Water availability and infrastructure development: cross-country econometric and neural network estimates Original Research Article

Freshwater shortage and pollution, and low coverage rates of adequate water infrastructures, pose serious obstacles to development in many countries. This analysis first briefly reviews current hypotheses and empirical evidence, and constraints in data availability and consistency. Econometric and neural network models are applied to a large number of indicators for developing and transition economies, with a view to examining the relationships among macroeconomic and environmental conditions, water infrastructure development, and availability, demand and quality of water resources. The principal components analysis suggests a trade-off between per capita internal freshwater and groundwater resources, on the one hand, and access to improved water supply services, on the other. Among other results, regression and neural network models indicate that access to these services has been lagging behind in some poor and scarcely populated countries. Less explained outlying cases are represented by heterogeneous groups of countries in terms of geographical location and other development features.