Association for Public Policy Analysis & Management

Jordan is one of the most water poor countries in the world (Haddadin 2006). Recent population growth, urbanization, and the influx of refugees have intensified the strain on Jordanian water resources (O. Jaber and S. Mohsen 2001; Haddadin 2006), a strain that is further exacerbated by high annual water losses. A particularly useful notion of water losses is that of non-revenue water (NRW, or units of water supply for which revenues are not generated), which reaches 50% in Jordan (Al-Ansari et al. 2013). NRW results from a variety of factors, including breakdown in pipes in the distribution system, broken meters, illegal water usage, poorly installed meters, or weak governance (Al-Ansari et al. 2014). These water losses contribute to poor cost recovery, low levels of maintenance, and degradation of the system performance (Jaber and Mohsen 2001), factors which reduce households’ confidence in the water reliability and quality (Haddadin 2006). 

This paper reports on the design and baseline survey results from an impact evaluation of an urban water infrastructure investment program being implemented in Zarqa, Jordan. The primary aim of the improvements is to increase incomes in Zarqa through reduced NRW and ameliorated water reliability, which in turn leads to cost savings and productivity gains among beneficiaries. A primary challenge for an evaluation of this type is to account for the non-random selection of neighborhoods for water network rehabilitation and sewer extension investments. Using census data, we implemented a propensity score matching procedure to identify observationally similar neighborhoods that will be differentially affected by the projected infrastructure improvements. Within matched zones, households were randomly selected for inclusion in the study. This design offers a practical approach to evaluation of complex and large infrastructure projects. It also fills an important gap in the literature on the economic benefits of water infrastructure.

Baseline data from our sample of 3468 households shows that households do not display statistically significant differences on a number of key variables. We then use these baseline data to estimate the social benefits to households of reliability improvements in Zarqa. We derive estimates of these social benefits using two methods: coping costs and contingent valuation willingness-to-pay (WTP) estimates. Coping costs refer to the costs of activities that households engage in to compensate for the unreliability and poor quality of existing water supplies. While 97% of our sample subscribes to WAJ services, a majority of households also purchase water from other water sources (e.g. water shops or water tankers). Coping costs include non-network water expenditures, collection time costs, storage costs, and water treatment costs. Baseline data shows that households in Jordan pay a high level of coping costs—on average, 7% of the monthly expenditures—implying potentially high social benefits from improving the network. However, contingent valuation WTP estimates imply a relatively low acceptability of increasing water tariff levels to obtain this increase in water reliability.