Using Satellite Data to Improve environmental Impact Evaluation

Faculty Member: 

Proposal Description:

In this paper I estimate the effects of a land tenure formalization program in Benin on developmental and environmental outcomes. I develop a new impact evaluation approach, combining machine learning methods for satellite remote sensing data with a double machine learning approach to inference. The approach conditions the estimated effect on the series of satellite images taken just before treatment was assigned. We plan to validate this method by showing that we can recover estimates of effects from the randomized control trial even when we do not account for selection into the study. We show that this approach can generate unbiased and more precise estimates at the village level, then demonstrate that it can be used to estimate the effects at the parcel level–something which has previously been impossible in land titling research without very strong assumptions. This will allow us to estimate outcomes which are directly of interest to policymakers. We plan to develop the method so that it can be used in other settings where confounding factors may be easiest to measure in satellite imagery.

Requisite Skills and Qualifications:

Applicants should have strong experience working with data in R, the ability to commit 10 hours per week to the project.

Previous work with spatial data, machine learning methods, or satellite imagery would be helpful, but is not required.

Award: 
  • Jackson Pullman