Working Papers

Pricing and Allocation of  New Agricultural Technologies - submitted

Paper: available here

World Bank Development Impact Blog post: available here

VoxDev article: available here

This study uses a two-stage experiment to examine whether lower prices allocate new agricultural technologies to farmers with lower returns. In stage one, I randomize a price subsidy, ranging from full to zero subsidies, for a new wheat seed variety. In stage two, I randomize free distribution across the self-selected sample of non-buyers from stage one. This design allows me to compare treatment effects across the entire population with treatment effects among the sample choosing not to buy the seed. If higher prices screen out farmers with low willingness to adopt, then the effect of stage-two free distribution on adoption by non-buyers should be trivial. Instead, I find that the stage-two free distribution increases adoption and wheat cultivation by an amount almost equal to the effect from stage one. In addition, farmers choosing not to buy in stage one do not realize lower returns to adoption --despite there being substantial heterogeneity in returns across the sample. Taken together, these findings imply that policy makers who aim to increase dissemination of agricultural technologies cannot rely on market prices as a mechanism for targeting high return farmers.

Targeting of Food Aid Programs: Evidence from Egypt (with Sikandra Kurdi) - R&R Journal of Development Economics

Paper: available here Presentation slides: available here

In-kind food aid programs remain prominent world-wide. Targeting in these programs is complex due to the potential for distortion of consumption choices and the need to balance competing objectives. This paper advances the literature on targeting in social protection by moving beyond poverty-based targeting to address nutritional objectives. Using data from a randomized controlled trial (RCT), we apply machine learning (ML) techniques to analyze heterogeneity in impacts across a range of nutritional outcome, with the aim of informing targeting based on observable characteristics. We find that such characteristics significantly predict heterogeneity in treatment effects, though the relevant predictors differ by outcome and treatment type. Building on recent literature advocating for the balancing of deprivation and expected impact, we show that in our context, the trade-off between targeting the most impacted versus the most deprived households is relatively limited. Instead, the main challenge is prioritizing among competing nutritional objectives. Our findings demonstrate that ML methods can identify targeting criteria dependent on specific policy goals.