Published on: 06/03/2015
Based on a review of the costs of 179 water and sanitation projects carried out by FHIS (Honduran Social Investment Fund) over the last five years, we established the unit cost ranges for different intervention models. We also identified the main factors driving the costs of water and sanitation projects.
In Honduras, water and sanitation investment programmes are carried out through a large number of intervention models. An intervention model refers to the set of definitions, criteria, procedures and rules that govern an investment programme. It entails the definition of: a) the intervention area, b) the types of projects to be carried out, c) the service level to be achieved, d) the form of project cycle management, e) the executive set-up, f) the procurement modality, and g) financial aspects, including unit reference costs, cut-off lines and co-financing rules.
The diversity in intervention models seeks to respond to the diversity in needs and context for water and sanitation investments in the country. But at times, the intervention models are also defined by development partners, without having clear criteria on which model (or component of a model) to apply under which circumstances. Neither is there any insight into which of the intervention models is most cost-effective in a given context.
The objective of this study is to identify the range of unit costs for rural water and sanitation investment projects for different intervention models. The specific objectives are:
The study was done by analysing the costs of 179 investment projects carried out in the scope of 11 water and sanitation programmes, carried out by FHIS (the Honduran Social Investment Fund) over the period 2006-2011, as well as the intervention models applied in those programmes. The cost analysis was done using the information contained in the project database of FHIS, complemented by a review of the project documents, and interviews with staff involved in managing those programmes. The latter also served to characterise the intervention models under which the projects were carried out. The main study limitation was the almost complete lack of information on service levels – both prior to and after the investments – in the FHIS database. Therefore we could not establish the change in service delivery that was achieved in the communities, nor relate that to the level of investment done. The study recommends to include such service level information into the database (linked to the SIASAR rural water and sanitation information system), and provides other specific recommendations for improvements in the database as well.
The study confirms the existence of a large number of intervention models. For each of the components an intervention model, typically three of four different options exist. For example, there are four different procurement modalities and four different forms of project cycle management in use within FHIS. Which of the options to apply depends for some of the components on national rules and regulations, for example the national procurement rules or design criteria. But for some components, there are no criteria which option to apply in which circumstances. Before unit costs can be defined per intervention model, there is need to define the intervention models as such more clearly. Particularly, there is a need to define per component the criteria under which to apply a certain option.
As far as unit costs are concerned, we found a high variability. The costs of water projects ranged from less than 50 US$/person to more than 450 US$/person, with a median of about 250 US$/person. Latrine projects had costs ranging from 50 to 250 US$/person, and a median of 140 US$/person.
The per capita costs are influenced by two main factors: the type of project (whether it refers to the development of a completely new system, an extension or a complete or partial replacement) and the size of the village in which the project is done. As can be expected, developing a completely new water system has higher per capita costs than an extension. The per capita costs of capital replacement projects are in between, with the costs depending on whether the project is a full system replacement or only of several components. In small communities (less than 200 persons), the per capita costs are much higher than in bigger ones – probably because in the latter economies of scale can be achieved in system development. A third factor that seems to influence the unit costs is the poverty level of the population. The projects done in municipalities with high poverty indices had higher per capita costs.
A factor that has indirect influence on unit costs is the total volume of the project. We found that FHIS has carried out a large number of project of relatively small size (less than 50,000 US$), those being mainly minor extensions of replacements in small communities. Given that fact that administrating a project represents a fixed cost to FHIS, managing a small project results in a relatively large overhead. The programmes that mainly carried out these small projects had bigger overhead costs.
An analysis of other factors that we thought could influence the unit costs, such as the form of project cycle management or the procurement modalities were not found to have that influence. Also technology type – such as the type of latrine – didn't have an influence on average unit costs.
We conclude that it is not possible to define unit costs per intervention model, for as long as the intervention models are not more clearly defined. Rather, we see potential in defining unit costs differentiating them according to the two factors that have most influence: type of project and size of the community.
We also conclude that unit costs are best defined in the form of ranges and not as a single number, in view of the wide variability of costs. These unit costs need to be a reference and not a cut-off point. In some cases, the costs of providing a certain community are just very high, due to topographic or other circumstances, as found in the study. If an investment project would have costs above the reference costs, this shouldn't mean that community would be without water and sanitation services; rather, it needs to act as an alert for FHIS, to critically review the project and explore alternatives in more detail.
Finally, the study concludes that unit costs are but one of the components of intervention models that need to be clarified. There is need and potential to define the other components more clearly, especially to define criteria which option per component to apply under which conditions. Priority ones – from a cost point of view – are a) the level of co-financing expected from different stakeholders, b) the feasibility of doing small projects and c) service levels to be achieved.
Below the full report and a summary presentation (both in Spanish) are available, as is a series of blogs discussing the findings.