Measuring the Impact of Convenient Water Supply on Household Time Use in Rural Ethiopia

What is the impact of providing convenient water supply on water carriers' pattern of time use? How much of the freed time is re-allocated to paid market work, education (for girls), agricultural labor, or leisure? Do women report spending more time on activities they enjoy? Does convenient water supply lead to a re-allocation of leisure time to other household members? These questions are an important, but largely missing, piece of the economic evidence base for investment in the water supply sector. Cairncross and Valdmanis (2007) observe that "given the relevance of the time-saving benefit to water supply policy and the fact that the benefit is usually uppermost in the mind of the consumer, it is remarkable how few data have been collected on the amounts of time spent collecting water". We address this gap by measuring changes in time use among female water carriers before and after new water systems are installed in three rural villages in the Oromia region of Ethiopia. The timing of completion of the projects in the three villages was staggered over time for logistical reasons, so our quasi-experimental design allows us to control for any region-wide changes in time use. Because of low literacy levels, we used a pictorial time use elicitation approach based on respondents' recall of the previous day as well as the standard questions used in the DHS and LSMS ("how many minutes..."). We measured time use for all household members over the age of 10. We use this unique panel dataset with both pre- and post-project time use data to examine not only the effect on water carriers' time use but also any intra-household reallocation of time savings. In total, we interviewed 454 randomly-selected households in the three villages over three rainy seasons, and collected time use information on 1,590 household members. Primary water carriers spend (pre-project) an average of 110 minutes per day collecting water, roughly representative of water collection times reported in the 2005 DHS for rural Ethiopia. Unsurprisingly, men and women in our setting show distinctly different patterns of time use. We analyze the data using both difference-in-difference and first-differences approaches. Although analysis is ongoing, our preliminary results indicate that the water projects reduced household water collection times, on average, by 27 minutes in one treatment village and 75 minutes in the second. These estimates are based on time use data from our pictorial approach; estimates using direct recall of minutes are roughly twice as high. We find a statistically-significant increase in time spent in food preparation, household chores and socializing among women with reduced water collection times. Non-water collectors who live in households with lower water collection times spent less time farming, and more time on miscellaneous work, socializing and bathing/hygiene.