Key Action that requires time Decision required
Figure 1 Flow Chart for Analysis of Downtime
The general framework above highlights the key elements of a sustainable water supply. This section now looks at the three case studies as a whole to discuss some of the details involved with sustainable supplies. We will then use these results and this discussion to comment on issues and processes that should be addressed for a successful exit strategy.
We have seen that while some policy and planning issues influence the sustainability of the water supply, the main focus of the research was on the community’s ability to sustain the supply. In particular we consider how the different variables[1] affect the functioning of the water supply. It has been stated that mechanical functioning of the borehole is taken as the primary indicator of sustainability, primarily the frequency of breakage and average downtime when broken (see Box 3.1.8).
The analysis starts with this primary indicator of sustainability, and presents the links within the context of a flow chart representing the decisions and delays encountered in getting a pump repaired. This is followed by the analysis of data relating to the three categories of social factors (Section 5.1).
There are two key parameters that describe the functionality or reliability of the borehole. They are “how frequently the pump breaks down” and the “average downtime”. We can now discuss what factors affect these two elements[2].
Regarding Frequency of breakdown:- logically this should be affected by the technical factors (depth of borehole, quality of the pump manufacture, suitability of the pump to its working conditions, age, etc). Social factors will only really come in to play in terms of the pressure on the pump (the amount it is used) and the quality of the operation, maintenance and repairs. If rough handling is allowed, or children can play on the pump, one might expect it to break down more often. If no preventative maintenance is undertaken one might expect the pump to breakdown more frequently as it gets older. A poor repair might only temporarily solve the problem, and a new breakdown might occur soon after a repair.
These variables were captured in the interview responses and will be discussed below. However before doing so, let us look at the process of repair. The diagram in Figure 1 attempts to present the decision making flow for repairing a rural pump as indicated by the interviews. Surprisingly, the literature survey did not identify an equivalent diagram.
Regarding the average downtime;- the decision flow diagram below has been created to assist the discussion of the time taken to repair the pump. The decisions required each affect the time it takes to repair and it is the cumulative time taken that is captured in the downtime variable.
Using this diagram, we may now consider each decision and using the data indicate whether this factor was of significance to the formation of the total downtime. For instance – “is finance available to fund the mechanic?” will probably be strongly influenced by whether the community collects money regularly or whether it waits for the pump to break and then collects the money (as many communities do).
The first element on the diagram is the preliminary diagnosis. This will be the chief, the committee or others in the community making a first guess at what might be wrong with the pump. They then have to decide who should repair the pump. Within the sample there are a variety of possible responses. Some of the early emergency pumps have been left with the impression they should call the Government. For some pumps a local area mechanic was created, sometimes a government employee and sometimes a private specialist (e.g. car mechanic). Other pumping communities had various numbers of people trained within their community, and the case studies have differences in the levels of training these groups or individuals received.
The first decision point is whether there is suitable technical competency on hand. Realistically, for those situations where there is an “area mechanic”, the mechanic might be within another community but within a bicycle ride – so no funds are required for the initial consultation. However, where the mechanic is a long distance away, perhaps at a district centre or further, there may be a need to raise transport costs for the messenger, and this may delay the start of the process. Data on “who usually repairs pump” is presented in Figure 2, and shows that the majority of pumps are repaired internally i.e. by committee or local mechanic.
Figure 2 Frequency Distribution: Who usually repairs pump
There may also be some delay if the community mechanic(s) is away. The early projects trained men, and sometimes only one man. Men often travel to find work and may be away from the village. There may, therefore, be a difference in time delay according to whether training was given to one man, or a sub group of two or more.
If the above really affects the downtime, one would expect a difference on “who repairs the pump” between external and internal (community) agents with respect to “downtime”. There is such a difference, and the Mann Whitney test shows a statistical significance. Looking at the data in more detail reveals that downtime is significantly shorter where committees repair pumps rather than local mechanics, which confirms the potential delay in getting hold of a single person.
We can also look at the difference between training one person and several persons. This to some extent is captured by the variable “Who was trained”. The average downtimes are lower where the whole committee was trained rather than just mechanics. (The discussion on the words “committee” or “caretaker” has been taken into account).
An element of finding the mechanic and asking he/she/them to look at the broken pump, is that the community must have, or have the ability to promise, payment to the mechanic. Where communities have internal mechanics, this may be in kind or by preferential treatment at the pump head when collecting water. In the interviews there were a wide range of responses as to how a community mechanic was paid.
Where external mechanics are required, there will probably be some element of payment. In some cases within the research interviews, a few respondents claimed that they could approach local or central government to repair the pump. These few cases stated that this was the “government’s responsibility and would be undertaken for no payment”. However respondents from the same village or community also talked about having to pay.
The availability of payment was difficult to measure. It was not possible to get clear answers on how much the committee had in hand. It varied between one treasurer who claimed to have several million under his bed at home, to other committee members who described how they gave the money to a wealthy man because he would not run off with it, and that he stored it in his own livestock – to be called upon as and when required. Transparency and accountability varies from village to village.
However if payment is a critical factor in the total downtime, one would expect a significant difference to be found on the variables “Pump maintenance payment” and “Payment of quota”. The former captures whether people make regular payments or not, the implication being that if people pay regularly then the committee might have reserves readily accessible. The latter variable is whether they make a payment or not. This latter variable is weaker in its implication that there are reserves held by the committee. Neither of these variables shows significance with respect to downtime. The implication of this is that the holding of reserves does not shorten downtime significantly; the corollary being that collecting money to fund the repair process does not delay it significantly.
Having got the mechanic to the pump, they may undertake a more detailed diagnosis of the problem to determine if it is within their capability to repair the system. It is possible that the person summoned does not have the capability to repair the problem. The case studies have trained people to varying levels of capability. A government service, or an area mechanic, will have the tools and skills to tackle the more difficult problems. The community mechanics were often only trained in maintenance and simple repair. The agencies in the study are undertaking training to upgrade the technical skills of the community mechanics.
If this inability to repair the pump by the local mechanic is critical to the total downtime, one might expect a difference with downtime (Mann Whitney MW) to be shown by the variables “Length of training”, “Level of training” and possibly “Follow-up training”. MW differences show on the length of training. It might be expected that a significant difference in downtime would result from the length of training received. The longer the training the shorter the downtime. Length of training should reflect the level of skill imparted. In the cases where training was provided for longer than a week there was a significantly shorter downtime when the pump was broken. This is logical, as one would expect the more highly trained to be able to reduce the period of downtime.
Neither the variable “Level of training” nor “Follow-up training” show on the average downtime. However, the level of training only has response from the committee members and therefore does not have a very high sample of responses. Similarly Follow-up training was questioned by the agencies as a potentially unreliable variable.
If the mechanic has the capability and willingness to repair, there is still the question of spares availability. This again depends to some extent on finance. Do the community have the financial resources available to purchase spares (or have they pre-purchased and stocked them)? This part of the process is very susceptible to spares availability generally, and this leads to the discussion on private suppliers.
Reported problems with pumps were divided into six categories, and indicate that, for the programmes covered by this study, the provision of spares is not foremost in the communities mind.
Figure 3 Incidence of reported problems divided into catagories
However, one must keep in mind that the community may not be aware of problems with spares. This is born out by the breakdown of stated problems by type of respondent. It shows chiefs to be feel technical, social and organisational problems more keenly. Not surprisingly, committee members see few problems with organisation and access to spare – presumably they are responsible; however, they are more aware of shortcomings in technical competence. Again, it is not surprising that pump mechanics feel their work is hindered by poor access to spares.
A small number (49) of comments explaining delays in pump repairs were recorded mainly from those involved in the actual repair. These can be categorised:
Figure 4 Reasons for delay in repairs
Figure 4 shows that materials availability is quoted as being responsible for most delays. Within organisational issues, comments referred exclusively to problems with funds. In one agency the comments predominantly referred to “finding a mechanic”. The other feature to consider is year of installation. Mechanisms for providing access to spares appear to have been sorted out in later installations. And we should note that later installations appear to have been installed in locations with high population pressure (to be discussed below)
Figure 5 Reported problems according to year of installation
While the above follows the logic imposed by the flow diagram, the data also highlighted a few other variables that seemed critical to the downtime.
Social Conflict. Figure3 shows how most of the problems reported can be attributed to social issues and population pressure, although these are related to some extent. Many of the social complaints relate to payment (e.g. people don’t pay; quarrelling over where to keep money collected), but other frequently mentioned social issues include conflict (e.g. fights), lack of motivation (e.g. community not active), and preference for alternative water sources. Most population pressure type complaints clearly relate to too many households. Other complaints tend to be consequences of the same problem e.g. long queues, insufficient water, need more pumps etc.
This confirms the importance of achieving good social organisation, especially for collecting money. However, it also shows that there are certain social factors which can disrupt sustainable functioning of management structures, irrespective of the amount of planning, participation, and the quality of organisational structures. In particular, the data highlights social conflict, and poverty, both of which are referred to in the literature.
Installation. It can be noted that downtime varies significantly with whether people were present at the installation or not. There are significant differences with downtime (MannWhitney) with the variable “present at the installation of the pump”. A slightly higher significance was recorded for the variable “participated in installation”. This may reflect the involvement of the community in the care of the pump and this can be explored in the following sections on organisation and participation.
Of those near to pump (proximity) 82% were present at installation, compared with 90% of those far from pump. Of those near, 69% participated in installation, compared with 71% of those far. These results are not statistically conclusive.
Participation in installation can be a relatively simple act. Responses to “Participation in Installation” were divided into the following categories:
1. labour in preparation
2. labour in installation
3. materials
4. organising
5. money matters
6. transport
We see from Figure 6 that the contribution of labour and materials (sand and gravel) were the main mechanisms for involvement.
Figure 6 Participation in installation (how)
Emergency installation. The variable “emergency installation” shows a significant difference. Since this is an important part of the research it is discussed in greater detail below.
Year of installation. The age of the pump appears to influence only the average downtime but not the frequency of breakage. Shorter downtimes for pumps installed from 1995 onwards probably reflects the shift towards greater community mobilisation in later programmes, in which case repairs carried out by local mechanics tend to be carried out more quickly.
Type of pump. Two main types of pump are compared - the Volanta for the deeper pump, and the Afridev for the shallower pumps, ideally those with pumps set at less than 30 meters. Only 5 Volanta pumps were captured in the sample. In Malawi 3 Nadia shallow pumps were also observed, however, in this comparison the Nadia have been excluded. The Volanta pumps breakdown more frequently, and are down for longer. It is interesting to note that the influence on the performance indicators remains the same even when only considering pumps set below 29 meters. This suggests that there is a major problem with the Volanta pump, particularly as it is specifically designed for the deeper setting compared to the Afridev.
The interesting issue is that with the Afridev deeper pumps the downtime when broken was not significantly different. The respondents using deep pumps expressed a greater sense of pressure which corresponds to the significant rise in time waiting to draw water.
Alternative unreliable source. Access to alternative sources of surface water, even if unreliable, has a significant impact on the attention to the care and urgency of repair of the pump at the time of breakage. Where no other water source exists, the care and speed of response to breakage is greater (average downtime is shorter). This suggests that in the long term health education focusing on reinforcing the dangers to health of drinking surface water might have an impact on the downtime of the pumps.
Quality of water: The salt content of the water appears to have an influence on the downtime. The downtime when broken is also significantly longer when the pump's water is more salty, . A number of interviews showed that people used salty boreholes for other domestic needs and not for drinking. If this is so then the slow response to repairing the borehole is a logical one.
Settlement. Where the respondents claim to have settled due to the presence of the pump (n = 30 or 16%), the average downtime when broken was significantly higher. The instance of settlement due to the presence of a pump does appear to influence the operational status and sustainability of the pump negatively. Interestingly, there is no significant correlation between settlement and felt pressure or the time waiting for water. This indicates that where settlement occurs due to the presence of a functioning pump the social cohesion may not be as strong and this may reflect a weaker commitment to the care of the pump.
Community Involvement. The average downtime was lower where committee meetings were held regularly, which confirms the positive effect of organisational capability within the community. However, whether people made payments regularly or irregularly had no significant influence on pump performance, so issues of community organisation have only a weak influence on performance.
Given the above, we may now look at the data surrounding frequency of breakdown. Most respondents seem to distinguish between maintenance activities (preventative) and an actual breakage requiring repair.
Depth and Type of pump. There is a very high difference on frequency of breakdown based on depth and type of pump. In this case the comparison was made between pumps set above and below 30 metres. The deeper the pump is set the more frequent the breakage . These findings are logical and are to be expected, especially when setting Afridev pumps below 29 metres.
The depth at which it is needed to set the pump was not open to much alteration as it was found that pump and static water depths only differed on average by 1.5 metres. Where water is deeper and its withdrawal more problematic, the community appears to respond positively to the challenge. The change of pump to one more appropriate for the depth does not appear to have provided a solution, as is discussed above in the context of average downtime under type of pump.
The Volanta pumps breakdown more frequently , and are down for longer than the Afridev pumps. The snapshot given from visits also shows a higher proportion of the Volanta pumps were out of action . The general status of pumpheads at Volanta pumps is, however, higher ; this shows that care of the pumphead cannot make good technical issues pertaining to pump design. It does however, demonstrates a positive response to problems by the community. Better care of the pumphead may have been seen as a way of attracting continued external help from the implementing agency or an expression of the importance of the pump to the community. They are doing what is in their power to maintain it.
It is worthwhile noting that Pump problems tend to be more severe (hardware) where pumpheads are better kept. This runs contrary to what might be expected and has some explanation in the predominance of good care in the Volanta pumps.
It is interesting to note that the influence on the performance indicators remains the same even when only considering pumps set below 29 meters. This suggests that there is a major problem with the Volanta pump, particularly as it is specifically designed for the deeper setting compared to the Afridev.
Who repairs the pump: As discussed above this is a comparison between external and internal (community) agents. This issue has a significant influence on both of the pump performance indicators, i.e. frequency of breakage and downtime.
Pumps tend to break down more frequently where they are repaired locally as opposed to by external staff , but downtimes tend to be shorter. As indicated on the decision flow diagram (Figure 1) this seems to indicate that the quality of repairs carried out by local, non-specialised, personnel may be poorer than those carried out by more highly trained external agents. It also shows that where a community has the internal resources to maintain a pump, they can be mobilised quickly resulting in shorter downtimes. These two important issues work against each other in making up total reliability or availability.
However, where it is a local personal who carries out repairs, their ability is held in high regard. This apparent anomaly indicates that perceived ability is more strongly influenced by the ability of mechanics to get a pump repaired quickly; people tend to forget how often it breaks down. It may also reflect the social cohesion amongst communities which have the capability to carry out repairs, and a certain pride in that ability. Conversely, it may reflect a lack of interest and absence of opinion in the ability of external mechanics. This confusion explains why perceived ability has no affect on the frequency of breakdowns.
Who was trained: This represents a comparison between cases where one or a few local mechanics were trained in pump maintenance as opposed to the whole committee. A significantly better performance appears to be achieved where training is given to the whole committee rather than just to mechanics - both frequency of breakage and average downtimes are lower.
One might expect that the way the community use the pump affects how often it breaks down. However the data seems to suggest a counter-intuitive response.
Where rules are applied there is a significant difference regarding frequency of breakage and the status of the pumphead. Existence of rules appears to correspond to a higher instance of breakdown . The issue of the status of the pumphead is counter-intuitive as the existence of rules appears to correspond to the less well-kept sites . This may also reflect a retroactive action on the part of the committee given the lack of care. A related variable is Sanctions, which does not appear to influence the performance of the pump, however, there is a tendency (not significant) for the communities who do not apply sanction to have fewer breakdowns. Breakdowns may motivate the committees to apply sanctions given the more constant demand for funds.
Payment of quota. Those who say they don’t make payments have lower frequency of breakage than those who do. How can this be possible? One explanation might be that those who are socially organised to make payments also use local community mechanics who, as we have seen above, have a lower quality of work than external mechanics. This link might explain the unusual significant difference of payments with respect to frequency of breakdown.
Pressure on the water source. The time spent waiting to draw water, which is an indicator of pressure on the borehole, shows that where people have to wait on average three or more hours, the frequency of breakage is higher . The pressure on the borehole, as expressed by the community, registers a difference with regard to status of the pump at the time of the research visit. Both of these indicate, as might be expected, that the borehole is more likely to break down where it is subject to higher pressure.
Pump requested: if the pump was requested by the community or provided without prior request. There is a significant difference such that more frequent breakage occurs in those instances where the pump has been requested. However, downtime tends to be higher where the pump was not requested (not significant).
Siting, who chooses the site for the pump - the community or an external agent. The frequency of breakage is significantly higher where the community claim to have sited the pump as opposed to the drillers. This suggests that with the issue of siting the pump it is better left to the experts with only some consultation with the community.
Prior contribution. It should be noted that this only applies to one agency, and was introduced with the more recently drilled pumps. Its value as a comparative variable is, therefore, questionable in this instance. In the cases where a prior contribution was made, there is a significantly lower instance of pump breakage . The time waiting for water is also significantly lower with those respondents who reported making a prior contribution . The correlation between prior contribution and proximity to the pump may suggest their closer association with the pump and shorter distance and time to draw water. There is a close correlation between the prior contribution and participation at the installation of the pump.
Ownership. It is also important to note variables that do not show a significant difference on pump functioning. The “ownership” variable, which was a stated response of “who owns this pump?” does not show any significant difference on either downtime or frequency of breakage. One might expect that those who stated that the village or village authorities owned the pump might respond quicker to pump repairs than those who state that the government owns the pump. However the data does not support this logic. Similarly Community involvement (in the committee), the number of women on the committee and the election process for the committee all do not show as significant to pump function. This will be discussed further below.
In the above discussion we have noted two key parameters of pump functioning. Whilst reliability (the combination of “frequency of breakdown” and “downtime”) may be the best indicator of whether a pump is effectively serving a community with water, NGOs take a wider view of interventions. The primary purpose of providing a clean water supply is to improve health, so pump installations in a development context are usually included as part of wider integrated development programmes, accompanied by hygiene and sanitation promotion (and possibly other development actions).
Therefore when we come to discuss ownership and the social mobilization to undertake Village Level Operation and maintenance (VLOM), we must consider the question from two angles. On the one hand we have literature stating that VLOM is the best practice for long term sustainability, and therefore we would expect to see that VLOM results in greater availability of water. Our analysis above questions whether this is actually so. Village or community level maintenance and repairs tend to lead to a greater frequency of breakage than if an area mechanic or government team take action. The downtime is lower for local repairs (as long as spares are available). However the combined effect of higher frequency of breakage and lower downtime for a VLOM response does not actually result in a higher availability of the pump when compared with area mechanic systems or even Government responses (in general).
On the other hand we must also consider the wider agenda of social mobilisation as an objective in its own right. If the community have a greater capacity to solve problems, then they may mobilize this capacity to be applied to other problems in the village. For water programmes health and sanitation are an integrated response and the agencies in the case study undertake health and sanitation training as part of their projects.
The following analysis considers issues influencing the three categories of social factors, and examines interactions between them. The primary relationships are represented in the figure below. This figure makes a distinction between preset conditions, over which project planners have little or no control, and programme design features, and shows how each makes an impact not only on borehole performance, but also on social factors. The historical issue of settlement is presented as a secondary level programme output.
Competence was considered as a whole by looking at the variables:-
The last variable considers people’s perceptions as to the competence of their village representatives, be they committee and/or pump mechanic. We have already seen how these factors influence pump functioning. We have noted that actually the greater skill that lies in external mechanics results in a lower frequency of breakdowns. However, even though in terms of pump availability the gains of lower downtime offset the loss on frequency of breakage, in this section we are considering local capacity a desirable end state. We consider local capacity to be part of the project’s wider objectives on health, sanitation and general development. However, it is difficult to extract the relationship between local competence and health and sanitation. A full Knowledge And Practice survey was not undertaken. Some insights are given through the care of the pumphead and these are explored in a section below.
As might be expected, training issues have the greatest influence on competence, especially who was trained, and when they were trained. Where committees received training they tend to carry out repairs, and where it was mechanics who were trained, it is they who carry out repairs. Providing training for committees has resulted in a larger number of trained mechanics, which helps reduce dependence on a single mechanic, as pointed out above. Where training was given some time after installation, fewer people were trained (concentrating on mechanics rather than committees), but they tend to be better supported by more frequent agency visits.
Where it is a local person who repairs the pump, the ability of the mechanics and committee are held in higher regard. Perceived ability is also higher where longer training has been given, as might be expected.
Training is a component of project design, and reflects changes in emphasis over the years. Later installations (post 1995) have a higher number of trained mechanics, and are repaired locally, and management personnel are regarded as very able, all of which reflect a high degree of internal capacity within the community. This indicates that later initiatives concentrated on sustainability issues such as management and maintenance payments. It is interesting that follow up training appears to have no affect on technical competence, indicating that it addressed other issues (e.g. community mobilisation , payment mechanisms etc.).
People tended to settle where there were more pump mechanics, where the last agency visit was more recent, and where the local committee and mechanics are regarded as competent. This indicates that technical competence and external support are important factors in attracting people to a water point. However, there is a strong correlation between people settling and year of installation such that people tended to settle due to pumps in later years, which indicates that insecurity factors may be involved.
It is interesting that certain entry issues (community requesting the borehole, and people being present at installation) have no affect on technical competence, yet others (prior contribution, and participation in installation) do show a link. Where boreholes were requested, agency pump maintenance visits tend to be more recent. Where a borehole is requested, this may be the first stage of a longer term and closer relationship with the community. This is confirmed by the large proportion of boreholes which were requested, that received follow-up training. It is only in cases where the committee repairs pumps that a significant proportion of prior contributions were paid, and people participated in installation. Prior contributions were only introduced in later installations (post 1995), and a greater proportion of these cases are repaired by committees. Therefore, this link more probably reflects changes in emphasis of project design in later years, with a greater degree of community focus – both prior contributions, training more mechanics, and committee activity.
On-going community involvement also appears to have little affect on technical competence. Perceived ownership has no link with any indicator, and the only link with community participation in committees is that where the time since last agency visit is shorter, confirming greater agency involvement keeps the communities motivated to take part in water programmes. This supports the idea that local social mobilisation is more about the wider development goals than about pump functioning.
The organisation performance variables chosen can be divided into two principal subsets - those dealing with payment, and those dealing with organisational function. The payment subset involves:
· frequency of payment schedule
· actual payment
· sanction for non-payment.
The function subset:
· existence of water committee
· number of members
· percentage of female membership,
· community election
· frequency of meetings
· existences of rules for pump use.
There are two main factors, which appear to have most influence on community organisation performance. The first is the perceived capability of the local mechanics or committee to repair the hardware as well as software problems of the pump. Where this higher ability exists the people tend to pay more regularly and sanctions for non-payment are enforced. The implication for future exit strategies is the apparent importance of ensuring that training is given to someone within the community to manage the hardware as well as software issues, particularly in the case of deeper pumps.
The second primary influence on community organisational performance is the level of community involvement in the management. Community involvement is implicit in many of the other influential factors such as the committee election process and siting of pump. Involvement is also implicit in whether the community requested the pump or not. It is assumed that where a request was made the community has already mobilised itself to address the water problem and delegated the task of identifying support, an indication of involvement and organisation from the outset. Age of installation has a strong influence on organisation indicators; again, this represents a shift in emphasis of later programmes to focus on community mobilisation issues.
The data raises some rather controversial issues regarding community participation in the committee. Firstly, the involvement of women on committees appears to have no affect on any aspect of technical competence. The development community has developed the logic that if women are on the committee then responses to water problems will be quicker and more effective. Although the data would seem to dispute this, care must be taken to consider points raised in the literature review. Two main areas of concern are evident - token representation of women, with resistance to genuine power sharing from a range of mostly cultural beliefs, and the need for training to enable women to contribute effectively, and to address the prejudices of men. Further study would be required to investigate detailed reasons why the involvement of women has had little affect. Secondly, the fact that the election process and community involvement in water committees have almost no affect on issues of technical competence confirms that this area is independent of community participation. To a certain extent this is backed up by comments in the literature to the effect that participation issues are complex, and traditional power structures are well established.
Training which took place before or at the time of pump installation appears to result in a more positive performance on the functioning of the committee. However, training, received sometime after the installation resulted in an improved performance on the payment subset of indicators (e.g. regular payments). This may be due to the later training placing greater emphasis on the development and management of a maintenance fund by a committee structure. In contrast, the earlier training's focus appears to have been on pump mechanics and pumphead maintenance.
Turnover of caretakers. One agency interviewed reported a frequent turnover of caretakers. The most common cause being marriages or death.
The type of pump affects organisational indicators. Communities with Volanta pumps make more regular payments and apply sanctions, yet their committees meet less often. This suggests that Volanta pumps are more expensive, and that the technically demanding nature of the pump requires external expertise to carry out repairs, so a lower level of participation exists. Dependency on the borehole as the only source of clean water encourages good organisation e.g. higher incidence of payments.
An interesting finding is that factors such as prior contribution to the cost of the pump, participation in the process of pump installation and the percentage women on the committee still appear to have little influence on the organisational performance indicators. This finding seems to contradict the expectations of implementing agencies and the importance placed on these issues within current exit strategies. It indicates that good organisation can be achieved independently of entry conditions.
Long term financial management could not be measured. There will be a need to replace the whole pump and possibly re-drill the borehole should it collapse with age. Complete Willingness To Pay (WTP) should include replacement, however none of the systems were old enough to have warranted such an action, and therefore no data was generated on the full WTP of the communities.
Participation in maintenance links to a range of organisation indicators e.g. where people take part in maintenance, committee meetings are held more frequently, it engenders payment of quotas and the application of sanctions and rules.
People also tended to settle where the maintenance of boreholes was better organised e.g. regular payments, more sanctions and rules. However, as noted above (discussion of competence issues), there is a strong correlation between people settling and year of installation such that people tended to settle due to pumps in later years, which indicates that insecurity factors may be involved.
When considering exit strategies. Follow-up training needs to ensure both local fund and hardware maintenance management capability. Each appears to reinforce the other. However, the most important aspect is the encouragement of community participation in the pump management structure and process.
In an attempt to look solely at participation of the wider community the following variables were investigated.
Date of installation is a significant issue and appears to highlight the difference between pumps provided with and without prior social mobilisation. It therefore draws attention to one of the initial assumptions of the research project, that where pumps have been installed rapidly during an emergency response there may be a level of dependency with minimal social commitment to maintaining the water point.
It is also evident that the entry strategy changed significantly between the pre '95 period and those pumps installed later. These changes in entry strategy do appear to have led to greater community participation in the installation and care of the water-point. The noted influential changes in entry strategies are the insistence on a prior contribution to the pump purchase before drilling is initiated (only one agency) and the training of the whole committee.
Training, particularly prior training and training which lifts the sense of local pump management competence, is possibly the most important stimulant to future participation. Prior training appears to act as a social mobiliser and correlates with the instances of contribution to the initial cost of the pump. (It is recognised that only one agency has introduced the concept of prior contribution and that this was combined with initial committee training.)
Issues of competence, such as who repairs the pump, and the perceived ability of local maintenance people, have an affect on participation indicators. Where demonstrated ability to deal with pump maintenance exists within a community (related to training mentioned above), participation will tend to higher e.g. greater community involvement in committees, participation in installation and on-going maintenance.
Responding to the request of the community as opposed to pre-empting it, is also an important factor in insuring future local participation. Possible related to this are the issues of social cohesion and the involvement in the siting of the pump. Where members were present at the time of installation there was a far greater propensity for future participation in the maintenance of the water-point.
It should be noted that one agency interviewed had difficulty to stimulate interest in Community Based Management before the community had seen the borehole. This project facilitated selection of the committee after the drilling and therefore the village headman did all the mobilisation during the construction phase. This, it is reported, tends to compromise community participation.
Proximity to the pump does have an influence on participation. Either special effort has to be invested to make those living more removed from the water-point part of the process or it should be accepted that their level of involvement will be less.
Within the five dependent participation indicators, the ownership and prior contribution variables could equally be considered as independent or causal variables.
Note: An issue worthy of note is that prior contribution does help stimulate initial participation but not necessarily the ongoing care of the pump.
So we have seen that while the social factors may be a goal in themselves, and that they have links to localisation and ownership, they do not directly affect the reliability of the pump. Competence is the main link and there seems to be no substitute for good technical training.
It has been pointed out that the indicator of care of the pumphead whilst not representative of the technical reliability of pumps does coincide with it. Where the observed status of the pump is good, the boreholes tend to have lower frequency of breakdown and lower downtime. But good care of the pumphead may also represent better understanding of health issues and a propensity to better hygiene. Since all the communities have had hygiene training, and the research did not include a full KAP study of the communities, it was not possible to test this supposition. However, all water supplies have been installed as part of integrated water supply and sanitation programmes, so data on this indicator may give some insights into these wider objectives.
Where pumps have been installed in emergency situations (due to drought) the general condition of the pumphead tends to be poorer. There is no significant difference in the condition of boreholes installed pre 1995 and those installed from 1995 onwards, so it does appear to be the emergency conditions rather than age that influence care of the pumphead. This reflects an increased emphasis on community involvement in maintenance and care in development installations.
The data highlights that good care of pumpheads is evident in communities which have a high degree of internal capability:
All of these indicate that community mobilisation and training initiatives which focus on local capability have made some impact on the care of the pump (keeping in mind our supposition that a cared for pump might represent some hygiene awareness).
The variable “Rules” highlights a counter intuitive relationship with observed status. The application of rules and sanctions correspond to other indicators of community organisation and participation e.g locally elected committee, frequency of meetings and payments, community involvement, participation in maintenance. However, whether repairs are carried out locally / externally has no significant bearing on rules and the data indicates that the care of the pumphead is better where there are no rules. This may indicate that it is the local capability of O&M that encourages care of the pumphead, rather than actual community mobilisation (ie participation factors). A good local mechanic or caretaker creates a well cared for pumphead, rather than peoples understanding and participation in the O&M of the pump.
 |
This is supported to a certain extent by the awareness of the rules. Figure 8 summarises the rules mentioned by the respondents. As we see, general hygiene and payments are mentioned much more than caring for the actual wellhead.
Another link that appears to contradict the first few points in this section is that the general status of pumpheads at Volanta pumps is higher. For Afridev pumps we find that :-
Closer inspection of the data reveals the influence of a number of issues. Volanta pumps:
This raises the importance of issues of dependency on water supply and pressure on the pump. Volanta installations have a high dependency on the borehole expressed by a higher WTP, they were requested and settlement. Also, people living closer to the pump combined with less pressure (shorter queuing times) make it easier to achieve good care of the pumphead. In themselves these do not provide clear indicators as to how hygiene awareness is increasing within these communities and it may be that the more recent agency visits also puts pressure on communities to take care of the pumphead.
The research project is particularly interested in the impact of entry strategies on long term sustainability and whether the shortfalls of a non participatory entry (caused by an emergency) can be made up by an exit strategy.
The issue of whether the pump was installed in response to an emergency significantly influences the frequency of breakage, the downtime when broken and the observed status of the pumphead. In all three cases the performance against these three indicators is more negative in the case of emergency installations. The pumps break more frequently, they take longer to mend and the pump head is less adequately cared for. Most emergency installations took place prior to 1995, but the fact that these issues are not reflected to the same degree when considering the date of installation suggests that there are other factors at play that are peculiar to the emergency response process. This is not surprising, as the whole point of a development approach is to ensure greater overall sustainability for the water supply. However, when we investigate the situation in more detail we can see the role of the “entry” and “exit” strategies.
By considering the cases of emergency / “development” installations as a whole, the data shows there are significant differences in entry strategy issues in terms of “participated in installation”, and “prior contribution”. Both confirm that there was less community mobilisation in emergency installations.
When looking at indicators of social factors, the following trends are evident in emergency installations:
Community organisation:
Competence:
Participation:
This confirms that social participation, organisation and training, which tend to be more prevalent in the non-emergency pump installation process, do indeed have a positive impact on social factors within a community.
“Development” pumps have the “advantage” of entry strategy (participatory with prior contribution), follow up (many of the development committees had follow up training) and availability of spares (no one at a development pump site complained about lack of spares being available locally). Emergency pumps have a mixture of some training at installation, a mix of entry involvements (people helping with siting, or helping form a road for the drilling rig), various levels of follow up and in some cases lack of available local spares.
In an attempt to investigate this complex relationship of variables, the data was sub divided. Since the Volanta pumps have particular difficulties in terms of reliability and spares, these were removed from the sample. MW tests of frequency of breakage, downtime and ownership were then run on the emergency holes only across all the variables that represent the design or strategy of the programme. . The five variables that show as statistically significant are siting of the pump, present at installation, spares availability, where spares are accessed and number of women on the committee. This would confirm that even the simplest of involvement during installation makes a long term difference, that the spares issue is important and influential, and that the involvement of women does affect ownership and reliability.
However, the data at this second level analysis is quite weak and statistical significance must be treated with caution. It is more helpful to look at the trends of the data as represented by the Mean Rank. While this must be handled qualitatively rather than statistically, it nevertheless confirms a number of important elements of the exit strategy. Table 1 summarises the results
Table 1 Independent variable (on emergency holes with Afridevs only)
|
|
Downtime |
Frequency of breakage |
Ownership |
|
Last agency pump maintenance visit |
- |
- |
= |
| Election of members |
- |
- |
+ |
| Follow-up training |
- |
+ |
+ |
| Sanitation and health training |
= |
= |
= |
| Length of training |
+ |
+ |
= |
| Level of training given |
+ |
+ |
+ |
| Number of members |
+ |
- |
+ |
| Number of trained pump mechanics |
+ |
+ |
+ |
| Participated in installation |
+ |
+ |
+ |
| Present at drill/installation * |
+ |
+ |
+ |
| Siting of well * |
+ |
+ |
+ |
| Number of women on committee * |
+ |
+ |
+ |
| Spares available * |
+ |
= |
= |
| Where are spares acquired * |
+ |
- |
= |
| Ability of local committee or trained pump maintenance rep's |
+ |
- |
+ |
| Water committee appointed |
= |
- |
= |
| Community involvement in water committee |
+ |
- |
+ |
Key
+ More of the variable, or localisation of the variable, results in a “better” outcome, eg less downtime, less breakage, or more ownership.
- More of the variable, or localisation of the variable, results in a “worse” outcome, eg more downtime, more breakage, or less ownership.
= No difference in the samples.
* statistical significance (but with borderline sample number – treated with caution)
Table 1 confirms two themes that have emerged from the wider analysis.
First, that technical training is important. Competence will be formed by technical knowledge, which in turn is influenced by the length of training, the level of training and the number of trained mechanics.
Second, that the more local mobilisation there is the greater the ownership and lower the downtime, but the higher the frequency of repair. This higher frequency seems to be due to local people having a lower level of technical competence and whilst local community mobilisation lowers downtime, it can increase frequency of breakage.
The two variables that are counter intuitive are “last agency visit” and “follow-up training”. The more frequent the last agency visit the longer the downtime and the higher the frequency of breakage. This may indicate that communities have come to rely on the agency, or it may mean that the agency is visiting frequently because this is a troublesome pump (or community). Similarly the more there is follow-up training the longer the downtime. The follow-up training given often emphasised the social mobilisation, the formation of committees and how to raise finance. One would expect that this would result in shorter downtime and more frequent breakages (given the rest of the analysis). In fact it results in longer downtime and less frequency of breakages. This again may be because follow-up training is given where the pump is maintained by area mechanics and the training has yet to take effect.
In general the analysis highlights what would be expected – that technical training affects competence which in turn affects pump reliability, and that social mobilisation results in an increasing level of ownership and local wider community participation.
Given that the length of training, level of training and number of trained mechanics have been highlighted as important elements of an exit strategy, the data was used to investigate whether these elements could compensate for a weak entry strategy. The data was compared using the following:-
I.e. comparing those emergency installations that have longer training with those that have shorter training, and comparing both with “development” pumps.
This approach confirms the following:-
Length of training – frequency of breakage and downtime are both reduced with longer training. More importantly; “emergency pumps” which have had longer training tend to have less frequent breakdowns and shorter downtimes than “development” pumps. Ownership was increased with the longer training for emergency pumps and the ownership levels were almost the same for emergency pumps with more than 1 week training and “development” pumps.
Level of training – the frequency of breakdown of emergency boreholes tends to be reduced with more technical training, although the downtime tends to be slightly shorter when sanitation and simple maintenance is emphasised. There was a trend to greater ownership of the pump with increased level of training which matched the “development” pump’s expression of ownership.
Number of mechanics – emergency pumps with more than 2 trained mechanics tend to have lower frequency of breakage and shorter downtimes. They do not as such match the “development” pumps who have even shorter downtimes and less breakages, but nevertheless this confirms the importance of the number of mechanics. The number of mechanics has no impact on levels of ownership of emergency pumps, which is as expected.
To summarise:- there are confirmed benefits to the availability of the pump from making trainings more than 1 week, including both soft (sanitation, health, community management and simple pump maintenance) and hard (complete pump maintenance and second level pump repair) subjects, and training as many mechanics or caretakers as possible. The expressions of ownership are also enhanced by the longer training and increased levels of training.
There are clear indications that there is no substitute for early involvement and that even simple involvement during a pump installation makes a long term difference to the sustainability of a borehole. However, while an emergency pump with weak entry strategy can probably never fully match the sustainability of the “development” pump, ensuring competence through technical training (over a suitable period of time and with enough people) can increase emergency pump sustainability.
[1] The data is treated as a whole and differences between agencies are not presented. All variables were checked for agency and country differences and the results taken into account. However, it was decided that since this is not a project based evaluation of an agency’s work, these results should not be presented. Therefore the data is always treated as a whole.
[2] Other pump function variables. It is important to note that although the pump is considered to be non-functional if the pump is out of action, the status of the pump-head in general could also be considered as part of the functioning aspect of the water point. The word function has been used as opposed to performance, as measurements of actual yields over time were not taken into account in the survey. Two other depended variables were taken into account alongside the two key indicators of pump function discussed above. These are:
· The operational status of the pump, i.e. whether it was functional at the time of the visit. This was a snapshot observation only. This is considered to be of only limited importance
· Observed status of the pumphead, i.e. observation by the researcher on the care taken to keep the pumphead and its surrounds in a clean and controlled state. This issue is questionable as a measure of function as many of the pumps that had never experienced problems also had some of the worst cared for pumpheads.
However, since these variables did not highlight any new insights or contradict those documented above, the full analysis is not presented here.
