Abstract:
Using data gathered by the WASHCost project in Ghana, this study uses cost functions to predict capital ex penditure for construction of piped water systems for small towns in Ghana. The data was collected from 45
small towns in peri-urban areas in the Volta, Ashanti and Northern regions of Ghana. Correlation and multiple
regression analyses are used to derive cost functions for both the complete and components of water systems.
The aim is to derive functions for the total capital expenditure of a small town water system and for the fixed
assets components: mechanised borehole, water reservoir, pipe work and stand post. The results are presented
as eight functions, representing the total capital expenditure for a water system and capital expenditure for the
infrastructure components water reservoirs, pipework and stand posts. Explaining variables identified are de sign population, length of pipeline, volume of reservoir and dummy variables for region (location). The func tions vary in accuracy with an explanation coefficient R2
varying from 0.42 to 0.88 with limited data, but the
functions are indicative for the cost of water systems with mechanised boreholes which are increasingly being
used as small town water supply systems in Ghana.
Key words – Ghana, small town water system, cost function, capital expenditure, multiple regression analysis
