I have written before about our work on life-cycle cost analysis (LCCA) in Honduras. The idea is to look into the real costs of investment programmes and projects in Honduras, so see which intervention model is the most cost-effective. The reason for that is that within the sector, and sometimes even within an institution, various intervention models are applied, each with their own type of project cycle, procedures, unit costs and maximum investment costs, and that without having always clarity what these unit costs are based on. The study should give better insight into the real costs and inform future guidelines on unit costs and help further standardize elements of intervention models. Last week saw our first effort of trying to understand what data is available and how we could possibly analyse that. That brought good news and bad news.
The good news is that FHIS, the Honduran Social Investment Fund, the main government body implementing water and sanitation projects, has an extensive database with all kinds of cost data of their programmes of the last 15 years. The bad news is that these cost data have not been classified into the life-cycle cost analysis categories, as proposed by WASHCost. Most investment projects done in fact are a mix of capital investment costs (CapEx to use the right acronym) and replacement of old infrastructure (CapManEx). Because the main type of water supply technology in Honduras are piped systems, often a project would consist of extending pipes to new houses (CapEx) with for example, the replacement of an old tank or main line (CapManEx). Getting a better idea of what percentage of project costs were one or the other would thus require looking into the technical design documents. That brought even worse news. None of the technical information is in the database. All project background documents, feasibility studies, designs and cost estimates are in hard-copy folder. The photo below depicts a typical folder for one water system! Luckily, one of the FHIS colleague could easily indicate us where the most relevant technical information could be found in each of these piles of paper. And with that, it is relatively easy to see what percentage of a project is capital investment and what is replacement.
The worst news though is that the project documents do have only limited data on service levels. Also on this blog, I have been talking positively about Honduras’ SIAR (Rural Water Information System), an information system which classifies all water systems in the country from A to D, with A being the ones performing very well, and D the ones needing a major overhaul. With all its imperfections, it gives a useful indication of the service level being provided by a water system. However, in none of the sample systems we studied, was there any information available on how the systems were classified before or after the intervention. One possible reason being that the SIAR was being managed by another agency, SANAA. And the two database apparently didn’t communicate to each other on these data. Hopefully, the upgraded version of SIAS, the SIASAR, that is being developed now in various countries in Central America, will overcome that problem as it is supposed to be an inter-institutional information system. Anyway, some of the project reports have some data on service levels, but this is not consistent across the projects: some mention the water quality before the intervention, but not after, whereas others only have information on continuity of supplies. So, linking costs to service levels will not be possible based on these secondary data.
Not all is gloomy though. Another piece of good news is that we will be able to get interesting data on overhead costs. It was revealing to see how many layers of overhead there are for any water system. There is a technical person within FHIS who typically manages these water projects, but then for each step in the project cycle, another department of the organisation is involved: to appraise projects, to review its finances, to manage the tender process, to carry out monitoring and above all to supervise. Each project has apart from its technical manager who supervises the whole project cycle, a supervisor who checks works on the ground. In addition, there is an inspector who controls the supervisor and confirms whether all is in order. In short, lots of checks, balances and controls, probably with the view towards reducing incorrect use of funds. But, all of this creates an overhead as well. Based on some first review of data sources, we should be able to quantify this for different types of projects. For example, in some projects, much of this cycle is decentralised to local governments, with the idea to empower them and let them contract works locally, and thereby probably reduce costs. However, for such projects, there is even more supervision needed. Hopefully our work will show whether this is most cost-effective.
There are no concrete results yet to share after one week of identifying information sources. But, it has already triggered lots of discussions on information management. The exercise raised question as: we have such a nice database with cost data, but we hardly ever use it to look at data across all project. Why not? And why doesn’t the database contain data on the technical design and only on the finances and contracting? And if so which data specifically? How could data on service levels, from for example SIAR or SIASAR be integrated? The database contains only the costs incurred by FHIS, but not the ones by municipalities or the village people; why not?
It is fun to work with the LCCA. And, hopefully we will get insight into the costs of the different intervention models. But I hope that also the questions raised on broader information management can be addressed. Much of the data is there, and probably the management of the information doesn’t need to change drastically. If it would only improve slightly, probably many pieces of bad news would turn to good news, not only serving the LCCA, but also many other analyses that can be done within FHIS and beyond.