PhD Defense: Vincent Nossek

Vincent Nossek
CERDI, Université Clermont Auvergne
Ferdi

Examiners

Vianney Dequiedt, Professor, CERDI, Université Clermont.
Flore Gubert, Research Director, Laboratoire d'économie de Dauphine, LEDa-DIAL, Institut de Recherche pour le Développement.
Ahmed Tritah, Professor, UFR Droit, Économie et Gestion, Université de Poitiers.
Michael Goujon, Professor, CERDI, Université Clermont Auvergne.
Stéphane Straub, Professor, Toulouse School of Economics, Université Toulouse Capitole.

Abstract

Ensuring access for all to reliable, sustainable and modern energy services at an affordable cost (SDG 7) is one of the major challenges of the United Nations' Agenda 2030. For many developing countries, access to electricity remains a major development challenge, and one that will be all the more difficult to achieve as it takes place under the constraint of the Paris Agreement committing countries to limiting their greenhouse gas emissions.

One of the technical solutions strongly adopted over the last decade to accelerate electrification in developing countries, with a still large rural population share, has been the use of decentralized electrification through mini-grid power plants. This flexible solution for rapidly electrifying villages has the advantage of offering an energy mix incorporating renewable energies, mainly solar and hydro. However, the short- and medium-term economic impact of these decentralized electrification projects has rarely been assessed, due to the additional costs involved.

One way of assessing the economic impact of these installed projects is to exploit satellite data such as night light, agriculture-related data, or urbanization and human activity data. An initial application of this approach to the evaluation of mini-grid electrification projects through the evolution of night light has validated the intuition behind this methodology (Berthelemy and Maurel 2021).

In the first chapter, we will extend the described approach to project evaluation using satellite data by testing different sources of satellite data, as well as applying staggered difference in differences approach to investigate the causal link through a quasi-experimental methodology comparing villages that received mini-grids and control villages.

In a subsequent chapter, we will examine the impacts of decentralized electrification projects deployed across six rural localities in Madagascar. Field surveys were conducted prior to the installation of electrification solutions and again several years post-installation to compare the evolution of various socio-economic indicators. The design of this intervention follows a randomized controlled trial (RCT) protocol, where six out of twelve villages were randomly selected to receive the decentralized electrification projects. An impact analysis methodology of the ANCOVA type is proposed to analyze the data collected from the questionnaires and to measure the effects on households living in the different localities.

Finally, a third chapter will focus on the demand side, by studying the planning aspect of public policy, through the construction of indicators of energy poverty at sub-national level, by cross-referencing different sources of satellite data, notably urbanization and night-light data. By cross-checking satellite data with household survey data such as DHS or LSMS, a supervised learning algorithm can be calibrated to produce a PRIO-GRID-type poverty database.

Keywords

Impact evaluation, energy, remote sensing, renewable energy, electricity, poverty,

https://theses.fr/s363459