Call Us: US - +1 845 478 5244 | UK - +44 20 7193 7850 | AUS - +61 2 8005 4826

the World Health Organization

We propose a new method to evaluate program impacts across the entire distribution of outcomes and changes in outcomes. The method applies stochastic dominance estimation to differences-in-differences across subgroups and time, thereby allowing us to look beyond the mean impact and to tease out program effects that differ across the distribution of changes in an outcome variable of interest. We apply this to a government program in Kenya and find that it was associated with less deterioration in children’s nutritional status among the worst-off children, thus, effectively functioning as a nutritional safety net. JEL classifications: I32,O22, H51 Keywords: Program evaluation, stochastic dominance, child nutrition, Kenya Word Count: 5273 words We would like to thank Nancy Johnson and Ayago Wambile for inspiration and comments, Xu Claire Cheng for excellent research assistance and the Arid Lands Resource Management Project of the Government of Kenya for making the data available. Financial support for this work was provided by the International Livestock Research Institute. Any views expressed and any remaining errors are ours alone.