Night-time lights as a proxy for socioeconomic indicators in Uganda

Abstract: A large amount of economic research is reliant upon census, survey, or national accounts as a source of data. For many developing nations, the quality of this data is widely recognised to be unreliable and is rarely available at the sub-national level. By using a new and improved source of night-time satellite imagery data, this study investigates to what extent this data can serve as a proxy to traditional means of measuring socioeconomic indicators in Uganda, as well as what improvements this new source of data offers over it’s more widely used predecessor. For the purpose of this study, an algorithm is constructed that sources, pre- processes, and converts satellite image data into a product ready for analysis. The relationship between night-time lights and the chosen indicators is investigated by conducting a series of linear regressions, and the ability of this new data source to detect small-scale change in night-time lights is explored through the creation of a “rate of change” map. This paper concludes that there exists a strong, substantial relationship between night-time lights and GDP, life expectancy, infant mortality rate, access to electricity, and urban population in Uganda. Furthermore, this paper concludes that this new source of satellite data allows for the detection of change in night-time lights at a spatial resolution far higher than what was previously possible with the older source of night-time light data.