned on service provision, since training is a fundamental component of community management (^{Harvey & Reed, 2007}). However, it has been reported in other rural contexts that communities have received managerial and technical responsibilities without being trained (^{Mandara et al., 2013}).

Table 2 Personnel and capacities of community organizations 

N.A.: Not Applicable

Source: Own elaboration

3.2. Technical aspects

The studied systems where gravity-piped systems, capturing water from protected springs, through a basic collective infrastructure that delivered water to household connections. These systems will be likely considered "improved" according to WHO (2012a).

Inspections to the systems helped to identify technical issues related to the infrastructure (Table 3). For most of these issues, water managers were aware and developed solutions, such as headwaters protection and artisanal repairs. However, these repairs were made using only common sense and despite the fact, they allowed continuous operation, they were highly vulnerable to new failures. Other situations were not seen by community organizations as areas for improvement. For example, the need of fittings and valves for optimum hydraulic operation, and the inadequate functioning of grit chambers. All these systems had more than 40 years use, and after they were built, builders left and the communities faced increasing demands for water with the same infrastructure capacity, and deterioration of units and pipes, due to the lack of resources for investment.

Table 3 Technical issues and local solutions 

Source: Own elaboration

3.3. Compliance with microbial water quality standards

Figure 2 shows ttc levels according to system based on who categories (who, 2012b). None of the systems met the criteria for microbiologically safe water of 0 cfu/100ml. These results suggest that even when the subscribers of these systems will be considered people with access to an "improved source", they not necessarily had access to safe water according to who standards (WHO, 2012a). However, currently there are criticisms about the adequacy of the "improved/unimproved" indicator that misses the water quality dimension of access to water (^{Bain, Wright, et al., 2014}; ^{Grady, Kipkorir, Nguyen, & Blatchley, 2015}; ^{Heitzinger et al., 2015}; who & Unicef, 2014); and on the real possibilities of rural systems to meet the 0 CFU/100ml who standard (Domínguez, Torres-López, Restrepo-Tarquino, Paterson, & Gowing, 2016; ^{Jensen, Jayasinghe, Van Der Hoek, Cairncross, & Dalsgaard, 2004}; ^{Parker et al., 2010}). Furthermore, the United Nations post-2015 agenda proposes changing the "improved/unimproved" indicator for an indicator of "safely managed drinking water services", which includes: use of a water source at the household that reliably delivers enough water to meet domestic needs; fulfil who guideline values for E. coli, arsenic and fluoride; and have a risk management plan (who & Unicef, 2014).

Figure 2 Percentage of water samples according to TTC level and WSS 

On the other hand, the results of the study highlight the urban-rural gap in access to safe water in Colombia, where data for 2007-2012 shows that only 41% of samples analysed from rural areas fulfil the microbiological standard compared to 91 % of urban samples (DNP, 2014). These data are from water quality surveillance by sanitary authorities that only covers 57% of the rural population (^{Instituto Nacional de Salud [INS], 2014}; ^{Wright et al., 2014}). This is consistent with systematic reviews that show that rural water supplies in low and middle income countries worldwide present higher levels of contamination compared to urban water supplies (^{Bain, Cronk, Hossain, et al., 2014}).

3.4. Perceived difficulties for service provision

Difficulties for service provision expressed by the managers included in WSS₁: lack of agreement among users to create an organization or join an existent; lack of participation of users in O&M; and lack of all kind of resources to facilitate service provision. In WSS₂, leaders identified as main issues the conflicts among users and between users and the Board due to insufficient water in dry season that demanded shifts for provision. In WSS₃, problems were related to: sabotage of systems by some users through valve manipulation during intense dry seasons; crops under agrochemical use upstream the intake; and increase in turbidity in rainy season. In WSS₄ difficulties identified were: dispute with the municipal government to control the system; conflicts with a forestry company, which owned lands at the headwaters; objection of some users to water chlorination; and community complaints for service outages in rainy season due to loss of water quality. Some of these aspects could be categorized among the factors raised in the common pool resources theory, as factors that influence the success of collective action (^{Ostrom, 2000}).

3.5. Post-construction support

There was no evidence of investments by the municipal government in any of the systems with the funds local governments in Colombia obtain by law from the national government to invest in water and sanitation. Leaders from WSS₃ expressed they eventually received support from the local government as materials for emergency repairs, as "payment for political favours". They also received government support to solve conflicts between users, when adjusting tariffs, or during dry season. In Colombia, the provision of economic resources from the government to community organizations through political patronage rather than from transparent and technical processes has been reported by other authors (^{Murtinho, Eakin, Lopez-Carr, & Hayes, 2013}). Other managers indicated support was not requested because they feared to lose control over the systems.

In general, managers wanted training in O&M, and accounting. However, the local government provided training on job skills for caretakers, but the training was offered to staff from the City Hall with no relation with the rural systems. One of the concerns was the systems" management was given to a joint public-private entity the municipal government tried to constitute. Three Municipal Agreements were issued to create this entity, but the community stopped these initiatives with mass demonstrations. In spite of this, according to the community, the threat persisted. For example, recent regulations from the Ministry of Housing, City and Territory (Decree 1077 of 2015 - [Ministerio de Vivienda, 2015]) have raised questions about whether legally organized communities are still authorized to provide public services in rural areas (^{Carrasco, 2015}).

The sanitary authority (Unidad Ejecutora de Saneamiento [UES]) had relations with WSS₃ and WSS₄, systems with the largest number of subscribers in the study, with two main activities: i) water quality monitoring twice a year; and ii) provision of disinfection units and one-year assistance. However, systems" managers felt UES made demands out of reach of small systems with precarious resources. Concerning the disinfection units, after the year of support, facilities provided to WSS₄ were working with monthly costs on chlorine around $266 US, accounting for 20% of total system costs. In contrast, the disinfection system installed in WSS was abandoned. The disinfection program may be justified on the potential health benefits that could be achieved when access increases from an improved source to an improved source with disinfection (^{Fewtrell et al., 2005}). Although, during the rainy season, water quality can be poor and disinfection can be interfered by turbidity, giving a false sense of security to the community. Therefore, training on administrative and financial aspects, plumbing, O&M, or undertake investments for the replacement of infrastructure, or the acquisition of land at the headwaters were perhaps more relevant strategies to help these organizations.

Due to the lack of support from the government, NGOs offer some assistance to communities to improve water services (^{Smits et al., 2013}). In the study area, the four systems were built by DCC in the 1970s during the coffee boom, being crucial to allow first-time access to water. However, from 1994 DCC reduced investments in water supply with their own funds due to a change on the service provision regulations in Colombia (Ley 142/1994 - "Ley de Servicios Publicos Domicialiarios," 1994) and the coffee crisis. The current approach of DDC is similar to that most NGOs operating in different parts of the developing world have adopted, which emphasizes on capacity development and strengthening community-based organizations, instead of building infrastructure. Despite criticisms on the results from this approach, it is considered more likely to contribute to decentralization (^{Galvin & Habib, 2003}).

Another NGO in the region was administered by a forestry company with plantations in the area. This NGO led initiatives on: education, income generation, health improvement, and sustainable use of natural resources. However, there was no evidence on projects from this NGO in the region. Conversely, people identified this company as an environmental problem. In many cases, the social investments these organizations undertake aim to mitigate the rejection of local people to the perceived impacts from their operation. For instance, in Australia people tended to view plantations as providing more benefits for owners than positive outcomes for the environment or the broader community (^{Williams, 2014}).

3.6. Synthesis of the challenges faced by rural community organizations

In summary, the challenges encountered in these systems are similar to those identified in community-managed organizations around the world: aging infrastructure, difficulty in carrying out replacements, low cost recovery, inadequate human resources, poor staff training and financial management (^{Harvey & Reed, 2007}; ^{Kayser et al., 2014}; ^{Moriarty et al., 2013}; ^{Smits et al., 2013}; ^{Wedgworth et al., 2014}), lack of information for decision making (^{Roa-Garcia & Brown, 2009}; Smits et al., 2013), inadequate water treatment and water quality monitoring (Kayser et al., 2014), leakage and microbial contamination; and at a lower level, particularly in dry season in WSy, intermittent supply (^{Lee & Schwab, 2005}; Moriarty et al., 2013; Smits et al., 2013). In addition, these communities had limited contact or positive relations with the local government. Some of these characteristics were identified by Harvey and Reed (2007), as common causes for the breakdown of community-managed systems.

The community providers in this study area emerged as most water user associations in the Colombian Andes, spontaneously and absent from outside help (^{Murtinho et al., 2013}). According to ^{Smits et al. (2013}), in this country, many small municipalities have not implemented any mechanism to help small organizations in rural areas under their jurisdiction. For them, systems with more than 300 subscribers, treatment plant or pumping station typically received more frequent support. The last statement contrasted with the situation of WSy, which despite of having more than 500 subscribers and treatment plant, had conflicting relations with the government. The absence of political support and inadequate or improper use of funds are considered part of the fundamental problems that must be addressed to improve the operation of water supply systems worldwide (^{Lee & Schwab, 2005}).

This case fits claims of ^{Smits et al. (2013}) who argue post-construction support should be an integral part of community management, because the limited number of users prevent achieving economies of scale to generate the revenues to access technical and financial expertise, and these costs are high to be paid via water charges. According to ^{Moriarty et al. (2013}), this support is necessary in cases like that reported here, where first-time access has been achieved and the need is sustainable services over time, through schemes in which all aspects of service could be funded. However, the best configuration to provide this support remains unknown (^{Kayser et al., 2014}). Thus, this is an area where research needs to be developed.

The current Colombian policy for provision of drinking water in rural areas (DNP, 2014) entitles municipalities to set their support schemes, a contradiction since the same policy recognizes the low capacity of these small municipalities as one of the main causes for the precarious conditions of rural water services. Besides the low capacity and resources, some authors note that corruption at the municipal level can hinder effective post-construction support schemes (van den Broek & Brown, 2015).

It has been considered that external support can come from multi-level organizations, which may be networks of community providers that besides supporting capacity development around financial, technical and managerial aspects, can gain access to domains of public policy formulation. In Colombia, such an organization is the Association of Community Providers of Water and Sanitation Services (Aquacol), which operates in the southwest of the country. Nevertheless, Aquacol growth has been highly influenced by external support from academia and international donors, and despite its achievements in strengthening community organizations, its financial sustainability in the long term is unclear (^{Dupuits & Bernal, 2015}). Furthermore, the support provided by this type of organizations to other community-based organizations could not address necessary aspects of post-construction support in rural communities such as transfer resources for rehabilitation and extension of infrastructure.

On the other hand, the Colombian policy also promotes the operation of urban providers in rural areas and the establishment of regional schemes. This raises questions about the real efforts to strength the more than 12,000 community organizations that for decades have filled the gaps left by the State in rural areas, such as those studied here. According to ^{Dupuits & Bernal (2015}), even when community organizations are recognized under the Colombian law, public policies are not fully oriented to support them.