Mapping Community Development Aid in Myanmar: Combining Satellite Imagery and Spatial Data

Aid policy has greater potential to combat global poverty when targeting areas of concentrated need. However, a gap exists between aid given and actual need due to data sparsity. These difficulties are exacerbated in fragile states experiencing armed conflict and civil war. Thus far, few aid-determinant studies have analyzed the characteristics of poverty at the subnational level, and even in those studies, the units of analysis are at a high administrative level such as the state. This study intends to fill this knowledge gap by examining aid across a range of spatial scales. This approach allows policymakers to portray poverty at the granular level, and promote the design, monitoring, and evaluation of aid towards the most marginalized.

The goal of this study is to explore the extent to which community-led-development (CLD) projects take place in poor villages, using the case of Myanmar. This study examines how much of the variance in CLD project allocation is explained by wealth and development-related measures. It also analyzes how two CLD models, NCDDP and SMU, differ by their targeting practice towards poor and conflict-prone villages. To collect outcome variables, I develop web scraping algorithms to create comprehensive and up-to-date locations of CLD participating villages (N=12,282). As for exploratory variables, nighttime satellite imagery is trained on the Demographic and Health Survey (DHS) to predict wealth in project villages and DHS village clusters. In addition to this, I spatially interpolate the DHS wealth index to make inference on poverty in aid sites. By matching geo-referenced aid and wealth data, I test factors that may explain variation in the distribution of CLD and different approaches to community development.

The results show limited evidence in poverty-oriented targeting. First, as each increment of the share of a vulnerable population rises, the likelihood of aid presence in that community declines by 4%. Next, as the intensity of nightlight increases, the density of community development projects also rises. One unit increase in the radiance value increases the number of projects by 86 within a two-degree radius of a DHS village cluster. Among villages of similar levels of nightlights and population, aid goes to less wealthy areas. A score higher wealth index yields a reduction of the 227 projects per unit area. Last, NCDDP, which emphasizes inclusion and collaboration, supports poorer villages farther away from conflict events. In contrast, SMU, which considers competition conductive to performance, favors more established areas not excluding villages near conflict zones.

This study shows that CLD in Myanmar disproportionately flows to better-off communities, as indicated by a lower share of vulnerable populations and lights shining brighter. However, a need-based allocation is also in place; CLD projects are deployed in villages with lower assets when holding other variables constant. In this study, nighttime luminosity at fine resolutions captures the variability of economic development in villages and improves prediction of aid allocation. Synthesizing new sources of data can be used to assess neighborhood-level interventions in the context of poverty and conflict where a traditional survey is too costly.